Cell Localization of DPI-Dependent Production of Superoxide in Reproductive Tissues of the Olive Tree ( Olea europaea L.)

: Reactive Oxygen Species (ROS) are compounds derived from oxygen with important implications in biological processes in plants, some of them related to reproduction. Among ROS, superoxide is the primary oxidant, since an array of other ROS are eventually derived from this anion. Therefore, analysis of the molecular systems able to generate this molecule and the cellular compartmentalization of these events is of paramount importance. We have used the ﬂuorochrome DCFH 2 -DA and the chromogenic substrate NBT in association with DPI (a speciﬁc inhibitor of Rboh enzymes generating superoxide in plants) in combination with confocal microscopy and stereomicroscopy, respectively to identify cell localization of ROS in general, and superoxide accumulation in olive reproductive tissues. A signiﬁcant production of both ROS and superoxide has been described, showing a fairly precise spatial and temporal location throughout olive ﬂower development. The reduction of the NBT signal after the addition of DPI suggests that the generation of superoxide is largely due to Rboh or other ﬂavin oxidase activity. At the subcellular level, accumulation of O 2 •− has been located in the plasma membrane of mature pollen and germinated pollen, as well as in the rough endoplasmic reticulum and in mitochondria.


Introduction
ROS and NO are currently considered central mechanisms controlling different steps of sexual plant reproduction, as they act as signaling molecules and are subjected to extensive regulation by antioxidant enzymes and molecules including glutathione, glutaredoxins or thioredoxins, among others. Thus, redox network has been described to control different aspects of the formation of male and female gametes, pollen-stigma interaction (including self-incompatibility), pollen germination and pollen tube growth and guidance through the female tissues, as well as fertilization (see reviews by [1][2][3][4]).
Among ROS, superoxide is of paramount significance because it represents the origin of other ROS through a cascade of reactions that have been described in different subcellular localizations including mitochondria, chloroplasts, peroxisomes and the endoplasmic reticulum [5]. Different superoxide sources have been described in plants like cell wall-bound peroxidases and amine, quinone and oxalate oxidases, together with the side reactions of metabolism and electron leakage (see review by [6]). However, plant oxidative burst has been described as a physiologically controlled and rapid source of ROS in response to different stresses and situations. The enzymes responsible for this O 2 •− production in plants are NADPH oxidases, members of the so-called Rboh (respiratory burst oxidase homolog) family of enzymes. Such Rboh enzymes are differentially inhibited by diphenyleneiodonium (DPI), which is able to specifically inhibit Rboh-dependent superoxide producing activity

Plant Material
Whole flowers at five different developmental stages and mature pollen were collected from olive trees (cv. Picual) placed at the Estación Experimental del Zaidín-CSIC (Granada, Spain). Five major developmental stages were established previously [24] to better dissect flower development in the olive. These were named "green bud" (stage 1), "white bud" (2), "recently opened flower" (3), "dehiscent anther (4) and finally "senescent flower" (5). Their main morphological characteristics, size and their duration were described by these authors.
Whole flowers were used straightforward for light microscopy (LM) or confocal laser scanning microscopy (CLSM) analyses, or stored at −80 • C until use. Mature olive pollen was isolated from dehiscent anthers by vigorously shaking the panicles inside large paper bags. Pollen was sequentially sieved through 150 and 50 µm mesh nylon filters to eliminate debris, and cultured in vitro or stored at −80 • C until use.

In Vitro Pollen Germination
In vitro olive pollen germination was induced as previously described [22,25]. Briefly, pollen prehydration was performed at 30 • C in the dark for 1 h in a humid chamber before transference to the culture medium [10% (w/v) sucrose, 0.03% (w/v) Ca(NO 3 ) 2 , 0.01% (w/v) KNO 3 , 0.02% (w/v) MgSO 4 and 0.03% (w/v) boric acid]. Pollen was cultured at RT in the dark and under continuous stirring, and samples were taken after different times of germination (5 min, 1, 3 and 6 h).

CLSM Detection of ROS Production
The presence of ROS was detected by using the DCFH 2 -DA (2 ,7 -diclorodihidrofluorescein diacetate) fluorochrome. This compound is able of traversing the cell membrane thanks to its ester group, which is hydrolyzed by esterases in the cell inside to generate DCFH 2 , that will remain within the cells. Different ROS may oxidize this compound to produce DCF (dichlorofluorescein), a fluorescent component emitting at 520 nm when excited with a 480 nm wavelength. Therefore, the intensity of such fluorescence can be correlated to the ROS content, although neither the origin nor the nature of the ROS involved can be differentiated.
Either pistils or whole flowers of olive were incubated with 50 µM DCFH 2 -DA in MES-KCl [5 µM KCl, 10 mM MES, 50 µM CaCl 2 (pH 6.15)] buffer for 10 min, then washed with MES-KCl buffer for 15 min. Such incubation conditions were similar to those previously described by us and elsewhere [23,24,26]. In order to assess the specific presence of DPIdependent production of O 2 •− in olive flowers, parallel samples were incubated with the DPI inhibitor to a concentration of 200 µM (prepared from a 10 mM stock in DMSO) for 1 h prior to the fluorescein treatment. Again, concentrations of chemicals and extension of treatments are based in prior analyses [14,15,22]. Observations were performed using a Nikon C1 confocal microscope (Nikon, Japan) by exciting at 488 nm (Ar-488 laser) and using a small (30 µm) pinhole. Z-stack capture of multiple optical sections was performed, followed by projection using the FreeViewer (v3.90.) software (Nikon). Observations were performed in at least five flowers of the same stage, for each one of three independent experiments. Intensity of the fluorescence was measured using the EZ-C1 FreeViewer software selecting the region of interest (ROI) and analyzing pixel intensity, which was referred to the area of the ROI, therefore calculating the ratio fluorescence intensity/area in arbitrary units. ROIs included separately stigma surface and the anthers, as other olive flower organs showed no fluorescence.

Light Microscopy (LM) Chromogenic Detection of O 2 •− Production
In order to determine the production of O 2 •− in olive flower organs, we used the ability of this anion to reduce the chromogenic compound nitro blue tetrazolium chloride (NBT), to form a colored precipitate (formazan-blue) localizing the places of O 2 •− production [14,27]. Whole flowers or isolated pistils were used from five developmental stages previously reported [24] on the basic of macroscopic characters. They were submerged in Tris-HCl buffer (pH 7.4) containing 1 mg/mL NBT for 10 min. For inhibition studies, an incubation with 200 µM DPI (from a 10 mM stock in DMSO) for 30 min was performed prior to the incubation with the NBT solution [14]. After NBT incubation, all samples were washed with ultrapure water and then imaged. Control samples were treated with Tris-HCl buffer (pH 7.4) only. Observations were carried out with a stereomicroscope (Motic Europe, Spain) fitted with a digital Nikon Coolpix 4500 camera (Nikon, Japan), using a resolution of 2272 × 1704 dpi. Intensity of the precipitate was measured using the EZ-C1 FreeViewer (v. 3.90) (Nikon) selecting the region of interest (ROI) and analyzing pixel intensity, which was referred to the area of the ROI, therefore calculating the ratio color intensity/area in arbitrary units. ROIs included separately stigma surface and the anthers, as other olive flower organs showed no fluorescence. Observations and measurements were performed in at least five flowers of the same developmental stage, for each one of three independent experiments.

Transmission Electron Microscopy (TEM) Localization of O 2 •− Production in Mature Pollen Grains and Pollen Tubes
In order to localize the subcellular production of O 2 •− in olive pollen grains, we also used the ability of this anion to reduce the chromogenic compound NBT, which was further derived to high-electron components by post-fixation with OsO 4 . To start the procedure, both mature pollen and germinated pollen samples were slightly fixed using 1% (w/v) paraformaldehyde in 0.1 M sodium-cacodylate buffer (pH 7.2) for 4 h at 4 • C. Fixation was stopped by centrifugation and removal of the supernatant, followed by three consecutive washes with 0.05 M sodium-cacodylate buffer (pH 7.2) for 30 min at 25 • C (washing solution). Next, pollen samples were incubated with a solution of 1 mg/mL NBT in 50 mM Tris-HCl (pH 7.4) buffer, containing 0.2 mM MgCl 2 and 1 mM CaCl 2 , for 5 min at RT. The incubation was stopped again by removing the supernatant after centrifugation, followed by three washes as above.
Samples were then post-fixed with 1% OsO 4 in washing solution for 1 h at 4 • C and further rinsed (3×) in washing solution. Samples were then dehydrated through an ethanol series and embedded in Epon resin using standard procedures. Polymerization, generation of sections and contrast of Ni grids with uranyl acetate and lead citrate were performed as described by [28]. Observations were carried out in a JEM-1011 transmission electron microscope (Jeol Ltd., Tokyo, Japan) operating at 80 kV. Images were captured with a Mega View III camera using the iTEM software v 5.0 (Build 1032) (Soft Imaging System GmbH, Muenster, Germany).
Control samples were carried out by omitting the NBT incubation step. In addition, a second control was performed by incubating the samples with 200 µM DPI (prepared from a stock of 10 mM DPI in DMSO) for 30 min at RT prior to the incubation with the NBT solution.
Densitometry of electron-dense deposits close to the plasma membrane in NBT-treated pollen samples were plotted using the tool "Plot Lane" of the Quantity One 4.6.2 software (Bio-Rad, Hercules, CA, USA), and were compared with control DPI-treated samples.

Statistical Methods
All data are presented as the mean values ± standard deviation (SD). The normality of variables was assessed using Q-Q graphs. Non-parametric analysis (Mann-Whitney U test) were used to assess differences between control and DPI-treated samples. Statistical analyses were conducted at p < 0.05 level using the SPSS Statistics v27 software (IBM, Armonk, NY, USA).

Fluorescence Detection of Developmental Production of ROS in Olive Floral Organs
Fluorescent signal was exclusively detected within the emission range of 515-560 nm (green color), indicating the oxidation of DCFH 2 -DA to DCF in those places of ROS production. Autofluorescence arising from the presence of chlorophyll and other pigments was also detected and labeled in red.
The incubation of the samples with DCFH 2 -DA revealed an accumulation of ROS in olive flower organs, mainly in the stigma surface and in the pollen grains attached to it. A changing pattern of fluorescence (i.e., ROS levels) in the stigma surface was observed, which varied throughout the different stages of flower development ( Figure 1). Thus, two peaks of ROS production were detected, a first one at stage 2 (white bud), and a second peak, similar in intensity to the first, at stage 4 (dehiscent anther) to reach a minimum at the last developmental stage (senescent flower). Green fluorescence was specific of the stigma since no signal was detected in either the style or the ovary, independently of the stage visualized. On the other hand, ROS production only appeared to be significant at the time of dehiscence in the developing anthers (stage 4).
The incubation of samples with 200 μM DPI, a Rboh inhibitor, prior to DCFH2-DA treatment resulted in a sharp decrease of the green fluorescence present in the olive stigma surface at all stages of flower development analyzed, as well as in the dehiscent anthers (stage 4).

Chromogenic Detection of Developmental Production of Superoxide in Olive Floral Organs
Superoxide anion (O2 •− ) formation sites in olive reproductive organs were detected by NBT reduction. This compound is reduced by O2 •− until it forms a blue formazan precipitate, indicating the places of superoxide production ( Figure 2). A pattern was observed that varied in the different stages of development of the olive flower. In the same way that occurred with the detection of ROS by fluorescence, the location of a dark brown-purple precipitate was limited to the stigma, with no labeling marking the style or the ovary within the female organ. The intensity of the precipitate in treated stigmas gradually

Superoxide anion (O 2
•− ) formation sites in olive reproductive organs were detected by NBT reduction. This compound is reduced by O 2 •− until it forms a blue formazan precipitate, indicating the places of superoxide production ( Figure 2). A pattern was observed that varied in the different stages of development of the olive flower. In the same way that occurred with the detection of ROS by fluorescence, the location of a dark brownpurple precipitate was limited to the stigma, with no labeling marking the style or the ovary within the female organ. The intensity of the precipitate in treated stigmas gradually increased from stage 1 (green bud) until reaching a maximum at stage 4 (dehiscent anther), and then to decrease again reaching a minimum at stage 5 (senescent flower). In all floral stages, the decrease in signal intensity was evident when samples were preincubated with the DPI inhibitor.
Oxygen 2022, 2, FOR PEER REVIEW 6 increased from stage 1 (green bud) until reaching a maximum at stage 4 (dehiscent anther), and then to decrease again reaching a minimum at stage 5 (senescent flower). In all floral stages, the decrease in signal intensity was evident when samples were preincubated with the DPI inhibitor.
In the anther, the presence of precipitate was only significant at stage 4 (dehiscent anther). Moreover, there was a noticeable decrease in the precipitate amount when incubating samples with DPI. Short and identical incubation times were used for all experiments, since an increase of the exposure time to tetrazolium salts produced the complete coloration of organs and its degradation.  In the anther, the presence of precipitate was only significant at stage 4 (dehiscent anther). Moreover, there was a noticeable decrease in the precipitate amount when incubating samples with DPI. Short and identical incubation times were used for all experiments, since an increase of the exposure time to tetrazolium salts produced the complete coloration of organs and its degradation.

Subcellular localization of the O 2
•− molecule in the mature pollen grain was carried out by treatment of the olive pollen samples with nitroblue tetrazolium (NBT). Too longtime exposure to NBT reagent led to blackened images, so an optimal incubation time with NBT was critical. Overall, NBT-treated samples gave more contrasted images than control samples, probably due to a basal pattern of O 2 •− presence in the cytoplasm of the vegetative cell. Increased accumulation of O 2 •− in several regions of the mature pollen grains, likely representing the superoxide-generation sites prior to diffusion or further metabolism was visualized at the transmission electron microscope as deposits of higher electron-density (Figure 3), which were mostly located in close proximity with the plasma membrane, although some precipitate appeared in association with the outer region of vesicles. Other identified structures like the vegetative nuclei and the generative cell, lipid bodies, mitochondria or the pollen wall (i.e., intine and exine) only exhibited background electron density. Neither the sections corresponding to samples untreated with NBT nor those corresponding to DPI+NBT-treated samples (not shown) presented precipitate. These two types of samples were practically indistinguishable.
gen 2022, 2, FOR PEER REVIEW precipitate is only relevant at stage 4. In addition, in this case there is a significant reduction in density of the precipitate when samples were incubated with the DPI inhibitor. Asterisks indicat significant difference at p < 0.05; ns, no significant difference. Bars = 0.5 mm; A: anthers; AU: ar trary units; O: ovary; S: stigma; St: style.

Ultrastructural Localization of Superoxide in Olive Pollen Grains
Subcellular localization of the O2 •− molecule in the mature pollen grain was carri out by treatment of the olive pollen samples with nitroblue tetrazolium (NBT). Too lon time exposure to NBT reagent led to blackened images, so an optimal incubation ti with NBT was critical. Overall, NBT-treated samples gave more contrasted images th control samples, probably due to a basal pattern of O2 •− presence in the cytoplasm of t vegetative cell. Increased accumulation of O2 •− in several regions of the mature poll grains, likely representing the superoxide-generation sites prior to diffusion or furth metabolism was visualized at the transmission electron microscope as deposits of high electron-density (Figure 3), which were mostly located in close proximity with the plasm membrane, although some precipitate appeared in association with the outer region vesicles. Other identified structures like the vegetative nuclei and the generative cell, lip bodies, mitochondria or the pollen wall (i.e., intine and exine) only exhibited backgrou electron density. Neither the sections corresponding to samples untreated with NBT n those corresponding to DPI+NBT-treated samples (not shown) presented precipita These two types of samples were practically indistinguishable.

Ultrastructural Localization of Superoxide in Olive Pollen Tubes
The accumulation of O 2 •− in the germinated pollen was visualized at the TEM also as electron-dense deposits (Figure 4), which were also highly abundant in the protruding vegetative cell cytoplasm close to the plasma membrane. In this case, other subcellular structures like the endoplasmic reticulum (alone, or associated to storage lipids), and some organelles similar to undifferentiated mitochondria were labeled over the background (Figure 4). The vegetative nucleus, the generative cell, other organelles like lipid bodies and the pollen tube wall were clean of enhanced precipitate. As in the mature pollen grain, neither the sections corresponding to control (untreated with NBT) samples nor those corresponding to DPI+NBT-treated samples (not shown) presented precipitate.

Ultrastructural Localization of Superoxide in Olive Pollen Tubes
The accumulation of O2 •− in the germinated pollen was visualized at the TEM also as electron-dense deposits (Figure 4), which were also highly abundant in the protruding vegetative cell cytoplasm close to the plasma membrane. In this case, other subcellular structures like the endoplasmic reticulum (alone, or associated to storage lipids), and some organelles similar to undifferentiated mitochondria were labeled over the background (Figure 4). The vegetative nucleus, the generative cell, other organelles like lipid bodies and the pollen tube wall were clean of enhanced precipitate. As in the mature pollen grain, neither the sections corresponding to control (untreated with NBT) samples nor those corresponding to DPI+NBT-treated samples (not shown) presented precipitate.

Discussion
The localization of ROS in the reproductive tissues of the olive tree by means of DCFH2-DA shows that the stigma and anther tissues present high levels of these mole-

Discussion
The localization of ROS in the reproductive tissues of the olive tree by means of DCFH 2 -DA shows that the stigma and anther tissues present high levels of these molecules. A large part of these ROS seems to correspond to superoxide anion (O 2 •− ), as shown by NBT reduction experiments, although the massive presence of other reactive oxygen species, mainly H 2 O 2 , cannot be ruled out [24]. In addition to this specific spatial distribution, ROS undergo a clear temporal evolution throughout the development of floral organs, which seems to indicate that these molecules play important roles in reproductive biology. One of the questions that arise in relation to the presence of ROS in reproductive tissues deals with its origin. The superoxide anion (O 2 •− ) is one of the most important ROS, as a product of the reduction of molecular oxygen by one electron, which occurs widely in nature. In the case of reproductive tissues, the marked reduction in the presence of ROS detectable with DCFH 2 -DA and of O 2 •− detectable with NBT after the addition of DPI suggests that much of this superoxide is generated by the action of a NADPH oxidase (Rboh) or other flavin oxidase activity. DPI has been described as a highly specific inhibitor of these enzymes in mammalian neutrophils by binding to the two structural components of the protein [29]. Our recent description of Rboh activity in olive pollen (OeRbohH, [22]), which displays sensitivity to DPI, may explain the changes in fluorescence and NBT histochemical analyses observed in pollen, as described here.
The present work describes differences between the localization patterns of ROS in general, and of superoxide in particular (as a marker of Rboh activity), throughout the stages of stigma development. The levels of both types of molecules does not maintain a parallel evolution throughout floral development. For example, the fluorescence due to ROS is maximum in the "white bud" stage and minimum in the "recently opened flower" stage, while in the case of superoxide, the maximum concentration is progressively reached in the "dehiscent anthers" stage. These differences may correspond essentially to the presence of other ROS such as H 2 O 2 and undoubtedly reflect the activity of other antioxidant enzymes such as superoxide dismutase and peroxidase, which have yet to be established. The biological significance of the presence of these ROS may differ between the earliest stages of flower development (defense function, similar to the Rboh activity present in the redox cycle of nectaries, in which it actively participates in the protection against pathogens [27], the stages in which there is an intense pollen-pistil interaction and the stages in which tissue senescence begins [24]. The results of the use of both methods (fluorescence and chromogenic detection of superoxide) can be considered comparable, and similar trends were observed after quantification. However, slight differences like the presence of two peaks of superoxide generation observed after fluorescence quantification, which were not detected with the chromogenic substrate, may indicate a higher ability of the fluorochrome to detect minor changes, due to the high dynamic range of fluorescence detection in comparison with color detection of chromogens.
Both methods, which are here used at low magnifications, provide useful information regarding the distribution of these ROS in the different organs of the flower. However, they can deliver limited subcellular information only, even when used at higher magnifications [24]. As an example, pollen grains over the stigma surface can only be poorly distinguished at stages 4 and 5 of Figures 1 and 2. This is the major reason of developing a specific method for identification of superoxide generation at ultrastructural level, based on the use of transmission electron microscopy.
At the subcellular level, the treatment of pollen with NBT allowed the localization of precipitate due to the presence of superoxide in the plasma membrane of the mature and germinated pollen, which has also been found in the endoplasmic reticulum and mitochondria. The origin of this O 2 •− can be diverse, and it is difficult to be determined due to the limited permeability of the pollen grain to inhibitors and even to NBT [30]. In this case, it is not excluded that the detection of superoxide generated in the electron transport chain of mitochondria may occur [31]. Such O 2 •− would not be a consequence of the activity of a NADPH oxidase in this case. The fractionation of plant tissues has shown that Rboh enzymes are located in the plasma membrane [19,32], and this distribution may be asymmetric [33]. Rbohs can even mediate the transfer of electrons to produce O 2 •− in the extracellular space, propagating these types of signals up to several cells away [34]. The plasma membrane is impermeable to superoxide in animals due to its negative charge at physiological pH (pKa is 4.8). In plants, the extracellular pH is close to 5, and 16% of the superoxide is in the form of hydroperoxyl capable of crossing the membrane. Once in the cytoplasm, SOD will dismutate superoxide to hydrogen peroxide. Our localization is fully compatible with these observations.
In humans, NOX5, which is the most structurally related form of Rboh in plants, has been detected inside the cell, specifically in the endoplasmic reticulum [35,36]. On the other hand, the main NOX2 subunits in endothelial cells colocalize with a perinuclear distribution, and exist as a functional intracellular complex associated with the cytoskeleton [37]. Rbohactive endosomes have also been identified in plants [38][39][40]. Such observations are also in good agreement with the localization of precipitate close to the vesicles boundaries that we describe here. Dark discrete electron-dense spots inside the vesicles should not be considered a consequence of superoxide generation, as they represent normal features of olive and other pollen grains ultrastructure, classically described before [41], and which are likely composed of calcium, magnesium and phosphorus salts [42].
The analysis of OeRbohH activity during pollen grain germination and pollen tube growth carried out by [22], offers a series of interesting results to determine the possible role of ROS in pollen functionality. Firstly, it is clear that OeRbohH activity is essential for pollen to germinate and generate an effective pollen tube, since treatment with the DPI inhibitor either prevents pollen grain germination or significantly reduces the elongation capacity of the pollen tube. Superoxide generation by OeRbohH is also quite polarized towards the apical growth zones of the tube, and is even focused to the apical end of the pollen tube segmented by callose plugs. This is a characteristic already observed in other species such as Nicotiana tabacum [14].
All these results probably relate the Rboh activity to the polar growth capacity of the pollen tube. There is a possible analogy with the growth of Arabidopsis root hairs, which requires the production of ROS by the Rboh homologue AtrbohC [33]. The methods described in the present work (e.g., subcellular analysis of superoxide accumulation) will enable further research into these aspects, and for example identify specific organelles involved in the accumulation of superoxide.
The future characterization of superoxide-generating activities in the stigma, whose tissues make the greatest contribution to the formation of this molecule (as shown in this report), is of crucial importance. The study of this activity is initially a challenge, given that there are a number of factors such as the limitation of the amount of material available, and the fact that gynoecium extracts contain a high level of secondary metabolites (phenols, lipids, polysaccharides) that require an additional fine-tuning of the procedures for biochemical and molecular analysis.

Conclusions
Several reproductive tissues (stigma and anthers) of the olive tree show high levels of reactive oxygen species, which are detectable either by fluorescence or by using histochemical stains, and display a specific spatial and temporal distribution. The intensity of the stain is highly modulated by DPI, a known inhibitor of NADPH oxidase (NOX) activity. Therefore, labeling can be associated to the generation of O 2 •− . The accumulation of superoxide anion in the pollen grain at the subcellular level takes place in the plasma membrane of the mature and germinated pollen. Hydration and activation of pollen metabolism through germination induces additional localization of O 2 •− in the endoplasmic reticulum and mitochondria. The origin of this O 2 •− might be diverse and not only attributable to NOX activity.