Polyamine metabolism is upregulated in response to tobacco mosaic virus in hypersensitive, but not in susceptible, tobacco

Authors


Author for correspondence: P. Torrigiani Tel: +30 0512091291 Fax: +39 051242576 Email:torrigia@alma.unibo.it

Abstract

  • • Change is reported in the biosynthetic and oxidative activity of hypersensitive (NN) and susceptible (nn) tobacco (Nicotiana tabacum) plants in response to tobacco mosaic virus (TMV).
  • • Mature leaves of nn and NN tobacco were collected over 0–72 h as uninoculated controls or after inoculation with TMV or phosphate buffer (mock-inoculation). The polyamine response to inoculation was analysed by measuring activity and gene expression of S-adenosylmethionine decarboxylase (SAMDC), arginine-(ADC) and ornithine decarboxylases (ODC); incorporation of labelled putrescine; and activity of diamine oxidase (DAO).
  • • In NN leaves SAMDC activity and transcript levels, and DAO activity increased in the TMV-inoculated plants but not in the other treatments; a two-fold increase in DAO activity was seen after 72 h. Both ADC and ODC activity increased in NN leaves at 72 h in TMV-inoculated plants; ADC mRNA increased with activity. The increase in SAMDC mRNA (24 h) preceded the rise in activity (72 h). [3H]putrescine added to NN leaves led to enhanced label recovery and incorporation into spermidine and spermine in TMV-inoculated plants. No significant changes in biosynthetic or oxidative activity occurred in nn plants.
  • • After TMV inoculation, NN, unlike nn, tobacco plants upgrade polyamine synthesis and oxidation; this leads to changes in cellular components which might induce programmed cell death.

Introduction

It has been purported for many years that the diamine putrescine and the polyamines spermidine and spermine, generally known as polyamines, are involved in the response to pathogen attack in that both fungal and bacterial pathogens elicit the production of their amide conjugates (Martin-Tanguy, 1985). The most studied function of free polyamines, which are polycationic components of prokaryotic and eukaryotic cells and also part of virus particles (Heby & Persson, 1990; Rabiti et al., 1994; Cohen, 1998), is that of plant growth regulators (Bagni & Torrigiani, 1992) due to their complex interactions mainly with nucleic acids, but also with phospholipids, proteins (Feuerstein & Marton, 1989) and cell wall anionic polysaccharides (Messiaen & Van Cutsem, 1999). In fact, they seem to cover a large spectrum of action in that they participate in the homeostatic adjustment of cells relative to environmental changes (Flores, 1990). In addition, the endogenous polyamine pool in pathogens can be modulated using analogues and this may represent a useful tool in controlling plant diseases (Walters & Mackintosh, 1997).

Among the various host-pathogen systems, the tobacco–TMV interaction offers advantages in that much is known about the molecular markers of the defence response, such as salicylic acid (SA), ethylene, jasmonate and pathogenesis-related (PR) proteins. Moreover, the N gene confers resistance to TMV in the hypersensitive (NN genotype) plants which develop a hypersensitive response (HR), while the absence of that gene (nn genotype) confers susceptibility to TMV and development of the systemic infection (Whitham et al., 1994). Previous studies were aimed at understanding the response to TMV in hypersensitive and susceptible tobacco (Nicotiana tabacum cv. Samsun) plants in terms of polyamine metabolism. An early positive polyamine response in TMV-inoculated leaves from NN plants is detectable already at 5 h, while not in nn plants (Rabiti et al., 1998). As far as the spatial and temporal patterns of polyamine response are concerned, in NN plants an increasing concentration gradient in free and conjugated putrescine and spermidine towards the centre of the hypersensitive lesion was reported in parallel to putrescine biosynthetic activity gradients at 3 and 5 d after inoculation (Torrigiani et al., 1997); opposite or no changes were observed in nn plants.

Recently, in TMV-inoculated tobacco leaves, a specific increase in free spermine in intercellular fluids (but not in the whole leaf tissue) was reported 4 d after inoculation (Yamakawa et al., 1998). The authors suggest a role for exogenous spermine in the salicylate-independent induction of acidic PR proteins.

Concerning the mechanism of action of free polyamines, it may be based on one or more of the following: (a) ligands for specific plasma membrane polyamine-binding proteins (Tassoni et al., 1998); (b) modulators of membrane permeability (Roberts et al., 1986); (c) precursors of hydroxycinnamoyl acid amides (HCA; Martin-Tanguy, 1985) and (d) producers of hydrogen peroxide through cell wall-located di-(DAO) and polyamine oxidases (PAO) (Federico & Angelini, 1991). The latter are involved in signalling programmed cell death (PCD, Møller & McPherson, 1998) which also mediates the HR (Mittler et al., 1997).

In this paper we further examined the time course of the polyamine response to TMV in NN and nn tobacco plants with respect to biosynthesis by analysing: activity and gene expression of S-adenosylmethionine decarboxylase (SAMDC), which leads to spermidine and spermine synthesis through the production of decarboxylated SAM (decaSAM) and the activity of specific synthases; incorporation of the labelled precursor putrescine into spermidine and spermine as an additional indication of biosynthetic activities; and activity and gene expression of arginine-(ADC) and ornithine decarboxylases (ODC) which lead to putrescine biosynthesis, the former indirectly, the latter directly. On the other hand, putrescine catabolism was evaluated through the activity of DAO which oxidatively deaminates diamines. These enzymes are involved in the modulation of diamine and polyamine levels both in the cytoplasm and cell wall. Results show that, upon inoculation with TMV, hypersensitive plants react in an opposite manner in comparison to susceptible ones by upregulating polyamine synthesis and oxidation.

Materials and methods

Plant material and TMV inoculation

TMV-susceptible (nn) and TMV-resistant (NN) plants of Nicotiana tabacum (L.) cv. Samsun were grown in the glasshouse with a photoperiod of 8 h (1.87 W m−2) at 25°C up to the vegetative stage of 15 ± 1 leaves (c. 6 wk).

The third-fifth mature leaf counting from the bottom of 5 resistant and susceptible tobacco plants was inoculated by treatment with carborundum (400 mesh size) and 200 µl purified TMV suspension (0.1 mg ml−1 0.01 M sodium phosphate buffer, pH 7; Torrigiani et al., 1995). An equal number of plants were mock-inoculated by treating each leaf with carborundum and 200 µl phosphate buffer. Noninoculated plants were considered as healthy controls. Leaves (5 per treatment) from inoculated-, mock-inoculated and control plants were collected, frozen in liquid nitrogen immediately (0 h) or at 5, 24 and 72 h after inoculation, and stored at −80°C until use.

Enzyme activity assays

All enzyme procedures were carried out in an ice bath. For SAMDC (EC 4.1.1.21) activity tissue (200–300 mg f. wt) was homogenised with 3 volumes of 100 mM Tris-HCl buffer, pH 7.6, containing 50 µM EDTA and 25 µM pyridoxal phosphate. The homogenate was centrifuged at 20 000 g for 30 min at 4°C and the supernatant was used for the enzyme assay. SAMDC activity was determined by incubating 0.2 ml aliquots of the supernatant with 3.7 kBq S-adenosyl-L-[carboxyl-14C]-methionine (2.22 TBq mol−1, Amersham Pharmacia Biotech Italia, Milano, Italy) and measuring the rate of 14CO2 evolution from the labelled substrate during a 2-h incubation at 37°C as previously described (Scaramagli et al., 1999a). The liberated 14CO2 was trapped in 1 M KOH and analysed by scintillation counting in a Beckman LS 6500 counter.

Putrescine oxidising (DAO, EC 1.4.3.6) activity was assayed by a radiometric method based on the production of [14C]Δ1-pyrroline from labelled putrescine as previously described (Scaramagli et al., 1999a). Samples were homogenised in 100 mM potassium phosphate buffer, pH 8, containing 2 mM dithiothreitol, and centrifuged at 20 000 g for 30 min at 4°C. Aliquots (0.2 ml) of supernatant or resuspended pellet were incubated at 37°C for 30 min in the presence of 7.4 kBq [1,4–14C]putrescine (4.03 GBq mmol−1, NEN, Boston, MA, USA) in the absence or in the presence of 100 µM unlabelled putrescine.

After adding 2% (w/v) sodium carbonate to stop the reaction, the [14C]pyrroline was immediately extracted in 1 ml toluene and 500 µl of the lipophilic phase were withdrawn, added to 2 ml scintillation liquid (Ultima Gold, Beckman Analytical, Milano, Italy) and the radioactivity counted. In preliminary experiments enzyme activity was measured in relation to putrescine concentration (0–1 mM) and the leaf homogenate or 20 000 g pellet were subjected to sonication with an MSE 150 W ultrasonic disintegrator for 3 cycles of 30 s each as described in Torrigiani et al. (1995).

ADC (EC 4.1.1.19) and ODC (EC 4.1.1.17) activities were measured as previously described by Rabiti et al. (1998). Samples were extracted in 5 volumes of ice-cold 0.1 M Tris-HCl buffer, pH 8.5, containing 50 µM pyridoxal 5-phosphate, and centrifuged at 20 000 g for 30 min at 4°C. Enzyme activity assays were performed by measuring the 14CO2 evolution from 7.4 kBq L-[1–14C]ornithine (2.11 GBq mmol−1, Amersham Pharmacia Biotech, Italia) or DL-[U-14C] arginine (11 GBq mmol−1, Amersham Pharmacia Biotech Italia), for ODC and ADC, respectively, in the presence of 2 mM unlabelled substrate during a 2-h incubation at 37°C. CO2 was entrapped in KOH and radioactivity counted.

In all cases protein content was measured according to Bradford’s method (Bradford, 1976), using bovine serum albumin as standard. Data represent the means ± SD of 2–3 experiments (3 replicates each); differences between means were analysed by Student’s t-test (significance level at 0.001, 0.01 or 0.05, as specified).

Polyamine analysis

Samples of 0.2–0.5 g were analysed for free and conjugated polyamines by homogenizing them in three volumes of cold 5% (w/v) trichloroacetic acid (TCA) and centrifuging for 10 min at 20 000 g. Replicate aliquots (0.3 ml) of the supernatant were placed in glass ampoules with the same volume of 12 N HCl; the ampoules were then flame sealed and incubated at 110°C for 18 h to allow the hydrolysis of covalent linkages between polyamines and other molecules. Standard polyamines were also subjected to the same procedure. The hydrolysates were taken to dryness and then resuspended in 0.3 ml TCA. Aliquots (0.1 ml) of the supernatant (free polyamines) and hydrolysed supernatant (conjugated polyamines) were dansylated, dansyl-polyamines extracted in benzene described by Torrigiani et al. (1987) and separated by thin layer chromatography (TLC) plates. Spots were visualised under UV light, scraped off the TLC plate, eluted in acetone, their fluorescence measured using a Jasco FP-770 (Tokyo, Japan) spectrofluorometer (excitation 360 nm, emission 506 nm) and compared with standard polyamines. Data represent the means ± SD of 2 experiments (3 replicates each).

Labelling procedure

For labelling experiments, [1,4(n)-3H]putrescine (0.37 MBq in 10 µl, specific activity 0.85 TBq nmol−1, Amersham Pharmacia Biotech Italia) was injected with a 1-ml-syringe into the main vein of control NN tobacco leaves and of leaves immediately after TMV-inoculation. Equal amounts of labelled putrescine were re-injected 6 and 24 h after the first injection. Radioactivity distribution was monitored during the first 6 h in control plants by homogenising the leaf tissue in 0.1 M Tris-HCl buffer, pH 7.6, and counting separately the various zones of the leaf after 2, 4 and 6 h from injection. For labelled di- and polyamine analysis leaves were collected after 6 and 72 h from control and virus-inoculated plants and stored at −80°C until use. Labelled free and conjugated polyamines were separated by TLC as described above and spots dissolved in acetone; the latter was placed in vials with scintillation cocktail and counted for radioactivity.

RNA extraction and northern blot

Total RNA was extracted from approx. 200–300 mg fresh weight explants using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RNA (15 µg per track) was size-fractionated on a 1.2% agarose formaldehyde gel and transferred in 10 × SSC (20 × SSC: 0.3 M sodium citrate, 3.0 M NaCl, pH 7) onto nylon membranes (Hybond-N, Amersham) overnight according to standard methods (Sambrook et al., 1989). RNA was cross-linked to the membrane by exposure to UV at 312 nm (Vilber Lourmat, Marné La Vallée, France) for 4 min.

For the tobacco probes (kindly supplied by A. J. Michael, Institute of Food Research, Norwich, UK), RNA isolation, PCR conditions, cloning of the PCR products and sequencing are described in Michael et al. (1996). RNA blots were prehybridised at 42°C for 2 h and hybridised at 42°C for 18–20 h with [32P]dCTP-labelled PCR fragments (random priming with a Rediprime DNA labelling Kit, Amersham) of SAMDC, ADC and ODC as described in Scaramagli et al. (1999a); they were amplified from tobacco by using specific oligonucleotide primers homologous to the 5′ and 3′ ends of the ORFs of the respective genes. The following primers were used to obtain approx. 1.0 kb, 2.0 kb and 1.2 kb PCR products of SAMDC, ADC and ODC, respectively: 5′-CTAATGGATTCGGCCTTGCCTGTC-3′ (sense) and 5′-CACAGCCCTCAAGACACTACTCC-3′ (antisense) for SAMDC; 5′-ATGCCGGCCTTAGGTTGTTGTGTAG-3′ (sense) and 5′-ACAACTTCAAGCGGTGCAATAGGACCA-3′ (antisense) for ADC; 5′-GGATGGCCGGCCAGACAGTCATCG-3′ (sense) and 5′-TAGAGGTGGTTCATCAGCTTGG-3′ (antisense) for ODC.

Following hybridization, membranes were washed as previously described by Scaramagli et al. (1999a) and then exposed to X-ray film at −80°C for 24 h (ADC, SAMDC) or 7 d (ODC) with intensifying screen (DuPont, Wilmington, DE, USA). Equal loading of RNA on gels was verified by ethidium bromide staining. Band densities were quantified in each sample using the image analysis Phoretix programme (Phoretix International Ltd, Newcastle upon Tyne, UK) and data are shown as relative intensity, normalised to the loading controls.

Results

Diamine oxidizing activity

Since diamine oxidizing enzymes (DAO) are mainly localised in the cell wall (Federico & Angelini, 1991) the DAO activity assay was checked for yield after sonication. Results show that sonication of the leaf homogenate or pellet did not affect the recovery of enzyme activity in those same fractions or in the derived supernatant and pellet (data not shown); the same activity measured in the homogenate, sonicated or not, was recovered in the supernatant fraction. Since part of the activity remained associated to the pellet fraction further experiments were performed using both supernatant (soluble activity) and pellet fractions (compartmented activity).

The analysis of DAO activity as a function of putrescine concentration and the Lineweaver–Burke plot (data not shown) yielded an apparent Km of 237 and 287 µM, respectively, in the soluble and pellet fractions. DAO activity was then analysed in NN and nn tobacco leaves at 0, 5, 24 and 72 h from treatment in control, mock- and TMV-inoculated samples in order to discriminate its involvement in mechanical stress due to carborundum from that in pathogen-induced stress.

In hypersensitive plants, DAO activity was present in the supernatant (Fig. 1a) and pellet (Fig. 1b) at comparable levels in all the treatments and did not change significantly in control and mock-inoculated leaves during the 72-h period in either fraction. In virus-inoculated leaves, DAO specific activity after 72 h was significantly (P < 0.001 for the soluble fraction and P < 0.01 for the particulate fraction) higher (two–three-fold) than in mock-inoculated and control samples.

Figure 1.

Soluble (a, c) and compartmented (b, d) diamine oxidizing enzyme (DAO) activity in control (C, closed columns), mock-inoculated (M, open columns) and TMV-inoculated (V, hatched columns) leaves from hypersensitive (a, b) and susceptible (c, d) tobacco plants at different times after inoculation. Values are the mean ± SD (n = 9). Asterisks indicate significant differences (**, P < 0.01; ***, P < 0.001) with respect to controls.

In nn plants, DAO activity was slightly lower in the supernatant (Fig. 1c) than in the pellet (Fig. 1d), although its magnitude was comparable with that of NN plants. No significant differences in enzyme activity were detected between control, mock- and TMV-inoculated samples at any time.

SAMDC activity and gene expression

The activity of SAMDC, whose localization is reported to be cytosolic (Cohen, 1998), was determined in NN and nn leaves at different times after inoculation. No significant differences were detected in NN samples at 0, 5, and 24 h between control, mock- or virus-inoculated tissues, while enzyme activity was significantly (P < 0.001) enhanced (about five-fold) after 72 h in TMV compared with mock-inoculated and control leaves (Fig. 2b). In nn plants SAMDC activity levels were comparable to those of NN plants and did not show any significant difference in terms of time course or treatment (Fig. 2c).

Figure 2.

S-adenosylmethionine decarboxylase (SAMDC) transcript levels and activity in control (C, closed columns), mock-inoculated (M, open columns) and TMV-inoculated (V, hatched columns) leaves from tobacco plants at different times from inoculation. (a) Upper panel: northern analysis of SAMDC mRNA levels in hypersensitive plants. Each track was loaded with 15 µg total RNA and after size-fractionation transferred to a nitrocellulose membrane and hybridized with 32P-labelled PCR-derived fragment of tobacco SAMDC. Middle panel: RNA loading control; the gel was stained with ethidium bromide. Lower panel: densitometric analysis of respective band intensities normalised to ethidium bromide-stained loading control. (b) SAMDC activity in hypersensitive plants. (c) SAMDC activity in susceptible plants. Values are the mean ± SD (n = 6). Asterisks indicate significant differences (***, P < 0.001) with respect to controls.

Since no changes were observed in SAMDC activity until 72 h, northern analysis of SAMDC was performed on control, mock- and virus-inoculated samples from NN plants at 24 and 72 h (Fig. 2a). At 24 h a transcript was detected whose signal intensity was comparable and higher in mock- and TMV-inoculated samples relative to controls. Two days later (72 h), SAMDC mRNA levels were similar in controls and mock-inoculated samples. In TMV-inoculated leaves transcript levels were strongly enhanced relative to controls and to 24-h inoculated samples (Fig. 2a).

Putrescine biosynthesis and free polyamine levels

Putrescine biosynthesis was evaluated in NN plants through the analysis of ODC and ADC transcript levels and activities in the three treatments at 24 and 72 h. As for SAMDC, ADC (Fig. 3b) and ODC (Fig. 3c) activities were not significantly different at 24 h in the three treatments while at 72 h in TMV-inoculated leaves, both activities were about double (significance P < 0.01) compared with control or mock-inoculated ones. Northern analysis of the same samples revealed that ADC mRNA levels were higher at 24 than at 72 h (Fig. 3a); at both times the signal was more intense in TMV-inoculated leaves relative to the other treatments. Under the same experimental conditions no ODC transcript was detectable even with a 7-d exposure.

Figure 3.

Arginine decarboxylase (ADC) transcript levels and activity and ODC activity in control (C, closed columns), mock-inoculated (M, open columns) and TMV-inoculated (V, hatched columns) leaves from hypersensitive tobacco plants at 24 and 72 h from inoculation. (a) Upper panel: northern analysis of ADC mRNA. Each track was loaded with 15 µg total RNA and after size-fractionation transferred to a nitrocellulose membrane and hybridized with 32P-labelled PCR-derived fragment of tobacco ADC. Middle panel: RNA loading control; the gel was stained with ethidium bromide. Lower panel: densitometric analysis of respective band intensities normalized to ethidium bromide-stained loading control. (b) ADC activity. (c) ODC activity. Values are the mean ± SD (n = 6). Asterisks indicate significant differences (**, P < 0.01) with respect to controls.

Free putrescine and spermidine, but only traces of spermine, were detected at 72 h in control, mock- and TMV-inoculated leaves both from hypersensitive (Fig. 4a) and susceptible (Fig. 4b) plants. In the former, putrescine accumulated approx. two-fold in TMV-compared to mock-inoculated tissues and approx. three-fold compared with the control ones. Minor increases (c. 30%) in spermidine content relative to mock-inoculated tissues were observed. The putrescine-to-spermidine ratio in NN leaves was close to the unit except for TMV-inoculated samples where putrescine accumulated almost two-fold relative to spermidine (Fig. 4a). TCA-soluble conjugated putrescine and spermidine were detected at 72 h (168.2 and 26.2 nmol g−1 f. wt, respectively) in TMV-inoculated leaves; in the other treatments these conjugates were one order of magnitude less (10.1 and 2.3 nmol g−1 f. wt, respectively, for mock-inoculated; 8.2 and 1.4, respectively, for controls). In nn plants, no changes were detected in either putrescine or spermidine levels and the putrescine-to-spermidine ratio was 1 in all the treatments (Fig. 4b).

Figure 4.

Free putrescine (Pu, closed columns) and spermidine (Sd, hatched columns) levels in control (C), mock-inoculated (M) and TMV-inoculated (V) leaves from hypersensitive (a) and susceptible (b) tobacco plants 72 h after inoculation. Values are the mean ± SD (n = 6).

Labelled putrescine incorporation into spermidine/spermine

In order to further evaluate spermidine and spermine synthesis, radioactive putrescine was injected into the leaf main vein in NN plants. Preliminary experiments conducted on control plants showed that the labelled precursor diffused throughout the lamina and formed a concentration gradient which decreased from the middle towards the apical and basal parts of the leaf (data not shown). Although most of the radioactivity remained confined to the main vein or in its immediate neighbourhood, part of it moved to the rest of the lamina within 2 h. Total label amount decreased with time probably due to transport to the rest of the plant.

After 6 h, label recovery in the various fractions (Fig. 5a) was higher in TMV-inoculated than in control samples although reciprocal ratios were comparable; it was higher in the TCA-supernatant (containing free and conjugated polyamines) than in the pellet (TCA-insoluble fraction). The radioactivity found in the supernatant was entirely recovered in the benzene fraction containing dansylated free polyamines. At 72 h label in virus-inoculated tissues was approx. seven- and fourfold higher in the acid soluble and insoluble fractions, respectively, than in controls (Fig. 5b). Labelled free putrescine recovery was higher in virus-inoculated (fourfold at 6 h and 10-fold at 72 h) than in control samples (Fig. 5a,b); putrescine incorporation into spermidine and spermine (even though endogenous levels of the latter were detectable only in trace amounts) was also higher in TMV-inoculated than in control samples (1.5-fold at 6 h, and ninefold and fourfold at 72 h for spermidine and spermine, respectively). Most of the label (approx. 70%) of the benzene fraction was represented by polyamines. Probably due to dilution exerted by endogenous conjugated putrescine, that exceeded 17-fold the free form, only trace amounts of label were found in conjugated forms.

Figure 5.

Label recovery in different fractions from control (C) and TMV-inoculated (V) leaves from hypersensitive tobacco after 6 (a) and 72 h (b) from [3H]putrescine administration. Sn, TCA-supernatant (open columns); Pt, TCA-pellet (shaded columns); B, benzene phase (hatched columns); Pu, putrescine (horizontal hatched columns); Sd, spermidine (dotted columns); Sm, spermine (closed columns). Values are the means ± SD (n = 6).

Discussion

Present results show that in NN tobacco plants, differently from nn plants, the response to TMV implies a substantial increase in polyamine turnover: SAMDC activity and gene expression sensibly increase and a two-fold rise in DAO activity occurs, in addition to stimulation of ADC and ODC. These events were detectable in TMV-inoculated leaves starting from 24 h while labelled putrescine incorporation into spermidine and spermine occurred earlier (6 h). These findings fit with previous results (Rabiti et al., 1998), indicating enhanced free putrescine and spermidine accumulation (5 h) and conjugation (24 h) following TMV inoculation in NN plants. In excised HR lesions the first changes in polyamine metabolism occurred at 3 d, but the TMV-inoculum was 20-fold lower (Torrigiani et al., 1997). On the contrary, in nn plants no changes were detected at any time confirming that they may be ‘genetically’ incapable of reacting to virus inoculation in terms of polyamine metabolism.

With the experimental protocol adopted here it is also shown that the only effect of mechanical stress imposed by carborundum was a transient SAMDC transcript accumulation in mock-inoculated NN plants relative to controls. This may indicate that hypersensitive plants promptly recover from a modest abiotic stress.

DAO activity, which is as crucial as the biosynthetic activities in controlling cellular (Scoccianti et al., 1991) as well as extracellular putrescine levels (Federico & Angelini, 1991), shows here an affinity for putrescine comparable with that previously reported in tobacco (Walton & McLauchlan, 1990). The ineffectiveness of sonication on the recovery of enzyme activity suggests that DAO enzyme is not completely released from the cell wall where it is probably tightly bound (Federico & Angelini, 1991). The increase in soluble and compartmented DAO activity after 72 h in TMV-inoculated NN plants may be explained on the basis of the processes, occurring at the cell wall level, such as polysaccharide-protein cross-linkings and lignification (Pellegrini et al., 1994), which require hydrogen peroxide and are involved in the establishment of the HR. DAO activity, in fact, has been reported to increase in the cell wall in response to wound stress (Scalet et al., 1991), and to fungal infection in chick-pea (Angelini et al., 1993). Such increases are spatially and temporally correlated to that of peroxidase activities which use as substrate the hydrogen peroxide resulting from diamine oxidation (Møller & McPherson, 1998). By producing hydrogen peroxide, polyamine oxidation also induces the formation of reactive oxygen species (ROS; Allan & Fluhr, 1997); hydrogen peroxide and ROS are key signalling factors in programmed cell death (PCD; Mittler et al., 1997), have a direct antimicrobial effect (Peng & Kuc, 1992) and contribute to protein-polysaccharide cross-linkings (Yang et al., 1997). In fact, DAO activity, with the consequent reduction in polyamine titres, has been reported to initiate PCD-driven developmental patterns in Arabidopsis (Møller & McPherson, 1998). In animal cells ODC over-expression (Poulin et al., 1995) but also polyamine depletion by analogues (Hu & Pegg, 1997) induce apoptosis, a form of PCD. Probably, the two findings are not in contradiction and simply confirm the contention that extremely regulated endogenous levels of polyamines are needed in all physiological processes; beyond these levels, cells prefer to commit ‘suicide’ rather than repair the damage. In the tobacco–TMV interaction, despite enhanced DAO activity, free putrescine prevails throughout (Torrigiani et al., 1997; Rabiti et al., 1998, present data), and thus increasing levels of free and conjugated putrescine and spermidine are associated with PCD.

In virus-inoculated hypersensitive plants the about five-fold increase in SAMDC activity at 72 h, which is preceded by transcript accumulation at 24 h, suggests that regulation is primarily at the transcriptional level. SAMDC activity, in fact, appears to be regulated both developmentally (Taylor et al., 1992; Lee et al., 1997) and post-translationally by cleavage of the proenzyme (Xiong et al., 1997). SAMDC transcript accumulation has been reported to be modulated both in dividing and differentiating tissues (Taylor et al., 1992; Lee et al., 1997; Scaramagli et al., 1999b) and in response to exogenous stimuli such as growth regulators or polyamine biosynthesis inhibitors (Scaramagli et al., 1999a). Moreover, previous studies report a twofold increase in SAMDC activity in Chinese cabbage infected with turnip yellow mosaic virus (Cohen et al., 1981), enhanced spermidine synthase activity and incorporation of the newly synthesised spermidine into the virus particles (Cohen, 1998). The rise in SAMDC gene expression and activity leads to decarboxylated S-adenosylmethionine (decaSAM) accumulation and this might stimulate spermine/spermidine synthase activities, since they depend upon substrate availability (Heby & Persson, 1990). At 72 h, however, under the present experimental conditions, free spermidine titres increased only slightly in virus-inoculated samples. By contrast, conjugated spermidine levels markedly increased (c. 10-fold), thus, depleting the free form; spermidine conjugation probably accounts for the modest free amine accumulation and the enhanced SAMDC activity. In fact, polyamine depletion by means of biosynthetic inhibitors has been reported to enhance the respective biosynthetic enzyme activity both in plant and animal cells (Hiatt et al., 1986; Heby & Persson, 1990). Another possible explanation for the lack of free spermidine accumulation is that it was oxidized. In legumes, in fact, the amine is also a substrate for DAO, although with a lower affinity than putrescine (Federico & Angelini, 1991). Spermidine oxidation has also been reported in cultured eggplant tissues (Scoccianti et al., 2000).

The observed rise in ADC and ODC activities can account for putrescine accumulation at 72 h in virus-inoculated samples. The latter results from the balance between synthesis, oxidation, conjugation and conversion to higher polyamines, with conjugation probably being the most active process. Both enzyme activities are regulated post-translationally (Cohen, 1998), and both respond to TMV but not to mechanical stress. In the present model system no cell division or elongation occur, but cell death and wall lignification do, processes which nevertheless require active cell metabolism. ADC could be implicated in the stress response (Flores, 1990) while ODC in hydroxycinnamoyl amide (HCA) synthesis (Burtin et al., 1989).

However, there is no positive relationship between the previously reported increase in free putrescine and spermidine levels at 5–24 h, and biosynthetic activities at these times (Rabiti et al., 1998 and present data). Since labelling experiments confirm early spermidine and spermine synthesis, we should consider the following: the enzyme activity assay is performed in vitro in a cell-free system while polyamine levels are measured in vivo; SAMDC does not directly lead to higher amine synthesis; endogenous decaSAM might be sufficient to support synthase activities for the first 24 h; and, most probably, transport of polyamines, which has been well documented in plants (Antognoni et al., 1998), could occur towards the injured leaf or the incipient HR lesion. In fact, labelling experiments also show that TMV-inoculated NN leaves retain more free putrescine (10-fold at 3 d) than controls indicating a kind of ‘sink’ effect for putrescine, in agreement with the fact that PCD, as well as HR, requires active metabolism (He et al., 1993).

In conclusion, in NN plants but not in nn plants, following TMV-inoculation, SAMDC gene expression and activity as well as DAO activity are upregulated and polyamine biosynthesis is detectable well before changes in biosynthetic enzyme activities. This leads to changes in cellular components, such as polyamines and hydrogen peroxide, which induce PCD. Further work is needed in order to establish whether polyamine transport occurs towards the HR lesion and which is the relative contribution of free and conjugated polyamines in the establishment of the HR.

Acknowledgements

We gratefully acknowledge Anthony J. Michael, Institute of Food Research, Norwich (UK) for kindly supplying the tobacco SAMDC, ADC and ODC probes. Funds are from the University of Bologna, Special Project Molecular Signalling in Differentiation (P. T.) and ex-60% MURST (L. B. and P. T.).

Ancillary