Biophysics and protein corona analysis of Janus cyclodextrin-DNA nanocomplexes. Efficient cellular transfection on cancer cells

1 Biophysics and protein corona analysis of Janus cyclodextrin-DNA nanocomplexes. Efficient cellular transfection on cancer cells M. Martínez-Negroa, G. Caracciolob, S. Palchettib, D. Pozzib, A. L. Capriottic, C. Cavalierec, A. Laganàc, C. Ortiz Melletd, J. M. Benitoe, J. M. García Fernándeze, E. Aicarta, E. Junqueraa,* aGrupo de Química Coloidal y Supramolecular, Departamento de Química Física I, Facultad de Ciencias Químicas, Universidad Complutense de Madrid, 28040-Madrid, Spain bDepartment of Molecular Medicine, “Sapienza” University of Rome, Viale Regina Elena 291, 00161-Rome, Italy cDepartment of Chemistry, “La Sapienza” University of Rome, Pzle Aldo Moro 5, 00185-Rome, Italy dDepartamento de Química Orgánica, Facultad de Química, Universidad de Sevilla, c/ Profesor García González 1, 41012-Sevilla, Spain eInstituto de Investigaciones Químicas (IIQ), CSIC – Universidad de Sevilla, Avda. Américo Vespucio 49, 41092-Sevilla, Spain Corresponding Author *Fax: +34913944135. E-mail: [email protected] 2 ABSTRACT The self-assembling processes underlining the capabilities of facially differentiated (“Janus”) polycationic amphiphilic cyclodextrins (paCDs) as non-viral gene nanocarriers have been investigated by a pluridisciplinary approach. Three representative Janus paCDs bearing a common tetradecahexanoyl multitail domain at the secondary face and differing in the topology of the cluster of amino groups at the primary side were selected for this study. All of them compact pEGFP-C3 plasmid DNA and promote transfection in HeLa and MCF-7 cells, both in absence and in presence of human serum. The electrochemical and structural characteristics of the paCD-pDNA complexes (CDplexes) have been studied by using zeta potential, DLS, SAXS, and cryo-TEM. paCDs and pDNA, when assembled in CDplexes, render effective charges that are lower than the nominal ones. The CDplexes show a selfassembling pattern corresponding to multilamellar lyotropic liquid crystal phases, characterized by a lamellar stacking of bilayers of the CD-based vectors with anionic pDNA sandwiched among them. When exposed to human serum, either in the absence or in the presence of pDNA, the surface of the cationic CD-based vector becomes coated by a protein corona (PC) whose composition has been analysed by nanoLC- MS/MS. Some of the CDplexes herein studied showed moderate-to-high transfection levels in HeLa and MCF-7 cancer cells combined with moderate-to-high cell viabilities, as determined by FACS and MTT reduction assays. The ensemble of data provides a detail picture of the paCD-pDNA-PC association processes and a rational base to exploit the protein corona for targeted gene delivery on future in vivo applications. KEYWORDS: CDplexes, effective charge ratio, multilamellar phases, cellular transfection, cytotoxicity, protein corona 3 1. Introduction Nucleic acids (DNA or RNA) have become in the last decade an attractive source of therapeutic agents [1]. The interplay of a direct structure-activity relationship and a highly specific mode of action theoretically permit exploiting the cellular machinery in a predictable fashion to either stimulate or silence the expression of virtually any protein, with reduced toxicity and fewer side effects as compared with classical drugs. This is the basis of gene therapy (GT), an exciting branch of medicine that tries to cure diseases (genetic, neurologic, cardiovascular, infectious, or carcinogenic, etc.) at a molecular level, by repairing damaged cellular DNA, either by introducing (by means of a plasmid DNA) and expressing a copy of the affected or missing gene into the cells [2-8] or by inserting a small interfering RNA (siRNA) to silence and stop the expression of an abnormal gene, responsible of the cell disorder [9-12]. Both approaches, either that one based on plasmid DNAs or on siRNAs, constitute nowadays the main strategies of GT on seeking for effectively restore healthy cell functioning. Conventional drugs consist of bioactive species and a carrier, the former being normally the key of the design strategies. However, in the case of biomolecular drugs, such as nucleic acids, the carrier becomes a decisive protagonist. In the free state, nucleic acids are easily degraded by nucleases in biological fluids [13,14], and their membrane-crossing abilities and cellular uptake are seriously limited by their negative charge, inherently large size, and rigidity [15]. Formulation with appropriate delivery systems is thus essential for nucleic acids to overcome the physiological barriers, reach the target in a fully functional form and carry out the designed therapeutic function. Viral gene vectors [16,17] were first developed towards this end due to their high effectiveness. Yet, the use of viral vectors bear inherent risks, including immunogenicity and oncogenicity, which has boosted research on the design, synthesis, and 4 characterization of non-viral carriers that may combine high transfection efficiencies with low cytotoxicity levels [18,19]. In addition, for in vivo applications, these vectors must remain stable in the presence of blood or serum and, ideally, should be amenable to surface decoration with selective groups in order to recognize the target cells. The majority of non-viral vector formulations on record are based on cationic lipids [19,20], polymers [21], nanoparticles [22], and polysaccharides [12,23]. More recently, gene delivery systems based on 3D molecular frameworks with precisely defined chemical structures, sizes, shape symmetry and functional group distributions have been developed [24-26]. Their intrinsic monodisperse character allows conducting systematic studies on the influence of vector modifications on the supramolecular interactions with nucleic acids and the consequences in transfection efficiencies [1]. Within the context of molecular gene vectors, facially differentiated (“Janus”) macrocyclic entities [24,26,27], among which cyclodextrin (CD)-based derivatives are paradigmatic examples, have conquered a prominent position [28,29]. CDs are biocompatible cyclic oligosaccharides consisting of α-1,4-linked glucopyranose units that define a cone-shape topology with well-differentiated faces. The Janus feature can be imparted by bidirectional functionalization with polycationic groups on one rim and hydrophobic chains on the opposite rim. Multihead-multitail polycationic amphiphilic CDs (paCDs) with self-assembling properties and biomimetic cell-membrane-crossing aptitudes, resembling both cationic lipids and cationic polymers, are thus accessed [30,31]. In the presence of nucleic acids, they spontaneously form well-defined supramolecular nanocomplexes (hereinafter referred to as CDplexes) where the gene material is protected from degradation by enzymatic agents. Analogously to the wellknown lipoplexes and polyplexes [32], CDplexes can elicit cellular uptake in physiological media and promote transfection [30,31,33,34]. 5 Whereas in vitro transfection of immortalized cultured cells can be generally conducted in serum-free medium, some ex-vivo and any in vivo GT protocol will imply exposing of the nucleic acid-entrapping nanosystems to other biomolecules present in biological fluids. Non-specific interactions with these biomolecules may then decisively affect their fate in a biological environment [35,36]. Indeed, once administered in vivo, the nanovectors are exposed to the biological fluids from which they adsorb proteins and other components, losing the bare vector identity to form a biologically active corona, known as the protein corona (PC) [37,38]. This PC plays an important role in both the cellular recognition [39] and cell-membrane crossing capabilities (internalization process) of the nanovector [40]. It has also a clear effect on its time of circulation in the blood, so that knowing the proteomic profile of the corona surrounding a gene vector is crucial to optimize the transfection process. For instance, certain proteins like opsonins and, specially, immunoglobulins, fibrinogen and complement proteins are recognized by the macrophages favoring the phagocytosis and the rapid clearance of the vector from the bloodstream. On the other hand, the adsorption of apolipoproteins and serum albumin (dyopsonins) promote prolonged blood circulation half-life [41]. Successful delivery of therapeutic genes into cells will therefore require not only a high control of the physical-chemical parameters of the vector and vector-DNA nanocomplex, such as effective charges, chemical composition, size and structure, but also a deep knowledge on their interaction with serum proteins, the composition of the resulting PC and its impact in transfection capabilities and toxicity. Notwithstanding, up to date, most studies in this field are centered on the PC that forms around inorganic and organic nanoparticles [40,42-44], with only a few investigations evaluating the PC adsorbed on vector-DNA complexes [45,46]. 6 This work is specifically focused on the biophysical and biochemical characterization of Janus paCD-pDNA CDplexes and their protein corona fingerprints as a function of the molecular paCD vector topology. We have chosen three different paCD representatives with a -cyclodextrin (CD) core, namely compounds ADM70, ADM105 and PBO234 (Scheme 1). The three molecular vectors share a C7- symmetrical skirt-type Janus architecture with a multitail lipophilic domain formed by fourteen hexanoyl groups that esterify the secondary hydroxyls, differing in the architecture of the multivalent cationic cluster installed at the primary positions. Thus, ADM70 displays a dendritic presentation of fourteen primary amino groups, ADM105 a linear arrangement of seven secondary and seven primary amino groups and PBO234 a heptavalent cluster of primary amino groups. ADM70 and ADM105 additionally incorporate a belt of seven thiourea groups at the primary face branches. All these structural elements have been previously found to impart good DNA compaction abilities to molecular gene vectors as well as good transfection capabilities to the corresponding vector-DNA complexes in cellulo and in vivo [33,34]. We have first conducted a deep electrochemical and structural characterization of the bare nanovector and pDNA-loaded Janus paCD assemblies by using high precision experimental techniques including zeta potential, agarose gel electrophoresis, dynamic light scattering (DLS), small angle X-ray scattering (SAXS) and cryo-transmission electron microscopy (cryo-TEM). Given that what cells “see” and process in in vivo situation is not the bare complexes but the CDplexes-PC bioentity formed, we have evaluated the PC developed at the surface of CDplexes formulated with the three paCDs herein used when they interact with human plasma (HP), by means of nano liquid chromatography tandem mass spectrometry (nanoLC-MS/MS). Finally, transfection performances of the CDplexes and levels of cellular toxicity were further investigated in vitro by 7 fluorescence-activated cell sorting (FACS) and reduction of 3-(4,5-dymethylthiazol-2- y1)-2,5-diphenyl tetrazolium bromide (MTT assay), respectively. The final goal of this pluridisciplinary approach is obtaining a global picture of the optimal requirements for the preparation of the CDplexes, in an attempt of opening alternative trails that provide non-viral vectors with improved outputs of bioavailability and transfection efficiency (TE). Scheme 1. Structural characteristics of ADM70, ADM105 and PBO234. 2. Experimental section 2.1 Materials The Janus paCDs used in this study, ADM70, ADM105 and PBO234 (Scheme 1), were prepared following the procedures previously reported [47]. Briefly, the synthesis of compound PBO234 was accomplished in four steps (i-iv) from commercially available -cyclodextrin by (i) heptabromination at the primary hydroxyl rim with the N-bromosuccinimide (NBS)/triphenylphosphine (TPP) system [48], (ii) cesium carbonate promoted nucleophilic displacement of bromine by N-Boc-cysteamine [49], = == = = + ADM70 ADM105 PBO234 ADM70 ADM105 PBO234 + CD paCD 8 (iii) acylation of the fourteen secondary hydroxyls by reaction with hexanoic anhydride in N,N-dimethylformamide (DMF) and N,N-dimethylaminopyridine (DMAP) as a nonnucleophilic base catalyst and (iv) final acid-promoted hydrolysis of the tert-butyl carbamate groups. Compounds ADM105 and ADM70 were obtained from PBO234 after isothiocyanation of the seven amino groups with thiophosgene [47] followed by thiourea coupling reactions with mono-N-Boc-ethylenediamine or N,N”-di-Boc- ethylenetriamine, respectively, and subsequent Boc removal. The physicochemical data for the three polycationic amphiphilic cyclodextrins were consistent with those previously reported [47]. Most significantly, the 1H and 13C NMR spectra showed the typical single-spin system for fully C7-symmetrical molecules. pEGFP-C3 Plasmid DNA (pDNA) was extracted from competent E. Coli bacteria previously transformed with pEGFP-C3, the extraction being carried out using GenElute HP Select plasmid Gigaprep Kit (Sigma Aldrich) following a protocol previously described [50,51]. Sodium salt of calf thymus DNA (ctDNA), provided by Sigma-Aldrich, was used as linear DNA to determine the effective charge ( ) of the cationic vector. Human plasma (HP) whole blood provided by the Department of Experimental Medicine of La Sapienza University of Rome, was obtained by venipuncture of ten healthy volunteers. Mixed plasma was aliquoted and stored at -80 ºC in protein LoBind tubes until further use. For analysis, the aliquots were thawed at 4 ºC and then allowed to warm at room temperature. Human cervical cancer cell line (HeLa), derived from human cervix adenocarcinoma, and human breast cancer cell line (MCF-7), derived from human pleural effusion breast cancer metastasis, were purchase from ATCC (Manassas, VA, USA). HeLa cells were maintained in Eagle’s Minimum Essential Medium (EMEM) supplemented with 2 mM L-glutamine, 100 IU/mL + CD q 9 penicillin/streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, 1.5 mg/L sodium bicarbonate, and 10% fetal bovine serum (FBS). MCF-7 cells were maintained in Eagle’s Minimum Essential Medium (EMEM) supplemented with 0.01 mg/ml human recombinant insulin and 10% fetal bovine serum (FBS). 2.2 Methods 2.2.1 Preparation of CDplexes. paCD-DNA complexes were formed by mixing the correct amounts of aqueous solutions of the paCDs and of DNAs (either ctDNA or pDNA) in HEPES 20 mM (pH = 7.4). The final solutions were left during 20 min prior to carry on the experiments. pDNA concentrations were optimized to fit the optimum conditions for each experimental technique. CDplex composition can be expressed either in terms of the masses ratio   CD DNA m / m  , between mass of the gene vector ( ) to plasmid DNA   DNA m , or the effective charge ratio (ρeff), between the paCD and pDNA effective charges. 2.2.2. Incubation of paCDs with HP, in absence and presence of pDNA. paCDs were mixed with HP (1:1 v/v) and were incubated at 37 ºC for 1 h. After incubation, the samples were centrifuged three times during 15 min at 14000 rpm in order to wash the sample and remove all the molecules not bound to the complex. The same procedure was followed with the CDplexes. 2.2.3. Methods for the characterization of paCD/pDNA CDplexes. Uncomplexed plasmid DNA along with the paCD/pDNA CDplexes (at several   CD pDNA m / m ratios) were loaded onto 1% agarose gel and run for 30 min at 100 mV in 1x TAE (Tris-HCl, Acetate and EDTA) buffer. In the Agarose gel electrophoresis experiments, fully paCD/pDNA complexes appeared as fluorescent band in the wells of the gel, while CD m 16 charged complex, as required to cross the negative cell membrane in an efficient cell transfection process. From Eq. (2), it is easily deduced that = 1 requires: (3) Thus, if the isoneutrality ratio   CD linearDNA m / m  is experimentally determined for a complex formed by a Janus CD-based vector and a commercial linear DNA (for which = -2/bp), the effective positive charge of the vector ( ) can be obtained. Once this charge is known, the negative effective charge of the plasmid ( ) can be straightforwardly determined from the experimental value of   CD pDNA m / m  for a complex formed by the same CD based vector but with a plasmid DNA instead of the linear DNA, on the same experimental conditions, as follows: (4) Electrochemical techniques are, among others, the most adequate tools to determine this electroneutrality ratio, agarose gel electrophoresis, and, more precisely, zeta potential, being the most recommended. Agarose gel electrophoresis informs about the compaction level of pDNA by the vectors. Fig.1 reports these experiments at three different   CD DNA m / m ratios (white numbers on lanes 2-4), for the three paCDs used in this work (ADM70, ADM105 and PBO234). Uncomplexed pDNA (lane 1) was used as a positive control. Results reported in Fig. 1 reveal that pDNA is efficiently compacted by the paCD molecular vectors, since the fluorescent bands disappear across the gel lanes as long as   CD DNA m / m ratio increases. As can be inferred from Fig. 1, the eff  1 , ,linear CD CD eff CD eff DNA bp linearDNA mM qq mM       ,lineareff DNA q ,eff CD q ,peff DNA q ,p , bp CD eff DNA eff CD pDNA CD mM qq mM       17 isoneutrality was reach at   CD DNA m / m ratios below 0.4, 0.7 and 0.5 for ADM70- pDNA, ADM105-pDNA and PBO234-pDNA CDplexes, respectively. With the aim of determining with higher precision the electroneutrality ratio, zeta potential was measured as a function of   CD DNA m / m (Fig. 2), covering the range within which pDNA is effectively compacted by each paCD, according to the results shown in Fig. 1. The electroneutrality ratio   CD DNA m / m  can be determined as the   CD DNA m / m value where a sign inversion on the charge occurs in the  -potential sigmoidal profiles. Fig. 1. Gel electrophoresis results for (a) ADM70-pDNA CDplexes, (b) ADM105-pDNA CDplexes, and (c) PBO234 -pDNA CDplexes. Lane 1: uncomplexed pDNA (positive control). Lanes 2-4: CDplexes at different mCD/mDNA mass ratios (white numbers at the bottom of the lanes). pDNA mCD/mDNA 0.4 0.7 1.6 0.7 1.6 3.0 0.5 1.0 2.0 b) c) a) 18 Fig. 2. Plot of ζ potential vs mCD/mDNA CDplex composition, for different samples with ctDNA (a) and pDNA (b). Black, red and blue symbols correspond to ADM70-pDNA, ADM105-pDNA and PBO234-pDNA CDplexes, respectively. Table 1 reports the results obtained for CDplexes formulated with ADM70, ADM105 or PBO234 and either ctDNA or pDNA. With these   CD DNA m / m  values and following the procedure above explained, the effective charges of both the Janus CD based cationic vectors and the pDNA herein used, and , were calculated and collected in Table 2. The differences encountered between nominal and effective charges in both the cationic vectors and the anionic pDNA are remarkable. Thus, the net positive charges available for interaction with pDNA are around 33% of the nominal one, assuming full protonation of the amino groups, for the tetradecaamine derivatives ADM70 and ADM105, and only 25% for heptamine PBO234. This scenario is similar to that found for some CD polyrotaxane-based vectors [12] but sharply different from that encountered for cationic lipid gene vectors bearing quaternary amino groups, which normally yield their total nominal positive charges within a range of 10 % of uncertainty [19,51,57]. On the other hand, the plasmid used herein seems to be quite supercoiled at the experimental conditions used, since its available negative charge per bp is far away from the nominal value (-2/bp), yielding ,eff CD q ,p  eff DNA q 0,0 0,2 0,4 0,6 0,8 -20 0 20  (mV) mCD/mDNA b) 0 3 6 9 -30 0 30  (mV) mCD/mDNA a) 19 around 7%, 13% and 5% of its nominal charge when being compacted by ADM70, ADM105 and PBO234, respectively. This behaviour, often found in lipoplexes, confirms that plasmids usually retain an important percentage of cationic sodium counter-ions (Na+). This is a priori a favourable feature for the use of these macrocyclic vectors as safer and effective vehicles of nucleic acids, since the weaker the anionic character of the DNA, the lower the amount of cationic vector needed to formulate the nanocomplexes, thus decreasing the risk of cytotoxicity of the vector. From the paCD and pDNA effective charges thus obtained, effective charge ratios (  eff) around 5-fold (for ADM70 and PBO234) or 2- to 3-fold (for ADM105) of the nominal ones were calculated using Eq. (2). Table 1. Values of isoneutrality ratios   CD DNA m / m  for the CDplexes formed by ADM70, ADM105 and PBO234 with either ctDNA or pDNA. ADM70-DNA ADM105-DNA PBO234-DNA ctDNA 3.0 2.9 5.7 pDNA 0.19 0.38 0.25 Values estimated with a 5% error Table 2. Nominal and effective charges of paCDs and pDNA. ADM70 ADM105 PBO234 nom,CD q 14 14 7 eff,CD q 4.6 4.7 1.7 eff,CD nom,CD q / q  0.33 0.33 0.25 nom,pDNA q / bp  -2 -2 -2 eff,pDNA q / bp  -0.13 -0.26 -0.09 eff,pDNA nom,pDNA q / q  0.07 0.13 0.05  eff/  nom 5.1 2.6 5.4 Values estimated with an 8% error Having information of the structure of the complex is important to find the best nonviral vectors and to promote its use in in vivo gene therapy. The structure of the paCD- pDNA CDplexes in concentrated samples was investigated by SAXS at several 20 effective charge ratios for which CDplexes are potentially active as gene transfection vectors (  eff > 1). Cryo-TEM was also used as a supporting technique. Fig. 3 shows the corresponding SAXS diffractograms (Intensity vs q factor) at  eff = 2, 4 and 8, with the Miller indexes being included in the plot. In all the cases, three peaks that index well to a lamellar lyotropic liquid crystal phase (L  ) were observed, regardless of  eff, with the characteristic inter-layer distance (d) directly related to the q factor (d = 2πn/qhkl, n is the diffraction order). This structure may be explained by considering that the hexanoyl chains linked to the wider entrance of the CD torus promote self-assembling in a lipidic bilayer fashion, CDplexes being adequately represented as alternating bilayers of paCDs molecules and an aqueous monolayer containing supercoiled pDNA, with thicknesses represented by dm, and dw, respectively, being d = dm + dw. Table 3 collects the values of d, calculated as an average of the data obtained from the more intense peaks (100 and 200) of the diffractograms at each  eff ratio. It is noticeable that the periodicity of the structure remains basically unaltered for a constant pDNA content when the proportion of the vector increases (i.e. d remains basically constant with  eff), for the three CDplexes studied. Considering these d values and the fact that pDNA supercoils needs around dw ~2.0-2.5 nm to be sandwiched by paCDs bilayers in a typical sandwich fashion, it can be deduced that the thickness of the bilayer (dm) must be 3.1, 3.4 and 2.5 nm for ADM70-pDNA, ADM105-pDNA and PBO234-pDNA CDplexes, respectively. Given the lengths of both the cationic tails and the lipid type chains, and the depth of the  -CD torus-shaped cavity, these d, dm and dw values are compatible with a L  structure only if: i) the cationic tails on the narrower entrance of the CD macrocycle are somehow in an open bouquet fashion (to avoid electrostatic repulsions among positive charges) and/or ii) the lipidic type chains that are linked to the wider entrance of -CD macrocycle are intercrossed. 21 Fig. 3. SAXS diffractograms of paCD-pDNA CDplexes at several effective charge ratios (ρeff): Red lines ρeff = 2, green lines ρeff = 4, and blue lines ρeff = 8 for: a) ADM70-pDNA; b) ADM105-pDNA; and c) PBO234-pDNA CDplexes. 1 2 3 4 5 (300) (200) (100) q (nm-1) (300) (200) (100) c) (300) (200) (100) (300) (200) (100) a) (300) (200) (100) (300) (200) (100) (300) (200) (100) (100) (100) (300) (200) Intensity (a.u.) (300) (200) b) 22 Table 3. Values of q and d of the lamellar (L  ) liquid crystal phase found for the paCD- pDNA CDplexes (where paCDs are ADM70, ADM105 and PBO234), at several effective charge ratios (  eff). ADM70 ADM105 PBO234  eff L  L  L  q100 1.12 1.06 1.28 2 q200 2.21 2.09 2.50 q300 3.30 3.12 - d 5.6 5.9 5.0 q100 1.13 1.09 1.31 4 q200 2.23 2.13 2.47 q300 3.33 3.16 - d 5.6 5.8 5.0 q100 1.15 1.07 1.30 8 q200 2.27 2.12 2.50 q300 3.39 3.18 - d 5.5 5.9 4.9 Values of q and d are reported in nm-1 and nm, respectively. Cryo-TEM experiments further confirmed the multilamellar character of the CDplexes studied in this work. Fig. 4 shows a selection of micrographs among those taken for ADM105-pDNA CDplex that are representative of the ensemble of data (see Fig. S1 in Supplementary data for a selection of micrographs of ADM70-pDNA and PBO234-pDNA CDplexes). The multilamellar arrangement found in SAXS is also seen in these micrographs. In fact, three types of nanostructures in coexistence are found: (i) CDplexes with a well-defined multilamellar profile and aggregated in a cluster type fashion (CT-type nanoaggregates; labelled with white asterisks in Fig. 4); (ii) CDplexes with a clear finger-print multilamellar pattern (FP-type nanoaggregates; labelled with white circles, and a zoom view in panel d), and; (iii) self-aggregated paCDs without pDNA compacted (labelled with white arrows), with a typical vesicle-type structure. The presence of aggregates of paCD molecules, without pDNA being compacted, is justified by the fact that the experiment is done for samples with a clear excess of cationic vector (notice that  eff = 4 in cryo-TEM experiments). These exceeding paCD 23 molecules tend to form unilamellar spherical vesicle-type nanostructures, the formation of the lipidic bilayer being promoted by the presence of 14 lipidic type chains linked to the wider entrance of each cyclodextrin torus (2 chains per glucose unit). It can be inferred that the unilamellar paCD vesicles above mentioned interact with pDNA through strong electrostatic interactions and tend to aggregate yielding the multilamellar phases, either CT or FP, by sandwiching the plasmid supercoils within the aqueous monolayer that remain between each two bilayers. Scheme 2 shows schematic drawings of both nanoaggregates: (a) unilamellar vesicles in the absence of pDNA, and (b) multilamellar complexes in the presence of pDNA. In the CT-type structures, the bilayers may be deformed when sandwiching pDNA supercoils with the adjacent bilayers, but they essentially keep their morphologies. However, FP-type nanoaggregates show the typical finger-print compaction pattern, and in contrast with the CT-type CDplexes, the bilayers tend to disrupt probably due to a more favourable paCD-pDNA interaction. This particular scenario has been also previously found for different lipoplexes (cationic lipids-pDNA complexes) [20,57]. Some of the multilamellar arrangements shown in these micrographs (and other not shown) have been chosen to analyze the presence of periodicity. As representative examples, the insets of panels b) and c) in Fig. 4, show Fast Fourier Transform (FFT) profiles where the diffraction spot corresponds to a typical lamellar pattern. 24 Fig. 4. A selection of cryo-TEM micrographs showing a general view of the ADM105- pDNA CDplexes at  eff = 4. Insets on panels b) and c) show the diffraction spots from FFT calculations over a selected area on the original micrograph (yellow square). FFT pattern reveals a multilamellar structure. Scale bars are 100 nm in panels a-c and 50 nm in panel d. Scheme 2. Schematic drawings of: a) Vesicles-type paCD self-aggregation pattern; and b) Multilamellar lyotropic liquid crystal phase (L  ) of the CDplexes studied in this work, showing the structural parameters, d, dm and dw. a) b) 25 Accordingly, the ensemble of electrochemical and structural results shows L selfassembling patterns irrespectively of  eff values, while the best compaction levels are obtained at  eff = 4. Additionally, it is known that CDplexes must be positively charged (  eff > 1), but with cationic vector content being as low as possible to diminish cytotoxicity. All these considerations point to  eff = 4 as a potentially adequate charge ratio to carry on both the proteomic and TE studies. Upon in contact with biological milieu, CDplexes will adsorb plasma proteins in a time-dependent manner. Most abundant proteins will bind to vector surface first and will be progressively substituted by proteins with high affinity for the surface of CDplexes. At the equilibrium (typically reached within 1 h exposure), the PC of CDplexes will be constituted by a longstanding protein layer referred to as the hard corona (HC), which will provide CDplexes with their biological identity, plus a soft corona (SC) made of proteins in dynamical exchange with the surrounding environment. In the present investigation, both paCDs and CDplexes were let to interact with HP for 1 h. The PC formation was confirmed by means of  -potential and size measurements of the CDplexes in the absence and in the presence of HP. The data collected in Table 4 indicate that the bare complexes are positively charged (  -potential around +30 mV), with mean hydrodynamic diameters of around 114 nm in the case of ADM70-pDNA CDplexes and around 150 nm for the other two CDplexes. After 1 h incubation with HP, a clear increase on the size of the complexes (hydrodynamic diameter increases around 40-45 nm in the three cases), and a marked drop in  - potential, shifting from positive to negative (from ca. +30 mV to ca. −17 mV), were observed. These evidences are clearly compatible with the formation of a ca. 20-nm- thick PC, mostly consisting of negatively charged proteins (i.e. pI < 7), that are located 32 to protein expression, including endosome scape, trafficking to the nucleus and pDNA release, are particularly favorable for PC-coated ADM70-pDNA CDplexes. Fig.7. Transfection efficiency (TE) values of paCDs-pDNA CDplexes, in terms of % GFP expressed (panels a-b) and mean fluorescence intensity (MFI) (panels c-d), at  eff = 4, 25 and 50, in the absence and presence of Human Plasma (HP), through HeLa and MCF-7 cells. CTR: cells alone. LFN: Lipofectamine, positive control. On the other hand, as shown in Fig. 8, the increase on effective charge ratio (  eff) from 4 to 50 seems to slightly decrease the viability of HeLa cells, in the absence of HP, while no appreciable decrease is observed in MCF-7 cells. In the presence of PC, cell viability of both cell lines was not appreciably diminished. Notice, nonetheless, that almost all the values are around or over 80%, which is considered a minimum threshold value. However, interpretation of the effect of PC on the cell viability of CDplexes must be done with caution. Whether PC reduce cell viability or protect the cells can be due to specific PC composition and, in turn, to the complex relationship between PC composition and nanocarrier processing by cell machinery [59]. Moreover, in the 33 absence of in vivo experiments, the evaluation of PC composition, although implies an important improvement with respect to the use of SDS PAGE experiments alone, does not allow one to depict strong conclusions about future in vivo applications (i.e. targeting ability, bio-distribution, cytotoxicity). In fact, in the literature, major works are discussing about mapping protein binding sites across the PC surface. In the absence of this piece of information, discussion about the potential role of corona proteins should be made, but kept at a minimum, as done in this work. Otherwise, the risk of speculation may be too high. Fig. 8. Cell viability of paCDS-pDNA CDplexes at  eff = 4, 25 and 50 for HeLa and MCF-7 cells, in the absence and presence of human plasma (HP). CTR: cells alone. CDS: cells in the presence of paCDS without pDNA. The above presented experimental evidences support that the Janus paCD structures ADM70 and ADM105 are promising nonviral vectors for pDNA delivery, overall 34 superior to PBO234 and to the Lipofectamine control. Differences in transfection efficiency as a function of the cell line might arise from differences in the preferred internalization routes. Thus, it has been reported that CDplexes can enter the cell by caveolae- and clathrin-mediated endocytosis, but only the first mechanism is productive regarding protein expression [30]. Moreover, CDplexes obtained from ADM70 seem to exhibit the most favourable features for in vitro gene therapy applications: high stability in the absence or presence of HP, homogeneous size distribution, lower impact (with respect to the other two paCDs) of the PC in the capacity of the CDplexes to efficiently mediate protein expression in cellulo, and very low cytotoxicity even at high  eff values, and independently of the presence or not of HP. These features are in agreement with the previously observed superiority of ADM70 formulations in in vivo transfection studies [33,34]. The body of biophysical and biochemical evidences obtained in this work provides a pathway for the rigorous characterization of pDNA-molecular vector formulations in view of their optimization for gene therapy applications. Broadening the current database on vector structures, electrochemical and structural properties of their aggregates, protein corona composition and transfection efficiencies in different in vitro and in vivo experimental settings should fund the basis for the rational design of second generation candidates. Work in that direction is currently sought in our laboratories. 4. Conclusions This work was aimed to provide new insights in the fascinating processes governing pDNA complexation by monodisperse vectors of the Janus paCD family, in an attempt to delineate the mechanisms whereby differences in molecular structure translates into differences in transfection capabilities. For that purpose, three Janus CD-based 35 compounds (ADM70, ADM105 and PBO234) have been thoroughly checked as potential nanocarriers of a pEGFP-C3 plasmid that codes for GFP expression to the interior of HeLa and MCF-7 cancer cells, both in the absence and presence of human serum. In a first level of organization, the vector molecular structure influences the effective positive charge available for interacting with the plasmid as well as the effective negative charge of the plasmid in the nanoaggregates. In fact, zeta potential study revealed that effective charges are lower than nominal ones both for the CD- based vectors (around 30% for ADMs and 25% for PBO234), and also for the pDNA, which renders a low percentage of its negative charge (less than 15%) when it is compacted by the nanocarriers. This is a potentially favourable finding since the weaker the anionic character of the DNA, the lower the amount of cationic vector needed to formulate the nanocomplexes, thus decreasing the cytotoxicity of the vector and increasing their potential output as safe and effective vehicles of nucleic acids. On the other hand, SAXS and cryo-TEM studies have shown that ADM70-pDNA, ADM105- pDNA and PBO234-pDNA CDplexes are structured according to a multilamellar lyotropic liquid crystal phase (L  . Two different multilamellar phases are distinguished, nonetheless, on cryo-TEM micrographs (CT-type and FP-type nanoaggregates), in coexistence with unilamellar vesicles of self-aggregated exceeding paCDs without pDNA compacted. Although the Janus feature seems to warrant a multilamellar arrangement for the paCD-pDNA nanocomplexes in all cases, their stability and homogeneity can be significantly different depending on the topology of the multivalent cationizable domain in the paCD entity. In the presence of human plasma, a second level of organization takes place involving the interaction with serum proteins, giving rise to multicomponent nanoassemblies equipped with a protein corona whose composition varies from a paCD formulation to another. The protein corona 36 (PC) characterization has also driven to interesting conclusions. The adsorbed proteins have mostly a negative charge in all the cases (pI < 7), pointing to the electrostatic interactions as the driven non-covalent forces between proteins and the cationic vectors. Complement, lipoprotein and coagulation were found to be the most abundant types of proteins within the corona of the Janus paCDs and the corresponding CDplexes, whereas immunoglobulins, tissue leakage and acute phase proteins constituted a minor fraction of the PC. At moderate-to-high charge ratios (  eff = 25 and 50), the bare CDplexes herein reported seem to transfect either HeLa or MCF-7 cancer cells more efficiently than Lipofectamine, and with high cell viabilities ranged from 80 to 100%. These transfection efficiencies were found to slightly decrease in the presence of biological media (HP), but in any case the values found were higher than those obtained with the control Lipofectamine. However, the effect of PC on the cell viability of CDplexes is not that easy to interpret. Probably, the specific PC composition and, in turn, its complex effect on nanocarrier processing by cell machinery, play a crucial role, although an accurate mapping of protein binding sites in the PC would be necessary to go further in these conclusions. In any case, the whole body of both biophysical and biochemical evidences obtained in this work allow us to conclude that the three paCDs proposed can be considered as potentially efficient nanocarriers in vitro and promising gen vectors for in vivo applications. Going further, among the three Janus paCDS nanovectors checked in this work, ADM70 seem to exhibit the most favourable features for in vitro gene therapy applications requiring serum-containing media, in total agreement with the already observed superiority of ADM70 formulations in in vivo transfection studies. 37 Acknowledgments MINECO of Spain, (contract numbers CTQ2012-30821, SAF2013-44021-R and CTQ2015-64425-C2-1-R), the Junta de Andalucía (contract number FQM2012-1467), University Complutense of Madrid (Spain) (project no. UCMA05-33-010) and the European Regional Development Funds (FEDER and FSE) for financial support. SAXS experiments were performed at NCD11 beamline at ALBA Synchrotron Light Facility with the collaboration of ALBA staff. Authors also thank C. Aicart-Ramos for carrying on amplification of plasmid DNA at the Departamento de Bioquímica y Biología Molecular I (UCM, Spain) and also P. Castro-Hartmann, Servei de Microscopia of UAB (Spain), for cryo-TEM experiments. Appendix A: Supplementary data Supplementary Information available: Additional cryo-TEM micrographs, details and tables of the top 25 most abundant proteins in the protein corona of the CDplexes, and additional TE and cell viability experiments at higher CDplexes concentrations per cell. This information can be found on line at http://dx.doi.org/ References [1] E. Junquera, E. Aicart, Recent progress in gene therapy to deliver nucleic acids with multivalent cationic vectors, Adv. Colloid Interface Sci., 233 (2016) 161-175. [2] I.M. Verma, M.D. Weitzman, Gene therapy: Twenty-first century medicine, Annu. Rev. Biochem., 74 (2005) 711-738. [3] E. Junquera, E. Aicart, Cationic lipids as transfecting agents of DNA in gene therapy, Curr. Topics Med. Chem., 14 (2014) 649-663. [4] T. Montier, T. Benvegnu, P.A. Jaffres, J.J. Yaouanc, P. Lehn, Progress in cationic lipid-mediated gene transfection: A series of bio-inspired lipids as an example, Curr. Gene Ther., 8 (2008) 296-312. [5] R.S. Dias, B. Lindman, DNA Interaction with Polymers and Surfactants, Wiley & Sons, Hoboken, NJ, 2008. [6] K. Ewert, N.L. Slack, A. Ahmad, H.M. Evans, A.J. Lin, C.E. Samuel, C.R. Safinya, Cationic lipid-DNA complexes for gene therapy: Understanding the relationship between complex structure and gene delivery pathways at the molecular level, Curr. Med. Chem., 11 (2004) 133-149. 38 [7] P.P. Karmali, A. Chaudhuri, Cationic liposomes as non-viral carriers of gene medicines: Resolved issues, open questions, and future promises, Med. Res. Rev., 27 (2007) 696-722. [8] A.J. Kirby, P. Camilleri, J. Engberts, M.C. Feiters, R.J.M. Nolte, O. Soderman, M. Bergsma, P.C. Bell, M.L. Fielden, C.L.G. Rodriguez, P. Guedat, A. Kremer, C. McGregor, C. Perrin, G. Ronsin, M.C.P. van Eijk, Gemini surfactants: New synthetic vectors for gene transfection, Angew. Chem., Int. Ed., 42 (2003) 1448- 1457. [9] P. Kesharwani, V. Gajbhiye, N.K. Jain, A review of nanocarriers for the delivery of small interfering RNA, Biomaterials, 33 (2012) 7138-7150. [10] B. Khurana, A.K. Goyal, A. Budhiraja, D. Arora, S.P. Vyas, siRNA delivery using nanocarriers - An efficient tool for gene silencing, Curr. Gene Ther., 10 (2010) 139- 155. [11] M.A. Islam, T.E. Park, B. Singh, S. Maharjan, J. Firdous, M.H. Cho, S.K. Kang, C.H. Yun, Y.J. Choi, C.S. Cho, Major degradable polycations as carriers for DNA and siRNA, J. Controlled Release, 193 (2014) 74-89. [12] V.D. Badwaik, E. Aicart, Y.A. Mondjinou, M.A. Johnson, V.D. Bowman, D.H. Thompson, Structure-property relationship for in vitro siRNA delivery performance of cationic 2-hydroxypropyl-β-cyclodextrin: PEG-PPG-PEG polyrotaxane vectors, Biomaterials, 84 (2016) 86-98. [13] W. Kusser, Chemically modified nucleic acid aptamers for in vitro selections: evolving evolution, Rev. Molec. Biotechnol., 74 (2000) 27-38. [14] B.E. Eaton, The joys of in vitro selection: chemically dressing oligonucleotides to satiate protein targets, Curr. Opin. Chem. Biol., 1 (1997) 10-16. [15] I.S. Blagbrough, A.A. Metwally, siRNA and Gene Formulation for Efficient Gene Therapy, InTech, 2013. [16] C.E. Thomas, A. Ehrhardt, M.A. Kay, Progress and problems with the use of viral vectors for gene therapy, Nature Rev. Gen., 4 (2003) 346-358. [17] P.D. Robbins, S.C. Ghivizzani, Viral vectors for gene therapy, Pharmacol Therapeut., 80 (1998) 35-47. [18] M.A. Mintzer, E.E. Simanek, Nonviral vectors for gene delivery, Chem. Rev., 109 (2008) 259-302. [19] K. Kumar, A.L. Barran-Berdon, S. Datta, M. Munoz-Ubeda, C. Aicart-Ramos, P. Kondaiah, E. Junquera, S. Bhattacharya, E. Aicart, A delocalizable cationic headgroup together with an oligo-oxyethylene spacer in gemini cationic lipids improves their biological activity as vectors of plasmid DNA, J. Mater. Chem. B, 3 (2015) 1495-1506. [20] A.L. Barran-Berdon, S.K. Misra, S. Datta, M. Muñoz-Ubeda, P. Kondaiah, E. Junquera, S. Bhattacharya, E. Aicart, Cationic gemini lipids containing polyoxyethylene spacers as improved transfecting agents of plasmid DNA in cancer cells, J. Mater. Chem. B, 2 (2014) 4640-4652. [21] D. Putnam, Polymers for gene delivery across length scales, Nature Mater., 5 (2006) 439-451. [22] D.J. Bharali, I. Klejbor, E.K. Stachowiak, P. Dutta, I. Roy, N. Kaur, E.J. Bergey, P.N. Prasad, M.K. Stachowiak, Organically modified silica nanoparticles: A nonviral vector for in vivo gene delivery and expression in the brain, PNAS, 102 (2005) 11539-11544. [23] C. Ortiz Mellet, J.M. Garcia Fernandez, J.M. Benito, Glycotransporters for gene delivery, in: A.P. Rauter, T.K. Lindhorst (Eds.) Carbohydrate Chemistry: Chemical and Biological Approaches vol. 38, 2012, pp. 338-375. 39 [24] C. Ortiz Mellet, J.M. Benito, J.M. García Fernandez, Preorganized, macromolecular, gene-delivery systems, Chem. Eur. J, 16 (2010) 6728-6742. [25] Y. Aoyama, Macrocyclic glycoclusters: From amphiphiles through nanoparticles to glycoviruses, Chem. Eur. J, 10 (2004) 588-593. [26] V. Bagnacani, V. Franceschi, M. Bassi, M. Lomazzi, G. Donofrio, F. Sansone, A. Casnati, R. Ungaro, Arginine clustering on calix-4-arene macrocycles for improved cell penetration and DNA delivery, Nature Commun., 4 (2013). [27] J.L. Jiménez Blanco, F. Ortega-Caballero, L. Blanco-Fernández, T. Carmona, G. Marcelo, M. Martínez-Negro, E. Aicart, E. Junquera, F. Mendicuti, C. Tros de Ilarduya, C. Ortiz Mellet, J.M. García Fernández, Trehalose-based Janus cyclooligosaccharides: the ‘‘click’’ synthesis and DNA-directed assembly into pH- sensitive transfectious nanoparticles., Chem. Commun., 52 (2016) 10117-10120. [28] S. Loethen, J.M. Kim, D.H. Thompson, Biomedical applications of cyclodextrin based polyrotaxanes, Polymer Reviews, 47 (2007) 383-418. [29] C. Ortiz Mellet, J.M. García Fernandez, J.M. Benito, Cyclodextrin-based gene delivery systems, Chem. Soc. Rev., 40 (2011) 1586-1608. [30] A. Diaz-Moscoso, D. Vercauteren, J. Rejman, J.M. Benito, C. Ortiz Mellet, S.C. De Smedt, J.M. García Fernandez, Insights in cellular uptake mechanisms of pDNA-polycationic amphiphilic cyclodextrin nanoparticles (CDplexes), J. Controlled Release, 143 (2010) 318-325. [31] A. McMahon, E. Gomez, R. Donohue, D. Forde, R. Darcy, C.M. O'Driscoll, Cyclodextrin gene vectors: cell trafficking and the influence of lipophilic chain length, J. Drug Deliv. Sci. Technol., 18 (2008) 303-307. [32] G. Caracciolo, H. Amenitsch, Cationic liposome/DNA complexes: from structure to interactions with cellular membranes, Eur. Biophys. J., 41 (2012) 815-829. [33] C. Aranda, K. Urbiola, A.M. Ardoy, J.M. García Fernandez, C. Ortiz Mellet, C. Tros de Ilarduya, Targeted gene delivery by new folate-polycationic amphiphilic cyclodextrin-DNA nanocomplexes in vitro and in vivo, Eur. J. Pharm. Biopharm., 85 (2013) 390-397. [34] A. Mendez-Ardoy, K. Urbiola, C. Aranda, C. Ortiz Mellet, J.M. Garcia Fernandez, C. Tros de Ilarduya, Polycationic amphiphilic cyclodextrin-based nanoparticles for therapeutic gene delivery, Nanomedicine, 6 (2011) 1697-1707. [35] S.D. Wettig, R.E. Verrall, M. Foldvari, Gemini surfactants: A new family of building blocks for non-viral gene delivery systems, Curr. Gene Ther., 8 (2008) 9- 23. [36] P.C. Bell, M. Bergsma, I.P. Dolbnya, W. Bras, M.C.A. Stuart, A.E. Rowan, M.C. Feiters, J. Engberts, Transfection mediated by gemini surfactants: Engineered escape from the endosomal compartment, J. Am. Chem. Soc., 125 (2003) 1551- 1558. [37] A.L. Barran-Berdon, D. Pozzi, G. Caracciolo, A.L. Capriotti, G. Caruso, C. Cavaliere, A. Riccioli, S. Palchetti, A. Lagana, Time evolution of nanoparticleprotein corona in human plasma: Relevance for targeted drug delivery, Langmuir, 29 (2013) 6485-6494. [38] D. Walczyk, F.B. Bombelli, M.P. Monopoli, I. Lynch, K.A. Dawson, What the cell “sees” in bionanoscience, J. Am. Chem. Soc., 132 (2010) 5761-5768. [39] G. Caracciolo, F. Cardarelli, D. Pozzi, F. Salomone, G. Maccari, G. Bardi, A.L. Capriotti, C. Cavaliere, M. Papi, A. Lagana, Selective targeting capability acquired with a protein corona adsorbed on the surface of 1,2-dioleoyl-3- trimethylammonium propane/DNA nanoparticles, ACS Appl. Mater. Interfaces, 5 (2013) 13171-13179. 40 [40] G. Caracciolo, S. Palchetti, V. Colapicchioni, L. Digiacomo, D. Pozzi, A.L. Capriotti, G. La Barbera, A. Laganà, Stealth effect of biomolecular corona on nanoparticle uptake by immune cells, Langmuir, 31 (2015) 10764-10773. 1 M. und vist, J. Stigler, . Cedervall, . ergg rd, M. . lanagan, . ynch, . Elia, K. Dawson, The evolution of the protein corona around nanoparticles: a test study, ACS Nano, 5 (2011) 7503-7509. [42] A.L. Capriotti, G. Caracciolo, C. Cavaliere, P. Foglia, D. Pozzi, R. Samperi, A. Lagana, Do plasma proteins distinguish between liposomes of varying charge density?, J. Proteomics, 75 (2012) 1924-1932. [43] J.D. Byrne, T. Betancourt, L. Brannon-Peppas, Active targeting schemes for nanoparticle systems in cancer therapeutics, Adv. Drug Delivery Rev., 60 (2008) 1615-1626. [44] V. Mirshafiee, M. Mahmoudi, K. Lou, J. Cheng, M.L. Kraft, Protein corona significantly reduces active targeting yield, Chem. Commun., 49 (2013) 2557-2559. [45] H. Amenitsch, G. Caracciolo, P. Foglia, V. Fuscoletti, P. Giansanti, C. Marianecci, D. Pozzi, A. Lagana, Existence of hybrid structures in cationic liposome/DNA complexes revealed by their interaction with plasma proteins, Colloids Surf. B, 82 (2011) 141-146. [46] A.L. Capriotti, G. Caracciolo, G. Caruso, P. Foglia, D. Pozzi, R. Samperi, A. Lagana, DNA affects the composition of lipoplex protein corona: A proteomics approach, Proteomics, 11 (2011) 3349-3358. [47] A. Diaz-Moscoso, L. Le Gourrierec, M. Gomez-Garcia, J.M. Benito, P. Balbuena, F. Ortega-Caballero, N. Guilloteau, C. Di Giorgio, P. Vierling, J. Defaye, C. Ortiz Mellet, J.M. García Fernandez, Polycationic amphiphilic cyclodextrins for gene delivery: synthesis and effect of structural modifications on plasmid DNA complex stability, cytotoxicity, and gene expression, Chem. Eur. J., 15 (2009) 12871-12888. [48] A. Gadelle, J. Defaye, Selektive halogenierung von cyclomaltooligosacchariden in C6‐position und synthese von Per (3, 6‐anhydro) cyclomaltooligosacchariden, Angew. Chem., 103 (1991) 94-95. [49] M. Gómez-García, J.M. Benito, D. Rodríguez-Lucena, J.-X. Yu, K. Chmurski, C. Ortiz Mellet, R. Gutiérrez Gallego, A. Maestre, J. Defaye, J.M. García Fernández, Probing secondary carbohydrate-protein interactions with highly dense cyclodextrin-centered heteroglycoclusters: the heterocluster effect, J. Am. Chem. Soc., 127 (2005) 7970-7971. [50] Y. Aoyama, T. Kanamori, T. Nakai, T. Sasaki, S. Horiuchi, S. Sando, T. Niidome, Artificial viruses and their application to gene delivery. Size-controlled gene coating with glycocluster nanoparticles, J. Am. Chem. Soc., 125 (2003) 3455-3457. [51] S.K. Misra, M. Muñoz-Ubeda, S. Datta, A.L. Barran-Berdon, C. Aicart-Ramos, P. Castro-Hartmann, P. Kondaiah, E. Junquera, S. Bhattacharya, E. Aicart, Effects of a delocalizable cation on the headgroup of gemini lipids on the lipoplex-type nanoaggregates directly formed from plasmid DNA, Biomacromolecules, 14 (2013) 3951-3963. [52] M. Muñoz-Ubeda, S.K. Misra, A.L. Barran-Berdon, S. Datta, C. Aicart-Ramos, P. Castro-Hartmann, P. Kondaiah, E. Junquera, S. Bhattacharya, E. Aicart, How does the spacer length of cationic gemini lipids influence the lipoplex formation with plasmid DNA? Physicochemical and biochemical characterizations and their relevance in gene therapy, Biomacromolecules, 13 (2012) 3926-3937. [53] M. Muñoz-Ubeda, S.K. Misra, A.L. Barran-Berdon, C. Aicart-Ramos, M.B. Sierra, J. Biswas, P. Kondaiah, E. Junquera, S. Bhattacharya, E. Aicart, Why is less 41 cationic lipid required to prepare lipoplexes from plasmid DNA than linear DNA in gene therapy?, J. Am. Chem. Soc., 133 (2011) 18014-18017. [54] J. Bednar, C.L. Woodcock, Chromatin, Academic Press Inc, San Diego, CA, 1999. [55] J. Dubochet, M. Adrian, J.J. Chang, J.C. Homo, J. Lepault, A.W. McDowall, P. Schultz, Cryo-electron microscopy of vitrified specimens, Q. Rev. Biophys., 21 (1988) 129-228. [56] J. Dubochet, B. Zuber, M. Eltsov, C. Bouchet-Marquis, A. Al-Amoudi, F. Livolant, How to "read" a vitreous section, in: Methods Cell. Biol., vol. 79, 2007, pp. 385- 406. [57] A.L. Barran-Berdon, M. Muñoz-Ubeda, C. Aicart-Ramos, L. Perez, M.R. Infante, P. Castro-Hartmann, A. Martin-Molina, E. Aicart, E. Junquera, Ribbon-type and cluster-type lipoplexes constituted by a chiral lysine based cationic gemini lipid and plasmid DNA, Soft Matter, 8 (2012) 7368-7380. [58] S. Palchetti, D. Pozzi, C. Marchini, A. Amici, C. Andreani, C. Bartolacci, L. Digiacomo, V. Gambini, F. Cardarelli, C. Di Rienzo, G. Peruzzi, H. Amenitsch, R. Palermo, I. Screpanti, G. Caracciolo, Manipulation of lipoplex concentration at the cell surface boosts transfection efficiency in hard-to-transfect cells, Nanomed. Nanotechnol, Biol. Med., (2016) in press, DOI: 10.1016/j.nano.2016.1008.1019. [59] D. Maiolo, P. Del Pino, P. Metrangolo, W.J. Parak, F.B. Bombelli, Nanomedicine delivery: does protein corona route to the target or off road?, Nanomedicine, 10 (2015) 3231-3247.