A

cpa(esotnh©

K

1

areis[

tb

c

0d

Available online at www.sciencedirect.com

Colloids and Surfaces B: Biointerfaces 64 (2008) 184–193

Nucleotides and nucleolipids derivatives interaction effectsduring multi-lamellar vesicles formation

Francesca Cuomo a,∗, Francesco Lopez a, Ruggero Angelico a,Giuseppe Colafemmina a,b, Andrea Ceglie a,∗

a Consorzio Interuniversitario per lo sviluppo dei Sistemi a Grande Interfase (CSGI), c/o Department of Food Technology (DISTAAM),Universita del Molise, I-86100 Campobasso, Italy

b Deptartment of Chemistry, Universita di Bari, I-70126 Bari, Italy

Received 28 September 2007; accepted 20 January 2008Available online 3 February 2008

bstract

In this paper a micellar interface, constituted by the cationic surfactant CTAB, in presence of 1,2-epoxydodecane and nucleotides was used foratanionic multi-lamellar vesicles (MLVs) formation. The micellar solution of CTAB is able to disperse the 1,2 epoxydodecane in the micellar coreromoting the reaction of this reagent with the nucleotide attracted by the positive surface charge of the micellar aggregates. The alkylation of AMPnd UMP nucleotides leads to the synthesis of nucleolipids. The behaviour of the supramolecular structures formed depends on the starting reagentsAMP, UMP and AMP + UMP) and on the assembly capabilities of the products. In particular nucleotides and nucleotides derivatives interactionffects are evaluated during the multi-lamellar vesicles formation. NMR spectroscopy and UV–vis measurements performed on MLVs showedtrong aryl interactions. Interestingly, NMR spectra revealed prevailing stacking interactions between complementary nucleolipids. The assemblyf complementary nucleotides affects the course of the reaction during the MLVs formation. Moreover the MLVs supramolecular stability has been

ested by means of turbidity and UV–vis measurements. In particular, an enhanced stability has been found in systems prepared with complementaryucleotides confirming that in these systems the self-assembly process is influenced by nucleolipids interactions. Furthermore by following theypocromic effect during the micellar catalysis, we showed that even in the earlier stages of the reaction significant differences are detectable.

2008 Elsevier B.V. All rights reserved.

ffect;

Ac[cewi[

eywords: Nucleotides; Nucleolipids; Microemulsion; MLVs; Hypochromic e

. Introduction

The study of the interactions between nucleotides and self-ssembled nanostructures such as micelles and vesicles hasecently been object of remarkable interest in the colloidal sci-nce. Moreover, a number of significant applications stresses themportance of employing liposomes and vesicles with increasedtability as drug delivery of non-viral vectors in gene therapy1,2].

The use of phospholipid vesicles presents some disadvan-ages; i.e., the need of sonication or extrusion of lipids to formilayers that finally may give rise to non-equilibrium structures.

∗ Corresponding author. Tel.: +39 0874404634; fax: +39 0874404652.E-mail addresses: francesca.cuomo@unimol.it (F. Cuomo),

eglie@unimol.it (A. Ceglie).

vansttte

927-7765/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfb.2008.01.020

NMR

n alternative to phospholipid vesicles is offered by surfactantatanionic vesicles formed by cationic and anionic surfactants3,4]. In these systems, the spontaneous vesicle formation pro-ess obtained by mixing these amphiphilic units [5] and thenhancement of the aggregates stability formed in this wayere emphasized. Moreover, favourable interactions of catan-

onic vesicles with nucleic acids have been recently studied6,7].

In this paper, we propose the spontaneous formation ofesicles by means of the “in situ” synthesis of one of themphiphilic components, namely, anionic surfactant bearingucleobase residues in its head polar group. The resultingelf-assembly properties are enriched because in addition to

he balance between hydrophilic and hydrophobic interactions,he aggregates, formation is influenced also by the contribu-ion of nucleobase interactions to the spontaneous curvaturenergy.

F. Cuomo et al. / Colloids and Surfaces B

Scheme 1. Schematic representation of micellar catalysis. 1,2-epoxydodecaneisp

aonOnbbrtctpfi

ruFhcapnmt

cc

b(steXhpemtip

cmttaaBe1Tdbi

cetiss

2

2

uaSsaXmaHD

2

s solubilized in the micellar core. Cationic trimethyl ammonium groups areurrounded by bromide and mononucleotidic anions (upper panel). In the loweranel the molecular structures of different alkylated nucleotides are reported.

In the present study, adenosine 5′-monophosphate (AMP)nd uridine 5′-monophosphate (UMP) have been used to ring-pen an hydrophobic oxirane, i.e., 1,2-epoxydodecane, owing toucleophilic attack of either the phosphate group, N-positions or-position of heterocyclic bases (Scheme 1) [8]. The reaction isot stereoselective and a mixture of four diastereoisomers shoulde obtained in an equimolar ratio. In order to favour the contactetween ionic nucleotides and the water insoluble organic oxi-ane, the cationic surfactant CTAB has been employed which inurn offers a suitable micellar interface where both the reactantsan meet and react. As a consequence, the reaction products inhe form of anionic surfactant monomers, accumulate into there-existent cationic surface to form catanionic bilayers in thenal state.

In this respect the formulation of these aggregates givesise to interesting elucidation for establishing the intermolec-lar interaction between nucleotides and amphiphilic species.or example molecular recognition, referred to Watson–Crickydrogen bond base pairing and aryl interaction between adja-ent base pairs [9], does not occur at room temperature inqueous solutions [10] of free nucleotide monomers but theresence of functionalised interfaces, characterised by a sort ofucleobase order, may represent an essential prerequisite. Thisight be realized with self-assembling molecules such as surfac-

ants having a nucleobase as hydrophilic moiety (nucleolipids).Basically, synthesis of lipids functionalized with nucleobases

an be performed through either enzymatic catalysis [11] orhemical synthesis [12–14] by linking covalently an hydropho-

2dM

: Biointerfaces 64 (2008) 184–193 185

ic residue (medium–long hydrocarbon chain) to a nucleobasepurine/pyrimidine) or by ion-exchange reaction on a cationicurfactant, where the ordinary counterion (Br−, Cl−, AcO−,he most common) is replaced by the anionic phosphate moi-ty of 5′-monophosphate-ribonucleosides (hereafter XMP with= adenine, uracil) [15]. Baglioni and co-workers [16–18]

ave carried out the synthesis of functionalized anionic phos-holiponucleosides by following the former procedure, i.e.,xchange of a choline headgroup with a nucleoside in a lecithinolecule catalyzed by phospholipase, and they have investigated

heir self-organization properties in aqueous solution by vary-ng hydrocarbon chain length and other chemical and physicalarameters [19].

In this paper we propose the possibility of using a non-onventional medium for the formation of catanionic giantulti-lamellar vesicles (MLVs). This scenario strongly suggests

o investigate the effect in terms of vesicles stability but above allo provide further information concerning various types of inter-ctions. We also investigated the behaviour of the reaction using1:1 mixture of the AMP and UMP nucleotides (MIX). In fact,aglioni and co-workers in recent papers reported experimentalvidences of preferential interactions in systems containing the:1 mixture of complementary phospholiponucleosides [16,18].he ability to synthesize bilayer forming dialkylphosphates fromirect reaction, catalyzed by cationic micelles (CTAB or DTAB),etween inorganic phosphate anion or dextran and the waternsoluble 1,2-epoxydes was previously reported [20,21].

The present paper in particular refers to the effects on theatanionic MLVs formation of different nucleotides and 1,2-poxydodecane in presence of a CTAB micellar solution andhe interactions between the new synthesized products werenvestigated by means of optical microscopy, dialysis, 1H NMRpectroscopy, UV–vis spectrophotometry and turbidity mea-urements.

. Experimental

.1. Materials

Adenosine 5′-monophosphate (AMP, ≥99%, disodium salt),ridine 5′-monophosphate (UMP, >99%, disodium salt), hex-decyltrimethylammonium bromide (CTAB, ≥99%) were fromigma. 1,2-epoxydodecane (≥95%) was from Aldrich. All thetudies have been performed at pH 7–8 where ribonucleotidesre expected to be predominantly in their fully ionized form:MP2− (X = A, U). Dialysis membranes (Spectra/Por 3) with aolecular weight cut off at 3500 Da (g mol−1) and an effective

rea of 49.2 cm2 were purchased from SPECTRUM (Lagunaills, CA, USA). Equilibrium dialysis cells were supplied byIANORM, Zurich, Switzerland.

.2. MLVs preparation

MLVs were prepared by adding 13 mM of nucleotides to a7 mM CTAB micellar solution and then 40 �L of 1,2 epoxy-odecane, for a final concentration of 18 mM, were added.LVs prepared with both the complementary nucleotides

1 ces B

(oc1tMm(sm

2

Nol2psiicpaawtMttcrca

2

oldatcca

2

8Cieri

L

asiCpt

2

Iaral

2

Ahfiwc2

3

3

cmottddMgstttoiw

86 F. Cuomo et al. / Colloids and Surfa

AMP + UMP) were prepared with 6.5 mM of AMP and 6.5 mMf UMP. Samples prepared with both nucleotides were indi-ated with the term MIX. The mixtures were kept at 25 ◦C forweek under stirring; within this period the systems showed

he appearance of an opaque suspension, thus indicating theLVs formation. Samples for hypochromic effects measure-ents were prepared by adding known amount of nucleotides

0.2 mM) to a water/CTAB/1,2-epoxydodecane microemul-ion ([CTAB] = 27 mM, [1,2-epoxydodecane] = 4.5 mM). Theicroemulsion was previously stirred for 40 min.

.3. Dynamic light scattering measurements

DLS measurements were carried out by a Zetasizer-ano-S90 (Malvern Instruments, Malvern U.K.) consistingf an Avalanche photodiode tube and a 4 mW He–Neaser (λ = 633 nm). The scattering cell was thermostated at5.0 ± 0.1 ◦C with a Peltier element. All the experiments wereerformed at the scattering angle of 90◦; other settings were:olvent (water) viscosity 0.887 mPa s and solvent refractivendex 1.33. Z-average hydrodynamic diameter Zave (nm), andntensity-weighted size distribution were calculated by theumulant [22] and CONTIN methods [23], respectively, usingrograms provided by the manufacturer. Before adding the rightliquot of 1,2 epoxydodecane (40 �L/10 mL of CTAB + XMPqueous solutions), weighted amounts of CTAB and nucleotidesere first dissolved in distilled water and the homogeneous solu-

ions filtered directly into a 10 mL reaction vessel through aillipore sterile 0.20 �m membrane filter to remove dust par-

icles. Sampling was carried out by withdrawing 1 mL fromhe reaction vessel and transferred into 10 mm square quartzuvette. After each measurement, the drawn volumes wereeintroduced into the mother 10 mL solutions, kept always atonstant stirring speed of about 350 rpm and controlled temper-ture (25 ◦C).

.4. Dialysis

MLVs were dialysed as follow: after 12 days reaction, 5 mLf MLVs suspensions were transferred in regenerated cellu-ose membranes. The diffusate chamber was filled with 0.8 L ofeionized water and placed on a magnetic stirrer at room temper-ture for 1 week to reach dialysis equilibrium. In order to monitorhe evolution of mass transport for molecular species able toross the membrane, 1-mL samples were taken from diffusateompartment to perform UV spectrophotometric measurementst 260 nm.

.5. Hypochromic effect and nucleolipids interactions

Spectrophotometric measurements were carried out in the00–200 nm range, under shearing conditions at 25 ◦C, with aary 100 Bio UV/Vis multicell spectrophotometer from Var-

an using quartz cells with a path length of 1 cm. Hypochromicffects in the earlier stages of the reaction were followed byecording the ε variations at intervals of 2 min. The start-ng values of ε were calculated in micellar solution through

dpmw

: Biointerfaces 64 (2008) 184–193

ambert–Beer relationship:

εAMPλmax

(mM−1 cm−1) = 11.93,

and εUMPλmax

(mM−1 cm−1) = 7.73.

Nucleolipids interactions were determined following the A260nd the scattering contribution at 400 nm for 5000 min undertirring condition. These interactions were investigated by dilut-ng the previously prepared MLVs (13 mM nucleotides, 27 mMTAB and 18 mM epoxydodecane). In details, the samples wererepared by diluting 45 �L of MLVs in 3 mL of deionized watero achieve the nucleotide surfactant concentration of 0.2 mM.

.6. NMR spectra

1H-NMR spectra were carried out by a high-resolution VarianNOVA 400 spectrometer equipped with an Oxford cryo-magnett 9.39 T and a Varian autoswitch probe of 5 mm. Spectra wereecorded with a presaturation sequence on water peak, using 6 ss acquisition time, an Ernst angle as pulse width and 0.3 Hz asine broadening.

.7. Polarized light optical microscopy

Images at optical microscope were obtained with a Zeissxioplan 2 using 10×, 40× and 100× magnifications on apoc-romatic lens (the last one is an “immersion lens”). Polarizedlter were used to detect birefringent structures. The microscopeas connected with a video camera (Moticam 2000 digital video

amera CCD of 16 mm), and the software Motic Images Plus.0 ML from Motic China Group Co. Ltd.

. Results and discussion

.1. Micelle–vesicles transition

The ability to induce a macroscopic-level morphologicalhange leading to the formation of micron-sized aggregates byeans of the cationic micellar interface (CTAB) in presence

f nucleotides was firstly investigated. The size evolution ofhe aggregate formation was determined by dynamic light scat-ering (DLS). Samples were checked out within a time of 12ays. The time courses of measured Z-average values of hydro-ynamic aggregate diameters for the system AMP, UMP, andIX (AMP–UMP in 1:1 ratio) are reported in Fig. 1. From this

raph it clearly appears that the rate of size evolution for theystem incubated with UMP results favoured and thus fasterhan the systems incubated with AMP or MIX. On the contraryhe increase of Z-average diameter in the analogue AMP sys-em occurs after a long threshold time. The initial fast growthf aggregates obtained with UMP was confirmed by visualnspection; in fact the turbidity of the sample becomes evidentithin 200 h. Moreover, in all cases the increase average sizes

emonstrated a process of formation of large aggregates. Theresence of large aggregates is confirmed by polarized lighticroscopy. After the reaction was run for 12 days, samplesere observed by polarized light microscopy. Fig. 2 depicts

F. Cuomo et al. / Colloids and Surfaces B

Fe

m1teAtfdp

Iii

ntscsq

3

odcmcwmtoob

Ft

ig. 1. Z-average evolution for CTAB micellar systems in presence 1,2-poxydodecane and different nucleotides: (�) AMP, (©) MIX, (�) UMP.

icrophotographs obtained from the reaction of XMP with,2 epoxydodecane in presence of the cationic interface. Pho-ographs present birefringent phases resulting by multifusionvents on MLVs. Fig. 2A–C are referred to MLVs formed withMP, UMP and MIX, respectively. As can be seen the micropho-

ographs reveal textures typical of multi-lamellar vesicles. Inact they displayed Malta crosses corresponding to the textureefect of a lamellar phase. Interestingly, the reaction takes alsolace in static conditions (without stirring) at room temperature.

wfet

ig. 2. Optical microphotograph of a birefringent phase in polarized light obtainedemperature. (A) Relative to AMP, (B) relative to UMP, (C) relative to MIX. In (D) th

: Biointerfaces 64 (2008) 184–193 187

n Fig. 2D the spontaneous MLVs formation obtained start-ng from CTAB/1,2-epoxydodecane/AMP in static conditionss also reported.

The synthesis of anionic amphiphilic derivatives ofucleotides during the reaction assisted by CTAB micellar solu-ion [8] allows the spontaneous formation of catanionic vesiclesince the anionic nucleolipids accumulate themselves within theationic pre-existent interface. These results indicate that theystem leads to a micelle–vesicles transition process as a conse-uence of “in situ” formation of alkylated anionic nucleotides.

.2. Dialysis and 1H NMR monodimensional spectra

In order to remove the 1,2-epoxydodecane excess and tobtain a rough purification of the products we performed theialysis of the MLVs followed by NMR analysis. For sake oflarity in Scheme 2 we report an illustration of the experi-ent. Nucleolipids, CTAB monomers and free mononucleotides

ross through the dialysis membrane. To clarify the terms in thisork the dialysate refers to the products that cross through theembrane and the retentate refers to products that remain in

he membrane volume. The enrichment of dialysate in nucle-lipids and free mononucleotides was monitored by carryingut absorption readings at 260 nm for almost 7 days. It shoulde stressed that at 260 nm both the salt and the surfactant form

ere followed; in fact in these experimental conditions both

orms can cross through the membrane. After dialysis the 1,2-poxydodecane excess remains in membrane. Fig. 3 shows theime evolution of the XMPtot (total concentration of free salts

from XMP, CTAB and 1,2-epoxydodecane in stirring conditions and at roome microphotograph is obtained in static condition using AMP as nucleotides.

188 F. Cuomo et al. / Colloids and Surfaces B

Scheme 2. Schematic representation of dialysis experiments. Free mononu-cleotides (XMP), nucleotides and CTAB monomers cross through the dialysismembrane. 1,2-epoxydodecane excess remains in membrane.

Fno

actd

ctMcvbromadidbttlyimr

boetb[bntstuscasa

T1

P

H

H

H

H

H

H

V

ig. 3. Time evolution of the XMPtot (total concentration of free salts anducleolipids) obtained from dialysed MLVs. In the insert: the time evolutionf nucleotides in sample prepared in absence of 1,2-epoxydodecane.

nd nucleolipids) obtained from the dialysate in terms of con-entration relative to initial [XMP], i.e., 13 mM, corrected forhe dilution factor 5/800 mL. This graph shows that only in theialysate of MLVs prepared with AMP the relative equilibrium

pp(a

able 1H NMR chemical shifts of base moiety protons

roton CTAB AMP�ν (Hz)

CTAB UMP�ν (Hz)

CTAB MIX�ν (Hz)

A(8) +31 +13

A(2) +13 −9

A(1′) +20 −3

U(5) +24 −2

U(6) +19 −2

U(1′) +19 −3

ariation in Hz referred to nucleotides solutions in D2O. Positive variations indicate

: Biointerfaces 64 (2008) 184–193

oncentration (95%) almost reaches the limit value. On the con-rary the limit value is not reached in the dialysate obtained from

LVs prepared with UMP; in fact, as shown the equilibriumoncentration in this case corresponds to the 72% of the limitalue. A similar behaviour is shown for the system containingoth complementary nucleotides, in this case a value of 85% iseached. The insert of Fig. 3 shows the evolution of nucleotidesbtained from the dialysate by performing a reference experi-ent in absence of 1,2-epoxydodecane (CTAB micellar solution

nd mononucleotides incubated for 12 days). As can be seen noifference has been shown between the different systems dur-ng the dialysis. The data collected in Fig. 3 indicate that theifferences of products that cross through the dialysis mem-rane are most probably due to the different reaction productshat self-assemble in several ways. Almost certainly, the reac-ions performed in presence of UMP and of the MIX produceess soluble surfactants that might be secluded inside the dial-sis membrane where a hydrophobic milieu is present. Thisnformation could be, in our opinion, highly valuable for infor-

ation concerning the solubility of products and therefore couldepresent a tool to discriminate the variety of the products.

Samples, roughly purified by dialysis, were then analysedy 1H NMR spectroscopy. This kind of investigation shed lightn interaction among the samples components and their influ-nce on the course of the reaction. 1H NMR spectroscopy is aool, commonly used, to investigate on the interactions occurringetween biologically relevant molecules such as nucleobases24]. In this section we compare the chemical shifts of thease protons dialysates with the analogoues resonances of freeucleotides in water and in presence of CTAB micellar solu-ions. In particular this technique allows discriminating betweentacking interactions and H-bonding, especially significant forhis kind of compounds. Stacking interactions are singled out bypfield shifts of the resonances, H-bonding implies downfieldhifts [25–27]. The shifts in Hz, obtained from the measuredhemical shifts of nucleotides in D2O, of the considered protonsre reported in Table 1. Positive variations indicate downfieldhifts, the negative ones upfield shifts. The numbering schemedopted for the protons is reported in Figs. 4–6. In Fig. 4 we show

artial protonic spectra (resonances of aromatic and anomericrotons) relative to AMP. 4a is the spectrum of a AMP solution13 mM) in D2O, 4b is relative to AMP (13 mM) in presence ofmicellar CTAB solution (27 mM), 4c is obtained analyzing a

AMP dialysate�ν (Hz)

UMP dialysate�ν (Hz)

MIX dialysate�ν (Hz)

−92, +37 −138

+27, +18 −29

+2, +23 −52

−49 −65

−123 −130

−39 −63

downfield shifts, the negative ones upfield shifts.

F. Cuomo et al. / Colloids and Surfaces B

Fig. 4. 1H NMR spectra (400 MHz, 298 K, D2O). Spectra (4a) is relative toAMP 13 mM, (4b) is obtained with AMP 13 mM and CTAB 27 mM, (4c) isfrom the AMP dialysate (5.6 mg/mL). For sake of clarity the structures reportedpresent the numbering scheme of protons.

Fig. 5. 1H NMR spectra (400 MHz, 298 K, D2O). Spectra (5a) is relative toUMP 13 mM, (5b) is obtained with UMP 13 mM and CTAB 27 mM, (5c) isfrom the UMP dialysate (5.6 mg/mL). For sake of clarity the structures reportedpresent the numbering scheme of protons.

diranrteapdriorrtm

bnatbmpH

Aatoatpsf

Fig. 6. 1H NMR spectra (400 MHz, 298 K, D2O). Spectra (6a) is relative to MIX 13 mdialysate (5.6 mg/mL). For sake of clarity the structures reported present the numberi

: Biointerfaces 64 (2008) 184–193 189

ialysed sample (5.6 mg/mL of freeze-dried powder dissolvedn D2O forming vesicles). By comparing spectra 4a and 4b itesults that all the protons in presence of the micellar solutionre downfield shifted (spectrum 4b), this behaviour is due toucleobase–micelle interactions; the protons, actually, are sur-ounded by a different local chemical environment induced byhe nucleotide in the hydrophobic region [28]. This experimentalvidence gives the indication that the nucleobase–micelle inter-ction is not negligible. In the dialysate system the aromaticortion of the spectrum (4c) shows two pairs of resonances withifferent intensities; this feature indicates that the same protonsesonate in two different sites. In fact, the resonances with lowerntensity have chemical shifts coinciding with aromatic signalsf free AMP + CTAB micelles while the higher intensity signalsesult upfield or downfield shifted. We assign the low-intensityesonances to the fraction of free non-alkylated nucleotides andhe higher to nucleolipids that are involved in MLVs bilayer and

ight be shifted owing to their mutual interactions.If we analyze the protons having the higher intensity, it can

e argued that on the supramolecular system, containing theew surfactant derivatives, both stacking and H-bonding inter-ctions hold. The upfield shift of H(8) is due to self-stacking,he downfield shift of H(2) is probably caused by the balance ofoth interactions. H(2) is the proton closer to N(1) position thatay be involved in hydrogen bonding, and this one is the only

roton that can be affected by H-boding effect. H(1′), as well(8), results upfield shifted compared with the spectrum 4b.In Fig. 5 spectra of the UMP containing systems are shown.

s can be seen the micellar solution induces downfield shifts forll the protons (spectrum 5b) also in this case. On the contraryhe dialysate’s spectrum (5c) results different from the dialysatef the previous system in two aspects: (1) each proton presentssingle chemical shift that results shifted from the reference,

herefore free nucleotide in salt form is not detected; (2) all therotons result upfield shifted and the only meaningful interactioneems to be via stacking. These differences could be due to theormation of different nucleolipids obtained as a consequence

M, (6b) is obtained with MIX 13 mM and CTAB 27 mM, (6c) is from the MIXng scheme of protons.

1 ces B: Biointerfaces 64 (2008) 184–193

oUNft

abbmsvstlpcnuAuUbntoi

Ctcmrtsst

3

optdTcIsmIp[vAvO

FU

ltAvsomaocahnetIcfibiTotrcS

3

nMMt

90 F. Cuomo et al. / Colloids and Surfa

f a different moiety involved in the reaction of alkylation; forMP nucleotides the alkylation site can be identified at level ofor O-positions of the pyrimidinic ring; therefore, no conditions

or H-bonding can be achieved. H(1′) proton is upfield shifted inhis case also.

In Fig. 6 spectra relative to MIX system are shown. Inbsence of interaction between nucleobases, the spectra woulde simply the superimposition of the single-base systems. Letegin to consider the micellar system. The protons are noore deshielded (except for H(8)) and are coincident with 6a

pectrum, probably because the inter-base interactions are pre-ailing on the nucleotide–micelle’s one. Also the dialysate’spectrum (6c) turn out sensitive to base–base interactions: (1)here are no resonances corresponding to AMP on the micel-ar system (the lower intensity protons in Fig. 4c); (2) all therotons show larger upfield shifts than the parent systems (thehemical shift data are reported in Table 1). For H(8) reso-ance, the upfield shift increases from 92 Hz (spectrum 4c)p to 138 Hz (6c); for both anomeric and H(2) protons ofMP, a chemical shift reversal from down- to up-field val-es has been detected; the same trend has been observed forMP protons. NMR data suggest that stacking interactionsetween complementary nucleolipids may represent the domi-ant effect; furthermore, the absence of downfield shift confirmshe fact that the pyrimidine ring is alkylated in UMP nucle-lipid monomers. Thus, only �–� type interaction can benvoked.

Remarkably, these data indicate that in absence of epoxideTAB–nucleotides interactions are effective but not selective for

he kind of nucleobase involved [29]; this effect is reduced if aomplementary combination of nucleotides interacts with CTABicelles because in this case the intra-nucleobase interactions

esult more effective. The first and more immediate informationhat arise from NMR analysis is a strong aryl interaction as justeen. Moreover, we can consider the hypothesis that in the MIXystem strong nucleotide- interactions do not allow us to detecthe protons of free AMP mononucleotides.

.3. Inter-molecular interactions: hypochromic effect

In order to investigate on the earlier interaction stagesccurring in these systems, UV–vis measurements wereerformed. For this purpose, the variation of ε values duringhe first hour of the reaction on the microemulsions containingifferent nucleobases (AMP, UMP, and MIX) was followed.hese measurements were achieved on dilute nucleotidesoncentration (0.2 mM) according to the instrumental limits.t is well known that UV spectra of aromatic rings are sen-itive to �–� interaction in fact, when two or more rings areutually close they give rise to hypochromic effect [9,30].

n literature the following staking tendency is reported:urine–purine > pyrimidine–purine > pyrimidine–pyrimidine31]. In Fig. 7 the time dependence of absolute variation of ε

alues in systems containing AMP, UMP and MIX are reported.s clearly indicated in the graph, a decrease of 10% of the ε

alue obtained by loading UMP 0.2 mM is easily observed.n the contrary the decrease of the ε in presence of AMP is

cccm

ig. 7. Absolute time variation of ε in microemulsion for AMP 0.2 mM (�);MP 0.2 mM (�); MIX 0.2 mM (AMP 0.1 mM + UMP 0.1 mM) (©).

ess meaningful resulting in a ε decrease of 2.5%. As concernhe MIX a variation of 5% is observed by loading 0.1 mMMP + 0.1 mM UMP. Notwithstanding, we verified that the ε

alues remain unchanged by loading nucleotides to the micellarolution (data not shown), demonstrating the important rolef 1,2-epoxydodecane in this reaction. Since the measure-ents were carried out at fixed nucleotides concentration, the

bsorption changes take into account of stacking interactionsccurring by reducing the aromatic rings’ mobility. This factould occur subsequently to the formation of nucleolipids thatrrange themselves to the pre-existent micellar interface. Theigher hypochromicism exhibited by UMP nucleotide doesot confirm the beforehand reported staking tendency. Thisvidence could be explained as follows: in presence of UMPhe reaction of nucleolipid formation seems to be favoured.n fact, if we discuss this behaviour in molecular terms wean state that having UMP a smaller group, apparently it bestts into the surface [32]. On the contrary being the adeninease of AMP larger requires more space in the lipid–waternterface of micelles, according to DLS data discussed before.he ε decrease recorded for the MIX system (5% variationbserved) could be described considering that the presence ofhe complementary nucleotides leads to a slow down of theeaction rate, due to interactions between the nucleobases thatould modify the rate of the reaction and the final products (seecheme 1).

.4. Supramolecular stability

To test whether the simultaneous presence of complementaryucleotides during the reaction could affect the stability of theLVs further investigations were performed on the new formedLVs. For this purpose measurements of turbidity (light scat-

ering at 400 nm), correlated with the aggregate stability [33],

ombined with measurements of absorption at 260 nm asso-iated with the stacking interactions, have been achieved. Weompare samples prepared in different ways: one was made byixing equal quantities of MLVs prepared separately with AMP

ces B

aot

tMas

napv5p

A

wRTt

FpUs

saolts7fcttrtnAtd

F. Cuomo et al. / Colloids and Surfa

nd UMP (we will indicate this sample with AMP/UMP), thether one was prepared in presence of both nucleotides simul-aneously (we indicated this sample with MIX).

It is worth noting that the MLVs in both of the cases con-ain the complementary nucleolipids. In particular aliquots of

LVs were first diluted, and measurements of turbidity and ofbsorption at 260 nm were recorded as function of time keepingamples under stirring conditions.

The absorption values measured were affected by a non-egligible contribution coming from the scattering of theggregates that was evaluated and eliminated following therocedure proposed on other similar systems [34]. Absorptionalues were corrected fitting the spectral region between 320 and00 nm, (where absorption from nucleic bases is absent) with aower law:

′ = K0λ−k (1)

here λ is the wavelength in the range defined above, k (forayleigh scatterers k ≈ 4) and K0 are adjustable parameters.hen, the calculated scattering curve has been extrapolated up

o 220 nm and finally subtracted to the measured absorption

ig. 8. Spectra and scattering correction as function of time in sample pre-ared by mixing equal quantities of MLVs prepared separately with AMP andMP (AMP/UMP) (A), and in sample prepared in presence of both nucleotides

imultaneously (MIX) (B).

oRtcod

(ostospvisaap

Miaicidsbd

vhiOsca[

: Biointerfaces 64 (2008) 184–193 191

pectrum thus showing true electronic absorption values. Asn explicative example in Fig. 8 the scattering curves extrap-lated from the spectra are reported at various times (dottedines). Panel A shows the absorption spectra for AMP/UMP inhe onset of the experiment (in black) and the correspondentcattering curve (black dotted line), and after 24 (in red) and5 h (in green). A decrease of scattering intensity is detected asunction of time. Panel B shows the MIX system spectra (coloursorrespond to the same times reported for the AMP/UMP sys-em): as clearly indicated, for the last system, the spectra relativeo 24 and 75 h are coincident because the aggregate stabilityesults enhanced compared with AMP/UMP. We report the rela-ive turbidity evolution for both systems in Fig. 9A: AMP/UMPucleolipids (full circles) and MIX derivatives (empty circles).s clearly reported the time courses of the measured relative

urbidity of the MLVs systems followed during the first hoursecrease to 40% in both of the systems. The initial decrementf turbidity takes into account, of course, of a dilution effect.emarkably MLVs obtained from MIX result more stable and

he relative turbidity remains almost the same (40%), on theontrary, on MLVs relative to AMP/UMP the initial stabilityn 40% of relative turbidity evolves, after 2200 min, in a linearecrease.

Moreover, if we switch to the absorbtion experimentFig. 9B), the correspondence with the experimental databtained by turbidity is easily deduced: when the turbidity is per-istent the value of A260 remains stable, when the turbidity startso slows down, simultaneously the absorption begins to increasewing to the loss of superorganization and a consequence oftacking interactions. Fig. 9B strikingly shows that with MLVsroduced in presence of complementary nucleotides the A260alues remain constant. On the contrary an absorption increases recorded in the sample realized with the separately preparedurfactants. These experimental evidences, combined together,re consistent with the hypothesis that on the AMP/UMP systemloss of superorganization and stacking interactions are takinglace.

It is worth to underline the role of nucleotides in catanionicLVs stability that is an important physical–chemical aspect

nfluenced also by nucleolipids interaction. The interactions,nd as a consequence the stability, as previously shown, aremproved in systems where the synthesis starts with both theomplementary nucleobases; on the other hand stronger stak-ng interactions in the complementary nucleotides system areetected by NMR analysis as well. The main important conclu-ion is that the enhanced stability of catanionic MLVs formedy complementary combination of nucleobases could be mostlyue to �–� interactions.

Comparing the absorption results (in Fig. 9B) with the pre-iously reported absorption experiments (Section 3.3) one thingas to be noticed: the change of absorption in the former cases faster than that observed starting within the vesicular system.n the MLVs, in fact, the process is slower than in the micellar

ystem because micelles do not present the double layer and theorresponding analogy with biological membranes and are char-cterized by higher dynamic due to a greater structural simplicity35].

192 F. Cuomo et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 184–193

F quantf ). (B

4

CovttTdcfsctstew

hpst

A

Rls

R

[

[

[

[

[

[

[[

[

[[[[[[

[

[

[

ig. 9. (A) Relative turbidity evolution for a sample prepared by mixing equalor a sample prepared in presence of both nucleotides simultaneously (MIX) (©

. Conclusions

A micellar interface constituted by the cationic surfactantTAB in presence of 1,2-epoxydodecane and various typef nucleotides can be used for the catanionic multi-lamellaresicles production. The driving force of the MLVs forma-ion depends on the alkylation of AMP and UMP nucleotideshat give the correspondent anionic amphiphilic derivatives.he nucleolipids synthesis obtained with this assembly processriven by inter-nucleobases interactions led to the production ofatanionic MLVs in which the catalytic role of the cationic inter-ace was evaluated in the formation of the nucleolipids. DLS andpectrophotometric measurements on hypochromic effect indi-ate that the reaction and the aggregates growth are faster forhe UMP containing system, and slower with AMP. 1H NMRpectra shed light on the prevailing of stacking interactions onhe supramolecular aggregate and on the hypothesis of a coop-rative process driven by nucleobases interactions in the systemith complementary nucleotides.The MLVs stability has been tested and in particular a

igher stability has been found in systems prepared with com-lementary nucleotides confirming that in these systems theelf-assembly process is influenced by nucleotides interac-ions.

cknowledgements

This work was supported by the Italian Ministero Universitaicerca (PRIN 2006: Self-assembling nanosystems with DNA-

ike addressability) and by Consorzio Interuniversitario per loviluppo dei Sistemi a Grande Interfase (CSGI-Firenze).

eferences

[1] R. Gardlık, R. Palffy, J. Hodosy, J. Lukacs, J. Turna, P. Celec, Med. Sci.Monit. 11 (2005) RA110.

[2] X. Wang, E.J. Danoff, N.A. Sinkov, J.-H. Lee, S.R. Ragahavan, D.D.English, Langmuir 22 (2006) 6461.

[3] S. Segota, S. Heimer, D. Tezak, Colloid Surf. A 274 (2006) 91.

[

[

ities of MLVs prepared separately with AMP and UMP (AMP/UMP) (�), and) A260 evolution for the same samples.

[4] S. Segota, D. Tezak, Adv. Colloid Int. Sci. 121 (2006) 51.[5] L. Zhai, X. Lu, W. Chen, C. Hu, L. Zheng, Colloid Surf. A 236 (2004) 1.[6] S.M. Mel’nikov, R. Dias, Y.S. Mel’nikova, E.F. Marques, M.G. Miguel, B.

Lindman, FEBS Lett. 453 (1999) 113.[7] M. Rosa, M. da Graca Miguel, B. Lindman, J. Colloid Int. Sci. 312 (2007)

87.[8] R. Angelico, A. Ceglie, F. Cuomo, C. Cardellicchio, G. Mascolo, G.

Colafemmina, Langmuir 24 (2008) 2348.[9] W. Saenger, Principles of Nucleic Acid Structure, Springer, New York,

1984.10] M. Raszka, N.O. Kaplan, Proc. Natl. Acad. Sci. U.S.A. 69 (1972)

2025.11] S. Shuto, H. Ito, S. Ueda, S. Imamura, K. Fukukawa, M. Tsujino, A.

Matsuda, T. Ueda, Chem. Pharm. Bull. 26 (1988) 209.12] T. Shimizu, R. Iwaura, M. Masuda, T. Hanada, K. Yase, J. Am. Chem. Soc.

123 (2001) 5947.13] L. Moreau, M.W. Grinstaff, P. Barthelemy, Tetrahedran Lett. 46 (2005)

1593.14] L. Moreau, P. Barthelemy, M. El Maataoui, M.W. Grinstaff, J. Am. Chem.

Soc. 126 (2004) 7533.15] Y. Wang, B. Desbat, S. Manet, C. Aime, T. Labrot, R. Oda, J. Colloid Int.

Sci. 283 (2005) 555.16] D. Berti, F. Pini, P. Baglioni, J. Phys. Chem. B 103 (1999) 1738.17] F. Baldelli Bombelli, D. Berti, U. Keiderling, P. Baglioni, J. Phys. Chem.

B 106 (2002) 11613.18] D. Berti, P. Baglioni, S. Bonaccio, G. Barsacchi-Bo, P.L. Luisi, J. Phys.

Chem. B 102 (1998) 303.19] P. Baglioni, D. Berti, Curr. Opin. Colloid Interf. Sci. 8 (2003) 55.20] K. Conde-Frieboes, E. Blochliger, Biosystems 61 (2001) 109.21] A. Duran, J. Mol. Catal. A-Chem. 256 (2006) 284.22] D.E. Koppel, J. Phys. Chem. 57 (1972) 4814.23] W. Provencher, Comput. Phys. Commun. 27 (1981) 213.24] G. Zandomeneghi, P.L. Luisi, L. Mannina, A. Segre, Helv. Chim. Acta 84

(2001) 3710.25] D. Berti, P. Barbaro, I. Bucci, P. Baglioni, J. Phys. Chem. B 103 (1999)

4916.26] J.S. Nowick, J.S. Chen, G. Noronha, J. Am. Chem. Soc. 115 (1993)

7636.27] M.P. Schweizer, A.D. Broom, P.O.P. Ts’o, D.P. Hollis, J. Am. Chem. Soc.

14 (1968) 1042.

28] I.L. Nantes, F.M. Correia, A. Faljoni-Alario, A.E. Kawanami, H.M. Ishiki,

A.T. Amaral, A.M. Carmona-Riberio, Arch. Biochem. Biophys. 416 (2003)25.

29] D.W. Armstrong, R. Sueguin, J.H. Fendler, J. Mol. Evol. 10 (1977)241.

ces B

[

[[

F. Cuomo et al. / Colloids and Surfa

30] C.R. Cantor, P.R. Shimmel, Biophysical Chemistry, W.H. Freeman and Co.,San Francisco, 1980.

31] P.R. Mitchell, H. Sigel, Eur. J. Biochem. 88 (1978) 149.32] A. Scheidt, W. Flasche, C. Cismas, M. Rost, A. Herrmann, J. Liebscher, D.

Huster, J. Phys. Chem. B 108 (2004) 16279.

[

[

[

: Biointerfaces 64 (2008) 184–193 193

33] E. Feitosa, G. Karlsson, K. Edwards, Chem. Phys. Lipids 140 (2006)66.

34] M.A.R.B. Castanho, N.C. Santos, L.M.S. Loura, Eur. Biophys. J. 26 (1997)253.

35] C. Heinz, U. Radler, P.L. Luisi, J. Phys. Chem. B 102 (1998) 8686.