• Here, the interaction of Melodoigyne incognita virulent and avirulent pathotypes with susceptible and Mi-resistant tomato (Solanum lycopersicon) has been studied. Significant differences in nematode penetration occurred 2 days postinoculation (dpi) and became stable from 3 dpi onwards. The hypersensitive cell response (HR) in resistant plants prevented the installation of the avirulent pathotype. The virulent pathotype overcame the Mi (nematode) resistance and induced feeding sites in root cells without triggering HR.
• Reactive oxygen species (ROS), visualized by subcellular reduction of nitroblue tetrazolium, accumulated in nematode penetrated cells. Quantitative analyses with dichlorofluorescein indicated that the oxidative burst occurred very early with both pathotypes, with an enhanced rate in hyper-responsive cells.
• Hydrogen peroxide (H2O2), detected by cerium chloride reaction, accumulated in the cell walls and especially in cells neighbouring HR. The apoplastic location of cerium perhydroxide indicated that either the plasma membrane or the cell wall was the primary site of the superoxide/H2O2 generator.
• The data provide evidence, for the first time, for ROS-generated signals and their spatiotemporal expression in the host and nonhost interaction of tomato with nematodes.
Plant–pathogen interactions, particularly those involving highly specialized biotrophic parasites, are governed by specific interactions between pathogen avirulence genes and plant resistance genes. In the idealized ‘gene-for-gene’ system, the recognition of an avirulence pathogen allele by the corresponding resistance host allele results in an incompatible interaction (i.e. disease resistance). In all other situations the compatible interaction occurs, leading to the disease (Flor, 1971; Keen, 1990). The phenotypic expression of plant resistance is often characterized by the so-called hypersensitive reaction (HR) consisting of the localized cell death of the plant tissues at the infection site, which contributes to the limitation of growth and spread of the pathogen (Gabriel & Rolfe, 1990). At the biochemical level, the rapid generation of reactive oxygen species (ROS), such as superoxide (O2−) and hydrogen peroxide (H2O2), the oxidative burst, is the first plant reaction in response to the attack by avirulent or virulent pathogen. In incompatible interactions between avirulent pathogens and resistant plants, this nonspecific, weak and transient ROS production is followed by a massive and prolonged ROS accumulation, and the latter is intimately associated with the HR response (Lamb & Dixon, 1997; De Gara et al., 2003). This two-phase kinetics of ROS production is typical of the incompatible interactions characterized by HR. Nitric oxide also acts as a signal molecule in plant–pathogen interactions, and cooperates with ROS in a balanced manner to trigger the HR (Bolwell, 1999; Delledonne et al., 2002; Melillo et al., 2003).
Plant-parasitic nematodes are obligate parasites that feed exclusively on the cytoplasm of living plant cells. Among them, the sedentary endoparasitic root-knot nematodes (RKN), Meloidogyne spp., have evolved a very complex and intimate relationship with their hosts, with the establishment of permanent feeding sites in the vascular cylinder of infected roots (Abad et al., 2003). Nematode resistance genes have been identified in several plant species, and most of them are associated with a typical HR in root cells surrounding the invading nematode. In addition, biochemical and histochemical studies suggested a concomitant activation of enzymes involved in ROS metabolism (Zacheo et al., 1997). In the incompatible interaction between Arabidopsis thaliana and the cyst nematode Heterodera glycines, symptoms of HR and production of H2O2 have been documented (Waetzig et al., 1999). However, there is little direct evidence that a typical oxidative burst in naturally occurring plant–nematode interactions actually determines resistance. One of the best characterized nematode resistance gene is Mi, which confers resistance to three RKN species in tomato (Williamson, 1999). The HR occurring in resistant tomatoes infected with avirulent RKN is associated with additional molecular changes, among which the modulation of enzymatic activities involved in the production/neutralization of ROS (e.g. peroxidases and superoxide dismutases) (Zacheo & Bleve-Zacheo, 1988; Zacheo et al., 1993). However, despite some fragmentary data, the occurrence and precise role of ROS, and more generally of an oxidative burst in plant defence during interactions between nematodes and their hosts, including RKN and tomato, is poorly understood.
Isogenic tomato lines, either susceptible to RKN or bearing the Mi resistance gene, are available (Laterrot, 1975). In addition, M. incognita avirulent and virulent pathotypes have been characterized in the laboratory at the biological and molecular level (Semblat et al., 2000). Therefore, this pathosystem indeed constitutes an excellent experimental material to study both compatible and incompatible situations. In this work, we designed time-course experiments to analyse the production and in planta localization of ROS and H2O2 during the early steps of the plant–nematode interaction, in order to compare the plant responses when susceptible and Mi-resistant tomatoes were challenged with avirulent and virulent M. incognita pathotypes. Hydrogen peroxide production was localized at the subcellular level to better elucidate the mechanisms and processes involved in pathogen-induced ROS production and its significance in the host defence. The results indicate that the timing of H2O2 generation is one of the most determinants in blocking the successful nematode development.
Materials and Methods
Two Meloidogyne incognita (Kofoid & White) Chitwood isolates from Morelos (Mexico) and Valbonne (France), avirulent and virulent against the tomato Mi resistance gene, respectively, were used. They were specifically identified according to their isoesterase electrophoretic pattern (Dalmasso & Bergé, 1978). The susceptible tomato (Solanum lycopersicon L.) cv. Roma and the near-isogenic resistant cv. Rossol carrying the Mi gene (Laterrot, 1975) were used in the experiments.
Penetration of avirulent and virulent nematodes in root tissues was evaluated on susceptible and resistant tomatoes. Nematodes used as inoculum sources consisted of infective second-stage juveniles (J2) collected from infested roots held in a mist chamber, and stored at 4°C before use within 2 d. The experiment was conducted in a growth chamber at 20 ± 1°C, with a light regime of 16 h day : 8 h night. Young plants individually grown in sterilized sand were inoculated with a water suspension of 125 J2. At 1, 2, 3 and 4 d postinoculation (dpi), plants were removed from pots and root systems were carefully washed and stained with acid fuchsin (McBryde, 1936). The root system of each plant was then examined under a stereo microscope, and penetration was evaluated by enumeration of the J2 stained into the root tissues. For each plant–nematode interaction, five replicate plants were analysed at 1, 2, 3 and 4 dpi, and the experiment was repeated three times.
An analysis of variance (anova) was performed on the data, followed my mean separation with Fisher's protected least significant difference (PLSD, at P = 0.05) to compare the rates of penetration. Statistical calculations were performed with statview software (Abacus Concepts, Inc., Berkely, CA, USA, 1992). Because no significant differences existed between the repeated experiments, data were pooled together for analysis.
Determination of ROS release
Fluorometric analyses were adopted for the detection of ROS accumulation using a membrane-permeable probe, dichlorofluoroscein (DCFH). This probe reacts with ROS and is oxidized to DCF, which is fluorescent. Root apices from infected (12, 24 and 48 h after nematode inoculation) and uninfected seedlings were excised and preincubated for 30 min in K-phosphate buffer (20 mm, pH 6) to remove preformed H2O2. The working solution of 50 µm 2′,7′-dichlorofluorescin (DCFH)-diacetate (Sigma, St Louis, MO, USA) in K-phosphate buffer 20 mm, pH 6 was incubated for 15 min at 25°C with 0.2 g ml−1 of porcine liver esterase (Sigma) for deacetylation (Schopfer et al., 2001). This solution was immediately used for ROS assay by incubating 50 mg ml−1 of each root sample for 30 min at 25°C on a shaker. A 1-ml aliquot of the solution was then removed and the increase in fluorescence (excitation 488 nm, emission 525 nm) caused by the oxidation of DCFH to DCF was measured within a few minutes using a fluorescence spectrophotometer (LS-5B; Perkin-Elmer, Beaconsfield, UK). Blanks without plant material were run in parallel and used for subtracting spontaneous fluorescence changes. For statistical purposes, fluorometry experiments were performed in triplicate, and fluorescence increases were expressed as a percentage of the maximal increase possible from the tissues.
In order to test the specificity of these reactions, ROS scavenging reagents were used. Root apices (50 mg) from uninfected and 12 h infected ‘Rossol’ roots were excised and preincubated for 30 min with 100 µg ml−1 catalase (H2O2 removal), 200 µm KCN (peroxidase inhibitor), and 100 µm diphenilene-iodonium (DPI, an NADPH oxidase inhibitor) followed by 30 min in the DCFH reaction medium also containing the respective test substances. The decrease in fluorescence in the incubation medium was measured as reported above, and using reagent blanks as reference.
In planta detection of ROS
For time-course detection of ROS, roots from ‘Rossol’ and ‘Roma’ tomato infected with both avirulent and virulent pathotypes, respectively, were collected at 12, 24 and 48 h after nematode inoculation and immersed in a solution containing 0.1% of nitroblue tetrazolium (NBT) in 10 mm potassium phosphate buffer pH 7.8 for 1 h. For controls, some samples were preincubated for 30 min in 100 µg ml−1 superoxide dismutase (SOD), one of the primary antioxidant enzymes, and then treated 1 h with the solution containing NBT. Root tissues were then fixed in 2% paraformaldehyde−2% glutaraldehyde in 50 mm sodium cacodylate buffer (pH 7.2) for 2 h, postfixed in a solution of 2% osmium tetroxide in the same buffer for 2 h and then dehydrated through an ethanol series to absolute ethanol and embedded in LR White resin (Sigma). Serial semithin sections were cut and then stained briefly in a solution of 0.02% toluidine blue in 50 mm citrate buffer pH 3.5, and observed under a light microscope. Production of ROS was specifically detected by the blue precipitate that was formed upon reduction of yellow NBT by superoxides (Zacheo & Bleve-Zacheo, 1988).
Subcellular localization of H2O2
Production of H2O2 was assessed cytochemically via determination of cerium perhydroxide formation after reaction of CeCl3 with H2O2 (Bestwick et al., 1997). Root tissues at 12, 24 and 48 h after nematode inoculation were excised and incubated into freshly prepared 5 mm CeCl3 in 50 mm 3-(N-morpholino)-propanesulfonic acid (MOPS) at pH 7.2 for 1 h. For controls, some roots were preincubated 15 min either with 10 mm sodium pyruvate, a molecule reported to be a strong hydrogen peroxide scavenger (Li et al., 1998), or with 2 mm potassium cyanide (peroxidase inhibitor). Samples were then incubated in CeCl3 solution as described above. Roots were fixed, dehydrated, and embedded in LR White resin as above reported. Ultrathin sections were made with a Leica Ultracut E, stained with aqueous solution of uranyl acetate and lead citrate and observed under a Philips 400T electron microscope.
Quantification of cerium perhydroxide deposits
To quantify the amount of cerium perhydroxide deposits, original unprocessed scans from electron micrographs of infection sites, taken with the same magnification (×22 000), were used. The Density Meter 1.0 Image Program differentiated cerium perhydroxide deposits from the background by the difference in contrast. Calculation of precipitates was based on areas randomly taken from the plasma membrane, cell wall, and intercellular spaces of injured cells, digitized at 300 dpi (dots per inch), and measured in pixels (30 × 30 pixel2). To eliminate artefacts from uneven light distribution in the original micrographs, a flatten background operation was performed. The intensity of CeCl3 staining was estimated from 10 infection sites at each time for three replicates. Statistical analyses were carried out to establish differences in deposits occurring between Rossol + avr and Rossol + vir interrelationships.
Penetration of Meloidogyne incognita in tomato roots
The avirulent and virulent infective J2 were able to penetrate roots of both tomato cultivars, but the anova revealed a significant influence of plant genotype (P < 0.0001), nematode line (P = 0.003), time after inoculation (P < 0.0001) and their interaction (P = 0.002) on the numbers of J2 found in root tissues (data not shown). For each plant–nematode interaction, the average penetration of J2 in roots is shown in Fig. 1. At 1 dpi, very few J2 were found in the root tissues, and no significant difference occurred between the four situations tested. From 2 dpi onward, a significant difference in nematode penetration could be observed between the interaction of the avirulent nematodes with the resistant tomatoes and the three other experimental situations (0.0028 < P < 0.0001), except at 2 dpi when virulent nematodes attacked the susceptible tomatoes. Clearly, when the avirulent nematodes were inoculated onto the resistant tomatoes, they could not invade the root tissues (i.e. incompatible interaction), unlike the other three interactions where the plant did not prevent nematode penetration (i.e. compatible interactions). However, significant differences could be seen in the level of J2 penetration in the three compatible interactions examined. At 2 dpi, virulent J2 invaded roots of the susceptible tomatoes less efficiently than avirulent J2 on susceptible tomatoes (P = 0.004) and virulent J2 on the resistant tomatoes (P = 0.0223). At 3 dpi and 4 dpi, the same overall response was observed: reduction of the ability of virulent nematodes to invade roots of resistant tomatoes compared with the same nematodes onto susceptible tomatoes (P = 0.042 and P = 0.001 at 3 dpi and 4 dpi, respectively) or the avirulent nematodes onto susceptible tomatoes (P = 0.002 and P = 0.003 at 3 dpi and 4 dpi, respectively).
Apart from intensity and time differences occurring in virulent and avirulent J2 penetration on both hosts, their early trophic action (at 1 dpi) induced different reactions in the cells selected as potential feeding sites. Figure 2 shows the incompatible interaction between the avirulent pathotype and the resistant tomato expressed as HR of cells all around the nematode (Fig. 2a) and incipient giant cells in the compatible interaction between the virulent pathotype and the resistant tomato (Fig. 2b). The same compatible interaction occurred when susceptible tomato was infested with both avirulent and virulent nematode populations (not shown).
Quantitative monitoring and histochemical localization of ROS production
In a time-course experiment, the occurrence of ROS was studied from early to late stages of infection of resistant and susceptible tomato with the avirulent and virulent pathotype, respectively. The main increase in ion leakage from infected roots, indicating that nematode penetration induced wounding, was observed within 12 h for all the four tomato–nematode relationships and a time-dependent pattern was observed (Fig. 3). The most significant increase in levels of ROS was recorded in ‘Rossol’ resistant roots infected with the avirulent pathotype at 12, 24 and 48 h postinoculation, with a slight decrease over time (Fig. 3). The peak of DCF reactivity in the DCFH assay indicated that ROS formation (61% compared with the control) occurred in the incompatible interaction Rossol + avr during the rising of the injured cell reaction. Differential situations were observed in the three compatible interactions (Fig. 3). At 12 h postinoculation, an equivalent increase of ROS levels was noted in Rossol + vir (34%), Roma + avr (29%) and Roma + vir (23%) compared with uninfected controls, respectively. At 24 h postinoculation, the level of ROS decreased significantly in the susceptible ‘Roma’ roots infected with both nematode populations. In the later stage of infection (48 h postinoculation), the levels of ROS were equally lowered in the three compatible interactions (from −20 to −25% compared with the uninfected controls).
Table 1 shows that the DCFH oxidation by ‘Rossol’ roots can be inhibited by scavengers of H2O2 (catalase), OH− (KCN), and O2− (DPI). Inhibition analysis confirmed enzymatic origins of a large part of ROS detected. The most efficient inhibition of ROS in infected roots was treatment with DPI (NADPH oxidase inhibitor) compared with infected tissues not treated with inhibitors and uninfected tissues. Catalase and KCN were less efficient, both inhibiting about 38% of ROS release. These data suggest that O2− and OH−, as well as H2O2 are involved in ROS production. Moreover, the differential sensitivity of different inhibitors between uninfected and infected tissues provides evidence that DPI in infected roots inactivates a cyanide resistant enzyme such as NADPH oxidase, which seems to be the most responsive for ROS (O2−) production.
Table 1. Effect of reactive oxygen species (ROS) scavengers on ROS release by uninfected and 12-h nematode infected tomato ‘Rossol’ roots
To examine the extent of nematode-induced superoxide production, the classical histological NBT assay was used in a time-course experiment for in vivo localization of superoxides in ‘Rossol’ roots upon avirulent and virulent nematode infection. As shown in Fig. 4, the number of cells displaying NBT staining adjacent to those injured by nematodes varied and seemed to be dependent on the intensity of the defence reaction. In the incompatible reaction, ‘Rossol’ roots were hyper-responsive when injured by the avirulent pathotype and revealed a stronger blue reaction than the compatible interaction Rossol + vir. Sections of ‘Rossol’ roots infected with avirulent pathotype showed heavy blue product in the rhyzodermis and outer cortical cells injured by J2 penetration 12 h after inoculation (Fig. 4a), and in the cells selected as feeding sites by the parasite at 48 h (Fig. 4b). A very heavy staining of both wall and cellular compartments, indicative of high ROS production, was detectable in rows of cells where the nematode was entrapped (Fig. 4c). ‘Rossol’ roots infected with the virulent pathotype showed a blue reaction in cortical cells at 12 h (i.e. the site of nematode penetration) (Fig. 4d). Blue staining was diffuse in 48 h tissue challenged with virulent nematodes (Fig. 4e). At this time, the juvenile was well established and a giant cell (nematode feeding site) was developing. Cells neighbouring the feeding site showed a faint blue reaction (Fig. 4f). Preincubation of root tissues with SOD, before the treatment with NBT, completely inhibited NBT formazan formation and blue staining development in nematode injured cells (data not shown).
Subcellular localization of H2O2 accumulation
To explore the hypothesis that H2O2 production at the initial stages of nematode infection might be an important component of host response to pathogen recognition, H2O2 localization was analysed either in ‘Rossol’ or ‘Roma’ roots challenged with avirulent and virulent M. incognita pathotypes, respectively. Samples of infected roots were treated with CeCl3 for visualization of H2O2 generation at 12, 24 and 48 h after nematode inoculation.
In resistant tomato roots injured by avirulent pathotype, the more widespread accumulation of H2O2 was first apparent 12 h after nematode inoculation and was not necessarily associated with visible cellular destruction. Heavy CeCl3 deposits were detected on the plasma membrane, cell wall and intercellular spaces of the root apex cells as a reaction to invading J2 (Fig. 5a). By 24 h, the attempt of J2 to establish their feeding sites induced hypersensitive reaction of a number of cells (Fig. 5b). In the cells located at the periphery of those undergoing the HR, H2O2 accumulation was detected on their plasma membranes and cell walls (Fig. 5b,c). At 48 h, the number of dead cells increased and H2O2 production was accumulated in their wall and cytoplasm and in the walls of neighbouring cells (Fig. 5d). At this point, confluent deposits of cerium perhydroxide were shifting from the cell walls to the vacuoles (Fig. 5d,e). Preincubation of 24 h infected roots with Na-pyruvate prevented the occurrence of cerium peroxide precipitates (Fig. 5f). Cell walls and plasma membrane were completely unstained (Fig. 5f, inset). Pyruvate has been shown to be a strong hydrogen peroxide scavenger in mammalian cells as it reacts stoichiometrically with H2O2 (Li et al., 1998). Treatment with KCN also caused a significant reduction in the cell walls of CeCl3 staining, which was still detectable in the vacuoles (Fig. 5g).
In the compatible interaction between ‘Rossol’ and the virulent pathotype, H2O2 production was seen in cells underlying the site of penetrating nematodes as early as 12 h postinoculation. Very extensive patches of CeCl3 precipitates completely covered the plasma membrane and the cell wall of the metabolically active cells (Fig. 6a). In Fig. 6b, a successful feeding site induction with a developing giant cell, in a 24 h inoculated root, is visible. At this time, H2O2 accumulation was clearly evident in the wall of the cell where the head of juvenile is located and on the outer surface of the plasma membrane, middle lamella and intercellular spaces of the modifying cells (Fig. 6b and inset). The intensity of perhydroxide formation indicates that H2O2 continued to be produced by 24 h injured tissues. At 48 h, the feeding action of the nematode induced a great development of giant multinucleate cells. Their structure differed from giant cells induced in compatible Meloidogyne–tomato interaction in that the most part of each cell was occupied by nuclei and a number of irregular and confluent vacuoles to detriment of the cytoplasm (Fig. 6c). The presence of H2O2 was weak and was detectable as randomly distributed dark spots along the cell wall and tonoplasts in the giant cells (Fig. 6c and inset).
The compatible interaction between susceptible tomato ‘Roma’ roots and the avirulent pathotype revealed that localized H2O2 generation, 12 h postinoculation, was detectable underlying the plasma membranes of cells neighbouring the nematode penetration site (Fig. 7a). Some stain was also observed in the middle lamella (Fig. 7a and inset). Staining became less intense by 24 h (Fig. 7b). At this point, a giant cell was developed and its vacuoles were lined by CeCl3 precipitates. By 48 h, when giant cells, whose dense cytoplasm indicated that organelles such as mitochondria, plastids, endoplasmic reticulum were actively synthesizing, were induced by the juvenile, cerium deposits were shifted from the plasma membrane and discarded into the vacuoles (Fig. 7c and Inset). H2O2 detection in the ‘Roma’–virulent pathotype interaction showed a similar localization than that observed in ‘Roma’-avirulent pathotype relationship (data not shown).
The specificity of CeCl3 staining for H2O2 was demonstrated by the almost complete reduction observed after treatment with pyruvate, which reacts with H2O2 producing the less toxic carbon dioxide and water. Treatment with cyanide also prevented CeCl3 deposits but to a lesser extent.
A decrease in cerium perhydroxide precipitates upon treatment with NADPH oxidase and peroxidase inhibitors suggests the involvement of these enzymes in H2O2 production and decomposition during infection. According to the results of this study, the rise in the amount of H2O2 in Rossol + avr is mainly caused by the combined action of NADPH oxidase and peroxidases.
Results on the quantification of CeCl3 deposits in the assessed sites by Density Meter 1.0 Image Program are presented in Table 2. Application of the χ2 test revealed a significant difference at P < 0.001 between incompatible and compatible tomato–nematode interactions, during the time-course of infection.
Table 2. Quantification of H2O2 amount within tomato ‘Rossol’ root cells infected with avirulent and virulent Melodoigyne incognita J2, respectively, using the imaging software density meter 1.0
avr, avirulent; vir, virulent. The presence of H2O2 was estimated by counting cerium perhydroxide deposits in the selected areas and statistically analysed with the χ2 test. Statistical analysis showed a significant reduced frequency of CeCl3 deposits in Rossol + vir J2 compared with Rossol + avr J2 (P < 0.001).
1 Precipitates per area (30 × 30 pixel2) digitized at 300 dpi.
Time after inoculation
Although a root-knot nematode infecting a root causes little damage as it migrates between the cells, ROS overproduction can be detected as early as initial J2 penetration. In general, the initial reaction of a susceptible cultivar is similar to that of a resistant host and may be the result of nematode secretions into the plant tissues during its migration (Davis et al., 2000; Huang et al., 2004). One of the main questions is whether overproduction and accumulation of ROS as a result of contact with the pathogen is a consequence of resistance or a hallmark of successful pathogenesis. There are strong suggestions that, in tomato plants, higher generation of ROS, especially H2O2, as a result of fungal infection, appears to be an important element of disease-resistance mechanisms (Borden & Higgins, 2002; Mellersh et al., 2002). A number of studies have also demonstrated that different ROS may trigger opposite effects in plants, depending on the intensity and time of their generation (Delledonne et al., 2002; Mellersh et al., 2002).
The present study showed that tomato plants reacted to root-knot nematodes by mounting defence responses. The timing and extent of the reaction differed, providing some clues as to which of these responses may be effective in defence. In the initial steps of the interaction between the plant and the nematode, J2 penetration into the root tissues was equally low in compatible and incompatible situations. Therefore, the ROS production detected at that time in the four interactions tested is most probably the result of some nonspecific mechanical perturbation of the external cells of the root, independently of the further development of the interaction. In the later stages, nematode penetration was blocked in the incompatible interaction only, while J2 can invade the roots to develop their feeding sites in the compatible interactions. At the same time, a second oxidative burst associated with HR was only observed in the incompatible interaction. In the compatible tomato–nematode relationships, ROS and H2O2 generation was seen at the time of pest invasion (12 h after inoculation) and became cytologically undetectable at 48 h, concomitantly with giant cell induction. These changes in the oxidative response of the plant are reminiscent of other plant–pathogen interactions where different levels and kinetics of ROS production activate different responses (Lamb & Dixon, 1997).
The noticeable differences in ROS production and the time-course of ROS generation observed in the different tomato–nematode interactions showed that superoxide induction is an early event during nematode infection. It may be supposed that the high concentration of ROS detected in the incompatible relationship can be related to the rapid induction of host cell death. High concentration of hydrogen peroxide, the essential precursor of hydroxyl radicals, also appears to play a key role in the oxidative events during early stages of infection. It is known that H2O2 acts as a signalling molecule that triggers gene activation, or as a cofactor in a process that requires new gene expression for both localized cell death and induction of defence genes in adjacent cells (Mellersh et al., 2002). Excess of H2O2, found to be produced during the HR in ‘Rossol’–avirulent pathotype interaction, suggests its direct role as an antimicrobial agent and as the cause of localized membrane damage at the site of nematode infection. The source of H2O2 is not clear, but the largely apoplastic location of cerium perhydroxide deposits in both internal cell wall regions and on the surface of the plasma membrane indicate that either the plasma membrane or the cell wall is the primary site of the ROS/H2O2 generator. Furthermore, the CeCl3 deposits on the plasma membrane have a distinct spatial pattern which suggests the presence of a single origin, presumably NADPH oxidase complex. Experiments with inhibitors of possible sources for H2O2 in the cell walls also seem to provide a good system in defining the subcellular site for ROS production. The pattern of ROS and H2O2 was completely different with both pathotypes between 12 h and 48 h after inoculation. It seems that the first 24 h in the plant–nematode interaction are critical for determining the plant response to avirulent or virulent nematodes. Thus, the oxidative burst at the early stage of infection appears to be involved in the nonspecific control of nematode infection. Because during the plant pathogen defence response the plant simultaneously produces more ROS while decreasing its ROS scavenging capacities, accumulation of ROS and activation of cell death occurs. The suppression of ROS detoxifying mechanisms is crucial for the onset of cell death (Apel & Hirt, 2004). According to the literature, during the incompatible reaction, when the nematode is detected and defence response, including cell death, is induced, the initial and rapid accumulation of H2O2 is followed by a prolonged burst of H2O2 production. Although the very early (i.e. poor J2 penetration and ROS production) and late events (i.e. high J2 penetration, no ROS production or HR and feeding site induction) are very similar in the three compatible interaction tested, significant differences can be seen when the resistant tomato is infected with the virulent nematode line. In particular, the kinetics of ROS production at the intermediary time (24 h postinoculation), and the cellular organization of the induced giant cells, significantly changed compared with the two compatible situations where the susceptible cultivar is involved. In addition, during the compatible ‘Rossol’–virulent pathotype interaction, only the first peak of H2O2 occurs, confirming that H2O2 not only acts as causal trigger for HR but also activates genes encoding enzymes that prevent cells from oxidative damage (Levine et al., 1994). This suggests that if general oxidation of the root tissues, attributable to H2O2 generation, had occurred, it was transient and reversed by 24 h after inoculation. Short duration of the burst, limited to the very first stage of infection, seems to be linked to the suppression of cell death or other defences and may be determined by the nematode which is able to overcome the defence lines and to infect the host plant.
Reactive oxygen species are known to be produced in different subcellular compartments and to influence the expression of a large number of genes in plants (Neill et al., 2002). Thus, ROS at interaction sites may have different roles as players in the elicitation or prevention of cell death, depending on their concentration, subcellular localization, and duration of the burst. This suggests that, depending on the kind of stimulus, cells have evolved strategies to utilize ROS with a high degree of specificity as signals to control different stress response (Laloi et al., 2004). During the incompatible–pathogen interaction, superoxides are produced enzymatically outside the cell and are rapidly converted to H2O2, which can cross the plasma membrane. Extracellular peroxidases (POXs) are considered to catalyse H2O2-dependent mechanism of ROS generation (Kawano, 2003). It is also well recognized that plants possess a plasma membrane (PM)-NADPH oxidase as a source of ROS production, and this suggests that both POXs and PM-NADPH oxidase are the major sources for ROS production as defence mechanisms during biotic stresses (Yoshioka et al., 2003). Evidence supporting a differential ROS generation where different stimuli induced ROS generation have also been described (Bolwell, 1999). Alternatively, different subcellular antioxidants might contribute to local redox changes visualized by distinct NBT staining patterns and DCFH reaction in the compatible interaction.
Overall, these results strongly suggest that spatio-temporal differences in the production and accumulation of ROS and H2O2 generation in the nematode–host interaction might form a discrete unit for defence response. However, no single defence mechanism has been unequivocally proven to operate in plant cells. In that respect, further investigations are currently addressing the questions of how other signalling pathways such as intracellular nitric oxide (NO) cooperates in the regulation of H2O2 accumulation and modulates the expression of defence-related genes during nematode infection, as reported in other plant–pathogen interactions (Tada et al., 2004).
The authors thank R. Lerario for his excellent assistance with photographs, and Drs P. L. Mazzeo and L. Capozzo, ISSIA, CNR, Bari, Italy for setting up the Density Meter 1.0 Image Program.