Article pubs.acs.org/Langmuir
Interaction of Cytotoxic and Cytoprotective Bile Acids with Model Membranes: Influence of the Membrane Composition M. Esteves,† M. J. Ferreira,† A. Kozica,†,‡ A. C. Fernandes,† A. Gonçalves da Silva,*,† and B. Saramago*,† †
Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal Wrocław University of Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland
‡
S Supporting Information *
ABSTRACT: To understand the role of bile acids (BAs) in cell function, many authors have investigated their effect on biomembrane models which are less complex systems, but there are still many open questions. The present study aims to contribute for the deepening of the knowledge of the interaction between BAs and model membranes, in particular, focusing on the effect of BA mixtures. The cytotoxic deoxycholic acid (DCA), the cytoprotective ursodeoxycholic acid (UDCA), and the equimolar mixture (DCA + UDCA) were investigated. Monolayers and liposomes were taken as model membranes with two lipid compositions: an equimolar mixture of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), sphingomyelin (SM), and cholesterol (Chol)) traditionally associated with the formation of lipid rafts and an equimolar POPC/SM binary mixture. The obtained results showed that DCA causes the fluidization of monolayers and bilayers, leading to the eventual rupture of POPC/SM liposomes at high concentration. UDCA may provide a stabilization of POPC/SM membranes but has a negligible effect on the Chol-containing liposomes. In the case of equimolar mixture DCA/UDCA, the interactions depend not only on the lipid composition but also on the design of the experiment. The BA mixture has a greater impact on the monolayers than do pure BAs, suggesting a cooperative DCA−UDCA interaction that enhances the penetration of UDCA in both POPC/SM and POPC/SM/Chol monolayers. For the bilayers, the presence of UDCA in the mixture decreases the disturbing effect of DCA.
■
and cholesterol (Chol). Zhou et al.11 claimed that both UDCA and DCA are toxic in the absence of Chol and that the presence of Chol enhances the membrane’s ability to resist their damaging action. Later, the same authors reported somewhat divergent results.5 They observed that, even in the presence of Chol, DCA solubilizes the model membranes at high concentrations, while at low concentrations it interacts preferentially with the liquid-disordered domains. Mello-Vieira et al.4 studied the interaction of DCA and UDCA with model liposome membranes of raftlike and nonraft compositions and found that both acids interact preferentially with liquiddisordered membranes (nonraft). They came to the conclusion that cytotoxic DCA remained at the surface, leading to membrane expansion, in contrast to the conclusions of Zhou et al.11 Moreover, they claimed that the simultaneous presence of both acids, DCA and UDCA, hindered the cytoprotective effect of UDCA and led to a cumulative disruptive effect on the membrane structure. The objective of the present work is to contribute to a deeper understanding of the mechanism of bile acid−lipid interaction. In particular, the effect of the simultaneous presence of both
INTRODUCTION Bile acids (BAs) are natural detergents, produced in the liver, whose major physiological functions include lipid solubilization in the lumen of the small intestine and control of the metabolism of cholesterol.1 The cytotoxicity of bile acids is known to be related to their hydrophobicity, which is dictated by the configuration of the hydroxyl groups.2 Two examples of bile acids with quite different hydrophilicity and toxicity are deoxycholic acid, DCA, and ursodeoxycholic acid, UDCA. While cytotoxic DCA has apoptotic effects on the plasma membrane of living cells, UDCA seems to protect these membranes against the toxic effects of hydrophobic bile acids.3,4 A large number of investigations have focused on the interaction of cytotoxic and cytoprotective BAs with model lipid membranes,5,6 but these studies have not reached a consensus. Some authors suggest that BAs adopt a similar position to cholesterol in membranes, intercalating with the lipid acyl chains,5 while others propose that they are adsorbed to the membrane surface.6 Several biomimetic synthetic systems, namely, monolayers and liposomes, have been used in those studies.4,7−11 Donovan et al.7,8 found that both hydrophilic and hydrophobic bile acids were able to condense the expanded 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) monolayers, but none of the bile acids were able to further condense the condensed mixed monolayers of POPC © XXXX American Chemical Society
Received: May 8, 2015 Revised: July 23, 2015
A
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
min for homogenization before spreading the lipid solution at the interface. The compression of the monolayer was started 30 min after the spreading of lipids, allowing the evaporation of the solvent and the stabilization of the initial surface pressure. In method 2, the acids were spread together with the lipids, with an SGE gastight microsyringe, on the subphase, and 15 min was allowed for solvent evaporation and stabilization of the initial surface pressure. Symmetric compression was performed at a rate of 10 mm/min. The temperature of the subphase was maintained at 25 °C. At least three π−A isotherms were obtained for each system. Quartz Crystal Microbalance (QCM-D) Measurements. A quartz crystal microbalance with dissipation, model E4, and gold-coated quartz crystals from Q-Sense (Gothenburg, Sweden) were used to study the interaction of BAs with the adsorbed liposomes. The normalized frequency change, Δf, of the sensors and their dissipation change, ΔD, were recorded, up to the ninth overtone, upon sequential addition of HEPES (baseline), liposome suspension, HEPES (rinsing), BA solution, and HEPES (rinsing) at 37 °C. The maximum BA concentration (1000 μM) is below the cmc of both BAs17 under the present experimental conditions. Immediately before use, the sensors were submitted to two cycles of UV−ozone treatment, 15 min each, separated by rinsing with ultrapure water, and finally dried with a stream of nitrogen. Data were processed using Qtools3 software for viscoelastic modeling. The viscosity and density of the fluid and the film were taken from Viitala et al.18 Differential Scanning Calorimetry. The experiments were performed in a VP-DSC microcalorimeter from MicroCal (Northampton, MA, USA). Heating/cooling cycles were used between 10 and 50 °C with a scan rate of 60 °C/h. All solutions were degassed except those containing liposomes to avoid liposome disruption. DSC curves were analyzed using Origin 7.0 software (OriginLab Corporation, Northampton, MA). It was not possible to draw realistic baselines for enthalpy calculations because the onset of the transitions was below 10 °C, which is the lowest temperature accessible in our calorimeter. Nuclear Magnetic Resonance. 31P NMR spectra of POPC/SM and POPC/SM/Chol liposome suspensions in HEPES were obtained in the absence and in the presence of BAs with two NMR Bruker Advance III spectrometers operating at 500 and 400 MHz, respectively. All of the spectra were acquired at 12.5, 25, and 37 °C without decoupling and without a lock. Lipid and bile acid concentrations of 12.5 and 2 mM, respectively, were used. Each spectrum is the result of 2000 scans with a repetition rate of 1.5 s and 90° pulses. 31P chemical shifts were externally referenced to phosphoric acid at 85%.
cytotoxic and cytoprotective bile acids needs further investigation. We studied the interactions of hydrophobic DCA and hydrophilic UDCA with two model membranes: phospholipid monolayers and bilayers (liposomes). The effect of the BAs on the structure of the phospholipid monolayers was studied through surface pressure−area measurements using the Langmuir technique. The bilayers were investigated with a quartz crystal microbalance with dissipation (QCM-D), differential scanning calorimetry (DSC), and nuclear magnetic resonance (31P NMR). Two lipid compositions were chosen: (1) ternary mixture of POPC, sphingomyelin (SM), and Chol (1:1:1, molar proportions) associated with the formation of lipid rafts which are microdomains in cellular membrane mediators of cell-signaling pathways, membrane trafficking, and signal transduction12,13 and (2) an equimolar mixture of SM with POPC, which is known to phase separate into more rigid SM-rich domains and more fluid POPC-rich domains.14,15 The chosen biomembrane models are multiphase complex systems. Although the presence of more than one phase may make the physical interpretation of the experimental results difficult, these models have the advantage of being more realistic than single-component models. In fact, phase separation plays a critical role in biochemical phenomena because the BAs can partition differently in liquid-disordered and liquid-ordered phases. Furthermore, the choice of binary mixture POPC/SM, without cholesterol, allows an investigation of the specific role of this lipid component in the interaction with bile acids.
■
EXPERIMENTAL SECTION
Materials. Sphingomyelin (brain SM, porcine) and POPC were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Cholesterol, deoxycholic acid, ursodeoxycholic acid, N-(2-hydroxyethyl)piperazineN-(2-ethanesulfonic acid) (HEPES), chloroform, and sodium dodecyl sulfate were supplied by Sigma−Aldrich (St. Louis, MO, USA). Methanol was acquired from Fisher Chemical (Loughborough, U.K.). Sodium chloride, ammonia, and hydrogen peroxide were obtained from Panreac (Barcelone, Spain), and Extran was obtained from Merck (Kenilworth, NJ, USA). Ultrapure water (resistivity ≥18.2 MΩ· cm) was purified with a Millipore Milli-Q system. HEPES buffer (10 mM HEPES and 0.1 M NaCl; pH 7.4) was used in the preparation of liposomes and BA solutions. Solutions of Hellmanex II 2% were obtained from Hellma GmbH (Müllheim, Germany). Methods. Preparation of Liposomes. Large unilamellar vesicles (LUV) were prepared using the protocol supplied by Avanti Polar Lipids.16 Appropriate amounts of the lipids were dissolved in chloroform to form equimolar mixtures of POPC/SM or POPC/ SM/Chol. Chloroform was removed afterward by drying in a vacuum oven for at least 3 h. The resulting film was hydrated with HEPES buffer at 65 °C, alternating with manual and vortex agitation for 1 h, and the suspension was then submitted to freeze−thaw cycles. The multilamellar vesicles formed were extruded in a homemade extruder through polycarbonate filters (Nuclepore from Whatman, Brentford, U.K.) of decreasing pore size. The final LUVs were stored at 4 °C for a maximum of 1 week. The mean diameter of the liposomes checked by dynamic light scattering was 102 nm. Suspensions of liposomes with a lipid concentration of 1.12 mM were used for QCM-D and DSC experiments. Surface Pressure−Area (π−A) Measurements. All assays were carried out on a KSV 2000 Langmuir−Blodgett system from KSV Instruments (Helsinki, Finland). Chloroform/methanol (4:1, v/v) solvent mixtures were used to dissolve the lipids at concentrations of about 1 mM each. Two distinct procedures were adopted: acids dissolved in a HEPES buffer subphase at pH 7 or added to the interface. In method 1, the aqueous solution of acids was injected into the subphase and left 15
■
RESULTS π−A Isotherms of Lipids. Experimental π−A isotherms obtained for the monolayers of both POPC/SM and POPC/ SM/Chol lipid mixtures, as well as the theoretical isotherms, based on the assumption of ideal mixing, are presented in Figure 1. Theoretical isotherms were calculated from the isotherms of single lipids, presented in the Supporting Information (Figure S1), weighted by their composition in the mixture. The POPC/SM equimolar binary mixture forms a liquid expanded (LE) monolayer (full line, 1), with lift off at ∼100 Å2/molecule and collapse at π ≈ 42 mN/m. The theoretical curve (dashed line, 1′) follows the experimental isotherm except at π > 20 mN/m, where a small shoulder shifts the calculated curve to smaller mean molecular areas (MMAs). This shoulder is derived from the contribution of the liquid expanded−liquid condensed (LE−LC) phase transition in the monolayer of pure SM (Figure S1). Thus, the small positive deviation exhibited by the POPC/SM binary mixture suggests that POPC hinders the LE−LC transition observed in pure SM. This behavior is compatible with the phase separation and
B
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
POPC/SM/Chol (curve 2) compares with the data reported by Wydro,23 who claimed that at the 1:1:1 molar composition, cholesterol molecules associate mainly with SM forming the SM/Chol-rich domains (LC phase) dispersed in fluid matrix of POPC (LE phase). Effect of DCA on Lipid Monolayers. To investigate the effect of BAs on both POPC/SM (1:1) and POPC/SM/Chol (1:1:1) monolayers, two kinds of experiments were designed. In method 1, BAs were dissolved in the subphase before the spreading of lipids at the interface (Figure 2A,B), while in method 2, the BAs were mixed together with lipids in the spreading solution (Figure 2C,D). A general trend was observed for both methods: the π−A curve of the lipid monolayer shifts to larger areas with the increase in BA content. This shift is more pronounced at low pressures, decreasing progressively as the surface pressure increases. This suggests that the BA incorporated into the lipid monolayer at low surface pressures is partially excluded or changes its orientation at high surface pressures. Figure 2A,B also shows the π−A isotherm of the 8 μM DCA content in the subphase (dashed line) in the absence of lipids at the interface. It is interesting that DCA dissolved in the subphase influences the surface pressure at the starting point, even before the spreading of the lipids. Upon compression, the surface pressure remains nearly constant (π = 5−7 mN/m) over the whole compression run. This behavior is consistent with the surfactant character of DCA. The effect of BA on the lipid monolayers is measured by the relative area deviation at constant surface pressure, calculated by eq 1, where AL is the area of the lipid monolayer in the absence of BA and ALBA is the area of the lipid monolayer in the presence of BA at the same surface pressure:
Figure 1. Representative π−A isotherms (solid lines) of the POPC/ SM (1:1) mixture (1) and of the POPC/SM/Chol (1:1:1) mixture (2) obtained at 25 °C on a HEPES buffer subphase. The theoretical curves (dashed lines, 1′ and 2′) were calculated on the basis of the assumption of ideal mixing.
partial miscibility of both components in the monolayer, where POPC-rich domains coexist with SM-rich domains.15 Differently, the experimental isotherm of the POPC/SM/ Chol equimolar ternary mixture (full line, 2) reveals a significant contraction of MMAs relatively to the theoretical curve (dashed line, 2′). This observation attests to the wellknown condensing effect of cholesterol on phospholipid monolayers.19−21 This effect can be attributed to various factors such as molecular packing, the ordering of chains, and the tilting of polar heads. In particular, sphingolipids and cholesterol tend to organize into liquid-ordered functional domains, the so-called lipid rafts.22The π−A isotherm of
Figure 2. π−A isotherms of POPC/SM (A) and POPC/SM/Chol (B) mixtures on HEPPES-buffered subphases at pH 7 in the presence of several contents (1.33, 2.63, 8.0, and 13.3 μM) of DCA dissolved in the subphase and for several DCA molar fractions (0.33, 0.5, 0.67, and 0.8) in the spreading solution of POPC/SM (C) and (0.33, 0.5, 0.73, and 0.8) in the spreading solution of POPC/SM/Chol (D). π−A isotherm of a 8.0 μM DCA subphase (dashed line) in the absence of the lipid monolayer but in the same range of total surface area (from the starting point at 530 cm2 to the ending point at ∼180 cm2). C
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir A − AL ΔA = LBA AL AL
significantly with the DCA content, one can assume that most of the DCA retained at the interface is segregated into DCArich domains coexisting with the lipid monolayer. Effect of DCA on Liposomes. As demonstrated previously, zwitterionic liposomes adsorb intact on the surface of oxidized gold-coated quartz crystals.24 At the temperature of the assay (37 °C), the adsorbed binary liposomes are in a liquid-disordered (Ld) phase, and for ternary liposomes, liquiddisordered (Ld) domains coexist with liquid-ordered (Lo) ones, the so-called lipid rafts.25 The smaller frequency shift observed after the addition of the POPC/SM liposomes in comparison to that of the POPC/SM/Chol (dark-gray bars in Figure 4)
(1)
Figure 3 compares the effect of DCA on POPC/SM and POPC/SM/Chol monolayers at π = 15 mN/m as a function of
Figure 3. Relative area increase (%) of POPC/SM (gray circles) and POPC/SM/Chol (black triangles) monolayers on a HEPES buffer subphase at π = 15 mN/m, pH 7, and 25 °C as a function of the concentration of DCA dissolved in the subphase (A) and of the molar fraction of DCA in the spreading solution (B). Curves are guides for the eye.
Figure 4. Average values, from at least three trials, for the frequency shifts (left) and dissipation shifts (right) of the third harmonic and their standard deviations obtained after the addition of liposomes (black), rinsing (dark gray), the addition of DCA (gray), and final rinsing (light gray).
the molar concentration of DCA (CDCA) dissolved in the subphase (panel A) and as a function of the DCA molar fraction (XDCA) in the spreading solution (panel B). At higher surface pressures (omitted) a similar trend was observed with lower deviations. These results show that the incorporation of DCA from the subphase is more pronounced in the POPC/SM expanded monolayer than in the POPC/SM/Chol monolayer. The opposite observation was found when DCA was added to the spreading solution, i.e., DCA promotes a relative larger expansion in the POPC/SM/Chol monolayer than it does in the POPC/SM monolayer. Furthermore, for low contents of DCA in the spreading solution (XDCA ≤ 0.5), the relative area deviation remains very low, while it increases when DCA becomes the major component in the spreading solution (XDCA > 0.5). This suggests that, at low contents (XDCA ≤ 0.5), part of the DCA molecules probably dissolve into the subphase and the remaining DCA adsorbs in the polar region of the lipid monolayer (forming an underneath adsorbed layer), which does not affect the area significantly. For higher contents (XDCA > 0.5), in addition to the dissolution into the subphase and the adsorption that saturates the polar region of the lipid monolayer, the excess DCA remains (incorporated/segregated) at the interface, thus contributing to the area increase. As the collapse surface of the lipid monolayer does not change
indicates a lower adsorbed mass per unit area which may be attributed to the flattening of the adsorbed liposomes that are less rigid. The results of the viscoelastic modeling show that the main difference between the liposome layers lies in the higher viscosity of the POPC/SM/Chol liposomes, 2.5 ± 0.3 mPa·s compared to 1.9 ± 0.2 mPa·s for the POPC/SM liposomes. Figure 4 shows the frequency and dissipation shifts obtained when different concentrations of DCA (500, 750, and 1000 μM) were added to adsorbed POPC/SM and POPC/SM/Chol liposomes. We must report that a wide range of DCA concentrations (down to 1 μM) was tested but that the observed effects were negligible. The general behavior was a positive shift in frequency and a decrease in dissipation, which are more significant for higher concentrations, with the exception of DCA (500 μM) added to the binary liposomes where the frequency decreases and the dissipation increases. The changes in thickness and viscosity resulting from the DCA addition to the adsorbed liposome layers, obtained with viscoelastic modeling, are reported in Figure 5. In the case of POPC/SM liposomes, the thickness of the adsorbed layer increased for the lower DCA concentrations and D
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
decreased sharply after the addition of 1000 μM, while the changes in the viscosity are negligible. These findings are corroborated by the results obtained in Langmuir monolayers by method 2 (Figure 3B), indicating that, at low surface contents, DCA forms an underneath adsorbed layer. The behavior of the POPC/SM/Chol liposomes is different. The addition of DCA led to a decrease in viscosity, which suggests that DCA might promote the fluidization of the raftlike domains in the liposome walls. The most significant effect was observed with DCA (1000 μM): the viscosity of the liposomes decreased markedly while the thickness suffered a small reduction. The fact that DCA decreases the viscosity of POPC/SM/ Chol bilayers slightly affecting its thickness while it decreases the thickness of POPC/SM bilayer without affecting its viscosity suggests that the key action of DCA may be the reduction of the bending stiffness of the bilayers, according to the findings of Rawicz et al.26 However, we could not confirm this hypothesis because we did not have access to the characterization of the mechanical properties of the membranes. The DSC curves of POPC/SM liposomes interacting with different concentrations of DCA (Figure S2A) reveal that increasing DCA concentration shifts the transition temperatures (Tt) to lower values, suggesting that DCA fluidizes the lipid bilayer. In the case of the POPC/SM/Chol liposomes where no transition was detected in the temperature range available, nothing could be concluded from the addition of
Figure 5. Changes in viscosity (A) and thickness (B) of the adsorbed liposome layers as a function of the concentration of the added DCA solution: POPC/SM (gray circles) and POPC/SM/Chol (black triangles).
Figure 6. 31P NMR spectra of DCA (A) and UDCA (B) interacting with POPC/SM (left) and POPC/SM/Chol (right) liposomes: liposomes (blue curve) and liposomes + BA (red curve). E
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir DCA. Since excess heat capacity values slightly decrease after the endothermic peak, partial lipid solubilization of the binary mixture may exist, although to much lower extent than suggested by QCM-D. It should be stressed that the obtained DSC curves allow only a qualitative comparison because, as explained before, the onset of the transition lies beyond the available temperature range. The interaction of both liposomes with DCA was studied with 31P NMR (Figure 6A) using a small relative concentration due to the sensitivity of the equipment. This means that the ratio between the number of DCA molecules and the number of liposomes is much smaller compared to those in the QCM-D and the DSC experiments. The addition of DCA to the POPC/SM liposomes enhances the low-field isotropic peak at 25 °C which may be attributed to some fluidization of the bilayer, in accordance with the DSC results, presumably due to the interaction with the solidordered gel SM-rich phase. At 37 °C, the POPC/SM liposomes are already in a liquid-disordered state and are practically not affected by DCA. In contrast, the effect of DCA on the POPC/ SM/Chol liposomes is very clear at all temperatures, being more intense at the highest temperature. In fact, at 37 °C, DCA yields not only the reduction of the anisotropy of the peaks but also the shift of the resonance position to higher fields. These results for the POPC/SM/Chol liposomes evidence that DCA, at low concentration, causes a perturbation in the membrane lipids, increasing the isotropic Ld phase, which gives rise to a higher mobility of their phosphate groups. The preferential interaction of DCA with monolayers of ternary composition was also verified when DCA was added to the spreading solution. Effect of UDCA on Monolayers. The experimental π−A isotherms obtained for the interaction of UDCA dissolved in the subphase (method 1) or spread at the interface (method 2) with both lipid monolayers (Figures S3) follow a similar trend to that observed with DCA in Figure 2, except in what concerns the amount of UDCA incorporated by method 1. In fact, the progressive deviation of the π−A isotherm of the lipid monolayer to larger areas with the increase in the BA content, is very modest when UDCA is dissolved in the subphase. This behavior is in agreement with the higher hydrophilic character of UDCA when compared to DCA. Actually, the π−A isotherm of UDCA (8 μM content in the subphase) in the absence of lipids at the surface remains at π ≈ 0 during the entire compression run (Figure S3A,B). Figure 7A shows the low incorporation of UDCA from the subphase in both lipid mixtures. However, the slight increase in the collapse surface pressure of the monolayer, observed in some cases, supports the interaction of UDCA with the polar heads of the phospholipids. When UDCA is spread on the interface with the lipid mixtures (Figure 7B), the relative deviation becomes more significant for the ternary mixture at XUDCA > 0.7. This observation is compatible with UDCA mainly adsorbing in the polar region of the lipid monolayer or dissolving into the subphase at low contents (XUDCA ≤ 0.5). For higher contents (XUDCA > 0.5), the exceeding UDCA molecules (that do not dissolve into the subphase and/or do not adsorb at the polar region) will contribute to the area increase of the lipid monolayer, with retention at the interface favored by the presence of cholesterol. Effect of UDCA on Liposomes. QCM-D experiments similar to those performed with DCA were done with UDCA
Figure 7. Relative area increase (%) of POPC/SM (gray circles) and POPC/SM/Chol (black triangles) monolayers on a HEPES buffer subphase at π = 15 mN/m, pH 7, and 25 °C as a function of the concentration of UDCA dissolved in the subphase (A) and of the molar fraction of UDCA in the spreading solution (B). The curves are guides for the eye.
(data not shown) and revealed practically no effect on the adsorbed liposomes of both compositions. The DSC curves (Figure S2B) show that both Tt and the enthalpy of the binary liposomes are slightly affected in the presence of the lowest concentration of UDCA (500 μM) but increase when the UDCA concentration increases. Analysis of the 31P NMR spectra (Figure 6B) shows that, at 37 °C, when UDCA is added to the POPC/SM liposomes, a high-field shoulder starts to develop which may be related to the slight increase in the bilayer stiffness detected in the DSC experiments. In contrast, no effect of UDCA can be detected for the POPC/SM/Chol liposomes. Effect of the DCA/UDCA (1:1) Mixture on Monolayers. The experimental π−A isotherms obtained for the interaction of the DCA/UDCA mixture dissolved in the subphase (method 1) or spread at the interface (method 2) with both lipid monolayers (Figures S4) follow a similar trend to that observed with DCA in Figure 2. The effect of the DCA/UDCA equimolar mixture on POPC/ SM and POPC/SM/Chol monolayers at π = 15 mN/m is shown in Figure 8A as a function of the molar concentration (CBA) in the subphase and in Figure 8B as a function of the molar fraction (XBA) in the spreading solution. The averaged effect calculated on the basis of the independent effect of each BA is also included (dashed lines). The incorporation of the BA mixture from the subphase is larger in the more expanded POPC/SM binary monolayer, mainly at high concentrations. The experimental deviations for the BA mixture are higher than the ones calculated by assuming the independent effect of each BA. F
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 9. Changes in viscosity (A) and thickness (B) of the adsorbed liposome layers as a function of the concentration of the added DCA/ UDCA (1:1) solution: POPC/SM (gray circles) and POPC/SM/Chol (black triangles). Curves are guides for the eye.
Figure 8. Relative area increase (%) of POPC/SM (gray circles) and POPC/SM/Chol (black triangles) monolayers on a HEPES buffer subphase at π = 15 mN/m, pH 7, and 25 °C as a function of the concentration of DCA/UDCA (1:1) dissolved in the subphase (A) and of the molar fraction of DCA/UDCA (1:1) in the spreading solution (B). The calculated average effect is also included (open symbols, dashed lines). The curves are guides for the eye.
However, when the BA mixture is added at the interface in the lipid spreading solution, the ternary mixed monolayer is slightly more affected (Figure 8B) for high contents of BAs. On the other hand, the experimental and calculated curves are close to each other, suggesting that the ability of UDCA, alone or mixed with DCA, to be incorporated by cospreading is similar. Effect of DCA/UDCA (1:1) on Liposomes. The interaction of the DCA/UDCA equimolar mixture with both types of liposomes leads to a decrease in frequency and an increase in dissipation for both types of liposomes (Figure S5). The great difference between both compositions is the effect of the final rinsing on the adsorbed layers. While the POPC/SM/ Chol liposomes after interacting with DCA/UDCA maintain their viscoelastic properties, the POPC/SM liposomes go back to their initial state. An analysis of the variation of the viscosity and thickness of the adsorbed liposomes (POPC/SM and POPC/SM/Chol) as a function of the concentration of the BA mixture (Figure 9) shows only a slight increase in the thickness for the ternary liposomes, which is more significant for the highest concentration. DSC experiments (data not shown) for POPC/SM liposomes show a slight increase in T t with increasing concentrations of BAs. A comparison of the effect of the BA mixture with that of pure BA is made in Figure 10. In the presence of 900 μM DCA/UDCA, Tt is greater than the one resulting from equal contributions of both BAs, suggesting that the rigidifying effect of UDCA in the mixture not only
Figure 10. Transition temperature of POPC/SM liposomes as a function of concentration of DCA, UDCA, and DCA/UDCA (1:1). The open squares represent the predicted transition temperature calculated from the average contribution of DCA and UDCA. The lines are guides for the eye.
superimposes the disorganizing effect of DCA but is larger than the effect of UDCA alone. The 31P NMR spectra (Figure 11) of POPC/SM liposomes interacting with the DCA/UDCA mixture, at 37 °C, reveal the same tendency as that observed with UDCA: a high-field shoulder suggesting a stabilizing effect of the liposomes which was also found in the DSC experiments. In the case of the POPC/SM/Chol liposomes, the effect of the BA mixture is similar to that observed with DCA (Figure 6A) but less intense, which should be due to the lower concentration of DCA, once no protecting effect of UDCA upon the ternary liposomes was found (Figure 6B). G
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
keeps the UDCA molecules mostly in solution. Surprisingly, the penetration of equimolar binary mixture DCA/UDCA in both lipid monolayers is much higher than the prediction based on the independent measurements of DCA and UDCA. This result suggests a cooperative DCA−UDCA interaction that enhances the penetration of UDCA in both POPC/SM and POPC/SM/Chol lipid monolayers. In method 2, the distribution of BA at the interface may be inferred, assuming that most of the BA added remains at the interface. For XBA < 0.5, the relative area deviation remains very low for the DCA, UDCA, or DCA/UDCA mixture. This suggests that BAs, single or mixed, preferentially interact with the lipid polar groups forming an underneath adsorbed layer which does not change the total area of the lipid monolayer significantly. For XBA > 0.5, the relative area deviation increases with the BA content, indicating that the BA molecules, in excess relatively to the lipid content, will contribute to the area increase of the lipid monolayer by incorporation or through the formation of BA segregated domains. A slightly larger effect was obtained for DCA due to its higher hydrophobicity. The increase in the relative area deviation is more pronounced for the ternary composition, suggesting that the cosolubilization of BAs with lipids favors the interaction with Chol in the lipid monolayer. On the other hand, the contribution of the DCA/ UDCA mixture is similar to the averaged contribution of single BAs, indicating that the lipid−BA interaction at the interface prevents BA dissolution into the subphase. These results confirm that the selectivity observed by method 1 is a consequence of the hydrophilic/hydrophobic balance in the whole system, subphase and monolayer, while in method 2 the contribution of the bulk subphase is less significant (CBA ≈ 0). The influence of BAs on adsorbed liposomes investigated by QCM-D agrees with the findings of Shubert et al.,31 who claimed that, in general, bile acids at very low concentrations bind to the outside layer of the liposomes, while at high concentrations they generate membrane stresses leading to eventual lipid solubilization. At low concentration, DCA led to a slight reduction of the viscosity of the POPC/SM/Chol liposome layer but did not affect that of the POPC/SM liposome layer. At high concentration, DCA significantly decreased the viscosity of the POPC/SM/Chol liposomes and disturbed the POPC/SM liposomes, by eventual solubilization of some lipids (Figure 5). These results suggest that monomers and/or aggregates of DCA penetrate the hydrophobic region of liposomes, increasing their hydration level and eventually damaging the continuous bilayer structure at high concentration. In fact, previous work32 using watersensing fluorescent probes in lipid bilayer membranes indicate that the submicellar concentration of bile salts induces the hydration of the lipid bilayer membrane in the core region, which explains the observed decrease in the viscosity. This effect was found to increase with the concentration and hydrophobicity of the bile salt as a result of a more efficient interaction with the bilayer membrane. Moreover, at 37 °C, the fluidity of the lipid bilayer of the POPC/SM liposomes (Tt ≈ 23 °C) is responsible for a higher degree of hydration as compared to the POPC/SM/Chol liposomes where the Lo phase predominates the Ld phase. Therefore, it is conceivable that the hydration effect of DCA is much smaller upon the prehydrated layer of POPC/SM liposomes. The addition of UDCA practically did not affect the adsorbed liposomes of both compositions (POPC/SM and POPC/SM/Chol), probably due to its lower hydrophobicity
Figure 11. 31P NMR spectra of DCA/UDCA (1:1) interacting with POPC/SM (left) and POPC/SM/Chol (right) liposomes: liposomes (blue curve) and liposomes + DCA/UDCA (red curve).
■
DISCUSSION The action of bile acids on membrane fluidity is known to depend strongly on the number and positioning of free hydroxyl groups.27,28 In the case of bile acids DCA and UDCA, the molecular composition is the same, differing only in the position of the two hydroxyl groups linked to the steroid nucleus: DCA has both groups on the same side, which favors the formation of dimers or oligomers in aqueous solution;29 UDCA has the OH groups on opposite sides. This slight difference has been known to confer different behaviors with respect to solubilization, emulsification, or interaction with lipid membranes.30 Generally, one can infer from this work that the relative impact of DCA and UDCA on the properties of the lipid membranes depends on the membrane model which is directly related to the experimental technique used. The Langmuir monolayer, being one-half of the bilayer structure of biomembranes, does not have the global integrity of a cell membrane, but it is suitable for localizing interactions on the polar or nonpolar regions. DCA penetrates the POPC/SM and POPC/SM/Chol lipid monolayers at low surface pressures using method 1. This penetration increases with the BA content in the subphase and is more pronounced in the absence of Chol. Under the same conditions, UDCA has a small impact on both lipid compositions probably because of its hydrophilicity that H
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
13 μM it is able to penetrate the monolayers. At low BA concentrations, our 31P NMR results are in agreement with the observations of Mello-Vieira et al.,4 who found that in the absence of Chol both cytotoxic and cytoprotective BAs were unable to affect acyl-chain dynamics. These authors also found that DCA led to an increase in the fluidity of Chol-containing membranes, which somehow does not seem compatible with their claim that this BA presents very limited insertion into the membrane interior. In our case, we concluded that DCA has a greater effect on the POPC/SM/Chol liposomes than on the POPC/SM liposomes. As the POPC/SM/Chol liposomes have a more condensed structure, this may mean that more important than the degree of cohesion of the membrane is the preferential interaction of DCA with Chol.
which weakens the hydrophobic interaction with the core region. The effect of the equimolar DCA/UDCA mixture is translated only into a slight increase in the thickness of adsorbed POPC/SM/Chol liposomes (Figure 9). This may reflect the cytoprotective action of UDCA: due to the preferential interactions with UDCA, DCA is retained in the adsorbed layer in the external polar region of liposomes. DSC measurements show that while the POPC/SM liposomes are disturbed by DCA, mainly at high concentrations, they become more organized in the presence of UDCA. The observed increases in both Tt and the enthalpy of the binary liposomes suggest that the adsorption of UDCA to the lipid polar heads may be responsible for the decrease in the hydrocarbon chain ordering, encouraging the formation of an interdigitated phase at high concentrations. Similar behavior was previously observed by Tomaia-Cotisel et al.33 for DPPC liposomes in the presence of UDCA. In the case of the mixture, the results show that the presence of DCA enhances the stabilizing effect of UDCA, contributing to the appearance of an interdigitated gel phase, even at a low UDCA concentration. 31 P NMR results, obtained at a low concentration of BAs, show that DCA did not significantly affect the 31P NMR spectra of POPC/SM liposomes, while both UDCA and the equimolar DCA/UDCA mixture slightly stabilize the binary liposomes. This suggests that, in the absence of Chol, the expanded organization of the phosphate groups is not affected by the small content of DCA, but it is slightly stabilized by the direct interaction with the adsorbed layer of UDCA or the DCA/ UDCA mixture. In contrast, the 31P NMR spectra of POPC/SM/Chol liposomes show that DCA increases the mobility of the phosphate groups in liposomes, while no effect was found for UDCA on these liposomes. If DCA is mixed with UDCA in a 1:1 proportion (retaining the total BA concentration), the impact on the ternary membrane is similar to but less intense than that of DCA, meaning that the presence of UDCA does not completely hinder the effect of DCA. The mechanism of the interaction of BAs with Chol-containing systems is complex because POPC/SM/Chol liposomes contain both Lo and Ld phases. The expansion of the polar heads region suggested by the observed increase in the phosphate mobility may be due to the penetration of DCA inside the condensed organization of lipids in the Lo (lipid raft) phase. Alternatively, DCA may lead to the fluidization of the Ld phase, which then promotes the formation of Lo domains, as suggested by Heerklotz et al.34 when studying the action of another detergent, Triton X-100, upon POPC/SM/Chol liposomes. On the contrary, the weak interaction of adsorbed UDCA with the polar region of the liposomes does not significantly affect the already low mobility of phosphates in the raft dense organization. A comparison of our results with others reported previously4,5 confirms a strong dependence on both the BA concentration and lipid composition of the membrane. In a recent work, Zhou et al.5 reported a quite different impact of BAs at cytotoxic concentrations (near the critical micellar concentration) or at subtoxic concentrations on liposomes of different compositions. For example, DCA at 1000 μM led to the solubilization of DOPC liposomes, but a 10 mM DCA treatment had only a disordering effect on the DOPC/Chol (1:1) liposomes and no effect on liquid-ordered phase DPPC/ Chol (1:1) liposomes. Our results reveal that DCA at 1000 μM disturbs the packing of the acyl chains of the POPC/SM liposomes, in the absence or in the presence of Chol, while at
■
CONCLUSIONS This work confirmed that the cytotoxic DCA at high concentrations has the capability to penetrate the lipid membranes (with and without Chol), leading to an increase in the fluidity of the acyl chains or even to lipid solubilization. At low concentrations, DCA has a low impact on the membrane properties, but it affects the dense structure of Chol-containing membranes more. The hydrophilicity of UDCA, commonly associated with its cytotoprotective nature, should promote its preferential interaction in the external polar region with a rigidifying effect on POPC/SM lipid membranes. The most surprising results were obtained with the DCA/ UDCA equimolar mixture. At high concentrations, the presence of UDCA protects both POPC/SM and POPC/SM/Chol liposomes against the destabilizing effect of DCA. However, in the case of the monolayers, the penetration of mixture DCA/ UDCA for both lipid compositions is much higher than the average of the independent effects of DCA and UDCA, suggesting a cooperative effect of both acids. Further studies are needed to understand the interaction mechanism between DCA and UDCA that causes a synergistic effect of the two bile acids on the stabilization of the lipid membrane.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01702. π−A isotherms of POPC, SM, and Chol on a HEPES buffer subphase. Effect of DCA concentration on the thermotropic behavior of POPC/SM liposomes. Effects of the UDCA concentration and the DCA/UDCA mixture on the π−A isotherms of POPC/SM and POPC/SM/Chol mixtures. Frequency shifts and dissipation shifts of the third harmonic obtained after the addition of DCA/UDCA to POPC/SM and of POPC/ SM/Chol liposomes. (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. *E-mail:
[email protected]. Notes
The authors declare no competing financial interest. I
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
■
(17) Roda, A.; Minutello, A.; Angellotti, M. A.; Fini, A. Bile acid structure-activity relationship: evaluation of bile acid lipophilicity using 1-octanol/water partition coefficient and reverse phase HPLC. J. Lipid Res. 1990, 31, 1433−1443. (18) Viitala, T.; Hautala, J. T.; Vuorinen, J.; Wiedmer, S. K. Structure of anionic phospho-lipid coatings on silica by dissipative quartz crystal microbalance. Langmuir 2007, 23, 609−618. (19) Jurak, M. Thermodynamic aspects of cholesterol effect on properties of phospholipid monolayers: Langmuir and Langmuir− Blodgett monolayer study. J. Phys. Chem. B 2013, 117, 3496−3502. (20) Stottrup, B. L.; Stevens, D. S.; Keller, S. L. Miscibility of ternary mixtures of phospholipids and cholesterol in monolayers, and application to bilayer systems. Biophys. J. 2005, 88, 269−276. (21) Ohvo-Rekilä, H.; Ramstedt, B.; Leppimäki, P.; Slotte, J. P. Cholesterol interactions with phospholipids in membranes. Prog. Lipid Res. 2002, 41, 66−97. (22) Wydro, P.; Flasiński, M.; Broniatowski, M. Does cholesterol preferentially pack in lipid domains with saturated sphingomyelin over phosphatidylcholine? A comprehensive monolayer study combined with grazing incidence X-ray diffraction and Brewster angle microscopy experiments. J. Colloid Interface Sci. 2013, 397, 122−130. (23) Wydro, P. Sphingomyelin/phosphatidylcholine/cholesterol monolayers − analysis of the interactions in model membranes and Brewster Angle Microscopy experiments. Colloids Surf., B 2012, 93, 174−179. (24) Serro, A. P.; Carapeto, A.; Paiva, G.; Farinha, J. P. S.; Colaço, R.; Saramago, B. Formation of an intact liposome layer adsorbed on oxidized gold confirmed by three complementary techniques: QCMD, AFM and confocal fluorescence microscopy. Surf. Interface Anal. 2012, 44, 426−433. (25) Bandeiras, C.; Serro, A. P.; Luzyanin, K.; Fernandes, A.; Saramago, B. Anesthetics interacting with lipid raft. Eur. J. Pharm. Sci. 2013, 48, 153−165. (26) Rawicz, W.; Olbrich, K. C.; McIntosh, T.; Needham, D.; Evans, E. Effect of chain length and unsaturation on elasticity of lipid bilayers. Biophys. J. 2000, 79, 328−339. (27) Sreekanth, V.; Bajaj, A. Number of free hydroxyl groups on bile acid phospholipids determines the fluidity and hydration of model membranes. J. Phys. Chem. B 2013, 117, 12135−12144. (28) Singh, M.; Bajaj, A. Unraveling the impact of hydroxylation on interactions of bile acid cationic lipids with model membranes by indepth calorimetry studies. Phys. Chem. Chem. Phys. 2014, 16, 19266− 19274. (29) Pártay, L. B.; Jedlovszky, P.; Sega, M. Molecular aggregates in aqueous solutions of bile acid salts. Molecular Dynamics simulation study. J. Phys. Chem. B 2007, 111, 9886−9896. (30) Sugihara, G.; Hirashima, T.; Lee, S.; Nagadome, S.; Takiguchi, H.; Sasaki, Y.; Igimi, H. The adsorption behavior of four bile salt species on cholesterol crystal in water − in connection with cholesterol solubilization. Colloids Surf., B 1995, 5, 63−73. (31) Schubert, R.; Schmidt, K. Structural changes in vesicle membranes and mixed micelles of various lipid compositions after binding of different bile salts. Biochemistry 1988, 27, 8787−8794. (32) Mohapatra, M.; Mishra, A. K. Effect of submicellar concentrations of conjugated and unconjugated bile salts on the lipid bilayer membrane. Langmuir 2011, 27, 13461−13467. (33) Tomoaia-Cotisel, M.; Levin, I. W. Thermodynamic study of the effects of ursodeoxycholic acid and ursodeoxycholate on aqueous dipalmitoyl phosphatidylcholine bilayer dispersions. J. Phys. Chem. B 1997, 101, 8477−8485. (34) Heerklotz, H. Triton promotes domain formation in lipid raft mixtures. Biophys. J. 2002, 83, 2693−2701.
ACKNOWLEDGMENTS This work was supported by Fundaçaõ para a Ciência e Tecnologia (FCT), National NMR Network (RECI/BBBBQB/0230/2012), and project UID/QUI/UI0100/2013.
■
REFERENCES
(1) Hofmann, A. F.; Hagey, L. R. Bile acids: chemistry, pathochemistry, biology, pathobiology, and therapeutics. Cell. Mol. Life Sci. 2008, 65, 2461−2483. (2) Attili, A. F.; Angelico, M.; Cantafora, A.; Alvaro, D.; Capocaccia, L. Bile acid-induced liver toxicity: relation to the hydrophobichydrophilic balance of bile acids. Med. Hypotheses 1986, 19, 57−69. (3) Im, E.; Martinez, J. D. Ursodeoxycholic acid (UDCA) can inhibit deoxycholic acid (DCA)-induced apoptosis via modulation of EGFR/ Raf-1/ERK signaling in human colon cancer cells. J. Nutr. 2004, 134, 483−486. (4) Mello-Vieira, J.; Sousa, T.; Coutinho, A.; Fedorov, A.; Lucas, S. D.; Moreira, R.; Castro, R. E.; Rodrigues, C. M. P.; Prieto, M.; Fernandes, F. Cytotoxic bile acids, but not cytoprotective species, inhibit the ordering effect of cholesterol in model membranes at physiologically active concentrations. Biochim. Biophys. Acta, Biomembr. 2013, 1828, 2152−2163. (5) Zhou, Y.; Maxwell, K. N.; Sezgin, E.; Lu, M.; Liang, H.; Hancock, J. F.; Dial, E. J.; Lichtenberger, L. M.; Levental, I. Bile acids modulate signaling by functional perturbation of plasma membrane domains. J. Biol. Chem. 2013, 288, 35660−35670. (6) Mouaz, A. B.; Lindheimer, M.; Montet, J.; Zajac, J.; Lagerge, S. A study of the adsorption of bile salts onto model lecithin membranes. Colloids Surf., B 2001, 20, 119−127. (7) Fahey, D. A.; Carey, M. C.; Donovan, J. M. Bile acid phosphatidylcholine interactions in mixed monomolecular layers − Differences in condensation effects but not interfacial orientation between hydrophobic and hydrophilic bile-acid species. Biochemistry 1995, 34, 10886−10897. (8) Leonard, M. R.; Bogle, M. A.; Carey, M. C.; Donovan, J. M. Spread monomolecular films of monohydroxy bile acids and their salts: Influence of hydroxyl position, bulk pH, and association with phosphatidylcholine. Biochemistry 2000, 39, 16064−16074. (9) Hildebrand, A.; Neubert, R.; Garidel, P.; Blume, A. Bile salt induced solubilization of synthetic phosphatidylcholine vesicles studied by isothermal titration calorimetry. Langmuir 2002, 18, 2836−2847. (10) Subuddhi, U.; Mishra, A. K. Effect of sodium deoxycholate and sodium cholate on DPPC vesicles: A fluorescence anisotropy study with diphenylhexatriene. Proc. - Indian Acad. Sci., Chem. Sci. 2007, 119, 169−174. (11) Zhou, Y.; Doyen, R.; Lichtenberger, L. M. The role of membrane cholesterol in determining bile acid cytotoxicity and cytoprotection of ursodeoxycholic acid. Biochim. Biophys. Acta, Biomembr. 2009, 1788, 507−513. (12) McMullen, T.; Lewis, R.; McElhaney, R. Cholesterolphospholipid interactions, the liquid ordered phase and lipid rafts in model and biological membranes. Curr. Opin. Colloid Interface Sci. 2004, 8, 459−468. (13) Holland, G. P.; McIntyre, S. K.; Alam, T. M. Distinguishing individual lipid headgroup mobility and phase transitions in raftforming lipid mixtures with 31P MAS NMR. Biophys. J. 2006, 90, 4248−4260. (14) Giocondi, M.; Boichot, S.; Plenat, T.; Le Grimellec, C. Structural diversity of sphingomyelin microdomains. Ultramicroscopy 2004, 100, 135−143. (15) Prenner, E.; Honsek, G.; Hönig, D.; Möbius, D.; Lohner, K. Imaging of the domain organization in sphingomyelin and phosphatidylcholine monolayers. Chem. Phys. Lipids 2007, 145, 106−118. (16) Avanti Polar Lipids-Preparation of Liposomes. http://www. avantilipids.com/index.php?option=com_content&view=article&id= 1384&Itemid=372. J
DOI: 10.1021/acs.langmuir.5b01702 Langmuir XXXX, XXX, XXX−XXX
