Plasma membrane lipids in the resurrection plant Ramonda serbica following dehydration and rehydration

Journal of Experimental Botany, Vol. 53, No. 378, pp. 2159±2166, November 2002 DOI: 10.1093/jxb/erf076 Plasma membrane lipids in the resurrection plant Ramonda serbica following dehydration and rehydration Mike F. Quartacci 1, Olivera GlisÏic 2, Branka Stevanovic 2 and Flavia Navari-Izzo 1,3 1 Dipartimento di Chimica e Biotecnologie Agrarie, UniversitaÁ di Pisa, Via del Borghetto, 80, 56124 Pisa, Italy 2 Institute of Botany, University of Belgrade, Takovska 43, 11000 Belgrade, Yugoslavia Received 10 December 2001; Accepted 3 July 2002 Abstract Plants of Ramonda serbica were dehydrated to 3.6% relative water content (RWC) by withholding water for 3 weeks, afterwards the plants were rehydrated for 1 week to 93.8% RWC. Plasma membranes were isolated from leaves using a two-phase aqueous polymer partition system. Compared with well-hydrated (control) leaves, dehydrated leaves suffered a reduction of about 75% in their plasma membrane lipid content, which returned to the control level following rewatering. Also the lipid to protein ratio decreased after dehydration, almost regaining the initial value after rehydration. Lipids extracted from the plasma membrane of fully-hydrated leaves were characterized by a high level of free sterols and a much lower level of phospholipids. Smaller amounts of cerebrosides, acylated steryl glycosides and steryl glycosides were also detected. The main phospholipids of control leaves were phosphatidylcholine and phosphatidylethanolamine, whereas sitosterol was the free sterol present in the highest amount. Following dehydration, leaf plasma membrane lipids showed a constant level of free sterols and a reduction in phospholipids compared with the well-hydrated leaves. Both phosphatidylcholine and phosphatidylethanolamine decreased following dehydration, their molar ratio remaining unchanged. Among free sterols, the remarkably high cholesterol level present in the control leaves (about 14 mol%) increased 2-fold as a result of dehydration. Dehydration caused a general decrease in the unsaturation level of individual phospholipids and total lipids as well. Upon rehydration the lipid composition of leaf plasma membranes restored very quickly approaching the levels of well-hydrated leaves. Key words: Dehydration, lipids, plasma membrane, Ramonda serbica, resurrection plants. Introduction Flowering plants growing in hot and arid regions usually survive the harsh environmental conditions either by avoiding the stressful events or by very promptly activating adaptative resistance mechanisms. Only a small number of higher plants, mostly originating from the southern hemisphere and called desiccation-tolerant or resurrection plants, are capable of surviving almost complete dehydration for prolonged periods. Ramonda spp, as well as other species belonging to the family Gesneriaceae, are among the resurrection plants which grow in the northern hemisphere. Ramonda serbica is a rare resurrection plant growing in the Balkan peninsula (Gaff, 1981; StevanovicÂ, 1986). This species is capable of surviving long dry periods between the wet periods, passing quickly from anabiosis, which can last much longer than three months depending on water de cit severity and temperatures, to the state of full biological activity in less than 8±10 h if the favourable water balance in the soil re-establishes suddenly. In spite of the fact that metabolic processes are almost stopped in resurrection plant dry leaves, the cell membranes as well as most of the enzymatic systems are protected in different ways (Bewley and Krochko, 1982; Oliver, 1996; Navari-Izzo and Rascio, 1999). It has been suggested that the rapid and ef cient recovery and full reconstitution of membrane organization and functionality, as well as the presence of effective membrane defence mechanisms, are the most important prerequisites for 3 To whom correspondence should be addressed. Fax: +39 050 598614. E-mail: fnavari@agr.unipi.it Abbreviations: ASG, acylated steryl glycosides; FS, free sterols; PA, phosphatidic acid; PL, phospholipids; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PG, phosphatidylglycerol; PM, plasma membrane; RWC, relative water content; SG, steryl glycosides. ã Society for Experimental Biology 2002

2160 Quartacci et al. survival upon rehydration (Stevanovic et al., 1992; Navari- Izzo et al., 1995; Navari-Izzo and Rascio, 1999; Sgherri et al., 2000). According to Oliver (1996), Ramonda serbica seems to belong to the group of resurrection plants which are able to withstand desiccation using both morphological and physiological mechanisms to slow down and, for a while, to control the rate of water loss. The results of some recent investigations indicate that Ramonda serbica has the ability to maintain cell membrane integrity, i.e. to preserve semipermeability during dehydration (Stevanovic et al., 1998), as well as to activate protective mechanisms that increase the level of zeaxanthin and the amounts of reduced ascorbate and glutathione, which are crucial for photoprotection during the dehydration/rehydration cycle (Augusti et al., 2001). Furthermore, it has been found that an increased amount of phenolic acids also protects Ramonda membranes during desiccation (Sgherri et al., 2000). In the last decade, only a few studies on lipids extracted from thylakoids or the whole plant have been undertaken in order to explain desiccation tolerance in resurrection plants (Stevanovic et al., 1992; Stefanov et al., 1992; Navari-Izzo et al., 1995; Quartacci et al., 1997). The investigations showed the general tendency of dehydrated plants to adapt their membranes to the altered conditions, and to recover quickly on rehydration both the lipid composition and the order parameters of the well hydrated plants. However, there is a complete lack of knowledge about changes of lipids during dehydration and rehydration for the plasma membrane (PM). The composition and organization of PM lipids are crucial for intracellular metabolism. Many vital activities of cells originate in the membrane, the structure and function of which are profoundly altered following water stress that leads to destructive events such as phase transition, fusion and increased permeability. The composition and physical state of the lipid bilayer in uence lipid±protein and protein±protein associations, membrane-bound enzyme activities and the carrier-mediated transport capacity of membranes (Navari-Izzo and Rascio, 1999; Leprince et al., 2000; Navari-Izzo et al., 2000; Kerkeb et al., 2001). Preserving membrane integrity, resurrection plants are capable of surviving during anabiosis and returning quickly to the complex and dynamic whole-organism functionality upon rehydration (Quartacci et al., 1997; Navari-Izzo et al., 1995, 2000). The aim of the present study was to examine the composition of lipids of plasma membranes isolated from Ramonda serbica leaves, as well as to determine changes during a dehydration/rehydration cycle in order to explain, at least in part, the plant's ability to reactivate its physiological functions so rapidly after rewatering. Materials and methods Plant material Specimens of the desiccation-tolerant plant Ramonda serbica PancÏ. & Petrov. were collected from their natural habitat in the south-east region of Serbia, in a gorge near the town of Nis. There, the plants grow on rocky slopes, exposed from north to north-east, on a thin layer of rich, mature, organo-mineral and dark mountain soil (ph 8.4) spread over limestone. During summer the habitat is characterized by high air temperatures and a remarkable decrease in air humidity and the plants pass to, and stay in, the anabiotic state, although they never receive sunlight directly. Plants of the same age were harvested together with the layer of soil on which they grew. After collection, plants were acclimated for 2 weeks keeping them fully watered until the beginning of the experiments. Plants were dehydrated for 3 weeks by withholding water at room temperature and ambient photoperiod. Rehydration was started by spraying the plants with water to simulate rainfall and keeping the soil damp. The rehydrated samples were collected after 1 week during which they were watered daily. Relative water content At regular intervals during the dehydration/rehydration cycle measurements of the relative water content (RWC) of the leaves were carried out as previously reported (Sgherri et al., 1994). For the analyses, mature and fully expanded leaves from the middle of the rosettes and comparable in size were selected. The RWC of the leaves was calculated according to the formula: 1003[(fresh weight±dry weight)/(saturated weight±dry weight)] and expressed as the mean value of ten replicates for each treatment. Solute leakage For solute leakage determination samples of the same weight were obtained from leaves comparable in size. The plant material was washed in double-distilled water to remove the contents of the cut cells, soaked in 25 ml of double-distilled water, shaken at room temperature for 24 h and aliquots for leachate measurements were taken. Samples were then immersed for 5 min in liquid N 2, placed again in the same vial containing the leachate and shaken for an additional hour prior to the measurement of the maximum conductivity (Metrohm 660 conductometer). The injury index was calculated according to the formula: injury index %=1±[(T±C)/T]3100, where T and C represent the conductivity of the leachate after and before liquid N 2 treatment, respectively. Plasma membrane preparation Plasma membranes were prepared using a two-phase aqueous polymer partition system. Leaves were cut into pieces and immediately homogenized in the isolation medium consisting of 50 mm Tris-HCl, ph 7.5, 0.25 M sucrose, 3 mm Na 2 EDTA, 10 mm ascorbic acid, and 5 mm diethyldithiocarbamic acid. The homogenate was ltered through four layers of a nylon cloth and centrifuged at 10 000 g for 10 min. The supernatant was further centrifugated at 65 000 g for 30 min to yield a microsomal pellet, which was resuspended in 2 ml of a resuspension buffer (5 mm K-phosphate, ph 7.8, 0.25 M sucrose, and 3 mm KCl). Plasma membranes were isolated by loading the microsomal suspension (1.0 g) onto an aqueous twophase polymer system to give a nal concentration of 6.6% (w/w) Dextran T500, 6.6% (w/w) polyethylenglycol, 5 mm K-phosphate (ph 7.8), 0.25 M sucrose, and 3 mm KCl. The PM was further puri ed using a two-step batch procedure. The resulting upper-phase was diluted 4-fold with 50 mm Tris-HCl, ph 7.5, containg 0.25 M sucrose, and centrifuged for 30 min at 100 000 g. The resultant PM pellet was resuspended in the same buffer containing 30% ethylenglycol and stored at ±80 C for lipid analyses. All steps of

Plasma membrane lipids in Ramonda 2161 Table 1. Distribution of marker enzymes (mmol mg ±1 protein min ±1 ) in the upper and lower phases of the partition system used for the isolation of plasma membrane from well hydrated and dehydrated leaves of Ramonda serbica Results are the means of three independent experiments 6SE. UP, upper phase; LP, lower phase. Marker enzymes Well hydrated leaves Dehydrated leaves UP LP UP LP ATPase 0.42160.021 0.26760.019 0.51460.022 0.31760.015 Vanadate-sensitive ATPase 0.06360.005 0.15560.014 0.03160.004 0.20160.012 NO ± 3 -sensitive ATPase 0.37460.026 0.02560.004 0.51260.020 0.06360.004 NADH-cyt c reductase 0.00660.001 0.18760.010 0.03260.003 0.22760.011 Cyt c oxidase 0.00460.001 0.03060.003 0.00460.001 0.09860.009 Latent IDPase 0.04160.005 0.10660.009 0.02160.003 0.07260.006 the isolation were carried out at 4 C. Plasma membrane pellets for enzyme activity determinations were used immediately. In order to check the purity of the PM, the activity of the vanadatesensitive ATPase as a marker enzyme was determined (Table 1). Cytochrome c oxidase, antimycin A-insensitive NADH cyt c reductase, NO 3 ± -sensitive ATPase and latent IDPase activities were used as markers of mitochondria, endoplasmic reticulum, tonoplast, and Golgi membranes, respectively (Navari-Izzo et al., 1993). Chlorophyll was not detected in the PM fraction. Tests with the markers showed that the partition behaviour of both PM and intracellular membranes may be in uenced by water contents since the net charge density of membranes is also related to their polar head group composition. The protein content was determined taking aliquots of the PM suspension. The analysis was performed according to Bradford (1976) with bovine serum albumin as a standard. Lipid extraction and separation Lipids were extracted from the PM suspension by the addition of boiling isopropanol followed by chloroform:methanol (2:1 v/v) containing butylhydroxytoluol (50 mg ml ±1 ) as an antioxidant. The solvent mixture was then washed with 0.88% KCl to separate the chloroform phase. The upper water phase was re-extracted with chloroform, the chloroform phases combined and dried under a stream of N 2. The lipid extracts were stored at ±20 C and retained for further separation. Lipids were fractionated into neutral lipid, glycolipid and phospholipid (PL) fractions on Sep-Pak cartridges (Waters) (Uemura and Steponkus, 1994). Lipid extracts dissolved in chloroform:acetic acid (100:1 v/v) were transferred to the Sep-Pak cartridges and sequentially eluted with 20 ml of chloroform:acetic acid (100:1 v/v) for neutral lipids, 10 ml of acetone and 10 ml of acetone:acetic acid (100:1 v/v) for glycolipids and 7.5 ml of methanol:chloroform:water (100:1 v/v) for PL. Chloroform (2.25 ml) and water (3 ml) were added successively to the eluate containing the PL to obtain a phase separation and to facilitate their recovery. The separation of individual lipids was performed by TLC (Silica Gel 60, 0.25 mm thickness; Merck) with the following solvent mixture: petroleum ether:ethyl ether:acetic acid (80:35:1 by vol.) for neutral lipids (free sterols and sterol esters); chloroform:- methanol:water (65:25:4 by vol.) for glycolipids (steryl glycosides and cerebrosides); chloroform:methanol:acetic acid:water (85:15:10:3.5 by vol.) for PL. After development, the bands were located with iodine vapour or spraying the plates with 0.1% Rhodamine 6G in ethanol. Individual lipids were identi ed by cochromatography with authentic standards. Quanti cation of lipids after TLC Quantitative analyses of sterols, cerebrosides (CER) and PL were performed as reported by Navari-Izzo et al. (1993) using cholesterol, glucose and KH 2 PO 4 as standards, respectively. All procedures were performed in the presence of silica gel from TLC. Sterol analysis Individual free sterol (FS) components were separated and quantiti ed by GLC as underivatized residues. The sterol moieties dissolved in ethyl acetate were analysed with a Perkin-Elmer Sigma 2B gas chromatograph using a ame ionization detector and a 30 m30.32 mm SPB-5 fused silica capillary column (Supelco). The operating conditions were: column temperature 250 C, injector and detector temperatures 280 C, N 2 was the carrier gas at 1 ml min ±1 (split ratio 1:70). Compound identi cation was made on the basis of the retention time relative to known standards. Cholestane was the internal standard, and corrections were made for differences in detector response. Fatty acid analysis The fatty acid methyl ester derivates from individual and total PL were obtained as previously described (Quartacci et al., 1997) and separated by GLC on a Dani 86.10 HT gas chromatograph equipped with a 60 m30.32 mm SP-2340 fused silica capillary column (Supelco) coupled to a ame ionization detector (column temperature 175 C). Both the injector and detector were maintained at 250 C. Nitrogen was used as the carrier gas at 0.9 ml min ±1 with a split injector system (split ratio 1:100). Statistical analysis A completely random experimental design was run in triplicate. Data from each experimental design determined in triplicate were analysed by a one-way analysis of variance. The signi cance of differences was determined according to Tukey's test. P values <0.05 are considered to be signi cant. Results Following dehydration by withholding water for 3 weeks the RWC decreased from 87.0% in the fully hydrated plants to the value of 3.6% in the desiccated ones, dehydrating very slowly especially in the rst 15 d. After rewetting, the plants regained quickly their hydration state reaching the RWC of 93.8% after a week (Fig. 1). The injury index calculated from the solute leakage measurements decreased from 13.0% in the well-watered plants to 6.8% in the desiccated ones, regaining the value of 13.2% in the rehydrated leaves (not shown).

2162 Quartacci et al. Fig. 1. Relative water content (RWC) of leaves of Ramonda serbica following dehydration and rehydration. Results are the means 6SE of ten measurements. In PM isolated from fully hydrated Ramonda leaves a total lipid content of 1.05 mmol mg ±1 protein was detected (Table 2). The amount of PM total lipids, as well as all the individual components of the dried leaves suffered a dramatic reduction and were reduced to one-quarter of the hydrated leaves. Upon rehydration, the PM lipid content of leaves was restored and approached the amount of the hydrated leaves (0.93 mmol mg ±1 protein). The lipid to protein ratio of PM showed a reduction in the dehydrated leaves from 3.5 to 2.1, but upon rehydration regained the value of 3.1. The same trend was followed also by the PL to FS molar ratio which decreased by 20% in the dehydrated leaves (Table 2). The main PM lipids of Ramonda serbica leaves were FS which accounted for more than half of the total lipids in both hydrated and desiccated plants (Table 2). Their proportion did not change during the dehydration/rehydration cycle. The other PM lipids included a relatively large amount of PL (which declined from 30.6 to 25.8 mol% during dehydration and regained the control value upon rehydration), and a smaller amount of CER (increasing from 6.6 to 10.9 mol% following dehydration and recovering the control value in rehydrayed leaves). Acylated steryl glycosides (ASG) decreased during desiccation from 5.1 to 2.0 mol% and were restored upon rehydration. A small proportion of steryl glycosides (SG) was also detected, which remained constant during the dehydration/rehydration cycle (Table 2). The predominant PM phospholipids of hydrated leaves were phosphatidylcholine (PC) and phosphatidylethanolamine (PE) (Table 3). During dehydration the proportions of these PLs declined by about 50%, regaining the control levels when rehydrated. The other PLs were present in smaller amounts and increased remarkably during dehydration, especially PA which almost doubled its level from Table 2. Lipid composition (mmol mg ±1 protein), PL to FS molar ratio and lipid to protein mass ratio of plasma membranes isolated from leaves of Ramonda serbica during dehydration and rehydration Lipid to protein ratio is expressed as mg mg ±1. In brackets, lipid class proportion (mol%) relative to total content. Results are the means of three independent experiments. For comparisons among means an analysis of variance was used. For each treatment means in rows followed by different letters are signi cantly different at P <0.05 level. tr, trace. Hydrated Dehydrated Rehydrated PL 0.32 b (30.6) 0.06 a (25.8) 0.29 b (31.3) FS 0.59 b (56.2) 0.14 a (60.1) 0.52 b (55.8) ASG 0.05 b (5.1) 0.01 a (2) 0.05 b (5.8) SG 0.02 a (1.5) tr (1.2) 0.02 a (1.7) CER 0.07 b (6.6) 0.03 a (10.9) 0.05 a (5.4) Total content 1.05 b 0.24 a 0.93 b PL/FS 0.54 b 0.43 a 0.56 b Lipid/protein 3.5 b 2.1 a 3.1 b 8.3 to 15 mol%. In the PM isolated from rehydrated leaves the individual PLs approached the values of the wellhydrated leaves with the exception of PG which further decreased. The changes in the PL composition due to water shortage did not cause any variation in the PC to PE molar ratio, which remained constant during the dehydration/ rehydration cycle (Table 3). The most abundant PM free sterol was sitosterol followed by campesterol and cholesterol with lesser amounts of stigmasterol (Table 4). The sitosterol level decreased constantly during the dehydration/rehydration cycle (from 54.8 to 45.3 mol%), whereas stigmasterol remained constant. Campesterol signi cantly decreased in the desiccated leaves (from 22.6 to 14.8 mol%) and exceeded the control value in the rehydrated ones (33.8 mol%). The cholesterol level signi cantly increased during dehydration, being in desiccated leaves 2-fold higher than in well-hydrated plants, and then regained the amount of 13.5 mol% upon rehydration (Table 4). The more planar (cholesterol+campesterol) to less planar (sitosterol+stigmasterol) molar ratio continuously increased during dehydration and rehydration. The main fatty acids of PM phospholipids were palmitic (16:0) and linoleic (18:2) acids, followed by lower amounts of oleic (18:1) and stearic (18:0) acids, and much smaller proportions of linolenic (18:3), miristic (14:0), and palmitoleic (16:1) fatty acids (Table 5). Palmitic acid was the most abundant fatty acid in all the individual PLs with the exception of PC, in which the main fatty acid, linoleic acid, drastically decreased by 3-fold or more in dried leaves. Almost all fatty acids, especially the most abundant ones, i.e. 16:0 and 18:2, quickly restored their levels in rehydrated leaves, returning almost to the same value of the fresh leaves, with the exception of 18:0

Plasma membrane lipids in Ramonda 2163 Table 3. Phospholipid composition (mol%) of plasma membranes isolated from leaves of Ramonda serbica during dehydration and rehydration Results are the means of three independent experiments. For comparisons among means an analysis of variance was used. The signi cance of the letters is the same as in Table 2. Hydrated Dehydrated Rehydrated PC 28.7 b 15.0 a 29.0 b PE 21.2 b 11.0 a 23.6 b PG 12.8 b 18.6 c 6.5 a PI 17.5 a 22.8 b 16.9 a PS 11.5 a 17.6 a 13.0 a PA 8.3 a 15.0 b 11.0 a PC/PE 1.35 a 1.36 a 1.23 a Table 4. Free sterol composition (mol%) and more planar to less planar molar ratio of plasma membranes isolated from leaves of Ramonda serbica during dehydration and rehydration Results are the means of three independent experiments. For comparisons among means an analysis of variance was used. The signi cance of the letters is the same as in Table 2. Hydrated Dehydrated Rehydrated Cholesterol 13.8 a 28.0 b 13.5 a Campesterol 22.6 b 14.8 a 33.8 c Stigmasterol 8.8 a 8.0 a 7.4 a Sitosterol 54.8 b 49.2 a 45.3 a More planar/less planar 0.57 a 0.75 b 0.90 c which generally remained in smaller amounts in rehydrated leaves. As for the degree of unsaturation, dehydration induced a higher saturation in the PM fatty acids compared with well-hydrated or rehydrated leaves (Table 5). Discussion The ability of R. serbica, as well as of other poikilohydric plants, to survive complete desiccation, i.e. to live in an anabiotic state, is the result of adaptations that both maintain the structure of membranes or allow it to be regained during rewatering and also prevent the functional impairment of cell metabolism during water loss and subsequent rehydration. Indeed, in spite of the changes observed in PM lipid composition, R. serbica was capable of pursuing its normal metabolic activity upon rehydration, rapidly recovering without accelerating physiological ageing as shown by its persistence over time, i.e. viability in the vegetative state, as well as by its capacity to ower and to enter again into anabiosis. A decrease in the lipid content is a common response of plants to water de cit and, in general, to environmental stresses (Navari-Izzo and Rascio, 1999). A similar behaviour has already been found in Ramonda species (Stevanovic et al., 1992) and in the resurrection plants Boea hygroscopica and Sporobolus stap anus following severe dehydration (Navari-Izzo et al., 1995, 2000; Quartacci et al., 1997). The reduction in lipids following dehydration (Table 2) is generally interpreted as causing a decrease in the total membrane area of the cells, and may alter the speci c interactions between lipids and membrane-intrinsic proteins, essential for the maintenance of membrane integrity (HernandeÂz and Cooke, 1997). The PM isolated from Ramonda leaves showed a relatively high FS level in comparison with PL (Table 2). High sterol contents are not a unique characteristic of the PM of this resurrection plant. Similar levels were also observed in other species such as rye, potato and barley (Lynch and Steponkus, 1987; Palta et al., 1993; Rochester et al., 1987) and in the halophyte species Spartina patens (Wu et al., 1998). The lipid bilayer is the major barrier to free diffusion in the selectively permeable membrane, and the permeability properties of the bilayer are greatly in uenced by its chemical composition and, in particular, by steryl lipids (Navari-Izzo et al., 1993). Under physiological conditions, FS act as the main lipid rigidi er by increasing the ef ciency of PL packing. Sterol enrichment of membranes has been interpreted as a mechanism of adaptation based on sterol-induced membrane rigidi cation (Yoshida and Uemura, 1990; Quartacci et al., 2001). The increase in the FS to PL molar ratio during dehydration may be an indication of reduced uidity of the PM, as also suggested indirectly by the lower lipid to protein ratio (Table 2) and the injury index, and might have altered the physical architecture and permeability of membranes. The increase in FS upon dehydration, observed earlier in water-stressed maize, soybean and sun ower (Navari-Izzo et al., 1988, 1989, 1990, 1993) may provide an advantage to plants growing under water-de cit conditions since it has been shown that higher sterol amounts in the bilayer reduce the rate of permeation by water (Schroeder, 1984). Besides the FS level, their composition (Table 4) also alters membrane status because of the speci c effect of the individual sterol involved (Navari-Izzo et al., 1993). It is worth mentioning that PM from leaves of R. serbica, irrespective of their hydration state, were characterized by a relatively high amount of cholesterol (Table 4) as already found in leaves of Boea hygroscopica (Navari-Izzo et al., 1995). Among free sterols, cholesterol has been found to be more effective in controlling membrane permeability and uidity due to the more planar con guration of the molecule (Grunwald, 1974). The more planar to less planar sterol molar ratio plays a fundamental role in allowing the plant to tolerate stress, as the ratio is considered to be an index of membrane permeability and functioning (Navari-Izzo et al., 1989; Surjus and Durand, 1996). The higher molar ratio value in the PM of dehydrated and, perhaps more

2164 Quartacci et al. Table 5. Fatty acid composition (mol%) of individual and total phospholipids in plasma membranes isolated from leaves of Ramonda serbica during dehydration and rehydration Results are the means of three independent experiments. For comparisons among means an analysis of variance was used. For each treatment means in columns followed by different letters are signi cantly different at P <0.05 level. tr, trace. 14:0 16:0 16:1 18:0 18:1 18:2 18:3 Unsaturation PC Hydrated tr 23.2 a 1.4 10.4 b 16.7 a 43.2 b 5.1 a 66.4 b Dehydrated 3 43.6 b tr 20.4 c 20.9 a 8.7 a 3.4 a 33.0 a Rehydrated tr 25.8 a tr 3.7 a 12.3 a 54.9 c 3.3 a 70.5b PE Hydrated 1.3 a 42.8 a 4.6 12.6 b 5.8 a 30.1 b 2.8 a 43.3 b Dehydrated tr 47.7 b tr 25.3 c 16.1 b 10.9 a tr 27.0 a Rehydrated 0.5 a 37.2 a tr 3.4 a 6.4 a 49.6 c 2.9 a 58.9 c PG Hydrated 2.6 a 49.6 a tr 19.0 b 10.3 a 18.0 b 0.5 a 28.8 a Dehydrated 8.0 b 46.4 a tr 22.7 b 11.5 a 8.8 a 2.6 b 22.9 a Rehydrated 1.7 a 46.9 a tr 14.2 a 14.1 a 20.5 b 2.6 b 37.2 b PI Hydrated 1.2 a 44.6 a 4.9 13.1 b 6.7 a 24.5 b 5.0 a 41.1 a Dehydrated 3.9 b 42.5 a tr 15.5 b 26.5 b 7.0 a 4.6 a 38.1 a Rehydrated tr 41.1 a tr 3.9 a 11.3 a 37.3 c 6.4 a 55.0 b PS Hydrated 1.0 a 39.2 a 7.5 17.9 a 9.9 a 23.3 b 1.2 a 41.9 b Dehydrated 3.0 a 47.5 b tr 17.1 a 21.4 b 7.6 a 3.4 a 32.4 a Rehydrated tr 32.9 a tr 23.8 a 22.7 b 20.6 b tr 43.3 b PA Hydrated 5.2 a 38.4 a 1.0 4.0 a 11.1 a 31.6 b 7.8 a 51.5 b Dehydrated 7.4 a 53.6 c tr 7.2 a 22.8 b 9.0 a tr 31.8 a Rehydrated tr 46.9 b tr 3.4 a 11.7 a 32.6 b 5.4 a 49.7 b Total Hydrated 1.3 a 33.5 a 3.2 12.7 b 10.5 a 35.0 b 3.8 a 52.5 b Dehydrated 4.4 b 46.4 b tr 19.4 c 20.2 b 7.0 a 2.6 a 29.8 a Rehydrated tr 35.6 a tr 6.9 a 12.1 a 41.9 b 3.3 a 57.3 b importantly, of rehydrated leaves compared with the control may have maintained membrane functionality due to the higher effectiveness of the at con gurations of cholesterol and campesterol in stabilizing the bilayer architecture (Navari-Izzo et al., 1989; Stalleart and Geuns, 1994). Changes in the ASG to SG ratio or an increase in the proportion of ASG plus FS at the expense of SG could be important in modulating the phase behaviour of PM during dehydration, and could have a relevant impact on the physical properties of the PM (Palta et al., 1993; Zhang et al., 1997). The change in the relative proportions of ASG and SG in the PM from dehydrated leaves (Table 2) may indicate that alterations in sterol conjugation play a role in R. serbica membrane functioning as previously observed in the PM of wheat roots subjected to copper stress (Quartacci et al., 2001). Cerebrosides are characterized by an extensive hydrogen bonding ability and high gel to liquid crystalline phase transition temperatures. This lipid class is, therefore, generally considered to stabilize the PM physically and to reduce ion permeability of the cells (Uemura and Steponkus, 1994). The increase in CER level following dehydration (Table 2) probably contributed to the overall membrane response to unfavourable events during the desiccation period. Indeed, it has been suggested that CER stabilize the interaction of bulk lipids and proteins by facilitating a tighter sealing of proteins into the lipid head groups, thus regulating the retention of water (Wu et al., 1998) and intracellular solutes. If this is the case, the increase in CER in the dehydrated PM should play an important role in stabilizing membrane structural integrity when bulk lipid content is reduced and speci c interactions between lipids and membrane-intrinsic proteins are altered (Table 2). Traditionally, alterations in fatty acid unsaturation degree are related to changes in bilayer thickness and uidity, and it is known that a decrease in fatty acid unsaturation results in a decrease in membrane uidity (Wu et al., 1998; Navari-Izzo et al., 2000), even though it is unlikely that the uidity of the bulk lipid phase has any important effect on the function of membrane proteins (Lee et al., 1989). The remarkable reduction of the acyl chain unsaturation detected in dehydrated PM (Table 5) may have contributed, together with the increase in FS and CER, to render the bilayer tighter and more rigid, as con rmed indirectly by the reduced solute leakage. A lower unsaturation compared with control plants has been observed in PM isolated from roots of wheat grown in excess copper (Quartacci et al., 2001), and the reduced PM unsaturation was demonstrated to be linked to a lower permeability to glucose and a tighter molecular packing

(Berglund et al., 2000). By contrast, PC and PE unsaturation of the resurrection plants B. hygroscopica and S. stap anus increased following dehydration (Navari- Izzo et al., 1995; Quartacci et al., 1997), indicating a different defence and/or adaptation mechanism depending on the species and the dehydration severity and time course. Lyotropic phase transitions and non-bilayer lipid structures have been reported in dehydrated cell membranes and liposomes (Uemura et al., 1995). The removal of water and the subsequent tightening of the lipid bilayers is suggested to induce lipid±lipid demixing, which facilitates the lamellar-to-hexagonal (H II ) transition due to the intrinsic curvature (bending energy) in membrane monolayers caused by dehydration-induced packing stress (Uemura and Steponkus, 1994). Increased membrane stability may be achieved by altering the lipid composition so that demixing and/or phase transitions are precluded or reduced. It has been shown that ASG are much more effective than FS in increasing the propensity for dehydration-induced formation of the H II phase (Webb et al., 1995), and that PE, especially polyunsaturated species of PE, is the PM lipid most likely to form the H II structure (Uemura and Steponkus, 1994). In this study, the lack of changes in the PE level following desiccation, its low content (2.8% of total PM lipids in the dehydrated leaves) and the decrease in its unsaturation (Tables 3, 4), together with ASG reduction (Table 2), may have limited the tendency to form non-lamellar con gurations which are otherwise favoured by the altered hydration characteristics of the membranes. Indeed, high proportions of highly hydrated species (e.g. PL) and poorly hydrated species (e.g. FS and CER) increase the tendency for dehydrationinduced lipid±lipid demixing and hence the tendency to form H II con gurations which, at low water contents (<20%, w/w), minimize the lateral pressure in the bilayer and thus the bending energy (Gruner, 1989). Nevertheless, it is likely that the tendency for dehydration-induced formation of the H II phase is in uenced by the collective changes in the various lipid classes rather than in a speci c lipid class or species (Uemura et al., 1995). Among PLs, the high PA amount present in the dehydrated PM (Table 2), besides the result of PL degradation, may be considered as a storage of PL precursors which can be readily used as soon as stress conditions are released. The full regaining of photosynthetic activity a few hours after rehydration (Augusti et al., 2001) as well as the active membrane defence systems against oxidative stress (Sgherri et al., 2000; Augusti et al., 2001) indicates that Ramonda leaves completely restored their thylakoid membrane integrity and functionality. In addition, the rapid recovery upon rehydration of the lipid PM composition indicates the importance of ef cient mechanisms that are necessary for repairing membranes after rewatering. Lipid modulation, and uidity as a consequence, seems to play a major role in the adaptation to altered conditions and in regenerating the original membrane structure and functioning. Acknowledgements This study was performed by collaboration between the University of Pisa (promoter F Navari-Izzo) and the University of Belgrade (promoter B StevanovicÂ). This paper is dedicated to the memory of O GlisÏicÂ. References Plasma membrane lipids in Ramonda 2165 Augusti A, Scartazza A, Navari-Izzo F, Sgherri CLM, Stevanovic B, Brugnoli E. 2001. Photosystem II photochemical ef ciency, zeaxanthin and antioxidant contents in the poikilohydric Ramonda serbica during dehydration and rehydration. Photosynthesis Research 67, 79±88. Berglund AH, Quartacci MF, Liljenberg C. 2000. Changes in plasma membrane lipid composition: a strategy for acclimation to copper stress. Biochemical Society Transactions 28, 905±907. Bewley JD, Krochko JE. 1982. Desiccation tolerance. In: Lange OL, Nobel PS, Osmond CB, Ziegler H, eds. Encyclopedia of plant physiologyðwater relations and carbon assimilation, Vol. 12. Berlin: Springer-Verlag, 325±378. Bradford MM. 1976. A rapid and sensitive method for the quanti cation of microgram quantities of protein utilizing the principle of protein±dye binding. Analytical Biochemistry 72, 248±254. Gaff DF. 1981. The biology of resurrection plants. In: Pate JS, McComb AJ, eds. The biology of Australian plants. Perth: University of Western Australia Press, 114±146. Gruner SM. 1989. Hydrocarbon chain conformation in the H II phase. Biophysical Journal 56, 1045±1046. Grunwald C. 1974. Sterol molecular modi cations in uencing membrane permeability. Plant Physiology 54, 624±628. HernandeÂz LE, Cooke DT. 1997. Modi cation of the root plasma membrane lipid composition of cadmium-treated Pisum sativum. Journal of Experimental Botany 48, 1375±1381. Kerkeb L, Donaire JP, Venema K, Rodriguez-Rosales MP. 2001. Tolerance to NaCl induces changes in plasma membrane lipid composition, uidity and H + -ATPase activity of tomato calli. Physiologia Plantarum 113, 217±224. Lee AG, Michelangeli F, East JM. 1989. Tests for the importance of uidity for the function of membrane proteins. Biochemical Society Transaction 17, 962±964. Leprince O, Harren FJ, Buitink J, Alberda M, Hoekstra F. 2000. Metabolic dysfunction and unabated respiration precede the loss of membrane integrity during dehydration of germinationg radicles. Plant Physiology 122, 597±608. Lynch DV, Steponkus PL. 1987. Plasma membrane lipid alterations associated with cold acclimation of winter rye seedlings (Secale cereale L. cv. Puma). Plant Physiology 83, 761±767. Navari-Izzo F, Izzo R, Quartacci MF. 1988. Phospholipid and sterol alterations associated with salinity and water stress in maize roots. Plant Physiology (Life Science Advance) 7, 137±142. Navari-Izzo F, Quartacci MF, Izzo R. 1989. Lipid changes in maize seedlings in response to eld water de cits. Journal of Experimental Botany 40, 675±680. Navari-Izzo F, Quartacci MF, Mel D, Izzo R. 1993. Lipid composition of plasma membranes isolated from sun ower seedlings grown under water-stress. Physiologia Plantarum 87, 508±514.

2166 Quartacci et al. Navari-Izzo F, Quartacci MF, Pinzino C, Rascio N, Vazzana C, Sgherri CLM. 2000. Protein dynamics in thylakoids of the desiccation-tolerant plant Boea hygroscopica during dehydration and rehydration. Plant Physiology 124, 1427±1436. Navari-Izzo F, Rascio N. 1999. Plant response to water-de cit conditions. In: Pessarakli M, ed. Handbook of plant and crop stress. New York: Marcel Dekker Inc, 231±270. Navari-Izzo F, Ricci F, Vazzana C, Quartacci MF. 1995. Unusual composition of thylakoid membranes of the resurrection plant Boea hygroscopica: changes in lipids upon dehydration and rehydration. Physiologia Plantarum 94, 135± 142. Navari-Izzo F, Vangioni N, Quartacci MF. 1990. Lipids of soybean and sun ower seedlings grown under drought conditions. Phytochemistry 29, 2119±2123. Oliver MJ. 1996. Desiccation tolerance in vegetative plant cells. Physiologia Plantarum 97, 779±787. Palta JP, Whitaker BD, Weiss LS. 1993. Plasma membrane lipids associated with genetic variability in freezing tolerance and cold acclimation of Solanum species. Plant Physiology 103, 793±803. Quartacci MF, Cosi E, Navari-Izzo F. 2001. Lipids and NADPHdependent superoxide production in plasma membrane vesicles from roots of wheat grown under copper de ciency or excess. Journal of Experimental Botany 52, 77±84. Quartacci MF, Forli M, Rascio N, Dalla Vecchia F, Bochicchio A, Navari-Izzo F. 1997. Desiccation-tolerant Sporobolus sta anus: lipid composition and cellular ultrastructure during dehydration and rehydration. Journal of Experimental Botany 48, 1269±1279. Rochester CP, Kjellbom P, Larsson C. 1987. Lipid composition of plasma membranes from barley leaves and roots, spinach leaves and cauli ower in orescences. Physiologia Plantarum 71, 257±263. Schroeder F. 1984. Fluorescent sterols probe molecules of membrane structure and function. Progress in Lipid Research 23, 97±113. Sgherri C, Stevanovic B, Navari-Izzo F. 2000. Role of phenolic acids during dehydration and rehydration of Ramonda serbica. Plant Physiology and Biochemistry 38, S196. Sgherri CLM, Loggini B, Bochicchio A, Navari-Izzo F. 1994. Antioxidant system in Boea hygroscopica: changes in response to desiccation and rehydration. Phytochemistry 37, 377±381. Stalleart VM, Geuns JM. 1994. Phospholipid and free sterol composition of hypocotyl plasma membranes of ageing mung bean seedlings. Phytochemistry 36, 1177±1180. Stefanov K, Markoska YK, Kimenov GP, Popov SS. 1992. Lipid and sterol changes in leaves of Haberlea rhodopensis and Ramonda species at transition from biosis into anabiosis and vice versa caused by water stress. Phytochemistry 30, 461±466. Stevanovic B. 1986. Ecophysiological characteristics of the species Ramonda serbica PancÏ. and Ramonda nathaliae PancÏ. et Petrov. Ekologija, Acta Biologica Iugoslavica 21, 119±134. Stevanovic B, Pham Thi AT, Monteiro de Paula F, Vieira da Silva J. 1992. Effect of dehydration and rehydration on the polar lipid and fatty acid composition of Ramonda species. Canadian Journal of Botany 70, 107±113. Stevanovic B, Sinzar J, GlisÏic O. 1998. Electrolyte leakage difference between poikilohydrous and homoiohydrous species of Gesneriaceae. Biologia Plantarum 40, 199±303. Surjus A, Durand M. 1996. Lipid changes in soybean rot membranes in response to salt treatment. Journal of Experimental Botany 47, 17±23. Uemura M, Joseph RA, Steponkus PL. 1995. Cold acclimation of Arabidopsis thaliana. Effect on plasma membrane lipid composition and freeze-induced lesions. Plant Physiology 109, 15±30. Uemura M, Steponkus P. 1994. A contrast of the plasma membrane lipid composition of oat and rye leaves in relation to freezing tolerance. Plant Physiology 104, 479±496. Webb MS, Irving TC, Steponkus PL. 1995. Effect on plant sterols on the hydration and phase behaviour of DOPE/DOPC mixtures. Biochimica Biophysica Acta 1239, 226±238. Wu J, Seliskar DM, Gallagher JL. 1998. Stress tolerance in the marsh plant Spartina patens: impact of NaCl on growth and root plasma membrane lipid composition. Physiologia Plantarum 102, 307±317. Yoshida S, Uemura M. 1990. Responses of the plasma membrane to cold acclimation and freezing stress. In: Larsson C, Mùller IM, eds. The plant plasma membrane. Berlin: Springer-Verlag, 293± 319. Zhang G, Slaski JJ, Archambault J, Taylor G. 1997. Alteration of plasma membrane lipids in aluminium-resistant and aluminium-sensitive wheat genotypes in response to aluminium stress. Physiologia Plantarum 99, 302±308.