- American Society of Plant Biologists
Abstract
Previously, it has been shown that Arabidopsis thaliana leaves exposed to high light accumulate hydrogen peroxide (H2O2) in bundle sheath cell (BSC) chloroplasts as part of a retrograde signaling network that induces ASCORBATE PEROXIDASE2 (APX2). Abscisic acid (ABA) signaling has been postulated to be involved in this network. To investigate the proposed role of ABA, a combination of physiological, pharmacological, bioinformatic, and molecular genetic approaches was used. ABA biosynthesis is initiated in vascular parenchyma and activates a signaling network in neighboring BSCs. This signaling network includes the Gα subunit of the heterotrimeric G protein complex, the OPEN STOMATA1 protein kinase, and extracellular H2O2, which together coordinate with a redox-retrograde signal from BSC chloroplasts to activate APX2 expression. High light–responsive genes expressed in other leaf tissues are subject to a coordination of chloroplast retrograde signaling and transcellular signaling activated by ABA synthesized in vascular cells. ABA is necessary for the successful adjustment of the leaf to repeated episodes of high light. This process involves maintenance of photochemical quenching, which is required for dissipation of excess excitation energy.
INTRODUCTION
In the natural environment, plants are frequently exposed to fluctuating light intensities and often absorb more light energy than can be consumed by photosynthetic metabolism and thus require that excess excitation energy be dissipated. Many abiotic and biotic stresses limit photosynthesis, which causes further increases in excess excitation energy needing to be dissipated (Long et al., 1994; Asada, 1999; Baker 2008). Failure to dissipate excitation energy results in overreduction of the photosynthetic chain components that direct linear electron flux (LEF) from water to NADPH (Baker et al., 2007). Part of the absorbed light energy is dissipated as heat in the light-harvesting complexes of photosystem II (PSII) through non photochemical quenching (NPQ; Horton et al., 1996; Müller et al., 2001). Additional dissipation of excitation energy is also achieved by photochemical quenching (Baker et al., 2007; Baker, 2008) and reflects that action of processes such as the reduction of molecular oxygen at photosystem I by the Mehler reaction (Asada, 1999; Ort and Baker, 2002; Baker et al., 2007) and through photorespiration (Asada, 1999; Douce and Neuberger, 1999). These two processes produce reactive oxygen species (ROS) that are scavenged by lipid- and water-soluble low molecular weight antioxidants and antioxidant enzymes (Mittler et al., 2004; Apel and Hirt, 2004; Van Breusegem et al., 2008). Sustained exposure to very high light intensities, well in excess of light intensities optimal for growth (hereafter called excess light), will exceed the antioxidant and excitation energy dissipation capacity of the leaf and cause oxidative damage to the photosynthetic apparatus (Aro et al., 1993; Asada, 1999; Krieger-Liszkay, 2005; Triantaphylides et al., 2008), photobleaching, and cell death (Karpinski et al., 1999; Mühlenbock et al., 2008).
In excess light–stressed plants, damaged chloroplasts initiate retrograde signaling to the nucleus (Nott et al., 2004; Pogson et al., 2008) to downregulate the expression of photosynthetic genes and upregulate stress defense genes to mitigate oxidative stress (Rossel et al., 2002, 2007; Kimura et al., 2003; Koussevitzky et al., 2007; Mühlenbock et al., 2008). In contrast with excess light treatments, Arabidopsis thaliana leaves that are exposed to a moderate increase (typically <10-fold) over growth light intensity (hereafter called high light [HL]) do not suffer oxidative stress (Fryer et al., 2003; Davletova et al., 2005) or irreversible photoinhibition (Russell et al., 1995; Karpinski et al., 1997; Fryer et al., 2003). However, these plants do accumulate H2O2 in the chloroplasts of bundle sheath cells (BSCs) and neighboring mesophyll cells (Fryer et al., 2003; Mullineaux et al., 2006). BSCs in Arabidopsis form a single layer of elongated cells around the vasculature (Kinsman and Pyke, 1998; Leegood, 2008). Limitations in the supply of CO2 to such cells (Morison et al., 2005) may cause them to more readily produce ROS via the photoreduction of O2 (Hibberd and Quick, 2002; Fryer et al., 2003; Leegood, 2008). The accumulation of H2O2 in BSC chloroplasts is associated with the rapid induction of the antioxidant gene ASCORBATE PEROXIDASE2 (APX2; Fryer et al., 2003; Ball et al., 2004; Karpinski et al., 1997, 1999). HL-mediated induction of APX2 requires LEF, redox signals from reduced glutathione, and abscisic acid (ABA; Karpinski et al., 1999; Fryer et al., 2003; Ball et al., 2004; Chang et al., 2004). These latter features are common to a majority of other HL-responsive genes examined to date (Rossel et al., 2002; Ball et al., 2004; Bechtold et al., 2008). However, in contrast with most HL-responsive genes so far examined, the induction of APX2 expression also requires extracellular H2O2 (Karpinski et al., 1999; Bechtold et al., 2008). The BSC-specific expression of APX2 in response to HL allows this gene to be used as a BSC-specific reporter (Mullineaux et al., 2006), in contrast with the expression of many HL-responsive genes that may not be confined to a single leaf tissue (Bechtold et al., 2008).
The induction of APX2 expression and increased capacity to dissipate excitation energy in BSCs can be prevented if the leaf is exposed to high humidity (Fryer et al., 2003). This observation suggests that at low humidity, HL-exposed BSCs experience a loss of water (Fryer et al., 2003). Water loss leads to the accumulation of ABA, which plays a central role in the regulation of plants' water status (Davies et al., 2002; Christmann et al., 2006). HL-mediated APX2 induction is attenuated in the ABA signaling mutants abi1-1 and abi2-1 (Fryer et al., 2003). In the mutant altered in APX2 expression8-1 that constitutively expresses APX2, ABA content is threefold elevated under nonstress conditions, providing a correlation between ABA accumulation and APX2 expression (Rossel et al., 2006). Supply of ABA to plants under low light conditions has shown that many HL-expressed genes are responsive to ABA (Fryer et al., 2003; Rossel et al., 2006; Bechtold et al., 2008).
Based on the requirement of APX2 induction for an extracellular source of H2O2 in BSCs (Karpinski et al., 1999; Bechtold et al., 2008), we speculate that an ABA-regulated plasma membrane NADPH oxidase as shown in guard cells (Pei et al., 2000; Murata et al., 2001; Mustilli et al., 2002; Kwak et al., 2003; Li et al., 2006) may be a source of ROS in BSCs (Mullineaux et al., 2006). BSCs may contain additional ABA signaling components that regulate extracellular ROS, such as the heterotrimeric G protein complex (Suharsono et al., 2002; Booker et al., 2004; Joo et al., 2005; Li et al., 2006), the OPEN STOMATA1 (OST1) protein kinase (Mustilli et al., 2002; Xie et al., 2006), and ABI1 (Murata et al., 2001).
ABA is synthesized in response to a reduction in water potential (Davies et al., 2002; Nambara and Marion-Poll, 2005; Christmann et al., 2006, 2007). ABA biosynthesis is partitioned between plastids and the cytosol (Qin and Zeevaart, 1999; Nambara and Marion-Poll, 2005). The oxidative cleavage of the precursor carotenoid 9′-cis-neoxanthin to xanthoxin, catalyzed by the plastidial enzyme 9′-cis-epoxycarotenoid dioxygenase (NCED), is the committed step for ABA biosynthesis (Nambara and Marion-Poll, 2005). Xanthoxin is then exported from the plastid, and the two remaining biosynthetic steps to ABA, catalyzed by enzymes coded by ABA DEFICIENT2 (ABA2) and ABA ALDEHYDE OXIDASE3 (AAO3), occur in the cytosol (Nambara and Marion-Poll, 2005).
In this study, we set out to establish how ABA, secreted from vascular parenchyma cells (Endo et al., 2008), regulates HL-responsive gene expression and integrates into H2O2- and redox-mediated retrograde signaling from chloroplasts in BSCs. We conclude that in HL-exposed leaves at ambient or lower humidity (1) paracrine (i.e., cell-to-nearby-cell) signaling occurs between vascular parenchyma cells and BSCs, (2) ABA signaling integrates into H2O2- and redox-mediated retrograde signaling from BSC chloroplasts, and (3) ABA is required for an effective physiological response of leaves to a fluctuating light environment.
RESULTS
Exposure to Moderate HL Treatment Does Not Induce Signaling Pathways Associated with Exposure to Excess Light
To determine the effects of exposure to HL in the BSCs, we measured the expression of the antioxidant gene APX2 and chlorophyll fluorescence as an indicator of photosynthetic performance and photoinhibition. The response of leaves was determined to a fivefold increase in PPFD over their growth PPFD (150 μmol m−2 s−1), hereafter referred to as HL. Exposure to HL caused detectable accumulation of APX2 transcript after 20 min, which continued to rise throughout the duration of the experiment (Figure 1A ). The exposure of leaves to HL resulted in a decline in the chlorophyll fluorescence parameter Fv/Fm, which defines the maximum quantum efficiency of PSII photochemistry (see Methods; Baker, 2008; see Supplemental Figure 1 online). Fv/Fm values reverted to pre-HL exposure values when plants were returned to their growth conditions (see Supplemental Figure 1 online). Under these conditions, the expression of genes controlled by retrograde signaling pathways responsive to excess light (Danon et al., 2005; Koussevitzky et al., 2007; Pryzbyla et al., 2008) showed no significant change (see Supplemental Figure 2 online). Taken together, these data indicate that the plant's responses to HL exposure did not induce permanent damage to leaves and did not activate retrograde signaling associated with such photooxidative stress.
ABA Levels and APX2 Expression in HL-Exposed Leaves and Osmotically Stressed Petioles.
(A) Increase in ABA content (closed symbols, dashed line) and APX2 transcript levels (open diamonds, solid line) in leaves attached to the rosette and exposed to HL (750 μmol m−2 s−1 PPFD) and 25% RH. Each leaf (one per plant) was clamped into a CIRAS leaf chamber to control humidity and temperature (see Methods). NCED3 transcript levels are shown (closed squares, dotted line). ABA levels were determined using a radioimmunoassay (see Methods) from fully expanded outer leaves attached to rosettes during HL exposure. The data presented, expressed as μg per gram dry weight (gm DWt−1), are the means (±se; n = 6) from one expanded leaf from each of three plants for each time point in two experiments. Foliar transcript levels were determined by quantitative real-time PCR on single-stranded cDNA prepared from total leaf RNA (see Methods) harvested at each time point. Each data point is the mean cDNA level (±se) relative to the zero time point, low-light cDNA level. Transcript levels were normalized with respect to CYCLOPHILIN transcript levels, which do not respond to HL (Rossel et al., 2006).
(B) Foliar ABA content in detached leaves exposed to HL (PPFD of 750 μmol m−2 s−1) for 45 min at either low humidity (25% RH) or high humidity (80% RH) or kept at the growth PPFD of 150 μmol m−2 s−1 (LL). The difference in ABA levels between the mean (±se) HL and LL samples at low humidity was significant (P = 0.024 from t test; n = 8 from two experiments). One fully expanded leaf from each rosette was used and clamped into a CIRAS leaf chamber to control humidity and temperature (see Methods).
(C) APX2LUC expression in osmotically stressed petioles is light dependent. Fully expanded leaves of APX2LUC/Col-0 plants (Karpinski et al., 1999) were detached and infiltrated with PEG-400 between 0 and 0.5 M for 2 h at a PPFD of 20 μmol m−2 s−1 and then petioles were detached and bathed in the same concentrations of PEG-400 plus 1 mM D (−) luciferin for a further 1.5 h either at growth PPFD (low light) or in the dark. At the end of this period, luciferase activity (top panels) and reflected light (bottom panels) were imaged using a CCD camera and false color images for luciferase activity generated (see Methods). The experiment shown was typical of replicates of these treatments.
(D) ABA levels in osmotically stressed petioles in the low light and the dark. Petioles of APX2LUC leaves were treated with 0.4 M PEG-400 or water as described in the legend for (C) and then their ABA content determined as in the legend for (A). Some additional petioles of low light–incubated petioles were imaged for luciferase activity, as shown in (C), to confirm their response to the PEG-400 treatment. The values are the means (±se) for three separate experiments from eight petioles pooled in each experiment from four plants.
(E) DCMU has no effect on HL-induced foliar ABA accumulation. Detached leaves were infiltrated with 30 μM DCMU for 2 h (see Methods) prior to exposure to HL (750 μmol m−2 s−1) at 25% RH. Leaf ABA content was then determined as in the legend for (A). These data are the means (±se) of two experiments with four leaves each from a separate plant (n = 8).
ABA Accumulation in HL-Exposed Leaves Is Due to Changes in Leaf Water Status
To examine the possible role of ABA in the response of leaves to HL, the ABA content of leaves was determined. ABA content increased in attached and detached leaves by 65 and 60%, respectively, when exposed to HL at 25% RH (hereafter called low humidity; see Methods). This indicates that the ABA accumulation is leaf autonomous; no other source of ABA, such as from roots (Nambara and Marion-Poll, 2005), is involved. The increase in ABA was observed after 15 min in HL and was followed by an increase in APX2 expression (Figure 1A). At 80% RH (hereafter termed high humidity; see Methods), HL did not increase the ABA level (Figure 1B).
In the first 10 min of exposure to HL under low humidity, the transpiration rate increases (Fryer et al., 2003), which transiently lowers leaf water status and provides the necessary cue to initiate ABA accumulation. However, the sudden increase in light intensity may also directly contribute (a) signal(s) that triggers ABA accumulation in leaves. To distinguish between these two possibilities, ABA accumulation was measured in petioles of Columbia-0 (Col-0) plants harboring an APX2 promoter LUCIFERASE gene fusion (APX2LUC; Karpinski et al., 1999) subjected to a range of osmotic stress treatments under low-light conditions and in the dark. Petioles contain BSCs, which express APX2 (Fryer et al., 2003; Ball et al., 2004), and vascular parenchyma that carry out ABA biosynthesis in the Arabidopsis leaf (Endo et al., 2008). Petioles of Col-0/APX2LUC were chosen as an experimental system because they could be bathed in solutions of different osmotic potential, and on exposure to light, APX2 expression could be rapidly monitored by imaging the luminescence produced by luciferase (Figure 1C). Incubation of isolated APX2LUC petioles in a range of polyethylene glycol (PEG-400) solutions of increasing osmotic potential under low-light conditions activated APX2 expression (Figure 1C), with osmotic potential thresholds of −0.5 to −0.8 MPa (calculations based on Money, 1989). Similar observations were made when petioles were treated with sorbitol or mannitol solutions of equivalent osmotic potential (data not shown). The 0.4 M PEG-400 that induced APX2 expression also caused a decrease in the chlorophyll fluorescence parameters Fq′/Fm′ and Fv/Fm and an increase in qL and NPQ (Table 1 ). Fq′/Fm′ provides an estimate of the quantum efficiency at which PSII operates under a given PPFD (Baker, 2008). qL and NPQ provide information of the redox state of the primary quinone electron acceptor of PSII (QA) and NPQ, respectively (Baker, 2008). Similar changes in these characteristics accompany the induction of APX2 expression (Karpinski et al., 1997, 1999; Fryer et al., 2003; Chang et al., 2004).
Chlorophyll Fluorescence Parameters for Water- and 0.4 M PEG-400–Treated Petioles.
ABA levels in petioles increased in low light and in the dark by 30 and 27%, respectively, when 0.4 M PEG-400, a concentration above the threshold for APX2 expression, was applied (Figure 1D). These data indicate that ABA accumulation in petioles is not dependent on a light-associated signal in addition to a change in water status. Similarly, treatment of HL-exposed leaves with the LEF inhibitor DCMU (Duysens, 1972) did not inhibit ABA accumulation (Figure 1E), confirming the observations in dark-incubated osmotically stressed petioles.
A Capacity for Foliar ABA Biosynthesis Is Required for Induction of HL-Responsive Genes
ABA has been implicated in the induction of BSC-specific APX2 expression and some additional HL-responsive genes whose expression occurs in several leaf tissues (Fryer et al., 2003; Rossel et al., 2006; Bechtold et al., 2008). To examine the role of the foliar ABA biosynthetic pathway, the compound ABAmineSG, a specific inhibitor of NCED activity that is involved in the synthesis of ABA (Kitahata et al., 2006), was applied to detached HL-illuminated leaves of Col-0/APX2LUC. ABAmineSG inhibited the rise in foliar ABA levels and the increase in APX2 expression in leaves exposed to HL at low humidity (Figure 2A ).
The Expression of HL-Responsive Genes Require a Foliar Capacity for ABA Biosynthesis.
(A) APX2LUC expression in leaves treated with the NCED inhibitor ABAmineSG (Kitahata et al., 2006). Image of luciferase activity from Col-0/APX2LUC infiltrated with ABAmineSG and D (−) luciferin (1 mM) or luciferin only for 2 h prior (see Methods) prior to exposure to HL (750 μmol m−2 s−1) at 25% RH for 45 min. Luciferase activity was imaged using a CCD camera system and image processing to color code luciferase activity (see Methods). Red indicates regions of highest luciferase activity. The effect of ABAmine on foliar ABA levels (Kitahata et al., 2006) was confirmed in a single experiment using four leaves from separate APX2LUC/Col-0 plants for each treatment as described in the legend of Figure 1 and Methods.
(B) APX2 transcript levels in HL-exposed leaves of ABA biosynthesis mutants. Whole rosettes of the wild type (Col-0), nced3-2, aba2-11, and aba2-14 were exposed to HL (750 μmol m−2 s−1) for 45 min at growth humidity (see Methods). APX2 expression was determined as transcript levels in the mutants relative to the wild type using real-time quantitative RT-PCR (see Methods and legend of Figure 1). Data were normalized to cDNA levels of CYCLOPHILIN, which shows no significant variation in expression under HL or where ABA levels differ between genotypes (Rossel et al., 2006). All data shown are the means (±se) combined from two biological replications of six plants. All differences between mutants and the wild type were significant (P < 0.001 from t test).
(C) Transcript levels of five HL-responsive genes in leaves of ABA biosynthesis mutants. The genes used in this study are to compare with the BSC-specific expression of APX2. The conditions and methods are as in legend for (B). Asterisks show where differences between mutants and the wild type were significant (P ≤ 0.05 from t test).
(D) Hierarchical clustering of shared gene expression responses to HL exposure and exogenous ABA application. A total of 816 genes were identified as representing a significant overlap between HL and ABA responses and were clustered with data from the Gene Expression Omnibus (GEO) and NASCARRAYS databases. The 30 min, 1 h, and 3 h ABA data come from NASCARRAYS-176, 4 h ABA from GSE7112, 3 h ABA #2 from GSE6171, and 3 h HL from GSE7743. In this TREEVIEW representation, red indicates upregulation relative to mock or control treatment, while green indicates downregulation. The scale bar indicates log (base 2) ratios of treatment to control for the heat map.
The predominant NCED gene expressed in Arabidopsis leaves is NCED3 (Iuchi et al., 2001; Tan et al., 2003; Ruggiero et al., 2004; Endo et al., 2008). When rosettes of a null mutant of NCED3 (nced3-2; see Supplemental Figure 3 online) were exposed to HL, induction of APX2 expression was attenuated (Figure 2B) consistent with the data obtained with ABAmineSG (Figure 2A). Leaves from nced3-2 plants did not show an increase in ABA content under HL (see Supplemental Figure 3 online), although prestress levels of ABA were no different from those of wild-type leaves (see Supplemental Figure 3 online). A second T-DNA insertion mutant allele of NCED3, sto1-1, which has reduced expression of the gene (Ruggiero et al., 2004; see Supplemental Figure 3 online), also showed attenuated induction of APX2 expression in HL-exposed leaves (Figure 2B), although not to the same degree as nced3-2 (Figure 1D). Null mutant alleles of ABA2, aba2-11, and aba2-14, which have low levels of foliar ABA under a range of conditions (Gonzalez-Guzman et al., 2002; Barrero et al., 2006), failed to induce APX2 expression in response to HL (Figure 2B). NCED3 expression did not increase upon HL and low humidity treatment (Figure 1A), suggesting that the mechanism of action is posttranscriptional.
Additional HL-responsive genes (Bechtold et al., 2008) were tested for their expression in HL-exposed ABA biosynthesis-deficient mutants. All mutants caused a significant reduction in the expression of the test genes compared with the wild-type controls (Figure 2C). A meta-analysis of publicly available microarray data for treatment of seedlings with ABA (Goda et al., 2008) compared with data from HL-exposed seedlings (Kleine et al., 2007; see Methods) revealed that 816 genes were coresponsive to ABA and HL (P value < 0.00001). Expression of 496 genes was induced under both conditions, while 320 were suppressed in response to both treatments (see Supplemental Data Set 1 online). When expression data for these genes were clustered with other publicly available ABA treatment data (Figure 2D), a strong correlation was observed between 3 h of HL exposure and plants 3 or 4 h after ABA application at a variety of concentrations (uncentered correlation = 0.780). Thus, a significant number of HL-responsive genes are also responsive to ABA levels and may require foliar ABA biosynthesis, implying an important role for this hormone in the response of the leaf to its prevailing light environment.
A Capacity for Foliar ABA Biosynthesis Is Required for Response to a Change in Light Intensity
Based on the above observations on the correlation between HL- and ABA-responsive gene expression, we hypothesized that a capacity to synthesize foliar ABA would be important for the leaf's ability to adjust to changing light conditions. To test this hypothesis, nced3-2 was subjected to HL and lowered humidity for 60 min at the same time each day for 5 d, and chlorophyll fluorescence parameters were imaged before, immediately after, and 2 h after each HL episode. The treatment protocol and times at which fluorescence parameters were imaged are shown in Figure 3A (see Methods).
A Foliar Capacity for ABA Biosynthesis Is Required for Fully Expanded Leaves to Adjust to Repeated Episodes of HL at Growth Humidity.
(A) The daily HL regime applied to and chlorophyll fluorescence measurements taken on nced3-2 and Col-0 plants. This daily regime was set up and run in a chlorophyll fluorescence imager programmed to apply the procedure automatically for a total of 130 min with plants transferred to the machine 2 h into their photoperiod. After transfer of plants into the Fluorimager cabinet, actinic lights were switched on and provided a PPFD equivalent to that in growth conditions (150 μmol m−2 s−1) and measurements of Fq′/Fm′ made every 5 min until unchanging values were obtained, typically after 20 min. The modulated (light adapted) fluorescence parameters required to calculate qL and NPQ (see Methods) were collected at point 1 before HL. Then the lights were switched off for a period of 20 min, at the end of which (point 2) dark-adapted fluorescence parameters were collected to calculate NPQ for point 1 and Fv/Fm (at point 2). The plants were then exposed for 60 min to a 10-fold increased light intensity with the same light source (PPFD of 1500 μmol m−2 s−1) and the response of plants monitored by measurements of Fq′/Fm′ every 10 min. At the end of the HL exposure (point 3), the modulated fluorescence parameters were measured to calculate qL and NPQ. Finally, a second 20-min dark period was applied, at the end of which (point 4) dark-adapted fluorescence parameters, required to calculate NPQ for point 3 and Fv/Fm (at point 4), were measured. After point 4, plants were returned to their growth environment.
(B) Outer leaves of nced3-2 show lower dark-adapted Fv/Fm values than Col-0 after five daily 60-min HL exposures. Whole rosette images for Fv/Fm are shown for one mutant and wild-type individual on day 1 before the first HL exposure (point 2 in [A]) and the same individuals at the end of the last HL exposure (point 4 in [A]) on day 5. Note how uniform Fv/Fm is across the rosettes at the start of the experiment on day 1. By contrast, on day 5, outer leaves of nced3-2 show lower values in at least half the leaves and overall more variation than the wild type. Note also that Col-0 has increased its rosette area in the 5-d period in contrast with the mutant.
(C) Daily dark-adapted Fv//Fm values after HL (point 4 in [A]) in nced3-2 and Col-0. The values are the means (±se) of two separate experiments each consisting of values collected daily from two outer leaves of three plants of each genotype (n = 30). The asterisks indicate a significant difference (P ≤ 0.05 from t test) between mutant and the wild type at the time point.
(D) Daily Fq′/Fm′, qL, and NPQ values before and after the HL exposure at points 2 and 4, respectively, in (A). The data are the means (±se) from the same experiments, and the asterisks denote the same threshold of significance as described for (C).
After a single HL treatment, only minor differences in dark-adapted Fv/Fm between mutant and wild-type plants were observed (Figures 3B, top panel and 3C). However, after 5 d, the outer, fully expanded leaves of nced3-2 rosettes had lower Fv/Fm values than the equivalent wild-type leaves (Figures 3B, bottom panel, and 3C), indicating that the mutant plants, unlike the wild type, experienced photoinhibition of PSII that was not recoverable during a 20-min dark period or after 22 min back in growth conditions. Measurements of the PSII operating efficiency at the growth and HL treatment PPFDs (Figure 3D) confirmed that nced3-2, but not wild-type leaves, experienced an increase in photoinhibition with increasing days of treatment. There were no significant differences in NPQ between the mutant and the wild type during the course of the HL treatments, but there were decreases in qL values for the mutant (Figure 3D). Thus, the photoinhibition observed in nced3-2 was attributable to a reduced ability for photochemical quenching since QA was more reduced in the mutant than in the wild type. This would imply that the mutant is less able than the wild type to use the products of LEF, ATP, and reductants. Consequently, nced3-2 leaves experienced increasing photoinhibition and, presumably, photodamage to PSII, with increasing periodic exposure to HL.
The same experiments were also conducted on aba2-11 and aba2-14, but the leaves wilted severely during the HL exposures, which made the accurate determination of chlorophyll fluorescence parameters impossible.
Taken together, these data indicate that induction of ABA biosynthesis is a key factor in the response of fully expanded leaves to repeated HL episodes at low RH.
HL Response Is Associated with Biphasic Accumulation of Extracellular H2O2 in Vein-Associated Cells and the Leaf Lamina
While previous studies reported a role for the production of extracellular ROS in the regulation of HL-responsive genes, direct observation of ROS accumulation during HL exposure was not reported. Real-time measurement of H2O2 accumulation in leaves can be achieved using a sensitive derivative of the H2O2-specific fluorogenic probe 10-acetyl-3,7-dihydroxyphenoxazine, called Amplex Red Ultra (ARU; see Methods). ARU is oxidized by H2O2 in a peroxidase-catalyzed reaction to resorufin (Zhou et al., 1997). ARU-resorufin fluorescence reports accumulation of extracellular H2O2 in leaf lamina tissues (Šnyrychová et al., 2008). At high magnification, ARU fluorescence was observed as a diffuse overlay of tissues of the veins and immediately adjacent leaf lamina (see Supplemental Figure 4 online), confirming the observations of Šnyrychová et al. (2008). Laser scanning confocal microscopy could not be used to establish more precise location of extracellular H2O2 in leaf lamina and periveinal tissue, since the laser light source used induced H2O2 production, rapidly caused bleaching of tissues and quenching of resorufin fluorescence.
To simultaneously image resorufin accumulation in periveinal and adjacent lamina, initial experiments relied on the colocalization of APX2LUC expression with ARU-resorufin fluorescence (Figure 4A ). These experiments showed that 30 min after HL exposure at low humidity, elongated cells displayed readily detectable resorufin fluorescence compared with adjacent lamina tissue (Figure 4A). Since APX2LUC expression is confined to BSCs of the periveinal region in HL-exposed leaves (Fryer et al., 2003; Ball et al., 2004), we conclude that extracellular H2O2 accumulation is associated with these cells, although we do not rule out that other vein-associated tissues would also produce extracellular H2O2.
Extracellular H2O2 Production in Veins and Adjacent Lamina Tissue in HL-Exposed Leaves Is Sensitive to Prevailing Humidity and the ABA Biosynthesis Inhibitor ABAmine.
(A) Colocalization of APX2 expression and H2O2 accumulation in the veinal tissue of HL-exposed leaves. A detached leaf from a Col-0/APX2LUC plant, previously infiltrated with ARU and luciferin (see Methods), was clamped into a CIRAS leaf chamber to maintain a constant temperature and low humidity (25% RH; see Methods). The leaf was then exposed to HL (PPFD 750 μmol m−2 s−1 for 45 min), after which images of reflected light from the leaf (a), resorufin fluorescence to locate H2O2 accumulation (b), and photon emission from luciferase activity from expression of APX2LUC (c) were captured (see Methods). In (a), the focal plane highlights the vein (V) and the surrounding darker area is the leaf lamina (L). In (b), the false red color denotes areas of high resorufin fluorescence, which is mainly in the vein but also may be immediately adjacent lamina; hence, the term periveinal is used. Note that APX2LUC expression in (c) is confined more tightly to the veins and includes the elongated cells seen in the reflected light image (a), consistent with expression of this gene in BSCs (Fryer et al., 2003; Ball et al., 2004). The images are from a single typical experiment. The yellow bar in the right panel denotes 100 μm.
(B) Resorufin fluorescence in HL leaves inhibited for foliar ABA biosynthesis. Col-0 detached leaves were infiltrated with 50 μM ABAmineSG and 40 μM ARU or ARU alone (closed and open circles, respectively) as in Methods. All leaves were then exposed to HL (750 μmol m−2 s−1 PPFD) for 60 min at low humidity in a leaf chamber as described for (A) and in Methods. Resorufin fluorescence from the veinal focal plane was quantified as described in Methods. The data are combined from three independent sets of measurements (±se; n = 3). The relative fluorescence values were normalized to the starting low light values for resorufin fluorescence in each leaf.
(C) Images of resorufin fluorescence in HL (750 μmol m−2 s−1) exposed leaves at low humidity (25% RH) and high humidity (80% RH) sealed in a leaf chamber as described for (A) and in Methods. The images are from a typical single experiment and leaves were infiltrated with ARU only as described for (A) and in Methods. The numbers under each panel refer to the time in minutes from the beginning of exposure to HL. In this focal plane, the periveinal region (V) and the adjacent leaf lamina (L) are as described for (A). False color yellow indicates resorufin fluorescence, and red color denotes higher resorufin fluorescence. The graphs show the mean values at each time point (±se; n = 3) for resorufin fluorescence, normalized to starting low light values, in the periveinal and lamina focal planes of HL-exposed leaves at low humidity (closed circles) and high humidity (open circles).
(D) Transverse sections of the mesophyll (M), bundle sheath (BS), and vascular parenchyma (VP; Kinsman and Pyke, 1998; Evert, 2006) of HL-exposed leaves at ×4000 magnification (the bar on each panel denotes 5 μm) infiltrated with CeCl3 prior to fixation, sectioning, and examination by transmission electron microscopy (see Methods). Cerium perhydroxide deposition (yellow arrows highlight prominent examples in the apoplast) denotes accumulation of H2O2. The leaves were exposed to HL (750 μmol m−2 s−1 PPFD) for 45 min at low humidity (right panel) or kept at growth PPFD at low humidity (left panel) as in the legend of (A), except that the leaves (one per plant) remained attached to the rest of plant when clamped into the leaf chamber. The two panels show typical images. More examples from two separate batches of plants can be seen in Supplemental Figure 5 online.
In a series of experiments with leaves exposed to HL at low humidity (Figures 4B and 4C), ARU-resorufin fluorescence in periveinal tissue was detected at 15 min after exposure to HL, although the peak of fluorescence varied between 15 and 30 min (cf. Figures 4B and 4C). A weaker secondary burst of resorufin fluorescence was detected at ∼90 min after HL exposure. In adjacent lamina tissue, a similar biphasic burst of fluorescence was detected; the second burst was of the same or greater magnitude than the first (Figure 4C). The biphasic changes in fluorescence were inhibited if leaves were placed in HL at high humidity (Figure 4C) or if the leaves were pretreated with ABAmineSG prior to HL exposure at low humidity (Figure 4B).
To verify accumulation of extracellular H2O2 in HL-exposed leaves, we used transmission electron microscopy to visualize cerium perhydroxide deposits, formed by reaction of infiltrated cerium trichloride (CeCl3) with H2O2 (Bestwick et al., 1997; Hu et al., 2005). Dark cerium perhydroxide deposits were detected in the apoplast adjoining cells of the bundle sheath and the vascular parenchyma (Figure 4D, right panel; see Supplemental Figure 5 online). At 150 μmols m−2 s−1 PPFD, weaker and less extensive deposits were observed, showing that the dense cerium perhydroxide accumulation was due to the HL treatments (Figure 4D, left panel; see Supplemental Figure 5 online). However, no staining was detected in the apoplast on the mesophyll side of BSCs nor in the apoplast surrounding mesophyll cells of the leaf lamina (Figure 4D; see Supplemental Figures 5 and 6 online). Previous work has shown that the electron-dense cerium perhydroxide deposits in Arabidopsis leaf cell walls are almost entirely caused by H2O2 (Solyu et al., 2005).
HL-Induced Accumulation of H2O2 in BSCs in Response to HL Is Rapid, Humidity Insensitive, and Directly Implicated in Signaling to the Nucleus
Previous studies showed that the chloroplasts principally of periveinal cells, but not exclusively of BSCs, accumulate H2O2 in response to HL (Fryer et al., 2003; Mullineaux et al., 2006). Infiltration of leaves using 3′ 3′ diaminobenzidine (DAB) penetrates cells and stains for H2O2 accumulation in chloroplasts (Fryer et al., 2003; Liu et al., 2007; Driever et al., 2008; Šnyrychová et al., 2008), appearing as a brown precipitate (Thordal-Christiansen et al., 1997). Using DAB, H2O2 accumulation in chloroplasts of periveinal cells was visualized after 10 min exposure (Figure 5A ), and the intensity of DAB staining in the veins of HL-exposed leaves was not susceptible to prevailing humidity (Figure 5B). Therefore, it was concluded that there was no direct role for leaf water status in the HL-induced H2O2 accumulation in the chloroplasts of periveinal cells, such as BSCs and flanking mesophyll cells.
Accumulation of Chloroplastic H2O2 in the Veins of HL-Exposed Leaves Is Humidity Independent but Does Direct Expression of APX2.
(A) DAB staining for H2O2 accumulation in veins of HL-exposed leaves at low humidity is visible after 10 min exposure to HL. Leaves were infiltrated with 5 mM DAB for 2 h prior to being clamped into a leaf chamber and exposed to HL (750 μmols m−2 s−1 PPFD) at low humidity (25% RH; see Methods and legend of Figure 4A; Fryer et al., 2002, 2003) for the times indicated. At the end of the HL exposure, leaves were fixed, destained, and photographed. The dotted line indicates the boundary between HL and LL exposed parts of the leaf.
(B) Higher magnification images from intact leaves shown in (A). The bar in the left panel denotes 20 μm. The elongated files of cells are BSCs (Kinsman and Pyke, 1998; Fryer et al., 2003) with stained chloroplasts clearly visible between 10 and 45 min of HL exposure.
(C) DAB staining in the veins of HL-exposed leaves is no different in low or high humidity. The leaves were treated as described in (A), except that they were kept at either 25% RH or 80% RH when clamped into the chamber and exposed to HL.
(D) APX2 expression is higher in mutants with compromised plastidial H2O2 scavenging capacity. APX2 transcript levels in srx1-1, srx1-2 (Rey et al., 2007), and vtc1-1 (Conklin et al., 1997) plants relative to the wild type (Col-0) exposed to HL (750 μmol m−2 s−1 PPFD) at growth humidity. APX2 cDNA levels were normalized against CYCLOPHILIN cDNA levels as described in Figure 1A and Methods. The data, expressed as normalized cDNA levels in the mutant relative to the wild type, are the means (±se) of three biological replicates totaling nine plants for Col-0, srx1-1, and vtc1-1 and 13 plants for srx1-2. The differences between mutants and the wild type were significant (P = 0.02, 0.058, and 0.022 from t tests for srx1-1, srx1-2, and vtc1-1, respectively).
BSC-specific APX2 expression was elevated in the sulfiredoxin 1 mutant srx1 and the vitamin C-1 deficient mutant vtc1, which are both partly defective in ROS scavenging capacity in the chloroplasts (Conklin et al., 1997; Rey et al., 2007; see Discussion) (Figure 5C). There was no greater degree of photoinhibition nor was it any less reversible than in wild-type plants (see Supplemental Figure 1 online). Other HL-expressed genes used in this study were unaffected in their expression in these mutants (see Supplemental Table 1 online). These observations show that while the accumulation of H2O2 in HL-exposed BSC chloroplasts plays no role in the activation of ABA accumulation, H2O2 has a BSC-specific function in the induction of APX2 expression under HL. The specificity of the effect of these mutants on APX2 expression may reflect the propensity for BSC chloroplasts to accumulate H2O2 more than other leaf tissues under HL conditions (Figure 5A; Fryer et al., 2003) and therefore affect BSC-specific gene expression more than genes expressed elsewhere in the leaf.
ABA Signaling Is Associated with Extracellular ROS Production and HL-Responsive Gene Expression
ABA signaling in guard cells involves the production of ROS at the plasma membrane (Murata et al., 2001; Kwak et al., 2003). At least two components of the guard cell ABA signaling network were proposed to be involved: OST1 (Mustilli et al., 2002; Li et al., 2006) and heterotrimeric G proteins (Joo et al., 2005; Li et al., 2006). Null mutant alleles of ost1 conferred a more than twofold reduction in APX2 expression under HL (Figure 6A ), but no consistent effect of the loss of OST1 was observed upon the expression of the five ABA- and HL-responsive genes studied in Figure 2C (see Supplemental Table 1 online). Throughout the HL treatment, stomatal conductance values for ost1-1 leaves were fourfold higher than for wild-type plants (see Supplemental Figure 7 online), which enhanced transpiration and a more rapid change in leaf water status than in the wild type.
HL-Responsive Gene Expression in ABA Signaling Mutants.
(A) Whole rosettes of ost1-1 and ost1-2 and wild-type plants (Landsberg erecta [Ler]) were exposed to HL (750 μmol m−2 s−1 for 45 min) at growth humidity, and APX2 cDNA levels were normalized with respect to those of CYCLOPHILIN (see legend of Figure 1A) and expressed as mutant relative to the wild type. All data shown (means ± se) are of two biological replications with a total of six to eight plants. The differences between the mutants and the wild type were significant; P = 0.03 and 0.02 from t tests for ost1-1 and ost1-2, respectively.
(B) APX2 transcript levels in HL exposed heterotrimeric G protein null mutants compared with wild-type plants. Rosettes of null mutants of GPA1 (gpa1-3 and gpa1-4), AGB1 (agb1-2 and agb1-9), and a double mutant (gpa1-4 agb1-2) were exposed to HL at growth humidity along with Col-0 and assayed for relative APX2 cDNA levels as described in (A). All data shown (means ± se) are of two biological replications with a total of six to eight plants. The differences between the mutants and the wild type are significant (P ≤ 0.05 from t test).
(C) to (F) Transcript levels of HL-responsive genes encoding lipocalin (C), RD20 (D), HSP17.6B-C1 (E), and HSP17.6C-C1 (F) assayed in HL-exposed rosettes (see legend of [A]) of gpa1-3, gpa1-4, agb1-2, and gpa1-4 agb1-2 relative to Col-0. All cDNA levels were normalized with respect to those of CYCLOPHILIN and expressed as mutant relative to the wild type. All data shown (means ± se) are of two biological replications with a total of six to eight plants. The histograms marked with asterisks are significant (P ≤ 0.05 from t test) between mutant and the wild type.
(G) Foliar H2O2 levels in low light and HL-exposed ost1-1 and wild-type (Ler) leaves. H2O2 levels were determined from cell-free acid extracts of fully expanded leaves of plants exposed to HL at growth humidity as described in (A) or from plants kept at growth PPFD and humidity (LL). The H2O2 amount in each sample was determined using an amplex red-based in vitro assay (see Methods). Data are the means (±se) of two separate experiments with four plants per experiment (n = 8). Three replicate determinations were carried per sample. The differences between the mutants and the wild type are significant (P ≤ 0.05 from t test).
(H) Foliar H2O2 levels in fully expanded outer leaves of gpa1-3, gpa1-4, and wild-type (Col-0) rosettes under low light and growth humidity conditions (see Methods). The methods were as in (G). Data are the means (±se) of two separate experiments with five plants per experiment (n = 10). Three replicate determinations were carried per sample. The differences between the mutants and the wild type are significant (P ≤ 0.05 from t test).
Total foliar H2O2 content in ost1-1 did not increase upon HL exposure, in contrast with an increase of 175% in wild-type plants (Figure 6G). Staining of HL-exposed leaves with DAB showed brown precipitate predominantly in periveinal tissue to equal levels in both ost1-1 and the wild type (see Supplemental Figure 8 online). We conclude that OST1 positively regulates both HL-induced APX2 expression and an increase in nonchloroplastic H2O2 levels, although this regulation may be confined to BSCs.
Null mutants (gpa1-3 and gpa1-4; Ullah et al., 2001; Jones et al., 2003) in genes encoding the Arabidopsis Gα subunit (G PROTEIN ALPHA1 [GPA1]) and the mutants agb1-2 (Jones et al., 2003), and agb1-9 in the gene ARABIDOPSIS G PROTEIN BETA1 (AGB1), encoding the Gβ subunit of the heterotrimeric G protein complex, all showed three- to fivefold stimulation of APX2 expression under HL conditions (Figure 6B). Furthermore, a gpa1-4 agb1-2 double mutant (Jones et al., 2003) showed the same increase in APX2 expression as gpa1-3 and gpa1-4 plants (Figure 6B). The HL-responsive genes coding for lipocalin, HSP17.6C-C1, HSP17.6B-C1, and RD20, showed a similar pattern of response to HL in the G protein mutants (Figures 6C to 6F). ELIP1 gene expression was not consistently significantly affected by the G protein null mutants (see Supplemental Table 1 online). There was no elevated expression of any of the test genes at ambient light intensity in the tested mutants, and during HL exposure, all tested mutants maintained stomatal conductance similar to that of wild-type plants (see Supplemental Figure 7 online).
Taken together, these data suggest that the heterotrimeric G protein complex is a negative regulator of HL-responsive gene expression in wild-type plants. Measurement of total foliar H2O2 content under low-light conditions revealed a 2.5-fold increase in the gpa1 null mutant (Figure 6H), and there was no increased DAB staining in the veins of low-light leaves of gpa1-4 compared to Col-0 (data not shown). We conclude that GPA1 negatively regulates the levels of extraplastidial H2O2 independent of any increase in light intensity. Since under HL conditions, the null mutants showed no differences from wild-type controls in foliar H2O2 content (data not shown) and DAB staining of periveinal chloroplasts (see Supplemental Figure 8 online), we conclude that HL acts by relieving the inhibitory effect by GPA1.
DISCUSSION
Rapid Initiation of Foliar ABA Biosynthesis in Response to HL under Low Humidity Conditions
We postulate that a rapid transient decline in leaf water content upon exposure to HL, at low humidity, initiates ABA biosynthesis and is a key step in coordinating the response of the leaf to these conditions. This is because leaves experience a sudden increase in transpiration rate brought about by exposure to HL (Fryer et al., 2003). This sudden increase in transpiration rate could cause a transient decline in leaf water potential since leaves have a high degree of hydraulic resistance to water flow, and hydraulic conductance is responsive to many environmental conditions, including prevailing light intensity and humidity (Sack and Holbrook, 2006; Levin et al., 2007; Sellin et al., 2008; Kim and Steudle, 2009).
In Arabidopsis leaves, NCED3, ABA2, AAO3, and ABA4 are prominently expressed in leaf veins (Nambara and Marion-Poll, 2005; Christmann et al., 2006; North et al., 2007), and recent immunolocalization studies showed that vascular parenchyma tissue is the site of foliar ABA synthesis (Endo et al., 2008). Petioles contain both vascular parenchyma cells and BSCs (Kinsman and Pyke, 1998; Evert, 2006; Endo et al., 2008) and have the capacity to synthesize ABA and express APX2 in these cell types, respectively (Fryer et al., 2003; Endo et al., 2008). ABA levels in osmotically stressed petioles in both low light and in the dark increased by a similar amount (Figure 1D; see Results), and DCMU had no effect on ABA accumulation in HL-exposed leaves at low humidity (Figure 1E). From these experiments, we conclude that in HL at low humidity, increased ABA accumulation requires a decline in leaf water content, but not active LEF or other light-associated signals.
The rise in foliar ABA levels in intact leaves exposed to HL and low humidity occurred at a rate ∼10-fold lower than in dehydration-stressed detached leaves (calculated from Endo et al., 2008 and data in Figure 1A). This difference is reflected in the strong induction of NCED3 expression during dehydration stress (Iuchi et al., 2001; Tan et al., 2003; Endo et al., 2008). By contrast, NCED3 transcript levels did not rise in HL-exposed leaves at growth or lower humidity (Figure 1A). However, in our experiments, NCED3 transcripts could be readily detected in well-watered plants at low light (see Supplemental Figure 3 online) using a similar RT-PCR technique (see Methods) to that of Endo et al. (2008), who could not detect this transcript under fully turgid conditions. This may reflect a difference in watering regimes, although it should be noted that in nced3-2, the levels of ABA under well-watered, low-light conditions were higher than in wild-type plants (see Supplemental Figure 3 online), suggesting that ABA levels can be maintained under nonstress conditions by NCED3-independent means. By contrast, under HL, foliar ABA levels did not rise in the mutant (see Supplemental Figure 3 online), underscoring the requirement for an extant capacity for foliar ABA biosynthesis for many aspects of the HL response (Figures 2 to 4⇑⇑).
The lack of NCED3 induction (Figure 1A), but the observed rise in foliar ABA levels (Figures 1A and 1B) and the requirement for an extant foliar ABA biosynthetic capability once HL is applied (Figures 2 to 4⇑⇑), suggest that NCED3 could also be subject to posttranslational regulation. This notion is supported by the observation that two different sized NCED3 isoforms of 64 and 58 kD were detected in dehydration-stressed Arabidopsis leaves (Endo et al., 2008), showing that posttranslational modification of NCED3 can occur. Posttranscriptional activation of foliar NCED by ethylene in cleavers (Gallium aparine) is proposed to occur concomitant with a slower induction of NCED transcription by auxin (Kraft et al., 2007).
Under high humidity conditions, ABA accumulation did not occur (Figure 1B). This could be due to a failure to activate ABA biosynthesis, since a transient change in leaf water content would be required. However, we cannot rule out that ABA catabolism is activated, since recently it has been shown that the vascular-located ABA 8'hydroxylase gene CYP707A3 is induced in 10 min of exposure to high humidity conditions (Okamoto et al., 2009).
A Capacity for Foliar ABA Biosynthesis Is Required for Maintenance of Photochemical Quenching and Acclimation to HL
The maximum expression of HL-responsive genes requires ABA (Figures 2A to 2C; Bechtold et al., 2008). This requirement may implicate >800 HL- and ABA-coresponsive genes (Figure 2D; see Supplemental Data Set 1 online), suggesting that ABA biosynthesis and signaling in different leaf tissues should be important for the response of the whole leaf to HL. This was the case since nced3-2 was unable to adjust to a daily exposure to 60 min of 10-fold HL (Figure 3A) and suffered a degree of irreversible photoinhibition (Figures 3B and 3C). NPQ was unaffected in the mutant (Figure 3D), but it displayed a lower capacity for photochemical quenching throughout these treatments (Figure 3D; see Results). Therefore, in wild-type plants, an extant capability for foliar ABA biosynthesis is required to maintain or induce additional metabolic capacity to dissipate excitation energy. Failure to do so results in damage to the photosynthetic apparatus, increased photooxidative stress, and accelerated foliar senescence. These symptoms were clearly observed in nced3-2 (Figure 3B) and are consistent with changes in chlorophyll fluorescence in senescing leaves (Jenkins et al., 1981). Increases in photochemical quenching observed in BSC chloroplasts after 30 min exposure to HL are humidity sensitive (Fryer et al., 2003), supporting the suggestion that the ABA-mediated induction of additional electron sink capacity in this tissue is occurring.
These observations provide a new role for foliar ABA biosynthesis in addition to those described for dehydration stress (Christmann et al., 2007; Endo et al., 2008), stomatal responses to low humidity (Xie et al., 2006), and exploitation by the bacterial pathogen Pseudomonas syringae to subvert host defenses (de Torres-Zabala et al., 2007).
ABA Signaling in Veins and Lamina in Response to HL at Low Humidity
For ABA to participate in the regulation of HL-expressed genes in a range of leaf tissues, it must be secreted from the vascular parenchyma and induce signaling responses in other cell types.
The interaction of ABA with BSCs is demonstrated by the ABA biosynthesis- and humidity-dependent biphasic production of extracellular H2O2 associated with these cells (Figure 4). This begins at 15 min for the first phase and a much weaker but evident second burst at 90 min after onset of HL (Figures 4B and 4C). This extracellular H2O2 production may be more precisely located in the intercellular spaces between BSCs and vascular parenchyma cells (Figure 4D; see Supplemental Figure 5 online). These observations are consistent in different experimental systems with descriptions of ABA-, excess light-, and ozone-stimulated signaling in guard cells, epidermal cells, and other leaf tissues requiring plasma membrane NADPH oxidase isoforms (Kwak et al., 2003; Davletova et al., 2005; Joo et al., 2005; Li et al., 2006). Moreover, APX2, which is inducible by exogenous ABA (Fryer et al., 2003; Rossel et al., 2006; Bechtold et al., 2008), is also inducible by provision of exogenous H2O2 to leaves (Karpinski et al., 1999; Bechtold et al., 2008), and is completely inhibited in excess light stressed leaves preinfiltrated with catalase (Karpinski et al., 1999), shows lowered expression in a double null mutant defective for NADPH oxidases D and F (Bechtold et al., 2008) and where NADPH oxidase activity is inhibited (Volkov et al., 2006). In other plant species, these observations are consistent with ABA treatment of maize (Zea mays) leaves, which causes an increase in apoplastic H2O2 levels (Hu et al., 2005; Jiang and Zhang, 2003) and similar distribution patterns of extracellular vascular H2O2 levels preceded by foliar ABA accumulation in the Mediterranean shrub Cistus albidus subject to summer drought (Jubany-Mari et al., 2009).
Mutants defective in OST1 protein kinase function (Mustilli et al., 2002) showed marked attenuation of APX2 induction in response to HL (Figure 6A), suggesting that OST1 positively regulates induction of APX2 expression under HL conditions. These observations are consistent with the ABI1 protein phosphatase 2C (PP2C)-mediated positive regulation of APX2 induction (Fryer et al., 2003) and its interaction with and activation of OST1 protein kinase (Yoshida et al., 2006). OST1 also may have a positive role in the production of extracellular H2O2 in HL-exposed leaves, similar to its proposed role in the production of ROS at the plasma membrane of guard cells (Mustilli et al., 2002). This is because total foliar H2O2 levels did not rise in response to HL treatment, in contrast with a near doubling in Col-0 (Figure 6G), but accumulation of H2O2 in BSC chloroplasts, estimated by DAB staining, appears to be no different between mutant and wild-type plants (see Supplemental Figure 8 online). Therefore, we suggest that OST1 positively regulates extracellular H2O2 production upon activation by ABA synthesized under HL conditions. In addition, in HL-exposed ost1-1, the normal induction of the HL-responsive genes (see Supplemental Table 1 online), other than APX2, and accumulation of H2O2 in BSC chloroplasts (see Supplemental Figure 8 online) rules out any pleiotropic affect in this mutant caused by enhanced stomatal conductance (see Supplemental Figure 7 online; Mustilli et al., 2002).
The induction of APX2 expression was enhanced by four- to fivefold over wild-type plants in HL-exposed GPA1 and AGB1 null mutants (Figure 6B). In the double mutant gpa1-4 agb1-2, there was no additive effect of these combined mutations (Figure 6B). These data suggest that in BSCs, the heterotrimeric G protein complex is a negative regulator of APX2 expression. Similar results were reported for Rab18 gene expression (Pandey et al., 2006). Under low-light conditions, the GPA1 null mutants had elevated levels of total foliar H2O2 (Figure 6H) in the absence of any increase in vein chloroplast H2O2 content (see Supplemental Figure 8 online). Thus, GPA1, and potentially the G protein complex, also negatively regulates extracellular H2O2 levels in veins under ambient light conditions.
In contrast with the situation in animal cells, activation of the Arabidopsis Gα subunit does not require a G-protein coupled receptor (i.e., specifically a receptor having guanine nucleotide exchange factor function since GPA1 is a rapid nucleotide exchanger) (Johnston et al., 2007; Temple and Jones, 2007). Indeed, the hydrolysis of GTP is probably the rate-limiting step in the plant G protein cycle. Therefore, induction of APX2 by ABA would include triggering a deactivation of some or all of the BSC GPA1 complement, which would release downstream signaling from negative regulation. This could include a release of extracellular H2O2 production from negative control, which could induce APX2 expression. RGS1, a seven-transmembrane-domain GTPase activating protein, controls the activation state of GPA1 (Chen et al., 2003). Based on the current body of evidence, control of the GTPase activating protein activity of RGS1 occurs upon glucose binding or by binding of some other, but related, photosynthate product.
H2O2- and Redox-Mediated Retrograde Signaling from Chloroplasts Is Independent of Humidity
The above considerations show that ABA and its biosynthesis in leaves are necessary, but not sufficient, for the induction of HL-responsive genes. BSC chloroplasts and those of flanking mesophyll cells accumulate H2O2 within 10 min of exposure of the leaf to HL (Figures 5A and 5B), and this accumulation is independent of prevailing humidity (Figure 5C). This suggests that any ABA signal from vascular parenchyma cells does not affect the accumulation of H2O2 in BSC chloroplasts, which is likely to result from the photoreduction of O2 at photosystem I (Fryer et al., 2003).
The induction of APX2 expression by mild osmotic stress of isolated petioles requires light (Figure 1C) and is associated with a lowering of the oxidation state of QA and the operating efficiency of PSII photochemistry (Table 1). The change in these parameters and the repeated observation that the LEF inhibitor DCMU suppresses the induction of a large majority of nuclear-encoded HL-responsive genes (Karpinski et al., 1997; Rossel et al., 2002; Yabuta et al., 2004; Bechtold et al., 2008), including APX2 expression in petioles (Chang et al., 2004), is consistent with a requirement for LEF to provide additional redox and ROS signals sourced from BSC chloroplasts.
Null mutants of SRX1, which codes for sulforedoxin, a key enzyme component of the plastidial peroxiredoxin-based H2O2 scavenging system in Arabidopsis leaves (Rey et al., 2007), showed a significant elevation of APX2 expression (Figure 5D). Similarly, vtc1-1, which has depleted levels of ascorbic acid (Conklin et al., 1997), also showed a marked stimulation of only APX2 expression upon exposure to HL (Figure 5C). Neither set of mutants showed any significantly increased sensitivity to the HL exposure as determined by chlorophyll fluorescence measurements (see Supplemental Figure 1 online). This discrepancy is explained by the observation that partial loss of ROS scavenging capacity occurs in every tissue, but under these moderate HL treatments, an effect is only noted for redox- and ROS-responsive APX2 in BSCs which readily accumulate chloroplastic H2O2.
The other HL-responsive genes used in this study did not give a consistent response across the mutants (see Supplemental Table 1 online), but this is consistent with the lack of accumulation of H2O2 in any other leaf tissue under these mild HL conditions. No doubt more extreme photooxidative stress would provoke alterations in the expression of a much wider group of ROS- and redox-responsive genes. Nevertheless, many of these HL-responsive genes still require a chloroplast-sourced redox- or LEF-associated signal (Karpinski et al., 1997; Rossel et al., 2002; Ball et al., 2004; Bechtold et al., 2008).
In summary, along with previous observations (Karpinski et al., 1997, 1999; Fryer et al., 2003; Ball et al., 2004; Chang et al., 2004), the data presented here confirm that in BSCs, APX2 expression requires an increased chloroplast oxidative state and active LEF to activate its expression, along with a requirement for an extracellular ABA signal.
Interdependence of a Transcellular ABA Signal and Chloroplast Retrograde Signals in Leaves Responding to HL
In the BSCs of HL exposed leaves, all the key signaling events involving H2O2 sourced from chloroplasts and ABA secreted from neighboring vascular parenchyma cells is complete within 30 min of exposure. This series of signaling events in HL-exposed BSCs supports the generic hypothesis proposed by Pfannschmidt et al. (2009), who suggested that retrograde signals from chloroplasts may typically act by merging with an external tissue-sourced signal.
In other leaf tissues, the topology of combined retrograde signaling and ABA signaling networks may be different. This is evidenced by less extensive and slower extracellular H2O2 production across leaf tissues (Figure 4B), which may reflect a later or less responsive interaction with external ABA, and no evidence for an involvement of OST1 in signaling and differing responses to G protein mutants (Figure 6). In addition, most leaf cell types do not accumulate chloroplastic H2O2, but most HL-responsive genes require an active LEF, supporting the argument that retrograde redox signals may vary from tissue to tissue (Ball et al., 2004).
METHODS
Growth of Plants
Arabidopsis thaliana genotypes (Col-0, Ler, or C24) and mutants used were grown in a peat-based compost in a controlled environment room under an 8/16-h light/dark cycle at a PPFD of 150 μmol m−2 s−1, 22°C ± 1°C, and a RH of 50%. All plants studied were at 5 to 6 weeks after germination.
Genotypes
The mutants used and their ecotype background are as follows: APX2LUC (Col-0; Karpinski et al., 1999); nced3-2 (Col-0; see below); sto1-1 (C24; Ruggiero et al., 2004); aba2-11 and aba2-14 (Col-0; Gonzalez-Guzman et al., 2002; Barrero et al., 2006); srx1-1 and srx1-2 (Col-0; Rey et al., 2007); vtc1-1 (Col-0; Conklin et al., 1997), ost1-1 and ost1-2 (Ler; Mustilli et al., 2002); gpa1-3 and gpa1-4 (Col-0; Jones et al., 2003); agb1-2 and agb1-9 (Col-0; Jones et al., 2003; see below); and gpa1-4 agb1-2 (Jones et al., 2003). All genotypes were confirmed.
Isolation of Mutants nced3-2 and agb1-9
The nced3-2 mutant line (N331021) for NCED3 was obtained from the Nottingham Arabidopsis Stock Centre. The T-DNA insertion site is in the coding region of the unique exon of NCED3 (see Supplemental Figure 3 online) and is identical to the mutant recently published by Urano et al. (2009).
The mutant agb1-9 was isolated from a screen of Arabidopsis displaying agb1-2 null mutant phenotypes. agb1-9 contains a W313-to-stop codon mutation that causes premature termination of AGB1 synthesis, rendering the truncated protein nonfunctional.
Exposure of Leaves to HL
Fully expanded outer leaves detached from the plant were exposed to a PPFD of 750 (±50) μmol m−2 s−1 at a RH of 25% (low humidity) or 80% (high humidity) using a fiber optic white light source of 1 cm diameter as described previously (Fryer et al., 2003). The temperature (27°C ± 1°C) and humidity during HL exposure was kept constant using a CIRAS leaf chamber (PP Systems) as described previously (Fryer et al., 2003). Prior to clamping in the chamber, detached leaves were infiltrated with luciferin (see below), DCMU (30 μM; Sigma-Aldrich), DAB, or ARU for staining H2O2 (see below) via their transpiration stream for 2.5 h in low-light conditions as described previously (Fryer et al., 2002).
Whole rosettes were exposed to a PPFD of 750 (±100) μmol m−2 s−1 at ambient (growth) RH. This was done using lamps and water filters as described previously (Karpinski et al., 1999). Under the lamps, leaf temperatures increased by ≤5°C over 90 min, the maximum time of exposure. Temperature increases of <6°C do not induce expression of any of the genes in this study (Bechtold et al., 2008).
Osmotic Stress Treatment of Petioles
Solutions of (0.1 to 0.5 M) PEG-400, sorbitol, mannitol, or water were infiltrated into detached fully expanded leaves of APX2LUC/Col-0 (Karpinski et al., 1999) for 2 h as described below for the infiltration of ARU. Then, an ∼15 to 20 mm length of petiole was cut from each leaf and immersed for a further 1.5 h in the same solution also containing 1 mM D (−) luciferin under low light or in the dark. At the end of this period, luciferase activity was imaged and the petioles harvested for determination of their ABA content.
Chlorophyll Fluorescence Imaging
Chlorophyll fluorescence parameters were measured using a Flourimager chlorophyll fluorescent imaging system (Technologica). The application of preprogammed regimes of actinic growth light exposure times, saturating light pulses, and dark periods and the calculation and imaging of the parameters Fv/Fm, Fq′/Fm′, Fv′/Fm′, and NPQ (Baker, 2008) were performed automatically by the Fluorimager's software. qL was calculated postmeasurement from images of Fq′/Fm′, Fo′, and F′.
Visualization of H2O2 Accumulation and Its Measurement in Total Foliar Extracts
Use of ARU
ARU (a proprietary derivative of 10-acetyl-3,7-dihydroxyphenoxazine; Invitrogen; Zhou et al., 1997) was used to detect accumulation of extracellular H2O2 in the veins and periveinal regions of Arabidopsis leaves. It should be noted that this dye is different in its penetration properties compared with Amplex Red (10-acetyl-3,7-dihydroxyphenoxazine) and depending on which tissues in the leaf are being examined (Šnyrychová et al., 2008). ARU was dissolved in dimethylsulfoxide and then diluted into 50 mM sodium phosphate buffer, pH 7.5, to give a final concentration of 40 μM for infiltration into leaves. A detailed description of how dyes can be infiltrated into leaves via the transpiration stream is given in two publications (Fryer et al., 2002; Driever et al., 2008). Infiltration was performed into fully expanded detached leaves for ARU for 2 h under low light (20 to 50 μmol m−2 s−1) conditions. The leaves were cut from the plant under liquid to ensure no airlock formed at the cut petiole surface. At the end of the infiltration period, the leaf was then clamped into the CIRAS leaf chamber and exposed to the HL conditions with the 500- to 570-nm region of the spectrum being removed using a Rose Pink filter (Lee Filters). This was to prevent absorption of these wavelengths by ARU during the course of the experiment and to prevent photodegradation of ARU. Resorufin (7-hydroxy-3H-phenoxazin-3-one) derivative, the fluorescent product obtained when ARU reacts with H2O2 (Zhou et al., 1997), was imaged using a Peltier-cooled CCD camera (Wright Instruments) with a custom-built LED lighting system (Technologica) that provides a constant and homogenous excitation light with a maximum at 466 nm. Detection of fluorescence emission from resorufin was achieved by placing a 590-nm band-pass filter (Edmund Optics) in front of the lens of the CCD camera. The camera was controlled by FluorImager V1.01 software (Technologica), which was purposely designed for image acquisition (576 × 384 pixels) and control of exposure time, which was 5 s for all experiments. Acquired images were processed using ImageJ software (Abramoff et al., 2004).
Use of DAB
DAB (5 mM at pH 3.8; Fryer et al., 2002; Driever et al., 2008) was infiltrated into detached leaves as described above for ARU. The leaves were placed in the CIRAS leaf chamber and subjected to HL at low or high humidity. At the end of the experiment, leaves were infiltrated with lacto-glycerol-ethanol to fix them and remove chlorophyll (Fryer et al., 2002; Driever et al., 2008) prior to imaging.
Use of CeCl3
CeCl3 reacts with hydrogen peroxide to form cerium perhydroxides, which forms electron-dense deposits that can be visualized by transmission electron microscopy (Bestwick et al., 1997). Control and HL-exposed leaves at low humidity were rapidly sliced into ∼10-mm strips and immediately vacuum infiltrated with 5 mM CeCl3 in 50 mM MOPS buffer, pH 7.2, and incubated for 2 h. Leaf segments (3 × 3 mm) were then prefixed in 2.5% (v/v) glutaraldehyde buffered with 0.1 M cacodylate buffer, pH 7.2, at 4°C overnight followed by three 10-min rinses in 0.1 M Na cacodylate buffer, pH 7.2. The leaves were then postfixed in 1% (v/v) buffered osmium tetroxide for 1 h at 4°C followed by three washes in distilled water. The leaves were then stained in 1% (w/v) uranyl acetate for 1 h followed by dehydration in an ethanol series (30, 50, 70, 90, and 100% ethanol for 10 min each). They were then treated with propylene oxide for 30 min prior to transfer to TAAB LV resin (TAAB Laboratories Equipment): propylene oxide (1:1) for 12 h in sealed vials. The leaves were then transferred to 100% TAAB LV for 2 h followed by fresh TAAB LV for 24 h. Finally, the leaf segments were transferred to fit molds with TAAB LV and cured for 16 h at 60°C. Sections (90 nm) were cut with a diamond knife, stained with lead citrate, and viewed with a Jeol JEM-1400 electron microscope at an accelerating voltage of 80 kV.
Total Foliar H2O2 Measurements
Hydrochloric acid (0.1 M) extracts from 100 mg of fresh Arabidopsis tissue ground in liquid nitrogen were prepared and purified over activated charcoal as described by Creissen et al. (1999). H2O2 concentrations in purified extracts were determined using an Amplex Red assay kit (Invitrogen) according to the manufacturer's instructions.
Imaging of Luciferase Activity in Leaf Veins
High-magnification imaging of HL-induced luciferase activity in Col-0/APX2LUC was as previously described (Fryer et al., 2003). Plants were sprayed with 1 mM D (−) luciferin (Biosynth), or it was infiltrated into detached leaves as described above for ARU and for HL experiments. Images were taken using the same CCD camera system and processed as described for resorufin (see above) using an image acquisition time of 2 min.
ABA Measurements
All measurements of foliar ABA content were performed on leaves clamped in the CIRAS chamber, and only one leaf per plant was used. After treatments, leaves were frozen in liquid nitrogen, freeze-dried, and ABA content measured using a radio-immunoassay procedure (Quarrie et al., 1988), as described previously (Xie et al., 2006). Following exposure to osmotic stress, individual petioles were immediately frozen in liquid nitrogen and similarly assayed for ABA content.
RNA Extraction and Quantitative Real-Time RT-PCR
RNA was extracted from 100 to 200 mg of fully expanded leaves using Triazol reagent (Sigma-Aldrich) according to the manufacturer's instructions, except that an additional ethanol precipitation step was included at the end of the procedure to ensure the RNA was of appropriate quality (A260/A280 ≈ 2.0). RNA (3 μg) was treated with RNase-free DNase1 (Ambion) and the absence of contaminating genomic DNA confirmed using a PCR test as described previously (Bechtold et al., 2008). RNA (2 μg) was used to make random-primed cDNA as previously described (Ball et al., 2004), except that the MuMLV reverse transcriptase was purchased from Fermentas. Quantitative real-time PCR was performed as described previously (Ball et al., 2004; Bechtold et al., 2008) using a cybergreen-fluorescence based procedure with reagents purchased in kit form from Sigma-Aldrich. Relative cDNA levels between two sets of threshold cycle (Ct) values were calculated using the ΔΔCT method (Kubista et al., 2006) and normalized with respect to relative cDNA levels for CYCLOPHILIN. This reference gene was chosen because it shows unchanging transcript levels in excess light-exposed leaves and in conditions where foliar ABA content varies (Rossel et al., 2006). The primers used in this study for quantitative RT-PCR are given in Supplemental Table 2 online.
Bioinformatics
Data for Affymetrix ATH1 GeneChips were downloaded from the GEO repository (http://www.ncbi.nlm.nih.gov/sites/entrez?db=gds) and the NASCArrays database (http://affymetrix.Arabidopsis.info/narrays/experimentbrowse.pl). The HL exposure data (GEO accession number GSE7743) and ABA treatment data (NASCARRAYS-176) were normalized using GCRMA procedures in the Bioconductor package within the R statistical environment (Gentleman et al., 2004; Wu et al., 2004). Values for replicate arrays were averaged and ratios calculated between treatment and control; these ratios were used to rank genes for the responses to 3 h HL exposure and 3 h following ABA treatment. Significant similarities between these two ranked lists were then calculated using the Ordered List Bioconductor package (Lottaz et al., 2006). The genes commonly upregulated and commonly downregulated that contributed to the significant weighted similarity score were clustered together with other ABA treatment time points from the NASCARRAYS-176 data set and additional ABA treatment data sets (GSE7112 and GSE6171). Hierarchical clustering was performed using CLUSTER (Eisen et al., 1998) and visualized with the program TREEVIEW (Eisen et al., 1998). Complete linkage clustering using an uncentered Pearson correlation was applied after the genes were first ordered by self-organizing maps to produce better-arranged clusters.
Accession Numbers
The Arabidopsis Genome Initiative locus identifiers of genes used or mentioned in this study are as follows: AAO3, At2g27150; ABA2, At1g52340; ABA4, At1g67080; ABI1, At4g26080; ABI2, At5g57050; AGB1, At4g34460; APX2, At3g09640; BAP1, At3g61190; CYCLOPHILIN, At2g29960; ELIP1, At3g22840; HSP17.6B-C1, At2g29500; HSP17.6C-C1, At1g53540; GPA1, At2g26300; LIPOCALIN, At5g58070; LHCB1.2, At1g29910; NCED3, At3g14440; OST1, At4g33950; PDF1.2, At5g44420; PR1, At2g14610; RD20, At2g33380; SRX1, At1g31170; VTC1, At2g39770; ZAT10, At1g27730; and ZAT12, At5g59820.
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. Dark-Adapted Fv/Fm Measurements on Col-0, srx1-1, srx1-2, and vtc1-1 before, Immediately after HL Exposure, and 24 h Later.
Supplemental Figure 2. Expression under HL Conditions of Genes Controlled by the GUN1/ABI4 and 1O2 - Related Chloroplast-to-Nucleus Retrograde Signaling Pathways.
Supplemental Figure 3. Characterization of nced3-2.
Supplemental Figure 4. Resorufin Fluorescence from the Periveinal Region of Amplex Red Ultrainfiltrated Leaves Exposed to Fivefold HL at Low Humidity.
Supplemental Figure 5. Further Electron Micrographs of CeCl3-Stained High Light and Control Leaf Sections through Vein Tissue.
Supplemental Figure 6. Further Electron Micrographs of CeCl3-Stained High Light and Control Leaf Sections through Vein Tissue.
Supplemental Figure 7. Stomatal Conductance of ABA Signaling Mutants.
Supplemental Figure 8. DAB Staining Showing H2O2 Accumulation in the Veins of ost1-1 and gpa1-4 Compared with Wild-Type Controls.
Supplemental Table 1. HL-Responsive Gene Expression in Mutants Used in This Study Where No Significant Effect Was Discerned.
Supplemental Table 2. A List of the Primers and Their Target Genes Used for qRT-PCR in This Study.
Supplemental Data Set 1A. A List of Genes Whose Expression Is Upregulated by Both High Light and ABA Treatment.
Supplemental Data Set 1B. A List of Genes Whose Expression Is Downregulated by Both High Light and ABA Treatment.
Acknowledgments
We thank Jeffrey Leung (Institut Science Vegètale, Gif-sur-Yvette, France), Hisashi Koiwa (University of Texas A&M), José Luis Micol (Universidad Miguel Hernández, Alicante, Spain), Pascal Rey (Centre National de la Recherche Scientifique, Commissariat à l'Énergie Atomique (CEA), Université de la Méditerranée, France), and the Nottingham Arabidopsis Stock Centre for the mutants used in this study and Nobutaka Kitahata (University of Tokyo, Japan) for the provision of ABAmineSG. We thank Julian Theobald (Lancaster University) for skilled technical assistance in the ABA measurements. Electron microscopy was carried out by Peter Splatt and Gavin Wakley at the Exeter Bioimaging Centre. A.M.J. thanks Paul Reeves for isolating the agb1-9 mutant. This work was supported by a grant to N.R.B., W.J.D., and P.M.M. from the UK Biotechnology and Biological Sciences Research Council (BBSRC). G.G.-V. acknowledges the support of a Ministerio de Educación y Ciencias (Spain; EX2005-1107) and a European Union Marie Curie Fellowship (041283). K.S. acknowledges the support of a BBSRC Research Studentship. This work was also supported by grants from National Institute of General Medical Sciences (R01GM065989), the Department of Energy (DE-FG02-05er15671), and the National Science Foundation (MCB-0723515 and 0718202) to A.M.J.
Footnotes
-
The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: Philip M. Mullineaux (mullin{at}essex.ac.uk).
-
↵1 These authors contributed equally to this work.
-
↵[W] Online version contains Web-only data.
- Received June 16, 2008.
- Revised July 1, 2009.
- Accepted July 8, 2009.
- Published July 28, 2009.
References
