DOI: 10.1021/la902916y 2665Langmuir 2010, 26(4), 2665–2670 Published on Web 11/03/2009

pubs.acs.org/Langmuir

© 2009 American Chemical Society

Study on the Correlation between Lateral Diffusion Effect and Effective

Charge in Neutral Liposomes

Elisa Galera-Cort�es,† Juan de Dios Solier,† Joan Estelrich,‡ and Roque Hidalgo-�Alvarez*,§

†Departamento de Fı́sica Aplicada, Universidad de Extremadura, Avda. de Elvas s/n�, 06071 Badajoz, Spain,‡Departament de Fisicoquı́mica, Facultat de Farm�acia, Universitat de Barcelona, Joan XXIII s/n�, 08028Barcelona, Spain, and §Departamento de Fı́sica Aplicada, Universidad de Granada, Campus Fuentenueva,

18071 Granada, Spain

Received August 7, 2009. Revised Manuscript Received October 16, 2009

An experimental investigation is described on the variables that affect the lateral diffusion coefficient (Dlat) ofdimyristoylphosphatidylcholine, a zwitterionic phospholipid, and the effective charge (Zef) on liposomes. The lateraldiffusion coefficient was obtained from the dielectric relaxation time of the zwitterionic phospholipids in the bilayer, andthe effective charge on the external monolayer was estimated from microelectrophoretic mobility measurements bymeans of the Henry and Coulomb equations. The measurements were performed at different pH values and salt(KBr) concentrations as well as in two physical states of the phospholipid: the liquid-crystalline phase and gel phase. TheZef and Dlat values in the gel phase are always lower than those in the fluid phase. A very small change of pH (∼0.5 pHunits) caused a pronounced variation of the effective charge and the lateral diffusion coefficient. Both variations arecorrelated, which demonstrates that the adsorption of the ions that determine the electrokinetic potential also controlsthe lateral diffusion of dipolar phospholipids in the bilayer and the effective charge on the external surface of theliposomes.

1. Introduction

Liposomes are lipid structures used as a model of biologicalmembranes. They can encapsulate biological molecules, such asproteins, enzymes, or drugs. Lipids used in the preparation ofliposomes are predominantly phospholipids (PLs), the samecomponents also found in biological membranes. Depending onthe processing conditions, liposomes are formed with one con-centric bilayer (the so-called unilamellar liposomes) or withseveral concentric bilayers (multilamellar liposomes).

The principal barrier to permeation in biological membranesis the lipid bilayer, and the lateral diffusion inside the bilayer isa process of importance for the diffusion of their componentsin the biological membranes, the lateral diffusion coefficientbeing one of the parameters that inform us about the dynamicstate of the membrane.1 Therefore, it is of great interest toknow which factors affect the diffusion of phospholipids intheir bilayer. This diffusion has fundamental implications infunctional coupling between membrane components throughcollisional mechanisms as, for example, in (a) visual transduc-tion, the process by which light initiates a nerve impulse;2 (b)receptor-mediated endocytosis, a process by which moleculesare internalized into a cell (endocytosis) by the inward buddingof plasmamembrane vesicles containing proteins with receptorsites specific to the molecules being internalized;3 and (c)intercellular adhesion molecules, which promote adhesionamong cells, for example, the adhesion of most white bloodcells, related to their immunological response to wound orbacterial infection.4

Lateral diffusion of phospholipids in membranes has beenstudied experimentally over the years by a variety of methods:fluorescence recovery after photobleaching (FRAP),5-7 electronspin resonance (ESR),8 nuclear magnetic resonance (NMR),9-11

and quasielastic neutron scattering (QENS).12,13 Lateral diffusionis also determined by dielectric spectroscopy.14,15 This determina-tion is possible due to the correlation existing between thetranslational diffusion process and the rotational relaxation ofthe phospholipids in the bilayer.14

In a previous study the dielectric spectrum of charged lipo-somes was analyzed.15 The liposomes were prepared with ananionic phospholipid (phosphatidylserine) and a zwitterionicphospholipid (phosphatidylcholine) at a molar ratio of 99:1,respectively. In that study, it was observed that the lateraldiffusion in the bilayer decreased with the ionic strength. In thepresent work, we have studied the electric response of liposomescomposed exclusively of a zwitterionic phospholipid, the dimyri-stoylphosphatidylcholine (DMPC), which possesses the commonchemical structure of a phospholipid: a phosphate group and acholine group as the headgroup. DMPC presents a transitiontemperature of 23.5 �C, which implies that, as a function of

*To whom correspondence should be addressed. E-mail: rhidalgo@ugr.es.(1) Tocanne, J. F.; Dupou-C�ezanne, L.; Lopez, A.Prog. Lipid Res. 1994, 33, 203.(2) Lamb, T. D. Biophys. J. 1994, 67, 1439.(3) Schlessinger, J. Biopolymers 1993, 22, 47.(4) Leckband, D. E.; Israelachvili, J. N.; Schmitt, F. J.; Knoll., W. Science 1992,

255, 1419–1421.

(5) Vaz, W. L. C.; Clegg, R. M.; Hallmann, D. Biochemistry 1985, 24, 78.(6) Rubenstein, J. L. R.; Smith, B. A.; McConnel, H. M. Proc. Natl. Acad. Sci.

U.S.A. 1979, 76, 15.(7) Merkel, R.; Sackmann, E.; Evans, E. J. Phys. France 1989, 50, 1535.(8) Devaux, P. F.; McConnel, H. M. J. Am. Chem. Soc. 1972, 94, 4475.(9) Bloom, M.; Burnell, E. E.; Mackay, A. L.; Nicol, C. P.; Valic, M. I.; Weeks,

G. Biochemistry 1978, 17, 5750.(10) Lindblomm, G.; Johansson, L. B. A.; Arvidson, G. Biochemistry 1981, 20,

2204.(11) Kuo, A. L.; Wade, C. G. Biochemistry 1979, 18, 2300.(12) Tabony, J.; Perly, B. Biochim. Biophys. Acta 1990, 1063, 67.(13) K€onig, S.; Pfeiffer, W.; Bayerl, T.M.; Richter, D.; Sackmann, E. J. Phys. II

1992, 2, 1598.(14) Haibel, A.; Nimtz, G.; Pelster, R.; Jaggi, R. Phys. Rev. E 1998, 57, 4838.(15) Solier, J. D.; Galera-Cort�es, E.; Sabat�e, R.; Estelrich, J. Colloids Surf., A

2005, 270-271, 88.

2666 DOI: 10.1021/la902916y Langmuir 2010, 26(4), 2665–2670

Article Galera-Cort�es et al.

temperature, we can have liposomes in a fluid phase (the crystal-liquid state) or in a rigid phase (the gel state). In these liposomes,we have analyzed the effect of the pH and the concentration ofcounterions in the bulk on the lateral diffusion of DMPC in thebilayer, when the phospholipid was in each one of the physicalstates mentioned above.

The aim of this work was to investigate a possible relationshipbetween the effective charge on the external surface of theliposomes and the lateral diffusion of the zwitterionic phospho-lipids, using DMPC as a model. Moreover, the effect of therigidity of the bilayer was also studied by performing experimentswell below and above the crystal-liquid-gel transition.

2. Materials and Methods

2.1. Materials. Dimyristoylphosphatidylcholine (DMPC)was purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.)and used without further purification. The cation- and anion-exchange resin Amberlite IRN-150 was from Supelco (Bellefonte,PA, U.S.A.). Organic solvents of ACS grade (methanol andchloroform) were obtained from Merck (Darmstadt, Germany)and usedwithout further purification.All inorganic reagents wereof analytical grade and aqueous solutions were prepared withdoubly distilled water.2.2. Methods. 2.2.1. Liposome Preparation. A total of

60 μmol of DMPC was dissolved in CHCl3/CH3OH (2:1, v/v),placed in a round-bottom flask, and dried in a rotary evaporatorunder reduced pressure at 40 �C to form a thin film on the innersurface of the flask. This film was hydrated with 2 mL of a 1 mMKBr solution to give a lipid concentration of 30 mM. Multi-lamellar vesicles (MLVs) were initially formed by constantvortexing for 4 min on a vortex mixer, followed by sonicationin a Ultrasonic Digitals bath sonifier (Elma, Germany) for10 min. MLVs were downsized to form oligolamellar vesicles byextrusion at 40 �C in a extruder device (Lipex Biomembranes,Canada) through polycarbonate membrane filters of variablepore size under nitrogen pressures up to 55 105 Pa.16 Briefly,liposomes were extruded in three steps: first, three consecutiveextrusions through a 0.8 μm pore diameter filter and three otherconsecutive extrusions through 0.4 μmmembranes. The resultinglipid suspension was then extruded three consecutive timesthrough 0.1 μm filters. The extrusion method has been demon-strated to produce unilamellar vesicles,17 and extruded DMPCliposomes are commonly used as a model of unilamellar vesiclesfor several biophysical studies.18 By means of freeze-fractureelectronic microscopy, these vesicles were shown to be spherical(see the Supporting Information, Figure SI1). As the headgroupof the DMPC is zwitterionic, at nonextreme pH values, theobtained liposomes have no net charge on their surface.

2.2.2. Characterization of the Liposomes. Particle size dis-tribution was determined at 25 �C by photon correlation spec-troscopy with a commercial light-scattering setup (ZetasizerNano ZS90, Malvern, U.K.) using a 5 mW He-Ne laser. Forviscosity and the refractive index, water values were used. Tomeasure the particle size distribution of the dispersion, a poly-dispersity index, ranging from 0.0 for an entirely monodispersesample up to 1.0 for a polydispersity sample, was used. It isgenerally accepted that a liposome sample with a polydispersityindex smaller than 0.2 can be considered practically a mono-disperse sample.

The volume fraction of the samples was determined from themass fraction using the equation proposed by Haro-P�erez et al.19

φ ¼ x4πða3 -R3

1Þ

3F04πa3

3ð1Þ

where a is the outer liposome radius, R1 = a - Δ is the innerradius, Δ is the thickness of the phospholipid shell (4.5 nm), x isthe phospholipid weight fraction used in each synthesis, and F0 isthe density of the phospholipid shell on the colloidal particle(1.015 g/cm3).

The liposomeswere dispersed in a 1mMKBr solution andwerekept over a bed of the Amberlite IRN-150 (0.1 mg of resin by mLof liposomal suspension) for different times. The particle sizedistribution of the liposomes was determined before (0 h) andafter mixing with the resins (24, 96, and 168 h). It was observedthat the contact of the liposomes with the resins did not affecteither the size or the polydispersity index of the vesicles, andtherefore, the vesicles can be considered stable in the colloidalsense.

The content of potassium in the vesicles incubated fordifferent times in resins was determined using an UNICAMPU 939 flame absorption spectrometer equipped with an acetyl-ene/air (1:1) burner and a selenium hollow cathode lamp(Photron) that was operated at 776.5 and 0.5 nm band pass. Areference potassium solution (0.5 μg/mL) was prepared bydilution of a 1000 μg/mL solution (VWR, Titrisol, Darmstadt,Germany) with doubly distilled water containing 1% (v/v) ofnitric acid (VWR, Titrisol, Darmstadt, Germany). Over theconcentration range of 0.1-0.8 μg/mL, the measurement ofpotassium was linear (R2 > 0.999). The reproducibility of theassaywas determined by repeatedmeasurements of the referencesolution as part of a run or from run to run. Each value is theaverage of three replicates. The experiments of atomic absorp-tion of potassium performed with the DMPC liposomes showedthat the concentration of this cation on the liposome surfaceremains unchanged and very close to zero, after 20 h on theresins.

Table 1 shows a summary of the properties of the liposomesused in the dielectric and electrophoresis measurements.

2.2.3. Impedance Measurements. To determine the complexdielectric permittivity, impedance measurements were carried outby means of an automated procedure covering the frequencyrange from 20 Hz to1 GHz. Three different Hewlett-Packardimpedance analyzers, HP4284A (20 Hz to 1MHz), HP4285A (75kHz to 30MHz), and HP4191A (1 MHz to 1 GHz) were used. Alow ac voltage with an amplitude of 22 mV was applied inall measurements; this voltage corresponds to a field strength of∼10 nV/nm. The data were obtained using a cell designed in sucha shape and size that they allowed for its insertion in a coaxialtransmission line of very short length,and for the whole setup tobe connected directly (without wire) to the impedance analyzers.

Table 1. Physical Characteristics of Extruded DMPC Liposomes

DMPC liposome

mean size (nm) 126 ( 1outside radius (nm) 63inside radius (nm) 58.5volume fraction (φ) 0.084fluid-gel transition (�C) 23.5mean polydispersity index 0.11 ( 0.01

(16) Bally, M. B.; Hope, M. J.; Van Echteld, C. J. A.; Cullis, P. R. Biochim.Biophys. Acta 1985, 812, 55–65.(17) Armengol, X.; Estelrich, J. J. Microencapsulation 1995, 12, 525–535.(18) Ku�cerka, N.; Kiselev, M. A.; Balgav�y. Eur. Biophys. J. 2004, 33, 328–334.

(19) Haro-P�erez, C.; Quesada-P�erez, M.; Callejas-Fern�andez, J.; Casals, E.;Estelrich, J.; Hidalgo-�Alvarez, R. J. Chem. Phys. 2003, 119, 628.

DOI: 10.1021/la902916y 2667Langmuir 2010, 26(4), 2665–2670

Galera-Cort�es et al. Article

With this setup, the contribution of parasite impedance at thehighest frequencies range (1MHz to 1 GHz) is avoided and, aftersome ad-hoc calibration procedure, the resolutionof themeasure-ments is improved considerably. The experimental procedureemployed in the dielectric measurements and the cell featureshave been described elsewhere.20 The entire system (impedanceanalyzers, cell, and samples) was kept at the desired temperatureinside a thermally insulating cabinet connected at all times to acirculation ultrathermostat (Selecta Frigiterm-30, Spain), whichallows making quick measurements.

The influence of salt concentration and pH on the dielectricalresponse and on the lateral diffusion of the PLs was assessed 48 hafter the liposome samples had been produced. Subsequentmeasurements were performed after having the liposome sampleson thematrix of the ion-exchange resin for 43, 50, 72, and 113 h. Itis important, however, to clarify that the dielectric and micro-electrophoretic mobility measurements were carried out in theabsence of ion-exchange resins.

2.2.4. pH Measurements. The pH measurements were car-ried out at 19 and 35 �C with a GLP22 pHmeter (Crison, Spain).This instrument is equipped with a glass electrode and automatictemperature compensation. It presents an accuracy of ( 0.01units and was calibrated with two buffer solutions of pH 7.02 and4.00.

2.2.5. Electrophoretic Mobility Measurements. The micro-electrophoretic mobility (μe) measurements were made with aZetasizer 4 device (Malvern, U.K.) based on the laser-Dopplershift. Results were the average of four measurements. Thestandard deviation of the microelectrophoretic mobility valueswas lower than 5% of the average value. Throughout thesemeasurements, liposomes were diluted with 10-3 MKBr in orderto obtain a suspension of 10-3 M lipid concentration. Allexperiments in this study were performed in triplicate.

Similar to the dielectric measurements, the influence of saltconcentration and pH on the electrophoretic response of theliposomes was assessed 48 h after the liposome samples had beenproduced. Subsequent measurements were performed after hav-ing the liposome samples on the matrix of the ion-exchange resinfor 63, 74, and 113 h. Themeasurements were carried out at 35 �C(liquid-crystalline phase) and at 19 �C (gel phase).

The values of the ζ-potential were obtained by means of theH€uckel-Onsager equation and considering the Henry’s correc-tion factor

ζ ¼ 3ημe2ε0εr

1

f ðKaÞ ð2Þ

where εr =78.5, ε0 = 8.85418 3 10-12 F 3m

-1, and η = 8.9 3 10-3

Pa 3 s. For the liposome sample at 10-3MKBr, aHenry’s factor of1.5 was used, whereas for the deionized samples, a Henry’s factorof 1 was used. Once the ζ-potential values were obtained, theywere converted into effective charge (Zef) using the Coulombequation.

All the experiments were carried out under atmospheric con-ditions and at constant temperature.

3. Results and Discussion

Figure 1 shows the variation of pH and ionic strength of aliposome sample in the fluid phase (T>23.5 �C) as a function ofthe ion-exchanging time. Evidently, the salt concentrationdecreases with increasing time. The pH, however, decreases

monotonously in the first 72 h, and subsequently, it increasesslightly. The time dependence of the pH will be commented onbelow.

The experimental dielectric permittivity spectra are obtainedfrom impedance spectra once the electrode effect has beenremoved. The procedure used to remove the electrode effect hasbeen described elsewhere.15,20 The experimental complex dielec-tric permittivity (ε*) is then

ε/ðωÞ ¼ ε0ðωÞ- j ε00 -σdc

ω

� �ð3Þ

where ε0 and ε00 are the real and imaginary part of the complexdielectric permittivity (CDP), σdc is the dc conductivity, and ω isthe angular frequency. The real part is related to the polarizationphenomena of the particle (electric double layer, polarization ofdipolar headgroup), and the imaginary part in eq 3 is associatedwith the dissipation of energy owing to the reorientation of thosedipoles and the electric conduction. In the frequency rangeanalyzed, the dissipation term is dominated by the dc ionicconduction (σdc/ω . ε0 0) so that the relaxation peaks are notobserved.

To remove the dc-conductivity contribution to the imaginarypart of the CDP, the Kramer-Kronig integral transform forε*(ω) can be used.15,20,21

ε0ðωÞ-ε0ð¥Þ ¼ 2

π

� � Z ¥

0

xε00ðωÞx2 -ω2

dx ð4Þ

ε00ðωÞ ¼ 2

π

� � Z ¥

0

xε0ðxÞx2 -ω2

dx ð5Þ

Figure 2 shows the ε0 and ε0 0 spectra (ε00 is obtained from ε0 usingeq 5) of the measurements for two DMPC samples, both in fluidphase: one without contact with the ion-exchange resin (0 h) andthe other after 72 h of ion exchanging (72 h). Two relaxations areobserved, one just below 105 rad/s and the other around 108 rad/s.The relaxation corresponding to the low frequency is associatedwith the diffusion of the counterions condensed into a relativelythin layer around the vesicle electric double layer. Because theliposome in our case has a bilayer formed predominantly of

Figure 1. Variation of the ionic strength (b) and pH (f) with theconditioning time of liposome samples in contact with ion-ex-change resins.

(20) Rold�an-Toro, R.; Solier, J. D. J. Colloid Interface Sci. 2004, 274, 76.(21) B€ottcher, C. J. F.; Bordewijk, P. Theory of Electric Polarization; Elsevier

Scientific: Amsterdam, 1978; Vol. II.

2668 DOI: 10.1021/la902916y Langmuir 2010, 26(4), 2665–2670

Article Galera-Cort�es et al.

zwitterionic phospholipids, the relaxation corresponding to thehigh frequency is associated with the lateral diffusion of zwitter-ionic molecules by rotation around the normal to the bilayer.Such rotation gives rise to the lateral movement14,15 mentionedabove. In the low-frequency range, the dielectric relaxationstrength decreases due to the decrease of the number of ionscondensed on the external surface of liposomeas a consequence ofthe deionization process caused by the ion-exchange resins.

The dielectric spectra of all samples are described by thesuperposition of two Havriliak-Negami (HN) equations:21,22

ε�ðωÞ ¼ ΔεL

ð1þðjωτLÞRLÞβLþ ΔεH

ð1þðjωτHÞRHÞβHþ ε¥ ð6Þ

where Δε is the relaxation strength or intensity, R and β arenumbers between 0 and 1, and τ is the characteristic timeassociated with the peaks of the imaginary part (L and H indi-cate low and high frequencies, respectively). Parameters R andβ aremeasures of the peak broadening due to the superposition ofthe dielectric response of different dipoles in the local structures.For a Debye-type process with a single relaxation time, one hasR= β= 1, and the frequency at which ε00 is maximum is a goodmeasure of the average relaxation time.23 In thiswork, in all cases,R ≈ β ≈ 1 for the peak at high frequencies that corresponds to

Debye-type behavior with a single relaxation time (τH). In a fluidlattice of a liposome, the Debye relaxation time gives an estimateof the 2D lateral self-diffusion coefficient, assuming the phos-pholipid molecules can be modeled as continuous cylinders withtwo degrees of freedom, one of rotation and the other oftranslation. In that case14

Dlat erm

2ffiffiffi2

pτH

ð7Þ

where rm is the molecule’s mean headgroup radius. This expres-sion is obtained by assuming that the fluid phase of themembranehas vacancy sites and also considering each degree of freedomtakes up the same thermal energy. Then, taking14 rm = 0.4 nmand the fitted relaxation times (τH) from each spectrum, thecorresponding average value of the lateral self-diffusion coeffi-cient Dlat can be obtained.3.1. DMPC in the Fluid Phase. Figure 3 shows the lateral

self-diffusion coefficient (Dlat) and pH values as a function ofionic strength. As can be seen in that figure, a close inversecorrelation exists between the values ofDlat and the solution pH.It seems that, under these experimental conditions, the pHcontrols the lateral self-diffusion of DMPC in the bilayer ofliposomes in fluid phase. The values of Dlat obtained are of theorder of 10-12-10-11 m2/s, and they are of a similar order tothose determined by pulsed field gradient NMR23 and a fluores-cence technique.24 In 100 nm diameter liposomes, the diffusion ofthe liposome itself should not be neglected. Considering that theliposome is a sphere with such diameter, the hydrodynamicdiffusionmodel predicts that the 100 nm liposome itself is movingwith a diffusion rate of∼5� 10-12m2/s at 30 �C.However, fromadielectric point of view, the corresponding characteristic ortypical relaxation time for this particle is of the order of ∼1 msand the relaxation times of the phospholipidmolecules are approxi-mately ∼10 ns. Obviously, these relaxation times correspond toranges of frequencies clearly separated and easy to differentiate.Therefore, the Dlat determined by dielectric spectroscopy in theMHzrangeof frequencieshasonly the contributionof the rotationalrelaxation of the phospholipid molecules. The dielectric spectrosco-py allows us to discriminate between both diffusive processes.

Regarding the dramatic change in Dlat of approximately1 order of magnitude with a relatively small change in pH, it

Figure 2. ε0 (a) and ε0 0 (b) spectra for DMPC samples in the fluidphase, without contact with the ion-exchange resin (0 h) and after72 h subjected to the effect of the ion-exchange resin in the liquidphase (72 h). Solid lines correspond to the fitting obtained byHavriliak-Negami equations.

Figure 3. Lateral diffusion coefficient (Dlat, b) and pH (f) as afunction of ionic strength for DMPC liposomes in the fluid phase.

(22) Havriliak, S.; Negami, S. Polymer 1961, 8, 161.(23) Filippov, A.; Or€add, G.; Lindblom, G. Biophys. J. 2003, 84, 3079.

(24) Almeida, P. F. F.; Vaz, W. L. C.; Thompson, T. E. Biochemistry 1992, 31,6739.

DOI: 10.1021/la902916y 2669Langmuir 2010, 26(4), 2665–2670

Galera-Cort�es et al. Article

should be noted that changes of the same order have beenobserved in the fluid phase of DMPC liposomes when the lateraldiffusion coefficient was measured as a function of the tempera-ture.14,23 These changes inDlat with temperature were even largerthan the change associatedwith the gel-to-liquid-crystalline phasetransition.14

Makino et al.25 showed experimentally some years ago thatneutral liposomes composed of DMPC phospholipids exhibit anonzero effective charge (Zef) when they are dispersed in aqueoussolutions at approximately neutral pH. Obviously, this effectivecharge is due to the selective adsorption of certain ions from thebulk solution. We have estimated the effective charge on theexternal surface of DMPC liposomes measuring their microelec-trophoretic mobility at different pH and ionic strength values.First, we converted the experimental values of mobility into ζ-potential using the classical theoretical treatment (Henry’s equa-tion). Second, we converted the ζ-potential into effective chargeusing the Coulomb equation. As can be seen in Figure 4, theeffective charge (Zef) of the DMPC liposomes is negative and thisindicates that anions are preferably adsorbed on these neutralliposomes. The Zef depends on both the solution pH and its ionicstrength. There is an important change of Zef when the ionicstrength varies 3 orders of magnitude approximately. A closeinverse correlation can be seen again between the values of bothZef and pH as in the case ofDlat (see Figure 3). Also, these resultsindicate that the effective chargeon theDMPC liposomes is a verysensitive function of the solution pH.A small change in pH (∼0.5)led to significant variations in the effective charge on the lipo-somes’ surface (see Figure 4). This, however, is not a new situationbecause similar changes of ζ-potential (equivalent to Zef) relatedto a similar range of variation of pHvalues have been observed byother authors.26 Hence, pH controls the negative effective chargeof DMPC liposomes, and this means that OH- is the ion thatdetermines the potential.

Therefore, the pH of the solution controls the lateral diffusionin the bilayer and the effective charge on the external surface ofthese liposomes. Both results can be explained assuming aselective adsorption of OH-. As a consequence of this anionadsorption, the dipolar headgroup of phospholipids varies itsorientation with respect to the direction normal to the surface of

the liposome; this means that dipoles are more or less perpendi-cular to the surface, depending on theOH- concentration close tothe surface, which is given by the pH changes. Depending on theorientation of the dipolar headgroup of the phospholipids withrespect to the external surface, the lateral diffusion of phospho-lipids ismore or less difficult. The correlation betweenZef andDlat

is shown in Figure 5 for different pH values. There is a goodagreement between both magnitudes for each value of pH.Accordingly, the effective charge of the liposomes is a macro-scopic effect of the dynamic or local structure of the DMPCphospholipids in the bilayer.27

It is quite evident that the pHplays a crucial role in the effectivecharge and the lateral diffusion of the PLs in DMPC liposomes.Now we can give a plausible explanation to the time dependenceof the pH (see Figure 1). Comparing Figures 1 and 4, we can seethat the pH of the solution is clearly determined by the effectivecharge on the liposomes’ surface and that the presence in the bulksolution of the potential-determining ions (OH- or Hþ) dependson their specific adsorption on the external surface of the lipo-some. Some authors have indicated that the decrease in the pH ofthe solutions is probably due to the diffusion toward the solution

Figure 4. Effective charge (Zef,b) and pH (f) as a function of theionic strength of the DMPC liposomes in the fluid phase.

Figure 5. Lateral diffusion coefficient (Dlat, b) and the effectivecharge (Zef,f) as a function of pH of the DMPC liposomes in thefluid phase.

Figure 6. Lateral diffusion coefficient (Dlat, b) and the effectivecharge (Zef,f) as a function of pH of the DMPC liposomes in thegel phase.

(25) Makino, K.; Yamada, T.; Kimura, M.; Oka, T.; Ohshima, M.; Kondo, T.Biophys. Chem. 1991, 41, 175.(26) Fatouros, D. G.; Klepetsanis, P.; Ioannou, P. V.; Antimisiaris, S. G. Int. J.

Pharm. 2005, 288, 151. (27) Schrader, W.; Kaatze, U. J. Phys. Chem. B 2001, 105, 6266.

2670 DOI: 10.1021/la902916y Langmuir 2010, 26(4), 2665–2670

Article Galera-Cort�es et al.

of the atmospheric CO2.28,29 In our case, this effect is probably

negligible because all the ionic impurities of this atmosphericcontamination would be removed from the solution by the ion-exchange resins.3.2. Comparison of Gel/Fluid Phases. We changed the

temperature from 25 to 19 �C to cause a phase transition fromthe fluid to the gel phase of the zwitterionic phospholipids in thebilayer. Bydoing this, wewere able to study the dielectric responseof a liposome with a gel structure in the bilayer.

Figure 6 shows the data of Zef and Dlat obtained with a gelstructure in the bilayer as a function of pH. As can be seen in thisfigure,whenZef increases in absolute value,Dlat also increases andwe observe basically the same behavior for both properties withthe pH. In this case, a clear inverse correlation also appears with

the pH, although the values ofZef andDlat are smaller for the gelphase than for the liquid phase. Hence, the local structure and the“lashing” of the phospholipids to each other in the bilayerdetermine the effective charge acquired by OH- adsorption/desorption as well as the lateral diffusion. Neither the frictionprocess and the dipolar orientation of the headgroup nor theireffects are simple, easy, and linear phenomena.

The effective charge in both phases has a similar behavior inrelation to pH, but the values in the gel phase are always smallerthan the corresponding values in the fluid phase (see Figure 7).14

The changes with pHare also smaller in the gel phase. This abruptchange of Zef is a consequence of higher stiffness of the bilayer inthe gel phase in comparison with the fluid phase. Furthermore,the different number of bonds changes the friction betweenphospholipids during their rotation and explains their behavior.The stiffness effect in the gel bilayer is more evident when lookingat the lateral diffusion versus pH (see Figure 8). The changesin the lateral diffusion in the fluid phase aremoremeaningful thanin the gel phase, although, in both cases, the pH effect is quitesimilar. The control of lateral diffusion by effective charge ismeaningfully higher in the fluid than in the gel phase. This is dueto a greater difficulty for the rotation/translation of the phos-pholipids in the gel phase.

4. Conclusion

Broad-band dielectric spectroscopy has been shown to be aneffective tool to study the dynamic and even the local structure ofphospholipids in the bilayer.Measurements of the dipolar relaxa-tion time provided the average diffusion coefficient of thesemolecules for the fluid as well as for the gel phases of theliposomes.

In the range of salt concentration analyzed (<10-3M), the pHplays a crucial role in the lateral translation of this molecule inboth phases of the vesicles. The dynamics of the lateral diffusionare very sensitive to very small changes of the pH, especially in thefluid phase. The dependence of the Zef and Dlat of the bilayer onthe pH of the solution is not simple and linear; they have shownsimilar behaviors in both phases.

The Zef and Dlat values in the gel phase are always lower thanthose in the fluid phase.

The translational motion of the phospholipid molecules asso-ciated with the dipolar headgroup rotation in the bilayer ofDMPC liposomes is correlated with the effective charge on theexternal layer. To the best of our knowledge, this is the first timethat this relation has been observed.

Acknowledgment.Thiswork has been supported byMICINNprojects MAT2006-12918-C05-01, -03, -05, “Plan Nacional deInvestigaci�on, Desarrollo e Innovaci�on Tecnologica (I þ D þ i)Ministerio de Ciencia e Innovaci�on” Spain, and by the EuropeanRegional Development Fund (ERDF).

Supporting Information Available: Micrograph of ex-truded DMPC liposomes obtained by freeze-fracture elec-tron microscopy. This material is available free of charge viathe Internet at http://pubs.acs.org.

Figure 7. Effective charge (Zef) of the DMPC liposomes in the gel(f) and in the fluid phases (b) versus pH.

Figure 8. Lateral diffusion coefficient (Dlat) of the DMPC lipo-somes in the gel (f) and in the fluid phases (b) versus pH.

(28) Carrique, F.; Ruiz-Reina, E. J. Phys. Chem. B 2009, 113, 8613.(29) Ruiz-Reina, E.; Carrique, F. J. Phys. Chem. B 2008, 112, 11960.