Introduction

Aldehydes occur as intermediates in various cellular pathways such as carbohydrate, amino acid, lipid and phenylpropanoid metabolisms. They can be formed also by lipid peroxidation, ascorbate autoxidation, or cytochrome P450s (O’Brien et al. 2005). Aldehydes are reactive molecules and have potential to modify proteins and nucleic acids (Burcham 1998; O’Brien et al. 2005). Due to this reactivity, they, at low concentrations, can act as signaling molecules for inducing stress defense genes (Alméras et al. 2003; Weber et al. 2004), while at higher concentrations they may exert cytotoxicity. In animal studies, aldehydes have been shown to cause mutation, cancer and cell-degenerating diseases (Comporti 1985; Esterbauer et al. 1991; Arlt et al. 2002; Nair et al. 2007).

In plants, there is increasing evidence for the toxicity of aldehydes in environmental and biotic stresses. Lipid peroxide (LOOH)-derived aldehydes, detected as thiobarbituric acid reactive substances, increased in association with biotic and abiotic stress-induced damages that were caused by bacterial infection (Muckenschnabel et al. 2001), ozone (Sakaki et al. 1983), chilling (Hodgson and Raison 1991), UV-B (Panagopoulos et al. 1992), heat (Mishra and Singhal 1992), intense light (Sharma and Singhal 1992), and drought (Moran et al. 1994). Such damage could be suppressed by aldehyde-scavenging enzymes when they were genetically enhanced in transgenic plants, as follows: Aldehyde reductase from Medicago sativa (alfalfa) overexpressed in Nicotiana tabacum (tobacco) improved their tolerance to drought stress (Oberschall et al. 2000), methyl viologen and UV-B (Hideg et al. 2003). Aldehyde dehydrogenase isozymes from Arabidopsis thaliana improved the tolerance of transgenic A. thaliana plants to NaCl, heavy metals, methyl viologen, and H2O2 (Sunkar et al. 2003). Vice versa, their knockout mutants were more sensitive to dehydration and salt than the wild type (Kotchoni et al. 2006). The observed protection was attributable to the scavenging of aldehydes because the level of thiobarbituric acid reactive substances was decreased by the overexpression of these enzymes and increased by the deficiency. Methylglyoxal, a by-product aldehyde in the triose-phosphate metabolism, is increased by NaCl stress (Yadav et al. 2005). Overexpression of the scavenging enzyme glyoxalase I in tobacco suppressed the increase in methylglyoxal (Yadav et al. 2005) and improved the tolerance to NaCl (Singla-Pareek et al. 2003). These results demonstrated a toxic action of methylglyoxal. A novel enzyme NADPH:2-alkenal reductase (AER; EC 1.3.1.74) found in A. thaliana (Mano et al. 2002) also improved the tolerance of transgenic tobaccos to intense light and to methyl viologen (Mano et al. 2005). This suggested that α,β-unsaturated aldehydes (2-alkenals) were produced on the photooxidative treatment and caused damage because AER specifically reduces the C–C double bond of a 2-alkenal to form a saturated aldehyde (Mano et al. 2002).

The targets of aldehyde toxicity in cells have been investigated mainly in animal cells. Several enzymes in the energy metabolism, including those in mitochondria (Chen et al. 1998; Humphries and Szweda 1998) are sensitive to 4-hydroxy-(E)-2-nonenal (HNE), one of the most reactive aldehydes produced from LOOH (Esterbauer et al. 1991). In plants also, mitochondrial enzymes are affected by HNE, as follows: respiratory reactions in pea mitochondria with various electron donors were inactivated by HNE, among which the glycine-dependent respiration was the most sensitive (Millar and Leaver 2000). The target site was the lipoate moiety of H-protein in the glycine decarboxylase complex and of other lipoate enzymes (Taylor et al. 2002). HNE also inactivated alternative oxidase (Winger et al. 2005). Recently, evidence has been provided for that the endogenously produced HNE modifies mitochondrial proteins in oxidative-stressed A. thaliana plants (Winger et al. 2007).

Chloroplast components would also be the potential targets of aldehydes. Chloroplasts produce reactive oxygen species such as singlet oxygen and superoxide radical in the vicinity or in the thylakoid membrane (Asada 2006). Considering that the thylakoid lipids are rich in polyunsaturated fatty acids, LOOH formation via these reactive oxygen species (Comporti 1985) is very likely, and from LOOH, many aldehydes such as malondialdehyde (MDA) can be formed by non-enzymatic mechanisms (Blée 1998). Chloroplasts have also enzymes to produce aldehydes; lipoxygenases catalyzing the oxygenation of polyunsaturated fatty acids to form LOOH, and hydroperoxide lyase to cleave it (Farmaki et al. 2007) to form C6-aldehydes such as (Z)-3-hexenal and n-hexanal, typical ‘green’ volatile organic compounds (Matsui 2006). Yamauchi et al. (2008) very recently reported that in heat-stressed plants the chloroplast proteins OEC33 and LHCII were modified with MDA. This is direct evidence for the formation and action of MDA in chloroplasts. Furthermore, the phototolerance of the AER-overproducing tobaccos (Mano et al. 2005) implies that 2-alkenals also would damage chloroplast components.

In order to obtain new insights into the action of aldehydes in photosynthesis, we investigated the toxicity of various aldehydes to chloroplasts. Several kinds of aldehydes including MDA, methylglyoxal and HNE were compared. Using 2-propenal (acrolein), the most toxic aldehyde found, targets in chloroplasts were investigated.

Materials and methods

Chemicals

Acrolein, HNE and MDA were prepared by acid hydrolysis of acrolein diethyl acetal, HNE-diethylacetal (Alexis Biochemicals, care of Biolinks, Tokyo, Japan) and 1,1,3,3-tetramethoxypropane, respectively. 4-Hydroxy-(E)-2-hexenal (HHE) was purchased from Cayman Chemical (Ann Arbor, MI, USA). Other aldehydes were of reagent grade. Glyceraldehyde 3-phosphate (GAP), ribose 5-phosphate, ribulose 5-phosphate and ribulose 1,5-bisphosphate were from Sigma Japan (Tokyo). Sedoheptulose 1,7-bisphosphate was enzymatically synthesized (Tamoi et al. 2005). Sodium [14C]-bicarbonate was from PerkinElmer Japan (Tokyo).

Preparation of chloroplasts and treatment of them with aldehydes

Intact chloroplasts were prepared from field-grown Spinacia oleracea L. (cv. Akution; Sakata Seed, Yokohama, Japan) leaves (Mano et al. 2001). All procedure was done at 4°C. After purified by Percoll (GE Biomedicals, Tokyo, Japan) density gradient centrifugation, chloroplasts were suspended in 0.3 M sorbitol, 10 mM NaCl, 1 mM MgCl2, 1 mM ascorbic acid (Asc), 0.5 mM diethylenetriamine-N,N,N′,N′′,N′′-pentaacetic acid, 0.5 mM sodium pyrophosphate, and 50 mM Hepes-NaOH, pH 7.6 (chloroplast medium). Chloroplast suspension (2 mg chlorophyll (Chl) ml−1) was incubated in 2 mM aldehyde in darkness at 25°C for 8 min, diluted with 9 volumes of chloroplast medium, and chilled on ice until assays.

Assays

Chloroplasts at 30 μg Chl ml−1 in chloroplast medium, supplemented with 10 mM NaHCO3, 0.05 mM Na-phosphate and 0.5 mM GAP, were illuminated at 2,000 μmol photons m−2 s−1 with white light from a tungsten lamp at 25°C, and the O2 evolution rate was monitored with a Clark-type O2 electrode (Hansatech, King’s Lynn, UK). GAP was added to start the CO2-fixation smoothly. For CO2-fixation assay, Na [14C]-HCO3 at 10 mM was used in the above medium. The reaction was stopped by adding 37 volumes of ethanol. After removing untreated CO2 by acidifying the medium with glacial acetic acid, incorporation of radioactivity into organic acids was determined. For the photoreduction of 3-phosphoglyceric acid (PGA), chloroplasts at 30 μg Chl ml−1 in chloroplast medium, supplemented with 0.05 mM Na-phosphate and 1 mM PGA, were illuminated and the O2 evolution was determined as described above. For the electron transport assay, 30 μg Chl aliquots of the chloroplasts ruptured in 5 mM Hepes-NaOH, pH 7.6, were transferred to the reaction medium containing 50 mM Tricine-KOH, pH 7.5, 20 mM NaCl, 0.5 μM nigericin, 0.5 mM NADP+ and 10 μM ferredoxin (Fd) and the O2 evolution under illumination was determined as described above. Aldolase activity was determined with an enzyme-coupled assay (Haake et al. 1998). Assay media for fructose-1,6-bisphosphatase (FBPase; Charles and Halliwell 1980) and sedoheptulose-1,7-bisphosphatase (SBPase; Harrison et al. 1998) contained DTT at 1.0 and 7.2 mM, respectively, for fully reducing the redox-regulated thiols. Enzymes were pretreated in the assay medium for 10 min, and then activities were determined in the same medium. Similarly, assay media for glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Mano et al. 2001) and phosphoribulokinase (PRK; Porter et al. 1986) contained the reduced form of glutathione (GSH) at 10 mM. Enzymes were pretreated for 10 min. Rubisco was activated with Mg2+ and CO2 prior to the assay, and the incorporation of [14C]-CO2 to PGA was determined (Shen and Ogren 1992). Total glutathione [GSH + oxidized form of glutathione (GSSG)] was determined as follows: Chloroplast suspension was mixed with four volumes of 0.5% sulfosalicylic acid. After centrifugation, the collected supernatant was mixed with nine volumes of 100 mM Hepes-KOH, pH 7.4, containing 0.5 mM EDTA. The content of glutathione was determined by a cycling assay using glutathione reductase and 5,5′-dithiobis-2-nitrobenzoic acid (Roberts and Francetic 1993).

Results

Comparison of the toxicity of aldehydes

Toxicity of aldehydes was evaluated as an inhibition of the CO2-supported electron transport activity as determined by O2 evolution (designated ‘CO2-photoreduction’) in chloroplasts. Intact chloroplasts isolated from spinach leaves showed a CO2-photoreduction activity at 40–80 μmol O2 mg Chl−1 h−1. This activity was inhibited in a time- and concentration-dependent manner by 2-propenal (acrolein), HNE, and (E)-2-butenal (crotonaldehyde), the most reactive aldehydes known (Esterbauer et al. 1991) (Fig. 1). The lowered activity was not recovered by washing the chloroplasts with chloroplast medium, indicating irreversible inactivation. The envelope intactness as determined by the ferricyanide method (Heber and Santarius 1970) was unaffected by acrolein or HNE (97% intactness, before and after the treatment), indicating that the loss of CO2-photoreducing activity was not due to the disintegration of envelopes. Because the CO2-photoreduction here was determined as O2 evolution, it was possible that the evaluation of inactivation was affected by changes, if any, in the O2 budget. For example, if Asc peroxidase was inactivated, H2O2 would accumulate, leading to an increase in the O2 consumption (Mano et al. 2001). In such a case, O2 evolution due to CO2 fixation could be partially masked by the O2 consumption, leading to an overestimation of the inactivation. We confirmed that the [14C]-CO2 incorporation in chloroplasts was inactivated by acrolein in a concentration dependency similar to that for the O2 measurement (data not shown). In addition, Asc peroxidase was insensitive to acrolein (J. Mano, Science Research Center, and H. Mizoguchi, Graduate School of Agriculture, both Yamaguchi University, unpublished result). Thus, the decrease in the O2 evolution observed in Fig. 1 represented that of CO2-fixation.

Fig. 1
figure 1

Concentration dependence of the inactivation of the CO2-photoreduction in chloroplasts by acrolein, HNE and crotonaldehyde. Chloroplasts (2 mg Chl ml−1) in chloroplast medium were treated with an aldehyde at indicated concentrations in darkness for 8 min at 25°C, diluted with 9 volumes of chloroplast medium, and chilled on ice. CO2-photoreducing activity was stable after dilution, i.e., the dilution and chilling virtually stopped the action of aldehydes

Full size image

We selected 17 species of aldehydes (carbon chain length 1–9) that have been known to occur in living cells and evaluated their toxicity to photosynthesis (Table 1). Their occurrence in plants is summarized in the Table legend. Ten of them were the aldehydes of saturated carbon chain, ranging from C1 to C9, some containing additional hydroxy- or oxo-groups. Glycolaldehyde and glyceraldehyde, by-products of sugar metabolisms, were added to the list because they are inhibitors of PRK (Miller and Canvin 1989), although their occurrence in plants is uncertain. Seven of them were unsaturated aldehydes of C3–C9. Most of the tested aldehydes inactivated the CO2-photoreduction significantly, except for glycolaldehyde, glyceraldehyde, MDA and butyraldehyde. The toxicity of tested aldehydes was attributable to their aldehyde moiety because their corresponding alcohols (methanol, ethanol, propanol, 2-propenol, butanol, 2-butenol, n-hexanol, (E)-2-hexenol, n-nonanol, and (E)-2-nonenol) or carboxylic acids (formic acid, acetic acid, acrylic acid, propionic acid, butyric acid, (Z)-3-hexenoic acid, hexanoic acid, and nonanoic acid) showed no toxicity at 2 mM (data not shown). As expected, toxicity of the 2-alkenals acrolein, crotonaldehyde, (E)-2-hexenal and (E)-2-nonenal was significantly higher than that of their corresponding n-alkanals, i.e. propionaldehyde, butyraldehyde, n-hexanal and n-nonanal, respectively. This explains the stress-defensive effects of AER overexpressed in tobaccos (Mano et al. 2005), which converts (E)-2-alkenals to n-alkanals (Mano et al. 2002). There was no obvious correlation between the toxicity and the carbon chain length; acrolein was the strongest, followed by HNE, (E)-2-hexenal and HHE. Among the tested 2-alkenals, crotonaldehyde and (E)-2-nonenal were the weakest, causing 29% inactivation.

Table 1 Toxicity of various aldehydes on chloroplast photosynthesis
Full size table

As for C6-unsaturated aldehydes, (Z)-3-hexenal was as toxic as (E)-2-hexenal and more than HHE. This order of toxicity in these aldehydes was unpredictable from their chemical reactivity, and suggests a rapid metabolism of C6-aldehydes in the chloroplast (see “Discussion”).

MDA is contained in A. thaliana leaves at ca. 3 nmol (g fresh weight)−1 (Weber et al. 2004). Methylglyoxal occurs at 50–70 μM in non-stressed leaves and roots of several species of plants, and is raised up to 200 μM upon NaCl stress (Yadav et al. 2005), to cause cellular injury (Singla-Pareek et al. 2003). Glyceraldehyde and glycolaldehyde inhibit PRK (Miller and Canvin 1989). These oxygenated aldehydes had relatively weak toxicity (5–17% inactivation) compared with 2-alkenals. In all tested aldehydes, acrolein showed the strongest inactivation, and HNE the second strongest.

Calvin cycle is more sensitive to 2-alkenals than electron transport chain

In order to identify the targets of 2-alkenals in chloroplasts, we employed acrolein as a representative compound. Chloroplasts were treated with acrolein at various concentrations and the activities of partial photosynthetic reactions were determined (Fig. 2). After the treatment with 2 mM acrolein, the CO2-photoreducing activity was totally lost, but the thylakoid electron transport chain (from H2O to Fd) retained 95% activity. Crotonaldehyde or HNE at 5 mM also resulted in the total inactivation of CO2-photoreduction with more than 95% electron transport activity retained. Thus, the thylakoid electron transport chain is rather insensitive to 2-alkenals.

Fig. 2
figure 2

Effects of acrolein on partial reactions of photosynthesis in the chloroplast. Chloroplasts were treated with acrolein at indicated concentrations as in Fig. 1. Assays were completed within 2 h after treatment. An aliquot of 30 μg of Chl was used for each assay, as determined by the O2-evolution, as described in “Materials and methods”. Average of 3 runs. Control rates of O2 evolution supported by distinct acceptors (in μmol (mg Chl)−1 h−1) were as follows: CO2, 40 ± 3; PGA, 105 ± 7; NADP+, 145 ± 11

Full size image

The Fd-photoreducing activity was unaffected in the above experiments, but there was a possibility that the endogenous Fd was inactivated, so that the photoproduction of NADPH in chloroplasts was impaired. It was also unclear whether or not photophosphorylation was functioning because in the above experiment electron transport was determined in the presence of an uncoupler. In order to verify the supply of both NADPH and ATP to the Calvin cycle enzymes in the acrolein-treated chloroplasts, we determined the photoreduction of PGA, in which ATP-dependent 3-phosphoglycerate kinase and NADPH-dependent GAPDH are involved (Fig. 4, “H2O → PGA”). The PGA photoreduction was inactivated by 40%, but not totally, by 2 mM acrolein. This partial loss of the PGA-photoreduction can be explained by a partial inactivation of GAPDH (described below). The activity remaining at a 60% level indicated that both NADP+-photoreduction and photophosphorylation were active, at least partially.

Saturated aliphatic aldehydes such as propionaldehyde and n-nonanal also caused loss of the PGA photoreduction by 21 and 13%, respectively, but glycolaldehyde and glyceraldehyde did not (Miller and Canvin 1989). These results indicate that the Calvin cycle is more sensitive than the thylakoid electron transport chain to the LOOH-derived aldehydes.

2-Alkenals inactivate multiple enzymes in the Calvin cycle

It was highly probable that acrolein targeted the thioredoxin-regulated enzymes such as FBPase and SBPase. 2-Alkenals electrophilically attack the thiol group and readily form a Michael adduct (Esterbauer et al. 1991). Acrolein could modify the redox-regulated cysteines (Cys) in these enzymes when they were in the reduced state. Another possibility was that 2-alkenals reacted with the imidazole group in histidine and the ε-amino group of lysine (Lys) (Uchida 2005), causing the loss of function of the target proteins. Therefore multiple enzymes in the Calvin cycle are candidates for covalent modification by acrolein. In order to evaluate the acrolein-mediated inactivation of the Calvin cycle enzymes, we extracted stroma fraction from acrolein-treated chloroplasts and determined their activities (Fig. 3).

Fig. 3
figure 3

Inactivation of the Calvin-cycle enzymes by the acrolein treatment of chloroplasts. Chloroplasts were treated with acrolein, and the CO2-photoreduction was assayed as in Fig. 1. For enzyme assays, treated chloroplasts were ruptured by dilution with 9 volumes of medium for the subsequent enzyme assay and centrifuged at 6,500×g for 1 min; the resulting supernatant was then collected. Enzyme activities were determined as described in “Materials and methods”. Activities of 100%, corrected for the Chl content of the original chloroplasts (average and standard deviation of 3 runs), were as follows: aldolase, 280 ± 27 μmol mg Chl−1 h−1; FBPase, 238 ± 14 μmol mg Chl−1 h−1; GAPDH, 640 ± 88 μmol NADPH mg Chl−1 h−1; PRK, 397 ± 9 μmol NADPH mg Chl−1 h−1; Rubisco, 57.7 ± 2.8 μmol CO2 mg Chl−1 h−1; and SBPase, 67.4 ± 0.4 μmol phosphate mg Chl−1 h−1. The CO2 photoreduction rate before treatment was 63 μmol mg Chl−1 h−1 (average of two runs)

Full size image

Activities of the thioredoxin-regulated enzymes PRK, GAPDH, SBPase, and FBPase were determined under reducing conditions made by GSH or DTT (see “Materials and methods”). The treatment with 2 mM acrolein decreased the activities of these enzymes by 90, 73, 63 and 20%, respectively, of the untreated controls, even under the reduced condition. This indicated that acrolein modified these enzymes to the inactive forms that were not recovered by reduction. Aldolase and Rubisco were also inactivated by 48 and 35%, respectively. Thus a wide range of enzymes in the Calvin cycle was affected by acrolein, most probably via covalent modification of Cys, Lys and His residues although detailed inactivation mechanisms for these enzymes have yet to be investigated. Other thioredoxin-regulated enzymes such as Rubisco activase can also be targets, and their inactivation by acrolein could contribute to the loss of CO2-photoreduction.

Glutathione prevents the acrolein toxicity

The above results clearly demonstrated that photosynthesis in chloroplasts is potentially susceptible to aldehydes. We then examined the possibility that GSH protected photosynthesis against acrolein because it has been reported that, in human plasma, GSH prevented the protein modification by 2-alkenals (O’Neill et al. 1994). When chloroplasts were treated with acrolein, the glutathione pool was decreased faster than was photosynthesis inactivated (Fig. 4). This was probably due to that GSH formed the Michael adduct with acrolein faster than were target proteins inactivated. Indeed, supplementation of GSH to the chloroplast suspension suppressed the acrolein-induced inactivation of photosynthesis (Fig. 5). We expected that a gluathione-S-transferase might mediate the scavenging of acrolein, but no enzyme activity to catalyze the glutathione-dependent scavenging of acrolein or HNE was detectable in spinach leaves (data not shown). Scavenging of acrolein by GSH was most probably due to its chemical action because another thiol compoud DTT at 10 mM also reduced the acrolein toxicity to photosynthesis i.e., 60% activity was retained after 2 mM acrolein treatment. In contrast, Asc at 10 mM did not show any protective effect (Fig. 5). These results indicate that GSH, but not Asc, provides as the primary defense against 2-alkenals in plant cells.

Fig. 4
figure 4

Acrolein decreases glutathione in chloroplasts. Chloroplasts were treated with 0.5 mM acrolein as in Fig. 1. At the indicated time points, 40 μl aliquots were sampled and mixed with 200 μl sulfosalicylic acid solution [5% (v/v)], for determination of total glutathione (see “Materials and methods”). In a separate incubation, CO2 photoreduction in treated chloroplasts was assayed as in Fig. 1. Control values for 100% were 57.1 μmol CO2 mg Chl−1 h−1 and 81.6 nmol glutathione (GSH + 2 × GSSG) mg Chl−1, which corresponded to the stromal concentration of 3.5 mM (stroma volume 23.0 μl mg Chl−1; Heldt et al. 1973), 90% of which in the reduced form (average of two runs)

Full size image
Fig. 5
figure 5

Effects of GSH and Asc on the acrolein-induced inhibition of photosynthesis. Chloroplasts were incubated in 10 mM GSH or 10 mM Asc in chloroplast medium for 2 min at 25°C, then acrolein was added to the medium, to give the indicated concentration. After 8 min incubation, CO2 photoreduction was determined as in Fig. 1. Control value for 100% was 50.6 μmol CO2 mg Chl−1 h−1. The CO2-photoreduction activity was not affected by 10 min-incubation in 10 mM GSH or Asc (average of two runs)

Full size image

Discussion

2-Alkenals are more toxic to photosynthesis than other types of stress-related aldehydes

The present results show that Calvin cycle enzymes are potential targets of LOOH-derived 2-alkenals, as are mitochondrial respiration enzymes (Taylor et al. 2002, Winger et al. 2005). When the aldehydes of the same carbon chain length were compared, 2-alkenals showed higher toxicity than n-alkanals for C3-, C4-, C6- and C9-aldehydes. These results can explain our previous observation that AER protected the transgenic tobaccos against photooxidative stress (Mano et al. 2005) if acrolein, crotonaldehyde, (E)-2-hexenal, HHE, (E)-2-nonenal or HNE was increased by the stress treatments.

C6-aldehydes are typical ‘chloroplast aldehydes’, and hence the effects of their individual species are quite important. One unexpected and interesting result was that HHE showed lower toxicity than (E)-2-hexenal (Table 1) in spite that the former has a higher electrophilicity due to the hydroxyl group at C4-position (Esterbauer et al. 1991). We infer that HHE might be scavenged in chloroplasts; there should be a scavenging mechanism specific to HHE because this highly reactive aldehyde is constitutively formed in chloroplast as an oxidized product of (Z)-3-hexenal (Kohlmann et al. 1999). Another unexpected result was that (Z)-3-hexenal showed a similar strength of toxicity to (E)-2-hexenal (Table 1) in spite that the former is obviously less electrophilic, because of the lack of conjugated double bonds, than the latter. One explanation is that (Z)-3-hexenal might be very rapidly converted to more reactive compounds such as 4-peroxy-(E)-2-hexenal by a peroxygenase (Kohlmann et al. 1999) or (E)-2-hexenal by the enzyme 3Z:2E-enal isomerase (Noordermeer et al. 1999), although their occurrence in chloroplasts has not been verified. Further investigations into the metabolism of C6 aldehydes and the action of (Z)-3-hexenal on photosynthetic reactions are required.

MDA has been recognized to be relevant to environmental stress in plants (see “Introduction”). As compared with 2-alkenals such as HNE, however, it showed only weak toxicity to chloroplasts. This is not surprising when one considers that the reactivity of MDA is much lower than that of HNE (Esterbauer et al. 1991). MDA can be as harmful as 2-alkenals when the tissue content of the former becomes 10 to 20-fold higher than those of the latter, as in the infected Phaseolus vulgaris leaves (Muckenschnabel et al. 2001). Another oxygenated C3 aldehyde methylglyoxal is obviously a major toxin in NaCl stress because glyoxalases improved the stress tolerance (Singla-Pareek et al. 2003). To chloroplast photosynthesis, however, methylgyoxal showed moderate toxicity, in comparison with 2-alkenals (Table 1). Probably the target sites of methylgyoxal in NaCl stress are different from chloroplast photosynthesis reactions.

Consequences of the toxicity of 2-alkenals

Current results with acrolein (Figs. 2–5) can be provided as models showing the consequences of enhanced 2-alkenal levels in leaves. The present data show that GSH is very likely more important than Asc in protecting chloroplast processes against acrolein. First, chloroplast GSH contents were rapidly decreased by acrolein (Fig. 4). Second, pre-addition of GSH but not Asc to chloroplast suspension decreased the sensitivity of photosynthesis to acrolein (Fig. 5). Although we cannot exclude that some of the protective effect of GSH may have been caused by a direct interaction with acrolein in the medium, the two observations together suggest that chloroplast GSH concentration is an important factor determining conjugation rate and protection of photosynthesis against this reactive aldehyde. This scavenging reaction through the formation of Michael adducts (conjugates), unlike the oxidation of GSH to GSSG, immediately leads to a decrease in the total glutathione pool (GSH + GSSG) (Fig. 4). Then GSH in chloroplasts should be supplemented by de novo synthesis within the plastid and by import from the cytosol. When the consumption exceeds the supplementation, the glutathione pool will be decreased. Two effects are expected. One is the inactivation of target enzymes by the enhanced 2-alkenals (Fig. 3; mechanism discussed below). The other is a loss of activity control through glutathionylation of aldolase, triose-phosphate isomerase, thioredoxin f, and GAPDH (Ito et al. 2003; Michelet et al. 2005; Zaffagnini et al. 2007).

The effects of acrolein on FBPase and PRK are interpreted in a common mechanism. These enzymes exhibited an acute inactivation phase below 0.5 mM acrolein and a relatively stable phase in its higher concentration (Fig. 3). The following mechanism explains such a biphasic inactivation: Acrolein reacts primarily with the redox-regulated thiols on the enzyme and converts it to an inactive form. The disulfide form of the enzyme is much less sensitive to acrolein (Fig. 6). When acrolein was added to chloroplasts, it should cause two effects on these thiol-regulated enzymes; the acceleration of the thiol oxidation to disulfide by consuming stromal GSH, and the modification of thiols. On assay of the resulting population of the enzyme, the disulfide form was reactivated by reduction, while the acrolein-modified forms did not restore the activity even by reduction. Higher sensitivity of PRK than that of FBPase (Fig. 3) can be explained by the difference of the ratio of these two forms, which could be ascribed to the difference of the midpoint potential of the thiols, as follows. The midpoint potential of the thiols are ca. –315 mV in PRK and –350 mV in FBPase (Hutchison et al. 2000). Therefore, when the stromal redox status becomes more oxidized due to the consumption of GSH, the former enzyme will stay in the reduced form longer than the latter. This would make PRK more susceptible to the Michael addition of acrolein.

Fig. 6
figure 6

Modification of a Cys residue(s) on a redox-regulated enzyme with acrolein

Full size image

Other tested enzymes did not show apparent biphasic inactivation although some of them also have redox-regulated thiols. The inactivation of these enzymes might be caused by the modification of not only the redox-regulated thiols but also other amino acids such as Lys and His, as observed for the inactivation of rabbit muscle GAPDH by HNE (Ishii et al. 2003). Thus, Calvin cycle enzymes are inactivated by acrolein in various modes, to different extents. We are investigating the inactivation mechanisms of several enzymes with various aldehydes. The above-mentioned mechanisms of acrolein’s action on photosynthesis can be basically applied to other 2-alkenals, although each aldehyde will show different strength of effect, as in Fig. 1, and there may be specific effects in certain combinations of aldehydes and targets.

In severe oxidative stress, not only the Calvin cycle but also mitochondrial respirations will be inactivated by 2-alkenals (Taylor et al. 2002; Winger et al. 2005). This will lead to the loss of energy-consumption capacity in the cell and exacerbate light-excess status and increase the production of reactive oxygen species. Scavenging of 2-alkenals would be thus critical to protect the target enzymes from inactivation at early stages of stress, and to prevent further development of oxidative injury, by preserving the electron sink capacity in leaf cells.

Physiological relevance

In what physiological situations the above-mentioned toxicity of 2-alkenals can be significant? As described in Table 1 legend, plant tissues can generate various 2-alkenals. Kohlmann et al. (1999) determined the HHE content in barley leaves as high as 17 nmol (g fresh weight)−1. This value corresponds to μM levels in cells when homogeneous distribution of the aldehyde in the tissue is assumed, and can be sub-mM level when its compartmentation in the cell is considered. Even higher aldehyde contents are expected under oxidative stress conditions. The HNE content in Phaseolus vulgaris leaves was increased 100-fold or more, during the oxidative stress induced by bacterial infection (Muckenschnabel et al. 2001). Similarly, in photooxidative status induced by various environmental stresses, 2-alkenal levels will be increased. Indeed, the emission of (E)-2-hexenal from leaves was increased by heat stress or strong light illumination to Phragmites australis plants, and by photoinhibition treatment of the A. thaliana NPQ1 mutant (Loreto et al. 2006). For evaluation of intracellular aldehyde levels we have recently developed an analysis method (Matsui et al. 2009) and obtained preliminary results that thylakoid membranes contained crotonaldehyde, (E)-2-pentenal and HHE, at sub-mM to mM on the thylakoid volume, in non-stressed preparations (S. Khorobrykh, J. Mano, Y. Iijima and D. Shibata (both Kazua DNA Institute, Kisaradzu, Japan), unpublished data), and that they were increased several fold by strong illumination of the leaves (J. Mano, S. Khorobrykh, K. Tokushige (Graduate School of Agriculture, Yamaguchi University), Y. Iijima and D. Shibata, unpublished data). Thus 2-alkenals are endogenously produced in chloroplasts and their levels can be increased to toxic levels by environmental stresses. In order to evaluate the in vivo toxicity of 2-alkenals, comprehensive analysis of the produced species and their levels will be necessary.