- Email: [email protected]
Effects of ion interactions with a cholesterol-rich bilayer
Biochemical and Biophysical Research Communications 468 (2015) 125e129
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Effects of ion interactions with a cholesterol-rich bilayer Lingxue Mao a, Linlin Yang b, Qiansen Zhang b, Hualiang Jiang a, b, **, Huaiyu Yang b, * a
School of Pharmacy, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China Drug Discovery and Design Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Science, 555 Zuchongzhi Road, Shanghai 201203, China
b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 October 2015 Accepted 27 October 2015 Available online 31 October 2015
Previous molecular dynamics (MD) simulations of ion-lipid interactions have focused on pure phospholipid bilayers. Many functional microdomains in membranes have a complex composition of cholesterol and phospholipids. Here, we reveal the distinctiveness of the interactions and the effects of the ions on a cholesterol-rich bilayer by performing MD simulations of a cholesterol-rich bilayer with a Naþ/Kþ mixture or a Naþ/Kþ/Ca2þ/Mg2þ mixture. The simulations reveal that Ca2þ maintains its dominant role in the interaction with the cholesterol-rich bilayer, but the binding affinity of Mg2þ to the cholesterol-rich bilayer is even weaker than the affinities of Naþ and Kþ, whereas its interaction with pure phospholipid bilayers is strong and is only slightly weaker than that of Ca2þ. Additionally, it was found that the presence of additional divalent cations induces the headgroups of phospholipids to be more perpendicular to the membrane surface, reducing the lateral movement of lipids and slightly altering the ordering and packing of the cholesterol-rich bilayer, different from divalent cations, which strongly influence that ordering and packing of pure phospholipid bilayers. Therefore, this study indicates that cholesterol in the membrane could affect the interactions between membrane and cations. The findings could be helpful in understanding the biological processes relevant to regulation of cations in cholesterol-rich regions. © 2015 Elsevier Inc. All rights reserved.
Keywords: Ion mixture Cholesterol-rich bilayer Molecular dynamics Binding affinity
1. Introduction Biological membranes in cells provide a barrier between the inside and outside environments and are involved in a variety of cellular functions such as signal transduction, substance exchange, and cellecell communication and recognition [1]. Biological membranes are complex assembled structures that consist mainly of phospholipids, proteins and cholesterol, their compositions and distributions regulated differently for different membranes [2]. In physiological environments, biological membranes are immersed in solutions of salt ions mainly including Naþ, Kþ, Ca2þ, Mg2þ, and Cl that are considered the most important biologically relevant salt ions. Simplified lipid bilayer models are often used to
Abbreviations: POPE, 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylethanola mine; POPS, 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine; Chol, Cholesterol. * Corresponding author. ** Corresponding author. School of Pharmacy, East China University of Science and Technology, 130 Meilong Road, Shanghai, 200237, China. E-mail addresses: [email protected] (H. Jiang), [email protected] (H. Yang). http://dx.doi.org/10.1016/j.bbrc.2015.10.149 0006-291X/© 2015 Elsevier Inc. All rights reserved.
investigate the interactions and the influence of salt ions on the membranes. The simplest case of a membrane based on pure phospholipid bilayers has been extensively studied [3e13]. Lipid mixtures were also investigated to extend the understanding of the specificity of ionic binding [14e21]. These studies have revealed that various membrane properties such as bilayer thickness, order parameters, lipid mobility, and surface charge density can be influenced by ion binding. Molecular dynamics (MD) simulation, a powerful tool for investigating the specificity of ion binding, has been widely and successfully used to probe the effects of ions on the lipid bilayers. The effects of ions on the properties of mixed lipid bilayers have been experimentally measured and numerically simulated [19e21]. However, cholesterol is also a key component of biological membranes and is known to modify membrane properties in a number of important ways [22]. The incorporation of certain levels of cholesterol in the membrane increases the packing density of phospholipids and reduces the mobility of lipids [22e24]. As a result, cholesterol may affect the extent of ion absorption and thus the strength of the influence on the membrane. Recently, Magarkar el al [25]. carried out MD simulations of interactions between Naþ
126
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
ions and phospholipid bilayers with increasing levels of cholesterol and showed that increasing the level of cholesterol decreases Naþ binding. However, the understanding of the interactions and effects of ions on the cholesterol-rich bilayer has not yet been fully elucidated. Here, we performed MD simulations to study the structures and dynamics of cholesterol-rich bilayers interacting with Naþ, Kþ, Ca2þ, and Mg2þ ions. A bilayer model consisting of 1-palmitoyl-2oleoyl-sn-glycero-3-phosphatidylethanolamine (POPE), 1palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine (POPS), and cholesterol (Chol) was designed. For clarity, this designed bilayer model was denoted as the PPC bilayer. Based on the PPC bilayer model, we first constructed a simulation system with a mixture of Naþ, Kþ, Ca2þ, and Mg2þ ions. Considering that divalent cations may play a dominant role in altering membrane properties [9], another system containing a mixture of Naþ and Kþ ions was also generated as a control for comparison. 2. Materials and methods To investigate the interactions and effects of cations on the cholesterol-rich bilayers, we constructed a PPC bilayer model composed of 62 POPE (~50 mol%), 22 POPS (~17 mol%), and 44 Chol (~33 mol%). Because the physiological concentrations of Naþ and Kþ ions are much higher than those of Ca2þ and Mg2þ ions, we constructed one system (PPCNaKCaMg) of cholesterol-rich bilayers with 80 Naþ, 80 Kþ, 16 Ca2þ, and 16 Mg2þ ions. To investigate whether divalent cations play the same dominant role in altering cholesterol-rich bilayers properties as they do in pure phospholipid bilayers reported in our previous studies [9,13], another system (PPCNaK) with 80 Naþ, 80 Kþ was constructed as a control for comparison with the PPCNaKCaMg system. Each system was solvated with ~28000 TIP3P [26] water molecules, and Cl ions were added to net-neutralize the systems. More details of the two simulation systems are shown in Fig. S1. Simulation parameters used in the present study are similar to our previous published simulations [13] and are given in the Supplementary Material. 3. Results 3.1. Cation-lipid interactions The effects of the cations on the structural and dynamic properties of a bilayer are directly associated with the cation-lipid interactions. Therefore, we start analysing the cation-lipid interactions for both systems by performing calculations of bound ions and ion-lipid bonds. The computed radial distribution functions (RDFs) of the lipid oxygen atoms relative to the cation position were used to define ion binding. An ion within the cut-off distance from the oxygen atoms was considered a bound ion. The cut-off distances for Naþ, Kþ, Ca2þ and Mg2þ ions are 0.31, 0.36, 0.30, and 0.24 Å, respectively [10]. Fig. 1 shows the time evolution of fractions of different bound ions. It was found that fractions of the bound Naþ and Kþ ions reached equilibrium quickly, whereas an increasing number of Ca2þ ions were observed to bind to lipid oxygen atoms before 700 ns in our simulation systems, indicating that a long time is necessary for the MD simulations of the divalent ions interacting with the cholesterol-rich bilayers. At the same time, the fluctuation of the numbers of bound Naþ and Kþ ions are much greater than that of Ca2þ ions, suggesting that the mean lifetime for a calcium-lipid complex is longer than that for the other cation-lipid complexes. Moreover, while the average percentage of bound Ca2þ ions in the last 200 ns of simulation for the PPCNaKCaMg system is ~100%, it was found that less than ~5% of Kþ and Naþ ions bind to lipids in all simulation systems, and all the Mg2þ ions are
distributed in aqueous solution in the PPCNaKCaMg system. To clearly elucidate the composition of the bound ions, Table S1 presents statistics for the bound ions in different systems based on the last 200 ns of MD trajectories. In the PPCNaK system, ~3.6 of 80 Naþ ions and ~4.0 of 80 Kþ ions are bound to the lipid bilayer. After divalent Mg2þ and Ca2þ ions were added into the bilayers, less bound Naþ and Kþ ions (~1.5 Naþ ions and ~1.5 Kþ ions in PPCNaKþ CaMg system) were found, demonstrating that the binding of Na , þ K ions with cholesterol-rich lipids can be affected by the bound divalent ions. All 16 Ca2þ ions bind the negatively charged PPC bilayer. However, no bound Mg2þ ion was observed. Therefore, of the four studied ions, Ca2þ ions are always the predominantly bound ions and should display dominant effects on the structure of the cholesterol-rich bilayers. We investigated the cation-lipid interaction modes that reflect the intrinsic properties of the binding between cations and lipids. We first obtained the statistics for the numbers of ions that interact with one, two, three, and four lipid molecules. While the major bound Naþ and Kþ ions coordinate with either one or two lipids, most of the bound Ca2þ ions interact with three or four lipids (Fig. 2AeB). To further elucidate which lipid oxygen atoms are usually involved in each ion-lipid interaction, the numbers of different ion-oxygen coordinated pairs were calculated. As shown in Fig. 2D, Ca2þ ions interact with both POPE and POPS oxygen atoms in the PPCNaKCaMg system. With respect to the specific location of bound oxygen atoms, Ca2þ ions interact with phosphate oxygen atoms of POPE and carboxyl oxygen atoms of POPS (Fig. 2D). For bound Naþ and Kþ ions, all four modes (ion-phosphatidyl, ioncarbonyl, ion-hydroxy, and ion-carboxyl), among which the ionhydroxy bonds are the least observed, contribute to the ion-lipid interaction (Fig. 2CeD). At the same time, it should be noticed that monovalent cations can coordinate with cholesterol by an ionhydroxy bond in the cholesterol-rich bilayers (Fig. 2CeD). 3.2. Effects of ion mixtures on the PPC bilayer The work described above elucidated different cation-lipid interactions for two simulation systems with different ion mixtures. To reveal the influences of ion mixtures on the structure of cholesterol-rich bilayers, we calculate six structural and dynamical parameters, i.e., headgroup orientation, bilayer thickness, area per lipid, ordering of hydrophobic tails, diffusion coefficient, and charge density distributions, as described below. The interaction of the cations with the membrane may alter the orientation of lipid headgroups. The orientation of a lipid headgroup was defined as the angle (q) between the P/N vector (from the phosphorus atom to the nitrogen atom in the headgroup) and the outward normal axis of the bilayer (Fig. 3A). This angle specifies whether the lipid headgroups are parallel or perpendicular to the membrane surface. Fig. 3A shows the distribution probability of q in the PPC bilayer for the two systems with different ion mixtures. The peak of the profiles for POPE and POPS molecules in the PPCNaK system are ~82.4 and ~72.7, with almost identical average values of q (~77.1 and ~77.6 ). Interestingly, in the PPCNaKCaMg system, two peaks of the angle probability distribution profiles for both POPE and POPS could be found. The two peaks of the profiles for POPE and POPS in the PPCNaKCaMg system are ~68.2 , ~87.1 and ~32.8 and ~52.9 , with the average q values of ~67.5 and ~58.2 , respectively. These results indicate that the presence of additional divalent cations results in a significant reorientation of the lipid headgroups. The angle probability distributions for both POPE and POPS were broadened and shifted toward lower angles, especially for anionic POPS lipids, suggesting that the headgroups of POPE and POPS in the PPCNaKCaMg system are more perpendicular to the membrane surface than those in the PPCNaK system. This result is
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
127
Fig. 1. Time evolution of the fractions of bound ions in PPCNaK (A) and PPCNaKCaMg (B) systems.
Fig. 2. Average numbers of ions binding to one, two, three, four lipids in PPCNaK (A) and PPCNaKCaMg systems (B). Number of ion-oxygen coordinated pairs (ion-phosphatidyl (PO4 ), ion-carboxyl (COO), ion-carbonyl (C]O), ion-hydroxy (OH)) for the PPCNaK (C) and PPCNaKCaMg (D) systems.
consistent with results of previous simulation that found that bound Ca2þ leads to lipid headgroups becoming oriented more perpendicular to the membrane surface [9,13]. The bilayer thickness mainly depends on the width of the membrane hydrophobic region. Here, the bilayer thickness (DP-P) was defined by the average distance between the centre of mass of the phosphate atoms in the two leaflets of the membrane at the normal direction. In the PPCNaK and PPCNaKCaMg systems, the DP-P values are ~48.1 and ~47.7 Å, respectively (Fig. S2). The area per lipid (AL) in the ternary mixture was calculated by the MEMBPLUGIN analysis tool in VMD [27], which is an automated tool for analysing MD simulations of complex membranes. The AL values of POPE, POPS, and Chol in the PPCNaK system are ~49.2, ~48.5, and ~25 Å2, respectively, and in the PPCNaKCaMg system these are ~48.7, ~49.4, and ~25.5 Å2 (Table S2). However, considering the margin of error
(Table S2), no definitive difference in the AL values can be observed. The ordering of the two hydrophobic tails of the lipid is usually characterized by the deuterium order parameter SCD that can be determined by NMR experiments [28]. The order parameter profiles of the four chains (Sn-1 and Sn-2 of POPE, Sn-1 and Sn-2 of POPS, as shown in Fig. S1A) for the PPCNaK and PPCNaKCaMg systems are plotted in Fig. 3B. Although all 16 Ca2þ ions bind to the negatively charged PPC bilayer as mentioned before, no profound differences were observed between the order parameter profiles of the PPCNaKePPCNaKCaMg (Fig. 3B). In fact, the deuterium order parameter profiles of phospholipid molecules in our simulation systems are globally larger than those of the pure phospholipid bilayers reported in previous studies [7,29]. This is in agreement with previous studies demonstrating that a significant ordering effect of the cholesterol on phospholipid chains can be observed
128
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
Fig. 3. The influence of Naþ, Kþ, Ca2þ, and Mg2þ ions on the cholesterol-rich bilayer. (A) Angular distribution of the headgroup PeN vector with respect to the bilayer normal. (B) Molecular order parameter (SCD) profiles in PPC bilayers calculated for the Sn-1 chain and Sn-2 chain of POPE (top) and POPS molecules (bottom), respectively.
[23]. To characterize the lateral mobility of lipid molecules that could be influenced by bound ions, we focus on the lipid molecules' lateral diffusion coefficient DL The DL of lipids was evaluated as DL ¼ limt/∞ ð1=4tÞ½rðtÞ2 and calculated from the long-time mean square displacement ½rðtÞ2 of the centre of mass positions of lipids versus time in the bilayer plane. The DL values of the lipids in the PPCNaK system are DPOPE ¼ 2.5 ± 0.3 108 cm2/s, DPOPS ¼ 2.3 ± 0.03 108 cm2/s, DChol ¼ 2.3 ± 0.03 108 cm2/s, while in the PPCNaKCaMg system are DPOPE ¼ 0.50 ± 0.07 108 cm2/s, DPOPS ¼ 0.57 ± 0.03 108 cm2/s, DChol ¼ 0.53 ± 0.07 108 cm2/s. The DL for the cholesterol-rich bilayers is within the wide range of experimentally values [30]. Clearly, the additional divalent cations that greatly decrease the DL values hinder the lateral movements of lipid molecules in the PPC bilayer. The charge density distribution perpendicular to the bilayer surface is also an important property of a plasma membrane [10]. Consistent with the chemical structure of the POPE and POPS molecules (Fig. S1A), the distributions of lipid charges exhibit two distinct regions: a negative one (2.7 nm), provided by the phosphate groups and, and a positive one caused by the choline groups (~3.0 nm) (Fig. S3). In the PPCNaK system, the compensation of nonzero charge for the phosphate and choline groups in the lipids is mainly contributed by water molecules rather than the monovalent ions owing to their negligible binding (Fig. S3). By contrast, in the PPCNaKCaMg system, Ca2þ ions penetrate deep into the bilayer and the charge density profile of the cations demonstrates a single prominent maximum for the compensation of a net negative charge of lipid headgroup (Fig. S3). At the same time, because bound Ca2þ ions alter the orientation of lipid headgroups (Fig. 3A), they may induce the reorientation of water molecules and thereby affecting the charge density profiles of lipids and water molecules, as shown in Fig. S3. As a consequence of these effects, the distribution of the membrane surface charges in the PPCNaK system is similar to that found in the PPCNaKCaMg system (Fig. S3). 4. Discussion It is well known that the specific interactions of cations with biological membranes alter the properties of membranes and affect the way the proteins bind or are inserted into membranes. MD simulations have contributed to providing a detailed microscopic picture of the ionebilayer interactions. Previous MD simulation studies on ionemembrane interactions were performed based on the lipid models of either pure phospholipid or phospholipid mixtures. In the present study, we designed a cholesterol-rich
bilayer model (negatively charged PPC bilayer). Two simulation systems were obtained by either adding 80 Naþ, 80 Kþ or 80 Naþ, 80 Kþ, 16 Ca2þ, and 16 Mg2þ into the cholesterol-rich bilayers and used to investigate the interactions and effects of ion mixtures on the cholesterol-rich bilayer. From the numbers of bound ions (Table S1), we conclude that Ca2þ ions have the highest affinity to the water/ membrane interface of cholesterol-rich bilayer compared to other cations (Fig. 1), similar to the conclusion regarding the interaction between cations and pure phospholipid bilayers [13]. In particular, whereas all 16 Ca2þ ions bind to the anionic PPC bilayer, none of the Mg2þ ions was found to bind to theses cholesterol-rich bilayers, and both Naþ and Kþ ions are very weakly bound, indicating that the interactions and effects of Mg2þ, Naþ, and Kþ ions are negligible in this cholesterol-rich bilayer. Clearly, Ca2þ ions show the highest affinity for the cholesterol-rich bilayers compared to other cations, making Ca2þ effects dominant for the structure of the cholesterolrich bilayers due to Ca2þ ions. However, previous studies have supposed that the binding affinity of Mg2þ ion was slightly weaker than that of the Ca2þ ion in pure phospholipid bilayers and that the binding affinity of Naþ ion should be stronger than that of the Kþ ion [3,10,13]. In our simulations, none of the Mg2þ ions was found to bind to the cholesterolrich bilayer, and both Naþ and Kþ ions are very weakly bound, indicating that the interactions and effects of Mg2þ, Naþ, and Kþ ions are negligible in the cholesterol-rich bilayer. We should note that cholesterol molecules are randomly distributed in the bilayer region, as presented in Fig. S4. Owing to its rigid ring systems, cholesterol could separate the neighbouring phospholipid molecules, constrain the movement of phospholipid molecules in the process of ion-lipid interaction, and further broaden the distance of neighbouring lipid headgroups in the cholesterol-rich bilayers. The weak binding affinities of the monovalent Naþ and Kþ ions for the cholesterol-rich bilayer could be explained by the negative effect of cholesterol on cation-lipid interactions as discussed above. The smaller size of the divalent Mg2þ ions has been previously hypothesized to cause them to fit less effectively between the headgroups compared to Ca2þ ions [3,13]. Moreover, when a Mg2þ ion binds to the cholesterol-rich bilayer, the coordination of the nearby phosphate oxygen atoms could be inhibited by the presence of neighbouring cholesterol molecules. Additionally, the competition of Ca2þ with Mg2þ for binding to the lipids [13] should also account for this effect. It is known that ion binding to the phospholipid bilayer produces more ordered bilayers accompanied with thicker bilayers and smaller areas per lipid [5,6,8]. Because Ca2þ shows the highest binding affinity to the water/membrane interface in the cholesterol-rich bilayer, and the binding of Mg2þ, Naþ, and Kþ ions
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
was found to be negligible in present study, we may expect that Ca2þ ions play a dominant role in altering bilayer properties. To reveal the influences of ion mixtures on the properties of cholesterol-rich bilayers, the structural and dynamic parameters of the two simulation systems were calculated and compared. However, we found that the additional divalent cations could only slightly alter the ordering and packing of cholesterol-rich bilayers. This result indicates that the bound cations may affect the structure of cholesterol-rich bilayer in a different way than those of the pure phospholipid bilayers. As reported, cholesterol could order the hydrocarbon chains of the neighbouring phospholipids and reduce the surface area of the lipid bilayer [23,24,31], indicating that the cholesterol-rich bilayers are already well packed and ordered. Consequently, less space is left for further condensation and the ordering of hydrophobic chains when the cations interact with this cholesterol-rich bilayer. We also discovered that Ca2þ binding induces the lipid headgroup to be more perpendicular to the membrane surface, and reduce the lipid mobility of the cholesterol-rich bilayer. The present MD simulations lead to the following conclusions. First, compared to other cations, Ca2þ ions have the highest binding affinity to the cholesterol-rich bilayers, consistent with the previous conclusion regarding pure phospholipid bilayers. Second, different from the results of our previous study regarding pure phospholipid bilayers, the interactions of Mg2þ, Naþ and Kþ ions are all relatively weak in the cholesterol-rich bilayer, indicating that cholesterol could negatively affect the binding affinities of cations to the cholesterol-rich bilayers. Third, the additional divalent cations induce phospholipid headgroups to be more perpendicular to the membrane surface and reduce the lateral movement of lipids but slightly alter the ordering and packing of the cholesterol-rich bilayer. This study gains a new and further insight on ionemembrane interactions and could be helpful in understanding the biological processes relevant to cations regulations in cholesterol-rich regions.
[4]
[5]
[6] [7]
[8]
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16] [17]
[18]
[19]
Conflicts of interest
[20]
The authors declare that there are no known conflicts of interest associated with this publication.
[21]
[22]
Acknowledgement [23]
This work was supported by the National Natural Science Foundation of China (grant numbers 21422208, 81173027, 81230076 and 21210003), and the SA-SIBS Scholarship Program. We thank the National Supercomputing Center in Tianjin (Tianhe 1A) and the National Supercomputing Center in Jinan for computational resources.
[25]
Appendix A. Supplementary data
[27]
Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.bbrc.2015.10.149.
[28]
References [1] M. Edidin, Lipids on the frontier: a century of cell-membrane bilayers, Nat. Rev. 4 (2003) 414e418. [2] S. Freeman, K. Quillin, L. Allison, Lipids, Membranes, and the First Cells, Biology and the Tree of Life, Benjamin Cummings, San Francisco, 2011, pp. 99e124. € rnig, The effect of metal cations on the phase behavior and [3] H. Binder, O. Zscho
[24]
[26]
[29] [30]
[31]
129
hydration characteristics of phospholipid membranes, Chem. Phys. Lipids 115 (2002) 39e61. S.A. Pandit, M.L. Berkowitz, Molecular dynamics simulation of dipalmitoylphosphatidylserine bilayer with Naþ counterions, Biophys. J. 82 (2002) 1818e1827. S.A. Pandit, D. Bostick, M.L. Berkowitz, molecular dynamics simulation of a dipalmitoylphosphatidylcholine bilayer with NaCl, Biophys. J. 84 (6) (2003) 3743e3750. €ckmann, A. Hac, T. Heimburg, H. Grubmüller, Effect of sodium chloride R.A. Bo on a lipid bilayer, Biophys. J. 85 (2003) 1647e1655. P. Mukhopadhyay, L. Monticelli, D.P. Tieleman, Molecular dynamics simulation of a palmitoyl-oleoyl phosphatidylserine bilayer with Naþ counterions and NaCl, Biophys. J. 86 (2004) 1601e1609. U.R. Pedersen, C. Leidy, P. Westh, G.H. Peters, The effect of calcium on the properties of charged phospholipid bilayers, Biochim. Biophys. Acta 1758 (2006) 573e582. H. Yang, Y. Xu, Z. Gao, Y. Mao, Y. Du, H. Jiang, Effects of Naþ ,Kþ, and Ca2þ on the structures of anionic lipid bilayers and biological implication, J. Phys. Chem. B 114 (2010) 16978e16988. A. Cordomi, O. Edholm, J.J. Perez, Effect of ions on a dipalmitoyl phosphatidylcholine bilayer. A molecular dynamics simulation study, J. Phys. Chem. B 112 (2008) 1397e1408. B. Klasczyk, V. Knecht, Validating affinities for ion-lipid association from simulation against experiment, J. Phys. Chem. A 115 (2011) 10587e10595. R.A. Bockmann, H. Grubmuller, Multistep binding of divalent cations to phospholipid bilayers: a molecular dynamics study, Angew. Chem. Int. Ed. Engl. 43 (2004) 1021e1024. Y. Mao, Y. Du, X. Cang, J. Wang, Z. Chen, H. Yang, H. Jiang, Binding competition to the POPG lipid bilayer of Ca2þ, Mg2þ, Naþ, and Kþ in different ion mixtures and biological implication, J. Phys. Chem. B 117 (2013) 850e858. M. Roux, M. Bloom, Ca2þ, Mg2þ,Liþ, Naþ, and Kþ Distributions in the headgroup region of binary membranes of phosphatidylcholine and phosphatidylserine as seen by deuterium, Biochemistry 29 (1990) 7077e7089. D. Huster, K. Arnold, K. Gawrisch, Strength of Ca2þ binding to retinal lipid membranes: consequences for lipid organization, Biophys. J. 78 (2000) 3011e3018. B.Y. Ha, Stabilization and destabilization of cell membranes by multivalent ions, Phys. Rev. E 64 (2001). S.A. Pandit, D. Bostick, M.L. Berkowitz, Mixed bilayer containing dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylserine: lipid complexation, ion binding, and electrostatics, Biophys. J. 85 (2003) 3120e3131. C.G. Sinn, M. Antonietti, R. Dimova, Binding of calcium to phosphatidylcholineephosphatidylserine membranes, Coll. Surf. Physicochem. Eng. Asp. 282e283 (2006) 410e419. I. Reviakine, A. Simon, A. Brisson, Effect of Ca2þ on the morphology of mixed DPPC-DOPS supported phospholipid bilayers, Langmuir 16 (2000) 1473e1477. P.T. Vernier, M.J. Ziegler, R. Dimova, Calcium binding and head group dipole angle in phosphatidylserine-phosphatidylcholine bilayers, Langmuir 25 (2009) 1020e1027. P. Jurkiewicz, L. Cwiklik, A. Vojtiskova, P. Jungwirth, M. Hof, Structure, dynamics, and hydration of POPC/POPS bilayers suspended in NaCl, KCl, and CsCl solutions, Biochim. Biophys. Acta 1818 (2012) 609e616. g, M. Pasenkiewicz-Gierula, I. Vattulainen, M. Karttunen, Ordering effects T. Ro of cholesterol and its analogues, Biochim. Biophys. Acta 1788 (2009) 97e121. H. Ohvo-Rekil€ a, B. Ramstedt, P. Leppim€ aki, J.P. Slotte, Cholesterol interactions with phospholipids in membranes, Prog. Lipid Res. 41 (2002) 66e97. g, M. Karttunen, I. Vattulainen, R. Reigada, Cholesterol H. Martinez-Seara, T. Ro induces specific spatial and orientational order in cholesterol/phospholipid membranes, PLoS One 5 (2010) e11162. g, A. Magarkar, V. Dhawan, P. Kallinteri, T. Viitala, M. Elmowafy, T. Ro A. Bunker, Cholesterol level affects surface charge of lipid membranes in saline solution, Sci. Rep. 4 (2014) 5005. W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, Comparison of simple potential functions for simulating liquid water, J. Chem. Phys. 79 (1983) 926e935. R. Guixa-Gonzalez, I. Rodriguez-Espigares, J.M. Ramirez-Anguita, P. CarrioGaspar, H. Martinez-Seara, T. Giorgino, J. Selent, MEMBPLUGIN: studying membrane complexity in VMD, Bioinformatics 30 (2014) 1478e1480. L.S. Vermeer, B.L. de Groot, V. Reat, A. Milon, J. Czaplicki, Acyl chain order parameter profiles in phospholipid bilayers: computation from molecular dynamics simulations and comparison with 2H NMR experiments, Eur. Biophys. J. 36 (2007) 919e931. C. Hong, D.P. Tieleman, Y. Wang, Microsecond molecular dynamics simulations of lipid mixing, Langmuir 30 (2014) 11993e12001. €dd, G. Lindblom, The effect of cholesterol on the lateral A. Filippov, G. Ora diffusion of phospholipids in oriented bilayers, Biophys. J. 84 (2003) 3079e3086. C. Hofs€ aß, E. Lindahl, O. Edholm, Molecular Dynamics Simulations of Phospholipid Bilayers with Cholesterol, Biophys. J. 84 (2003) 2192e2206.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Effects of ion interactions with a cholesterol-rich bilayer Lingxue Mao a, Linlin Yang b, Qiansen Zhang b, Hualiang Jiang a, b, **, Huaiyu Yang b, * a
School of Pharmacy, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China Drug Discovery and Design Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Science, 555 Zuchongzhi Road, Shanghai 201203, China
b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 October 2015 Accepted 27 October 2015 Available online 31 October 2015
Previous molecular dynamics (MD) simulations of ion-lipid interactions have focused on pure phospholipid bilayers. Many functional microdomains in membranes have a complex composition of cholesterol and phospholipids. Here, we reveal the distinctiveness of the interactions and the effects of the ions on a cholesterol-rich bilayer by performing MD simulations of a cholesterol-rich bilayer with a Naþ/Kþ mixture or a Naþ/Kþ/Ca2þ/Mg2þ mixture. The simulations reveal that Ca2þ maintains its dominant role in the interaction with the cholesterol-rich bilayer, but the binding affinity of Mg2þ to the cholesterol-rich bilayer is even weaker than the affinities of Naþ and Kþ, whereas its interaction with pure phospholipid bilayers is strong and is only slightly weaker than that of Ca2þ. Additionally, it was found that the presence of additional divalent cations induces the headgroups of phospholipids to be more perpendicular to the membrane surface, reducing the lateral movement of lipids and slightly altering the ordering and packing of the cholesterol-rich bilayer, different from divalent cations, which strongly influence that ordering and packing of pure phospholipid bilayers. Therefore, this study indicates that cholesterol in the membrane could affect the interactions between membrane and cations. The findings could be helpful in understanding the biological processes relevant to regulation of cations in cholesterol-rich regions. © 2015 Elsevier Inc. All rights reserved.
Keywords: Ion mixture Cholesterol-rich bilayer Molecular dynamics Binding affinity
1. Introduction Biological membranes in cells provide a barrier between the inside and outside environments and are involved in a variety of cellular functions such as signal transduction, substance exchange, and cellecell communication and recognition [1]. Biological membranes are complex assembled structures that consist mainly of phospholipids, proteins and cholesterol, their compositions and distributions regulated differently for different membranes [2]. In physiological environments, biological membranes are immersed in solutions of salt ions mainly including Naþ, Kþ, Ca2þ, Mg2þ, and Cl that are considered the most important biologically relevant salt ions. Simplified lipid bilayer models are often used to
Abbreviations: POPE, 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylethanola mine; POPS, 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine; Chol, Cholesterol. * Corresponding author. ** Corresponding author. School of Pharmacy, East China University of Science and Technology, 130 Meilong Road, Shanghai, 200237, China. E-mail addresses: [email protected] (H. Jiang), [email protected] (H. Yang). http://dx.doi.org/10.1016/j.bbrc.2015.10.149 0006-291X/© 2015 Elsevier Inc. All rights reserved.
investigate the interactions and the influence of salt ions on the membranes. The simplest case of a membrane based on pure phospholipid bilayers has been extensively studied [3e13]. Lipid mixtures were also investigated to extend the understanding of the specificity of ionic binding [14e21]. These studies have revealed that various membrane properties such as bilayer thickness, order parameters, lipid mobility, and surface charge density can be influenced by ion binding. Molecular dynamics (MD) simulation, a powerful tool for investigating the specificity of ion binding, has been widely and successfully used to probe the effects of ions on the lipid bilayers. The effects of ions on the properties of mixed lipid bilayers have been experimentally measured and numerically simulated [19e21]. However, cholesterol is also a key component of biological membranes and is known to modify membrane properties in a number of important ways [22]. The incorporation of certain levels of cholesterol in the membrane increases the packing density of phospholipids and reduces the mobility of lipids [22e24]. As a result, cholesterol may affect the extent of ion absorption and thus the strength of the influence on the membrane. Recently, Magarkar el al [25]. carried out MD simulations of interactions between Naþ
126
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
ions and phospholipid bilayers with increasing levels of cholesterol and showed that increasing the level of cholesterol decreases Naþ binding. However, the understanding of the interactions and effects of ions on the cholesterol-rich bilayer has not yet been fully elucidated. Here, we performed MD simulations to study the structures and dynamics of cholesterol-rich bilayers interacting with Naþ, Kþ, Ca2þ, and Mg2þ ions. A bilayer model consisting of 1-palmitoyl-2oleoyl-sn-glycero-3-phosphatidylethanolamine (POPE), 1palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine (POPS), and cholesterol (Chol) was designed. For clarity, this designed bilayer model was denoted as the PPC bilayer. Based on the PPC bilayer model, we first constructed a simulation system with a mixture of Naþ, Kþ, Ca2þ, and Mg2þ ions. Considering that divalent cations may play a dominant role in altering membrane properties [9], another system containing a mixture of Naþ and Kþ ions was also generated as a control for comparison. 2. Materials and methods To investigate the interactions and effects of cations on the cholesterol-rich bilayers, we constructed a PPC bilayer model composed of 62 POPE (~50 mol%), 22 POPS (~17 mol%), and 44 Chol (~33 mol%). Because the physiological concentrations of Naþ and Kþ ions are much higher than those of Ca2þ and Mg2þ ions, we constructed one system (PPCNaKCaMg) of cholesterol-rich bilayers with 80 Naþ, 80 Kþ, 16 Ca2þ, and 16 Mg2þ ions. To investigate whether divalent cations play the same dominant role in altering cholesterol-rich bilayers properties as they do in pure phospholipid bilayers reported in our previous studies [9,13], another system (PPCNaK) with 80 Naþ, 80 Kþ was constructed as a control for comparison with the PPCNaKCaMg system. Each system was solvated with ~28000 TIP3P [26] water molecules, and Cl ions were added to net-neutralize the systems. More details of the two simulation systems are shown in Fig. S1. Simulation parameters used in the present study are similar to our previous published simulations [13] and are given in the Supplementary Material. 3. Results 3.1. Cation-lipid interactions The effects of the cations on the structural and dynamic properties of a bilayer are directly associated with the cation-lipid interactions. Therefore, we start analysing the cation-lipid interactions for both systems by performing calculations of bound ions and ion-lipid bonds. The computed radial distribution functions (RDFs) of the lipid oxygen atoms relative to the cation position were used to define ion binding. An ion within the cut-off distance from the oxygen atoms was considered a bound ion. The cut-off distances for Naþ, Kþ, Ca2þ and Mg2þ ions are 0.31, 0.36, 0.30, and 0.24 Å, respectively [10]. Fig. 1 shows the time evolution of fractions of different bound ions. It was found that fractions of the bound Naþ and Kþ ions reached equilibrium quickly, whereas an increasing number of Ca2þ ions were observed to bind to lipid oxygen atoms before 700 ns in our simulation systems, indicating that a long time is necessary for the MD simulations of the divalent ions interacting with the cholesterol-rich bilayers. At the same time, the fluctuation of the numbers of bound Naþ and Kþ ions are much greater than that of Ca2þ ions, suggesting that the mean lifetime for a calcium-lipid complex is longer than that for the other cation-lipid complexes. Moreover, while the average percentage of bound Ca2þ ions in the last 200 ns of simulation for the PPCNaKCaMg system is ~100%, it was found that less than ~5% of Kþ and Naþ ions bind to lipids in all simulation systems, and all the Mg2þ ions are
distributed in aqueous solution in the PPCNaKCaMg system. To clearly elucidate the composition of the bound ions, Table S1 presents statistics for the bound ions in different systems based on the last 200 ns of MD trajectories. In the PPCNaK system, ~3.6 of 80 Naþ ions and ~4.0 of 80 Kþ ions are bound to the lipid bilayer. After divalent Mg2þ and Ca2þ ions were added into the bilayers, less bound Naþ and Kþ ions (~1.5 Naþ ions and ~1.5 Kþ ions in PPCNaKþ CaMg system) were found, demonstrating that the binding of Na , þ K ions with cholesterol-rich lipids can be affected by the bound divalent ions. All 16 Ca2þ ions bind the negatively charged PPC bilayer. However, no bound Mg2þ ion was observed. Therefore, of the four studied ions, Ca2þ ions are always the predominantly bound ions and should display dominant effects on the structure of the cholesterol-rich bilayers. We investigated the cation-lipid interaction modes that reflect the intrinsic properties of the binding between cations and lipids. We first obtained the statistics for the numbers of ions that interact with one, two, three, and four lipid molecules. While the major bound Naþ and Kþ ions coordinate with either one or two lipids, most of the bound Ca2þ ions interact with three or four lipids (Fig. 2AeB). To further elucidate which lipid oxygen atoms are usually involved in each ion-lipid interaction, the numbers of different ion-oxygen coordinated pairs were calculated. As shown in Fig. 2D, Ca2þ ions interact with both POPE and POPS oxygen atoms in the PPCNaKCaMg system. With respect to the specific location of bound oxygen atoms, Ca2þ ions interact with phosphate oxygen atoms of POPE and carboxyl oxygen atoms of POPS (Fig. 2D). For bound Naþ and Kþ ions, all four modes (ion-phosphatidyl, ioncarbonyl, ion-hydroxy, and ion-carboxyl), among which the ionhydroxy bonds are the least observed, contribute to the ion-lipid interaction (Fig. 2CeD). At the same time, it should be noticed that monovalent cations can coordinate with cholesterol by an ionhydroxy bond in the cholesterol-rich bilayers (Fig. 2CeD). 3.2. Effects of ion mixtures on the PPC bilayer The work described above elucidated different cation-lipid interactions for two simulation systems with different ion mixtures. To reveal the influences of ion mixtures on the structure of cholesterol-rich bilayers, we calculate six structural and dynamical parameters, i.e., headgroup orientation, bilayer thickness, area per lipid, ordering of hydrophobic tails, diffusion coefficient, and charge density distributions, as described below. The interaction of the cations with the membrane may alter the orientation of lipid headgroups. The orientation of a lipid headgroup was defined as the angle (q) between the P/N vector (from the phosphorus atom to the nitrogen atom in the headgroup) and the outward normal axis of the bilayer (Fig. 3A). This angle specifies whether the lipid headgroups are parallel or perpendicular to the membrane surface. Fig. 3A shows the distribution probability of q in the PPC bilayer for the two systems with different ion mixtures. The peak of the profiles for POPE and POPS molecules in the PPCNaK system are ~82.4 and ~72.7, with almost identical average values of q (~77.1 and ~77.6 ). Interestingly, in the PPCNaKCaMg system, two peaks of the angle probability distribution profiles for both POPE and POPS could be found. The two peaks of the profiles for POPE and POPS in the PPCNaKCaMg system are ~68.2 , ~87.1 and ~32.8 and ~52.9 , with the average q values of ~67.5 and ~58.2 , respectively. These results indicate that the presence of additional divalent cations results in a significant reorientation of the lipid headgroups. The angle probability distributions for both POPE and POPS were broadened and shifted toward lower angles, especially for anionic POPS lipids, suggesting that the headgroups of POPE and POPS in the PPCNaKCaMg system are more perpendicular to the membrane surface than those in the PPCNaK system. This result is
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
127
Fig. 1. Time evolution of the fractions of bound ions in PPCNaK (A) and PPCNaKCaMg (B) systems.
Fig. 2. Average numbers of ions binding to one, two, three, four lipids in PPCNaK (A) and PPCNaKCaMg systems (B). Number of ion-oxygen coordinated pairs (ion-phosphatidyl (PO4 ), ion-carboxyl (COO), ion-carbonyl (C]O), ion-hydroxy (OH)) for the PPCNaK (C) and PPCNaKCaMg (D) systems.
consistent with results of previous simulation that found that bound Ca2þ leads to lipid headgroups becoming oriented more perpendicular to the membrane surface [9,13]. The bilayer thickness mainly depends on the width of the membrane hydrophobic region. Here, the bilayer thickness (DP-P) was defined by the average distance between the centre of mass of the phosphate atoms in the two leaflets of the membrane at the normal direction. In the PPCNaK and PPCNaKCaMg systems, the DP-P values are ~48.1 and ~47.7 Å, respectively (Fig. S2). The area per lipid (AL) in the ternary mixture was calculated by the MEMBPLUGIN analysis tool in VMD [27], which is an automated tool for analysing MD simulations of complex membranes. The AL values of POPE, POPS, and Chol in the PPCNaK system are ~49.2, ~48.5, and ~25 Å2, respectively, and in the PPCNaKCaMg system these are ~48.7, ~49.4, and ~25.5 Å2 (Table S2). However, considering the margin of error
(Table S2), no definitive difference in the AL values can be observed. The ordering of the two hydrophobic tails of the lipid is usually characterized by the deuterium order parameter SCD that can be determined by NMR experiments [28]. The order parameter profiles of the four chains (Sn-1 and Sn-2 of POPE, Sn-1 and Sn-2 of POPS, as shown in Fig. S1A) for the PPCNaK and PPCNaKCaMg systems are plotted in Fig. 3B. Although all 16 Ca2þ ions bind to the negatively charged PPC bilayer as mentioned before, no profound differences were observed between the order parameter profiles of the PPCNaKePPCNaKCaMg (Fig. 3B). In fact, the deuterium order parameter profiles of phospholipid molecules in our simulation systems are globally larger than those of the pure phospholipid bilayers reported in previous studies [7,29]. This is in agreement with previous studies demonstrating that a significant ordering effect of the cholesterol on phospholipid chains can be observed
128
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
Fig. 3. The influence of Naþ, Kþ, Ca2þ, and Mg2þ ions on the cholesterol-rich bilayer. (A) Angular distribution of the headgroup PeN vector with respect to the bilayer normal. (B) Molecular order parameter (SCD) profiles in PPC bilayers calculated for the Sn-1 chain and Sn-2 chain of POPE (top) and POPS molecules (bottom), respectively.
[23]. To characterize the lateral mobility of lipid molecules that could be influenced by bound ions, we focus on the lipid molecules' lateral diffusion coefficient DL The DL of lipids was evaluated as DL ¼ limt/∞ ð1=4tÞ½rðtÞ2 and calculated from the long-time mean square displacement ½rðtÞ2 of the centre of mass positions of lipids versus time in the bilayer plane. The DL values of the lipids in the PPCNaK system are DPOPE ¼ 2.5 ± 0.3 108 cm2/s, DPOPS ¼ 2.3 ± 0.03 108 cm2/s, DChol ¼ 2.3 ± 0.03 108 cm2/s, while in the PPCNaKCaMg system are DPOPE ¼ 0.50 ± 0.07 108 cm2/s, DPOPS ¼ 0.57 ± 0.03 108 cm2/s, DChol ¼ 0.53 ± 0.07 108 cm2/s. The DL for the cholesterol-rich bilayers is within the wide range of experimentally values [30]. Clearly, the additional divalent cations that greatly decrease the DL values hinder the lateral movements of lipid molecules in the PPC bilayer. The charge density distribution perpendicular to the bilayer surface is also an important property of a plasma membrane [10]. Consistent with the chemical structure of the POPE and POPS molecules (Fig. S1A), the distributions of lipid charges exhibit two distinct regions: a negative one (2.7 nm), provided by the phosphate groups and, and a positive one caused by the choline groups (~3.0 nm) (Fig. S3). In the PPCNaK system, the compensation of nonzero charge for the phosphate and choline groups in the lipids is mainly contributed by water molecules rather than the monovalent ions owing to their negligible binding (Fig. S3). By contrast, in the PPCNaKCaMg system, Ca2þ ions penetrate deep into the bilayer and the charge density profile of the cations demonstrates a single prominent maximum for the compensation of a net negative charge of lipid headgroup (Fig. S3). At the same time, because bound Ca2þ ions alter the orientation of lipid headgroups (Fig. 3A), they may induce the reorientation of water molecules and thereby affecting the charge density profiles of lipids and water molecules, as shown in Fig. S3. As a consequence of these effects, the distribution of the membrane surface charges in the PPCNaK system is similar to that found in the PPCNaKCaMg system (Fig. S3). 4. Discussion It is well known that the specific interactions of cations with biological membranes alter the properties of membranes and affect the way the proteins bind or are inserted into membranes. MD simulations have contributed to providing a detailed microscopic picture of the ionebilayer interactions. Previous MD simulation studies on ionemembrane interactions were performed based on the lipid models of either pure phospholipid or phospholipid mixtures. In the present study, we designed a cholesterol-rich
bilayer model (negatively charged PPC bilayer). Two simulation systems were obtained by either adding 80 Naþ, 80 Kþ or 80 Naþ, 80 Kþ, 16 Ca2þ, and 16 Mg2þ into the cholesterol-rich bilayers and used to investigate the interactions and effects of ion mixtures on the cholesterol-rich bilayer. From the numbers of bound ions (Table S1), we conclude that Ca2þ ions have the highest affinity to the water/ membrane interface of cholesterol-rich bilayer compared to other cations (Fig. 1), similar to the conclusion regarding the interaction between cations and pure phospholipid bilayers [13]. In particular, whereas all 16 Ca2þ ions bind to the anionic PPC bilayer, none of the Mg2þ ions was found to bind to theses cholesterol-rich bilayers, and both Naþ and Kþ ions are very weakly bound, indicating that the interactions and effects of Mg2þ, Naþ, and Kþ ions are negligible in this cholesterol-rich bilayer. Clearly, Ca2þ ions show the highest affinity for the cholesterol-rich bilayers compared to other cations, making Ca2þ effects dominant for the structure of the cholesterolrich bilayers due to Ca2þ ions. However, previous studies have supposed that the binding affinity of Mg2þ ion was slightly weaker than that of the Ca2þ ion in pure phospholipid bilayers and that the binding affinity of Naþ ion should be stronger than that of the Kþ ion [3,10,13]. In our simulations, none of the Mg2þ ions was found to bind to the cholesterolrich bilayer, and both Naþ and Kþ ions are very weakly bound, indicating that the interactions and effects of Mg2þ, Naþ, and Kþ ions are negligible in the cholesterol-rich bilayer. We should note that cholesterol molecules are randomly distributed in the bilayer region, as presented in Fig. S4. Owing to its rigid ring systems, cholesterol could separate the neighbouring phospholipid molecules, constrain the movement of phospholipid molecules in the process of ion-lipid interaction, and further broaden the distance of neighbouring lipid headgroups in the cholesterol-rich bilayers. The weak binding affinities of the monovalent Naþ and Kþ ions for the cholesterol-rich bilayer could be explained by the negative effect of cholesterol on cation-lipid interactions as discussed above. The smaller size of the divalent Mg2þ ions has been previously hypothesized to cause them to fit less effectively between the headgroups compared to Ca2þ ions [3,13]. Moreover, when a Mg2þ ion binds to the cholesterol-rich bilayer, the coordination of the nearby phosphate oxygen atoms could be inhibited by the presence of neighbouring cholesterol molecules. Additionally, the competition of Ca2þ with Mg2þ for binding to the lipids [13] should also account for this effect. It is known that ion binding to the phospholipid bilayer produces more ordered bilayers accompanied with thicker bilayers and smaller areas per lipid [5,6,8]. Because Ca2þ shows the highest binding affinity to the water/membrane interface in the cholesterol-rich bilayer, and the binding of Mg2þ, Naþ, and Kþ ions
L. Mao et al. / Biochemical and Biophysical Research Communications 468 (2015) 125e129
was found to be negligible in present study, we may expect that Ca2þ ions play a dominant role in altering bilayer properties. To reveal the influences of ion mixtures on the properties of cholesterol-rich bilayers, the structural and dynamic parameters of the two simulation systems were calculated and compared. However, we found that the additional divalent cations could only slightly alter the ordering and packing of cholesterol-rich bilayers. This result indicates that the bound cations may affect the structure of cholesterol-rich bilayer in a different way than those of the pure phospholipid bilayers. As reported, cholesterol could order the hydrocarbon chains of the neighbouring phospholipids and reduce the surface area of the lipid bilayer [23,24,31], indicating that the cholesterol-rich bilayers are already well packed and ordered. Consequently, less space is left for further condensation and the ordering of hydrophobic chains when the cations interact with this cholesterol-rich bilayer. We also discovered that Ca2þ binding induces the lipid headgroup to be more perpendicular to the membrane surface, and reduce the lipid mobility of the cholesterol-rich bilayer. The present MD simulations lead to the following conclusions. First, compared to other cations, Ca2þ ions have the highest binding affinity to the cholesterol-rich bilayers, consistent with the previous conclusion regarding pure phospholipid bilayers. Second, different from the results of our previous study regarding pure phospholipid bilayers, the interactions of Mg2þ, Naþ and Kþ ions are all relatively weak in the cholesterol-rich bilayer, indicating that cholesterol could negatively affect the binding affinities of cations to the cholesterol-rich bilayers. Third, the additional divalent cations induce phospholipid headgroups to be more perpendicular to the membrane surface and reduce the lateral movement of lipids but slightly alter the ordering and packing of the cholesterol-rich bilayer. This study gains a new and further insight on ionemembrane interactions and could be helpful in understanding the biological processes relevant to cations regulations in cholesterol-rich regions.
[4]
[5]
[6] [7]
[8]
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16] [17]
[18]
[19]
Conflicts of interest
[20]
The authors declare that there are no known conflicts of interest associated with this publication.
[21]
[22]
Acknowledgement [23]
This work was supported by the National Natural Science Foundation of China (grant numbers 21422208, 81173027, 81230076 and 21210003), and the SA-SIBS Scholarship Program. We thank the National Supercomputing Center in Tianjin (Tianhe 1A) and the National Supercomputing Center in Jinan for computational resources.
[25]
Appendix A. Supplementary data
[27]
Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.bbrc.2015.10.149.
[28]
References [1] M. Edidin, Lipids on the frontier: a century of cell-membrane bilayers, Nat. Rev. 4 (2003) 414e418. [2] S. Freeman, K. Quillin, L. Allison, Lipids, Membranes, and the First Cells, Biology and the Tree of Life, Benjamin Cummings, San Francisco, 2011, pp. 99e124. € rnig, The effect of metal cations on the phase behavior and [3] H. Binder, O. Zscho
[24]
[26]
[29] [30]
[31]
129
hydration characteristics of phospholipid membranes, Chem. Phys. Lipids 115 (2002) 39e61. S.A. Pandit, M.L. Berkowitz, Molecular dynamics simulation of dipalmitoylphosphatidylserine bilayer with Naþ counterions, Biophys. J. 82 (2002) 1818e1827. S.A. Pandit, D. Bostick, M.L. Berkowitz, molecular dynamics simulation of a dipalmitoylphosphatidylcholine bilayer with NaCl, Biophys. J. 84 (6) (2003) 3743e3750. €ckmann, A. Hac, T. Heimburg, H. Grubmüller, Effect of sodium chloride R.A. Bo on a lipid bilayer, Biophys. J. 85 (2003) 1647e1655. P. Mukhopadhyay, L. Monticelli, D.P. Tieleman, Molecular dynamics simulation of a palmitoyl-oleoyl phosphatidylserine bilayer with Naþ counterions and NaCl, Biophys. J. 86 (2004) 1601e1609. U.R. Pedersen, C. Leidy, P. Westh, G.H. Peters, The effect of calcium on the properties of charged phospholipid bilayers, Biochim. Biophys. Acta 1758 (2006) 573e582. H. Yang, Y. Xu, Z. Gao, Y. Mao, Y. Du, H. Jiang, Effects of Naþ ,Kþ, and Ca2þ on the structures of anionic lipid bilayers and biological implication, J. Phys. Chem. B 114 (2010) 16978e16988. A. Cordomi, O. Edholm, J.J. Perez, Effect of ions on a dipalmitoyl phosphatidylcholine bilayer. A molecular dynamics simulation study, J. Phys. Chem. B 112 (2008) 1397e1408. B. Klasczyk, V. Knecht, Validating affinities for ion-lipid association from simulation against experiment, J. Phys. Chem. A 115 (2011) 10587e10595. R.A. Bockmann, H. Grubmuller, Multistep binding of divalent cations to phospholipid bilayers: a molecular dynamics study, Angew. Chem. Int. Ed. Engl. 43 (2004) 1021e1024. Y. Mao, Y. Du, X. Cang, J. Wang, Z. Chen, H. Yang, H. Jiang, Binding competition to the POPG lipid bilayer of Ca2þ, Mg2þ, Naþ, and Kþ in different ion mixtures and biological implication, J. Phys. Chem. B 117 (2013) 850e858. M. Roux, M. Bloom, Ca2þ, Mg2þ,Liþ, Naþ, and Kþ Distributions in the headgroup region of binary membranes of phosphatidylcholine and phosphatidylserine as seen by deuterium, Biochemistry 29 (1990) 7077e7089. D. Huster, K. Arnold, K. Gawrisch, Strength of Ca2þ binding to retinal lipid membranes: consequences for lipid organization, Biophys. J. 78 (2000) 3011e3018. B.Y. Ha, Stabilization and destabilization of cell membranes by multivalent ions, Phys. Rev. E 64 (2001). S.A. Pandit, D. Bostick, M.L. Berkowitz, Mixed bilayer containing dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylserine: lipid complexation, ion binding, and electrostatics, Biophys. J. 85 (2003) 3120e3131. C.G. Sinn, M. Antonietti, R. Dimova, Binding of calcium to phosphatidylcholineephosphatidylserine membranes, Coll. Surf. Physicochem. Eng. Asp. 282e283 (2006) 410e419. I. Reviakine, A. Simon, A. Brisson, Effect of Ca2þ on the morphology of mixed DPPC-DOPS supported phospholipid bilayers, Langmuir 16 (2000) 1473e1477. P.T. Vernier, M.J. Ziegler, R. Dimova, Calcium binding and head group dipole angle in phosphatidylserine-phosphatidylcholine bilayers, Langmuir 25 (2009) 1020e1027. P. Jurkiewicz, L. Cwiklik, A. Vojtiskova, P. Jungwirth, M. Hof, Structure, dynamics, and hydration of POPC/POPS bilayers suspended in NaCl, KCl, and CsCl solutions, Biochim. Biophys. Acta 1818 (2012) 609e616. g, M. Pasenkiewicz-Gierula, I. Vattulainen, M. Karttunen, Ordering effects T. Ro of cholesterol and its analogues, Biochim. Biophys. Acta 1788 (2009) 97e121. H. Ohvo-Rekil€ a, B. Ramstedt, P. Leppim€ aki, J.P. Slotte, Cholesterol interactions with phospholipids in membranes, Prog. Lipid Res. 41 (2002) 66e97. g, M. Karttunen, I. Vattulainen, R. Reigada, Cholesterol H. Martinez-Seara, T. Ro induces specific spatial and orientational order in cholesterol/phospholipid membranes, PLoS One 5 (2010) e11162. g, A. Magarkar, V. Dhawan, P. Kallinteri, T. Viitala, M. Elmowafy, T. Ro A. Bunker, Cholesterol level affects surface charge of lipid membranes in saline solution, Sci. Rep. 4 (2014) 5005. W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, Comparison of simple potential functions for simulating liquid water, J. Chem. Phys. 79 (1983) 926e935. R. Guixa-Gonzalez, I. Rodriguez-Espigares, J.M. Ramirez-Anguita, P. CarrioGaspar, H. Martinez-Seara, T. Giorgino, J. Selent, MEMBPLUGIN: studying membrane complexity in VMD, Bioinformatics 30 (2014) 1478e1480. L.S. Vermeer, B.L. de Groot, V. Reat, A. Milon, J. Czaplicki, Acyl chain order parameter profiles in phospholipid bilayers: computation from molecular dynamics simulations and comparison with 2H NMR experiments, Eur. Biophys. J. 36 (2007) 919e931. C. Hong, D.P. Tieleman, Y. Wang, Microsecond molecular dynamics simulations of lipid mixing, Langmuir 30 (2014) 11993e12001. €dd, G. Lindblom, The effect of cholesterol on the lateral A. Filippov, G. Ora diffusion of phospholipids in oriented bilayers, Biophys. J. 84 (2003) 3079e3086. C. Hofs€ aß, E. Lindahl, O. Edholm, Molecular Dynamics Simulations of Phospholipid Bilayers with Cholesterol, Biophys. J. 84 (2003) 2192e2206.