Review TRENDS in Plant Science Vol.11 No.10

Autoluminescence imaging: a non-invasive tool for mapping

Michel Havaux, Christian Triantaphylide` s and Bernard Genty

CEA/Cadarache, DSV, DEVM, Laboratoire d’Ecophysiologie Mole´ culaire des Plantes, UMR 6191 CEA-CNRS-Aix Marseille University, F-13108 Saint-Paul-lez-Durance, France

Living organisms spontaneously emit faint light, and includes carotenoids or terpenoids, which can produce this autoluminescence is stimulated in response to many endoperoxides upon oxidation [8,9]. Singlet oxygen is pro- stresses. This phenomenon is attributable to the duced in reactions between two lipid peroxy radicals, and endogenous production of excited states during the light emission peaks in the red (at 634 and 703 nm; oxidative reactions, particularly during peroxidation of Figure 1). By contrast, the bimolecular reaction of alkoxy lipids, which generates light-emitting molecules such as radicals and the cleavage of dioxetanes lead to the forma- triplet carbonyls and singlet oxygen. Using highly sen- tion of excited triplet carbonyls, which in turn decay and sitive cameras, it is now possible to remotely image yield light in the range 350 to 480 nm (blue). Consistently, spontaneous luminescence with a good spatial spectral analyses of the luminescence emission of oxidized resolution, providing a new non-invasive tool for lipids in vitro showed the expected peaks in the blue and mapping oxidative stress and lipid peroxidation in red-near infrared domains [10]. However, in plants, triplet plants. carbonyls can also transfer excitation energy to chlorophyll molecules [11] whose deactivation is associated with The types of luminescence of living organisms emission at around 730 nm (far-red light) [12,13]. Plants, animals and bacteria glow continuously. This spon- Formation of activated carbonyls from lipid hydroper- taneous photon emission (SPE) occurs in the oxide decomposition is stimulated by heat [14], and this and visible range of wavelengths with a low intensity of phenomenon is the basis of the luminescence burst À19 À16 À2 10 to 10 Wcm [1]. This is about three orders of observed at 135 8C in plant leaves (Figure 2a) [3,4,15]. magnitude less than the luminescence arising from enzy- The amplitude of the luminescence measured at this tem- matic reactions in fireflies, jellyfishes and some fishes in a perature is two orders of magnitude higher than the room- transitory phenomenon commonly called temperature autoluminescence amplitude. Numerous [2]. It is different from the luminescence of chlorophyll, thermoluminescence studies have validated the use of emitted by photosynthetic organisms following exposure to the luminescence measured at 135 8C as an index of lipid light, which results from recombination of positively peroxidation by correlating the signal amplitude with the charged carriers at the donor side of photosystem II and extent of lipid peroxidation, which can be measured using a electrons on the acceptor side [3,4]. The chlorophyll signal variety of biochemical methods [15–18]. The luminescence decays rapidly in the dark within seconds to minutes. Here, at 135 8C has been used in plants to quantify lipid we provide background information about SPE, with damage after oxidative stress [18–20] or to characterize emphasis on the origin of the signal, and discuss in more the stress resistance of mutants [16,17,21]. Importantly, detail imaging of SPE in plants in comparison with other when the luminescence signal emitted by Arabidopsis available techniques. leaves at 25 8C was plotted against the luminescence emission at 135 8C, a linear regression was obtained Spontaneous photon emission, a signature of biological (Figure 2b), confirming that both signals reflect oxidative oxidation status stress damage and supporting the hypothesis that they SPE, often called biophoton emission, is attributed to the could have a common origin in plants [15]. endogenous production of excited states during oxidative Light produced during lipid peroxidation therefore con- metabolic reactions [5,6] and it is therefore closely related stitutes an internal probe of oxidative stress that has many to the oxidation status of the organism. The major source of potential applications. Lipid peroxidation and oxidative SPE is the slow spontaneous decomposition of lipid per- stress have been measured by room-temperature SPE in oxides and endoperoxides, such as dioxetanes, which gen- a variety of samples, from whole organisms to isolated erates the light-emitting byproducts singlet oxygen and membranes or lipoproteins, and even in human breath excited carbonyl groups [7]. The term ‘lipids’ is used here in (reviewed in Refs [1,15]). In many studies, the intensity a broad sense and is not restricted to fatty acids but also of SPE correlated with other indexes of lipid peroxidative

Corresponding author: Havaux, M. ([email protected]). degradation, such as the level of conjugated dienes, the Available online 7 September 2006. malondialdehyde content or the concentration of www.sciencedirect.com 1360-1385/$ – see front matter ß 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2006.08.001 Review TRENDS in Plant Science Vol.11 No.10 481

1 Figure 1. associated with lipid peroxidation. The emission of light is related to the production of excited carbonyl compounds and singlet oxygen ( O2) that occurs during lipid peroxide degradation. Self-reaction of lipid peroxy radicals (the Russel reaction) gives a tetroxide intermediate, which decomposes either into a triplet state carbonyl compound (reaction 1a; the triplet state is indicated by a star) or into singlet oxygen (reaction 1b). Self-reaction of alkoxy radicals produces excited carbonyl compounds in a redox reaction (reaction 2). The singlet oxygen that is produced during lipid peroxidation (such as reaction 1b) reacts with unsaturated compounds, leading to the production of dioxetanes, which yield triplet state carbonyl compounds upon decomposition (reaction 3). Upon relaxation, the excited carbonyl compounds emit light in the blue region (380–460 nm; reaction 4a) whereas bimolecular emission of singlet oxygen occurs in the red region (634 and 703 nm; reaction 4b). Abbreviation: hy, energy of a single photon. Adapted from Refs [6,7]. exogenously added lipid hydroperoxides [22–24]. This Imaging of spontaneous photon emission non-invasive technique is useful for real-time monitoring A currently emerging and promising application of SPE is of rapid perturbations of the cellular oxidation state and is the imaging of oxidative stress in intact tissues. Early performed without an exogenous probe. attempts at autoluminescence imaging in the 1990s used

Figure 2. Luminescence of Arabidopsis leaves and its variation with temperature. (a) High-temperature luminescence emission from Arabidopsis leaves exposed to photo-oxidative stress. The curves are of wild type (Columbia; 1) and a vte1 npq1 double mutant deficient in the carotenoid zeaxanthin and in vitamin E (2) [17].Plantswere exposed for 3 days to high light stress (1200 mmol mÀ2 sÀ1, 8 h photoperiod) at low temperature (8 8C). The luminescence band peaking at around 135 8Cisattributedtolight- emitting molecules produced during lipid peroxidation [15]. (b) Correlation between the luminescence emission at 25–30 8C (commonly used for imaging) and the amplitude of the luminescence band at 135 8C. The results shown were obtained in a series of experiments on a range of Arabidopsis mutants exposed to various stress conditions. www.sciencedirect.com 482 Review TRENDS in Plant Science Vol.11 No.10

Figure 3. Imaging of the spontaneous luminescence emission from Arabidopsis plants at room temperature (1, double mutant vtc2 npq1 deficient in both ascorbate and zeaxanthin; 2, single mutant vtc2 deficient in ascorbate; 3, wild type) pre-exposed for 4 days to high light stress (4 8C, 1600 mmol mÀ2 sÀ1, 8 h/16 h day/night regime). These mutants are differently sensitive to high light stress [21]. (a) Luminescence, as imaged with a liquid nitrogen-cooled CCD camera (VersArray LN/CCD 1340–1300B, Roper Scientific (as in Ref. [32]), mounted with a back-illuminated CCD from E2V CCD36–40). The emitting sample was imaged on the sensor by a 50-mm focal distance lens with an f-number of 1.2 (F mount Nikkor 50-mm, f:1.2, Nikon) for maximizing light collection. Integration time was 60 min using the full resolution of the CCD (1300 Â 1340 pixels). Plants were adapted to darkness for 2 h before measurements. Randomly distributed white spots are the results of stray radiation in the range of gamma and cosmic rays. The color scale indicates signal intensity from 0 (blue) to 250 (white). (b) Photograph of the corresponding plants. Asterisks indicate the type of leaves used for the malondialdehyde (MDA) analysis in (c) (i.e. larger, mature leaves). (c) MDA level in leaves. MDA levels in mature leaves of the vtc2 npq1 double mutant (1), vtc2 mutant (2) and wild type (3) plants that had been exposed to high light stress for 4 days were determined by HPLC as described in Ref. [17]. Data are mean values of a minimum of three experiments + standard deviation. single photon counting imaging techniques that If leaf autoluminescence does reflect lipid oxidative demonstrated the feasibility of this approach. However, damage, mutants known to be sensitive to oxidative stress most of the images obtained with these systems had low should emit more than do wild types. This is spatial resolution. Nevertheless, SPE from the brain of a shown in Figure 3, in which SPE from whole Arabidopsis rat was visualized and the imaged signal correlated with plants was imaged using a liquid nitrogen-cooled CCD cerebral energy metabolism and oxidative stress [25]. SPE camera similar to that described in Ref. [32]. The plants in plant tissues infected with a pathogen, injured mechani- were pre-adapted to darkness for 2 h to allow photosynth- cally or exposed to drought stress were also visualized with esis-related chlorophyll luminescence to decay [13]. Two this technique [26–28]. These pioneering studies generated photosensitive mutants, the vtc2 single mutant deficient in interest in SPE and prompted researchers to use alter- ascorbate and the vtc2 npq1 double mutant deficient in native technologies to improve the imaging of this signal. both ascorbate and zeaxanthin [21], were compared with Recent studies on plants, animals and humans have the photoresistant wild type after high light stress. Both shown that autoluminescence imaging can be improved mutants showed a much higher stimulation of SPE after using highly sensitive charge-coupled device (CCD) cam- photostress than did the wild type (Figure 3a), and SPE of eras. Cooling of the sensor, using peltier devices (thermo- the mutants was proportional to their relative sensitivity electric modules) or liquid nitrogen, is required for to oxidative stress [21]. Mature Arabidopsis leaves are well reducing thermal noise and enabling measurement of faint known to be less tolerant to photo-oxidative stress than light by signal integration. Using this approach, SPE from young leaves at the center of the rosette, and their SPE mouse tumors and from the human body were recently reflected this difference. Furthermore, the SPE intensity of imaged [29,30]. In plant science, the first application of mature leaves was related to the concentration of malon- CCD to SPE imaging was made using germinating seed- dialdehyde, a biochemical marker of lipid peroxidation lings [31]. High-resolution images of SPE from detached (Figure 3c). Another Arabidopsis double mutant, vte1 leaves were also reported [32]. So far, the best images of npq1, deficient in two major of the plastid plant autoluminescence allowing quantitative mapping membranes (vitamin E and zeaxanthin), is highly sensitive have been obtained by Michel Flor-Henry and co-workers to high light, with leaves suffering extensive lipid perox- [13], who used a cooled CCD detector coupled to optical idation, which was measured by various methods [17]. The fibers on which detached leaves were placed. This config- sensitivity of the vte1 npq1 double mutant to lipid perox- uration enables maximal capture of light and high sensi- idative damage can be readily monitored by autolumines- tivity, but it is applicable to flat samples only (e.g. detached cence imaging (Figure 4a): stimulation of SPE occurs much leaves). Wound-induced luminescence of leaves was more rapidly and more intensively in mutant leaves com- imaged with a good spatial resolution (222 000 pixels pared with wild-type leaves. Importantly, stimulated SPE per image). Although the origin of the signal was not from vte1 npq1 double mutant leaves was detected after 6 h investigated in detail in this study, lipid peroxidation in high light, before symptoms of oxidative damage were products are probable candidates [7,33]. Because wound- visible on the leaves (>26 h). ing-induced photon emission is quenched by sodium azide A field that has much to gain from this new imaging and amplified by deuterium oxide [34], singlet oxygen is method is phytopathology. For example, infiltration of a likely to be involved in this light emission. tobacco leaf with the elicitor cryptogein in the dark induced www.sciencedirect.com Review TRENDS in Plant Science Vol.11 No.10 483

Figure 4. Measurement of the effects of photooxidative stress and the hypersensitive defense response on SPE. (a) Autoluminescence imaging of Arabidopsis plants [wild type (WT) and vte1 npq1 double mutant deficient in vitamin E and zeaxanthin] pre-exposed to high light stress at low temperature (4 8C, 1600 mmol mÀ2 sÀ1, on a regime of 14 h day and 10 h night) for 0 to 54 h. The vte1 npq1 double mutant is highly sensitive to high light stress [17], and symptoms of oxidative stress (leaf bleaching) were observed visually after 26 h in the mutant and after 30 h in the wild type. Plants were adapted to darkness for 2 h before measurements. (b) Autoluminescence imaging of a tobacco leaf infiltrated with the elicitor cryptogein (2 mg) for 0–25 h in the dark. The cryptogein treatment was performed as described in Ref. [35]. Integration time was 20 min and on-CCD binning of 2 Â 2 pixels was used to increase detection sensitivity (the resulting resolution is 650 Â 670 pixels). The color scale indicates signal intensity from 0 (blue) to 250 (white). hypersensitive programmed cell death [35] and an changes in living cells [40–42]. Redox sensing is enhancement of SPE [19] that can be quantified by CCD achieved by introducing cysteine residues close to the imaging, as shown in Figure 4b. Increased SPE was chromophore –their oxidation causes changes in the GFP detected 18 h after infiltration, matching the time course fluorescence characteristics. Although application of this of the massive enzymatic lipid hydroperoxide accumula- technique to imaging in plants is not yet fully documented tion [35]. Recently, Mark Bennett and co-workers [36] [43], redox-sensitive GFPs are promising tools for visualiz- showed bursts of light over the Arabidopsis leaf surface ing the oxidation of specific proteins or organelles. However, induced during the hypersensitive response to a pathogen. this approach is technically more complex than SPE It seems that this photon emission is different to the SPE imaging because it requires genetic transformation of the associated with lipid peroxidation because it was transi- biological material. ent, beginning around 2 h after inoculation and lasted only 2–3 h. This transient light signal, which presumably Concluding remarks reflects some signaling phenomena related to nitric oxide Autoluminescence imaging is likely to become an in the plant defense response [36], also has great potential important tool in stress physiology to measure the for applications in phytopathological studies. oxidation status of plants non-invasively. Oxidative stress Because SPE imaging can be performed on intact imaged by SPE is useful not only for quantification pur- plants and does not require an exogenously added probe poses, but also for the early detection of oxidative stress or sample preparation (except adaptation to darkness), it is before damage becomes visible. SPE can be measured at expected to provide realistic mapping of lipid peroxidative different integration levels, from small populations of damage and oxidative stress in plants. This new approach organisms to cells. Because of its high sensitivity and easy is complementary to more direct measures of reactive use, SPE imaging should enable the rapid screening of oxygen species such as fluorescence imaging using fluor- mutants, including algae, cyanobacteria and vascular escent probes sensitive to singlet oxygen and/or reduced plants, with variable tolerance to oxidative stress. Even forms of oxygen [16,37–39]. Fluorescence imaging is useful though the sensitive equipment required for acquiring SPE because it provides information on specific reactive images is still relatively expensive, we predict that this oxygen species and their sites of production in tissues. new imaging approach will successfully complement the However, fluorescence imaging of reactive oxygen assortment of functional imaging tools available to plant species requires pre-infiltration of plant tissues with an biologists [44]. exogenous probe and it is therefore usually restricted to Acknowledgements cells or organs (e.g. detached leaves). A new approach Financial support from CEA (program ‘Toxicologie Nucle´aire Environne- based on redox-sensitive green fluorescent proteins (GFPs) mentale’) and EU project IHP-RTN-00–2001 ‘STRESSIMAGING’ is has recently emerged to monitor non-invasively redox acknowledged. We thank Ilja Reiter for helpful discussions. www.sciencedirect.com 484 Review TRENDS in Plant Science Vol.11 No.10

References 23 Doi, H. et al. (2002) Chemiluminescence associated with the oxidative 1 Devaraj, B. et al. (1997) Biophotons: ultraweak light emission from metabolism of salicylic acid in rat liver microsomes. Chem. Biol. living systems. Curr. Opin. Solid State Mater. Sci. 2, 188–193 Interact. 140, 109–119 2 Wilson, T. and Hastings, J.W. (1998) Bioluminescence. Annu. Rev. Cell 24 Guajardo, M.H. et al. (2002) Retinal fatty acid binding proteins reduce Dev. Biol. 14, 197–230 lipid peroxidation stimulated by long-chain fatty acid hydroperoxides 3 Ducruet, J-M. (2003) Chlorophyll thermoluminescence of leaf discs: on rod outer segments. Biochim. Biophys. Acta 1581, 65–74 simple instruments and progress in signal interpretation open the way 25 Kobayashi, M. et al. (1999) In vivo imaging of spontaneous ultraweak to new ecophysiological indicators. J. Exp. Bot. 54, 2419–2430 photon emission from a rat’s brain correlated with cerebral energy 4 Gilbert, M. et al. (2004) A new type of thermoluminometer: a highly metabolism and oxidative stress. Neurosci. Res. 34, 103–113 sensitive tool in applied photosynthesis research and plant stress 26 Suzuki, S. et al. (1991) Two-dimensional imaging and counting of physiology. J. Plant Physiol. 161, 641–651 ultraweak emission patterns from injured plant seedlings. 5 Abeles, F.B. (1986) Plant chemiluminescence. Annu. Rev. Plant J. Photochem. Photobiol. B. Biol 9, 211–217 Physiol. 37, 49–72 27 Makino, T. et al. (1996) Ultraweak luminescence generated by sweet 6 Sies, H. (1987) Intact organ spectrophotometry and single-photon potato and Fusarium oxysporum interactions associated with a defense counting. Arch. Toxicol. 60, 138–143 response. Photochem. Photobiol. 64, 953–956 7 Halliwell, B. and Gutteridge, J.M.C. (1989) Free Radicals in Biology 28 Ohya, T. et al. (2002) Biophoton emission due to drought injury in red and Medicine. Clarendon Press beans: possibility of early detection of drought injury. Jpn. J. Appl. 8 Stratton, S.P. et al. (1993) Isolation and identification of singlet oxygen Phys. 41, 4766–4771 oxidation products of beta-carotene. Chem. Res. Toxicol. 6, 542–547 29 Takeda, M. et al. (2004) Biophoton detection as a novel technique for 9 Baker, D.L. et al. (1999) Reactions of beta-carotene with cigarette cancer imaging. Cancer Sci. 95, 656–661 smoke oxidants. Identification of carotenoid oxidation products and 30 Van Wijk, R. et al. (2006) Anatomic characterization of human ultra- evaluation of the prooxidant/ effect. Chem. Res. Toxicol. 12, weak photon emission with a moveable photomultiplier and CCD 535–543 imaging. J. Photochem. Photobiol. B. Biol 83, 69–76 10 Miyazawa, T. et al. (1982) Spectroscopic analysis of the weak light 31 Kobayashi, M. et al. (1997) Two-dimensional imaging of ultraweak generated in autoxidation of linseed oil. Agric. Biol. Chem. 46, 1671– photon emission from germinating soybean seedlings with a highly 1672 sensitive CCD camera. Photochem. Photobiol. 65, 535–537 11 Bohne, C. et al. (1986) Chlorophyll: an efficient detector of 32 Creath, K. and Schwartz, G.E. (2004) Biophoton images of plants: electronically excited species in biochemical systems. Anal. Biochem. revealing the light within. J. Altern. Complement. Med. 10, 23–26 155, 1–9 33 Reverberi, M. et al. (2005) Relationship among lipoperoxides, 12 Hideg, E. and Vass, I. (1993) The 75 8C thermoluminescence band of jasmonates and indole-3-acetic acid formation in potato tubers after green tissues: chemiluminescence from membrane-chlorphyll wounding. Free Radic. Res. 39, 637–647 interaction. Photochem. Photobiol. 58, 280–283 34 Chen, W.L. et al. (2003) Imaging of ultraweak bio-chemiluminescence 13 Flor-Henry, M. et al. (2004) Use of a highly sensitive two-dimensional and singlet oxygen generation in germinating soybean in response to luminescence imaging system to monitor endogenous bioluminescence wounding. Luminescence 18, 37–41 in plant leaves. BMC Plant Biol. 4, 19 35 Ruste´rucci, C. et al. (1999) Involvement of -dependent 14 Vavilin, D.V. and Ducruet, J.M. (1998) The origin of 115–130 8C production of fatty acid hydroperoxides in the development of thermoluminescence bands in chlorophyll-containing material. the hypersensitive cell death induced by cryptogein on tobacco Photochem. Photobiol. 68, 191–198 leaves. J. Biol. Chem. 274, 36446–36455 15 Havaux, M. (2003) Spontaneous and thermoinduced photon emission: 36 Bennett, M. et al. (2005) Biophoton imaging: a nondestructive method new methods to detect and quantify oxidative stress in plants. Trends for assaying R gene responses. Mol. Plant Microbe Interact. 18, 95–102 Plant Sci. 8, 409–413 37 Fryer, M.J. et al. (2002) Imaging of photo-oxidative stress responses in 16 Baroli, I. et al. (2004) Photo-oxidative stress in a xanthophyll-deficient leaves. J. Exp. Bot. 53, 1249–1254 mutant of Chlamydomonas. J. Biol. Chem. 279, 6337–6344 38 Flors, C. et al. (2006) Imaging the production of singlet oxygen in vivo 17 Havaux, M. et al. (2005) Vitamin E protects against photoinhibition using a new fluorescent sensor. Singlet Oxygen Sensor Green. J. Exp. and photooxidative stress in Arabidopsis thaliana. Plant Cell 17, 3451– Bot. 57, 1725–1734 3469 39 Hideg, E. et al. (2006) Singlet oxygen in plants – its significance and 18 Vavilin, D.V. et al. (1998) Membrane lipid peroxidation, cell viability possible detection with double (fluorescent and spin) indicator and Photosystem II activity in the green alga Chlorella pyrenoidosa reagents. Photochem. Photobiol DOI: 10.1562/2006-02-06-RA-797 subjected to various stress conditions. J. Photochem. Photobiol. B. Biol (http://phot.allenpress.com) 42, 233–239 40 Ostergaard, H. et al. (2001) Shedding light on disulfide bond formation: 19 Stallaert, V.M. et al. (1995) Lipid peroxidation in tobacco leaves treated engineering a redox switch in green fluorescent protein. EMBO J. 20, with the elicitor cryptogein: evaluation by high-temperature 5853–5862 thermoluminescence emission and chlorophyll fluorescence. Biochim. 41 Dooley, C.T. et al. (2004) Imaging dynamic redox changes in Biophys. Acta 1229, 290–295 mammalian cells with green fluorescent protein indicators. J. Biol. 20 Marder, J.B. et al. (1998) Light-independent thermoluminescence from Chem. 279, 22284–22293 thylakoids of greening barley leaves. Evidence of oxygen radicals and 42 Bjo¨rnberg, O. et al. (2006) Measuring intracellular redox conditions free chlorophyll. Physiol. Plant. 104, 713–719 using GFP-based sensors. Antioxid. Redox Signal. 8, 354–361 21 Mu¨ ller-Moule´,P.et al. (2003) Zeaxanthin deficiency enhances the high 43 Jiang, K. et al. (2006) Expression and characterization of a redox- light sensitivity of an ascorbate-deficient mutant of Arabidopsis. Plant sensing green fluorescent protein (reduction-oxidation-sensitive Physiol. 133, 748–760 green fluorescence protein) in Arabidopsis. Plant Physiol. 141, 397– 22 Albertini, R. and Abuja, P.M. (1998) Monitoring of low-density 403 lipoprotein oxidation by low-level chemiluminescence. Free Radic. 44 Chaerle, L. and Van der Straeten, D. (2000) Imaging techniques and Res. 29, 75–83 the early detection of plant stress. Trends Plant Sci. 5, 495–501

www.sciencedirect.com