A salivary EF-hand calcium-binding protein of the brown planthopper Nilaparvata lugens functions as an effector for defense responses in rice

The brown planthopper (BPH), Nilaparvata lugens (Stål) (Hemiptera: Delphacidae), a major pest of rice in Asia, is able to successfully puncture sieve tubes in rice with its piercing stylet and then to ingest phloem sap. How BPH manages to continuously feed on rice remains unclear. Here, we cloned the gene NlSEF1, which is highly expressed in the salivary glands of BPH. The NlSEF1 protein has EF-hand Ca2+-binding activity and can be secreted into rice plants when BPH feed. Infestation of rice by BPH nymphs whose NlSEF1 was knocked down elicited higher levels of Ca2+ and H2O2 but not jasmonic acid, jasmonoyl-isoleucine (JA-Ile) and SA in rice than did infestation by control nymphs; Consistently, wounding plus the recombination protein NlSEF1 suppressed the production of H2O2 in rice. Bioassays revealed that NlSEF1-knockdown BPH nymphs had a higher mortality rate and lower feeding capacity on rice than control nymphs. These results indicate that the salivary protein in BPH, NlSEF1, functions as an effector and plays important roles in interactions between BPH and rice by mediating the plant’s defense responses.

In nature, plants are constantly threatened by herbivorous insects. Consequently, plants have evolved both constitutive and induced defenses that appear after herbivore attack 1 . Inducible defenses begin with the recognition of specific herbivore-associated molecular patterns (HAMPs) and are followed by the elicitation of a complex signaling network, consisting mainly of mitogen-activated protein kinase (MAPK) cascades, and jasmonic acid (JA), salicylic acid (SA), and ethylene (ET) signaling pathways, which subsequently results in the reconfiguration of the transcriptome and proteasome as well as the biosynthesis of defensive chemicals 2,3 . This HAMP-triggered immunity (HTI) is effective with some herbivore populations but not with herbivores that secrete effectors which can effectively suppress the HTI 3 . In that case, plant genotypes that contain resistance genes can recognize these effectors and thereby result in effective and specific effector-triggered immunity 3,4 . Thus, HAMPs and effectors derived from herbivore saliva and egg secretions play important roles in plant-herbivore interactions 3 .
Thus far, the chemicals in many HAMPs from herbivore saliva and egg secretions have been identified, and these include proteins, peptides, lipids and other small molecule compounds, such as fatty acid-amino acid conjugates, β -glucosidase, inceptins and caeliferins 2,4 . Moreover, the mechanisms underlying HAMP-elicited defense responses have also been extensively studied 4,5 . In contrast, herbivore effectors have been less well studied. It has been reported that some salivary proteins from herbivores, such as glucose oxidase from Helicoverpa zea and C002, Mp10, Me10, Me23 from aphids 6,7 , function as effectors of plant defense. These effectors also include salivary calcium binding proteins, which can bind to Ca 2+ and thus suppress a plant's defenses. For example, saliva secreted by the aphid Megoura viciae contains Ca 2+ -binding proteins which can bind to Ca 2+ and then lead to the contraction of forisomes; such contraction prevents the plugging of sieve elements and facilitates an aphid's continuous ingestion from sieve tubes 7,8 . The green rice leafhopper Nephotettix cincticeps has been observed to secrete an 84-kDa salivary EF-hand Ca 2+ -binding protein into sieve tubes during leafhopper feeding; this protein

Results
Isolation and Characterization of NlSEF1. Based on the data from transcriptomes of BPH salivary glands 23 , the open reading frame (ORF) (726 bp) of the gene NlSEF1 was obtained by reverse-transcription-polymerase chain reaction (RT-PCR) (Fig. 1, GenBank: KT698079). Sequence analysis revealed that NlSEF1 encodes a 241-amino-acid protein with a predicted molecular weight of 28.06 kDa and a pI of 5.93. The protein possesses an extracellular signal peptide and has no transmembrane domains, suggesting that NlSEF1 is a putative secreted protein. The mass of the predicted mature protein is 26.04 kDa and the protein has a pI of 5.94. There was no potential O-glycosylation or N-glycosylation site in the protein. The predicted secondary structure of NlSEF1 contained eight coil regions, seven α -helixes and no β -sheet (Supplemental Fig. S1). The PROSITE scan indicated two EF-hand calcium-binding domains (EF_HAND_1, PS00018) in the C-terminus of NlSEF1. Both domains contained a 12-residue loop that begins with aspartic acid and ends with glutamic acid, and each loop was surrounded by two α -helixes (Supplemental Fig. S1), showing a canonical helix-loop-helix EF-hand motif.
Protein alignment revealed that NlSEF1 is homologous to a predicted multiple coagulation factor deficiency protein 2 (MCFD2) that belongs to the EF-hand 7 family (PF13499) and is characterized by a pair of EF-hand domains. In humans, MCFD2 is a soluble luminal protein and a part of a cargo-specific endoplasmic reticulum-to-Golgi transport complex 30 . However, no evidence shows that homologs in insects play similar roles. NlSEF1 has the highest identity to a predicted MCFD2 of Pediculus humanus corporis (XP_002429496.1, 77% identity) but very low query coverage (46%). The best blast hit is a predicted MCFD2 of Solenopsis invicta (XP_011158313.1) when sorted by max or total blast score (96% coverage, 55% identity). Unlike the high similarity of C-terminals, the similarity of N-terminals is low except the 11 amino acid residues after the signal peptide (Supplemental Fig. S2).
NlSEF1 was highly expressed in the salivary glands, but low expression levels (approximately 10-fold lower) were detected in other tissues, including wing, leg, thorax muscle, midgut, cuticle, ovary, and fat body ( Fig. 2A).
Recombinant protein NlSEF1 (with a mass of about 37 kDa) was produced in a bacterial expression system (Fig. 3). The Ca 2+ -binding capability of NlSEF1 was verified by a gel mobility shift assay. Purified NlSEF1 mixed with different concentrations of CaCl 2 was subjected to SDS-PAGE. Compared with the mobility of NlSEF1 in the presence of 0.5 mM EDTA, the mobility of NlSEF1 was slowed by the addition of 0.015 to 0.5 mM CaCl 2 . The migration of NlSEF1 appeared to slow when the concentration of CaCl 2 was high (Fig. 4A), suggesting that NlSEF1 has the Ca 2+ -binding capability.

NlEG1 Is a Secreted Protein that Enters Rice Plants When BPH Feeds. Western blot analysis with
NlSEF1 antibodies was performed to verify whether the protein was secreted into the rice plants. Approximately 200 fifth-instar nymphs of BPH were placed on individual rice plants for 24 h. The total proteins from plant stems were extracted and Western blot analysis was performed using polyclonal anti-NlSEF1 rabbit antibodies. A band of about 26 kDa was detected in extracts from BPH salivary glands ( Fig. 4B; lane 1). The same band was also detected in plants infested by BPH ( Fig. 4B; lanes 2 and 3). In contrast, no NlSEF1 band was detected in plants that had not been exposed to BPH (Fig. 4B; lane 4). These results suggested that NlSEF1 was injected as a salivary component into the rice plants when BPH fed.

NlSEF1
Decreases the Levels of Cytosolic Ca 2+ in Rice. Since NlSEF1 has the Ca 2+ -binding capability, we were interested in whether NlSEF1 could influence the cytosolic Ca 2+ content, [Ca 2+ ] cyt , in rice plants. To investigate this issue, we used RNAi as described in Liu et al. 31 to obtain a BPH population in which NlSEF1 had been silenced (dsSEF-BPH) and analyzed [Ca 2+ ] cyt in leaves of plants that were infested with dsSEF-BPH or BPH that were kept non-manipulated (C-BPH). Injecting BPH with the dsRNA decreased the transcript levels of NlSEF1 in the whole body and salivary gland of the insect by 62-88% over a period of 10 days (Supplemental Fig. S3). Silencing NlSEF1 in BPH did not influence the growth phenotype of individuals (Supplemental Fig. S4). The  [Ca 2+ ] cyt was investigated by confocal laser scanning microscopy using Fluo-3 AM (acetoxy-methyl ester of Fluo-3) as the Ca 2+ -selective fluorescent indicator. Clear fluorescence was observed around BPH feeding sites (Fig. 4C), suggesting that BPH feeding activates cytosolic Ca 2+ transient in rice. Moreover, fluorescence intensity around feeding sites by dsSEF-BPH was significantly higher than that around feeding sites by C-BPH at 1 and 3 h but not 6 h after BPH infestation ( Fig. 4C and D); this demonstrates that the [Ca 2+ ] cyt at dsSEF-BPH feeding sites was higher than those at C-BPH feeding sites. These results suggested that the Ca 2+ -binding protein NlSEF1 from BPH seems to be the determinant of [Ca 2+ ] cyt in plants when they were infested by BPH.

NlSEF1 Suppresses the Production of H 2 O 2 in Rice.
The JA, JA-Ile, SA, and H 2 O 2 signaling pathways are reported to play central roles in plant defense responses in many plant species, including rice [32][33][34][35] . Therefore, we asked whether NlSEF1 influenced the biosynthesis of these phytohormones and signals. Our results revealed that C-BPH nymph infestation did not induce the production of these defense-related phytohormones and signals, except the case of JA whose levels were enhanced 8 h after infestation (Fig. 5A-D), whereas mechanical wounding enhanced levels of these phytohormones and signals (Fig. 5E-H). The levels of JA, JA-Ile and SA were similar among rice plants infested with dsSEF-BPH, BPH that had been injected with the dsRNA of GFP (dsGFP-BPH) and C-BPH ( Fig. 5A-C), whereas H 2 O 2 levels were significantly higher in the dsSEF-BPH-infested plants than in the dsGFP-BPH-and C-BPH-infested plants at 8 h (Fig. 5D). Consistently, the exogenous application of the recombinant protein NlSEF1 on plants, compared to mechanical wounding and the exogenous application of the purified products of the empty vector, had no influence on JA, JA-Ile and SA biosynthesis, but suppressed the production of H 2 O 2 0.5 h after the start of the treatment (Fig. 5E-H). Together, our data demonstrate that the NlSEF1 of BPH is likely to modulate H 2 O 2 levels in rice plants.
Knockdown of NlSEF1 Decreases the Feeding Capacity but Increases Mortality of BPH Nymphs. We investigated whether NlSEF1 could influence the feeding capacity and mortality rate of BPH and whether its effects changed with the rice varieties on which BPH fed. Mortality among BPH fed on rice varied: compared with dsGFP-BPH and C-BPH, dsSEF-BPH had a significantly higher mortality rate over 1 to 12 days, and dsSEF-BPH fed on the variety Mudgo than on the susceptible variety TN1 had higher mortality rates (Fig. 6). Moreover, the amounts of honeydew, an index for food intake, secreted by dsSEF-BPH were significantly lower than those secreted by dsGFP-BPH and C-BPH, whereas there was no significant difference in the amounts of honeydew between the dsGFP-BPH and C-BPH (Fig. 6C). These results suggest that NlSEF1 is essential to the feeding and survival of BPH.

Discussion
So far, several salivary Ca 2+ -binding proteins in insects, such as NcSP84 9 and Armet 36 , have been reported. Some of these proteins can bind Ca 2+ and then decrease host defense by preventing sieve tube plugging 3 . Here we found that NlSEF1 is most highly expressed in BPH salivary glands and its mRNA level remains the same in immature and mature stages of BPH (Fig. 2). Moreover, NlSEF1, containing two EF-hand motifs in its C-terminus, can be injected into rice plants when BPH are feeding (Fig. 4B). This result implies that NlSEF1 may play an important role in rice-BPH interactions. Indeed, the mobility of NlSEF1 on the SDS-PAGE gel was gradually reduced by increasing the concentration of calcium from 0.015 to 5 mM (Fig. 4A). Moreover, rice plants infested by dsSEF-BPH had higher cytoplasmic Ca 2+ contents than did plants infested by C-BPH at 1 and 3 h after infestation (Fig. 4C,D). These data suggest that BPH secrete NlSEF1, a salivary protein which can regulate calcium signaling pathway in rice by binding to cytosolic Ca 2+ . Interestingly, the observed molecular mass (about 37 kDa) of the purified recombinant NlSEF1, using the E. coli expression system, was about 11 kDa bigger than the predicted mass of mature NlSEF1, 26.04 kDa. The mass may be explained by the 51 additional amino acids from the expression vector (causing the mass of the recombinant protein to increase to 31.63 kDa) and strong positive charge of the His-tag. Some His-tag fusion proteins have been reported to show higher apparent molecular weight in SDS-PAGE 37,38 .
We investigated the influence of NlSEF1 on defense-related signaling pathways in rice and found that rice plants infested by dsSEF-BPH had higher H 2 O 2 levels than those infested by dsGFP-BPH or C-BPH (Fig. 5D). Consistently, the exogenous application of recombinant NlSEF1 on rice plants decreased the levels of H 2 O 2 (Fig. 5H). These findings suggest that NlSEF1 secreted by BPH saliva can modulate the H 2 O 2 -mediated signaling pathways in rice plants. Ca 2+ transients in plants elicited by herbivore infestation are well known to directly or indirectly regulate diverse signaling pathways, such as those mediated by JA, SA, reactive oxygen species (ROS) and NO, which subsequently regulate plant defenses [39][40][41] . AtSR1, a CaM-binding protein, for example, has been reported to negatively regulate the production of SA by suppressing the expression of EDS1, a positive modulator of SA biosynthesis 42 . In Solanum tuberosum, CDPK4 and CDPK5 have been reported to regulate the pathogen-induced ROS production 43 . Moreover, in Arabidopsis, CDPK4, CDPK5, CDPK6 and CDPK11 were found to be involved in ROS production 44 . In rice, BPH infestation has been observed to cause a Ca 2+ flux and induce the accumulation of JA, SA, ethylene and H 2 O 2 33,34,45,46 . Therefore, the regulatory effect of NlSEF1 on the level of H 2 O 2 in rice (Fig. 5) is probably due to its Ca 2+ -binding activity. Further researches should elucidate how NlSEF1-mediated the change in Ca 2+ levels regulates the biosynthesis of H 2 O 2 and why this change does not influence the production of other defense-related signals.
Bioassays showed that the knockdown of NlSEF1 significantly reduced the amount of honeydew secreted by female BPH adults (Fig. 6C). Moreover, when BPH fed on the resistance variety Mudgo, the mortality rate of dsSEF-BPH was higher than that of dsGFP-BPH, whereas, when fed the susceptible variety TN1, the mortality rate of dsSEF-BPH did not differ from that of dsGFP-BPH until the last three days of the experiment (Fig. 6). Our data suggest that NlSEF1 mediates the mortality rate and feeding of BPH not by influencing the insect but by suppressing defense in the plant, suggesting that the protein plays an effector role in rice defense responses. It has been reported that H 2 O 2 signaling pathways positively regulate resistance in rice to BPH, whereas the JA signaling pathway negatively mediates BPH resistance 32,35 . Our experiments showed that rice plants infested by BPH with the knockdown of NlSEF1 had higher H 2 O 2 level than those infested by dsGFP-BPH or C-BPH. Thus, the lower performance of dsSEF-BPH on rice plants, compared with the performance of dsGFP-BPH and C-BPH, might be related to the low activity of NlSEF1 in dsSEF-BPH; low activity of NlSEF1 resulted in a high level of H 2 O 2 in rice plants and thus enhanced the resistance of rice to BPH. Phloem plugging, including callose deposition, which can be induced by Ca 2+ , was also reported to be one of the mechanisms that rice plants use to defend themselves against BPH 47 . Given that rice plants infested by dsSEF-BPH had higher cytoplasmic Ca 2+ levels than did plants infested by C-BPH (Fig. 4C), it is reasonable to think that phloem plugging might also cause dsSEF-BPH to perform poorly. Further research should elucidate this question.
In summary, NlSEF1 is a salivary EF-hand Ca 2+ -binding protein that can be secreted into rice plants when BPH feed. NlSEF1 binds to Ca 2+ and then causes a decrease in the content of cytosolic Ca 2+ in rice; this decrease subsequently reduces H 2 O 2 levels and possible phloem plugging, which finally impairs resistance in rice to BPH. These findings suggest that the secreted salivary protein NlSEF1 functions as an effector, suppressing defense response in rice plants and allowing BPH to continuously suck the juice from rice phloem.

Tissue-specific and Developmental Stage Expression Patterns of NlSEF1.
Total RNA was extracted from the following materials: (1) eight different BPH tissue samples (salivary gland, wing, leg, thorax muscle, midgut, cuticle, ovary, and fat body) isolated from newly emerged brachypterous female adults; (2) whole bodies of BPH at different developmental stages (from first-to fifth-instar nymphs, newly emerged to 12-d-old brachypterous females). Total RNA was isolated using the SV Total RNA Isolation System (Promega) according to the manufacturer's instructions. Each RNA sample was reverse-transcribed into cDNA using the Takara Primescript ™ RT reagent kit. The QPCR assay was performed on CFX96 TM Real-Time system (Bio-Rad, Hercules, CA, USA) using iQ SYBRGreen Supermix (Bio-Rad). A BPH 18S rRNA gene (GenBank accession no. JN662398) was used as an internal standard to normalize cDNA concentrations. The primers used for qRT-PCR are listed in Supplemental Table S1. Three independent biological replicates were analyzed in each experiment.
Expression of NlSEF1 in Escherichia coli. The ORF of NlSEF1 was amplified by PCR using the primers listed in Supplemental Table S1. The PCR product was ligated into the pMD19-T vector, sequenced, and then cloned into vector pET-30a. The recombinant vector NlSEF1:pET-30a and empty vector pET-30a (used as a control) were transformed into E. coli BL21 (DE3) strain. Expression was induced by adding IPTG (1 mM final concentration). The products from recombinant and empty vector were purified by using Ni-NTA columns (Qiagen, Venlo, Netherlands) according to the manufacturer's instructions. The purified products were concentrated with a YM-10 Centricon membrane (Millipore, Billerica, MA, USA) to remove imidazole. The final purified concentrated products from E. coli cells with the empty vector and recombinant vector were mixed with 2× SDS loading buffer, respectively, separated by SDS/PAGE in a 12.5% acrylamide gel (ISC Bioexpress, Kaysville, UT, USA), and stained with 0.025% Coomassie blue R-250 in water. The predicted mass of the mature recombinant protein NlSEF1 including 6 N-terminal His-tags is 31.63 kDa.

Polyclonal Antibody Preparation and Western Blot Analysis. A polypeptide (QQKAAKPPPPQQHS)
of NlSEF1 was selected as the antigen, and the polyclonal rabbit antibodies of NlSEF1 were made and purified by GenScript TM . The following protein samples used for Western blot analysis were prepared: (1) Proteins extracted from the salivary glands of BPH. The salivary glands of 200 newly emerged adult females of BPH were dissected and homogenized separately in 1 mL PBS. The extract was centrifuged at 12,000 × g for 5 min at 4 °C, and the supernatant used as a sample. (2) Proteins from rice leaf sheaths infested by BPH or not. Rice stems of Mudgo were individually confined within glass cylinders (diameter 4 cm, height 8 cm, with 48 small holes, diameter 0.8 mm) in which approximately 200 fifth-instar nymphs of BPH were released and, after 24 h, removed. Plants in empty glass cylinders were used as controls. The outer three leaf sheaths from three rice stems (about 0.9 g) from plants that had undergone the same treatment were harvested and merged, then ground in liquid nitrogen, homogenized in 4 ml PBS, and centrifuged at 12,000 × g for 5 min at 4 °C, after which the supernatant was concentrated to 200 μ L using a YM-10 Microcon centrifugal filter device (Millipore). SDS-PAGE sample buffer (2× ) was added to this extract, which was then subjected to SDS-PAGE on 12% gradient gels (ISC Bioexpress) and transferred onto a BioTrace TM pure nitrocellulose blotting membrane (PALL Life Sciences). Nonspecific binding sites were blocked with 5% instant nonfat dry milk (Yili company, Hohehot municipality, Huhehaote, China), and the membrane was incubated with purified polyclonal antibody (1:200 dilution) overnight in 4 °C followed by extensive washing for 30 min with frequent changes of TBST. The antigen-antibody complexes were visualized with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:5000 dilution; Multisciences, Hangzhou, China) at 37 °C for 1 h followed by extensive washing for 20 mins with frequent changes of TBST and detected with ECL substrates A and B (Multisciences) by FluorChem FC2 (Alpha Innotech).  Table S1). The PCR products were used to synthesize dsRNA by using MEGAscript T7 High Yield Transcription Kit (Ambion, Austin, TX, USA). Thirdor fifth-instar nymphs were injected at the conjunction of the prothorax and mesothorax 31 using the FemtoJet (Eppendorf, Hamburg, Germany) microinjection device. Each nymph was injected with about 0.25 μ g dsRNA of NlSEF1 or GFP, or was not injected (control). The levels of NlSEF1 transcripts in the whole body (third-instar nymphs were injected) and in the salivary gland (fifth-instar nymphs were injected) of the insect that had been injected with NlSEF1 or GFP dsRNA, or kept non-injected (C-BPH), were investigated at 2, 4, 6, 8 and 10 days after injection.

BPH Bioassays.
To measure the effect of the knockdown of NlSEF1 on mortality rates of BPH, third-instar BPH nymphs injected with NlSEF1 or GFP dsRNA, or kept non-injected (C-BPH), were allowed to feed on Mudgo or TN1. The treated BPH were first allowed to recover on rice seedlings at 27 ± 1 °C with 70 ± 10% RH and 14:10 h (light/dark) photoperiod for 1 day, then the healthy ones were moved onto the tillering stage of Mudgo or TN1 rice plant for the bioassay. Stems of rice plants (one plant per pot) were individually confined within glass cylinders as stated above into which 20 third-instar nymphs were released. Each treatment had 5 replicates. The number of dead BPH nymphs in each cylinder was recorded every day.
The effect of NlSEF1-knockdown on BPH feeding was also investigated. A newly emerged brachypterous female adult at 3 days after the injection of NlSEF1 or GFP dsRNA into fifth-instar nymphs, or no injection, was placed into a small parafilm bag (6 × 5 cm), which was then fixed onto the stem of rice variety Mudgo. The amount of excreted honeydew was weighed (to an accuracy of 0.1 mg) 24 h after the start of the experiment. Each treatment was replicated 40 times.
Intracellular Calcium Variation Determination. The intracellular calcium variation of Mudgo plants was determined by using Fluo-3 AM (acetoxy-methyl ester of Fluo-3) as the Ca 2+ -sensitive fluorescent indicator following the method described in Maffei et al. 48 . Briefly, Fluo-3 AM (stock solution in dimethyl sulfoxide; Molecular Probes, Eugene, OR, USA) was diluted in 50 mM MES buffer (pH 6.0) containing 0.5 mM calcium sulfate and 2.5 μ M 3-(3,4-dichlorophenyl)-1,1-dimethylurea (Sigma-Aldrich, Steinheim, Germany) to a final concentration of 5 μ M. Rice leaves of Mudgo were individually confined within glass cylinders (diameter 3 cm, height 1.5 cm, with 10 small holes, diameter 0.8 mm) into which 15 newly emerged female adults of dsSEF-BPH (2 d after injection) or C-BPH were released. Infested portions of leaves were individually harvested 1, 3 and 6 h after infestation and were immediately incubated in 500 μ L of 5 μ M Fluo-3 AM solution as described above. Thirty minutes after incubation, leaves were mounted on a Zeiss LSM 780 confocal laser scanning microscope and were observed at 488 nm excitation wavelength. Images generated by the Zen 2010 software were analyzed by using the ImageJ software (https://imagej.nih.gov/ij). The experiment was replicated three times and the measurement for fluorescence intensity at BPH feeding sites was repeated at least 10 times.
JA, JA-Ile, SA and Hydrogen Peroxide Analysis. Potted plants (one per pot) of Mudgo were randomly assigned to the following treatments: (1) infestation by different BPH nymph groups. Plant stems were individually confined in glass cylinders into which 20 (dsSEF-BPH) or 15 (dsGFP-BPH and C-BPH) fifth-instar nymphs (1 day after injection) were released, according to the variation of feeding ability between BPH in response to different treatments. (2) NlSEF1 treatment. Plant stems (lower part, about 2 cm long) were individually pierced 200 times with a fine needle and treated with 20 μ L of either the recombinant protein NlSEF1 (0.15 μ g/μ L) or the purified products of the empty vector (EV), or kept non-manipulated (W). The outer three leaf sheaths of stems were harvested at different time points (Fig. 5) after the start of the treatment. JA, JA-Ile and SA levels were analyzed by high performance liquid chromatography-mass spectroscopy using labeled internal standards ( 2 D 4 -SA, 2 D 6 -JA and 2 D 6 -JA-Ile) following the method as described in Lu et al. 49 . The H 2 O 2 concentrations in the sheath were determined using Amplex-Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen, Carlsbad, CA, USA) as described previously by Lou and Baldwin 50 . Each treatment and time point was replicated five times. Data Analysis. Differences in experiments involving three treatments were analyzed by one-way ANOVAs.
If the ANOVA was significant (P < 0.05), Duncan's multiple range tests were used to detect significant differences between treatments. All tests were carried out with Statistica (SAS Institute, Inc., http://www.sas.com/).