Advances in Colloid and Interface Science 205 (2014) 230–239

Contents lists available at ScienceDirect

Advances in Colloid and Interface Science

j ourna l homepage: www.e lsev ie r .com/ locate /c i s

Structure and kinetics of lipid–nucleic acid complexes

Nily Dan a,⁎, Dganit Danino b,c,⁎⁎a Department of Chemical and Biological Engineering, Drexel University, Philadelphia, PA 19104, USAb Department of Biotechnology and Food Engineering, Technion, Haifa 32000, Israelc The Russell Berrie Nanotechnology Institute, Technion, Haifa 32000, Israel

⁎ Corresponding author. Tel.: +1 215 895-6624; fax: +⁎⁎ Correspondence to: D. Danino, Department of Biotech

E-mail addresses: dan@coe.drexel.edu (N. Dan), dgani

0001-8686/$ – see front matter © 2014 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.cis.2014.01.013

a b s t r a c t

a r t i c l e i n f o

Available online 28 January 2014

Keywords:LipoplexesGene therapyNon-viral carriersKineticsDNA/lipidsiRNACryo-TEM

The structure and function of lipid-based complexes (lipoplexes) have been widely investigated as cellulardelivery vehicles for nucleic acids—DNA and siRNA. Transfection efficiency in applications such as gene therapyand gene silencing has been clearly linked to the local, nano-scale organization of the nucleic acid in the vehicle,as well as to the global properties (e.g. size) of the carriers. This review focuses on both the structure of DNA andsiRNA complexes with cationic lipids, and the kinetics of structure evolution during complex formation.The local organization of the lipoplexes is largely set by thermodynamic, equilibrium forces, dominated bythe lipid preferred phase. As a result, complexation of linear lambda-phage DNA, circular plasmid DNA, orsiRNA with lamellae-favoring lipids (or lipid mixtures) forms multi-lamellar LαC liquid crystalline arrays.Complexes created with lipids that have bulky tail groups may form inverted hexagonal HII

C phases, orbicontinuous cubic QII

C phases.The kinetics of complex formation dominates the large-scale, global structure and the properties of lipoplexes.Furthermore, the time-scales required for the evolution of the equilibrium structure may be much longer thanexpected. In general, the process may be divided into three distinct stages: An initial binding, or adsorptionstep, where the nucleic acid binds onto the surface of the cationic vesicles. This step is relatively rapid, occurringon time scales of order of milliseconds, and largely insensitive to system parameters. In the second step, vesiclescarrying adsorbed nucleic acid aggregate to form larger complexes. This step is sensitive to the lipid characteris-tics, in particular the bilayer rigidity and propensity to rupture, and to the lipid to nucleic acid (L/D) charge ratio,and is characterized by time scales of order seconds. The last and final step is that of internal rearrangement,where the overall global structure remains constant while local adjustment of the nucleic acid/lipid organizationtakes place. This stepmay occur on unusually long time scales of order hours or longer. This rate, aswell, is highlysensitive to lipid characteristics, includingmembrane fluidity and rigidity. While the three step process is consis-tent with many experimental observations to date, improving the performance of these non-viral vectors re-quires better understanding of the correlations between the parameters that influence lipoplexes' formationand stability and the specific rate constants i.e., the timescales required to obtain the equilibrium structures.Moreover, new types of cellular delivery agents are now emerging, such as antimicrobial peptide complexeswith anionic lipids, and other proteins and small-molecule lipid carriers, suggesting that better understandingof lipoplex kinetics would apply to a variety of new systems in biotechnology and nanomedicine.

© 2014 Elsevier B.V. All rights reserved.

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2311. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2312. The structure of cationic lipid–nucleic acid complexes (lipoplexes) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2313. Studies of the kinetics of lipoplex formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2334. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2355. Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

1 215 895-5837.nology and Food Engineering, Technion, Haifa 32000, Israel. Tel.: +972 4 829 2143; fax: +972 4 829 3399.td@tx.technion.ac.il (D. Danino).

ghts reserved.

231N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

Preface

For many decades Professor Bjorn Lindman has been a prolific andinsightful leader in thefield of complex fluids. His seminal contributionsilluminated the fundamentals of self-assembly of surfactants and blockcopolymers [1] that had major impact on both the basic science andapplications of soft condensed matter.

Another field investigated by Professor Lindman is that of DNA–lipidcomplexes. For example, his study of the solubilization of highMWDNAcomplexeswith cationic lipids in organic solvents [2] demonstrated thatcomplex hydrophobicity is sensitive to the lipid used, providing insightsinto the methodology by which such complexes interact with mem-branes. When most studies focused on DNA complexes with cationiclipids, he has shown that DNA can form complexes with zwitterioniclipids through the use of divalent ions [3].

In this paper we focus on an aspect of nucleic–acid complexeswith cationic lipids, to which Professor Lindman's contributions havebeen invaluable: The kinetics of complex formation. A decade after itspublication, his experimental work on the kinetics of DNA–lipid com-plexation remains one of the few systematic studies [4]. The model heproposed [5] reconciled seemingly contradictory results, and has beenendorsed by more recent work.

We dedicate this paper to Professor Lindman on the occasion of his70th birthday, as a small token of our deep appreciation for his manycontributions and scientific achievements.

1. Introduction

Modification of cellular functions through the transfer of nucleicacids, transfection, is highly inefficient when based on ‘naked’ nucleicacids. Effective biomedical application requires, therefore, complexationof the nucleic acids into supramolecular ‘carriers’ that can increasedelivery efficiency and ensure activity [6–9]. Nucleic acid carriers mustperform several, somewhat contradictory functions. They must, firstand foremost, condense and stabilize the nucleic acid. In addition,carriers must bind to the cell surface, transport into the cell cytosol, es-cape from the endosome and release the nucleic acid. Some require-ments are specific to the material delivered—DNA must be importedinto the nucleus for transcription, and siRNA requires targeting tospecific tissues. In vivo systemic intravenous administration requires,in addition, the ability to circulate in the bloodstream for a prolongedperiod of time while resisting degradation by immune-agents. As aresult, carriers must be stable under some conditions, yet responsiveto the biological environment to be functional.

Early studies of gene therapy utilized viral-based vectors as idealgene delivery vehicles, but were found to cause side-effects and posehealth risks [10]. This has focused efforts on synthetic analogues, includ-ing lipid-based (lipoplex) [11–15] and polymer-based (polyplex) struc-tures (recently detailed in [6]). The majority of carriers are based onsupramolecular assemblies between nucleic acids and cationic lipids.These lipoplexes form (self-assemble) spontaneously and display a richarray of structures. While earlier studies focused on gene delivery, thepotential therapeutic advantages of RNA interference focused intereston small interfering (si)RNA vehicles. These were found to be similarin structure and properties to the DNA complexes [8,16–19].

Lipoplexes self-assembly is driven by electrostatic attraction be-tween the oppositely charged nucleic acid and lipids. The equilibriumstructures resulting from these interactions are controlled by the lipidtype (and in case of mixtures, composition), and by the charge ratio,defined as the ratio between the cationic lipid and DNA charges(L/D, +/−), and the specific interactions between DNA molecules(hydrophobic end-to-end and electrostatic). The kinetics of the assem-bly, however, is likely to vary as a function of additional parameters,such as the overall concentration of the components, and systemtemperature. Complex kinetics is also expected to strongly depend onthe initial form of the components (e.g. vesicle size) and complex

preparation route (e.g. order and rate of mixing) [4,20,21]. Thus, whilethe ultimate molecular lipoplex structure may be set by equilibrium,the route and rate by which the structure would be obtained are setby the kinetics. Furthermore, some complexation rates may be ex-tremely long, on the order of hours or even days—much longer thantypical self-assembly rated; as a result, complexes studied on shortertime scales may not be at equilibrium bur rather kinetically trappedstructures.

Understanding the kinetics of lipoplex self-assembly is of fundamen-tal interest: As discussed in this review, lipoplex formation is a multi-stage, multi-time scale and multi-length-scale process that encom-passes many aspects of self-assembly that may be relevant to othersystems (e.g. polyelectrolye–amphiphile complexes).

The sensitivity of carrier efficiency to the equilibrium, nano-scalecomplex structure [7,8,22] suggests that understanding structure evolu-tion is necessary for effective application. For example, nucleic acid per-formance in the cell environment requires its release from the carrier.Thus, understanding the rate and path of complex formation wouldelucidate the process of complex dissociation and nucleic acid release.

Another issue to be emphasized is the correlation between the globallipoplex structure and its efficacy.While local, equilibriumproperties onthe nano-scale affect the order and density of nucleic acid in the com-plex and the method and rate of nucleic acid release from the complex,transfection steps such as circulation in vivo or interactionswith cellularmembranes are set by the overall complex size, shape and net charge.These properties are dominated by kinetics rather than equilibrium, sothat controlling them requires control of the kinetic process.

Recent reviews emphasized the properties of equilibrium structuresin relation to transfection efficiency and biological performance[7,8,22–25]. The present review is focused on the kinetics of complexformation. In the next section we summarize the ‘phase diagram’ ofnucleic acid–lipid supramolecular complexes and the prime parametersdetermining lipoplex geometry and properties. Experimental studies oflipid–nucleic acid assembly kinetics are reviewed in detail in Section 2,followed by a discussion of the results in Section 3. We conclude with abrief outlook.

2. The structureof cationic lipid–nucleic acid complexes (lipoplexes)

The synthesis of complex biomaterials is achieved, in general, viatwo routes: ‘Top down’, where a more complex structure is modifiedor simplified until the target is reached, or ‘bottom up’wheremoleculesor simple structures are assembled into supramolecular arrays. The lat-ter can be further divided into directed assembly, where the placementof components is imposed (e.g. through 3D printing) or throughspontaneous self-assembly. The former yields structures that arenon-equilibrium assemblies, andmay therefore be unstable, while spon-taneous self-assembly leads to meta-stable or equilibrium structures.

Cationic lipids interact spontaneously with anionic DNA (andsiRNA), thereby leading to the formation of lipoplexes [11,12]. The ther-modynamic driving forces are electrostatic interactions between the op-positely charged lipids and the nucleic acids, and increase in entropydue to dehydration and release of bound counterions from the DNAand the lipid surfaces [11,12,26–31]. The binding free energy gain pernucleotide is rather small (~kT); This has been shown by calculations(see, for example, [32]) and later confirmed by calorimetry analysis ofDNA interactions with both single lipids and lipid mixtures [20,33,34].Nevertheless, complexes form since the large number of parallelbinding events per nucleic acidmolecule counteracts the loss of entropyassociated with the nucleic acid immobilization. The result is a sponta-neous and rapid complexation process that is generally irreversible, ina manner similar to that of polyelectrolyte adsorption on oppositelycharged surfaces [35]. However, due to the low binding energy persite, the binding is locally reversible, namely, there is a finite probabilitythat at any given time any given bondmay be broken as a result of ther-mal fluctuations, giving rise to molecular rearrangements within the

232 N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

complex. Moreover, as in any electrostatically-driven binding, bondscan be dissolved through changes in system parameters such as pH orsalinity. This dissolution is essential for the utilization of lipoplexes fornucleic acid delivery into cells, since processes such as transcriptioncannot take place with the DNA bound to the vehicle.

The geometry of lipoplexes is largely dominated by the phase pre-ferred by the pure lipid. Generally, bilayer-forming lipids tend to formlamellar complexes, and cylindrical micelle-forming lipids form hexag-onal phases [36]. However, these structures may be modified by thepresence, concentration and length of the oppositely charged geneticmaterial. Mixing of bilayer-forming lipids and lipids favoring non-lamellar phases (such as inverted hexagonal or bicontinuous cubic)can lead to other types of geometries [27,29,30,32,36,37], with theorder and packing of the nucleic acid within the lipid phase vary as afunction of parameters such as the overall charge ratio or, in lipidmixtures, the mixture composition and charge density.

Accordingly, the structure of complexes formed by DNA and bilayer-favoring lipids, whether pure cationic lipid systems e.g., dioleoyl-trimethylammonium propane (DOTAP) or mixtures of this lipid with aneutral lipid (dioleoyl-phosphadtidycholin, DOPC), was identified as alamellar Lα

C phase [11,38]. As shown in Fig. 1A LαC complexes are

composed of DNA layers intercalated between lipid bilayers [11], withthe DNA strands within a layer regularly spaced, organized in a two-dimensional liquid-crystal smectic array [11,20,25,27–31,37]. Charac-teristic dimensions for LαC lipoplexes are membrane thickness of≈40 Å, and water gap accommodating a DNA monolayer betweenlipid bilayers ≈ 25 Å. Unlike the overall geometry, the typical spacingbetween DNA strands within a monolayer varies, e.g., between 25 and

Fig. 1. The structure of the twomost common lipoplex geometries: Lamellar LαC (top) and invertand lipid thickness, given by d= δw+ δm; the DNA spacingwithin amonolayer is dDNA, as demright show isoelectric lipoplexes of short DNA fragments of ~146 basepairs and mixtures of DOcorded by freeze-fracture cryo-transmission electronmicroscopy (FF cryo-TEM). Top image shoof the ordered inverted hexagonal phase, composed ofmultiple arrays of densely packed elongain the image by black arrowheads. Bar = 100 nm.Reproduced from Ref. [40] by permission of Cell press.

50 Å for isoelectric complexes, depending onmembrane charge densityand L/D charge ratio [11,25].

In mixtures of bilayer-forming and micellar-favoring lipids, othertypes of structures were occasionally identified. Complexes contain-ing dioleoyl-phosphadtidylethanolamine (DOPE), a non-ionic lipidwith a relatively small headgroup that favors the inverse hexagonalmesophase, were found by Koltover et al. to display a rich phasebehavior that was not observed in the lamellae-favoring DOPC[12]. Mixtures of DOTAP, DOPC and DOPE yielded LαC lipoplexeswhen the fraction of DOTAP was high. However, in mixtures whereDOPE dominated (fraction of 0.7–0.85), complexes formed a two-columnar, inverted hexagonal phase, HII

C—a honeycomb phase. In thisstructure (Fig. 1B) DNA is coated by an inverse lipid monolayer into cy-lindrical micelles, which, in turn, are arranged to a hexagonal lattice,similar to the inverted hexagonal HII phase of pure DOPE in water[39]. Typical diameter of the inverse micelles in HII

C lipoplexes is≈28 Å [12,25].

The structure of equilibrium LαC andHIIC lipoplexesmay bemanipulat-

ed by changes in the environment. Studies demonstrated a dynamictransition from lamellar to hexagonal lipoplexes in isoelectric com-plexes of DNA and mixed neutral and cationic lipids, in response to anosmotic stress exerted by polyethylene glycol (PEG) [40]. The processwas shown to be reversible by, for example, Scarzello et al. [41], sothat HII

C lipoplexes converted back to lamellar ones upon PEG dilution.In another study, ofmixtures of N-methyl-4-(dioleyl)methylpyridinium(SAINT) and DOPE, complexes exhibiting a lamellar phase were con-verted to stable non-lamellar HII structures at high temperatureswhen suspended in a salt solution of physiological concentration.

ed hexagonal HIIC (bottom). The interlayer spacing of the lamellar phase is the sumof water

oted in the drawing (Reproduced from Ref. [11,12] by permission of AAAS). Images on thePC and dioleoylphosphatidyl tri-methylammonium propane at a mole ratio of 3:2, as re-ws the lamellar layered arrangement of the LαC phase. Bottom image discloses the structureted cylinders of DNA coatedwith a lipidmonolayer. Few edges of the cylinders aremarked

233N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

Although lamellar and hexagonal lipoplex structures dominate,some other symmetries were also reported: Complexes of DNA withmixtures of themultivalent cationic lipidMVLBG2 (dendritic headgroupwith +16 charges) and DOPC were found by Ewert et al. to exhibit thelamellar LαC phase at lowMVLBG2 fraction. However, in a narrow rangeof compositions, a dual-lattice structure was detected of hexagonallyarranged normal cylindrical lipid micelles surrounded by DNA, forminga continuous substructure with a hexagonal symmetry (HI

C) [42]. Atlarger mole fractions of MVLBG2 a distorted hexagonal HI

C phase wasfound by Zidovska et al. [43]. Similarly, in sugar-based Gemini surfac-tants, lamellar DNA complexes transitioned into HI

C phase upon acidifi-cation [24,44]. Another phase, the bicontinuous cubic phase (termed QII

and sometimes VII), was found by Bilalov et al. in DNA complexes withlecithin and dodecyltrimethylammonium (DTA) in a narrow composi-tion range [45]. Cubic bicontinuous structures are known to be interme-diates of membrane fusion thus it was envisioned that complexes withthis symmetry will present better biological activity, as indeed wasrecently indicated. Complexes formed between short DNAandmixturesof DOTAP and the nonionic lipid monoolein (GMO, a low water-solubility nonionic lipid known to extensively form non-lamellarmesophases—hexagonal and bicontinuous cubic [46]) were also foundto form cubic bicontinuous phases with short DNA. These phasescould reversibly transition to the inverted hexagonal phase by changein temperature, as found by Leal at al. [47].

The guiding principle that lipoplex geometry is dominated by thelipid-preferred phase is further substantiated by the fact that nucleicacid properties have relatively little effect on the overall structure ofLαC lipoplexes. Similar arrays were reported for both linear lambda-phage DNA and circular plasmid DNA. However, nucleic acid arrange-ment in the intercalating layers depends on such parameters as theDNA length, with a packing density (the repeat spacing betweenadjacent DNA strands) that depends on the lipid mix composition andthe charge ratio [12,25], or the presence of additives [40].

Theoretical and coarse-grained simulation papers investigated thestructure and thermodynamic stability of DNA-containing lipoplexes[27,29,30,32,36,37]. They predicted that complexes structure shouldgenerally match that of the lipid phase (i.e. lamellar or hexagonal)[36] although the presence of DNA affects the lipid packing and, poten-tially, lead to local ‘buckling’ [32], as indeed evident later by experi-ments. The packing density of DNA in lamellar phases was predictedto increase monotonically with the cationic lipid content when theDNA concentration is low [27,29,30,32,36,37]. However, in higher DNApacking densities, Dan [32] predicted potential coexistence betweenclosely packed and more dispersed domains, which was indeed ob-served later by Farago et al. [48].

The discovery of RNA interference (RNAi) as a post-transcriptionalgene-silencing pathway by Fire et al. [49] led to exploration ofcomplexes between small-interfering RNA (siRNA) and cationic lipids(or cationic polymers) for gene therapy of numerous chronic diseasesand cancer [50–53]. Qualitatively, it is expected that siRNA lipoplexeswould be similar to the DNA–lipid complexes. Indeed, numerous stud-ied indicated that like for DNA–lipid systems, the lipid compositionand L/D charge ratio are primary factors affecting the siRNA–lipid struc-tures. Desigaux et al. [19] reported on lamellar lipoplexes formed inmixtures of custom-made aminoglycoside derivatives and siRNA.Small lipoplexes (b200 nm) formed in excess of one of the components,siRNA or cationic lipid, and large and unstable complexes at concentra-tions near charge neutralization (~700 nm) [19]. Two main lamellarmorphologies were detected—of concentric multilamellar structuresand of ordered lamellar microdomains, depending on the class of theaminoglucoside derivative. Importantly, both were shown to form alsoin DNA lipoplexes. Like for DNA, the presence of GMO induced forma-tion of multiple siRNA lipoplex structures [25,54,55]. Lipoplexesdisplayed layered structure (LαsiRNA) in mixtures of DOTAP and GMOat very lowGMOconcentrations [25], and a two-dimensional hexagonalarray of inverse cylindrical micelles decorating siRNA (HII

siRNA) at higher

GMO concentrations. The transition between these two phases tookplace at a given GMO concentration, where both phases coexisted[25,54]. In mixtures containing a pentavalent cationic lipid and highexcess of GMO the bicontinuous, Leal et al. found that a doublegyroid cubic phase formed, with the siRNA molecules incorporatedin the water channels [55]. Characteristic multilamellar siRNA lipidlipoplexes were also found by Aytar et al. [56] in mixtures containingthe reduced form of the cationic lipid BFDMA (bis(11-ferrocenylundecyl)dimethylammonium bromide). Interestingly, lipoplexes of oxidizedBFDMA and siRNA showed primarily amorphous structures.

However, there are some key differences between DNA and siRNA[8]. First, siRNA is hundred times smaller than DNA, which may influ-ence the carriers' size. Further, experiments show that optimizedsiRNA complexes need to be small, typically below 100 nm [8]. Second,siRNA much greater sensitivity to enzymatic degradation stimulatedvarious modification to its backbone and conjugation strategies for im-proved stability [16,57,58]. Finally, several challenges associated withthe biological interfaces—including targeting siRNA to specific tissuesand improving circulation, led to furthermodifications such as the addi-tion of targeting ligands, addition of PEGmoieties for steric stabilization,pH-sensitive groups to promote siRNA release (reviewed recently in[6,8]). Such considerations havemajor effects on siRNA complex design.

3. Studies of the kinetics of lipoplex formation

The complexation of DNA with cationic lipid assemblies is thoughtto consist of at least two stages. The first step is a relatively rapid,electrostatically-driven binding between DNA's negatively chargedphosphates and the lipid cationic charges. This step is followed by a sec-ond, longer stage where rearrangement and condensation take place,resulting in the ordered lipoplex microstructure. The overwhelmingmajority of studies of lipoplex formation kinetics focused on DNA LαC

complexes, where DNA (typically plasmid) is mixed with cationic vesi-cles. However, it is expected that the kinetics of other lipoplex phaseformation (e.g. the hexagonal HII

C), or complexes with siRNA, will followthe same steps (although with different rate constants).

Lai and van Zanten [59] examined lipoplex formation frommixturesof monovalent cationic lipids (1-[2-(9-(Z)-octadecenoyloxy)ethyl]-2-(8-(Z)-heptadecenyl)-3-(hydroxyethyl)imidazolinium chloride(DOTIM), and 1,2-dimyristoyl-sn-glycero-3-ethylphosphocholine(EDMPC) with two neutral lipids (cholesterol and diphytanoyl phos-phatidylethanolamine—DipPE). Time-resolved multi-angle laserlight scattering (TR-MALLS) indicated L/D charge ratio as the primaryfactor determining both the initial rate of lipoplex formation kinetics,and the final lipoplex size, mass and density. These properties werefound to increase with concentration, and were maximal, for a givensystem concentration, at charge ratios near unity. Quite surprisingly,though, the rate at 2:1 charge ratio was not the same as for the 1:2 sys-tem, an asymmetry attributed to the smaller number of complexesformed at excess DNA when compared to the number of complexesformed at excess lipid [59]. It was noted that only the charge ratio, butnot the type of cationic or nonionic lipid, affected the kinetics.

While complexes formed in solutions with high or low charge ratiowere found to reach a steady-state quickly, complexes formed in mix-tures with a charge ratio close to unity displayed a continued growthfor order 10 min following the initial complexation [60]. These findingsare qualitatively consistent with early work by Minsky [61] on DOTAP/DNA systems, showing that complexes that form in excess lipid or ex-cess DNA (charge ratios of 0.5 and 1.5) are stabilized within seconds,while lipoplexes of nearly equal charge ratios (1.1) become stable onlyafter 40min. Lai and van Zanten [60] interpreted the growth as aggrega-tion of the initial (primary) complexes. The growth stage was thoughtto be associatedwith (the lack of) electrostatic repulsion between com-plexes: In highly asymmetric charge composition, excess charge stabi-lizes lipoplexes and prevents their aggregation; Isoelectric aggregatesare more likely to flocculate and coalesce. Indeed, Lai and van Zanten

234 N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

[59] find that decreasing the effects of electrostatic repulsion by usingDMEM/F-12 media (a commonly used cell growth and transfectionmedium which contains, among other components, divalent Ca2+ andMg2+ ions) led to increase in lipoplex size when compared to systemsimmersed in dextrose.

Pozharski andMacDonald [21] examined the effect of vesicle charac-teristics (extruded vs. vortexed) andmixingmethodology on the kinet-ics, using flow fluorometry. Consistent with other reports [59], theyidentified the first, rapid lipoplex formation step as that of DNA coatingcationic lipid vesicles, followed by a longer stage where rearrangementleads to themultilamellar structure of DNA layers alternating with lipidbilayers. However, the initial vesicle coating step occurred on timescales shorter than the method's threshold of 1 min. In equal-chargeratio mixtures of DNA and dioleoyl-O-ethylphosphocholine (EDOPC),DNA-coated vesicles from uniform, extruded vesicles were found to bestable over periods of order hours, while vortexed large vesicles con-verted to multilamellar lipoplexes within several minutes. In systemswhere lipid charges were in excess (charge ratio of 2:1), the stabilityof the coated vesicles was reduced compared to that of isoelectric(1:1) complexes. It was further suggested that in excess DNA, competi-tion between vesicle rupture and vesicle coating determines theamount of lipid that converts into multilamellar lipoplex.

Zhang et al. [62] studied lipoplex formation between DOTAP–DOPEand DOTAP–cholesterol mixtures and plasmid DNA using fluorescenceresonance energy transfer (FRET). As in previous studies, they findthat DNA binding to the cationic vesicles occurs on time scales below1min. However, a subsequent reorganization stage required time scaleson order several hours, although the most significant changes occurredwithin the first 60 min. The time scale for the second stage varied withthe charge ratio: Increasing the ratio of cationic lipid to DNA charges re-duced the time required to reach stability. The effect of the charge ratio,however,was found to differ between complexes containing cholesteroland those containing DOPE. In cholesterol-containing systems, theeffect of charge ratio was weaker than in those containing DOPE, sug-gesting that the uncharged bilayer component in lipoplexes may playan active role in complex formation kinetics.

The kinetics of lipoplex formation by dimethyldioctadecylammoniumbromide (DDAB) and DOTAP with plasmid DNA was studied by Braunel al. [63] as a function of temperature. Two sequential steps were distin-guished for the assembly: One thought to be a binding/dehydration step,and the second a condensation one. The initial rate constant, of the asso-ciation step, was found to decrease with temperature in both DDAB andDOTAP; indeed, above 40 °C it became too rapid for measurement forboth systems.However, the sensitivity of the rate constant to temperaturewasmuchweaker in DDAB (0.086 s at 10 °C and 0.061 s at 30 °C) than inDOTAP (0.25 s at 10 °C to 0.05 s at 30 °C). Thiswas translated to activationenergy for DOTAP that is nearly 5 times higher than that of DDAB. Thetime scales for the second, condensation stage, also decreased withtemperature; however, the activation energy was three times higherfor DDAB than for DOTAP. Unlike other studies, however, the timescales found were similar for the two stages. These measurementsalso suggested that DOTAP condenses DNA more than DDAB, most

Table 1Rate constants and time constants in lipoplex formation. The table is modified after Barreleiro

Charge ratio(+/−)

DODABfraction

1st rate constantk′1 (1/s)

1st time constantτ1 (s)

2nd ratk′2 (1/s

0.25 0.5 13.8 ± 0.7 0.072 ± 0.004 0.08 ±0.75 0.5 12.8 ± 1.1 0.078 ± 0.007 0.17 ±1.5 0.5 28.6 ± 2.7 0.035 ± 0.003 0.14 ±3 0.5 23.7 ± 2.2 0.042 ± 0.004 0.49 ±5 0.5 24.5 ± 1.5 0.041 ± 0.002 0.64 ±3 0.33 27.6 ± 2.5 0.036 ± 0.003 0.45 ±3 0.5 23.7 ± 2.2 0.042 ± 0.004 0.49 ±3 0.67 16.5 ± 1.3 0.061 ± 0.005 0.24 ±3 1.0 16.6 ± 1.4 0.060 ± 0.005 0.13 ±

likely due to differences between the lipid headgroup interactionswith DNA.

Contrary to the two-stage process discussed above, Lindman [4,5]suggested a three-step mechanism that follows first-order reactionsfor the formation of multilamellar lipoplexes. At first, over a period oforder milliseconds, DNA adsorbs onto the oppositely charged mem-brane surface (process is controlled by electrostatic interactions). Inthe next stage DNA–lipid complexes aggregate and grow at time scalesof seconds (possibly by another cationic bilayer surrounding the DNAmolecule). During these two stages, the secondary structure of theDNA becomes more compact, so that tertiary folding exposes morebinding sites to the cationic bilayer. The third, final stage, which occursover long time scales (ranging from 100 s to hours), involves furtherchanges in DNA conformations, and rearrangement of complexes, prob-ably into multilamellar structures. Time constants from stopped-flowturbidity and fluorescence experiments [4] were found to vary withthe DNA to lipid charge ratio and the lipid composition (extruded vesi-cle mixtures of the cationic dioctadecyldimethlammonium bromide(DODAB) and DOPE) as listed in Table 1.

In systems with a fixed lipid composition, the time constant for theinitial complex formation depends on the charge ratio: Complexesformed with excess DNA (0.24 and 0.75 charge ratio in Table 1) have atime constant that is about twice as much as that of complexes withthe same lipid composition but excess lipids (1.5, 3 and 5). The timeconstant for the second stage, however, decreases with increasing L/Dcharge ratio. In contrast, both first and second stage constants increasewith the fraction of cationic lipid, DODAB (under excess cationic lipidconditions). This was attributed to the fact that increasing DODABfraction increases the cationic charge density and decreases the bilayerfluidity, thereby hindering rearrangement (pure DODAB bilayers are inthe gel state). The rate constant for the third stage is almost indepen-dent of both charge ratio and lipid composition.

Neutron scattering experiments by Barreleiro et al. of DODAB/DOPElipoplexes supported the 3-step model, extending the 3rd time scale ofcomplexes reorganization into multilamellar structures to order of~10 min. The authors correlated the second step to the formation ofan intermediate complex with a locally cylindrical structure developingvia vesicle rupture events and rolling up of the cationic bilayers aroundthe DNA double helix axis [5]. The third step was explained by furtherlayer-to-layer association of previously ruptured vesicles leading tocontinuous aggregation and growth of the complexes into, eventually,multilamellar structures [5]. The 3-step mechanism and lipoplex devel-opment through binding and rupture events are detailed in Fig. 2.

As noted, the last stage of lipoplex formation requires significant re-arrangement of the lipids and nucleic acid within the aggregate. Lealet al. [64] investigated the mobility of lipids and DNA in lipoplexes,using both solid state and diffusion NMR in systems composed ofcationic and zwitterionic lipids. They find that the zwitterionic lipiddisplayed diffusion coefficients typical for lipids in bilayers, while thecationic species diffusion rate was an order of magnitude slower. TheDNA was found to be relatively static [64]. Their findings suggest thatthe DNA induced lipid segregation within the bilayer, resulting in

and Lindman [4], with permission of ACS publications.

e constant)

2nd time constantτ2 (s)

3rd rate constantk′3 (1/s)

3rd time constantτ3 (s)

0.02 12.5 ± 3.1 0.006 ± 0.001 167 ± 280.03 5.9 ± 1.0 0.005 ± 0.002 200 ± 800.02 7.1 ± 1.0 0.005 ± 0.002 200 ± 800.05 2.0 ± 0.2 0.006 ± 0.001 167 ± 280.06 1.6 ± 0.1 0.006 ± 0.001 167 ± 280.02 2.2 ± 0.1 0.007 ± 0.001 143 ± 200.05 2.0 ± 0.2 0.006 ± 0.001 167 ± 280.04 4.2 ± 0.7 0.007 ± 0.002 143 ± 200.03 7.7 ± 1.7 0.006 ± 0.001 167 ± 28

Fig. 2. Proposedmechanism for the formation of DNA–cationic vesicle lipoplexes in excesslipid, showingmultilamellar structure by vesicle rupture, and the three time scales associ-ated with the 3-step mechanism.Reproduced from [5] by permission of RSC.

235N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

regions of pure lipid (mostly the zwitterionic component) with typicalmobility, and regions of DNA-bound (cationic) lipids, where the DNAis immobilized, and the lipid mobility is suppressed.

4. Discussion

Previous studies of lipoplex formation have clearly establishedthat the structure of equilibrium lipoplexes is governed by threemain parameters (i) the nature of the lipids, (ii) lipid composition(mole fraction of the cationic lipid in the lipid mixture, type of helperlipid), and (iii) the charge ratio between the lipids and the nucleicacid [13].

Fig. 3. The initial kinetic steps in lipoplex formation: Illustration (modified after [67]) and timevesicles, and subsequent growth by successive binding events. (A) A partially decorated vesiclbetween two vesicles, while another part of the bound DNA is exposed to the bulk solution (time (t ~ 30 s), this primary complex grows through further binding events, with (the majoriratio 1:1; excess lipid (Danino D, Talmon Y and Barenholz Y, unpublished results). Bars = 100

The initial step in lamellar lipoplexes formation is thought to consistof electrostatically driven binding between the anionic nucleic acid andthe cationic vesicles. This rate was shown by Braun et al. to follow Ar-rhenius kinetics, with an activation energy that depends on lipid type[63]. Lai and van Zanten found that the rate increases linearly with theconcentration (for fixed L/D ratio) [59], but varied non-monotonicallywith the charge ratio, reaching a maximum in systems with chargeratio close to unity. These findings are consistent with the view that,in this initial step, the nucleic acid ‘adsorbs’ on the outer surface of thevesicles, as shown in Fig. 3.

As in the adsorption of polymers on oppositely charged surfaces, thisinitial step is a rapid and irreversible process occurring on a time scale ofmilliseconds. Since the probability of binding is high once oppositelycharged groups come within a specific distance, increasing the concen-tration of charges increases the rate of binding. This can be obtained byeither increasing the overall concentration of the components (nucleicacid and lipids), or by increasing the density of charges in the lipidassemblies. However, the system is not symmetric: Binding occursbetween the nucleic acid charges (which is proportional to the nucleicacid concentration) and the lipid charges available on the outer surfaceof the vesicles. As a result, for a given overall lipid concentration,increasing the concentration of nucleic acid should increase the rate ofbinding by a factor of 2 when compared to increasing the cationiccharges by the same amount—as indeed observed. ‘Diluting’ the cationiclipids by mixing with a nonionic component reduces the membranecharge density, i.e., the density of potential adsorption site. This reducesthe probability – and rate – of adsorption, and lowers the efficiency ofDNA neutralization, even if the charge ratio is held fixed [13,65,66].

Studies of the kinetics of lipoplex formation focused on lamellae–forming lipid mixtures that yield LαC liquid crystalline complexes. Theinitial step in the formation of hexagonal HII

C and other non-lamellartypes of complexes is expected to be similar, for all cases where thelipids are introduced into the mix in the form of vesicles.

While the initial adsorption step is rapid and largely independent ofmost system parameters, the preparation route plays a significant rolein both the kinetics of the subsequent lipoplex formation steps andthe global properties (overall size and heterogeneity) of the final com-plexes. Surprisingly, this issue is under-emphasized in the literature.The evolving complexes will depend on the local ‘environment’ at anygiven time point, controlled by the overall concentration and masstransfer. As a result, lipoplexes obtained via rapid mixing of vesicle

-resolution cryo-TEM data following the initial binding of DNA to the outer surface of lipide found immediately after addition of DNA to lipid vesicles. Part of the DNA is condensedwhite arrowhead). The vesicle itself is also only partially decorated with DNA. (B) Overty of) lipid vesicles remaining intact. Data of the system composed of DOTAP/DOPE molenm.

Fig. 4. Cryo-TEM images of DOTAP/DOPE (as in Fig. 3), presenting characteristic structures~1 min after addition of DNA. Upper panel is showing a large lipoplex with some vesiclesthat remained intact, and other vesicles that underwent rupture into thick ring-like struc-tures (see model in Fig. 6). Lower image is showing a lipoplex, N1 μm in length just beforeprecipitation, and at the background smaller vesicles and ‘flower-like’ structures charac-teristics of the DOPE hexagonal phase. (Danino D, Talmon Y and Barenholz Y, unpublishedresults).

Fig. 5.A typical lipoplex found inDOSPA/DOPEmixtures (total lipid concentration 2.5mM,cationic-to-helper ratio is 1:1, L/D = 4). Part of the DNA is condensed between lipidbilayers (arrowheads), as often seen with other lipids. Large regions of DNA and DOSPAmicelles are also found between the lipid bilayers. Arrows point to individual spheroidalmicelles as those forming by DOSPA in aqueous solutions. Schematic drawing from Ref[68], with permission; Cryo-TEM image is from Ref [13], with permission.

236 N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

solutions with nucleic acid solutions will display different kinetics andlarger-scale structures than those obtained by the drop-wise additionof one component to the other. When solutions are mixed togetherthe local nucleic acid and lipid concentrations are similar to the global,end-result values, while when pipetting solution of one component tothe solution of the other, a commonly applied protocol, complexationstarts with highly asymmetric, and potentially inhomogeneous condi-tions, which is also very different from the final solution composition.Some of the resulting effects are irreversible. Furthermore, propertiesof the vesicles themselves as size and polydispersity may significantlyaffect the kinetics: under identical complexation conditions DNA-coatedvesicles from uniform, extruded vesicles were found by Pozharskiet al. to be stable over periods of order hours, while large (and likely

polydispersed) vortexed vesicles converted to multilamellar lipoplexeswithin several minutes [21].

In the next kinetic stage of lipoplex formation, DNA–lipid complexesaggregate and grow. This stage, which occurs on the order of seconds tominutes, is highly sensitive to the properties of the lipids, and in partic-ular to the rigidity, or resistance to rupture, of the bilayer. As shown inFigs. 3 to 7, the process of complex aggregation differs greatly for vesi-cles/nucleic acid complexes that remain intact (Figs. 3–5), and those un-dergoing rupture (Figs. 6 and 7). The first typically leads to growth oflarge unorganized complexes, while the second to more ordered andrelatively uniform structures. In either process, the irreversibility ofthe initial binding step means that various binding sites (both on thevesicles and on the nucleic acid) may remain exposed; these act as‘sticky’ points for further events.

Figs. 3 and 4 present cryo-TEM results that follow the complexationof DNA with extruded DOTAP/DOPE vesicles (1:1 mole ratio betweencationic and helper lipids). In a primary complex shown in Fig. 3Abound DNA is partially condensed between two intact vesicles, andpartially exposed to the bulk. Such a complex forms immediatelyand continues to evolve by successive similar binding events(Fig. 3B). Complexation is rapid and within less than a minute theaggregates are big (see Fig. 4), and they continue to grow and precip-itate out.

Fig. 5 presents and even more limiting case, of 1:1 mole ratio ofDOPE and DOSPA (2,3-dioleyloxy-N-[2(sperminecarboxamido)ethyl]-N,N-dimethyl-1-propanammonium trifluoroacetate) at charge ratio of4. DOSPA is a polycationic lipid that in solution forms small micelles(~10 nm in diameter). Even the presence of the potent helper lipidDOPE can mediate the formation of lamellar structures (vesicles), butnot of inverted hexagonal structures due to the high resistance ofDOSPA to bending. DNA binds to the lipids, but the following steps inlipoplex formation are hindered—the vesicles remain intact, hardlychange their curvature, and complexes grow but do not arrange to anordered mesophase [13].

In contrast, membranes which undergo rupture readily form or-dered, uniform structures, relatively rapidly. Models for complexationvia vesicle rupture are shown in Figs. 6 and 7 along with cryo-TEMdata confirming the formation of the suggested lipoplexes. Specifically,the model proposed in Fig. 7 (based on the model presented in Fig. 2[5]) considers the formation of multi-domain lipoplexes via vesicle rup-ture events and binding through ‘sticky’points on the surface of vesicles,DNA and small multilayered lipoplexes. Note that some concentricmultilamellar lipoplexes assembly on top of the primary vesicle, whilein other complexes even the primary vesicle underwent rupture.

A B

Fig. 6. Proposed model for the formation of concentric multilamellar lipoplexes trough successive events of DNA binding and vesicle rupture [67]. (A) Cryo-TEM image of concentricmultilamellar DNA lipoplexes with DMPC/DC-Cholesterol (3β[N-(N,s-dimethyl aminoethane)-carbamoyl] cholesterol, mole ratio = 3:2); reproduced with permission from Ref [69].(B) Complexes of similar global and local organization that form between the positively charged Aminoglycoside (AG) antibiotic gentamicin and the anionic lipid DOPS(dioleoylphosphatidylserine) (Danino D, Hollander A and Baasov T, unpublished results). Note in (B) stages in the assembly of the multilamellar complex. Inset shows small complexesas those described in stages 3 and 4 in the drawing. Bars = 100 nm.

237N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

In the last final stage of lipoplex formation, the local order of thenucleic acid develops. During this stage, which can occur on the orderof hours, the global characteristics of the lipoplex remain unchanged:The number of nucleic acid and lipid molecules in the complex isfixed. However, their local organization evolves towards the preferred,equilibrium nanostructure.

As in any irreversible adsorption process of multi-charged compo-nents, the initial binding results in a random distribution of the nucleicacid. The mobility of the lipids in the bilayers, however, allows rear-rangement until reaching the equilibrium (lamellar or hexagonal)array. In the case of lamellar structures, the nucleic acid must rearrangeinto the two-dimensional, long range ordered array shown in Fig. 2. Asfound by Leal et al. [64], the close binding betweenDNAand the cationiclipids inhibits themobility of the DNA in the complex, and slows the dif-fusion rate of the associated cationic lipids. In fluidmembranes, lipid re-arrangement in the bilayer plane can still lead to the equilibrium orderrelatively quickly (on time scales similar to the lipid two-dimensional

A B

Fig. 7. (A) Model proposing the development of multi-domain lamellar lipoplexes through succ(or nucleic acid and vesicles to such small complexes) through binding sites on their surfaces. NBFDMA) lipoplexes (B), and siRNA/lipid (DOST, dioleyl succinyl tobramycin) lipoplexes (C, D). (B= 50 nm.

diffusion), as found by Maier and Radler [71]. However, in bilayerswith low fluidity, or that are in the gel phase, this reorganization maybe suppressed.

Overall, the kinetics of non-lamellar complexes is expected to followthe same three-step process: First, DNA adsorption onto the lipidaggregates, followed by aggregate–aggregate condensation, andthen by internal rearrangement. Thus, the time scales for the first twosteps should be relatively similar regardless of the end-complex ge-ometry. However, the time scale for the third stage may differ great-ly, depending on the type of lipids and the end-phase: The transitionfrom the nucleic-acid adsorbed onto vesicle surfaces obtained in thesecond stage, to hexagonal or bicontinuous cubic structures, requiresmore radical reorganization than the formation of the multi-lamellarcomplexes. The bilayer must bend, locally, to ‘wrap’ around thenucleic acid. In the case of the hexagonal arrays, a decoupling ofthe two monolayers composing the bilayer is also required. Indeed,rigid membrane would require extremely long time scales to reach

C D

essive binding and rupture events, and possible ‘sticking’ of small lipoplexes to each otherote in the cryo-TEM images the similar global and internal structure of DNA/lipid (reduced)Modified after [70]with permission, (C andD)Modified after [19], with permission. Bars

238 N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

such equilibrium structure, or may even be kinetically trapped in anintermediate form.

The type and properties of the nucleic acid are also expected to affectcomplex formation. In the first stage where the nucleic acid binds to thesurface of lipid aggregates, the decrease in the diffusion rate of chainmolecules with increasing chain length [35] may slow the rate ofassociation to some degree, although the time scales are still expectedto be of the same order of magnitude. The second stage is dominatedby the mobility of the nucleic-acid covered aggregates, and is thus notexpected to be affected significantly by the nucleic acid characteristics.However, in the third, re-arrangement stage, shorter nucleic acids areexpected to be more mobile, and inhibit the associated lipid mobilityless than that of longer chains. Thus, the time scales for the last stagemay differ greatly.

5. Summary and outlook

The equilibrium structure of nucleic acid complexes with cationiclipids or lipid mixtures, has been widely studied. Generally, lipoplexstructure is set by the lipid properties and is less dependent on thenucleic acid characteristics. Transfection efficacy has been clearly linkedto the internal, nano-scale organization of the complexes (e.g. lamellarvs. hexagonal arrays), as well as to more global characteristics such asthe complex size and net charge.

Understanding the kinetics of lipoplex formation is of fundamentalinterest, as a multi-component, multi-length scale and multi-timescale system. It also has significant implications for the utilization oflipoplexes as carriers for gene delivery and gene silencing, due to thestructure–function efficacy correlations: Lipoplex kinetics determineboth the rate of nanostructure evolution, and the global properties ofthe complex.

The data clearly suggests a three-step process. Initially, the nucleicacid adsorbs onto the surface of the cationic vesicles. This step is rapid(occurringwithin less than aminute in all systems studied), and largelyinsensitive to system parameters. The second step is one where thenucleic-acid carrying vesicles flocculate to form larger aggregates. Thisstep occurs on intermediate time scales, and is sensitive to both thecharge ratio and lipid properties. In particular, the rate and type oflipoplex formed dependon thepropensity of the lipid bilayer to rupture.In the third step, local rearrangement occurs, while the global propertiesof the lipoplex remain constant. This process may require hours (or evenlonger), and is largely set by the bilayer characteristics. Fluid membraneswith a tendency to rupture would allow rapid rearrangement, whilerigid, gel-like ones may potentially be permanently trapped in a non-equilibrium structure. As a result, time scales may vary greatly.

Distinguishing between the flocculation stage and the rearrange-ment rate also requires understanding of the measurement type:Methods that track the size andmass of aggregatesmeasureflocculation(aggregation). As a result, ‘steady-state’ under such conditions indicatesreaching the end of overall complex growth, namely the second stage.However, such methods cannot distinguish the third stage, which maycontinue beyond the time frame of complex growth. In contrast,molecular-type probes can distinguish and follow rearrangement, andthus the third stage, but are not sensitive to the second step flocculationand growth. This explains the large discrepancy in time scales reportedfor the second/third steps.

It should be emphasized that, although the three-stage processshould be universal, and is expected to apply even when mixing non-vesicle lipid phases with nucleic acids, the specific rates are highly sen-sitive to the lipid properties, as discussed.

Understanding the kinetics of lipoplex formation is not limited toimproving the (arguably broad) efficiency of nucleic-acid transfection.Similar complexes are currently being investigated for addressingchallenges of current medicine, e.g., fighting microbial resistance toantibiotics [72–74], or treating genetic diseases (e.g., cystic fibrosis(CF) and Duchenne muscular dystrophy (DMD)) caused by alterations

in the genetic code that lead to the production of non-functional pro-teins. An example of vehicles designed for the latter is presented inFig. 6B. These lipoplexes formed by the positively charged aminogly-coside antibiotic gentamicin and the anionic lipid DOPS (dioleoylphosphatidylserine), drug/lipid lipoplexes formed spontaneously,into concentric andmultilamellar lipoplexes resembling in morphol-ogy and order nucleic acid/lipid lipoplexes, and guided by the sameinteractions that lead to the creation of nucleic acid/lipid lipoplexes.

Better understanding of lipoplex kinetics will allow better design ofeffective nucleic acid carriers, aswell as shed light onmethodologies forimproving any type of similar lipid complexes, thereby opening excitingopportunities for new applications in biotechnology and nanomedicine.The rapid advances inmicrofluidicmethodologies promise the potentialfor production of homogeneous, small, uniform and effective carriers[75,76] thereby overcoming heterogeneity and reproducibility difficul-ties of traditional preparation routes.

This review emphasizes the kinetics of lipoplex formation, whichdetermines the physicochemical properties of the complexes, and even-tually the efficacy of the therapeutic application. Many additional im-portant factors come into play and affect transfection efficiency underphysiological conditions or even in transfection cell medium; thesecan be considered as another phase in the kinetics of complexes,which is of great practical importance. The interaction of lipoplexeswith the biological interface is summarized in numerous recent reviews(e.g., Tros de Ilarduya et al., [7]; Schroeder et al., [8]; Wasungu andHoekstra [24], and references therein).

References

[1] Lindman B, Wennerstrom H. Miceles. Amphiphile aggregation in aqueous solution.Top Curr Chem 1980;87:1–87.

[2] Mel'nikov SM, Lindman B. Solubilization of DNA–cationic lipid complexes in hy-drophobic solvents. A single-molecule visualization by fluorescence microscopy.Langmuir 1999;15(6):1923–8.

[3] McLoughlin D, Dias R, Lindman B, Cardenas M, Nylander T, Dawson K, et al. Surfacecomplexation of DNA with insoluble monolayers. Influence of divalent counterions.Langmuir 2005;21(5):1900–7.

[4] Barreleiro PCA, Lindman B. The kinetics of DNA-cationic vesicle complex formation.J Phys Chem B 2003;107(25):6208–13.

[5] Barreleiro PC, May RP, Lindman B. Mechanism of formation of DNA–cationic vesiclecomplexes. Faraday Discuss 2003;122:191–201.

[6] Mintzer MA, Simanek EE. Nonviral vectors for gene delivery. Chem Rev2009;109(2):259–302.

[7] Tros de Ilarduya C, Sun Y, Duzgunes N. Gene delivery by lipoplexes and polyplexes.Eur J Pharm Sci 2010;40(3):159–70.

[8] Schroeder A, Levins CG, Cortez C, Langer R, Anderson DG. Lipid-based nanotherapeuticsfor siRNA delivery. J Intern Med 2010;267(1):9–21.

[9] Ewert KK, Zidovska A, AhmadA, BouxseinNF, Evans HM,McAllister CS, et al. Cationicliposome–nucleic acid complexes for gene delivery and silencing: pathways andmechanisms for plasmid DNA and siRNA. Top Curr Chem 2010;296:191–226.

[10] Li S, Huang L. Nonviral gene therapy: promises and challenges. Gene Ther2000;7(1):31–4.

[11] Radler JO, Koltover I, Salditt T, Safinya CR. Structure of DNA–cationic liposome com-plexes: DNA intercalation in multilamellar membranes in distinct interhelical pack-ing regimes. Science 1997;275(5301):810–4.

[12] Koltover I, Salditt T, Radler JO, Safinya CR. An inverted hexagonal phase of cat-ionic liposome–DNA complexes related to DNA release and delivery. Science1998;281(5373):78–81.

[13] Simberg D, Danino D, Talmon Y, Minsky A, Ferrari ME, Wheeler CJ, et al. Phase be-havior, DNA ordering, and size instability of cationic lipoplexes—relevance to opti-mal transfection activity. J Biol Chem 2001;276(50):47453–9.

[14] EvansHM, AhmadA, Ewert K, Pfohl T,Martin-HerranzA, BruinsmaRF, et al. Structur-al polymorphism of DNA–dendrimer complexes. Phys Rev Lett 2003;91(7):075501.

[15] Merkel OM, Mintzer MA, Sitterberg J, Bakowsky U, Simanek EE, Kissel T. Triazinedendrimers as nonviral gene delivery systems: effects of molecular structure on bi-ological activity. Bioconjug Chem 2009;20(9):1799–806.

[16] Juliano R, AlamMR, Dixit V, Kang H. Mechanisms and strategies for effective deliveryof antisense and siRNA oligonucleotides. Nucleic Acids Res 2008;36(12):4158–71.

[17] de Fougerolles A, Vornlocher HP, Maraganore J, Lieberman J. Interfering with disease: aprogress report on siRNA-based therapeutics. Nat Rev Drug Discov 2007;6(6):443–53.

[18] Akinc A, Goldberg M, Qin J, Dorkin JR, Gamba-Vitalo C, Maier M, et al. Developmentof lipidoid–siRNA formulations for systemic delivery to the liver. Mol Ther2009;17(5):872–9.

[19] Desigaux L, Sainlos M, Lambert O, Chevre R, Letrou-Bonneval E, Vigneron JP, et al.Self-assembled lamellar complexes of siRNA with lipidic aminoglycoside derivativespromote efficient siRNA delivery and interference. Proc Natl Acad Sci U S A2007;104(42):16534–9.

239N. Dan, D. Danino / Advances in Colloid and Interface Science 205 (2014) 230–239

[20] Kennedy MT, Pozharski EV, Rakhmanova VA, MacDonald RC. Factors governing theassembly of cationic phospholipid–DNA complexes. Biophys J 2000;78(3):1620–33.

[21] Pozharski EV, MacDonald RC. Single lipoplex study of cationic lipoid–DNA, self-assembled complexes. Mol Pharm 2007;4(6):962–74.

[22] MaBC, Zhang SB, JiangHM, ZhaoBD, LvHT. Lipoplexmorphologies and their influenceson transfection efficiency in gene delivery. J Control Release 2007;123(3):184–94.

[23] KapoorM, Burgess DJ, Patil SD. Physicochemical characterization techniques for lipidbased delivery systems for siRNA. Int J Pharm 2012;427(1):35–57.

[24] Wasungu L, Hoekstra D. Cationic lipids, lipoplexes and intracellular delivery ofgenes. J Control Release 2006;116(2):255–64.

[25] Safinya CR, Ewert KK, Leal C. Cationic liposome–nucleic acid complexes: liquid crys-tal phases with applications in gene therapy. Liq Cryst 2011;38(11–12):1715–23.

[26] Ross PD, Shapiro JT. Heat of interaction of DNA with polylysine, spermine, andMg++. Biopolymers 1974;13(2):415–6.

[27] May S, Ben-Shaul A. DNA–lipid complexes: stability of honeycomb-like andspaghetti-like structures. Biophys J 1997;73(5):2427–40.

[28] Hirsch-Lerner D, Barenholz Y. Hydration of lipoplexes commonly used in gene deliv-ery: follow-up by laurdan fluorescence changes and quantification by differentialscanning calorimetry. BBA Biomembr 1999;1461(1):47–57.

[29] Harries D, May S, Gelbart WM, Ben-Shaul A. Structure, stability, and thermodynam-ics of lamellar DNA–lipid complexes. Biophys J 1998;75(1):159–73.

[30] Bruinsma R. Electrostatics of DNA cationic lipid complexes: isoelectric instability. EurPhys J B 1998;4(1):75–88.

[31] Ahsan A, Rudnick J, Bruinsma R. Elasticity theory of the B-DNA to S-DNA transition.Biophys J 1998;74(1):132–7.

[32] Dan N. Formation of ordered domains in membrane-bound DNA. Biophys J1996;71(3):1267–72.

[33] Pozharski E, MacDonald RC. Thermodynamics of cationic lipid–DNA complex forma-tion as studied by isothermal titration calorimetry. Biophys J 2002;83(1):556–65.

[34] Pozharski E, MacDonald RC. Lipoplex thermodynamics: determination of DNA–cationic lipoid interaction energies. Biophys J 2003;85(6):3969–78.

[35] Netz RR, Andelman D. Neutral and charged polymers at interfaces. Phys Rep2003;380(1–2):1–95.

[36] Dan N. The structure of DNA complexes with cationic liposomes—cylindrical or flatbilayers? BBA Biomembr 1998;1369(1):34–8.

[37] Dan NL. Multilamellar structures of DNA complexes with cationic liposomes.Biophys J 1997;73(4):1842–6.

[38] Koltover I, Salditt T, Safinya CR. Phase diagram, stability, and overcharging of lamel-lar cationic lipid–DNA self-assembled complexes. Biophys J 1999;77(2):915–24.

[39] Tilcock CP, Bally MB, Farren SB, Cullis PR. Influence of cholesterol on the structuralpreferences of dioleoylphosphatidylethanolamine-dioleoylphosphatidylcholine sys-tems: a phosphorus-31 and deuterium nuclear magnetic resonance study. Biochem-istry 1982;21(19):4596–601.

[40] Danino D, Kesselman E, Saper G, Petrache HI, Harries D. Osmotically induced revers-ible transitions in lipid–DNA mesophases. Biophys J 2009;96(7):L43–5.

[41] Scarzello M, Chupin V, Wagenaar A, Stuart MC, Engberts JB, Hulst R. Polymorphismof pyridinium amphiphiles for gene delivery: influence of ionic strength, helper lipidcontent, and plasmid DNA complexation. Biophys J 2005;88(3):2104–13.

[42] Ewert KK, Evans HM, Zidovska A, Bouxsein NF, Ahmad A, Safinya CR. A columnarphase of dendritic lipid-based cationic liposome–DNA complexes for gene delivery:hexagonally ordered cylindrical micelles embedded in a DNA honeycomb lattice. JAm Chem Soc 2006;128(12):3998–4006.

[43] Zidovska A, Evans HM, Ewert KK, Quispe J, Carragher B, Potter CS, et al. Liquid crys-talline phases of dendritic lipid–DNA self-assemblies: lamellar, hexagonal, and DNAbundles. J Phys Chem B 2009;113(12):3694–703.

[44] Wasungu L, StuartMC, ScarzelloM, Engberts JB, Hoekstra D. Lipoplexes formed fromsugar-based gemini surfactants undergo a lamellar-to-micellar phase transition atacidic pH. Evidence for a non-inverted membrane-destabilizing hexagonal phaseof lipoplexes. Biochim Biophys Acta 2006;1758(10):1677–84.

[45] Bilalov A, Olsson U, Lindman B. A cubic DNA–lipid complex. Soft Matter2009;5(20):3827–30.

[46] Larsson K. Two cubic phases in monoolein–water system. Nature 1983;304:5927.[47] Leal C, Ewert KK, Bouxsein NF, Shirazi RS, Li Y, Safinya CR. Stacking of short DNA

induces the gyroid cubic-to-inverted hexagonal phase transition in lipid–DNA com-plexes. Soft Matter 2013;9(3):795–804.

[48] Farago O, Ewert K, Ahmad A, Evans HM, Gronbech-Jensen N, Safinya CR. Transitionsbetween distinct compaction regimes in complexes of multivalent cationic lipidsand DNA. Biophys J 2008;95(2):836–46.

[49] Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello CC. Potent and specificgenetic interference by double-stranded RNA in Caenorhabditis elegans. Nature1998;391(6669):806–11.

[50] Davidson BL, Paulson HL. Molecular medicine for the brain: silencing of diseasegenes with RNA interference. Lancet Neurol 2004;3(3):145–9.

[51] Ponnappa BC. siRNA for inflammatory diseases. Curr Opin Investig Drugs2009;10(5):418–24.

[52] Whitehead KA, Langer R, Anderson DG. Knocking down barriers: advances in siRNAdelivery. Nat Rev Drug Discov 2009;8(2):129–38.

[53] Takeshita F, Ochiya T. Therapeutic potential of RNA interference against cancer. Can-cer Sci 2006;97(8):689–96.

[54] Leal C, Bouxsein NF, Ewert KK, Safinya CR. Highly efficient gene silencing activity ofsiRNA embedded in a nanostructured gyroid cubic lipid matrix. J Am Chem Soc2010;132(47):16841–7.

[55] Leal C, Ewert KK, Shirazi RS, Bouxsein NF, Safinya CR. Nanogyroids incorporatingmultivalent lipids: enhanced membrane charge density and pore forming abilityfor gene silencing. Langmuir 2011;27(12):7691–7.

[56] Aytar BS, Muller JP, Kondo Y, Talmon Y, Abbott NL, Lynn DM. Redox-based control ofthe transformation and activation of siRNA complexes in extracellular environmentsusing ferrocenyl lipids. J Am Chem Soc 2013;135(24):9111–20.

[57] Kim DH, Rossi JJ. Strategies for silencing human disease using RNA interference. NatRev Genet 2007;8(3):173–84.

[58] De Paula D, Bentley MV, Mahato RI. Hydrophobization and bioconjugation for en-hanced siRNA delivery and targeting. RNA 2007;13(4):431–56.

[59] Lai E, van Zanten JH. Real time monitoring of lipoplex molar mass, size and density. JControl Release 2002;82(1):149–58.

[60] Lai E, van Zanten JH. Monitoring DNA/poly-L-lysine polyplex formation with time-resolved multiangle laser light scattering. Biophys J 2001;80(2):864–73.

[61] Gershon H, Ghirlando R, Guttman SB, Minsky A. Mode of formation and structuralfeatures of DNA–cationic liposome complexes used for transfection. Biochemistry1993;32(28):7143–51.

[62] Zhang Y, Garzon-Rodriguez W, Manning MC, Anchordoquy TJ. The use of fluores-cence resonance energy transfer to monitor dynamic changes of lipid–DNA interac-tions during lipoplex formation. Biochim Biophys Acta 2003;1614(2):182–92.

[63] Braun CS, Fisher MT, Tomalia DA, Koe GS, Koe JG, Middaugh CR. A stopped-flow ki-netic study of the assembly of nonviral gene delivery complexes. Biophys J2005;88(6):4146–58.

[64] Leal C, Sandstrom D, Nevsten P, Topgaard D. Local and translational dynamics inDNA–lipid assemblies monitored by solid-state and diffusion NMR. BBA Biomembr2008;1778(1):214–28.

[65] Zuidam NJ, Barenholz Y. Electrostatic parameters of cationic liposomes commonlyused for gene delivery as determined by 4-heptadecyl-7-hydroxycoumarin. BiochimBiophys Acta 1997;1329(2):211–22.

[66] Zuidam NJ, Barenholz Y. Electrostatic and structural properties of complexes involv-ing plasmid DNA and cationic lipids commonly used for gene delivery. BiochimBiophys Acta 1998;1368(1):115–28.

[67] Huebner S, Battersby BJ, Grimm R, Cevc G. Lipid–DNA complex formation: reorgani-zation and rupture of lipid vesicles in the presence of DNA as observed bycryoelectron microscopy. Biophys J 1999;76(6):3158–66.

[68] Bertino G, Benazzo M, Gatti P, Bernardo G, Corbella F, Tinelli C, et al. Curative andorgan-preserving treatment with intra-arterial carboplatin induction followed bysurgery and/or radiotherapy for advanced head and neck cancer: single-centerfive-year results. BMC Cancer 2007;7:62.

[69] Battersby BJ, Grimm R, Huebner S, Cevc G. Evidence for three-dimensional interlayercorrelations in cationic lipid–DNA complexes as observed by cryo-electron micros-copy. Biochim Biophys Acta 1998;1372(2):379–83.

[70] Aytar BS, Muller JPE, Golan S, Kondo Y, Talmon Y, Abbott NL, et al. Chemical oxida-tion of a redox-active, ferrocene-containing cationic lipid: influence on interactionswith DNA and characterization in the context of cell transfection. J Colloid InterfaceSci 2012;387:56–64.

[71] Maier B, Radler JO. Conformation and self-diffusion of single DNA molecules con-fined to two dimensions. Phys Rev Lett 1999;82(9):1911–4.

[72] Epand RF, Sarig H, Ohana D, Papahadjopoulos-Sternberg B, Mor A, Epand RM. Phys-ical properties affecting cochleate formation and morphology using antimicrobialoligo-acyl-lysyl peptidemimetics andmixturesmimicking the composition of bacterialmembranes in the absence of divalent cations. J Phys Chem B 2011;115(10):2287–93.

[73] Epand RF, Mor A, Epand RM. Lipid complexes with cationic peptides and oaks; theirrole in antimicrobial action and in the delivery of antimicrobial agents. Cell Mol LifeSci 2011;68(13):2177–88.

[74] Sarig H, Ohana D, Epand RF, Mor A, Epand RM. Functional studies of cochleate as-semblies of an oligo-acyl-lysyl with lipid mixtures for combating bacterial multidrugresistance. FASEB J 2011;25(10):3336–43.

[75] Chen DL, Love KT, Chen Y, Eltoukhy AA, Kastrup C, Sahay G, et al. Rapid discovery ofpotent siRNA-containing lipid nanoparticles enabled by controlled microfluidic for-mulation. J Am Chem Soc 2012;134(16):6948–51.

[76] Leung AKK, Hafez IM, Baoukina S, Belliveau NM, Zhigaltsev IV, Afshinmanesh E, et al.Lipid nanoparticles containing siRNA synthesized by microfluidic mixing exhibit anelectron-dense nanostructured core. J Phys Chem C 2012;116(41):22104.