Differential Function of Endogenous and Exogenous Abscisic Acid during Bacterial Pattern-Induced Production of Reactive Oxygen Species in Arabidopsis

Abscisic acid (ABA) plays important roles in positively or negatively regulating plant disease resistance to pathogens. Here, we reassess the role of endogenous and exogenous ABA by using: 35S::ABA2, a previously reported transgenic Arabidopsis line with increased endogenous ABA levels; aba2-1, a previously reported ABA2 mutant with reduced endogenous ABA levels; and exogenous application of ABA. We found that bacterial susceptibility promoted by exogenous ABA was suppressed in 35S::ABA2 plants. The 35S::ABA2 and aba2-1 plants displayed elevated and reduced levels, respectively, of bacterial flagellin peptide (flg22)-induced H2O2. Surprisingly, ABA pre-treatment reduced flg22-induced H2O2 generation. Exogenous, but not endogenous ABA, increased catalase activity. Loss of nicotinamide adenine dinucleotide phosphate oxidase genes, RBOHD and RBOHF, restored exogenous ABA-promoted bacterial susceptibility of 35S::ABA2 transgenic plants. In addition, endogenous and exogenous ABA had similar effects on callose deposition and salicylic acid (SA) signaling. These results reveal an underlying difference between endogenous and exogenous ABA in regulating plant defense responses. Given that some plant pathogens are able to synthesize ABA and affect endogenous ABA levels in plants, our results highlight the importance of reactive oxygen species in the dual function of ABA during plant-pathogen interactions.


Introduction
Abscisic acid (ABA) is a class of metabolites known as isoprenoids that controls seed maturation and germination, as well as adaptation to abiotic stress [1,2]. As a stress-related hormone, ABA plays a negative or a positive role in plant disease resistance, depending on pathogen lifestyles. The ABA is formed by cleavage of C 40 carotenoids in the 2-C-methyl-d-erythritol-4-phosphate pathway [3,4]. Zeaxanthin is produced as a trans-isomer via β-carotene. Zeaxanthin epoxidase (ZEP) catalyzes the conversion of zeaxanthin and antheraxanthin to violaxanthin in plastids [5,6]. The formation of cis-isomers of violaxanthin and neoxanthin may require a neoxanthin synthase and an isomerase [2]. The cis-isomers of violaxanthin and neoxanthin are oxidized by nine-cis-epoxycarotenoid dioxygenase (NCED) to generate a C 15 product, xanthoxin, and a C 25 metabolite [7]. The Xanthoxin is further converted by a short-chain alcohol dehydrogenase/reductase 1 to abscisic aldehyde [8,9]. The final conversion of abscisic aldehyde to ABA is catalyzed by abscisic aldehyde oxidase (AAO) and a molybdenum co-factor sulfurase is required to ensure AAO activity [10][11][12][13].
ABA biosynthetic mutants have been widely used to investigate the role of ABA in plant defense responses [14][15][16]. Following Leptosphaeria maculans infection, the Arabidopsis ZEP mutant aba1-3 showed enhanced susceptibility, impaired callose deposition, enhanced expression of the SA-dependent pathogenesis-related protein 1 gene PR1, and reduced expression of the jasmonic acid and ethylene (JA/ET)-responsive defensin gene PDF1.2 [17]. Compared with wild-type Arabidopsis, many defense-related genes regulated by SA, JA, and ET were constitutively up-regulated in aba1-6 [18]. Compared with wild-type tomato, PR1a expression in the AAO gene mutant sitiens was induced with a very low concentration of the SA-analog benzo(1,2,3)thiadiazole-7-carbothioic acid S-methyl ester (BTH) [19]. Compared with wild-type Arabidopsis, JA accumulation induced by the virulent Pseudomonas syringae pv. tomato (Pst) strain DC3000 was lower in the molybdenum co-factor sulfurase ABA3 gene mutant aba3-1 [20]. Recently, it has been reported that aba3-1 exhibited enhanced expression of SA-responsive genes when challenged with Golovinomyces cichoracearum [21]. Nevertheless, it must be noted that ABA deficiency often leads to growth or developmental defects [16], as seen with aba2-2 exhibiting a wilting phenotype [22]. Developmental differences between mutant and wild type might result in a biased evaluation of ABA function.
An ABA2 transgenic Arabidopsis line 35S::ABA2 has been demonstrated to enhance endogenous ABA levels and grow normally [23], compared to wild-type plants. In this study, the transgenic line 35S::ABA2, the ABA2 gene mutant aba2-1 exhibiting a dwarf and wilting phenotype [10], and exogenous application of ABA, were together used for reassessment of endogenous and exogenous ABA in plant defense responses.

Endogenous ABA Antagonizes Exogenous ABA-Aggravated Susceptibility
ABA negatively regulates resistance to biotrophic bacteria P. syringae in Arabidopsis and exogenous application of ABA restores disease susceptibility of ABA-deficient mutants [14][15][16]. To better understand the role of ABA in plant disease resistance, wild-type (Col-0), 35S::ABA2, and aba2-1 Arabidopsis plants were pre-treated with distilled water (as control) and 50 µM ABA, prior to inoculation with virulent PstDC3000. Pathogenicity analysis showed that in the water controls, 35S::ABA2 plants displayed severe chlorosis and enhanced bacterial growth, whereas aba2-1 had no observed symptoms and had lower bacterial numbers, compared to wild-type ( Figure 1). Following exogenous ABA pre-treatment, aba2-1 exhibited severe necrosis and strongly enhanced bacterial numbers compared to wild-type ( Figure 1). Meanwhile, enhanced disease symptoms, but not bacterial numbers, were exhibited in 35S::ABA2 under ABA treatment ( Figure 1). These results indicate that endogenous and exogenous ABA overlap with each other in repressing bacterial resistance, but it is also possible that endogenous ABA antagonizes exogenous ABA-promoted susceptibility to PstDC3000, at least partly, in Arabidopsis.

Endogenous ABA Enhances, but Exogenous ABA Represses Bacterial Pathogen-Associated Molecular Pattern (PAMP)-Induced ROS Burst
ROS, such as hydrogen peroxide (H 2 O 2 ), play an important role in plant defense responses [24]. We next examined the effect of endogenous and exogenous ABA on the bacterial flagellin peptide flg22-induced H 2 O 2 production. The results showed that 35S::ABA2 had higher H 2 O 2 levels under induction with 1 µM flg22 than wild-type, whereas aba2-1 had the lowest H 2 O 2 levels ( Figure 2A). Interestingly, exogenous application of ABA reduced flg22-induced H 2 O 2 production in wild-type plants when compared to water controls ( Figure 2B). These results indicate that endogenous and exogenous ABA function differently during PAMP-induced ROS burst in Arabidopsis.

Exogenous ABA, but not Endogenous ABA, Enhanced Catalase Activity under PstDC3000 Infection in Arabidopsis
In Arabidopsis, the nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, such as RBOHD and RBOHF, are mostly responsible for ROS generation during plant-pathogen interactions [25]. We examined whether endogenous and exogenous ABA differently regulate RBOHD and RBOHF transcripts in Arabidopsis under PstDC3000 infection. Compared to wild-type, 35S::ABA2 plants displayed an increase in the transcript levels of RBOHD and RBOHF ( Figure S1). In contrast, aba2-1 showed a reduced level of expression of these two genes under PstDC3000 infection ( Figure S1). ABA treatment also enhanced the expression of RBOHD and RBOHF, compared to water treatment ( Figure S2). These results indicate that both endogenous and exogenous ABA upregulate RBOHD and RBOHF expression in Arabidopsis under PstDC3000 infection.
The alleviation of ROS-induced oxidative damage is often balanced via an efficient antioxidant defense system; for example, catalase (CAT) reacts with H 2 O 2 to form oxygen and water [26]. We further detected whether CAT is differentially regulated by endogenous and exogenous ABA under PstDC3000 infection. The results showed that CAT activity was not different between wild-type, 35S::ABA2, and aba2-1 plants ( Figure 3A). However, ABA treatment significantly enhanced CAT activity in wild-type plants under PstDC3000 infection ( Figure 3B).

RBOHD and RBOHF are Required for Endogenous ABA to Antagonize Exogenous ABA-Enhanced Bacterial Susceptibility in Arabidopsis
To investigate the mechanism of antagonism of 35S::ABA2 plants to exogenous ABA-promoted PstDC3000 susceptibility, we generated 35S::ABA2rbohD/F plants by crossing 35S::ABA2 and the rbohD/F double mutant. The 35S::ABA2rbohD/F plants were pre-treated with ABA prior to inoculation with PstDC3000. The results showed that ABA treatment enhanced disease symptoms and bacterial numbers in 35S::ABA2rbohD/F (Figure 4), indicating that RBOHD and RBOHF are required for endogenous ABA to antagonize exogenous ABA-aggravated bacterial susceptibility. Error bars indicate the standard error, and values were based on eight independent replicates. Asterisk indicates a significant difference (t-test, p < 0.05). These experiments were performed at least three times.

Both Endogenous and Exogenous ABA Repress Callose Formation and SA Signaling
In Arabidopsis, PAMP-induced callose formation has been widely used as a marker to characterize plant defense responses [27,28]. We assessed the effects of endogenous and exogenous ABA on flg22-induced callose generation by using wild type, 35S::ABA2, and aba2-1 plants. With water treatment, 35S::ABA2 showed no flg22-induced callose deposition, whereas aba2-1 exhibited enhanced callose deposition compared to wild type ( Figure 5A). Under ABA treatment, all tested plants showed no flg22-induced callose deposition ( Figure 5A). These results suggest that both endogenous and exogenous ABA have similar functions in PAMP-induced callose deposition.
SA-dependent defense signaling has also been used as a marker to characterize plant defense resistance [29]. We next investigated the effect of endogenous and exogenous ABA on SA defense signaling during PstDC3000 infection. qRT-PCR analysis showed that 35S::ABA2 had lower SA-dependent PR1 gene expression than wild-type plants, whereas aba2-1 had higher PR1 expression than wild type ( Figure 5B). In addition, following PstDC3000 infection, ABA treatment reduced PR1 transcripts in wild-type plants, compared to water treatment ( Figure 5C). These results indicate that endogenous and exogenous ABA function similarly in SA defense signaling during pathogen infection.

Discussion
Our data demonstrated that the ABA2 overexpressing transgenic line 35S::ABA2 showed no apparent response to exogenous ABA on PstDC3000 multiplication (Figure 1). This indicates that endogenous ABA overlaps with exogenous ABA in regulating bacterial resistance. However, it is also possible that endogenous ABA antagonize exogenous ABA-promoted bacterial susceptibility. Interestingly, we found that endogenous and exogenous ABA function in opposing ways during flg22-induced H 2 O 2 production ( Figure 2). Analysis of gene expression demonstrated that both endogenous and exogenous ABA increased RBOHD and RBOHF transcripts (Figures S1 and S2). In addition, exogenous, but not endogenous ABA, enhanced CAT activity (Figure 3), thus giving an explanation for the difference between endogenous and exogenous ABA action on flg22-induced ROS burst. Loss of RBOHD and RBOHF restored the susceptibility of 35S::ABA2 to exogenous ABA-promoted PstDC3000 infection (Figure 4). These findings reveal that differential regulation of ROS production by endogenous and exogenous ABA is a main reason for the antagonistic effect of endogenous ABA on exogenous ABA-enhanced PstDC3000 susceptibility in Arabidopsis.
Among ten NADPH oxidases in Arabidopsis, RBOHD and RBOHF play a main role in ROS generation during plant-pathogen interactions. Under recognition of PAMPs by the plasma membrane-localized receptor kinases (RKs) in plants, rapid cellular responses, such a burst of ROS are activated [25]. One of the best-studied RK signaling pathways is that the recognition of flg22 by FLAGELLIN-SENSING-2 (FLS2) [30] recruits BRI1-ASSOCIATED RECEPTOR KINASE 1 (BAK1), another RK as a co-receptor [31], to trans-phosphorylation of BOTRYTIS-INDUCED KINASE1 (BIK1), a cytoplasmic protein [32]. Phosphorylation of RBOHD by BIK1 is required for RBOHD activation [31]. In parallel to PAMPs-induced ROS burst, ABA also regulates ROS generation through RBOHD and RBOHF 24]. OPEN STOMATA 1 (OST1), one of ABA receptors, is required for ABA-induced ROS generation [33]. In addition, phosphorylation of OST1 by BAK1 is required for ABA-induced ROS generation in guard cells [34]. These findings indicate a crosstalk between PAMP and ABA-induced ROS production and imply that activation of RK signaling contributes to ABA-induced ROS generation. Since 35S::ABA2 and aba2-1 separately enhanced and reduced flg22-induced H 2 O 2 production (Figure 2A), it is apparent that endogenous ABA contributes to PAMPs-induced ROS burst, and this contribution might be resulted from an underdetermined mechanism regulated by endogenous ABA. Despite ROS generation is required for both PAMPs-and ABA-induced stomatal immunity, but BIK1 is only required for stomatal closure induced by flg22 [24], indicating the difference between PAMP-and ABA-induced ROS generation. Our results that ABA treatment reduced flg22-induced H 2 O 2 generation ( Figure 2B) further reveal that exogenous ABA antagonizes PAMP-induced ROS production.
Indeed, it is not necessarily surprising for the different role of endogenous and exogenous ABA on ROS burst. On the one hand, ABA-responsive signal transduction is often involved in accumulation of H 2 O 2 under abiotic stress [35]. It is likely that flg22 functions as abiotic stimulus. On the other hand, ABA limits H 2 O 2 generation in some physiological processes, such as seed germination [36], and the transient alteration of ABA levels in the systemic leaf triggers acclimation of the systemic tissue [37]. It is possible that exogenous application of ABA prioritizes physiological processes by repressing flg22-induced H 2 O 2 production.
ABA pre-treatment completely suppresses flg22-induced callose deposition in Arabidopsis [38]. ABA-hypersensitive lines and ABA treatment destroyed, whereas ABA-insensitive lines augmented, PstDC3000-induced callose formation [39]. Our data that 35S::ABA2 and aba2-1 separately reduced and increased flg22-induced callose deposition, respectively ( Figure 5A). Meanwhile, ABA treatment repressed flg22-induced callose deposition in wild-type plants ( Figure 5A). These findings suggest that endogenous and exogenous ABA function as repressors in PAMP-induced callose deposition. A recent report showed that the expression of SA-responsive genes is higher in the ABA-deficient mutant aba3-1 than in wild type, whereas ABA treatment restores the expression of SA-responsive genes in aba3-1 to the levels seen in wild-type plants [21]. Our data that aba2-1 increased, but 35S::ABA2 and wild-type plants treated with ABA reduced the expression of SA-dependent PR1 gene expression under PstDC3000 infection ( Figure 5B,C), indicate the similarity between endogenous and exogenous ABA in SA signaling, and support the antagonism of ABA in SA signaling in Arabidopsis. Considering the antagonism of endogenous ABA to exogenous ABA in disease susceptibility, our data further reveal that callose and SA are not related to this antagonistic relationship in PstDC3000 susceptibility in Arabidopsis. Repression of callose deposition and SA signaling by both endogenous and exogenous ABA explains why 35S::ABA2 and ABA-treated wild-type plants and aba2-1 are all susceptible to PstDC3000.
In conclusion, aba2-1 and 35S::ABA2 characterized conversely on PAMP-induced ROS burst and callose deposition, as well as SA signaling, which; thus, removes the supposed effect of dwarfish of aba2-1 on endogenous ABA function in plant defense responses. Particularly, we found that endogenous and exogenous ABA had completely converse function on PAMP-induced ROS burst, which is probably derived from differential regulation of ROS levels by endogenous and exogenous ABA affecting CAT activity. Since some pathogens are able to synthesize ABA and mediate endogenous ABA levels in planta [39][40][41], our results; thus, provide new insights into understanding the dual function of ABA in plants defending against bacterial pathogens.

ABA Treatment and Bacterial Inoculation
For analyzing bacterial pathogenicity, ROS burst, callose deposition, and the expression of defense-related marker gene PR1, four-week-old Arabidopsis plants were sprayed with distilled water (H 2 O) as a control or 50 µM ABA. After 4 h, treated leaves were infiltrated with a suspension of 10 6 colony forming units (cfu) mL −1 PstDC3000 in 10 mM MgCl 2 . Photographs of leaves were taken at 3 dpi. These experiments were repeated at least three times.

Pathogenicity Analysis
After 3 days post inoculation (dpi) with virulent PstDC3000, eight leaf discs (0.74 cm 2 per disc, two leaf discs from one plant) per treatment were ground in 10 mM MgCl 2 and used as one biological replicate. Dilutions of leaf homogenates with 10 mM MgCl 2 were spotted onto KB agar medium containing 100 µg mL −1 Rifampicin. Eight biological replicates were harvested for each genotype used and each disease test was repeated at least three times.

Oxidative Burst
Four-week-old Arabidopsis leaves were sliced into strips of approximately 1 mm length and incubated overnight in distilled water in a 96-well plate. After removal of water from the 96-well plate, 1 µM flg22 in distilled water (H 2 O), supplemented with 20 mM luminol and 1 µg horseradish peroxidase (Sigma-Aldrich, Saint Louis, MO, USA), were added. Luminescence was recorded immediately with a GloMax ® 96 Microplate Luminometer (Promega, Madison, WI, USA).

Antioxidant Enzyme Activity
Total CAT activity of four-week-old Arabidopsis leaves was detected as previously reported [46].

qRT-PCR
Total RNA was extracted with RNeasy Plant Mini kit (Qiagen, Shanghai, China). RNA samples were treated with DNase Turbo DNAfree (Promega Madison, WI, USA). Total RNA (2 µg) was used for reverse transcription with SuperScript III reverse transcriptase (Invitrogen, Shanghai, China). qRT-PCR was performed using an Applied Biosystems 7500 Real-Time PCR System (Thermoscientific, Waltham, MA, USA) and SYBR Premix Ex Taq kit (TaKaRa, Otsu, Shiga, Japan). Gene transcripts were standardized using ACT2 [47] as a control for PR1, RBOHD, and RBOHF. The primers are listed in Table S1.
Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/10/ 2544/s1. Figure S1: Effects of endogenous ABA on RBOHD (A) and RBOHF (B) transcripts under PstDC3000 infection in Arabidopsis. Error bars indicate the standard error, and values were based on at least three independent replicates. Different letters indicate values that are significantly different (p < 0.05) from each other as determined by one-way ANOVA. These experiments were performed at least three times. Figure S2: Effects of exogenous ABA on RBOHD (A) and RBOHF (B) transcripts under PstDC3000 infection in Arabidopsis. Wild-type (Col-0) plants were treated with distilled water (H2O) or ABA before inoculation with PstDC3000. Error bars indicate the standard error, and values were based on at least three independent replicates. Different letters indicate values that are significantly different (p < 0.05) from each other as determined by one-way ANOVA. These experiments were performed at least three times. Table S1: List of genotyping primers.