Alternative Oxidase in Resistance to Biotic Stresses: Nicotiana attenuata AOX Contributes to Resistance to a Pathogen and a Piercing-Sucking Insect But Not Manduca sexta Larvae^,,

Abstract

The role of the alternative respiratory pathway in the protection of plants against biotic stress was examined in transgenic tobacco (Nicotiana attenuata) plants (irAOX) silenced in the expression of ALTERNATIVE OXIDASE (AOX) gene. Wild-type and irAOX plants were independently challenged with (1) chewing herbivores (Manduca sexta), (2) piercing-sucking insects (Empoasca spp.), and (3) bacterial pathogens (Pseudomonas syringae pv tomato DC3000), showing that all these treatments can strongly elicit accumulation of AOX gene transcripts in wild-type plants. When N. attenuata chemical defenses and resistance were examined, irAOX plants showed wild-type levels of defense-related phytohormones, secondary metabolites, and resistance to M. sexta. In contrast, piercing-sucking leafhoppers (Empoasca spp.) caused more leaf damage and induced significantly higher salicylic acid levels in irAOX compared with wild-type plants in the field and/or glasshouse. Subsequently, irAOX plants accumulated lower levels of defense metabolites, 17-hydroxygeranyllinalool diterpene glycosides, caffeoylputrescine, and nicotine compared with wild-type plants under prolonged attack of Empoasca spp. in the glasshouse. Finally, an accelerated cell death phenotype was observed in irAOX plants infected with P. syringae, which correlated with higher levels of salicylic acid and hydrogen peroxide levels in pathogen-infected irAOX compared with wild-type leaves. Overall, the AOX-associated changes in phytohormone and/or redox levels appear to support the resistance of N. attenuata plants against cell piercing-sucking insects and modulate the progression of cell death in pathogen-infected tissues but are not effective against rapidly feeding specialist herbivore M. sexta.

Metabolic plasticity allows plants to adapt to variable environmental stress conditions. Plants rapidly reconfigure their primary metabolism during stress to cope with the increased metabolic demands of resistance responses, and mitochondria are known to play a central role in this reconfiguration (for review, see Bolton, 2009; Millar et al., 2011). Under stress, electrons are frequently rerouted through an alternative respiratory pathway branching from the cytochrome (Cyt) pathway at the level of ubiquinone (UQ) pool to an enzyme known as the alternative oxidase (AOX; McDonald, 2008). In contrast with the Cyt pathway, the AOX pathway is resistant to cyanide and therefore known as cyanide-resistant respiration (Henry and Nyns, 1975). The alternative pathway bypasses two important energy conservation sites (complex III and IV) and avoids oxidative phosphorylation and ATP synthesis: energy is dissipated as heat after reduction of oxygen to water by AOX enzyme (Moore et al., 1978). Whereas alternative respiration is widespread in plants, fungi, and some prokaryotes (McDonald, 2008), its physiological function and seemingly energy-wasting character are not well understood.

One of the most-investigated functions of AOX is heat production, which can volatilize odoriferous compounds during flowering for the attraction of pollinators in Arum lily (Sauromatum guttatum) flowers, for example (Meeuse, 1975; Raskin et al., 1987). Heat production is only associated with a limited number of thermogenic plant species (Seymour, 1997), and alternative functions of AOX have been extensively investigated. Subsequently, AOX has been established as one of the essential defense components in plant response to acute stress (Arnholdt-Schmitt et al., 2006). Abiotic stresses such as drought (Bartoli et al., 2005), high salt (Costa et al., 2007; Feng et al., 2010a), low temperature (Vanlerberghe and McIntosh, 1992; Popov et al., 2011; Wang et al., 2011), and wounding (Hiser and McIntosh, 1990) are known to stimulate the activity of alternative respiratory pathways or at least increase AOX transcripts and/or protein levels. Phytochrome, phototropin, and cryptochrome photoreceptors have been shown to mediate the light-responsiveness of the AOX1a gene in Arabidopsis (Arabidopsis thaliana) seedlings, and aox1a mutants showed increased photobleaching damage under high light conditions (Zhang et al., 2010). Several studies have indicated that AOX may also function in resistance to biotic stress such as infection by pathogens. For example, AOX protein levels increased in local and systemic Nicotiana tabacum, and in tomato (Solanum lycopersicum) leaves infected with tobacco mosaic virus (TMV; Lennon et al., 1997; Liao et al., 2012). The overexpression of AOX protein resulted in smaller hypersensitive response lesions, suggesting that AOX may act as a suppressor of programmed cell death (PCD) in virus-infected leaves (Ordog et al., 2002). The induction of AOX in leafy mustard (Brassica juncea) during infection with turnip mosaic virus was suppressed in the presence of ethylene biosynthesis inhibitor aminoethoxyvinylglycine, suggesting that ethylene and AOX are likely involved in leafy mustard’s systemic resistance to virus infection (Zhu et al., 2011).

Bacterial pathogens and their elicitors also strongly elicit AOX transcript accumulations and protein levels, both in leaves and cell suspensions (Krause and Durner, 2004; Mizuno et al., 2005; Kiba et al., 2008). In Arabidopsis plants, infection with virulent and avirulent strains of Pseudomonas syringae pv tomato (Pst) elevated AOX transcript and proteins levels in infected leaves (Simons et al., 1999). An accurate oxygen isotope discrimination method was used to determine the partitioning of electrons between Cyt and the alternative pathways in pathogen-infected plants and documented a 5-fold increase in in vivo AOX activity 4 h after application of the bacterial elicitor harpin N_Ea to Nicotiana sylvestris leaves (Vidal et al., 2007). In Paracoccidioides brasiliensis, a human thermal dimorphic pathogenic fungus, PbAOX played an important role in fungal defense response against oxidative stress imposed by immune cells and virulence of the pathogen (Ruiz et al., 2011).

Therefore, control of oxidative damage is essential for living organisms, including plants, to survive under and recover from stresses (Hanqing et al., 2010). Here, AOX is very likely to counteract the overreduction of mitochondrial UQ pools by consuming excess reducing equivalents, particularly when the flow of electrons through the Cyt pathway is impaired by acute stress. Because insects and microbial attacks are both knowingly associated with oxidative bursts (Torres et al., 2002; De Vos et al., 2007; Maffei et al., 2007), AOX and its redox-controlling function could contribute to a plant’s resistance to herbivores and pathogens. To test this expectation, we transformed Nicotiana attenuata plants with an RNA interference (RNAi) construct harboring a fragment of NaAOX gene in an inverted repeat orientation (irAOX) to silence the expression of endogenous NaAOX genes and selected three mechanistically distinct biotic stress models to thoroughly examine the role of AOX in plant defense against biotic stresses. Whereas NaAOX transcript and protein levels significantly increased in response to all three stress conditions, AOX played different roles in response of plants to these different biotic stresses.

RESULTS

Identification of AOX Gene Sequences in N. attenuata

Two AOX genes, NaAOX1 and NaAOX2, from N. attenuata of nearly identical nucleotide sequence, have been previously deposited in public National Center for Biotechnology Information databases (AY422688, AY422689). Both genes show a high level of identity to members of the Arabidopsis AOX1 gene family: NaAOX1 and NaAOX2 proteins are 70.7% and 70.3% identical to AtAOX1a from Arabidopsis (Supplemental Fig. S1A). The high degree of identity between NaAOX1 and NaAOX2 genes at the nucleotide level did not allow for efficient monitoring and gene-specific silencing of individual NaAOX isoforms by quantitative reverse transcription PCR (qRT-PCR) and RNAi, respectively. Therefore, primers detecting transcripts of both genes were used and reported herein as NaAOX transcript abundance. Subsequently, both genes were targeted with a single silencing construct to down-regulate AOX expression. Despite our attempts to identify additional AOX genes from N. attenuata, we were not able to find any additional conserved AOX proteins in this diploid Nicotiana species.

Analysis of NaAOX Transcript Levels in Response to an Attack from a Chewing Herbivore

Abiotic stress and pathogen elicitors are known to increase transcript levels of AOX genes in plants. However, much less is known about the effects of herbivory. We examined NaAOX transcript levels in wild-type leaves subjected to wounding or simulated herbivory treatment, and in N. attenuata leaves after direct feeding to Manduca sexta neonates. The NaAOX transcripts strongly increased 1 h after wounding the N. attenuata leaves with a pattern wheel and immediately applying either water to mechanical wounds (W+W) or M. sexta’s oral secretions (OS) to simulate herbivory (W+OS; Fig. 1A; Supplemental Fig. S1B). It is interesting that NaAOX transcript and protein levels appeared higher in W+OS- compared with W+W-treated leaves at 1 to 3 h (Fig. 1A; Supplemental Fig. S2A). Similarly, direct M. sexta feeding significantly elevated basal NaAOX transcript and protein levels at 1 and 2 d after placing neonates on the leaves (Fig. 1B; Supplemental Fig. S2B). Because NaAOX transcript levels increased significantly more in response to M. sexta’s salivary elicitors, we hypothesized that AOX may function in direct defense of N. attenuata against chewing herbivores. In addition to the herbivore- and wound-regulated patterns shown in Figure 1, an apparent oscillation in NaAOX transcript abundance was observed in untreated control leaves, suggesting a possible circadian control of AOX transcription (Fig. 1A). In particular, AOX transcripts appeared to be higher during the light periods than during dark periods, which could be associated with a previously reported AOX function in buffering high photosynthetic and respiration rates in response to high light (Dinakar et al., 2010).

Figure 1.

NaAOX gene transcripts increase in response to simulated herbivory and feeding of M. sexta caterpillars in N. attenuata. A, Technical replicate means (± se) of NaAOX transcript relative abundances quantified with qRT-PCR in pooled samples using five independent controls, W+W- and W+OS-elicited N. attenuata plants. The leaves were elicited with a pattern wheel (wounding) and immediately applying either water to mechanical wounds (W+W) or M. sexta’s OS to simulate herbivory (W+OS). Gray areas indicate dark periods showing low NaAOX transcript levels. B, Means (± se) of NaAOX transcripts quantified with qRT-PCR from control and M. sexta-attacked rosette leaves at 0, 1, 2, and 3 d after placing three M. sexta neonates on each leaf. Asterisk indicates significant differences between control and insect-fed plants at respective time points (unpaired Student’s t test, *P < 0.05; n = 3).

NaAOX Silencing Does Not Compromise Defense of N. attenuata against M. sexta Caterpillars

To examine the function of NaAOX in resistance of N. attenuata against its specialist herbivore M. sexta caterpillars, we generated stably AOX-silenced N. attenuata plants (irAOX) using RNAi and inverted repeat fragment of the NaAOX1 gene (Supplemental Fig. S3A). Two independent homozygous diploid lines (referred to as irAOX-200 and irAOX-203) carrying a single insertion of transfer DNA were selected for further experiments (Supplemental Fig. S3B). In both lines, the levels of NaAOX transcripts were strongly reduced compared with identically treated wild-type plants (Supplemental Fig. S3C). Alternative respiration capacity of the leaves was also significantly suppressed in irAOX plants compared with wild-type plants (Supplemental Fig. S4).

Feeding of M. sexta larvae elicits burst of jasmonic acid (JA) and JA-Ile in N. attenuata leaves, which triggers defense signaling pathways and deploys defense responses (Wu and Baldwin, 2010). To examine if NaAOX mediates changes in phytohormone levels, we determined JA and JA-Ile contents in irAOX and wild-type plants at 0, 45, 90, and 180 min after W+W- and W+OS-elicitation. Similar analysis was conducted using M. sexta-attacked leaves before and 1, 2, and 3 d after placing neonates on the leaves. Whereas both treatments elicited significantly higher levels of JA and JA-Ile in the leaves compared with untreated controls, we did not find any significant differences in JA or JA-Ile content between wild-type and irAOX plants (Fig. 2A). The levels of another stress hormone, salicylic acid (SA) remained comparable between wild-type and irAOX plants (Supplemental Fig. S5, A and B) and staining of W+OS and control leaves with 3,3'-diaminobenzidine (DAB) showed comparable levels of hydrogen peroxide (H₂O₂) accumulation in wild-type and irAOX leaves (Supplemental Fig. S5C). Because important defense metabolites such as nicotine (Steppuhn et al., 2004) and 17-hydroxygeranyllinalool diterpene glycosides (HGL-DTGs; Jassbi et al., 2008; Heiling et al., 2010) were not differentially regulated in irAOX compared with wild-type plants (Supplemental Fig. S6), it was not surprising that caterpillar performance in feeding bioassays did not differ (Fig. 2B). These results suggest that despite increased transcript levels of NaAOX in response to M. sexta attack, NaAOX genes do not play any significant role in the direct defense of N. attenuata plants against chewing specialist herbivores. The induction of NaAOX transcripts and protein levels by insect feeding could be a consequence of/and defense against mechanical wounding, similar to previously described accumulation of AOX protein in mechanically wounded potato (Solanum tuberosum) tubers (Hiser and McIntosh, 1990).

Figure 2.

irAOX OS- and herbivory-induced plants have wild-type JA levels and AOX silencing does not affect M. sexta caterpillar performance. A, Mean (± se) levels of JA and JA-Ile in samples collected at 0, 0.75, 1.5, and 3 h after wounding and immediately applying OS from M. sexta (W+OS; n = 5, left), and mean (± se) levels of JA and JA-Ile in samples collected after 0, 1, 2, and 3 d of M. sexta feeding (n = 3, right). B, Mean (± se) mass of M. sexta caterpillars determined after neonates placed on 30 independent wild-type and irAOX plants were allowed to feed for 12 d. No statistically significant differences in caterpillar mass were detected between wild-type and NaAOX gene-silenced lines.

We also considered an alternative hypothesis that increased AOX expression during herbivory could facilitate higher emissions of volatile organic compounds, mediators of indirect defenses, from herbivore-attacked leaf surfaces of N. attenuata. However, in our previous experiments, we did not find any consistent differences in α-bergamotene or benzyl acetone emissions from wild-type and irAOX plants, even when the plants were maintained at less than ambient temperatures (data not shown).

NaAOX Silencing Makes Plants More Susceptible to Piercing-Sucking Empoasca spp.

In the natural environment, plants are exposed to a large variety of pests and pathogens. When irAOX-203 plants were transplanted as size-matched pairs with N. attenuata wild-type plants into their native habitat in the Great Basin Desert of southwestern Utah, we noticed that leafhoppers (Empoasca spp.) caused significantly more damage on irAOX-203 compared with wild-type plants (Fig. 3A). We also observed mirid (Tupiocoris notatus)- and flea beetle (Epitrix spp.)-associated damage on the plants in the field; in contrast to Empoasca spp. damage, there were no significant differences in damage caused by these insects (data not shown).

Figure 3.

irAOX plants are more susceptible to Empoasca spp. damage in the glasshouse and natural environment. A, Mean (± se) average leaf damage by Empoasca spp. in wild-type and irAOX plants planted in a pairwise design in the natural environment of N. attenuata in the Great Basin Desert (n = 25). B, Leaves from wild-type and irAOX plants that were placed for 12 d in a closed cage in the glasshouse with Empoasca spp. (left), and mean (± se) leaf damage by Empoasca spp. feeding estimated at the end of the experiment (n = 10, right). Asterisk indicates significant differences between wild-type and irAOX plants at respective time points determined by unpaired Student’s t test (*P < 0.05; ***P < 0.001). Inset in A, Empoasca spp. photographed in the field.

The Empoasca spp. are cell rupture feeders that are often associated with microbes transmitted during herbivore feeding (Backus et al., 2005). Although Candidatus Phytoplasma spp. have not been identified in Empoasca spp. occurring in the Utah field or in our in-house colony derived from Utah-collected insects (Kallenbach et al., 2012), it suggests that AOX is important for protection of plants against sucking-piercing insects and/or other microbes, such as viruses and bacteria, vectored by these insects. To examine the robustness of our initial field observations, a choice-test experiment with irAOX and wild-type plants was conducted in the glasshouse. An equal number of wild-type and irAOX-203 plants were placed in a randomized design in the sealed glass container with Empoasca spp. that were allowed to feed on the plants for 12 d. Consistent with our field observations, irAOX plants were significantly more damaged compared with wild-type plants (Fig. 3B).

The NaAOX transcript and protein levels were determined before and after 4, 8, and 12 d of exposure of wild-type plants to Empoasca spp. feeding (Fig. 4; Supplemental Fig. S2C). Empoasca spp. feeding induced NaAOX transcripts that became 6-fold more abundant compared with control plants after 12 d of exposure (Fig. 4). The consistently higher damage on attacked irAOX plants (Fig. 3B) and inducible character of NaAOX gene shown in Figure 4 prompted a more detailed analysis of N. attenuata phytohormone and defense metabolite profiles after Empoasca spp. attack.

Figure 4.

Leaf NaAOX gene transcript levels increase in response to Empoasca spp. attack. Means (± se) of NaAOX relative transcript abundances quantified with qRT-PCR in control and Empoasca spp.-attacked wild-type leaves at 0, 4, 8, and 12 d. Asterisk indicates significant differences between control and Empoasca spp.-damaged plants at respective time points (unpaired Student’s t test, **P < 0.01, ***P < 0.001; n = 6).

irAOX Plants Contain Higher Levels of SA after Empoasca spp. Attack

Piercing-sucking insects such as aphids, mites, and leafhoppers can activate both salicylate- and JA-dependent defense signaling pathways in plants (Ament et al., 2004; Mozoruk et al., 2006; Kempema et al., 2007). The lack of AOX activity in irAOX N. attenuata plants clearly disturbed the homeostasis of SA after Empoasca spp. attack. Although SA levels did not differ in the control nonattacked wild-type and irAOX-203 set of plants, SA levels increased in response to Empoasca spp. attack in irAOX, but remained unchanged in wild-type plants (Fig. 5A). In contrast, JA and JA-Ile levels were not differentially elicited in wild-type and irAOX plants (Fig. 5B). It is likely that JA levels after Empoasca spp. attack increased only transiently and/or locally, which was not captured in our experimental setup (first measurement after 4 d). Recently, a significant increase in JA levels was reported in the leaves of N. attenuata challenged with 25 Empoasca spp. individuals allowed to feed on the leaf for 2 d (Kallenbach et al., 2012). In addition, reduced resistance to Empoasca spp. was found in the field and glasshouse experiments with JA biosynthesis-deficient N. attenuata plants (Kessler et al., 2004; Kallenbach et al., 2012). In addition, irLOX3 plants silenced in the accumulation of JA were significantly more damaged by Empoasca spp. compared with wild-type plants in our glasshouse choice test experiments (Supplemental Fig. S7), emphasizing a central role for JA in Empoasca spp. feeding behavior.

Figure 5.

Long-term Empoasca spp. feeding results in differential SA but not JA levels in irAOX plants. A, Mean (± se) levels of SA. B, JA and JA-Ile in samples collected after 0, 4, 8, and 12 d of Empoasca spp. feeding. Asterisk indicates significant differences between wild-type and irAOX plants at respective time points determined by unpaired Student’s t test (*P < 0.05, **P < 0.01; n = 6).

Empoasca spp.-Attacked irAOX Plants Have Attenuated Defense Metabolite Accumulations after Prolonged Feeding

Plants use a large variety of secondary metabolites to defend against herbivores and pathogens (Baldwin 2001; Metlen et al., 2009). Nicotine, caffeoylputrescine (CP), and HGL-DTGs are typical secondary metabolites found in N. attenuata plants elicited with biotic stresses (Steppuhn et al., 2004; Heiling et al., 2010; Kaur et al., 2010). Whereas nicotine and HGL-DTGs are known to strongly inhibit M. sexta larvae feeding, hydroxycinnamic acid-polyamine conjugates like CP can function against both herbivores and microbes (for review, see Bassard et al., 2010). In our experiments, nicotine (P < 0.001), CP (P < 0.05), and HGL-DTGs (P < 0.001) accumulated significantly less in irAOX compared with wild-type plants exposed for 8 to 12 d to Empoasca spp. feeding (Fig. 6), suggesting a positive role of AOX in efficient accumulation of these metabolites after prolonged Empoasca spp. feeding. The initial feeding of Empoasca spp., however, is known to be independent of nicotine, HGL-DTGs, and pathogen infections in N. attenuata plants (Kallenbach et al., 2012). Consistently, after 4 d of Empoasca spp. feeding, defense metabolites were not yet elicited (Fig. 6).

Figure 6.

Secondary metabolite levels significantly differ in Empoasca spp.-attacked wild-type and irAOX leaves. Mean (± se) levels of defensive secondary metabolites in wild-type and irAOX-203 leaf samples collected after 0, 4, 8, and 12 d of exposure to Empoasca spp. Asterisk indicates significant differences between wild-type and irAOX plants at respective time points determined by unpaired Student’s t test (*P < 0.05, **P < 0.01, *** P < 0.001; n = 6).

NaAOX Transcripts Increase in Response to Pathogen Infection

The treatment of N. sylvestris cells with pathogen-associated elicitors is known to alter partitioning between Cyt- and AOX-dependent respiratory pathways, and transiently increase AOX expression (Vidal et al., 2007). We examined if Pst DC3000 strain infection can induce AOX expression in N. attenuata; Pst DC3000 causes local chlorosis followed by leaf necrosis in N. attenuata (Hettenhausen et al., 2012). The leaves of rosette stage plants were infiltrated with 1 × 10⁵ colony forming units suspension of Pst DC3000, and samples were collected from infiltrated leaf areas before and 45 and 90 min and 3, 24, 48, and 72 h after pathogen inoculation. Both mock- and pathogen-inoculated leaves experienced a transient burst of NaAOX transcripts at 45 min post inoculation (Fig. 7), which was likely caused by wounding during infiltration process. However, transcript and protein levels of NaAOX dramatically up-regulated in pathogen-inoculated compared with mock-treated leaves at the later stages of infection (Fig. 7; Supplemental Fig. S2D), suggesting a direct involvement of AOX in plant defense against pathogen-imposed stress. The levels of NaAOX protein in Pst-infected leaves of irAOX-200 and irAOX-203 transgenic lines remained low at 2 and 3 d (Supplemental Fig. S2D), as expected from AOX-silencing efficiency shown in Supplemental Fig. S3C.

Figure 7.

NaAOX gene transcripts increase in response to infection with Pst DC3000 in N. attenuata leaves. Mean (± se) NaAOX relative transcript abundances quantified with qRT-PCR in three independent mock- and Pst DC3000-inoculated wild-type plants at indicated time points. Asterisk indicates significant differences between mock- and Pst DC3000-infected leaves at respective time points determined by unpaired Student’s t test (*P < 0.05, **P < 0.01, *** P < 0.001; n = 3).

NaAOX Silencing Accelerates Pst DC3000-Induced Cell Death

In the following experiment, we observed the progress of necrosis in wild-type and irAOX leaves inoculated with Pst DC3000. Both irAOX-200 and irAOX-203 lines exhibited moderately but reproducibly accelerated development of necrotic symptoms, which typically occurred 1 d earlier in irAOX compared with wild-type plants (Fig. 8A, left). Microscopic examination of trypan blue-stained leaves confirmed more extensive cell death in irAOX leaves 2 d after inoculation with the pathogen compared with identically treated wild-type leaves (Fig. 8A, right). When the extent of cell death was estimated by quantitative electrolyte leakage of the cells (Pike et al., 1998), the levels of leakage were significantly higher in both irAOX lines compared with wild-type plants at 24, 48, and 72 h after Pst DC3000 infiltration (Fig. 8B).

Figure 8.

NaAOX gene silencing accelerates cell death after infection of plants with Pst DC3000 pathogen. A, Pst DC3000-infected leaves from wild-type and two irAOX independent lines (200 and 203) at 1, 2, and 3 d after inoculation (left), and trypan blue-stained leaves at the corresponding time points (right). B, Mean (± se) electrolyte leakage from mock and Pst-infected leaves in wild-type and two irAOX lines determined at 1, 2, and 3 d after inoculation. Significant differences were determined for each time point by one-way ANOVA (*P < 0.05, **P < 0.01, *** P < 0.001; n = 5). C, Mean (± se) relative transcript levels of cell death marker gene NaHIN1 in mock and Pst DC3000-inoculated leaves of wild-type and two irAOX lines. Significant differences at respective time points were determined by one-way ANOVA (*P < 0.05, **P < 0.01; n = 5).

To characterize Pst-induced cell death at a molecular level, the expression of a N. attenuata hypersensitive response marker HAIRPIN-INDUCED1 (HIN1) gene was examined. HIN1 is known to be highly expressed during incompatible plant-pathogen interactions in N. tabacum (Takahashi et al., 2004a, 2004b). NaHIN1 transcripts were significantly induced by Pst DC3000 in N. attenuata, and they were significantly higher in two irAOX lines compared with wild-type plants 3 d after infection (Fig. 8C). Accelerated cell death process only slightly decelerated bacterial growth in irAOX leaves at 1 (irAOX-200, 203) and 2 (irAOX-203) d after inoculation with the pathogen (Supplemental Fig. S8A).

irAOX Plants Accumulate Higher Levels of H₂O₂ after Pst Infection

Cellular reactive oxygen species (ROS) is an important signaling molecule in eukaryotic cells (Rhoads et al., 2006). One of the proposed functions of AOX in abiotic stress resistance is the prevention of overreduction of the mitochondrial UQ pool that can counteract the formation of mitochondrial ROS (Maxwell et al., 1999; Umbach et al., 2005).

We first examined the levels of H₂O₂ by semiquantitative histochemical DAB staining of wild-type and irAOX leaves after Pst DC3000 infection. A strong brown precipitate of oxidized DAB was observed in both the wild type and two irAOX lines at 1 to 3 d after inoculation, but the staining intensity at 2 d appeared stronger in irAOX compared with wild-type leaves (Fig. 9A). To determine H₂O₂ contents more precisely, we used sensitive quantitative Amplex red hydrogen peroxide/peroxidase assay kit and determined H₂O₂ levels in the leaves. Pst DC3000 infiltration resulted in a strong increase in H₂O₂ content at 2 d post inoculation in both wild-type and irAOX plants, and these levels were significantly higher in irAOX leaves (Fig. 9B), coinciding with a more rapid development of cell death symptoms in these plants (Fig. 8A).

Figure 9.

Pst-infected leaves of irAOX plants show higher levels of H₂O₂ and increased expression of the NaRbohD gene. A, DAB-stained Pst DC3000-infected leaves from wild-type and two irAOX lines at 1, 2, and 3 d post inoculation. B, Mean (± se) H₂O₂ levels in mock- and Pst DC3000-infected wild-type and irAOX leaves determined at 24 and 48 h post inoculation. Asterisk indicates significant differences between wild-type and irAOX plants at respective time points determined by unpaired Student’s t test (**P < 0.01; n = 5) C, Mean (± se) of NaRbohD relative transcript abundances quantified with qRT-PCR in mock- and Pst DC3000-inoculated leaves at indicated time points (n = 3). Asterisk shows significant differences between wild-type and irAOX plants at respective time points determined by one-way ANOVA (*P < 0.05, **P < 0.01, *** P < 0.001).

Previously, the Arabidopsis respiratory burst oxidase homolog D (RbohD) was shown to produce a majority of ROS found in pathogen-infected Arabidopsis leaves (Torres et al., 2002). We therefore examined the expression of a putative NaRbohD homolog in N. attenuata after Pst DC3000 inoculation. The transcripts of this NaRbohD gene increased both in the wild type and irAOX (lines 200 and 203) but appeared significantly higher in irAOX-infected leaves (Fig. 9C). This suggests that this RbohD homolog may contribute to the higher ROS levels observed in irAOX lines, possibly in response to a retrograde mitochondrial signal and/or the accumulation of SA, as proposed below.

SA Levels Increase Significantly More in Pathogen-Infected irAOX Leaves

Accumulation of SA is one of the critical features in plant defense against pathogens (for review, see Vlot et al., 2009). As expected, SA levels dramatically increased in both wild-type and irAOX plants after Pst DC3000 infection, showing maximum accumulations at 24 h in Pst-infected wild-type leaves (Fig. 10A). SA levels steadily increased in irAOX plants and reached their putative maximum at 72 h, which was 2 times higher compared with SA levels in wild-type plants (Fig. 10A). It is interesting that higher levels of SA in pathogen-infected irAOX plants negatively correlated with lower levels of induced JA in these leaves, suggesting a negative cross talk between Pst DC3000-induced SA and JA accumulations (Fig. 10A).

Figure 10.

NaPAL gene transcript accumulation correlates with increased SA accumulation in Pst DC3000-infected irAOX leaves. A, Mean (± se) levels of SA and JA in wild-type and irAOX leaves determined 0, 24, 48, and 72 h after Pst DC3000 infection (n = 5). Note large differences in scale between mock- and Pst DC3000-inoculated SA levels. B, Mean (± se) of NaPAL relative transcript abundances quantified with qRT-PCR in mock- and Pst DC3000-infected wild-type and two irAOX lines at indicated time points (n = 3). Asterisk indicates significant differences between wild-type and irAOX plants at respective time points determined by one-way ANOVA (*P < 0.05, ** P < 0.01).

In agreement with the higher SA levels, transcripts of two known PHE AMMONIA-LYASE (PAL) genes in N. attenuata, NaPAL1 and NaPAL2, were significantly more induced in irAOX compared with wild-type plants, attaining approximately 3-fold higher levels 72 h after inoculation (Fig. 10B). This observation provides a potential explanation for higher accumulation of SA in irAOX plants via benzoic acid-dependent SA biosynthetic pathway (Vlot et al., 2009). In contrast to PAL expression, the transcripts of isochorismate synthase (Vlot et al., 2009) encoding another SA biosynthetic pathway, although induced by the pathogen, were not differentially regulated in irAOX and wild-type plants (Supplemental Fig. S8B). Similar to Empoasca spp.-attacked plants, higher levels of SA in irAOX plants were associated with the lower accumulation of JA-dependent HGL-DTGs (Supplemental Fig. S9). In contrast, the levels of a typical PAL-dependent phenolic compound, chlorogenic acid (CGA), increased in wild-type and irAOX plants (Supplemental Fig. S9), and this increase was pronounced in plants that lacked AOX gene function.

DISCUSSION

Activation of AOX is an important marker of stress-adaptive cell reprogramming of plants (for review, see Arnholdt-Schmitt et al., 2006). It is also one of the best known examples of mitochondrial retrograde regulation (MRR; for review, see Rhoads and Subbaiah, 2007) when nuclear-encoded genes like AOX or antioxidant enzymes are induced in response to perturbation of mitochondrial homeostasis. To understand the role of AOX and MRR in defense of N. attenuata against biotic stress, we examined NaAOX function in three biotic stress situations using AOX-silenced N. attenuata plants. The NaAOX transcript and protein levels significantly increased in all cases; however, the lack of AOX transcripts in irAOX plants showed case-specific modifications of plant responses to biotic stress.

AOX-Modulated Responses to Abiotic Stress

In contrast with the well-known role of AOX in the production of heat in thermogenic plants (Meeuse, 1975), the role of the AOX in nonthermogenic plants remains elusive. Fiorani et al. (2005) observed smaller and larger plants in AOX1a antisense- and sense-transformed Arabidopsis plants, respectively, which were maintained at a low temperature (12°C). In an independent experiment, silencing of the cold-inducible AOX gene did not result in any visible growth phenotype; however, several antioxidant defense genes were strongly up-regulated in the cold-exposed aox1a mutant plants, pointing to stronger cold-induced oxidative stress in these plants (Watanabe et al., 2008). Under low nitrogen supply, aox1a plants showed slight increase in gene expression of antioxidant enzymes (Watanabe et al., 2010). In N. tabacum plants with antisense-targeted AOX gene, higher amounts of H₂O₂ were produced under cold stress (Zhang et al., 2009). From these examples, the control of redox status and ROS levels in plants seems to be a unifying stress-related function of AOX in abiotic stresses. As discussed further, a similar concept can be applied to plants under biotic stress.

Reported Roles of AOX in Plant Resistance to Biotic Stress

To further complement the role of AOX in stress responses, the role of AOX in resistance against two types of herbivores and one pathogen was examined in N. attenuata plants. Previously, the role of AOX in response to biotic stress was described in N. tabacum ‘Xanthi’ (nn genotype) plants infected with TMV. Whereas the pretreatment of leaf discs with SA suppressed TMV replication, application of AOX inhibitor salicylhydroxamic acid (SHAM) together with SA counteracted the effect of the hormone on TMV (Chivasa et al., 1997). In addition, application of inhibitors of the Cyt pathway (potassium cyanide [KCN] or antimycin) that elicit AOX transcription via MRR, displayed levels of resistance to TMV that were comparable with SA-induced resistance (Chivasa and Carr, 1998), suggesting a role of SA as potential integrator of mitochondrial function, AOX expression, and MRR. Recently, a nitric oxide-dependent systemic activation of AOX in tomato inoculated with TMV was reported by Fu et al. (2010), showing that AOX is essential for reduced accumulation of viral mRNA and systemic basal resistance to TMV in tomato plants (Fu et al., 2010). In contrast, the existence of SA-dependent and AOX-independent resistance to viruses in N. tabacum plants was proposed by Gilliland et al. (2003). In addition, the overexpression of AOX in transgenic Nicotiana benthamiana plants infected with potato virus X increased susceptibility of plants to systemic disease induction and virus accumulation, and AOX modulated SA-induced resistance to potato virus X (Lee et al., 2011).

AOX and Herbivory in N. attenuata

Here we show that AOX transcription is efficiently induced by the feeding of a chewing herbivore. However, the suppression of AOX in irAOX plants did not impair resistance of transgenic plants to M. sexta, suggesting that these JA-mediated responses are robust and independent of AOX expression. In contrast, the irAOX plants interacting with piercing-sucking insect Empoasca spp. (and Pst DC3000) showed increased SA levels, suggesting that SA could act as an important intermediate between AOX-regulated processes and JA-mediated defense responses in plants that are attacked by insects associated with minimal tissue damage and/or low JA accumulations.

Even at low concentrations, SA induced transient accumulations of AOX protein and transcripts in cultured cells (Norman et al., 2004), presumably due to the interference with normal mitochondrial function. It has been shown that SA at concentrations as low as 20 μm can rapidly inhibit both ATP synthesis and oxygen uptake in the plant cells; however, the site of SA action has not been clearly elucidated (Xie and Chen, 1999). The effect of SA on respiration by intact cells and isolated mitochondria from N. tabacum suspension cell cultures was therefore reexamined by Norman et al. (2004). In these experiments, SA stimulated ADP-limited electron transport at less than 1 mm concentrations, acting as an uncoupler. At higher concentrations, between 1 and 5 mm, SA inhibited respiration in isolated mitochondria by preventing electron flow from the substrate dehydrogenases to the UQ pool, being a potent inhibitor of electron transport. Both the uncoupling and inhibitory effects of SA can explain lower ATP levels in cells observed by Xie and Chen (1999).

Because some of the negative effects of SA were substantially reduced in the presence of strong antioxidant N-acetyl-Cys (Xie and Chen, 1999), ROS was likely involved in these responses. Because SA strongly induces AOX (Chivasa et al., 1997; Norman et al., 2004), a self-regulated interaction network involving SA, AOX, and ROS may be proposed in plants. The Arabidopsis aox1a mutants showed enhanced ROS production in the roots after direct inhibition of Cyt respiratory pathway by KCN (Umbach et al., 2005). The lack of AOX function moderately increased H₂O₂ levels in TMV-inoculated antisense AOX-targeted N. tabacum plants, providing a biologically relevant connection from SA accumulation (typical metabolite induced by TMV; Yalpani et al., 1991) to AOX-controlled ROS accumulation, similar to that found in Pst DC3000-infected irAOX N. attenuata plants.

Possible Mechanisms of Accelerated PCD in irAOX Plants

Mitochondria represent one of the most important sources of ROS in plant cells. Recently, it has been shown that mitochondrial complex II of the electron transport chain is a key contributor to localized ROS accumulation during plant stress and defense responses (Gleason et al., 2011). Potato cells treated with β-glucan elicitor displayed dramatic bursts of H₂O₂, disruption of mitochondrial membrane potential, and PCD when the function of catalase and AOX were simultaneously inhibited by 3-amino-1,2,4-triazol and SHAM, respectively (Mizuno et al., 2005); AOX, therefore, was essential for counteraction of PCD. The application of harpin elicitors strongly induced expression of Arabidopsis AOX1a (Krause and Durner, 2004), further pointing to involvement of AOX in ROS and PCD regulation. In another experiment, harpin treatment rapidly inhibited ATP synthesis in N. tabacum cells, similar to the effect of direct SA application (Xie and Chen, 1999), which induced PCD (Xie and Chen, 2000). In addition, N. tabacum cells lacking AOX by antisense suppression of the gene showed increased sensitivity to PCD-inducing agents such as H₂O₂, SA, and protein phosphatase inhibitor cantharidin (Robson and Vanlerberghe, 2002).

At the whole plant level, treatment of Arabidopsis leaves with SA increased the accumulation of H₂O₂, lipid peroxidation, and oxidative damage to proteins (Rao et al., 1997); however, direct application of H₂O₂ failed to mimic these events. Taking into account the effects of SA on mitochondria, SA is very likely to control mitochondrial ROS in concert with AOX to fine tune the progression of PCD in plants, supported by our findings in N. attenuata plants challenged with Pst DC3000. The expression of HIN1, a known PCD marker in N. tabacum (Takahashi et al., 2004a) increased significantly more in irAOX leaves compared with the wild type after Pst infection, supporting our view that AOX functionally counteracts the early development of PCD in pathogen-infected leaves. Recently, Kiba et al. (2008) reported enhanced cell death in Pst-infected lettuce (Lactuca sativa) leaves after octyl gallate treatment, an effective inhibitor of AOX.

AOX Affecting JA-SA Hormonal Cross Talk

In our experiments with piercing-sucking insects, irAOX plants became more susceptible to Empoasca spp. (leafhopper) feeding, both in the glasshouse and under Utah field conditions. The greater susceptibility to piercing-sucking insects correlated with higher accumulation of SA that was associated with a lower content of late-accumulating JA-dependent defense metabolites in irAOX plants. Recently, a detailed analysis of Empoasca spp. host choice was conducted, revealing a key role of JA in initial determination of plant hosts by these insects (Kallenbach et al., 2012); Empoasca spp. damage negatively correlated with the ability of these plants to perceive wound-induced JAs (Kallenbach et al., 2012). Similarly, the ectopically higher levels of SA in irAOX plants at later time points could interfere with downstream JA signaling in N. attenuata plants, resulting in enhanced feeding of Empoasca spp. on these plants. Recently, Leon-Reyes et al. (2010) showed that SA strongly inhibits JA-mediated gene expression in Arabidopsis, leading to a significant down-regulation of typical JA-responsive genes like PLANT DEFENSIN1.2 and VEGETATIVE STORAGE PROTEIN2 by mechanisms located downstream of JA biosynthesis. Although we did not find directly reduced JA levels in Empoasca spp.-attacked irAOX leaves (associated with moderate AOX-dependent increase of SA), we found significantly reduced JA levels in Pst-infected plants (associated with greatly elevated SA levels). Remarkably, lower levels of JA-dependent, defense-related HGL-DTGs were identified in both interactions. These data suggest that the amounts as well as ratios between SA and JAs may be important to determine the outcome of stress-modulated JA-SA cross talk in plants.

Insects frequently serve as vectors/hosts for microbial pathogens that cause diseases in plants (Weintraub and Beanland, 2006). Although Empoasca spp.-associated Candidatus Phytoplasma spp. have not been detected in our insect colony or from insects collected in the field (Kallenbach et al., 2012), other Empoasca spp.-transmitted microbes, such as viruses or bacteria, could have been responsible for ectopic accumulation of SA and activation of SA-dependent signaling in AOX-deficient N. attenuata plants. Plant antiherbivore defenses countered by pathogenic infections associated with higher levels of SA were recently reported by Thaler et al. (2010). The effect of TMV infection on herbivore performance was tested in the wild type and SA-deficient (NahG) plant backgrounds, showing that SA was required for TMV to increase herbivore vector performance.

Priming and activation of plant immune responses triggered by some beneficial microbes point to an existence of complex interactions mediating both positive and negative roles of microbes in ecological interactions (for review, see Van Wees et al., 2008). It was previously shown that methyl salicylic acid (MeSA) derived from SA is required for systemic acquired resistance (SAR) signal perception in systemic N. tabacum tissues (Park et al., 2007). In the same study, it was proposed that a lipid-derived SAR signal could work with (or upstream of) MeSA to activate SAR. Indeed, the local production of MeSA from SA in Arabidopsis by methyltransferase enzyme was shown to be dependent on the Pst virulence factor coronatine (functional analog of JA-Ile; Attaran et al., 2009). Because most MeSA was directly emitted into the atmosphere, and only a small amount was retained in the leaves, it was proposed that the phytopathogen is using JA signaling-dependent, coronatine-mediated volatilization of MeSA from the leaves to attenuate the SA-based defense pathway.

Although antagonistic interactions prevail in examples of JA-SA cross talk, the effects of both signaling pathways need not be mutually exclusive when these responses become temporally or spatially separated. Several JA perception, biosynthetic, and signaling mutants were attenuated in RPM1-mediated systemic immunity (SAR) in Arabidopsis plants challenged with an avirulent strain of Pst, consistent with a positive role of JA in systemic immunity (Truman et al., 2007). In addition, it was noticed that SA often accumulates 1 d after JA accumulation in natural plant’s innate immune response to viruses (Shang et al., 2011), suggesting that JA and SA can act in synergism and confer optimal virus resistance when applied in appropriate concentrations and with a time delay. Indeed, when these changes were simulated by exogenous application of JA and SA, the strong inhibitory efficiency to virus replication was achieved (80%–90%). It will be interesting to see in the future if AOX could be directly involved in these interactions.

CONCLUSION

Although AOX expression is induced by numerous biotic stress factors, the induction of AOX and protection against stress may not always be in a positively linear relationship. N. attenuata plants clearly benefited from the activity of AOX and alternative respiration to cope with Empoasca spp. attack, and controlled progression of PCD during Pst infection. However, they did not gain any clear benefit from inducing AOX after being attacked by M. sexta caterpillars. These results prompt further investigations, and the use of additional plant-biotic stress models to unravel additional ecological and physiological roles of alternative respiratory pathways in plants.

MATERIALS AND METHODS

Plant Growth and Treatments

Wild-type Nicotiana attenuata plants (22nd inbred generation) seeds, originally collected from a native population in Utah, were used for all experiments, including transformation and generation of transgenic lines. Wild-type plants were transformed with constructs carrying fragments in an inverted repeat orientation of N. attenuata AOX1 (AY422688) and lipoxygenase-3 (LOX3), which mediates JA production (Halitschke and Baldwin, 2003). All transformed plants were diploid as determined by flow cytometry and homozygous for a single transgene insertion. Plants from T2 or T3 generations were used in experiments. Seeds were germinated on agar plates supplemented with Gamborg B5 media (Duchefa) according to the procedures described by Krügel et al. (2002). All treatments were performed on rosette stage plants grown in 1-L pots with soil, unless otherwise noted. For W+W and W+OS treatments, leaves were wounded with a pattern wheel and 20 µL of water, or Manduca sexta OS (one-fifth diluted in water) was rubbed into wounds, respectively. For M. sexta feeding assays, two freshly hatched neonate larvae were placed on the leaves of each wild-type and irAOX plant. Clip cages were used to restrict movement of larvae on the plant, and samples were collected after 0, 1, 2, and 3 d. Clip-caged leaves from separate plants without larvae were used as controls. For Empoasca spp. feeding assays, wild-type and transgenic lines were placed into a glass container (110 × 50 × 80 cm) with Empoasca spp. reared on Cucurbita moschata, Cucurbita maxima var Goldnugget, and Cucurbita pepo plants, and Empoasca spp. were allowed to choose the host plants. The estimated colony size at the time of experiment was 800 to 1000 Empoasca spp. in the glass box. Samples from Empoasca spp.-infested plants were collected after 0, 4, 8, and 12 d. A set of plants growing next to the box with Empoasca spp. were used to collect control samples. Bacterial strain of Pseudomonas syringae pv tomato (Pst) DC3000 was cultivated on Luria-Bertani plates supported with agar containing 25 mg/L rifampicin and 5 mg/L tetracyclin at 28°C. Bacterial growth assays and inoculation of the leaves were carried out essentially as described in Rayapuram et al. (2008).

In the field, 15-d-old seedlings were transferred into previously hydrated 50-mm peat pellets (Jiffy 703; http://www.jiffypot.com) and grown in shade tents for 2 weeks to gradually adjust the plants to the high sun and low relative humidity conditions of the Great Basin Desert. Three to four weeks later, when plants were in the early rosette stage of growth, size-matched wild-type/irAOX-203 plant pairs were transplanted into an irrigated field plot (at a distance of 1.5 m between the pairs) in the 2008 growing season at the Lytle Ranch Preserve near Santa Clara, Utah. The release of transgenics was carried under Animal and Plant Health Inspection Service notification 06–242–3r and the seeds were imported under 07–341–101n.

Generation and Characterization of NaAOX-Silenced Plants (ir-AOX)

A 510-bp fragment of the complementary DNA (cDNA) sequence of the NaAOX1 gene (position 338–848 in AY422688) was inserted into the pSOL3 transformation vector (Bubner et al., 2006) as an inverted-repeat construct to generate pSOL3AOX vector (Supplemental Fig. S3A). N. attenuata wild-type plants were transformed with pSOL3AOX vector using an Agrobacterium tumefaciens-mediated transformation procedure previously described by Krügel et al. (2002). The gene for hygromycin resistance (hptII) allowed selecting hygromycin-resistant transgenic plants on the agar plates supplemented with 35 mg/L hygromycin (Krügel et al., 2002). Southern-blot hybridization of genomic DNA from independently transformed T1 generation plants and from the wild-type line was carried out according to a standard protocol, using ³²P-labeled PCR fragment of the hptII gene as a probe. Independent transgenic lines harboring a single copy of the transgene (Supplemental Fig. S3B) were further screened for homozygosity after germination on hygromycin-containing GB5 media. Two independently transformed single-insert homozygous lines, 200 and 203, strongly suppressed in their NaAOX transcript accumulation, were selected for further experiments.

Expression Analysis by qRT-PCR

Total RNA was extracted from approximately 100 mg of leaf tissue using TRIZOL reagent (Invitrogen; http://www.invitrogen.com), and total RNA was treated with RQ1 RNAse-free DNAse to remove all DNA contaminants (Promega; http://www.promega.com). RNA samples were diluted to a final concentration 0.5 µg/µL, and 1 μg of total RNA was used for cDNA synthesis with oligo(dT) primer (Fermentas; http://www.fermentas.com) and Superscript II reverse transcriptase (Invitrogen; http://www.invitrogen.com) following a standard protocol. qRT-PCR assays were performed on a Stratagene Mx3005P real-time PCR system (http://www.stratagene.com) using qPCR kit for SYBR Green I and qPCR Core kit for Taqman assays (Eurogentec GmbH; http://www.eurogentec.com). To determine NaAOX gene transcript levels in irAOX lines, gene specific primers were designed outside the region used in the ir-silencing construct. Gene-specific primers were designed with Primer3 software (http://frodo.wi.mit.edu/primer3). Relative gene expression was calculated from linear standard curves prepared by serial dilutions of a specific cDNA standard sample. An actin housekeeping gene from N. attenuata (EU273278) was used as endogenous reference to normalize the gene expression. The primers, Taqman probe sequences, and primer sequences for SYBR Green-based qRT-PCR are described in Supplemental Table S1.

AOX Protein Determination by Western Blot

Leaves were ground to a fine powder in liquid nitrogen, and 100 µL of 2× Laemmli buffer with 2% (v/v) β-mercaptoethanol was added to 150 mg of tissue before protein extractions. The samples were centrifuged for 20 min at 16,100g, and protein concentration in the cleared supernatants was determined by the Bradford reagent (Sigma; http://www.sigmaaldrich.com). The aliquots (25 µg) of protein were boiled for 3 min in Eppendorf tubes and separated on 10% SDS-PAGE gel. Proteins from the gel were electroblotted onto a polyvinylidene difluoride membrane (GE Healthcare Life Sciences; http://www.gelifesciences.com) in Tris-Gly buffer and 20% (v/v) methanol. AOX protein was immunodetected using a commercial polyclonal antibody against Arabidopsis (Arabidopsis thaliana) AOX (Agrisera; http://www.agrisera.com) at a dilution of 1:1000 and WesternBreeze immunodetection kit (Invitrogen; http://www.invitrogen.com). The manufacturer’s instructions were followed except for the following modifications: membranes were blocked overnight at 4°C, and washing steps were prolonged to 10 min.

Measurement of Respiration in N. attenuata Leaves

The plants were placed in the dark for 15 min before the start of respiration measurements. Leaf discs were excised by an 8-mm-diameter cork borer from the leaves and placed in the plastic syringe with assay buffer (20 mm HEPES, 0.2 mm CaCl₂, pH 7.2; Simons et al., 1999) supplied with inhibitors or without inhibitors for determination of total respiration. The leaf discs were briefly infiltrated by creating a partial vacuum in the syringe and incubated in the dark for 15 min. Four infiltrated discs (approximately 80 mg fresh mass) were transferred into a 4-mL glass vial filled with fresh O₂-saturated assay buffer (2 mL) equipped with a miniaturized Clark-type oxygen electrode (Oxygen Microsensor, Unisense). After initial equilibration of the electrode for several minutes, O₂ consumption was continuously measured in the dark for 2 to 3 min and recorded using Sensor Trace Basic 3.1.1 software (Unisense) in 10-s intervals. Total respiration rates (V_TOTAL) were determined as leaf disc O₂ uptake (μmol min⁻¹ g fresh mass⁻¹) in the assay buffer without inhibitors. The cyanide-sensitive Cyt-dependent respiration (V_CYT) was determined in the assay buffer containing 1 mm KCN and calculated as V_TOTAL – V_KCN. Finally, cyanide-insensitive but SHAM-sensitive alternative respiration (V_ALT) was determined after addition of 5 mm SHAM to 1 mm KCN-containing buffer and calculated as difference V_KCN – V_KCN+SHAM (Furuhashi et al., 1989).

Phytohormone Analysis

Harvested leaves were immediately frozen in liquid nitrogen and stored at –80°C until analysis. Phytohormones SA, JA, and JA-Ile were extracted following the procedure described in Wu et al. (2007). In brief, 150 mg of leaf material was extracted with 1 mL of ethyl acetate spiked with 200 ng D₄-SA, ¹³C₂-JA, and ¹³C₆-JA-Ile as internal standards in FastPrep tubes containing 0.9 g of FastPrep Matrix (Sili GmbH; http://www.sigmund-lindner.com). The tissues were homogenized in 2000 GENO/GRINDER homogenizer (SPEX CertiPrep; http://www.spexcsp.com) for 2 min, and samples were centrifuged at 13,000g for 20 min at 4°C. The supernatants were transferred to clean 2-mL Eppendorf tubes, and each pellet was reextracted with 1 mL of ethyl acetate and centrifuged. Combined supernatants were dried in a vacuum concentrator to near dryness (Eppendorf; http://www.eppendorf.com) before resuspending the samples in 0.5 mL of 70% methanol (v/v). Ten microliters of extract was subjected to reverse-phase HPLC coupled to tandem mass spectrometry using the 1200LC LC/MS-MS-MS system (Varian; http://www.varianinc.com) according to the procedure described in Wu et al. (2007). The levels of each phytohormone were calculated based on the peak area of the labeled internal standards and fresh mass of the sample.

Analysis of Secondary Metabolites

Secondary metabolites were analyzed by HPLC coupled to diode array detector as described in Kaur et al. (2010). Approximately 100 mg of leaf material was extracted with 1 mL of extraction buffer containing 40% (v/v) methanol and 0.5% (v/v) acetic acid. After centrifugation, the supernatants were collected and injected into an Agilent 1100 series HPLC (http://www.chem.agilent.com) equipped with ODS Inertsil C18 column (3 μm, 150 × 4.6 mm i.d.) that was protected by a Phenomenex Security Guard C18 precolumn (Phenomenex; http://www.phenomenex.com). Nicotine, rutin, and CGA standards were used to construct corresponding calibration curves. CP and dicaffeoylspermidine contents were calculated on the basis of CGA calibration that showed nearly identical UV-absorption profile.

Determination of Cell Death

Trypan blue staining specific to dead cells with permeabilized cytoplasmic membranes was carried out as described in Koch et al. (1991). Electrolyte leakage was measured as described in Pike et al. (1998) using a conductivity meter (Mettler; http://us.mt.com/). Eight discs (0.785 cm² each) were punched from the treated leaves using a cork borer and placed in 15-mL Falcon tubes (BD Biosciences; http://www.bdbiosciences.com) together with 10 mL of sterile distilled water. Samples were incubated for 4 h under slow rotation at 20°C, and the conductivity of resulting water solutions was measured by conductivity meter.

H₂O₂ Visualization and Determinations

DAB staining was performed on the leaves inoculated with Pst DC3000 or mock solutions as described by Thordal-Christensen et al. (1997) with some modifications. Leaves with petioles were excised from the plants at each point in time after bacteria infiltration and supplied with 1 mg/mL DAB solution through the petioles until a brown precipitate was observed in the leaf blades (5 h). Leaves were destained in boiling ethanol (96%) for 10 min and fixed in 3:1:1 ethanol/lactic acid/glycerol solution. To determine endogenous concentrations of H₂O₂, liquid nitrogen-stored leaves were ground in liquid nitrogen, and 100 mg of tissue powders were extracted in 200 µL of 25 mm HCl supplemented with 50 mg of activated charcoal. After 1 to 2 min vortexing, samples were kept on ice for 10 min and centrifuged at 12,000g for 20 min at 4°C. Cleared supernatants were transferred to clean tubes, and H₂O₂ concentrations were determined using Amplex red hydrogen peroxide/peroxidase assay kit (Invitrogen; http://www.invitrogen.com). H₂O₂ was used to construct calibration curves. Prior to each measurement, 10 µL of each extract was neutralized with 90 µL of 1× reaction buffer, and 50 µL of neutralized extract was loaded on a 96-well Nunc multiwell plate (Thermo Scientific; http://www.nuncbrand.com). Next, 50 µL of reaction buffer was dispensed into each well to initiate the reaction. Samples were incubated for 30 min at room temperature before measurement of fluorescent intensities at Excitation/Emission 530/590 nm in an Infinite 200 multimode reader (Tecan Group Ltd.; http://www.tecan.com).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. AOX genes from N. attenuata are highly homologous to AOX1 a gene from Arabidopsis.

• Supplemental Figure S2. NaAOX proteins increase in response to simulated herbivory, feeding of M. sexta caterpillars, Empoasca spp. attack, and Pst DC3000 infection in N. attenuata.

• Supplemental Figure S3. Structure of plant transformation vector pSOL3AOX used for silencing of AOX genes in N. attenuata and characterization of irAOX transgenic lines.

• Supplemental Figure S4. Functional characterization of irAOX lines.

• Supplemental Figure S5. irAOX WOS- and herbivory-induced plants have salicylate levels and H₂O₂ levels comparable to wild-type plants.

• Supplemental Figure S6. Exposure to M. sexta feeding does not change a majority of the secondary metabolite profiles in irAOX plants compared with the wild type.

• Supplemental Figure S7. Silencing of NaLOX3 compromises defense of N. attenuata against Empoasca spp.

• Supplemental Figure S8. Multiplication of Pst DC3000 is slightly suppressed in irAOX plants and NaICS transcript levels are not differentially regulated in wild-type and irAOX plants infected with Pst DC3000.

• Supplemental Figure S9. Secondary metabolites levels are altered in irAOX plants relative to the wild type after Pst DC3000 infection.

• Supplemental Table S1. Real-time PCR primers and probes.

Glossary

Cyt

cytochrome

UQ

ubiquinone

TMV

tobacco mosaic virus

PCD

programmed cell death

qRT-PCR

quantitative reverse transcription PCR

OS

oral secretions

RNAi

RNA interference

SA

salicylic acid

JA

jasmonic acid

Pst

Pseudomonas syringae pv tomato

DAB

3,3'-diaminobenzidine

HGL-DTG

17-hydroxygeranyllinalool diterpene glycoside

ROS

reactive oxygen species

CGA

chlorogenic acid

MRR

mitochondrial retrograde regulation

SHAM

salicylhydroxamic acid

SAR

systemic acquired resistance

cDNA

complementary DNA

CP

caffeoylputrescine

MeSA

methyl salicylic acid

H₂O₂

hydrogen peroxide