An NADPH-Oxidase/Polyamine Oxidase Feedback Loop Controls ... · An NADPH-Oxidase/Polyamine Oxidase Feedback Loop Controls Oxidative Burst Under Salinity1 Katalin Gémes2, Yu Jung

An NADPH-Oxidase/Polyamine Oxidase Feedback LoopControls Oxidative Burst Under Salinity1

Katalin Gémes2, Yu Jung Kim2, Ky Young Park*, Panagiotis N. Moschou*, Efthimios Andronis,Chryssanthi Valassaki, Andreas Roussis, and Kalliopi A. Roubelakis-Angelakis*

Department of Biology, University of Crete, Voutes University Campus, Heraklion 700 13, Greece (K.G., E.A.,K.A.R.-A.); Department of Biology, Sunchon National University, 57922 Chonnam, South Korea (Y.J.K., K.Y.P.);Department of Plant Biology and Linnean Center of Plant Sciences, Swedish University of AgriculturalSciences, Uppsala BioCentrum, 750 07 Uppsala, Sweden (P.N.M.); Department of Biology, National andKapodistrian University of Athens, Panepistimioupolis, Ilissia, Athens 118 55, Greece (C.V., A.R.); andBiological Research Centre, Hungarian Academy of Sciences, H–6726 Szeged, Hungary (K.G.)

ORCID IDs: 0000-0002-5631-6843 (K.G.); 0000-0002-1233-250X (Y.J.K.); 0000-0001-7212-0595 (P.N.M.); 0000-0003-4072-3599 (E.A.);0000-0001-8188-6656 (K.A.R.-A.).

The apoplastic polyamine oxidase (PAO) catalyzes the oxidation of the higher polyamines spermidine and spermine, contributing tohydrogen peroxide (H2O2) accumulation. However, it is yet unclear whether apoplastic PAO is part of a network that coordinates theaccumulation of reactive oxygen species (ROS) under salinity or if it acts independently. Here, we unravel that NADPH oxidase andapoplastic PAO cooperate to control the accumulation of H2O2 and superoxides (O2

$2) in tobacco (Nicotiana tabacum). To examine towhat extent apoplastic PAO constitutes part of a ROS-generating network, we examined ROS accumulation in guard cells of plantsoverexpressing or down-regulating apoplastic PAO (lines S2.2 and A2, respectively) or down-regulating NADPH oxidase (lineAS-NtRbohD/F). The H2O2-specific probe benzene sulfonyl-H2O2 showed that, under salinity, H2O2 increased in S2.2 and decreasedin A2 compared with the wild type. Surprisingly, the O2

$2-specific probe benzene sulfonyl-So showed that O2$2 levels correlated

positively with that of apoplastic PAO (i.e. showed high and low levels in S2.2 and A2, respectively). By using AS-NtRbohD/F linesand a pharmacological approach, we could show that H2O2 and O2

$2 accumulation at the onset of salinity stress was dependent onNADPH oxidase, indicating that NADPH oxidase is upstream of apoplastic PAO. Our results suggest that NADPH oxidase and theapoplastic PAO form a feed-forward ROS amplification loop, which impinges on oxidative state and culminates in the execution ofprogrammed cell death. We propose that the PAO/NADPH oxidase loop is a central hub in the plethora of responses controlling saltstress tolerance, with potential functions extending beyond stress tolerance.

Several enzymatic and nonenzymatic reactions controlthe production of reactive oxygen species (ROS; Gilroyet al., 2014; Foyer and Noctor, 2016). Superoxide ions(O2

$2) are generated mainly by the respiratory burst ox-idase homologs NADPH oxidases (encoded by the Rbohgenes), and O2

$2 dismutation by superoxide dismutase

is considered one of the major routes for subsequenthydrogen peroxide (H2O2) production (Torres et al., 2002;Kwak et al., 2003; Wang et al., 2013; Baxter et al., 2014).

The homeostasis of ROS is controlled by low-Mr in-termolecular and intramolecular compounds, such asthe polyamines (PAs). PAs are highly reactive aliphaticpolycations; the main PAs in plants are the diamineputrescine (Put) and the so-called higher PAs, spermi-dine (Spd; triamine) and spermine (Spm; tetramine;Tiburcio et al., 2014; Saha et al., 2015, and refs. therein).PA homeostasis affects a vast range of dynamic devel-opmental and metabolic processes (Paschalidis andRoubelakis-Angelakis, 2005a, 2005b; Moschou et al.,2009; Wu et al., 2010; Moschou and Roubelakis-Angelakis, 2014; Tiburcio et al., 2014; Pal et al., 2015).The oxidation of PAs is catalyzed by amine oxidases(AOs). AOs, such as the diamine oxidases (DAOs; orcopper-containing AOs) and the flavin-containing poly-amine oxidases (PAOs), localize either intercellularly(i.e. apoplast) or intracellularly (i.e. cytoplasm andperoxisomes). DAOs oxidize mainly Put, but also Spdand Spm (with much lower efficiency), yielding H2O2and aminoaldehydes. The apoplastic PAOs terminallyoxidize Spd and Spm, yielding aminoaldehydes andH2O2, while the intracellular PAOs (also referred to as

1 This work was supported by European Union and Greek Na-tional funds, by the Research Funding Program THALES (grant no.MIS 377281 to K.A.R.-A.), by the Korea Research Institute of Biosci-ence and Biotechnology (to K.Y.P.), and by the Swedish ResearchCouncil V.R. and Carl Tryggers Stiftelse för Vetenskaplig Forskning(to P.N.M.), and it was implemented in the frame of COST ActionsFA1106 and BM1307.

2 These authors contributed equally to the article.* Address correspondence to [email protected], panagiotis.

[email protected], and [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Kalliopi A. Roubelakis-Angelakis ([email protected]).

Conceptualization was by K.A.R.-A. and P.N.M.; research designwas by K.A.R.-A., K.Y.P., and P.N.M.; research execution was byK.G., Y.J.K., P.N.M., E.A., C.V., and A.R.; article writing and revisionwere by K.A.R.-A., K.G., P.N.M., and K.Y.P.

www.plantphysiol.org/cgi/doi/10.1104/pp.16.01118

1418 Plant Physiology�, November 2016, Vol. 172, pp. 1418–1431, www.plantphysiol.org � 2016 American Society of Plant Biologists. All Rights Reserved. www.plantphysiol.orgon March 14, 2020 - Published by Downloaded from

Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://orcid.org/0000-0002-5631-6843

http://orcid.org/0000-0002-1233-250X

http://orcid.org/0000-0001-7212-0595

http://orcid.org/0000-0003-4072-3599

http://orcid.org/0000-0001-8188-6656

mailto:[email protected]




http://www.plantphysiol.org


http://www.plantphysiol.org/cgi/doi/10.1104/pp.16.01118


back-converting PAOs) oxidize PAs to produce H2O2,an aminoaldehyde, and a PAwith one less amino group(in the order tetramine→triamine→diamine; Angeliniet al., 2010; Pottosin and Shabala, 2014). Through theircatabolic oxidative deamination, PAs increase the in-tracellular and extracellular H2O2 load.Under physiological or stress conditions, the rate of ROS

generation/scavenging determines their steady level; thisrate is integrated into a multitude of vital signaling cues.ROS seem to be multifaced players: at low levels, they areefficiently scavenged by enzymatic and nonenzymaticantioxidants, present in nearly all cellular compartments(Mittler et al., 2004; Miller et al., 2010; Suzuki and Mittler,2012; Baxter et al., 2014; Foyer and Noctor, 2015); at me-dium levels and up to a threshold signature, ROS partic-ipate in downstream signaling cascades that activatestress-protective effector genes/mechanisms. When acertain high level is reached, oxidative stress is establishedand ROS participate in a plethora of destructive pathwaysthat culminate in the induction of programmed cell death(PCD; Moschou et al., 2008a, 2008b; Gémes et al., 2011;Moschou and Roubelakis-Angelakis, 2014).PAOs and NADPH oxidases, major ROS generators,

have been mostly studied separately, and it remainsunknown whether they are functionally linked. Theirinvolvement in similar processes points at their possibleinterplay. Perhaps the best example of the convergentaction of PAOs and NADPH oxidases is the control ofstomatal aperture. In Arabidopsis (Arabidopsis thaliana)guard cells, abscisic acid (ABA) induces the production ofH2O2 arising from O2

$2 generated by NADPH oxidases.The produced H2O2 activates, among others, down-stream ROS-dependent Ca2+ channels contributing tocytosolic Ca2+ increase (Kwak et al., 2003; Desikan et al.,2004; Baxter et al., 2014). Likewise, ABA induces the in-crease of H2O2 in the apoplast through the up-regulationof peroxidase and apoplastic PAO (Zhu et al., 2006).In an attempt to increase our understanding of how

PAOs can contribute to processes where NADPH oxi-dases are involved, we examined the interplay betweenthese genes/enzymes. To this end, we used tobacco(Nicotiana tabacum) plants up-/down regulating apo-plastic PAO (lines S2.2 and A2, respectively; Moschouet al., 2008a, 2008b) and tobacco plants down-regulatingtwo NADPH oxidase genes (AS-NtRbohD andAS-NtRbohF; Ji and Park, 2011). We used guard cells forreal-time in vivomonitoring of apoplastic PAO/NADPHoxidase-derived H2O2 and O2

$2 intracellularly andintercellularly (Song et al., 2014). Our results provideevidence for an interplay of PAO/NADPH oxidase thatis important for balancing the ratio of intracellular andintercellular O2

$2 and H2O2 levels.

RESULTS

Apoplastic PAO Represents the Main SpdOxidation Source

Considering the large number of AOs in plants(Moschou et al., 2008c), we aimed at determining the

relative contribution of apoplastic versus intracellularPA oxidation to H2O2 production during salinity. Wepreviously established that, during salt stress, Spd issecreted into the apoplast, where it is oxidized by theapoplastic PAO (Moschou et al., 2008b). However, thecontribution of intracellular AOs to Spd oxidation un-der the same conditions was not examined. In an at-tempt to dissect the contribution of different AOs toH2O2 production, we used tobacco transgenic linesoverexpressing or down-regulating ZmPAO (S-ZmPAO[line S2.2] and AS-ZmPAO [line A2], respectively;Moschou et al., 2008a, 2008b). Line S2.2 shows in-creased while A2 shows reduced apoplastic PAO ac-tivity (Moschou et al., 2008b). In contrast to ourprevious works, here we used leaves that were not fullyexpanded, in order to take into consideration the im-portance of PAOs in developmental processes, such asleaf expansion during salt stress (Rodríguez et al.,2009). At this stage, the profile of PAs in the wild type,A2, and S2.2 was somewhat different from what hasbeen described previously (Supplemental Fig. S1;Moschou et al., 2008a, 2008b). However, the observedexpected increase of PAs in A2 and the decrease ofhigher PAs (Spd and Spm) in S2.2 suggest that theapoplastic PAO controls PA levels in expanding leaves,as was the case for the fully expanded ones (Moschouet al., 2008a, 2008b).

Next, we determined the total cellular capacity of Spdoxidation (terminal plus back conversion) versus ter-minal Spd oxidation in the wild type, A2, and S2.2. Toachieve this, we developed an in-gel Spd oxidation as-say that determines total Spd oxidation activity. Wecompared the results obtained from this in-gel Spdoxidation assay with those obtained from a colori-metric assay that determines terminal Spd oxidation(Supplemental Fig. S2A). The in-gel assay is based onthe fact that H2O2 produced by Spd oxidation reactswith 3,39-diaminobenzidine (DAB), forming a brown-ish adduct that denotes the gel regions (bands) enrichedin Spd oxidase activity. In the wild type, Spd oxidasecan be visualized as multiple bands (three main ones),with a major isoenzyme (greater than 50%) showinghigh mobility (referred to hereafter as anodal). Thisisoenzyme pattern is consistent with the large numberof predicted PAOs andDAOs in the tobacco genome (atleast one apoplastic and four intracellular PAOs andmore than 12 DAOs; Supplemental File S1). However,we could not define a large number of bands in the wildtype, suggesting that some isoenzymes may showsimilar mobility on the gel, preventing their separation,may not be present in leaves, or could be refractory tothis analytical method. In A2, the major anodal Spdoxidase isoenzymes were depleted, suggesting thatthey most likely correspond to apoplastic PAO iso-forms (Supplemental Fig. S2, A and B). In S2.2, we ob-served a significant increase of the in-gel Spd oxidasepotential and, inparticular, the appearance of an additionalfast-migrating band that could not be seen in thewild typeand A2. Although we could only achieve a fair resolu-tion of isoenzymes, we assume that the fast-migrating

Plant Physiol. Vol. 172, 2016 1419

A PAO/NADPH Oxidase Nexus

www.plantphysiol.orgon March 14, 2020 - Published by Downloaded from Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org/cgi/content/full/pp.16.01118/DC1





band corresponds to the apoplastic maize (Zea mays)PAO isoenzyme (ZmPAO), which is overexpressed inS2.2 (predicted mass approximately 53 kD). We alsoobserved in S2.2 an increase of additional bands, whichwere significantly less mobile than the band that pre-sumably corresponds to maize PAO. These isoenzymescould correspond to a posttranslationallymodifiedmaizePAO or different maize PAO fractions (Cona et al., 2006).Alternatively, the increase in apoplastic PAO may signalthe up-regulation of other AOs, or simply the DAB ad-duct, due to its higher production in S2.2, may diffuse,producing erroneous bands. The quantification of bandsin the three genotypes showed that the overall Spd oxi-dase activity in S2.2 increased significantly by 2-fold,mostly due to the increase of the anodal isoenzymes; A2lines showed a 2-fold decrease due to the absence of themajor anodal isoenzyme. Taken together, these resultssuggest that the apoplastic PAO represents themajor Spdoxidase activity.

In the colorimetric assay, DAO activity (terminaloxidation of Put) was not increased significantly amongthe three genotypes (Supplemental Fig. S2C). On theother hand, the terminal Spd oxidase activity (mainlyapoplastic PAO) was reduced significantly in A2 lines,while in S2.2 it increased by 3-fold (Supplemental Fig.S2C). In addition, apoplastic PAO activity was highlyresponsive to 200 mM salt treatment (referred to here-after as NaCl treatment), exhibiting a significant in-crease (Moschou et al., 2008b), whereas the cathodaltotal Spd oxidase activity responded moderately toNaCl treatment (Supplemental Fig. S2, B and C).

To further substantiate the previous finding, we exam-ined the Spd oxidase activity of protoplasts by the colori-metric 4-aminopterine oxidation assay used to determinethe activity of both terminal and back-converting PAOsand DAOs (Tavladoraki et al., 2006). The activity of Spdoxidase in wild-type protoplasts was negligible (close tobackground levels [as a positive control, purified AtPAO3was used in these assays]; Moschou et al., 2008c), sug-gesting that the main Spd oxidase activity resides in theapoplastic compartment. Taken together, the data pro-duced through the in-gel and in vitro assays suggest thatthe apoplastic PAO accounts for at least 50% of the totalSpd oxidase activity in expanding tobacco leaves; there-fore, it is the major Spd oxidase activity during salinity.

Apoplastic PAO Impacts O2$2 Production

Previously, we found that S2.2 plants show increasedsuperoxide dismutase activity, suggesting that O2

$2

homeostasis may be compromised in these plants(Moschou et al., 2008a). NaCl treatment can be used toexamine the contribution of apoplastic PAO to H2O2levels, and the in situ ROS detection assay is a powerfultool in the estimation of PAO-derived H2O2 levels(Moschou et al., 2008a, 2008b). Under control condi-tions, we could not detect significant differences in thestaining intensities for O2

$2 and H2O2 among the threegenotypes (Fig. 1, A and B, 0 h). NaCl treatment in-duced the increase of both ROS in a time-dependent

manner. One to 24 h posttreatment, A2 leaves con-tained lower, while S2.2 leaves contained higher, levelsof O2

$2 and H2O2 than the wild type (Fig, 1, A and B,1 h). These results were confirmed using an in vitroquantification assay for H2O2 (Fig. 1C) and suggest thatapoplastic PAO influences the production of not onlyH2O2 but also of O2

$2 under stress conditions.

The Apoplastic PAO-Dependent ROS Accumulation IsSufficient to Induce PCD within the First Few Hours ofNaCl Treatment

We have shown that apoplastic PAO is critically re-quired for PCD execution during prolonged NaCl stress(stress treatment in the range of several days; Moschouet al., 2008b). Here, we examined to what extent undershort-term NaCl treatments (in the range of hours) apo-plastic PAO-generated ROS are sufficient to induce PCDhallmarks. The array of events that precede PCD execu-tion during NaCl stress are yet unclear and might becontext/species specific. S2.2 showed an early accumu-lation of oxidized proteins (Supplemental Fig. S3, A andB; 1 h posttreatment), in contrast to A2. Significant ac-cumulation of necrotic cells was observed 6 h posttreat-ment and onward (Supplemental Fig. S3C). Thus, theaccumulation of oxidized proteins and ROS seems toprecede PCD. Our results suggest that short NaCl treat-ments (i.e. less than 24 h) are enough to induce apoplasticPAO-derived ROS accumulation of sufficient amounts toinduce PCD hallmarks. In addition, our results suggestthat protein oxidation and the accumulation of ROS areupstream events in the execution of NaCl-induced PCD,at least under the described conditions.

Guard Cells Reflect the Real-Time ROS AccumulationPost-NaCl Treatment

Guard cells have been used to study real-time ROSaccumulation (Song et al., 2014). In these cells, theNADPH oxidase genes RbohD and RbohF are involved inABA-mediated stomatal closure (Zhang et al., 2001;Kwak et al., 2003; Song et al., 2014). Similarly, apoplasticPAO contributes to ABA-induced H2O2 production inmaize under control conditions (Xue et al., 2008).

First, we used the unspecific ROS probe 29,79-dichlorofluorescein diacetate (DCFDA) to determineROS production in guard cells. DCFDA is hydrolyzedby cellular esterases to formDCFH,which is oxidized inthe presence of peroxidases by hydroxyl or organicperoxyl radicals and the reactive nitrogen species NO$and ONOO2 to form the fluorescent dye dichloro-fluorescein (DCF; Myhre et al., 2003). The intensity ofDCF reflects the formation of general reactive species(RS; the sum of nitrogen and oxygen reactive species)rather than specific ones, providing a rough estimate ofROS production. In guard cells of thewild type, A2, andS2.2, the fluorescence of DCF coincided with the totalH2O2 and O2

$2 production determined using the in situdetection method (Fig. 2A). In particular, under control

1420 Plant Physiol. Vol. 172, 2016

Gémes et al.










conditions, no significant differences were observed inDCF fluorescence in guard cells among the three geno-types (Fig. 2, 0 h). Thus, under control conditions,apoplastic PAO does not seem to influence the RS levels.However, 1 and 6 h posttreatment, S2.2 contained higher,while A2 contained lower, DCF compared with the wildtype (Fig. 2, 1 and 6 h). DCF accumulated mainly in thenucleus and chloroplasts, but also at the cell margins, of

S2.2 guard cells (Fig. 2C, 6 h). This accumulation patterndoes not necessarily reflect the RS production sites. Inaccordance, previous studies suggested that differentROS probes tend to accumulate to distinct intracellularsites that may not coincide with the ROS-producing sites(Snyrychova et al., 2009).

Next, we used more specific dyes to estimate H2O2levels in guard cells. To this end, we evaluated two dif-ferent sets of fluorescent probes. First, we used the H2O2probes Amplex Red (AR) and Amplex Ultra Red (AUR;Ashtamker et al., 2007), which are used to estimate H2O2levels intracellularly and extracellularly, respectively.Under control conditions, no significant differencescould be observed among the three genotypes inAR andAUR fluorescence intensities (Supplemental Fig. S4, Aand B, 0 h). One and 6 h posttreatment with NaCl, anincrease in AR and AUR fluorescence was detected in allthree genotypes, mostly in S2.2 plants (SupplementalFig. S4A). Significant AR and AUR fluorescent signalsaccumulated in chloroplasts. A2 plants showed reducedAR and AUR fluorescence (6 h), preceded by a transientincrease of AUR 1 h posttreatment. This transient in-crease may reflect the presence of high levels of per-oxidase in the apoplast of A2 plants or the interferenceof the probe with a cellular metabolite. Snyrychovaet al. (2009) showed that AR and AUR are highlysensitive to peroxidase levels, similar to DCF andDAB, which also are highly sensitive to peroxidase(Noctor et al., 2016).

Then,we employed a peroxidase-independentmethodfor the estimation of H2O2 levels. We used the highlyspecific benzene sulfonyl (BES)-H2O2 and BES-H2O2-Acprobes to estimate intracellular/extracellularH2O2 levels,respectively (Fig. 3). This probe pair is converted to fluo-rescent molecules in the presence of esterases and mightbemore specific thanAR andAUR,which are usedmoreextensively in the in vitro determinations of H2O2 whereperoxidases are added in surplus (Noctor et al., 2016). ByusingBES-H2O2 andBES-H2O2-Ac,weobserved a similartrend of H2O2 accumulation in S2.2 (Fig. 3). However, inthis case, we did not observe the transient increase ofH2O2 in A2 1 h posttreatment (compare Fig. 3B withSupplemental Fig. S4B). Taken together, our resultsconfirm that guard cells can be used efficiently tomonitorreal-time ROS accumulation. In addition, guard cells of-fer some unique advantages over other cell tissues/typesfor ROS detection. They are homogenous, readily acces-sible for microscopic observation, and they show a pro-found physiological responsiveness to short-term NaCltreatment. In addition, we confirm that BES-H2O2 andBES-H2O2-Ac aremore specific probes for the detection ofH2O2 levels in plants. Nevertheless, a careful assessmentof different probes might be required depending on thecontext/tissue.

PAO-Derived H2O2 Coincides with O2$2 Production in

Guard Cells

Intracellular generation of O2$2 was detected using

BES-So-AM, a highly specific fluorescent probe for O2$2

Figure 1. In situ ROS detection in leaves of wild-type (WT), A2, and S2.2plants post-NaCl treatment. A, In situ detection of O2

$2 (blue) and H2O2

(brown) levels 1, 6, and24hpost-NaCl treatment. Images are representativeof three independent experimentswith six leaf images per genotype in eachtime point. B, Quantification of O2

$2 (blue) and H2O2 (brown) signal fromthe in situ detection. NBT, Nitroblue tetrazolium; RU, relative units. C,H2O2 levels in leaves 3 and 24 h post-NaCl treatment. FW, Fresh weight;n.d., not detected. Data in B and C are means 6 SE of three independentexperiments with three technical replicates each. Different letters indicatesignificant differences of Duncan’s multiple comparisons (P , 0.05).










(Maeda et al., 2007). Under control conditions, no sig-nificant accumulation of O2

$2 could be detected in thethree genotypes (Fig. 4A, 0 h). One and 6 h posttreat-ment, the levels of intracellular O2

$2 were increasedsignificantly in guard cells of wild-type and S2.2 plantscompared with A2 plants (Fig. 4A, 1 and 6 h). Particu-larly, fluorescent BES-So-AM accumulated in the nu-cleus and chloroplasts of thewild type. BES-So-AMalsowas detected in cell margins of S2.2 guard cells. Thus,although the 1-h post-NaCl treatment pixel intensity ofBES-So-AM fluorescence differed marginally betweenthe wild type and S2.2, the difference in the total in-tracellular levels of fluorescent BES-So-AM was verybig, as estimated by counting the total number of pixelspseudocolored green (in arbitrary units: 506 10 for thewild type, 10 6 2 for A2, and 153 6 32 for S2.2; see“Materials andMethods”). The previous result is due toadditional BES-So-AM in the cell margins of S2.2 plants.

Next, we used BES-So to detect extracellular O2$2.

Similar to the intracellular O2$2, no significant accumu-

lation of BES-So could be detected under control condi-tions in the three genotypes (Fig. 4B, 0 h). One hourposttreatment, the extracellular BES-So fluorescence in-creased significantly in S2.2 and the wild type, while itincreased moderately in A2 (Fig. 4B, 1 h). Six hoursposttreatment, BES-So fluorescence increased further inthe wild type and mainly in S2.2, but not in A2 (Fig. 4B,6 h). Our results indicate that apoplastic PAO levelspositively correlate with O2

$2 levels in guard cells.

Apoplastic PAO Levels Correlate with NADPHOxidase Activity

Τhe correlation between PAO and O2$2 levels in our

experiments prompted us to examine the genetic in-teraction between PAO and two of the major NADPH

Figure 2. RS detection in guard cells of wild-type (WT), A2, and S2.2 plants post-NaCl treatment. A, Confocal laser scanningmicroscopy (CLSM) images of DCF fluorescence (green;DCFDA staining) and chlorophyll (Chl) autofluorescence (red) at 0, 1, and6 h post-NaCl treatment. White boxes (black in merged images) denote nuclei. Images at right show enlarged versions of wild-type and S2.2 guard cells (6 h). The arrow indicates the signal accumulation on the cell margins. Images are representative of threeindependent experiments with six micrographs per genotype in each time point. Bars = 20 mm. B, DCF fluorescence quantifi-cation in leaf extracts. C, Time-course quantification of DCF fluorescence in A. Data in B and C are means 6 SE of three inde-pendent experiments with three technical replicates each. Different letters indicate significant differences of Duncan’s multiplecomparisons (P , 0.05). RU, Relative units.


Gémes et al.



oxidase genes in guard cells, RbohD and RbohF (Songet al., 2014). Under control conditions, mRNA levels ofRbohD/Fwere increased significantly in S2.2 comparedwith the wild type, but not in A2 (Fig. 5A). One and 6 hpost-NaCl treatment, the mRNA levels of RbohD ten-ded to increase in all genotypes (Fig. 5A, 1 and 6 h). Thesame trend, although to a lesser extent, was observed inall genotypes for mRNA levels of RbohF. However, 6 hpost-NaCl treatment, the mRNA levels of RbohF de-creased slightly in all genotypes compared with 1 h.Under both control and NaCl treatments, the highermRNA levels of RbohD/F in S2.2 were accompanied byincreased in-gel activity of NADPH oxidase, while A2showed a marked decrease (Fig. 5, B and C, 1 h).

PAO-Mediated ROS Production Depends onNADPH Oxidase

Furthermore, we examined the physiological effect ofRbohD/F down-regulation in ROS production usingplants with silenced RbohD or RbohF (AS-NtRbohD andAS-NtRbohF; Ji and Park, 2011).We observed thatRbohF

and RbohD mRNA also were reduced in AS-NtRbohDand AS-NtRbohF (Supplemental Fig. S5), respectively.The mRNA of RbohD and RbohF share high sequencesimilarity (81%; query coverage, 89%), suggesting thatthe antisense cDNA of RbohD and RbohF can down-regulate RbohF and RbohD, respectively. Therefore, werefer to these transgenics hereafter as AS-NtRbohD/F.Importantly, under control and post-NaCl treatmentconditions, the AS-NtRbohD/F plants showed apo-plastic NtPAO similar to the wild type (SupplementalFig. S5). Interestingly, neither O2

$2, as expected, norH2O2 significantly accumulated post-NaCl treatment inthe two transgenic genotypes under control and stressconditions (Figs. 6 and 7). These results point to theimportance of NADPH oxidase in the production ofROS under short-term NaCl treatment.

In order to confirm the previous result and examinethe contribution of PAO/NADPH oxidase to a pre-sumably sustained H2O2 accumulation, we used apharmacological approach. We used the potent inhibi-tors diphenyleneiodonium (DPI; 50 mM) and guazatine(Guaz; 5 mM) to inhibit NADPH oxidase and PAO,

Figure 3. Intracellular/extracellular H2O2 in guard cells of wild-type (WT), A2, and S2.2 plants post-NaCl treatment. A, Rep-resentative CLSM images of intracellular BES-H2O2-Ac fluorescence (green) and chlorophyll (Chl) autofluorescence (red) at 0, 1,and 6 h post-NaCl treatment. White boxes (black in merged images) denote nuclei. Images are representative of three inde-pendent experimentswith sixmicrographs per genotype in each time point. Quantification of green signal is shown at right. Bars =20 mm. B, CLSM images of intercellular BES-H2O2 fluorescence (green) and chlorophyll autofluorescence (red) at 0, 1, and 6 hpost-NaCl treatment. White boxes (black in merged images) denote nuclei. Images are representative of three independent ex-periments with six micrographs per genotype in each time point. Quantification of green signal is shown at right. Bars = 20 mm.Data in charts at right are means 6 SE of three independent experiments with three technical replicates each. Different lettersindicate significant differences of Duncan’s multiple comparisons (P , 0.05). RU, Relative units.








respectively. Our guard cell assay cannot be used toassay sustained H2O2 accumulation, since even theuntreated leaf strips die out after approximately 12 h. Inorder to estimateH2O2 for a prolonged time (up to 72 h),we used whole leaves. In all genotypes, DPI amelio-rated NaCl-induced H2O2 production (SupplementalFig. S6). These data point out that NADPH oxidasecontributes significantly to the accumulation of H2O2.In the presence of Guaz and NaCl, H2O2 accumulationwas induced relative to the control, albeit to a lesserextent. The strong effect of DPI at early time points(compare 6 h with 72 h) indicates the importance ofNADPH oxidase for ROS homeostasis at the onset ofstress. As expected, the accumulation of H2O2 wasfurther inhibited by the simultaneous addition of bothDPI and Guaz, supporting the notion that the two en-zymes cooperate in constituting a feed-forward ROSamplification loop.

We should note that DPI inhibits PAO activityamong others; however, the potency of this inhibition ismuch weaker than that of Guaz (Moschou et al., 2008c).We estimated the activity of PAO in the presence of DPIor Guaz. Under our experimental conditions, in the

wild type and S2.2, DPI slightly inhibited the apoplasticPAO activity (approximately 15%; Supplemental Fig.S7). However, Guaz nullified the activity of PAO inboth genotypes within 6 h. We assume that the weakinhibitory effect of DPI on PAO is not significant.

DISCUSSION

In this work, we studied the contribution of theapoplastic PAO and the plasma membrane NADPHoxidase to ROS accumulation and how their cross talkregulates ROS homeostasis. Building on the unexpectedobservation that PAO regulates O2

$2 accumulation, theresults presented here allow us to propose a model inwhich a feed-forward amplification loop that involvesapoplastic PAO and NADPH oxidase controls ROSaccumulation. Our model integrates the observationsthat apoplastic PAO positively influences the activity ofNADPH oxidase and that NADPH oxidase is upstreamof PAO in the relay of events that control ROS accu-mulation. By detailing the relationship between PAOand NADPH oxidase, we could show the absolute re-quirement of NADPH oxidase for ROS production

Figure 4. Intracellular/extracellular O2$2 in guard cells of wild-type (WT), A2, and S2.2 plants post-NaCl treatment. A, CLSM

images of intracellular BES-So-Am fluorescence (green) and chlorophyll (Chl) autofluorescence (red) at 0, 1, and 6 h post-NaCltreatment.White boxes (black inmerged images) denote nuclei. Images next to the 1-h time point show enlarged versions of wild-type and S2.2 guard cells. The arrow indicates the signal accumulation on the cell margins. Images are representative of threeindependent experiments with six micrographs per genotype in each time point. Quantification of green signal is shown at right.Bars = 20mm. B, CLSM images of intercellular BES-So fluorescence (green) and chlorophyll autofluorescence (red) at 0, 1, and 6 hpost-NaCl treatment. White boxes (black in merged images) denote nuclei. Images are representative of three independent ex-periments with six micrographs per genotype in each time point. Quantification of green signal is shown at right. Bars = 20 mm.Data in charts at right are means 6 SE of three independent experiments with three technical replicates each. Different lettersindicate significant differences of Duncan’s multiple comparisons (P , 0.05). RU, Relative units.


Gémes et al.







within the first few hours of NaCl treatment. The apo-plastic PAO functions as an amplifier of the initial ROSaccumulation controlled by NADPH oxidase. Takentogether, our model suggests that the apoplastic PAOfeeds a stress-inducible ROS amplification loop that canlead to ROS accumulation above a toxicity threshold,culminating in PCD. Our findings allow us to extendour understanding of how apoplastic PAOs controltolerance responses during stresses. Notably, the tissue-wide role of NADPH oxidase and apoplastic PAO inROS regulation can be detailed in a single-cell context,the guard cells, by the careful selection of specific ROSprobes. The observed positive correlation between O2

$2

and apoplastic PAO levels upon short-term NaCltreatment at an organ level (leaf; Fig. 1) could be ex-trapolated in guard cells (Figs. 2–4). This finding sim-plifies analyses of ROS accumulation, considering theunique advantages of guard cells as a study system:accessibility for microscopic studies and homogenicity.The latter reason can be quite important consideringthat different cell types can have different contributionsto ROS levels.

But to what extent are the NADPH oxidase andapoplastic PAO important for guard cell physiology?It is well established that both of them contribute tothe regulation of stomatal aperture, and this role is

Figure 5. mRNA levels and activity of NADPH oxidase in wild-type (WT), A2, and S2.2 leaves post-NaCl treatment. A, Abun-dance of mRNA levels of RbohD (left) and RbohF (right) in leaves post-NaCl treatment with 200mMNaCl. B, Gel images showingthe in gel activity assay of NADPH oxidase 1 h post-NaCl treatment with 200 mM NaCl. Images are representative of three in-dependent experiments with one technical replicate in each (one gel). C, Quantification of anodal and cathodal isoenzymes ofNADPH oxidase. A similar isoenzyme pattern has been reported previously in tobacco (Sagi and Fluhr, 2001). Data in A and C aremeans 6 SE of three independent experiments with three technical replicates. Different letters indicate significant differences ofDuncan’s multiple comparisons relative to the wild type (P , 0.05). RU, Relative units.





executed through their intrinsic relation to ROS (Zhanget al., 2009; Fincato et al., 2012, and refs. therein). Loss ofRBOHF in Arabidopsis leads to the partial impairmentof ABA-induced stomatal closure, which is further re-duced, and ROS production is abolished in anAtRbohD/F mutant, suggesting that the two genes act redun-dantly in the control of stomatal aperture (Chater et al.,2015). In addition, AOs positively contribute to sto-matal closure in grapevine (Vitis vinifera; Paschalidiset al., 2010). In contrast, the acetylation of 1,3-dia-minopropane, a product of apoplastic PAO byN-ACETYLTRANSFERASE ACTIVITY1 in Arabidopsis,can result in the slowing of stomatal closure (Jammeset al., 2014). Thus, both enzymes are of critical impor-tance to the physiology of stomatal aperture and mayact redundantly or cooperatively in the same ROSnetwork.

Feed-forward loops offer an evolutionarily con-served solution to the problem of signal amplification

(Cordero and Hogeweg, 2006). Their overabundance insignaling networks most likely reflects their incremen-tal acquisition of adaptive single interactions betweendifferent components within the network. Plants haveevolved a wide array of feedback loops to control avariety of physiological responses upon various exog-enous or endogenous signals. For example, salicylateoperates in a feed-forward ROS loop that culminates incell death (Yun and Chen, 2006). Feed-forward loopsfor ROS amplification have been described in nonplantsas well, between NADPH oxidase and mitochondria-derived ROS (Graham et al., 2012). These loops aresubordinate to additional signals, such as metabolicperturbations (e.g. Glc deprivation). Likewise, thePAO/NADPH oxidase loop is subordinate to exoge-nous stress; activation of this loop requires NaCltreatment. In the absence of NaCl, the loop could not beinitiated, even though in S2.2, NADPH oxidase wasincreased in the controls (Fig. 5). Indeed, under control

Figure 6. Intracellular/extracellular H2O2 in guard cells of wild-type (WT), AS-NtRbohD, and AS-NtRbohF plants post-NaCltreatment. A, CLSM images of intracellular BES-H2O2-Ac fluorescence (green) and chlorophyll (Chl) autofluorescence (red) at 0,1, and 6 h post-NaCl treatment. White boxes denote nuclei. Images are representative of three independent experiments with sixmicrographs per genotype in each time point. Bars = 20 mm. B, CLSM images of intercellular AUR fluorescence (red) at 0, 1, and6 h post-NaCl treatment. Images are representative of three independent experiments with six micrographs per genotype in eachtime point. Bars = 20 mm.


Gémes et al.



conditions, the cellular content of O2$2 and H2O2 does

not differ significantly among the wild type, A2, andS2.2, as well among the wild type and AS-NtRbohD/F(Figs. 1–4). On the contrary, NaCl treatment dramati-cally increases both H2O2 and O2

$2 in S2.2; these ROSincrease moderately in the wild type and at very lowlevels in A2, both intracellularly and extracellularly.Taken together, these findings suggest that the PAO/NADPH oxidase loop is subordinate to yet unidentifiedsignals.What is the nature of the signals that bring about the

activation of the PAO/NADPH oxidase loop? Consid-ering that this loop is activated early after the onset ofsalinity, it is highly unlikely that it is activated by time-consuming pathways, such as lengthy transcriptionalcascade(s). In fact, accumulating evidence supports thatNADPH oxidase is amenable to several regulatoryposttranslational modifications (Li et al., 2014). Like-wise, apoplastic PAO activity also may be controlled byposttranslational modifications. In maize, apoplastic

PAO activity is controlled by its phosphorylation status(Cona et al., 2006). An alternative scenario would bethat the loop is not induced at all but its effect is maskedby the ROS-scavenging machinery. In accordance, anadaptive regulation of the ROS-scavenging machineryhas been suggested to dispose of surplus H2O2 pro-duced by apoplastic PAOduring development (Moschouet al., 2008a). This is supported by the absence of signifi-cant ROS accumulation in S2.2, although NADPH oxi-dase is preinduced in this line (Fig. 5). During stress, atransient decrease of the antioxidant machinery may leadto the unmasking of the effect of the PAO/NADPHoxidase loop that is further enhanced by additionalsignaling pathways. These two scenarios are not mu-tually exclusive and may both be plausible, perhaps atdifferent times/phases.

Taking into consideration the potency of the PAO/NADPH oxidase loop to the overall ROS contribution,the next question is to what extent these ROS signaldownstream events. A dedicated set of sensor proteins

Figure 7. Intracellular/extracellular O2$2 in guard cells of wild-type (WT), AS-NtRbohD, and AS-NtRbohF plants post-NaCl

treatment. A, CLSM images of intracellular BES-So-Am fluorescence (green) and chlorophyll (Chl) autofluorescence (red) at 0, 1,and 6 h post-NaCl treatment. White boxes (black in merged images) denote nuclei. Images are representative of three inde-pendent experimentswith sixmicrographs per genotype in each time point. Quantification of green signal is shown at right. Bars =20 mm. B, CLSM images of intercellular BES-So fluorescence (green) and chlorophyll autofluorescence (red) at 0, 1, and 6 h post-NaCl treatment.White boxes denote nuclei. Images are representative of three independent experimentswith sixmicrographs pergenotype in each time point.





is involved in the perception of ROS signals (Boschet al., 2014). These proteins are clustered in networksthat mediate signaling events leading to downstreamresponses, including changes in gene expression andthe activation of cell death programs. Our work high-lights that the PAO/NADPH oxidase loop has the po-tential to trigger cell death. Indeed, this loop producesROS of sufficient quantity to drive protein oxidationand to reach a level of cellular toxicity (SupplementalFig. S3). Protein oxidationmight be the tip of the icebergin myriad additional cell-wide consequences broughtabout by the PAO/NADPH oxidase loop, which is setin motion by NaCl treatment and may affect manydownstream processes that culminate in cell death ex-ecution. Certainly, this loop might just be a hub in aplethora of additional pathways that refine the decisiontoward cell death. However, it seems likely that thePAO/NADPH oxidase loop possesses a central regu-latory role in the execution of cell death, taking intoconsideration the tight association between apoplasticPAO levels and cell death levels.

An interesting twist to our story is the possible tem-poral dependence for a PAO/NADPH oxidase loop.The application of DPI significantly affected H2O2levels mostly at early time points (6 h), while Guaz hada minor effect that was escalated with time (more than24 h; Supplemental Fig. S6).We speculate that this time-resolved effect of the two inhibitors may indicate theinitial importance of the PAO/NADPH oxidase loop;then, PAO is uncoupled from NADPH oxidase and isrequired to sustain ROS levels. In support of this,AS-NtRbohD/F failed to accumulate O2

$2 and H2O2(Figs. 6 and 7) during the early stages of salinity, al-though they contained wild-type-like levels of apo-plastic PAO. This finding suggests that NADPHoxidase is upstream of the apoplastic PAO in ROSregulation and that an initial ROS accumulation byNADPH oxidase might be important for triggering theactivation of the apoplastic PAO pathway. However,we should note that the interaction between PAO/NADPH oxidases and their feed-forward relationshipsdo not allow us at this stage to efficiently disentangletheir distinct contribution to ROS levels. Consideringthat inhibitors may be imposed to differential uptakeduring different stages of stress, our model regardingthe temporal emergence of the loop requires furtherrefinement.

Overall, our data suggest that NADPH oxidase andthe apoplastic PAO are not parallel pathways for ROSproduction. Instead, they form a nexus and cross talk inthe frame of the strategy of plant cells to regulate ROShomeostasis. In addition, NADPH oxidase and apo-plastic PAO show a feed-forward relationship in whichhigh PAO levels correlate with high NADPH oxidaseactivity. Therefore, the two proteins are part of the sameROS homeostatic regulatory module, which affects theextracellular and intracellular cross talk of ROS regu-latory mechanisms. However, it is still unclear to whatextent intracellular PAOs affect this module. We pre-viously established that, in Arabidopsis, a peroxisomal

PAO cross talks with NADPH oxidase to activate themitochondrial alternative oxidase pathway (Androniset al., 2014). To advance our understanding of PAO/NADPHoxidase cross talk, the next critical step could beto explore how ROS signals are transduced/perceivedfor the fine orchestration of this cross talk and the rela-tionship between apoplastic and intracellular PAOs inthis regulation.

MATERIALS AND METHODS

Preparation of Transgenic Plants and Growth Conditions

The preparation of transgenic tobacco (Nicotiana tabacum ‘Xanthi’) plantswith altered expression of the maize (Zea mays) POLYAMINE OXIDASE(ZmPAO) gene (lines A2 and S2.2) has been described previously (Moschouet al., 2008a, 2008b). The preparation of transgenic tobacco specifically down-regulating the two genes coding NADPH oxidase, RbohD and RbohF, wasdescribed by Ji and Park (2011). Surface-sterilized transgenic seeds (T3 homo-zygous) were cultured on solid Murashige and Skoog medium (pH 5.8) andthen transferred to soil under light (16/8-h photoperiod and 100 mmol pho-tons m22 s21) at 25°C 6 5°C. Two- to 3-week old-plants were used.

RNA Extraction Quantitative PCR

Total RNA preparation was performed as described previously (Wi andPark, 2002). The primers used (Bionics) are shown in Supplemental Table S1.One microgram of total RNA from leaves was reverse transcribed for 30 min at42°C in a 20-mL reaction volume using the High Fidelity PrimeScript RT-PCRkit (Takara) according to the manufacturer’s instructions. Quantitative PCRwas carried in the Chromo 4 Continuous Fluorescence Detector (Bio-Rad).Cycle threshold values were analyzed using MJ Opticon Monitor Softwareversion 3.1 (Bio-Rad) and then exported to Microsoft Excel for further analysis.The reference gene b-ACTIN was used.

Protein Extraction, Western Blotting, in Gel EnzymaticAssays, and Electrophoresis

Proteins were extracted and treated as described by Papadakis andRoubelakis-Angelakis (2005). For NADPH oxidase activity staining, the pro-cedure was carried out according to Carter et al. (2007). An aliquot containing100 mg of protein from each tissue homogenate was electrophoresed on a 10%native PAGE gel. The gel was then incubated in 0.5mgmL21 NBT in 10mMTris,pH 7.4, supplied with 134 mM NADPH until bands were detected. For PAOactivity staining, 50 mg of protein extracts was electrophoretically resolved on a10% polyacrylamide gel. Subsequently, the gel was incubated in 50 mM phos-phate buffer (pH 7) for 30 min, to which 10 mM Spd was added for a further10min. The gelwas rinsed and then incubated in 50mMphosphate buffer (pH 7)containing 1 mg mL21 DAB. Protein samples that were incubated with 1 mM

Guaz prior to electrophoresis were used as negative controls.

PAO and DAO Enzymatic Assay

The spectrophotometric method developed by Federico et al. (1985) wasused to determine apoplastic PAO and DAO activities. Absorbance was read at460 nm.

Determination of Endogenous PAs

PAs were analyzed as described by Goren et al. (1982). Leaves (0.2 g) werehomogenized in 0.5 mL of 5% (v/v) perchloric acid and centrifuged at15,000 rpm for 20 min. Then, 0.2 mL of saturated sodium carbonate and 0.4 mLof dimethylaminonaphthalene-1-sulfonyl chloride (1 mg mL21 in acetone) wereadded to 0.2 mL of the supernatant, and the mixture was incubated at roomtemperature for 24 h in the dark. The dansylated products were extracted withbenzene and separated by thin-layer chromatography in chloroform:triethylamine(25:2, v/v). The separated PAs were scraped off and quantified using a spec-trophotofluorimeter (RF-1501; Shimadzu; http://www.shimadzu.com), by which


Gémes et al.






http://www.shimadzu.com


the emission at 495 nmwas recorded after excitation at 350 nm. Alternatively, PAswere determined as described previously (Kotzabasis et al., 1993) using an HP1100 high-performance liquid chromatograph (Hewlett-Packard).

Photometric Determination of H2O2 Levels

The endogenous levels of H2O2 content of the tissues were determined asdescribed by Sahebani and Hadavi (2009). Fresh leaf material (100 mg) washomogenized in an ice bath with 0.375mL of 0.1% (w/v) TCA. The homogenatewas centrifuged at 7,000 rpm for 20 min, and 0.25 mL of the supernatant wasadded to 0.25mL of 10mM potassium phosphate buffer (pH 7) and 0.5mL of 1 M

KI. The absorbance of the supernatant was read at 390 nm. The content of H2O2was determined using a standard curve.

In Situ Detection of ROS

In situ accumulation of H2O2 was detected using the method of Thordal-Christensen et al. (1997) and that of O2

$2 according to Jabs et al. (1996). In ad-dition, NaCl-treated tobacco leaves were incubated for 2 h in NBT stainingsolution (1 mg mL21, pH 7.8, 10 mM potassium phosphate buffer) at roomtemperature. To detect the in situ accumulation of H2O2, NaCl-treated tobaccoleaves were incubated for 2 h in DAB staining solution (1 mg mL21, pH 3.8) atroom temperature. Tobacco leaves were destained by boiling in 96% (v/v)ethanol and then photographed using a digital camera.

Confocal Microscopy Detection of ROS in Guard Cells

For fluorescent detection of ROS, leaf epidermal strips were used. ForDCFDA (Sigma Chemical), strips were floated on a solution of 10 mM in 20 mM

potassiumphosphate buffer (pH 6) for 10min (excitation, 4506 490 nm; barrier,5206 560 nm). AR andAUR (Invitrogen) were used at a concentration of 50mM

in 50 mM sodium phosphate buffer (pH 6) for 1 h in the dark (for AR, excitation,571 nm; emission, 585 nm; for AUR, excitation, 568 nm; emission, 581 nm). BES-H2O2-Ac and BES-H2O2 (WAKOChemicals) were used at a concentration of 50mM

in 20mMpotassiumphosphate buffer (pH 6) for 1 h in the dark (excitation, 485 nm;emission, 530 nm). BES-So-Am and BES-So (WAKO Chemicals) were used at aconcentration of 20 mM potassium phosphate buffer (pH 6) for 1 h in the dark(excitation, 505 nm; emission, 544 nm). Fluorescence was observed using theconfocal laser scanning microscope FluoView 300 (FV 300; Olympus).

Quantification of DCF in Plant Extracts

Plant leaves were homogenized with 10 mM Tris buffer (pH 7.2) and thencentrifuged at 2,000g for 5 min. The supernatant was incubated with DCFDA atroom temperature for 10 min in the dark. DCF fluorescence was detected by aspectrofluorophotometer (excitation, 485 nm; emission, 525 nm; RF-1501; Shimadzu).Data were expressed as relative fluorescence per milligram of protein.

Detection of Carbonylated Proteins

Total proteins from tobacco leaves were extracted from frozen samples bygrinding the tissue to afinepowder and resuspended inprotein extractionbuffer(50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1 mM phenyl-methylsulfonyl fluoride, and protease inhibitor cocktail [Sigma Chemical]). TheOxyBlot procedure (Millipore) was used to perform immunoblot detection ofoxidatively modified proteins by the generation of carbonyl groups. Car-bonylated proteins were detected and analyzed following the derivatization ofprotein carbonyl groups with 2,4-dinitrophenylhydrazine (DNP). Total pro-teins from tobacco leaves post-NaCl treatment (10 mg) were separated by SDS-PAGE. Following transfer to a nitrocellulose membrane, DNP-derivatizedproteins were detected by an anti-DNP antibody. Oxidation index was calcu-lated by the ratio between total proteins and standard protein of pixel-basedintegrated densitometric values using the OxyBlot procedure.

Trypan Blue Staining

To monitor cell death, NaCl-treated tobacco leaf discs were immersed for1min inaboiling solutionof 10mLof lactic acid, 10mLofglycerol, 10gofphenol,and 0.4% (w/v) Trypan Blue. After leaf discs had cooled down to room tem-perature, the solution was replaced with 70% (w/v) chloral hydrate. Leaf discswere destained overnight and then photographed using a digital camera.

Statistical and Image Analyses

Statistical analysis was carried out with SigmaPlot 12.0 statistical software.After ANOVA, Duncan’s multiple comparisons were performed. Image anal-ysis was done using FIJI software (Schindelin et al., 2012). For image quantifi-cations of NBT and DAB, we selected 10 regions of interest (five on each leafside) of the same area (rectangular) and quantified the integrated density ininverted color images. These 10 measurements corresponded to a technicalreplicate. For quantification of fluorescent signals, the same approachwas used.For the total green pixel count, we used the Adjust . Color Threshold in FIJIand regions of interest that included the guard cells. For FIJI analyses, themethods described by Moschou et al. (2013, 2016) were used.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL datalibraries under the following accession numbers: Zea mays POLYAMINE OXI-DASE, NC_024468; Nicotiana tabacum POLYAMINE OXIDASE, NP_001313211;Nicotiana tabacum RESPIRATORY BURST OXIDASE HOMOLOG F, EU072744;and Nicotiana tabacum RESPIRATORY BURST OXIDASE HOMOLOG D,EF366670.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Endogenous PA levels in the leaves of wild-type,A2, and S2.2 transgenic plants under control and 24 h post-NaCl treat-ment.

Supplemental Figure S2. PA catalytic genes/enzymes in wild-type, A2,and S2.2 transgenic plants under control and post-NaCl treatment.

Supplemental Figure S3. PCD hallmarks in wild-type, A2, and S2.2 leavespost-NaCl treatment.

Supplemental Figure S4. Intracellular/extracellular H2O2 in guard cells ofwild-type, A2, and S2.2 plants post-NaCl treatment.

Supplemental Figure S5. Relative mRNA levels of PAO, RbohD, and RbohFgenes in AS-NtRbohD and AS-NtRbohF plants post-NaCl treatment.

Supplemental Figure S6. H2O2 levels in leaves 6, 24, 48, and 72 h post-NaCl treatment in the absence or presence of DPI, Guaz, or both.

Supplemental Figure S7. Apoplastic PAO activity in the presence of DPIpost-NaCl treatment.

Supplemental File S1. PAO and DAO genes in tobacco.

ACKNOWLEDGMENTS

We thank Dr. Imene Toumi for assistance in the laboratory of K.A.R.-A.

Received July 21, 2016; accepted August 30, 2016; published September 6, 2016.

LITERATURE CITED

Andronis EA, Moschou PN, Toumi I, Roubelakis-Angelakis KA (2014)Peroxisomal polyamine oxidase and NADPH-oxidase cross-talk for ROShomeostasis which affects respiration rate in Arabidopsis thaliana. FrontPlant Sci 5: 132

Angelini R, Cona A, Federico R, Fincato P, Tavladoraki P, Tisi A (2010)Plant amine oxidases “on the move”: an update. Plant Physiol Biochem48: 560–564

Ashtamker C, Kiss V, Sagi M, Davydov O, Fluhr R (2007) Diverse sub-cellular locations of cryptogein-induced reactive oxygen species pro-duction in tobacco Bright Yellow-2 cells. Plant Physiol 143: 1817–1826

Baxter A, Mittler R, Suzuki N (2014) ROS as key players in plant stresssignaling. J Exp Bot 65: 1229–1240

Bosch M, Berger S, Schaller A, Stintzi A (2014) Jasmonate-dependent in-duction of polyphenol oxidase activity in tomato foliage is important fordefense against Spodoptera exigua but not against Manduca sexta. BMCPlant Biol 14: 257













Carter C, Healy R, O’Tool NM, Naqvi SMS, Ren G, Park S, Beattie GA,Horner HT, Thornburg RW (2007) Tobacco nectaries express a novelNADPH oxidase implicated in the defense of floral reproductive tissuesagainst microorganisms. Plant Physiol 143: 389–399

Chater C, Peng K, Movahedi M, Dunn JA, Walker HJ, Liang YK, McLachlanDH, Casson S, Isner JC, Wilson I, et al (2015) Elevated CO2-inducedresponses in stomata require ABA and ABA signaling. Curr Biol 25:2709–2716

Cona A, Rea G, Botta M, Corelli F, Federico R, Angelini R (2006) Flavin-containing polyamine oxidase is a hydrogen peroxide source in theoxidative response to the protein phosphatase inhibitor cantharidin inZea mays L. J Exp Bot 57: 2277–2289

Cordero OX, Hogeweg P (2006) Feed-forward loop circuits as a side effectof genome evolution. Mol Biol Evol 23: 1931–1936

Desikan R, Cheung MK, Bright J, Henson D, Hancock JT, Neill SJ (2004)ABA, hydrogen peroxide and nitric oxide signaling in stomatal guardcells. J Exp Bot 55: 205–212

Federico R, Angelini R, Cesta A, Pini C (1985) Determination of diamineoxidase in lentil seedlings by enzymic activity and immunoreactivity.Plant Physiol 79: 62–64

Fincato P, Moschou PN, Ahou A, Angelini R, Roubelakis-Angelakis KA,Federico R, Tavladoraki P (2012) The members of Arabidopsis thalianaPAO gene family exhibit distinct tissue- and organ-specific expressionpattern during seedling growth and flower development. Amino Acids42: 831–841

Foyer CH, Noctor G (2016) Stress-triggered redox signaling: what’s inpROSpect? Plant Cell Environ 39: 951–964

Gémes K, Poór P, Horváth E, Kolbert ZS, Szopkó D, Szepesi Á, Tari I(2011) Cross-talk between salicylic acid and NaCl-generated reactiveoxygen species and nitric oxide in tomato during acclimation to highsalinity. Physiol Plant 142: 179–192

Gilroy S, Suzuki N, Miller G, Choi WG, Toyota M, Devireddy AR,Mittler R (2014) A tidal wave of signals: calcium and ROS at the fore-front of rapid systemic signaling. Trends Plant Sci 19: 623–630

Goren R, Palavan N, Flores H, Galston AW (1982) Changes in polyaminetiter in etiolated pea seedlings following red light treatment. Plant CellPhysiol 23: 19–26

Graham NA, Tahmasian M, Kohli B, Komisopoulou E, Zhu M, Vivanco I,Teitell MA, Wu H, Ribas A, Lo RS, et al (2012) Glucose deprivationactivates a metabolic and signaling amplification loop leading to celldeath. Mol Syst Biol 26: 589

Jabs T, Dietrich RA, Dangl JL (1996) Initiation of runaway cell death in anArabidopsis mutant by extracellular superoxide. Science 273: 1853–1856

Jammes F, Leonhardt N, Tran D, Bousserouel H, Véry AA, Renou JP,Vavasseur A, Kwak JM, Sentenac H, Bouteau F, et al (2014) Acetylated1,3-diaminopropane antagonizes abscisic acid-mediated stomatal clos-ing in Arabidopsis. Plant J 79: 322–333

Ji NR, Park KY (2011) Stress-induced biphasic ethylene and ROS biosyn-thesis are synergistically interacted in cell damage. J Plant Biotechnol 38:22–29

Kotzabasis K, Christakis-Hampsas MD, Roubelakis-Angelakis KA (1993)A narrow-bore HPLC method for the identification and quantification offree, conjugated, and bound polyamines. Anal Biochem 214: 484–489

Kwak JM, Mori IC, Pei ZM, Leonhardt N, Torres M, Dangl JL (2003)NADPH oxidase Atrbohd and Atrbohf genes function in ROS-dependent ABA signaling in Arabidopsis. EMBO J 22: 2623–2633

Li L, Li M, Yu L, Zhou Z, Liang X, Liu Z, Cai G, Gao L, Zhang X, Wang Y,et al (2014) The FLS2-associated kinase BIK1 directly phosphorylates theNADPH oxidase RbohD to control plant immunity. Cell Host Microbe15: 329–338

Maeda H, Yamamoto K, Kohno I, Hafsi L, Itoh N, Nakagawa S, Kana-gawa N, Suzuki K, Uno T (2007) Design of a practical fluorescent probefor superoxide based on protection-deprotection chemistry of fluores-ceins with benzenesulfonyl protecting groups. Chemistry 3: 1946–1954

Miller G, Suzuki N, Ciftci-Yilmaz S, Mittler R (2010) Reactive oxygenspecies homeostasis and signaling during drought and salinity stresses.Plant Cell Environ 33: 453–467

Mittler R, Vanderauwera S, Gollery M, Van Breusegem F (2004) Reactiveoxygen gene network of plants. Trends Plant Sci 9: 490–498

Moschou PN, Delis ID, Paschalidis KA, Roubelakis-Angelakis KA(2008a) Transgenic tobacco plants overexpressing polyamine oxidase arenot able to cope with oxidative burst generated by abiotic factors.Physiol Plant 133: 140–156

Moschou PN, Gutierrez-Beltran E, Bozhkov PV, Smertenko A (2016)Separase promotes microtubule polymerization by activating CENP-E-related kinesin Kin7. Dev Cell 37: 350–361

Moschou PN, Paschalidis KA, Delis ID, Andriopoulou AH, Lagiotis GD,Yakoumakis DI, Roubelakis-Angelakis KA (2008b) Spermidine exodusand oxidation in the apoplast induced by abiotic stress is responsible forH2O2 signatures that direct tolerance responses in tobacco. Plant Cell 20:1708–1724

Moschou PN, Roubelakis-Angelakis KA (2014) Polyamines and pro-grammed cell death. J Exp Bot 65: 1285–1296

Moschou PN, Sanmartin M, Andriopoulou AH, Rojo E, Sanchez-SerranoJJ, Roubelakis-Angelakis KA (2008c) Bridging the gap between plantand mammalian polyamine catabolism: a novel peroxisomal polyamineoxidase responsible for a full back-conversion pathway in Arabidopsis.Plant Physiol 147: 1845–1857

Moschou PN, Sarris PF, Skandalis N, Andriopoulou AH, Paschalidis KA,Panopoulos NJ, Roubelakis-Angelakis KA (2009) Engineered poly-amine catabolism preinduces tolerance of tobacco to bacteria and oo-mycetes. Plant Physiol 149: 1970–1981

Moschou PN, Smertenko AP, Minina EA, Fukada K, Savenkov EI, RobertS, Hussey PJ, Bozhkov PV (2013) The caspase-related protease separase(extra spindle poles) regulates cell polarity and cytokinesis in Arabi-dopsis. Plant Cell 25: 2171–2186

Myhre O, Andersen JM, Aarnes H, Fonnum F (2003) Evaluation of theprobes 29,79-dichlorofluorescin diacetate, luminol, and lucigenin as in-dicators of reactive species formation. Biochem Pharmacol 65: 1575–1582

Noctor G, Mhamdi A, Foyer CH (2016) Oxidative stress and antioxidativesystems: recipes for successful data collection and interpretation. PlantCell Environ 39: 1140–1160

Pal M, Szalai G, Janda T (2015) Speculation: polyamines are important inabiotic stress signaling. Plant Sci 237: 16–23

Papadakis AK, Roubelakis-Angelakis KA (2005) Polyamines inhibitNADPH oxidase-mediated superoxide generation and putrescine pre-vents programmed cell death induced by polyamine oxidase-generatedhydrogen peroxide. Planta 220: 826–837

Paschalidis KA, Roubelakis-Angelakis KA (2005a) Spatial and temporaldistribution of polyamine levels and polyamine anabolism in differentorgans/tissues of the tobacco plant: correlations with age, cell division/expansion, and differentiation. Plant Physiol 138: 142–152

Paschalidis KA, Roubelakis-Angelakis KA (2005b) Sites and regulation ofpolyamine catabolism in the tobacco plant: correlations with cell divi-sion/expansion, cell cycle progression, and vascular development. PlantPhysiol 138: 2174–2184

Paschalidis KA, Toumi I, Moschou PN, Roubelakis-Angelakis KA (2010)ABA-dependent amine oxidases-derived H2O2 affects stomata conduc-tance. Plant Signal Behav 5: 1153–1156

Pottosin I, Shabala S (2014) Polyamines control of cation transport acrossplant membranes: implications for ion homeostasis and abiotic stresssignaling. Front Plant Sci 5: 154

Rodríguez AA, Maiale SJ, Menéndez AB, Ruiz OA (2009) Polyamineoxidase activity contributes to sustain maize leaf elongation under salinestress. J Exp Bot 60: 4249–4262

Sagi M, Fluhr R (2001) Superoxide production by plant homologues of thegp91(phox) NADPH oxidase: modulation of activity by calcium and bytobacco mosaic virus infection. Plant Physiol 126: 1281–1290

Saha J, Brauer EK, Sengupta A, Popescu SC, Gupta K, Gupta B (2015)Polyamines as redox homeostasis regulators during salt stress in plants.Front Environ Sci 3: 21

Sahebani N, Hadavi NS (2009) Induction of H2O2 and related enzymes intomato roots infected with root knot nematode (M. javanica) by severalchemical and microbial elicitors. Biocontrol Sci Technol 19: 301–313

Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, PietzschT, Preibisch S, Rueden C, Saalfeld S, Schmid B, et al (2012) Fiji: anopen-source platform for biological-image analysis. Nat Methods 9: 676–682

Snyrychova I, Ayaydin F, Hideg E (2009) Detecting hydrogen peroxide inleaves in vivo: a comparison of methods. Physiol Plant 135: 1–18

Song Y, Miao Y, Song CP (2014) Behind the scenes: the roles of reactiveoxygen species in guard cells. New Phytol 201: 1121–1140

Suzuki N, Mittler R (2012) Reactive oxygen species-dependent woundresponses in animals and plants. Free Radic Biol Med 53: 2269–2276

Tavladoraki P, Rossi MN, Saccuti G, Perez-Amador MA, Polticelli F,Angelini R, Federico R (2006) Heterologous expression and biochemical


Gémes et al.



characterization of a polyamine oxidase from Arabidopsis involved inpolyamine back conversion. Plant Physiol 141: 1519–1532

Thordal-Christensen H, Zhang Z, Wei Y, Collinge DB (1997) Subcellularlocalization of H2O2 in plants: H2O2 accumulation in papillae and hy-persensitive response during the barley-powdery mildew interaction.Plant J 11: 1187–1194

Tiburcio AF, Altabella T, Bitrián M, Alcázar R (2014) The roles of poly-amines during the lifespan of plants: from development to stress. Planta240: 1–18

Torres MA, Dangl JL, Jones JD (2002) Arabidopsis gp91phox homologuesAtrbohD and AtrbohF are required for accumulation of reactive oxygenintermediates in the plant defense response. Proc Natl Acad Sci USA 99:517–522

Wang GF, Li WQ, Li WY, Wu GL, Zhou CY, Chen KM (2013) Character-ization of rice NADPH oxidase genes and their expression under variousenvironmental conditions. J Mol Sci 14: 9440–9458

Wi SJ, Park KY (2002) Antisense expression of carnation cDNA encodingACC synthase or ACC oxidase enhances polyamine content and abioticstress tolerance in transgenic tobacco plants. Mol Cells 13: 209–220

Wu JY, Shang ZL, Wu J, Jiang XT, Moschou PN, Sun WD, Roubelakis-Angelakis KA, Zhang SL (2010) Spermidine oxidase-derived H2O2 regulatespollen plasma membrane hyperpolarization-activated Ca2+-permeable chan-nels and pollen tube growth. Plant J 63: 1042–1053

Xue B, Zhang A, Jiang M (2008) Involvement of polyamine oxidase inabscisic acid-induced cytosolic antioxidant defense in leaves of maize. JIntegr Plant Biol 51: 225–234

Yun LJ, Chen WL (2006) SA and ROS are involved in methyl salicylate-induced programmed cell death in Arabidopsis thaliana. Plant Cell Rep30: 1231–1239

Zhang X, Dong FC, Gao JF, Song CP (2001) Hydrogen peroxide-induced changesin intracellular pH of guard cells precede stomatal closure. Cell Res 11: 37–43

Zhang Y, Zhu H, Zhang Q, Li M, Yan M, Wang R, Wang L, Welti R, ZhangW, Wang X (2009) Phospholipase Da1 and phosphatidic acid regulateNADPH oxidase activity and production of reactive oxygen species inABA-mediated stomatal closure in Arabidopsis. Plant Cell 21: 2357–2377

Zhu G, Ding GH, Fang K, Zhao FG, Qin P (2006) New perspective on themechanism of alleviating salt stress by spermidine in barley seedlings.Plant Growth Regul 49: 147–156





Documents

An NADPH-Oxidase/Polyamine Oxidase Feedback Loop Controls ... · An NADPH-Oxidase/Polyamine Oxidase Feedback Loop Controls Oxidative Burst Under Salinity1 Katalin Gémes2, Yu Jung