Headgroup Mediated Water Insertion into the DPPC Bilayer: A Molecular Dynamics Study

Published: March 08, 2011

r 2011 American Chemical Society 3155 dx.doi.org/10.1021/jp1090203 | J. Phys. Chem. B 2011, 115, 3155–3163

ARTICLE

pubs.acs.org/JPCB

Headgroup Mediated Water Insertion into the DPPC Bilayer:A Molecular Dynamics StudyPrithvi Raj Pandey and Sudip Roy*

Physical Chemistry Division, National Chemical Laboratory, Pune-411008, India

bS Supporting Information

1. INTRODUCTION

S. J. Singer and G. L. Nicolson proposed in their famous fluidmosaic model1 of the biological membranes that it containsoriented globular proteins and lipids. Water is present near bothsides of the bilayer in the extracellular and intracellular fluid.Lipid molecules contain polar headgroups and nonpolar hydro-carbon tails. They assemble to form a lipid bilayer. Being polar,the headgroups are hydrophilic in nature and point toward theextracellular fluid and the cytoplasm. The tails of the two layersface each other and form a hydrophobic region.

Phospholipids are the major kind of lipid molecules present inthe biological systems. Various kinds of phospholipids present inbiological systems are phosphatidylserine, phosphatidylethano-lamine, and phosphatidylcholine. They differ in the structure oftheir headgroups. 1,2-Dipalmitoyl-sn-phosphocholine (DPPC,Figure 1) contains a choline group at its headgroup. Sincephospholipids are the building blocks of biological membranes,their structure and dynamics have been studied extensively.2-7 Inthese experimental studies, X-ray crystallography and NMR havelargely been used and proved to be helpful techniques forelucidating the structure of various kinds of crystalline andamorphous phospholipids.

This has been a subject of interest for almost two decades, butinteresting questions still remain unsolved. How the headgroupof the lipid molecules in the bilayer interact with water is oneamong them? Or, in other words, how at the lipid-water

interface the hydrophilic headgroup of lipid molecules interactwith adjacent water molecules? In this regard, NMR8-12 andneutron scattering13-15 and diffraction16 experiments haveshown that water molecules permeate the bilayer interface witha steeply decreasing concentration as they proceed toward the innerhydrophobic region of the bilayer. Studies with other experimentaltechniques (e.g., FTIR) have also provided similar insight.44

For a quantitative understanding of the interactions of watermolecules with polar headgroups at the lipid-water interface atthe molecular level, molecular dynamics (MD) simulations haveproved to be useful. MD simulations have also enforced the facts,just like experimental studies, that the majority of the watermolecules residing in the polar region of the membrane arehydrogen (H-) bonded to the phosphate groups, with a smallerfraction binding to the carbonyl groups located deeper in thebilayer but with more extensive structural details.17-19,43,45 Oneof the initial reports that presented the study of the lipid-waterinterface on the dilauroylphosphatidylethanolamine lipid bilayerwith MD simulations was by Damodaran et al.20 They calculatedthe velocity autocorrelation function for both lipid and water andalso the orientational correlation function for water to under-stand the dynamics of the system and diffusive properties of the

Received: September 21, 2010Revised: February 13, 2011

ABSTRACT: Molecular dynamics simulation was performedon the 1,2-dipalmitoyl-sn-phosphocholine (DPPC) bilayer-water system using the GROMOS96 53a6 united atom forcefield. The transferability of force field was tested by reproducingthe area per lipid within 3% accuracy from the experimentalvalue. The simulation shows that water can penetrate muchdeeper inside the bilayer almost up to the starting point of thealiphatic chain. There is significant evidence from experimentsthat water goes deep in the DPPC bilayer, but it has not beenreported from theoretical work. The mechanism of insertion ofwater deep inside the lipid bilayer is still not clear. In this report,for the first time, themechanism of water insertion deep into thebilayer has been proposed. Water transport occurs by the headgroup and its first solvation shell. The trimethyl ammonium (NMe3)group (headgroup of DPPC) has two stable conformations at the bilayer-water interface, one outside the bilayer and another insideit. The NMe3 group has a large clustering of water around it and takes the water molecules inside the bilayer with it during its entryinto the bilayer. The water molecules penetrate into the bilayer with the help of the NMe3 group present at the headgroup of DPPCand eventually form hydrogen bonds with carbonyl oxygen present deep inside the bilayer. Structural characteristics at the bilayer-water interface region are also reported.

3156 dx.doi.org/10.1021/jp1090203 |J. Phys. Chem. B 2011, 115, 3155–3163

The Journal of Physical Chemistry B ARTICLE

system, respectively. MD simulations have also shown thatorganization of water molecules at the lipid-water interfacedepends upon the type of lipid headgroup.21

The most critical component of any MD simulation is theforce field used, which consists of a set of mathematical functionsand parameters that describes the bonded and nonbondedinteractions between the particles (atoms, united atoms) of thesystem under study. Recently, Kukol in his study22 has proposeda model for DPPC, which reproduces the area per lipid of thelipid bilayer within 3% error without the assumption of a constantsurface area or the inclusion of surface pressure as provided byexperimental study.26,27 In Kukol’s model, modification of theoriginal model included in the GROMOS96 53a628 force fieldhas been done with two changes in the topology. First, the partialcharges on the lipid headgroup due to Chiu et al.33 wereimplemented, with a subdivision into four charge groups assuggested by Chandrasekhar et al.41 Second, ester-carbonylcarbon atom type was changed to “CHO” from “C”, resultingin an increase in the van der Waals radius for ester-carbonylcarbon atom to 0.664 nm as opposed to 0.336 nm before. Thesechanges resulted in better values for area per lipid and alsoincreased the penetration of water into the lipid bilayer head-group region. Gierula et al. have studied the H-bonding dynamicsbetween water and dimyristoylphosphatidylcholine (DMPC) usingMD simulation.46 In their study, they have observed that, among allthe DMPC oxygens, the largest ordering of water is around non-ester phosphate oxygen but to a much lesser extent around ester-carbonyl oxygens. The dynamics of water molecules at the fullyhydrated 1-palmitoyl-2-oleoyl-phosphatidylcholine (POPC), 1-pal-mitoyl-2-oleoyl-phosphatidyl ethanolamine (POPE), and 1-palmi-toyl-2-oleoyl-phosphatidylglycerol (POPG) bilayers and the effectof headgroups on this motion have been studied by Murzyn et al.31

They have shown that the headgroup plays an important role in thedynamics of water molecules at the interface. Inter- and intralipidinteractions were studied by Zhao et al. for anionic palmitoylo-leoylphosphatidylglycerol (POPG) lipid.32

A recent experimental study by Sovago et al.42 showed that, forlipid monolayers of lauric acid (LA), octadecyl trimethyl ammo-nium bromide (OTAB), 1,2-myristoyl-sn-glycerol-3-phosphoserine(DMPS), DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphoethanola-mine (DPPE), and 1,2-dipamitoyl-3-trimethyl ammonium-propane(DPTAP), buried water molecules exist between the lipid head-groups and their alkyl tail. Also, they have shown that these watermolecules have a weaker H-bond network than that of bulk water.As water penetrated much deeper in the bilayer has been proved byexperiment and theory,22 in this work, we have studied for the firsttime to the best of our knowledge the mechanism of furtherinsertion of water molecules, which cross through the headgroupregion of the DPPC bilayer and remain in between the head and tailof lipid molecules. We have also studied the change in hydrogenbonding environment, using the modified topology proposed byKukol, around different electronegative groups of DPPCmolecules.This has been studied by clustering of water molecules and also the

H-bonding of the water molecules attached to non-ester carbonyloxygens, phosphoryl oxygens, andNMe3 of the lipid molecules. Wehave shown that the insertion of these water molecules inside thebilayer occurs with the help of the NMe3 group present at theheadgroup of DPPC. All the structural details obtained points to themechanism as proposed in this work.

2. COMPUTATIONAL METHODS

The initial structure of the DPPC bilayer and water systemwastaken from the last frame of the 40 ns run from Kukol’s work.Each DPPC molecule was taken as the united atom model with50 atoms as discussed by Kukol. However, the all atomistic modelof DPPC contains a total of 130 atoms. The choice of the unitedatom model was made, as it is the best known model at this stagefor DPPC and is even better than any other atomistic models, andcertainly provides faster and longer time scale simulation, andhence better statistics. The initial coordinates consisted of aDPPC bilayer having 128 DPPC molecules hydrated with 3655water molecules (Figure 2). The previous model of DPPC basedonGROMOS-53a6 has failed to reproduce the experimental datawith high accuracy, which has already been shown in previousstudies.55 The force field by Berger et al.23-25 has also been usedfor DPPC but without much success in the case of membraneprotein. In this work, the DPPC bilayer was simulated for 40 nsusing the GROMACS 4.0.729,30 package and the GROMOS-53a628 force field, with the slightly modified topology proposedby Kukol,22 which happens to be the best parameters available forthe united atomDPPC system and tested for proteins and bilayerassemblies. The SPC water model was considered for thesimulation. The NPT ensemble and periodic boundary condi-tions were used. Temperature was kept constant at 325 K using av-rescale thermostat for the whole simulation time. At thistemperature, DPPC exists in liquid crystalline state, as its phasetransition temperature is 314.5 K.34,35 Semi-isotropic pressurecoupling was applied using a Berendsen barostat36 with separatecoupling to the xy-plane and the z-direction (the bilayer normal).The pressure coupling time constant was 2.0 ps in order tomaintain a constant pressure of 1.0 bar. For Lennard-Jonesinteraction, a cutoff at 1.4 nm was applied; electrostatic interac-tions were taken care of with the particle mesh Ewald (PME)37

method and a real space cutoff of 0.9 nm. Electrostatic interac-tions were treated with the PME method, which did notintroduce artificial ordering like reaction field cutoff methods.

Figure 1. Stucture of the DPPCmolecule with atom numbering as usedin the text.

Figure 2. DPPC-water interface.



The lipid molecule bonds were constrained with the LINCSalgorithm. All of these parameters were in agreement withKukol’s work. The MD simulation for 40 ns was carried out,and a trajectory was written every 0.5 ps. The trajectory wasfurther analyzed to calculate radial distribution functions be-tween different atoms of DPPC and water. Partial density profilesare reported as a two-dimensional plot where we have consideredthe interface along the z-axis. Dihedral distributions for a fewspecific dihedrals are reported to illustrate the mechanism ofwater insertion into the bilayer. All other nonconvetional analysismethods are described along with the results and discussions.

3. RESULTS AND DISCUSSION

3.1. Area per Lipid. Area per lipid is the most widely usedproperty for characterization of lipid bilayers, as it can also bemeasured by experiments. Also, area per lipid is related to variousother properties like lateral diffusion, membrane elasticity, etc. Inthe present work, the lipid bilayer was along the Z-axis. Thus, thearea per lipid was calculated from the lengths of the box in the Xand Y direction by the following equation

area per lipid ¼ ðbox length in XÞðbox length in YÞno: of lipids in one lamella of bilayer ði:e:; 64Þ

ð1ÞThe plot of area per lipid as a function of time for a total of 40 ns isshown in Figure 3. The trend of the area per lipid was similar tothat reported in Kukol’s work. The average area per lipid forthe 40 ns trajectory was calculated to be 0.64063 ( 0.02 nm2.This value is within a range of 3% of the experimental value of0.64 nm2. This provides the validation of the force field andalso shows the reproducibility of our simulation. We have alsoperformed a self-assembly simulation starting from randomlyoriented DPPC in water, using the same force field, and wehave obtained the same result for area per lipid after self-assembly.3.2. Partial Density Profile. We have plotted the partial

density profile for DPPC, water, and various atoms present atthe head and neck region of DPPC versus the length of

simulation box in the Z-direction in Figure 4. Interfaces betweenDPPC and water are clearly visible in the plot, one from 0.5 to2.5 nm and the other from 4 to 6 nm approximately. A small dipin the plot of DPPC (at around 3.5 nm) shows the intersectionbetween the hydrophobic regions of the two monolayers. Thepeak of nitrogen (N) appears nearest to the water molecules atthe interface. This shows that N or the NMe3 group faces most ofthe bulk of water, which is quite obvious. For the phosphorylgroup (PdO), both of the non-ester oxygens of PdO weretreated equivalently. The partial density of PdO is highestamong all the atoms in the head and neck region of DPPC. Thisis possibly because we have treated both of the phosphoryloxygens as equivalent atoms. The difference between the posi-tions of the peaks of N and PdO is negligible. However, with Nbeing the outmost atom, we expect the peak of N in the partialdensity plot to be present more toward the bulk water at thebilayer-water interface. This issue has been clarified and

Figure 3. Area per lipid.

Figure 4. Partial density profile.



discussed later. The position of the peak of O16 (atom numberused as in Figure 1) appears more toward the inside of the bilayerthan N and PdO. The position of the peak of O35 appears moretoward the tail region of the bilayer than the others. O35 insteadof being present at a difference of one carbon atom (C32) thanO16 from the headgroup, the partial density plot shows that O35is still present at the water-DPPC interface region. This showsthat water penetrates deep into the DPPC bilayer.3.3. Environment of the Headgroup (NMe3) and Its Neigh-

boring Group (PdO). 3.3.1. Clustering of Water Molecules.Figure 5 shows the radial distribution functions (RDFs) ofoxygen atoms of water molecules with carbon atoms (C15,C34) of carbonyl groups, phosphorus (P), and nitrogen (N)atoms present in DPPC. Thus, the RDFs represent the clusteringof water molecules around these atoms in DPPC. Integration ofthe first RDF peak shows that clustering of water molecules islargest around N (as integration of its first peak has a value of13.64), which is expected as N is located at the external part of theheadgroup of DPPC. As a result, it sees most of the bulk of water.The heights of the first RDF peaks for P and C15 with oxygen

of water molecules appear at similar values, whereas the integra-tion of the first peaks, which is themeasure of the average numberof water molecules present in the first solvation shell, is 3.27 for Pand 1.87 for C15. This shows that a considerable number ofwater molecules penetrated into the bilayer and there is a rapiddecrease in water concentration from NMe3 to C15. This is alsovisible from the partial density profile, where we have observedthat O35, despite being situated at the beginning of the tail ofDPPC, is present in the interfacial region. The peak of C34 issmallest of all and solvated by only an average of 1.02 watermolecules. This is because C34 is located deeper inside thebilayer where the hydrophobic region starts to form and is also ata difference of one carbon atom (C32) with respect to C15.Thus, clustering of water molecules around various atoms in theheadgroup and neck region is in the following order as inter-preted from the values of the number of water molecules presentin the first solvation shell.

N > P >C15 >C34

The size of the solvation shells around these atoms (asinterpreted from the first minima of the peaks) are in the

following order

N > P � C15 > C34

3.3.2. Water Environment around O16 (Carbonyl Oxygen),PdO (non-ester Oxygen of Phosphorus), and O35 (CarbonylOxygen). Gierula et al. in their study suggested that the largestordering of water in the DMPC bilayer is around phosphorylnon-ester oxygens (PdO). They observed less ordering aroundnon-ester carbonyl oxygens. Also, Lopez et al.19 have obtainedsimilar results with DMPC from a longer trajectory. We haveplotted RDFs between hydrogen atoms of water with non-esteroxygens of PdO (O9 and O10), O35, and O16 (Figure 6a). Theheight of the first peak in the RDF plot is maximum for O16,which suggests ordering of a greater number of water moleculesaroundO16 thanO9 andO10 (non-ester oxygens of phosphorusatom). This is in direct contradiction to the results obtained byGierula et al. The first peak of O9 and O10 of phosphorus issmaller than O16, and hence suggests a lesser ordering of watermolecules around PdO than O16. As expected, the number of

Figure 5. RDFs for clustering of oxygens (water) with headgroupnitrogen, phosphorus, and carbonyl carbons.

Figure 6. (a) RDFs between carbonyl oxygens and non-ester oxygensof phosphorus with hydrogen of water molecules. (b) RDFs betweennitrogen-O16, N-O (water), O16-H (water), and carbonyl C-(C15)-O (water).



water molecules present in the first solvation shell of O35 is theleast among others present at the interface. It appears to beobvious as O35 is located near the starting of the tail and is at adifference of one carbon atom (C32) with respect to O16 fromthe headgroup region. Also, the peak integrations indicate thesame fact. Integration of first RDF peaks for O9þO10, O16, andO35 are 1.04, 1.49, and 0.61, respectively.Although we have observed and has also been proposed by

Kukol22 that water penetrates considerably into the bilayer-water interface region, the higher number of water moleculesnear O16 than O9 þ O10 (non-ester oxygen of phosphorus) isquite interesting. A question arises of howmuch water moleculesbypass the non-ester oxygens of phosphorus and take part in thesolvation shell of O16 which is located deeper inside the bilayerthan the non-ester oxygens of phosphorus? Since PdO seesmore of the interfacial water as compared to O16, water isexpected to be more around PdO as compared to O16, asobserved by Gierula et al. for DMPC. Thus, the question is whydo we get a higher first peak for O16 than PdO in the RDF plot?To address this issue and find the mechanism of water

insertion deep inside the bilayer, we have compared the RDFplots between N and O16 (combined intermolecular andintramolecular (inter þ intra) RDF and only intermolecular(inter) RDF) with RDFs between N and water O (N-Owater),O16 and water H (O16-Hwater), and C15 and water O (C15-Owater) in Figure 6b.We can readily observe that the minimum ofthe first RDF peak of N and water O appears at a similar distanceas that of inter þ intra and only inter RDF plot between N andO16. It shows that the solvation shell of O16 merges with theclustering shell of the NMe3 group present in the same moleculeor on a neighboring molecule. Similarly, the minimum of the firstpeak between C15 (carbonyl carbon) and water O appearswithin the first peak of N and O16. This shows that the size ofthe clustering shell of C15 is within the distance of closeness of Nand O16 irrespective of whether they belong to the same DPPCmolecule (intramolecular) or to neighboring DPPC molecules(intermolecular). Thus, it appears that the solvation shell of O16,which is attached to C15, merges with the first solvation shell ofthe NMe3 group. The peaks with respect to N appear at a higherdistance than the O16 peak because N is surrounded by three-CH3 (Me3) groups, and hence water molecules can arrangethemselves just outside the-CH3 groups. Similar is the case withC15, where C is attached with O16 so water molecules arrangearound O16, whereas O16 is naked and directly sees waterwithout any intervening group. At a first intuition, it might lookunusual that water molecules arrange around hydrophobic -CH3 groups, but actually it is rather common. Nitrogen beingpositively charged (in DPPC) attracts the polar water moleculestoward itself. The Coulombic attraction between N and watermolecules outweighs the hydrophobic mismatch between-CH3

and water. Lum et al.39 and Chandler et al.40 have calculated thehydrophobic interactions between solute and water. Also, God-awat et al. have shown the density fluctuations at the interface ofwater and a range of hydrophilic and hydrophobic solutes.38

Integration of the peaks also shows that N has a larger firstsolvation shell than O16.Revisiting our question of why the first RDF peak is higher for

O16 than PdO in Figure 6a, it can be suggested that, since thefirst solvation shell of C15 (and hence O16) merges with the firstsolvation shell of NMe3 (Figure 6b), also we have seen thatclustering of water around N is the highest (Figure 5), it ispossible that O16 shares the solvation environment of the NMe3

group in certain intervals. This possibility can only be true if theNMe3 group periodically bends toward O16, and as a result ofwhich O16 shares the solvating water molecules of NMe3.Dihedral bending of NMe3 is represented by the dihedral angledistribution plot of N-C5-C6-O7 sampled from the 40 nstrajectory in Figure 7a, where C5 and C6 are the connectingmethyl groups between headgroup N and ester oxygen of thephosphoryl group. Two peaks in the plot suggest two conforma-tions of NMe3, one of which corresponding to toward O16 andthe other away from it. Krishnamurty et al.47 have shown usingdensity functional theory (DFT) calculations that, for DMPC inthe lowest energy conformation, the headgroup (choline group)is bent toward carbonyl in the neck region. Also, it has beenproposed previously49-53 that the bent headgroup configurationshould minimize the intramolecular electrostatic interaction inthe polar region of phospholipids. In other words, the internalelectrostatic energy resulting from repulsion between the phos-phoryl group and carbonyl oxygen is minimized by attractionbetween choline and carbonyl groups. In Figure 6a, we have alsoobserved that the first RDF peak of PdO (non-ester oxygen ofphosphorus) with water H is in the region of the first RDF peakof O16 with water H. Hence, from this, we can also say that O16(carbonyl oxygen near the phosphorus) shares the solvation shellwith PdO. Leekumjorn et al. have shown by varying thepercentage of DPPE in DPPC that the amine group in DPPEstrongly interacts with phosphoryl and carbonyl groups throughintra- or intermolecular H-bonding.54 Also, a surface sum-frequency generation spectroscopy study of buried water mole-cules inside phospholipid membranes by Sovago et al.42 sug-gested that water molecules interact with the zwitterionic lipidDPPC with its O-H group pointing toward bulk water and ispositioned just below the choline group. This leads to the factthat the interfacial water may have a variety of hydrogen bondingenvironments. This fact is further confirmed by IR spectroscopicstudy of various kinds of lipids present in the biological system byHubner et al.48

Further, in RDF plots between N and O16 (inter þ intra andinter) sampled for a 40 ns trajectory (Figure 6b), we haveobserved two peaks: one large peak at around 0.4 nm andanother peak at around 0.85 nm. The origin of the predominanttwo peaks is due to the sampling of dihedrals (N-C5-C6-O7)at different conformations. There is a second dihedral C5-C6-O7-P which may also have some effect on the bending of theNMe3 group toward the carbonyl oxygen (O16). The dihedralangle distributions for the above dihedral angles are plotted in thesame figure (Figure 7a). We observe that the C5-C6-O7-Pdihedral is sampling much more angular space than N-C5-C6-O7. To understand the time span of each conformationproduced by the combination of these two dihedrals (N-C5-C6-O7 and C5-C6-O7-P), we have plotted the dihedralangles for a single DPPC molecule as a function of time inFigure 7b. In the same figure, we have also depicted the distancebetween N and O16 (carbonyl oxygen). From the figure, it isobserved that indeed the dihedral related to the headgroupsamples only two major angles and it is spending a maximumof 0.5 ns in each conformation, whereas the other dihedral (C5-C6-O7-P) samples angles from-180 to 180�withmuchmorerapid motion than the other. Therefore, the distance between Nand O16 changes mainly due to the headgroup dihedral motionassisted by the much more frequent dihedral angle C5-C6-O7-P motion. Taking these observations for one molecule intoaccount, distributions of the C5-C6-O7-P dihedral were



plotted (Figure 7c) to understand the motion of the C5-C6-O7-P dihedral with respect to the N-C5-C6-O7 dihedralunder two conditions, first when the N-C5-C6-O7 dihedrallies between-180 and 0� (i.e., negative) and second when it liesbetween 0 and 180� (i.e., positive). These dihedral distributionsshow that the C5-C6-O7-P dihedral samples dihedral anglesfrom -180 to 180� in both positive and negative regions of theN-C5-C6-O7 dihedral. It is also interesting to note that theC5-C6-O7-P dihedral distribution has a small peak around90� when the N-C5-C6-O7 dihedral lies in the positiveregion and at -90� when the N-C5-C6-O7 dihedral lies inthe negative region. This shows the selectivity of a conformation.In Figure 7b, we have observed that the N-C5-C6-O7dihedral stays in the positive or negative region maximum for0.5 ns for one molecule. Whereas in this time the C5-C6-O7-P dihedral fluctuates very frequently and aquires a larger space.To make this fact statistically more acceptable, we have plottedthe distribution of residence time of both the dihedrals (N-C5-C6-O7 andC5-C6-O7-P) in the positive and negative spacein Figure 7d. We recorded the time for which a dihedral stays inone direction for all DPPC molecules and counted how manydihedrals stay in a particular direction for a particular span oftime. Finally, we have converted it into a distribution which isnormalized by the number of frames and number of DPPCmolecules. It is clearly visible from the plot that the N-C5-C6-O7 dihedral distribution decays muchmore rapidly than the

C5-C6-O7-P dihedral. That is, the N-C5-C6-O7 dihe-dral stays in either the positive or negative direction for muchlonger time than the C5-C6-O7-P dihedral. As a result ofthese observations, we predict that the distances obtained fordifferent dihedrals and dihedral angle conformations correspondto the closest distance of approach of N toward O16 of the samemolecule. This means that the second peak of the RDF plotbetween N and O16 in Figure 6b has a contribution from bothintramolecular and intermolecular distances between them. Thatis, at the headgroup region of the DPPC bilayer, N and O16 ofsamemolecule and also of the neighboring molecules come closeto each other. As a result, O16 not only sees the solvation sphereof N of the same DPPCmolecule but also the solvation sphere ofN of neighboring DPPC molecules. This fact also contributes tothe higher RDF peak of O16 with water H than PdOwith waterH in Figure 6a. This results in a larger solvation shell for O16.A two-dimensional densitymap of DPPC andwater (Figure 8)

shows that the distribution of water is continuous at the interfaceregion from the 40 ns simulation. This continuous environmentof water indicates the mobile nature of the water molecules evendeeper in the interface. Bending of the headgroup toward insidethe bilayer is further proved by plotting the angle between theO7-N bond vector and the z-axis with time (Figure 9). Whenthe angle is 90�, i.e., the vector is perpendicular to the z-axis, it isparallel to the interface. Also, an angle greater than 90�means thevector and hence the NMe3 group is inside the bilayer. In

Figure 7. (a) Dihedral angle distribution N-C5-C6-O7 and C5-C6-O7-P. (b) Dihedral angles (N-C5-C6-O7 and C5-C6-O7-P) anddistance between N and O16 as a function of time. (c) Conditional probability distribution of dihedral angle C5-C6-O7-P when N-C5-C6-O7dihedral angle samples positive and negative angles. (d) Distribution of dihedral residence time for N-C5-C6-O7 and C5-C6-O7-P for bothpositive or negative directions.



Figure 9, we observe that the concerned angle is greater than 90�(shown by a straight line) for a considerable amount of time. Thisis also a very firm proof of bending of the NMe3 group inside thebilayer.In Figure 10, a distribution of the common water molecules

between the solvation shell of NMe3 and O16 has been depicted.The distribution is made using a small part of the trajectory (25-30 ns) and was normalized with respect to number of frames inthat time. From each frame number of common water moleculesbetween the two solvation shells were counted and divided by thetotal number of DPPC, i.e., 128 in the system. Then, thedistribution of this number of common water molecules perDPPC was plotted. The plot shows one broad peak at 0.5, whichproves the presence of 0.5 commonwater molecules between thesolvation shell of NMe3 and O16 molecules. The plot acts as a

proof of our postulate for the mechanism of penetration of watermolecules into the bilayer from bulk.In addition to the water insertion deep into the bilayer, we

have also observed the bridging H-bonded water moleculesbetween O16 and non-ester oxygen attached to the phosphorusatom. These water molecules simultaneously form H-bonds toboth O16 and PdO. The average number of such H-bonds ineach frame is 15.6. Thus, out of 128 DPPC molecules present inthe system, around 16 DPPC molecules are involved in suchH-bonding. To calculate this number, we counted the number ofwater molecules appearing within H-bonding distance (0.28 nmobserved from Figure 6a) of O16 and PdO for each DPPCmolecule in each frame. From this, we counted the number ofwater molecules common to both O16 and PdO and divided itby the number of frames. The number appears to be low withrespect to the total number of DPPC molecules present in thesystem.It has been mentioned in the text that various experimental

studies have shown that water has a different H-bondingenvironment around the headgroup of the lipid bilayer. Meansquare displacement (MSD) of different water molecules, pre-sent initially in different regions, as a function of time has beenchecked. For this plot, we first recorded the water moleculepresent near the carbonyl group (CdO), PdO, and in bulkseparately. Then, we tracked these water molecules in the Z-direction for 500 ps of the trajectory and plotted MSD in the Z-direction, XY-plane, and XYZ-directions separately, for thosewho stayed in their initial region with respect to the Z-axis (i.e.,the axis along the interface) for more that 200 ps. Finally, we havereported ameanMSD plot out of these variousMSDs plotted forthe movement of water in the XY-plane and all XYZ-directions inthe Supporting Infomation. In the two plots, namely, the averageMSD of water in the XY-plane (SI1) and in all XYZ-directions(SI2), a clear distinction between the three regions is visible. Thisindicates different environments of water and its confinementnear the headgroup region and also after entering the bilayer.Dynamics and entropy of such water molecules have beenstudied recently by Debnath et al.56 which supports our finding.

4. CONCLUSION

Molecular dynamics simulation for 40 ns was performed usingGROMOS96 53a6 force field parameters but with modified

Figure 9. Angle between the O7-N vector and the z-axis of the boxwith time (90� angle is shown by a horizontal line).

Figure 10. Distribution of the number of common water moleculesbetween the solvation shells of N and O16.

Figure 8. Two-dimensional density map of the DPPC-water interfacein the XZ-plane.



topology, as suggested by Kukol22 on the DPPC-water bilayersystem. The simulations provided the area per lipid value forDPPC in 3% error range from experimental values. A bettermatch to the experimental area per lipid leads to the penetrationof water molecules deep into the bilayer, giving us the scope tounveil themechanism of water insertion into the bilayer. We haveobserved from our study that the NMe3 group at the headgroupof DPPC moves inside and outside of the bilayer. Density plotsshow that water is distributed in a continuous fashion at thebilayer-water interface and water penetrates deep into thebilayer up to the start of the hydrophobic tail of DPPCmolecules.Water keeps moving from the bilayer-water interface region tobulk water and vice versa. Due to this, it might be possible that,during the process of insertion of the headgroup, water moleculesclustered around the NMe3 group enter into the deep interfacialregion. Taking this as a postulate, we propose that, as a result ofbending, the NMe3 group acts as a water carrier and pumps inand out water molecules in the interfacial region of the bilayer.The mechanism is shown in the cartoon (Figure 11) which isdrawn from a snapshot taken from the simulation. The cartoondepicts the mechanism of how the headgroup bends inside thebilayer interface and overlaps with the solvation shell of carbonyloxygen (O16). The NMe3 group bends inside the bilayer withwater molecules solvating it. At the inside conformation of theNMe3 group, the solvation shells of NMe3 and O16 merge.There is a possibility of exchange of water molecules inside thebilayer-water interface with the solvation shell of headgroupatoms; i.e., water molecules clustering around the NMe3 groupget hydrogen bonded to O16 or O9 and O10. This exchange ispossible because, around the NMe3 group, water molecules areclustered outside the methyl groups, and hence in addition to theattraction with N due to its positive charge, there exists ahydrophobic mismatch with the methyl groups. However, whenthe water molecules approach near O16 or O9 or O10, they seeonly bare oxygen and are involved in H-bonding with it. Thisexchange is not necessarily between the solvation shells of NMe3and O16 of the same molecule, but exchange between neighbor-ing molecules is also possible. Complex motion of proteins orlipids is strongly affected by the amount of solvent interactingwith them. Our mechanismmay help to verify howmuch water isinteracting during its entry into the bilayer mediated by head-group. Hydration at the protein-lipid interface is required forenzymatic activity; therefore, the effect of water on the enzymaticactivity might be dependent on the amount of water present atthe interface, which can be verified with our mechanism. Inaddition to the mechanism proposed, observation of bridging

H-bonding water molecules points to the various H-bondingenvironments in the headgroup region of DPPC. This study alsoopens a scope to investigate the dynamics of such bridgingH-bonds using the same force field as used in this work. Sharppeaks in the RDF plots of O16 with water H and also PdO withwater H point to the fact that some ordered structure of watermight be present in the continuous distribution of water at theinterface. Further studies in this direction are required to studythe possibility and nature of the ordering of the water moleculespresent at the interface.

’ASSOCIATED CONTENT

bS Supporting Information. Average mean square displa-cement (MSD) plots for water molecules staying near CdO,PdO, and in bulk for more than 200 ps. Also table containingvalues of diffusion constants. This material is available free ofcharge via the Internet at http://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected]. Phone: þ91 (020) 25902735. Fax:þ91 (020) 25902636.

’ACKNOWLEDGMENT

We gratefully acknowledge CSIR and NCL for financialsupport for the project.

’REFERENCES

(1) Singer, S. J.; Nicolson, G. L. Science 1972, 175, 720–731.(2) Hauser, H.; Pascher, I.; Pearson, R. H.; Sundell, S. Biochim.

Biophys. Acta 1981, 650, 21–51.(3) Hauser, H.; Guyer, W.; Pascher, I.; Skrabal, P.; Sundell, S.

Biochemistry 1980, 19, 366–373.(4) Hauser, H.; Pascher, I.; Sundell, S. Biochemistry 1988,

27, 9166–9174.(5) Hong, M.; Zimmermann Schmidt-Rohr, H. Biochemistry 1996,

35, 8335–8341.(6) Bruzik, K.; Harwood, S. J. Am. Chem. Soc. 1997, 119, 6629–6637.(7) Aussenac, F.; Laguerre, M.; Schmitter, J. M.; Dufourc, E. J.

Langmuir 2003, 19, 10468–10479.(8) Yeagle, P. L.; Martin, R. B. Biochem. Biophys. Res. Commun. 1976,

69, 775.(9) Schmidt, C. F.; Barenholz, Y.; Huang, C.; Thompson, T. E.

Biochemistry 1977, 16, 3948–3954.(10) Pope, J. M.; Comell, B. A. Chem. Phys. Lipids 1979, 24, 27.(11) Smaby, J. M.; Hermetter, A.; Schmid, P. C.; Paltauf, F.; Brock-

man, H. L. Biochemistry 1983, 22, 5808–5813.(12) Zhou, Z.; Sayer, B. G.; Hughes, D.W.; Stark, R. E.; Epand, R.M.

Biophys. J. 1999, 76, 387.(13) Fitter, J.; Lechner, R. E.; Dencher, N. A. J. Phys. Chem. B 1999,

103, 8036–8050.(14) Sotomayor, C. P.; Aguilar, L. F.; Cuevas, F. J.; Helms, M. K.;

Jameson, D. M. Biochemistry 2000, 39, 10928–10935.(15) Poolman, B.; Spitzer, J. J.; Wood, J. M. Biochim. Biophys. Acta

2004, 88, 1666.(16) Jacobs, R. E.; White, S. H. Biochemistry 1989, 28, 3421–3437.(17) Takaoka, Y.; M. Pasenkiewicz-Gierula, M.; Miyagawa, H.;

Kitamura, K.; Y. Tamura, Y.; Kusumi, A. Biophys. J. 2000, 79, 3118.(18) Pasenkiewicz-Gierula; Takaoka, M. Y.; Miyagawa, H.; Kita-

mura, K.; Kusumi, A. Biophys. J. 1999, 76, 1228.(19) Lopez, C. F.; Nielsen, S. O.; Klein, M. L.; Moore, P. B. J. Phys.

Chem. B 2004, 108, 6603–6610.

Figure 11. Cartoon depicting the overlap of the solvation shells ofNMe3 and O16. Blue and red spheres represent the solvation shells of Nand O16, respectively.



(20) Damodaran, K. V.; Kenneth, M. Merz, Jr.; Gaber, B. P. Bio-chemistry 1992, 31, 7656–7664.(21) Polyansky, A. A.; Volynsky, P. E.; Nolde, D. E.; Arseniev, A. S.;

Efremov, R. G. J. Phys. Chem. B 2005, 109, 15052–15059.(22) Kukol, A. J. Chem. Theory Comput. 2009, 5, 615–626.(23) Beevers, A. J.; Kukol, A. J. Mol. Graphics Modell. 2006,

25, 226–233.(24) Liu, X.; Xu, Y.; Li, H.; Wang, X.; Jiang, H.; Barrantes, F. J. PLoS

Comput. Biol. 2008, 4, e19.(25) Cuthbertson, J. M.; Bond, P. J.; Sansom, M. S. P. Biochemistry

2006, 45, 14298–14310.(26) Nagle, J. F.; Tristram-Nagle, S. Curr. Opin. Struct. Biol. 2000,

10, 474–480.(27) Kucerka, N.; Tristram-Nagle, S.; Nagle, J. F. Biophys. J. 2006,

90, L83–L85.(28) Oostenbrink, C.; Villa, A.; Mark, A. E.; Van Gunsteren, W. F. J.

Comput. Chem. 2004, 25 (13), 1656–1676.(29) Van Der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G. J.

Comput. Chem. 2005, 26, 1701–1718.(30) Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. J. Chem.

Theory Comput. 2008, 4, 435–447.(31) Murzyn, K.; Zhao, W.; Karttunen, M.; Kurdziel, M.; R�og, T.

Biointerphases 2006, 1 (3), 98–105.(32) Zhao, W.; R�og, T.; Gurtovenko, A. A.; Vattulainen, I.; Karttu-

nen, M. Biophys. J. 2007, 92 (4), 1114–1124.(33) Chiu, S. W.; Clark, M.; Subramaniam, B. S.; Scott, H. L.;

Jacobson, E. Biophys. J. 1995, 69 (10), 1230–1245.(34) Bakouche, O.; Gerlier, D.; Letoffe, J. M.; Claudy, P. Biophys. J.

1986, 50 (1), 1–4.(35) Lopez, J. J.; Cascales; Otero, T. F.; Romero, A. J. F.; Camacho,

L. Langmuir 2006, 22, 5818–5824.(36) Berendsen, H. J. C.; Postma, J. P. M.; Vangunsteren, W. F.;

Dinola, A.; Haak, J. R. J. Chem. Phys. 1984, 81 (8), 3684–3690.(37) Patra, M.; Karttunen, M.; Hyvonen, M. T.; Falck, E.; Lindqvist,

P.; Vattulainen, I. Biophys. J. 2003, 84 (6), 3636–3645.(38) Godawat, R.; Jamadagni, S. N.; Garde, S. Proc. Natl. Acad. Sci. U.

S.A. 2009, 106, 15119–15124.(39) Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999,

103, 4570–4577.(40) Chandler, D. Nature 2002, 417, 491.(41) Chandrasekhar, I.; Kastenholz, M.; Lins, R. D.; Oostenbrink,

C.; Schuler, L. D.; Tieleman, D. P.; van Gunsteren, W. F. Eur. Biophys. J.2003, 32 (1), 67–77.(42) Sovago, M.; Vartiainen, E.; Bonn, M. J. Chem Phys. 2009,

131, 161107.(43) Zhou, F.; Schulten, K. J. Phys. Chem. 1995, 99, 2194.(44) Volkov, V. V.; Palmer, D. J.; Righini, R. J. Phys Chem. B 2007,

111, 1377.(45) Damodaran, K. V.;Merz, K.M., Jr.Langmuir 1993, 9, 1179–1183.(46) Pasenkiewicz-Gierula, M.; Takaoka, Y.; Miyagawa, H.; Kita-

mura, K.; Kusumi, A. J. Phys. Chem. A 1997, 101, 3677–3691.(47) Krishnamurty, S.; Stefanov, M.; Mineva, T.; Bgu, S.; Devois-

selle, J. M.; Goursot, A.; Zhu, R.; Salahub, D. R. J. Phys. Chem. B 2008,112, 13433–13442.(48) Hubner, W.; Blume, A. Chem. Phys. Lipids 1998, 96, 99–123.(49) Hauser, H.; Pascher, I.; Pearson, R. H.; Sundell, S. Biochim.

Biophys. Acta 1981, 650, 21–51.(50) Hauser, H.; Pascher, I.; Sundell, S. Biochemistry 1988,

27, 9166–9174.(51) HongK, M.; Zimmermann; Schmidt-Rohr, H. Biochemistry

1996, 35, 8335–8341.(52) Bruzik, K. S.; Harwood, J. S. J. Am. Chem. Soc. 1997,

119, 6629–6637.(53) Aussenac, F.; Laguerre, M.; Schmitter, Jean-Marie; Dufourc,

E. J. Langmuir 2003, 19, 10468–10479.(54) Leekumjorn, S.; Sum, A. K. Biophys. J. 2006, 90 (11), 3951.(55) Chandrasekhar, I.; Bakowies, D.; Glattli, A.; Hunenberger, P.;

Pereira, C.; Van Gunsteren, W. F. Mol. Simul. 2005, 31, 543–548.

(56) Debnath., A.; Mukherjee, B.; Ayappa, K. G.; Maiti, P. K.; Lin, S.-T. J. Chem. Phys. 2010, 133, 174704.

Date post:	11-Dec-2016
Category:	Documents
Upload:	sudip
View:	212 times
Download:	0 times

Headgroup Mediated Water Insertion into the DPPC Bilayer: A Molecular Dynamics Study

Documents