Calcium-Induced Formation of Subdomains in

4648

J. Phys. Chem. B 2009, 113, 4648–4655

Calcium-Induced Formation of Subdomains in Phosphatidylethanolamine-Phosphatidylglycerol Bilayers: A Combined DSC, 31P NMR, and AFM Study Laura Picas,† M. Teresa Montero,†,‡ Antoni Morros,§,| Miquel E. Cabanãs,⊥ Bastien Seantier,# Pierre-Emmanuel Milhiet,#,∇ and Jordi Hernańdez-Borrell*,†,‡ Departament de Fisicoquı´mica, Facultat de Farma`cia, UniVersitat de Barcelona (UB), and Institut de Nanocie`ncia i Nanotecnologia de la UniVersitat de Barcelona (IN2UB), E-08028 Barcelona, Spain, Unitat de Biofı´sica, Departament de Bioquı´mica i Biologia Molecular, Facultat de Medicina, Centre d’Estudis en Biofı´sica (CEB), and SerVei de Ressona`ncia Magne`tica Nuclear (SeRMN), UAB, E-08193 Bellaterra, Barcelona, Spain, and Inserm, Unite´ 554, Montpellier, France, and Centre de Biochimie Structurale, UniVersite´ de Montpellier, CNRS, UMR 5048, Montpellier, France. ReceiVed: NoVember 21, 2008; ReVised Manuscript ReceiVed: January 26, 2009

We study the effect of Ca2+ on the lateral segregation of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) (3:1, mol/mol). Supported lipid bilayers (SLBs) were observed by atomic force microscopy (AFM). Since SLBs are formed from liposomes of POPE:POPG, we examined the effect of calcium on these suspensions by differential scanning calorimetry (DSC) and 31P nuclear magnetic resonance spectroscopy (31P NMR). AFM images revealed the existence of two separated phases, the higher showing a region with protruding subdomains. Force spectroscopy (FS) was applied to clarify the nature of each phase. The values of breakthrough force (Fy), adhesion force (Fadh), and height extracted from the force curves were assigned to the corresponding gel (Lβ) and fluid (LR) phase. The endotherms obtained by DSC suggest that, in the presence of Ca2+, phase separation already exists in the suspensions of POPE:POPG used to form SLBs. Due to the temperature changes applied during preparation of SLBs a 31P NMR study was performed to assess the lamellar nature of the samples before spreading them onto mica. With in situ AFM experiments we showed that the binding of Ca2+ to POPG-enriched domains only induces the formation of subdomains in the Lβ phase. 1. Introduction The physiological activity of transmembrane proteins may be influenced by or depend on the physicochemical properties of the neighboring phospholipids. According to the matching principle,1 such phospholipids provide a hydrophobic environment suitable to host the protein in its natural folded state. Thus, lactose permease (LacY) from Escherichia coli (E. coli) requires, at least, phosphatidylethanolamine (PE) for appropriate folding and activity.2-4 Although LacY has been reconstituted into liposomes or crystallized in two dimensions with zwitterionic phospholipids,5 it is normally reconstituted in a biomimetic environment such as 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleoyl-sn-glycero3-phosphoglycerol (POPG) or the total lipid extract of E. coli membranes.6 In previous studies, we used atomic force microscopy (AFM) to reveal the nanostructure of the supported lipid bilayers (SLBs) of POPE:POPG (3:1, mol/mol) and we observed laterally segregated domains.7 These SLBs were obtained by spreading liposomes on a mica substrate, using a buffer containing 0.19 * To whom correspondence should be addressed. [email protected]. † Departament de Fisicoquı´mica, UB. ‡ IN2UB. § Departament de Bioquı´mica i Biologia Molecular, UAB. | CEB, UAB. ⊥ SeRMN, UAB. # Inserm. ∇ Universite´ de Montpellier, CNRS, UMR 5048.

mM CaCl2. Ca2+ favors not only the extension of SLBs but also interaction between phospholipid species.8 Although Ca2+ improves the spreading of SLBs on mica,9 the concentration used was higher than physiological levels.10 This could affect domain formation and, eventually, protein insertion. Bacterial membrane domains can be studied by fluorescence11 and AFM.12 The role of domains in the membrane physiology of both eukaryotic and prokaryotic cells has now been recognized.13 However, there is little information on the physicochemical properties that induce their formation. In the present study, before attempting protein insertion, we studied the effect of Ca2+ on the lateral segregation of POPE and POPG by observing the SLBs with AFM and their nanomechanical properties using force spectroscopy (FS). The SLBs are formed from liposomes of POPE:POPG (3:1, mol/ mol); therefore, we examined the effect of calcium on these suspensions by differential scanning calorimetry (DSC) and 31P nuclear magnetic resonance spectroscopy (31P NMR). Since the lateral separation of phospholipid components in different domains may affect the insertion of proteins such as LacY into the SLBs of POPE:POPG (3:1, mol/mol), it is important to study the influence of Ca2+ on the origin of these domains.

E-mail:

2. Materials and Methods 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), specified as 99% pure, and 1-palmitoyl-2-oleoyl-sn-glycero-3phosphoglycerol (POPG) were purchased from Avanti Polar Lipids (Alabaster, AL). Buffer A consisted of 20 mM Hepes

10.1021/jp8102468 CCC: $40.75  2009 American Chemical Society Published on Web 03/10/2009

Subdomain Formation in POPE-POPG Bilayers (pH 7.40) and 150 mM NaCl; buffer B, of 20 mM Hepes (pH 7.40), 150 mM NaCl, and 10 mM CaCl2; buffer C, of 20 mM Hepes (pH 7.40), 150 mM NaCl, and 20 mM CaCl2; and buffer D, of 20 mM Hepes (pH 7.40), 150 mM NaCl, and 10 mM ethylenediaminetetraacetic acid (EDTA). All buffers were prepared in ultrapure water (Milli Q reverse osmosis system, 18.3 MΩ · cm resistivity). HPLC-grade chloroform and methanol and n-dodecyl-β-D-maltoside (DDM) were purchased from SIGMA (St. Louis, MO). 2.1. Differential Scanning Calorimetry. Multilamellar vesicles (MLVs) for DSC studies were prepared at 50 °C in 20 mM Hepes buffer and 150 mM NaCl, pH 7.4 (buffer A), in the absence or presence of 10 mM CaCl2 to a final concentration of 2.74 mM. A MicroCal MC-2 calorimeter was used for DSC analyses, following procedures described elsewhere.14,15 The data were analyzed by the original calorimeter software. The temperature of maximum excess specific heat (Tm) was measured to the nearest 0.5 °C. Transition enthalpy was calculated in line with previously described methods.16 The calorimetry accuracy for Tm and for enthalpy changes was (0.1 °C and (0.2 kcal · mol-1, respectively. Each sample was scanned in triplicate over the temperature range 5-80 °C at a scan rate of 0.44 °C · min-1. Tm and the intensities of the overlapping components of the DSC endotherm in the presence of 10 mM Ca2+ were assessed as follows. The second derivative, calculated using the SavitskyGolay procedure (third-grade polynomial, 27° of smoothing), and Fourier self-deconvolution (γ factor, 0.3) were used to determine the minimum number of components and to locate their peak positions by means of the program GRAMS/32 (Galactic Industries). The curve was fitted by using the four parameters of the Weibull curve, a built-in function of the Sigma Plot 7.0 program (SPSS Inc.) used to quantify greatly overlapping asymmetrical chromatographic peaks.17,18 The sum of two Weibull functions was fitted to the profile of the complex endotherm by least-squares iteration. 2.2. 31P NMR Spectroscopy. MLVs for 31P NMR studies were prepared at 50 °C in 20 mM Hepes buffer, 150 mM NaCl, and 10 mM CaCl2, pH 7.4 (buffer B), to a final concentration of 2.74 mM. A 25 mL aliquot of the resulting unsonicated MLV suspension was pelleted by ultracentrifugation at 115000g for 1 h at 10 °C. The hydrated pellet was then resuspended in 300 µL of supernatant and placed in a conventional 5 mm NMR tube. A capillary tube containing 2H2O was added for fieldfrequency stabilization and the NMR tube was sealed after a purge with N2. 31 P NMR spectra were recorded as previously described19 on a Bruker AvanceIII-400 spectrometer (Bruker Espanõla, S.A., Madrid, Spain) operating at 161.98 MHz using a single 90° pulse sequence, with proton-decoupling during signal sampling by means of a Waltz-16 composite pulse sequence.20,21 A single pulse sequence was used instead of the phase-cycled Hahn echo pulse sequence22 to obtain spectra with higher signal-to-noise ratios. Each spectrum was the result of accumulating 3072 scans sampled using 2048 complex data points, with a 90° pulse of 9 µs (Beff ) 27.78 kHz), an interpulse delay of 2.1 s, and a spectral width of 50 kHz. Exponential multiplication resulting in a line broadening of 50 Hz was applied before Fourier transformation to improve the signal-to-noise ratio.23 Spectra were processed on a personal computer running the TopSpin v. 2.0 software (Bruker Biospin GmbH, Germany) on Debian GNU/Linux v. 3.1. All chemical shift values are quoted in parts per million (ppm) with reference to external 85% phosphoric acid in H2O (0 ppm), with positive values referring to low-field shifts.

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4649 2.3. SLB Formation. SLBs were prepared according to a method described elsewhere.24 Briefly, large unilamellar vesicles (LUVs) were prepared by extrusion of MLVs of POPE:POPG in 20 mM Hepes buffer, 150 mM NaCl, and 10 mM CaCl2, pH 7.4 (buffer B), through 100 nm pore filters (Nucleopore). LUVs were deposited onto freshly cleaved mica disks mounted on a Teflon O-ring and incubated at 50 °C for 2 h. The bilayers were always kept in an aqueous environment and carefully rinsed with the same buffer before imaging. 2.4. AFM Imaging. AFM experiments were performed with a multimode microscope controlled by Nanoscope IIIa electronics (Digital Instruments, Santa Barbara, CA). Images were acquired in tapping mode (TM-AFM) at minimum vertical force, maximizing the amplitude setpoint value and maintaining the vibration amplitude as low as possible. V-shaped Si3N4 cantilevers (MLCT-AUNM, Veeco) with a nominal spring constant of 0.10 N · m-1 were used in liquid operation. 2.5. Force Spectroscopy. Force spectroscopy experiments were carried out in liquid media using V-shaped Si3N4 tips (MLCT-AUNM) with a nominal spring constant of 0.10 N · m-1. However, individual spring constants were calibrated using the thermal noise method.25 Typically, 300 force plots were performed and recorded, with the laser spot maintained in the same position on the cantilever in order to keep the corresponding photodetector sensitivity constant (V · nm-1).26 All spectroscopy experiments were performed at a constant cantilever linear velocity of 0.5 µm · s-1 in order to avoid any velocity-dependent effect. Applied forces F are given by F ) kc∆, where ∆ is the cantilever deflection and kc is the cantilever spring constant. 3. Results and Discussion Divalent cations are normally used to enhance SLB formation.9,27 Thus, both adsorption and rupture of liposomes on a mica support are favored by the presence of Ca2+, and they are both specific requirements for screening the charge of mica for those samples containing negatively charged phospholipids.28 Both Ca2+ and Mg2+ neutralize the charge of the mica surface and may form bridges with anionic phospholipids,28 inducing major reorganization in charged and zwitterionic phospholipids.29 Such interactions can modify the behavior of phospholipids at the surface, leading to phase separation,30 or they can increase the nanomechanical resistance of the membrane.9 Figure 1 illustrates the effect of Ca2+ on the formation of SLBs of POPE:POPG (3:1, mol/mol). Concentrations of CaCl2 ranging from 0 to 20 mM were used. As can be seen, no planar structures are formed either in the absence (Figure 1A) or in the presence of 5 mM (Figure 1B) CaCl2, and only phospholipid aggregates and nonfused liposomes are found. As reported elsewhere,9 vesicles may remain intact on the substrate, yielding supported vesicular layers. Thus, once a critical vesicular coverage is reached, vesicle rupture and the formation of bilayer patches are favored, promoting interaction with adjacent lipid material and leading to the formation of extended bilayer patches. This can be seen from Figure 1A,B, where the area formed by lipid patches in the absence of calcium is almost negligible, but it increases substantially with 5 mM of CaCl2. At 10 (Figure 1C) and 20 (Figure 1D) mM CaCl2 a complete coverage of the substrate is achieved. Judging only from topography, these two images could be interpreted as continuous planar bilayers constituted by two separated phases: the darker and the brighter regions. The step height differences between the darker and brighter regions in Figure 1C,D can be established by line profile analysis

4650 J. Phys. Chem. B, Vol. 113, No. 14, 2009

Picas et al.

Figure 1. TM-AFM images acquired after spreading POPE:POPG (3:1, mol/mol) liposomes in 20 mM Hepes, 150 mM NaCl (pH 7.40) buffer supplemented with the following: 0 (A), 5 (B), 10 (C), and 20 mM (D) CaCl 2. The z color scale is 20 nm, and the scale bar is 2 µm. A cross-section taken along the white line in images C and D is presented beneath in the topographic images (E and F, respectively).

Figure 2. TM-AFM images showing the in situ effect of adding DDM at 2(CMC) to an SLB of POPE:POPG (3:1, mol/mol) in the presence of 10 mM CaCl2 (A). After adding the detergent, successive images (B and C) were obtained. The z color scale is 20 nm, and the scale bar is 1 µm.

as 3.5 ( 0.5 nm (n ) 10; Figure 1E) and 3.25 ( 0.25 nm (n ) 10; Figure 1F), respectively. It is worth mentioning that the brighter region contains subdomains that protrude 0.50 ( 0.07 nm (n ) 10). These height measurements give rise to three hypotheses: (i) SLBs are constituted only by the brighter regions that cover a small part of the substrate (dark background); (ii) the images show a continuous bilayer (dark region) with another bilayer fragment superimposed (brighter region); or, alternatively, (iii) as phospholipids are thermotropic and show welldefined phase transition behavior from gel (Lβ) to fluid (LR) phases, the images reflect a single continuous bilayer that displays the Lβ and the LR phases. To study the nature of the background region and confirm or refute its lipid nature, a DDM solution was injected into the AFM fluid cell. Figure 2 shows the topographic images before (A) and after (B and C) detergent injection. As can be seen,

while the brighter regions (see white arrows in Figure 2A,B) are clearly disrupted by DDM, the effect on the darker region is much more discrete. Remarkably, the solubilization effect on the background region becomes more significant with time (Figure 2C). Hence, we conclude that hypothesis (i) is false and that the dark background in Figures 1C,D and 2 is indeed lipid material, most probably an SLB that covers the substrate. However, the heights in Figure 1E,F for the brighter regions compare well with the size obtained by molecular dynamics calculations31 and may suggest that we have in fact obtained a continuous bilayer (darker region) with a bilayer structure superimposed32 (brighter region). Nevertheless, under our experimental conditions, the step height values between the regions may be overestimated as a result of the repulsion between the tip and the negative charge conferred by POPG on the bilayer. This effect has previously been observed in other systems33 and would tentatively lead us to exclude hypothesis (ii) in favor of hypothesis (iii). Thus, parts C and D of Figure 1 would reflect a single continuous bilayer that displays the Lβ and LR phases at a certain stage of phase transition. To further study the nature of these two lipid regions and to unambiguously assign the corresponding phase to the brighter or darker regions, force curve analysis was performed. FS has previously been applied to the characterization of phase separation of phospholipid monolayers and bilayers,34-36 and it can provide quantitative information on the nanostructure and nanomechanics of two separated lipid phases.37 Figure 3A shows a continuous SLB of the lipid mixture at 10 mM calcium concentration, and parts B and C of Figure 3 are two examples of the vertical force versus piezo displacement curve (Fν-D) from the highlighted darker and brighter regions, respectively. The first information we obtain comes from the approaching part of the FS curves (see magnifications in Figure 3D,E). Thus, when the gradient of the van der Waals attractive forces exceeds the gradient of the tip spring constant and the repulsive forces, the tip jumps onto the surface.38 In our FS experiments, while this jump to contact can be observed when the tip approaches the Lβ phase (arrow in Figure 3D), it is no longer seen when it approaches the LR phase (arrow in Figure 3E). Therefore, we can assume that a certain electrostatic repulsion occurs when the tip approaches the Lβ phase, which suggests that this region is actually enriched in the negatively charged phospholipid, POPG. This provides an explanation of the overestimation of the height of the brighter regions reported above and supports hypothesis iii. As reported elsewhere,33 and in further support of this interpretation, deviations in height determination should be attributed to differences in surface charge densities that would affect the electrostatic double-layer

Subdomain Formation in POPE-POPG Bilayers

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4651

Figure 3. TM-AFM image of an SLB of POPE:POPG (3:1, mol/mol) acquired in the presence of 10 mM CaCl2 (A). The z color scale is 15 nm, and the scale bar is 2 µm. The pattern of the Fν-D curves in the highlighted areas is depicted in B and C, corresponding to the liquid (darker region) and gel (brighter region) phases. The approach region of the curve is depicted in blue, whereas the retracting of the cantilever is presented in red. A magnification of the jump-to-contact region of approach Fν-D curves in images B and C is shown in D and E, respectively.

Figure 4. Histograms corresponding to the experimental Fy (A), Fadh (B), and bilayer height (C) values obtained for the two lipid phases studied: LR in gray and Lβ in black.

TABLE 1: Mean Values of Fy, Fadh, and Bilayer Height Obtained from Experimental Fν-D Curves for the Lr (Fluid) and Lβ (Gel) Phases of a Continuous SLB of POPE:POPG (3:1, mol/mol) lipid phase

Fy (nN)

Fadh (nN)

bilayer height (nm)

LR Lβ

0.243 ( 0.070 0.915 ( 0.190

1.580 ( 0.125 0.243 ( 0.060

3.300 ( 0.184 4.440 ( 0.440

repulsion between the tip and the sample. In fact, more hydrated cations, such Ca2+ or Mg2+, could exert repulsive forces when adsorbed to the surfaces.38 In fact, Ca2+ bonds specifically to POPG,39 promoting the formation of solid “clusters” that coexist within the bulk fluid lipid phase. Hence, the repulsion between the tip and the Lβ phase could result from these hydration forces. Fν-D curves provide the threshold force (Fy), that is, the force required to indentate a bilayer down to the substrate.37 The values of Fy obtained from 250 Fν-D plots are shown in histogram A in Figure 4. As can be seen, two mean Fy values are found at 0.243 ( 0.070 and 0.915 ( 0.190 nN (Table 1), which can be attributed to the LR and Lβ phases, respectively. Since two subdomains are present in the Lβ phase, a wider distribution of the Fy values is observed. The mean Fy values found for both the LR and Lβ phases are slightly smaller than those reported for other SLBs;8,40-42 however, they indicate that the two lipid phases exhibit different nanomechanical behavior.

The adhesive force (Fadh) can be determined from the magnitude of the pull of jump in the retraction portion of the Fν-D curves34-36 (see arrows in Figure 3B,C). The magnitude of Fadh for the LR and Lβ phases and their frequencies are displayed in histogram B in Figure 4. The mean value of Fadh was 1.580 ( 0.125 and 0.243 ( 0.060 nm for the LR and Lβ phases, respectively (Table 1). These results are in agreement with the observation that the fluidity of a bilayer is proportional to the adhesion force observed.43 Furthermore, the value of Fadh for the LR phase is similar to the value reported for POPS,44 a heteroacid phospholipid that, like POPG, binds Ca2+ and which is in fluid phase at room temperature. Finally, by representing Fy versus the tip-sample distance, we can obtain the depth of penetration and, hence, estimate the bilayer height.37,45 The height values obtained by this method are plotted in histogram C in Figure 4. The mean height values were 3.300 ( 0.184 and 4.400 ( 0.440 nm for the LR and Lβ phases, respectively (Table 1). Thus, the step height difference between the phases that results from FS curves is lower than that obtained from line profile analysis (Figure 1E,F) and is in agreement with the step height differences between fluid and solid phases found in mixtures of other phospholipids. This result provides an additional explanation of the overestimation of the height of the brighter regions reported above and reinforces the interpretation of the images given in hypothesis iii. The whole of the FS


Picas et al. TABLE 2: Thermodynamic Parameters of the Phase Transitions of POPE:POPG (3:1) Mixtures (Mean ( SEM)a sample

n

Tm (°C)

∆T (°C)

∆H (kJ · mol-1)

absence of Ca2+ 10 mM Ca2+

4 2 2

20.0 ( 0.6 23.5 ( 0.1 (61.1 ( 1.1)

4.5 ( 0.2 2.8 ( 0.2 (5.8 ( 0.5)

19.8 ( 1.2 17.5 ( 0.4 (1.7 ( 0.2)

a The thermodynamic parameters of the bilayer to non-bilayer phase transition are shown in brackets.

Figure 5. Normalized DSC thermograms of POPE:POPG (3:1, mol/ mol) multilamellar liposomes in the absence of Ca2+ (lower line) and with 10 mM CaCl2 (upper line) added to 20 mM Hepes buffer, 150 mM NaCl, pH 7.4. Only the first heating scan is shown for each study, as there were no significant differences in the second and third scans. A 10× magnified profile is included as an inset to show the bilayer to nonbilayer phase transition in the presence of 10 mM Ca2+. The total phospholipid concentration was 2.0 mg · mL-1 and the heating scan rate was 0.44 °C · min-1 (A).Resolution enhancement of the main phase transition profile in the absence of Ca2+ (B, D, and F) and in the presence of 10 mM Ca2+ (C, E, and G). Parts B and C show the expanded experimental trace of the main transition (continuous line) and, in the presence of 10 mM Ca2+ (C), the best fit (dotted lines) with a sum of two Weibull asymmetrical peaks (broken lines). Parts D and E show the second derivatives; and F and G show the Fourier selfdeconvoluted traces.

results support the idea that the SLBs shown are a single bilayer displaying the LR and Lβ phases. To study whether the nanostructure of the POPE:POPG SPBs originates on the mica surface or whether it already exists in the liposomes used to prepare the SLBs, we performed DSC and a 31P NMR study of the POPE:POPG (3:1, mol/mol) multilamellar liposomes. Figure 5A shows the normalized DSC endotherms for this phospholipid mixture in the absence (lower trace) and in the presence (upper trace) of 10 mM Ca2+. Table 2 shows values of several thermodynamic parameters of the observed phase transitions: the temperature of maximal excess heat capacity (Tm, °C), the peak width at half-height (∆T, °C) and the transition enthalpy change (∆H, kJ · mol-1). In the absence of Ca2+, only one phase transition appeared in the temperature

Figure 6. Solid-state 31P NMR spectra of POPE:POPG (3:1, mol/mol) multilamellar liposomes in 10 mM CaCl2 added to 20 mM Hepes buffer, 150 mM NaCl, pH 7.4. Spectra are shown at full scale. All chemical shift values are quoted in parts per million with reference to external 85% phosphoric acid in H2O (0 ppm), with positive values referring to low-field shifts. By comparison with the DSC thermogram (see Figure 5), these spectra can be assigned as follows: 5 and 10 °C, gel phase; 23 °C, main phase transition; 30 °C, liquid-crystal phase; 62 and 65 °C, very low enthalpy second endotherm.

range studied, corresponding to the Lβ-to-LR main phase transition; its Tm occurred at 20.0 °C. The heat capacity curve is asymmetrical, skewed toward the low-temperature side. However, in the presence of 10 mM Ca2+ this main transition occurred at 23.5 °C and the peak was significantly narrower, whereas the enthalpy change was only slightly smaller than it was in the absence of the divalent cation (Table 2). We may conclude that the binding of Ca2+ to POPG induces a reduction in its surface charge and area and causes tighter packing of the lipid lattice, thus increasing the phase transition temperature. In addition to the major transition mentioned, in the presence of 10 mM Ca2+ a very low enthalpy second endotherm also appeared with a peak at 61.1 °C. The shape of the main transition peak in the presence of Ca2+ was characteristic of a complex endotherm. This indicates a phase separation in the POPE:POPG mixture. Fitting the results to two asymmetrical curves (Weibull functions; see Materials and Methods) skewed to low temperatures yielded a much better fit (Figure 5C) than that obtained with symmetrical curves (Gauss functions, results not shown). Conversely, the application of the same resolution enhancement techniques to the main transition peak in the absence of Ca2+ did not reveal overlapping peaks (Figure 5B,D,F). Thus, we conclude that, on binding to POPG, Ca2+ induces lipid phase separation. A solid phase consisting of a Ca2+-POPG complex is probably formed within the fluid phase of POPE.39

Subdomain Formation in POPE-POPG Bilayers

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4653

Figure 7. Sequence of TM-AFM images of SLBs of POPE:POPG (3:1, mol/mol) in 20 mM Hepes buffer, 150 mM NaCl (pH 7.40), showing the effect of Ca2+: in the presence of 10 mM CaCl2 (A), after calcium removal with CaCl2-free buffer (B), after adding 20 mM CaCl2 buffer (C), and after removing Ca2+ with a calcium-free buffer (D). The z color scale is 20 nm, and the scale bar is 2 µm.

In the process of SLB formation (see Materials and Methods, subsection 2.3), the sample was incubated at 50 °C for 2 h and observation was carried out at room temperature. This means that the sample has a thermal history and that phases other than the lamellar phase may appear (see inset Figure 5A). So we performed a 31P NMR study in the range of temperatures used in the formation of the SLBs. Thus, Figure 6 shows the corresponding 31P NMR powder pattern spectra of POPE:POPG (3:1, mol/mol) liposomes in the presence of 10 mM CaCl2 at six different temperatures. By comparison with the DSC thermogram in Figure 5, the spectra at 0 and 10 °C can be assigned to the Lβ phase, for which 23 °C corresponds to the main phase transition temperature and the 30 °C spectrum, therefore, corresponds to the LR phase. The shapes of all these spectra are characteristic of lamellar organization,46 which indicates that the presence of non-bilayer phases can be ruled out in this temperature range. However, the spectrum at 62 °C corresponds to the temperature of the very low enthalpy second endotherm (inset in Figure 5A) and shows the coexistence of the hexagonal phase (HII) and an isotropic phase, probably a cubic phase. The spectrum at 50 °C shows a minor shoulder between 5 and 10 ppm, which could be attributed to a small amount of hexagonal and/or isotropic phases. This shoulder

would explain the observed increase at half-height from 10.4 ppm at 30 °C to 15.7 ppm at 50 °C (see Figure 6). Transition temperatures in liposomes and planar-supported bilayers do not coincide due to the effect of the underlying substrate surface on this property. Thus, because of the uncoupling of the two leaflets of the bilayer, the range of temperatures that encompasses the transition is wider in supported bilayers than in liposomes. Therefore, when preparing the SLBs, at least during incubation at 50 °C, the occurrence of non-bilayer structures (HII phases) can be expected. However, as recently demonstrated,42 when the sample is cooled to room temperature for AFM observations, only lamellar structures remain. According to the information obtained from DSC, phase separation in SLBs (Figure 5B,C) can be expected. However, the existence of subdomains within one of the phases is still unexplained. To examine how Ca2+ affects the organization of POPE:POPG (3:1, mol/mol) at the surface, we performed in situ AFM experiments. Thus, Figure 7A shows an SLB of POPE:POPG (3:1, mol/mol) formed in the presence of 10 mM CaCl2 (buffer B; see Material and Methods), where the Lβ phase (brighter regions) with protruding subdomains and the LR phase (darker regions) are observed. Once the image was obtained,


Picas et al.

TABLE 3: Height (h) Values Obtained from the Cross-Section Profile Analysis of SLBs of POPE:POPG (3:1, mol/mol) in the Absence of Calcium (Figure 7B) and with 10 mM Calcium (Figure 7A)a h (nm) Darker domain Brighter domain

n

calcium-free buffer

10 mM calcium buffer

10 10

3.75

3.67 4.32

The h (nm) values are the mean of 10 different sections (n ) 10) and were obtained in tapping mode. a

the tip was lifted up a few nanometers and calcium was removed by exchanging the liquid with a calcium-free buffer (buffer A). As can be seen, after 10 min for tip stabilization, there is no evidence of any segregated subdomain on top of the Lβ phase (Figure 7B). These observations demonstrate that only the subdomains are calcium-dependent. Remarkably, when buffer A was exchanged for a buffer supplemented with 20 mM calcium (buffer C), the subdomains reappear in the Lβ phase (Figure 7C). In corroboration, when buffer C is exchanged for buffer A (calcium-free), the subdomains disappear again and flat featureless SLBs are obtained (Figure 7D). In conclusion, while DSC experiments indicated that there is lateral segregation of POPE:POPG liposomes in the presence of Ca2+, the direct observation of SLBs suggests a more complex situation. Thus, according to DSC and AFM, we conclude that there is a lateral phase separation between the Lβ and LR phases in the POPE:POPG binary mixture. However, AFM reveals Ca2+ subdomains in the Lβ phase. In addition, we have evidence of the reversibility of subdomain formation. Quantitative information obtained from the height profile analysis shown in Figure 7A,B is given in Table 3. The most striking result is that the occurrence and reversibility of subdomains were only observed in the Lβ phase and not in the LR phase. Since the phospholipids that constitute the LR phase did not bind Ca2+, the distal leaflet in the SLBs should be mainly constituted of POPE and almost depleted of POPG. In fact, it has been shown elsewhere28,47 that when liposomes composed of anionic phospholipids are deposited on a charged substrate, there is an asymmetric distribution of phospholipids. The POPE: POPG (3:1, mol/mol) mixture may behave in the same way. Hence, in the fluid phase, POPG may preferentially be adsorbed onto the underlying mica, mainly driven by the electrostatic bridging effect of Ca2+. Consequently, the leaflet in immediate contact with the mica becomes enriched in POPG, while the distal leaflet becomes enriched in POPE (the darker background domain in Figures 1 and 7). Conversely, in the Lβ phase, POPE and POPG in the presence of Ca2+ are immiscible, leading to the formation of subdomains (bright regions in Figures 1 and 7). These subdomains reflect the fact that the Lβ phase is much richer in POPG than the LR phase is, which is the reason why only this phase responds to Ca2+. Once the SLBs are washed out with Ca2+-free buffer, both phospholipid components mix ideally and the subdomains are no longer observed (Figure 7B,D). The capacity of phospholipids to segregate into domains is of physiological interest, since it points to microenvironments in the vicinity of membrane proteins.48 Although the Ca2+ concentration used in this study is above physiological levels, concentrations could be higher in some pathological conditions, such as in the formation of kidney stones, where the presence of negatively charged phospholipids enhances the formation of calcium oxalate crystals.49

Acknowledgment. L.P. is the recipient of an R&D fellowship from the University of Barcelona. Special thanks to Dr. Gerard Oncins for helpful discussions. This study is supported by Hispano-French bilateral action (Grant HF 2007-0028) and the Grant CTQ-2008-03922/BQU from the Spanish Government. References and Notes (1) Jensen, M. O.; Mouritsen, O. G. Lipids Do Influence Protein Function-The Hydrophobic Matching Hypothesis Revisited. Biochim. Biophys. Acta 2004, 1666, 205–226. (2) Bogdanov, M.; Dowhan, W. Phosphatidylethanolamine is Required for in Vivo Function of the Membrane-Associated Lactose Permease of Escherichia coli. J. Biol. Chem. 1995, 270, 732–739. (3) Bogdanov, M.; Dowhan, W. Phospholipid-Assisted Protein Folding: Phosphatidylethanolamine Is Required at a Late Step of the Conformational Maturation of the Polytopic Membrane Protein Lactose Permease. EMBO J. 1998, 17, 5255–5264. (4) Bogdanov, M.; Xie, J.; Heacock, P.; Dowhan, W. To Flip Or Not to Flip: Lipid-Protein Charge Interactions Are a Determinant of Final Membrane Protein Topology. J. Cell Biol. 2008, 182, 925–935. (5) Zhuang, J.; Prive, G. G.; Werner, G. E.; Ringler, P.; Kaback, H. R.; Engel, A. Two-Dimensional Crystallization of Escherichia coli Lactose Permease. J. Struct. Biol. 1999, 125, 63–75. (6) Vazquez-Ibar, J. L.; Weinglass, A. B.; Kaback, H. R. Engineering a Terbium-Binding Site into an Integral Membrane Protein for Luminescence Energy Transfer. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 3487–3492. (7) Merino, S.; Domenech, O.; Vinãs, M.; Montero, M. T.; HernandezBorrell, J. Effects of Lactose Permease on the Phospholipid Environment in Which It Is Reconstituted: A Fluorescence and Atomic Force Microscopy Study. Langmuir 2005, 21, 4642–4647. (8) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of pH and Ionic Strength on Phospholipid Nanomechanics and on Deposition Process onto Hydrophilic Surfaces Measured by AFM. Electrochim. Acta 2006, 51, 5029– 5036. (9) Richter, R. P.; Berat, R.; Brisson, A. R. Formation of SolidSupported Lipid Bilayers: An Integrated View. Langmuir 2006, 22, 3497– 3505. (10) Naseem, R.; Holland, I. B.; Jacq, A.; Wann, K. T.; Campbell, A. K. pH and Monovalent Cations Regulate Cytosolic Free Ca(2+) in E. coli. Biochim. Biophys. Acta 2008, 1778, 1415–1422. (11) Vanounou, S.; Parola, A. H.; Fishov, I. Phosphatidylethanolamine and Phosphatidylglycerol Are Segregated into Different Domains in Bacterial Membrane. A Study with Pyrene-Labelled Phospholipids. Mol. Microbiol. 2003, 49, 1067–1079. (12) Zerrouk, Z.; Alexandre, S.; Lafontaine, C.; Norris, V.; Valleton, J. M. Inner Membrane Lipids of Escherichia coli Form Domains. Colloids Surf., B 2008, 63, 306–310. (13) Milhiet, P. E.; Domec, C.; Giocondi, M. C.; Van Mau, N.; Heitz, F.; Le Grimellec, C. Domain Formation in Models of the Renal Brush Border Membrane Outer Leaflet. Biophys. J. 2001, 81, 547–555. (14) Villaverde, J.; Cladera, J.; Padros, E.; Rigaud, J. L.; Dunãch, M. Effect of Nucleotides on the Thermal Stability and on the Deuteration Kinetics of the Thermophilic F0F1 ATP Synthase. Eur. J. Biochem. 1997, 244, 441–448. (15) Villaverde, J.; Cladera, J.; Hartog, A.; Berden, J.; Padros, E.; Dunãch, M. Nucleotide and Mg2+ Dependency of the Thermal Denaturation of Mitochondrial F1-ATPase. Biophys. J. 1998, 75, 1980–1988. (16) Tilcock, C. P.; Fisher, D. Interaction of Phospholipid Membranes with Poly(ethylene glycol)s. Biochim. Biophys. Acta 1979, 557, 53–61. (17) Hervas, C.; Martinez, A. C.; Silva, M.; Serrano, J. M. Improving the Quantification of Highly Overlapping Chromatographic Peaks by Using Product Unit Neural Networks Modeled by an Evolutionary Algorithm. J. Chem. Inf. Model. 2005, 45, 894–903. (18) Grimalt, J.; Iturriaga, H.; Tomas, X. The Resolution of Chromatograms with Overlapping Peaks by Means of Different Statistical Functions. Anal. Chim. Acta 1982, 139, 155–166. (19) Grancelli, A.; Morros, A.; Cabanãs, M. E.; Dome`nech, O.; Merino, S.; Va´zquez, J. L.; Montero, M. T.; Vinãs, M.; Hernandez-Borrell, J. Interaction of 6-Fluoroquinolones with Dipalmitoylphosphatidylcholine Monolayers and Liposomes. Langmuir 2002, 18, 9177–9182. (20) Killian, J. A.; Fabrie, C. H.; Baart, W.; Morein, S.; de Kruijff, B. Effects of Temperature Variation and Phenethyl Alcohol Addition on Acyl Chain Order and Lipid Organization in Escherichia coli Derived Membrane Systems. A 2H- and 31P-NMR Study. Biochim. Biophys. Acta 1992, 1105, 253–262. (21) Gomez-Fernandez, J. C.; Llamas, M. A.; Aranda, F. J. The Interaction of Coenzyme Q with Phosphatidylethanolamine Membranes. Eur. J. Biochem. 1999, 259, 739–746. (22) Rance, M.; Byrd, R. A. Obtaining High-Fidelity Spin-1. J. Magn. Reson. 1983, 52, 221–240.

Subdomain Formation in POPE-POPG Bilayers (23) Lewis, R. N.; Sykes, B. D.; McElhaney, R. N. Thermotropic Phase Behavior of Model Membranes Composed of Phosphatidylcholines Containing Cis-Monounsaturated Acyl Chain Homologues of Oleic Acid: Differential Scanning Calorimetric and 31P NMR Spectroscopic Studies. Biochemistry 1988, 27, 880–887. (24) Milhiet, P. E.; Vie, V.; Giocondi, M. C.; Le Grimellec, C. AFM Characterization of Model Rafts in Supported Bilayers. Single Mol. 2001, 2, 109–112. (25) Florin, E. L.; Rief, M.; Lehmann, H.; Ludwig, M.; Dornmair, C.; Moy, V. T.; Gaub, H. E. Sensing Specific Molecular Interactions with the Atomic Force Microscope. Biosens. Bioelectron. 1995, 10, 895–901. (26) Proksch, R.; Schaffer, T. E.; Cleveland, J. P.; Callahan, R. C.; Viani, M. B. Finite Optical Spot Size and Position Corrections in Thermal Spring Constant Calibration. Nanotechnology 2004, 15, 1344–1350. (27) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. Dissipation-Enhanced Quartz Crystal Microbalance Studies on the Experimental Parameters Controlling the Formation of Supported Lipid Bilayers. J. Phys. Chem. B 2005, 109, 21755–21765. (28) Rossetti, F. F.; Textor, M.; Reviakine, I. Asymmetric Distribution of Phosphatidyl Serine in Supported Phospholipid Bilayers on Titanium Dioxide. Langmuir 2006, 22, 3467–3473. (29) Cha, T.; Guo, A.; Zhu, X. Y. Formation of Supported Phospholipid Bilayers on Molecular Surfaces: Role of Surface Charge Density and Electrostatic Interaction. Biophys. J. 2006, 90, 1270–1274. (30) Binder, H.; Zschornig, O. The Effect of Metal Cations on the Phase Behavior and Hydration Characteristics of Phospholipid Membranes. Chem. Phys. Lipids 2002, 115, 39–61. (31) Murzyn, K.; Rog, T.; Pasenkiewicz-Gierula, M. Phosphatidylethanolamine-Phosphatidylglycerol Bilayer as a Model of the Inner Bacterial Membrane. Biophys. J. 2005, 88, 1091–1103. (32) Egawa, H.; Furusawa, K. Liposome Adhesion on Mica Surface Studied by Atomic Force Microscopy. Langmuir 1999, 15, 1660–1666. (33) Muller, D. J.; Engel, A. The Height of Biomolecules Measured with the Atomic Force Microscope Depends on Electrostatic Interactions. Biophys. J. 1997, 73, 1633–1644. (34) Dufrene, Y. F.; Barger, W. R.; Green, J. B. D.; Lee, G. U. Nanometer-Scale Surface Properties of Mixed Phospholipid Monolayers and Bilayers. Langmuir 1997, 13, 4779–4784. (35) Dufrene, Y. F.; Boland, T.; Schneider, J. W.; Barger, W. R.; Lee, G. U. Characterization of the Physical Properties of Model Biomembranes at the Nanometer Scale with the Atomic Force Microscope. Faraday Discuss. 1998, 79, 94. (36) Dufrene, Y. F.; Lee, G. U. Advances in the Characterization of Supported Lipid Films with the Atomic Force Microscope. Biochim. Biophys. Acta 2000, 1509, 14–41.

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4655 (37) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of Ion-Binding and Chemical Phospholipid Structure on the Nanomechanics of Lipid Bilayers Studied by Force Spectroscopy. Biophys. J. 2005, 89, 1812–1826. (38) Butt, H. J. Measuring Electrostatic, Van der Waals, and Hydration Forces in Electrolyte-Solutions with an Atomic Force Microscope. Biophys. J. 1991, 60, 1438–1444. (39) Houslay, M. D.; Stanley, K. K. Dynamics of Biological Membranes, influence on synthesis, structure and function; Wiley Interscience: New York, 1982. (40) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of Temperature on the Nanomechanics of Lipid Bilayers Studied by Force Spectroscopy. Biophys. J. 2005, 89, 4261–4274. (41) Oncins, G.; Picas, L.; Hernandez-Borrell, J.; Garcia-Manyes, S.; Sanz, F. Thermal Response of Langmuir-Blodgett Films of Dipalmitoylphosphatidylcholine Studied by Atomic Force Microscopy and Force Spectroscopy. Biophys. J. 2007, 93, 2713–2725. (42) Picas, L.; Montero, M. T.; Morros, A.; Oncins, G.; HernandezBorrell, J. Phase Changes in Supported Planar Bilayers of 1-Palmitoyl-2oleoyl-sn-glycero-3-phosphoethanolamine. J. Phys. Chem. B 2008, 112, 10181–10187. (43) Leonenko, Z. V.; Finot, E.; Ma, H.; Dahms, T. E.; Cramb, D. T. Investigation of Temperature-Induced Phase Transitions in DOPC and DPPC Phospholipid Bilayers Using Temperature-Controlled Scanning Force Microscopy. Biophys. J. 2004, 86, 3783–3793. (44) Kunneke, S.; Kruger, D.; Janshoff, A. Scrutiny of the Failure of Lipid Membranes as a Function of Headgroups, Chain Length, and Lamellarity Measured by Scanning Force Microscopy. Biophys. J. 2004, 86, 1545–1553. (45) Franz, V.; Loi, S.; Muller, H.; Bamberg, E.; Butt, H. H. Tip Penetration through Lipid Bilayers in Atomic Force Microscopy. Colloids Surf., B 2002, 23, 191–200. (46) Cullis, R. R.; de Kruijff, B.; Hope, M. J.; Verkleij, A. J.; Nayar, R.; Farren, S. B.; Tilcock, C. P. S.; Madden, T. D.; Bally, M. B. In Structural Properties of Lipids and Their Functional Roles in Biological Membranes; Aloia, R. C., Ed.; Academic Press: New York, 1983; Vol. 1, pp39-81. (47) Richter, R. P.; Maury, N.; Brisson, A. R. On the Effect of the Solid Support on the Interleaflet Distribution of Lipids in Supported Lipid Bilayers. Langmuir 2005, 21, 299–304. (48) Jacobson, K.; Papahadjopoulos, D. Phase-Transitions and Phase Separations in Phospholipid Membranes Induced by Changes in Temperature, pH, and Concentration of Bivalent-Cations. Biochemistry 1975, 14, 152–161. (49) Khan, S. R.; Glenton, P. A.; Backov, R.; Talham, D. R. Presence of Lipids in Urine, Crystals and Stones: Implications for the Formation of Kidney Stones. Kidney Int. 2002, 62, 2062–2072.

JP8102468

Calcium-Induced Formation of Subdomains in

Recommend Documents