|Home | About | Journals | Submit | Contact Us | Français|
Haberlea rhodopensis is a perennial, herbaceous, saxicolous, poikilohydric flowering plant that is able to survive desiccation to air-dried state under irradiance below 30 µmol m−2 s−1. However, desiccation at irradiance of 350 µmol m−2 s−1 induced irreversible changes in the photosynthetic apparatus, and mature leaves did not recover after rehydration. The aim here was to establish the causes and mechanisms of irreversible damage of the photosynthetic apparatus due to dehydration at high irradiance, and to elucidate the mechanisms determining recovery.
Changes in chloroplast structure, CO2 assimilation, chlorophyll fluorescence parameters, fluorescence imaging and the polypeptide patterns during desiccation of Haberlea under medium (100 µmol m−2 s−1; ML) irradiance were compared with those under low (30 µmol m−2 s−1; LL) irradiance.
Well-watered plants (control) at 100 µmol m−2 s−1 were not damaged. Plants desiccated at LL or ML had similar rates of water loss. Dehydration at ML decreased the quantum efficiency of photosystem II photochemistry, and particularly the CO2 assimilation rate, more rapidly than at LL. Dehydration induced accumulation of stress proteins in leaves under both LL and ML. Photosynthetic activity and polypeptide composition were completely restored in LL plants after 1 week of rehydration, but changes persisted under ML conditions. Electron microscopy of structural changes in the chloroplast showed that the thylakoid lumen is filled with an electron-dense substance (dense luminal substance, DLS), while the thylakoid membranes are lightly stained. Upon dehydration and rehydration the DLS thinned and disappeared, the time course largely depending on the illumination: whereas DLS persisted during desiccation and started to disappear during late recovery under LL, it disappeared from the onset of dehydration and later was completely lost under ML.
Accumulation of DLS (possibly phenolics) in the thylakoid lumen is demonstrated and is proposed as a mechanism protecting the thylakoid membranes of H. rhodopensis during desiccation and recovery under LL. Disappearance of DLS during desiccation in ML could leave the thylakoid membranes without protection, allowing oxidative damage during dehydration and the initial rehydration, thus preventing recovery of photosynthesis.
Haberlea rhodopensis is a perennial, herbaceous, rock-dwelling, desiccation-tolerant, poikilohydric flowering plant that is homoiochlorophyllous (i.e. chlorophyll-retaining) and able to survive desiccation to air-dried state. Upon watering, the plants rapidly rehydrate and metabolism is restored to its former state. H. rhodopensis, an endemic species of the Balkan Peninsula, is a tertiary relict descended from the tropical–subtropical family Gesneriaceae. It inhabits mostly shaded, northern, chiefly limestone slopes in mountain zones with relatively high humidity. Previous investigations have shown that detached Haberlea leaves, as well as whole plants, survived desiccation to water content below 10 % in the dark or at low irradiance (about 30 µmol m−2 s−1) and that photosynthetic activity recovered fully after rehydration (Georgieva et al., 2005, 2007). Drought resistance and rapid recovery after rehydration were attributed to unchanged chlorophyll (Chl) content and the maintenance of Chl–protein complexes, reversible modifications in photosystem II (PSII) electron transport, and enhanced dissipation of non-radiative energy. Also, increased synthesis of polyphenolics during desiccation of Haberlea may contribute to resistance and recovery (Georgieva et al., 2007).
Under field conditions, drought is frequently accompanied by high irradiance, which strongly affects the vitality of plants. Energy in the thylakoid membranes, chloroplasts and cells, in excess of that needed for CO2 assimilation, may arise either from high irradiance incident on the leaf or as a consequence of drought stress, which closes the stomata and restricts CO2 supply. Excess energy is potentially harmful and can induce enhanced production of reactive oxygen species (ROS) (Mittler, 2002) leading to photoinhibition of PSII reaction centres (Yuan et al., 2005), damage of ATP synthase (Lawlor and Tezara, 2009), and severe inhibition of photosynthesis and growth (Lawlor and Cornic, 2002; Chaves et al., 2003). Plants have mechanisms that protect the photosynthetic apparatus against excess light during dehydration as it is very sensitive and liable to injury (Tuba et al., 1996). Angiosperm resurrection plants can prevent the absorption of excess light and accumulation of free radicals by leaf movements, and by the production of ‘sunscreen’ pigments (e.g. carotenoids and anthocyanins) (Sherwin and Farrant, 1998). Upregulation of the synthesis of free radical-quenching molecules and enzymes is of great importance in protection against the deleterious effects of excess light (Sherwin and Farrant, 1998). Ascorbate, glutathione, carotenoids, anthocyanins, plant phenols, osmolytes, proteins (e.g. peroxiredoxin) and amphiphilic molecules (e.g. tocopherol) can function as ROS scavengers (Smirnoff and Cumbes, 1989; Apel and Hirt, 2004; Kruk et al., 2005). Compatible solutes can also act by directly stabilizing membranes and/or proteins (Hoekstra et al., 2004).
Craterostigma wilmsii and Myrothamnus flabellifolius, desiccation-tolerant resurrection species, are able to decrease free radical formation, in part, by leaf folding, which minimizes light–chlorophyll interaction in inner leaves of the former and adaxial surfaces of the latter (Sherwin and Farrant, 1998). Light stress during dehydration contributed significantly to the loss of viability in leaves of C. wilmsii, which were prevented from folding, and both the outermost leaves and the inner leaves died when dried under exposure to light (Farrant et al., 2003). A similar result was obtained when stems of the desert resurrection plant Selaginella lepidophylla were restrained from curling during dehydration (Casper et al., 1993). However, in M. flabellifolius, restraining does not cause subcellular damage or loss of viability of leaves dried in the light. Thus, in this species it appears that both adaxial and abaxial cells are able to protect against light stress. Farrant et al. (2003) have shown that the leaves of this species are well protected: ascorbate peroxidase and glutathione reductase activity, and anthocyanin concentrations, were far higher in M. flabellifolius than in C. wilmsii. In addition, thylakoid membranes become unstacked during drying, a feature suggested to stabilize photosynthesis and minimize ROS formation (Farrant, 2000). The resurrection fern Polypodium polypoides suffers more severe damage and recovery takes longer when plants were dried under high compared with low irradiance (Muslin and Homann, 1992). Haberlea rhodopensis was also found to be very sensitive to photoinhibition (Georgieva and Maslenkova, 2006). Desiccation of plants at 350 µmol m−2 s−1 compared with at approx. 30 µmol m−2 s−1 in the natural habitat, below trees, induced irreversible changes in the photosynthetic apparatus, and leaves (except the youngest ones) did not recover after rehydration (Georgieva et al., 2008). Thus, there is interaction between desiccation and irradiance. On this basis, the aim of the present study was to examine the changes in chloroplast structure, photosynthetic activity and polypeptide patterns during desiccation of Haberlea under medium compared with low irradiance (100 vs. 30 µmol m−2 s−1), in order to elucidate changes associated with dehydration and the reasons for the irreversible damage of the photosynthetic apparatus as a result of dehydration at high irradiance. Also, the possible role of polyphenolic compounds and their synthesis in protection of the photosynthetic system was examined by measuring blue and green fluorescence as criteria for polyphenolic synthesis.
Well-hydrated plants of Haberlea rhodopensis Friv. were collected from their natural habitat growing on rocks below trees under very low irradiance (the vicinity of Asenovgrad, Bulgaria). Adult rosettes of similar size and appearance were selected for the experiments, from the same locality. The tufts with naturally occurring thin soil layers were planted in peat-soil and transferred to growth-chambers at 22–23 °C, irradiance of either 30 or 100 µmol m−2 s−1 (either at low or medium light, respectively), 12/12-h day/night cycle and relative humidity of 60 %. After 10 d acclimation, the plants were dried by not watering: the leaves were fully desiccated after 7 d. They were then rehydrated for 7 d by spraying water on the leaves to simulate rainfall and keeping the soil damp. Measurements were made after 2 d [stage D1; relative water content (RWC) about 70 %], 4 d (stage D2; RWC about 25 %) and 7 d (stage D3; RWC about 6 %) of dehydration, and after 1 d (stage R1) and 7 d (stage R7) of rehydration. Control plants kept at either 30 or 100 µmol m−2 s−1 were regularly watered throughout the experiment and samples were always taken together with the treated samples. Mature but not old leaves of similar developmental stage were chosen during the whole period of the experiment. The different parameters (see below) were measured taking samples of the same leaf or the same group of leaves depending on the amount of sample needed. Control values did not differ significantly during the course of the experiment.
The RWC of Haberlea leaves was determined gravimetrically by weighing them before and after oven drying at 80 °C to a constant mass and expressed in a percentage of water content in water-saturated tissues, using the equation:
CO2 exchange rates were measured on intact leaves at 100 µmol m−2 s−1 photosynthetically active radiation (PAR) using an IRGA system (LCA-2, ADC, Hoddesdon, UK) operated in differential mode at ambient CO2 concentration of 360 µmol mol−1 and at 25 °C air temperature. CO2 assimilation (µmol CO2 m−2 s−1), transpiration and stomatal conductance (mmol m−2 s−1) were calculated by the IRGA system according to von Caemmerer and Farquhar (1981). Six leaves from different plants of similar age were used for the measurements.
Fluorescence induction was measured on leaves with a PAM Chlorophyll Fluorometer (Walz, Effeltrich, Germany). Leaf discs were placed into a leaf section chamber and dark-adapted for 20 min. Basal fluorescence, F0, was determined by switching on the measuring light (modulation frequency of 1·6 kHz and PFD less than 1 µmol m−2 s−1). The maximum fluorescence yield in the dark-adapted state, Fm, and in the light-adapted state Fm′ were measured by applying a 0·7-s pulse of white light (PPFD of 3500 µmol m−2 s−1; light source: KL 1500 electronic, Schott, Mainz, Germany). For the quenching analysis, actinic white light (PPFD of 100 µmol m−2 s−1, KL 1500 electronic) was provided. F0′ was measured by turning off the actinic light and applying 3 s of weak far-red light (102-FR, Walz, emission peak at 730 nm). Measurements were repeated six times.
For the fluorescence imaging at 440, 520, 690 and 740 nm, a compact flash-lamp fluorescence imaging system was used (Lichtenthaler and Babani, 2000). The light of the exciting flashing xenon lamp (16·7 Hz) was filtered via a DUG 11 (Schott) filter to ensure the appropriate exciting wavelength (λexc. = 360–370 nm). Fluorescence images at the four wavelengths were detected via a CCD video camera using appropriate interference filters. Four hundred images were accumulated to obtain the accuracy required, and were corrected for the filter sensitivity parameters, background and inhomogeneity of the exciting light, with Camille 1·05 software (Photonetics, Kehl, Germany). Measurements were repeated six times.
The chlorophyll content of leaves and thylakoids was determined in 80 % acetone according to Porra et al. (1989) with three replicates for each. Isolation of thylakoid membranes was as described by Sárvári and Nyitrai (1994). Thylakoids were solubilized in the solubilization buffer of Laemmli (1970) at room temperature for 30 min. Leaf proteins were directly extracted in solubilization buffer. Sodium dodecylsulfate–polyacrylamide gel electrophoresis was performed according to Laemmli (1970), but using a 10–18 % acrylamide gradient with 10 % glycerol in the gels (BioRad mini Protean GE apparatus, 1·5 mm). Amounts of polypeptides in the scanned bands were determined by Phoretix software (Phoretix International, Newcastle upon Tyne, UK). Different lanes were compared after normalization to equal amounts of total stained protein. Gel electrophoresis was repeated three times.
For electron microscopy, leaf pieces were fixed in 2·5 % glutaraldehyde (65 mm K–Na phosphate buffer, pH 7·2) for 2 h at room temperature. After thorough washing with the above buffer, they were post-fixed in 1 % OsO4 for 1·5 h, followed by dehydration in an ethanol series. Samples were embedded in Durcupan ACM, sectioned with a Reichert-Jung Ultracut E ultramicrotome, then stained with uranyl acetate and lead citrate. The sections were examined in a Hitachi 7100 (Hitachi Ltd, Tokyo, Japan) electron microscope. Micrographs were taken with a MegaView III camera (Soft Imaging System, Münster, Germany).
The data for control and water stress treatments were analysed statistically. Means from six replications (six leaves from different plants) were compared with Student's t-test.
The decrease in RWC of Haberlea leaves during dehydration under low (LL) or medium light (ML) irradiance (Fig. 1) was very similar and plants became air-dried in 7 d (6 % RWC, stage D3). Following application of water, the RWC of Haberlea plants desiccated at LL increased very rapidly to about 80 % of the control value within 24 h (stage R1). However, despite a small increase in RWC of plants desiccated at ML 24 h after watering (17 % RWC, stage R1), their RWC was only 10 % of the control after 7 d recovery (stage R7). Exposure of well-watered (control, C7) Haberlea plants to ML did not influence leaf water content.
The maximum quantum efficiency of PSII photochemistry, estimated by the ratio variable (Fv=Fm−F0) to maximum chlorophyll fluorescence, Fv/Fm, decreased more in plants desiccated at ML than at LL: the difference was most pronounced at 25 % RWC (stage D2; Fig. 2A). The reduction of Fv/Fm was due to a decrease in both the basal fluorescence, F0, and especially that of the maximum Chl fluorescence, Fm (data not shown). The quantum efficiency of PSII of plants desiccated at LL almost returned to the control value after 24 h rehydration and recovered fully after 7 d. However, rehydration of severely desiccated plants at ML led to an initial enhancement of the Fv/Fm ratio (stage R1), but it decreased upon further rehydration (stage R7). Values of Fv/Fm of control plants kept either at LL or at ML during the experiment were very close to the initial values (see C and C7, Fig. 2A).
The actual quantum efficiency of PSII linear electron transport during illumination, ΦPSII (Genty et al., 1989), decreased more than primary PSII photochemistry (Fv/Fm) as a result of dehydration of Haberlea plants under both LL and ML (Fig. 2B). Inhibition of ΦPSII during dehydration of Haberlea leaves was due to decreased efficiency of both the excitation capture by open PSII reaction centres (Fv′/Fm′) and, particularly, of photochemical quenching, qP (data not shown). Dehydration of plants to 25 % RWC (stage D2) under LL reduced Fv′/Fm′ and qP by 10 and 50 %, respectively, while their values were decreased by 50 and 75 % under ML. Similarly to Fv/Fm, ΦPSII recovered rapidly following rehydration of plants desiccated at LL but despite some increase after 24 h rehydration (37 % of the control value in stage R1), finally ΦPSII did not recover under ML. The values of ΦPSII of control plants exposed to LL or ML were 0·642 ± 0·006 and 0·690 ± 0·009, respectively, at the end of the experiment (C7, Fig. 2B).
The reduction of ΦPSII during dehydration was accompanied by a corresponding increase in (1 – Fv′/Fm′), which is indicative of an increased proportion of thermal energy dissipation in the antenna (Demmig-Adams et al., 1996; Fig. 2C). This increase in thermal energy dissipation started earlier during drying at ML than at LL, despite the similar RWC.
The fluorescence ratios blue/red and blue/far-red proved to be very good stress indicators (Buschmann et al., 2000). The blue (F440) and green (F520) fluorescence emission increased in the course of desiccation of H. rhodopensis at LL and, together with reduced Chl fluorescence, enhanced the blue/red (F440/F690) and green/red (F520/F690) ratios. The blue/red ratio (F440/F690) was five times greater than the control value in LL dried leaves (6 % RWC, stage D3) but rapidly decreased upon rehydration (Fig. 3). The green/red (F520/F690), blue/far-red (F440/F740) and green/far-red (F520/F740) ratios changed in a similar way (data not shown). The stronger enhancement of the investigated ratios in plants desiccated at higher irradiance was mainly due to greater reduction of red and far-red Chl fluorescence, whereas blue and green fluorescence increased similarly to plants desiccated at LL. The ratios recovered totally at LL but did not change significantly after rehydration at ML. Exposure of well-watered (control) Haberlea plants to ML did not significantly influence the blue–green fluorescence emission.
The rate of net CO2 assimilation (PN) decreased sharply as a result of dehydration, particularly under ML (Fig. 4). During the first phase of dehydration under LL and ML (70 % RWC, stage D1), CO2 assimilation declined by 50 and 70 %, respectively, and after 7 d desiccation (6 % RWC, stage D3) there was no net CO2 assimilation, but the illuminated desiccated leaves showed respiratory activity. Seven days after rehydration (R7) of plants air-dried at LL, the CO2 assimilation rate was similar to the control (Fig. 4A). However, there was hardly any recovery of CO2 assimilation after rehydration of plants desiccated under ML. Transpiration rate and stomatal conductance changed in a similar way to CO2 assimilation during dehydration under both LL and ML, but they recovered earlier than CO2 assimilation during rehydration at LL (Fig. 4). In contrast to the unaffected photochemical activity, the CO2 assimilation declined in control plants kept at ML and was reduced by about 30 % at the end of experiment (C7; from 3·7 ± 0·4 to 2·6 ± 0·3 µmol CO2 m−2 s−1). A similar decrease was also shown in the stomatal conductance of these leaves.
The total Chl content remained constant during desiccation of H. rhodopensis at both LL and ML (Table 1). The Chl a/b ratio of leaves and the corresponding isolated thylakoids did not differ significantly. It decreased slightly during desiccation, and recovered or showed some recovery during rehydration at LL and ML, respectively.
In accordance with the minimal changes in Chl content and Chl a/b ratio, only non-significant minor changes of thylakoid composition were observed during dehydration (Fig. 5). A tendency for a relative decrease in PSII and the low-molecular-weight region and increase in the light-harvesting complex (LHC) region was observed, particularly in ML plants.
In contrast to the polypeptide patterns of thylakoids, those of whole leaves changed considerably. Drying increased the amounts of the bands at 70, 47, 33, 22 and 16 kDa under both LL and ML conditions (Fig. 6). In addition, ML plants also accumulated 55-, 40- and 38-kDa polypeptides. Band heights of the above-mentioned polypeptides on the densitograms were 1·5–2 times greater in the D3 stage than those of the corresponding controls. Amount of the above-mentioned polypeptides, however, started to increase at different times during treatment, the exact resolution of which will require more detailed study. These changes were completely reversed in LL plants 1 week after rewatering, but persisted in plants dried under ML. Although the polypeptide pattern did not recover in ML plants, the increased amount of the 47-, 33- and 22-kDa polypeptides was reduced during rehydration. The polypeptide pattern of control leaves did not differ significantly either with irradiance or during treatment.
Figures 7 and and88 show the chloroplast structure in control, desiccated and rehydrated plants under LL and ML. In accordance with the findings of Markovska et al. (1994), rounding up and dislocation of the chloroplasts toward the cell interior during desiccation was observed (not shown). Although this process proved to be reversible under LL conditions, ML inhibited recovery of the normal cellular arrangement. Viewing the chloroplasts at a higher magnification revealed ‘inverse contrast’ (Keresztes and Sárvári, 2001a) in the thylakoids of this plant species, which has not previously been reported for the staining properties of thylakoids or for Haberlea chloroplasts. This ‘inverse’ staining pattern indicates that the thylakoid lumina are filled with an electron-dense substance (dense luminal substance, DLS), while the thylakoid membrane around it is lightly stained, i.e. it seems almost entirely electron-transparent. This was observed both in the initial and 7-d control samples whether kept at LL or ML (Figs 7A, E and 8A, E). Upon desiccation or rehydration the DLS thinned and disappeared, but the time course of this change depended considerably on the intensity of illumination. Under LL the DLS persisted during dehydration (Fig. 7B), and started to disappear from some of the mesophyll cells just at the end of rehydration (Fig. 7C, D). Under ML, however, the DLS started to disappear from the thylakoid lumina as early as in the D1 stage (Fig. 8B), and this continued during the D2 stage. In D3 the thylakoid lumina were all transparent, and the thylakoid membranes were all strongly stained (Fig. 8C). Under ML, rewatering did not result in recovery, some thylakoids became swollen and there were no starch grains, but instead, electron-dense plastoglobuli accumulated (Fig. 8D).
In order to elucidate the causes of the irreversible changes seen in the structure and function of the photosynthetic apparatus of the resurrection plant Haberlea rhodopensis during desiccation at high (350 µmol m−2 s−1) light intensity (Georgieva et al., 2008), the different stages of dehydration under low (about 30 µmol m−2 s−1) and medium (100 µmol m−2 s−1) light irradiance were compared. Plants were known to recover perfectly under LL (Georgieva et al., 2007). In the present study, however, ML was used instead of the strongly damaging high light (350 µmol m−2 s−1) to slow down the changes that occurred at high light intensity.
Well-watered (control) plants kept at 100 µmol m−2 s−1 (ML) throughout the experiment showed no evidence of damage except that their CO2 assimilation decreased. However, regardless of the similar rate of water loss of plants desiccated either at LL or at ML (Fig. 1), all the damaging effects were more pronounced at ML. Dehydration at ML decreased the quantum efficiency of PSII photochemistry (Fig. 2) and the rate of net CO2 assimilation (Fig. 4) more than dehydration at LL. During the first stage of dehydration (stage D1, 70 % RWC), CO2 assimilation was decreased due to stomatal closure (Fig. 4). The further decrease and inhibition of CO2 assimilation was caused by the decrease in both stomatal conductance and photochemical activity. The presence of light during dehydration can be extremely damaging to actively photosynthesizing tissues. Under (even mild) water stress conditions, closure of stomata can result in excitation energy being transferred from chlorophyll to oxygen with the subsequent formation of oxygen-free radicals (Smirnoff, 1993; Sherwin and Farrant, 1998). After inactivation of electron transport by desiccation, photon energy continues to be absorbed by photosynthetic pigments. The retention of chlorophyll molecules during desiccation in H. rhodopensis (i.e. it is a homoiochlorophyllous, desiccation-tolerant plant) could also be a source of potentially harmful singlet oxygen production. In spite of the changes in photosynthetic activity, only minor differences were observed in thylakoid composition (Fig. 5) and structure (Fig. 8) during drying even at ML. In contrast, the amount of some proteins greatly increased in drying leaves, particularly under ML conditions. Their accumulation was strictly parallel to a decrease in the photosynthetic activity of leaves, and clearly indicated stress conditions. Although they have not been identified, these proteins may include dehydrins and LEA (late embryogenesis abundant) proteins, which protect membranes/enzymes against dehydration (Swayze, 2004; Vicré et al., 2004), ROS-scavenging enzymes and chaperones reactivating damaged proteins (Bartels, 2005; Jiang et al., 2007).
Both photosynthetic activity and compositional changes of Haberlea leaves reversed rapidly during rehydration at LL, which suggests that the photosynthetic apparatus had not been damaged by desiccation per se (Figs 2 and and4).4). Loss of photosynthetic activity was faster under drying at ML than at LL, which suggests light-dependent damage during dehydration. However, the transitory recovery of water content and photosynthetic activity in leaves of plants re-watered at ML demonstrated that not only was desiccation strongly affected by the light conditions but so too was the recovery of plants after rehydration. The lack of recovery of photosynthetic activity at ML may be related to the poor re-hydration of the tissues caused by ROS-induced damage of membranes and/or stomatal functioning (Farrant, 2000).
Angiosperm resurrection plants appear to have a number of mechanisms, which vary among species, to minimize photo-oxidative damage. The increased non-radiative energy dissipation during desiccation contributed to protecting PSII from photoinhibition both at LL and at ML (Fig. 2C). Furthermore, extensive shrinkage and some folding of Haberlea leaves also occurred during desiccation, which would decrease energy interception and avoid light-induced damage. Leaf folding started when the RWC was less than 50 % and shrinkage started when the leaves lost most of their water: leaf area decreases by about 40 % (Georgieva et al., 2008). Moreover, during dehydration of Haberlea leaves blue and green fluorescence emission increased, as did the blue–green/red–far red fluorescence ratios (Fig. 3). The blue and green fluorescence has been generally attributed to cell-wall-bound ferulic acid, and thus this increase may be connected to the shrinkage of leaves. However, the contribution of flavonoid compounds to the blue and green fluorescence was also suggested by Lichtenthaler and Schweiger (1998); Apostol et al. (2003) confirmed this by using a mutant deficient in flavonoids (Hideg et al., 2002). The extensive shrinkage and folding of Haberlea leaves and the increased polyphenolic synthesis and thermal energy dissipation, which occurred at both LL and ML under desiccation, was not enough for successful protection of photosynthetic apparatus in plants dehydrated at ML, although some transitional regeneration has been observed (stage R1; Fig. 2).
The greater resistance of the photosynthetic apparatus to LL than to ML, not only under recovery but also during desiccation, may be connected with the functioning of another protective mechanism. We propose that accumulation of a DLS in the thylakoid lumen, as revealed by inverse staining, is an important mechanism in Haberlea. Although the DLS has not been chemically identified, on the basis of fixation, staining and extraction probes it is probably a phenolic compound (van Steveninck and van Steveninck, 1980; Keresztes and Sárvári, 2001a). Some phenolics are considered to be antioxidants (Sgherri et al., 2004) and ROS scavengers (Kruk et al., 2005). However, the dense phenolic substance, whether present in the thylakoid lumen or in the stroma, i.e. being on either side of the membrane, is thought to prevent conformational changes of membrane lipids (Olesen, 1978; Keresztes and Sárvári, 2001b). It is tempting to suppose that those plant species or families that are able to produce and accumulate soluble phenolics may also use these special substances for the stabilization of thylakoid membranes. With this supposition, the presence and absence of DLS in the thylakoid lumen in Haberlea during dehydration–rehydration may explain the functional survival of chloroplasts and the lack of recovery, respectively. At LL the DLS is continuously present during dehydration, so the thylakoid membrane is protected. The DLS disappears just by the end of rehydration. Photosynthesis resumes, as judged from gas exchange measurements and from the reappearance of starch grains. This is accompanied by the disappearance of proteins accumulated as a consequence of desiccation. At ML the DLS started to disappear from the onset of dehydration, so the thylakoid membranes remained without protection. Thus, the disappearance of DLS may have contributed to the stronger damage and lack of recovery of photosynthesis. The changes in the amount of DLS are not reflected in the changes of the blue–green fluorescence of leaves (Fig. 3), the latter rather being determined by substances in the surface layers. The chemical nature of the DLS remains to be identified.
In Haberlea rhodopensis during drying and rehydration under illumination, the extent of loss and reappearance of photosynthetic activity correlates strongly with the presence or absence of DLS. This is possibly a phenolic substance that provides protection against oxidative damage caused by the excess light under drought stress and the early phase of recovery.