European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798

Contents lists available at ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics

journal homepage: www.elsevier .com/locate /e jpb

Research paper

Use of 1H NMR STD, WaterLOGSY, and Langmuir monolayer techniquesfor characterization of drug–zein protein complexes

0939-6411/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.ejpb.2013.07.008

⇑ Corresponding author. Department of Physical-Chemistry, School of Pharmacy,University of Santiago de Compostela, Campus Vida s/n, 15782 Santiago deCompostela, Spain.

E-mail addresses: Phabio_oliveira@yahoo.com.br (F.F.O. Sousa), asteriam.luzardo@usc.es (A. Luzardo-Álvarez), jose.blanco.mendez@usc.es(J. Blanco-Méndez), francisco.otero@usc.es (F.J. Otero-Espinar), manuel.martin@usc.es (M. Martín-Pastor), bebel.sandez@usc.es (I. Sández Macho).

F.F.O. Sousa a,b, A. Luzardo-Álvarez a, J. Blanco-Méndez a,c, F.J. Otero-Espinar a,c, M. Martín-Pastor d,I. Sández Macho e,⇑a Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Santiago de Compostela, Santiago de Compostela, Spainb Department of Biological and Health Sciences, School of Pharmacy, University Federal of Amapá, Macapá, Brazilc Institute of Industrial Pharmacy, University of Santiago de Compostela, Santiago de Compostela, Spaind Unidade de Resonancia Magnética, RIAIDT, University of Santiago de Compostela, Santiago de Compostela, Spaine Department of Physical-Chemistry, School of Pharmacy, University of Santiago de Compostela, Santiago de Compostela, Spain

a r t i c l e i n f o

Article history:Received 10 April 2013Accepted in revised form 14 July 2013Available online 24 July 2013

Keywords:BiopolymersZeinMaize proteinTetracyclineIndomethacinDrug–protein complexesNMRp–A isothermBrewster angle microscopy

a b s t r a c t

Zein is a protein based natural biopolymer containing a large amount of nonpolar amino acids, which hasshown the ability to form aggregates and entrap solutes, such as drugs and amino acids to form stableprotein–drug complexes. In this work, p–A isotherm, NMR, and Dynamic light scattering were used todetect the formation of protein aggregates and the affinity between zein and two different drugs: tetra-cycline and indomethacin. An effective interaction of zein and the two drugs was evidenced by means ofliquid NMR reinforced by means of changes in the surface pressure by p–A isotherm.

The effective interactions zein/drugs under air/water interface were evidenced as a change in the sur-face pressure of the p–A isotherm of zein in the presence of drug solutions. The presence of tetracycline inthe subphase decreased the area occupied by the monolayer at the expanded region until pressures of12 mN/m were the areas became similar, but indomethacin produces an increment of the area in bothexpanded and collapsed region.

The feasible methodology employed, focused in the functionality of the protein–drug interaction, canbe very promising in the drug delivery field.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

The technological advances in drug delivery demand constantinnovation on biomaterials and polymer science, mostly regardingspecial functionalities that may interest for the transport and re-lease of the ligand from the carrier systems. Recently, zein hasgained attention because of its potential applications as a biopoly-mer [1], being considered in the development of drug deliverysystems (DDS) such as microspheres, films, and other devices forthe release of antimicrobials, anticancer, anti-coagulants, and alsoparasitic drugs, mainly due to its cost and good standards ofbiocompatibility [2–8].

Zein, a prolamin rich protein, is found in protein bodies in theendosperm of the corn kernel (Zea mays L.). There are four classes

of this protein as follows: a, b, c, and d zein. The component a-zeinrepresents 35% of these proteins that include two prominent bandsof 22 and 24 kDa. Reducing SDS–PAGE analysis shows that b-zeinhas three major bands of 24, 22, and 14 kDa. The amino acid se-quences have been published by Phillips et col [9]. Approximately70% of zein is composed by glutamine (20%), leucine (20%), alanine(14%), and proline (9%), and due the high proportion of nonpolaramino acids, it is neither soluble in pure water or alcohol, but sol-uble in hydroalcoholic and acetone water mixtures. The amphiph-illic nature comes from the hydrophilic and hydrophobic residuesdistributed in the protein structure.

Its properties and structure allow zein to be a filmogenic mate-rial that is able to form stable, biodegradable, and biocompatiblefilms with great potential for they use in DDS, implants, food pack-aging, and electrochemical devices (selective sensors, biosensors,etc.). The average hydrophobicity of zein is reported to be 50 timeslarger than albumin, fibrinogen, etc. [10]. The properties of zeinfilms, such as biodegradability, mechanical resistance, waterabsorption, and barrier ability, largely depend on the interactionamong protein, plasticizers, and other functional groups [11]. Ithas been shown the ability of zein to form aggregates and entrap

F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798 791

solutes, drugs, and also amino acids [12]. This capability is directlyinfluenced by different conditions, such as the medium polarity[13], steric and electronic parameters [14], pH and charge of cat-ionic amino acids [15].

The monomer structure of zein protein as reported byMatsushima et al. [16] based on small-angle X-ray scattering(SAXS), presents 10 successive helical segments folded uponeach other in an antiparallel arrangement, stabilized by hydro-gen bonds. The helical segments are linked at each end by glu-tamine-rich turns or loops (Fig. 1). Elements of secondarystructure of zein have been disclosed, not only by SAXS but alsovia Infrared spectroscopy, Circular dichroism (CD), SDS–PAGEelectrophoresis, and Dynamic light scattering (DLS), suggestingthat the monomer adopts an extended structure in solution, aresult that is also consistent with an NMR qualitative study ofH/D exchange [17–29]. However, the structure of this proteinhas not been fully elucidated in the literature neither by highresolution methods such as X-ray crystallography or liquidNMR.

Due to the complex structure of zein, its amphiphilic characterand solubility characteristics, it is difficult to establish a homoge-neous distribution of drugs in zein-based DDS. Drugs are able toestablish interactions with zein and form complexes [17]; thus,gaining insight into these interactions would improve in the devel-opment of zein devices, concerning the drug distribution or theirrelease behavior. Its amphiphilic nature also permits to use p–Aisotherm to study the interaction in the air/water interface be-tween drugs and zein. Most recent, advances in NMR techniqueshave provided tools to probe specificity, affinity, and structural as-pects of receptor–ligand interaction [30], an adequate approach tounderstand the molecular recognition between drugs andbiomaterials.

Our main focus in this paper is to characterize by the aforemen-tioned NMR, p–A isotherm techniques, and Dynamic lightscattering the interactions of zein with small molecules of pharma-cological interest and study the potential of zein as biomaterial toelaborate multiparticulate drug delivery systems.

2. Materials and methods

2.1. Materials

Zein isolated from maize and tetracycline hydrochloride (TCL)were purchased from Sigma. Indomethacin (IND) was obtainedfrom Guinama. Ultrapure water was obtained by reverse osmosis(Millipore Corp.) and purified with a 0.22 lm zetapore filter. Theresistivity of the purified water was checked and kept at18 MX cm. All the other reagents were pure grade and used as re-ceived. Deuterated solvents D2O 99.9%, CD3OD 99.8%, and CDCl3

99.8% were purchased from Eurisotop.

Fig. 1. Structure of monomeric unit of zein (after Ref. [15]).

2.2. NMR Spectroscopy

2.2.1. Sample analysisSolution NMR spectra were obtained at 25 �C in an Agilent

(Varian) INOVA-750 operating at 750 MHz (proton frequency)spectrometer or a Bruker DRX-500 spectrometer operating at500 MHz (proton frequency). Both spectrometers were equippedwith a proton inverse detection probe with shielded pulse fieldgradients. The exact spectrometer used and the sample prepara-tion are indicated in each case. NMR data processing and analysiswas performed with MestreNova� software (Mestrelab ResearchInc.) [31].

2.2.2. 1H NMR signal assignmentProton NMR signal assignment of the pure drugs TCL or IND was

performed using a combination of three standard NMR spectra: 1D1H, TOCSY (Total Correlation Spectroscopy), and 2D NOESY (Nucle-ar Overhauser Enhancement Spectroscopy) acquired in the DRX-500 spectrometer. For these experiments, a sample of the puredrug IND or TCL was prepared at 7.5 mM concentration in the mix-ture CDCl3:CD3OD 90:10 (v/v). The 1H NMR signal assignment wasobtained from the analysis of the spectra assisted by the compari-son with previous assignments found in the literature for thesemolecules [34] and the theoretical prediction based in the molec-ular structure using MestreNova� software. The signal assignmentand numbering scheme used thorough the text for TCL and INDmolecules is given in Fig. 2.

2.2.3. Saturation Transferred Difference (STD) and water-ligand-observed via gradient spectroscopy (waterLOGSY)

These spectra were acquired in the Inova-750 spectrometer.Samples containing a mixture of a drug and zein protein, IND:zeinand TCL:zein, were dispersed in H2O:CD3OD 4:5 v/v. For discerningthe role of water, control samples IND:zein and TCL:zein wereprepared identically by replacing the protic solvent H2O by D2O(i.e. mixture of solvents D2O:CD3OD 4:5 v/v). The drug and proteinconcentrations used were 7.5 mM and 0.15 mM, respectively,equivalent to the ligand:protein molar ratio of �50:1.

Samples IND:zein were prepared by dissolving IND in 0.5 mLCD3OD. After stirring, 0.1 mL of water (either D2O or H2O) wasadded, and then, zein was added. Finally, 0.3 mL of water (eitherD2O or H2O) was added. The final product was transferred to anNMR tube for measurement.

Samples TCL:zein were prepared by dissolving of TCL in 0.1 mLof water (either H2O or D2O). After stirring, 0.5 mL of CD3OD wasadded and then zein was added. Finally, 0.3 mL of water (eitherH2O or D2O) was added. The final product was transferred to anNMR tube for measurement. A sample was prepared with the mix-ture of the two drugs, TCL and IND, and zein protein dissolved inH2O:CD3OD 4:5 v/v. The total ratio drug:zein in this sample wasthe same as in the other four samples.

Saturation Transferred Difference spectra (STD) [32,33] wereacquired for each one of the four samples prepared drug:zein.The selective saturation was placed at 0.87 ppm, a region of theproton spectrum where exclusively appear signals of zein but notTCL or IND. It provided the so-called STDon spectrum. Besides, a ref-erence 1D STD spectrum was acquired under the same conditionsexcept for the selective saturation that was placed in an empty re-gion of the spectrum at 15 ppm, providing the so-called STDoff

spectrum. The total duration of the saturation was 2 s. It consistedin a train of low power selective Gaussian pulses of 50 ms sepa-rated by a 0.1 ms delay. The scans corresponding to the STDon

and STDoff experiments were interleaved during the acquisition,and the correspondingly, FIDs are subtracted automatically bythe phase cycling providing the STDoff–on spectrum that is finally

(a)

(b)

Fig. 2. NMR proton signal assignment and numbering scheme for TLC and IND molécules used in the text.

792 F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798

Fig. 3. NMR spectra of the mixture IND:zein dissolved in H2O:CD3OD 4:5 v/v. (a) 1H reference spectrum. (b) STDoff–on spectrum and (c) water-LOGSY spectrum. The signalselectively saturated or inverted is indicated with a ray symbol. The assignment of the signals of IND is indicated, it is based in the NMR assignment of the pure compound inFig. 2.

F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798 793

reported. Each STDoff–on spectrum was acquired with 128 scans anda total duration of each scan of 4 s.

WaterLOGSY experiments [34,35] were acquired for the threesamples dissolved in H2O:CD3OD 4:5 v/v. They are the two samplesthat contain zein and a single drug, TCL or IND, and the sample thatcontains the mixture of two drugs and zein. The experiments wereperformed with a 180� inversion pulse applied over the watersignal at �4.7 ppm by means of a sinc shaped selective pulse of25 ms. The mixing time of the experiment was set to 2 s. EachwaterLOGSY spectrum was acquired with 128 scans and a totalduration of each scan of 5 s.

2.3. Compression isotherms

Isotherms were carried out using a surface balance singlebarrier NIMA 611 (U.K.) with total area 550 cm2 placed on ananti-vibration table. Prior to experiments, the trough was cleanedwith chloroform and ethanol and further rinsed with water. Themonolayer stability was verified by monitoring the change in sur-face pressure while holding the area constant. The subphases werepurified water, TCL, or IND solutions. Also, the temperature waskept at 20 �C. Solutions of zein (0.1 mg mL�1) were prepared inchloroform/methanol 9:1 v/v, To record the p–A isotherms, a vol-ume of 30 lL of the protein solution was deposited by means ofa microsyringe (Hamilton, USA) at the air/water interface and al-lowed to stand for at least 10 min in order to ensure the completeevaporation of the solvent. The monolayers were compressed at aspeed of 15 cm2 min�1. The surface pressure was measured withan accuracy of ±0.1 mN/m, using a Wilhelmy plate as a pressure

sensor, and the surface pressure, p, was recorded as a function ofthe area of the monolayer. To carry out relaxation measurements,monolayers were compressed at a speed of 15 cm2/min until thedesired surface pressure was achieved. The relationship betweenthe time and the area was recorded as the surface pressure wasmaintained automatically.

2.4. BAM images and thickness of monolayers

Brewster angle microscopy (BAM) images and ellipsometricmeasurements were performed with a BAM 2 Plus (NFT, Göttingen,Germany) equipped with a 30 mW laser emitting p-polarized lightat 532 nm wavelength which was reflected off the air/water inter-face at the Brewster angle (53.1�). This reflected beam pass througha focal lens, into an analyzer at a known angle of incident polariza-tion and finally to a CCD camera to measure gray levels (GL) in-stead of relative intensity (I). The light intensity at each point inthe BAM image depends on the local thickness and film opticalproperties. At the Brewster angle: I ¼ jR2

p j ¼ C � d2, where I is therelative intensity or relative reflectivity (defined as the ratio ofthe reflected intensity (Ir) and the incident intensity (I0), I = Ir/I0),Rp is the p-component of the light, C is a constant, and d is therelative film thickness. The relative reflectivity of the film wasmeasured with a calibrated charge-coupled device (CCD) camera.Therefore, it was determined the relationship between the graylevels (GL) and the relative reflectivity (I). The procedure used forthis calibration was described in previous articles [36], but cur-rently, with the BAM 2 Plus equipment, the calibration is automat-ically performed for each measurement at different shutter speeds,

(b)

(a)

(c)

Fig. 4. NMR spectra of the mixture TCL:zein dissolved in H2O:CD3OD 4:5 v/v. (a) 1H reference spectrum. (b) STD off–on spectrum and (c) water-LOGSY spectrum. The signalselectively saturated or inverted is indicated with a ray symbol. The assignment of the signals of TCL is indicated, it is based in the NMR assignment of the pure compound inFig. 2.

794 F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798

providing the relative thickness values of the monolayer as it iscompressed. The lateral resolution of the microscope was 2 lm,the shutter speed used was 1/50 s, and the images were digitalized(768 � 572 pixels) and processed to optimize the quality (Iris v5.34software, Christian Buil, France).

2.5. Dynamic light scattering

Drug–protein aggregates were characterized by means of sizeand zeta potential measurements using Dynamic light scattering(Zetasizer� Nano-ZS, Malvern Instruments). In order to obtain theaggregates, we prepared a hydroalcoholic (70%) solution of zeinof a concentration of 8 mM. From this solution, an aliquot of60 lL was dropped to a vial containing a drug solution of 8 mM(TCL or IND) and stirred vigorously. By this means, it was obtaineda ratio zein:drug at approximately 1:10 each.

3. Results and discussion

There are available several NMR experiments for the detectionor characterization of macromolecule-ligand or supramolecular-li-gand interactions that relies in the observation of the NMR signalsof the ligand [32–34,37–39]. The Saturation Transfer Difference(STD) [32,33] and water-LOGSY [34,38] experiments are two sensi-tive and reliable experiments of this class. Both experiments arebased in the Nuclear Overhauser Effect (NOE) and are able to detectligand binding to a macromolecule or supramolecular aggregate in

the range of medium-to-weak affinity, corresponding to an equilib-rium dissociation constant (Kd) from lM to mM [33].

The STD experiment provides a simpler access to the ligandbinding epitope and to other relevant aspects such as the ligandbinding affinity [40]. A previous study performed in our grouphad used the STD methodology and found that in the complexesTCL–zein and IND–zein, the aromatic ring of TCL and the aromaticring containing chlorine of IND are, respectively, the most relevantparts of these drugs for its interaction with zein in D2O:CD3ODsolutions [17]. The same study had also determined that bothdrugs present medium-to-weak affinity for zein, whereas the affin-ity of IND is an order of magnitude higher than TCL.

The water-LOGSY experiment is an alternative method to STDfor the detection of binding for ligands to macromolecularreceptors [34,35]. It relies in the transference of NOE from thewater resonance. In a sample containing a small molecule and amacromolecular receptor, the appearance in the spectrum of sig-nals of the small molecule with the same phase than the waterpeak proves that there is binding affinity among the two moleculesand that water is present at the macromolecule binding site [41].

In this section, solution NMR methods are used to gain insightabout the drug:protein mechanism of interaction and the role ofthe water co-solvent in the complexes TCL–zein and IND–zein.STD and water-LOGSY spectra were used to obtain informationabout the intermolecular interactions between a drug, TCL orIND, and zein under the relevant conditions in which the sampleis forming a colloidal dispersion, as it is required for the formation

(a)

(b)

(c)

Fig. 5. Aromatic region of the water-LOGSY spectrum of colloidal samples drug:zein in the mixture of solvents H2O:CD3OD. (a) IND:zein, (b) TCL:zein, and (c) IND:TCL:zein ina ratio IND:TCL 1:1. The signals of TCL and IND are indicated with filled squares and open triangles, respectively.

F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798 795

of monolayers. For this mean, samples IND:zein and TCL:zein wereprepared in the mixture of solvents H2O:CD3OD 4:5 v/v. To discernthe role of water, two analog samples were also prepared replacingthe water protic solvent (H2O) by deuterated water (D2O).

0 5 10 15 20 25 300

5

10

15

20

25

30INDTCLWater

A2/residue

(mN

/m)

Water TLC

IND

Water TLC

IND

Water TLC

IND

Water TLC

IND

Fig. 6. Zein p–A isotherm and BAM images obtained in water and in presence of thedrugs TCL and IND.

The STDoff–on spectra obtained for samples IND:zein andTCL:zein by placing the on-saturation in the aliphatic signal of zeinat�0.87 ppm are given in Figs. 3 and 4, respectively. It is clear fromthe comparison of the spectra of Fig. 3b and c (or Fig. 4b and c) thatthe STD signals corresponding to the drug are much weaker whenthe sample is prepared co-solvent H2O than when they are pre-pared in co-solvent D2O. This unexpected result indicates that forthese colloidal samples, the water co-solvent remarkably affectsto the mode of interaction of the drugs TCL or IND with zein. Themost plausible interpretation suggests that the form of the waterco-solvent, D2O or H2O, affects to the conformation of zein. Sinceit is known that the strength of hydrogen-bonding interactionsweakens in D2O respect to H2O, zein could be more unfolded inD2O than in H2O co-solvent and therefore have different residuesexposed for the interaction with the drug. The STD spectra suggestthat the direct interaction of the drug with zein is diminished inH2O respect to D2O co-solvent.

To obtain further information about the mode of binding of thedrug, waterLOGSY spectra were acquired for the colloidal disper-sions drug:zein prepared with the H2O co-solvent. In the water-LOGSY spectrum of samples IND:zein and TCL:zein (Fig. 5a andb), most of the signals of the drug have considerably intensityand the same phase as the water peak. This is an indication thatwith the H2O co-solvent, there is binding between the drug andthe zein mediated by a layer of water molecules that have relativelong residence times attached to the hydrophilic surface of the pro-tein (bound water fraction). Qualitatively, considering the poortransference of saturation from the protein to the drug in theSTDoff–on spectra obtained with the H2O co-solvent (Fig. 3a and

0 60 120 180 240 300 360 420 480 540 6000

5

10

15

20

0

5

10

15

20

25

30

waterTCLIND

Isoterm-time

Thickness-time

Time (sec.)

thic

knes

s (n

m)

(mN

/m)

Fig. 7. Compressibility values and thickness–time curves of zein molecules extends in the air/water, air/IND and air/TCL solution interface.

796 F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798

b), the mode of binding mediated by the water molecules must bepredominant when the sample is prepared in H2O co-solvent.

A waterLOGSY competition experiment was performed with asample containing the two drugs, TCL and IND, and zein in a colloi-dal dispersion in H2O co-solvent (Fig. 5c). This spectrum shows thepresence of the signals the two drugs with the same phase as thewater peak. Moreover, the signals of each drug appear with similarintensity as that previously seen in the waterLOGSY spectrum ofthe sample containing each individual drug (Figs. 3 and 4). This re-sult indicates that the two drugs IND and TCL do not compete forthe binding to zein in the colloidal dispersions with H2O co-sol-vent, a very different situation of what it was observed by us inthe co-solvent D2O [17]. The later suggests that the binding ofthe drug in H2O co-solvent is rather unspecific, and possibly to anumber of water solvated and relative amphiphilic residues ex-posed at the surface of zein.

Zein molecules are able to extend in the A/W interface forminga stable monolayer. It is represented in Fig. 6, the compressioncurve obtained for this protein. The shape of the isotherm denotes

Fig. 8. Model of structure of monomeric unit of zein at the air water

the existence of a liquid condensate state monolayer in the rangeof surface pressure between 12 mN/m and 27 mN/m, as indicatedin the value of the compressibility modulus (C�1

s , Fig. 7) [2] whichrepresents the slope of the p–A curve. The area liftoff (where thesurface pressure starts to increase) is 26 Å2/residue. The limit areaof the monolayer (extrapolated area to pressure 0) is 20 Å2/aminoacid residue. This is a typical value for proteins monolayers inroom conditions when they are well extended in air/water inter-face (A/W) [42,43]. Assuming an average molecular weight of27 kDa for the protein molecule and 113 Da and for each aminoacid residue and a number of residues of amino acid by proteinmolecule of 239, the liftoff value of 26 Å2/residue correspondsapproximately to an area of 600 Å2/molecule [44]. In accordancewith Mitchell [45], a value of 26 Å2/residue for the area liftoff cor-responds to the area of zein molecules when a reorientation ofthe side chains occurs due to the effect of the monolayer com-pression: polar residues are oriented toward the water subphaseunder the main chain and nonpolar residues toward the air(Fig. 8a).

interface when is deposited over (a) water, (b) IND and (c) TCL.

10 15 20 250

5

10

15

20

25

firstsecondthirdfourth

A2/residue

(mN

/m)

π

Fig. 9. Hysteresis cycle of p–A isotherms of zein molecules extends in the air/waterinterface.

F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798 797

Above the equilibrium extension pressure, estimated as19.5 mN/m [45], an inflection in the p–A isotherm is observed,remaining relatively stable until pressures of 24.5 mN/m. Thisbehavior could be due that helical parties and unroll chains of zeinthat are rearranged to an upright position up to an area at whichthey start to compress, or due to the formation of a bilayer of coilparts. Such behavior was already described for cyclic moleculessuch as cyclosporine, which featured a ‘‘quasi plateau’’ [46]. In or-der to determinate the transition values more accurately, the com-pressibility modulus was obtained [47], showing a reduction ofabout a quarter in the value of the area from the beginning untilthe end of the plateau which is incompatible with the formationof the bilayer. To gain more insight into the interpretation, thethickness-time curves were obtained (Fig. 7), showing a constantthickness during the entire plateau, indicating that a bilayer wasnot formed. BAM images taken throughout the plateau (Fig. 6)show a homogeneous system without appearance of collapsedmonolayer, confirming that the plateau can be attributed to a chainreorientation, forming ‘‘loops’’ and ‘‘tails’’ in the air/water inter-face. This behavior can also explain the quasi-reversible processesobserved in the hysteresis cycles of compression curves (Fig. 9).

Zein p–A isotherm obtained when TCL and IND are added to thesubphase showed significant changes compared to water alone(Fig. 6). These modifications are not related to the physical stateof monolayer, but rather in the occupied area. TCL produces adeclining in the area occupied by the monolayer at low pressureuntil values around 12 mN/m when the area becomes similar tothe one obtained with pure water. Contrary to TCL, IND producesan increment in the area of the monolayer compared to water,along all the pressure values assayed. These differences in mono-layer area are indicative that both TCL and IND interact with zein,but possibly at different places in the structure of the protein, as ithas been confirmed by the NMR studies.

In the air interface, zein molecules oriented their polar residuestoward water subphase and nonpolar residues toward the air. Inthis situation, IND molecules interact with the hydrophobic partsof the protein (Fig. 8b) bringing the area occupied by themonolayer of zein larger than the same in water alone. So the

Table 1Mean size and superficial charge properties of zein and zein/drug aggregates. Valuesare means ± SD (n = 3).

Formulation Particle size (nm) Zeta potential (mV)

Zein 288.0 ± 3.7 17.1 ± 0.1Zein–TCL 380.0 ± 25.5 9.1 ± 0.1Zein–IND 501.5 ± 9.1 13.9 ± 0.8

hydrophobic parts of IND are interacting to de zein residues inthe protein–air surface avowing the contact with water, this situa-tion can strongly influence the drug entrapment in zein micropar-ticles and their subsequent release. However, TCL, which is morehydrophilic, interacts with the residues oriented toward water(Fig. 8c) favoring the hydrophilic attractions between chain andlowering the area occupied by the monolayer; nevertheless, whenthe monolayer is compressed, both compression of side chains andchains folding cannot be increased, and so the isotherm in presenceof TCL becomes similar to water on values of pressure higher than12 mN/m.

Zein–drug aggregates were also studied in size and superficialcharge. The results obtained using DLS show that aggregates havedifferences in diameter depending to the drug entrapped (Table 1)that is in accordance with the assembly of the drug in the proteincomplex proposed in monolayer experiments. Thus, it seems thatIND is involved in the expansion of the polymeric chains of the pro-tein, while TCL keeps the morphology and size distribution closerto what is found with the zein aggregates in absence of drugs. Inany case, the aggregates obtained were uniform in size, with poly-dispersity index (PdI) ranging from 0.027 to 0.070. Zeta potentialwas also determined found to be positive and quite similar inaggregates of zein and in IND/zein aggregates but lower in the caseof TCL/zein aggregates.

4. Conclusion

The affinity between a drug and its carrier is a key parameter onthe development of a drug delivery system since it affects impor-tant aspects such as loading efficiency, release, and distribution.The formation of a protein–drug complex is the interplay of a num-ber of molecular interactions that utterly contributes to the ther-modynamics of the process such as the establishment ofhydrogen bonds, van der Waals forces [48], the release of solventmolecules and the entropic balance. The net contribution of all ofthem must be favorably in order to form the complex. Even whenthe structure of the two counterparts is known experimentally, theprediction of the affinity is still limited [49]. Some of the reasonsthat make difficult such prediction are the fact that the interactionsdoes not only depends on the structure but also depends on thedynamics of both components. In addition, there are difficultiesfor the evaluation of the enthalpy and entropic contribution ofthe solvent. In this study is show that at difference that in solution[17], were aromatic–aromatic interactions are the driving force forthe stabilization of drug/zein complexes, in colloidal state theinteractions drug–protein are rather unspecific and stronglydependent on drug hydrophilic–hydrophobic residues, the hydra-tion of the water solvated amphiphilic residues of zein and alsothe interaction with the hydrophobic parts of the protein in theair interphase. Moreover, molecular affinity and the protein/drug/water interactions are important parameters affecting the incorpo-ration of the drug into the protein complex.

Current NMR and p–A Isotherms techniques are suitable toolsto presume and determine binding interactions and thermody-namic parameters such as the affinity among molecules. Thisstrategy can be useful for predicting the drug incorporation charac-teristics from a polymeric drug delivery system. This feasiblemethodology, focused in the functionality of the protein–drugcomplex, does not require any knowledge of its three-dimensionalstructure and can be very promising within the drug delivery field.

Acknowledgement

Francisco Fábio Oliveira de Sousa was supported by the Brazil-ian Agency Capes, Grants No. 4668/06-5.

798 F.F.O. Sousa et al. / European Journal of Pharmaceutics and Biopharmaceutics 85 (2013) 790–798

References

[1] A. Gennadios, L.C. Weller, Edible films and coatings from wheat and cornproteins, Food Technol. 44 (1990) 63–69.

[2] Y. Matsuda, T. Suzuki, E. Sato, M. Sato, S. Koizumi, K. Unno, T. Kato, K. Nakai,Novel preparation of zein microspheres conjugated with P S-K available forcancer immunotherapy, Chem. Pharm. Bull. 37 (1989) 757–759.

[3] T. Suzuki, E. Sato, Y. Matsuda, H. Tada, K. Unno, T. Kato, Preparation of zeinmicrospheres conjugated with antitumor drugs available for selective cancerchemotherapy and development of a simple colorimetric determination ofdrugs in microspheres, Chem. Pharm. Bull. 37 (1989) 1051–1054.

[4] X. Liu, Q. Sun, H. Wang, L. Zhang, J.Y. Wang, Microspheres of corn protein, zein,for an ivermectin drug delivery system, Biomaterials 26 (2005) 109–115.

[5] H.J. Wang, Z.X. Lina, X.M. Liub, S.Y. Shenga, J.Y. Wang, Heparin-loaded zeinmicrosphere film and hemocompatibility, J. Control. Release 105 (2005) 120–131.

[6] J. Dong, Q. Sung, J.Y. Wang, Basic study of corn protein, zein, as a biomaterial intissue engineering, surface, morphology and biocompatibility, Biomaterials 25(2004) 4691–4697.

[7] F.F. Sousa, A. Luzardo-Alvarez, A Pérez-Estévez, R. Seoane-Prado, J Blanco-Méndez, Development of a novel AMX-loaded PLGA/zein microsphere for rootcanal disinfection, Biomed. Mater. 5 (2010) 1–10. 055008.

[8] F.F. Sousa, J. Blanco-Méndez, A. Pérez-Estévez, R. Seoane-Prado, A. Luzardo-Álvarez, Effect of zein on biodegradable inserts for the delivery of tetracyclinewithin periodontal pockets, J. Biomater. Appl. 2 (2012) 187–200.

[9] R.L. Phillips, B.A. McClure, Elevated protein-bound methionine in seeds of amaize line resistant to lysine plus threonine, Cereal Chem. 62 (1985) 213–218.

[10] K.A. Tilley, R.E. Benjamin, K.E. Bagorogoza, B.M. Okot-Kotber, O. Prakash, H.Kwen, Tyrosine cross-links: molecular basis of gluten structure and function, J.Agric. Food Chem. 49 (2001) 2627–2632.

[11] Y. Wang, A.M. Rakotonirainy, G.W. Padua, Thermal behavior of zein-basedbiodegradable films, Starch – Stärke 55 (2003) 25–29.

[12] B.L. Hsu, Y.M. Weng, Y.H. Liao, W. Chen, Structural investigation of edible zeinfilms/coatings and directly determining their thickness by, J. Agric. Food Chem.53 (2005) 5089–5095.

[13] Q. Wang, P Geil, G. Padua, Role of hydrophilic and hydrophobic interactions instructure development of zein films, J. Polym. Environ. 12 (2004) 197–202.

[14] E. Forgács, T. Cserháti, Z. Deyl, I. Miksík, Binding of substituted phenol andaniline derivatives to the corn protein zein studied by high-performance liquidchromatography, J. Chromatogr. B, Biomed. Sci. 753 (2001) 79–86.

[15] T. Cserháti, E. Forgács, Effect of pH and salts on the binding of free amino acidsto the corn protein zein studied by thin-layer chromatography, Amino acids 28(2005) 99–103.

[16] N. Matsushima, G. Danno, H. Takezawa, Y. Izumi, Three-dimensional structureof maize alpha-zein proteins studied by small-angle X-ray scattering, Biochim.Biophys. Acta 1339 (1997) 14–22.

[17] F.F. Sousa, A. Luzardo-Álvarez, J. Blanco-Méndez, M. Martí-n-Pastor, NMRtechniques in drug delivery: application to zein protein complexes, Int. J.Pharm. 439 (2012) 41–48.

[18] B.L. Hsu, Y.M. Weng, Y.H. Liao, W. Chen, Structural investigation of edible zeinfilms/coatings and directly determining their thickness by FT-Ramanspectroscopy, J Agric. Food Chem. 53 (2005) 5089–5095.

[19] M.R. Bugs, L.A. Forato, R.K. Bortoleto-Bugs, H. Fischer, Y.P. Mascarenhas, R.J.Ward, L.A. Colnago, Spectroscopic characterization and structural modeling ofprolamin from maize and pearl millet, Eur. Biophys. J. 33 (2004) 335–343.

[20] L.A. Forato, C. Bicudo Tde, L.A. Colnago, Conformation of alpha zeins in solidstate by Fourier transform IR, Biopolymers 72 (2003) 421–426.

[21] V. Cabra, R. Arreguin, R. Vazquez-Duhalt, A. Farres, Effect of temperature andpH on the secondary structure and processes of oligomerization of 19 kDaalpha-zein, Biochim. Biophys. Acta 1764 (2006) 1110–1118.

[22] C. Gao, M. Stading, N. Wellner, M.L. Parker, T.R. Noel, E.N. Mills, P.S. Belton,Plasticization of a protein-based film by glycerol: a spectroscopic, mechanical,and thermal study, J. Agric. Food Chem. 54 (2006) 4611–4616.

[23] L.A. Forato, R. Bernardes-Filho, L.A. Colnago, Protein structure in KBr pellets byinfrared spectroscopy, Anal. Biochem. 259 (1998) 136–141.

[24] D.J. Sessa, A. Mohamed, J.A. Byars, Chemistry and physical properties of melt-processed and solution-cross-linked corn zein, J. Agric. Food Chem. 56 (2008)7067–7075.

[25] K. Shi, J.L. Kokini, Q. Huang, Engineering zein films with controlled surfacemorphology and hydrophilicity, J. Agric. Food Chem. 57 (2009) 2186–2192.

[26] D.M. Georget, S.A. Barker, P.S. Belton, A study on maize proteins as a potentialnew tablet excipient, Eur. J. Pharm. Biopharm. 69 (2008) 718–726.

[27] Y. Mizutani, Y. Matsumura, H. Murakami, T. Mori, Effects of heating on theinteraction of lipid and zein in a dry powder system, J. Agric. Food Chem. 52(2004) 3570–3576.

[28] S. Subramanian, S. Sampath, Adsorption of zein on surfaces with controlledwettability and thermal stability of adsorbed zein films, Biomacromolecules 8(2007) 2120–2128.

[29] P. Argos, K. Pedersen, M.D. Marks, B.A. Larkins, A structural model for maizezein proteins, J. Biol. Chem. 257 (1982) 9984–9990.

[30] B. Meyer, T. Peters, NMR spectroscopy techniques for screening and identifyingligand binding to protein receptors, Angew. Chem. Int. Ed. Engl. 42 (2003)864–890.

[31] J.C. Cobas, F. Sardina, Nuclear magnetic resonance data processing. MestRe-C:a software package for desktop computers., J, Concepts Magn. Reson. 19A(2003) 80–96.

[32] M. Mayer, B. Meyer, Characterization of ligand binding by saturation transferdifference NMR spectroscopy, Angew. Chem. Int. Ed. 38 (1999) 1784–1788.

[33] B. Meyer, T. Peters, NMR spectroscopy techniques for screening andidentifying ligand binding to protein receptors, Angew. Chem. Int. Ed. 42(2003) 864–890.

[34] C. Dalvit, G. Fogliatto, A. Stewart, M. Veronesi, B. Stockman, WaterLOGSY as amethod for primary NMR screening: practical aspects and range ofapplicability, J. Biomol. NMR 21 (2001) 349–359.

[35] E.C. Johnson, V.A. Feher, J.W. Peng, J.M. Moore, J.R. Williamson, Application ofNMR SHAPES screening to an RNA target, J. Am. Chem. Soc. 125 (2003) 15724–15725.

[36] P. Toimil, G. Prieto, J. Miñones Jr., J.M. Trillo, F.l. Sarmiento, Monolayer andBrewster angle microscopy study of human serum albumin/dipalmitoylphosphatidyl choline mixtures at the air/water interface, Colloids Surf. B:Biointerf. 92 (2012) 64–73.

[37] C.B. Post, Exchange-transferred NOE spectroscopy and bound ligand structuredetermination, Curr. Opin. Struct. Biol. 13 (2003) 581–588.

[38] C. Dalvit, P. Pevarello, M. Tato, M. Veronesi, A. Vulpetti, M. Sundström,Identification of compounds with binding affinity to proteins viamagnetization transfer from bulk water, J. Biomol. NMR 18 (2000) 65–68.

[39] P. Thordarson, Determining association constants from titration experimentsin supramolecular chemistry, Chem. Soc. Rev. 40 (2011) 1305–1323.

[40] J. Angulo, P.M. Enríquez-Navas, P.M. Nieto, Ligand–receptor binding affinitiesfrom saturation transfer difference (STD) NMR spectroscopy: the bindingisotherm of std initial growth rates, Chem. – Eur. J. 16 (2010) 7803–7812.

[41] C. Ludwig, P.J.A. Michiels, X. Wu, K.L. Kavanagh, E. Pilka, A. Jansson, U.Oppermann, U.L. Günther, SALMON: solvent accessibility, ligand binding, andmapping of ligand orientation by NMR spectroscopy, J. Med. Chem. 51 (2007)1–3.

[42] D.F. Cheesman, J.T. Davies, M.L. Anson, T.E. John, Physicochemical andbiological aspects of proteins at interfaces, in: Advances in ProteinChemistry, Academic Press, 1954, pp. 439–501.

[43] J.M. Trillo, E.I. Jado, S.G.a. Fernández, S.S. Pedrero, Monolayers of human serumalbumin, Kolloid-Z. Z. Polym. 250 (1972) 318–324.

[44] I. Pezron, L. Galet, D. Clausse, Surface interaction between a protein monolayerand surfactants and its correlation with skin irritation by surfactants, J. ColloidInterface Sci. 180 (1996) 285–289.

[45] J.S. Mitchell, The structure of protein monolayers, Trans. Faraday Soc. 33(1937) 1129–1139.

[46] T.S. Wiedmann, T. Trouard, S.C. Shekar, M. Polikandritou, Y.-E. Rahman,Interaction of cyclosporin A with dipalmitoylphosphatidylcholine, Biochim.Biophys. Acta (BBA) – Biomembr. 1023 (1990) 12–18.

[47] J.T. Davies, E.K. Rideal, Interfacial Phenomena, Academic Press, 1963.[48] F.A. Quiocho, Carbohydrate-binding proteins: tertiary structures and protein–

sugar interactions, Annu. Rev. Biochem. 55 (1986) 287–315.[49] N.R. Syme, C. Dennis, A. Bronowska, G.C. Paesen, S.W. Homans, Comparison of

entropic contributions to binding in a ‘‘hydrophilic’’ versus ‘‘hydrophobic’’ligand–protein interaction, J. Am. Chem. Soc. 132 (2010) 8682–8689.