Introduction

Plants have evolved specific defense systems to protect themselves against attack from a wide range of pathogens (Staskawicz et al. 1995; Heath 2000). Upon pathogen challenge, a necrotic lesion is formed at the site of pathogen entry, thereby preventing further spread of disease. This event is referred to as the hypersensitive response (HR).

The HR is initiated by recognition of pathogen attack, followed by an oxidative burst, induction of defense-related gene expression and hypersensitive cell death. There are two types of HR; one is cultivar-specific HR, in which a plant directly or indirectly recognizes the particular protein (avirulent protein) derived from the pathogen through the corresponding resistance (R) gene product (Hammond-Kosack and Jones 1997). This type of HR, often called host HR, is limited to a particular pathogen, and is referred to as a “gene-for-gene” mechanism (Keen 1990). Another type of HR is non-specific to particular pathogens. In this type of resistance, most plant species exhibit resistance to pathogens by responding to biochemical changes upon infection (Heath 2000). This type of HR, known as nonhost HR, is commonly observed in many plants in response to a variety of phytopathogens. As nonhost HR is effective against a broad spectrum of pathogens and is the predominant form of pathogen resistance among plants, its application might be useful to generate resistant crops. Its molecular mechanisms, however, are not yet completely understood, although similarities between host- and nonhost HRs have been documented. For example, an oxidative burst is induced during both HRs (Huckelhoven et al. 2001). Several defense signaling components functioning in host HR were also found in nonhost HR, including phytohormones (salicylic acid and ethylene) (Mellersh and Heath 2003; Knoester et al. 1998), MAP kinases (WIPK and SIPK) (Sharma et al. 2003), and ubiquitin ligase-associated protein SGT1 (Peart et al. 2002). It is thus conceivable that host- and nonhost HRs share at least some mechanisms.

Polyamines are small, positively charged aliphatic amines. They are involved in a multitude of cellular functions in plants, including growth, development and stress responses (Walden et al. 1997). Although the exact molecular basis of their biological function is still to be determined, their levels are known to fluctuate during plant–microbe interactions (Walters 2000). For example, during the incompatible interaction between barley and powdery mildew, polyamine levels greatly increased, accompanied by increased activity of the catabolic enzymes diamine oxidase and polyamine oxidase (Cowley and Walters 2002). Previously, we demonstrated that polyamines serve as the major source for hydrogen peroxide during host HR upon infection by tobacco mosaic virus (TMV) in tobacco plants carrying the resistant (N) gene (Yoda et al. 2003). We also showed that synthesized polyamines are efficiently degraded by polyamine oxidase to release hydrogen peroxide, which largely contributes to the second phase of the oxidative burst to induce hypersensitive cell death (Yoda et al. 2003, 2006). Thus, it was clear that polyamines, and their degradation product, hydrogen peroxide, were one of the key elements to induce host HR in tobacco plants.

Questions then arose whether or not a similar mechanism functions in nonhost HR in tobacco, and whether or not polyamines are commonly used during resistance in other plant species. In order to address these questions, we analyzed the interaction between tobacco plants (Nicotiana benthamiana) and its nonhost pathogen Pseudomonas cichorii (Sharma et al. 2003). We also examined the contribution of polyamines to host HR in Arabidopsis and rice. We show here that polyamine-derived hydrogen peroxide is critical not only for host HR but also for nonhost HR in tobacco plants, and that polyamines are commonly employed in other plant species in defense against pathogens.

Materials and methods

Plant materials and treatments

Tobacco (Nicotiana benthamiana) and rice (Oryza sativa cv Kinmaze) plants were grown in a growth cabinet at 23°C and 30°C, respectively, under a 14/10 h light/dark cycle. Arabidopsis thaliana (Colombia) was grown in a growth cabinet at 22°C under a 9/15 h light/dark cycle. Pseudomonas cichorii and P. syringae DC3000 (avrRpm1) were cultured in King’s B medium containing 20 g/l Bacto peptone, 20 g/l Bacto tryptone, 20 g/l glycerol, 0.5 g/l K2HPO4 and 0.5 g/l KH2PO4, and diluted with 10 mM MgCl2 (OD600 = 0.01 and 0.05, respectively) before inoculation. Tobacco leaves were infected with P. cichorii and Arabidopsis leaves were infected with P. syringae DC3000 (avrRpm1). Polyamines were infiltrated into tobacco and Arabidopsis leaves at a final concentration of 1 mM or 10 mM using a syringe without a needle, and leaves incubated at 23°C for appropriate time intervals under continuous light. Rice leaves were inoculated with Magnaporthe grisea (race 031) (1 × 105 conidia/ml) as described (Iwai et al. 2006) and incubated at 30°C under a 14/10 h light/dark cycle. Polyamines were absorbed at the cutting site of rice leaves and incubated at 30°C for appropriate time intervals under continuous light.

Polyamine quantification

Apoplastic fluids were extracted from leaf disks prepared from tobacco or Arabidopsis (15 mm in diameter), or from rice leaf segments (2–3 cm in length). Each leaf disk or leaf segment was weighed and submerged in water in vacuo. Leaf extract was collected by a brief centrifugation by placing samples in a 10-ml syringe, which was set in a 50-ml Falcon tube. Amines in extracted fluids were quantified either by HPLC as previously described (Yoda et al. 2003) or by capillary electrophoresis mass spectrometry (Soga and Heiger 2000).

RNA gel-blot analysis

Total RNA was isolated from leaf samples by the guanidium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi 1987), and gel-blot hybridization was performed as previously described (Yoda et al. 2002). Probes for each gene were synthesized with a pair of specific primers, which were designed based on the cDNA from N. tabacum: forward (5′-GGCTCATTACCAAGGTGTTTATCC-3′) and reverse (5′-CCAAGCCCACCAATATCAATG-3′) for NbADC; forward (5′-TCAGCTTGGATAAGAATAAGCGAGG-3′) and reverse (5′-ATGGCCGGCCAAACAATCATCG-3′) for NbODC; forward (5′-GGCCTTGCCTGTCTCTGCCATTGG-3′) and reverse (5′-TTGGTAGCAACATCAGCATGCAAAGC-3′) for NbSAMDC; forward (5′-TGAGATTAGCCCCCTTTGGCCTGGGG-3′) and reverse (5′-GCAAATGATGGCAAACAGAATGATGC-3′) for NbSPDS; forward (5′-GGAAAAGTTCTAGTGCTAGATGG-3′) and reverse (5′-GGATCTGATGAATCAACTATGATAGC-3′) for NbSPMS; and forward (5′-AAGAAAGCGGTGGACTCGGC-3′) and reverse (5′-TCGTGGGCGTAGATGAAGAACTCC-3′) for NbPAO.

Virus-induced gene silencing

Using cDNA derived from P. cichorii-infected N. benthamiana leaves as a template, a partial fragment of NbPAO was amplified by PCR with a pair of synthesized primers; forward containing BamHI site (5′-GGATCCAAGAAAGCGGTAGACTCGGC-3′) and reverse containing KpnI site (5′-GGTACCTCGTGGGCGTAGATGAAGAAC-3′). The resulting fragment was introduced into corresponding sites of a pTV00 vector derived from Tobacco rattle virus (TRV), which was transformed into A. tumefaciens GV3101 strain. The resultant GV3101 strain and A. tumefaciens C58C1 strain containing pBINTRA6 (Ratcliff et al. 2001) were grown to saturation in LB medium. Cells were collected by centrifugation and resuspended in a solution containing 10 mM MgCl2, 10 mM MES and 100 μM acetosyringone (OD600 = 0.5), and kept at room temperature for 2 h. Each culture containing pBINTRA6 or pTV00-derived constructs (or pTV00 as a control vector) was mixed in a 1:1 ratio, and infiltrated into a lower leaf of 3-week old tobacco plants using a syringe without a needle. Infiltrated tobacco plants were incubated at 25°C under a 15/9 h light/dark cycle for 3 weeks and used for further experiments.

Detection of hydrogen peroxide

Leaf samples infected with pathogen or infiltrated with polyamines were incubated for indicated time intervals and then submerged in 3,3′-diaminobenzidine (DAB) solution (1 mg/ml, pH 3.8) for 8 h. DAB deposits indicative of DAB oxidation products were visualized after brief boiling in ethanol as previously described (Yoda et al. 2003).

Bacteria growth assay

A leaf disk (15 mm in diameter) from infected leaves with P. cichorii was ground in 10 mM MgSO4 solution. The resultant crude extract was diluted to the appropriate concentration, and spread onto King’s B medium. Bacterial numbers in leaves were calculated from colony units formed on the medium.

DNA fragmentation assay

DNA was extracted from leaves infected with P. cichorii by the cetyltrimethyl ammonium bromide method (Murray and Thompson 1980), fractionated on a 2% agarose gel and visualized by staining with ethidium bromide as previously described (Yoda et al. 2006).

Estimation of cell death

Leaf samples treated with amines (infiltrated or absorbed) were submerged in a solution containing 2.5 mg/ml trypan blue, 25% lactic acid, 23% acid phenol, and 25% glycerol at 70°C for 15 min in vacuo. This step was repeated twice. After boiling for 2 min, samples were incubated at room temperature for 1 h. Samples were recolorized with chloral hydrate (2.5 g/ml) several times and fixed with 70% glycerol to visualize stained regions.

Results

Polyamine accumulation during nonhost HR

In our previous studies with TMV and tobacco plants (N. tabacum) carrying the resistant (N) gene, by which the host HR is induced, we demonstrated that polyamines were accumulated in apoplasts of infected leaves, and degraded by polyamine oxidase, releasing hydrogen peroxide that causes hypersensitive cell death (Yoda et al. 2003). In the present study, we examined whether or not polyamines similarly function during the nonhost HR. P. cichorii was previously reported to induce a nonhost HR on N. benthamiana (Sharma et al. 2003). When bacteria were inoculated onto healthy leaves of N. benthamiana, an HR lesion began to appear approximately 12 h later, and hypersensitive cell death was evident at ~24 h (Fig. 1a). To determine whether or not polyamines were involved in this response, apoplastic fluids were collected and subjected to the HPLC analysis. Putrescine was found to be present in apoplasts of intact leaves at a low level, and increased up to 10-fold 24 h after infection (Fig. 1b). Spermidine was scarcely detectable before infection, but began to accumulate after infection up to 24 h (Fig. 1b). The control plants infiltrated only with MgCl2 solution (mock treatment) also exhibited a slight increase of both amines, this being probably due to infiltration stress (Fig. 1b). The amount of polyamines in intracellular space was then examined. The basal level of putrescine was high, showing approximately 80 nmole per gram fresh weight, and increased up to 3-fold in infected leaves 24 h later (Fig. 1c). In control samples (mock-treated leaves), such an increase was not observed (Fig. 1c). The basal level of spermidine was also high, showing 100 nmole per gram fresh weight, and this level was maintained for 24 h in treated and untreated leaves (Fig. 1c). Spermine was not detectable at any experimental time points in treated and untreated leaves in both apoplasts and intracellular spaces (data not shown). To take account of amines from P. cichorii, the amount of each amine was examined in excess P. cichorii (1 × 108 cfu). The amount of spermidine derived from P. cichorii was 0.92 ± 0.05 nmol, while those of putrescine and spermine were undetectable. This result indicated that amine compounds derived from P. cichorii were negligible to estimate the total amine contents in inoculated leaves.

Fig. 1
figure 1

Nonhost HR and polyamine accumulation. a Formation of necrotic lesion. A healthy leaf of wild-type tobacco (N. benthamiana) was infiltrated with P. cichorii, and observed for visible necrotic lesion formation after 12 h and 24 h. b, c Accumulation of putrescine and spermidine in apoplast (b) and in intracellular space (c). Apoplastic fluid was extracted from leaves infected with P. cichorii (closed circle) or mock-treated (open circle) at indicated time points. Intracellular amines were extracted from remaining leaf-disks after extraction of apoplastic fluid. Values are means from four independent assays with standard deviations with the significance of difference of the 95% confidence intervals (P < 0.05)

Full size image

Induction of transcripts for polyamine metabolic enzymes

Polyamines are synthesized and catabolized through multiple steps, each step being catalyzed by a specific enzyme. To examine expression profile, cDNAs encoding these enzymes were newly isolated, and designated as NbADC, NbODC, NbSAMDC, NbSPDS, NbSPMS and NbPAO. When healthy leaves were challenged by the nonhost pathogen, P. cichorii, all transcripts for these genes were simultaneously induced (Fig. 2). Transcripts for NbADC, NbODC, NbSAMDC, NbSPDS and NbPAO began to accumulate 6 h after infection, reaching a plateau 12 h later with a gradual decline thereafter (Fig. 2). Transcripts for NbSPMS were low and began to accumulate 12 h after infection (Fig. 2). These results suggest that increase of amines during the nonhost HR is due to an increase in expression of the relevant genes.

Fig. 2
figure 2

Accumulation profile of transcripts for enzymes involved in polyamine metabolism during the nonhost HR. Healthy leaves of wild-type tobacco were treated with P. cichorii or with water (mock-treated) for indicated time period. Total RNA was extracted from leaves, and subjected to RNA gel-blot analysis with the indicated probe labeled with 32P. rRNA was used as loading control. Nucleotide sequence data reported here are available in the DDBJ/EMBL/GenBank databases under the accession numbers: NbADC; AB304777, NbODC; AB304778, NbSAMDC; AB304782, NbSPDS; AB304779, NbSPMS; AB304780, and NbPAO; AB304781

Full size image

Hydrogen peroxide production by polyamine oxidase

Polyamine oxidase is essential for generating hydrogen peroxide to induce programmed cell death during the host HR (Yoda et al. 2003). To investigate the possibility that polyamine oxidase is also required in the nonhost HR, the virus-induced gene silencing (VIGS) method was employed to transiently silence the gene encoding polyamine oxidase from N. benthamiana (NbPAO). A partial cDNA fragment for NbPAO was isolated with a pair of primers, which were designed based on the cDNA for a polyamine oxidase from N. tabacum (NtPAO). The deduced polypeptide with 188 amino acids of NbPAO was similar to that of NtPAO (Fig. 3a), indicating the isolated cDNA was an ortholog in N. benthamiana. Total suppression of NbPAO expression was confirmed by RT-PCR (Fig. 3b). Using these materials, we first directly infiltrated polyamines into leaves, and observed lesion formation. In wild-type plants, spermidine and spermine at a concentration of 1 mM induced HR-like cell death 72 h and 24 h after infiltration, respectively, whereas putrescine did not (Fig. 4a). The effect appeared to be dose-dependent, as HR-like cell death was observed faster and clearer upon infiltration with 10 mM polyamines in comparison with 1 mM (Fig. 4a). Hence we performed further experiments with a fixed concentration of amines at 10 mM. In NbPAO-silenced tobacco plants, hydrogen peroxide was not produced and cell death was not induced either (Fig. 4b). Subsequently, we examined effects of P. cichorii infection. In wild-type plants, hydrogen peroxide was produced and hypersensitive cell death was induced (Fig. 4c). In NbPAO-silenced tobacco plants, neither hydrogen peroxide nor cell death was observed (Fig. 4c). The number of bacteria grown during infection was then examined. When 107 cfu/ml of P. cichorii was used as the inoculum, the number equally increased up to 2 × 108 cfu/cm2 during the first 2 days in both control and NbPAO-silenced plants. However, the growth thereafter was markedly suppressed in the control plants due to the nonhost HR, whereas growth in NbPAO-silenced plants increased up to 6 × 108 cfu/cm2 5 days after infection (Fig. 4d). Further analysis indicated that, upon P. cichorii infection, genomic DNA of wild-type plants was markedly degraded, showing fragmentation patterns, while that of NbPAO-silenced plants apparently remained intact (Fig. 5). These results strongly suggested that hypersensitive cell death took place in the wild-type but not in the transgenic plants.

Fig. 3
figure 3

Construction of NbPAO silencing tobacco. a Partial amino acid sequence alignment of NbPAO with NtPAO (accession no. AB200262). Numbers on the left side indicate amino acid position. b Confirmation of VIGS silencing. Three-week old tobacco plants (N. benthamiana) were infected with A. tumefaciens (GV3101 strain) containing the pTV00-NbPAO construct or empty vector (control) and cultivated under standard conditions. Leaves were detached from control (control) or NbPAO-silenced (VIGS-NbPAO) plants 3 weeks later and infected with P. cichorii. Total RNA was extracted from each infected leaf 24 h after infection and subjected to RT-PCR. rRNA was used as the internal control

Full size image
Fig. 4
figure 4

Effects of NbPAO silencing on the nonhost HR. a Lesion development by polyamine infiltration. Healthy leaves of wild-type (WT) tobacco were infiltrated with water or indicated amines at 1 mM (left panel) or 10 mM concentration (right panel) using a syringe without a needle. Necrotic lesion due to cell death was observed 72 h (1 mM) and 24 h (10 mM) after infiltration, respectively. Put putrescine, Spd spermidine, Spm spermine. b Polyamine-induced cell death and hydrogen peroxide production. Healthy leaves of wild-type tobacco treated with empty vector (control, upper panel) or NbPAO-silenced plants (VIGS-NbPAO, lower panel) were infiltrated with water or indicated amines (left panel). A necrotic lesion was observed 24 h later (middle panel), and hydrogen peroxide production was observed 12 h later (right panel). c Pathogen-induced cell death and hydrogen peroxide production. Healthy leaves of wild-type tobacco treated with empty vector (control, upper panel) or NbPAO-silenced plants (VIGS-NbPAO, lower panel) were infected with P. cichorii (left panel), and nonhost hypersensitive cell death was observed 24 h later (middle panel). Hydrogen peroxide production was observed 12 h after infection (right panel). d Bacterial population in inoculated region. Healthy leaves of control (Control; open circle) or NbPAO-silenced plants (VIGS-NbPAO; closed circle) were infected with P. cichorii, and the number of bacteria determined at indicated time points. Values are means from three independent assays with standard deviations

Full size image
Fig. 5
figure 5

DNA fragmentation. Healthy leaves of wild-type tobacco treated with the empty vector (control) or NbPAO-silenced plants (VIGS-NbPAO) were infected with P. cichorii, and genomic DNA was extracted 24 h later. After fractionation by 2% agarose gel electrophoresis, DNA was visualized by staining with ethidium bromide

Full size image

Polyamines in other plant species

The above-mentioned experiments indicated that polyamines are a direct substrate for hydrogen peroxide production in both the host- and nonhost HRs in tobacco plants. We then aimed to determine whether polyamines similarly function in the HR of other plant species. Healthy leaves from Arabidopsis (dicot) and rice (monocot) plants were inoculated with respective host pathogens, P. syringae DC3000 (avrRpm1) and M. grisea (race 031), and examined for polyamine production. Results showed that putrescine and spermidine accumulated in apoplasts after inoculation in both plant species (Fig. 6a, b). In Arabidopsis, the basal levels of putrescine and spermidine were low and reached 12 nmol and 11 nmol per gram fresh weight 24 h after inoculation, respectively (Fig. 6a). In rice, putrescine was hardly detectable before inoculation, but increased to more than 10 nmol per gram fresh weight 21 days after inoculation (Fig. 6b). The basal level of spermidine in rice was undetectable, but began to increase upon infection, reaching 1.7 nmol per gram fresh weight 7 days later (Fig. 6b). This level was maintained up to 21 days, although its total amount was lower than that of putrescine. The absolute level of spermidine in rice was 1/7 that of the level in Arabidopsis (Fig. 6a, b). Polyamine level in rice was undetectable under mock-treated leaves (data not shown).

Fig. 6
figure 6

Requirement of polyamines for HR induction in Arabidopsis and rice. a Accumulation profile of amines in Arabidopsis. Healthy leaves of Arabidopsis were inoculated with (closed circle) or without (open circle) the host pathogen, P. syringae DC3000 (avrRpm1), and amounts of putrescine (upper panel) and spermidine (lower panel) in apoplasts was determined. Values are from three independent assays with standard deviations. b Accumulation profile of amines in rice. Healthy leaves of rice plants were inoculated with the host pathogen M. grisea (race 031), and amounts of putrescine (open circle) and spermidine (closed circle) in apoplasts were determined at indicated time points. Values are from three independent assays with standard deviation. c, d Polyamine-induced cell death in Arabidopsis (c) and in rice (d). Indicated amine compounds were infiltrated into leaves, and hydrogen peroxide production (left panel) and cell death (right panel) were observed 12 h and 24 h later, respectively. Put putrescine, Spd spermidine, Spm spermine

Full size image

Induction of hydrogen peroxide and cell death by polyamines

Effects of exogenously applied polyamines were then examined in a similar manner as performed with tobacco plants (Fig. 4b). Healthy leaves from Arabidopsis and rice plants were infiltrated with putrescine, spermidine or spermine, and hydrogen peroxide production and cell death were analyzed. Results showed that spermidine and spermine efficiently induced production of hydrogen peroxide and concomitant cell death in both plants, whereas putrescine did not (Fig. 6c, d).

Discussion

Plants defend themselves against the majority of pathogens through the combination of host HR and nonhost HR systems. Since the host HR is usually race-specific, it is restricted to particular pathogen species. In contrast, the nonhost HR is observed across all plant species, and is effective against a broad spectrum of pathogens. It is more common and more durable than the host HR (Mysore and Ryu 2004). In this article, we documented that polyamine-derived hydrogen peroxide catalyzed by polyamine oxidase is sufficient to induce the hypersensitive cell death during the nonhost HR in tobacco plants. This observation is consistent with the case of the host HR against TMV (Yoda et al. 2003), suggesting that both host- and nonhost HRs share a common system utilizing polyamines as the source of hydrogen peroxide.

Experimentally we showed that silencing of NbPAO results in the suppression of hydrogen peroxide production and eventually of the HR onset. We also found that the in vivo levels of putrescine and spermidine are temporally and spacially controlled during bacterial infections. Spermidine and spermine, but not putrescine, are considered to be substrates of NbPAO as in the case for NtPAO (Yoda et al. 2003), both showing a close similarity in sequence. Although the in vivo substrates of PAO from the various organisms may be diverse, including polyamine derivatives and conjugates, we speculated that in N. benthamiana spermidine and spermine possibly serve as a source of hydrogen peroxide during the HR through their oxidation. Recent findings have shown that in Arabidopsis spermidine and spermine are oxidized through a polyamine back-conversion pathway resulting in the production of putrescine and spermidine, respectively (Tavladoraki et al.; Moschou et al. 2008; Kamada-Nobusada et al. 2008). Thus, in the case of tobacco plants, putrescine and spermidine accumulated in apoplast are highly probable to be produced through this pathway. This idea accounts for the undetectable level of spermine in apoplasts, suggesting its complete back-conversion to spermidine by NbPAO. Although spermine was reported to be accumulated in extracellular space during HR in tobacco plants upon TMV infection (Yamakawa et al. 1998), the difference might be due to the different protocols for polyamine extraction from the extracellular space, being under non denaturing conditions in our present work. Since polyamine oxidase from barley was reported to be in vacuole and that from Ararbidopsis in peroxisome (Cervelli et al. 2004; Moschou et al. 2008; Kamada-Nobusada et al. 2008), it is conceivable that polyamine dependent production of hydrogen peroxide also takes place in the intracellular space. It remains to be determined, however, whether or not thus-produced hydrogen peroxide contributes to the HR. Putrescine is generally considered to be degraded by copper amine oxidases (CuAO) (Cona et al. 2006), but in the present case, accumulated putrescine appeared not to be degraded, since NbPAO-silenced plants did not produce detectable level of hydrogen peroxide when infiltrated with putrescine. This suggested the absence of CuAO activity in apoplasts during the early phase of nonhost HR, being consistent with a previous report showing the absence of extracellular enzymes catabolizing putrescine in tobacco protoplasts (Papadakis and Roubelakis-Angelakis 2005). Since putrescine is known to serve as an antioxidant (Groppa and Benavides 2008), and/or to prevent programmed cell death syndrome (Papadakis and Roubelakis-Angelakis 2005), one of the biological functions of accumulated putrescine could be to prevent excess hypersensitive cell death.

Expression of genes for polyamine metabolism is temporally controlled, transcripts beginning to accumulate 6 h after inoculation. This induction profile is similar to our previous observation on the host HR (Yoda et al. 2003). It is also consistent with our previous observation that polyamine-derived hydrogen peroxide is generated at the late phase of the oxidative burst in BY2 cells that were elicited by cryptogein (Yoda et al. 2006). Such a late and simultaneous induction of gene expression suggested that some upstream signaling molecule(s) are involved. Methyl jasmonate is one of the candidates, as it has been shown to induce expression of ODC, ADC, SAMDC and SPDS, whereas these are not induced by salicylic acid or ethylene (Imanishi et al. 1998; Walters et al. 2002; Yoda unpublished observation).

In the host HR, we proposed that, upon pathogen attack, genes for polyamine metabolism are transcriptionally activated. This results in biosynthesis and accumulation of polyamines in apoplasts, and the polyamine-derived hydrogen peroxide directly contributes to hypersensitive cell death (Yoda et al. 2003, 2006). In the nonhost HR, we also found that hydrogen peroxide generated from polyamines is the major signaling molecule for inducing hypersensitive cell death. These observations strongly suggest that the induction mechanism of the oxidative burst is common between the host- and nonhost HRs. Since other sources for oxidative burst, for instance NADPH oxidase, have also been reported for both host- and nonhost HRs (Bestwick et al. 1998), the oxidative burst could possibly be differentially regulated in temporal and spatial manners during hypersensitive cell death.

A notable feature of the present study is that polyamines accumulated in apoplasts of Arabidopsis and rice during the host HR, and that exogenously applied polyamines induced HR-like cell death in both plant species. These findings are consistent with our previous observation in tobacco plants (Yoda et al. 2003). It appears that polyamines are commonly utilized among plant species, although the accumulation profiles and amounts of polyamines varied in each plant species. The level of putrescine was comparable in both plants, although its accumulation rate in rice was much slower than in Arabidopsis. The level of spermidine in Arabidopsis increased after pathogen infection, showing a similar profile as in tobacco. In contrast, its level in rice was lower and increasing rate was slower than in Arabidopsis and tobacco plants. The low level of spermidine in rice suggests that rice may be more sensitive to low levels of hydrogen peroxide in comparison with Arabidopsis and tobacco plants. It is also conceivable that the kinetics of polyamine metabolism varies depending on the pathogen species; virus, bacteria and fungus, of which infection mode differs from each other. Overall, we propose that polyamines are the common source for hydrogen peroxide during the host HR among plant species. It is also tempting to speculate that, judging from the case in tobacco plants, polyamines may also commonly serve as the source for hydrogen peroxide in the nonhost HR in several plant species.