Harpins are proteins secreted by the type-three secretion system of phytopathogenic bacteria. They are known to induce a hypersensitive response (HR) in non-host plant leaf tissue. Erwinia amylovora, the fire blight pathogen of pear and apple trees, secretes two different harpins, HrpNea and HrpWea. In the present study, we showed that an Erwinia amylovora hrpWea mutant induces stronger electrolyte leakages in Arabidopsis thaliana foliar disks than the wild-type strain, thus suggesting that HrpWea could function as a HR negative modulator. We confirmed this result by using purified HrpWea and HrpNea. HrpWea has dual effects depending on its concentration. At 200 nM, HrpWea, like HrpNea, provoked the classical defense response – active oxygen species (AOS) production and cell death. However, at 0.2 nM, HrpWea inhibited cell death and AOS production provoked by HrpNea. HrpWea probably inhibits HrpNea-induced cell death by preventing anion channel inhibition, confirming that anion channel regulation is a determinant feature of the plant response to harpins. Collectively our data show that the HrpWea harpin can act antagonistically to the classical HrpNea harpin by suppressing plant defense mechanisms.
Gram-negative pathogenic bacteria use protein secretion machinery, called the type-three secretion system (TTSS), to increase their virulence in host tissue. TTSSs were reported initially from the mammalian pathogen Yersinia enterocolitica, but have also been identified in various plant pathogenic bacteria such as Erwinia, Pseudomonas, Xanthomonas and Ralstonia (Hueck, 1998). TTSSs allow delivery of specific effector proteins to host cell (Hueck, 1998). Bacterial TTSS effectors are highly diverse and collectively manipulate host cell metabolism to improve bacterial nutrition or to encounter plant cell defense mechanisms (Hauck et al., 2003; Abramovitch et al., 2003; DebRoy et al., 2004; Kim et al., 2005).
Harpins are TTSS-delivered proteins from plant pathogens and constitute a group of proteins that are secreted in the intercellular space of plant tissue during interaction, rather than being injected (Perino et al., 1999; Tampakaki and Panopoulos, 2000). Collectively, harpins could be involved in the release of nutrients from the host cells or they could facilitate the delivery of other bacterial proteins in the host cell (Holeva et al., 2004; Lindeberg et al., 2006). Harpins are not homologous in primary sequence but have common physiochemical characteristics: they are heat-stable, glycine rich, do not possess cysteine, and above all, trigger a hypersensitive response (HR) cell death, requiring an active plant metabolism, in non-host plants (Wei et al., 1992; He et al., 1993; He et al., 1994; Arlat et al., 1994). The mechanisms by which harpins provoke HR cell death have been extensively studied. Harpins trigger closely related signal transduction mechanisms such as mitogen-activated protein kinase (MAPK) activation (Adam et al., 1997; Desikan et al., 1999), cytosolic [Ca2+] elevation (He et al., 1994; Pike et al., 1998; Blume et al., 2000; Cessna et al., 2001), active oxygen species (AOS) production (Baker et al., 1993; Desikan et al., 1996; Reboutier et al., 2007) and ion flux modulations (Desikan et al., 1999; Popham et al., 1995; El-Maarouf et al., 2001; Reboutier et al., 2005; Reboutier et al., 2007). Particular harpins possess a C-terminal domain that is homologous to class III pectate lyases. These HrpW harpins are widespread among phytopathogenic bacteria (Gaudriault et al., 1998; Kim and Beer, 1998; Charkowski et al., 1998). HrpWpst from Pseudomonas syringae pv. tomato was shown to bind to pectate (Charkowski et al., 1998), but lacks detectable pectate lyase activity (Gaudriault et al., 1998; Charkowski et al., 1998). Purified HrpW harpins do not macerate plant tissue and elicit a HR on non-host plants (Gaudriault et al., 1998; Kim and Beer, 1998; Charkowski et al., 1998). Two harpins – HrpNea and HrpWea – have been characterized in Erwinia amylovora, the fire blight pathogen of pear and apple trees (Wei et al., 1992; Gaudriault et al., 1998; Kim and Beer, 1998). Mutant analysis indicates that HrpNea is an important virulence factor involved in the generation of oxidative stress in planta (Wei et al., 1992; Barny, 1995), whereas HrpWea is not required for full virulence (Gaudriault et al., 1998). Furthermore, on non-host plants, such as tobacco, although hrpNea mutants elicited a weaker HR than the wild-type strain, hrpWea mutants elicited a stronger HR than the wild-type strain. HrpWea could thus function as a HR-negative modulator (Gaudriault et al., 1998). The aim of our study was to test this hypothesis. We first confirmed the phenotype of E. amylovora hrpWea and hrpNea on the model plant A. thaliana. Then, we tested the effects of purified HrpWea and HrpNea on A. thaliana cells in suspension to elucidate the putative role or control of early cellular mechanisms involved in cell death regulation.
hrpWea induces strong electrolyte leakages in Arabidopsis thaliana foliar disks
To check whether, as already observed in tobacco (Gaudriault et al., 1998), HrpWea behaves as a negative HR modulator in A. thaliana, we quantified cell death induced by E. amylovora wild-type strain, hrpNea, hrpWea or double hrpNea-hrpWea mutants, by measuring electrolyte leakage of A. thaliana foliar disks. Although the hrpNea mutant was severely impaired in its ability to induce electrolyte leakage, the hrpWea mutant induced stronger electrolyte leakage than observed in the wild-type strain (Fig. 1). The double hrpNea-hrpWea mutant triggered electrolyte leakages similar to levels in the hrpNea mutant. Therefore, in A. thaliana, as in tobacco, HrpWea behaves in planta as a negative modulator of the cell death induced by E. amylovora. Furthermore, its effects are opposed to those induced by HrpNea, which appears to be necessary for E. amylovora to induce cell death in A. thaliana.
HrpWea, like HrpNea, induces cell death and a transient oxidative burst but no cytosolic [Ca2+] increase
The preceding result prompted us to check whether purified HrpWea was able to induce cell death in A. thaliana suspension cells, a convenient system that allows the identification of early physiological events induced by pathogen-derived elicitors on single cells and also allows us to monitor the behavior of large populations of cells. HrpWea, like HrpNea, provoked dose-dependent cell death in A. thaliana cells (Fig. 2A,B). The cell death level was similar for the two harpins, and reached its maximum at a concentration near 200 nM. To have a better understanding of the HrpWea-induced cell death process, we compared the effects of purified HrpWea and HrpNea on physiological responses classically associated with cell death.
At 200 nM, both HrpWea and HrpNea induced large transient oxidative bursts sharing the same characteristics (Fig. 3A). H2O2 production reached similar levels for the two harpins (Fig. 3B). Oxidative bursts reached their maximum between 30 minutes and 1 hour, then, H2O2 levels decreased and returned to control levels after 2 hours (Fig. 3A). In both cases, H2O2 production was blocked by the NADPH oxidase inhibitor DPI (10 μM) (Fig. 3B), suggesting that a plasma membrane NADPH oxidase was involved in H2O2 production.
To measure [Ca2+]cyt variations, we used transgenic A. thaliana cells expressing aequorin protein (Brault et al., 2004). Hypo-osmotic stress, used as a positive control, triggered an instantaneous and transient [Ca2+]cyt peak reaching 2.5 μM, followed by a second slower transient [Ca2+]cyt peak appearing within 1 minute and reaching 2 μM (Fig. 3C). However, neither HrpWea nor HrpNea triggered any [Ca2+]cyt increase (Fig. 3C) over 60 minutes (data not shown). The absence of a Ca2+ response after addition of HrpWea or HrpNea indicated that [Ca2+]cyt elevation is probably not involved in signal transduction mechanisms activated in A. thaliana in response to HrpNea or HrpWea.
HrpWea provokes ion channel modulations opposed to those induced by HrpNea
Recently, we showed that HrpNea probably triggers cell death by inhibiting anion channel activity (Reboutier et al., 2005). Therefore, we checked the effects of HrpWea on ion channel modulation. Within the first minute of treatment, HrpWea (200 nM) decreased time- and voltage-dependent outward rectifying currents previously characterized as K+ outward rectifying currents (KORCs) (Reboutier et al., 2002) (Fig. 4C-E) and increased deactivating currents previously characterized as anion currents (Reboutier et al., 2002) (Fig. 4F-H). The HrpWea-induced increase in anion current was reduced after addition of 10 μM glibenclamide or 40 μM 9AC, two potent anion channel inhibitors (Fig. 4H), confirming the anionic nature of these currents. K+ and anion channel modulations led to depolarization of the plasma membrane [+32±5 mV (n=10)], from a mean membrane potential of –42±7 mV (n=72) (Fig. 4A,B). These effects were opposed to those triggered by 200 nM HrpNea, which increased KORC (Fig. 4E) and inhibited anion current (Fig. 4H), thus leading to hyperpolarization of the membrane [–18±6 mV (n=9)] (Fig. 4A,B). KORC increase induced by HrpNea was reduced after addition of the K+ channel inhibitor TEA (10 mM) (Fig. 4E) confirming the K+ nature of these currents.
Low concentrations of HrpWea counteracts cell death, AOS production and anion channel reduction
Since HrpWea had opposite effects to those triggered by HrpNea on anion channel modulation and because HrpNea induced cell death through anion channel activity decrease (Reboutier et al., 2005), we checked whether increasing concentrations of HrpWea could modify HrpNea-induced cell death in A. thaliana cells.
HrpWea inhibited HrpNea (200 nM)-induced cell death in a dose-dependent manner in the range 0.002 to 2 nM (Fig. 5A). The maximum inhibitory concentration was 0.2 nM. Concentrations greater than 0.2 nM did not inhibit HrpNea-induced cell death. On the contrary, addition of 200 nM HrpWea to 200 nM HrpNea increases cell death indicating that at high concentration HrpNea and HrpWea effects on cell death are additional. Reciprocally, we tested whether increasing concentrations of HrpNea could inhibit 200 nM HrpWea-induced cell death (Fig. 5B); however, HrpNea did not inhibit HrpWea-induced cell death. These results indicate that HrpWea effects are dual: HrpWea induces cell death at concentrations greater than 20 nM and inhibits HrpNea-induced cell death at 0.2 nM.
As the strongest inhibition of HrpNea-induced cell death was observed at 0.2 nM HrpWea, we checked the effect of addition of 0.2 nM HrpWea to 200 nM HrpNea on AOS production and ion channel modulation. HrpWea inhibited HrpNea-induced AOS increase (Fig. 6A). Furthermore, HrpWea totally prevented anion current decrease and partially inhibited the KORC increase induced by HrpNea (Fig. 6B,C). A control experiment with 0.2 nM HrpWea alone showed that it did not modify AOS production or ion channel activities (Fig. 6A-C).
Low concentrations of HrpWea did not counteract HrpZpph-induced cell death
Since HrpWea could counteract the effect of HrpNea, we checked the ability of HrpWea to inhibit cell death triggered by HrpZpph, a harpin similar to HrpN produced by Pseudomonas syringae pv. Phaseolicola (Li et al., 2005). HrpZpph induced cell death in A. thaliana suspension cells (Fig. 7), yet HrpWea had no significant effect on HrpZpph-induced cell death (Fig. 7).
Here, we studied the effects of HrpWea on non-host A. thaliana. The effect of HrpWea on cell death appeared dual because 200 nM HrpWea triggered cell death, as already reported for tobacco (Kim and Beer, 1998), whereas 0.2 nM HrpWea inhibited cell death provoked by 200 nM HrpNea. This result confirmed the data we obtained in planta, since an hrpWea mutant induced stronger electrolyte leakages compared with those observed in the wild-type strain, suggesting that HrpWea, when delivered to A. thaliana leaves by the bacteria, rather acts as a cell death negative modulator. Such a dual effect was not observed with the classical harpin HrpNea, for which cell death induced by the purified protein correlates with a hrpNea mutant altered in its cell-death-inducing capacity.
To gain a better understanding of the dual effects of HrpWea, we compared the effects of purified HrpWea with the well-characterized effects of HrpNea on physiological responses classically associated with cell death. At 200 nM, purified HrpWea or HrpNea had strictly the same effects on AOS production and [Ca2+]cyt variation. Both harpins triggered a strong transient H2O2 production probably through plasma membrane NADPH-oxidase activation. Neither 200 nM HrpWea nor 200 nM HrpNea provoked [Ca2+]cyt variation in A. thaliana cells. Moreover, Ca2+-channel inhibitors or Ca2+ surrogates were inefficient at reducing harpin-induced cell death and harpin-induced effects on ion channels (not shown). These data indicated that [Ca2+]cyt elevation is probably not involved in signal transduction mechanisms activated in response to these harpins in A. thaliana. These results differ from previous data suggesting that Ca2+ was involved in signal transduction mechanisms triggered by harpins (He et al., 1993; He et al., 1994; Blume et al., 2000; Cessna et al., 2001), but are in accordance with data from Chandra et al. (Chandra et al., 1997) showing that HrpNea could activate AOS production without triggering any [Ca2+]cyt variation. It is possible that the use of diverse plant species, such as Nicotiana tabaccum (He et al., 1993; He et al., 1994; Cessna et al., 2001), Nicotiana plumbaginifolia (Chandra et al., 1997) or A. thaliana (the present study) would be responsible for the discrepancy observed. Interestingly, purified HrpWea provoked opposed KORC and anion channel modulations when compared with those triggered by HrpNea: 200 nM HrpWea decreased KORC activity and strongly increased anion channel activity. Because HrpNea was previously shown to trigger cell death by decreasing anion channel activity (Reboutier et al., 2005), it suggested that HrpWea and HrpNea provoke cell death by two distinct mechanisms. Moreover, it allowed us to hypothesize that HrpWea could counteract HrpNea-induced cell death, by having an opposed effect on anion channel modulation. This hypothesis was confirmed when 0.2 nM HrpWea was found to inhibit HrpNea-induced cell death. Although the PopArs harpin was recently found to form oligomers through the conserved GxxxG amino acid motif (Racape et al., 2005), this inhibition is unlikely to be the result of the direct interaction between HrpWea and HrpNea because GST pull-down technology did not show any interaction between HrpWea and HrpNea (data not shown). In addition, the fact that the most effective concentration of HrpWea to inhibit HrpNea-induced cell death was one-thousandth of that of HrpNea, also constitutes evidence against a direct protein-protein interaction. Inhibition of HrpNea-induced cell death by HrpWea more probably acts through integration of the two signal transduction pathways triggered by HrpWea or HrpNea.
Because anion channel inhibition was shown to be involved in HrpNea-induced cell death (Reboutier et al., 2005; Reboutier et al., 2007), we checked the effect of addition of 0.2 nM HrpWea to 200 nM HrpNea on ion channel modulation. At 0.2 nM, HrpWea was able to reduce KORC activation and totally prevented anion channel inhibition provoked by HrpNea. The fact that HrpWea could prevent anion channel inhibition provoked by HrpNea probably explains the decrease in cell death. Taken together, all these results confirmed that anion channels are a determinant feature of the plant response to harpins and participate in the decisions leading to cell death.
At 0.2 nM, HrpWea was also able to strongly decrease AOS production provoked by 200 nM HrpNea. Anion channels were shown to be involved in cell death and AOS production in response to cryptogein (Wendehenne et al., 2002). Consequently, we checked the effect of anion channel modulators, which regulate cell death in A. thaliana (Reboutier et al., 2005), on H2O2 production in response to HrpWea and HrpNea. Anion channel modulators did not modify HrpWea- or HrpNea-induced AOS production (see supplementary material Fig. S1A,B). These results suggested that the H2O2 production induced by HrpNea or HrpWea does not depend on anion current modulation. Moreover, the anion channel activator bromotetramisole, which was shown to inhibit HrpNea-induced cell death (Reboutier et al., 2005), did not decrease AOS production (see supplementary material Fig. S1B). This result indicated that AOS production in response to HrpNea was not sufficient to trigger cell death. Ion channel modulations and AOS production in response to harpins are more likely to be concomitant than linked. Many studies suggest that AOSs are involved in cell death observed in response to pathogens (Levine et al., 1994) or elicitors such as the HrpZpss harpin (Desikan et al., 1996; Desikan et al., 1998). However, data are controversial and other studies propose that AOSs are not involved in harpin-induced cell death (Ichinose et al., 2001; Xie and Chen, 2000). More generally, AOS production appears to be a generic defense response triggered by pathogens or elicitors (Desikan et al., 1998; Mehdy, 1994; Bradley et al., 1992; Samuel et al., 2005; Jabs et al., 1997), which could participate in limiting the spread of pathogens.
HrpWea inhibited cell death triggered by 200 nM HrpNea at a very low concentration (0.2 nM). Yet, HrpWea and HrpNea are roughly secreted in the same amounts by the bacteria (Gaudriault et al., 1998). It is possible that free concentrations of each protein differ in the plant cell apoplasm. Indeed HrpW proteins have been reported to bind to pectate through to their C-terminal domain homologous to pectate lyase (Gaudriault et al., 1998; Kim and Beer, 1998; Charkowski et al., 1998), whereas classical harpin proteins, such as HrpNea, did not. Furthermore, we cannot exclude the fact that each protein could be present in the apoplasm at different amounts because of differential proteolytic degradation in planta.
We report here for the first time that HrpWea, a protein from the harpin family, inhibits cell death. The question remains why E. amylovora, considered as a necrotroph, possesses the cell death inhibitor HrpWea? Bretz et al. (Bretz et al., 2003) suggested that HopPtoD2, an injected TTSS effector of P. syringae, acts as an inhibitor of defense responses by suppressing cell death and AOS production. When delivered by E. amylovora, HrpWea could in the same manner be considered not only as a cell death inhibitor but, more widely, as a defense response inhibitor. HrpWea seemed not be required for pathogenicity, bacause a mutant strain was as pathogenic as the wild-type strain on apple trees (Gaudriault et al., 1998); however, HrpWea could be required for survival in non-host plants during the epiphytic period. The fact that HrpWea was not able to inhibit HrpZpph-induced cell death in Arabidopsis suggests that HrpWea cell death inhibition is specific for HrpNea-hypersensitive reaction, thus confirming the high specificity acquired by secreted proteins from plant pathogens during evolution. Several injected TTSS effectors were shown to inhibit cell death (Abramovitch et al., 2003; Axtell and Staskawicz, 2003; Espinosa et al., 2003; Mackey et al., 2003) when injected into plant cells. To our knowledge, HrpWea is the first TTSS-delivered protein that can inhibit plant cell defense mechanisms when secreted in the plant cell apoplasm. These results suggest that inhibitors of defense mechanisms could act when secreted in the extracellular medium.
In conclusion, this work highlights the role of anion channels in programmed cell death. Using CFTR modulators, we recently proposed that anion current decrease was a prerequisite to HrpNea-induced cell death (Reboutier et al., 2005), as previously described with glibenclamide in animal cells (Kim et al., 1999). However, the signaling pathway(s) leading to this cell death remain largely unknown. HrpWea-induced cell death could be compared with the mechanisms described during apoptotis volume decrease (AVD) in animal cells (Okada and Maeno, 2001) or cryptogein-induced cell death in tobacco cells (Wendehenne et al., 2002; Gauthier et al., 2007), which involve strong anion efflux through the activation of anion channels. Our work also proposes a new role for the HrpWea harpin that was mainly described as a cell death inducer. Here, we show by using in planta and cellular approaches that this harpin is more likely to be a plant defense inhibitor. These new data bring new tracks concerning modulation of plant defenses by phytopathogenic bacteria.
Materials and Methods
Electrolyte leakage determination
A. thaliana (ecotype Columbia) 5-week-old leaves were infiltrated with bacterial suspension Ea321 (WT Erwinia amylovora), Ea321-T5 (hrpNea mutant), Ea321-G204 (hrpWea mutant) and Ea321-T5-G204 (double hrpNea-hrpWea mutant) (supplementary material Table S1) of 3×108 cells/ml in assay medium (0.5 mM MES, 0.5 mM CaCl2, pH 6). Electrolyte leakage assays were initiated by removing six infiltrated leaf samples using a boring tool with an inner diameter of 0.5 cm. These samples were washed in deionized water to remove surface-adhered electrolytes. Then they were placed in a test tube and vacuum-infiltrated in 10 ml distilled water. Samples were incubated under light conditions, at 22°C, with continuous rotation shaking. Conductivity was measured at 0, 5, 10, 24, 29 and 39 hours after bacteria infiltration with a SevenMulti conductimeter (Mettler Toledo, France).
HrpNea and HrpWea were PCR amplified with high-fidelity TAQ (Roche) using pMAB 40 (Gaudriault et al., 1997) as a template with the following primers: AACGGATCCATGAGTCTGAATACAAGTGGG and AAACTCGAGTTAAGCCGCGCCCAGCTTGCC for HrpNea; AAGGATCCTAGTCAATTCTTACGCTTTAAC and AAACTCGAGTTATTGGCATCTTCGCTGTG for HrpWea. The PCR-amplified products were digested with BamHI and XhoI, cloned into the pGEX6P vector (Amersham) and sequenced. The plasmid containing HrpNea and HrpWea, respectively named pMAB164 and pMAB165 were used to transform Escherichia coli BL21. E. coli strains were grown to an optical density of 0.5 at 600 nm and the overproduction of GST-HrpWea or GST-HrpNea was induced with 1 mM IPTG. Bacterial cells were harvested, resuspended in 12 ml PBS buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, DTT 1 mM, PMSF 1 mM, pH 7.3) and sonicated for 5 minutes. After centrifugation, the supernatant was injected into an affinity chromatography column (GST-trap FF, 1 ml, Amersham). The GST-tag was removed by using the Prescission Protease enzyme (Amersham). Then HrpWea or HrpNea were eluted and desalted with a Hi-Trap desalting column (Amersham). For this column, the protein buffer was changed to 5 mM KNO3, 1 mM MES (pH 5.8). This desalting buffer was used as a control in all experiments.
Cell death quantification
Cell death was quantified using the fluorescein diacetate (FDA) spectrofluorimetric method (Amano et al., 2003). Four-day-old A. thaliana cells were collected and washed by filtration in a medium containing 175 mM mannitol, 0.5 mM CaCl2, 0.5 mM K2SO4 and 10 mM HEPES (H10 medium) adjusted with KOH to pH 5.8. 1 ml of cell suspension was incubated in the presence of HrpWea and/or HrpNea alone or with the appropriate pharmocological effector. After 24 hours of treatment, 500 μl suspension was diluted in 1.5 ml H10 medium in a quartz cuvette. Cells were gently stirred with a magnetic stirrer. Then, FDA was added at a final concentration of 12 μM and the fluorescence increase was monitored for 120 seconds using a F-2000 spectrofluorimeter (Hitachi, Japan). We verified the linearity between the percentage of dead cells and FDA-detected esterase activity by melting different amounts of control living cells and heated dead cells. A 100% esterase activity corresponds roughly to 100% living cells. 0% esterase activity corresponds to 100% dead cells. Cell death was thus calculated as follows: % of cell death=(slope of treated cells/slope of non-treated cells) × 100.
Quantification of H2O2 in culture medium
H2O2 release in the culture medium was quantified as described (Bouizgarne et al., 2006). Briefly, 1.5 ml of the cell suspension (stabilized for 4 hours in H10 medium) was inoculated with HrpWea and/or HrpNea alone or with the appropriate chemical effector. Before each measurement, 200 μl cell culture was added to 600 μl phosphate buffer (50 mM, pH 7.9) before addition of 100 μl of 1.1 mM luminol. Then 100 μl of 14 mM K3Fe(CN)6 was added as an electron acceptor. Chemiluminescence was monitored at 30-minute intervals with a FB12-Berthold luminometer (signal integrating time 0.2 second).
Aequorin luminescence measurements
Cytoplasmic Ca2+ variations were recorded with an A. thaliana cell suspension expressing the aequorin gene (Brault et al., 2004). Aequorin was reconstituted by overnight incubation of the cell suspension in Gamborg medium (containing 1 mM Ca2+) supplemented with 30 g l-1 sucrose and 2.5 μM native coelenterazine. Cell-culture aliquots (250 μl) were transferred carefully to a luminometer glass tube and the luminescence was recorded continuously, at 0.2-second intervals, with a FB12-Berthold luminometer (Berthold Technologies, Bad Wildbad, Germany). Treatments were performed by addition of harpin directly into the luminometer tube. At the end of each experiment, the residual aequorin was discharged by the addition of 10% (v/v) ethanol and 1 M CaCl2 (final concentration). The resulting luminescence was used to estimate the total amount of aequorin in each experiment. Calibration of the Ca2+ measurement was performed by using the equation pCa=0.332588(–log k)+5.5593, where k is a rate constant equal to luminescence counts per second divided by the total remaining counts (Knight et al., 1996).
For electrophysiological measurements, cells were impaled in the culture medium as previously described (Reboutier et al., 2005; Bouizgarne et al., 2006). Individual cells were voltage clamped using an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA). Voltage and current were digitized with a personal computer fitted with a Digidata 1320A acquisition board (Axon Instruments, USA). The electrometer was driven by pClamp software (pCLAMP8, Axon Instruments, USA). All experiments were performed at 22±2°C.
This work is dedicated to Prof. Claude Grignon on the occasion of his retirement. We are grateful to A. Pugin and O. Lamotte for help with some experiments and S. V. Beer for kindly providing the E. amylovora mutant strains. We thank A. Pugin, H. Keller, J. M. Frachisse, S. V. Beer and C. S. Oh for fruitful discussions. This work was supported by funds from the MENESR (EA 3514 and ACI 5078) and INRA. D.R. was supported by a grant from MENESR.