Probing the Modes of Antibacterial Activity of Chitosan. Effects of pHand Molecular Weight on Chitosan Interactions with MembraneLipids in Langmuir FilmsBarbara Krajewska,* Paweł Wydro, and Agnieszka Janczyk

Faculty of Chemistry, Jagiellonian University, 30-060 Krakow, Ingardena 3, Poland

ABSTRACT: Chitosan, a cationic biopolymer derived fromchitin, has been described as having antibacterial activity. Themodes of this activity, however, have not been established.One mode proposed is that chitosan perturbs bacterial cellmembranes. To validate this proposal, in this study weinvestigated chitosan interactions with lipids in Langmuirmonolayers as model membranes. The interactions wereassessed by monitoring differences in the shape of the com-pression isotherms measured in the absence and presence ofchitosan in the subphase (acetate buffer). To appraise thecontribution of electrostatic interactions versus hydrogen bonding and hydrophobic interactions, three membrane lipids differingin charge were studiedanionic dipalmitoylphosphatidylglycerol (DPPG), zwitterionic dipalmitoylphosphatidylcholine (DPPC),and neutral cholesteroland the pH of the subphase was varied between 3.5 and 6.0. In addition, the impact of the molecularweight of chitosan on the interactions was assessed at pH 3.5. It was found that while chitosan had a negligible effect on DPPCmonolayers over the pH range studied, it distinctly affected DPPG and cholesterol monolayers. The effect on DPPG was foundto decrease with increasing pH, that at pH 3.5 being ascribed to the charge-mediating action of chitosan on the local ionicenvironment and that at higher pHs to the intercalation of chitosan to the monolayers. Practically independent of pH, the effectof chitosan on cholesterol was accounted for by the formation of cholesterol−chitosan hydrogen bonds. Chitosan of lowermolecular weight facilitated the interactions with all the three lipids studied. The results obtained may be helpful in identifyingthe mode of antibacterial activity of chitosan versus other modes that involve the disturbance of cell life cycles.

■ INTRODUCTION

Owing to their availability, biological benignity, and biodegr-ability, biopolymers are increasingly researched for theapplication in medicine, among others as antibacterial agents.One such biopolymer with proposed antibacterial activity ischitosan. Chitosan is a polyaminosaccharide derived fromchitin.1−3 Chitin, one of the most plentiful organic resources innature, fabricated mainly from the shells of crustaceans,chemically is a linear polymer composed of N-acetylglucos-amine units. Chitosan is obtained from chitin by N-deacetylation and consequently possesses multiple aminogroups in its molecules (Scheme 1). The groups make chitosan

a cationic polyelectrolyte (pKa ≈ 6.5), owing to which inaqueous acidic media chitosan has high positive charge on−NH3

+ groups, thereby being able to adhere to or aggregatewith negatively charged molecules. The amino groups are alsocapable of chelating metal ions. These properties can beexploited in a variety of processes, medical treatments

included,1−4 and these are possible due to excellentbiocompatibility and physiological inertness of chitosan.Chitosan has shown antimicrobial activity against both

Gram-positive and Gram-negative bacteria.5,6 Though demon-strated experimentally in bacterial tests, the mechanism of thisactivity is still not well understood. Several possible modes ofthis activity have been considered. In one, electrostaticinteractions between chitosan polycationic molecules andnegatively charged cell membrane components are seen asresponsible for the loosening of membrane structures and forthe consequent loss of cell components. In another one,chitosan is viewed as an agent penetrating through the cell wallinside the cells, where its binding to DNA results in theinterference with synthesis of mRNA and proteins. In the thirdmode by contrast, it is proposed that chitosan chelates metalions and subsequently inhibits the production of toxins and themicrobial growth. In other modes, dehydration, gelling,enzyme-inhibiting, and flocculating properties of chitosan areproposed to be involved.In this work we addressed the first proposal that involves the

disturbance of cell membranes by chitosan. One way to study

Received: September 2, 2011Revised: September 17, 2011Published: September 21, 2011

Scheme 1. Structure of Chitosan

Article

pubs.acs.org/Biomac

© 2011 American Chemical Society 4144 dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−4152

such disturbances is by using the Langmuir film balancetechnique, where biological membranes are mimicked by modelmembranes prepared as monolayers of membrane lipids onaqueous subphases.In the Langmuir film balance technique, the surface pressure

π is measured as a function of the area A available for themolecules in the monolayer of a water-insoluble amphiphilespread at the air−aqueous interface.7,8 The resulting sur-face pressure−area (π−A) isotherm evinces the phase behaviorof the monolayer in the course of compression, when themolecular packing gradually increases. Typically at large areas,monolayers are in the gaseous phase and zero surface pressureis observed. Further compression changes the gaseous phaseinto the liquid-expanded (LE) phase, and the isotherm liftsfrom zero pressure. Next, the monolayer undergoes a first-ordertransition to the liquid-condensed (LC) phase that is seen as aplateau in the isotherm, where LE and LC phases coexist. Oncein the LC phase, the surface pressure of the monolayer growsrapidly with the compression. Depending on the amphiphileand experimental conditions, a monolayer can also condensedirectly from the gaseous phase to the LC phase. A kink in theisotherm at higher surface pressure, by contrast, indicates asecond-order phase transition to the condensed phase. In eithercase the isotherms continue to rise up to the collapse of themonolayer marked by an irregularity in the isotherm.Importantly, the detailed characteristics of the π−A isothermsare indicative of the interactions both between molecules in themonolayer and between molecules of the monolayer and of thesubphase. These in turn depend on the molecular features ofthe amphiphile (ionization of the headgroup, length andsaturation degree of the hydrocarbon tails) and on thecomposition of the subphase (pH, presence of cations andother compounds, ionic strength).The technique has been widely applied in studies of various

membrane disturbants, such as drugs,9,10 peptides andproteins,7,11 surfactants,12 and, importantly, polyelectro-lytes.13−17 Among the latter chitosan was extensively studiedby Pavinatto et al.,18−21 however, under one experimentalcondition of pH and molecular weight of chitosan.In this context, we report here the results of a study

performed to validate the proposal that chitosan disturbsbacterial cell membranes, where both pH conditions and themolecular weight of chitosan were varied. The interactions ofchitosan with membrane lipids in Langmuir monolayers wereassessed by monitoring differences in the shape of the com-pression isotherms measured in the absence and presence ofchitosan in the subphase (acetate buffer). To appraise thecontribution of electrostatic interactions versus hydrogenbonding and hydrophobic interactions, three lipids differingin charge were studiedanionic dipalmitoylphosphatidylgly-cerol (DPPG), zwitterionic dipalmitoylphosphatidylcholine(DPPC), and neutral cholesteroland the pH of the subphasewas varied between 3.5 and 6.0. The pH range of the subphase< pH 6.0 was chosen to ensure the solubility of chitosan (pKa ≈6.5). This pH variation assured a change of charge at theinterface. In addition, the impact of the molecular weight ofchitosan on the interactions was assessed at pH 3.5.

■ EXPERIMENTAL SECTIONMaterials. Three chitosans of different molecular weights were

purchased from Sigma. Low molecular weight chitosan (chitosan L)was from chitin of shrimp shells, its nominal viscosity being 20−300cP for 1% solution in 1% acetic acid at 20 °C, whereas medium and

high molecular weight chitosans (chitosan M and chitosan H) werefrom chitin of crab shells, with nominal viscosities in the sameconditions, of 200−800 and >400 cP, respectively. The samples weredetermined to contain moisture in the following amounts: chitosan L,10.3%; chitosan M, 8.3%; and chitosan H, 4.3%, which is why thesamples were always dried overnight at 65 °C before use. All stockchitosan solutions were prepared along the following scheme.22 Anamount of chitosan was added to a chosen aqueous solution. Thesuspension was heated to 60 °C and mixed with a magnetic stirreruntil the complete dissolution of chitosan was achieved. Further thesolution was stirred for another 2 h at 60 °C and allowed to stand for20 h at room temperature. Ultrapure water obtained from a Simplicity185 Millipore water purification system (resistivity 18.2 MΩ cm) wasused throughout the experiments for solution preparation andmeasurements. The lipids studieddipalmitoylphosphatidylglycerol(DPPG), dipalmitoylphosphatidylcholine (DPPC), and cholesterolwere purchased from Sigma. They all had purities ≥99% and wereused as received. The spreading solutions of DPPG and DPPC wereprepared in chloroform/methanol 4:1 (v/v) mixture, while cholesterolwas dissolved in chloroform. Chloroform and methanol were of HPLCgrade, purity ≥99.9% (Sigma-Aldrich).Characterization of Chitosans (Molecular Weight and

Deacetylation Degree). The viscosity-average molecular weightsof chitosans, Mv, were determined using the Mark−Houwink equationfor which the measurements were carried out in an Ubbelohde-typecapillary viscometer, Visco System AVS 350, Schott-Gerate, Germany,at 25 °C. For the measurements chitosans were purified. Thepurification consisted of filtering 1% (w/v) chitosan solutionsprepared in 0.8% (v/v) acetic acid to remove solid contaminants.This was followed by precipitating chitosans with a NaOH solutionand filtering them again. The precipitates were washed with water untilneutral followed by washing with methanol and next dried overnight at65 °C. The stock chitosan solutions for the viscosity measurementswere prepared at a concentration of 0.1 g/100 mL in 0.1 MCH3COOH containing 0.2 M NaCl.23 The dilutions for the desiredconcentrations were done in the viscometer with the same solvent.The viscometer was installed in a temperature-controlled water bathmaintained at 25 ± 0.1 °C. Each flow in the viscometer was repeatedat least five times. The flow time of the solvent was 90.52 ± 0.03 s.

The deacetylation degrees (DD) of chitosans were determined byIR using chitosan in membrane form. The membranes were preparedby a solvent evaporation method, for which volumes of 1% (w/v)chitosan solution in 0.8% (v/v) acetic acid were cast on polyethyleneplates and dried overnight at 65 °C. The dry membranes were nextneutralized with 4% NaOH solution for 30 min, rinsed with water untilneutral, washed with methanol, and dried again for 40 min.24,25 The IRspectra were recorded with Bruker Equinox 55 FTIR spectrometer.Langmuir Monolayer Measurements. Stock chitosan solutions

of the concentration 1 g/100 mL in 0.8% (v/v) acetic acid wereprepared. Chitosan solutions to be used as subphase solutions in theLangmuir monolayer measurements were prepared by diluting thestock solutions with 0.6 M acetate buffer of a given pH. Acetate buffers0.6 M with pHs 3.5, 4.1, 4.75, 5.35, and 6.0 were prepared by adjustingpH of acetic acid with NaOH.

The Langmuir monolayer measurements were performed with aLangmuir balance (NIMA, model type 301, UK) mounted on anantivibration table and housed in a Plexiglass box. The balance had aTeflon trough of total area 300 cm2. The trough was filled with asubphase solution that was either acetate buffer alone or the samebuffer containing chitosans at concentrations from the range between0.0125 and 1 mg/mL. The lipid spreading solutions were deposited onthe subphase in the trough with use of a Hamilton microsyringe(accuracy ±1.0 μL). The concentrations of the lipids and thespreading volumes were as follows: DPPG, 0.224 mg/mL, 110 μL;DPPC, 0.258 mg/mL, 120 μL; cholesterol, 0.311 mg/mL, 70 μL.Upon deposition the solutions were left for 20 min to allow thesolvents to evaporate, and subsequently the monolayers werecompressed by a movable barrier at a rate of 20 cm2/min (thecompression velocity between 5 and 35 cm2/min was found not toinfluence the investigated compounds). The surface pressure (π)−area

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524145

(A) isotherms were recorded, where the pressure was measured with aWilhelmy plate made of filter paper (ashless Whatman Ch1)connected to an electrobalance, to an accuracy of ±0.1 mN/m. Themeasurements were performed at 25 ± 0.1 °C, at pHs 3.5, 4.1, 4.75,5.35, and 6.0.

To characterize the monolayers of the lipids, the recorded isothermswere processed to obtain the following parameters: (i) A lift-offmolecular area A0 that is the area at which the isotherm begins to risefrom zero surface pressure on compression. (ii) A limiting moleculararea Alim that is the area occupied by one molecule in a highlycompressed monolayer. Alim was estimated by extrapolation of thesteep linear part of the isotherm to zero surface pressure. (iii) Acollapse pressure π coll that is a pressure at which a monolayer breaks.π coll was determined from the point where the rise of an isotherm wasarrested. (iv) A compression modulus CS

−1 (also called elasticity). Themodulus is a reciprocal of the monolayer compressibility and is ameasure of the monolayer resistance to a change in area. It can becalculated from the slope of the π−A isotherm as26

The importance of the modulus lies in the fact that its maximal valuein the plot of CS

−1 vs π allows the phases of the monolayer observed inthe compression isotherm to be classified as26 liquid-expanded(CS,max

−1 ca. 12.5−50 mN/m), liquid (CS,max−1 ca. 50−100 mN/m),

liquid-condensed (CS,max−1 ca. 100−250 mN/m), or as condensed

(CS,max−1 >250 mN/m). The minima in the plots of CS

−1 vs π, on theother hand, correspond to phase transitions.

In this study the uncertainty of A0 and Alim determinations was±1 Å2, and that of π coll and CS

−1 was ±1 and ±10 mN/m, respectively.

■ RESULTS AND DISCUSSION

Characterization of Chitosans. For the determination ofthe viscosity-average molecular weights of chitosans, the flowtimes, t, of chitosan solutions of concentrations in the range0.01−0.05 g/100 mL in 0.1 M CH3COOH/0.2 M NaCl weremeasured in a Ubbelohde viscometer. Viscosity numbers werecalculated from the formula

(1)

where t0 is the flow time of the solvent alone and c is thechitosan concentration in the solutions expressed in g/mL. Theviscosity numbers η r measured are presented in Figure 1a as afunction of c. The limiting viscosity numbers [η] were obtainedby extrapolating η r to zero concentration, and the viscosity-average molecular weights Mv were calculated using Mark−Houwink equation:

(2)

where the constants for the solvent used, 0.1 M CH3COOH/0.2 M NaCl at 25 °C, were K = 1.81 × 10−3 mL/g and α =0.93.23 The Mv values of chitosans studied were found to rangefrom 360 kDa for chitosan L to 2506 kDa for chitosan H(Table 1).

The deacetylation degree (DD) of chitosan is defined as theaverage number of glucosamine units per 100 monomeric unitsin chitosan chains, expressed as a percentage.27 For thedetermination of DD of the studied chitosans, IR spectra of thedry chitosan membranes were recorded. Representative spectraare shown in Figure 1b. For the calculation of DD, the A1655/A3450,

28 and A1655/A2867 ratios29 were used. Both methodsutilize the absorbance of amide I band at 1655 cm−1

corresponding to the concentration of N -acetyl groups, theinternal reference in the first method being the absorption bandof hydroxyl groups at 3450 cm−1, and that in the latter methodbeing the CH stretching band at 2867 cm−1. The baselinesindicated in Figure 1b with dashed and dotted lines werechosen for the calculations in keeping with the suggestions ofthe authors. In the first method the DD values were calculatedfrom the formula28

(3)

while in the second one they were read out from the plot ofA1655/A2867 versus DD offered by the author.29 The DD’s listedin Table 1 are means of the values obtained with both methodsfrom several independently recorded spectra. The values showthat all three chitosans used in this study have their DD close invalue around 80%.Langmuir Monolayer Measurements. The lipids chosen

for this study were DPPG, DPPC, and cholesterol. Theirstructural formulas are presented in Scheme 2. In terms of thecharge, the lipids differ in that DPPG is anionic, DPPC iszwitterionic, and cholesterol is uncharged.The monolayers of the lipids were formed on 0.6 M acetate

buffer at five pHs between 3.5 and 6.0, in the absence and

Figure 1. Characterization of chitosans: (a) viscosity numbers measured in Ubbelohde viscometer plotted as a function of chitosan concentration;(b) IR spectra of chitosans (baselines: dashed;28 dotted29).

Table 1. Physical Properties of Chitosans

property chitosan L chitosan M chitosan H

Mv, kDa 360 ± 1 1202 ± 2 2506 ± 14DD, % 81 ± 2 83 ± 5 77 ± 2

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524146

presence of chitosan M in the subphase at concentrations fromthe range 0.0125−1 mg/mL, and their compression isotherms(surface pressure−area, π−A) were recorded. An overview ofthe recorded data is presented in Figure 2, where representativeisotherms at three pHs, 3.5, 4.75, and 6.0 are shown. For clarity,only selected isotherms are illustrated at each pH.Exhibiting no surface activity on its own,30 chitosan present

in the subphase in increasing concentration shifted thecompression isotherms of all the lipids studied toward largermolecular areas at all pHs studied (Figure 2), meaning that themonolayers were expanded and that chitosan action waseffectively that of a membrane disturbant. Although withdifferent magnitudes, the expansions observed had onecommon feature when expressed in terms of the limitingmolecular area Alim and the maximal compression modulusCS,max

−1. The feature is that on increasing chitosan concen-tration from 0 to 1 mg/mL Alim of all the monolayers at eachpH studied increased and CS,max

−1 decreased to reach stablevalues. This shows that at certain chitosan concentrations<1 mg/mL no more chitosan molecules could interact with thelipids molecules to further expand their monolayers. Therepresentative dependencies of Alim and CS,max

−1 of themonolayers on chitosan concentration, recorded for chitosanM at pH 3.5, are presented in Figure 3. In view of the aboveobservations, in further discussion we adopted the values of Alimand CS,max

−1 determined for 1 mg/mL chitosan concentration,as those characterizing the maximal possible expansive effect ofchitosan on the lipid monolayers.Major forces operative in the formation of lipid monolayers

on aqueous subphases are electrostatic interactions, hydro-phobic interactions, and hydrogen bonding. These in this studywere tuned by varying the pH of the subphase and addingchitosan to the subphase. The isotherms recorded in thesevarying conditions had their principal parameters, Alim andCS,max

−1, determined. The parameters are plotted as a functionof pH in Figure 4 for the pure lipids on the buffers and for thelipids on the buffers containing chitosan at a concentrationcausing the maximal expansive effects (1 mg/mL). Thesedependencies together with other results will be used in thediscussion of pH and chitosan effects on monolayer phasebehavior of each individual lipid at the interface. To better

analyze the effects of chitosan, in each case we will first discussthe behavior of the monolayers of pure lipids on the buffers.

DPPG Monolayers. The representative π−A isotherms ofpure DPPG monolayers recorded on acetate buffers in the pHrange 3.5−6.0 are shown in Figure 2a−c (solid lines). Aspresented in Figure 2, pH had a marked effect on the phasebehavior of DPPG monolayers, significantly changing the shapeand location of the isotherms. At pH 3.5, the isotherm lifts offat a molecular area of 70 Å2 and the gaseous phase transformsdirectly into the LC phase. This is followed by the second-orderphase transition at a surface pressure between 30 and 35 mN/mto the condensed phase, where molecules occupy the smallestarea compared to the monolayers at higher pHs. Themonolayer collapsed at π coll ≈ 47 mN/m. In contrast to pH3.5, at higher pHs the isotherms are located at larger molecularareas, lifting off at 115 Å2/molecule. More importantly, absentat pH 3.5 and 4.1, the first-order LE−LC phase transition wasdeveloped by increasing pH. It appeared at pH ≥ 4.75 as aplateau in the isotherms. The surface pressure of this transitiongrew with pH to reach the values between 12.5 and 16.5 mN/mat pH 6.0. Additionally, the increasing pH gradually increasedthe monolayer collapse pressure to π coll ≈ 67 mN/m at pH 6.0,this being indicative of a better stability of the monolayers athigher pHs. Similar dependencies of the isotherm shapes on thesubphase pH, though sometimes with different numericalcharacteristics depending on the buffer system applied, werereported by other authors.31,32

The changes in the structure of DPPG monolayers can alsobe readily followed by the changes of Alim and CS,max

−1 valuesresulting from the isotherms. These are plotted as a function ofpH in Figures 4a and 4d, respectively (solid circles). The plotsshow that the values of Alim markedly grew with pH between3.5 and 6.0, from 42 to 54 Å2/molecule, and those of CS,max

−1

markedly fell from 615 to 225 mN/m, both plots leveling off atpH ≥ 4.75. The latter plot also reveals that on compression atpH < 4.75 the monolayer exists as a condensed phase, while atpH > 4.75 as a LC phase.The foregoing observations prove that on increasing pH of

the subphase from 3.5 to 6.0 DPPG monolayers underwentexpansion and fluidization. This behavior originates from thedeprotonation of the phosphate group of the lipid,31,32 that atpH 3.5 being typical of the protonated DPPG. Thedeprotonation results in increasing repulsive Coulombic forcesbetween negatively charged headgroups, thereby giving riseto looser molecular packing in DPPG monolayers, ofwhich characteristic is the observed LE−LC phase transition.From the data obtained (Figure 2a−c and Figure 4a,d) it canbe inferred that the pKapp (referred to the bulk pH) ofthe phosphate group of DPPG in the monolayer under theexperimental conditions applied in this work is ca. 4. The valueis in a good agreement with other pKapp values suggested inearlier reports, i.e., ∼4−5,32 3.52,33 and 3.9−4.1.34 Clearly, atpH below this value the electrostatic repulsion between themolecules is lost due to protonation of the headgroups. Underthese conditions, a strong intermolecular hydrogen bondingbetween the glycerol hydroxyl and the phosphate of theneighboring lipid molecules was shown to be prevailing,32,35,36

resulting in dense packing, rigidity, and brittleness of themonolayer.The compression isotherms for the DPPG monolayers on

the subphase containing chitosan M at different pH valueswere further obtained. The representative curves at increasingchitosan concentration are shown in Figure 2a−c. At all studied

Scheme 2. Structural Formulas of the Investigated Lipids

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524147

pHs in the presence of chitosan, the expansion of themonolayers is observed, very interestingly, this expansive effectdistinctly decreasing with the rising of pH. The recordedisotherms had their Alim and CS,max

−1 determined. Importantly,as the Alim are reported per number of initially spread moleculesof the lipid, any change in their values is reflective of thepolymer effects on the monolayer. As shown in Figure 4a,d, like

for the pure DPPG monolayers, Alim of the isotherms in thepresence of chitosan 1 mg/mL grows with pH to level off at pH> ∼4.75, but what is important is that the effect of chitosan onthe magnitude of Alim is larger at low than at higher pHs, ΔAlim

at pH 3.5 being 22 Å2/molecule and that at pH 6.0, 14 Å2/molecule. Remarkably, the CS,max

−1 underwent reduction toreach a stable value of 130 mN/m on average, independent of

Figure 3. Effect of chitosan concentration in the subphase, determined for chitosan M at pH 3.5, on the parameters of lipids monolayers: (a) limitingmolecular area Alim and (b) the maximal compression modulus CS,max

−1.

Figure 2. Compression surface pressure−area isotherms of DPPG (a−c), DPPC (d−f), and cholesterol (g−i) monolayers spread on acetate buffersubphase recorded at pH 3.5, 4.75, and 6.0 in the absence and presence of dissolved chitosan M (chitosan concentration: 0 mg/mL, solid line (purelipids); 0.1 mg/mL, dotted line; 0.3 mg/mL, short-dashed line; 1 mg/mL, dashed line).

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524148

pH, meaning that in the presence of chitosan all the DPPGmonolayers are in the LC state. The magnitude of thisreduction was highest at pH 3.5, amounting to 480 mN/m, andlowest at pH 6.0, 80 mN/m.In view of the fact that with an increase in pH both the lipid

monolayer and the polyelectrolyte alter their properties, themonolayer becoming negatively charged as described in theforegoing section, and chitosan losing its positive charge (pKa ≈6.5), it can be suggested that an intricate interplay betweenvarious interdependent forces operative in the system isresponsible for the modification of the monolayer induced bythe presence of chitosan. The expansion itself, consisting offorcing the molecules in the monolayer away from each other,can basically be achieved either by building in polymersegments between the lipid headgroups in the monolayer orby modifying the molecules in the monolayer. In either case theexpansion depends on the monolayer state characteristic ateach pH, prior to the coupling of chitosan.The most striking expansion of the DPPG monolayer was

noted at pH 3.5. Intriguingly, while at this pH the pure DPPGmonolayer is protonated and the corresponding isotherm steepand lacking the first-order transition, in the presence of chitosanit acquired the shape and location much the same as that of thepure deprotonated DPPG monolayer observed at higher pHs.While clearly, this could not be brought about by direct bindingof chitosan to the monolayer, as at this pH the monolayer lackscharge, this behavior is suggestive that the monolayerunderwent deprotonation, chitosan being a modifier of theionic environment at the interface. It is known that the degreeof dissociation of ionic lipids in monolayers depends not onlyon pH but also on the presence of ions.31,33,34,37 Chitosancarries a positive charge, which is why its action can be viewedas that of simple salts of monovalent cations that were shown toinduce the dissociation of lipids in monolayers and, accordingly,their expansion. Depending on pH and ionic strength of the

subphase, the concentration causing the maximum filmexpansion was reported to be ca. 0.1 M NaCl, its furtherincrease frequently developing either a screening of themonolayer charge or the penetration of cations into themonolayer. These result in that the monolayer is not furtherexpanded or may be contracted.31,33,34,37

As shown in Figure 3, the observed change in the monolayerbehavior at pH 3.5 was brought about practically by as low achitosan concentration as 0.0125 g/mL. Using the determinedmolecular weight and deacetylation degree of chitosan M(Table 1), and assuming a complete protonation of chitosan atpH 3.5,38 the calculated concentration of positive chargesbrought to the solution by 0.0125 g/mL chitosan is 0.06 M, andthat by 1 mg/mL, 4.9 M, thus supporting the charge-mediatingaction of chitosan on the immediate ionic environment ofthe DPPG monolayer. Also, the high concentration of positivecharges at 1 mg/mL chitosan could be viewed as responsiblefor the smoothing of the LE−LC transition plateau seen in theisotherms at all pHs at this chitosan concentration, in that byreducing the surface potential it reduces electrostatic repulsionbetween the molecules in the monolayer.The question that remains unanswered is if at pH 3.5

chitosan intercalated the DPPG monolayer. Whereas thiscannot be ruled out at lower surface pressures, uponcompression all the isotherms in the presence of chitosanultimately approached that of pure DPPG and showed collapseat the same surface pressure (Figure 2a), thus supporting a viewthat at higher surface pressures chitosan was not present withinthe monolayers. Therefore, we can conclude that at pH 3.5 thepresence of chitosan in the solution led to the expansion ofDPPG monolayer through a charge-mediated action rather thanthrough intercalation into the monolayer. Similar studies of theeffects of chitosan on monolayers of anionic lipids (DPPG,DMPA) at pH 3 were performed by Pavinatto et al.18−20

Although general and qualitative descriptions of these effects in

Figure 4. Dependencies of the limiting molecular area Alim and the maximal compression modulus CS,max−1 of the lipid monolayers on pH of the

subphase in the absence (solid circles) and presence of 1 mg/mL chitosan in the subphase (open circles).

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524149

our works agree, the interpretations differ in that we looked atthe effects as resulting from the interplay of charges.As we further move on to higher pHs, the situation changes

in that the DPPG monolayer becomes negatively charged as aresult of dissociation. Under these conditions the effect ofaddition of positive charges into the solution should cause ascreening of the charges on the monolayer, which should inturn result in contraction.37 However, as observed in Figure2b,c the monolayers did undergo expansion. Consequently, itmay reasonably be assumed that the expansions were inducedby direct electrostatic attractions between the charged polyionsites and charged lipid headgroups. A possible scenario of whathappens at the interface is that the interactions lead to theinsertion of segments of the polymer chain between theheadgroups, inducing a larger separation between them as wellas preventing forces between hydrocarbon chains from beingoperative. Conducive to this action of chitosan should be theloss of its positive charge on approaching pH 6.5, an effectthat allows chitosan to become more flexible and to build athicker adsorption layer with more loops.13,16,39 Additionally, itmay be observed in Figure 2b,c that the DPPG monolayersin the presence of chitosan collapse at surface pressures that areincreasingly lower than those of pure DPPG, with an increasein pH, thus indicating that they are less stable than the puremonolayers.10 This in turn further supports the notion thatchitosan intercalated the monolayers.The foregoing discussion has attempted to overview the

behaviors of DPPG monolayers in the presence of chitosan.However, importantly, in analyzing these varying behaviors itshould not be overlooked that the surface pressure of cellmembranes is 30−35 mN/m.19 At this pressure in all theconditions of pH studied, the disturbance of the monolayers bychitosan was significantly more pronounced than at higherpressures where the monolayers were forced to dense packing.Likewise, it should not be overlooked that the behaviors of bothDPPG and chitosan presented here for 25 °C may not be thesame at higher temperatures.

DPPC Monolayers. The compression π−A isotherms ofpure DPPC on acetate buffers were further measured, andrepresentative examples are presented in Figure 2d−f (solidlines). The figure shows that, unlike on DPPG, the pH hadconsiderably less pronounced effects on the pure DPPCmonolayers. At all pHs in the experimental conditions used inthis study, the isotherms exhibited three characteristicregions:13,32,40,41 a LE phase below π < 9 mN/m, followed bya LE/LC phase transition at π between 9 and 12 mN/m, theLC phase being further compressed up to the collapse. Thisbehavior invariable with pH is consistent with the fact thatDPPC is a zwitterionic lipid bearing a negative charge on thephosphate group and a positive charge on the quaternaryammonium group, thus showing the properties of unchargedmonolayers in a broad range of pH, reported to be 3−13, 3−7,or 4−8 (summarized in ref 42). Although looking very muchalike, a closer inspection of the recorded isotherms revealssome small changes in the structure of the monolayers thatimply that the monolayers were expanded on an increase in pH.For instance, the lift-off area of the isotherms increased from 90to 98 Å2/molecule between pH 3.5 and 6.0 and likewise grewπ coll from 57 to 67 mN/m, although the latter data are lessreliable due to difficult handling of DPPC monolayers uponovercompression. More importantly, as shown in Figure 4b,Alim grew from 48 Å2/molecule by 2 Å2/molecule, but theCS,max

−1 modulus, typical of the LC phase, remained practically

constant between pH 3.5 and 6.0. It was suggested in theliterature that this small expansion of DPPC monolayer maypossibly derive from the formation of hydrogen bonds betweenthe headgroups that may have some partial positive charge atlow pHs close to pKa1 of DPPC (3.8−4.0)40 and from therepulsion of the zwitterionic headgroups (parallel-oriented) atpHs toward the isoelectric point of DPPC (6−7).43As shown in Figure 2d−f, the presence of chitosan M in the

subphase produced very little effect on DPPC monolayers.Only at low pH 3.5 is the shift of the isotherm at 1 mg/mLchitosan appreciable, which may be a result of the repulsionbetween the partial positive charge on the DPPC headgroups atthis pH and fully protonated chitosan. At higher pHs bycontrast, all the isotherms in the presence of chitosan, althoughslightly enlarged at pressures lower than 30 mN/m, approachedthe slope of the isotherms of pure DPPC and collapsed at thesame pressures. Also, all the isotherms retained the plateautypical of the first-order LE/LC transition. As shown in Figure4b,e (open cirles), in the whole pH range studied, themonolayer experienced rather small structural loosening; i.e.,Alim grew to one value ca. 52 Å2/molecule accompanied by aslight drop in the maximum compression modulus of ca. 30mN/m, independent of pH. Noteworthily, our observations ofthe behavior of DPPC differ from those reported by Pavinattoet al.,18 where the authors recorded the isotherms considerablyenlarged by the presence of chitosan at pH 3.0, this disparitypossibly arising from the buffer effect. However, that couplingpolyelectrolytes to monolayers of zwitterionic lipids results innegligible effects was also reported for DPPC entered intocontact with anionic polyacrylate in the subphase13 and forDPPE (dipalmitoylphosphatidylethanolamine) with the cati-onic polymer PDADMAC.16

Crucially, in view of the fact that DPPG is a lipid typical ofbacterial cell membranes, while DPPC is typical of mammaliancell membranes,11 our findings that show that DPPGmonolayers are significantly and DPPC monolayers insignif-icantly disturbed by chitosan are of vital importance for theantibacterial activity of chitosan.

Cholesterol Monolayers. The π−A isotherms measured forcholesterol spread on the buffer subphases are shown in Figure2g−i (solid lines). Owing to its uncharged character, thecholesterol monolayers showed very little dependence on pH.Located at the limiting area Alim = 42 Å2/molecule independentof pH (Figure 4c, solid circles), the isotherms revealed a steepshape typical of condensed monolayers. Their maximal slope,however, was found to grow with increasing pH asdemonstrated by the values of the maximal compressionmodulus CS,max

−1 that grew from 650 to 750 mN/m in the pHrange 3.5−6.0 (Figure 4f (solid circles)), and π coll increasedfrom 37 to 42 mN/m. Similar pH effects were reported forcholesterol monolayers in the pH range 2−10.44As presented in Figure 2g−i, when added to the subphase,

chitosan M gave rise to the expansion of cholesterolmonolayers. Interestingly, at each pH studied the shift of theisotherms was nearly parallel to the isotherms of purecholesterol with a slight reduction of the slope on increasingchitosan concentration. The same behavior of cholesterol wasreported earlier for pH 3.145 and for pH 4.0 in our pre-vious work.46 Very importantly, such a pattern of isothermsproves that, in coupling to cholesterol monolayers, chitosan wasaccommodated within the monolayers, from which it was notsqueezed out upon complete compression. Moreover, at eachpH, the π coll of the isotherms on chitosan solutions was slightly

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524150

higher than the π coll of the parent isotherm, an effect especiallyvisible at pH 6.0, thus revealing that upon chitosan coupling themonolayers acquired stability. By contrast, Figure 4c,d showsthat the limiting area Alim at 1 mg/mL chitosan increased to aconstant value 46 Å2/molecule, and the CS,max

−1 modulus of thecholesterol monolayers also dropped to a constant value 400mN/m independent of pH, but the relative drop was highest atpH 6.0. Thus, the monolayer remained in the condensed state,but its structure changed into less compact. These findingsshow that although uncharged and thus lacking the possibilityof developing electrostatic interactions with chitosan, choles-terol did couple to chitosan to form stable assemblies. Theprincipal cholesterol−chitosan interaction mode should bethrough the formation of hydrogen bonds between thehydroxyls of cholesterol and the hydroxyls and amines ofchitosan, the latter supported by the fact that the highest effectof chitosan on cholesterol monolayer was observed as a drop ofCS,max

−1 at pH 6.0, where −NH3+ groups are close to

deprotonation. Additionally, having said that chitosan staysstably in cholesterol monolayers, hydrophobic interactionsbetween the steroid part of cholesterol and chitosan backboneintercalated into the monolayer in the form of loops, tangles,and tails cannot be excluded. Thus, it can be concluded thateven though not showing vast effects, the interactions ofcholesterol with chitosan play an important role in chitosandisturbing the cell membranes20 and, crucially, that thismembrane disturbing activity of chitosan is not solely due toelectrostatic interactions.21

Effect of Chitosan Molecular Weight on the LipidMonolayers. In addition to chitosan M (Mv 1202 kDa) usedin all of the earlier experiments, the isotherms of the three lipidsstudied were also measured for chitosan L and H (Mv 360 kDaand 2506 kDa, respectively) at pH 3.5. The resultingparameters observed at a chitosan concentration of 1 mg/mL,Alim, and CS,max

−1 are compared in Figure 5, where they are

plotted against chitosan molecular weight. The plots clearlydemonstrate that a lower molecular weight of chitosan favorsthe destabilization of all the monolayers; namely, the lower isthe molecular weight, the higher Alim and lower CS,max

−1 areobserved. A particularly pronounced effect of chitosanmolecular weight is seen for cholesterol, whose monolayer onchitosan L is in a liquid-condensed state in contrast to itscondensed state on chitosans M and H. These effects cannot beascribed to different positive charges of individual chitosansamples. It is because the charges depend on deacetylation

degree, whose values were shown to be similar for the chitosansunder study (Table 1). On the contrary, they apparently stemfrom a higher mobility of shorter polymers, which approachand couple the monolayers more readily than heavier polymers.Supportive to this is that for lower-molecular-weight chitosansthe diffusion coefficient was determined to be higher and theconformation more stretched.47 Like ours, marked effects ofmolecular weight of polyelectrolytes on the expansion of lipidmonolayers were reported for instance for a cationic lipidDOTAP (dioleoyltrimethylammonium propane) coupled toanionic polyacrylate14 and for a cationic surfactant DODAB(dioctadecyldimethylammonium bromide) in the presence ofanionic poly(styrenesulfonate).48

■ CONCLUSIONSIn this study the Langmuir film technique was used tocharacterize the interactions of chitosan with lipids in an effortto assess the impact of chitosan on cell membranes as a possiblemode of its antibacterial activity. Compression surfacepressure−area isotherms were measured for the monolayersof three membrane lipidsanionic dipalmitoylphosphatidyl-glycerol (DPPG), zwitterionic dipalmitoylphosphatidylcholine(DPPC), and neutral cholesterolformed on acetate buffersubphases containing different concentrations of chitosan atpHs between 3.5 and 6.0. It was observed that, though todifferent extents, chitosan changed the lipids monolayers, theirexpansion and fluidization being the measured effects. Thisindicates that chitosan has an effect on the monolayers, and itsaction is effectively that of a membrane disturbant. Morespecifically, it was found that chitosan had a very small effect onDPPC monolayers over the pH range studied, compared toDPPG. This finding is important in that DPPG is a lipid typicalof bacterial cell membranes, while DPPC is typical ofmammalian cell membranes. The effect on DPPG was foundto decrease with increasing pH, that at pH 3.5 being ascribed tothe charge-mediating action of chitosan on the local environ-ment and that at higher pHs to the intercalation of chitosan tothe monolayers. Practically independent of pH, the effect ofchitosan on cholesterol was accounted for by the formation ofcholesterol−chitosan hydrogen bonds. Chitosan of lowermolecular weight facilitated the interactions with all the threelipids studied. Overall, the study provides particulars to theunderstanding of the interactions of chitosan with membranelipids under different pH and molecular weight conditions andthus may prove useful in clarifying the mechanisms of theantibacterial activity of chitosan.

■ AUTHOR INFORMATION

Corresponding Author*Tel: +48-12-6632235. Fax: +48-12-6340515. E-mail:krajewsk@chemia.uj.edu.pl.

■ ACKNOWLEDGMENTSThis work was supported by grant K/PBW/000764.

■ REFERENCES(1) Rinaudo, M. Prog. Polym. Sci. 2006, 31, 603−620.(2) Aranaz, I.; Mengibar, M.; Harris, R.; Panos, I.; Miralles, B.;

Acosta, N.; Galed, G.; Heras, A. Curr. Chem. Biol. 2009, 3, 203−230.(3) Honarkar, H.; Barikani, M. Monatsh. Chem. 2009, 140, 1403−

1420.(4) Krajewska, B. Sep. Purif. Technol. 2005, 41, 305−312.

Figure 5. Dependencies of the parameters of the lipid monolayersmeasured for 1 mg/mL chitosan in the subphase at pH 3.5, onchitosan molecular weight: (a) Alim and (b) CS,max

−1.

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524151

(5) Badawy, M. E. I.; Rabea, E. I. Int. J. Carbohydr. Chem. 2011,DOI: ArticleID 460381.(6) Raafat, D.; Sahl, H.-G. Microb. Biotechnol. 2009, 2, 186−201.(7) Maget-Dana, R. Biochim. Biophys. Acta 1999, 1462, 109−140.(8) Takamoto, D. Y.; Lipp, M. M.; von Nahmen, A.; Lee, K. Y. C.;

Waring, A. J.; Zasadzinski, J. A. Biophys. J. 2001, 81, 153−169.(9) Casas-Sanchez, J.; Alsina, M. A.; Herrlein, M. K.; Mestres, C.

Colloid Polym. Sci. 2007, 285, 1351−1360.(10) Steinkopf, S.; Simeunovic, A.; Bustad, H. J.; Ngo, T. H.; Sveaass,

H.; Gjerde, A. U.; Holmsen, H. Biophys. Chem. 2010, 152, 65−73.(11) Won, A.; Ianoul, A. Biochim. Biophys. Acta 2009, 1788, 2277−

2283.(12) McConlogue, C. W.; Malamud, D.; Vanderlick, T. K. Biochim.

Biophys. Acta 1998, 1372, 124−134.(13) Bordi, F.; Cametti, C.; De Luca, F.; Gili, T.; Gaudino, D.;

Sennato, S. Colloids Surf., B 2003, 29, 149−157.(14) Truzzolillo, D.; Bordi, F.; Cametti, C.; Sennato, S. Colloids Surf.,

A 2008, 319, 51−61.(15) de Meijere, K.; Brezesinski, G.; Mo hwald, H. Macromolecules

1997, 30, 2337−2342.(16) de Meijere, K.; Brezesinski, G.; Kjaer, K.; Mo hwald, H.

Langmuir 1998, 14, 4204−4209.(17) de Meijere, K.; Brezesinski, G.; Pfohl, T.; Mo hwald, H. J. Phys.

Chem. B 1999, 103, 8888−8893.(18) Pavinatto, F. J.; Pavinatto, A.; Caseli, L.; dos Santos, D. S. Jr.;

Nobre, T. M.; Zaniquelli, M. E. D.; Oliveira, O. N. Jr.Biomacromolecules 2007, 8, 1633−1640.(19) Pavinatto, F. J.; Caseli, L.; Pavinatto, A.; dos Santos, D. S. Jr.;

Nobre, T. M.; Zaniquelli, M. E. D.; Silva, H. S.; Miranda, P. B.;Oliveira, O. N. Jr. Langmuir 2007, 23, 7666−7671.(20) Pavinatto, F. J.; Pacholatti, C. P.; Montanha, E. A.; Caseli, L.;

Silva, H. S.; Miranda, P. B.; Viitala, T.; Oliveira, O. N. Jr. Langmuir2009, 25, 10051−10061.(21) Pavinatto, A.; Pavinatto, F. J.; Barros-Timmons, A.; Oliveira, O.

N. Jr. ACS Appl. Mater. Interfaces 2010, 2, 246−251.(22) Krajewska, B.; Leszko, M.; Zaborska, W. Polim. Med. 1990, 20,

31−41.(23) Roberts., G. A. F.; Domszy, J. G. Int. J. Biol. Macromol. 1982, 4,

374−377.(24) Ehrlich, H.; Krajewska, B.; Hanke, T.; Born., R.; Heinemann, S.;

Knieb, C.; Worch, H. J. Membr. Sci. 2006, 273, 124−128.(25) Krajewska, B.; Zaborska, W.; Leszko, M. J. Mol. Catal. B: Enzym.

1997, 3, 231−238.(26) Davies, J. T.; Rideal, E. K. Interfacial Phenomena; Academic

Press: New York, 1963.(27) Sabnis, S.; Block, L. H. Polym. Bull. 1997, 39, 67−71.(28) Domszy, J. G.; Roberts, G. A. F. Makromol. Chem. 1985, 186,

1671−1677.(29) Miya, M.; Iwamoto, R.; Yoshikawa, S.; Mima, S. Int. J. Biol.

Macromol. 1980, 2, 323−324.(30) Elsabee, M. Z.; Morsi, R. E.; Al-Sabagh, A. M. Colloids Surf., B

2009, 74, 1−16.(31) Maltseva, E.; Shapovalov, V. L.; Mo hwald, H.; Brezesinski, G.

J. Phys. Chem. B 2006, 110, 919−926.(32) Seifert, M.; Breitenstein, D.; Klenz, U.; Meyer, M. C.; Galla, H.-

J. Biophys. J. 2007, 93, 1192−1203.(33) Grigoriev, D.; Krustev, R.; Miller, R.; Pison, U. J. Phys. Chem. B

1999, 103, 1013−1018.(34) Garidel, P.; Blume, A. Chem. Phys. Lipids 2005, 138, 50−59.(35) Zhang, Y.-P.; Lewis, R. N. A. H.; McElhany, R. N. Biophys. J.

1997, 72, 779−793.(36) Dicko, A.; Bourque, H.; Pezolet, M. Chem. Phys. Lipids 1998,

96, 125−139.(37) Helm, C. A.; Laxhuber, L.; Losche, M.; Mohwald, H. Colloid

Polym. Sci. 1986, 246, 46−55.(38) Nystrom, B.; Kjøniksen, A.-L.; Iversen, C. Adv. Colloid Interface

Sci. 1999, 79, 81−103.(39) Alatorre-Meda, M.; Taboada, P.; Sabin, J.; Krajewska, B.; Varela,

L. M.; Rodriguez, J. R. Colloids Surf., A 2009, 339, 145−152.

(40) Lucero, A.; Rodriguez Nino, M. R.; Gunning, A. P.; Morris, V.J.; Wilde, P. J.; Rodriguez Patino, J. M. J. Phys. Chem. B 2008, 112,7651−7661.(41) Koppenol, S.; Yu, H.; Zografi, G. J. Colloid Interface Sci. 1997,

189, 158−166.(42) Minones, J.; Sandez Macho, M. I.; Iribarnegaray, E.; Sanz

Pedrero, P. Colloid Polym. Sci. 1981, 259, 450−461.(43) Minones, J.; Sandez Macho, M. I.; Iribarnegaray, E.; Sanz

Pedrero, P. Colloid Polym. Sci. 1981, 259, 382−390.(44) Rey Gomez-Serranillos, I.; Min ones, J. Jr.; Dynarowicz-Łatka,

P.; Minones, J.; Iribarnegaray, E. Langmuir 2004, 20, 928−933.(45) Wydro, P.; Krajewska, B.; Hac-Wydro, K. Biomacromolecules

2007, 8, 2611−2617.(46) Parra-Barraza, H.; Burboa, M. G.; Sanchez-Vazquez, M.; Juarez,

J.; Goyocoolea, F. M.; Valdez, M. A. Biomacromolecules 2005, 6, 2416−2426.(47) Tsaih, M. L.; Chen, R. H. J. Appl. Polym. Sci. 1999, 73, 2041−

2050.(48) Ahrens, H.; Baltes, H.; Schmitt, J.; Mo hwald, H.; Helm, C. A.

Macromolecules 2001, 34, 4504−4512.

Biomacromolecules Article

dx.doi.org/10.1021/bm2012295 |Biomacromolecules 2011, 12, 4144−41524152