Chemistry and Physics of Lipids, 62 (1992) 39-43 Elsevier Scientific Publishers Ireland Ltd.

39

Activation energy and entropy for intramolecular excimer formation in a dipyrenylphosphatidylcholine probe in lamellar and hexagonal lipid phases Peter Butko and Kwan

Hon Cheng

Department of Physics, Texas Tech University, Lubbock, TX 79409 (USA)

(Received December 31st, 1991; revision received March 16th, 1992; accepted March 31st, 19921

lntramolecular excimer formation in pyrene-labeled phosphatidylcholine was used as a tool to determine thermodynamic characteristics of the lamellar to hexagonal phase transitions in a binary lipid system dilinoleoylphosphatidylethanolamine (DLPE)/palmitoyloleoylphosphatidylcholine(POPC). Upon an LdHn phase transition, the activation energy Ea for excimer formation increased from 5.6 ± 0.2 kcal/mol to 6.3 ± 0.2 kcal/mol, while the activation entropy AS¢ decreased from --40.0 ± 0.8 cal/K.mol to -38.4 ± 0.8 cal/K.mol. The results are consistent with the idea of molecular splaying of the acyl chains in the hexagonal phase. It is estimated that the molecular area at the terminal carbon of the lipid acyl chains increases by a factor of 2.2 upon the I.aHn transition in DLPE/POPC. Key words: pyrene-labeled lipid; excimer; phase transition

Introduction Di-(1 '-pyrenemyristoyl)-phosphatidylcholine (dipyPC) is a fluorescent phospholipid in which two fluorophores (pyrenes) are conjugated to the terminal carbons o f both acyl chains. When one of the fluorophores is excited, there is a probability that, during the fluorescence lifetime, the excitation will be shared with the other fluorophore, creating an excited-state dimer, an excimer. Intermolecular excimer formation, where the excimer is formed by two fluorophores in two different molecules [1], has been used to study the dynamic and structural properties of lipid membranes [2-4]. The rate of intermolecular excimer formation depends on the fluorescent probe concentration and its diffusibility in the lipid bilayer, which is determined by temperature and the membrane fluidity. Correspondence to: Peter Butko, Institute for Biological Sciences, M-54, National Research Council, Ottawa, Canada KIA 0R6.

Intramolecular excimer formation, where the excimer is formed by two fluorophores within the same molecule, was introduced into membrane studies only recently. Vauhkonen et al. [5] used dipyrenylphosphatidylcholine as a membrane fluidity probe and Cheng et al, [6] showed that dipyPC fluorescence responds to the lamellar to hexagonal phase transition of the matrix lipid. The rate of intramolecular excimer formation is concentration-independent, but is a function of temperature and the geometry of lipid packing (the free volume of a membrane, i.e. the fractional volume not occupied by the matrix molecules). In this study we determined activation energies and entropies for intramolecular excimer formation in dipyPC in a binary system consisting of dilinoleoylphosphatidylethanolamine (DLPE) and l-palmitoyl-2-oleoylphosphatidylcholine (POPC). We also presented a simple theory based on basic thermodynamics that allowed us to estimate the geometric change in lipid packing upon a phase transition from lamellar L~ to inverted hexagonal HII phase.

0009-3084/92/$05.00 © 1992 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland

40

Materials and Methods

-4.2 0

DLPE and POPC were purchased from Avanti Polar Lipids (Birmingham, AL) and used without further purification. Fluorescent lipids di-(l'pyrenemyristoyl)-phosphatidylcholine (dipyPC) and 1-palmitoyl-2- [ 14-(1-pyrenyl)tetradecanoyl]phosphatidylcholine were a generous gift from Dr. P. Somerharju (University of Helsinki, Finland). The samples - vortexed and mildly sonicated lipid dispersions in a buffer (100 mM NaCI, 10 mM N-tris-(hydroxy-methyl)-methyl-2-aminoethanesulfonic acid (TES) and 2 mM EDTA, pH 7.4)were prepared as in Ref. 6. The final concentration of lipid was about 100/~g/ml, the dipyPC to lipid molar ratio 1:1000. Fluorescence was excited with a Xe arc lamp (ILC Technology, Sunnyvale, CA) at 325 nm. The spectra were recorded with an ISS multifrequency cross-correlation fluorometer (Urbana, IL). Since the samples were quite turbid, a correction for light scattering was essential. An identical sample which did not contain the fluorescent lipid was prepared in parallel and used as a blank; its signal was subtracted from the fluorescence spectra. The values of the steady-state excimer to monomer intensity ratios (E/M) were calculated as the ratios of fluorescence intensities at 475 and 392 nm from spectra taken at different temperatures in two systems: DLPE/POPC in molar ratios 50:50 and 100:0. The former system is in the lamellar L~ phase and the latter in the inverted hexagonal Hn phase throughout the temperature region investigated [7]. Results and Discussion

Arrhenius plots of the dipyPC E/M ratio in both DLPE/POPC compositions are shown in Fig. 1. The points fit straight lines with the correlation coefficients r = 0.99 and 0.98 for the lamellar and hexagonal phases, respectively. The shift between the two lines confirms our previous finding [61: the E/M ratio in the hexagonal phase is lower than that in the lamellar phase at each single temperature. In addition, one-tailed Student's ttest showed that the difference d (± standard error) between the two slopes (absolute values 3.17

0

-4.6

• ,--

-5.0

0

0

•

0

s~ ,ll v

C

-5.4

-5.8

-6.23.2

'

31.3

I 3.4

IO00/T

,

I 35.

, 3.6

[l/K]

Fig. 1. Arrhenius plot of the excimer to monomer fluorescence intensity ratio E/M of dipyPC in lameUar ((3, 50 tool% phosphatidylethanolamine (PE)) and inverted hexagonal (@, 100 tool% PE) phases of DLPE/I~PC.

and 2.83) differs from zero: d = 0.34 ± 0.23. The hypothesis that the absolute value of the slope in the hexagonal phase is not larger than that in the lamellar phase could be rejected at the level of significance P < 0.08. The E/M ratio is related to the rate constants in an excimer-forming system [1,2,5] as: E/M = (kFD/kFM) kDM (k D + kMD) -1

(i)

where kFD and kFM are the fluorescence decay parameters of excimer and monomer, respectively, kDM and kMD are the rate constants of excimer formation and dissociation and (k D + kMo) -l is the lifetime rD of the excimer (ko is the sum of radiative and non-radiative decay rates of the excimer). In order to analyse our results within the framework of the Birks' kinetic model, several assumptions were made. (i) The ratio kvo/kFM is independent of temperature. This was found to be true for intermolecular excimer formation by pyrene in isotropic solutions [1] and for intramolecular excimer formation in a series of dipyrenylmethoxycarbonylalkanes [8]. The ratio was also found constant and nearly constant in the liquidcrystalline phases of DPPC and POPC, respectively [2]. (ii) The excimer lifetime 1"o is phase- and

41

temperature-independent above the Lt~/L~ transition temperature (above the melting point of acyl chains). This was confirmed in preliminary timeresolved measurements in the DLPE/POPC system (rD = 50 ns) [9]. In addition, Hresko et al. [2] reported that the excimer lifetime in DPPC and POPC is only weakly dependent on temperature (about -0.2 ns/degree) above the acyl chains melting point. (iii) The rate of fluorescence quenching by oxygen is similar in both L~ and Hn phases. This is important to assume since our measurements were performed in the presence of air. It is known that oxygen is more soluble in the liquid-crystalline phases than in the gel phases of lipids [10], which contributes to large changes in excimer lifetimes upon melting of the lipid acyl chains. However, the present study deals with the L d H n transition which takes place above the melting point temperature. In the lack of data on oxygen solubility and diffusibility in H n phase, it is reasonable to assume that the determining factor for oxygen solubility and diffusibility is the physical state of acyl chains: when the chains are frozen (in La phase) oxygen is excluded from the lipid, whereas in the phases where the chains are melted (both L~ and Hn phases), oxygen can diffuse in the lipid. Indirect support for the above assumptions comes from the data in Fig. 2, where we measured temperature dependence of the E/M ratio of the monopyrenyl PC probe. In such a probe where only one of the acyl chains is labeled with pyrene, the rate of excimer formation does not depend as much on the geometry of lipid packing, but rather on the probe collision frequency, that is on the probe concentration and diffusibility. It is seen that the two data sets - monopyrenyl E/M ratios in lamellar and hexagonal p h a s e - a r e indistinguishable within experimental error. Since photophysical properties of an excimer do not depend on whether the excimer was formed intra- or intermolecularly, this result has a direct bearing for the experiment with dipyPC. Namely, it indicates that there are no significant changes upon LJHII phase transition in the parameters of the system, other than the change in the free volume. Any possible changes occurring upon the phase

-5.8 -6.0 -6.2

I--6.4 ~ ¢-

-8.6

-6.8 -7.0 -7.2 3.2

i

i

3.3

,

34

IO00/T

i

3.5

i

3.6

3.7

[l/K]

Fig. 2. Arrhenius plot of the excimer to monomer fluorescence intensity ratio E/M of monopyrenylPC in lamellar (O, 40 mol% PE) and inverted hexagonal (@, 100 mol% PE) phases of DLPE/POPC.

transition, be it due to changes in the kFD/kFM ratio, excimer lifetime or the rate of oxygen quenching, would reveal themselves in the Arrhenius plots of the monopyrenyl PC E/M ratio in Fig. 2 in the form of a shift between the two lines or a difference in the slopes. Thus, with the above assumptions, the predominant temperature-dependent term in Eqn. 1 will be the excimer formation rate kDM and the E/M ratio will be proportional to kDM. Indeed, the plots of E/M and kDM as a function of temperature are very similar, as observed by Hresko et al. [2] in DPPC and POPC and by Cheng et al. [6] in DOPC and DOPE. Figure 1 shows that the E/M ratio and consequently, also koM, can be written in the form of an Eyring equation: kDM ---- C E/M = C A T exp(-Ea/RT)

(2)

where C (= rD-l kFM/kFD) is a constant of proportionality between koM and E/M, and A and Ea have their usual thermodynamic meanings, i.e., frequency factor and activation energy, respectively. Based on the reported measurements of To [9] and kFM/kVD[2], the value of C is in the order of 2 x l 0 7 s -1. The values of A and Ea can be determined from an Arrhenius plot as the intercept at l/T = 0 and the slope, respectively.

42 For a bimolecular rate-determining step in liquid phase, Ea = A H s + R T [11], where AH s is the activation enthalpy. The relationship between AH s and the Gibbs energy AGs of activation, AGs = AH s - T S, together with the frequency factor derived from the activated complex theory [11], yield the relationship kDM = CK (kT/h) exp(ASS/R) exp(-Ea/RT)

(3)

where Kis the transmission coefficient, assumed to be 1 in this work, e is the base of natural logarithms and AS s the activation entropy. The latter can be calculated from the experimental data as AS s = R (lnA - ln(k/h) - 1), where k and h are the Boltzrnann and Planck constants, respectively. Note that the two Ea's in Eqns. 2 and 3 are the same. Thus, the activation energy determined from the steady-state E/M ratio measurements is the true activation energy for the rate of excimer formation. The activation energies and entropies for the excimer formation, as determined from Fig. 1, are 5.6 ± 0.2 kcal/mol and -40.0 ± 0.8 cal/K.mol in the lamellar and 6.3 ± 0.2 kcal/mol and -38.4 ± 0.8 cal/K.mol in the hexagonal phase, respectively. (The standard errors were determined by error propagation method.) The energies are within the range of the values measured in various systems: 3.2 kcal/mol for pyrene in cyclohexane [1], 6.2-11.6 kcal/mol for dipyrenylPC in PC bilayers [5], and about 5 kcal/mol for dipyrenylmethoxycarbonylalkanes in 2-methyltetrahydrofuran [8] and for dipyrenylpropane in toluene [12]. All these values are higher than that required for rotational motion about carbon-carbon bonds in alkanes [13]. We previously suggested [6], that there are two important factors determining the facility of attaining the proper orientation of the two pyrenes in the probe molecule: namely, the free volume of a membrane (a geometrical factor) and the intramolecular thermal motion (a dynamical factor). The larger the free volume, the greater the number of microstates that can be attained by the two acyl chains in a lipid, the further the average distance between the two terminal pyrenes and consequently, the lower the rate of excimer formation. On the other hand, the higher the

temperature, the higher the rate of sampling through all the microstates and consequently, the higher the rate of excimer formation. These two competing processes make the thermotropic phase transitions usually difficult to interpret. However, in the DLPE/POPC systems at constant temperature, but of different composition, the change in the free volume of the membrane (or the change in the geometry of lipid packing) is solely responsible for the change in the excimer formation rate. A similar conclusion has been arrived at by Vauhkonen et al. [5] and by Sassaroli et al. [14]. The principle of equivalence between the thermodynamic and statistical entropy definitions allows us to write the following equation exp(AASS/R) = exp((ASSn- ASSL)/R) = W H / W L

(4) where ~ S s is the difference between the activation entropies in the hexagonal and lamellar phases and WH and WL are the numbers of available microstates in the hexagonal and lamellar phases, respectively. Thus, by attributing a certain physical reality to the activated complex, we are able to estimate the change in the number of microstates of the terminal pyrene moieties in dipyPC upon the transition from lamellar to hexagonal phase. The values of activation entropies in the two phases imply that the number of microstates in the hexagonal phase is 2.2 times larger than that in the lamellar L~ phase of DLPE/ POPC. Recent results of De Loof et al. [15] on stochastic dynamics simulation of a phospholipid molecule indicate that, regarding movement in the direction of the long molecular axis, the terminal carbon of an acyl chain spends about 2/3 of the time within 0.5 nm around its equilibrium position. Therefore, in a crude approximation we can assume that the terminal pyrenes of dipyPC only move within a plane perpendicular to the lipid molecular axis. Then the ratio of the numbers of microstates W n / W L equals the ratio of the areas occupied by the ends of the acyl chains in the two phases. Consequently, our result suggests that the area occupied by the ends of the acyl chains is 2.2 times larger in the hexagonal phase than in the

43

lamellar phase. This is consistent with the idea of molecular splaying of the acyl chains in the hexagonal phase [6,16-18]. Unfortunately, the values of intercepts determined from Arrhenius plots are always very sensitive to the data scattering. Due to relatively large scattering in the data points of Fig. 1 and a limited temperature range that was dictated by the optical properties of the samples (considerable aggregation of lipids occurred above 312 K) the values of AAS* and WH/WLare burdened with large uncertainties. The latter were determined through error propagation as 1.6 cal/K.mol and 1.7, respectively. Nevertheless, our best estimate for the area ratio (2.2) is very close to the value of 2 determined by X-ray diffraction studies on the dioleoylphosphatidylethanolamine system (R.P. Rand, pers. commun.). In the present study we estimated both activation energies and entropies for the intramolecular excimer formation of dipyPC in lamellar and hexagonal phases. The negative values of the activation entropies are relatively high, which indicates that entropy, not energy, may be the determining factor in the process studied, at least under certain conditions. Further support for the ideas put forward in this work may come from time-resolved fluorescence measurements and from computer simulations (molecular dynamics and/or Monte Carlo) of the acyl chains in the lipid polymorphic phases.

Acknowledgements This work was supported by grants from the National Cancer Institute (PHS CA 47610) and the Robert A. Welch Research Foundation (D-1158) to K.H.C. Thanks are due to Dr. M.

Sassaroli for critical comments and suggestions for future work.

References 1 J.B. Birks (1970) Photophysics of Aromatic Molecules, Wiley Interscience, New York. 2 R.C. Hresko, I.P. Sugar, Y. Barenholz and T.E. Thompson (1986) Biochemistry 25, 3813-3823. 3 S.Y. Chen, K.H. Cheng and D.M. Ortalano (1990) Chem. Phys. Lipids 53, 321-330. 4 M. Sassaroli, M. Vauhkonen, D. Perry and J. Eisingcr (1990) Biophys. J. 57, 281-290. 5 M. Vauhkonen, M. Sassaroli, P. Somerharju and J. Eisinger (1990) Biophys. J. 57, 291-300. 6 K.H. Cheng, S.Y. Chen, P. Butko, B.W. van der Meet and P. Somerharju (1991) Biophys. Chem. 39, 137-144. 7 T.L. Boni and S.W. Hui (1983) Biochim. Biophys. Acta 731, 177-185. 8 T. Kanaya, K. Goshiki, M. Yamamoto and Y. Nishijima (1982) J. Am. Chem. Soc. 104, 3580-3587. 9 L.I. Liu, K.H. Cheng, S.Y. Chen, B.W. van der Meet and P. Somerharju (1991) Biophys. J. 59, 505a. l0 S. Fischkoffand J.M. Vanderkooi (1975) J. Gen. Physiol. 65, 663-676. 11 P.W. Atkins (1978) Physical Chemistry, W.H. Freeman and Co., San Francisco. 12 M.J. Snare, P.J. Thistlethwaite and K.P. Ghiggino (1983) J. Am. Chem. Soc. 105, 3328-3332. 13 O.J. Sovers, C.W. Kern, R.M. Pitzcr and M. Karplus (1968) J. Chem. Phys. 49, 2592-2599. 14 M. Sassaroli, S. Scadata, M. Vauhkonen and P. Somerharju (1990) Biophys. J. 57, 483a. 15 H. De Loof, S.C. Harvey, J.P. Segrest and R.W. Pastor (1991) Biochemistry 30, 2099-2113. 16 P. Butko, S.Y. Chert, K.H. Cheng, B.W. van der Meet and P. Somerharju (1990) Biophys. J. 57, 485a. 17 S.M. Gruner, P.R. Cullis, M.J. Hope and C.P.S. Tilcock (1985) Annu. Rev. Biophys. Biophys. Chem. 14, 211-238. 18 R.P. Rand, N.L. Fuller, S.M. Grnner and V.A. Parsegian (1990) Biochemistry 29, 76-87.