Effect of Vesicle Size on the Prodan Fluorescence in

Jul 22, 2010 - see the location of Prodan in the bilayer membrane as blue valleys were constructed for LMV and GMV under atmospheric pressure...
0 downloads 0 Views 4MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Effect of Vesicle Size on the Prodan Fluorescence in Diheptadecanoylphosphatidylcholine Bilayer Membrane under Atmospheric and High Pressures Masaki Goto,† Hiroshi Sawaguchi,‡ Nobutake Tamai,† Hitoshi Matsuki,*,† and Shoji Kaneshina† †

Department of Life System, Institute of Technology and Science and ‡Department of Biological Science and Technology, Faculty of Engineering, The University of Tokushima, 2-1 minamijosanjima-cho, Tokushima 770-8506, Japan Received March 2, 2010. Revised Manuscript Received July 3, 2010

The bilayer phase behavior of diheptadecanoylphosphatidylcholine (C17PC) with different vesicle sizes (large multilamellar vesicle (LMV) and giant multilamellar vesicle (GMV)) was investigated by fluorescence spectroscopy using a polarity-sensitive fluorescent probe Prodan under atmospheric and high pressures. The difference in phase transitions and thermodynamic quantities of the transition was hardly observed between LMV and GMV used here. On the contrary, the Prodan fluorescence in the bilayer membranes changed depending on the size of vesicles as well as on the phase states. From the second derivative of fluorescence spectra, the three-dimensional image plots in which we can see the location of Prodan in the bilayer membrane as blue valleys were constructed for LMV and GMV under atmospheric pressure. The following characteristic behavior was found: (1) the Prodan molecules in GMV can be distributed to not only adjacent glycerol backbone region, but also near bulk-water region in the lamellar gel or ripple gel phase; (2) the blue valleys of GMV became deeper than those of LMV because of the greater surface density of the Prodan molecules per unit area of GMV than LMV; (3) the liquid crystalline phase of the bilayer excludes the Prodan molecules to a more hydrophilic region at the membrane surface with an increase in vesicle size; (4) the accurate information as to the phase transitions is gradually lost with increasing vesicle size. Under the high-pressure condition, the difference in Prodan fluorescence between LMV and GMV was essentially the same as the difference under atmospheric pressure except for the existence of the pressure-induced interdigitated gel phase. Further, we found that Prodan fluorescence spectra in the interdigitated gel phase were especially affected by the size of vesicles. This study revealed that the Prodan molecules can move around the headgroup region by responding not only to the phase state but also to the vesicle size, and they become a useful membrane probe, detecting important membrane properties such as the packing stress.

1. Introduction Phospholipids form bilayer aggregates in aqueous solutions. The aggregates induce structural changes called phase transitions depending on environmental variables such as temperature, pressure, pH, and concentrations of added salts. We have focused our attention on the effect of pressure on the phospholipid bilayer membranes because a high-pressure study has the following merits: (1) complete thermodynamic quantities of bilayer phase transitions including volume information in addition to thermal information can be obtained, (2) the bilayer phase behavior can be discussed by the two-dimensional temperature (T)-pressure (p) phase diagram, and (3) low-temperature transitions, which cannot be detected by conventional experimental methods, can be observed by applying pressure due to elevation of phase-transition temperatures. Most pressure studies of phospholipid bilayer membranes have been performed on a dipalmitoylphosphatidylcholine (DPPC) bilayer membrane by use of several physicochemical techniques including dilatometry, Raman spectroscopy, calorimetry, X-ray *Correspondence author. Address: Department of Life System, Institute of Technology and Science, The University of Tokushima, 2-1 minamijosanjima-cho, Tokushima 770-8506, Japan. TEL: þ81-88-656-7513; FAX þ81-88-655-3162; E-mail address: [email protected]. (1) (2) (3) (4) (5) (6) 371.

Utoh, S.; Takemura, T. Jpn. J. Appl. Phys. 1985, 24, 356–360. Yager, P.; Peticolas, W. L. Biophys. J. 1980, 31, 359–370. Wong, P. T. T.; Mantsch, H. H. Biochemistry 1985, 24, 4091–4096. Russell, N. D.; Collings, P. J. J. Chem. Phys. 1982, 77, 5766–5770. Chong, P. L. G.; Weber, G. Biochemistry 1983, 22, 5544–5550. Lakowicz, J. R.; Thompson, R. B. Biochim. Biophys. Acta 1983, 732, 359–

Langmuir 2010, 26(16), 13377–13384

diffraction, neutron diffraction, and NMR.1-10 We have also examined barotropic phase transitions of bilayer membranes of diacylphosphatidylcholines (diacyl-PCs) by the light-transmittance measurements and constructed their T-p phase diagrams.11-13 Fluorescence probe techniques have been proven to be useful to reveal the effect of hydrostatic pressure on model and cell membranes. A variety of probes such as diphenylhexatrien, perylene, pyren, dipyrenylpropane, and 12-(9-anthroyloxy) sterate have been used to examine the molecular order of the phospholipids under high pressure.14-17 Probe locations in lipid membranes under high pressure were also investigated by using the polarity-sensitive fluorescent probe 6-propionyl-2-(dimethylamino)naphthalene (Prodan).18 Unlike the former probes, (7) Braganza, L. F.; Worcester, D. L. Biochemistry 1986, 25, 2591–2596. (8) Winter, R.; Pilgrim, W. C. Ber. Bunsenges. Phys. Chem. 1989, 93, 708–717. (9) Maruyama, S.; Matsuki, H.; Ichimori, H.; Kaneshina, S. Chem. Phys. Lipids 1996, 82, 125–132. (10) Driscoll, D. A.; Jonas, J.; Jonas, A. Chem. Phys. Lipids 1991, 58, 97–104. (11) Ichimori, H.; Hata, T.; Matsuki, H.; Kaneshina, S. Biochim. Biophys. Acta 1998, 1414, 165–174. (12) Kusube, M.; Goto, M.; Tamai, N.; Matsuki, H.; Kaneshina, S. Chem. Phys. Lipids 2006, 142, 94–102. (13) Kusube, M.; Matsuki, H.; Kaneshina, S. Biochim. Biophys. Acta 2005, 1668, 25–32. (14) Chong, P.L.-G.; van der Meer, B. W.; Thompson, T. E. Biochim. Biophys. Acta 1985, 813, 253–265. (15) Flamm, M.; Okubo, T.; Turro, N. J.; Schachter, D. Biochim. Biophys. Acta 1982, 687, 101–104. (16) Turley, W. D.; Offen, H. W. J. Phys. Chem. 1986, 90, 1967–1970. (17) Chong, P.L.-G.; Fortes, P. A. G.; Jameson, D. M. J. Biol. Chem. 1985, 260, 14484–14490. (18) Chong, P.L.-G. Biochemistry 1988, 27, 399–404.

Published on Web 07/22/2010

DOI: 10.1021/la100871z

13377

Article

Prodan responds sensitively to its surrounding microscopic environments in a phospholipid bilayer with low polarity in the hydrocarbon region and high polarity at the water-lipid interface. The emission spectra of Prodan in the bilayer membrane change markedly at the main transition between the gel and liquid crystalline (LR) phases and at the pressure-induced bilayer interdigitation.18 The rearrangement of the Prodan molecules in a bilayer membrane causes a significant change of its surrounding micropolarity, and as a result leads to the variation of the emission spectra. In the previous study,19 we also used the Prodan fluorescence technique to investigate the dependence of Prodan fluorescence on acyl chain length for the bilayers of several DPPC homologues and showed that the fluorescence intensity of Prodan in the gel phase increases with a decrease in acyl chain length or an increase in pressure. Moreover, we established an advantageous method of analyzing the emission spectrum to estimate the location of the Prodan molecules oriented in bilayer membranes in terms of the second derivative of the original emission spectrum,20 and thereby determined the wavelengths of the emission maxima characteristic of the LR, ripple gel (Pβ0 ), lamellar gel (Lβ0 ), and interdigitated gel (LβI) phases for bilayer membranes of symmetric diheptadecanoyl-PC (C17PC) and asymmetric palmitoylstearoyl-PC and stearoylpalmitoyl-PC, all of which are PCs with the same total chain length. The analysis method enables us to detect the phase transitions much more precisely, and we clarified the difference in interdigitation between symmetric and asymmetric PC bilayers. On the other hand, the size of bilayer aggregates is well-known to affect the bilayer properties markedly. The effect of the vesicle size on the thermotropic phase behavior of disaturated PCs has already been investigated using small unilamellar vesicles (SUVs) by means of differential scanning calorimetry (DSC),21,22 dynamic heat capacity measurement,23 and solvent relaxation technique,24 demonstrating that exhaustive sonication abolishes the pretransition and drastically reduces the enthalpy change (ΔH) and cooperativity of the main transition. We are interested in how different the bilayer properties of the micrometer-scale vesicle, generally known as a giant vesicle, are under high pressure as compared with those of the vesicle of the submicrometer order. As far as we know, however, there has been almost no report of the effect of the vesicle size on the bilayer properties in that range of the vesicle size under high pressure, although increasing use of giant vesicles has been recently made for the purpose of observing its morphological changes directly by means of various kinds of microscopy including the fluorescent microscopy. In this study, we examined the bilayer phase transitions of C17PC with different vesicle sizes under high pressure by using the Prodan fluorescence. Although C17PC is not a naturally occurring phospholipid, we have revealed that the bilayer phase behavior of C17PC is essentially similar to that of a naturally occurring phospholipid like DPPC.11 Therefore, we chose C17PC also in this study as used in the previous study.20 (19) Kusube, M.; Matsuki, H.; Kaneshina, S. Colloids Surf., B: Biointerfaces 2005, 42, 79–88. (20) Goto, M.; Kusube, M.; Tamai, N.; Matsuki, H.; Kaneshina, S. Biochim. Biophys. Acta 2008, 1778, 1067–1078. (21) de Kruijff, B.; Cullis, P. R.; Radda, G. K. Biochim. Biophys. Acta 1975, 406, 6–20. (22) van Dijck, P. W. M.; de Kruijff, B.; Aarts, P. A. M. M.; Verkleij, A. J.; de Gier, J. Biochim. Biophys. Acta 1978, 506, 183–191. (23) Nagano, H.; Nakanishi, T.; Yao, H.; Ema, K. Phys. Rev. E 1995, 52, 4244– 4250. (24) Sykora, J.; Jurkiewicz, P.; Epand, R. M.; Kraayenhof, R.; Langner, M.; Hof, M. Chem. Phys. Lipids 2005, 135, 213–221.

13378 DOI: 10.1021/la100871z

Goto et al.

2. Experimental Section 2.1. Materials and Sample Preparation. A synthetic phospholipid, 1,2-diheptadecanoyl-sn-glycero-3-phosphocholine, was purchased from Avanti Polar Lipids, Inc. (Alabaster, AL) and used as received. The fluorescent probe, 6-propionyl-2-(dimethylamino)naphthalene, was obtained from Molecular Probes, Inc. (Eugene, OR). Water treated by a ultrafiltration system (Millipore Corp., MA) and distilled twice from a dilute alkaline permanganate solution was used for dynamic light scattering and fluorescence measurements, respectively. We employed Bangham’s method to prepare multilamellar vesicles containing fluorescence probe Prodan.25 The chloroform stock solution of a lipid was mixed with an ethanol solution of Prodan. The mixed solution was dried under vacuum to remove all residual solvents and finally to get a dry film. Water was added to the dry film, and the lipid samples were hydrated by a treatment of vortex. Subsequently, the suspensions were rotated in the gyratory shaker for ca. 8 h at a temperature above the main transition. We prepared two kinds of different-sized vesicles: one is an intact sample, which we call giant multilamellar vesicle (GMV), and the other is a sample with short-term sonication (0.5-3 min) by a water-bath-type sonicator (Branson 3510J-DTH with output 130 W) immediately after preparation, which we call large multilamellar vesicle (LMV). The mean diameters of LMVs and GMVs were 170 ( 15 nm and 5000 ( 800 nm, respectively, the values of which were determined by the dynamic light scattering (DLS) measurements.26 The size of GMVs was reduced to about 1/30 of the original size by short-term sonication, and the LMVs and GMVs were both multilamellar. Here, the GMV solution contained 10% LMV of ca. 550 nm diameter. We confirmed that the mean diameter of GMV was 4000-5000 nm by directly counting GMVs (at least 30 particles) using a differential interference microscope. The total concentration of a lipid was 1.0 mmol kg-1 and the molar ratio of Prodan to the lipid was 1:500. The sample solutions were protected from light until measurement. A method of the sample preparation for the DLS measurements was the same as that for the fluorescence measurements except for no addition of Prodan. We paid much attention to not contaminating sample solutions by dust. 2.2. Fluorescence Measurements. The fluorescence measurements under ambient and high pressures were carried out with an F-2500 fluorescence spectrophotometer (Hitachi HighTechnology Corp., Tokyo, Japan), which was equipped with a high-pressure cell assembly PCI-400 (Syn Corporation Ltd., Kyoto, Japan). Fluorescence spectra of Prodan in lipid bilayers of various phases under high pressure were observed by an isobaric thermotropic phase transition measurement described previously.20,27,28 Pressures were generated by a hand-operated KP-3B hydraulic pump (Hikari High Pressure Instruments, Hiroshima, Japan) and monitored within an accuracy of 0.2 MPa by using a Heise gauge. The temperature of the high-pressure cell was controlled within (0.1 °C by circulating water from a water bath through the jacket enclosing the measurement cell. The heating rate was 0.50 K min-1. The excitation wavelength was 361 nm, and the emission spectra were recorded in the wavelength range from 400 to 600 nm. The second derivative of the emission spectrum was obtained by attached software (FL-solutions) of the apparatus and Origin 7.0.

3. Results and Discussion 3.1. Differences in Bilayer Properties between LMV and GMV. The T-p phase diagram of the C17PC bilayer membrane, (25) Bangham, A. D.; DeGier, J.; Grevill, G. D. Chem. Phys. Lipids 1967, 1, 225– 246. (26) Takeda, K.; Okuno, H.; Hata, T.; Nishimoto, M.; Matsuki, H.; Kaneshina, S. Colloids Surf., B: Biointerfaces 2009, 72, 135–140. (27) Goto, M.; Ishida, S.; Tamai, N.; Matsuki, H.; Kaneshina, S. Chem. Phys. Lipids 2009, 161, 65–76. (28) Broniec, A.; Goto, M.; Matsuki, H. Langmuir 2009, 25, 11265–11268.

Langmuir 2010, 26(16), 13377–13384

Goto et al.

Figure 1. Temperature-pressure phase diagrams for C17PC bilayer membrane: (O) LMV, (b) GMV. A broken arrow in the diagram indicates the process of temperature elevation for measurements of fluorescence spectra in Figures 5(A) and 6(A).

which was constructed on the basis of the phase-transition data determined by the methods of DSC and high-pressure light transmittance,11 is depicted in Figure 1. In the phase diagram, closed symbols represent the data for GMV of the C17PC bilayer membranes, which are consistent with the transition data for LMV, already reported in our previous paper.11 Under ambient pressure, the C17PC bilayer membrane undergoes the pretransition from the Lβ0 phase to the Pβ0 phase at 42.9 °C and subsequently the main transition from the Pβ0 phase to the LR phase at 49.1 °C as the temperature rises. The temperatures of the pre- and main transitions increased with increasing pressure, and moreover, under high pressure above 79.1 MPa, the pressure-induced LβI phase appeared in the temperature range between the Lβ0 and Pβ0 phases. The temperature of the transition from the Lβ0 phase to the LβI phase initially reduced with increasing pressure up to ca. 120 MPa, but above this pressure, it elevated again with increasing pressure. As is clear from the phase diagram, no significant difference in the phase boundaries was observed between the GMV and LMV samples, meaning that the phase behavior of the C17PC bilayer membrane is hardly affected by the difference in vesicle size between GMV and LMV. The thermodynamic quantities of the pre- and main transitions for LMV and GMV of the C17PC bilayer membrane summarized in Table 1. These data were obtained from the DSC measurements and thermodynamic calculations.11 Although it is known that long-term sonication on multilamellar vesicles brings about the disappearance of the pretransition and decreases in main-transition properties,21,22 such as ΔH and the cooperativity of the transition, LMV and GMV used in this study show no significant difference in thermodynamic properties of the phase transitions. 3.2. Prodan Fluorescence for LMV and GMV under Atmospheric Pressure. For the Prodan-containing bilayer membranes of PCs,19,20,27 we have shown that fluorescence spectra of Prodan are greatly dependent on the phase states of the PC bilayer membranes, that is, the maximum emission wavelengths (λmax) drastically varies with their phase states: λmax of ca. 430 nm corresponds to the gel phase, ca. 480 nm to the LR phase, and ca. 500 nm to the LβI phase. Since Prodan is a polarity-sensitive fluorescent probe, the λmax values depend primarily on the dielectric constant of a solvent around the probe molecules at constant temperature and pressure despite that other factors, such as a change in microviscosity around the probe Langmuir 2010, 26(16), 13377–13384

Article

molecules, are known to affect the Prodan fluorescence spectra. The local dielectric constant or the water density along a lipid molecule in the bilayer decreases monotonously from the bilayer surface to the bilayer core, which has been demonstrated as the zone model.29,30 A higher value of λmax indicates that the probes exist in a polar or hydrophilic environment, while a lower value of λmax signifies that the probes exist in a less polar or hydrophobic environment. Therefore, the location of the Prodan molecules in the bilayer membrane of each phase can be reasonably estimated from the λmax values as a less polar region around glycerol backbone for the gel phase (i.e., ca. 430 nm), an intermediate region around phosphate group for the LR phase (i.e., ca. 480 nm), and a hydrophilic region around lipid headgroup for the LβI phase (i.e., ca. 500 nm). This means that the location of the Prodan molecules in the PC bilayer membrane is closely related to the phase state of the bilayer membrane. The fluorescence spectra of Prodan in LMV of the C17PC bilayer membrane were observed under atmospheric pressure as a function of temperature, which are shown in Figure 2A. Here the temperature dependence of the λmax value, which was extracted from the spectra, is also shown as an inset in the figure. During the process of the temperature elevation under atmospheric pressure, the C17PC bilayer membrane undergoes pre- and main transitions. It was observed that the fluorescence spectra notably changed with these phase transitions. The λmax value at 39.8 °C was found to be 434 nm, which corresponds to the Prodan spectrum for the Lβ0 phase. The λmax values were almost constant with increasing temperature and then suddenly increased at a temperature of the pretransition (43.7 °C) up to 442 nm, which corresponds to the Prodan spectrum for the Pβ0 phase. Further increase in temperature caused the shift of λmax from 442 to 477 nm at 50.6 °C due to the main transition. The spectrum with λmax of ca. 480 nm is caused by Prodan in the bilayer membrane of the LR phase. Here, we have measured the Prodan fluorescence also for the SUV with diameters of ca. 50 nm for comparison. In the case of SUV, only one peak at ca. 520 nm was observed in the fluorescence spectra, and even when the main transition occurred, only a slight increase of the intensity around 480 nm was observed (data not shown). This result is consistent with the previous results21,22 that the pretransition is abolished and the cooperativity of the main transition is reduced for SUV. Recently, we have developed the analysis method of employing the second derivatives of the fluorescence spectra to investigate the distribution behavior of Prodan in the bilayer membranes. The second-derivative spectra are useful for the peak separation of spectra including multiple peaks and mainly used in UV/vis spectroscopy. We applied the technique to fluorescence spectroscopy and determined the oriented positions of the Prodan molecules in each phase of the bilayer on the basis of the T-p phase diagram in Figure 1. The second-derivative spectrum can be often employed to detect other minor components included as shoulders in the original emission spectrum.20,27,28 The main and shoulder peaks in the original spectrum can be converted to separate valleys in its second-derivative spectrum, and at the same time, the λmax in the former spectrum is replaced by the wavelength (λ00 min) of minimum intensity in the latter spectrum. It should be noted that the value of λ00 min is not always exactly identical to that of λmax. The second-derivative curves are more useful than the original ones for the purpose of estimating the location of Prodan in the bilayer membrane in detail and much (29) Okamura, E. Nakahara, M. In Liquid interfaces in chemical, biological, and pharmaceutical applications; Volkov, A. G., Ed.; Marcel Dekker: New York, 2000; p 775. (30) Okamura, E.; Nakahara, M. Int. Congr. Ser. 2005, 1283, 203–206.

DOI: 10.1021/la100871z

13379

Article

Goto et al.

Table 1. Thermodynamic Properties of Phase Transitions for LMV and GMV of C17PC Obtained from DSC and Light-Transmittance Measurements sample

transition

temp./°C

temp./K

dT/dp/K MPa-1

ΔH/kJ mol-1

ΔS/J K-1 mol-1

ΔV/cm3 mol-1

Lβ0 /Pβ0 42.9 ( 0.16a 316.1 0.13b 3.8 ( 0.25a 12 1.6 49.1 ( 0.01a 322.3 0.224b 37.6 ( 0.61a 117 26.1 Pβ0 /LR 42.7 ( 0.20a 315.9 0.16c 3.4 ( 0.12a 11 1.7 GMV Lβ0 /Pβ0 49.0 ( 0.01a 322.2 0.235c 36.5 ( 0.12a 113 26.6 Pβ0 /LR a Each value is denoted as (average ( standard deviation) over 3 measurements. b Data from Ichimori et al.11 c Estimated from temperature-pressure phase diagram for GMV solution (Figure 1). LMV

Figure 2. (A) Fluorescence spectra of Prodan for LMV of C17PC bilayer membrane obtained at every 1 °C from 36.3 to 54.5 °C under atmospheric pressure. The inset shows temperature dependence of the maximum emission wavelengths of Prodan in the C17PC bilayer. The temperatures of phase transitions taken from the diagram (Figure 1) are indicated by arrows. (B) Secondderivative spectra of the fluorescence spectra given in (A).

more informative than generalized polarization31 for studies of bilayer packing. The second-derivative spectra for the emission spectra given in Figure 2A are shown in Figure 2B. We found at least three minima in the second-derivative spectra, wavelengths of which are 430 ( 1, 431 ( 1, and 484 ( 1, respectively. This finding indicates that the original spectra are composed of multiple components, that is, the Prodan molecules can be distributed into multiple sites in the bilayer membrane. Taking into consideration the phase sequence with increasing temperature, we determined that the wavelengths of the emission maxima characteristic of the three phases of the C17PC bilayer under atmospheric pressure, namely, the Lβ0 , Pβ0 , and LR phases, are 430, 431, (31) Parasassi, T.; Stasio, G. D.; d’Ubaldo., A.; Gratton, E. Biophys. J. 1990, 57, 1179–1186.

13380 DOI: 10.1021/la100871z

and 484 nm, respectively. It should be noted that the λ00 min values were determined to the respective phases which are already assigned in the phase diagram (Figure 1). Since a series of Prodan spectra obtained at different temperatures contains two independent variables, i.e., wavelength and temperature, we constructed a three-dimensional image plot (hereafter referred to as a 3-D image plot or simply an image plot) of the isobaric thermotropic change of the second derivatives of Prodan spectra shown in Figure 2B to clarify the change of the location of the Prodan molecules in the bilayer with temperature in detail. The resulting 3-D image plots for LMV and GMV of the C17PC bilayer membrane under atmospheric pressure are demonstrated in Figures 3(A) and 4(A), respectively. Here, the fluorescence intensity is drawn by a color scale as an indicator in these figures: the change from low intensity to high intensity is indicated by change of the color from blue to red in the order of the rainbow colors. The image plot can exhibit the change of λ00 min values much more sharply, which expresses the exact relationship between the fluorescence spectra and the bilayer phases. Then, we can clearly see the location of Prodan in the bilayer membrane as blue valleys in the image plots. The λ00 min value changed from 431 to 484 nm at ca. 50 °C for LMV and ca. 47 °C for GMV, which indicates that the bilayers of LMV and GMV underwent the main transition at those temperatures. This apparent difference of the main transition temperatures between LMV and GMV will be discussed later. Regarding the pretransition, a variation in width of a blue band near 430 nm at ca. 43 °C was observed for LMV, while they were hardly discernible for GMV at the pretransition. Comparing the image plot of GMV to that of LMV, we immediately notice the peculiar behavior in GMV. A clear green valley band corresponding to another minimum in the secondderivative spectra appeared in the wavelength range between ca. 500 and 525 nm unlike in the case of LMV. The fluorescence spectra for each gel phase have two local minima, that is, the wavelengths of the local minima characteristic of the Lβ0 phase are 430 and 520 nm and those of the Pβ0 are 431 and 508 nm. Since the λmax value of Prodan dissolved in water is 528 nm,19 these minima of 520 and 508 nm in the second-derivative spectra may result from the fact that the Prodan molecules are excluded into a more hydrophilic region at the membrane surface. It is only for GMV that the 520 nm or 508 nm minimum was observed simultaneously with the 430 nm minimum of the Lβ0 phase or the 431 nm minimum of the Pβ0 phase, respectively. This means that GMV in the Lβ0 or Pβ0 phase has at least two distinct major sites where the Prodan molecules can reside in contrast to LMV with one site. In LMV, it is easier for the Prodan molecules to distribute into the outer layers of the bilayer membrane than the inner layers in the gel phase because of difference in packing stress between outer and inner layers.5 Since it originates from the mismatch between the curvature of the membrane and its spontaneous curvature, the packing stress is relaxed with increasing vesicle size considering the critical packing parameter of C17PC. As a result, the lipid molecules in the bilayer can be packed more densely and tightly in GMV than in LMV, and a part of Prodan molecules are squeezed Langmuir 2010, 26(16), 13377–13384

Goto et al.

Article

Figure 3. (A) 3-D image plot for the second derivatives of fluorescence spectra of Prodan for LMV of C17PC bilayer obtained under atmospheric pressure in the temperature range 36.354.5 °C. (B) Ratio of intensity of the second-derivative spectra at 484 nm to that at 430 nm, F00 484/F00 430 versus temperature plot for LMV of C17PC bilayers under atmospheric pressure. The inset is magnifications of the low-value region of F00 484/F00 430. The phase transition temperatures, taken from the temperature-pressure phase diagram (Figure 1), are shown by arrows.

Figure 4. (A) 3-D image plot for the second derivatives of fluorescence spectra of Prodan for GMV of C17PC bilayer obtained under atmospheric pressure in the temperature range 36.354.5 °C. (B) Ratio of intensity of the second-derivative spectra at 491 nm to that at 430 nm, F00 491/F00 430 versus temperature plot for GMV of C17PC bilayers under atmospheric pressure. The inset is magnifications of the low-value region of F00 491/F00 430. The phase transition temperatures, taken from the temperature-pressure phase diagram (Figure 1), are shown by arrows.

out of the glycerol backbone region to move near to the bulkwater region due to the dense and tight packing. From the comparison of the shapes of the spectra for GMV with those for LMV, we can see that the depths of the minima corresponding to the three phases were more enhanced in GMV than in LMV. The enhancement of the minima is caused by an increase in fluorescence intensity. Taking into consideration that the Prodan molecules are distributed into a wide region from a region adjacent to the glycerol backbone to a region around the polar headgroup, the intensity increase may be attributed to the increase in quantity or in density of the Prodan molecules partitioned in the bilayer membrane per one vesicle due to the decrease in the total surface area of the vesicles. Since the partitioning of the Prodan molecules does not seem to change appreciably at the same molar ratio of Prodan and lipid, we speculate the latter case that the enhancement of the minima is attributable to the increase in surface density of the Prodan molecules per unit area in GMV as compared with that in LMV. Further, the blue valley corresponding to the Pβ0 phase for GMV shrunk in the direction of the temperature axis, while that

to the LR phase extended as compared with the case of LMV. This means that the main transition occurs at a lower temperature in GMV than in LMV. As described above, however, no difference was observed in the bilayer phase behavior between LMV and GMV. The shrinkage of the Pβ0 phase region or extension of the LR phase region may mean that the Prodan molecules in the Pβ0 phase distributed into the two locations in the GMV bilayer in/of the Pβ0 phase can transfer to a single location during the course of the phase transition from the Pβ0 phase to the LR phase more easily than those in the bilayer of LMV due to the tighter packing of the lipid molecules in the bilayer of GMV, which tends to squeeze them out; hence, the main transition apparently occurs more rapidly than LMV. It should be noted that this means that the transition temperatures cannot be determined accurately from the Prodan fluorescence spectra for GMV. In addition, judging from the difference in λ00 min value between LMV (484 nm) and GMV (491 nm), the Prodan molecules in the bilayer of the LR phase are located in a slightly more polar region in GMV than in LMV. This may indicate the slight difference of the exclusion effect of the Prodan molecules in the bilayer of the LR phase, which is caused

Langmuir 2010, 26(16), 13377–13384

DOI: 10.1021/la100871z

13381

Article

Figure 5. (A) 3-D image plot for the second derivatives of fluorescence spectra of Prodan for LMV of C17PC bilayer at 105.6 MPa in the temperature range 20.0-84.5 °C. (B) Ratio of intensity of the second-derivative spectra at 497 nm to that at 430 nm, F 00 497/F 00 430 versus temperature plot for LMV of C17PC at 105.6 MPa in the temperature range 20.0-84.5 °C. The inset is magnification of the low-value region of F 00 497/F 00 430. The phase transition temperatures, taken from the temperature-pressure phase diagram (Figure 1), are shown by arrows.

by the tighter packing of the phospholipid molecules with an increase in vesicle size. These results clearly show that the curvature of vesicles markedly affects the distribution of the Prodan molecules in the bilayer membrane of each phase, signifying that Prodan can be an indicator of the change of the packing stress or the curvature of the bilayer membrane instead of losing an ability to detect the bilayer phase transitions accurately. The bilayer phase transitions cause the migration of Prodan molecules between polar and less polar environments, which is reflected in the change of the λ00 min value. Therefore, a ratio of the intensities of the second-derivative spectrum (F 00 X) at two specific wavelengths characteristic of respective phases (i.e., X = 430 nm for the gel phase, and 484 or 491 nm for the LR phase) is available for us to determine the phase-transition temperatures. The ratios F 00 484/F 00 430 for LMV and F 00 491/F 00 430 for GMV are plotted against temperature in Figures 3B and 4B, respectively. In the figures, the phase transition temperatures are indicated by arrows, 13382 DOI: 10.1021/la100871z

Goto et al.

which were taken from the T-p phase diagram in Figure 1. The F 00 484/F 00 430 value for LMV increased slightly at the pretransition temperature of 42.8 °C. After a subsequent plateau, it showed an abrupt increase at the main transition temperature of 49.6 °C. Both transition temperatures agreed with those obtained from DSC measurements.11 As for GMV, unlike the case of LMV, the F 00 491/F 00 430 values did not show any change in the vicinity of the pretransition temperature and increased abruptly at 46.7 °C, which was not in accord with the main transition temperature. We usually use the plots of second derivative ratios of for the determination of phase transitions, especially under high-pressure condition (where DSC cannot be used). Here, we exhibited the plots to emphasize that the phase-transition determination by the Prodan fluorescence cannot be applied for GMV. The behavior may result from the facts that the Prodan molecules locate themselves around the glycerol backbone for the Lβ0 phase and hardly change their own location in the bilayer at the pretransition and also that they start to migrate to the region adjacent to the phosphate groups faster, as described above. These are mainly attributable to the tighter and denser packing of the lipid molecules in GMV than in LMV. Accordingly, we can conclude that it is surely difficult to observe the bilayer phase transitions accurately by the Prodan fluorescence as the vesicle size increases, but instead, it can be possible to obtain the information as to the change of the packing of the bilayer with vesicle size from the difference in the behavior of the Prodan fluorescence. 3.3. Prodan Fluorescence for LMV and GMV under High Pressure. The fluorescence spectra of Prodan in LMV and GMV of the C17PC bilayer membrane under high pressure were measured along the process of the temperature elevation indicated as a broken arrow in Figure 1, and the 3-D image plots for the second derivatives of them were constructed in the same manner as those for the observation under atmospheric pressure. Figures 5A and 6A show the resulting image plots for LMV and GMV under high pressure, respectively (also see Figure S1 inSupporting Information). Since the C17PC bilayer membrane undergoes the bilayer interdigitation, one more minimum (497 ( 1 nm for LMV, 501 ( 1 nm for GMV) characteristic of the nonbilayer LβI phase was found in the second-derivative spectra in addition to the minima characteristic of bilayer Lβ0 phase (430 ( 1 nm for LMV and GMV), Pβ0 phase (433 ( 1 nm for LMV and GMV), and LR phase (480 ( 1 nm for LMV, 482 nm for GMV).20 We can clearly see four blue valleys corresponding to these minima in both image plots for LMV and GMV, meaning that the Prodan molecules can be distributed into the specific locations according to the phase state on the high-pressure condition as well. The difference in the Prodan fluorescence between LMV and GMV observed under high pressure was essentially the same as the difference observed under atmospheric pressure except for the difference as to the LβI phase. Namely, (1) a green valley band corresponding to another local minimum (518 nm) was found in the image plot for GMV below ca. 45 °C, indicating that, in the case of GMV of the Lβ0 phase, the Prodan molecules can be distributed into not only a region adjacent to the glycerol backbone, but also a region near the bulk water also under high pressure. The local minimum at 518 nm gradually shifted to ca. 500 nm with increasing temperature from 39.6 to 49.8 °C, while the minimum local of 430 nm jumped to ca. 500 nm suddenly at the temperature of the Lβ0 /LβI transition. This may suggest that the Prodan molecules distributed around the bilayer surface region are not so restricted and thus have higher mobility than those in the glycerol backbone region. (2) Each minimum of GMV became deeper as compared with that of LMV, indicating Langmuir 2010, 26(16), 13377–13384

Goto et al.

Article

Figure 7. Fluorescence spectra of Prodan in (A) LMV and (B) GMV of C17PC bilayer membrane obtained at 52.0 °C and 151.9 MPa. It took 15 min to complete each measurement. The insets show the second derivatives of these fluorescence spectra.

Figure 6. (A) 3-D image plot for the second derivatives of fluorescence spectra of Prodan for GMV of C17PC bilayer at 119.9 MPa in the temperature range 20.0-87.7 °C. (B) Ratio of intensity of the second-derivative spectra at 501 nm to that at 430 nm, F 00 501/F 00 430 versus temperature plot for GMV of C17PC at 119.9 MPa in the temperature range 20.0-87.7 °C. The inset is magnifications of the low-value region of F00 501/F00 430. The phase transition temperatures, taken from the temperature-pressure phase diagram (Figure 1), are shown by arrows.

that the surface density of the Prodan molecules per unit area in GMV may be greater than that in LMV. (3) The λ00 min value of the LR phase changed from 484 nm for LMV and 491 nm for GMV under atmospheric pressure to 480 nm for LMV and 482 nm for GMV under high pressure, respectively, although the λ00 min values of the gel phases were scarcely affected by pressure for both LMV and GMV. As described above, the difference in λ00 min between LMV and GMV under atmospheric pressure probably means that the Prodan molecules can reside in a more polar region for GMV than for LMV because of the exclusion effect by the tighter packing. A similar tendency was observed also under high pressure, but evidently, the difference in λ00 min between LMV and GMV became smaller under high pressure. This may imply that the high pressure almost cancels out the exclusion effect. It can be said from these facts that the bilayer of the LR phase can exquisitely include or exclude another species of small molecule in response to the change in pressure and vesicle size, and Prodan has a high ability to detect such small changes of the microstructure of the bilayer membrane. The ratio of the intensities of the second-derivative spectra at two specific wavelengths was examined to evaluate the temperatures of Langmuir 2010, 26(16), 13377–13384

the bilayer phase transitions under high pressure, especially the transitions relating to the pressure-induce LβI phase.20,27 We selected the λ00 min values of the Lβ0 phase and the LβI phase as the two specific wavelengths for this purpose, that is, 430 and 497 nm for LMV, and 430 and 501 nm for GMV. The plot of F 00 497/ F 00 430 against temperature for LMV and that of F 00 501/F 00 430 against temperature for GMV are depicted in Figures 5B and 6B, respectively. The temperatures of the three inflection points observed in the curve for LMV were in good agreement with the phase transition temperatures along the heating process given in Figure 1. This method of determining the phase transition temperatures was proven to be superior for LMV to the method by using the ratio of two λmax values, which were proposed previously.20,27 Contrary to this, the temperatures of the three inflection points obtained from the curve for GMV were not consistent with the transition temperatures at all, although the F 00 501/F 00 430 ratios for GMV had much higher values as compared to the corresponding F 00 497/F 00 430 values for LMV due to the increase in distribution density of Prodan in the bilayer. We can say that the Prodan fluorescence provides detailed information about the packing of the bilayer membranes, which is altered in response to the changes of the temperature, pressure, and vesicle size, although the accurate information as to the bilayer phase transitions is gradually lost with increasing the vesicle size. 3.4. Effect of Vesicle Size on the Prodan Fluorescence in the LβI Phase. Since one characteristic feature of the saturated PC bilayer membranes under high pressure is the appearance of the pressure-induced LβI phase, it is interesting how the vesicle size affects the fluorescence spectra of Prodan in the bilayer membrane of the LβI phase. The emission spectra of Prodan for LMV and GMV at 52.0 °C and 151.9 MPa, at which the membrane state is the LβI phase, are shown in Figure 7 together DOI: 10.1021/la100871z

13383

Article

Goto et al.

the packing state in the formation of the LβI phase and become a useful indicator in examining the behavior of the LβI phase in more detail.

4. Conclusions

Figure 8. Time dependence of wavelengths of local minima (λ00 min) in the second derivatives of fluorescence spectra of Prodan in LMV (open circle) and GMV (closed circle) of C17PC bilayer membrane obtained at 58.7 °C and 147.0 MPa.

with the corresponding second derivative spectra. Here, a longer sonication time (ca. 10 min) than the time conventionally used in our previous studies was adopted in the preparation of the sample of LMV to clarify the effect of the sonication on the fluorescence spectrum. The peak at 497 nm for LMV originates from the bilayer interdigitation. From the figure, we found that the fluorescence intensity curve has a broad peak, which consists of two peak components centered at 430 and 497 nm, as is evident from its second-derivative spectrum, indicating that, in LMV of the LβI phase, Prodan molecules can be distributed around the glycerol backbone region and headgroup region. On the other hand, the fluorescence spectra of Prodan for GMV showed one peak at 501 nm with a shoulder peak at 430 nm. This indicates that most Prodan molecules are squeezed out of the glycerol backbone region due to the stronger packing of the bilayer membrane: GMV with a small curvature does not appreciably produce enough space around the glycerol backbone for Prodan molecules to be partitioned. Therefore, the enhancement of the peak at 430 nm in LMV probably indicates that the bilayer interdigitation in LMV produces more space between adjacent PC molecules than that in GMV, which enables the Prodan molecules to remain around the glycerol backbone even in the LβI phase. Furthermore, judging from the λ00 min values between 497 nm (LMV) and 501 nm (GMV), the Prodan molecules are pushed out into a more hydrophilic region in GMV than in LMV. Figure 8 shows the time dependence of the λ00 min values for LMV and GMV under the condition of the LβI phase at 58.7 °C and 147.0 MPa. With increasing time, the λ00 min values for LMV shifted slightly toward the longer wavelength region (501.5 nm), whereas those for GMV shifted greatly to 512.0 nm. This behavior was observed only for the LβI phase and not for the other phases. The large shift for GMV implies that the Prodan molecules are being further squeezed out of the monolayer in the LβI phase with time. The more the vesicle sizes increase, the flatter the structure of the bilayer membranes becomes. Then, the packing of the bilayer membrane tends to become tighter and denser. Therefore, we speculate that an increase in the λ00 min value is responsible for the Prodan molecules continuously being squeezed out of the monolayer with the growth of the flat structure of the monolayer. This effect is small in LMV but appreciably emphasized in GMV as expected from the critical packing parameter. These results clearly suggest that Prodan can also detect the time-dependent change of

13384 DOI: 10.1021/la100871z

The curvature of vesicles affects membrane properties of lipid bilayer membranes; however, it is difficult to evaluate the effect of size on the bilayer properties quantitatively. The phase transitions and thermodynamic quantities of the transition did not exhibit any differences between LMV and GMV in this study. However, we could obviously reveal that the fluorescence spectra of Prodan in the bilayer membrane are affected by not only the phase state of the bilayer membrane, but also the size of the vesicle. This indicates that the Prodan molecules in the bilayer membranes change their location depending on the size of vesicles, as well as on the phase states. From the analysis of the spectra by employing its second derivative, we clarified the distribution of the Prodan molecules in the bilayer membrane in detail and the change of the distribution according to the bilayer phase states. In the case of LMV, the change of the distribution was in good accordance with the bilayer phase states: the Prodan molecules are distributed mainly into a less polar region around the glycerol backbone for the gel phase, an intermediate region around the phosphate group for the LR phase, and a hydrophilic region around the polar headgroup for the LβI phase. As for GMV, on the other hand, the Prodan molecules are distributed into the two specific regions in the bilayer membrane of the gel phases: the regions adjacent to the glycerol backbone and near the bulk water. The presence of the Prodan molecules in the region near the bulk water is thought to be caused by the exclusion effect of the bilayer membrane due to the tighter and denser packing of the lipid molecules. When the bilayer phase state turns into the LR or LβI phase, the Prodan molecules around the glycerol backbone region are easily pushed out to a more hydrophilic region. The change of the Prodan fluorescence observed for GMV was not consistent with the bilayer phase transition. Further, we clarified the difference between LMV and GMV in the formation of the pressure-induced LβI phase from the time-dependent change of the Prodan fluorescence spectra. We can conclude that the fluorescent probe Prodan has a higher sensitivity to detect the alteration of the microstructure of the bilayer membranes, such as small changes of the packing stress and the membrane curvature, although it loses the ability to monitor the bilayer phase transition accurately with increasing vesicle size. This suggests the possibility that the Prodan fluorescence can be utilized to investigate the microstructure of the bilayer membranes in further detail. Supporting Information Available: Figure S1. (A) The second-derivative spectra for LMV solution of C17PC at 105.6 MPa in the temperature range 20.0-84.5 °C. The insets show the extraction of second derivative spectra for the LMV solution of C17PC: (1) 20.0 °C, (2) 30.2 °C, (3) 55.3 °C, (4) 72.1 °C, (5) 84.5 °C. (B) The second-derivative spectra for GMV solution of C17PC at 119.9 MPa in the temperature range 20.0-87.7 °C. The insets show the extraction of second derivative spectra for the SMV solutions of C17PC: (1) 20.0 °C, (2) 30.3 °C, (3) 55.4 °C, (4) 74.0 °C, (5) 87.7 °C. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(16), 13377–13384