Lipid Organization in Mixed Lipid Membranes Driven by Intrinsic Curvature Difference

Radha Ranganathan; Intisar Alshammri; Miroslav Peric

doi:10.1016/j.bpj.2020.03.009

. 2020 Mar 29;118(8):1830–1837. doi: 10.1016/j.bpj.2020.03.009

Lipid Organization in Mixed Lipid Membranes Driven by Intrinsic Curvature Difference

Radha Ranganathan ^1,^∗, Intisar Alshammri ¹, Miroslav Peric ¹

PMCID: PMC7175694 PMID: 32246900

Abstract

Laurdan fluorescence, novel spectral fitting, and dynamic light scattering were combined to determine lateral lipid organization in mixed lipid membranes of the oxidized lipid, 1-palmitoyl-2-glutaryl-sn-glycero-3-phosphocholine (PGPC), and each of the three bilayer lipids, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), and 1-palmitoyl-2-oleoylphosphatidylcholine (POPC). Second harmonic spectra were computed to determine the number of elementary emissions present. All mixtures indicated two emissions. Accordingly, spectra were fit to two log-normal distributions. Changes with PGPC mole fraction, X_PGPC, of the area of the shorter wavelength line and of dynamic light scattering-derived aggregate sizes show that: DPPC and PGPC form component-separated mixed vesicles for X_PGPC ≤ 0.2 and coexisting vesicles and micelles for X_PGPC > 0.2 in gel and liquid-ordered phases and for all X_PGPC in the liquid-disordered phase; POPC and PGPC form randomly mixed vesicles for X_PGPC ≤ 0.2 and component-separated mixed vesicles for X_PGPC > 0.2. DOPC and PGPC separate into vesicles and micelles. Component segregation is due to unstable inhomogeneous membrane curvature stemming from lipid-specific intrinsic curvature differences between mixing molecules. PGPC is inverse cone-shaped because its truncated tail with a terminal polar group points into the interface. It is similar to and mixes with POPC, also an inverse cone because of mobility of its unsaturated tail. PGPC is least similar to DOPC because mobilities of both unsaturated tails confer a cone shape to DOPC, and PGPC separates form DOPC. DPPC and PGPC do not mix in the liquid-disordered phase because mobility of both tails in this phase renders DPPC a cone. DPPC is a cylinder in the gel phase and of moderate similarity to PGPC and mixes moderately with PGPC.

Significance

A mechanism for membrane lipid organization induced by oxidized lipids (OxPL) that is specific to the OxPL and to the bilayer lipid is presented. Key molecular features responsible as deduced from the data are the contrasting intrinsic curvatures specific to the bilayer phospholipid and to the OxPL and its terminal chemical group. The particular organizations produced can account for stimulation of inflammation, triggering sPLA₂ activity in oxidized membranes and are nonspecific pathways to membrane fission, endo/exocytosis and enhanced binding of curvature sensing proteins. This work is thus of significance to elucidating bioactivity of OxPL. Present spectral fitting analysis involving computed second harmonic spectra of Laurdan fluorescence is a novel approach to determine lipid organization.

Introduction

Biological activity of oxidized phospholipids (OxPL) is an emerging area of research. Mechanisms of OxPL activities remain open for elucidation. Some of OxPL activity can stem from its membrane-altering properties. Intrinsic curvature differences between component molecules in a mixed lipid assembly cause curvature stresses that are known to modulate the binding of some proteins and are potentially nonspecific pathways to membrane fission, exocytosis, and endocytosis (1). Evidence exists for an OxPL-induced increase in membrane polarity, hydration, lateral lipid mobility, membrane thinning, aggravated flip-flop, decrease in membrane order, and bilayer chain melting temperature (2). Physical effects like destruction of lipid order and packing and lipid segregation can result in impairment of membrane protein functions and ion transport (3). Lipid packing, order and spatial distribution play a critical role in cell stiffening, endocytosis, apoptosis, and cell signaling (3, 4, 5). Membrane oxidation is a leading trigger of sPLA₂ activity, a reaction that produces inflammation-triggering molecules (6). This bioactivity of OxPL is due to the physical effect of inducing curvature in the membrane.

The subject of this research is spatial lipid organization in phospholipid (PL) bilayers induced by the OxPL 1-palmitoyl-2-glutaryl-sn-glycero-3-phosphocholine (PGPC) with the aim of gaining mechanistic knowledge on lipid organization in OxPL/bilayer PL mixed assemblies. Mixtures of PGPC with each of the three bilayer PLs, 1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC), and 1-palmitoyl-2-oleoylphosphatidylcholine (POPC), were investigated by novel spectral line fitting of Laurdan fluorescence emission (7).

OxPL are products of oxidation of polyunsaturated fatty acids. Upon oxidation, the sn-2 tail is truncated and terminates in a reactive group. An OxPL is thus asymmetric and inverse cone-shaped in the bilayer. The chemical structure of PGPC is shown in Fig. 1. The three bilayer PLs differ in their intrinsic shape from each other and from the asymmetric PGPC. The question addressed in this article is the relation between particular intrinsic curvature differences and lateral lipid organization.

Intrinsic curvature differences between component molecules result in inhomogeneous bilayer curvature, leading to inhomogeneous composition and segregation (8,9). Thus, key molecular features responsible for the lipid-specific organization deduced from observed data in this work are hypothesized to be the intrinsic curvatures of the bilayer PL and PGPC molecules and the OxPL terminal chemical group, which can bring about activity specific to the OxPL.

Laurdan emission spectra were deconvoluted into elementary emissions by fitting to log-normal Gaussian distributions (7,10, 11, 12). Second harmonic (SH) of the experimental spectra were computed to obtain higher spectral resolution and thus to better define the characteristics of the spectra. Features of the SH provided input to the fitting of spectra. This set of experiments together with previously completed dynamic light scattering (DLS) experiments brings new insight into the lateral lipid organization in the mixed assemblies (13). Similarities and differences observed between the lateral lipid organization in mixed lipid systems of PGPC/POPC, PGPC/DOPC, and PGPC/DPPC motivate this model for a mechanism of curvature-driven lipid organization in these systems.

Materials and Methods

Materials

Mixtures studied in this work were the oxidized PL, PGPC, with each of the bilayer PLs, DPPC, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), and DOPC; 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), and 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC). The lipids were from Avanti Polar Lipids (Alabaster, AL), and Laurdan was from AnaSpec (Fremont, CA).

Sample preparation

2 mM stock solutions of bilayer PL multilamellar vesicles and PGPC micelles in water with 10 μM Laurdan were prepared by the dry film hydration method (7). Sample compositions are expressed in mole fraction of PGPC, X_PGPC. Solutions with 0 < X_PGPC < 0.5 were prepared by mixing bilayer PL and PGPC solutions.

Experiments

Fluorescence emission spectra were recorded from 365 to 650 nm at a step size of 1 nm with a FluoroMax-4 Spectroflourometer (Horiba Scientific, Kyoto, Japan). The excitation wavelength was 360 nm. Slit widths were 1 nm. Sample temperatures, measured by a thermocouple, ranged from 5 to 60°C and were maintained by a thermostated water bath.

Spectral fitting

Wavelength spectra were converted to wavenumber (k) spectra by Jacobian transformation (11,14). SH of spectra were computed first as described in previous work (7). In all cases in this work, the SH indicated two peaks. Therefore, Laurdan emission spectra as functions of wavenumber k were fit to two log-normal distributions each in the form of (7)

{\begin{matrix} I (k) = I_{m} e x p [- \frac{\ln 2}{l n^{2} (ρ)} l n^{2} (\frac{k_{m} + \frac{w ρ}{ρ^{2} - 1} - k}{\frac{w ρ}{ρ^{2} - 1}})] & i f k < k_{m} + \frac{w ρ}{ρ^{2} - 1} \\ I (k) = 0 & i f k \geq k_{m} + \frac{w ρ}{ρ^{2} - 1} a c s \end{matrix}

(1)

The parameters are the peak height I_m, peak position k_m, and asymmetry function $ρ = \frac{k_{m} - k_{m i n}}{k_{m a x} - k_{m}}$ , where k_min and k_max are the half-maximal positions, and width w = (k_max − k_min). Fits were performed by a Trust Region Reflective algorithm in MATLAB (The MathWorks, Natick, MA).

Fluorescence spectra of complex molecules in homogeneous solvents have been fit to single log-normal functions, and the positions k_min and k_max were reported to bear a linear relation to the fluorescence peak position, k_m (11,12). These equations have been used as constraints to fit spectra of the same fluorophore in cases in which there are more elementary emissions than one to reduce the number of fitting parameters. In recent work we showed that only the spectra in aprotic polar solvents could be fit to a single line. k_min and k_max do vary linearly with k_m; however, using these as constraints did not yield better fits to the two line spectra in nonpolar and protic polar solvents and lipid bilayers than when parameters were unconstrained (7). Results presented here are those with unconstrained parameters. Initial guess values and search interval were changed until a convergence and optimum in the confidence interval of the fitted parameters, flatness in residuals of the fit to the spectra and of the SH of fit and SH of data, and R² were obtained. SDs (1/2 of confidence interval) of the fitted parameters are presented as error bars. Areas of the individual lines, representing their contributions to the total emission, were calculated from the fit parameters using the equation for the area of a log-normal distribution (7,15). SDs in areas, k_max, and k_min due to errors propagated from those in k_m, ρ, w, and I_m were calculated using error propagation formulae (7,16). R² in all cases were better than 0.999.

Data analyses methodology

A complementary combination of Laurdan fluorescence emission experiments, fit results, and associations between previously published results of DLS (13) is used to determine lipid organization and develop a mechanistic view of molecular interactions that favor a particular organization. Although DLS can readily confirm the presence of coexisting vesicles and micelles, it cannot give information on the organization of lipids within the mixed vesicle. Probes that partition preferentially in one domain over another are useful in this context (17,18). Most probes have been reported to not prefer one phase over another. Some other physicochemical property drives distribution (18). In the case of Laurdan, membrane curvature differences appear to be that property (18,19). This work takes advantage of Laurdan’s preference for regions of lower curvature over those of higher curvature.

Laurdan fluorescence in bilayers is from unrelaxed and relaxed charge-transfer states (10,19, 20, 21). Complementary fitting of spectra and their SH showed the presence of two emissions, even in the gel phase, with both of the emissions in the blue region of the spectrum (7). The weaker shorter wavelength line at ∼433 nm in the gel phase, distinguishable in the SH, was identified as a unrelaxed charge-transfer state emission from Laurdan in tighter packed lipid regions and the dominant line at ∼447 nm in the gel phase shifting to 484 nm in the liquid phase as emission from less tightly packed regions. The latter is from an emissive state that is in varying stages of relaxation, beginning as unrelaxed at low temperatures in the gel phase and becoming a relaxed charge-transfer state in the liquid phase. Continuous rapid red shift of the longer wavelength line from the less packed region of the gel phase shows that this region becomes liquid as the bilayer changes phase. The 433-nm weak line shifts minimally while its area increases in the transition region where the bilayer is in a gel-liquid mixed-phase state. Peak positions show that Laurdan is present in both regions (18). There are two reasons that could account for the increase in the area fraction of the 433-nm line: 1) The flatness of gel or bilayer PL-rich region increases to relieve the stresses due to the high curvature region. Higher curvature drives Laurdan toward tighter packed lower curvature regions (19,22). 2) Uneven diffusion of oxygen between liquid and gel phases quenches the Laurdan in the liquid phase more than in the gel phase, which can lead to an apparent increase in area fraction of the blue line in coexisting gel-liquid bilayers (23). In any case, the A_b increase is associated with demixing and phase separation.

Composition trajectories of peak positions and their areas in mixed lipid bilayers were analyzed for insight into the evolution of lipid organization and packing. Contribution of an elementary emission is expressed as area fraction or percentage of the total emission. In mixed vesicles with component-separated domains, Laurdan is more likely to be found in a curvature-induced flatter bilayer PL-rich region than in the more curved micelle forming PGPC-rich domains. Consequently, area fraction of the shorter wavelength line, denoted by A_b, will increase. A decrease in or constant A_b with an increase in X_PGPC would suggest random mixing. Peak positions will inform whether Laurdan is predominantly present in one or the other region or distributed.

Results

POPC/PGPC

Measurements were conducted for 0 ≤ X_PGPC ≤ 0.5 at temperatures in the liquid phase from 5 to 45°C. Spectra, SH, fits, and residuals in Fig. 2, a and b are for X_PGPC = 0.35 at 15°C. Two peaks were present in the SH at all compositions and temperatures investigated. In this temperature and composition region, DLS showed the presence of only vesicles of a radius ∼40 nm as represented by sample DLS data in Fig. 2 c. Fluorescence spectra were fitted to two log-normal distributions. The resolved peak positions and area fraction, A_b, of the resolved blue peak are presented in Fig. 3 d. A_b decreases or is about constant with X_PGPC until about X_PGPC ≅ 0.15–0.3 and then increases. The blue peak position at ∼430 nm does not vary significantly, but the red peak at ∼472–480 does shift to a longer wavelength when X_PGPC = 0.5 by ∼6–10 nm at 10 and 15°C. An expanded view of the peak region given as Supporting Materials and Methods illustrates change with X_PGPC in the relative contributions of the two lines.

The decrease or lack of change in A_b for POPC/PGPC, together with the observation of the presence of only vesicles by DLS experiments, prompts this conclusion that PGPC mixes with POPC randomly and induces a decrease in bilayer packing in POPC up to X_PGPC ≈ 0.2. Component segregation within the mixed vesicle begins at this point. There is more presence of Laurdan in the tighter-packed, lower curvature POPC-rich region. Consequently, A_b increases for X_PGPC ≥ 0.2. Peak positions are representative of POPC bilayer liquid phases and, together with A_b increase, support the argument for the increased presence of Laurdan in POPC-rich domains. Fluorescence peaks in PGPC micelles, shown in Supporting Materials and Methods, are at 437 and 495 nm. The combination of higher liquid-phase temperatures and greater PGPC compositions increases the statistical likelihood of some increased Laurdan in PGPC-rich regions, and this can account for the red shift at X_PGPC = 0.5.

DPPC/PGPC

Laurdan emission spectra, SH, fits, and residuals for DPPC/PGPC for X_PGPC = 0.35 at 25°C are shown in Fig. 3, a and b. A reference-simulated SH of a single emission with parameters of the dominant 449-nm resolved line of DPPC is included in Fig. 3 b. The SH of the data is skewed on the red side as compared to the reference spectrum. This indicates two closely spaced peaks from two differently packed regions (7). Fit to two log-normal distributions showed the presence of a weak peak at 434 nm and a strong peak at 449 nm. The weak peak causes skewness on the long-wavelength side because of its width rather than appearing as a peak on the blue side.

Fig. 3 c shows an expanded view of spectra in the shorter wavelength region. An isoemissive point in the area-normalized spectra suggests that the two emissive states are interconverting states or, in this case, that Laurdan favors the tighter packed region (24,25). Statistical likelihood of Laurdan migration to PGPC micelles is to be expected with an increase in PGPC mole fraction as well as temperature. This effect, most likely, is responsible for the emission not sharing the isoemissive point at 50%. However, the minority presence of Laurdan in the more polar PGPC micelles does not produce a well-resolved peak at 495 nm observed in PGPC micelles.

Area fraction of the blue line and peak positions vary with X_PGPC as shown in Fig. 3 d in the gel phase at 25°C and liquid crystalline phase at 46°C. From DLS experiments, it is known that PGPC separates as micelles from DPPC when its mixture composition is >0.2 at both of these temperatures (13). An increase in A_b appears to be a behavior associated with demixing. In the gel phase, A_b increases for X_PGPC > 0.1, and the peak positions remain about constant with X_PGPC. The initial dip could then suggest random mixing and the beginning of demixing within the mixed vesicle at X_PGPC ≈ 0.1 followed by separation into vesicles and micelles for X_PGPC > 0.2. The two peak positions are in the range 432–434 and 447–450 nm. These are typical of positions obtained with a two-line fit in purely bilayer lipid gel phases (7). The nearly constant peak positions at values representative of bilayer PL gel phases together with the increase in A_b is significant because it shows that DPPC and PGPC separate, and Laurdan is selective toward lower curvature bilayer lipid-rich regions.

At 46°C, in the liquid crystalline phase, A_b generally increases with X_PGPC, implying separation even within the mixed vesicle for 0 < X_PGPC < 0.2. The peak positions are at 427–432 and 480–485 nm.

DOPC/PGPC

Fluorescence spectra of DOPC/PGPC at 10°C and their SH are shown in Fig. 4, a and b for X_PGPC = 0.35. SH of the measured spectrum indicates two peaks. An increase in the blue peak contribution with X_PGPC and peak positions are as presented in Fig. 4 c. Coexistence of vesicles and micelles extend to even lower X_PGPC in this system than in DPPC/PGPC as evidenced by DLS (13). DOPC and PGPC separate into DOPC-rich bilayers and PGPC-rich mixed micelles, of radii ∼9 nm, containing some DOPC. The increase in the blue peak area with X_PGPC together with the nonmixing of DOPC and PGPC confirm the proposed association between the two behaviors of an increase in area of blue peak and component separation and that PGPC induces increased lipid packing in DOPC. Peak positions are constant at ∼434 and 487 nm.

DSPC/DMPC and POPC/DOPC

Laurdan fluorescence measurements and line fitting were conducted on the known phase separating DSPC/DMPC and nonseparating POPC/DOPC mixed lipid bilayers to test this method of using peak areas and positions to determine lipid spatial organization. At 37°C, DSPC/DMPC mixed bilayer separates into coexisting domains of DMPC liquid and DSPC gel phases, and POPC/DOPC are randomly mixed at 25°C (26,27). A_b and peak position of DSPC/DMPC vary with mixture composition, as shown in Fig. 5 a, in a manner similar to variation with temperature in the biphasic region of bilayer gel-liquid transitions (7). Fluorescence peak positions indicate that Laurdan is present in both regions, emitting the blue line from the DSPC-rich gel regions and the red line from the DSPC-poor regions, with, however, a preference for the gel phase. As X_DMPC increases, some Laurdan migration to the lower-curvature DSPC-rich gel region causes A_b to increase to a peak value at X_DMPC ≈ 0.4. Laurdan distribution shifts back toward the DMPC-rich liquid phase with a further increase in X_DMPC.

Composition dependence of the area fraction of the blue line (●) and peak positions of the blue (Δ) and red (∎) lines. (a) Shown are DSPC/DMPC at 37°C and (b) DOPC/POPC at 25°C. X refers to mole fraction of DMPC in (a) and of POPC in (b). The lines are to guide the eye. Total lipid concentration was 2 mM. Error bars are SDs from the fits (7). To see this figure in color, go online.

In liquid-liquid POPC/DOPC mixtures at 25°C, there is no separation between the components (27). A_b values are at ∼29.1 ± 1.2, unlike in the nonideal mixing case of DSPC and DMPC. No change in peak positions is found as shown in Fig. 5 b.

Positions of the red line in DSPC/DMPC and in the bilayer gel-liquid transition region in pure bilayers vis-a-vis their positions in DPPC/PGPC is significant in its implication that Laurdan is present in both regions in biphasic DSPC/DMPC and in biphasic pure lipid bilayers, albeit with a preference for flat gel regions, whereas in DPPC/PGPC at 25°C, it is present in only the DPPC gel phase part of the mixture. A lipid organization is suggested in which the PGPC forms buds on the bilayer before separating as micelles. The high curvature buds exclude Laurdan into the DPPC bilayer, and therefore, the emission observed is characteristic of DPPC gel phase.

Discussion

The presence of two peaks in both gel and liquid bilayers at ∼434 and 449 nm in the gel phase that shifts to ∼480 nm in the liquid phase is an expression of ever-present heterogeneous packing. Rapid but continuous red shift of the 449-nm peak upon bilayer melting or upon inclusion of a lower melting bilayer lipid like DMPC in a higher melting bilayer like DSPC gel phase suggests a route through which mixing or melting takes place. DMPC mixes in the looser packed region, and in the case of phase transition, melting begins in this region. Laurdan in this region senses decreased packing or increased polarity because of water entering, and its emission peak shifts to longer wavelengths. Stress due to curvature difference between liquid and gel domains induces Laurdan migration to the lower-polarity, lower-curvature regions, and the peak area of the emission from this region increases. However, curvature difference is perhaps not large enough to drive significant Laurdan redistribution.

The addition of PGPC to DPPC gel phase does not have the same effect as the addition of DMPC to DSPC on the peak positions. PGPC, because of the incompatibility of its intrinsic curvature with that of DPPC, does not enter the bilayer completely but forms buds on the bilayer before separating as micelles. The curvature difference between the buds and the flat bilayer is large enough to tilt the Laurdan distribution in favor of the bilayer PL region. Upon increasing X_PGPC beyond 0.2, curvature inhomogeneity in the mixed vesicle state is no longer sustainable, and component segregation begins. The size of the smaller aggregates measured in DLS is ∼3.5 nm, typical of pure PGPC micelles (13). This determines that the 3.5-nm aggregates are PGPC micelles. In the vesicle/micelle coexistence regime of X_PGPC > 0.2 or in segregated mixed vesicles, had Laurdan nonpreferentially distributed between the phases or between micelles and vesicles, a peak or shoulder would be expected in the red part of the spectrum at ∼491–495 nm. Contrary observations are consistent with the placement of Laurdan in the lower curvature regions, in agreement with published research (19).

PGPC distributes between the bilayer phase and water initially up until X_PGPC = 0.2, and thereupon, PGPC separates and forms micelles. After a further increase, PGPC continues to distribute between the bilayer and micelle phases, and PGPC content in the bilayer also increases. In the demixed state of coexisting vesicles and micelles, the vesicle is a mixed vesicle of DPPC and PGPC. Stresses accompanying curvature inhomogeneity also increase with PGPC, which lead to segregation into lower-curvature DPPC-rich and higher-curvature PGPC domains within the mixed vesicle. Boundary stresses propagate through the domains, increasing the lipid packing that is sensed by Laurdan. This would explain the increase in A_b with X_PGPC in the coexistence region at 25°C.

In DPPC/PGPC at 46 and 60°C and in liquid-phase DOPC/PGPC, micelle radii were ∼9 nm, indicative of mixed PGPC + DOPC or DPPC micelles. The randomizing effect of entropy at the higher temperature and increased lipid mobility in the liquid phase structure can favor a more even distribution of Laurdan between the micelles and bilayer (13). Accordingly, in the coexistence region, the emission peaks are at 432–434 nm from the DPPC- or DOPC-rich domain of the bilayer and at 486 nm from the bilayer liquid and PGPC-rich domains and micelles.

These results show clear differences between mixtures of PGPC with each of the bilayer lipids POPC, DPPC, and DOPC and also between gel, liquid-ordered, and disordered bilayer phases. In summary, 1) PGPC mixes with POPC to high proportions of X_PGPC of at least up to 0.5, forming mixed vesicles. There is a separation of PGPC and POPC into domains within the mixed vesicle from about X_PGPC = 0.2. 2) PGPC mixes with DPPC up to X_PGPC of ∼0.2 in the gel but does not mix in the liquid phase. 3) PGPC does not mix with DOPC. Demixing in liquid-disordered phases while mixing in liquid-ordered phases is lipid specific and has been reported in the case of other lipid mixtures as well (28).

Lateral lipid organization is founded on the intrinsic curvature difference between mixing molecules among other properties. Application of the theory that intrinsic curvature difference leads to inhomogeneous curvature and eventually to component segregation gives the conclusion that PGPC is most similar in shape to POPC, less to DPPC in the gel, and least similar to DPPC and DOPC in the liquid phase.

Examination of the chemical architectures of the molecules supports this conclusion. The OxPL terminal chemical group, bilayer conformation of the sn-2 tail of OxPL, and the intrinsic curvature difference between OxPL and bilayer PL as the physicochemical bases of OxPL segregation provides new insight into OxPL-induced membrane lipid spatial organization. The model proposed here is based on present and published results (13).

The shape of PGPC and those of the bilayer PL are visualized in Fig. 6. Lateral mobility of the chains also play a role in shape determination. Lateral mobility increases with unsaturation and temperature. The carboxylic (COOH) end group of the truncated tail in the OxPL has a hydrophilic character. Coarse-grained simulations show that the sn-2 tail of PGPC is in the interface with a tilt angle of ∼105° with respect to the bilayer normal (3). In the bilayer, OxPL are inverse cone shaped. POPC is asymmetrical and is closer in shape to PGPC. The saturated chains of POPC and PGPC are similar. The ability of the unsaturated tail of POPC to fold back makes it asymmetrical (29). Molecular dynamics simulations show that such a back folding with the terminal tail group turned toward the interface reduces packing stress. Thus, POPC with one unsaturated chain is rendered an inverse cone in the bilayer similar to PGPC. Any shape dissimilarity between POPC and PGPC may not be enough to drive demixing into micelles and vesicles but enough to just separate them into domains within the vesicle. DPPC with its saturated tails are rigid cylinders in the gel phase and cones, due to increased lateral chain mobility, in the liquid phase. The difference in shapes between the cylindrical DPPC and inverse cone PGPC limits the region of mixing in the gel and liquid crystalline phase. DOPC is a cone due to the increased mobility of both of its unsaturated tails. An even larger difference between the inverse cone PGPC and the cone-shaped DOPC or liquid-phase DPPC drives demixing into vesicles and micelles.

Conclusions

Mixtures of the oxidized PL PGPC with each of the bilayer lipids POPC, DPPC, and DOPC exhibit a variety of lateral lipid heterogeneity in organization, depending on lipid tail saturation state and bilayer phase. The type of lipid organization present was determined by complementary Laurdan fluorescence experiments and DLS. Fitting of the fluorescence spectra to two log-normal functions was justified by the shape of the SH. The results of fitting together with the size of aggregates gave new insight into the lipid organization in the mixed lipid bilayers.

Lipid packing in POPC/PGPC has a minimum at X_PGPC ≈ 0.2. Packing in DPPC/PGPC and DOPC/PGPC increase with PGPC content. DLS shows the existence of vesicles and micelles at conditions in which fluorescence shows an increase in the blue part of the spectrum. Thus, an increase in the area fraction of the blue peak is associated with component separation and a decrease or invariance with random mixing. Based on this observation, the conclusions made are as follows:

PGPC and POPC mix randomly to form mixed vesicles up to X_PGPC ≈ 0.2. For higher X_PGPC, aggregates are still mixed vesicles but with POPC-rich and PGPC-rich domains within the mixed vesicle.

PGPC mixes with DPPC up to X_PGPC = 0.2, forming DPPC-rich and PGPC-rich segregated mixed vesicles in the gel and liquid crystalline phases. For X_PGPC > 0.2, DPPC and PGPC separate into mixed vesicles and PGPC micelles.

PGPC does not mix with DOPC or DPPC in the liquid phase, forming instead coexisting PGPC-rich mixed micelles and mixed vesicles of PGPC and DOPC or DPPC.

Fluorescence peak positions are indicative of Laurdan in the bilayer PL-rich phase. A lipid organization is suggested in which the PGPC forms buds on the bilayer before separating as micelles. The high curvature buds exclude Laurdan into the DPPC bilayer, and therefore, the emission observed is characteristic of the bilayer lipid phase.

Grouping of lipids in mixed lipid membranes according to their compatibility in shape is one of the bases for heterogeneity in lipid spatial organization. Accordingly, between the three bilayer lipids studied, POPC is closest in shape to PGPC and DOPC is most unlike PGPC in shape. The mechanistic model proposed for the compatibility of intrinsic curvatures petitions conformational differences between mixing molecules that arise from the polarity of the PGPC terminal group, tail unsaturation, and chain mobility of the bilayer lipids. The polarity of the carboxyl terminal group of PGPC raises the truncated tail toward the bilayer interface, whereas the longer tail remains normal to the interface, thus creating a nonzero angle between the two tails. The enhanced mobility and back folding of the unsaturated tail in POPC over the other tail has a similar effect that confers inverse cone shape to POPC as in PGPC. The mobility of both of the unsaturated chains renders DOPC a cone, and it is most incompatible with the inverse cone PGPC. DPPC with both saturated tails that are rigid in the gel phase and more mobile in the liquid crystalline phase is cylindrical to narrow cone, and its curvature compatibility with PGPC is somewhere between that of POPC and DOPC, limiting its mixing behavior with PGPC. At higher temperature in the liquid phase, the cone angle increases, and DPPC is similar to DOPC in this respect and does not mix with PGPC.

Author Contributions

R.R. designed the research and experiments, analyzed data, developed the mechanistic model, and wrote the manuscript. I.A. conducted the experiments. M.P. carried out spectral fitting.

Acknowledgments

The authors gratefully acknowledge National Institutes of Health for their support through grant contract 1SC3GM122499-01A1.

Editor: Ilya Levental.

Footnotes

Supporting Material can be found online at https://doi.org/10.1016/j.bpj.2020.03.009.

Supporting Material

Document S1. Supporting Materials and Methods and Figs. S1–S3

mmc1.pdf^{(465.6KB, pdf)}

Document S2. Article plus Supporting Material

mmc2.pdf^{(1.9MB, pdf)}

References

1.McDonald C., Jovanovic G., Buck M. Membrane stored curvature elastic stress modulates recruitment of maintenance proteins PspA and Vipp1. mBio. 2015;6:e01188-15. doi: 10.1128/mBio.01188-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Jurkiewicz P., Olżyńska A., Hof M. Biophysics of lipid bilayers containing oxidatively modified phospholipids: insights from fluorescence and EPR experiments and from MD simulations. Biochim. Biophys. Acta. 2012;1818:2388–2402. doi: 10.1016/j.bbamem.2012.05.020. [DOI] [PubMed] [Google Scholar]
3.Ayee M.A.A., LeMaster E., Levitan I. Molecular-scale biophysical modulation of an endothelial membrane by oxidized phospholipids. Biophys. J. 2017;112:325–338. doi: 10.1016/j.bpj.2016.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Salvayre R., Auge N., Negre-Salvayre A. Oxidized low-density lipoprotein-induced apoptosis. Biochim. Biophys. Acta. 2002;1585:213–221. doi: 10.1016/s1388-1981(02)00343-8. [DOI] [PubMed] [Google Scholar]
5.Fruhwirth G.O., Moumtzi A., Hermetter A. The oxidized phospholipids POVPC and PGPC inhibit growth and induce apoptosis in vascular smooth muscle cells. Biochim. Biophys. Acta. 2006;1761:1060–1069. doi: 10.1016/j.bbalip.2006.06.001. [DOI] [PubMed] [Google Scholar]
6.Gibbons E., Nelson J., Bell J.D. Role of membrane oxidation in controlling the activity of human group IIa secretory phospholipase A(2) toward apoptotic lymphoma cells. Biochim. Biophys. Acta. 2013;1828:670–676. doi: 10.1016/j.bbamem.2012.09.013. [DOI] [PubMed] [Google Scholar]
7.Ranganathan R., Burkin A.J., Peric M. Complementary fluorescence emission and second harmonic spectra improve bilayer characterization. J. Fluoresc. 2020;30:205–212. doi: 10.1007/s10895-020-02487-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Góźdź W.T., Bobrovska N., Ciach A. Separation of components in lipid membranes induced by shape transformation. J. Chem. Phys. 2012;137:015101. doi: 10.1063/1.4731646. [DOI] [PubMed] [Google Scholar]
9.Seifert U. Curvature-induced lateral phase segregation in two-component vesicles. Phys. Rev. Lett. 1993;70:1335–1338. doi: 10.1103/PhysRevLett.70.1335. [DOI] [PubMed] [Google Scholar]
10.Vincent M., de Foresta B., Gallay J. Nanosecond dynamics of a mimicked membrane-water interface observed by time-resolved stokes shift of LAURDAN. Biophys. J. 2005;88:4337–4350. doi: 10.1529/biophysj.104.057497. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Bacalum M., Zorilă B., Radu M. Fluorescence spectra decomposition by asymmetric functions: laurdan spectrum revisited. Anal. Biochem. 2013;440:123–129. doi: 10.1016/j.ab.2013.05.031. [DOI] [PubMed] [Google Scholar]
12.Burstein E.A., Emelyanenko V.I. Log-normal description of fluorescence spectra of organic fluorophores. Photochem. Photobiol. 1996;64:316–320. [Google Scholar]
13.Singh J., Ranganathan R. Mixing of oxidized and bilayer phospholipids. Biochim. Biophys. Acta. 2015;1848:1472–1480. doi: 10.1016/j.bbamem.2015.03.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Mooney J., Kambhampati P. Get the basics right: jacobian conversion of wavelength and energy scales for quantitative analysis of emission spectra. J. Phys. Chem. Lett. 2013;4:3316–3318. doi: 10.1021/jz502066v. [DOI] [PubMed] [Google Scholar]
15.Siano D.B., Metzler D.E. Band shapes of the electronic spectra of complex molecules. J. Chem. Phys. 1969;51:1856–1861. [Google Scholar]
16.Ku H.H. Notes on the use of propagation of error formulas. Engineering and Instrumentation. 1966;70C:263–273. [Google Scholar]
17.Baumgart T., Hunt G., Feigenson G.W. Fluorescence probe partitioning between Lo/Ld phases in lipid membranes. Biochim. Biophys. Acta. 2007;1768:2182–2194. doi: 10.1016/j.bbamem.2007.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Bagatolli L.A. To see or not to see: lateral organization of biological membranes and fluorescence microscopy. Biochim. Biophys. Acta. 2006;1758:1541–1556. doi: 10.1016/j.bbamem.2006.05.019. [DOI] [PubMed] [Google Scholar]
19.Viard M., Gallay J., Paternostre M. Origin of laurdan sensitivity to the vesicle-to-micelle transition of phospholipid-octylglucoside system: a time-resolved fluorescence study. Biophys. J. 2001;80:347–359. doi: 10.1016/S0006-3495(01)76019-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Bagatolli L.A., Parasassi T., Gratton E. A model for the interaction of 6-lauroyl-2-(N,N-dimethylamino)naphthalene with lipid environments: implications for spectral properties. Photochem. Photobiol. 1999;70:557–564. [PubMed] [Google Scholar]
21.Parasassi T., De Stasio G., Gratton E. Quantitation of lipid phases in phospholipid vesicles by the generalized polarization of Laurdan fluorescence. Biophys. J. 1991;60:179–189. doi: 10.1016/S0006-3495(91)82041-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Bagatolli L.A., Gratton E. Direct observation of lipid domains in free-standing bilayers using two-photon excitation fluorescence microscopy. J. Fluoresc. 2001;11:141–160. [Google Scholar]
23.Parasassi T., Gratton E. Packing of phospholipid vesicles studied by oxygen quenching of Laurdan fluorescence. J. Fluoresc. 1992;2:167–174. doi: 10.1007/BF00866931. [DOI] [PubMed] [Google Scholar]
24.Viard M., Gallay J., Paternostre M. Laurdan solvatochromism: solvent dielectric relaxation and intramolecular excited-state reaction. Biophys. J. 1997;73:2221–2234. doi: 10.1016/S0006-3495(97)78253-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Merlo S., Yager P. Optical method for monitoring the concentration of general anesthetics and other small organic molecules. An example of phase transition sensing. Anal. Chem. 1990;62:2728–2735. doi: 10.1021/ac00223a015. [DOI] [PubMed] [Google Scholar]
26.Michonova-Alexova E.I., Sugár I.P. Component and state separation in DMPC/DSPC lipid bilayers: a Monte Carlo simulation study. Biophys. J. 2002;83:1820–1833. doi: 10.1016/S0006-3495(02)73947-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Shimokawa N., Nagata M., Takagi M. Physical properties of the hybrid lipid POPC on micrometer-sized domains in mixed lipid membranes. Phys. Chem. Chem. Phys. 2015;17:20882–20888. doi: 10.1039/c5cp03377b. [DOI] [PubMed] [Google Scholar]
28.Krause M.R., Daly T.A., Regen S.L. Push-pull mechanism for lipid raft formation. Langmuir. 2014;30:3285–3289. doi: 10.1021/la500510s. [DOI] [PubMed] [Google Scholar]
29.Risselada H.J., Marrink S.J. Curvature effects on lipid packing and dynamics in liposomes revealed by coarse grained molecular dynamics simulations. Phys. Chem. Chem. Phys. 2009;11:2056–2067. doi: 10.1039/b818782g. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Supporting Materials and Methods and Figs. S1–S3

mmc1.pdf^{(465.6KB, pdf)}

Document S2. Article plus Supporting Material

mmc2.pdf^{(1.9MB, pdf)}

[bib1] 1.McDonald C., Jovanovic G., Buck M. Membrane stored curvature elastic stress modulates recruitment of maintenance proteins PspA and Vipp1. mBio. 2015;6:e01188-15. doi: 10.1128/mBio.01188-15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Jurkiewicz P., Olżyńska A., Hof M. Biophysics of lipid bilayers containing oxidatively modified phospholipids: insights from fluorescence and EPR experiments and from MD simulations. Biochim. Biophys. Acta. 2012;1818:2388–2402. doi: 10.1016/j.bbamem.2012.05.020. [DOI] [PubMed] [Google Scholar]

[bib3] 3.Ayee M.A.A., LeMaster E., Levitan I. Molecular-scale biophysical modulation of an endothelial membrane by oxidized phospholipids. Biophys. J. 2017;112:325–338. doi: 10.1016/j.bpj.2016.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Salvayre R., Auge N., Negre-Salvayre A. Oxidized low-density lipoprotein-induced apoptosis. Biochim. Biophys. Acta. 2002;1585:213–221. doi: 10.1016/s1388-1981(02)00343-8. [DOI] [PubMed] [Google Scholar]

[bib5] 5.Fruhwirth G.O., Moumtzi A., Hermetter A. The oxidized phospholipids POVPC and PGPC inhibit growth and induce apoptosis in vascular smooth muscle cells. Biochim. Biophys. Acta. 2006;1761:1060–1069. doi: 10.1016/j.bbalip.2006.06.001. [DOI] [PubMed] [Google Scholar]

[bib6] 6.Gibbons E., Nelson J., Bell J.D. Role of membrane oxidation in controlling the activity of human group IIa secretory phospholipase A(2) toward apoptotic lymphoma cells. Biochim. Biophys. Acta. 2013;1828:670–676. doi: 10.1016/j.bbamem.2012.09.013. [DOI] [PubMed] [Google Scholar]

[bib7] 7.Ranganathan R., Burkin A.J., Peric M. Complementary fluorescence emission and second harmonic spectra improve bilayer characterization. J. Fluoresc. 2020;30:205–212. doi: 10.1007/s10895-020-02487-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Góźdź W.T., Bobrovska N., Ciach A. Separation of components in lipid membranes induced by shape transformation. J. Chem. Phys. 2012;137:015101. doi: 10.1063/1.4731646. [DOI] [PubMed] [Google Scholar]

[bib9] 9.Seifert U. Curvature-induced lateral phase segregation in two-component vesicles. Phys. Rev. Lett. 1993;70:1335–1338. doi: 10.1103/PhysRevLett.70.1335. [DOI] [PubMed] [Google Scholar]

[bib10] 10.Vincent M., de Foresta B., Gallay J. Nanosecond dynamics of a mimicked membrane-water interface observed by time-resolved stokes shift of LAURDAN. Biophys. J. 2005;88:4337–4350. doi: 10.1529/biophysj.104.057497. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Bacalum M., Zorilă B., Radu M. Fluorescence spectra decomposition by asymmetric functions: laurdan spectrum revisited. Anal. Biochem. 2013;440:123–129. doi: 10.1016/j.ab.2013.05.031. [DOI] [PubMed] [Google Scholar]

[bib12] 12.Burstein E.A., Emelyanenko V.I. Log-normal description of fluorescence spectra of organic fluorophores. Photochem. Photobiol. 1996;64:316–320. [Google Scholar]

[bib13] 13.Singh J., Ranganathan R. Mixing of oxidized and bilayer phospholipids. Biochim. Biophys. Acta. 2015;1848:1472–1480. doi: 10.1016/j.bbamem.2015.03.027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Mooney J., Kambhampati P. Get the basics right: jacobian conversion of wavelength and energy scales for quantitative analysis of emission spectra. J. Phys. Chem. Lett. 2013;4:3316–3318. doi: 10.1021/jz502066v. [DOI] [PubMed] [Google Scholar]

[bib15] 15.Siano D.B., Metzler D.E. Band shapes of the electronic spectra of complex molecules. J. Chem. Phys. 1969;51:1856–1861. [Google Scholar]

[bib16] 16.Ku H.H. Notes on the use of propagation of error formulas. Engineering and Instrumentation. 1966;70C:263–273. [Google Scholar]

[bib17] 17.Baumgart T., Hunt G., Feigenson G.W. Fluorescence probe partitioning between Lo/Ld phases in lipid membranes. Biochim. Biophys. Acta. 2007;1768:2182–2194. doi: 10.1016/j.bbamem.2007.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Bagatolli L.A. To see or not to see: lateral organization of biological membranes and fluorescence microscopy. Biochim. Biophys. Acta. 2006;1758:1541–1556. doi: 10.1016/j.bbamem.2006.05.019. [DOI] [PubMed] [Google Scholar]

[bib19] 19.Viard M., Gallay J., Paternostre M. Origin of laurdan sensitivity to the vesicle-to-micelle transition of phospholipid-octylglucoside system: a time-resolved fluorescence study. Biophys. J. 2001;80:347–359. doi: 10.1016/S0006-3495(01)76019-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] 20.Bagatolli L.A., Parasassi T., Gratton E. A model for the interaction of 6-lauroyl-2-(N,N-dimethylamino)naphthalene with lipid environments: implications for spectral properties. Photochem. Photobiol. 1999;70:557–564. [PubMed] [Google Scholar]

[bib21] 21.Parasassi T., De Stasio G., Gratton E. Quantitation of lipid phases in phospholipid vesicles by the generalized polarization of Laurdan fluorescence. Biophys. J. 1991;60:179–189. doi: 10.1016/S0006-3495(91)82041-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Bagatolli L.A., Gratton E. Direct observation of lipid domains in free-standing bilayers using two-photon excitation fluorescence microscopy. J. Fluoresc. 2001;11:141–160. [Google Scholar]

[bib23] 23.Parasassi T., Gratton E. Packing of phospholipid vesicles studied by oxygen quenching of Laurdan fluorescence. J. Fluoresc. 1992;2:167–174. doi: 10.1007/BF00866931. [DOI] [PubMed] [Google Scholar]

[bib24] 24.Viard M., Gallay J., Paternostre M. Laurdan solvatochromism: solvent dielectric relaxation and intramolecular excited-state reaction. Biophys. J. 1997;73:2221–2234. doi: 10.1016/S0006-3495(97)78253-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] 25.Merlo S., Yager P. Optical method for monitoring the concentration of general anesthetics and other small organic molecules. An example of phase transition sensing. Anal. Chem. 1990;62:2728–2735. doi: 10.1021/ac00223a015. [DOI] [PubMed] [Google Scholar]

[bib26] 26.Michonova-Alexova E.I., Sugár I.P. Component and state separation in DMPC/DSPC lipid bilayers: a Monte Carlo simulation study. Biophys. J. 2002;83:1820–1833. doi: 10.1016/S0006-3495(02)73947-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Shimokawa N., Nagata M., Takagi M. Physical properties of the hybrid lipid POPC on micrometer-sized domains in mixed lipid membranes. Phys. Chem. Chem. Phys. 2015;17:20882–20888. doi: 10.1039/c5cp03377b. [DOI] [PubMed] [Google Scholar]

[bib28] 28.Krause M.R., Daly T.A., Regen S.L. Push-pull mechanism for lipid raft formation. Langmuir. 2014;30:3285–3289. doi: 10.1021/la500510s. [DOI] [PubMed] [Google Scholar]

[bib29] 29.Risselada H.J., Marrink S.J. Curvature effects on lipid packing and dynamics in liposomes revealed by coarse grained molecular dynamics simulations. Phys. Chem. Chem. Phys. 2009;11:2056–2067. doi: 10.1039/b818782g. [DOI] [PubMed] [Google Scholar]

PERMALINK

Lipid Organization in Mixed Lipid Membranes Driven by Intrinsic Curvature Difference

Radha Ranganathan

Intisar Alshammri

Miroslav Peric

Abstract

Significance

Introduction

Figure 1.