Biochimica et Biophysica Acta 1848 (2015) 2154–2162

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamem

brought to you by COREView metadata, citation and similar papers at core.ac.uk

provided by Elsevier - Publisher Connector

Antagonistic effects of α-tocopherol and ursolic acid on modelbacterial membranes

Marcin Broniatowski ⁎, Michał Flasiński, Katarzyna Hąc-WydroDepartment of Environmental Chemistry, Faculty of Chemistry, Jagiellonian University, ul. Gronostajowa 3, 30-387 Kraków, Poland

⁎ Corresponding author. Tel.: +48 126646795.E-mail address: broniato@chemia.uj.edu.pl (M. Bronia

http://dx.doi.org/10.1016/j.bbamem.2015.05.0090005-2736/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 18 February 2015Received in revised form 27 April 2015Accepted 13 May 2015Available online 21 May 2015

Keywords:TocopherolUrsolic acidLangmuir monolayersE. coli membranePentacyclic triterpenes

α-tocopherol (Toc), the most active component of vitamin E can exert antagonistic effects disabling the therapyof cancers and bacterial infections. Such antagonisms were observed also between Toc and bioactive pentacyclictriterpenes (PT) exhibiting anticancer and antibacterial properties. Both Toc and PT are water-insoluble mem-brane active substances. Thus, our ideawas to emulate their interactionswithmodel Escherichia colimembranes.E. coli inner membranes were selected for the experiments because their lipid composition is quite simple andwell characterized and the two main components are phosphatidylethanolamine and phosphatidylglycerol. Asa model of E. colimembranes we applied Langmuir monolayers formed by the E. coli total extract of polar lipids(Etotal) as well as by the main lipid components: phosphatidylethanolamine (POPE) and phosphatidylglycerol(ECPG). The antagonistic effects of ursolic acid (Urs) and Tocwere investigatedwith the applicationof ternary Lang-muirmonolayers formed by Urs, Toc and one of the phospholipids POPE or ECPG. Our studies indicated that the af-finities of Urs and Toc towards the POPE molecule are comparable; whereas there are profound differences in theinteractions of Urs and Toc with ECPG. Thus, the model experiments prove that in the case of E. coli membrane,the differences in the interactions between Urs and Toc with the anionic bacterial phosphatidylglycerol can bethe key factor responsible for the antagonistic effects observed between PT and Toc in vivo.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Ursolic acid (Urs) is a representative of ursane-type pentacyclictriterpenes (PT), a subfamily of the vast group of pentacyclic bioactiveplant secondary metabolites [1]. The biodiversity of PT synthesizedby plants, fungi and algae is enormous [2]; however, only some of thecompounds have been proved in laboratory and clinical tests to be phar-maceutically active [3]. Ursolic acid exhibits wide spectrum of antican-cer [3–5] as well as antimicrobial activities [6–9]. The anticancer andchemopreventive properties of Urs are often linked to the triggering ofapoptosis via the so calledmitochondrial tract [10–12]. The mechanismof the antimicrobial activity of Urs is less elucidated; however, it is as-sumed that there are common steps in the interactions of this com-pound with cancer cells and bacteria. The triggering of the apoptosisprocess, on one hand, is beneficial as far as the elimination of cancercells is concerned, on the other hand, the apoptosis or necrosis of thesomatic body cells can be the starting point of different diseases [13,14]. Reactive oxygen species (ROS) can be indicated as one of the maincause of degenerative diseases initiated by the uncontrolled apoptosis[13,14]. The living organisms combat the uncontrolled apoptosis by theapplication of different antioxidants, such as ubiquinone, glutathione,

towski).

vitamins C or E [15,16]. The application of antioxidant vitamins can bebeneficial in the therapies of different degenerative diseases; however,it can be deleterious in the treatment of cancer or bacterial infections.An example can be the work by Gopal and co-authors [17], in which itwas indicated thatα-tocopherol (Toc) eliminates the proapoptotic effectof the active pentacyclic triterpene, betulinic acid. It was also reportedthat the therapeutic effects of the application of some antibiotics can belost in the case of simultaneous application ofα-tocopherol [18]. Themo-lecularmechanismof PT pharmacological activity is not entirely elucidat-ed; however, one of its most important stages is the incorporation of PTinto the mitochondrial or bacterial membrane [6,19,20]. On the otherhand Toc also exhibits its antioxidative and regulatory functions in themembrane environment [21–23]. Thus, the antagonistic effects occurringbetween α-tocopherol and PT can be connected with their interactionswith membrane phospholipids. In our previous papers we reported theresults of the comparative studies regarding the behavior of different bi-ologically active PT in artificial mitochondrial and bacterial membranes[24,25]. As membrane models we applied Langmuir monolayers beingquaternary or ternary mixtures of mitochondrial or E. coli membranephospholipids, respectively. It turned out that Urs and its reduced formα-amyrin had the highest ability to destruct the model E. coli innermembrane.

In the present contribution we report the results regarding theinteractions of Urs and Toc with model E. coli membrane and its

2155M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

phospholipid components. As amodel of E. colimembranewe proposeda Langmuir monolayer formed by the E. coli polar lipids total extract.At the first step we wanted to compare the interactions of Urs vsToc in the mixed monolayers with E. total. The composition of E.coli inner membrane lipids is quite simple: it contains 75% of phos-phatidylethanolamine (PE), 20–25% of phosphatidylglycerol (PG)and 0–5% of cardiolipin [26,27]. The proportion of the saturated tounsaturated chains in E. coli PE is roughly 1/1 and the dominatingfatty acid chains are palmitic and oleic [27], so 1-palmitoyl, 2-oleoyl-sn-3-phosphatidylethanolamine (POPE) was selected by us toour research, whereas E. coli PG (ECPG) applied by us was a naturalphospholipid extracted from E. coli. After the characterization of the sys-tems with Etotal we studied the interactions of Urs and Toc with POPEand ECPG in binary systems. At the final step we studied ternary Lang-muir monolayers formed by equimolar amounts of Urs and Toc withPOPE or ECPG to search the possible antagonistic effects in the simpli-fied model systems. In our studies we applied the Langmuir techniquebased on the recording of surface pressure (π) – mean molecular area(A) isotherms supplemented by Brewster angle microscopy (BAM).We analyzed the changes in monolayer condensation interpreting thecompression modulus (CS

−1) – surface pressure (π) curves anddiscussed the thermodynamic effects of the intermolecular interactionsin the model systems performing the analysis of the excess free energyof mixing ΔGexc – composition dependences. The Langmuir monolayertechnique was previously applied for the studies of Toc-phospholipidinteractions; however, the systems intended to mimic animal [28,29]or plant membranes [30]. The interactions of Toc with bacterial mem-brane phospholipids has not been modeled by now. Thus, the main ob-jectives of our research were the studies of the interactions of Toc withmixed and separated E. coli lipids in Langmuir monolayers, the compar-ison of the interactions between Toc and Urs with the bacterial lipids,and the application of model systems for the elucidation of the antago-nistic effects observed betweenUrs and Toc in pharmacological applica-tions. The understanding of the antagonistic effects of Toc and somedrugs is of high importance as such phenomena diminish the efficacyof antibacterial and anticancer therapies. Our studies revealed signifi-cant differences between Urs and Toc in the investigated systems andit is probable that the observed antagonistic effects are membrane-related.

2. Experimental

2.1. Materials

All the applied phospholipids: E. coli polar lipids total extract(Etotal), E. coli natural phosphatidylglycerol (ECPG) sodium saltand 1-palmitoyl-2-oleoyl-sn-glycerophosphatidylethanolamine (POPE)were purchased from Avanti Polar Lipids. Ursolic acid (Urs, 99%) and α-tocopherol (DL-all-rac-α-tocopherol, Toc, N99%, analytical standard)were purchased from Sigma Aldrich. HPLC-grade chloroform (ethanolstabilized) and HPLC-grade methanol were bought also from SigmaAldrich. Ultra-pure water (resistivity ≥18.2 MΩ∙cm) was producedin our laboratory with the application of the Synergy Millipore waterpurification system. Structural formulas of POPE, ECPG, Urs and Toc arepresented in Supporting materials.

2.2. Preparation of surfactant solutions

The investigated phospholipids, ursolic acid and tocopherol weredissolved in chloroform/methanol 9/1 v/vmixture and kept refrigeratedin −20 °C. The concentrations of the substances ranged from 0.2 to0.3 mg/ml, which for the phospholipids gave molar concentrations ofca. 3 to 4°10−4 M, for Urs and Toc 4 to 6°10−4 M. To prepare the worksolutions appropriate volumes of the stock solutions were mixed indark-glass vials just before the experiments. For the mixtures of Ursand Toc with Etotal, POPE and ECPG we investigated the following

mole ratios of Urs (or Toc): 0.20, 0.34, 0.50, 0.66 and 0.8. For the ternarysystems containing both Urs, Toc and one of the phospholipids (POPEor ECPG) the above mole proportions of Urs + Toc/phospholipid werepreserved. For all the ternary mixtures the number of Urs and Toc mol-ecules at the air/water was equal, e.g. in the system Urs/Toc/POPE atX(POPE) = 0.50 X(Urs) was 0.25 and X(Toc) was also 0.25.

It should be underlined that in biomembranes the tocopherol/phospholipids ratio is small and do not exceed 0.1. In our paper the in-vestigated Toc and Urs proportions are significantly higher ranging upto 0.8. However, it should be taken into consideration that our studiesare performed on artificial systems,which are composedof some select-ed membrane lipids but are far from the real membrane compositionand dynamics (for example no proteins are present in the investigatedsystems).We are mainly interested in themutual interactions of select-ed lipids in artificial matrices of Langmuir monolayers; therefore for theunderstanding of the intermolecular interactions muchwider ranges ofthe component proportions were applied.

2.3. Methods

2.3.1. Langmuir techniqueIn our experiments we applied KSV-NIMA Langmuir trough of the

nominal area of 273 cm2 equipped with two Delrin barriers enablingsymmetrical compression of the monolayers. BAM experiments wereperformed on a larger instrument of the area 840 cm2 designated byKSV for microscopic experiments. Surface pressure (π) was monitoredwith the application of Wilhelmy-type tensiometer (NIMA, KSV) withthe accuracy of 0.1 mN/m. The Wilhelmy sensor plate was made of fil-tration paper (Whatman, ashless). After each experiment the Teflontrough was cleaned with chloroform and ethanol and rinsed withultra-pure water. Appropriate amounts of the work solutions were de-posited at the air/water interface with the application of Hamilton vol-umetric gas-tight micro-syringe. 10 minutes were left for chloroformevaporation after which the monolayer compression was started. Inthe experiments we applied a rather slow compression rate of 10 mm/minute/barrier. The experiments for a given components proportionwere replied at least three times. The subphase temperature was keptconstant at 20 °C using the Julabo water circulating bath (accuracy0.1 °C).

2.4. Calculations

Compression modulus CS−1 was calculated according to its defini-tion: CS−1 = -Adπ/dA. The excess free energy of mixing ΔGexc [31] wascalculated at given π values (5, 10, 15 and 20 mN/m) according to itsdefinition:

ΔGexc ¼Zπ

0

Aexcdπ;

where the excess molecular area Aexc is defined as follows:

Aexc ¼ A1X1 þ A2X2 þ…þ AnXnð Þ–A1 ::n;

whereA1…n is the mean molecular area observed in an n-componentLangmuir monolayer at a given π value; A1 to An are molecular areasobserved in one-component monolayers of components 1 to n at thesame π; whereas X1 to Xn are mole ratios of components 1 to n in then-component Langmuir monolayer. The sum of the surface mole ratios:X1 + X2 + … + Xn = 1.

2.5. BAM

Brewster angle microscopy experiments were performed withultraBAM instrument (Accurion GmbH, Goettingen, Germany) equipped

2156 M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

with a 50mWlaser emitting p-polarized light at awavelength of 658 nm,a 10 xmagnification objective, polarizer, analyzer and a CCD camera. Thespatial resolution of the BAMwas2 μm. The instrumentwas coupledwiththe KSV 2000 Langmuir trough and installed on an antivibration table.

3. Results and discussion

3.1. One component monolayers

At the beginning of the results section we would like to summarizethe data regarding one-component monolayers formed by the investi-gated compounds. With the exception of Etotal the data has alreadybeen published but are scattered in different publications [25,30] andwe think that it is reasonable to collect them together herein. Fig. 1gathers π-A isotherms, CS−1-π curves and representative BAM imagesfor the investigated substances.

The π-A isotherm registered by us for pure tocopherol monolayershas an expanded character so the calculated CS−1 values do not exceed100 mN/m. The monolayers were completely homogeneous in BAMimages. Toc was investigated previously in Langmuir monolayers andthe most complete data together with thermodynamic parameters ofthe spreading can be found in the paper by Capuzzi and co-authors[32]. The isotherm presented in the cited paper has also an expandedcharacter and the CS

−1 – π values are also lower than 100 mN/mwhich according to the Davis and Rideal criterion [33] classifies themonolayers as liquid-expanded. Capuzzi et al. [32] undertook specialprecautions in their studies –the chloroform solutions of Tocwere sealedunder argon, thewater applied as a subphasewas de-oxygenized and theexperimentwas performed in a nitrogen-enriched atmosphere. Thus, theπ-A isotherm registered by Capuzzi et al. can be treated as a reference forToc monolayer studies. The parameters of the isotherm are as follows:lift-off area at 60 Å2/molecule, collapse at 20 mN.m at 48 Å2/molecule.In comparison the parameters of Toc isotherm registered by us are: lift-off at 72 mN/m and collapse at 20.4 mN/m at 54.2 Å2/molecule. Patiland Cornwell [34] investigated the oxidation of α-tocopherol in Lang-muir monolayers. According to their results in acidic pH Toc is graduallyoxidized to tocopherolquinone; whereas in alkaline conditions a dimericspirodienone is the product of Toc oxidation. Tocopherolquinone formsalso expanded Langmuir monolayers but its π-A isotherm is shifted tohigher mean molecular areas (lift-off at ca. 100 mN/m). It should beunderlined that acidic or alkaline subphase pH stimulates Toc oxidation;whereas in neutral pH this process is rather low. Generally, there are twoattitudes: 1) experiments with special precautions: degassed water assubphase, nitrogen or argon flow over the Langmuir trough (however,nobody performed the experiments in a gas-tight container), 2) experi-ments with no precautions: application of ultrapure water and no inert

Fig. 1. π-A isotherms, CS−1 – π curves and BAM images for one-compon

gases flushing the water/air interface. The second attitude dominates inthe newer papers [28–30]. In our opinion the possibility of Toc oxidationin monolayers should always be taken under consideration; however, ifmixed systems are investigated at neutral pH, the stock solutions are pre-pared just before the experiments and the compression starts in a shorttime after film spreading the presence of Toc osidation products can beneglected.

Urs behaves differently at the water/air interface. As it is visible inthe course of the CS−1-π curve, upon the film compression CS−1 achievesthe values exceeding 100mN/m,which is characteristic of the LC (liquidcondensed) state of Langmuirmonolayers. However, the phase diagramof the PT acids in monolayers is more complicated. These compoundsare bolaamphiphiles, that is, surfactants having two polar headgroupsat both termini of their hydrophobic moiety. As it was discussed in sci-entific literature [35,36] a PT acidmolecule can achieve two competitiveorientations at the interface: upright with the –OH group in contactwith water and tilted with the –COOH group immersed in water. Itwas reported [36] that the uprightmolecules group and form hexagonalcrystalline domains, the state of which is solid (S); whereas the tiltedmolecules form LE regionswithin the film. As a result of the coexistenceof the crystalline and amorphous phase the microscopic texture of thefilms is heterogeneous. Large crystalline domains are surrounded bydark areas identified as the LE phase (see BAM image in Fig. 1). Regard-ing the investigated phospholipids, their monolayers are highly com-pressible and at room temperature are in the LE state. For Etotal atsurface pressures greater than 20 mN/m CS

−1 exceeds 100 mN/m,which is a threshold of the LC state [33]; however, no condense domainswere observed in BAM images and the film remained homogeneous,even behind the collapse. Only for POPE a LE/LC phase transition wasobserved both in the π-A isotherm course and in BAM images (Fig. 1).At ca. 37 mN/m characteristic flower-like domains nucleated in thehomogeneousfilm. For ECPG themonolayerswere completely homoge-neous upon their compression and even behind the collapse no textureswere visible. All the investigated monolayers were stable and the regis-tered π-A isotherms were highly reproducible.

4. Interaction of Toc and Urs with E. coli polar lipids total extract

At the first step of our studies we focused attention on the interac-tions of Toc and Urs with Etotal. The isotherms are presented in theSupplementary materials (SFig. 1.), here in Fig. 2. we present the dataextracted from the isotherms, that is CS−1-π curves and collapse pressure(πcoll) – X dependences.

Both the CS−1-π and πcoll – X curves indicate that the systems withEtotal differ significantly. In the systemToc/Etotal the collapse pressuresfall down systematically with growing X(Toc). The strong dependence

ent Langmuir monolayers formed by the investigated substances.

Fig. 2. CS−1-π curves and πcoll-X dependencies in the systems with Etotal.

2157M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

of πcoll on the binary film composition corroborates themutualmiscibil-ity of the components within the film. No condensation effect was ob-served after mixing both components. The maximal values of CS

−1 inthis system can roughly be considered weighted averages of theCS−1max values observed for Toc and Etotal. On the contrary, in the sys-tem Urs/Etotal strong condensation effect can be observed. The com-pression modulus values for the mixtures are higher than for the purecomponents. The condensation effect is pronounced especially at higherUrs proportions, as for X(Urs) = 0.8 CS−1 max achieves the value of 450mN/m. Such high values of the compression modulus indicate that themolecules in the monolayers achieved the tight packing characteristicfor the solid (S) state of the surface film. For this system also the collapsepressures are composition dependent – a slow increase of πcoll withgrowing X(Urs) can be observed. To have the thermodynamic informa-tion about the properties of the binary systemswe calculated the 2D ex-cess free energies ofmixingΔGexc. TheΔGexc-X plots at different surfacepressures are shown in Supportingmaterials (SFig. 2), here in a columnplot (Fig. 3) we compare theΔGexc values for both systems at 20mN/m.

In the system Toc/Etotal the ΔGexc values for all the investigatedcompositions are close to 0. This together with the strong πcoll-composition dependence corroborates the ideal miscibility of the com-ponents in these mixtures. In the system Urs/Etotal strong negative de-viations from ideality can be observed. This means that the substancesare miscible and that the interactions between Urs and the componentsof Etotal are more energetically beneficial than the interactions in thepure Urs monolayer.

Fig. 3. The comparison of the ΔGexc values at 20 mN/m for the systems with Etotal. Thelabels on the plot mean the mole ratio of Toc or Urs.

5. Interactions of Toc and Urs with main membrane phospholipidsof E. coli

In the previous sections the character of the interactions of Toc andUrs with the total extract of E. coli polar phospholipids have been re-vealed. To go further it was reasonable to perform similar investigationsin binary systems composed of a selected membrane phospholipid andUrs or Toc. E. coli is the most thoroughly investigated organism ofall bacteria, so the composition of its membrane was also discussed inmultiple publications. It is composed mainly of phosphatidylethanol-amine (75%), phosphatidylglycerol (20–25%), and the third component(1–5%) is cardiolipin [26,27,37]. Multiple studies reported that the pro-portion of saturated to unsaturated chains in E. coli PE is roughly 1/1 andthat the dominating saturated fatty acid is palmitic and themain unsat-urated oleic [27]; therefore, POPE can be treated as the bestmodel E. coliPE. On the other hand, E. coli anionic lipids usually contain non-typicalcarboxylic chains in their hydrophobic moiety. Often the cyclopropanering is incorporated into the chain instead of the cis-double bond typicalto eukaryotic organisms [38,39]. Therefore, for our studies we appliedsynthetic PE, but natural PG (ECPG) isolated from E. coli. ECPG is anexample of bacterial phospholipids containing the cyclopropane ringin one of its chains (For formulas see Supporting materials, Scheme 1).For the four systems – mixtures of Toc or Urs with POPE or ECPGwe alsomeasured π-A isotherms, performed the thermodynamic calcu-lations and visualized the monolayers by BAM microscopy duringtheir compression. Theπ-A and CS−1-π plots are presented in Supportingmaterials (SFigs. 3 and 4). In Fig. 4. we compare the maximal CS

−1

values at a given X in the investigated systems and show the πcoll – Xdependences.

The interpretation of Fig. 4 leads to the conclusion that there areimportant differences in the interactions of Urs and Toc with particularE. colimembrane lipids. Starting from CS−1max, for the systemswith Toca disorganizing effect of Toc incorporation can be observed. In both sys-tems Toc/POPE and Toc/ECPG CS−1max falls with X(Toc) and for most ofthe compositions remains below the threshold of 100 mN/m indicatingthe LE state of the binary monolayers. In the systems with Urs strongcondensation effect can be observed. It is the most pronounced atX(Urs)= 0.66 or 0.80, as for thesemole ratios CS−1 max achieves highervalues than 250 mN/m proving close packing of the molecules and thesolid (S) state of the Langmuir monolayers. The above mentioned effecthas been previously observed by us in other Urs/phospholipid systems[25,40]. Regarding the πcoll-X plots significant differences between thesystems with Toc and Urs can be described. In the system Toc/ECPGthe πcoll-X plot is roughly linear, which can be interpreted as a proof of

Fig. 4. CS−1max-X (column plot) and πcoll – X plots (insets) for the systems with POPE

and ECPG.

2158 M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

mutual miscibility of both components. On the contrary, in the systemToc/POPE a sudden decrease of πcoll occurs at X(Toc) = 0.50, while forhigher X(Toc) the collapse pressure is even lower than for pure Tocmonolayer. The monolayers of both systems were visualized by BAMupon their compression. In the system Toc/ECPG the films were homo-geneous regardless their composition. The BAM images of the mono-layers in the system Toc/POPE strongly depended on the componentsproportions: for X(Urs) = 0.20 and X(Urs) = 0.34 the films werecompletely homogeneous and even behind the collapse it was difficultto observe 3D aggregates. On the contrary, for X(Toc)=0.50 and highera sudden separation of small (resolution limit) 3D nuclei was observedat ca. 16–18 mN/m. An example of such image is shown in Fig. 4. It isprobable that the separated phase is enriched in Toc, as the collapsepressure value is close to πcoll observed for pure Toc monolayers.

In the systems with Urs the collapse pressure remains on a constantlevel regardless the composition. The collapse pressure is comparablewith πcoll of the pure phospholipid and much higher than πcoll of Ursfilm. The monolayers of both Urs-containing systems were also visual-ized by BAM and it turned out that theywere completely homogeneousfor all components proportions. This fact together with the above de-scribed condensation effect proves the miscibility of Urs with POPEand ECPG.

For all the systems the excess free energies of mixing were also cal-culated. TheΔGexc-X curves are shown in Supportingmaterials (SFig. 5),in Fig. 5 in the form of a column plot we present a comparison of ΔGexc

in the systems with POPE and ECPG.In the systemswith POPE important differences can be observed be-

tween Toc and Urs. In both systemsΔGexc have negative sign; however,in the system Toc/POPE the absolute values of ΔGexc are small and closeto 0, whereas in the system Urs/POPE the absolute values are large (for

Fig. 5. The comparison of ΔGexc at 20 mN/m for the systems with POPE (left panel) and EC

some X(Urs) even 10 times greater than in the system with Toc). Itmeans that the Urs-POPE interactions are energetically more beneficialthan in the pair Toc-POPE. Regarding the systemswith ECPG, the resultsare interesting and the differences between Toc and Urs are here evenmore drastic. In the systemUrs/ECPG theΔGexc absolute values are rath-er low, and oscillate around 0, being positive for X(Urs)= 0.20 and 0.33and negative for higher Urs proportions. On the contrary, in the systemToc/ECPG the values of ΔGexc are high and positive which means thatthe mutual interactions between Toc and ECPG molecules are unfavor-able in the light of the ΔGexc criterion.

5.1. Investigation of the antagonism between Toc and Urs

In the experiments described in the previous sections we providedevidence that Toc and Urs differ significantly in their interactionswith E. coli polar lipids total extract as well as with the main membranephospholipids POPE and ECPG. However, in nature E. coli can be simul-taneously exposed to both the substances. In such a situation antagonis-tic effects can occur, which was described in scientific literature [17].Thus, our ideawas to simulate such a situation inmodel systems. There-fore, in the further experiments we studied ternary Langmuir mono-layers containing Toc, Urs and POPE or ECPG. We remained thepreviously used mole ratios of Urs or Toc, i.e. 0.20, 0.33, 0.50, 0.66 and0.80. That time; however, these proportions were a sum of both Tocand Urs, so for example, the case described in the figure below X =0.50 regards such a monolayer in which X(Toc) = X(Urs) = 0.25, andthe sum X(Toc) + X(Urs) = 0.50. The π-A isotherms recorded for theternary films are shown in the Supporting materials (SFig. 6), whereasthe CS−1 –π and πcoll – X dependencies are shown in Fig. 6 and discussedin the article.

Regarding the CS−1 – π curves the most striking difference betweenthe ternary and the above described binary systems is the lack of con-densation effect observed previously in the systemswith Urs. In the sys-tem Toc/Urs/POPE (Fig. 6 left panel) the maximal values of CS−1 remainon a stable level of ca. 150 mN/m. At the systems with lower Toc + Ursproportion (X = 0.20 and 0.33) the LE-LC phase transition, manifestedas a sharp minimum in the POPE CS−1-π curve, disappears. For higherX a local minimum is observed in the course of the CS

−1 – π curve atca. 20–25 mN/m. All the monolayers were visualized by BAM upontheir compression. At X = 0.20 and 0.30 the films were homogeneouseven behind the collapse achieved on a relatively high π value of ca.45 mN/m. For higher X multiple small bright 3D nuclei appeared sud-denly at ca. 20 mN/m. The BAM images were similar to the photo pre-sented in Fig. 4, so it can be assumed that at these conditions also aToc-enriched phase separates from POPE-rich phase. In the systemToc/Urs/ECPG the maximal CS−1 values oscillated around 100 mN/m,only for X=0.80 the value of 140mN/mwas observed. The BAM imagesof the monolayers within the system were homogeneous, regardless the

PG (right panel). The X values are mole ratios of Toc or Urs, depending on the system.

Fig. 6. CS−1 – π and πcoll – X dependencies for the investigated ternary systems. X values regard the sum of Toc and Urs mole ratios.

2159M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

film components proportion. In this system the separation of 3D aggre-gates was not observed. As far as the πcoll-X curves are concerned, inthe system Toc/Urs/POPE a roughly linear reduction in πcoll value can beobserved till X = 0.50 after which an abrupt fall of πcoll is observed.This phenomenon can be correlatedwith the separation of 3D aggregatesobserved in BAM images. In the system Toc/Urs/ECPG collapse pressuredecreases uniformly with X, and no sudden falls in the πcall – X curvecan be observed.

Similarly to the previously discussed systems, also for the ternarymixtures we performed the calculations of ΔGexc. The ΔGexc – X depen-dences at different surface pressures are gathered in Supporting mate-rials (SFig. 7). Here we had a different idea – we compared the ΔGexc

values observed at 20 mN/m for the ternary systems at a given X withthe averages of ΔGexc observed for the binary systems with the samephospholipid – Fig. 7.

For the system Toc/Urs/POPE theΔGexc values have negative sign forall the proportions of the components. Similarly the average values arenegative, because the ΔGexc values in the systems Toc/POPE and Urs/POPE were negative. For X = 0.20 and X = 0.80 the ΔGexc valuesobserved for the ternary system are comparable with the averages, forthe middle mole ratios they are visibly lower. However, interpreting

Fig. 7. The comparison of ΔGexc values at 20 mN/m calculated for the ternary systems with a

these results it is impossible to indicate noticeable antagonistic effectsof Toc andUrs in their interactionwith POPE. The situation is completelydifferent for the system Toc/Urs/ECPG. All the ΔGexc observed for theternary system are negative whereas the averages are positive. More-over, the ΔGexc values in the ternary system Toc/Urs/ECPG are morenegative than in the system Urs/ECPG, even though in the system Toc/ECPG ΔGexc are high and positive. Therefore in the interactions of boththe investigated compounds with ECPG strong antagonistic effectcan be observed. The nature of the effect and its significance will bediscussed in the following section.

6. Discussion

α-tocopherol is the most important component of vitamin E,the main role of which is the protection of unsaturated lipids againstROS-induced peroxidation. Toc exerts its function in the membraneenvironment. The plant cells synthesize Toc and incorporate it into thecellular and organelle membranes, whereas in animal organisms Tocis absorbed from food and distributed into the cells [22,41]. Thus, theproblemof Toc incorporation into the animalmembraneswas discussedinmultiple papers. Themost important issuewas the distribution of Toc

verages of the ΔGexc values observed for binary systems with the same phospholipids.

2160 M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

in the lipid matrix, that is whether it is distributed uniformly or formsspecific lipid domains [22,41,42]. Some authors reported that Toc inter-acts specifically with phospholipids possessing unsaturated and poly-unsaturated fatty acid chains in their hydrophobic moieties, which islogical as Toc should protect them against peroxidation [22,41,42]. Onthe other hand, cholesterol is well known to form so-called lipid raftswith saturated phospholipids,mainly sphingolipids [43,44]. Urs is struc-turally similar to cholesterol; thus, it is probable that Ursmore preferen-tially interacts in biomembrane with different lipids than Toc and thatboth compounds segregate to different phospholipids domains. The fol-lowing lipid binary systems: cholesterol/POPE, Urs/POPE and Toc/POPEhave been investigated with the application of different approaches ofsurface chemistry. It turned out that cholesterol is miscible with POPEin Langmuir monolayers [45] and in vesicles [46]. The hydrophobicchain of PE must be much more unsaturated to force cholesterol tophase separate [47]. On the other hand the interactions of cholesterolwith POPE are less strong than with phosphatidylcholine (POPC) orsphingomyelin (SM) as it was much easier to desorb cholesterol bythe application of β-cyclodextrin from the Chol/POPE 1/1 monolayerthan from similar mixtures with POPC or SM [50]. Toc was also investi-gated in the mixtures with POPE [48,49]. Small angle synchrotron dif-fraction studies evidenced that in solutions Toc induces in POPErippled lamellar phases [48]. The experiments were also performed onPOPE bilayers mimicking E. coli membranes proving that considerableamounts of Toc can be incorporated into such model system [49]Thus, on the basis of scientific literature it is difficult to concludewhich of the compounds Toc or Urs (or cholesterol) has higher affinityto POPE. This assessment would be of help as POPE is the main compo-nent of Etotal. There were two main observations regarding the Urs/Etotal and Toc/Etotal mixtures. Urs can condensate Etotal films andToc has not such ability, but this result was rather predictable and canbe explained by the structure of the hydrophobic parts of both mole-cules. Urs with its five stiff fused cycloalkane rings and no long acylchains at peripheries can behave similarly to cholesterol in lipid mono-layers and increase the packing of phospholipid molecules. This phe-nomenon was described by us in previous publications [40]. Thesecond important difference in the systems with Etotal regards theΔGexc values. In both systems they are negative but the absolute valuesare much higher in the system Urs/Etotal indicating that the interac-tions of Urs with E. coli membrane lipids are energetically more favor-able than their interactions with Toc. To gain a better understandingof the observations collected for the systems with Etotal we moved tobinary systems composed by the main E. coli membrane phospholipidsPOPE or ECPG and Toc or Urs. The results for the systems with POPEwere qualitatively similar to these observed for the systemswith Etotal.Namely, we observed strong condensation effects in the system Urs/POPE, especially at higher X(Urs), and a moderate expansion effect inthe case of the systems Toc/POPE. Regarding the system Toc/POPE, aninteresting observation was the phase separation at 17–20 mN/m inthe mixed films at moderate and high Toc proportions (starting fromX(Toc) = 0.50). Such an effect was not observed in the system Urs/POPE. It should be underlined here that in normal cellular membranesToc comprises only 0.1 to 1 % of all the membrane lipids [21], somodel membranes with Toc comprising 50 or more percent of all thelipids are far from the real conditions [29]. On the other hand suchmodels are of help in the investigation of themutual interactions of sep-arated lipids in the model membrane environments. Our results provethat only one cis-double bond in the hydrophobic moiety of the phos-pholipid molecule is not enough for the effective interactions with Tocand for the formation of a stable surface complex. On the other hand,one cis-double bond in the oleoyl chain is not a substantial obstaclefor Urs to prevent its condensing abilities. Similarly to cholesterol [45,50], Urs can force mixed acyl phospholipids to form closely packedmolecular arrays. As it was previously described [45], only a system ofmultiple cis-double bonds in a fatty acid chain can prevent cholesterolfrom the condensing action. Regarding the ΔGexc- composition plots,

the dependences observed for the systems with POPE resemble thesediscussed previously for the systemswith Etotal, even the values at par-ticular film compositions are comparable. It means that globally the in-teractions with POPE have the highest contribution into the discussedenergetic effects, which is in accordance with the proportion of POPEin Etotal. The most important differences between Toc and Urs can beobserved in the ΔGexc – composition plots for the binary systems withECPG. For the systemUrs/ECPG theΔGexc values are close to 0which to-gether with the CS−1-π curves and the BAM observations prove the idealmiscibility of the components in this system. On the contrary the ΔGexc

values observed in the system Toc/ECPG are high and have positive sign,which means that the Toc/ECPG interactions are energetically unfavor-able. On the other hand, we had no evidence in BAM images indicatingphase separation in these systems. The location of Toc in biologicalmembranes was a subject of heated debate in scientific literature. Espe-cially, it was discussed how deep this molecule is immersed withinthe phospholipids layer and with which part of the phospholipidsheadgroup the tocopherol polar headgroup is in contact [21]. It was sug-gested that polyunsaturated fatty acids can form specific pockets for theside methyl branches of the Toc phytyl chain [51,52] and that the chainis not a stiff structure, so the formation of gauche defects is highly prob-able. By the phytyl chain deformation Toc acquires the optimal confor-mation to fit to a particular polyunsaturated hydrocarbon chain. Thisobservation can help in the interpretation of the differences in the inter-actions in the pairs Toc/ECPG and Toc/POPE. POPE has one cis-doublebond and can form one “pocket” interacting with a side methyl groupof Toc. Effective arrangement of Toc and POPE requires at leastone gauche defect in the Toc phytyl chain. Such an arrangement is ener-getically beneficial, but the increased probability of gauche defectformation leads to the lowering of themonolayer condensation as com-pared with pure POPE monolayer.

The last part of our experiments regarded the antagonistic effects be-tween Urs and Toc in ternary monolayers with both main E. coli mem-brane phospholipids. Such effects are of utmost importance in theliving systems andwere stimulus of ourmodel studies. Themost impor-tant effect of our studieswas the corroboration that such an effect is alsovisible in themodel systems and that its interpretation can shed light onthemutual action of tocopherols and pentacyclic terpenes in real bacte-rialmembranes. The first important observation for the ternary systemswas the lack of condensation effect, which was exerted by Urs on POPEand ECPGmonolayers. This observation can be of utmost importance, asthemembrane fluidity is one of its most important properties. Urs anti-bacterial action can be connected with important changes in the bacte-rial membrane fluidity connected with the PT molecule incorporation.Elimination of this effect by Toc can be responsible for the lack of phar-maceutical activity of pentacyclic triterpenes when applied togetherwith vitamin E. Very interesting are also the thermodynamic results cal-culated for the ternary systems. Our idea was as follows: if there are notantagonistic effects the ΔGexc value at a given phospholipids proportionin a ternary system should be an average of the values observed in bina-ry systems with the same phospholipids. The results for the systemswith POPE roughly follow this trend. The negative sign ofΔGexcwas pre-served and the absolute values of theΔGexc for the ternary systemswerecomparable (with the exception of X(POPE) = 0.50) to the calculatedaverages. On the contrary, the trend observed for the systems withECPG was completely different. The average ΔGexc values had positivesign; whereas the real values calculated for the ternary systems werenegative. These results prove that in the ternary systems strong antago-nistic effects are present. These effects should be connected with theseparation of the molecules within the monolayer. As it was alreadymentioned the affinity of Urs towards ECPG is much higher than inthe case of Toc. Thus it is probable that ECPG molecules interact mainlywith Urs and are separated from Toc. In absence of Toc POPE and ECPGmolecules compete in their interactions with Urs; whereas in the pres-ence of Toc another competition appears. The interactions with POPEcan be energetically beneficial both for Urs and Toc. The ternary Toc/

2161M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

Urs/phospholipids system strives for the energetic minimum. Thus, themost energetically convenient situation is when ECPG interacts mainlywith Urs and Toc mainly with POPE. This can lead to microdomain for-mation and to a completely different arrangement of the membranelipids than in the case when only Urs or Toc, respectively is present.

7. Conclusions

It was indicated in multiple studies that vitamin E and its main com-ponent α-tocopherol can exert antagonistic effects to antibiotics orpentacyclic triterpene drugs preventing their antibacterial activity.Both Toc and Urs are water-insoluble bioactive substances which canbe incorporated into biological membranes. In our studies we modeledEscherichia coli membranes by the application of Langmuir monolayersformed by the total extract of E. coli polar lipids aswell as themain com-ponents of E. coli inner membrane, namely POPE and ECPG. Our resultsevidence that Urs and Toc behave differently in binary monolayers withEtotal. The main differences were reflected in the condensation of themonolayers and in the values of ΔGexc. Namely, the increase in Urs pro-portion leads to a significant rise of the compressionmodulus value and,on the other hand, mixing of Urs with Etotal is more energetically ben-eficial than with Toc. Similar trends were repeated for the binary sys-tems with POPE, which was in accordance with the composition ofEtotal, as POPE is its dominating component. Profound differenceswere observed in the interactions of Toc and Urs with ECPG. Mixing ofUrs with ECPG is energetically beneficial; whereas for binary Toc/ECPGmonolayers large positiveΔGexc valueswere observed proving unfavor-able energetic effects. The antagonistic effects of Toc and Urs observedin vivo were emulated in the applied model systems; namely, ternaryLangmuir monolayers composed of Toc, Urs and ECPG or POPE. Thepresence of both Toc and Urs in a ternary monolayer leads to the elim-ination of the condensation effects observed in the binary systems com-posed of Urs and bacterial phospholipids. On the other hand, POPE andECPG interact differently with both substances. The affinities of Urs andToc towards POPE are similar; thus, the ΔGexc values observed for theternary Toc/Urs/POPE system were comparable with the average ofthe values observed for the binary systems Toc/POPE and Urs/POPE.The affinities of Toc and Urs towards ECPG are completely differentand it seems that in ternary systems Urs interacts preferentially withECPGwhich can lead to phase separation andmicrodomains formation.Our model studies support the idea that the antagonistic effects ob-served in the pharmacological activity of Urs and Toc can be connectedwith their membrane action. In the case of E. coli inner membrane thedifferences in the interaction of Toc and Urs with the anionic phospho-lipid ECPG can be the key factor responsible for the extinguishing of Urspharmacological activity in presence of Toc.

Conflict of Interest

I as a corresponding author and on behalf of my co-authors wouldlike to certify that there are no conflicts of interest between us andany physical persons or institutions.

Acknowledgement

This project was financed by the National Science Centre (No DEC-2012/05/B/ST5/00287).

The research was carried out with the equipment (UltraBAM)purchased thanks to the financial support of the European Regional De-velopment Fund in the framework of the Polish Innovation EconomyOperational Program (contract no. POIG.02.01.00-12-023/08).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamem.2015.05.009.

References

[1] E. Breitmaier, Terpenes, Wiley-VCH, Weinheim, 2006.[2] S. Jager, H. Trojan, T. Kopp, M. Laszczyk, A. Scheffler, Pentacyclic triterpene distribu-

tion in various plants – Rich sources for a new group of multi-potent plant extracts,Molecules 14 (2009) 2016–2031.

[3] M.N. Laszczyk, Pentacyclic triterpenes of the lupane, oleanane and urwane group astools in kancer therapy, Planta Med. 75 (2009) 1549–1560.

[4] L. Zhang, B. Wu, Y. Lin, J. Wang, F. Lei, Z. Tang, Research progress of ursolic acid’santi-tumor actions, Chin. J. Integr. Med. 20 (2014) 72–79.

[5] V.R. Yadav, S. Prasad, B. Sung, R. Kannappan, B.B. Aggarwal, Targeting inflammatorypathways by triterpenoids for prevention and treatment of cancer, Toxins 2 (2010)2428–2466.

[6] A. Kurek, A.M. Grudniak, M. Szwed, A. Klicka, L. Samouk, K.L. Wolska, K.W.Janiszowska,M. Pomowska, Oleanolic acid and ursolic acid affect peptidoglycanme-tabolism in Listeria monocytogenes, Antonie Van Leeuwenhoek 97 (2010) 61–68.

[7] S. Fontanay, M. Grare, J. Mayer, C. Finance, R.E. Duval, Ursolic, oleanolic and betulinicacids: Antibacterial spectra and selectivity indexes, J. Ethnopharmacol. 120 (2008)272–276.

[8] S.G. Kim, M.J. Kim, D. Jin, S.N. Park, E. Cho, M. Oliveira Freire, S.J. Jang, J.K. Kook, An-timicrobial effect of ursolic acid and oleanolic acid against methicillin-resistantStaphylococcus aureus, Korean J. Microbio. 48 (2012) 212–215.

[9] A. Filocamo, C. Bisignano, M. D’Arrigo, G. Ginestra, G. Mandalari, E.M. Galati,Norfloxacin and ursolic acid: in vitro association and postantibiotic effect againstStaphylococcus aureus, Lett. Appl. Microbiol. 53 (2011) 193–197.

[10] P.O. Harmand, R.J. Duval, C. Delage, A. Simon, Int. J. Cancer 114 (2005) 1–11.[11] Y.H. Lee, E. Wang, N. Kumar, R.D. Glickman, Ursolic acid differentially modulates ap-

optosis in skin melanoma and retinal pigment epithelial cells exposed to UV–VISbroadband radiation, Apoptosis 19 (2014) 816–828.

[12] M. Mahmoudi, S. Zamani, T. Rabe, M. Balali-Mood, G. Karimi, N. Tabaki, B. Riahi-Zanjani, Ursolic acid induced apoptotic cell death following activation of caspasesin isolated human melanoma cells, Cell Biol. Int. 39 (2015) 230–236.

[13] S. Saeidnia, M. Abdollahi, Toxicological and pharmacological concerns on oxi-dative stress and related diseases, Toxicol. Appl. Pharmacol. 273 (2013)442–455.

[14] M. Maes, P. Galecki, Y.S. Chang, M. Berk, A review on the oxidative and nitrosativestress (O&NS) pathways in major depression and their possible contribution tothe (neuro)degenerative processes in that illness, Prog. Neuro-Psychopharmacol.Biol. Psychiatry 35 (2011) 676–692.

[15] G.W. Burton, Vitamin E: molecular and biological function, Proc. Nutr. Soc. 53(1994) 251–262.

[16] A. Azzi, Molecular mechanism of α-tocopherol action, Free Radic. Biol. Med. 43(2007) 16–21.

[17] D.V. Raghuvar Gopal, A.A. Narkar, Y. Badrinath, K.P. Mishra, D.S. Joshi, Protection ofEwing’s sarcoma family tumor (ESFT) cell line SK-N-MC from betulinic acid inducedapoptosis by α-dl-tocopherol, Toxicol. Lett. 153 (2004) 201–212.

[18] M.M. Masadeh, N.M. Mhaidat, K.H. Alzoubi-Sayer, I. Al-Azzam, A.I. Shaweesh,Ciprofloxacin-induced antibacterial activity is reversed by vitamin E and vitaminC, Curr. Microbiol. 64 (2012) 457–462.

[19] J. Prades, O. Vogler, R. Alemany, M. Gomez-Florit, S.S. Funari, V. Ruiz-Gutierrez, F.Barcelo, Plant pentacyclic triterpenic acids as modulators of lipid membrane physi-cal properties, Biochim. Biophys. Acta 1808 (2011) 752–760.

[20] S.K. Han, Y.I. Ko, S.J. Park, I.J. Jin, Y.M. Kim, Oleanolic acid and ursolic acid stabilizeliposomal membranes, Lipids 32 (1997) 769–773.

[21] J. Atkinson, R.F. Epand, R.M. Epand, Tocopherols and tocotrienols in membranes: Acritical review, Free Radic. Biol. Med. 44 (2008) 739–764.

[22] X. Wang, P.J. Quinn, Vitamin E and its function in membranes, Prog. Lipid Res. 38(1999) 309–336.

[23] D. Marquardt, J.A. Williams, N. Kučerka, J. Atkinson, S.R. Wassall, J. Katsaras, T.A.Harroun, Tocopherol activity correlates with its location in a membrane: A newperspective on the antioxidant vitamin E, J. Am. Chem. Soc. 135 (2013)7523–7533.

[24] M. Flasiński, K. Hąc-Wydro, M. Broniatowski, Incorporation of pentacyclictriterpenes into mitochondrial membrane – Studies on the interactions in model2D lipid systems, J. Phys. Chem. B 118 (2014) 12927–12937.

[25] M. Broniatowski, P. Mastalerz, M. Flasiński, Studies of the interactions of ursane-type bioactive terpenes with the model of Escherichia coli inner membrane – Lang-muir monolayer approach, Biochim. Biophys. Acta, Biomembranes 1848 (2015)469–476.

[26] R.M. Epand, S. Rotem, A. Mor, B. Berno, R.F. Epand, Bacterial membranes as predic-tors of antimicrobial potency, J. Am. Chem. Soc. 130 (2008) 14346–14352.

[27] S. Morein, A.S. Andersson, L. Rilfors, G. Lindblom,Wild-type Escherichia coli cells reg-ulate themembrane lipid composition in a “window” between gel and non-lamellarstructures, J. Biol. Chem. 271 (1996) 6801–6809.

[28] C.M. Castro, M. Pinheiro, M. Lucio, J.J. Giner-Casares, L. Camacho, J.L.F.C. Lima, S. Reis,M.A. Segundo, Insights about α-tocopherol and Trolox interaction with phosphati-dylcholine monolayers under peroxidation conditions through Brewster anglemicroscopy, Colloids Surf. B: Biointerfaces 111 (2013) 626–635.

[29] M. Jurak, J. Miñones Conde, Characterization of the binary mixed monolayers ofα-tocopherol with phospholipids at the air-water interface, Biochim. Biophys.Acta 1828 (2013) 2410–2418.

[30] B. Gzyl-Malcher, M. Zembala, M. Filek, Effect of tocopherol on surface properties ofplastid lipids originating from wheat calli cultivated in cadmium presence, Chem.Phys. Lipids 163 (2010) 74–81.

[31] G.L. Gaines Jr., Insolublemonolayers at liquid–gas interfaces, Interscience, New York,1966.

2162 M. Broniatowski et al. / Biochimica et Biophysica Acta 1848 (2015) 2154–2162

[32] G. Capuzzi, P. Lo Nostro, K. Kulkarni, J.E. Fernandez, Mixtures of strearoyl-6-O-ascor-bic acid and α-tocopherol: A monolayer study at the gas/water interface, Langmuir12 (1996) 3957–3963.

[33] J.T. Davies, E.K. Rideal, Interfacial phenomena, 2nd ed Academic Press, New York,1963.

[34] G.S. Patil, D.G. Cornwell, J. Lipid Res. 19 (1978) 416–422.[35] G. Brezesinski, D. Vollhard, Model studies of the interfacial ordering of oleanolic acid

in the cuticula, ChemPhysChem 9 (2008) 1670–1672.[36] M. Broniatowski, M. Flasiński, P. Wydro, Lupane-type pentacyclic triterpenes

in Langmuir monolayers: A synchrotron radiation scattering study, Langmuir 28(2012) 5201–5210.

[37] R.F. Epand, P.B. Savage, R.M. Epand, Bacterial lipid composition and the antimicrobi-al efficacy of cationic steroid compounds (Ceragenins), Biochim. Biophys. Acta 1768(2007) 2500–2509.

[38] D.W. Grogan, J.E. Cronan Jr., Cyclopropane ring formation in membrane lipids ofbacteria, Microbiol. Mol. Biol. Rev. 61 (1997) 429–441.

[39] Y.Y. Chang, J.E. Cronan Jr., Membrane cyclopropane fatty acid content is a majorfactor in acid resistance of Escherichia coli, Mol. Microbiol. 33 (1999) 249–259.

[40] M. Broniatowski, M. Flasiński, K. Zięba, P. Miśkowiec, Interactions of pentacyclictriterpene acids with cardiolipins and related phosphatidylglycerols in modelsystems, Biochim. Biophys. Acta 1838 (2014) 2530–2538.

[41] S. Lemaire-Ewing, C. Desrumaux, D. Neel, L. Lagrost, Vitamin E transport, membraneincorporation and cell metabolism: Is a-tocopherol in lipid rafts an oar in the life-boat? Mol. Nutr. Food Res. 54 (2010) 631–640.

[42] J. Atkinson, T. Harroun, S.R. Wassall, W. Stillwell, J. Katsaras, The location and behav-ior of α-tocopherol in membranes, Mol. Nutr. Food Res. 54 (2010) 641–651.

[43] K. Simons, M.J. Gerl, Revitalizing membrane rafts: new tool and insights, Nat. Rev.Mol. Cell Biol. 11 (2010) 688–699.

[44] D. Lingwood, K. Simons, Lipid rafts as a membrane-organizing principle, Science 327(2010) 46–50.

[45] S.R. Shaikh, M.R. Brzustowicz, N. Gustafson, W. Stillwell, S.R. Wassall, Monounsatu-rated PE does not phase-separate from the lipid raft molecules sphingomyelin andcholesterol: Role for polyunsaturation, Biochem. 41 (2002) 10593–10602.

[46] X. Wang, P.J. Quinn, Cubic phase is induced by cholesterol in the dispersion of 1-palmitoyl-2-oleoyl-phosphatidylethanolamine, Biochim. Biophys. Acta 1564 (2002)66–72.

[47] B. Terova, G. Petersen, H.S. Hansen, J.P. Slotte, N-acyl phosphatidylethanolamines af-fect the lateral distribution of cholesterol in membranes, Biochim. Biophys. Acta1715 (2005) 49–56.

[48] X. Wang, P.J. Quinn, The interaction of a-tocopherol with bilayers of 1-palmitoyl-2-oleoyl-phosphatidylcholine, Biochim. Biophys. Acta 1567 (2002) 6–12.

[49] X. Wang, P.J. Quinn, The structure and phase behaviour of α-tocopherol-rich do-mains in 1-palmitoyl-2-oleoyl-phosphatidylethanolamine, Biochemie 88 (2006)1883–1888.

[50] K. Hąc-Wydro, R. Lenartowicz, P. Dynarowicz-Łątka, The influence of plant stanol(_-sitostanol) on inner leaflet of human erythrocytes membrane modeled with theLangmuir monolayer technique, Colloids Surf. B: Biointerfaces 102 (2013) 178–188.

[51] M. Afri, B. Ehrenberg, Y. Talmon, J. Schmidt, Y. Cohen, A.A. Frimer, Active oxygenchemistry within the liposomal bilayer. Part III: Locating vitamin E, ubiquinoland ubiquinone and their derivatives in the lipid bilayer, Chem. Phys. Lipids 131(2004) 107–121.

[52] A.N. Erin, M.M. Spirin, L.V. Tabidze, V.E. Kagan, Formation of alpha-tocopherol com-plexes with fatty acids. A hypothetical mechanism of stabilization of biomembranesby vitamin E. Biochim. Biophys. Acta 774 (1984) 96–102.