Diverse Subcellular Locations of Cryptogein-Induced Reactive Oxygen Species Production in Tobacco Bright Yellow-2 Cells 1[W][OA]

Reactive oxygen species (ROS) play a crucial role in many cellular responses and signaling pathways, including the oxidative burst defense response to pathogens. We have examined very early events in cryptogein-induced ROS production in tobacco ( Nicotiana tabacum ) Bright Yellow-2 suspension cells. Using Amplex Red and Amplex Ultra Red reagents, which report real- time H 2 O 2 accumulation in cell populations, we show that the internal signal for H 2 O 2 develops more rapidly than the external apoplastic signal. Subcellular accumulation of H 2 O 2 was also followed in individual cells using the 2 # ,7 # -dichloroﬂuorescein diacetate ﬂuorescent probe. Major accumulation was detected in endomembrane, cytoplasmic, and nuclear compartments. When cryptogein was added, the signal developed ﬁrst in the nuclear region and, after a short delay, in the cell periphery. Interestingly, isolated nuclei were capable of producing H 2 O 2 in a calcium-dependent manner, implying that nuclei can serve as a potential active source of ROS production. These results show complex spatial compartmentalization for ROS accumulation and an unexpected temporal sequence of events that occurs after cryptogein application, suggesting novel intricacy in ROS-signaling cascades.

Reactive oxygen species (ROS) are versatile molecules that play an indispensable role in mediating a diversity of cellular responses in plant cells, including necrotic reactions, programmed cell death, development, gravitropism, and hormonal signaling. The membrane-bound NADPH oxidase or respiratory burst oxidase homologs (Rboh; Groom et al., 1996;Keller et al., 1998;Torres et al., 2002) are key enzymes that generate superoxide radicals in plant cells, which are converted into H 2 O 2 (Pugin et al., 1997;Sagi and Fluhr, 2006). Rboh are localized to plasma membrane fractions (Keller et al., 1998) and are stimulated directly by Ca 21 (Sagi and Fluhr, 2001). This stimulation is likely mediated by their N-terminal extension containing EF-hand calcium-binding motifs. These enzymes are known to participate in different plant processes and responses. For example, developmental processes, such as cell expansion and root growth (Foreman et al., 2003;Renew et al., 2005), xylem differentiation (Barcelo, 2005), and apical dominance and leaf shape (Sagi et al., 2004), are regulated by Rboh activity. Accumulating evidence suggests a role for Rboh in hormonal signaling pathways in plants, including abscisic acid-mediated stomatal closure (Kwak et al., 2003) and ethylene regulation of cell death (Overmyer et al., 2003), and as an auxin-mediated signal transduction controlling root development (Nagata et al., 2004). Rboh are also involved in different defense responses to wounding (Sagi et al., 2004), abiotic stress (e.g. water stress and ozone; Jiang and Zhang, 2002;Joo et al., 2005), and hypersensitive cell death in response to avirulent pathogen attack (Sagi and Fluhr, 2001;Torres et al., 2002).
A well-studied example of the role of Rboh in plant pathogen interactions is exemplified in the cryptogeininduced defense response in tobacco (Nicotiana tabacum). Cryptogein, a proteinaceous elicitor, produced by the pathogenic fungus Phytophthora cryptogea, belongs to a family of 10-kD proteins called elicitins (Panabieres et al., 1995). In tobacco plants and suspension cells, cryptogein induces a hypersensitive response (Ricci et al., 1989;Binet et al., 2001) and systemic acquired resistance (Keller et al., 1996). A mode of action for cryptogein has been suggested that starts with recognition of the elicitor by an unidentified plasma membrane receptor that leads to a cascade of events, including calcium influx, potassium and chloride effluxes, cytosolic acidification, and plasma membrane depolarization (Pugin et al., 1997;Bourque et al., 1999). These events lead to the activation of Rboh (Simon-Plas et al., 2002), which is responsible for ROS production in the oxidative burst, and to the activation of mitogenactivated protein kinase homologs (Bourque et al., 1999;Nurnberger and Scheel, 2001;Kadota et al., 2004). The production of apoplastic ROS in the oxidative defense-related burst promoted by cryptogein and other pathogens is biphasic: a rapid and transient production of ROS, which occurs within minutes of the perception of pathogens, and a second, much later phase of ROS production within hours, which is specific to hypersensitive response-inducing pathogens (Grant and Loake, 2000).
The spatial, temporal, and quantitative components of ROS signaling dictate its specificity. These components include the precise location and magnitude of ROS production coupled to its cellular scavenging (Mullineaux and Karpinski, 2002) and modulation by nitric oxide (Delledonne et al., 2001;Zaninotto et al., 2006). Rboh enzymatic activity mediates the vectoral transfer of electrons to produce extracellular shortlived superoxides. The relationship of the external signal produced by Rboh to modulation of the intracellular ROS signal and its localization are fundamental questions that remain unanswered. In this study, we examine very early events in cryptogein-induced H 2 O 2 production, its localization and kinetics. We have identified subcellular sites for H 2 O 2 accumulation in the cytoplasm, endoplasmic reticulum (ER), and nuclear region, and have identified a novel site for ROS production in nuclei. Furthermore, our results suggest that the early cryptogein-induced response in Bright Yellow (BY)-2 cells is actually initiated within the cells, suggesting that extracellular ROS production may serve to amplify preceding internal ROS signals.

Rapid H 2 O 2 Response in Cryptogein-Treated BY-2 Cells
The low autofluorescence of BY-2 cells, due to the lack of developed chloroplasts, facilitates their use in real-time kinetic observation. Amplex Red (AR) and Amplex Ultra Red (AUR) reagents were employed due to their high sensitivity and have been used previously to measure H 2 O 2 production in human leukocytes and enzymatic assays (Mohanty et al., 1997;Mazura et al., 2006). When AUR was used to assay the response of BY-2 cells to cryptogein (100 nM), rapid accumulation of H 2 O 2 was detected. Catalase, an enzyme that catalyzes the breakdown of H 2 O 2 , and diphenylene iodonium (DPI), which inhibits flavincontaining enzymes, were incorporated to examine the nature of the signal. As shown in Figure 1A, the AUR response was sensitive to catalase and DPI, indicating that it was H 2 O 2 that was measured and that a flavindependent enzyme is a source for this activity. To localize the origin of the signal, optical sections of individual cells were sampled using confocal laserscanning microscopy (CLSM). The image showed that, in the case of AUR reagent, the signal emanated from the surrounding medium, leaving the cells as a dark shadow (Fig. 1A, left inset). This result indicates that AUR does not cross the cell membrane and reports extracellular H 2 O 2 accumulation, presumably from a DPI-sensitive source like Rboh.
A similar assay was carried out with AR, which resulted in comparable reaction patterns, but with significant kinetic differences and a markedly different image pattern (Fig. 1B). Significantly, when the optical section of individual cells was examined, it showed that the fluorescent signal emanates from inside the Figure 1. Generation of H 2 O 2 in cryptogein-treated BY-2 cells as reported by AR and AUR. A, AUR assay of cryptogein-treated cells. Left, Representative kinetics graph of H 2 O 2 response to cryptogein (100 nM) as measured by fluorescence emission in the absence or presence of DPI (2 mM) and catalase (628 units/mL). Arrow indicates the addition of cryptogein. Left inset, Representative confocal image of cryptogeintreated cells in the presence of AUR reagent. Right, Maximal reaction response rates before and after cryptogein addition (average of five experiments 6 SD). B, AR assay of cryptogein-treated cells as described in A. C, Measurement of the kinetics of H 2 O 2 accumulation in cryptogein response. Left, Accumulation of H 2 O 2 with AR and AUR after the addition (arrow) of cryptogein (n 5 5, average 6 SD). Right, Calculated time response to achieve maximal linear accumulation after the addition of cryptogein (6SD). cell (Fig. 1B, left inset). The physicochemical basis for the differences in cell permeability displayed between AUR and AR is unknown. In the case of AR, the effective inhibition by catalase with respect to that observed in the AUR measurement was reduced (80% in the AUR measurement, 50% in the AR measurement). Similarly, the effectiveness of inhibition by DPI was slightly reduced (100% in the AUR measurement, 90% in the AR measurement). These results show that, under the conditions used, AR is cell permeable and reports internal changes in H 2 O 2 apparently with the support of a cellular peroxidase activity. This scenario is consistent with the observed weaker inhibition of the signal by exogenously applied catalase relative to inhibition of catalase during AUR measurement. The fact that external catalase can affect internal H 2 O 2 levels has been observed previously (Allan and Fluhr, 1997).
Rapid scanning of H 2 O 2 accumulation in BY-2 cells was carried out as shown in Figure 1C. Normalized emission data were analyzed for response kinetics by measuring the response time from the addition of cryptogein (Fig. 1C, arrow) and the commencement of the linear phase of H 2 O 2 accumulation. Analysis of these curves revealed that AR response time was more rapid than that measured for AUR (1.3 6 0.3 min, using AR, compared to 2.5 6 0.2 min, using AUR). The longer response time shown by AUR may be an intrinsic feature of this reporter agent. To examine this possibility, we employed the calmodulin antagonist N-(6-aminohexyl)-5-chloro-1-napthelenesulfonamidhydrochloride (W7), which induces rapid ROS production in BY-2 cells (this work; C. Ashtamker, unpublished data). The response time after W7 addition was immediate, as measured by AUR (Supplemental Fig. S1), and was identical to the rapid response measured by AR (data not shown), indicating that AUR can measure rapid kinetic changes as well. Thus, it appears that the internal signal from cells treated by cryptogein, as reported by AR, develops more rapidly than the external signal, as reported by AUR.

Cellular Accumulation of H 2 O 2
The kinetics of the internal H 2 O 2 signal was further examined by CLSM using dichlorofluorescein (DCF; 2#,7#-DCF diacetate [H 2 DCFDA]). The initial fluorescent signals of the AR and DCF probes showed complete colocalization after the addition of H 2 O 2 (Supplemental Fig. S2). For kinetic CLSM measurements of individual cells, the AR reagent proved to be less useful because it showed laser-induced instability (Zhou et al., 1997;Towne et al., 2004). As in the case of AR, the DCF signal is dependent on its concentration in a particular compartment and on the intrinsic cellular peroxidase activity (Gerber and Dubery, 2003). However, the kinetics of DCF oxidation within a certain compartment should represent actual real-time changes in H 2 O 2 concentrations that result from production in the compartment or from H 2 O 2 influx to the compartment. DCF-treated cells showed a basal-level signal that accumulated in all cytoplasmic areas, including the nucleoplasm ( Fig. 2A, left; basal level). After exogenous excitation of the cells with 100 nM cryptogein ( Fig. 2A, left; 1Cryp), the DCF signal was significantly elevated in most cellular localizations. Cells that were treated with DPI showed significant reduction of the basal level of the DCF signal to a level in which the basal signal cannot be detected (data not shown). The inhibition indicates that the basal level of ROS is at least partially produced by flavin-containing enzymes (e.g. Rboh). A time course of the response to cryptogein in the absence or presence of DPI was conducted and measurements of the signal intensity were taken every 15 s for a total reaction time of 15 min. A representative normalized graph ( Fig. 2A, right) shows an increase in DCF signal in response to cryptogein application (arrow). Interestingly, DPI inhibition was significant as observed in the AR assay ( Fig. 1B). This result is consistent with the identification of Rboh as the source of cryptogein-induced ROS response (Simon-Plas et al., 2002).
To further clarify the results obtained with the AR/ AUR assays that suggested an earlier internal ROS signal, kinetics of H 2 O 2 accumulation as reported by DCF were followed by collecting data separately from periphery and nuclear regions in individual immobilized cells. Measurements were collected every 0.3 s for a total reaction time of 2 to 5 min from both regions (Fig. 2B). As a control, 0.1 mM H 2 O 2 was added to the cells. The exogenously added H 2 O 2 was shown to promote a rapid rise in the DCF signal (Fig. 2, B and C; compare image of basal level to 1H 2 O 2 ). Quantitative analysis of five independent experiments showed an average of 2.6 6 0.3 s delay between the maximal rate change of the signal measured at the periphery, which occurred first, compared to the maximal rate change of the signal measured from the nuclear region, which followed (Fig. 2D). The results are consistent with the external application of H 2 O 2 and indicate the time needed for inward diffusion of H 2 O 2 . These results show the sensitivity of the system to discern cellular flux directions of H 2 O 2 . The same experiments were carried out after the addition of 100 nM cryptogein (Fig. 2, B and C, right insets). In this case, measurements showed that the flux direction was reversed, with the nuclear region showing an earlier change in rate than the periphery (7.4 6 2.2 s; n 5 6; Fig. 2D). These results suggest that internal sources of H 2 O 2 may contribute to the cellular response to cryptogein and are consistent with the finding that internal response times reported by AR occurred more rapidly than external response times reported by AUR (Fig. 1).

Cellular Compartmentalization of DCF Accumulation
The imaging characteristics of the basal and induced DCF signal suggest that it is accessible to many cellular compartments. To identify DCF subcellular localization, double-staining experiments with a variety of cellular probes were conducted. FM 4-64, a hydrophobic fluorescent stain, has been shown to rapidly stain the cell plasma membrane and progress by endocytosis to label internal cellular membranes (Bolte et al., 2004). Double-stained cells with DCF and FM 4-64 showed distinct localization (Fig. 3A). In comparison, double staining with a mitochondria-specific probe, MitoTracker Red CMXRos (Zanella et al., 2002), showed a great deal of fluorescent overlap that colocalized in all cytoplasmic areas in which mitochondria are present, consistent with the respiratory function of these organelles (Fig. 3B). Cells double stained with DCF and the ER-specific probe ER-Tracker Blue-White DPX (Richter-Unruh et al., 2004) showed that the signals were completely interspersed (Fig. 3C). The prominent envelopment of the plant nucleus by the ER is also evident, as has been shown previously (Staehelin, 1997;Okushima et al., 2002). In addition to the localization noted above, the DCF signal emanates from the nucleus as well. Taken together, the results suggest that DCF reports the accumulation of H 2 O 2 in the vicinity of the ER, mitochondria, and nucleus.

Brefeldin A Disruption of Cellular Compartments
To further differentiate the localization of the DCF signal with reference to endomembranes, cells were examined after the application of the fungal metabolite brefeldin A (BFA). Exposing cells to BFA has been shown to modify intracellular protein traffic from the ER to the Golgi apparatus and to induce remodeling of ER-Golgi apparatus morphology (Fujiwara et al., 1988;Klausner et al., 1992;Driouich et al., 1993). The rationale for following DCF distribution in BFA-treated cells is the expectation that, if the DCF signal is associated with the ER, its distribution would be radically altered in BFA-treated cells.
Initially, the influence of BFA on ER structure was followed with DiOC 5 (3), an ER marker (Zhao et al., 2006). Cells were treated with 40 mM BFA for different time periods and then stained with DiOC 5 (3). Treated and untreated cells were visualized by CLSM and optical sections of cells were obtained to produce the three-dimensional (3D) reconstructed images that are presented in Figure 4A (left). Untreated cells showed that the stain is spread throughout the cytoplasm as thin filaments that form a net and concentrated especially around the nucleus (Fig. 4A, top left; control). This staining by DiOC 5 (3) is consistent with both ER and Golgi structures, as shown before (Ritzenthaler  , 2002). After 30 to 90 min of exposure to BFA, the ER was reorganized as patch-like structures observed in the periphery of the cells (Fig. 4A, bottom left; 1BFA). The patch-like structures are thought to represent novel blending of ER and Golgi membranes . A parallel experiment was conducted with DCF-stained cells. The basal level of H 2 O 2 accumulation yielded similar images (Fig. 4A, right insets; compare control and 1BFA). Untreated cells showed the prevalent normal DCF signal distribution; however, after BFA treatment, cells with exceptional morphology, including large smears and patches at the cell periphery, were detected.
To further establish the influence of BFA on the distribution of the DCF signal, a similar experiment was conducted by double staining with DCF and ER-Tracker DPX. Optical sections of cells were collected and the 3D reconstruction images are shown (Fig. 4B, top inset). The images yielded patterns similar to that detected with DiOC 5 (3) (Fig. 4A, left inset) and DPX (Fig. 3C). After 30 min of exposure to BFA (Fig. 4B, middle inset), the area surrounding the nucleus and some patch-like areas were stained by DPX, which is in concordance with the DiOC 5 (3) stain (Fig. 4A). DCF signals were seen to colocalize to the same areas, with the nucleus as an additional DCF-stained region. After 1 h of treatment with BFA (Fig. 4B, bottom inset), a patch-like appearance was observed (Fig. 4B, bottom  inset). These results show that BFA treatment dramatically shifts the pattern of DCF staining into ER-Golgi hybrid stacks and suggests that part of the DCF signal emanates from the endomembrane system.

Nuclei Are a Potential Source of ROS Production
Nuclear regions showed intense DCF fluorescence (Figs. 1-4). This could be a result of ROS that originate from other cellular locations and cross into the nucleoplasm, where they react with DCF. Alternatively, the nuclear region may itself be a source. To examine this, intact nuclei were purified as previously described (Stoppin et al., 1994). When viewed by confocal microscopy, isolated nuclei were largely free of surrounding ER and mitochondria as established by DiOC 5 (3) and MitoTracker staining (data not shown) and appeared to be intact (Fig. 5, A and C, transmission insets). To see whether DCF is present in the nuclei, high excitation regimes were employed. Such treatment was shown previously to excite DCF in BY-2 cells (Dixit and Cyr, 2003). Under these conditions, DCF labeled the nucleolus (Fig. 5A, green color). Triple staining with the membrane-specific dye FM 4-64 and with the DNA-specific dye diamidino-2-phenylindole (DAPI; Fig. 5A, red and blue colors, respectively) revealed an intact nuclear membrane surrounding an area with chromatin and complex unstained substructure within and around the nucleolus.
Cryptogein addition to isolated nuclei showed no effect (data not shown), indicating that signal transduction for ROS induction was disrupted. Consequently,  we sought a different mode of activation of ROS production. Calcium, a downstream component of the cryptogein-signaling pathway that is also known to induce ROS production by Rboh, was employed as a possible activator. As shown in Figure 5B, addition of 1 mM Ca 21 was found to induce a DCF signal in EGTApretreated BY-2 cells. In isolated nuclei, the same concentration of Ca 21 induced a DCF signal that was chiefly localized to the nucleolus (Fig. 5C, top). Similar results were obtained using AUR as a reporter, which in this case was found to be competent to enter isolated nuclei (Fig. 5C, bottom). The rate of the H 2 O 2generated signal as measured by DCF was found to correlate with Ca 21 concentrations but not to the addition of NADPH (Fig. 5D), which is a potential electron donor (e.g. to Rboh). High concentrations of calcium necessary to induce H 2 O 2 accumulation in isolated nuclei may reflect their relative impermeability to calcium, as has been reported previously (Pauly et al., 2000). The signal was partially inhibited by DPI (60% inhibition) and exogenous catalase (30% inhibition), as shown in Figure 5E. The results imply that nuclei can be an independent source of H 2 O 2 .

DISCUSSION
Real-time rapid changes in elicitor-induced ROS accumulation were followed in BY-2 cell populations and in individual cells. The results revealed unexpected complex kinetics in which ROS accumulated within the cell more rapidly than the response outside the cell. The result is not due to limitations in the measurement of external response times as shown by the immediate reporting of H 2 O 2 accumulation after W7 application (Supplemental Fig. S1). The exact mode by which W7 induces ROS accumulation is not known. W7 serves as a calmodulin antagonist and has been used to study calcium/calmodulin-mediated pathways in animals and plants. The ROS response in BY-2 cells induced by W7 is likely a result of rapid calcium bursts (Kaplan et al., 2006). Such bursts are known to precede the accumulation of H 2 O 2 in different signaling pathways (Neill et al., 2002;Jiang and Zhang, 2003). Thus, W7 appears to short circuit cellular ROS activation and shows that different modes of ROS activation will yield different action patterns.
The discrepancy between the external and internal cellular response times after cryptogein addition is inconsistent with the conventional pattern expected for a peripheral extracellular origin of ROS by Rboh. Superoxides are thought to be produced on the outer surface of the cell by a plasma membrane-bound Rboh activity. The negatively charged superoxide product could traverse the plasma membrane as the neutral hydroperoxyl (HO 2 , pKa 5 4.88) or by being converted to the membrane-permeable H 2 O 2 (Sagi and Fluhr, 2006). In any case, this scenario predicts prior extracellular superoxide/H 2 O 2 accumulation followed by a membrane and hence diffusion-limited increase in intracellular H 2 O 2 . The resultant opposite findings reported here suggest a delay time for the extracellular response that may be due to inherent qualities of the cellular response to cryptogein. It is of interest that, in cryptogein-treated epidermal tissue, cytosolic and nuclear accumulation of ROS is more prominent than peripheral accumulation (Allan and Fluhr, 1997).
Changes in H 2 O 2 accumulation in cell populations reported by AR/AUR reagents are consistent with dynamic intracellular processes of ROS as measured by DCF. Thus, when the cell area was roughly divided into peripheral and nuclear spatial cellular compartments, the response to cryptogein as compared to the response to the exogenous addition of H 2 O 2 suggested that the nuclear region (i.e. internal region) responded before the periphery of the cell. By use of double-stain techniques, the regions for DCF accumulation were identified as cytoplasm, mitochondria, ER, and the nucleoplasm, all of which indicate that these compartments could be possible sources for intracellular ROS production or that they are compartments in which the ROS levels were modified due to elicitor addition. Mitochondria are a known source of ROS (Moller, 2001), but have not been implicated in the rapid cryptogein response that is associated with Rboh (Simon-Plas et al., 2002). In contrast to the measurement of the initial bursts of ROS featured here (seconds to minutes), longer term analysis may implicate the involvement of mitochondria and other cellular compartments. For example, cadmium treatment of BY-2 cells elicited calcium currents and Rboh-dependent ROS bursts (Garnier et al., 2006). However, later waves (beyond 30 min) of ROS response and subsequent cadmium toxicity were shown to emanate from ROS activity in the mitochondria.
An interesting finding is that one of the major compartments in plant cells reported by DCF is the endomembrane system. BFA treatment (Fig. 4) radically altered the distribution of the ER and concomitantly dramatically shifted the normal localization pattern of the DCF signal into ER-Golgi hybrid stacks. The finding that a major DCF signal emanates from around the nucleus is consistent with endomembrane localization because the ER is known to surround this organelle. The ER compartment is substantially more oxidizing than the surrounding cytoplasm (Hwang et al., 1992;Nardai et al., 2003); this oxidizing environment ensures the correct formation and isomerization of disulfide bonds facilitated by special protein disulfide isomerases that act in protein folding and assembly (Freedman et al., 1994). In the context of endomembrane reactivity to the cryptogein elicitor, this may indicate that, despite intrinsic buffering capacity, a wave of ROS is either initiated at, or floods through, endomembrane compartments. Treatment of cells with H 2 O 2 was shown to promote or inhibit cytoplasmic disulfide bond formation of select proteins in a concentration-dependent manner (Cumming et al., 2004). This oxidation can function as a cellular messenger that is likely to modulate the activity of many different proteins, leading to a variety of responses (Mori and Schroeder, 2004). The results here indicate that the endomembrane and nuclear compartments are likely targets or sources for ROS signaling.
DCF reported rapid accumulation of ROS in the nucleoplasm as a result of cryptogein addition. A major ramification of this observation is that ROS signals initiated by cryptogein permeate the cell and would impinge on the nucleoplasm redox state. This could have direct impact on a variety of transcription factors that are known to be redox regulated (Fedoroff, 2006). Visualization of H 2 O 2 transients in the cell does not necessarily indicate that it directly interacts with these targets, but that its effect may be funneled through thioredoxin-like redox transducing mechanisms (Dos Santos and Rey, 2006). Plasma membranes Figure 5. Imaging of H 2 O 2 signal in BY-2 cells and isolated nuclei in response to calcium application, as reported by DCF and AUR. A, Merged optical section of isolated BY-2 nucleus triple stained with FM 4-64, DAPI, and DCF. B, Optical section of cells stained with DCF before and after the addition of 1 mM calcium. C, Optical section of isolated nuclei stained with DCF (top inset) and AUR (bottom inset) before (control) and after the addition of 1 mM calcium (1Ca 21 ), and transmission images of the isolated nuclei. Scale bar 5 10 mm. D, Representative graph of calcium-dependent DCF response in isolated nuclei. E, Reaction rates after 1 mM calcium addition (average of at least five experiments 6 SD). and organelles, such as mitochondria, peroxisomes, and chloroplasts have been shown to act as ROS generators. Here, we show that isolated nuclei can also generate H 2 O 2 in response to calcium addition. This is consistent with the suggestion that plant nuclei are capable of generating their own calcium currents that could contribute to organelle signaling (Pauly et al., 2000). Furthermore, animal nuclei were shown to regulate calcium signals in localized subnuclear regions, indicating that calcium can simultaneously regulate independent processes in the nucleus (Echevarria et al., 2003).
Due to the isolated state of the nuclei, we do not know whether they normally function as a physiological source of ROS. In contrast to the endomembrane system, the nucleus is considered to be in a relatively reduced state (Hansen et al., 2006); thus, the signal measured may be part of the homeostasis mechanism used to maintain that state. The molecular source of this ROS is unknown. However, because the signal was partially inhibited by DPI (60%), it could be from an enzymatic source that uses flavin. RbohD and RbohF were shown to localize mainly to the plasma membrane (Keller et al., 1998;Sagi and Fluhr, 2001); nevertheless, other members of the Rboh family could reside in other cellular locations (e.g. the nuclear membrane). For example, human Nox4 (NADPH oxidase 4) from endothelial cells of the umbilical vein showed preferential localization to the nucleus (Kuroda et al., 2005). Interestingly, all plant rboh genes carry EF-hands that bind Ca 21 , and plant Rboh proteins were shown to be stimulated directly by Ca 21 (Sagi and Fluhr, 2001). Alternatively, a completely novel source could be responsible for nuclear ROS production. Taken together, the implications are that nuclei are not only reservoirs of ROS accumulation, but may actively be a source of ROS production. Whether they play a role in the early cryptogein-induced intracellular signal remains to be seen.
Our results suggest that an internal ROS signal precedes the external signal, implying that, in intact cells, the cryptogein-signaling cascade stimulates intracellular components that produce an initial H 2 O 2 signal. To what extent the later extracellular burst depends on this event is unknown. In calcium signaling, primary currents of calcium are known to stimulate other compartments to increase calcium release and thus set up an integrated signaling network of calcium currents (Sanders et al., 2002). It is intriguing that the ROS-signaling pathway appears to have adopted similar attributes.

Plant Material and Isolation of BY-2 Nuclei
Tobacco (Nicotiana tabacum) cv BY-2 cells were maintained under continuous light conditions at 26°C on a rotary platform shaker at 120 rpm and subcultured every 7 d at a 1:50-mL dilution in Murashige and Skoog medium (Murashige and Skoog, 1962) supplemented with 3% Suc and thiamine (adjusted to pH 6.2). All experiments with BY-2 cells were performed using cells in their log phase of growth after a 3:50-mL (v/v) dilution and after a further 3 d of subcultivation. Isolation of nuclei was performed as described previously (Stoppin et al., 1994). (T-3166) was a 10 mM stock solution in sterile water and was kept at room temperature. DPI (D-7909) and BFA (B-7450) were kept as 20 mM stock solution in DMSO at 220°C. Other chemicals were supplied by Sigma-Aldrich. DAPI dihydrochloride (D-8417) was a 5 mg/mL stock solution in sterile water and was kept protected from light in a desiccator at 220°C.

Expression and Purification of Cryptogein
Pichia pastoris (strain GS115) bearing the plasmid pLEP3 was a gift from Dr. Toyoki Amano (Department of Biology, Shizuoka University, Japan), and was used for cryptogein production (O'Donohue et al., 1996). Cryptogein was dissolved in distilled water as a 50 mM stock solution.

AUR and AR Assays
AR hydrogen peroxide/peroxidase assay kit (A-22188) and AUR reagent (A-36006) were from Molecular Probes. In the presence of horseradish peroxidase (HRP), these reagents react in a 1:1 stoichiometry with H 2 O 2 to produce the highly fluorescent resorufin. In all experiments, BY-2 cells were washed and resuspended in fresh medium in a final packed volume of 1:10, and then the suspension was incubated for 90 min at 50 rpm for adaptation at room temperature. In AUR assays, a working solution of 50 mL (100 mM reagent and 0.2 units/mL HRP in 13 reaction buffer) was used in a 1:1 ratio with the cells. In AR assays, cells were preloaded in the working solution (without HRP) for 10 min in the dark, then washed and resuspended in fresh medium at a 1:10 ratio. Then cells were used in a 1:1 ratio with fresh medium. All experiments were preformed in a final volume of 100 mL/microplate well. A microplate fluorescence reader (Synergy HT; Bio-Tek) was used for excitation at 520 6 20 nm and emission detection at 590 6 20 nm. When used, DPI (2 mM) was added 35 min prior to the beginning of each experiment, while 628 units/mL catalase were added at the beginning of each experiment. The addition of these reagents did not affect the fluorescent properties of AR and AUR (Supplemental Fig. S3).

Cell Labeling and Viability Tests
A modified slide was used to facilitate microscopy flow-through experiments on nonadherent live cells, such as plant BY-2 cells. In all experiments, a volume of 50 mL of suspension BY-2 cells was applied to the modified slide. All staining, washing, and reagent supplementation procedures were done directly on the slide using the incubation times and final concentrations as described in Table I. For each reagent, viability tests were conducted using Trypan blue 0.4% solutions (Sigma-Aldrich; T-8154). A final concentration of 0.1% Trypan blue was added to the specimens and incubated for 1 min. At least 500 cells were counted for each treatment tested. All viability tests showed more then 95% viability during the time spans measured.

BY-2 Cell Confocal Imaging
All images were taken with the 403 and 603 objectives on an Olympus IX70 microscope equipped with FLUOVIEW 500 CLSM. For all probes, laser intensity was adjusted to the lowest level that retained a significant signalto-noise ratio. Laser and filter combinations used are summarized in Table I. For 3D reconstructions, Z series were done with 0.5-mm steps.

Kinetics Measurements and Analysis
For fluorescence measurements of cell populations using AR and AUR reagents, the response delay time was calculated from the time of the addition of cryptogein to the time of linear accumulation. The first derivative was taken from the region of maximal linear accumulation of H 2 O 2 and the slope was extrapolated to the time axis. For kinetics measurement of individual cells using the DCF reporter, confocal signal measurements were taken from the whole cell area provided by FLUOVIEW 500 software. In the case of kinetics of ROS production from different compartments, intensity measurements were taken from defined areas of the nucleus and the cell periphery as described in Figure 2B. To calculate the time gap between signals gathered from each compartment in a given experiment, the first-order derivative of the kinetic data provided by FLUOVIEW 500 software was calculated and smoothed by the moving average function provided in an Excel worksheet (period 5 12). The peak in the first-order derivative graph was used to define the start point of the maximal rate change of signal acquisition. For each experiment, the difference between the time points to maximal rate of signal acquisition in the periphery to the maximal rate of signal acquisition in the nucleus was calculated.

BFA Treatment
Cells were treated with 40 mM BFA for different time periods. Samples of cells were taken, stained with different probes, washed three times with medium containing the same BFA concentration, and immediately imaged.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Generation of H 2 O 2 in cryptogein or W7-treated BY-2 cells as reported by AUR assay.
Supplemental Figure S2. Imaging of H 2 O 2 accumulation in BY-2 cells stained with DCF and AR.
Supplemental Figure S3. Examination of the fluorescent properties of AR and AUR reagents in the presence of DPI and catalase.