INTRODUCTION

The oxidative burst constitutes one of the more rapid responses of a plant cell to pathogen attack (1). Within minutes of pathogen recognition, reactive oxygen species are generated that may promote cross-linking and lignification of the cell wall (2, 3), transcription of defense-related genes (4, 5), secondary metabolite biosynthesis (6), the hypersensitive response (4, 7, 8), and direct pathogen cytotoxicity (9, 10), depending on the plant species examined. In cultured cell systems, the oxidative burst can be promoted by isolated elicitors including harpin (11, 12), oligouronides (13), elicitins (14), purified fungal peptides (15), and other molecules from the extracts of pathogens (7, 16). Abiotic stimuli such as mechanical stress (17), the pesticide fenthion (18), cold shock (19), phosphatase inhibitors (4, 20), and hypo-osmotic stress (17) can also induce biosynthesis of reactive oxygen species in plant cells.

Because of its rapid expression, ease of assay, and similarity to the analogous defense response in human neutrophils (21), the oxidative burst has recently served as a model for exploring signal transduction pathways in plants. As the numbers of elicitors examined have risen, it has become increasingly clear that different elicitors may inaugurate independent signal transduction pathways that employ distinct sets of second messengers (1, 22). Indeed, the pathway initiated by oligouronides appears to require participation of phospholipase C but not phospholipase A (12, 23), whereas the pathway triggered by an elicitor from Verticillium dahliae requires exactly the converse (12). Although all of the independent signaling pathways are believed to converge with the assembly of an oxidase complex on the plasma membrane (24-26), it remains uncertain as to whether the various pathways also share a need for specific kinases or Ca²⁺.

Ca²⁺ is already believed to serve as a second messenger in such important plant processes as stomatal closure (27, 28), geotropism (29), control of circadian rhythms (30), pollen tube elongation (31, 32), regulation of plasmadesmata aperture (33), response to physical stress (34), and phytochrome phototransduction (35). Nevertheless, its accurate quantitation during cellular signaling events has until recently proven very difficult. Thus, ⁴⁵Ca²⁺ binds avidly to cell walls, often masking the small changes in cytoplasmic ⁴⁵Ca²⁺ that could indicate its function as a second messenger (36). Further, signal transduction pathways that might trigger release of Ca²⁺ from intracellular organelles are largely invisible to ⁴⁵Ca²⁺-based methodologies, because no net change in cellular ⁴⁵Ca²⁺ content occurs. Although Ca²⁺-sensitive fluorescent dyes have yielded useful information in a few plant systems (37), their general resistance to entry into plant cells has greatly limited their application (37). Indeed, most membrane-permeant esters of Ca²⁺-dependent fluorophores hydrolyze to their impermeant forms in the cell wall, preventing their penetration to their intracellular destinations. Fortunately in 1991, Knight et al. (38) transformed tobacco plants with a Ca²⁺-sensitive luminescent protein, termed aequorin, and demonstrated that resting levels of Ca²⁺ were insufficient to induce photon emission but that stimulus-induced Ca²⁺ increases generated readily measurable levels of bioluminescence. Since this initial discovery, aequorin-transformed plants have been exploited to quantitate intracellular Ca²⁺ fluxes accompanying such diverse processes/stimuli as direct mechanical perturbation (34, 38), gusts of wind (39), cold shock (38, 40), and circadian oscillations (30). Because the aequorin protein is expressed in the plant cell cytoplasm and because the intensity of light emitted by aequorin is directly proportional to Ca²⁺ concentration, measurement of aequorin-derived bioluminescence enables direct quantitation of Ca²⁺ transients in the transformed cytoplasm of the cell. In this report, we have employed suspension cultures of aequorin-transformed tobacco plants to monitor Ca²⁺ transients during elicitation of the oxidative burst. We report here that several but not all elicitors of the oxidative burst induce a rapid and quantifiable influx of Ca²⁺ and that this Ca²⁺ is required for expression of the subsequent reactive oxygen species.

EXPERIMENTAL PROCEDURES

BAPTA-AM and ionomycin were obtained from Calbiochem (La Jolla, CA), whereas ruthenium red was purchased from Sigma. Coelenterazine was obtained from Molecular Probes (Eugene, OR). All other chemicals were reagent grade or of higher purity and were obtained from major chemical suppliers.

An OGA¹ fraction that elicits the oxidative burst in all plant species tested to date was prepared as described previously (13). The OGA preparation used in these studies contained 0.5 mg/ml galacturonic equivalents as determined by the method of Blumenkrantz and Asboe-Hansen (41). Harpin, a potent inducer of both the oxidative burst (11, 12) and the hypersensitive response (42), was a kind gift of Dr. Steven Beer (Cornell University). Mas-7, an active analog of mastoparan, and Mas-17, its inactive counterpart, were purchased from Calbiochem.

Generation of aequorin-transformed tobacco (Nicotiana plumbaginofolia) plants has been previously described (38, 43). Transgenic seeds from an F3 generation were a generous gift from Dr. Anthony J. Trewavas (University of Edinburgh). The seeds were sprouted on wet filter paper, and the seedlings were potted in soil. Leaves from these plants were surface sterilized using a solution of 8% calcium hypochlorite and 0.1% Tween 20 and transferred to W-38 agar to promote callus formation, as described previously (18). The calli were fragmented, transferred to liquid W-38 medium, and maintained in suspension by continuous shaking. The resulting suspension cultures were expanded by transferring 3 ml of culture to 25 ml of fresh media every 10 days. Except where noted in Fig. 4, all cell suspensions were studied between 16 and 30 h after transfer to fresh media, when the cells responded maximally to elicitor stimulation but displayed no activated characteristics in the absence of elicitation.

[View Larger Version of this Image (12K GIF file)]

In vivo reconstitution of aequorin was carried out by adding 25 µl of coelenterazine (5 µM final concentration) in ethanol to 3 ml of suspension cultured cells (24 h old) and incubating overnight at 22 °C in the dark. Control cells were treated similarly with an equivalent volume of ethanol alone.

Luminescence measurements were made using a digital luminometer (LKB BioOrbit model 1250, Gaithersburg, MD). The luminometer was calibrated by setting the background counts to zero and a 0.26 µCi of ¹⁴C internal standard to 10 mV. Coelenterazine-treated cultures (0.5 ml) were transferred to a luminometer cuvette and maintained in suspension by mild stirring. Luminescence was measured every 0.1 s for the duration of the experiment. All inhibitors and chelators were added at the start of stirring, whereas elicitors were added at the times indicated in the figures. At the end of the experiment, all of the unconsumed aequorin was discharged by injecting 250 µl of a solution of 10% Nonidet P-40 and 100 mM CaCl₂ into the luminometer cuvette to determine the total light output.

The kinetic profile of intracellular Ca²⁺ concentration was obtained by transforming data of each luminescence curve using the equation described by Allen et al. (44), [Ca²⁺] = {(L/L_max) + [118(L/L_max)]  1}/{7 × 10⁶  [7 × 10⁶(L/L_max)]}. To determine the calcium concentration at each time point, a computer program was written to calculate L, the light intensity at time intervals of 0.1 s. L is obtained by integrating the luminometer output (mV) over these 0.1-s intervals. The sum of all L values is defined as L_total (the total area under the curve). The sum of all L values from t = 0 to any given time point is L_cumulative and represents the aequorin that has already been discharged. L_max is defined as L_total  L_cumulative at each time point, and L_max represents the residual aequorin yet to be discharged. L_max is not a constant but decreases steadily as aequorin is consumed during the experiment. Using the parameters L and L_max, we calculated the Ca²⁺ concentration corresponding to each time point in the luminescence curves. The figures present the raw computer transformations of the digitized luminescence data without any smoothing or noise reduction.

H₂O₂ production in aequorin-transformed cells was monitored spectrofluorimeterically, as described earlier (13, 45). Briefly, 1 ml of cultured tobacco cells were transferred to a gently stirred fluorimeter cuvette and treated with 7 µl of pyranine (8-hydroxypyrene-1,3,6-trisulfonic acid trisodium salt; 0.2 mg/ml). Desired elicitors and inhibitors were added at the start of stirring. Loss of fluorescence due to peroxidase-dependent quenching of the dye pyranine (_ex at 405 nm, _em at 512 nm) was continuously monitored. Incubation of cells with coelenterazine had no effect on either the oxidative burst or peroxidase activity.

RESULTS

To study the role of Ca²⁺ in signaling the oxidative burst, we generated suspension cultures from tobacco plants expressing the aequorin transgene, as described under "Experimental Procedures." However, before undertaking to measure elicitor-stimulated Ca²⁺ transients, it was necessary to document that the suspension cultures behaved similarly to the transgenic plants from which they were derived. Three comparisons were undertaken to establish this similarity. First, earlier studies using seedlings from the transgenic tobacco plants established that cold shock is accompanied by a flux of cytoplasmic Ca²⁺ that has a characteristic signature, i.e. a sharp rise that returns to basal levels within a minute (38). As can be seen in Fig. 1A, rapidly lowering the temperature of the suspension cultures (from 23 to 16 °C) also induces a large increase in aequorin luminescence with an overall profile that is remarkably similar to that described for whole plants. Second, because the Ca²⁺ transient in the seedlings could be partially inhibited by soaking the seedlings in millimolar concentrations of either La⁺³ or EGTA (46), we next examined the sensitivity of our suspension cultures to the same reagents. As also shown in Fig. 1A, both La⁺³ and EGTA blocked most of the cold shock-induced increase in cytoplasmic Ca²⁺ concentration. Third, a crude proportionality was previously observed in the transgenic seedlings between the magnitude of the temperature decrease and the amplitude of the Ca²⁺ pulse. Using the transgenic cell cultures, we observed an analogous proportionality, wherein the intensity of the luminescence spike increased with the magnitude of the temperature drop (data not shown). Together these data argue that the transgenic tobacco cell suspension cultures constitute a reasonable model of the transgenic plant.

[View Larger Version of this Image (12K GIF file)]

Because the relationship between luminescence and Ca²⁺ concentration is double logarithmic, presenting data simply as a change in luminescence can greatly exaggerate the extent of any change in intracellular Ca²⁺ concentration. Therefore, we wrote a computer program to convert the observed luminescence signal directly into intracellular Ca²⁺ concentration using the equation described under "Experimental Procedures." Fig. 1B presents the transformation of the data presented in Fig. 1A. This transformation reveals that in response to cold shock, the cytoplasmic Ca²⁺ concentration peaks at around 2.7 µM, a value in the range observed for whole plants. Furthermore, addition of La³⁺ and EGTA reduced the peak Ca²⁺ level to less than 1 µM, once again comparable with the results obtained by Knight et al. (46).

With the ability to characterize the kinetics and magnitude of Ca²⁺ fluxes in cultured tobacco cells established, it was of interest to examine whether an intracellular Ca²⁺ pulse might participate in an elicitor-induced oxidative burst. We focused our investigation on the elicitors OGA, Mas-7, hypo-osmotic shock, and harpin primarily because of our previous experience with these elicitors. Specifically OGA, a component of the plant cell wall thought to be released during pathogen attack or wounding, was selected because it rapidly induces activation of phospholipase C (23) and, presumably, release of intracellular Ca²⁺ through the action of inositol trisphosphate. Mas-7, an activator of heterotrimeric G proteins, was also chosen because it similarly stimulates the oxidative burst via activation of phospholipase C (23). Hypo-osmotic shock, a novel inducer of the burst (17), was selected primarily because it is known to trigger oxidant production via some type of mechanosensor (17), i.e. a pathway commonly inaugurated by Ca²⁺ channels (47). Finally, harpin, a heat stable protein from the bacteria E. amylovora was chosen, because it induces the oxidative burst (11, 12) and may promote the hypersensitive response (42, 48) in tobacco. Importantly, in addition to serving as potent inducers of the oxidative burst, all of the above elicitors are available in relatively pure form, confirming that any Ca²⁺ transients observed will have been induced by the same component in the preparation that is responsible for stimulating the oxidative burst.

The impact of elicitors on intracellular Ca²⁺ levels was evaluated by mixing the elicitors with reconstituted aequorin cells and monitoring the increase in aequorin luminescence. As seen in Fig. 2A, OGA induces an influx of Ca²⁺ into the cytoplasm that begins immediately following OGA addition and ends within 20 s of its initiation. The peak Ca²⁺ concentration achieved in the experiment displayed here was 0.67 µM; however, this value varied from 0.28 to 0.9 µM, depending on the responsiveness of the particular flask of cells employed. Mas-7 also induces a rapid increase in cytoplasmic Ca²⁺, although the magnitude (1.4 µM) and duration of the pulse are much larger than observed with OGA (Fig. 2). Mas-17, an inactive analog of Mas-7, has little effect on aequorin luminescence, confirming that the Mas-7 effect is indeed specific. Osmotic shock, unlike the other elicitors, triggers two contiguous phases of Ca²⁺ release that last for a total of ~4-5 min (Fig. 2). The peak Ca²⁺ concentration is 1.34 ± 0.39 µM for the first pulse and 1.44 ± 0.12 µM for the second (n = 5). Interestingly, harpin, even at concentrations sufficient to induce both the burst and the hypersensitive response does not cause an increase in aequorin luminescence, even when the luminescence measurements are extended to 40 min following stimulation. Because the signal to noise ratio in these studies is greater than 50:1 (actual data are shown without any noise reduction), it can be concluded that Ca²⁺ is not necessary for transduction of the harpin-stimulated oxidative burst.

[View Larger Version of this Image (18K GIF file)]

Using untreated cells from the same flasks, we simultaneously monitored the oxidative burst initiated by the above elicitors using a fluorimeteric assay. Although the oxidative burst patterns for these elicitors have been previously published (12), they are presented here to allow a direct comparison of the kinetics of the Ca²⁺ pulse with the kinetics of oxidant production by the same cells. Fig. 2B shows that the different elicitors, including harpin, induce bursts characterized by different lag periods following elicitor addition and by different rates of H₂O₂ production. Nevertheless, in all cases where a Ca²⁺ transient is observed (Fig. 2A), the Ca²⁺ influx precedes the biosynthesis of H₂O₂ (Fig. 2B). In fact, for each of the three Ca²⁺-mediated pathways, the concentration of cytoplasmic Ca²⁺ is seen to return to its unstimulated level before production of the oxidative species terminates. Importantly, this rapid expulsion of Ca²⁺ from the cytoplasm cannot be an artifact of aequorin depletion, because cell permeabilization at the end of each experiment demonstrated that only a small fraction of the intracellular aequorin had been consumed (see "Experimental Procedures"). Further, the Ca²⁺ influx cannot be dependent on oxidant production, because addition of sufficient catalase to eliminate any detectable H₂O₂ had no effect on the Ca²⁺ influx (data not shown). Taken together, these data suggest that the Ca²⁺ pulses associated with elicitor stimulation exhibit the properties of a true second messenger in that they precede initiation of the biological response but are not required to maintain it, nor do they depend on it.

The observation that harpin can induce a burst without participation of Ca²⁺ as a second messenger raises the question of whether the Ca²⁺ influx stimulated by the other three elicitors is an epiphenomenon or an essential step in signaling their oxidative bursts. To investigate this issue, we decided to block the influx of Ca²⁺ induced by OGA, Mas-7, and osmotic shock, and then monitor the effect on H₂O₂ biosynthesis. As shown in Fig. 3, A-F, incubation of cells with ruthenium red, a Ca²⁺ channel inhibitor known to be active in plant cells (49), eliminates most of the elicitor-evoked Ca²⁺ flux as well as most of the oxidative burst for all three elicitors. Furthermore, treatment of the cell suspensions with BAPTA-AM, a membrane-permeable intracellular Ca²⁺ chelator, eliminated all oxidant production (data not shown). As expected, neither treatment with ruthenium red nor Ca²⁺ chelator affected the harpin-induced burst, i.e. consistent with its inability to mobilize Ca²⁺. These data suggest that Ca²⁺ is indeed necessary for the Mas-7, OGA, and hypo-osmotic stress-stimulated oxidative bursts and that prevention of its influx is sufficient to block oxidant biosynthesis.

[View Larger Version of this Image (25K GIF file)]

To learn whether Ca²⁺ influx is itself sufficient for induction of the oxidative burst, we examined the effect of ionomycin, a Ca²⁺ ionophore known to be active in both plants and animals, on generation of H₂O₂ by the coelenterazine-treated tobacco cells. As seen in Fig. 4, A and B, addition of 20 µM ionomycin to the aequorin-transformed cells induces both a rise in intracellular Ca²⁺ levels (>4 µM) as well as a burst in H₂O₂ biosynthesis. In contrast, when EGTA is present in the culture medium, ionomycin-mediated induction of the oxidative burst is prevented (data not shown). These data suggest that Ca²⁺ influx alone is sufficient to initiate a burst in responsive tobacco cells. However, in the course of these studies, we also observed that as the suspension cultures age (>36 h after transfer to fresh medium) and enter their elicitor-insensitive stage (13, 45), they also lose their ability to respond to ionomycin with oxidant biosynthesis (data not shown). That is, although ionomycin is still capable of inducing the anticipated Ca²⁺ influx in these insensitive cells (as is also hypo-osmotic stress), the cells fail to respond with an oxidative burst. These data suggest that competency factors downstream of the Ca²⁺ influx remain important in regulating the response of a plant to elicitation.

DISCUSSION

The advent of transgenic aequorin plants has significantly improved the technology for monitoring Ca²⁺ fluxes in suspension cultured cells. Thus, in contrast to the more conventional use of ⁴⁵Ca²⁺ and/or Ca²⁺-sensitive fluorophores, the aequorin methodology (i) requires no harsh methods for loading cells or preparing them for observation, (ii) is neither toxic to the plant cells nor to the investigator, (iii) allows for the collection of quantitative data at short (0.1 s) time intervals, thereby permitting accurate evaluation of the kinetics of a Ca²⁺ transient, (iv) permits localization of the Ca²⁺-sensitive protein in desired internal organelles by use of organelle targeting sequences, and (v) generates crude data with a signal to noise ratio generally exceeding 50:1, enabling characterization of processes involving even very small changes in cytoplasmic Ca²⁺. With such desirable attributes, the aequorin technology should serve as the method of choice for investigating the involvement of Ca²⁺ in many cellular signaling events.

Since the use of suspension cultured plant cells to investigate molecular aspects of plant behavior, the question has frequently arisen regarding the relevance of single cell studies to whole plant physiology. Unfortunately, responses to this concern have been largely inadequate, primarily because valid comparisons between whole plants and single cells/cell clusters have been difficult to design. Although much still remains to be resolved on this issue, the similar responses to cold shock of the aequorin transgenic plants and their derived suspension cultures suggests that single cells can in some cases yield data that very accurately reflect the behavior of the whole plants. Thus, Knight et al. (46) observed a Ca²⁺ transient in cold-shocked tobacco seedlings that was characterized by a 10-s width at half-height that rose abruptly to ~2.3 µM Ca²⁺ and fell slowly to its basal (nanomolar) level over a period of ~30 s. We also observed a cold shock-induced Ca²⁺ pulse in the transgenic suspension cultures with an ~8-s width at half-height that increased precipitously to 2.6 µM Ca²⁺ and then declined more slowly to basal levels over a ~30 s total time period. Ca²⁺ fluxes in both the cell suspension cultures and whole seedlings were similarly inhibited by EGTA and La³⁺, and in both systems the magnitudes of the cold shocks and Ca²⁺ transients were roughly proportional. Although physical barriers/inequities prevent similar comparisons of Ca²⁺ transients following mechanical perturbation, elicitor stimulation, or osmotic stress, the observation that these signals also trigger Ca²⁺ fluxes in both single cell and whole plant systems (38, 39, 50, 51) adds further weight to the contention that cultured cells can serve as useful models when properly validated. With this motivation in mind, it should be possible to explore in greater detail the role of Ca²⁺ pulses in signaling a variety of defense-related responses.

Although the studies presented here provide the first quantitative data on the kinetics and magnitudes of the Ca²⁺ transients induced by elicitors of a defense response, they represent by no means the first evidence for Ca²⁺ involvement in defense-related signaling. Indeed, several groups have reported that extracellular EGTA can inhibit the mobilization of selected disease resistance mechanisms (52, 53). Other workers have shown that introduction of exogenous Ca²⁺ into resting cells can initiate a defense response in the absence of the usually required elicitors (16, 54). Still other laboratories have demonstrated that Ca²⁺ channel blockers such as La³⁺ can prevent expression of a normal pathogen resistance mechanism when conditions are otherwise designed to stimulate its induction (52, 54). Nevertheless, it has never been determined wherein Ca²⁺ actually exerts control over processes leading to disease resistance. In the case of the oxidative burst, data presented here demonstrate that Ca²⁺ must enter the tobacco cell cytoplasm to initiate the burst, but the cation must not remain to sustain it. Thus, for the OGA-induced burst, cytoplasmic Ca²⁺ rises to ~0.7 µM and returns to basal levels within ~20 s (Fig. 2). H₂O₂ biosynthesis, in contrast, is initially detected ~2 min following OGA addition, i.e. well after the Ca²⁺ concentration has returned to its resting level. Even with ionomycin-treated cells, where the primary stimulus for oxidant production is the ionophore-catalyzed entry of Ca²⁺ into the cell, the Ca²⁺ stimulus is seen to disappear long before the end product of the pathway (i.e. H₂O₂) is observed (Fig. 4). Clearly, these data indicate that Ca²⁺ serves only a transient role of communicating an upstream signal to a downstream effector. After performing this messenger function, Ca²⁺ appears to be no longer needed. Moreover, it can be argued that the proximal Ca²⁺-dependent effector in the signaling pathway may be similarly transiently activated, returning to its resting state when the activating Ca²⁺ is removed. Signaling components such as the Ca²⁺-dependent protein kinases (55, 56), calmodulin (57) and its responsive kinases (58, 59), Ca²⁺-dependent phosphatases (60), Ca²⁺-gated channels (61), and Ca²⁺-activated phospholipases (62) are obvious candidates for this downstream effector.

As mentioned previously, elevation of elicitor concentration stimulates a roughly proportional increase in the amplitude of the Ca²⁺ transient and the magnitude of its downstream oxidative burst. In contrast, when the same comparison is conducted among unrelated elicitors of the oxidative burst, no such proportionality is observed. Indeed, Mas-7 was found to induce the largest Ca²⁺ influx yet stimulate the weakest oxidative burst. Because Ca²⁺ appears to be essential for transduction of the burst signals of at least three of the elicitors, it must be concluded that factors in addition to Ca²⁺ can significantly modulate the strength of a burst signal and that these unidentified factors are differentially activated by the various elicitors. Such regulatory characteristics require the involvement of bifurcating and reconverging signal transduction pathways.

During preparation of this manuscript, another article appeared describing the use of aequorin-expressing suspension cultured cells to evaluate Ca²⁺ transients during hypo-osmotic shock (51). Although their aequorin transgene was introduced directly into the cultured tobacco cells by infection with Agrobacterium tumefaciens, the response to low osmolarity was nevertheless very similar to that shown in Fig. 2, except the first Ca²⁺ peak was very small and no quantitative data on Ca²⁺ were presented. Because no evaluation of the consequent oxidative burst was conducted (51), further comparisons between the two studies are not possible.

Finally, it was admittedly surprising that harpin was found to trigger no Ca²⁺ influx during its stimulation of the oxidative burst (Fig. 2) and presumed induction of hypersensitive cell death (4). Obviously, its signaling pathway must be very different from those initiated by OGA, Mas-7, and hypo-osmotic stress. Previous work has also shown that harpin signaling is also distinct from that of a V. dahliae elicitor (12). Nevertheless, Ca²⁺-independent transduction of an oxidative burst signal is not without precedence, because concanavalin A, in contrast to most other stimuli, induces human neutrophils to synthesize reactive oxidants in the absence of any change in cytoplasmic Ca²⁺ (63). Clearly much additional research will be required before the diversity of mechanisms that initiate and regulate the plant oxidative burst is fully understood.