Kinetics of violaxanthin de-epoxidation by violaxanthin de-epoxidase, a xanthophyll cycle enzyme, is regulated by membrane fluidity in model lipid bilayers Dariusz Latowski 1 , Jerzy Kruk 1 , Kvetoslava Burda 2 , Marta Skrzynecka-Jaskier 1 , Anna Kostecka-Gugała 1 and Kazimierz Strzałka 1 1 Department of Plant Physiology and Biochemistry, The Jan Zurzycki Institute of Molecular Biology and Biotechnology, Jagiellonian University, Krako ´ w, Poland; 2 H. Niewodniczanski Institute of Nuclear Physics, Krako ´ w, Poland This paper describes violaxanthin de-epoxidation in model lipid bilayers. Unilamellar egg yolk phosphatidylcholine (PtdCho) vesicles supplemented with monogalactosyldi- acylglycerol were found to be a suitable system for studying this reaction. Such a system resembles more the native thylakoid membrane and offers better possibilities for studying kinetics and factors controlling de-epoxidation of violaxanthin than a system composed only of monogalacto- syldiacylglycerol and is commonly used in xanthophyll cycle studies. The activity of violaxanthin de-epoxidase (VDE) strongly depended on the ratio of monogalactosyldiacyl- glycerol to PtdCho in liposomes. The mathematical model of violaxanthin de-epoxidation was applied to calculate the probability of violaxanthin to zeaxanthin conversion at different phases of de-epoxidation reactions. Measurements of deepoxidation rate and EPR-spin label study at different temperatures revealed that dynamic properties of the membrane are important factors that might control con- version of violaxanthin to antheraxanthin. A model of the molecular mechanism of violaxanthin de-epoxidation where the reversed hexagonal structures (mainly created by monogalactosyldiacylglycerol) are assumed to be required for violaxanthin conversion to zeaxanthin is proposed. The presence of monogalactosyldiacylglycerol reversed hexa- gonal phase was detected in the PtdCho/monogalactosyl- diacylglycerol liposomes membrane by 31 P-NMR studies. The availability of violaxanthin for de-epoxidation is a dif- fusion-dependent process controlled by membrane fluidity. The significance of the presented results for understanding the mechanism of violaxanthin de-epoxidation in native thylakoid membranes is discussed. Keywords: xanthophyll cycle; de-epoxidation; liposomes; violaxanthin; zeaxanthin Xanthophyll cycle is a photoprotective mechanism wide- spread in nature operating in the thylakoid membranes of all higher plants, ferns, mosses and several algal groups [1]. This cycle involves two reversible reactions, light-dependent de-epoxidation of violaxanthin to zeaxanthin via anther- axanthin as an intermediate and light-independent epoxi- dation of zeaxanthin to anteraxanthin and violaxanthin [2]. The conversion of violaxanthin to zeaxanthin is catalysed by violaxanthin de-epoxidase (VDE) and the reverse reaction of violaxanthin formation from zeaxanthin is catalysed by another enzyme, zeaxanthin epoxidase. VDE has been isolated from spinach and lettuce chloroplasts and the molecularmassofthenativeenzymewasestimatedas 43 kDa [3–5]. The gene encoding VDE has been already isolated and cloned [6]. VDE is located on the lumenal side of the thylakoid membrane, shows an optimum activity at pH 4.8 when present in chloroplasts and at 5.2 for the isolated enzyme [7] and requires ascorbate as a reductant [8]. In the dark, when the pH in thylakoid lumen is neutral or alkaline, VDE is inactive, whereas under strong light conditions, pH in the thylakoid lumen decreases, the enzyme binds to the membrane, becomes active and converts violaxanthin to zeaxanthin [8,9]. The inhibition of the enzyme activity by zeaxanthin has been reported [3]. For optimal activity, VDE requires the presence of monogal- actosyldiacylglycerol, the major lipid of the thylakoid membrane [10–12]. With its small head-group area and critical packing parameter value superior to one, monogal- actosyldiacylglycerol in water forms reversed hexagonal phase instead of bilayer structures [13]. It is known that monogalactosyldiacylglycerol forms hexagonal phases over a wide temperature range of )15 °Cto80°C at concentra- tions higher than 50% lipid in water and this process also depends on the degree of unsaturation of the acyl chains. Until now, all in vitro studies on the VDE activity have been carried out using largely undefined systems of buffered suspension of monogalactosyldiacylglycerol aggregates containing violaxanthin as substrate. Here, we present a Correspondence to K. Strzalka, Department of Plant Physiology and Biochemistry, The Jan Zurzycki Institute of Molecular Biology and Biotechnology, Jagiellonian University, ul. Gronostajowa 7, 30-387 Krako ´ w, Poland. Fax: + 48 12 252 69 02, Tel.: + 48 12 252 65 09, E-mail: strzalka@awe.mol.uj.edu.pl Abbreviations: LHC, light harvesting complex; PSI, photosystem I; PSII, photosystem II; VDE, violaxanthin de-epoxidase; PtdCho, phosphatidylcholine; PtdGro, phosphatidylglycerol; VA, probability of violaxanthin to antheraxanthin conversion; AZ, probability of antheraxanthin to zeaxanthin conversion; VV, probability that violaxanthin remains violaxanthin; AA, probability that anthera- xanthin remains antheraxanthin; ZZ, probability that zeaxanthin remains zeaxanthin; S VA , the constant rate of VA 0 decrease; S AZ , the constant rate of AZ 0 decrease. (Received 12 April 2002, revised 18 July 2002, accepted 5 August 2002) Eur. J. Biochem. 269, 4656–4665 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03166.x new approach in the study of VDE activity employing violaxanthin-containing liposomes as an experimental sys- tem, which is a closer to the native thylakoid membrane. The use of lipid bilayers instead of monogalactosyldiacyl- glycerol aggregates offers new possibilities in the investiga- tion of the kinetic parameters and mechanism of violaxanthin de-epoxidation. One of the advantages of such system is the defined orientation of violaxanthin molecules in the lipid bilayer, which, according to various sources [14– 16], is perpendicular to the plane of the membrane. Violaxanthin-supplemented unilamellar liposomes with VDE present only outside the vesicles are also a good system to study the flip-flop rate of antheraxanthin which is probably a necessary step preceding zeaxanthin formation in membrane. Additionally, experiments carried out at different temperatures and application of a mathematical model of de-epoxidation for the analysis of the obtained results provide important information on the influence of membrane physical properties, kinetic parameters of vio- laxanthin into zeaxanthin conversion and flip-flop rate of antheraxanthin. A possible molecular mechanism of vio- laxanthin de-epoxidation is proposed. MATERIALS AND METHODS Preparation of unilamellar liposomes The mixture of lipids with violaxanthin in chloroform was evaporated under stream of nitrogen to form a thin film and dried under vacuum for 1 h. The dried lipids were dissolved in ethanol and the solution was injected slowly with a Hamilton syringe into 0.1 M sodium citrate buffer, pH 5.1, under continuous bubbling with nitrogen. The final ethanol concentration did not exceed 1.25%. Subsequently, the liposome suspension was extruded through a polycarbonate membrane with a pore diameter of 100 nm [17]. The final lipid concentration in a liposome suspension was 43 l M and violaxanthin concentration was 0.33 l M . Egg yolk phosphatidylcholine (PtdCho) was purchased from Sigma (P2772) and plant monogalactosyldiacylgly- cerol was obtained from Lipid Products. Electron microscopy One drop of PtdCho/monogalactosyldiacylglycerol lipo- somes (350 l M lipid concentration) or monogalactosyldi- acylglycerol structures (12.9 l M lipid concentration) in citrate buffer (pH 5.1) was placed on a Formvar coated grid and after 30 s one drop of staining solution was added. Negative staining was performed with uranyl acetate at room temperature [18]. After 30 s, excess solution was drained off with filter paper and the grid was allowed to dry in the air. The grids were examined in a JEM 100SX electron microscope operated at 80 kV. Photon correlation spectroscopy (PCS) analysis Diameter of PtdCho/monogalactosyldiacylglycerol lipo- somes and monogalactosyldiacylglycerol reversed hexa- gonal phase was measured by PCS analysis. The 10 mW He-Ne laser (633 nm) was used as a light source. The selected angle was 90°, the viscosity was 0.890 centipoise and refractive index 1.333. All analyses were performed at 25 °C and at the equilibration time of 2 min. Total lipid concen- tration in the case of PtdCho/monogalactosyldiacylgly- cerol liposomes was 43 l M (30.1 l M PtdCho, 12.9 l M monogalactosyldiacylglycerol) and 12.9 l M for monogal- actosyldiacylglycerol structures. Both liposomes and monogalactosyldiacylglycerol reversed hexagonal phase were suspended in 0.1 M sodium citrate buffer (pH 5.1). Isolation of violaxanthin Violaxanthin was isolated from dark-stored leaves of lucerne (Medicago sativa) by pigment extraction with acetone, saponification of the lipid extract [19], followed by column chromatography on Silica Gel F254 (Merck) in petroleum ether : acetone (4 : 1, v/v). Isolation and purification of VDE VDE was isolated and purified from 7-day-old wheat leaves grown at 28 °C according to the method described by Hager and Holocher [9]. Additionally, the enzyme was purified by gel filtration on Sephadex G100. The gel electrophoretic and ion-exchange chromatography analysis of VDE preparation showed two other minor proteins apart from VDE (data not shown). The enzyme activity was determined by dual- wavelength measurements (502–540 nm) using DW-2000 SLM Aminco spectrophotometer at 25 °C according to Yamamoto [20]. The reaction mixture contained 0.33 l M violaxanthin, 12.9 l M monogalactosyldiacylglycerol and 30 m M sodium ascorbate in 0.1 M sodium citrate buffer (pH 5.1). Measurement of violaxanthin de-epoxidation De-epoxidation of violaxanthin was measured at 4, 12 and 25 °C both in a monogalactosyldiacylglycerol reversed hexagonal phase and in liposomes. The composition of the reaction mixture of the monogalactosyldiacylglycerol systemwasthesameasthatusedfortheenzymeactivity determination. The liposomes (30.1 l M PtdCho, 12.9 l M monogalactosyldiacylglycerol, 0.33 l M violaxanthin) were prepared in 30 m M sodium ascorbate, 0.1 M sodium citrate buffer (pH 5.1). In another series of experiments, liposomes with constant concentration of monogalactosyldiacylgly- cerol (12.9 l M ) and violaxanthin (0.33 l M )wereusedand PtdCho was changed in order to obtain following mono- galactosyldiacylglycerol proportions: 5 mol%, 15 mol% and 30 mol%. All mixtures were placed in darkness and gently stirred. The de-epoxidation reaction was initiated by addition of saturating amount of VDE, the activity of which correspon- ded to 4 nmol de-epoxidated violaxanthin per min per mL. The reaction was terminated and pigments were extracted by mixing 750 lL of the reaction medium with 750 lLof the extraction solution containing chloroform/methanol/ ammonia (1 : 2 : 0.004, v/v/v). Xanthophyll pigments were extracted by vigorous shaking and centrifugation for 10 min at 10 000 g in Micro-Centrifuge Type-320. After centrifu- gation, the chloroform fraction (200 lL) was evaporated to dryness under stream of nitrogen. Subsequently, pigments were dissolved in 50 lL tetrahydrofuran and 550 lLofthe following solvent mixture, acetonitrile/methanol/water (360 : 40 : 40, v/v/v). Ó FEBS 2002 Violaxanthin de-epoxidation in liposomes (Eur. J. Biochem. 269) 4657 Pigment separation was performed by reverse phase HPLC using a RP-18 column, 5 lm particle size, according to the modified method of Gilmore and Yamamoto [21] at the flow rate of 3 mLÆmin )1 . The eluted pigments were monitored at 440 nm and quantitatively determined. Analysis of de-epoxidation kinetics We have applied a new mathematical model [22] to analyse the kinetics of conversion of violaxanthin to zeaxanthin. The model allowed us to follow independently the kinetics of the two de-epoxidation steps: the conver- sion of violaxanthin into antheraxanthin (VfiA) with a probability VA and antheraxanthin into zeaxanthin (AfiZ) with a probability AZ. It is known from experimental data that these two steps reach equilibrium. It means that the parameters VA and AZ must vanish. In themodelwehaveassumedalineardecreaseofthe conversion probabilities: VA ¼ VA 0 À nðS VA ÁDtÞ for VA > 0 AZ ¼ AZ 0 À nðS AZ ÁDtÞ for AZ > 0 where VA 0 and AZ 0 are the initial values of the conversion probabilities, S VA and S AZ are the constant rates and nÆDt is thetimeofreaction(Dt ¼ a constant time interval, n ¼ number of time intervals). Measurement of the order parameter in liposome membrane Temperature dependent changes of the order parameter of lipid fatty acyl chains in liposome membranes were recorded by EPR spin label measurements using a spin label 5-doxyl- stearic acid reporting on dynamics of membrane regions close to the headgroup area. The spin label was added to the chloroform mixture of monogalactosyldiacylglycerol, Ptd- Cho and V, dried under stream of nitrogen and stored under vacuum for 1 h. After this time, the dried mixture was suspendedin0.1 M sodium citrate buffer pH 5.2 by vortexing. The final concentration of 5-doxyl-stearic acid was 10 )4 M . The final concentration of lipids was 10 )2 M and their proportions were the same as in the section on ÔPreparation of unilamellar liposomesÕ. EPR spectra of the spin label as a function of temperature were recorded using a Bruker ESP-300E spectrometer fitted with TM 110 cavity. The modulation amplitude was 1 G, microwave power was 2 or 8 mW. The measurements were performed within the temperature range of 0–40 °C. All measurements were performed in a heating mode. Temperature was stabilized using Brucker temperature controller. Spin label was purchased from Sigma. 31 P-NMR studies 31 P-NMR spectra of liposomes suspended in the citrate buffer pH (5.1), containing 10% D 2 O were recorded at 202.5 MHz using a Bruker AMX-500. Generally, a sweep width of 41.7 kHz and a repetition 2.6 s using 30° radio frequency pulses were used. The exponential multiplication of the free induction decay resulted in a 100-Hz line broadening. The number of scans was 28 000. All spectra were recorded at 17 °C. RESULTS Effect of monogalactosyldiacylglycerol proportion in liposomes on violaxanthin deepoxidation Our initial attempts to use liposomes with a lipid compo- sition similar to that of the thylakoid membrane were unsuccessful because the chemical instability of violaxanthin related to the presence of phosphatidylglycerol (PtdGro) and sulphoquinovosyldiacyloglycerol, which complicates the quantitative measurements [17]. Therefore, we applied PtdCho, as the lipid that most readily forms bilayers [13], supplemented with monogalactosyldiacylglycerol which was found necessary for VDE activity. With the rise in monogalactosyldiacylglycerol proportion in PtdCho lipo- somes to a certain level, the percentage of transformed violaxanthin also increased (Fig. 1). However, at 35 mol% of monogalactosyldiacylglycerol the de-epoxidation rate became significantly lower due to liposome aggregation and the suspension became turbid. The increase in turbidity was followed by sedimentation of the lipid aggregates formed. These changes were caused probably by fusion of liposomes or the appearance of monogalactosyldiacylglycerol aggre- gates at its high proportion to PtdCho in the lipid mixture [13,23]. The liposome suspension with monogalactosyldi- acylglycerol content < 30 mol% was transparent and showed no tendency to aggregate. The presence of liposomes and absence of aggregates in such a suspension was confirmed by electron microscopy and PCS (data not shown). On the other hand, 31 P-NMR measurements Fig. 1. The effect of monogalactosyldiacylglycerol proportion in Ptd- Cho/monogalactosyldiacylglycerol liposomes on the level of xanthophylls after 20 min of the violaxanthin de-epoxidation reaction at room tem- perature. 4658 D. Latowski et al. (Eur. J. Biochem. 269) Ó FEBS 2002 revealed formation of the reversed hexagonal phase domains existing in PtdCho/monogalactosyldiacylglycerol liposomes (Fig. 7). Violaxanthin de-epoxidation was found to be strongly dependent not only on the concentration of monogalacto- syldiacylglycerol but also on the ratio of monogalactosyl- diacylglycerol to PtdCho in liposomes, even if the absolute amount of monogalactosyldiacylglycerol in the reaction mixture and its proportion to violaxanthin and VDE were constant (Fig. 2). The values of transition probabilities of the violaxanthin conversion into antheraxanthin (VA)and antheraxanthin conversion into zeaxanthin (AZ) show that the varying amounts of PtdCho, which result in changes in monogalactosyldiacylglycerol/PtdCho ratio, have much stronger effect on conversion of violaxanthin to anther- axanthin than on conversion of antheraxanthin to zeaxan- thin (Table 1). At low monogalactosyldiacylglycerol concentration (5 mol%), probability of violaxanthin to antheraxanthin conversion is very low (VA 0 is 0.006 only). However, once antheraxanthin has been formed, its con- version to zeaxanthin occurs at relatively fast rate (AZ 0 ¼ 0.548). VA 0 and S VA parameters are very sensitive to an increase in relative proportion of monogalactosyldi- acylglycerol; at 30 mol% of this lipid, their values increase 43 and 76 times, respectively, while values of corresponding parameters describing kinetics of antheraxathin to zeaxan- thin conversion (AZ 0 and S AZ parameters) increase only 1.5 and 2.2 times, respectively. For further studies on violaxanthin de-epoxidation, PtdCho liposomes with 30 mol% content of monogalacto- syldiacylglycerol were used as an optimal system. Comparison of violaxanthin de-epoxidation in monogalactosyldiacylglycerol and liposomal systems and the effect of temperature The temperature dependence of de-epoxidation reaction was measured in unilamellar PtdCho/monogalactosyldi- acylglycerol liposomes and in the monogalactosyldiacylglyc- erol reversed hexagonal phase system with the composition given in Materials and methods. The concentrations of monogalactosyldiacylglycerol, violaxanthin and VDE (saturating amount) were the same both in PtdCho/ monogalactosyldiacylglycerol liposomes and monogalacto- syldiacylglycerol systems. It was found that kinetics of de-epoxidation reaction were different in liposomes and in the monogalactosyldiacylglycerol system, and that tempera- ture has a strong influence on the reaction rate in both systems studied (Figs 3 and 4). At the three temperatures studied (4, 12 and 25 °C), the initial rate of violaxanthin de-epoxidation was always faster in liposomes than in monogalactosyldiacylglycerol system; this difference was most evident at 25 °C. However, changes Monogalctosyldiacylglycerol 5 mol% Monogalctosyldiacylglycerol 15 mol% Monogalctosyldiacylglycerol 30 mol% Fig. 2. Time course of violaxanthin to zeaxanthin conversion in PtdCho/ monogalactosyldiacylglycerol liposomes at 25 °C when PtdCho and monogalactosyldiacylglycerol concentrations were, respectively: 245.1 l M and 12.9 l M (5 mol% of monogalactosyldiacylglycerol); 73.1 l M and 12.9 l M (15 mol% of monogalactosyldiacylglycerol); 30.1 l M and 12.9 l M (30 mol% of monogalactosyldiacylglycerol) and violaxanthin concentration was 0.33 l M . Table 1. Kinetic parameters of a de-epoxidation reaction calculated for the experimental data presented in Fig. 2 by means of the mathematical model. Monogalactosyldiacylglycerol (mol %) VA 0 S VA · 10 )3 (min )1 ) AZ 0 S AZ · 10 )3 (min )1 ) 5 0.006 0.26 0.548 35.50 15 0.045 5.52 0.600 50.29 30 0.260 19.75 0.810 76.80 Ó FEBS 2002 Violaxanthin de-epoxidation in liposomes (Eur. J. Biochem. 269) 4659 in the temperature had different effects on the reaction plateau reached in both systems. At 25 °C, plateau was achieved in 20 min in PtdCho/monogalactosyldiacylgly- cerol liposomes and in 40 min in monogalactosyldiacylgly- cerol system. At this stage, about 93% and 83% of initial violaxanthin amount were de-epoxidated in liposomes and monogalactosyldiacylglycerol system, respectively; corres- ponding zeaxanthin levels amounted about to 86% and 80% of total xanthophyll pigments. The plateau levels of anteraxanthin were about 3.8% in liposomes and 2.5% in monogalactosyldiacylglycerol reversed hexagonal phase. In our experimental systems the complete de-epoxidation of violaxanthin was not observed. A possible reason for this may be an inhibitory effect of accumulating zeaxanthin on VDE activity as reported previously [3]. On the other hand, it may be also connected with the presence in the violax- anthin pool of small amount of cis isomers that cannot serve as a substrate for VDE [24]. At 25 °C, the rate of antheraxanthin formation was considerably faster in liposomes than in the monogalacto- syldiacylglycerol system. Its maximum level in liposomes was achieved after 2 min and amounted to about 23% of the total xanthophyll pool, whereas in monogalactosyldi- acylglycerol system the maximum level of antheraxanthin was detected after 10 min and it accounted only for about 16% of all xanthophylls. At 12 °C, in spite of the higher initial de-epoxidation rate of violaxanthin in liposomes, more violaxanthin was de- epoxidated and more zeaxanthin was formed in the mono- galactosyldiacylglycerol system at the plateau stage of the reaction. The same was found at 4 °C, although at this temperature the plateau in monogalactosyldiacylglycerol system had not been reached during 180 min reaction time. The kinetic parameters of violaxanthin de-epoxidation calculated for the exeperimental data obtained from the liposome and monogalactosyldiacylglycerol systems by Fig. 4. Time course of violaxanthin to zeaxanthin conversion at different temperatures in monogalactosyldiacylglycerol systems. Fig. 3. Time course of violaxanthin to zeaxanthin conversion at different temperatures in liposomes. Monogalactosyldiacylglycerol was present at (A) 5mol%, (B) 15mol% and (C) 30mol%. 4660 D. Latowski et al. (Eur. J. Biochem. 269) Ó FEBS 2002 means of the mathematical model are compared in Table 2. The rates of the de-epoxidation reactions are more sensitive to temperature in the liposomal system than in the monogalactosyldiacylglycerol system. When rising the tem- perature from 4 to 25 °C, zeaxanthin level at the plateau stage increases 2.7-fold, violaxanthin level decreases eight- fold and antheraxanthin maximal level increases twofold in the liposomal system, whereas corresponding values for monogalactosyldiacylglycerol system are 2.1, 3.5 and 1.7. When analysing the effect of temperature on probabilities of VA 0 and AZ 0 transition in both systems studied (Table 2), it is evident that these values are higher in the liposome system than in the monogalactosyldiacylglycerol system at all temperatures studied. The increase in the temperature from 4to25°CincreasestheVA 0 value from 0.005 to 0.26 (52-fold) in liposomes, whereas VA 0 increases from 0.004 to 0.085 (21-fold) in monogalactosyldiacylglycerol system. The AZ 0 transition increases proportionally in both systems on elevating the temperature from 4 to 25 °C and its value rises 12.5-fold in liposomal and 13.3-fold in micellar systems, respectively. On the other hand, the values S VA and S AZ coefficients increase much more in monogalactosyldiacyl- glycerol reversed hexagonal phase than in the liposomal system when inceasing the temperature from 4 to 25 °C. Temperature-dependent changes in the value of VA 0 parameter in PtdCho/monogalactosyldiacylglycerol lipo- somes correlate well with the corresponding changes in the value of the order parameter as found by the use of EPR spectrometry and a spin probe 5-doxyl-stearic acid (Fig. 5). The lower value of the order parameter the higher the violaxanthin de-epoxidation rate is observed. DISCUSSION This paper is the first work where VDE has been isolated from a monocotyledonous plant. The action and properties of this enzyme are the same as VDE isolated previously from dicotyledonous plants [25]. The presented results show that VDE-mediated conver- sion of violaxanthin via antheraxanthin into zeaxanthin can occur in PtdCho/monogalactosyldiacylglycerol liposomes. It is worth noting that the VDE enzyme added to initiate the reaction was present on the external and not internal side of the liposome membrane and that similar kinetics and decline in violaxanthin amount as in the measurements performed with thylakoids [15] were observed. For this reason, the PtdCho/monogalactosyldiacylglycerol unilamel- lar liposome system used in this work is a good model of the native photosynthetic membrane for studying the VDE activity. The presence of monogalactosyldiacylglycerol in Ptd- Cho-liposomes was found to be indispensable for the violaxanthin de-epoxidation reaction. As we have demon- strated, the rate of violaxanthin to antheraxanthin conver- sion depends on monogalactosyldiacylglycerol/PtdCho ratio in the liposome membrane even if the absolute amount of monogalactosyldiacylglycerol in the reaction mixture and its proportion to violaxanthin and VDE remains constant (Fig. 2, Table 1). On the basis of these results, we postulate that VDE binds only to certain membrane domains that are rich in monogalactosyldiacylglycerol and the de-epoxida- tion reactions take place in these domains. Violaxanthin being distributed homogeneously in the lipid bilayer has to Table 2. Kinetic parameters of a de-epoxidation reaction calculated for the experimental data presented in Figs 3 and 4 by means of the mathematical model. Temp. (°C) VA 0 S VA · 10 )3 (min )1 ) AZ 0 · 10 )3 (min )1 )S AZ Liposomes 4 0.005 0.024 0.065 0.371 12 0.026 0.34 0.150 1.808 25 0.260 19.76 0.810 50.29 Monogalactosyldiacylglycerol system 4 0.004 0.012 0.030 0.0004 12 0.020 0.133 0.100 0.681 25 0.085 21.76 0.400 8.51 Fig. 5. PtdCho/monogalactosyldiacylglycerol liposome membrane flui- dity and percent of violaxanthin converted after 1 min de-epoxidation reaction at different temperatures. Ó FEBS 2002 Violaxanthin de-epoxidation in liposomes (Eur. J. Biochem. 269) 4661 enter the monogalactosyldiacylglycerol-enriched domains by lateral diffusion to be converted to antheraxanthin. The higher monogalactosyldiacylglycerol/PtdCho ratio, the higher the amount of such domains in the liposomal membrane. This shortens the diffusion path of violaxanthin molecules to these domains and results in higher rate of violaxanthin de-epoxidation (see the values of VA 0 in Table 1). It is well known that nonbilayer prone lipids (e.g. monogalactosyldiacylglycerol) may form reversed hexa- gonal phase in model lipid membranes and it has been reported that such structures exist in biological membranes [26–28]. 31 P-NMR spectra shown in Fig. 7 clearly demon- strate the existence of the reversed hexagonal phase domains in our system of PtdCho/monogalactosyldiacylglycerol liposomes. The presence of the reversed hexagonal phase in thylakoid membranes has been also reported in our earlier papers and by other authors using 31 P-NMR and freeze-fracturing techniques [23,29–31]. These observa- tions give a sound basis for our model of violaxanthin de-epoxidation in liposomes and thylakoid membranes. According to the presented model, the second reaction of de-epoxidation, i.e. conversion of antheraxanthin to zea- xanthin, also occurs in the monogalactosyldiacylglycerol rich domains, and is greatly facilitated because antheraxan- thin, formed in such domains, has an immediate access to the VDE enzyme. Therefore, and in contrast to the de-epoxidation of violaxanthin to antheraxanthin, the conversion of antheraxanthin to zeaxanthin seems to be not limited by diffusion process. The conclusion that the conversion of violaxanthin to antheraxanthin is more sensitive to monogalactosyldiacylglycerol concentration than the conversion of antheraxanthin to zeaxanthin is supported by relatively high value of the AZ 0 parameter even in conditions of very low VA 0 (e.g. at 5 mol% of monogalactosyldiacylglycerol, Table 1). The model assuming the existence of monogalactosyldi- acylglycerol reversed hexagonal phase domains in the membrane to which VDE binds and rather homogeneous distribution of violaxanthin molecules explains also the strong correlation between the rate of violaxanthin de-epoxidation and value of the membrane lipid order parameter. It seems that decreasing value of the order parameter permits faster lateral diffusion of violaxanthin in the membrane and the molecules of this xanthophyll may reach sooner the monogalactosyldiacylglycerol rich domains where they are de-epoxidated (Fig. 5, Tables 2 and 3). This model can also explain a clear temperature effect on the level and the time of antheraxanthin appear- ance in the liposome system. The conversion of anther- axanthin to zeaxanthin is less dependent on the changes in membrane physical properties (Tables 2 and 3) for the reasons already discussed. Application of the proposed model to the results obtained shows why the conversion of violaxanthin to antheraxanthin is much slower and more sensitive to temperature than transition from the anther- axanthin to zeaxanthin. On the basis of our results and literature data [15,22], we postulate that changes in mem- brane fluidity may play an important role in regulation of the violaxanthin de-epoxidation rate in membranes. The higher rate of violaxanthin de-epoxidation to antheraxanthin and stronger temperature effect on this process in liposomes than in monogalactosyldiacylglycerol systems is probably related to the different availability of violaxanthin for VDE in both systems studied. As revealed by PCS and electron microscopy, the size of monogalacto- syldiacylglycerol aggregates differed greatly from that of PtdCho/monogalactosyldiacylglycerol liposomes. The monogalactosyldiacylglycerol structures were found as large, heterogeneous aggregates with a mean diameter of % 600 nm and a large standard deviation. Thus, the previous assumption that monogalactosyldiacylglycerol creates small micelles with only one molecule of violaxan- thin inside [32] was not confirmed in our study. The average diameter of liposomes was % 110 nm (as expected) with narrow standard deviation. Apparently, the availability of violaxanthin for VDE is higher in liposomes than in the monogalactosyldiacylglycerol system where access of the enzyme to its substrate may be impeded by large scale aggregation of monogalactosyldiacylglycerol structures. Neither the molecular arrangement of monogalactosyldi- acylglycerol in such aggregates nor the orientation of violaxanthin in these structures have been precisely deter- mined [23]. To convert violaxanthin into zeaxanthin, VDE has to remove two epoxy groups attached to two rings of the violaxanthin molecule. In the unilamellar liposome system, where all the xanthophylls are oriented perpendicularly to the plane of the membrane and VDE is present only outside the vesicles, the formed antheraxanthin molecule, to be converted into zeaxanthin, has to reverse its orientation as a whole in such a way that the end group containing the ring with the remaining epoxy group appears on the other side of membrane. Such a Ôflip-flopÕ of antheraxanthin molecule is a necessary step assuming that VDE cannot penetrate through the lipid bilayer and has access to the outer surface of the liposome only. Table 3 shows the maximal time in which all molecules of violaxanthin are converted to antheraxanthin and all molecules of antheraxanthin are converted to zeaxanthin at a given temperature at saturating amount of VDE and at the initial reaction rate. While the time for VfiA conversion is shortened considerably on the increase of temperature from 4 to 25 °C,thetimeforAfiZ transition changes is in a much more narrow range. In the liposome system studied, the time for AfiZ transition takes 3.1 min at the temperature of 25 °C and 2.9 and 2.0 min at 12 and 4 °C, respectively. This means that flip-flop of antheraxanthin in PtdCho/monogalactosyldiacylglycerol liposomes at a given temperature has to be shorter than the times specified in Table 3. Moreover, faster conversion of antheraxanthin to zeaxanthin (AZ 0 values) than violax- anthin to antheraxanthin (VA 0 values) was observed at all temperatures studied (Tables 2 and 3) suggesting that flip- flop of antheraxanthin is not the limiting step in the transformation of violaxanthin into zeaxanthin in the Table 3. The maximal time required for the conversion of all violaxan- thin molecules into antheraxanthin (T VA ) and all molecules of anther- axanthin into zeaxanthin (T AZ ) at three different temperatures in PtdCho/monogalactosyldiacylglycerol liposomes. Temp. (°C) T VA (min) T AZ (min) 4 217.4 2 12 30 2.9 25 5.6 3.1 4662 D. Latowski et al. (Eur. J. Biochem. 269) Ó FEBS 2002 membrane system investigated. Our results are in agreement with those of Arvidsson et al. [15] who suggested that in the isolated thylakoids the flip-flop of antheraxanthin is not the limiting factor in zeaxanthin formation. The apparently longer time necessary for antheraxanthin to zeaxanthin conversion at higher temperatures (Table 3) can be explained in terms of our model assuming diffusion controlled rate of violaxanthin to antheraxanthin conver- sion. Violaxanthin and antheraxanthin compete for the same active site of the VDE enzyme. At elevated tempera- ture, when violaxanthin lateral diffusion in the membrane is faster, more violaxanthin molecules reach the monogalacto- syldiacylglycerol rich domains in time unit. In such a situation, violaxanthin competes successfully with anther- axanthin for the VDE active site; this results in enlargement of the antheraxanthin pool. As a consequence, a relatively lower number of antheraxanthin molecules of the pool can reach the VDE active site and become converted to zeaxanthin. This conclusion is supported by the data presented in Fig. 6, which shows that at higher temperatures a lower proportion of total antheraxanthin pool is conver- ted into zeaxanthin. It should be also added that because VDE was present in excess in the reaction mixture, only part of it could be bound to the membrane, depending on the size and number of monogalactosyldiacylglycerol-rich domains. Some conclusions drawn from our results obtained with model lipid bilayer can be extrapolated to describe the role of the xanthophyll cycle in the regulation of thylakoid membrane fluidity. In the darkness, due to zeaxanthin epoxidase activity, violaxanthin accumulates in thylakoids. Illumination of plants with strong light causes acidification of thylakoid lumen, which is a prerequisite for VDE binding to thylakoid membrane, and also it usually increases leaf temperature, which results in the increase of the membrane dynamics. A temperature-induced increase of thylakoid membranes dynamics facilitates diffusion of violaxanthin molecules into monogalactosyldiacylglycerol-VDE domains where it is converted into antheraxanthin and zeaxanthin. This conclusion is supported by the results of Sarry et al. [33] who found that illumination of plants at low tempera- ture results in a lower amount of zeaxanthin formed than at higher temperature. There are reports that show that zeaxanthin may act like cholesterol and play important role in the regulation of thylakoid membrane arrangement. Gruszecki and Strzalka [34] showed that light induced accumulation of zeaxanthin affects membrane fluidity. Tardy and Havaux [35] found that decreased value of the thylakoid membrane order parameter was proportional to the amount of zeaxanthin present in the membrane. The rigidifying effect of this xanthophyll was also found upon incorporation of exogenous zeaxanthin into isolated thyl- akoid membranes [36]. Fig. 6. The percentage of total antheraxanthin pool converted into zeaxanthin during violaxanthin de-epoxidation reaction in PtdCho/ monogalactosyldiacylglycerol liposomes at three different temperatures. Fig. 7. 31P-NMR spectra of PtdCho liposomes (A) without mono- galactosyldiacylglycerol and (B) with 30 mol% monogalactosyldiacyl- glycerol. Ó FEBS 2002 Violaxanthin de-epoxidation in liposomes (Eur. J. Biochem. 269) 4663 Zeaxanthin formed in the hexagonal phase domains can probably leave these regions and, due to its membrane rigidifying properties, it regulates the molecular dynamics of thylakoid membranes and protects them at elevated temperatures resulting from intense irradiation. In conclusion, the PtdCho/monogalactosyldiacylglycerol liposome system described in this work is more appropriate than monogalactosyldiacylglycerol aggregates for studying the mechanism of violaxanthin de-epoxidation catalysed by VDE in vitro because it approaches the native photosyn- thetic membranes. The existence of de-epoxidation reactions in liposomes opens new possibilities in the investigation of the xanthophyll cycle, which might contribute to a better understanding of this process. ACKNOWLEDGEMENTS This work was supported by a grant no. 6P04A 02819 from Committee for Scientific Research (KBN) of Poland. We wish to thank Maria Kozlowska for electron microscopy pictures, Dr Maria Zembala for PCS measurements and Dr F. Szneler for 31 P- NMR analysis. We are very grateful to Dr Fabrice Franck from University of Liege, Belgium for helpful discussion. REFERENCES 1. Stransky, H. & Hager, A. (1970) The carotenoid pattern and the occurence of the light induced xanthophyll cycle in various classes of algae part 6 chemo systematic study. Arch. Microbiol. 73,315– 323. 2. Yamamoto, H.Y., Nakayama, T.O.H. & Chichester, C.O. (1962) Studies on the light and dark interconversions of leaf xantho- phylls. Arch. Biochem. Biophys. 97, 168–173. 3. Havir, E.A., Tausta, L.S. & Peterson, R.B. (1997) Purification and properties of violaxanthin de-epoxidase from spinach. Plant Sci. 123, 57–66. 4. Rockholm, D.C. & Yamamoto, H.Y. (1996) Purification of a 43-kilodalton lumen protein from lettuce by lipid-affinity preci- pitation with monogalactosyldiacylglyceride. Plant Physiol. 110, 697–703. 5. A ˚ kerlund, H E., Arvidson, P O., Bratt, C.E. & Carlsson, M. (1995) Partial purification of the violaxanthin de-epoxidase. In Photosynthesis from Light to Biosphere (P.Mathis,ed.),Vol.IV, pp. 103–106. Kluwer Academic Publishers, Dortrecht. 6. Bugos, R.C. & Yamamoto, H.Y. (1996) Molecular cloning of violaxanthin de-epoxidase from romaine lettuce and expression in Escherichia coli. Proc. Natl Acad. Sci. USA 93, 6320–6325. 7. Hager. A. (1969) Lichtbedingte pH-Erniedringung in einem Chloroplasten-Kompartiment als Ursache der Enzymatichen Violaxanthin-zu Zeaxanthin – Umwandlung Beziehungen zur Photophosphorylierung. Planta 89, 224–243. 8. Bratt, C.E., Arvidsson, P O., Carlsson, M. & A ˚ kerlund, H E. (1995) Regulation of violaxanthin de-epoxidase activity by pH and ascorbate concentration. Photosynth. Res. 45, 169–175. 9. Hager, A. & Holocher, K. (1994) Localization of the xanthophyll cycle enzyme violaxanthin de-epoxidase within the thylakoid lumen and abolition of its mobility by a (light-dependent) pH decrease. Planta 192, 581–589. 10. Yamamoto,H.Y.,Chenchin,E.E.&Yamada,D.K.(1974)Effect of chloroplast lipids on violaxanthin de-epoxidase activity. Pro- ceedings of the 3rd International Cong. Photosyn, pp. 1999–2006. Elsevier, Amsterdam. 11. Siefierman-Harms, D., Ninnemann, H. & Yamamoto, H.Y. (1987) Reassembly of solubilized chlorophyll-protein complexes in proteolipid particles-comparison of monogalactosyldiacylglycerol and two phospholipids. Biochim. Biophys. Acta 892, 303–313. 12. Webb, M. & Green, B. (1991) Biochemical and biophysical properties of thylakoid acyl lipids. Biochim. Biophys. Acta 1060, 133–158. 13. Israelachvili, J.N. & Mitchell, D.J. (1975) A model for the packing of lipids in bilayer membranes. Biochim. Biophys. Acta. 389,13– 19. 14. Gruszecki, W.I. (1995) Different aspects of protective of the xan- thophyll cycle under stress conditions. Acta Physiol. Plant. 17, 145–152. 15. Arvidsson, P O., Carlsson, H., Stefansson, H., Albertsson, P A. &A ˚ kerlund, H E. (1997) Violaxanthin accessibility and tem- perature dependency for de-epoxidation in spinach thylakoid membranes. Photosynth. Res. 52, 39–48. 16. Milton, A., Wolff, G., Ourisson, G. & Nakatani, Y. (1986) Organization of carotenoid-phospholipid bilayer systems incorporation of zeaxanthin and their C-50 homologues into dimyristoylphosphatidylcholine vesicles. Helv. Chim. Acta 69, 12–24. 17. Latowski, D., Kostecka, A. & Strzalka, K. (2000) Effect of monogalactosyldiacylglycerol and other thylakoid lipids on vio- laxanthin de-epoxidation in liposomes. Bioch. Soc. Trans.,Vol.28 Part. 6, 810–812. 18. Enoch, H.G. & Stritmatter, P. (1979) Formation and properties of 1000 A ˚ -diameter, single-bilayer phospholipidvesicles. Proc. Natl Acad. Sci. USA 76, 145–149. 19. Davies, B.H. (1976) Carotenoids. In Chemistry and Biochemistry of Plant Pigments (Goodwin, T.W., ed.), pp. 65–66. Academic Press, London New York San Francisco. 20. Yamamoto, H.Y. (1995) Xanthophyll cycle. Meth Enzymol. 110, 303–312. 21. Gilmore, A.M. & Yamamoto, H.Y. (1991) Resolution of lutein and zeaxanthin using a non-endcappted, lightly carbon-loaded C 18 high-performance liquid chromatographic column. J. Chrom. 543, 137–145. 22. Latowski, D., Burda, K. & Strzalka, K. (2000) A mathematical model describing kinetics of conversion of violaxanthin to zeaxanthin via intermediate antheraxanthin by the xanthophyll cycle enzyme violaxanthin de-epoxidase. J. Theor. Biol. 206,507– 514. 23. Sprague, G.S. & Staehelin, L.A. (1984) Effect of Reconstitution Method on the structural organization of isolated chloroplast membrane lipids. Biochim. Biophys. Acta 777, 306–322. 24. Phillip, D., Molnar, P., Toth, G. & Young, A.J. (1999) Light- induced formation of 13-cis violaxanthin in leaves of Hordeum vulgarum. J. Photochem. Photobiol. B: Biol. 49, 89–95. 25. Hieber, A.D., Bugos, R.C. & Yamamoto, H.Y. (2000) Plant lipocalins: violaxanthin de-epoxidase and zeaxanthin epoxidase. Biochim. Biophys. Acta. 1482, 84–91. 26. De Kruijff, B., Verkleij, A.J., Van Echteld, C.J.A., Gerritsen, W.J., Mombers, C., Noordam, P.C. & Gier, J. (1979) The occurance of lipid particles in lipid bilayers as seen by 31 P NMR and freeze- fracture electron-microscopy. Biochim. Biophys. Acta. 555,200– 209. 27. Walde, P., Giuliani, A.M., Boicelli, C.A. & Luisi, P.L. (1990) Phospholipid-based reversed micelles. Chem. Phys. Lipids 53,265– 288. 28. Venetie ¨ , R.V. & Verkleij, A.J. (1981) Analysis of the hexagonal II phase and its relations to lipidic particles and the lamellar phase. A freeze-fracture study. Biochim. Biophys. Acta. 645,262– 269. 29. Haran ˜ czyk, H., Strza 3 ka,K.,Dietrich,W.&Blicharski,J.S.(1995) 31 P-NMR observation of the temperature and glycerol induced non-lamellar phase formation in wheat thylakoid membranes. J. Biol. Phys. 21, 125–139. 30. Quinn, P.J. & Williams, W.P. (1983) The structural role of lipids in photosynthetic membranes. Biochim. Biophys. Acta. 737,223– 266. 4664 D. Latowski et al. (Eur. J. Biochem. 269) Ó FEBS 2002 31. Gounaris, K., Sen, A., Brain, A.P.R., Quinn, P. & Williams, W.P. (1983) The formation of non-bilayer structures in total polar lipid extracts of chloroplast membranes. Biochim. Biophys. Acta. 728, 129–139. 32. Yamamoto, H.Y. & Higashi, R.M. (1978) Violaxanthin de-epo- xidase. Lipid composition and substrate specifity. Arch. Bioch. Biophys. 190, 514–522. 33. Sarry, J E., Montillet, J L., Sauvaire, Y. & Havaux, M. (1994) The protective function of the xanthophyll cycle in photosynthesis. FEBS Lett. 353, 147–150. 34. Gruszecki, W.I. & Strzalka, K. (1991) Does the xanthophyll cycle take part in the regulation of fluidity of the thylakoid membrane? Biochim. Biophys. Acta. 1060, 310–314. 35. Tardy, F. & Havaux, H. (1997) Thylakoid membrane fluidity and thermostability during the operation of the xanthophyll cycle in higher-plant chloroplasts. Biochim. Biophys. Acta 1330, 179–193. 36. Strzalka, K. & Gruszecki, W. (1997) Modulation of thylakoid membrane fluidity by exogenously added carotenoids. J. Biochem. Mol. Biol. Biophys. 1, 103–108. Ó FEBS 2002 Violaxanthin de-epoxidation in liposomes (Eur. J. Biochem. 269) 4665 . Kinetics of violaxanthin de-epoxidation by violaxanthin de-epoxidase, a xanthophyll cycle enzyme, is regulated by membrane fluidity in model lipid bilayers Dariusz. light-independent epoxi- dation of zeaxanthin to anteraxanthin and violaxanthin [2]. The conversion of violaxanthin to zeaxanthin is catalysed by violaxanthin