Oriented Confined Water Induced by Cationic Lipids - Langmuir (ACS

Article pubs.acs.org/Langmuir

Oriented Confined Water Induced by Cationic Lipids Lydia Woiterski,*,† David W. Britt,‡ Josef A. Kas̈ ,† and Carsten Selle† †

Institut für Experimentelle Physik I, Abteilung Physik Weicher Materie, Universität Leipzig, Linnéstrasse 5, 04103 Leipzig, Germany Department of Biological Engineering, 4105 Old Main Hill, Utah State University, Logan, Utah 84322-8200, United States

‡

S Supporting Information *

ABSTRACT: We report on attenuated total reflection Fouriertransform infrared (ATR FTIR) spectroscopic measurements on oriented lipid multilayers of N,N-dimethyl-N,N-dioctadecyl-ammonium halides (DODAX, X = F, Cl, Br, I). The main goal of this study is the investigation of the structure and spectroscopic properties of water absorbed to these model membranes. Intensities of the water stretch absorptions were used to determine the amount of bound water. At high water activity, DODAF membranes bind ∼11 water molecules/lipid while DODAC and DODAB adsorb 1−2 water/lipid and DODAI was hydrophobic. By adjustment of DODAF hydration to ∼2 water molecules, stretching absorptions from water of the first hydration shell were accessible for the fluoride, chloride, and bromide analogs. The polarized measurements demonstrate highly confined and oriented water with infrared (IR) order parameters ranging from 0.2 to −0.4. Resolved IR water band components are attributed to different hydrogen-bonded populations. Complementary molecular dynamics simulations of DODAB strongly support the existence of differently hydrogen-bonded and oriented water within DODAB multilayers. A combination of both techniques was used for an assignment of water stretch band components to structures. The described cationic lipid systems are a prototype for a bottom-up approach to understand the IR spectroscopy of structured water at biological interfaces since they permit a defined increase of hydrophilic water−anionic interactions leading to extended water networks at membranes.

1. INTRODUCTION Water has outstanding solvent properties permitting important applications in chemistry, engineering, and industrial technologies. In biological systems water is intricately coupled to numerous processes such as energy transduction, enzyme activity, and solute transport through pore proteins. The unique properties of water arise from the dynamic intermolecular hydrogen-bonding network.1,2 Water molecules in the liquid state form between two to four hydrogen bonds (HBs) to neighboring molecules.3−5 The broad and strong stretching vibration absorption is a superposition of several Gaussian bands of different frequency and width,4,6−8 ascribed to diverse hydrogen-bonded environments and/or vibrational modes derived from antisymmetric or symmetric stretching and to a bending overtone absorption.4,9 Assignment of the bands is challenging and still controversial as vibrational stretching modes involve complex intra- and intermolecular couplings.2,4,7,10−12 Despite the large body of experimental data accumulated over the last decades and recent progress on the theoretical description, there is still no consensus in the assignment of stretching bands of bulk liquid water.4 Spectra of uncoupled OH or OD oscillators occurring in dilute HOD in H2O or D2O are easier to interpret in terms of changed localized interactions if H bonding in liquid water is alternated by effects of temperature or solutes.13 Ultrafast vibrational spectroscopy methods for investigation of dilute HOD led to a deeper understanding on the dynamical behavior of liquid water.2 © 2012 American Chemical Society

A tremendous interest persists in the interactions between biomolecules and water. Interfacial water strongly differs from the bulk liquid. It has a profound effect on the folding of proteins and DNA as well as on structure and function of biological membranes.14,15 Lipid−water interactions were studied during the last 30 years by thermodynamic (differential scanning calorimetry), scattering (X-ray and neutron diffraction), and spectroscopic (nuclear magnetic resonance, NMR, and IR) methods.9,16−20 They have been often examined via phospholipid headgroup IR absorptions such as carbonyl or phosphate bands.10,20,21 Alternatively, H bonding structures can be monitored by changes of water stretching IR absorptions.10,22−24 However, these changes are relatively small, leaving stretch bands similar to those of pure liquid water, indicating a flexible HB network as proposed for the bulk liquid.9 Differences in the dynamics between membrane-bound and bulk water were demonstrated by molecular dynamics (MD) simulations,25,26 ultrafast vibrational spectroscopy methods,22,27 and sum frequency generation (SFG) spectroscopy.28 Cationic lipids and their biophysical properties are widely studied due to their potential application in gene therapy. 29−33 N,N-Dimethyl-N,N-dioctadecyl-ammonium (DODA) halides are structurally simple two-chain lipids spontaneously forming bilayer assemblies in aqueous media.34 Received: May 30, 2011 Revised: February 7, 2012 Published: February 16, 2012 4712

dx.doi.org/10.1021/la205043x | Langmuir 2012, 28, 4712−4722

Langmuir

Article

2. EXPERIMENTAL SECTION AND SIMULATION DETAILS

The thermotropic and lyotropic phase behavior of model membranes from chloride and bromide analogs have been investigated by a variety of methods.35−41 The water capacity of DODAX (X = Cl, Br) phases under controlled laboratory conditions is one to three water molecules/lipid.36,37 Thus, the presence of bulk water can be ruled out as only interfacial water can exist in DODAX multilayers. A thorough study on the structure and orientation of water at DODAX model membranes has not been published. We present a detailed attenuated total reflection (ATR) Fourier transform infrared (FTIR) spectroscopy study of the lipid-associated water structure as a function of the counterion, allowing us to resolve and assign water absorption bands to discrete H bonding modes. To this end, DODAX (X = F, Cl, Br, I) compounds with ions other than bromide were produced by substituting bromide in DODAB by other halide ions, purified, and analytically characterized. The counterions belong to the Hofmeister series, which ranks the ions according to their ability to modify the HB network of bulk water.42 Current research aims at unraveling direct interactions between ions and water molecules of the first hydration shell of macromolecules.43 However, the Hofmeister series is not further discussed here. We used polarized ATR FTIR spectroscopy, which is a powerful tool to investigate molecular orientation and ordering of membranes and associated water.9,44,45 Prior characterization of the membrane matrix, i.e., of the lipids’ phase behavior, was essential for the main goal of the present work: a structural study of adhered water. The stretching absorption from water at cationic lipids was used to probe the molecular orientation and H bonding within the membrane environment. At low hydration, an anisotropic orientation was found for membrane-bound water. Corresponding stretch absorptions demonstrate clearly visible subbands which are attributed to differently H bonded populations. Within DODAX membranes, halide ions and water molecules represent two different species of HB acceptors. Variation of halide counterions leads to different H bonding strengths, changes of water stretch absorptions, and corresponding molecular orientations that follow the Hofmeister series for anions. Complementary information about the structure and dynamics of lipids and associated water was obtained from molecular dynamics (MD) simulations.25,26,46−49 The lipid bilayer structure was extrapolated from crystal data of DODAB. Simulations confirm the existence of differently H bonded and oriented water along the DODAB bilayer. Relatively slow rotation and confined diffusion were found for the lipidassociated water molecules. Our results on IR stretching absorptions in conjunction with MD data lead to a molecular picture of the structure and dynamics of confined and oriented water at cationic lipid membranes. The described cationic lipid systems are ideal for the pursuit of a bottom-up approach to understand the IR spectroscopic properties of water structures at biological interfaces since they allow the gradual variation of water−anionic interactions and sequential build up of water networks within membranes. We illustrate this through experiments on the phase behavior of the lipid membranes, the water absorption capability, and the orientation of water within DODAX membranes with supporting MD simulations of DODAB membranes that reveal distinct HB environments of water molecules within DODAB through analysis of their orientation distributions.

2.1. Materials. Solvents and KI were from Merck (purity ≥ 99.5% Darmstadt, Germany). DODAB and DMPC were purchased from Sigma-Aldrich (≥98%, Schnelldorf, Germany). The other three DODA halides were either not commercially available or not of adequate purity. Thus, DODAB was converted into its chloride or iodide analogs in a water−isopropanol solution with up to 20-fold excess of KCl (>99.5%, Roth, Karlsruhe, Germany) or KI, respectively. Analog ion exchange failed for DODAF; thus, it was produced by precipitation of insoluble AgBr after addition of an equimolar amount of AgF (≥99%, Fluka, Buchs, Switzerland). This suggests that strongly bound water molecules shield the fluoride charge, preventing substitution of bromide ions. In the Supporting Information preparation details and analytical EDX (Figures S1−S4), MALDITOF (Figure S5), and NMR (Figures S6 and S7) data of all lipids are given. The structure of the lipids used is given in Scheme 1.

Scheme 1. Chemical Structure of the Dioctadecyldimethylammonium Halides DODAX (X = F, Cl, Br, I) According to the Crystal Structure of DODAB from Okuyama et al.50

2.2. IR Spectroscopic Measurements. Samples were prepared as lipid bilayer multistacks by spreading of 80 μL of chloroformic lipid (10 mg/mL) solution on a clean ZnSe ATR crystal (72 × 10 × 6 mm3, incident angle 45°, LOT Oriel, Darmstadt, Germany) to yield a film thickness of ∼2 μm. The crystal with the sample was mounted on a horizontal Benchmark unit (Specac, Orpington, U.K.). After solvent evaporation, ATR FTIR measurements were carried out by a Digilab Excalibur FTS 3100 spectrometer (Varian, Darmstadt, Germany). Absorbance spectra A∥(ν) and A⊥(ν) were measured under parallel and perpendicular polarization (KRS-5 grid polarizer, LOT Oriel, Darmstadt, Germany) with regard to the incident angle of the IR beam. Relative humidity (RH) and temperature were adjusted by a moisture generator10 (Humivar, Leipzig, Germany) and a circulating water bath (Julabo, Seelbach, Germany). Lipid films at 50% RH and 25 °C were successively heated in steps of ΔT = 2−5 K. Chain melting transition temperatures of DODAX (X = F, Cl, Br) were determined at the point of maximal slope of the νs(CH2) wavenumber as a function of temperature (cf. Figure S8 in the Supporting Information) and well reproducible after rehydration. Samples were hydrated with either H2O or D2O (Chemotrade, Leipzig, Germany). The low hydration of DODAX lipids prevented a structural study of isotopically diluted water (HOD). For a mixture of D2O:H2O 1:20 and maximal hydration, the intensity of the HOD stretching band in the ATR spectra (data not shown) was so low that it was impossible to determine molecular orientations. Further details on the measurements are given in the Supporting Information. 2.3. Evaluation of Dichroic Data. Polarized ATR FTIR spectroscopic measurements are classified by the sample thickness d with respect to the penetration depth λd of the evanescent wave of the IR light.44 Thick films fulfilling the condition d > λd were used in this study. Their advantage over thin films was discussed previously.51 The IR order parameter SIR derived from the absorbances A∥ and A⊥ using the thick film approximation relates to the mean orientation of an individual transition dipole moment μ⃗ as follows10,44 SIR = 4713

1 ⟨3 cos2 θμ − 1⟩ 2

(1)


Langmuir

Article

θμ is the angle enclosed by the transition dipole moment μ⃗ and the normal of the ATR crystal. The tilt order parameter Sθ of polymethylene chains is given as a linear combination of the IR order parameters of the symmetric and antisymmetric CH2 stretching mode52

Sθ = − [SIR (νs(CH2)) + SIR (νas(CH2))]

samples above as well as below Tm. More details on the phase behavior of the lipids are given in the Supporting Information. 3.2. Water Capacity of DODAX Membranes. Figure 1 shows the OD stretching band of DODAX-associated D2O.

(2)

SIR values of alkyl chains can be derived from the sum of the average orientations of individual CH2 transition dipole moments of the chain.53 Further details on band fitting and analysis are given in the Supporting Information. 2.4. Simulations. MD simulations were performed with the GROMACS 3.3.3 software package. Water was modeled using the simple point charge (SPC) model.54,55 DODAB was modeled using the parameters provided by Siu et al. to implement the Berger lipid force field within GROMOS96 (ffG53a6).26,47 The parameters were slightly modified to accommodate the structure of DODAB. The bromide ion parametrization from OPLS (Optimized Potentials for Liquid Simulations)56 was transferred accordingly. A DODAB bilayer of lipids, bromide ions, and water molecules, 128 constituents each, was generated from X-ray data50 and energy minimized by a conjugate gradient. The cut offs for van der Waals and Coulomb interactions were set to 1 nm. The particle-mesh Ewald method was used to model long-range electrostatic interactions. Bond lengths were constrained with the LINCS algorithm57 for lipids and the SETTLE algorithm for water.58 Run parameters for our MD simulations were chosen according to similar studies.26,47,59,60 Temperatures of lipids, water, and ions were separately kept at 300 K using the Berendsen thermostat and a coupling constant of 0.1 ps. Semi-isotropic pressure coupling was performed with the Berendsen barostat and a coupling rate of 2.5 ps at 1 bar. The modeled bilayer was equilibrated for 200 ns with a sampling rate of 500 fs. Detailed evaluation was carried out with a continuative 1 ns simulation run performed with a faster sampling rate of 50 fs. Trajectories were analyzed for translational and orientational diffusion, H bonding, and molecular orientations of SPC water.

Figure 1. ATR spectra of ν(OD) absorptions from D2O bound to oriented DODAX multilayers (X = F, Cl, Br, I) and DMPC at 25 °C and 80% RH recorded with p-polarization (solid lines) and spolarization (dotted), respectively. All spectra were normalized to the lipids’ CH stretching intensities to compare for the water content of the different lipids.

Pronounced differences in ν(OD) absorption intensities, band structures, and bandwidths were observed. H2O-hydrated films demonstrated analogous properties (not shown). To quantify the water adsorbed per lipid, the broad OH stretching bands were related to the intense CH stretching bands at ∼3000 cm−1 due to the lipid alkyl chains.63,64 Using the ratio of the integrated absorbances of the (OH) and (CH) absorptions and data on the hydration of phospholipids by Karl Fischer titration64 provided an average hydration of 11.0 ± 2.6 water molecules/lipid for DODAF, 1.2 ± 0.2 for DODAC, and 0.8 ± 0.2 for DODAB. DODAI films resisted hydration and were excluded from further analysis. Due to low hydration of the DODAX samples, the effect of membrane swelling is considered small and was ignored in the estimations of the water content. The water capacity monotonously decreases with increasing halide radius, correlating with the halide anions’ hydration free energies.65 Indeed, DODAF could not be completely dehydrated under our experimental conditions. DODAF-bound water shows a strong ν(OD) band centered at 2501 cm−1 and a shoulder at 2395 cm−1 attributed to the second and higher hydration shells. However, it does not show substantial differences compared to spectra at low hydration (∼2 water/lipid, cf. Figure 2). In contrast, water at DODAC (and DODAB) exhibits three distinct peaks at 2569 (2617), 2475 (2560), and 2385 cm−1 (2463 cm−1). These results agree with IR spectra of low-hydrated DODAC samples36,37 and our previous measurements, where a clearly split ν(OH) band of DODAB-associated water was found.66 The striking substructures in the OD/OH stretch bands of DODAC and DODAB were never observed for phospholipid samples (e.g., DMPC in Figure 1), even at very low hydration.10,22 DODAB vesicles dispersed in excess water displayed a slightly structured ν(OH) band typical for the solvent.38 The narrow bandwidths of the ν(OD) subbands suggest an underlying water population having confined H bonding and mobility.67 The broadest OD stretching band of water within DODAF (2000−2700 cm−1) resembles the one of pure bulk water.68 Bandwidths found for DODAC- and DODAB-bound water (2200−2700 cm−1) are significantly smaller. The absorption intensity at lower

3. RESULTS AND DISCUSSION 3.1. DODAX Phase State. The phase behavior of hydrated DODAX model membranes is briefly characterized. Chain melting transition temperatures (Tm) of 65 ± 1 °C for DODAF, 50 ± 1 °C for DODAC, and 47 ± 1 °C for DODAB were found. These values are for DODAC slightly higher and for DODAB in good accordance with previous results.37,39 DODAI showed a continuous, relatively small wavenumber increase for the νs(CH2) band during heating, indicating less ordered alkyl chains at all temperatures. Its low water capacity impeded the study of associated water. For low-hydrated (∼1 water/lipid) DODAB, DODAC, and DODAF and dry DODAI samples we employed eqs 1 and 2 to find Sθ values of 0.37 ± 0.06, 0.22 ± 0.05, 0.39 ± 0.06, and 0.34 ± 0.09, respectively. Corresponding chain tilt angles θμ are 40.4 ± 5.5°, 46.1 ± 10.0°, 39.6 ± 6.1°, and 41.6 ± 10.7° with regard to the ATR crystal. For DODAB θμ agrees well with earlier data.50,61 At our conditions (25 °C; 80 RH%, if not stated otherwise) the lipids are in the gel phase. We did not observe a coagel phase, as the DODAX δsciss(CH2) wavenumber at 1470 cm−1 is well below the previously reported coagel phase absorption (1471.2−1473 cm−1).35,36,41 DODAX headgroups directly interact with interbilayer water, which is reflected by the νs(CNC) stretching vibrations between 880 and 930 cm−1.10,62 IR spectra of hydrated DODAX (X = F, Cl, Br) gel-phase samples exhibited two peaks at 910 and 920 cm−1, which disappeared for dehydrated samples at the same temperature (cf. Figure S9, Supporting Information). At the cost of those two bands, the integral intensity of the band at 889 cm−1 increased for dry 4714


Langmuir

Article

Figure 2. Band-fitting results for OH (top) and OD (bottom) stretching bands of water bound to DODAF (at 25 °C, 3% RH), DODAC, and DODAB (both at 25 °C, 80% RH) from left to right. ATR spectra were recorded with p-polarized (black solid) and s-polarized (dotted) light. Deconvolved Gaussian components were depicted for spectra measured with p-polarized IR light together with their sums drawn by solid red lines.

Table 1. Band-Fitting Results of ν(OH) and ν(OD) Absorptions of DODAX-Bound Water DODAF

H2O

D2O

1 2 3 4 5 6 1 2 3 4 5 6

DODAC

DODAB

νmax/ cm−1

Δν1/2/ cm−1

SIR

θ/deg

% area

νmax/ cm−1

Δν1/2/ cm−1

SIR

θ/deg

% area

νmax/ cm−1

Δν1/2/ cm−1

SIR

θ/deg

% area

2928 3188 3409

248 278 174

−0.13 −0.05 −0.14

60.2 56.8 60.7

20 64 16

2307 2385 2484 2546

198 153 40 129

0.16 −0.21 0.69 −0.24

48.4 63.9 27.0 65.4

8 69 1 22

3238 3244 3368 3411 3446 3549 2368 2380 2475 2524 2569 2622

117 25 29 162 41 137 123 26 23 168 34 87

−0.17 −0.44 −0.47 −0.05 0.22 0.22 −0.15 −0.15 −0.42 −0.09 0.01 0.13

62.0 78.5 81.9 56.8 46.1 46.1 61.1 61.1 76.6 58.5 54.3 49.6

11 1 3 73 9 3 11 1 4 59 13 12

3237 3338 3394 3444 3519 3580 2379 2461 2489 2561 2623 2657

105 66 155 65 60 67 104 37 108 73 42 42

−0.19 −0.11 −0.17 −0.06 −0.14 −0.16 −0.22 −0.26 −0.19 −0.13 −0.11 −0.21

63.0 59.3 62.0 57.2 60.7 61.6 64.4 66.4 63.0 60.2 59.3 63.9

7 10 42 22 15 4 9 7 25 40 14 5

3.3.2. Band Analysis and Orientation Measurements for DODAF. DODAF’s water stretching band (Figure 2) is less resolved into subbands than for DODAC and DODAB. The low ν(OH) and ν(OD) band maxima at 3190 and 2330 cm−1 indicate stronger H bonding than in bulk water with a band maximum at ∼3400 cm−1.8 Fitting of these bands for samples hydrated by H2O and D2O resulted in three and four subbands, respectively (see Table 1). The low-frequency wing of the ν(OH) band (3050−2800 cm−1) overlaps with methylene/ methyl stretching bands. Thus, analysis of the ν(OH) band required prior subtraction of the alkyl absorptions leading to a higher error in the resulting ν(OH) band and in fitted subbands within this spectral range. Apparent differences of the D2O and H2O IR stretch bands can arise from this correction. The strongest components (no. 2) occur at 3188 (H2O) and 2385 cm−1 (D2O). The peaks at 3409 (H2O, no. 3) and 2546 cm−1 (D2O, no. 4) approximately agree in their order parameters, which suggest analog vibrational groups for these components. Observations for DODAC and DODAB samples support this interpretation (see below). A relatively high error can be expected for Sθ and θ of peak no. 3 (D2O) due to the poor absorbance signal-to-noise ratio. Spectra recorded from adsorbed H2O did not show a significant shoulder at a comparable frequency range. Occurrence of the additional weak peak no. 3 can be connected to specific properties of D2O.70 3.3.3. Band Analysis and Orientation Measurements for DODAC. Spectra of water adhered to DODAC bilayer stacks

wavenumbers is concomitantly less prominent, which clearly depends on the anions interacting with water. 3.3. Stretching Bands of Water Adsorbed to DODAX. 3.3.1 General Remarks on Samples and Band Assignment. To compare counterion effects, spectra of DODAX samples with a constant hydration of ∼1 water/lipid were analyzed (Figure 2). These were measured at 25 °C and 3% RH for DODAF and 80% RH for DODAC and DODAB. We used initial fit parameters like peak positions, widths, and intensities for the OH stretch bands according to a study by Liu et al.8 They measured ATR FTIR spectra of neat water and aqueous sodium halide solutions and assigned ν(OH) subbands at 3249, 3420, and 3591 cm−1 to an intermolecularly coupled symmetric stretching vibration within a symmetric HB network, to a weaker coupled asymmetric HB network, and to a coupled antisymmetric ν(OH) mode, respectively.8 Spectra of the ν(OH) and ν(OD) regions recorded for DODAX samples (X = F, Cl, Br) and their fitted Gaussian subbands are shown in Figure 2 for p- and s-polarization (dotted), respectively. Differences between D2O and H2O stretching absorptions reflecting the dissimilar properties due to isotopic substitution have been a widely investigated phenomenon and are still not completely understood.2,7,68,69 All band-fitting results are given in Table 1. At 3% RH DODAF samples had a minimal water content of ∼2 water/lipid. Further removal of water by heating was avoided because the ν(OH) band structure was shown to be affected by temperature changes.7,70 4715


Langmuir

Article

OD stretching band at ∼2460 cm−1 with a larger bandwidth than observed for neat water.72 Thus, the findings for the DODAF water stretch bands indicate the presence of a significant though minor fraction (16−22% of adsorbed water) of similar vibrational groups as reported for bulk aqueous fluoride solutions. Due to the low proportion of water in DODAF exhibiting vibrations similar to that of bulk fluoride solutions, a major fraction of water molecules should be confined at the polar surfaces between DODAF bilayers. A comparison with IR spectroscopic data of fluoride−water clusters allows us to interpret the ν(OH) spectra. Water−fluoride (1:1) complexes exhibit very low water stretch band frequencies.73 (H2O)2·F− and (H2O)3·F− clusters give rise to IHB (ionic hydrogen bond) stretching absorptions at 2490 and 2890 cm−1, respectively.74,75 F−·(H2O)4 and F−·(H2O)5 clusters exhibited IHB stretch bands at 3072 and 3143, 3389 cm−1, respectively, and little or no indications for H bonding between water molecules were found.74 Fluoride clusters with 3−5 water molecules have the strongest vibrational absorptions in the range of 2890 and 3143 cm−1. IR absorptions of DODAF water in this region (H2O, peak nos. 1 and 2), which comprise 77−84% of adsorbed water, may suggest a major fraction of related structures within DODAF bilayers. DODAC. The strongest peak of H2O (no. 4) is similar to a ν(OH) component at 3396 cm−1 reported for bulk water, which was assigned to a stretching absorption of a weaker coupled asymmetric HB network.8 The corresponding ν(OD) component agrees well with a chloride hydration band (0.442 M NaCl in 8% D2O/H2O) and an uncoupled ν(OD) stretch band (4% D2O/H2O at high chloride concentrations) which were both found at 2530 cm−1.71,72 This suggests that either a large fraction of water molecules in DODAC includes uncoupled OH or OD oscillators or asymmetric H bonded water in these samples exhibits similar spectral features as decoupled OH or OD oscillators. As for DODAF, the measured spectra of water at DODAC are compared with water−chloride cluster spectra because similar interactions can be assumed. Absorption bands at 3441 and 3408 cm−1 reported for Cl−·(H2O)5 and Cl−·(H2O)4 clusters were assigned to IHB water stretch vibrations,76 respectively, and highly agree with ν(OH) component nos. 5 and 4. Cl−·(H2O)3 clusters show three bands between 3200 and 3400 cm−1. IR spectra of Cl−·(H2O)2·Ar3 clusters show five bands. Their peaks at 3375 and 3130 cm−1 are assigned to a DD and acceptor−donor (AD) IHB OH stretching.75,77 A relation may exist between the DD IHB OH stretching band and the weak component at 3368 cm−1. All correlations point to similarities of DODAC-associated water and anionic clusters with 2−5 water molecules. Our findings agree well with earlier IR spectroscopic studies of H2O bound to DODAC gel phases.35,36 However, ν(OD) absorptions presented here differ strongly from those results possibly due to the better signal-tonoise ratio in our measurements.36 DODAB. Peak nos. 1 and 6 of the ν(OH) band are shifted by 12 cm−1 to lower frequencies compared to peak positions in spectra of bulk aqueous sodium bromide (NaBr) solutions.8 The position of the strongest component (no. 3) agrees with the corresponding ν(OH) band of pure bulk water and is 30 cm−1 lower than that of NaBr solutions.8 This is supported by the frequency of the analog ν(OD) component no. 3 which is close to the decoupled ν(OD) stretching absorption in bulk neat water.70,71 In peak system II, the frequency of the

were fitted with six Gaussian components (cf. Figure 2, Table 1) assorted into two different peak systems. Peak system I comprises three broader components (nos. 1, 4, and 6). The three remaining narrower components (nos. 2, 3, and 5) are denoted as peak system II. Due to the high similarity of the H2O and D2O stretching spectra, analog peak systems I and II were found for the observed ν(OD) bands. The ν(OH) component wavenumbers of peak system I deviate only slightly from values reported for NaCl solutions.8 The three peaks represent 72−87% of the integral absorbance of the entire ν(OH) or ν(OD) stretching bands. This can reflect a major water fraction within DODAC bilayers, which retains similar properties or interactions as water within aqueous chloride solutions. The broad central peaks (no. 4) at 3411 and 2524 cm−1 strongly correlate with each other. Moreover, the orientation angles of the other ν(OH) and ν(OD) components (nos. 1 and 6, respectively) of peak system I agree well with each other but show a large difference between component nos. 1 and 6. This supports a strong correlation to the fundamental symmetric (no. 6) and antisymmetric stretching (no. 1) vibrations of water. The narrow ν(OH) and ν(OD) components of peak system II match in their widths, intensities, and order parameters. Small widths indicate a narrow distribution of H bonding of involved vibrational groups. Peak no. 5 at 3446 cm−1 can be assigned to less H bonded water molecules due to its comparatively higher peak frequency. Peak no. 5 and peak nos. 2 and 3 show a large difference in the order parameters, which suggests a drastic change in orientation of the related vibrational transition moments. Therefore, component nos. 2 and 3 can be linked to oscillators belonging to a uniformly oriented but differently H bonded water fraction, varying in accepted HBs. 3.3.4. Band Analysis and Orientation Measurements for DODAB. DODAB-associated water exhibits very similar IR spectroscopic properties as water within DODAC (Figure 2). As for DODAC, peaks and additional visible shoulders are summarized in peak systems I and II including the broader (nos. 1, 3, and 6) and narrower (nos. 2, 4, and 5) components, respectively. ν(OH) spectra agree mostly with ν(OD) spectra regarding the widths, intensities, and order parameters. Peak system I contributes ∼39−53% to the entire integral stretching absorption. The average orientation of water in DODAB appears more homogeneous than within DODAC. A possible explanation is that water molecules within DODAC samples could be sterically less confined than in DODAB, giving rise to more variation in orientation compared with bromide ions, which require more room than chlorides. The higher water capacity of DODAC supports this assumption. 3.3.5. Comparison of Water Stretch Bands from Aqueous Salt Solutions, Ion−Water Clusters, and DODAX Membranes. DODAF. DODAF ν(OH) peak no. 3 resembles the peak at 3412 cm−1 recorded for aqueous sodium fluoride (NaF) solutions (molar fraction x = 0.015 ≈ 0.83 M) regarding its frequency and bandwidth.8 As the latter showed little deviations from neat water, one can also assume weakly modified H bonding for water at DODAF.8 Furthermore, salt solutions with isotopically diluted water exhibit ν(OD) absorptions similar to D2O adhered to DODAF multilayers. IR spectra of NaF solutions (0.642 M in 8% D2O/H2O) showed a fluoride hydration band at 2472 cm−1, which was interpreted as a marker of an anion-bonded isolated OD stretch oscillator within HOD molecules.71 In another study, concentrated NaF (molar fraction, x = 0.25) in 4% D2O/H2O mixtures yielded an 4716


Langmuir

Article

strongest ν(OH) component at 3444 cm−1 (no. 4) deviates by 24 cm−1 from the position of a related subband of aqueous NaBr solutions and its width is significantly smaller (65 vs 257 cm−1).8 The analog ν(OD) peak frequency agrees well with that of a bromide hydration band observed for uncoupled OD oscillators,71 although the bandwidth is much smaller (73 vs 147 cm−1), indicating less variation of the H bonded and/or dipolar environment within DODAB. The results for DODAB are also related to stretch bands of small water clusters. Bromide−water clusters with 1−4 water molecules show IHB OH stretch bands at 3368, 3373, 3422, and 3466 cm−1.78 Spectra of Br−·H2O and Br−·(H2O)2 clusters give rise to IHB absorption peaks at 3270 and 3438 cm−1, respectively. The latter band was assigned to an OH oscillator of a DD water molecule with HBs to bromide and another water.73,77 Such a cluster-like water structure can occur within DODAB bilayers and may be reflected by component no. 4. Otherwise, it can be described as an asymmetric unit including an OH stretching group with HB to bromide.8 Component no. 3 can be due to a similar water fraction accepting a water molecule, leading to a red shift compared with component no. 4. 3.3.6. Correlations between Water Stretching Bands from the Varied DODAX Species. Hydrated DODAX samples exhibit highly similar OH stretching component frequencies. DODAF peak nos. 3 (H2O) and 4 (D2O) are related with DODAC peak no. 4 and DODAB peak no. 3. All of these components are connected to absorptions of isolated water stretching oscillators (HOD in H2O) in aqueous halide solutions. DODAF (no. 4) and DODAC (no. 3) peaks are also similar to stretch band components observed for bulk water (H 2 O or D 2 O) halide solutions. Furthermore, component no. 1 of spectra from DODAC and DODAB samples agrees in frequency and width. We find that an asymmetric H bonded water population in DODAF and DODAC is split in DODAB bilayers into two fractions, which strongly and weakly interact with bromide anions, respectively. Further similarities are found for ν(OH) component nos. 5 (DODAC) and 4 (DODAB); however, they are not as apparent for component nos. 2 and 3 (DODAC) and 1 and 2 (DODAB), which differ in the order parameter. The common features in spectra point to water fractions in similar polar/H bonded environments. The fraction of bulk-like water can be in DODAC larger than in DODAB because of the high integral absorbance of component nos. 1, 4, and 6. A possible reason is the higher water content of the DODAC samples leading to less interaction of water with the ammonium salt components. It should be further noted that the bandwidths of peak system II of DODAC water are noticeably smaller than those of analog DODAB values. This indicates the existence of a water fraction whose molecules are less similar to bulk in DODAC than in DODAB. FTIR analysis of interfacial water as a function of DODAX counterion, hydration state, and temperature yielded a detailed picture of the distribution of the water structure and corresponding lipid order parameters. 3.4. Simulations of DODAB Bilayers. 3.4.1 Model Building and Running, Extraction of Spectroscopy-Related Data. The orientation and nature of molecular oscillators giving rise to significant features of OH stretching spectra can be better interpreted if the structure of interacting water and lipid molecules is known.26,48,60 MD simulations may provide an atom-resolved static and dynamic description of those

structures. Appropriate sets of atomic positions to be used in MD simulations can be derived from the crystal structure of the lipids which form a two-dimensional crystal powder.10,79 Coordinates of single lipid crystals are multiplied in space according to the crystal unit’s cell parameters defining a simulation box.46,59 DODAB’s crystal structure agrees with both the chain tilt angles (cf. section 3.1) as well as the water/ lipid ratio (cf. section 3.2) determined in our study.50,61 Thus, it was used to set up coordinates of a cationic bilayer with 128 DODAB lipids including 128 bromide ions and 128 water molecules. Crystal-structure-based coordinates for other DODAX analogs were not accessible from the literature. A Berger force field of DOPC provided by R. Böckmann was modified for DODAB bilayers.26,47 According to Okuyama et al., two possible oxygen positions of water denoted as OW1 and OW2 were used with added hydrogen for SPC water molecules.50 Energy minimization (EM) using conjugate gradient resulted in two different DODAB/water bilayer structures referred to as EM−OW1 and EM−OW2. They were used as starting configurations for MD simulations as NPT ensembles (298 K and 1 bar). Energy minimization and MD simulations yielded atomic positions of DODAB/water membranes and information on alkyl chain orientation, water orientation, and H bonding. The chain segment order parameter ⟨SCD⟩ can be established from the orientation of every methylene group.26 It is equivalent to chain order parameters obtained by NMR spectroscopy for deuterated lipids.80 ⟨SCD⟩ values obtained from simulations can be converted into average CH2 orientation angles, which can be used to compute a corresponding IR order parameter Sθ reflecting the average alkyl tilt. Likewise, positions of simulated water molecules can be used for calculation of molecular orientations and corresponding IR order parameters. The orientations of the transition dipole moments of pure (uncoupled) symmetric and antisymmetric stretching modes of isolated water and of the stretching vibration along isolated (uncoupled) water OH bonds agree with the orientations of water dipoles μ, the HH vectors connecting the H atoms of each water molecule, and the OH vectors defined by the direction and length of OH bonds. Corresponding orientation angles were calculated relative to the bilayer plane. Order parameters Sμ, SHH, and SOH were computed for comparison with IR spectroscopic measurements. Finally, simulated water and bromide positions were analyzed to characterize involved HBs. Due to the simplicity of the SPC water model, a geometric definition of H bonding was used in this study. The maximum HB length defined by the first minimum of the radial distribution function was found to be 0.30 nm between two water oxygen atoms and 0.40 nm between oxygen and bromide. The maximum HB angle is defined to 30°.11 The EM bilayers show very similar configurations of the DODAB lipids but striking differences of the adsorbed water. The DODAB alkyl chains of EM−OW1 and EM−OW2 have ⟨SCD⟩ values of 0.109 and 0.117, respectively. Sθ values are 0.390 and 0.383 corresponding to chain tilt angles of 39.6° and 39.9°, respectively. These Sθ values for simulated DODAB lipids agree roughly with a previous experimental study61 and are consistent with our measurements (cf. section 3.1). Detailed results on the structural organization of water bound to both EM bilayers are given in the Supporting Information. MD simulations at 300 K and 1 bar were set up using both EM bilayers to obtain a more realistic cationic bilayer model. 4717


Langmuir

Article

Markers for the equilibration of simulated membranes are weak time-dependent fluctuations of the average area per lipid and stable Coulomb interaction energies.81 Average lipid areas for 100 ns simulation times were extracted from trajectories using both EM bilayers. The extent of the molecular area fluctuations was independent of the starting bilayer (cf. Supporting Information, Figure S11) and similar to previous studies.26,47 Equally, Coulomb interaction energies for water and bromide ions dependent on simulation times were monitored, which differed by ∼4 kJ mol−1 at the beginning. Convergence occurs at times t > 0.5 ns (cf. Supporting Information, Figure S12). Both interaction energies slightly fluctuate at ∼−8.0 kJ mol−1 at all simulation times t > 0.5 ps. Thus, we assume equilibrium of DODAB bilayers at a simulation time of 100 ns independent of the EM bilayer used as the start configuration. Simulations of 100 ns length were continued for 1 ns at higher sampling rate (50 fs) under otherwise unchanged conditions. Translational and orientational motion of the simulated SPC water was analyzed from trajectory parts of 200 ps length and compared with reference MD simulations of bulk SPC water performed under the same conditions. The results for bulk SPC water agree very well with earlier studies.55

Figure 4. Snapshot of a DODAB bilayer MD simulation at T = 300 K and p = 1 bar after 100.2 ns. View along the bilayer plane (b axis pointing into the plane of the picture): (A) entire lipids within the bilayer, (B) hydrophilic residues including water, (C) detailed view on water molecules and bromide ions highlighting the HBs between these species (other atoms omitted, different angle of view).

octadecyl chains is kinked, whereas the major part of the alkyl chains is an all-trans conformation as found in the gel phase. Chain segment order parameters ⟨SCD⟩ were averaged over all methylene groups and all frames recorded for a 200 ps trajectory. We found a value of ⟨SCD⟩ = 0.098 corresponding to an averaged CH2 orientation angle ⟨θCD⟩ of 50.83°, which implies an average chain tilt angle ⟨θ⟩ of 39.17° that relates to an IR order parameter Sθ = 0.402. For the octadecyl chains of DODAB, this agrees very well with our measured Sθ. A small deviation can result from disorder of the investigated bilayer stacks due to imperfections in the experimental preparation. The orientation angle distributions of θμ, θHH, and θOH of water within a DODAB bilayer collected from a 200 ps trajectory are shown in Figure 5. They significantly differ from results for bulk SPC water shown as a reference within the same figure. The observed steady and continuous distribution of angles clearly displays that the chosen time interval was large enough to achieve good statistics. It is evident, especially from the small angle region, that the amount of randomly oriented water in simulated cationic membranes is negligible. The θHH angle distribution exhibits a maximum at ∼49°. The θOH distribution has a broader shape with a maximum at ∼60°, whereas the θμ angle distribution is relatively broad and without a clear maximum. The peak-like shape of the θHH angle distribution suggests a water population with reduced rotational freedom, which can originate from its different interactions, especially different H bonding.1,11 If a single water molecule is considered as HB donor only, there are five H bonded water species possible within DODAB. Three DD water species exist, which form HBs to two bromide ions (denoted as BWB), to two different water molecules (OWO, because water oxygen acts as acceptor) and to a bromide ion and another water (BWO). Further, there are two possible single D water species with water H bonded to one bromide (BW) or to another water molecule (OW). A sixth water species is represented by water (W) with two free protons. Each of these six water species can accept up to two water hydrogens by its oxygen. Evaluation of water and bromide positions collected from 200 ps trajectories yielded that (65.0 ± 5)% of the entire DODAB bilayer water belong to BWB. WB ((40 ± 1)%) molecules act as single acceptors (A) and only 0.1% as DA. Hence, BWB water molecules form on average 2.4 HBs. BWO formed the second largest fraction comprising (25 ± 3)% of water within DODAB. Only 6% form an additional HB acting as A, giving rise to an average number of 2.06 HBs. BW

Figure 3. MSD vs t plot for bulk and membrane-associated water. (Inset) Evidence for confined diffusion of the DODAB (SPC) water.

Figure 3 shows the mean-square displacement (MSD) as a function of time for both water in DODAB bilayers and the bulk water. Slow diffusion of DODAB-associated water is clearly evident by its almost constant MSD(t). The inset of Figure 3 shows the MSD’s saturation within short times, indicating confined diffusion of water within DODAB. On longer time scales, the MSD of DODAB water grows only slightly compared to bulk water, which is reflected by a small diffusion coefficient D (D = (1.35 ± 0.74) × 10−7 cm2 s−1). In contrast, for bulk SPC water a much larger D (D = (4.02 ± 0.05) × 10−5 cm2 s−1) is found. The orientation correlation function of SPC water in DODAB bilayers decays much slower than that of bulk water (cf. Supporting Information, Figure S13), indicating also reduced rotational motion of water within DODAB. Experimental and simulation studies of waterdepleted phospholipid membranes agree well with our observations.23,27,48 Detailed analysis of simulation results may provide structural insights on the behavior of water in DODAB bilayers. 3.4.2. Hydrogen Bonding and Orientation of DODAB Water. Before giving details regarding the water molecules, the structure of the lipid octadecyl chains obtained from MD simulations is briefly described. A snapshot of a DODAB bilayer’s MD trajectory after 100.2 ns simulation time is shown in Figure 4. The lipid structure resembles the findings for EM bilayers (cf. Supporting Information, Figure S10). One of the 4718


Langmuir

Article

Figure 5. Orientation angle distributions of the HH vectors defined by the hydrogen atoms (A), OH bonds (B), and dipole moments (C) of SPC water in DODAB for a trajectory of 200 ps length: (crosses) data of randomly oriented bulk SPC water; (blue) orientations of all DODAB water molecules. Water acts as DD to bromide (BWB, yellow), DD to water and bromide (BWO, gray), and single D to bromide (BW, red). BWO and BW are both split for the OH vectors into a gray (bound to bromide) and a white (bound to O of water) component for BWO and components due to two free and bonded OH groups for BW, respectively. The three components include 96.5% of the entire DODAB water.

Table 2. Mean Orientations and Order Parameters (in parentheses) of SPC Water in Simulated DODAB Bilayersa water fraction

⟨θHH⟩/deg (SHH)

all water (average) BWB

55.3° (−0.01) 49.0 (0.15) 48.4 (0.16) 49.6 (0.13) 69.3 (−0.31) 70.1 (−0.33) 69.3 (−0.31) 57.1 (−0.06) 56.7 (−0.05) 58.1 (−0.08)

BWO

BW

total DD ADD total DD ADD total D AD

⟨θOH⟩/deg (SOH) 60.0 56.3 56.1 56.4 65.3

(−0.13) (−0.04) (−0.03) (−0.04) (−0.24)⊖; 72.3 (−0.36)‡

53.2 (0.04)c; 68.1(−0.29)b

⟨θμ⟩/deg (Sμ)

molecular fraction/%

69.8° (−0.32) 72.2 (−0.36) 70.9 (−0.34) 70.3 (−0.33) 66.0 (−0.25) 66.0 (−0.25) 67.2 (−0.28) 66.6 (−0.26) 65.8 (−0.25) 69.4 (−0.31)

100 65 39 26 25 23.5 1.5 6.5 4.8 1.7

a

The major fractions (96.5%) are BWB (DD to two bromides), BWO (DD to water and bromide), and BW (single D to bromide). Subfractions are defined by nonacceptance/acceptance of another HB. Orientation angles of two differently bonded OH bonds (defined by the vector from O to H) were separately analyzed. ⊖OH groups bonded to water oxygen. bOH bonded to bromide. cFree OH groups.

comprises (6.4 ± 1.5)% of DODAB water. A proton of a nearby water molecule is accepted by 25%, giving an average number of 1.25 HBs. The remaining 3.5% of DODAB water molecules are split into HB D fractions comprising water with a single water molecule (2.5 ± 0.3%) with two water molecules (0.5 ± 0.2%) and free water molecules (0.5 ± 0.2%). Further details on these species are omitted due to their small contribution to the entire water population. The coordinates of the three major SPC water fractions were analyzed to obtain molecular orientations that are summarized in Table 2. BWB. BWB contributes a Gaussian-shaped, large component at ∼50° to the θHH angle distribution (Figure 5) and a similar broader shape centered at ∼55° to the θOH angle distribution. Its contribution to the total θμ angle distribution has the broadest width and cannot be separated from other contributions. The uniform shapes of the BWB contributions, especially those of the θHH and θOH angle distributions, suggest a homogeneous water population. BWB water molecules with no (DD) and with one accepted HB (ADD) were found to differ only slightly in their mean orientation angles and order parameters (cf. Table 2); thus, an additional HB does not significantly modify the rotational freedom of water. Oxygen atoms comparably less involved in interactions can explain a greater rotational mobility of water dipoles, which agrees with the increased widths of the θμ and θOH angle contributions of BWB water. BWO. BWO molecules also contribute a Gaussian-like component with a peak at ∼71° to the θHH angle distribution (Figure 5). The contributions from water−oxygen-bonded and

bromide-bonded OH groups appear with different shapes. However, the associated mean orientation angles can be judged as similar (Table 2). The θμ angle contribution of BWO is similar to the entire θμ distribution and those of BWB and BW regarding width and maximum, which is reflected by comparable mean orientation angles θμ. Since 94% of BWO were found to act as DD and only 6% as ADD, the ADD subfraction contributes very weakly to the entire water behavior. BW. Contributions of BW water are similarly broad as BWO (Figure 5). The maximum of the θHH angle contribution appears at ∼50° similar to that of BWB. However, it is the broadest of the three major water fractions. There are two θOH angle contributions from bromide-bonded and free OH groups which occur very broad, without a well-defined maximum and differ in their mean orientation angles. The θμ angle contribution is similar to those from BWB or BWO, but all of the orientation angle contributions from BW are broader, reflecting a higher degree of rotational freedom within BW. Probably BW water molecules are only bonded at one of their OH groups, whereas the other is free. The separation into D (75%) and AD (25%) subfractions causes a similar broadening. The differences of mean orientation angles θμ, θHH, and θOH between the three subfractions are the largest found. Thus, in contrast to BWB and BWO, there is an effect of an additional HB on rotational motion for a D water molecule. Another accepted HB has only a visible effect on angle distributions if the molecular rotation was not already restricted by double H bonding of the water OH groups. 4719


Langmuir

Article

corresponding SOH value of OH bond directors. For the D2O stretch components the relative intensities roughly agree with the relative sizes of the modeled DDA and DD BWB fractions. However, the order parameters SIR and SOH show a stronger deviation between experiment and simulation. The behavior of the D2O stretching modes within these membranes can differ from H2O. A generic correlation between average orientations of simulated water fractions and experimentally observed average orientations of all water populations was not found, possibly due to molecular coupling. Water molecules in DODAB samples cannot be considered isolated, since H bonding to one neighbored water molecule occurs for about 50−60%. Coupled oscillations have dipole moments different from those of the underlying molecular fundamentals. The difference between experiment and simulation regarding average orientations of DDA BWB molecules is significantly larger than observed for the DD BWB. While the latter might approximate uncoupled OH stretch oscillators, increased intermolecular coupling in DDA molecules is assumed to induce stronger deviation from the fundamental. In synopsis, by controlling the degree of hydration and through counterion exchange the unusually strong structured water in the proximity of the model lipid membranes was revealed, demonstrating the dramatic extent to which biological assemblies can influence the surrounding water layer, which in turn influences subsequent interfacial events.

Analysis of the HBs revealed a broad continuum of bond lengths covering an interval with a width of 0.1 nm. The most frequent HB lengths determined for water−bromide HBs and for two connected water molecules occurred at about 0.33 and 0.29 nm, respectively. HB angles of bonds with bromide and water are broadly distributed and have their maximal occurrence at 9° and 10°, respectively.

4. CONCLUDING REMARKS In this paper we report on the interaction of water molecules with cationic DODAX (X = F, Cl, Br, I) lipid membranes as studied by polarized ATR FTIR spectroscopy and MD simulations of DODAB bilayers. The thermotropic phase behavior of DODAC and DODAB samples found agrees with previous experimental observations.35,37,39 Furthermore, we detected hydration-dependent changes of DODAX headgroup absorptions directly pointing to headgroup−water interactions. The lipids’ water capacity, as monitored by the band intensity, was demonstrated to be strongly altered by exchange of halide anions. The OH stretch absorptions were analyzed at equal sample hydration (∼1 water/lipid), which is a very low hydration compared to phospholipids under similar conditions.19 OH/OD stretch bands of water within DODAB and DODAC exhibit striking band shapes, which have not been reported for other water/lipid interfaces so far. Common peak wavenumbers were found for water at different DODAX membranes, especially for DODAB and DODAC. Our IR data demonstrate anisotropic orientations of differently H bonded water fractions for almost all subbands throughout the different lipid samples. Furthermore, component frequencies of DODAX samples are comparable with IR spectroscopic data from small water−halide clusters. This suggests the presence of cluster-like structures within DODAX membranes forming similar types of HBs including DD and DDA IHBs. Moreover, the frequencies reported in this study resemble findings earlier reported for uncoupled OH and OD oscillators within aqueous halide solutions. Simulations of DODAB bilayers yielded lipid chain tilt angles, agreeing well with our experimental results. MD simulations revealed slow translational and rotational diffusion of DODAB-associated SPC−water compared to simulated bulk water as reported for water at low-hydrated phospholipids.23,27,48 Three major water fractions (BWB, BWO, and BW) were found, where water acts as HB DD to bromide, DD to bromide and water, and single D to bromide, respectively. To relate experiments with MD simulations, average orientations of the differently H-bonded water fractions were computed from trajectories of DODAB−water membranes. Angle distributions of water’s dipole moments, HH vectors, and OH vectors, which are correlated to fundamental water stretching modes, clearly indicate an anisotropic water orientation with different H bonding. We conclude that water molecules in DODAB are rotationally confined by varied H bonding to bromide ions, resulting in differently oriented populations, which is in good qualitative agreement with the spectroscopic measurements. The most intense stretching band components of DODAB water at 3394 and 3444 cm−1 are loosely assigned to DDA and DD subpopulations of the BWB water fraction, which is the largest found in simulated DODAB. Although the fractional sizes of these subpopulations do not agree with the relative intensities of the band components, there is a good accordance between the experimental order parameter SIR and the

■

ASSOCIATED CONTENT

S Supporting Information *

Further experimental details; MALDI-TOF mass spectra, micro-XRF spectra, H NMR, and F NMR spectra of the lipids; results on the phase behavior, structural changes of the headgroup of DODAX lipids, and structural organization of SPC water bound to simulated EM DODAB bilayers; graphs of the simulated temporal fluctuations of the area per DODAB lipid, time dependence of the Coulomb interaction energies for both EM bilayers, and rotation correlation function of SPC water in DODAB and bulk liquid. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (0049)341 9732713. Fax: (0049)341 9732479. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the Deutsche Forschungsgemeinschaft (FOR877). The authors thank J. Schiller for the MALDI-TOF, M. Findeisen for the NMR, and J. Lenzner for the XRF measurements that were performed to characterize the substituted DODAX lipids and B. Kohlstrunk for technical support. L.W. thanks the Graduate School InterNeuro (GK1097), and D.W.B. acknowledges the American Heart Association BGIA Program for financial support.

■

REFERENCES

(1) Ball, P. Nature 2008, 452, 291−292. (2) Bakker, H. J.; Skinner, J. L. Chem. Rev. 2010, 110, 1498−1517. (3) Ball, P. Chem. Rev. 2008, 108, 74−108.

4720


Langmuir

Article

(4) Paesani, F.; Xantheas, S. S.; Voth, G. A. J. Phys. Chem. B 2009, 113, 13118−13130. (5) Wernet, P.; Nordlund, D.; Bergmann, U.; Cavalleri, M.; Odelius, M.; Ogasawara, H.; La Naslund; Hirsch, T. K.; Ojamae, L.; Glatzel, P.; Pettersson, L. G.; Nilsson, A. Science 2004, 304, 995−999. (6) Bertie, J. E.; Ahmed, M. K.; Eysel, H. H. J. Phys. Chem. 1989, 93, 2210−2218. (7) Brubach, J. B.; Mermet, A.; Filabozzi, A.; Gerschel, A.; Roy, P. J. Chem. Phys. 2005, 122. (8) Liu, D. F.; Ma, G.; Levering, L. M.; Allen, H. C. J. Phys. Chem. B 2004, 108, 2252−2260. (9) Binder, H. Eur. Biophys. J. Biophys. Lett. 2007, 36, 265−279. (10) Binder, H. Appl. Spectrosc. Rev. 2003, 38, 15−69. (11) Auer, B. M.; Skinner, J. L. Chem. Phys. Lett. 2009, 470, 13−20. (12) Sovago, M.; Campen, R. K.; Wurpel, G. W. H.; Muller, M.; Bakker, H. J.; Bonn, M. Phys. Rev. Lett. 2008, 100. (13) Bergström, P. A., Lindgren, J. Inorg. Chem. 1992, 31, 1529− 1533. (14) Bhattacharyya, K. Chem. Commun. 2008, 2848−2857. (15) Bonn, M.; Campen, R. K. Surf. Sci. 2009, 603, 1945−1952. (16) Bechinger, B.; Seelig. J. Chem. Phys. Lipids 1991, 58, 1−5. (17) Gawrisch, K.; Ruston, D.; Zimmerberg, J.; Parsegian, V. A.; Rand, R. P.; Fuller, N. Biophys. J. 1992, 61, 1213−1223. (18) McMullen, T. P. W; Lewis, R. N. A. H.; McElhaney, R. N. Biochemistry 1993, 32, 516−522. (19) Nagle, J. F.; Tristram-Nagle, S. Biochim. Biophys. Acta, Rev. Biomembr. 2000, 1469, 159−195. (20) Lewis, R. N. A. H., McElhaney, R. N. In Methods in Membrane Lipids; Dopico, A., Ed.; Humana Press Inc.: Totowa, NJ, 2007; Vol. 400, pp 207−226. (21) Mantsch, H. H.; McElhaney, R. N. Chem. Phys. Lipids 1991, 57, 213−226. (22) Tielrooij, K. J.; Paparo, D.; Piatkowski, L.; Bakker, H. J.; Bonn, M. Biophys. J. 2009, 97, 2484−2492. (23) Zhao, W.; Moilanen, D. E.; Fenn, E. E.; Fayer, M. D. J. Am. Chem. Soc. 2008, 130, 13927−13937. (24) Volkov, V. V.; Takaoka, Y.; Righini, R. J. Phys. Chem. B 2009, 113, 4119−4124. (25) PasenkiewiczGierula, M.; Takaoka, Y.; Miyagawa, H.; Kitamura, K.; Kusumi, A. J. Phys. Chem. A 1997, 101, 3677−3691. (26) Siu, S. W.; Vacha, R.; Jungwirth, P.; Böckmann, R. A. J. Chem. Phys. 2008, 128. (27) Volkov, V. V.; Palmer, D. J.; Righini, R. J. Chem. Phys. B 2007, 111, 1377−1383. (28) Bonn, M.; Bakker, H. J.; Ghosh, A.; Yamamoto, S.; Sovago, M.; Campen, R. K. J. Am. Chem. Soc. 2010, 132, 14971−14978. (29) Felgner, P. L.; Gadek, T. R.; Holm, M.; Roman, R.; Chan, H. W.; Wenz, M.; Northrop, J. P.; Ringold, G. M.; Danielsen, M. Proc. Natl. Acad. Sci., U.S.A. 1987, 84, 7413−7417. (30) Zuhorn, I. S.; Engberts, J. B. F. N.; Hoekstra, D. Eur. Biophys. J. 2007, 36, 349−362. (31) Rädler, J. O.; Koltover, I.; Salditt, T.; Safinya, C. R. Science 1997, 275, 810−814. (32) Ewert, K. K.; Evans, H. M.; Zidovska, A.; Bouxsein, N. F.; Ahmad, A.; Safinya, C. R. J. Am. Chem. Soc. 2006, 128, 3998−4006. (33) Rodriguez-Pulido, A.; Martin-Molina, A.; Rodriguez-Beas, C.; Llorca, O.; Aicart, E.; Junquera, E. J. Phys. Chem. B 2009, 113, 15648− 15661. (34) Kunitake, T.; Okahata, Y.; Tamaki, K.; Kumamaru, F.; Takayanagi, M. Chem. Lett. 1977, 387−390. (35) Umemura, J.; Kawai, T.; Takenaka, T.; Kodama, M.; Ogawa, Y.; Seki, S. Mol. Cryst. Liq. Cryst. 1984, 112, 293−309. (36) Kawai, T.; Umemura, J.; Takenaka, T.; Kodama, M.; Ogawa, Y.; Seki, S. Langmuir 1986, 2, 739−743. (37) Laughlin, R. G.; Munyon, R. L.; Fu, Y. C.; Fehl, A. J. J. Phys. Chem. 1990, 94, 2546−2552. (38) Schulz, P. C.; Rodriguez, J. L.; Soltero-Martinez, F. A.; Puig, J. E.; Proverbio, Z. E. J. Therm. Anal. Cal. 1998, 51, 49−62.

(39) Feitosa, E.; Barreleiro, P. C. A.; Olofsson, G. Chem. Phys. Lipids 2000, 105, 201−213. (40) Saveyn, P.; van der Meeren, P.; Zackrisson, M.; Narayanan, T.; Olsson, U. Soft Matter 2009, 5, 1735−1742. (41) Wu, F. G.; Wang, N. N.; Yu, Z. W. Langmuir 2009, 25, 13394− 13401. (42) Hofmeister, F. Arch. Exp. Pathol. Pharmakol. 1888, 24, 247−260. (43) Zhang, Y. J.; Cremer, P. S. Curr. Opin. Chem. Biol. 2006, 10, 658−663. (44) Harrick, N. J. Internal reflection spectroscopy; John Wiley & Sons Inc.: New York, 1967. (45) Goormaghtigh, E.; Raussens, V.; Ruysschaert, J. M. Biochim. Biophys. Acta, Rev. Biomembr. 1999, 1422, 105−185. (46) Egberts, E.; Marrink, S. J.; Berendsen, H. J. C. Eur. Biophys. J. Biophys. Lett. 1994, 22, 423−436. (47) Berger, O.; Edholm, O.; Jahnig, F. Biophys. J. 1997, 72, 2002− 2013. (48) Berkowitz, M. L.; Bostick, D. L.; Pandit, S. Chem. Rev. 2006, 106, 1527−1539. (49) Marrink, S. J.; Vries, A. H.; de, Tieleman, D. P. Biochim. Biophys. Acta, Rev. Biomembr. 2009, 1788, 149−168. (50) Okuyama, K.; Soboi, Y.; Iijima, N.; Hirabayashi, K.; Kunitake, T.; Kajiyama, T. B. Chem. Soc. Jpn. 1988, 61, 1485−1490. (51) Picard, F.; Buffeteau, T.; Desbat, B.; Auger, M.; Pezolet, M. Biophys. J. 1999, 76, 539−551. (52) Binder, H.; Schmiedel, H. Vib. Spectrosc. 1999, 21, 51−73. (53) Binder, H.; Gawrisch, K. J. Phys. Chem. B 2001, 105, 12378− 12390. (54) Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F., Hermans, J. In Intermolecular Forces. Proceedings; Pullman, A., Ed.; D. Reidel Publishing Co.: Dordrecht, 1981; Vol. 14, pp 331−342. (55) van der Spoel, D.; van Maaren, P. J.; Berendsen, H. J. J. Chem. Phys. 1998, 108, 10220−10230. (56) Jorgensen, W. L.; Tirado-Rives, J. J. Am. Chem. Soc. 1988, 110, 1657−1666. (57) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. J. Comput. Chem. 1997, 18, 1463−1472. (58) Miyamoto, S.; Kollman, P. A. J. Comput. Chem. 1992, 13, 952− 962. (59) Jamroz, D.; Kepczynski, M.; Nowakowska, M. Langmuir 2010, 26, 15076−15079. (60) Murzyn, K.; Zhao, W.; Karttunen, M.; Kurdziel, M.; Rog, T. Biointerphases 2006, 1, 98−105. (61) Ionov, R.; Angelova, A. Thin Solid Films 1996, 284−285, 809− 812. (62) Pohle, W.; Selle, C. Chem. Phys. Lipids 1996, 82, 191−198. (63) Kint, S.; Wermer, P. H.; Scherer, J. R. J. Phys. Chem. 1992, 96, 446−452. (64) Gauger, D. R.; Selle, C.; Hahn, M.; Pohle, W. Anal. Biochem. 2001, 299, 108−110. (65) Lamoureux, G.; Roux, B. J. Phys. Chem. B 2006, 110, 3308− 3322. (66) Britt, D. W.; Goodman, T.; Selle, C. Materialwiss. Werkstofftech. 2003, 34, 1133−1137. (67) Sammon, C.; Mura, C.; Yarwood, J.; Everall, N.; Swart, R.; Hodge, D. J. Phys. Chem. B 1998, 102, 3402−3411. (68) Venyaminov, S. Y.; Prendergast, F. G. U. Anal. Biochem. 1997, 248, 234−245. (69) Paesani, F.; Voth, G. A. J. Phys. Chem. B 2009, 113, 5702−5719. (70) Riemenschneider, J.; Wulf, A.; Ludwig, R. Z. Phys. Chem. 2009, 223, 1011−1022. (71) Bergström, P. A.; Lindgren, J.; Kristiansson, O. J. Phys. Chem. 1991, 95, 8575−8580. (72) Nickolov, Z. S.; Miller, J. D. J. Colloid Interface Sci. 2005, 287, 572−580. (73) Roscioli, J. R.; Diken, E. G.; Johnson, M. A.; Horvath, S.; McCoy, A. B. J. Phys. Chem. A 2006, 110, 4943−4952. (74) Cabarcos, O. M.; Weinheimer, C. J.; Lisy, J. M.; Xantheas, S. S. J. Chem. Phys. 1999, 110, 5−8. 4721


Langmuir

Article

(75) Robertson, W. H.; Johnson, M. A. Annu. Rev. Phys. Chem. 2003, 54, 173−213. (76) Choi, J. H.; Kuwata, K. T.; Cao, Y. B.; Okumura, M. J. Phys. Chem. A 1998, 102, 503−507. (77) Ayotte, P.; Nielsen, S. B.; Weddle, G. H.; Johnson, M. A.; Xantheas, S. S. J. Phys. Chem. A 1999, 103, 10665−10669. (78) Ayotte, P.; Bailey, C. G.; Weddle, G. H.; Johnson, M. A. J. Phys. Chem. A 1998, 102, 3067−3071. (79) Fringeli, U. P.; Günthard, H. H. Mol. Biol., Biochem. Biophys. 1981, 31, 270−332. (80) Seelig, J. Q. Rev. Biophys. 1977, 10, 353−418. (81) Kandt, C.; Ash, W. L.; Tieleman, D. P. Methods 2007, 41, 475− 488.

4722


Oriented Confined Water Induced by Cationic Lipids - Langmuir (ACS

Recommend Documents