Guido Viscardi,‡ Pierluigi Quagliotto,‡ Gaetano Donofrio,§ Bozenna Rozycka-Roszak,∥ Paweł Misiak,∥
Edyta Wozniak,∥,⊥ and Francesco Sansone#
†Department of Pharmacy, University of Parma, Parco Area delle Scienze 27/A, 43124 Parma, Italy‡Department of Chemistry, Interdepartmental “Nanostructured Surfaces and Interfaces” NIS Centre, University of Torino, Via P.Giuria 7, 10125 Torino, Italy§Department of Veterinary Sciences, University of Parma, Via del Taglio 10, 43126 Parma, Italy∥Department of Physics and Biophysics and ⊥Department of Chemistry, Wrocław University of Environmental and Life Sciences,Norwida 25, 50-375 Wrocław, Poland#Department of Chemistry, University of Parma, Parco Area delle Scienze 17/A, 43124 Parma, Italy
*S Supporting Information
ABSTRACT: The interaction with a model membrane, the formationof DNA nanoparticles, and the transfection ability of a homologousseries of bispyridinium dihexadecyl cationic gemini surfactants, differingin the length of the alkyl spacer bridging the two pyridinium polar headsin the 1 and 1′ positions (P16-n with n = 3, 4, 8, 12), have been studiedby means of differential scanning calorimetry (DSC), atomic forcemicroscopy, electrophoresis mobility shift assay, and transient trans-fection assay measurements. The results presented here show that theirperformance in gene delivery is strictly related to their structure insolution. For the first time the different transfection activities of thecompounds can be explained by referring to their thermodynamicproperties in solution, previously studied. The compound with a spacerformed by four carbon atoms, showing unexpected enthalpic propertiesvs concentration in solution, is the only one giving rise to a transfectionactivity comparable to that of the commercial reagent, when formulated with L-α-dioleoylphosphatidylethanolamine. We suggestthat P16-4 behaves like molecular tongs able to grip basic groups near each other, allowing the formation of compact and nearlyspherical DNA particles. The compound with the longest spacer gives rise to loosely condensed structures by forming a sort ofbow, not able to give rise to transfection notwithstanding the double positive charge of the molecule. On the other hand, DSCmeasurements on synthetic membranes show that the compounds with the shortest spacers (three and four methylene groups)practically do not interact with the 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine membrane, while compounds P16-8 and,particularly, P16-12 induce the formation of surfactant-rich and surfactant-poor domains in the membrane, without showing anypeculiarity for compound P16-4. This could suggest that the mechanisms involved in the interaction with the model membraneand in gene delivery are substantially different and could strike a blow for an endocytosis mechanism for the internalization in thecell of the DNA nanoparticles.
■ INTRODUCTIONBy the name gemini surfactants are indicated those surfactantsin which at least two identical moieties are bound together by aspacer at the polar head level. They show peculiar properties incomparison with the monomeric ones owing to their increasedsurface activity, lower critical micelle concentration (cmc), anduseful viscoelastic properties.1−8 Recently, a comprehensivereview has been published, focused on positively chargedheterocyclic gemini surfactants: their major synthetic accessroutes are presented, and the impact of structural elements ontheir physicochemical and aggregation properties is examined.8
In recent years, the scientific and applicative interest in cationic
gemini surfactants has increased, also because of theirpotentialities in the pharmaceutical field, both as noncovalentfunctionalization agents for carbon-nanotube-based formula-tions for drug delivery9 and as nonviral vectors in genetherapy.10−18 Gene therapy is used to treat diseases caused by amissing, defective, or overexpressing gene. It is based on thedelivery of a therapeutic by means of a specially designedvector. To reach this goal, the DNA must be packaged in a viral
Received: August 7, 2014Revised: October 17, 2014Published: October 23, 2014
or nonviral delivery system.17 Since viral vectors, currently themost efficient, are not without the risk of adverse orimmunogenic reaction or replication, depending on the virusbeing used, cationic lipids have in many cases become thepreferred means of gene delivery into eukaryotic cells. Cationiclipids constitute the building blocks to compact and encapsulatethe DNA into soft nanoparticles, delivered to specific sites bysize-dependent passive targeting or by active targeting, that is,by means of ligands attached onto their structure to achieve cellsurface specificity. Moreover, nanoparticles are able to protectthe DNA from enzymatic degradation, and their tunable sizeallows for building nanoparticles large enough for preventingrapid leakage into blood capillaries but small enough forescaping macrophages of the reticuloendothelial system.However, nonviral delivery systems with high transfectionefficiency must still be realized. We have obtained encouragingresults in gene delivery by using very simple bis(quaternaryammonium) gemini surfactants, derivatives of N,N,N′,N′-tetramethyl-1,2-ethanediamine of general formula CnH2n+1-OOCCH2(CH3)2N
+CH2CH2N+(CH3)2CH2COOCnH2n+1·
2Cl− (bis-CnBEC), where the subscript n stands for the numberof carbon atoms of the alkyl chain bound to the carboxyl group,when formulated with DOPE [L-α-dioleoylphosphatidyletha-nolamine (C18:1,[cis]-9)].15 These results urged us to designand characterize new gemini compounds starting from thosehaving, as polar heads, two pyridinium groups, bridged togetherby aliphatic chains of different lengths.19 In this paper we reportfor the first time physicochemical and biological dataconcerning this new homologous series of gemini surfactants,with the aim to achieve better insight into the interaction of thecationic surfactants with membranes and DNA. In fact, we havebeen collecting these data for some time with the idea ofcorrelating, in a quantitative way, the structures of thesurfactants with their biological activities, particularly theirtransfection abilities.15,20−22 In this perspective, syntheticvectors present the advantage that their constituent parts canbe quite easily modified, thereby facilitating the elucidation ofstructure−activity relationships.8 A kind of gemini pyridiniumsurfactant, structurally different from the compounds presentedhere with the spacer bridging the heteroaromatic nitrogens, hasbeen recently proposed in the literature for transfectionpurposes.18
■ EXPERIMENTAL METHODSCompounds. The series of bispyridinium cationic gemini
surfactants under study was prepared by us, as described in ref19 for the dodecyl compounds (see also the SupportingInformation). In this paper the compounds having chloride as acounterion, are named Pm-n, where m indicates the number ofcarbon atoms of the alkyl chain and n the spacer length and Pstands for bispyridinium. The compounds studied are 1,1′-dihexadecyl-2,2′-trimethylenebispyridinium dichloride (P16-3),1,1′-dihexadecyl-2,2′-tetramethylenebispyridinium dichloride(P16-4), 1,1′-dihexadecyl-2,2′-octamethylenebispyridinium di-chloride (P16-8), and 1,1′-dihexadecyl-2,2′-dodecamethylene-bispyridinium dichloride (P16-12). The structure of compoundP16-3 is shown in Figure 1.Moreover, P12-8 (1,1′-didodecyl-2,2′-octamethylenebispyr-
idinium dichloride) and P12-8-MS (1,1′-didodecyl-2,2′-octa-methylenebispyridinium dimethanesulfonate) were also synthe-sized by us, as described in ref 19.The purity was checked by NMR, elemental analysis, and
acid:water = 4:1:5, organic phase) on a silica gel plate (Merck).The solutions were prepared by weight using freshly boiledbidistilled water and stored under nitrogen. The solutionconcentrations are expressed as molality, m (mol kg−1).1,2-Dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC)
was purchased from Sigma-Aldrich, Steinheim, Germany. Thelipid purity was greater than 99%.
DNA Preparation and Storage. Plasmid DNA waspurified through cesium chloride gradient centrifugation. Astock solution of the plasmid (0.7 μM) in Milli-Q water(Millipore Corp., Burlington, MA) was stored at −20 °C.Linearized plasmid DNA (pEGFP-C1) was obtained by cuttingwith EcoRI restriction enzyme (Roche), column purified(Genomed), and alcohol precipitated. Linearized plasmidDNA pellet was washed with 70% ethanol, air-dried, anddissolved in distilled H2O at a final concentration of 1 μg/μL.
Cell Culture. The human rhabdomyosarcoma cell line RD-4, obtained from David Derse, National Cancer Institute,Frederick, MD, was maintained as a monolayer using growthmedium containing 90% Dulbecco’s modified Eagle’s medium(DMEM), 10% fetal bovine serum (FBS), 2 mM L-glutamine,100 IU/mL penicillin, and 10 μg/mL streptomycin. The cellswere subcultured to a fresh culture vessel when growth reached70−90% confluence (i.e., every 3−5 days) and incubated at 37°C in a humidified atmosphere of 95% air/5% CO2.
Electrophoresis Mobility Shift Assay (EMSA). Bindingreactions were performed in a final volume of 14 μL with 10 μLof 20 mM Tris/HCl (pH 8), 1 μL of plasmid (1 μg of pEGFP-C1), and 3 μL of P16-n with n = 3, 4, 8, and 12 at different finalconcentrations ranging from 25 to 200 μM. The bindingreaction was allowed to take place at room temperature for 1 h,5 μL of 1 g/mL glycerol in H2O was added to each reactionmixture, and the resulting mixture was loaded onto a TA (40mM Tris−acetate) 1% agarose gel. The gel was run for 2.5 h inTA buffer at 10 V/cm. EDTA was omitted from the buffersbecause it competes with DNA in the reaction.
Transient Trasfection Assay. Transfections were per-formed in six-well plates when the cells were 80% confluent(approximately 3 × 105 cells). A 3 μg portion of plasmid P16-nwith n = 3, 4, 8, and 12 was added to 1 mL of serum-freemedium at a final concentration of 15 μM, and the resultingmixture was mixed rapidly and incubated at room temperaturefor 20 min. Each mixture was carefully added to the cellsfollowing the aspiration of the culture medium from the cells.Lipoplex formulations were performed by adding DOPE to theplasmid−surfactant mixture at different surfactant:DOPE molarratios (1:1, 1:2, and 2:1), where the surfactant concentrationwas kept to 15 μM. GenePORTER transfection reagent, a
Figure 1. Example of the optimized conformation of compound P16-3: C, green; N, blue; H, not shown.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113184
neutral lipid transfection reagent, was used as a positivetransfection control.The mixture and cells were incubated at 37 °C in a
humidified atmosphere of 95% air/5% CO2 for 5 h. Finally, 1mL of medium containing 20% FBS was added to eachtransfected well, and the contents were allowed to incubate for72 h.15,22
Transfected cells were observed under a fluorescencemicroscope for enhanced green fluorescent protein (EGFP)expression. Five random fields were examined from each well,and each experiment was repeated three times. Statisticaldifferences among treatments were calculated with Student’stest and multifactorial analysis of variance (ANOVA).Differential Scanning Calorimetry (DSC) Measure-
ments. DSC studies were performed according to the protocoldescribed earlier using the Mettler Toledo thermal analysissystem DSC 821e.15 The cycles were performed three times.The experimental uncertainty of temperature was ±0.2 °C.Sample Preparation and Atomic Force Microscopy
(AFM) Imaging. DNA samples were prepared by diluting the
plasmid DNA to a final concentration of 0.5 nM in depositionbuffer (4 mM Hepes, 10 mM NaCl, 2 mM MgCl2, pH 7.4)either in the absence or in the presence of P16-n with n = 3, 4,8, and 12 in the same DNA:surfactant ratio used in thetransient transfection assay experiments. The mixture wasincubated for 5 min at room temperature, and then a 20 μLdroplet was deposited onto freshly cleaved ruby mica (TedPella, Redding, CA) for 1 min. The mica disk was rinsed withMilli-Q water and dried with a weak stream of nitrogen.AFM imaging was performed on the dried sample with a
Nanoscope IIIA microscope (Digital Instruments Inc., SantaBarbara, CA) operating in tapping mode. Commercial divingboard silicon cantilevers (NSC-15, Micromash Corp., Estonia)were used. Images of 512 × 512 pixels were collected with ascan size of 2 μm at a scan rate of 3−4 lines/s and wereflattened after being recorded using Nanoscope software.
Modeling. The geometry optimization for dications ofcompounds P16-n with n = 3, 4, 8, and 12 and P12-8 wasperformed using semiempirical quantum computations with theMOPAC2012, version 12.309W,23 program via the VEGA ZZ
Figure 2. DSC heating curves of multilamellar vesicles (MLVs) with increasing molar ratios of the compounds under study to DPPC. The curves arenormalized for the amount of DPPC, scan rate 2 °C/min.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113185
package.24 The computations were done for the dications in avacuum.
■ RESULTSInteractions with Model Lipid Membrane and
Modeling. The influence of gemini surfactants on thethermotropic phase behavior of a dipalmitoylphosphatidylcho-line bilayer is shown in Figure 2. In the absence of a surfactant,DPPC exhibited two endothermal transitions upon heating, thepretransition at 35 °C and the main gel-to-liquid-crystallinephase transition at 41 °C. Compounds P16-3 and P16-4 havevirtually no effect on the transitions. For compound P16-8 themain phase transition temperature (Tm) remains practicallyunchanged, but the values of the transition enthalpy (ΔHm)and the half-maximum width (T1/2) increase (Table 1). Themain transition is slightly broadened and asymmetrical, whichsuggests a lateral phase separation. Probably surfactant-rich andsurfactant-poor domains are formed. Compound P16-8, likeP16-3 and P16-4, does not suppress the pretransition.Compound P16-12 affects both phase transitions much morethan the previous compounds. At a molar ratio of 0.01, ashoulder appears on the left side of the main phase transition(I). This suggests that also compound P16-12 causes phaseseparation and surfactant-rich and surfactant-poor domains areformed. At a molar ratio of 0.03, the shoulder transforms into anew peak (II) and the pretransition disappears. At higherconcentrations both peaks are shifted to lower temperatures,their half-maximum widths increase, and the enthalpy of peak Idecreases. At a molar ratio of 0.1, the peaks overlap.The interaction of compound P16-12 is stronger and phase
separation is much more enhanced in comparison to the effectsof compound P16-8. The compounds with shorter alkyl chains,P12-8 and P12-8-MS (having methanesulfonate as thecounterion), affect the phase transitions in a similar way.They both abolish pretransition at the lowest concentration anddecrease Tm and ΔHm much more than the former ones,showing a strong fluidization effect.The dications of compounds P16-n with n = 3, 4, 8, and 12
and P12-8 (the same as P12-8-MS) were prepared in differentinitial conformations for the quantum calculations. Then theywere optimized using the two parametrizations PM7 andRM1.25 An example of the optimized geometry is shown inFigure 1 for P16-3 (other examples can be found in theSupporting Information). The main effect observed in thecalculations is that in a vacuum the pyridinium rings tend toposition away from each other, which is understandable giventhat they bear the positive charge in the vicinity of thenitrogens. In some cases for the compounds with long spacers(8 and 12 carbon atoms) the hydrocarbon tails tend to alignwith the spacer chain. This can take place especially in thewater solution, considering the hydrophobic interactionsbetween the alkyl chains. However, one can expect that thedications entering the lipid bilayer will take conformations withthe alkyl tails immersed in the hydrophobic layer of themembrane and the pyridinium rings placed at the level of thenegatively charged phosphate groups. The calculated distancebetween the pyridinium rings of the optimized dications,measured as the average distance between nitrogens, is about0.69, 0.81, 1.27, 1.5−1.7, and 1.25 nm for compounds P16-nwith n = 3, 4, 8, and 12 and P12-8, respectively.Biological Assays. The interaction of P16-n with n = 3, 4,
8, and 12 with plasmid DNA pEGFP-C1 (Clontech) wasmonitored by agarose gel EMSA (Figure 3). The shift activity
was observed for all the compounds investigated, able tomodify the mobility of DNA at the lowest concentration tested(25 μM). Finally, to test the capability of the same compoundsto deliver DNA inside the cells, a transient transfection assaywas performed with a plasmid carrying an EGFP expressioncassette under the control of the Cytomegalovirus (CMV)immediate early promoter (pEGFP-C1, Clontech) to monitorEGFP expression under a fluorescence microscope.RD-4 cells were chosen among a large panel of several cell
lines because they are a good compromise between verydifficult to transfect cells and very easy to transfect cells withtraditional methods (electroporation, lypofection, and calciumphosphate precipitation). Moreover, they are easy to handle,
Table 1. Temperature Tm, Half-Width of the Peaks T1/2, andEnthalpy Change ΔH of the Main Phase Transition of theLipid Bilayer vs the Surfactant-to-Lipid Molar Ratio (nX/nDPPC)
aThe values marked I and II are related to the peaks marked I and II inthe DSC thermograms.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113186
are fast growing, and are derived from a nasty human cancer.This later reason makes the data achieved from the moleculesemployed in the present paper relevant for oncology genetransfer studies. P16-4 was also the only one able to deliverDNA inside the cells, as shown by EGFP expression, albeit atan efficiency much lower than that obtained with a standardcommercial transfection reagent (GenePORTER, Gene Ther-apy System), which we used as a positive transfection control.It appears that DOPE enhances the transfection activity of
cationic formulations through the stabilization of the DNA/lipid complex26,27 and facilitates the transfer of DNA in thecontext of endosomal escape, owing to its fusogenic property.Previous studies have demonstrated that DOPE significantlyaffects the polymorphic features of lipoplexes promotingtransition from a lamellar to a hexagonal phase, thus facilitatingendosomal escape.28 We therefore tested in vitro thetransfection efficiency of P16-n with n = 3, 4, 8, and 12formulated with DOPE at different P16-n:DOPE molar ratios(1:1, 1:2, and 2:1). P16-n concentrations were fixed at 15 μM.The addition of DOPE increases the transfection efficiency ofP16-4 and P16-8 a little, but it is ineffective with P16-3 andP16-12 (Figure 4). In particular, P16-4 shows, when formulatedwith DOPE, a transfection ability comparable with that of thestandard commercial transfection reagent (Figure 4). In
contrast, absence of transfection is observed only for theDOPE transfection control.
AFM Experiments. To verify the ability of the compoundsunder investigation to induce structural changes in the DNA,we employed AFM.29,30 This technique has been successfullyused to study the interaction of both synthetic ligands31,32 andproteins33 with DNA. AFM experiments were carried out usingcircular DNA imaged in air in the tapping mode. Figure 5a
shows a typical image of the plasmid DNA alone depositedonto freshly cleaved mica. Single plasmids and concatamers areseen in their plectonemic form with several supercoils whichcause the double helix to cross itself a number of times. Besidesthe topological constraint, plectonemes appear well extended allover the mica surface. Parts b−d of Figure 5 show the plasmidDNA after incubation with P16-n with n = 4, 8, and 12,respectively. According to the EMSA results, only P16-4(Figure 5b) is able to condense all DNA into nearly sphericalnanoparticles less than 0.1 μm in diameter. After addition of
Figure 3. EMSA experiments showing complexation of P16-n (n = 3,4, 8, 12) with circular plasmid pEGFP-C1. Shifting is observable as afunction of the concentration (μM). As a negative control, only theplasmid was used, which is completely unshifted.
Figure 4. Transfection of RD-4 cells with P16-n (n = 3, 4, 8, 12) without and with DOPE in different molar ratios, only DOPE, and the positivecontrol of a standard commercial transfection reagent.
Figure 5. AFM images showing the effect induced on DNA plasmid byincubation with P16-n (n = 4, 8, 12). Each image represents a 2 × 2μm scan (scale bar 0.2 μm). All images were obtained with supercoiled0.5 nM pEGFP-C1 plasmid deposited onto mica and with themicroscope operating in tapping mode in air. (a) DNA plasmid alone.Plasmid incubated with (b) 2 nM P16-4, (c) 2 nM P16-8, and (d) 2nM P16-12. For P16-3 see the Supporting Information. The chromaticbar inside the figure refers to the thickness of the objects.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113187
P16-8 (Figure 5c) only part of the DNA condenses intonanoparticles homogeneous neither in size nor in shape. In thecase of P16-12, the AFM images show that partially condensedstructures, if any, are formed, looking like bows (Figure 5d).P16-3 (not shown), interacting with DNA as well, is unable
to condense the DNA into nanoparticles, but only reduces theextension of the loops formed by plasmid DNA by overlappingof the opposite sides.
■ DISCUSSIONThe use of gemini surfactants as nonviral vectors in genetherapy has been proposed, on account of the possibility oftaking advantage of their multiple cationic charge, necessary forbinding and compacting DNA, and of their superior surfaceactivity.8−18 DNA nanoparticles, in fact, are generally composedof specific cationic lipids to ensure efficient DNA condensationand cellular uptake of the complexes and of helper lipids suchas DOPE.17 As said before, we have obtained encouragingresults in gene delivery by using very simple bis(quaternaryammonium) gemini surfactants, bis-CnBEC, when formulatedwith DOPE.15 Along the same lines, we designed andcharacterized new gemini compounds having, as polar heads,two pyridinium groups, bridged together by an aliphatic chain,with the idea that the presence of two aromatic pyridiniumrings could enhance the interaction with DNA and their abilityof compacting and encapsulating DNA into nanoparticlesreadily internalized by cells.19 The cmc of the hexadecylderivatives having a chloride counterion was found to be lowerthan 1 × 10−4 M by means of conductometric techniques. As anexample, the cmc of P16-3 was 8.51 × 10−5 M, which is about17-fold lower than that of the corresponding dodecyl derivativeP12-3 (cmc = 1.45 × 10−3 M). The cmc is little affected by theaddition of a methylene group in the spacer. It is worth notingthat the compound with four carbons in the spacer (P16-4) hasa cmc value of 9.4 × 10−5 M, a little higher than that of P16-3(see the Supporting Information). With the aim of enrichingthe fundamental understanding of the self-aggregationthermodynamics of this class of surfactants, we have measuredthe apparent and partial molar enthalpies at 298 K of theaqueous solutions of the homologous series of cationic geminisurfactants 1,1′-didodecyl-2,2′-alkylenebispyridinium chloridesand methanesulfonates, differing in the spacer length.20,21 Wemeasured the dilution enthalpies of the didodecyl compoundsbecause the dihexadecyl compounds have too low a cmc toallow for an accurate determination in both the premicellar andpostmicellar concentration regions.20,21 We were amazed bytheir very peculiar behavior as a function of the spacer length,never found before in the literature, not allowing for thedetermination of a −CH2− group contribution when this groupis added to the spacer. In fact, as a rule, the inclusion of amethylene group in the structure of the surfactant gives rise to amonotonic change in the trends of the apparent and partialmolar thermodynamic quantities as a function of theconcentration of the solute, so that it is possible to extract a−CH2− group contribution of the solute, useful to foresee theproperties of new compounds. This is not the case for thehomologous series of gemini compounds under study in whichthe methylene groups are added to the spacer. In fact, the curveof the compound with a spacer formed by four carbon atomslies between those of the compound with a spacer of threecarbon atoms, not below the latter, as expected, and thisbehavior, not affected by the counterion, suggests thatsomething happens in the structure of the molecule in solution
when the spacer is four carbon atoms long.21 We interpretedthis surprising behavior as evidence of a conformational changeof the molecule caused by stacking interactions between thetwo pyridinium rings, mediated by the counterion andappearing at an optimum length of the spacer. The hypothesiswas also supported by the data obtained from the surfacetension vs log c curves, showing that Amin, the minimum areataken on the surface by the molecule, is significantly lower forthe compound with a spacer four carbon atoms long than thatof the other compounds of the same homologous series andthat the same compound has a greater tendency to formmicelles instead of adsorbing at the air/water interface.20,21 It isinteresting to understand if this different behavior outlined bythe trend of apparent and partial molar enthalpies vsconcentration also affects the biological properties of thesurfactants. This has been done by studying the interaction ofthe gemini surfactants with model membranes and their genedelivery ability.
Interactions with Model Membranes. Measurementsshow that compounds P16-3 and P16-4 with the shortest spacer(three and four methylene groups, respectively) practically donot alter the ordering of the DPPC membrane. The othercompounds interact with the DPPC membrane differently,depending on the length of the spacer. Compounds P16-8 andP16-12 induce the formation of surfactant-rich and surfactant-poor domains in the membrane, but the process is much morepronounced for P16-12.We have also evaluated the effect on the membrane ordering
of the alkyl chain length of the surfactants by comparingcompounds P16-8 and P12-8 and of the counterion bycomparing P12-8 and P12-8-MS, having methanesulfonate(MS) instead of chloride as the counterion. P12-8, with dodecylchains, shows a much more destructive effect on the membranethan the compound with hexadecyl tails. Compounds P12-8-MS and P12-8 show very similar effects on membranes, andhence, the kind of counterion has almost no effect on theinteraction with model membranes, although this is not ageneral case.5−7,34−36
The effect of the compounds under study on the phasetransitions of the DPPC bilayer can be elucidated in part inview of commensurability of the dications with the structure ofthe bilayer. The calculated distance between the pyridiniumrings of the compounds is given in the Results. For P16-12, thespacer is most flexible; thus, the N−N distance in the optimizedgeometries varies over a fairly wide range. In the crystallinestate, the average distance between the DPPC molecules isabout 0.8 nm, which is close to the distance between thepyridinium rings of P16-4 and about half that of compoundP16-12. The calculated N−N distance for compounds P16-8and P12-8 is incommensurate with the bilayer lattice constant;thus, it has the effect of a stronger disturbance of the bilayerstructure. Moreover, the alkyl tails of compounds P12-8-MSand P12-8 are distinctly incommensurate with the palmitoylchains of DPPC molecules, in contrast to the tails of compoundP16-n with n = 3, 4, 8, and 12. This is another factor which mayexplain the stronger impact of compounds P12-8-MS and P12-8 than compound P16-8 on the bilayer ordering.
Gene Delivery Ability. To obtain DNA nanoparticles,cationic lipids (cytofectins)15,27,37 or cationic polymers38 areneeded to neutralize the anionic charges of the DNA phosphategroups, thus compacting the DNA and obtaining complexes ofapproximately spherical shape having a small dimensioncompared to that of the naked DNA. The amphiphilic nature
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113188
of such molecules allows strong interaction with membranephospholipids and release of the DNA in due time. DNAnanoparticles are therefore composed of specific cationic lipidsto ensure efficient DNA condensation and cellular uptake andhelper lipids such as DOPE.39 We expect that the compoundsunder investigations have to interact with the DNA, owing totheir double positive charges carried by their heterocyclic polarheads. Moreover, the aromatic nature of the polar heads couldgive rise to stacking interaction with DNA bases byintercalation.For gene delivery studies, we have chosen to use the
compounds with alkyl chains of 16 carbon atoms, in accordwith our previous experience with this subject.15,22 It isreported that the biological activity of cationic surfactantsincreases with the chain length up to a critical point.39 In thecase of the homologous series of alkanediyl-α,ω-bis-(dimethylalkylammonium bromide), the member with twoalkyl chains of 16 carbon atoms is generally the mostbiologically active.40 The same seems to hold also for single-chained sugar-based cationic surfactants, for example, withglucopyranosyl22 and gluconamide41,42 moieties in the head-group.The results of biological assays show that the compounds
studied here are all able to interact specifically with the DNA,completely shifted at concentrations above 50 μM. This isparticularly true for P16-4 and P16-8, doing so at halvedconcentration. It is very surprising to compare the DNAcondensation ability of the compounds under study. In fact, asshown in Figures 4 and 5, the addition of one methylene groupin the spacer is enough to go from a compound unable toefficiently compact the DNA to a compound able to transformplasmid DNA in nanoparticles and give rise to transfection.Without thermodynamic data, it would be very difficult tounderstand this sudden change in behavior. As outlined before,P16-4 shows an anomalous behavior as far as, for instance,enthalpic properties are concerned. To explain this, we haveassumed a conformational change of the molecule: when thespacer reaches the right size, the molecule in solution doublesup, like a book, due to stacking interactions between the twopyridinium rings and the hydrophobic interactions of the alkylchains, independently of the counterion.20,21 This arrangement,not envisaged by quantum mechanical calculation in a vacuumin the absence of counterions and hydrophobic interactions,because the electrostatic repulsion prevails, is possible neitherwhen the spacer is too short (two or three carbon atoms),because of the lack of enough conformational freedom, norwhen the spacer is too long, because the pyridinium rings aretoo far apart. In this way, P16-4 behaves like molecular tongs,able to grip the aromatic bases of the DNA or phosphategroups. The distance between two residues of DNA is 0.34 nm:if the phosphate groups substitute the counterions between thetwo aromatic rings, reducing the electrostatic repulsion, themolecule could assume the stacked conformation suggested bythermodynamic data, with a medium distance betweenpyridinium rings much lower than that calculated in a vacuum(0.81 nm) for the dications. Then the hydrophobic interactionsbetween the chains of the surfactants cause the efficientformation of nanoparticles. When the length of the spacerincreases, this arrangement becomes more difficult and thepyridinium rings are prone to interact with DNA sites far fromeach other. This explains why P16-8 is still able to form someless compact nanoparticles and to give rise to a littletransfection, while the compound with a spacer 12 carbon
atoms long is not. In fact, the latter probably assumes anextended conformation in solution with the pyridinium rings farapart, also because the hydrophobic interactions in solutioncould involve the spacer, and the positive polar heads interactwith DNA, bridging remote bases and therefore giving rise tobows, as shown in Figure 5d. In this way the structures formedare not compact enough to be able to penetrate into the cellularmembrane, the first step of transfection. Similarly, we suggestedthe mechanism of cytotoxicity of gluconamide-based single-chained cationic surfactants involving an early effect on themetabolic activity of the cells.42
The coformulation with DOPE does not considerably changethe gene delivery ability of the compounds.In short, the result is that P16-4, showing the highest gene
delivery ability, is unable to affect the ordering of DPPCmembranes. The above results could suggest that themechanisms involved in the interaction with a modelmembrane and in gene delivery are substantially different andcould strike a blow for an endocytosis mechanism for theinternalization in the cell of the DNA nanoparticles. This is, toenhance transfection ability, we do not necessarily have to lookfor carriers giving rise to strong interactions with the cellmembrane, but we have to optimize the electrostaticinteractions between the cationic lipid and DNA.
■ CONCLUSIONS
The results presented here show that pyridinium geminisurfactants could be a valuable tool for gene delivery purposes,but also that their performance is strictly related to theirstructure in solution. For the first time the transfection activityof the compounds is strictly related to their thermodynamicproperties in solution. The compound with a spacer formed byfour carbon atoms, showing unexpected enthalpic properties vsconcentration in solution, is the only one giving rise to atransfection activity comparable to that of the commercialreagent, when formulated with DOPE. We suggest that P16-4behaves like molecular tongs able to grip basic group near eachother. The compound with the longest spacer gives rise to aloosely condensed structure by forming a sort of bow, not ableto give rise to transfection notwithstanding the double positivecharge of the molecule. On the other hand, DSC measurementson synthetic membranes show that the compounds with theshortest spacers (three and four methylene groups) practicallydo not alter the ordering of the DPPC membrane, whilecompounds P16-8 and, particularly, P16-12 induce theformation of surfactant-rich and surfactant-poor domains inthe membrane, without showing any peculiarity for compoundP16-4. This could suggest that the mechanisms involved in theinteraction with a model membrane and in gene delivery aresubstantially different and could strike a blow for anendocytosis mechanism for the internalization in the cell ofthe DNA nanoparticles.
■ ASSOCIATED CONTENT
*S Supporting InformationComplete refs 11 and 22, schematic description of the synthesisof the compounds under study together with details of thepaper under consideration, reporting synthesis and tensidiccharacterization of P16-n gemini surfactants, and examples ofoptimized conformations of the compounds under study. Thismaterial is available free of charge via the Internet at http://pubs.acs.org.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113189
NotesThe authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
This work was supported by the Italian Ministry of Education,University and Research (MIUR), PRIN2011 “Nano MolecularTechnologies for Drug DeliveryNANOMED”, by theUniversity of Torino (ex 60% - Ricerca Locale 2013), and, inpart, by Grant No. N305 361739 from the Polish Ministry ofScience and Higher Education. We thank the InterdipartimentalCentre of Measurements (CIM) of the University of Parma forallowing the use of the AFM facilities. The use of a differentialscanning calorimeter at the Institute of Agricultural Engineeringof Wrocław University of Environmental and Life Sciences isgratefully acknowledged. The modeling computations were, inpart, performed using the software and the computingresources provided by the Wrocław Centre for Networkingand Supercomputing (WCSS).
■ REFERENCES(1) Rosen, M. J.; Tracy, D. J. J. Gemini Surfactants. J. SurfactantsDeterg. 1998, 1, 547−554.(2) Rosen, M. J. Geminis: A NewG of Surfactants. CHEMTECH1993, 23 (3), 30−3.(3) Menger, F. M.; Keiper, J. S. Gemini Surfactants. Angew. Chem.,Int. Ed. 2000, 39, 1906−1920.(4) Zana, R. In Novel SurfactantsPreparation, Applications, andBiodegradability. Dimeric (Gemini) Surfactants; Holmberg, K., Eds.;Surfactant Science Series, Vol. 74; Marcel Dekker, Inc.: New York,1998; p 241.(5) Zana, R.; Xia, J. Gemini Surfactants, Interfacial and Solution PhaseBehavior, and Applications; Marcel Dekker, Inc.: New York, 2004.(6) Fisicaro, E. Gemini Surfactants: Chemico-Physical and BiologicalProperties. Cell. Mol. Biol. Lett. 1997, 2, 45−63.(7) Han, Y.; Wang, Y. Aggregation Behavior of Gemini Surfactantsand Their Interaction with Macromolecules in Aqueous Solution. Phys.Chem. Chem. Phys. 2011, 13, 1939−1956.(8) Sharma, V. D.; Ilies, M. A. Heterocyclic Cationic GeminiSurfactants: A Comparative Overview of Their Synthesis, Self-Assembling, Physicochemical, and Biological Properties. Med. Res.Rev. 2014, 34, 1−44.(9) Lamprecht, C. J.; Huzil, T.; Ivanova, M. V.; Foldvari, M. Non-Covalent Functionalization of Carbon Nanotubes with Surfactants forPharmaceutical ApplicationsA Critical Mini-Review. Drug DeliveryLett. 2011, 1, 45−57.(10) Bell, P. C.; Bergsma, M.; Dolbnya, I. P.; Bras, W.; Stuart, M. C.A.; Rowan, A. E.; Feiters, M. C.; Engberts, J. B. F. N. TransfectionMediated by Gemini Surfactants: Engineered Escape from theEndosomal Compartment. J. Am. Chem. Soc. 2003, 125, 1551−1558.(11) Kirby, A. J.; Camilleri, P.; Engberts, J. B. F. N.; Feiters, M. C.;Nolte, R. J. M.; Soderman, O.; Bergsma, M.; Bell, P. C.; Fielden, M. L.;Garcia Rodriguez, C. L.; et al. Gemini Surfactants: New SyntheticVectors for Gene Transfection. Angew. Chem., Int. Ed. 2003, 42, 1448−1457.(12) Donkuru, M. D.; Wettig, S. D.; Verrall, R. E.; Badea, I.; Foldvari,M. Designing pH-Sensitive Gemini Nanoparticles for Non-Viral GeneDelivery Intokeratinocytes. J. Mater. Chem. 2012, 22, 6232−6244.(13) Badea, I.; Shaterian, N.; Foldvari, M. Topical Gene Delivery inMice Using Gemini SurfactantLipid Nanoparticles with and withoutTape Electrode Electroporation. Drug Delivery Lett. 2011, 1, 62−66.(14) Yang, P.; Singh, J.; Wettig, S.; Foldvari, M.; Verrall, R. E.; Badea,I. Enhanced Gene Expression in Epithelial Cells Transfected with
Amino Acid-Substituted Gemini Nanoparticles. Eur. J. Pharm.Biopharm. 2010, 75, 311−320.(15) Fisicaro, E.; Compari, C.; Duce, E.; Donofrio, G.; Rozycka-Roszak, B.; Wozniak, E. Biologically Active Bisquaternary AmmoniumChlorides: Physico-Chemical Properties of Long Chain Amphiphilesand Their Evaluation as Non-Viral Vectors for Gene Delivery. Biochim.Biophys. Acta, Gen. Subj. 2005, 1722, 224−233.(16) Sing, J.; Michel, D.; Chitanda, J. M.; Verral, R.; Badea, I.Evaluation of Cellular Uptake and Intracellular Trafficking asDetermining Factors of Gene Expression for Amino Acid-substitutedGemini Surfactant-Based DNA Nanoparticles. J. Nanobiotechnol. 2012,10:7, 1−14.(17) Donkuru, M. D.; Badea, I.; Wettig, S.; Verrall, E.; Elsabahy, M.;Foldvari, M. Advancing Nonviral Gene Delivery: Lipid- andSurfactant-Based Nanoparticle Design Strategies. Nanomedicine 2010,5 (7), 1103−1127.(18) Sharma, V. D.; Lees, J.; Hoffman, N. E.; Brailoiu, E.; Madesh,M.; Wunder, S. L.; Ilies, M. A. Modulation of Pyridinium CationicLipid-DNA Complex Properties by Pyridinium Gemini Surfactantsand Its Impact on Lipoplex Transfection Properties. Mol. Pharmaceu-tics 2014, 11, 545−559.(19) Quagliotto, P.; Viscardi, G.; Barolo, C.; Barni, E.; Bellinvia, S.;Fisicaro, E.; Compari, C. Gemini Pyridinium Surfactants: Synthesisand Conductimetric Study of a Novel Class of Amphiphiles. J. Org.Chem. 2003, 68 (20), 7651−7660.(20) Fisicaro, E.; Compari, C.; Biemmi, M.; Duce, E.; Peroni, M.;Barbero, N.; Viscardi, G.; Quagliotto, P. L. The Unusual Behaviour ofthe Aqueous Solutions of Gemini Bispyridinium Surfactants: Apparentand Partial Molar Enthalpies of the Dimethanesulfonates. J. Phys.Chem. B 2008, 112, 12312−12317.(21) Fisicaro, E.; Compari, C.; Bacciottini, F.; Barbero, N.; Viscardi,G.; Quagliotto, P. L. Is the Counterion Responsible for the UnusualThermodynamic Behavior of the Aqueous Solutions of GeminiBispyridinium Surfactants? Colloids Surf., A 2014, 443, 249−254.(22) Fisicaro, E.; Compari, C.; Biemmi, M.; Duce, E.; Peroni, M.;Donofrio, G.; Sansone, F.; Rozycka-Roszak, B.; Pruchnik, H.; Barbero,N.; et al. Thermodynamics and Biological Properties of the AqueousSolutions of New Gluco-Cationic Surfactants. J. Phys. Chem. B 2008,112, 9360−9370.(23) Stewart, J. J. P. MOPAC2012, version 12.309W, StewartComputational Chemistry, Colorado Springs, CO. http://openmopac.net (accessed February 2013).(24) Pedretti, A.; Villa, L.; Vistoli, G. VEGAAn Open Platform ToDevelop Chemo-Bio-Informatics Applications, Using Plug-in Archi-tecture and Script Programming. J. Comput.-Aided Mol. Des. 2004, 18,167−173.(25) Rocha, G. B.; Freire, R. O.; Simas, A. M.; Stewart, J. J. P. RM1:A Reparameterization of AM1 for H, C, N, O, P, S, F, Cl, Br, and I. J.Comput. Chem. 2006, 27, 1101−1111.(26) Liu, F.; Yang, J.; Huang, L.; Liu, D. Effect of Non-IonicSurfactants on the Formation of DNA/Emulsion Complexes andEmulsion-Mediated Gene Transfer. Pharm. Res. 1996, 13, 1642−1646.(27) Hara, T.; Liu, F.; Liu, D.; Huang, L. Emulsion Formulations as aVector for Gene Delivery in Vitro and Vivo. Adv. Drug Delivery Rev.1997, 24, 265−271.(28) Chen, Y.; Lu, Y.; Tao, C.; Huang, J.; Zhang, H.; Yu, Y.; Zou, H.;Gao, J.; Zhong, Y. Complexes Containing Cationic and Anionic pH-Sensitive Liposomes: Comparative Study of Factors InfluencingPlasmid DNA Gene Delivery to Tumors. Int. J. Nanomed. 2013, 8,1573−1593.(29) Bustamante, C.; Rivetti, C. Visualizing Protein-Nucleic AcidInteractions on a Large Scale with the Scanning Force Microscope.Annu. Rev. Biophys. Biomol. Struct. 1996, 25, 395−429.(30) Hansma, H. G. Surface Biology of DNA by Atomic ForceMicroscopy. Annu. Rev. Phys. Chem. 2001, 52, 71−92.(31) Sansone, F.; Dudic, M.; Donofrio, G.; Rivetti, C.; Baldini, L.;Casnati, A.; Cellai, S.; Ungaro, R. DNA Condensation and CellTransfection Properties of Guanidinium Calixarenes: Dependence on
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113190
Macrocycle Lipophilicity, Size, and Conformation. J. Am. Chem. Soc.2006, 128, 14528−14536.(32) Volcke, C.; Pirotton, S.; Grandfils, C.; Humbert, C.; Thiry, P. A.;Ydens, I.; Dubois, P.; Raes, M. Influence of DNA Condensation Stateon Transfection Efficiency in DNA/Polymer Complexes: An AFM andDLS Comparative Study. J. Biotechnol. 2006, 125, 11−21.(33) Ceci, P.; Cellai, S.; Falvo, E.; Rivetti, C.; Rossi, G. L.; Chiancone,E. DNA Condensation and Self-Aggregation of Escherichia coli DpsAre Coupled Phenomena Related to the Properties of the N-Terminus. Nucleic Acids Res. 2004, 32, 5935−5944.(34) Rozycka-Roszak, B.; Przyczyna, A. Interaction between N-Dodecyl-N,N-dimethyl-N-benzylammonium Halides and Phosphati-dylcholine BilayersThe Effect of Counterions. Chem. Phys. Lipids2003, 123, 209−221.(35) Rozycka-Roszak, B.; Przyczyna, A.; Pernak, A. A Study on theInteraction between 1-Decyloxymethyl-3-carbamoylpyridinium Saltsand Model MembranesThe Effect of Counterions. Biophys. Chem.2004, 109, 271−279.(36) Rozycka-Roszak, B.; Przyczyna, A.; Misiak, P.; Pruchnik, H. TheInfluence of Cholesterol on the Interaction between N-Dodecyl-N,N-dimethyl-N-benzylammonium Halides and Phosphatidylcholine Bi-layers. Z. Naturforsch. 2006, 61c, 302−308.(37) Schmidt-Wolf, G. D.; Schmidt-Wolf, I. G. Non-Viral and HybridVectors in Human Gene Therapy: An Update. Trends Mol. Med. 2003,9, 67−72.(38) Mansouri, S.; Lavigne, P.; Corsi, K.; Benderdour, M.; Beaumont,E.; Fernandes, J. Chitosan-DNA Nanoparticles as Non-Viral Vectors inGene Therapy: Strategies To Improve Transfection Efficacy. Eur. J.Pharm. Biopharm. 2004, 57, 1−8.(39) Balgavy, P.; Devinsky, F. Cut-Off Effects in Biological Activitiesof Surfactants. Adv. Colloid Interface Sci. 1996, 66, 23−63.(40) Jennings, K. H.; Marshall, I. C. B.; Wilkinson, M. J.; Kremer, A.;Kirby, A. J.; Camilleri, P. Aggregation Properties of a Novel Class ofCationic Gemini Surfactants Correlate with Their Efficiency as GeneTransfection Agent. Langmuir 2002, 18, 2426−2429.(41) Misiak, P.; Wilk, K. A.; Kral, T.; Wozniak, E.; Pruchnik, H.;Frackowiak, R.; Hof, M.; Rozycka-Roszak, B. New Gluconamide-TypeCationic Surfactants: Interactions with DNA and Lipid Membranes.Biophys. Chem. 2013, 180−181, 44−54.(42) Rozycka-Roszak, B.; Misiak, P.; Wozniak, E.; Zaczyn ska, E.;Czarny, A. Effect of Biocompatible Gluconamide-Type CationicSurfactants on Thermotropic Phase Behavior of Phosphatidylcho-line/Cholesterol Bilayers. Thermochim. Acta 2014, 590, 219−225.
The Journal of Physical Chemistry B Article
dx.doi.org/10.1021/jp507999g | J. Phys. Chem. B 2014, 118, 13183−1319113191
