Adsorption of Polylysine on the Surface of the DMPS Monolayer
Yurii A. Ermakova, Victor E. Asadchikov, Yurii O. Volkov, Alexander D. Nuzhdin, Boris S. Roshchin, Veijo Honkimaki, Aleksey M. Tikhonov
aa r X i v : . [ c ond - m a t . s o f t ] A p r Adsorption of Polylysine on the Surface of the DMPSMonolayer
Yurii A. Ermakova a , Victor E. Asadchikov b , Yurii O. Volkov b , Alexander D. Nuzhdin b , Boris S. Roshchin b ,Veijo Honkim¨aki c , and Aleksey M. Tikhonov d ∗ a Frumkin Institute of Physical Chemistry and Electrochemistry, Russian Academy of Sciences, Moscow, 119071 Russia b Shubnikov Institute of Crystallography, Federal Research Center Crystallography and Photonics, Russian Academy of Sciences,Moscow, 119333 Russia c European Synchrotron Radiation Facility, 38000 Grenoble, France d Kapitza Institute for Physical Problems, Russian Academy of Sciences, Moscow, 119334 Russia
Abstract
The effect of the adsorption of a polypeptide on the lat-eral interaction of dimyristoylphosphatidylserine moleculesin different phase states on the surface of a 10 mM KClaqueous solution has been studied. Changes in the surfacepressure and Volta potential induced by the adsorption oflarge poly-D-lysine molecules (about 200 links in a chain)have been determined at different areas per lipid molecule ina monolayer. The adsorption of macromolecules noticeablyincreases the elasticity of the monolayer under lateral com-pression in the liquid expanded state of lipid and reduces theeffective dipole moment from 0.48 to 0.38 D. These prop-erties are in qualitative agreement with X-ray reflectometrydata for the lipid monolayer obtained with synchrotron radi-ation with a photon energy of ≈ keV. The electron den-sity profiles perpendicular to the surface of the aqueous sub-phase have been reconstructed from reflectometry data withina model approach to the structure of an interface with twoand three layers. These profiles indicate the existence of awide diffuse polymer layer (150 ± ˚A in width at the inter-face of the monolayer in both the liquid expanded and liquidcondensed states. A decrease in the area per molecule inthe monolayer by a factor of 2 results in the doubling of thesurface density of the macromolecule film. The adsorptionof the polymer also affects the integral density of the layerof polar phospholipid groups, which decreases by a factor of ≈ in the liquid expanded phase and by ∼ % in the liquidcondensed phase. INTRODUCTION
Lipid models of biomembranes have been widely used fora long time to study structural factors in the interaction ofvarious membrane active compounds with the surface of cells[1, 2, 3]. Special attention is focused on the effect of variouspolypeptide macromolecules, which were applied in variousbiomedical applications, on the structure of the lipid matrixof cell walls. The interpretation of experimental data usually ∗ [email protected] includes an important fundamental property of saturatedphospholipids to demonstrate a first-order phase transitionfrom a liquid expanded (LE) state to a liquid condensed(LC) one under the variation of the temperature T or thelateral pressure Π [4]. In the general case, in the L monolayerstate, diverse lipid mesophases both crystal and hexatic areobserved; they differ between each other in, e.g., area permolecule A and the angle of canting of aliphatic tails withrespect to the normal to the surface [5, 6, 7, 8, 9, 10]. Wepreviously showed that change in the lateral packing of neg-atively charged phosphatidylserine molecules in model lipidmembranes is also initiated by the adsorption of multiva-lent cations and large polypeptide molecules on their surface[11, 12, 13]. A good signature of this phenomenon appearedto be change in the dipole component of the boundary po-tential [14]. It can be expected that the clustering of lipids inthe presence of polypeptide and the appearance of the dipolecomponent of the boundary potential have many features ofthe molecular nature of the LE-LC phase transition.In this work, to study the effect of the adsorption ofpolypeptide on the lateral interaction of lipid molecules indifferent phase states, we use dimyristoylphosphatidylserine(DMPS), which is saturated analog of phosphatidylserine.We report experimental data indicating structural changesin the DMPS monolayer in both the LE and the LC statein the presence of poly-D-lysine hydrobromide molecules atits water interface. These data were obtained when study-ing the compressibility of the lipid monolayer using theLangmuir monolayer technique and from X-ray reflectom-etry with synchrotron radiation. EXPERIMENT C H NO PNa sodium salt of DMPS phospholipid(Avanti Polar Lipids) was prepared at a concentration of ≈ . C N O] n × HBr poly-D-lysine hydrobromide,1 where n ≈
200 (P7886, Sigma-Aldrich). The concentrationof polylysine added to the subphase was = 0 . ∼ . µ M. The DMPS monolayer was deposited by aHamilton syringe on the surface of the aqueous solution withthe necessary amount of lipid, which corresponded to areasof 36 ˚A and 72 ˚A per molecule in the monolayer at themeasurement of X-ray scattering. Pressurearea diagramsand the Volta potential were measured simultaneously at aMicroTrough XS, V.4.0 setup (Kibron Inc., Finland) witha 20 × ∼
20 minafter the deposition of lipid and at a compression rate of ∼
10 mm /min. The typical reproducibility of compressiondiagrams was no worse than 1-3%.The compression diagrams of DMPS monolayers shown inFig. 1a reproduce well the results reported in our previousworks [11, 15, 16]. Both curves clearly demonstrate a sectionof a smooth increase in the lateral pressure in the regionof the LE state of the monolayer. According to previousstudies and molecular dynamics simulations of such systems,the aqueous surface in this region includes lipid domainswhose total area increases as barriers approach each otherand the lateral pressure increases. This region is followedby a small plateau with an intermediate state and a sharperincrease in the pressure at the transition to the LC statewith the maximally dense packing of lipids near collapse.Thus, the compression of the lipid monolayer successivelychanges its physical state beginning with a two-dimensionalgas at pressures below 12 mN/m to the LE and LC statesand, finally, to the collapse of the monolayer when lipidsgo out the bath or/and from the monolayer to the aqueousmedium.The data presented in Fig. 1b make it possible to iden-tify three segments of the variation of the Volta potentialwhose positions correlate with the Π − A diagram, where Πis the surface pressure and A is the area per molecule. Thepotential in the LE region varies according to an almost lin-ear law, whereas the poten tial in the LC region increasesmore sharply to the plateau, which characterizes the maxi-mally dense monolayer immediately before the collapse. Allphases listed above are manifested in the variation of theedge Volta potential. In a very strongly diluted monolayer, aKelvin vibrating electrode detects random lipid spots, whichare located under its surface, whereas the surface tension ofthis surface remains as a whole unchanged and close to thesurface tension of pure water γ (under normal conditions, γ ≈ . Figure 1. (a) Compression diagrams and (b) Volta potential ofthe DMPS monolayer measured at the deposition of lipid on thesurface of the 10 mM KCl aqueous solution in the (1) absence and(2) presence of poly-D-lysine molecules with the concentration c = 0 . mg/mL in this solution. The reflection coefficient of X rays, R , for LangmuirDMPS monolayers at the waterair interface was measuredon the ID31 beamline of the European Synchrotron Radia-tion Facility (Grenoble, France) [17]. The intensity of thefocused monochromatic photon beam with a wavelength of λ ≈ .
18 ˚A(the photon energy E ≈
70 keV, ∆
E/E = 0 . ∼ µ m in height and ∼ µ min horizontal plane was ∼ photons/s. The method ofmeasurement of the reflection coefficient R was described in[18, 19, 20, 21]. Monolayer samples of DMPS phospholipidwere prepared under conditions similar to those at the mea-surement of compression diagrams but in a circular tetraflu-oroethylene bath 10 cm in diameter, which was placed in ahermetic thermostat with X-ray transparent windows. THEORY
In the case of specular reflection, the scattering vector q has only one nonzero component q z = (4 π/λ ) sin α , where α is the glancing angle in the plane normal to the sur-face (see the inset of Fig. 2). The total external reflec-tion angle for the surface of water α c ≈ λ p r e ρ w /π ≈ . ◦ ( q c = (4 π/λ ) sin α c ≈ .
022 ˚A − ) is specified by the electrondensity in it ρ w ≈ . e − / ˚A , where r e = 2 . · − ˚Ais the classical electron radius. Figures 2 and 3 show theexperimental dependences R ( q z ) obtained for two LC andLE surface states with A = 36 ˚A and 72 ˚A , respectively.It is seen that the addition of the polymer to the subphaseaffects the shape of the reflection line, which indicates the Figure 2.
Dependence R ( q z ) for the DMPS monolayer on thesurface of water at an area of 72 ˚A for the lipid monolayer (cir-cles) without and (squares) with the adsorption polymer layer.Lines 1 and 2 are the respective model calculations. The insetshows the kinematics of scattering in the coordinate system wherethe xy plane coincides with the interface between the monolayerand water, the Ox axis is perpendicular to the beam direction,and the Oz axis is opposite to the gravitational force perpendicu-lar to the surface; k in and k sc are the wave vectors of the incidentand scattered beams, respectively; correspondingly, q = k in - k sc is the scattering vector. Figure 3.
Dependence R ( q z ) for the DMPS monolayer on thesurface of water at an area of 36 ˚A for the lipid monolayer (cir-cles) without and (squares) with the adsorption polymer layer.Lines 1 and 2 are the respective model calculations. The insetshows the parameterization of the structure of the air airwaterinterface. formation of a lipidpolymer film.The reflectometry data were analyzed within a model ap-proach, where the structure of the film at the airwater inter-face was divided into N layers. Each layer has the thickness L j and the electron density ρ j , and the widths of the inter-faces in the multilayer are described by the parameters σ j (see the inset of Fig. 4). The electron density profile ρ ( z )in the surface structure along the normal to the interface isparameterized as [22, 23, 24] ρ = 12 ρ + 12 N X j =0 ( ρ j +1 − ρ j )erf z j σ j √ ! , (1)where erf( x ) is the error function, z j = z + P jn =0 L n , ( L =0), and ρ = 0 and ρ N +1 = ρ w are the electron densitiesin air and in the bulk of water. In the distorted wave Bornapproximation (DWBA), the reflection coefficient R ( q z ) isexpressed in terms of the parameters of electron density (1)as [25] R ( q z ) ≈ (cid:12)(cid:12)(cid:12)(cid:12) q z − q tz q z + q tz (cid:12)(cid:12)(cid:12)(cid:12) × (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) j = N X j =0 ρ j +1 − ρ j ρ w e − q z q tz σ j − iz j p q z q tz !(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) . (2)where q tz = p q z − q c . In the general case, the desired struc-ture is found by fitting Eq. (2) with 3 N + 1 free parameters ρ j , l j , and σ j to experimental data for R ( q z ) .In the experiment, thermal fluctuations of the surface(capillary waves) in the illuminated region ( ∼ ) makea contribution to the observed structure, which smearsjumps in the electron density profile [26]. In the calculations,the capillary width σ = k B T / (2 πγ ( A )) ln( Q max /Q min ),where k B is the Boltzmann constant, γ ( A ) = γ − Π( A ), Q max = 2 π/a ( a ≈
10 ˚A is the intermolecular distance) isthe short wavelength limit in the capillary wave spectrum,and Q min = q maxz ∆ β (∆ β ≈ . ◦ is the angular resolutionof the detector) is the long-wavelength limit of surface fluc-tuations involved in the experiment, was used as the param-eter σ j for the boundaries of the monolayer. We successfullyused such an approach to analysis of reflectometry data, e.g.,to study structures and phase transitions in adsorption am-phiphilic films on macroscopically planar oilwater interfaces[27, 28, 29, 30, 31]. RESULTS
All experimental dependences R ( q z ) can be quite well de-scribed within two- and three-layer models. The structure ofthe waterlipidsair interface can be conditionally representedin the form of two layers ( N = 2). In view of the chemicalstructure of the DMPS molecule, it is reasonable to suggestthat the first layer with the thickness L = 10 ÷
15 ˚A andelectron density ρ = 0 . ÷ . ρ w is formed by aliphaticsegments of the molecule and its C H tails. The sec-ond layer with the thickness L ∼
10 ˚A and electron den-
Figure 4.
Electron density profiles ρ ( z ) in units of the electrondensity in water under normal condition ρ w ≈ . e − /˚A forthe monolayer (dashed lines) without and (solid lines) with thepolymer. The point z = 0 is chosen at the interface betweenlipid molecules with air. Lines a and b correspond to the liquid(LE, A = 72 ˚A ) and condensed (LC, A = 36 ˚A ) states of themonolayer. For convenient comparison, lines b are shifted alongthe y axis by 0.75. sity ρ = 1 . ÷ . ρ w is formed by polar groups of phos-phatidylserine. The parameter σ = 3 ÷ σ , σ ,and σ . There parameters of the DMPS monolayer in theLE and LC states are in agreement with [16].The adsorption of the polymer is clearly manifested inchanges in the reflectivity curves for both states of the mono-layer. This adsorption is described well by the three-layermodel ( N = 3), where the packing parameters ρ and L ofaliphatic tails of the phospholipid are equal to the respec-tive parameters of the pure monolayer in the correspondingstate. Good agreement of the calculated curves with theexperiment is reached by varying the parameters ρ and L of the layer of polar groups, the thickness L and density ρ of the polymer film, and the width σ of the polymerfilmsubphase interface.According to this procedure, the adsorption of the poly-mer first affects the integral density ρ L of the layer ofpolar groups: this density decreases noticeably (by a factorof ≈
2) in the LE phase, whereas this decrease is much less( ∼
30 %) in the LC phase. Second, the fitting ρ value de-pends on the density of the monolayer: the densities of thepolymer film in the LE and LC states are ρ ≈ . ρ w and ρ b ≈ . ρ w , respectively. The parameter L = 150 ±
40 ˚Ais determined with a large error and weakly depends on thedensity of the monolayer. Consequently, a decrease in thearea per molecule in the monolayer from 72 ˚A to 36 ˚A leadsto an almost doubling of the surface density, ρ , of the ad- Figure 5. (Solid lines) Volta potential of DMPS monolayers(see the notation in Fig. 1) versus the density of lipid in themonolayer on the surface of the corresponding solutions. Dot-ted straight lines are approximations of the linear segments ofthe curves. The effective dipole moments estimated from theHelmholtz relation on the linear segments of curves 1 and 2 are0.48 and 0.38 D, respectively. sorption film of macromolecules: (( ρ b − ρ w ) / ( ρ a − ρ w ) ≈ σ ≈
40 ˚A is much larger than the capillarywidth σ , which indicates the diffuse structure of the polymerfilmsubphase interface. The calculated curves for R ( q z ) areshown by solid lines in Figs. 2 and 3 and the correspondingmodel profiles ρ ( z ) are shown in Fig. 4. DISCUSSION
Changes in the boundary potential of monolayers in theliquid crystal state described in many works can be at-tributed to a simple increase in the density of effective dipolemoments of lipid molecules. These changes satisfy well thesimple Helmholtz relation ∆ V = 12 πµ/A , where V is thechange in the boundary potential in millivolts, µ is the ef-fective dipole moment in millidebyes, and A is the area permolecule in angstroms squared [32]. This linear dependenceis observed in the segments of curves shown in Fig. 1b thatcorrespond to the LE state of the monolayer and are shownin Fig. 5. The slope of this dependence becomes muchsmaller in the presence of high molecular polypeptide in thesolution. This means that the effective dipole moment inthese segments of compression of the monolayer decreasesfrom 0.48 to 0.38 D. This value is obviously the sum of alldipole moments that are determined by the structure of thephospholipid and are due to the partial immersion of cationsof the electrolyte in the polar region and to the orientation ofwater molecules associated with this region. For this reason,the molecular interpretation of the decrease in the effectivedipole moment is a complex problem, which can be solved, e.g., by numerical molecular dynamics simulation of thesesystems.According to the measurement and the compression dia-gram of DMPS monolayers on the surface of KCl solutions inthe presence of polylysine in it (Fig. 1), the effect of poly-D-lysine is the most pronounced in the region of the LEstate. A larger slope of experimental curves reflects an in-crease in the stiffness of the monolayer under compression.At the same time, the phase transition in the condensedstate begins at the same area per lipid molecule below 70 ˚A per molecule. The compression curve in the condensed statehas a slope close to that of the pure DMPS monolayer, but isshifted toward larger areas per lipid molecule. This area ata lateral pressure of ∼
40 mN/m increases from 39 to 46 ˚A .The polymer most probably groups lipids to denser clusters.Indeed, if the polymer filled aqueous cavities in the mono-layer, the compression curves would be shifted toward largerareas. Since polylysine in test experiments did not exhibitsurface activity, the shift of the curves toward larger areasper molecule in the LC state cannot be attributed to the in-troduction of links of the polymer between lipid molecules,as is possible for hydrophobic molecules. CONCLUSION
To summarize the above results, we note that the choiceof the polypeptide, its density, and a sufficiently large molec-ular mass ensures the maximum expected filling of the sur-face by macromolecules. Furthermore, according to studiesof the adsorption of polylysines on the surface of bilayerlipid membranes, in particular, black lipid membranes andthose in the suspension of liposomes, their adsorption occursonly on negatively charged surfaces and with a very highefficiency, filling the entire surface available for adsorption[13, 33]. The character of the filling of the surface dependson the polymerbilayer equilibrium constant. This constantfor polylysines with a long chain estimated within a com-paratively simple theoretical model is no more than 10 ,which means their irreversible coupling with phospholipids[34]. However, it was found that the presence of a layerof polylysines does not affect the stability of membranes,in contrast to polymers with hydrophobic segments, whichcan be partially incorporated into a membrane and createpores penetrable for ions. These facts mean that the shiftof the compression curves toward larger areas per moleculeseen in Fig. 1 is due not to the introduction of the polymerinto the structure of the monolayer but is due to an increasein its stiffness in the LE state. The compressibility of themonolayer in the LC state hardly depends on the presenceof the polymer in the aqueous substrate. This conclusion isconfirmed by X-ray reflectometry data in Figs. 2 and 3 (theshape of the reflection curve for the LC state of the mono-layer at the adsorption of poly-D-lysine generally does notchange) and by the reconstructed electron density profiles inFig. 4. Changes in the reflectivity curves are the most sig-nificant in the region where the monolayer behaves as a two-dimensional liquid. The structural parameters of the inter-face estimated within the two-layer model are in good agree-ment with the geometrical characteristics of the amphiphiliclipid molecule. We believe that the described methods for the study of complex polymerlipid systems at the waterairinterface are useful for various charged macromolecules hav-ing biomedical applications. Corresponding experimentalresults will allow confirming or rejecting molecular struc-tures actively analyzed by molecular dynamics methods.
ACKNOWLEDGEMENTS
The experiments were performed on the ID31 beamlineof the European Synchrotron Radiation Facility (Grenoble,France) within the SC-4845 project. We are grateful to He-lena Isern and Florian Russello (ESRF) for assistance in thepreparation of the experiments. This work was supported bythe Ministry of Science and Higher Education of the RussianFederation (state assignments for the indicated institutions)and by the Russian Foundation for Basic Research (projectno. 16-04-00556).
References [1] H. Mohwald,
Phospholipid monolayers . In:
Handbook of Bi-ological Physics , Ed. R. Lipowsky, E. Sackmann. ElsevierScience B.V., pp. 161-211 (1995).[2] C. Stefaniu, G. Brezesinski, H. Mohwald, Adv. Colloid In-terface Sci 208, 197 (2014).[3] G. Cevc, D. Marsh,
Phospholipid Bilayers. Physical Princi-ples and Models , Cell Biology: A Series of Monographs, Eds.E. E. Bittar, V. 5, p. 442, New York: Wiley-IntersciencePublication, 1987.[4] M. Loesche, E. Sackmann, and H. Mohwald, Ber. Bunsenges.Phys. Chem. 87, 848 (1983).[5] V. M. Kaganer, H. Mohwald, and P. Dutta, Rev. Mod. Phys.71, 779 (1999).[6] M. C. Shih, T. M. Bohanon, J. M. Mikrut, P. Zschack, andP. Dutta, Phys. Rev. A 45, 5734 (1992).[7] I. R. Peterson, V. Brzezinski, R. M. Kenn, and R. Steitz,Langmuir 8, 2995 (1992).[8] C. A. Helm, P. Tippmann-Krayer, H. Mohwald, J. Als-Nielsen, and K. Kjaer, Biophys. J. 60, 1457 (1991).[9] R. M. Kenn, K. Kjaer, and H. Mohwald, Colloids Surf. A117, 171 (1996).[10] K. de Meijere, G. Brezesinski, and H. Mohwald, Macro-molecules 30, 2337 (1997).[11] Y. A. Ermakov, K. Kamaraju, K. Sengupta, S. Sukharev,Biophys J. 98(6), 1018 (2010).[12] Yu. Ermakov, K. Kamaraju, A. Dunina-Barkovskaya,K. Vishnyakova, Y. Egorov, A. Anishkin, S. Sukharev, Bio-chemistry 56, 5457 (2017).[13] N. Marukovich, M. McMurray, O. Finogenova,A. Nesterenko, O. Batishchev, Yu. Ermakov,
Interac-tion of polylysines with the surface of lipid membranes:the electrostatic and structural aspects . In: