Accumulation of styrene oligomers alters lipid membrane phase order and miscibility

Significance

Accumulation of nanopollutants in lifeforms has become the subject of increasing concern, but effects are not understood. Biological systems are inherently complex, and therefore any small variations on the membrane properties can potentially perturb cell functionality. In this study we observe that the presence of styrene oligomers in lipid membranes alters their phase behavior, with stronger changes occurring in more complex systems, which could translate into potential cellular disruption. This study provides evidence that the presence of nanopollutants in the membrane may alter its fundamental properties and affect cell viability.

Abstract

Growth of plastic waste in the natural environment, and in particular in the oceans, has raised the accumulation of polystyrene and other polymeric species in eukyarotic cells to the level of a credible and systemic threat. Oligomers, the smallest products of polymer degradation or incomplete polymerization reactions, are the first species to leach out of macroscopic or nanoscopic plastic materials. However, the fundamental mechanisms of interaction between oligomers and polymers with the different cell components are yet to be elucidated. Simulations performed on lipid bilayers showed changes in membrane mechanical properties induced by polystyrene, but experimental results performed on cell membranes or on cell membrane models are still missing. We focus here on understanding how embedded styrene oligomers affect the phase behavior of model membranes using a combination of scattering, fluorescence, and calorimetric techniques. Our results show that styrene oligomers disrupt the phase behavior of lipid membranes, modifying the thermodynamics of the transition through a spatial modulation of lipid composition.

The increasing amount of plastic present in sea waters has become a major issue in recent years, with increasing concerns regarding the potential hazardous effects it may have on living systems (1, 2). Annual production of plastic has reached almost 300 million tons/y, of which 5 to 13 million tons are estimated to reach the oceans by different means (3). While initially the main concern for plastic contamination was the presence of microplastic, produced by polymeric degradation, recently the focus has shifted to nanoplastic (4, 5). Objects of this scale can easily enter the food chain via digestion and there is increasing evidence of plastic micro- and nanoobjects found in marine life forms (6–9). Moreover, nanometer-size polymer particles are also produced industrially for specific research and technological applications, such as imaging, sensing, and preparation of nanocomposites (10), providing a second route, besides degradation, for plastic-derived nanoparticles’ entry into sea waters. Despite a current lack of evidence on the presence of nanoobjects (4), studies have indeed shown that plastic nanoparticles can accumulate in the tissues of living organisms and disrupt their metabolism (11–13) and that size plays an important role in determining their accumulation (12, 13). However, a physicochemical characterization of the interaction between plastic nanoobjects and living organisms is still lacking, especially regarding the mechanisms of potential toxicity. As the first barrier encountered by any foreign object entering an organism, the cell membrane is the primary candidate of investigation in assessing possible toxicity of plastic nanofragments. In particular, the membrane lipid lateral organization plays a crucial role in many cellular signaling processes due to the presence in the membrane of small transient domains called “lipid rafts,” and even minute changes in membrane lipid organization can result in a potential alteration of these processes and pose a threat to cellular viability.

Polystyrene is one of the most commonly used plastics in the world, contributing to a significant fraction of marine plastic wastes in the form of styrene oligomers (SO) (14), and it has been shown in several studies to directly affect lipid bilayer physical properties when accumulated within the membrane. Accumulation of styrene oligomers and polymers in 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine bilayers indeed showed, in numerical simulations, to change the membrane mechanical properties and lateral lipid organization (15) and a stabilization of cholesterol-induced domains (16). However, experimental studies on the effects of polystyrene and oligomers in the cell membrane and model membranes are still lacking. It was previously reported that incorporation of styrene monomers into lipid membranes significantly changes the fluidity of the membrane (17), and a similar effect was indeed shown for charged polystyrene sulfonated chains in surfactant bilayers (18), but from an experimental point of view many aspects of the interaction are still elusive. In this work we investigate the effects of styrene oligomers ($M_n$ = 500 Da) on the phase transition of lipid bilayers composed of unsaturated lipids or a mixture of unsaturated/saturated lipids, to obtain a more comprehensive picture of the role of membrane complexity on the effects of styrene oligomer accumulation. Indeed, small styrene chains display the largest mobility; they are the most likely to be transferred to the lipid membranes upon contact with a degrading piece of plastic. Also, when one considers the important role played by the gel–liquid transition point of the saturated lipid in the formation of domains in ternary lipid mixtures relevant for biomembranes, it becomes clear that the most fundamental question to be asked in this context is that of the influence of the oligomers on the gel–liquid transition of the saturated lipid. We investigated the changes in transition using differential scanning calorimetry (DSC), small-angle neutron scattering (SANS), and Laurdan fluorescence spectra to extract information on the structure and the thermodynamics. Moreover, we directly visualized the changes on the membranes at the micrometric scale using epifluorescence microscopy.

Results

Our experimental results focus on two main aspects of styrene oligomer–lipid interaction: First, we employed SANS both to confirm the presence of the oligomers within the lipid membrane and to obtain information about the changes on the bilayer structure they induce. Subsequently, we focused on changes in the thermodynamics properties and phase behavior of the membrane caused by inclusion of these short styrene chains.

Fluid and Gel Phases of 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine Lipid Membranes Can Host a High Amount of Styrene Oligomers.

For SANS, we utilized liposomes formed either of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) in D₂O as a solvent, to maximize scattering contrast, or ${\mathrm{d}}_{62}$-DPPC (DPPC lipids with all of the hydrogen molecules in the carbon chains replaced by deuterium) vesicles in a H₂O/D₂O mixture to mask the signal from the lipids and maximize polymer signal. In both lipidic compositions we compared changes between lipid-only liposomes and liposomes formed with 30% mol of styrene oligomers, and scattering curves were acquired at both $\mathrm{20}\hspace{0.17em}°$C and $\mathrm{50}\hspace{0.17em}°$C to investigate the differences between gel and fluid phase. Herein we report a brief description of the results; a more comprehensive analysis of the SANS data can be found in SI Appendix. Fig. 1 shows the scattering curves of DPPC:SO 70:30 and ${\mathrm{d}}_{62}$-DPPC:SO 70:30 with their respective model fitting. In the case of DPPC:SO (Fig. 1 A and B), we observe a good agreement between our data and a bilayer model, particularly for scattering curves at $\mathrm{50}\hspace{0.17em}°$C (Fig. 1B). Contrarywise, for DPPC:SO at $\mathrm{25}\hspace{0.17em}°$C the fitting strongly deviates from the scattering curve in the intermediate region, indicating a membrane structural change occurring in the gel phase due to the presence of SO. A similar feature has been reported before by Pencer et al. (19) for membranes displaying liquid domain coexistence, and they suggested that changes in the scattering curve in that region might be linked to membrane lateral segregation and heterogeneities in scattering length density. The fitting deviates in the intermediate $q$ region. The good values of model fitting for the DPPC:SO system suggest that upon incorporation of oligomers we still maintain a liposomal suspension with no additional structures, such as aggregates or micelles. This result was also confirmed via dynamic light scattering (DLS), where we observe one single peak in the size distribution of the liposomes (SI Appendix, Fig. S6). Scattering curves for ${\mathrm{d}}_{62}$-DPPC vesicles further confirm the presence of styrene oligomers within the membrane. Fig. 1 C and D shows the data obtained when the lipid signal is masked. The clear increase in scattering intensity in systems containing SO compared to ${\mathrm{d}}_{62}$-DPPC-only vesicles (Fig. 1 C and D, Insets) is due to the presence of oligomers in the membrane.

Fig. 1.

SANS scattering curves for 50-nm LUVs and relative fittings. (A) Fitting with a bilayer model of DPPC:SO scattering curves at $\mathrm{25}\hspace{0.17em}°$C. (B) Fitting with a bilayer model of DPPC:SO scattering curves at $\mathrm{50}\hspace{0.17em}°$C. (C) Fitting of SANS scattering curves for d₆₂-DPPC:SO 70:30 at $\mathrm{20}\hspace{0.17em}°$C. Inset shows comparison of scattering signal between d₆₂-DPPC and d₆₂-DPPC:SO liposomes at $\mathrm{20}\hspace{0.17em}°$C. (D) Comparison of d₆₂-DPPC and d₆₂-DPPC:SO liposomes at $\mathrm{50}\hspace{0.17em}°$C. Inset shows comparison of scattering signal between d₆₂-DPPC and d₆₂-DPPC:SO liposomes at $\mathrm{50}\hspace{0.17em}°$C. All plots are represented as experimental data (circles) and relative fitting (gray line). Both fittings for C and D were performed using the Sasfit implemented thin disk model. The signal arising from the pure lipid samples displays very low intensity due to the matching of scattering density between solvent and lipids. In the case of $\mathrm{20}\hspace{0.17em}°$C, the signal appears slightly more intense due to changes in the scattering length density of lipids.

Phase Changes of the Lipid Bilayer Result in Styrene Oligomer Reorganization.

To better understand the oligomer distribution within the lipid bilayer in both the gel ($S_o$) and fluid ($L_{\alpha }$) phases, we extrapolated the structural data for the bilayer under lipid-masking conditions in the presence of styrene oligomers. Fitting of the experimental curves was performed using a disk model, motivated by previous analyses of phase separation in model membranes using SANS, where domains were fitted using a similar form factor (20), and obtained good agreement with the data (Fig. 1 C and D). We observe that in the gel phase polystyrene appears to aggregate with an average diameter of 349 $\pm$ 0.3 Å, smaller than the liposome size, and with a thickness of 38.3 $\pm$ 0.1 Å, comparable to the hydrophobic region of the bilayer. The fitting values are compatible with a heterogenous lateral distribution of the polymer, as was hinted by the strong deviation from a bilayer scattering in DPPC:SO in ${\mathrm{D}}_2$O (Fig. 1A). Moreover, the resulting thickness is comparable with the values obtained for DPPC using a model-free approach (Kratky–Porod and Modified Kratky–Porod at $\mathrm{25}\hspace{0.17em}°$C (SI Appendix, Figs. S3 and S4). In the case of the polymer distribution in the fluid phase, we argue that the lack of a visible inflection point in the scattering curve for $\mathrm{50}\hspace{0.17em}°$C reflects a change in lateral distribution of the polymer, which becomes more homogeneously dispersed within the bilayer, and not a real change in diameter of the vesicles. However, since it was not possible to verify the values for the diameter of the disk, we took into consideration only the value for the thickness, which resulted to be 34.0 $\pm$ 0.1 Å. This is once again comparable with the bilayer thickness obtained for DPPC:SO liposomes (SI Appendix, Table S1), hinting that both in the gel and in the fluid phase the polymer is intercalated in the acyl chains.

Vesicles Containing SO Have Higher Membrane Order.

The Laurdan emission spectra for DPPC liposomal suspensions containing increasing molar fractions of SO (SI Appendix, Fig. S7), and the calculated corresponding general polarization (GP) values for each temperature probed (Fig. 2A), highlight changes in the membrane order upon incorporation of the oligomers. The GP values obtained for pure DPPC liposomes in the gel and the fluid phase are consistent with previously reported GP curves for DPPC (21), with a GP in the gel phase remaining almost constant at 0.50 $\pm$ 0.01, and with a sharp decrease after ${\mathrm{T}}_m$, to a negative value of −0.33 $\pm$ 0.01 at $\mathrm{60}\hspace{0.17em}°$C. Incorporation of the polymer in the membrane does not significantly induce any variation at low temperatures, with a slight increase to 0.54 $\pm$ 0.01 for 30% molar fraction of styrene oligomers being observed. In contrast, presence of the polymer strongly shifts the GP to more positive values in the fluid phase, up to −0.13 $\pm$ 0.01 for the maximal polymer fraction investigated. We observe a clear dependence of the GP variation on the polymer content (Fig. 2 A, Inset) that we here argue is due to a uniform distribution of the oligomers within the membrane. The difference in variation of GP, minimal in the gel phase and maximal for the liquid crystalline phase, indicates a different behavior of the polymer between the $S_o$ and $L_{\alpha }$. Laurdan spectral properties in different phases of the bilayer have been directly linked to the hydration in the glycerol backbone region of the lipids and to the degree of tail alignment (22). Therefore, the increase in GP suggests a lower number of water molecules in the fluid phase of the bilayer due to the presence of polystyrene, hinting also at a higher packing order of the lipids induced by SO. Such an effect might be indicative of intercalation of styrene oligomers through the lipid chains in the fluid phase. A similar effect was reported for incorporation of hydrophobic monoterpenes, where there was an increase of the order parameter of the acyl chains, particularly the carbon groups closer to the interface (23). In the case of styrene oligomers, the lack of a hydrophilic headgroup prevents the polymer from being exposed to the water–acyl chain interface; however, the observed shift in GP, indicative of a higher order of the chain closer to the headgroup region, suggests that at least a significant portion of the oligomers is not confined in the midplane of the hydrophobic region. This picture is consistent with neutron scattering results reported by Richter et al. (24) in the case of styrene monomers interacting with the 1,2-dimyristoyl-sn-glycero-3-phosphocholine vesicle, where the distribution of the monomers was found to be a coexistence of molecules highly segregated in the midplane and molecules aligned with the hydrocarbon tails.

Fig. 2.

(A) GP curves over temperature for LUVs of DPPC formed in water at 0% (black squares), 10% (orange circles), 20% (blue triangles), and 30% (green inverted triangles) molar fraction of styrene oligomers. Inset shows the variation of GP between fluid and gel phase for each SO molar fraction, calculated as $\mathrm{\Delta }$GP = GP[$\mathrm{50}\hspace{0.17em}°$C] − GP[$\mathrm{20}\hspace{0.17em}°$C]. Liposomes were extruded to a size of 100 nm. Each data point and error bars represent the average and standard deviation from three (n = 3) separate samples, with four technical repeats each. (B) DSC thermographs of DPPC MLVs containing increasing amounts of styrene oligomers. Each curve represents the second thermographic signal from the full run experiment. Thermographs provide a determination of the melting temperature $T_m$ (maximum), enthalpy of melting $\mathrm{\Delta }H$ (area comprised between the baseline and the peak), and transition width $T_{1/2}$ (half height width).

Gel-to-Fluid Transition Is Inhibited by Styrene Oligomers.

We employed differential scanning calorimetry to measure the gel-to-fluid transition, specifically changes in the enthalpy and transition temperature which could hint at effects on the membrane induced by styrene oligomers. DSC thermographs (Fig. 2B) show for DPPC-only bilayers a sharp transition peak centered at 41.8 $\pm$ 0.2 $°$C, in good agreement with data from the literature (25). With an increasing amount of styrene oligomers incorporated the transition temperature slightly decreases, to a final value of 40.8 $\pm$ 0.4 $°$C for 30% polymer molar fraction. The peak also significantly decreases in intensity and broadens, suggesting a loss in enthalpy and cooperativity. Calculations of $\mathrm{\Delta }H$ of the transition yield a value of 38.5 $\pm$ 0.6 kJ$\cdot {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ for pure DPPC, consistent with previously reported data (25). Incorporation of polymer within the bilayer results in a decrease of enthalpy and cooperativity with increase of the amount of styrene oligomers, as shown in Table 1 and SI Appendix, Fig. S8. In particular, we observe a linear decrease of $\mathrm{\Delta }H$ and increase of $T_{1/2}$ with respect to polymer content, whereas $T_m$ does not vary significantly (SI Appendix, Fig. S9). Our results differ significantly from the previously reported trend for dioctadecyldimethylammonium bromide vesicles incorporating 60% molar fraction of styrene monomers, where a strong decrease in transition temperature and only slight variation of enthalpic contribution were observed (17). However, the effects we observe of decrease in $T_m$ and broadening of the transition peak are in agreement with previously reported studies of incorporation of hydrophobic molecules in lipid bilayers (26, 27).

Table 1.

Calculated $\mathrm{\Delta }H$, $T_m$, and $T_{1/2}$ for DPPC with increasing molar fractions of styrene oligomers

Styrene oligomers fraction, % mol	$\mathrm{\Delta }H$, kJ$\cdot {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$	$T_m$, $°$C	$T_{1/2}$, $°$C
0	38.5 $\pm$ 0.7	41.8 $\pm$ 0.2	0.27 $\pm$ 0.01
10	30.7 $\pm$ 3.8	41.5 $\pm$ 0.4	0.29 $\pm$ 0.01
20	27.9 $\pm$ 1.1	41.2 $\pm$ 0.3	0.45 $\pm$ 0.01
30	25.1 $\pm$ 0.9	40.8 $\pm$ 0.4	0.66 $\pm$ 0.01

Values are presented as average and standard deviation from two (n = 2) separate samples, with six technical repeats each.

The overall trends of depression of the melting temperature and broadening of the transition are consistent with the effects reported for hydrophobe/lipid bilayer interactions. Wolka et al. (28) and Rolland et al. (26) reported that incorporation of penetration enhancers reduces the transition temperature of DPPC to $\mathrm{40}\hspace{0.17em}°$C at 10% molar fraction of molecule, as well as causing a significant increase in the width of the transition peak. Similar results have been found for membranes containing flavonoids, with increasing hydrophobicity of the molecule producing a stronger effect (29, 30). Borsacchi et al. (31) also reported similar behavior for incorporation of pheromones in 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) bilayers.

Styrene Oligomers Increase Miscibility between $S_o$ and $L_{\alpha }$ Phases.

To better understand the effects induced by the styrene oligomers on the phase behavior of lipid bilayers, we probed changes in the Laurdan spectra in liposomes composed of DOPC and DPPC at different compositions (SI Appendix, Fig. S9). Calculation of general polarization yielded GP curves illustrated in Fig. 3. For pure DOPC:DPPC liposomes we observe a broad transition, starting from a value of 0.27 $\pm$ 0.01 and 0.45 $\pm$ 0.01 at $\mathrm{15}\hspace{0.17em}°$C to −0.25 $\pm$ 0.02 and −0.21 $\pm$ 0.01 at $\mathrm{45}\hspace{0.17em}°$C for DPPC molar fractions of 0.4 and 0.6, respectively. The values at low temperature are much lower than the GP of pure DPPC bilayer (21), indicating a gel/fluid phase coexistence. The miscibility temperature $T_m$, measured from the intersection of GP curves obtained from excitation wavelengths 350 and 400 nm (SI Appendix, Fig. S10), is approximately 28.3 $\pm$ 1.5 $°$C for $X_{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}}$ = 0.4 and 34.9 $\pm$ 0.9 $°$C for $X_{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}}$ = 0.6, consistent with known values of miscibility for this system (32–36). The values of GP and curve broadness are in agreement with a membrane displaying $L_d$/$S_o$ coexistence at low temperature and a homogeneous liquid disordered phase above $T_m$. In the case of systems with the same lipid ratio and additional incorporation of styrene oligomers, we observe significant changes in the Laurdan GP values. For 0.4 DPPC molar fraction the initial value is highly increased to 0.33 $\pm$ 0.01 at $\mathrm{15}\hspace{0.17em}°$C, as well as the value in the liquid crystalline phase which displays higher GP, with a value of 0.22 $\pm$ 0.02 at $\mathrm{45}\hspace{0.17em}°$C. Moreover, the transition temperature decreases to 23.4 $\pm$ 0.8 $°$C and the curve becomes sharper. Increasing the fraction of DPPC in the membrane seems to slightly reduce these effects, as for $X_{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}}$ = 0.6 the initial values of GP are comparable, while at 45 $°$C we observe a slight increase to −0.16 $\pm$ 0.02.

Fig. 3.

(A) Variation of general polarization over temperature for LUVs of DOPC:DPPC 60:40 (black squares) and DOPC:DPPC 60:40 + 10 mol% SO (orange circles). (B) Variation of general polarization over temperature for LUVs of DOPC:DPPC 40:60 (blue triangles) and DOPC:DPPC:SO 40:60 + 10 mol% SO (green inverted triangles). Liposomes were extruded to a size of 100 nm. Each data point and error bars represent the average and standard deviation from two (n = 2) separate samples, with four technical repeats each.

We further confirmed the lowering of the liquidus line observed in the GP plots upon addition of 10 mol% SO, by analyzing under epifluorescence microscopy the area coverage of gel phase domains in giant unilamellar vesicles (GUVs) composed of different ratios of DOPC and DPPC. Giant unilamellar vesicles at both DOPC:DPPC compositions display fluid/gel coexistence at room temperature, with characteristic $S_o$ domains of irregular or hexagonal morphology (Fig. 4), consistent with previous reported studies (35). For each vesicle we calculated first the area fraction of $S_o$ domains, $f_b$, in the pure lipid case, and obtained a total area coverage of 0.17 $\pm$ 0.03 and 0.47 $\pm$ 0.03 for DPPC molar fraction of 0.4 and 0.6, respectively. It is possible to compare measured values for area coverage with theoretically predicted fractions of gel phase by the lever rule, benefiting from the almost vertical shape of the solidus line in this region of the phase diagram (32). Assuming thus a constant solidus line at 0.95 for the considered DPPC molar fraction interval, our values are consistent with theoretical predictions (Table 2), as well as with previously reported coverage of DPPC gel domains in DOPC membranes (35). Since the $S_o$ fraction in GUVs agrees with DOPC:DPPC phase diagrams, we can now measure the gel phase coverage upon addition of 10 mol% of styrene oligomers, while maintaining the same DOPC:DPPC composition. This overall allows to obtain an experimental phase diagram by combining the values obtained by Laurdan and GUVs. Vesicles incorporating 10% mol SO still exhibit $S_o$/$L_{\alpha }$ coexistence, with intact domain morphology (Fig. 4); however, the solid domain fraction, $f_b^{SO}$, is greatly reduced. Analysis of the domain coverage yielded a fraction for the solid phase of 0.077 $\pm$ 0.021 and 0.28 $\pm$ 0.06 for 60:40 and 40:60 mixtures, respectively. The values of solid area fraction obtained with GUVs containing SO allow us to estimate the liquidus line in presence of 10% mol oligomers. For that reason, we have visualized GUVs containing styrene oligomers at two separate temperatures, namely $\mathrm{20}\hspace{0.17em}°$C and $\mathrm{23}\hspace{0.17em}°$C. Using a solidus line value for each temperature based on previously reported data by Schmidt et al. (32), and using the lever rule, we obtained values for $X_a^{SO}$ of 0.35 $\pm$ 0.05 for $\mathrm{20}\hspace{0.17em}°$C and 0.38 $\pm$ 0.06 for $\mathrm{23}\hspace{0.17em}°$C. We observe that the miscibility temperatures obtained both from Laurdan spectra and from GUVs confirm the lowering of the liquidus line between $S_o$-$L_{\alpha }$ coexistence and pure $L_{\alpha }$ phase (Fig. 5).

Fig. 4.

Representative images of DOPC:DPPC GUVs labeled with 1 mol% DiI, without and with 10 mol% fraction of SO incorporated, visualized using epifluorescence microscopy. The bright area represents presence of fluorophore, which partitions in the $L_{\alpha }$ phase. Dark spots are $S_o$ domains. The 60:40 vesicles with and without SO were imaged at $\mathrm{20}\hspace{0.17em}°$C, and 40:60 systems with and without SO were imaged at $\mathrm{23}\hspace{0.17em}°$C. (Scale bars, 5 $\mathrm{\mu }$m.)

Table 2.

Summary of $S_o$ area coverage for DOPC:DPPC GUVs at different composition

DPPC molar fraction	Experimental $S_o$ fraction ($f_b$)	Theoretical fraction pure DPPC	Theoretical fraction 0.95 DPPC	$S_o$ fraction with 10% SO ($f_b^{SO}$)
0.4	0.17 $\pm$ 0.03	0.18 $\pm$ 0.01	0.20 $\pm$ 0.01	0.08 $\pm$ 0.02
0.6	0.47 $\pm$ 0.03	0.42 $\pm$ 0.01	0.47 $\pm$ 0.01	0.28 $\pm$ 0.06

Each value represents average and standard deviation of two separate samples of 40 vesicles each. Vesicles with 0.4 DPPC molar fraction with and without SO were imaged at 20°C, and systems with 0.6 DPPC molar fraction with and without SO were imaged at 23°C.

Fig. 5.

Comparison between experimental data points for the liquidus line estimated by results from Chen and Santore (35) (orange), experimental data points for DOPC:DPPC (black squares) and DOPC:DPPC:SO (blue triangles) obtained from Laurdan emission spectra, and theoretical predictions of liquidus lines for DOPC:DPPC (red) and DOPC:DPPC + 10% SO (green).

A Theoretical Model Explains the Changes in DOPC:DPPC Miscibility.

Furthermore, we compared our experimental results with a thermodynamics model for binary phase diagrams developed by Wolff et al. (36), which considers the gel-to-liquid transition in the framework of a mean-field Ising (two states) model. Simply, for a binary lipid mixture the thermodynamics of the gel-to-liquid transition of a binary mixtures can be written as

$$\begin{array}{cc}\hfill {\mathcal{G}}^{\mathrm{m}\mathrm{i}\mathrm{x}}& \left(T,{\varphi }_1,{\varphi }_2,m\right)=-m\left[h_1\left(T\right){\varphi }_1+h_2\left(T\right){\varphi }_2\right]-2J{m}^2\hfill \\ \hfill & +\left(\frac{1+m}{2}\right)\ln \left(\frac{1+m}{2}\right)+\left(\frac{1-m}{2}\right)\ln \left(\frac{1-m}{2}\right)\hfill \\ \hfill & +{\varphi }_1\ln \left({\varphi }_1\right)+{\varphi }_2\ln \left({\varphi }_2\right)\hfill \end{array}$$ [1]

with

$$h_1\left(T\right)=\frac{\mathrm{\Delta }{H}_1}{2R{T}_1^2}\left(T-T_1\right),\hspace{0.17em}{h}_2\left(T\right)=\frac{\mathrm{\Delta }{H}_2}{2R{T}_2^2}\left(T-T_2\right),$$ [2]

where ${\varphi }_1$ and ${\varphi }_2$ are the area fractions of the two lipid species, $J$ is the mismatch energy associated with the interaction between two neighboring lipids, $m$ is a nonconserved scalar order parameter restricted to the interval $\left[-\mathrm{1,1}\right]$, $R$ is the gas constant, $\mathrm{\Delta }{H}_1$ and $\mathrm{\Delta }{H}_2$ are the respective gel-to-liquid enthalpies of melting, and $T_1$ and $T_2$ are the respective transition temperatures. Changes in area per lipid at the transition are not taken into account. The thermodynamic potential ${\mathcal{G}}^{\mathrm{m}\mathrm{i}\mathrm{x}}$ represents the ratio $G/{\mathcal{A}}_{l}RT$ of the Gibbs free energy of mixing of the hydrated lipid bilayer divided by the total area of lipids ${\mathcal{A}}_l$ and temperature $T$, while $m$ interpolates continuously between negative (gel state) and positive (fluid state) values. Phase coexistence results from minimization of ${\mathcal{G}}^{\mathrm{m}\mathrm{i}\mathrm{x}}$ with respect to $m$ followed by convex minimization with respect to ${\varphi }_1$ and ${\varphi }_2$.

For DOPC:DPPC bilayers we calculated the phase diagram using values of $\mathrm{\Delta }{H}^{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}}=38.5$ kJ$\cdot {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ and $T_m^{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}}=41.8\hspace{0.17em}°$C for DPPC, obtained from our DSC experiments, and $\mathrm{\Delta }{H}^{\mathrm{D}\mathrm{O}\mathrm{P}\mathrm{C}}=7.7$ kJ$\cdot {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ and $T_m^{\mathrm{D}\mathrm{O}\mathrm{P}\mathrm{C}}=-21.3\hspace{0.17em}°$C obtained from the literature. The mismatch energy of the interaction between neighboring lipids was set at $J$ = 0.31. The theoretical model is in good agreement with our experimental data from Laurdan emission spectra and with the DOPC:DPPC phase diagram reported by Chen and Santore (35) (Fig. 5).

For systems containing styrene oligomers, we kept all model parameters constant, except for the different enthalpy and transition temperature obtained from calorimetry experiments, namely $\mathrm{\Delta }{H}^{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}:\mathrm{S}\mathrm{O}}=30$.7 kJ$\cdot {\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ and $T_m^{\mathrm{D}\mathrm{P}\mathrm{P}\mathrm{C}:\mathrm{S}\mathrm{O}}=41.5\hspace{0.17em}°$C, to calculate the new boundary line. The boundary line in presence of styrene oligomers indeed shows good agreement with our experimental observation (Fig. 5). This suggests that the thermodynamic changes induced by SO in DPPC bilayers are the main driving force behind depression of the liquidus line toward lower temperature.

Conclusions

Our results show that the incorporation of styrene oligomers in lipid bilayers is strongly coupled to membrane phase behavior. In the low temperature gel phase, styrene short chains are laterally segregated in the membrane, as shown by SANS experiments. This segregation is due to the poor solubility of the oligomers in the tightly packed acyl chain region. Neutron scattering experiments also show that, as the bilayer melts from the gel to the $L_{\alpha }$ phase upon an increase of temperature, the oligomers become more uniformly distributed and intercalate between the acyl chains toward the water interface. This scenario is also supported by the strong variation of GP in the fluid phase, as well as by the close to linear dependence with polymer content of the Laurdan emission and of the transition enthalpy. Incorporation of the oligomers was also found to alter the gel-to-liquid main transition and the lipid packing of the membrane in the fluid phase, as shown by the Laurdan emission spectra. These changes can be ascribed to a preferential partition of the styrene chains into the fluid phase, thus depressing the gel-to-fluid transition temperature of the membrane. The evolution of the thermodynamic coexistence lines between pure bilayers and bilayers containing the short styrene chains is correlated with the change in enthalpy at the gel–fluid transition. In the case of the binary lipid system DOPC:DPPC the presence of the styrene oligomers shifts the $S_o$ and $L_{\alpha }$ miscibility line toward lower temperatures: Incorporated styrene oligomers have thus the potential to preclude domain formation. Our findings demonstrate that the presence of styrene chains in lipid bilayers affects the membranes’ phase behavior and point to a likely disruption of biomembranes’ functionality by polymeric nanopollutants.

Materials and Methods

Materials.

Chloroform solutions of DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine, ${\mathrm{C}}_{44}{\mathrm{H}}_{84}{\mathrm{N}\mathrm{O}}_8$P, ${\mathrm{M}}_w$ 786.11) and DPPC (1,2-dipalmitoyl-sn-glycero-3-phosphocholine, C₄₀H₈₀NO₈P, ${\mathrm{M}}_w$ = 734.039) were purchased from Avanti Polar Lipids. DiI stain (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate ${\mathrm{C}}_{59}$H₉₇ClN₂O₄, M_w = 933.8793) was provided by ThermoFisher Scientific. Sucrose (C₁₂H₂₂O₁₁, M_w = 342.3) and Laurdan (6-dodecanoyl-N,N-dimethyl-2-naphthylamine) were purchased from Sigma-Aldrich. Atactic styrene oligomers ((C₈H₈)_n) M_n = 500 Da) were purchased from Polymer Source Inc. All chemicals had high purity and were used without further purification. The osmolarities of the sucrose solutions were measured with a cryoscopy osmometer (Osmomat 030; Gonotec).

Multilamellar Vesicles and Large Unilamellar Vesicles (LUVs) Preparation.

For DSC experiments, multilamellar vesicles (MLVs) were prepared by placing 1.5 mg of the desired lipid (with and without styrene oligomers) composition in chloroform in a glass vial, and organic solvent was evaporated using first an argon stream for 20 min, followed by 8 h of vacuum pumping. The lipid film was then hydrated with MilliQ distilled water (18.2 M$\mathrm{\Omega }$) at $\mathrm{70}\hspace{0.17em}°$C to reach the desired concentration and gently vortexed. The resulting MLV suspensions were sonicated for 15 min to disperse larger aggregates. In the case of Laurdan measurements, large unilamellar vesicles (LUVs) were prepared similarly to the MLV protocol with the following modifications. The lipid solution in chloroform was mixed with 1 mol% Laurdan in chloroform prior to evaporation. Upon swelling with MilliQ water and sonication, the MLV solution was extruded using an Avanti Mini-Extruder 21 times through a 100-nm diameter pore polycarbonate filter.

Small Unilamellar Vesicles Preparation.

Liposomes of the desired lipid composition were formed using MilliQ distilled water (18.2 M$\mathrm{\Omega }$), ${\mathrm{H}}_2$O/${\mathrm{D}}_2$O 92:8, or ${\mathrm{D}}_2$O. For fluorescence measurements, the lipids were stained with 1 mol% Laurdan in chloroform prior to evaporation. Liposomal solutions remained stable over a period of days. In the case of liposomal suspension used in SANS experiments liposomes were then extruded using an Avanti Mini-Extruder. The sample was first extruded 21 times through 200-nm diameter, subsequently 21 times through 100-nm diameter, and finally 21 times through 50-nm diameter pore polycarbonate filters.

SANS.

The neutron wavelength resolution was $2\cdot \mathrm{\Delta }\lambda /\lambda =0.1$. All scattering data were corrected for background, and incoherent scatterings from 1-mm-thick D₂O or H₂O solutions were used to correct for the deviation in uniformity of the detector response. The final data were converted to absolute scale. All experiments were first performed at low temperature (25$\hspace{0.17em}°$C) and then at $\mathrm{50}\hspace{0.17em}°$C. SANS measurements for 100% D₂O contrast were performed at 24.0-, 8.0-, and 1.5-m detector distances at the D11 small-angle instrument at Institut Laue-Langevin, Grenoble, France. The wavelength of neutrons was set to 6 Å. The instrument was configured to provide an effective q range of $\sim$0.01 to 0.4 Å⁻¹. All SANS measurements were performed at $\mathrm{20}\hspace{0.17em}°$C and $\mathrm{50}\hspace{0.17em}°$C, well above the phase transition temperature of DPPC ($\sim$42$\hspace{0.17em}°$C). The SANS data from the position-sensitive two-dimensional detector were reduced to one-dimensional profiles I(q) vs. q, by using a facility-supplied data reduction software lamp. Preliminary scattering data and measurements to determine the contrast matching conditions were obtained on the PACE spectrometer at the Laboratory Leon Brillouin (LLB) in Saclay, France.

Steady-State Fluorescence.

A total of 3 mL of liposomal suspension stained with Laurdan of total concentration 3 mg/mL was placed in a quartz silica cuvette with 1-mm path length. Acquisition of Laurdan emission spectra was performed with a Jobin Horiba FluoroMax equipped with a Peltier unit to control temperature. Excitation wavelength was set at 350 nm with a bandpass of 1 nm and emission was also recorded with slit of 1 nm. The solution was equilibrated at a given temperature for 10 min before each acquisition. For DPPC:SO Laurdan experiments, each sample (n = 3) was probed for two cycles of heating and cooling. For DOPC:DPPC:SO LUVs experiments, we performed two cycles of heating and cooling at excitation 350 nm and one cycle of heating and cooling at excitation 400 nm on two (n = 2) separate samples. GP was calculated using the standard expression provided by Parasassi et al. (37). To quantify the miscibility temperature for each sample, we compared the GP plot from excitations 350 and 400 nm and extrapolated the temperature range at which the two curves intersect. A full description of quantification is provided in SI Appendix.

Giant Unilamellar Vesicles Preparation.

GUVs composed of DOPC:DPPC, both in absence and presence of styrene oligomers, were prepared by electroformation following the protocol introduced by Angelova and Dimitrov (38). Simply, 5 $\mathrm{\mu }$L of 2 mg/mL solution of DOPC:DPPC or DOPC:DPPC:SO at the desired molar ratio, stained with 1% mol of DiI, in chloroform was spread on each cathode of a custom-made electroformation stage. The stage was kept under vacuum for at least 1 h to ensure complete evaporation of solvent and subsequently the lipid film was hydrated using sucrose solution (100 mosm/kg) at $\mathrm{55}\hspace{0.17em}°$C. We applied a sinusoidal electric field of 1-V peak–peak intensity at 10 kHz for 1 h while keeping the sample heated above the transition temperature. The resulting GUV suspension was kept at $\mathrm{20}\hspace{0.17em}°$C in a water bath to ensure complete stabilization of the sample. Vesicles were used on the same day of preparation.

Optical Microscopy and $S_o$ Phase-Domain Quantification.

Imaging of GUVs labeled with 1 mol% DiI was performed using a Nikon Eclipse TE2000-E microscope equipped with a Diagnostic Instruments NDIAG1800 camera and a Nikon 60× water immersion, NA 1.2 objective. Observation was performed in the epifluorescence mode with a Hg lamp, with 100 W (Intensilight; Nikon) as excitation source and adapted filtering TE/TRITC Ex 543/22 nm, DM 562 nm, and Em 593/40 nm. GUV samples were initially swelled by diluting the external medium with $\sim$5 vol% of pure water. Prior to experimental observation, GUVs were kept at the desired temperature for at least 1 h to stabilize. A total of 100 $\mathrm{\mu }$L of a GUV solution was placed in a chamber. Quantification of $S_o$ phase domains in vesicles was performed by measuring the radius of the domains, which was then corrected for by taking into account the spherical nature of the GUVs. A more detailed description of the quantification approach can be found in SI Appendix. For each lipid composition we quantified the domain area coverage from two separate samples, with 40 vesicles from each replicate.

Differential Scanning Calorimetry.

The calorimetry measurements were performed with a high-sensitivity differential scanning calorimeter ($\mathrm{\mu }$DSC; Setaram). The measuring cell was filled with the sonicated sample, while the reference cell was filled with MilliQ water. The temperature of the cells was changed with a constant rate (heating rate, 0.5 K$\cdot {\mathrm{m}\mathrm{i}\mathrm{n}}^{-1}$; cooling rate, 0.3 K$\cdot {\mathrm{m}\mathrm{i}\mathrm{n}}^{-1}$). The system was equilibrated $\sim$20 min before each heating or cooling ramp. The analysis of DSC data was performed using OriginPro 9.0. For each sample (n = 2) we performed three cycles of heating and cooling.

DLS.

Liposomal suspensions of lipid and lipid:SO vesicles were characterized by dynamic light scattering using a Malvern Zetasizer Nano ZS. Simply, 1 mL of liposomal suspsension at 1 mg/mL was placed in a disposable plastic cuvette and light scattering was recorded. Measurements were repeated at least three times for statistics.

Data Availability.

SANS, Laurdan, and DSC are ASCII data files arranged as tables. GUV images are provided as TIFF files. Data have been deposited in Zenodo (DOI: 10.5281/zenodo.4118224).