Trapping of short-lived intermediates in phospholipid phase transitions The La* phase Peter Laggner Heinz Amenitsch Manfred Kriechbaum Georg Pabst and Michael Rappolt Institute of Biophysics and X-ray Structure Research Austrian Academy of Sciences Steyrergasse 17 A-8010 Graz Austria Receiøed 12th August 1998 a Time-resolved small-angle X-ray diÜraction of liquid-crystalline phospholipid»water systems under temperature or pressure jump conditions has demonstrated the existence of an ordered intermediate L phase with a sub-second lifetime designated as the L*-phase. The lamellar repeat spacing is universally 0.3 nm smaller than that of the parent phase irrespective of the lipid composition and of the jump conditions provided that the jump leads to a net volume expansion of the phase.The presence of salts most notably LiCl leads to a prolongation of the lifetime. The results suggest a non-monotonic potential function for the interbilayer water thickness. a Faraday Discuss. 1998 111 31»40 I. Introduction Among the manifold molecular interactions which govern the intra- and intercellular communication in biomembranes the supramolecular structure and dynamics of lipid constituents deserves particular attention as they provide the structural variability and controlled one- or twodimensional —uidity essential for membrane function. The wide chemical diversity of membrane lipid constituents e.g. phospho- or glycolipids with varying hydrocarbon chain lengths saturation or branching sterols confers a powerful mechanism for the cell to control the physico-chemical properties mechanical electrostatic and chemical of membranes.1h4 A crucial point in the discussion of supramolecular lipid aggregates is their cooperativity,5 i.e.the fact that properties and mechanisms are governed by the collective behaviour of cooperative units rather than by individual molecules. In the extreme this leads to the concept of generic physical properties and interactions of bilayer membranes where the individual molecular structure plays only a subordinate role.6,7 One of the most interesting features of phospholipid»water systems is their polymorphism i.e. the existence of various ordered crystalline gel or liquid-crystalline phases.8 In the transitions between these phases the above-mentioned supramolecular nature becomes particularly apparent.9 Most transition processes are highly cooperative such that several hundreds of individual lipid molecules transit simultaneously and discontinuously from one phase to the other.It has long been an open question how one ordered phase changes into another with diÜerent geometry without disruption of lattice order.10 This problem is schematically illustrated in Fig. 1 for the simplest case of a one-dimensional lattice as given in principle by a multilamellar liposome structure. It seems that for highly cooperative transitions between two ordered phases with a minimisation of structural disruption and disorder there can only be highly symmetric and localised transition mechanisms as de–ned by the martensitic transition type originally de–ned for metallurgical phases11,12 [Fig.1(b)]. 31 Fig. 1 Scheme of two alternative transition mechanisms in lamellar phases. (a) Transition mechanism with zones of disorder resulting in a loss of coherence. (b) Transition mechanism with minimal loss of order and coherence (martensitic transition). In addition to the concept of cooperativity the non-equilibrium nature of transition processes gains relevance in the discussion of supramolecular processes and transitions. Molecular dynamics diÜusivity —exibility are normally treated as equilibrium phenomena.13,14 Close to equilibrium at minor elevations from the thermodynamic equilibrium potential well the kinetic and mechanistic behaviour can still be described classically by single-exponential energy and entropy terms.However under strong jump conditions at large elevations from the equilibrium potential trough the system may respond non-linearly no longer able to be described by single exponential terms.15,16 The simplest consequence of this notion is that transition processes may follow diÜerent pathways depending on whether they are guided in a slow isothermal or in a fast adiabatic fashion. Structural intermediates of diÜerent lifetime play a signi–cant role in this concept. Such intermediates have been postulated and de–ned some time ago in studies of membrane fusion by several groups.17v19 By time-resolved X-ray diÜraction experiments on various diÜerent lipid phases such intermediates have been demonstrated.20v22 Although the results have made it possible to propose hypothetical models it has not been possible so far to provide a sound structural analysis of these intermediates to good resolution.a In the present work this jump»relaxation approach has been focussed on intermediates within one thermodynamic equilibrium phase region the L -phase of phosphatidylcholines. The experiments performed with advanced jump»relaxation techniques and synchrotron X-ray diÜraction lead to the notion of a modulated potential versus bilayer-distance function and a universal thickness increment for the interbilayer water space which closely corresponds to the molecular dimensions of water. II. Materials and methods Sample preparation 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) egg-yolk phosphatidylcholine (EYPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphoethanolamine (POPE) were purchased from Avanti Polar Lipids Birmingham Alabama and used without further puri–cation.Multilamellar liposomes were prepared by dispersing weighed amounts of dry lipids typically 20»30 wt.% in bidistilled water and in solutions of 0.1 and 0.3 M LiCl respectively. To ensure complete hydration the lipid dispersions were incubated for ca. 1»2 h at least 10 °C above the main transition temperature and thereafter vigorously vortexed under a N atmosphere to prevent oxidation. Aqueous dispersions of these lipids dis- 2 played narrow cooperative melting transitions within the limits of published values,23 thus proving that the lipid purity corresponded to the claimed one of 99%.Faraday Discuss. 1998 111 31»40 32 Fig. 2 Schematic view of the set-up for T- and P-jump relaxation experiments. Throughout fast time-resolved X-ray diÜraction recordings aqueous suspensions can be studied far from the equilibrium situation. An IR laser with a pulse-characteristic of 1»2 J ms~1 provides T-jumps up to 20 °C and with an in-house-built pressure cell P-jumps up to 3 kbar can be carried out. Experimental protocol Fast time-resolved X-ray diÜraction experiments on lipid dispersions were carried out at the Austrian small-angle X-ray scattering station at ELETTRA Trieste.24v26 During rapid excitation of the lipid»water systems using temperature- (T-) and pressure (P-) -jump techniques the relaxation processes were monitored with a millisecond time resolution (Fig.2). For temperature-jump (T-jump) experiments the lipid dispersions were sealed in a thin-walled 1 mm diameter Mark capillary held in a steel cuvette which provides good thermal contact to the Peltier heating unit. T-jumps were generated with an erbium-glass laser27 with an pulse length of 2 ms and a maximum emitted energy of 4 J. Barotropic phase transitions were investigated with a high-pressure X-ray cell28 using jump amplitudes up to 3 kbar (0.3 GPa) within 10 ms. In particular P-jump induced phase transitions of the phospholipid DOPE within the temperature region 5»70°C and pressure region 1»3000 bar were performed. Data analysis The raw data of the time-resolved experiments were normalised for the integration time of each time-frame.Each small-angle X-ray diÜraction pattern was analysed by –tting the –rst-order Bragg re—ections using a least-squares method based on the Levenberg»Marquardt algorithm. Depending on the number of given phases the model function was given by one or the sum of two Lorentzians respectively. The lamellar repeat distances were determined from the corresponding peak positions. The relaxation kinetics of the d spacings are best described by a doubleexponential model (1) B B 0[A expA[ q t [B expA[ q t d(t)\d B A For comparison single and triple exponential models were also checked but have been proven by statistical tests (Variance-analysis F-Test) not to describe the relaxation kinetics as well as the two-component model.33 Faraday Discuss. 1998 111 31»40 a III. Results The intermediate La*-phase in diÜerent transitions The observation of a thin ordered lamellar structure which we denote as the L*-phase as a transient intermediate has been made in diÜerent jump»relaxation experiments on various lipid classes. In the following these shall be described separately. Faraday Discuss. 1998 111 31»40 34 a The La«La*«La transition (single phase). Fig. 3 shows the results of a typical experiment in a which a phosphatidylcholine lipid has been subjected to a T-jump starting from the single L phase which it attains under equilibrium conditions at the starting temperature. The jump amplitude was ca.15 °C and the maximum temperature reached was well within the L -phase region. In the jump experiment the d value –rst decreases discontinuously by ca. 0.3 nm and then relaxes back within less than 15 s to the equilibrium spacing. The minimum d spacing with a value of Fig. 3 L a»L a *»L a transition in POPC (20 wt.%) induced by a 15 °C(2 ms)~1T-jump (initial temperature 30 °C). (a) A series of time-sliced diÜraction patterns shows the temporal development of the –rst-order Bragg peaks (maximum resolution 5 ms). Raw data are given. (b) The evolution of the d spacings in fully hydrated POPC is displayed. Each single diÜraction pattern was –tted by a Lorentzian distribution ([). The line gives the best –t to the relaxation model of the L a * phase [eqn.(1)]. For comparison the temperature dependence of the lamellar repeat distance d of POPC under equilibrium conditions is depicted in the insert. ca. 6.2 nm is clearly thinner than the corresponding lattice parameter under near-equilibrium conditions which even at a temperature of 70 °C does not drop below 6.37 nm [see insert Fig. 3(b)]. Similar La ]La*]La experiments have been performed with other phosphatidylcholine and ethanolamine lipids and the resulting parameters are listed in Table 1. The result that with all lipids studied the non-equilibrium decrease in d value is ca. 0.3 nm irrespective of the hydrocarbon-chain or head-group composition emerges as a salient result. The La«La*«La-transition in two coexisting phases. Alkali-metal ions especially Li` can induce an equilibrium phase separation of two coexisting L -phases.29 These phases of which the structural nature and the origin of their coexistence is not yet quite clear diÜer in d spacings by ca.0.6 nm. It was of particular interest therefore to examine their behaviour under nonequilibrium jump conditions. Fig. 4 shows that both phases as demonstrated by the time-course of the lamellar repeat spacings respond in a parallel fashion to the T-jump and relax with similar kinetics. As with the single-phase transitions presented above the discontinuous changes in d spacing amount to ca. 0.3 nm and the relaxation times are of the order of 8»15 s i.e. still faster than the thermal equilibration within the sample cell. However the thermal equilibration does in—uence the relaxation times thus the relative behaviour is more relevant than the absolute values in the discussion.a Lipid a/degrees d /nm 0 T /°C f T /°C i 18 18 16 6.53 6.43 6.55 45 76.5 31.5 POPC DPPC EYPC POPE 0.1 M KCl 0.3 M LiCl 0.1 M LiCl 0.1 M MgCl2 a Table 2 The –tting parameters of the relaxation curves of POPC according to eqn. (1) phase 2 phase 1 phase 1 phase 1 phase 2 phase 1 no salt a In view of the dramatic eÜects of Li on the L -phase structure it is necessary to evaluate also the eÜects of other alkali-metal ions. This has not yet been done comprehensively due to the excessive requirements on synchrotron beam-time. Some interesting features of salt eÜects are already visible from Table 2 which summarises the results from all jump»relaxation experiments so far performed in the presence of salts.The general feature of the parameters obtained is the Table 1 Summary of laser T-jump experiments in the liquidcrystalline phase of various phospholipids *d/nm 0.31 0.32 0.26 30 70 25 30 36.5 5.37 0.37 21 All samples were equilibrated in the liquid-crystalline phase at the initial temperature T and jumps of 15 and 6.5 °C respectively were performed. The d spacing of the L phase d reduces i 0[*d. The theoreti- d directly after the laser-pulse to the value 0 cal declination angle a between the parent (L and nascentL a ) ( a *) phase respectively is described through eqn. (2). 0.06 0.09 0.06 0.06 0.10 0.09 0.13 A/s 0.5 0.8 0.6 0.3 0.3 1.0 0.5 0.23 0.25 0.23 0.27 0.27 0.29 0.18 3.1 8.3 8.1 15.8 14. 1 6.9 6.5 5.89 6.37 5.95 6.37 6.07 6.73 6.53 A/nm q B/nm q d /nm B/s 0 *d/nm 0.29 0.33 0.37 0.38 0.31 0.31 Increasing LiCl concentration results in longer lifetimes of the intermediate phase L (see A]qB). The a * kinetics of the salt-induced phases 1 and 2 respectively show similar relaxation behaviour. Independent from the salt concentration the d spacings of the liquid crystalline phases always decrease by ca. 0.3 nm directly after the laser pulse (*d\A]B). The errors in the parameters are of the order of the last digit given. 0. q 32 Faraday Discuss. 1998 111 31»40 35 Fig. 4 L a»L a *»L a transition in POPC (20 wt.% in 0.1 M LiCl) induced by a 15 °C(2 ms)~1 T-jump (initial temperature 30 °C).(a) A series of time-resolved diÜraction patterns given in the form of a contour plot demonstrates the kinetics of the salt-induced L phases (maximum resolution 5 ms). (b) The d spaca1 ings for L a1 and L a2 determined from the –tted Bragg-peak positions (sum of two Lorenztians). Both phases and L a2 exhibit similar temporal and structural behaviour (see also Table 2). prolongation of the lifetimes of the intermediates by alkali-metal salts. The prolongation factors reach a value of 5 for the addition of 0.3 M LiCl. a a a The La«La*«HII transition of phosphatidyl-ethanolamines pressure jumps. The existence of an intermediate thin L*-phase has been previously observed by T-jump experiments.20 This intermediate structure provides the contact conditions of opposing bilayers necessary for fusion and the formation of tubular structures.By P-jump experiments this result has been fully veri–ed in all details [Fig. 5(a)]. The initial fast step of the transition is the formation of a thin L*-phase which disappears in parallel with the formation of the H -phase taking comparatively long times of II several seconds to develop fully. In contrast to the T-jump technique the P-jump technique has the advantage of allowing jumps in either direction. This makes it possible to investigate the reversibility of the transitions and their pathways. Fig. 5(b) shows the pressure-drop experiment from the H phase –rst into the L -phase II and then onwards into the L phase.Two observations are particularly noteworthy. First the transition from the hexagonal phase into the lamellar phase proceeds without intermediates i.e. b Faraday Discuss. 1998 111 31»40 36 a a * Hphase. (b) II b a b a Fig. 5 Time-resolved X-ray diÜractograms of fully hydrated DOPE exposed to pressure-jumps recorded with a time resolution of 5 50 and 500 ms for each frame respectively. (a) Starting from the lamellar L phase at high pressure (T \41 °C and p\2300»155 bar) the lattice transforms immediately after a depressurizing jump into the intermediate lamellar L phase which coexists then with the emerging hexagonal Starting in the hexagonal H phase at low pressure the lattice transforms after a P-jump (T \20 °C and II p\1»2940 bar) immediately into the lamellar L phase.After 500 ms the –nal L phase begins to appear growing at the expense of the coexisting L phase until the phase transformation into L is completed. The lattice spacing of all phases remains constant over the entire time interval. b a a a there appears no L*-phase in this direction. Second the L -and L -phases coexist over a long period unlike in the other direction where the two phases coexist only for the period of the temperature jump i.e. 1»2 ms as described previously.30 No thin intermediate L*-phase can be detected in either direction of the La%Lb transition. a a a IV. Discussion a The programme under which the present investigation was performed is primarily aimed at the exploration of methods for prolonging the lifetimes of structural intermediates in phospholipid phase transitions.A bene–t from such an achievement could be the better structural description of the intermediates because longer lifetimes would lead to better precision of diÜraction data. Also there are likely to be biomedical bene–ts from such results since the development of agents modulating the dynamics of membrane transformations such as fusion is likely to play an important role in many medical applications e.g. liposome-based gene therapy fertility modulation or percutaneous drug applications to name but a few.31 Strong interest in such intermediates comes also from the –eld of nanomaterial research,32 where such structures could serve as templates for new materials which cannot be obtained under equilibrium conditions.The main discovery made in this search for trapped intermediates is the demonstration of a discrete ordered transition state the L*-phase which occurs rapidly and cooperatively upon a T- or P-jump from the normal equilibrium L -phase. This phase is always characterised by a ca. 0.3 nm lower d spacing than the parent phase independent of the hydrocarbon chain composition of the phosphatidylcholine species ; it also occurs transiently during the P- or T-jump from the L into the H phase. These two facts the constant *d and the composition independence II particularly that of hydrocarbon chain composition are taken to indicate that the La ]La* transition involves primarily a change in interbilayer water thickness.Indirect support for this idea comes from the fact that the same seems to be the case in the thin L phases found in the presence of LiCl (also separated by ca. 0.3 nm29) where the Fourier analysis of the coexisting diÜraction patterns of normal and thinner phases results in an invariant bilayer thickness (paper in preparation). It remains uncertain whether the transient L*-phase in the jump experiments is indeed the same as one of the equilibrium structures in the presence of LiCl. However it is tempting to assume that the two represent the same discrete secondary minimum in the hydration separation of bilayers and that the constant *d of ca. 0.3 nm relates to a change in water thickness a Faraday Discuss. 1998 111 31»40 37 by one molecular layer.Since there is presently no evidence for a diÜerent interpretation e.g. in terms of a discrete thinning of the hydrocarbon chain thickness or a change in head-group conformation we adhere to the hypothesis of a discontinuous hydration change. Two questions follow immediately from this hypothesis –rst how is the quasi-immediate thinning of the interbilayer water space achieved while the lattice order is fully preserved ? Second how does the intermediate L lattice return into the equilibrium a* La structure ? A tentative answer to the –rst question is presented by the martensitic lattice-disclination mechanism as shown schematically in Fig. 1(b). The transition would be localised in a discontinuous transition plane linking the parent to the nascent phase and moving rapidly with the speed of sound through the liposome.At the transition plane the two lattice planes would be disclinated at the angle a [Fig. 1(b)] which is given simply by the cosine relation between parent and nascent d spacings (2) a\arccosAd[*d \arccosA1[ *d d d B B where *d\0.3 nm. This mechanism results in a minimum disruption of lattice order and involves as a diÜusion component only the rapid movement of the transition plane thus providing for maximum transition speed. It should be emphasised that this behaviour is indeed demonstrated by Fig. 3(a) where the Bragg peaks immediately after the jump are very sharp and no intermediate disordering can be observed. What happens to the water where does it disappear to while the bilayer separation decreases ? The relative decrease in water layer thickness amounts to ca.15% assuming a value of 2 nm for the water layer in the fully hydrated L structure.33 A transient increase in water density for the lifetime of the intermediate (of the order of 0.1 s) seems unlikely considering the typical ps relaxation times of water. An efflux from the liposome structure into the excess water phase through transient defects in the lamellar lattice might be more plausible but it is again the observation of the very sharp Bragg re—ections right after the jump which indicates that such defects are not increased and suggests that this is not a likely mechanism. Another possibility would be an increase in bilayer surface area by ca. 15% and a concomitant reduction in water layer thickness thus conserving the interbilayer water volume.As a consequence the molar phospholipid volume would have to increase by the same proportion to conserve the bilayer thickness implying an increase in lateral headgroup separation of 7%. Perhaps the simplest way to dispose of the water is through the formation of localized ìì lentils œœ or cavities which would not gravely perturb the multibilayer order (Fig. 6). This would avoid the necessity of increasing the molecular surface area to conserve the bilayer thickness. Faraday Discuss. 1998 111 31»40 a While the formation mechanism of the intermediate L lattice appears to be best described by a* the discontinuous martensitic mechanism sketched in Fig. 1(b) the return to the equilibrium La structure follows a diÜerent pathway.As Fig. 3(b) and the decay parameters in Table 2 show this follows slow (on the experimental timescale) bi-exponential kinetics and most signi–cantly passes through a relatively disordered lattice situation as indicated by the broadening and decrease in intensity of the Bragg peaks only to increase again with complete re-equilibration. This can be interpreted qualitatively in terms of a model as shown in Fig. 1(a) where zones of disorder link the parent thin lattice with the nascent thicker one. The process could be analysed in terms of a nucleation-and-growth mechanism but a quantitative evaluation would require more detailed information on the morphology of the liposome particles during the process and is beyond the scope of the present work.The results of the T-jump experiments in the presence of LiCl (Fig. 4) indicate that the La*- lattice is not a limiting one to smaller thicknesses. There the initial equilibrium structure is Fig. 6 Schematic view of ìì lentils œœ of bulk water in multilamellar liposomes. 38 a* already separated into three coexisting phases with discrete d spacings diÜering by ca. 0.3 nm.29 The thickest and thinnest ones respectively diÜering in d spacing by ca. 0.6 nm are the dominant ones. The T-jump experiment shows an essentially identical behaviour for these two coexisting phases. Again immediately after the jump a discontinuous shift of both Bragg peaks to smaller (by 0.3 nm) d spacings is followed by a slow continuous return to the equilibrium situation.We assume that both parent phase structures have the normal molar phospholipid surface areas of approximately 0.63 nm2.34 The jump-induced thinning of the water-layer thickness could be again compensated by a transient increase in surface area or by the formation of ìì lentils œœ. The parallel return kinetics suggest the same mechanism and thermodynamic driving force for the two phases. The universality of the 0.3 nm increment in the jump response of hydrated phospholipids is further con–rmed by the behaviour of ethanolamine phospholipids in the La ]HII transition in a pressure-release-jump experiment [Fig. 5(a)]. There again the intermediate L phase shows a d spacing which is by the same amount thinner as was observed in all other cases so far investigated.An analogous observation has been made previously by a T-jump experiment. Hence *d seems to be largely independent of the chemical structure of the phospholipid and also of the physical nature of the jump as long as it leads to an overall volume expansion. In the pressurising jump through the HII ]La transition on the other hand no intermediate was observed [Fig. 5(b)]. a Having thus established the model for the L*-phase it remains to be examined what are the reasons for its formation. Our previous results of an LiCl-induced L -phase separation in liquidcrystalline phosphatidylcholines29 have suggested that there exists a modulation in the interbilayer separation potential with secondary minima at distances of ca.one molecular water diameter. The present results seem to support this notion although they are derived from a strongly non-equilibrium approach. This is in contrast to the hitherto generally held belief of a smooth continuous potential function with only one potential well at the equilibrium separation. 35 Indeed none of the studies on the hydration dependence of the lamellar repeat spacing has shown any sign of secondary minima.36h38 The reason might simply lie in the dynamic undulations of the individual bilayers such that at thermal equilibrium all states diÜering in energies by a few kT only are simultaneously occupied and the Bragg peaks represent the average bilayer separation. The simultaneous energy supplied by the T-jump would lead to a uniform occupation of the secondary potential minimum and hence to the observed sharp Bragg peak of the L phase.a a* The interbilayer separation in the L phase is ca. 2 nm corresponding hypothetically to 7 water layers if one layer is ca. 0.3 nm thick. Despite the highly dynamic state of the water between the bilayers it is obvious that the restrictions to mobility are much more stringent in the direction normal to the bilayer plane than in the parallel direction. These restrictions together with a modulated interaction potential between opposing bilayers would become more pronounced the closer the bilayers approach simply from a steric consideration of the discrete water molecule dimensions despite the —exibility and vertical ììbobbingœœ motions of individual phospholipids.The fact that the interbilayer water spacing decreases with temperature even isothermally and not just in the adiabatic jump experiments shows that the water volume expands thermally in the two dimensions of the bilayer plane. This creates an increased interface area which is counteracted by the manifold intermolecular phospholipid interactions the water bridges in the backbone region and by the hydrophobic eÜect which however decreases with temperature. We want to speculate therefore that the 0.3 nm increment is indeed the average thickness of one water layer in the sense that the interbilayer water space as it decreases under non-equilibrium conditions tends to prefer integral multiples of this thickness. Concerning the underlying intention to explore ways to prolong and eventually trap ordered Faraday Discuss.1998 111 31»40 a intermediate structures the present experiments with salts hold some promise although the obtained changes in relaxation times are not as dramatic as one would wish. The most important observation is that the salts used do not change the 0.3 nm increment by which the d spacings vary under jump conditions. The mechanism by which they prolong the lifetimes of the intermediates are still unclear and therefore no predictions in the direction of more potent trapping agents can be made. However since we assume as has been elaborated above that the cause for the discrete intermediate formation is the water structure it can be speculated that cosmotropic rather than chaotropic agents might be suitable candidates.If one considers the results obtained with 39 LiCl under equilibrium conditions,29 with Li` being at the cosmotropic end of the lyotropic series of monovalent cations this speculation gains in substance. Acknowledgements This work has been supported by the ììElettra-Projectœœ of the Austrian Academy of Sciences. M. Rappolt is the recipient of a long-term grant from the European Commission under the programme ììTraining and Mobility of Researchersœœ [Contract no. SMT4-CT97-9024(DG12-CZJU)]. Phasenué bergaé nge und Kritische Phaé nomene F. Vieweg & Sohn Braunschweig Paper 8/06384B References 1 M. Bloom and O. G. Mouritsen in Structure and Dynamics of Membranes ed. R. Lipowsky and E. Sackmann Elsevier Amsterdam 1995 p.65. 2 M. Bloom E. Evans and O. G. Mouritsen Q. Rev. Biophys. 1991 24 293. 3 E. Sackmann in ref. 1 p. 1. 4 P. Kinnunen Chem. Phys. L ipids 1991 57 375. 5 R. Biltonen J. Chem. T hermodyn. 1990 22 1. 6 R. Lipowski in Festkoé rperprobleme Advances in Solid State Physics ed. U. Roessler Vieweg Braunschweig/Wiesbaden 1992 vol. 32 p. 19. 7 W. Helfrich in ref. 1 p. 691. 8 J. M. Seddon and G. Cevc in Phospholipids Handbook ed. G. Cevc Marcel Dekker New York 1993 p. 403. 9 D. Marsh Chem. Phys. L ipids 1991 57 109. 10 P. Laggner and M. Kriechbaum Chem. Phys. L ipids 1991 57 121. 11 J. W. Christian in Physical Metallurgy ed. R. W. Cahn North-Holland Amsterdam 1970 p. 471. 12 Z. Nishiyama Martensitic T ransformations Material Science Series ed.M. Fine M. Meshii and C. Wayman Academic Press New York 1978. 13 A. Blume in ref. 8 p. 455. 14 P. F. F. Almeida and W. L. C. Vaz in ref. 1 p. 305. 15 W. Gebhardt and U. Krey 1980. 16 D. Lou J. Casas-Vasquez and G. Lebon Extended Irreversible T hermodynamics Springer Berlin 1993. 17 P. R. Cullis M. J. Hope B. deCruijÜ A. Verkleij and C. P. S. Tilcock in Phospholipids and Cellular Recognition ed. J. F. Kuo CRC Press Boca Raton 1985 p. 1. 18 D. P. Siegel Biophys. J. 1986 49 1155. 19 D. P. Siegel Biophys. J. 1986 49 1171. 20 P. Laggner M. Kriechbaum and G.Rapp J. Appl. Crystallogr. 1991 24 836. 21 P. Laggner J. Phys. IV 1993 3 259. 22 M. Rappolt PhD Thesis Hamburg 1995. 23 M. CaÜrey L ipid T hermotropic Phase T ransition Database (L IPIDAT 2) NIST 1994.24 H. Amenitsch S. BernstorÜ M. Kriechbaum D. Lombardo H. Mio M. Rappolt and P. Laggner J. Appl. Crystallogr. 1997 30 872. 25 H. Amenitsch S. BernstorÜ and P. Laggner Rev. Sci. Instrum. 1995 66 1624. 26 S. BernstorÜ H. Amenitsch and P. Laggner J. Synchrotron Rad. 1998 5 1215. 27 G. Rapp M. Rappolt and P. Laggner Prog. Colloid Polym. Sci. 1993 93 25. 28 K. Pressl M. Kriechbaum M. Steinhart and P. Laggner Rev. Sci. Instrum. 1997 68 4588. 29 M. Rappolt K. Pressl G. Pabst and P. Laggner Biochim. Biophys. Acta 1998 1372 389. 30 M. Kriechbaum G. Rapp J. Hendrix and P. Laggner Rev. Sci. Instrum. 1989 60 2541. 31 D. D. Lasic L iposomes From Physics to Applications Elsevier Amsterdam 1993. 32 S. Mann Nature (L ondon) 1988 332 119. 33 J. F. Nagle R. Zhang S. Tristram-Nagle W. Sun H. I. Petrache and R. M. Suter Biophys. J. 1996 70 1419. 34 S. Tristram-Nagle H. I. Petrache and J. F. Nagle Biophys. J. 1998 75 917. 35 J. Israelachvili in Intermolecular and Surface Forces Academic Press London 1992 p. 176. 36 R. P. Rand and V. A. Parsegian Biochim. Biophys. Acta 1989 988 351. 37 T. J. McIntosh and S. A. Simon Biochemistry 1993 32 8374. 38 B. Koenig H. H. Strey and K. Gawrisch Biophys. J. 1997 73 1954. Faraday Discuss. 1998 111 31»40 40