Carotenoid oxidation products as stress signals in plants


  • Michel Havaux

    Corresponding author
    1. Laboratoire d'Ecophysiologie Moléculaire des Plantes, CEA, DSV, IBEB, Saint-Paul-lez-Durance, France
    2. CNRS, UMR 7265 Biologie Végétale et Microbiologie Environnementales, Saint-Paul-lez-Durance, France
    3. Aix-Marseille Université, Saint-Paul-lez-Durance, France
    Search for more papers by this author


Carotenoids are known to play important roles in plants as antioxidants, accessory light-harvesting pigments, and attractants for pollinators and seed dispersers. A new function for carotenoids has recently emerged, which relates to the response of plants to environmental stresses. Reactive oxygen species, especially singlet oxygen, produced in the chloroplasts under stress conditions, can oxidize carotenoids leading to a variety of oxidized products, including aldehydes, ketones, endoperoxides and lactones. Some of those carotenoid derivatives, such as volatile β–cyclocitral, derived from the oxidation of β–carotene, are reactive electrophile species that are bioactive and can induce changes in gene expression leading to acclimation to stress conditions. This review summarizes the current knowledge on the non-enzymatic oxidation of carotenoids, the bioactivity of the resulting cleavage compounds and their functions as stress signals in plants.


Carotenoids are a group of isoprenoid compounds comprising a wide range of structures. They are constituted by eight isoprene units, with the vast majority of carotenoids deriving from the linear tetraterpene phytoene (Davies, 1976; Cunningham and Gantt, 1998; Hirschberg, 2001; DellaPenna and Pogson, 2006; Ruiz-Sola and Rodriguez-Concepcion, 2012). The introduction of double bonds in the latter 40–carbon basal structure by multiple desaturation reactions leads to the lycopene pigment, which confers the red color to ripe fruits of Solanum lycopersicum (tomato). A number of modifications of the linear backbone by cyclases, hydroxylases, ketolases and other enzymes give rise to the wide spectrum of carotenoid diversity. Over 700 chemical structures of carotenoid molecules have been determined from plants, algae, bacteria and invertebrates. Two main classes of carotenoids can be distinguished (Figure 1): the unoxygenated carotenoids, called carotenes (e.g. β–carotene and lycopene), and their oxygen-containing derivatives, called xanthophylls (e.g. lutein and zeaxanthin).

Figure 1.

Some of the compounds formed during enzymatic or non-enzymatic oxidative cleavage of the carotenoid β–carotene: blue arrows, oxidation of the double bonds in the carotenoid backbone; green arrows, secondary oxidation of β–ionone. Oxidation of zeaxanthin leads to similar compounds with an additional hydroxyl group. Red frame: structure of β–carotene (from the carotene group), zeaxanthin (from the xanthophyll group) and their precursor isoprene.

Carotenoids are synthesized by a large variety of phototrophic and non-phototrophic organisms, fulfilling a multitude of functions. The presence of between three and 13 conjugated double bonds in the carotenoid skeleton confers the pigment character to this family of molecules, allowing them to absorb light in a wavelength range of 450–570–nm, in the absorption gap of chlorophyll. Consequently, carotenoids can function as accessory pigments in photosynthesis, enhancing light harvesting in the blue–green spectral domain. In addition, photoprotective properties are brought about by the low-lying triplet state, allowing quenching of both triplet chlorophyll (3Chl*) and singlet oxygen (1O2) (Edge et al., 1997; Triantaphylidès and Havaux, 2009). The latter function is crucial because 1O2 is supposed to be the main reactive oxygen species (ROS) produced in photosynthetic organisms during strong illumination (González-Pérez et al., 2011), and it is instrumental in the development of photooxidative damage to plant leaves induced by (a)biotic stress conditions (Grun et al., 2007; Triantaphylidès et al., 2008; Vellosillo et al., 2010). Carotenoid-deficient photosynthetic organisms are highly photosensitive, suffering extensive photodamage (Aluru et al., 2009) and displaying high-frequency mutagenesis as a result of 1O2 overproduction in light (Ouchane et al., 1997). It has long been known that herbicides targeting phytoene desaturation induce plant bleaching (e.g. Sandmann et al., 1990; Jung et al., 2000). Additional protective functions of carotenoids include the xanthophyll cycle and the associated regulation of light harvesting by the chlorophyll antenna system of photosystem II (PSII; Horton and Ruban, 2005; Jahns and Holzwarth, 2012), scavenging of free radicals and protection against membrane lipid peroxidation (Lim et al., 1992; Johnson et al., 2007) and stabilization of membrane lipid bilayers (Havaux, 1998). Actually, the latter function is believed to be the initial role played by carotenoids in primitive organisms, in which they probably emerged as membrane reinforcers owing to their rigid conjugated double-bond backbone (Rohmer et al., 1979; Ourisson and Nakatani, 1994). Carotenoids also fulfill ecological roles in plants by attracting pollinators and seed dispersers to brightly colored flowers and fruits (Zhu et al., 2010; Khoo et al., 2011).

Animals cannot synthesize carotenoids, and must obtain them from their diet; however, striking exceptions to this paradigm have recently been discovered in three arthropod species (pea aphid, spider mite and gall midge; Moran and Jarvik, 2010; Altincicek et al., 2011; Cobbs et al., 2013). Analysis of the sequenced genome of these insects has led to the discovery of an endogenous carotenoid biosynthetic gene cluster, presumably introduced via horizontal gene transfer from a fungal donor (Moran and Jarvik, 2010). The molecular basis of the red–green polymorphism of pea aphids was identified to be an amino acid substitution within a carotenoid desaturase gene, leading to a loss of torulene formation, which is connected with red body color. Interestingly, the free electrons generated by photo-activated carotenoids in carotenoid-synthesizing insects appeared to be transferrable to the reducing power machinery, followed by mitochondrial ATP synthesis, possibly providing an archaic photosynthetic mechanism (Valmalette et al., 2012). The health benefits and canonical functions of carotenoids in animals and humans are well described (Krinsky and Johnson, 2005; Rao and Rao, 2007; Sommer and Vyas, 2012). These functions include their general role as antioxidants, the provitamin A property of β–carotene and related carotenoids to enable vitamin A-mediated color vision, roles in preventing macular degeneration of the eye and the regulatory functions of retinoids. Ingested carotenoids are also used as pigments that furnish many birds, fishes and invertebrates with their characteristic color (Brush, 1990). Interestingly, birds with carotenoid-based plumage coloration appeared to have a specific appetite for carotenoid-enriched food, possibly through the tasting or smelling of volatile degradation products of carotenoids (Senar et al., 2010).

New functions for carotenoids have recently emerged in plants following the discovery of new apocarotenoid phytohormones (strigolactones; Yoneyama et al., 2009; Xie et al., 2010; Cheng et al., 2013). The term apocarotenoid refers to any cleavage product of a parent 40–carbon carotenoid. Figure 1 shows all the possible cleavage products of β–carotene that can arise from oxidative cleavage of the double bonds in the polyene chain. All products shown in Figure 1 were found in β–carotene solutions oxidized by 1O2 (Ramel et al., 2012a,b). The initial products are aldehydes and ketones. β–Apo-8′-carotenal is the longer fragment that arises from cleavage of the 7′,8′ double bond of β–carotene, with the shorter fragment being β–cyclocitral. A similar reaction on the xanthophyll zeaxanthin yields 3–hydroxy-β-apo-8′-carotenal and 3–hydroxy-β-cyclocitral. Recent reviews have covered the enzymatic mechanisms of oxidative cleavage of carotenoids that lead to hormonal compounds (strigolactone, abscisic acid), and their regulatory functions in plant development and responses to the environment (Nambara and Marion-Poll, 2005; Yoneyama et al., 2009; Raghavendra et al., 2010; Xie et al., 2010; Cheng et al., 2013). The present review focuses on the direct oxidation of carotenoids by ROS in plants exposed to environmental stresses and the biological functions of the resulting carotenoid derivatives.

Specific and Unspecific Oxidation of Carotenoids in Plants

In vitro experiments have shown that carotenoid solutions can be oxidized by 1O2, producing a large variety of oxidized derivatives. As shown in Figure 1, each double bond in the β–carotene backbone can be oxidized by 1O2, primarily leading to various aldehydes and ketones (Stratton et al., 1993; Ramel et al., 2012a), but endoperoxides, epoxides and lactones can also be produced (Yamauchi et al., 1998; Fiedor et al., 2001, 2005; Bando et al., 2004). Although less studied than 1O2 oxidation of β–carotene, the 1O2-induced cleavage of xanthophylls can lead to the formation of similar types of oxidized compounds (Ramel et al., 2012a). Of particular interest are the carotenoid endoperoxides because they cannot be formed enzymatically or by free radicals, and they are specific to 1O2 attack on the carotenoid molecules. In fact, β–carotene-5,8-endoperoxide (Figure 2), resulting from the 1,4–cycloaddition of 1O2 with cis-diene (Gollnick and Kuhn, 1979), is considered to be the primary compound formed during photosensitized oxidation of β–carotene solutions by 1O2 (Montenegro et al., 2002). This 1O2-specific compound was detected in vivo in different light-exposed animal and human tissues, such as the skin and the eye (Bernstein et al., 2001; Bando et al., 2004; Bhosale and Bernstein, 2005), and also in plant leaves (Ramel et al., 2012a, 2013b). In particular, β–carotene-5,8-endoperoxide was measured in plant leaves in the absence of any light stress, suggesting chronic oxidation of β–carotene by 1O2. This compound was observed to rapidly accumulate in Arabidopsis leaves exposed to high light stress (Figure 2), in correlation with the levels of PSII photoinhibition and the losses of β–carotene (Ramel et al., 2012a). Moreover, the β–carotene endoperoxide content in Arabidopsis leaves was correlated with 1O2 levels measured by various techniques (Ramel et al., 2013b). In contrast, lutein/zeaxanthin endoperoxide levels were very low in plant leaves, and did not increase significantly during high light stress. This difference in the light response of β–carotene endoperoxide and xanthophyll endoperoxide cannot be attributed to a differential reactivity of β–carotene and xanthophylls with 1O2. Indeed, in vitro oxidation of β–carotene and of the xanthophylls lutein and zeaxanthin was associated with similar accumulation rates of the respective endoperoxides (Ramel et al., 2012a). In plants, β–carotene and xanthophylls are distributed differently in PSII: β–carotene is present exclusively in the reaction centers, whereas xanthophylls are located in the light-harvesting antenna complexes (Siefermann-Harms, 1985), suggesting different functions in photosynthesis. The selective accumulation of β–carotene endoperoxide leaves exposed to high light stress suggests that the PSII reaction center (containing β–carotene), rather than the PSII chlorophyll antennae (binding xanthophylls), is a major site of 1O2 accumulation in photosynthetic organisms. It is likely that this effect results from the close proximity of xanthophylls and chlorophylls in the PSII light-harvesting complexes, allowing the quenching of 3Chl* by chlorophyll–carotenoid triplet transfer (Mozzo et al., 2008). In contrast, the distance between the β–carotene molecules and the chlorophylls in the PSII reaction center is relatively large (Ferreira et al., 2004), making the direct quenching of 3Chl* unlikely. As a consequence, the probability of 1O2 production is enhanced in the PSII centers relative to the chlorophyll antennae, and therefore the primary function of β–carotene in PSII is believed to be limited to the quenching of 1O2 (Telfer, 2005). Consistently, 14C pulse-chase experiments showed a continuous flux of newly fixed carbon into β–carotene in leaves exposed to high light conditions, indicating a rapid turnover of this carotenoid (Beisel et al., 2010). In contrast, similar 14C incorporation was not found for xanthophylls, suggesting a slow turnover compared with β–carotene.

Figure 2.

Accumulation of the 1O2-specific β–carotene-5,8-endoperoxide compound in Arabidopsis leaves exposed to high light stress (Light) and decay in the dark (Dark). At time 0, plants were transferred from low light (250 μmol photons m−2 sec−1) to high light (1400 μmol m−2 sec−1). The grey zone, starting at 7 h, corresponds to the transfer of the plants from high light to darkness. Redrawn from Ramel et al. (2012a).

Various apocarotenoids, besides β–carotene endoperoxide, were found in different plant tissues. β–Apocarotenoids are present in the fruits of some Rubus species (raspberry; Beekwilder et al., 2008), Solanum lycopersicum (tomato) or Citrullus lanatus (watermelon; Lewinsohn et al., 2005), and, in melon, their concentration reaches approximately 1.5% of the level of β–carotene (Fleshman et al., 2011). β–Carotene oxidation products including β–carotene epoxides, diepoxides and apocarotenals are also present in processed fruits (juice and dried fruits; Rodriguez and Rodriguez-Amaya, 2007). 3–Hydroxy-β-ionone, a xanthophyll oxidation product, was detected in the hypocotyls of bean seedlings (Kato-Noguchi, 1994). β–Ionone was measured in Zea mays (sweetcorn), and its concentration was substantially increased in a carotenoid-enhanced variety (Gallon et al., 2013). Exposure to photooxidative stress conditions can also bring about the production of various volatile aldehydic or ketonic derivatives of β–carotene, such as β–ionone or β–cyclocitral (Figure 1), in photosynthetic microorganisms (Walsh et al., 1998) and in vascular plants (Ramel et al., 2012b). The accumulation of those volatile short-chain products derived from the oxidation of β–carotene was fast (i.e. took minutes) in high light-exposed Arabidopsis plants (Ramel et al., 2012b). There is no report of an equivalent accumulation of volatile oxidation products derived from xanthophylls under high light conditions.

Carotenoids can also be oxidized enzymatically by the carotenoid cleavage dioxygenases (CCDs). The latter enzymes cleave carotenoid molecules at particular positions, leading to the formation of specific apocarotenoids (Schwartz et al., 2004; Vogel et al., 2008; Floss and Walter, 2009). Apocarotenoids can also be formed enzymatically by co-oxidation initiated through lipoxygenase (Wu et al., 1999). Four true CCDs, named AtCCD1, AtCCD4, AtCCD7 and AtCCD8, have been identified in Arabidopsis. The AtCCD7 enzyme has been shown to cleave carotenoids at the 9,10 and/or 9′,10′ double bonds of the chromophore, producing β–ionone and β–apo-10′-carotenal from β–carotene. The latter apocarotenal can be subsequently converted into β–apo-13′-carotenol and a dialdehyde by AtCCD8. AtCCD1 has been shown to produce two β–ionone molecules and a C13 dialdehyde from β–carotene. CCD and CCD-related enzymes play an important role in the synthesis of plant hormones, such as abscisic acid and strigolactones, and they are involved in the synthesis of the aroma and flavor of a variety of flowers and foods (Auldridge et al., 2006). CCDs are ancestral enzymes that are also present in bacteria and animals. In animals, the CCD enzyme BCDO1 plays a role in retinoid synthesis, whereas BCDO2 is involved in carotenoid homeostasis in mitochondria, limiting mitochondrial dysfunction and oxidative damage (Amengual et al., 2011).

The regulation of CCD gene expression by abiotic or biotic stresses is not well documented. High light-induced upregulation of the NosCCD gene was shown in the cyanobacterium Nostoc (Scherzinger and Al-Babili, 2008), but similar effects were not reported in vascular plants. In Arabidopsis, none of the four AtCCD genes displayed an increase in expression during high-light stress (Ramel et al., 2013a). In addition, the levels of β–cyclocitral and β–ionone were not reduced in Arabidopsis mutants deficient in the AtCCD genes, compared with the wild type (Ramel et al., 2013a). Taken together, those results do not support the idea that CCD-catalyzed oxidation is a major component of the production of oxidized carotenoid metabolites during photooxidative stress (Ramel et al., 2013a). Accordingly, Simkin et al. (2003) evaluated the photooxidative degradation of carotenoids in pepper leaves after a dark-to-light transfer to around 1 mg per g of dry weight per day, and they estimated that the light-independent degradation or conversion of carotenoids, e.g. to abscisic acid, is a very minor process.

Carotenoid Oxidation Products are Bioactive

A large number of compounds derived from the oxidation of carotenoids by ROS, such as the volatile carotenoid cleavage products shown in Figure 1, are reactive electrophile species: they contain a carbonyl function adjacent to a double bond that is able to react with nucleophilic (electron donor) atoms (such as S and N) common to many biological molecules (Farmer and Davoine, 2007; Mueller and Berger, 2009; Farmer and Mueller, 2013). Such binding is often reversible (Davoine et al., 2006; Dueckershoff et al., 2008). In particular, reactive electrophile species (RES) have a high reactivity towards thiols and, consequently, can modify proteins in vivo. Thiol modification by electrophilic lipids can activate transcription factors, hence inducing gene responses (Levonen et al., 2004). Similarly, metabolites generated by the oxidation of lycopene are active in the regulation of gene expression, and an overlap was observed with the gene expression profile of retinoic acid (Gouranton et al., 2011; Reynaud et al., 2011). RES have also been shown to form adducts with nucleic acids (Blair, 2001). In addition, RES can conjugate with glutathione (Davoine et al., 2005), and some of the effects of RES may be related to the depletion of reductants. In animal cells, oxidized carotenoid derivatives have been reported to have a multitude of effects, including cytotoxicity (Lakshminarayana et al., 2010), induction of aptotosis (Janakiram et al., 2008; Kalariya et al., 2008; Liu et al., 2008), enzyme inactivation and inhibition of mitochondrial respiration (Siems et al., 2000, 2002; Hurst et al., 2005), induction of P450 cytochrome (Jeong et al., 1998), damage to DNA (Yeh and Wu, 2006; Kalariya et al., 2009) and retinoid signaling antagonism or promotion (Kuntz et al., 2006; Eroglu et al., 2012). In addition, carotenoid oxidation products, such as dihydroactinidiolide (Figure 1), are active components of pheromones in several insects (Rocca et al., 1983), and act as cat attractants in some plant species (Zhao et al., 2006).

Photosynthesis can be inhibited by high concentrations of carotenoid RES in cyanobacteria (Shao et al., 2011). Very recently, it was shown that the cyanobacterium Synechocystis contains an aldehyde dehydrogenase that can mediate the conversion of apocarotenals into their corresponding acids (Trautmann et al., 2013). Moreover, the gene encoding this enzyme is inducible by high light levels and cold, indicating a role in stress responses. This recently discovered aldehyde dehydrogenase is probably part of a detoxification mechanism scavenging toxic apocarotenals; however, β–cyclocitral can also have beneficial effects in those organisms, such as repelling grazers (Jüttner et al., 2010). β–Ionone and related compounds were found to have antimicrobial and antifungal properties (Salt et al., 1986; Utama et al., 2002), although most studies used high, non-physiological concentrations. The production of the carotenoid metabolite 3–hydroxy-β-ionone in some mosses inhibits the growth of neighboring organisms, thus allowing pure colonies (Kato-Noguchi and Seki, 2010). Retinal, the visual pigment derived from β–carotene (Figure 1), can be found in plants (Lorenzi et al., 1994), and retinal-containing rhodopsins play a role in the phototactism of the green alga Chlamydomonas (Heintzen, 2012). β–Ionone and β–cyclocitral were also reported to act as feeding deterrents of various herbivorous insects (Wang et al., 1999; Gruber et al., 2009; Nyalala et al., 2013). Transgenic Arabidopsis plants overexpressing AtCCD1 exhibited enhanced β–ionone emission compared with the wild type (Wei et al., 2011). Interestingly, the transgenics showed 50% less leaf area damage by flea beetles, confirming the beneficial effect of high β–ionone concentrations against herbivorous insects.

It is known that products derived from the enzymatic oxidation of carotenoids can have important signaling functions in plants. On the one hand, the hormone abscisic acid is derived from the enzymatic oxidation of the xanthophyll neoxanthin (Nambara and Marion-Poll, 2005). This hormone is involved in plant responses to environmental stress and pathogens, and it also plays a role in seed germination, early embryon development and stomatal regulation. On the other hand, the carotenoid-derived strigolactones are terpenoid lactones constituting a class of hormones that regulate shoot branching (Gomez-Roldan et al., 2008; Umehara et al., 2008). Apocarotenoids are also believed to provide signaling compounds for the regulation of root colonization by arbuscular mycorrhizal fungi (Strack and Fester, 2006).

ROS Oxidation of Carotenoids and Stress Signaling in Plants

As RES, carotenoid derivatives, such as β–cyclocitral or β–ionone, are potential signal molecules the concentration of which increases in plants exposed to photoxidative stress through ROS oxidation of β–carotene (Ramel et al., 2012b). The function of RES in cellular signaling is known mainly from the study of fatty acid-derived RES, such as malondialdehyde, hexenal, oxophytodienoic acid (OPDA) or cyclopentenone phytoprostanes. Exogenously applied oxylipin RES have been shown to affect the transcription of genes involved in cell survival and stress responses (Alméras et al., 2003; Weber et al., 2004; Loeffler et al., 2005; Mueller et al., 2008). These effects are accompanied by changes in protein abundances (Dueckershoff et al., 2008). Concerning the effects of phytoprostane A1 and OPDA, 17% of the induced genes were related to detoxification processes (Mueller et al., 2008). The dominant gene families encoded glutathione-S-transferases, UDP-glucosyl transferases, cytochrome P450 and transporters. The genes downregulated by lipid RES affected cell walls, cell division and auxin signaling.

In a recent study by Ramel et al. (2012b), the treatment of Arabidopsis plants with low concentrations of volatile β–cyclocitral, leading to internal concentrations of this compound in the range measured in leaves of high light-treated plants, was demonstrated to modify the expression of a large set of genes (Ramel et al., 2012b). A large fraction of the genes induced by β–cyclocitral was related to cellular defense against stress and to metabolism, whereas the repressed genes were linked mainly to development and biogenesis. Interestingly, more than 80% of the genes affected by β–cyclocitral were identified as 1O2-responsive genes, and a rather large subset of the induced genes were even classified as specific to 1O2, leading to the conclusion that β–cyclocitral is an intermediate in the signaling of this ROS in Arabidopsis. Moreover, there was a certain specificity in these effects, as the related molecule β–ionone was not able to induce or repress the expression of 1O2-responsive genes. A difference between the two carotenoid RES is the fact that the β–carbon in the α,β–unsaturated carbonyl group is methylated in β–cyclocitral, suggesting a weaker electrophilicity for this metabolite. In addition, only a partial overlap was observed between the transcription changes induced by β–cyclocitral and the gene expression profile induced by lipidic RES. This suggests that the effects of β–cyclocitral on gene expression are more specific than a general response to RES. An important observation was the fact that β–cyclocitral-induced changes in gene expression were associated with an enhancement of the photoresistance of the treated plants. However, in a screening of carotenoid oxidation products we recently observed that other β–carotene derivatives are able to induce changes in the expression of 1O2-responsive genes (L. Shumbe and M. Havaux, unpublished results), suggesting that β–cyclocitral is probably not the unique carotenoid-derived messenger involved in the 1O2 signaling pathway. It is thus possible that β–cyclocitral is part of a larger group of RES that collectively stimulate 1O2-specific responses and activate acclimation to photooxidative stress. Taken together, those results lead to the conclusion that molecules generated by 1O2 oxidation of a carotenoid act as chloroplastic messengers, pointing to an important signaling function for carotenoids in stressed plants (Ramel et al., 2013a). β–Cyclocitral is a volatile, lipid-soluble molecule that can cross lipid membrane bilayers, and is therefore a potential candidate for the transfer of information out of the chloroplast. In addition, it is known that RES adducts to glutathione can be actively transported across membranes (Klein et al., 2006). In this context, it is important to note that glutathione-S-transferase genes, which catalyze the conjugation of RES with glutathione, were strongly induced by 1O2 in algae (Ledford et al., 2007), and also by β–cyclocitral in higher plants (Ramel et al., 2012b).

Isoprene, the fundamental building block of carotenoids, is volatile and can be emitted into the atmosphere by some plant species (Sharkey et al., 2008). This phenomenon has been related to thermotolerance and resistance to oxidative stress. The isoprene molecule contains two double bonds (Figure 1), making it sensitive to oxidation. It is known that, once emitted, isoprene is oxidized in the atmosphere to RES, such as methyl vinyl ketone and methacrolein (Atkinson and Arey, 2003). Interestingly, these oxidation products can also be formed within plants (Jardine et al., 2012), and abiotic stresses were reported to promote the emission of a number of isoprene oxidation products by isoprene-synthesizing plants (Jardine et al., 2013). Consequently, isoprene oxidation could represent an antioxidant mechanism through depletion of the oxidant pool and through RES-mediated signaling processes.

Gene expression reprogramming by carotenoid RES was observed to be independent of the EXECUTER 1 protein, and to lead to an increase in the tolerance towards photooxidative stress (Ramel et al., 2012b). The EXECUTER proteins are involved in the response of the flu Arabidopsis mutant to 1O2 (Wagner et al., 2004). This mutant accumulates a chlorophyll precursor, protochlorophyllide, in the dark that acts as a photosensitizer in the light, causing massive 1O2 production. Under those conditions, the 1O2-induced changes in gene expression were shown to induce a cell death program that was dependent on the EXECUTER 1 protein. Possibly, this difference between the β–cyclocitral effect and the EXECUTER 1-dependent effect relies on the intensity of the gene responses, which was much lower with volatile β–cyclocitral (Ramel et al., 2012b) relative to the gene responses reported in the flu mutant (Wagner et al., 2004). In fact, different 1O2 levels should be distinguished when considering the effects on gene expression: programmed cell death could be a response to high 1O2 concentrations, such as the massive 1O2 production occurring from a chlorophyll precursor in flu mutant leaves after a transition from darkness to light, whereas low 1O2 levels and modest changes in gene expression, such as those induced by treatments with volatile β–cyclocitral, triggers acclimation to 1O2 stress. Changes in gene expression triggered by 1O2 were also found to occur in a xanthophyll mutant of Arabidopsis without causing cell death (Alboresi et al., 2011). An Arabidopsis mutant deficient in two plastoglobule-localized kinases was recently observed to display a rapid chlorosis upon high light stress (Lundquist et al., 2013). This senescence-like phenotype was correlated with increased levels of β–cyclocitral, but was independent of the EXECUTER pathway. Kim and Apel (2013) suggested that the responses to 1O2 may be complex, particularly in the wild-type background, because direct toxicity of this ROS can superimpose on the 1O2 signaling pathway, hence masking the EXECUTER-dependent responses.

Acclimation to 1O2 stress induced by low 1O2 concentrations in the medium was previously reported in the green microalga Chlamydomonas reinhardtii (Ledford et al., 2007). This phenomenon involves the activation of an RES-induced defense response (Fischer et al., 2012). An electrophile response element was identified in the promoter region of many genes induced during the 1O2 acclimation process in this microalgal species. Furthermore, the lipidic RES compound 1–chloro-2,4-dinitrobenzene can substitute for 1O2 produced by a photosensitizing dye in enhancing the tolerance of Chlamydomonas cells to 1O2. A different mechanism appears to prevail in vascular plants: the promoters of many genes induced by oxylipins were found to contain TGA motifs, which constitute putative binding sites for basic leucine zipper transcription factors of the TGA family (Mueller et al., 2008; Stotz et al., 2013). Analyses of Arabidopsis mutants affected in the expression of TGA transcription factors confirmed that indeed those transcription factors are involved in mediating gene regulation by electrophilic oxylipins.

Acclimation to 1O2 was recently demonstrated to occur in higher plants (Ramel et al., 2013b). The tolerance to 1O2 toxicity and photooxidative stress increased in a 1O2-overproducing Arabidopsis mutant (ch1) after exposure to moderately elevated light intensity that promoted the controlled accumulation of 1O2 in the leaves (Ramel et al., 2013b). This phenomenon was accompanied by an accumulation of β–cyclocitral and also by changes in the expression of a number of genes, but was independent of EXECUTER 1. In particular, acclimation to 1O2 was accompanied by a downregulation of the jasmonate biosynthesis pathway, with most of the genes involved in this pathway being strongly repressed. In contrast, in the non-acclimated state, high light levels induced jasmonate synthesis and cell death in the 1O2-producing ch1 and flu mutants (Przybyla et al., 2008; Ramel et al., 2013b). Those results indicate a potential interaction between carotenoid RES responses and oxylipin signaling, and suggest that jasmonate could function as a decision maker between cell death and acclimation in the response of a plant to 1O2 stress (Ramel et al., 2013c). In this context, it is important to remember that oxylipin hormones are components of the defense responses of plants to pathogens and herbivores (Thaler et al., 2004; Howe and Jander, 2008; Bari and Jones, 2009). As recently pointed out by Demmig-Adams et al. (2013), a decrease in jasmonate concentration associated with an increased tolerance to abiotic stresses may therefore have negative consequences for plant tolerance towards biotic stresses. This possible trade-off between abiotic and biotic plant defenses is an important issue that deserves to be addressed in the future.

The finding that plastid carotenoid metabolites can affect the expression of nuclear genes points at an additional function for carotenoids, besides their well-known roles in light harvesting and tolerance to oxidative stress (Ramel et al., 2013a). Because β–carotene is localized very close to the primary site of 1O2 production in chloroplasts (i.e. the PSII reaction center), β–carotene oxidation can be considered as an early event during photostress, and therefore β–carotene oxidation metabolites may constitute primary sensors of light stress in plants. The identification of carotenoid oxidation products as components of the photooxidative stress signaling pathway opens exciting avenues for the future (Figure 3). In particular, the elucidation of the primary targets and the exact mode of action of carotenoid RES in the acclimatory response to 1O2 are major challenges for future work. Another important future issue concerns the mechanisms involved in the metabolization of carotenoid oxidation products. β–Carotene endoperoxide has been shown to have a rather short lifetime in vivo, declining in Arabidopsis leaves with a half-life of around 6 h (Figure 2). Whether the turnover of this metabolite and other oxidized carotenoid derivatives in plants is spontaneous or catalyzed by enzymes remains to be determined. The recent discovery of a cyanobacterial aldehyde dehydrogenase that is able to act on apocarotenals (Trautmann et al., 2013) indicates the existence of specialized enzymes for the metabolization of carotenoid RES in photosynthetic organisms.

Figure 3.

Oxidation of β–carotene by 1O2 in the photosysten II (PSII) reaction centers produces oxidized derivatives, such as β–cyclocitral or β–ionone, which can have various biological effects, including toxicity, protection against herbivores, and stress signalling, leading to 1O2 tolerance. TF, transcription factor; in red, pathway specific to β–cyclocitral.