Cationic gemini lipids containing polyoxyethylene spacers as improved transfecting agents of plasmid DNA in cancer cells

Article abstract of DOI:10.1039/c4tb00389f

Lipoplex nano-aggregates have been analyzed through biophysical characterization (electrostatics, structure, size and morphology), and biological studies (transfection efficiency and cell viability) in five cancer cell lines. Lipoplexes were prepared from pEGFP-C3 plasmid DNA (pDNA) and mixed liposomes, constituted by a zwitterionic lipid (DOPE) and a gemini cationic lipid (GCL) synthesized in this work, [bis(hexadecyl dimethyl ammonium) oxyethylene], referred to as (C₁₆Am)₂(C₂O) _n, (where n is the oxyethylene spacer length, n = 1, 2 or 3, between the ammonium heads). Cryo-TEM micrographs show nano-aggregates with two multilamellar structures, a cluster-type (at low-to-medium GCL composition) and a fingerprint-type that coexists with the cluster-type at medium GCL composition and appears alone at high GCL composition. SAXS diffractograms show that these lipoplexes present three lamellar structures, two of them coexisting at low and high GCL composition. The optimized transfection efficiency (TE) of pDNA was higher for lipoplexes containing GCLs with a longer (n = 3) or shorter (n = 1) polyoxyethylene spacer, at high GCL composition (α = 0.7) with low charge ratio (ρ_eff = 2). In the all cancer cell lines studied, the TE of the optimized formulations was much better than those of both lipofectamine 2000 and lipoplexes with GCLs of the bis(hexadecyl dimethyl ammonium) alkane series recently reported. Probably, (a) the coexistence of two lamellar structures at high GCL composition synergizes the TE of these lipid vectors, (b) the orientation of the polyoxyethylene region in (C₁₆Am) ₂(C₂O)₃/DOPE may occur in such a way that the spacing between two cationic heads becomes smaller than that in (C ₁₆Am)₂(C₂O)₂/DOPE which is poor in terms of TE, and (c) the synergistic interactions between serum proteins and (C₁₆Am)₂(C₂O)_n/DOPE-pDNA lipoplexes containing a polyoxyethylene spacer improve TE, especially at high GCL content. Lipoplexes studied here show very low levels of toxicity, which confirm them as improved vectors of pDNA in gene therapy.

Full text of DOI:10.1039/c4tb00389f

Journal of
Materials Chemistry B
View Article Online
View Journal | View Issue
PAPER
Cationic gemini lipids containing polyoxyethylene
spacers as improved transfecting agents of plasmid
DNA in cancer cells†
Cite this: J. Mater. Chem. B, 2014, 2,
4640
Ana L. Barran-Berdon,‡a Santosh K. Misra,‡b Sougata Datta,b Monica Munoz-U´beda,a
´
´
´
˜
c
a
b
a
*
*
Paturu Kondaiah, Elena Junquera, Santanu Bhattacharya and Emilio Aicart
Lipoplex nano-aggregates have been analyzed through biophysical characterization (electrostatics,
structure, size and morphology), and biological studies (transfection efficiency and cell viability) in five
cancer cell lines. Lipoplexes were prepared from pEGFP-C3 plasmid DNA (pDNA) and mixed liposomes,
constituted by a zwitterionic lipid (DOPE) and a gemini cationic lipid (GCL) synthesized in this work,
[bis(hexadecyl dimethyl ammonium) oxyethylene], referred to as (C16Am)2(C2O)n, (where n is the
oxyethylene spacer length, n ¼ 1, 2 or 3, between the ammonium heads). Cryo-TEM micrographs show
nano-aggregates with two multilamellar structures, a cluster-type (at low-to-medium GCL composition)
and a fingerprint-type that coexists with the cluster-type at medium GCL composition and appears
alone at high GCL composition. SAXS diffractograms show that these lipoplexes present three lamellar
structures, two of them coexisting at low and high GCL composition. The optimized transfection
efficiency (TE) of pDNA was higher for lipoplexes containing GCLs with a longer (n ¼ 3) or shorter (n ¼ 1)
polyoxyethylene spacer, at high GCL composition (a ¼ 0.7) with low charge ratio (reff ¼ 2). In the all
cancer cell lines studied, the TE of the optimized formulations was much better than those of both
lipofectamine 2000 and lipoplexes with GCLs of the bis(hexadecyl dimethyl ammonium) alkane series
recently reported. Probably, (a) the coexistence of two lamellar structures at high GCL composition
synergizes the TE of these lipid vectors, (b) the orientation of the polyoxyethylene region in
(C16Am)2(C2O)3/DOPE may occur in such a way that the spacing between two cationic heads becomes
smaller than that in (C16Am)2(C2O)2/DOPE which is poor in terms of TE, and (c) the synergistic
Received 10th March 2014
Accepted 7th May 2014
interactions between serum proteins and (C16Am)2(C2O)n/DOPE-pDNA lipoplexes containing
a
DOI: 10.1039/c4tb00389f
polyoxyethylene spacer improve TE, especially at high GCL content. Lipoplexes studied here show very
www.rsc.org/MaterialsB
low levels of toxicity, which confirm them as improved vectors of pDNA in gene therapy.
1. Introduction
a
´
´
´
Grupo de Quımica Coloidal y Supramolecular, Departamento de Quımica Fısica I,
´
Facultad de Ciencias Quımicas, Universidad Complutense de Madrid, 28040
Nucleic acid delivery into the cells for therapeutic use is a
promising eld of modern medicine. However, there are a
variety of factors that prevent obtaining efficient and promising
outcomes and transfection.1–5 Cationic lipids (CLs) are
commonly used vectors to improve the efficiency in nucleic acid
delivery.6–10 Their structural similarity to the cell membranes
makes them appropriate delivery agents. CLs are known to pack
a large variety of nucleotides provoking a low immunogenic
response; however, their use is limited due to cytotoxicity
concerns, and moderate transfection efficiency (TE).10–13 In this
regard, recent efforts have been addressed to nd cationic
lipids, as gemini cationic lipids (GCLs), with higher transfection
capability and more biocompatibility.2,14–16 GCLs are amphi-
philic molecules containing two head groups linked by a rigid
or exible spacer.17–19 Compared to monomeric CLs, GCLs have
superior properties such as a lower critical micelle concentra-
tion (CMC), the presence of double charge per monomer, a
Madrid, Spain. E-mail: aicart@quim.ucm.es; Fax: +34 913944135
bDepartments of Organic Chemistry, Bangalore, Karnataka, India. E-mail: sb@
orgchem.iisc.ernet.in; Fax: +91 8023600529
cMolecular Reproduction Development and Genetics, Indian Institute of Science,
560012 Bangalore, Karnataka, India
† Electronic supplementary information (ESI) available: The procedure related to
the synthesis of the divalent cationic gemini lipids (C16Am)2(C2O)n, with the
oxyethylene spacer n ¼ 1, 2 or 3; description of the experimental methods;
tables of the effective charges for the three GCLs and plasmid DNA; the particle
size of mixed liposomes (C16Am)2(C2O)n/DOPE at several GCL compositions;
SAXS diffractograms and parameters for the three lamellar structures
determined from them; zeta potential plots; cryo-TEM images; uorescence
anisotropy results; transfection efficiency plots in the absence (ꢀFBS ꢀ FBS)
and presence (ꢀFBS + FBS) of serum in ve cancer cell lines; uorescence and
confocal uorescence images of transfected HEK293T cells with the optimized
formulations of the lipoplexes; cell viability in the presence (ꢀFBS + FBS) of
serum, with the optimized formulations, in ve cancer cell lines. See DOI:
10.1039/c4tb00389f
‡ These authors contributed equally.
4640 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
lower toxicity, a wider possibility of structural modications
and an unusual ability to induce inter-bilayer DNA correla-
tions,20 factors that present them as promising transfecting
agents.21–24 The aggregation properties of gemini lipids strongly
depend on the nature and length of the spacer between the two
head groups.25–29 The introduction of oxygen into the spacer can
affect the orientation in a different fashion; thus, the spacer
chain containing oxyethylene functionalities folds in associa-
tion with water molecules of the interfacial region, this addi-
tional hydration reduces the coulumbic repulsions between two
cationic heads30 and might improve the interaction between the
GCLs and DNA. Such interfacial polarity has been found to
inuence the membrane properties of lipid aggregates as well
as many crucial biological events.9 We have observed that
modications of interfacial hydration in cationic liposomes
could be of signicant importance to improve the nucleic acid
delivery efficiency.10–12 Thus, the extent of the variation at the
head groups,13 spacers14 or combination of both (head groups
and spacers)15 of cationic lipid assemblies could be of great
importance during their interaction with nucleic acids and, in
turn, may produce great changes in TE.
2. Experimental
2.1. Materials
The divalent cationic gemini lipids of the bis(hexadecyl
dimethyl ammonium) oxyethylene series, referred to as
(C16Am)2(C2O)n, with the oxyethylene spacer n ¼ 1, 2 or 3, were
synthesized according to a reported procedure detailed in the
1
ESI.† All the compounds were characterized by H-NMR spec-
troscopy, mass spectrometry and elemental analysis. Chemical
shis (d) reported in ppm downeld from the internal standard
(TMS) and the pertinent details are also given in the ESI.† The
zwitterionic lipid, 1,2-dioleoyl-sn-glycero-3-phosphatidyletha-
nolamine (DOPE), was purchased from Avanti Polar Lipids. The
sodium salt of calf thymus DNA (ctDNA), provided by Sigma-
Aldrich, was used as linear DNA to determine the lipid charge
(qL þ) of the GCL. pEGFP-C3 plasmid DNA (pDNA) was extrac-
þ
ted from competent E. coli bacteria previously transformed with
pEGFP-C3; the extraction was carried out using a GenElute HP
Select plasmid Gigaprep Kit (Sigma Aldrich) following a
protocol previously described.31,33
In previous studies, we have reported that supercoiled
plasmid DNA is compacted with a large number of its Na+
counterions which are not released when the lipoplex is
formed, thus, yielding a much lower effective negative charge
compared to the linear DNA form.16 In addition to this, the
2.2. Preparation of lipoplexes
Films of the lipid mixtures were prepared by dissolving appro-
priate amounts of gemini cationic lipid (L+) and DOPE (L0) in
chloroform to obtain the desired molar fraction composition (a)
(thus, L ¼ L+ + L0, is the total mass of lipid). Aer brief vor-
texing, the solvent was evaporated under high vacuum to yield a
dry lipid lm that was nally hydrated with HEPES buffer
40 mM, pH ¼ 7.4 and sequentially extruded according to a
procedure previously described.31,33 The desired lipoplex
composition, either in terms of the mass ratio (L/D), between
total lipid (L) to plasmid DNA (D), or the effective charge ratio
(reff), between the GCL and pDNA charges, was obtained by
mixing equal volumes of pDNA and (C16Am)2(C2O)n liposomal
suspensions34 (see ESI for details†). Concentrations of pDNA
were used to t the optimum conditions for each experimental
technique, as follows: 0.05 mg mLꢀ1 for zeta potential and
uorescence anisotropy, 1 mg mLꢀ1 for cryo-TEM, 200 mg
per capillary for SAXS, 200 ng per well for gel electrophoresis,
4 mg mLꢀ1 for each of TE determination and confocal uores-
cence microscopy, and 1 mg mLꢀ1 for the cell viability assays.
cationic lipid composition (a), the effective charge ratio (reff
)
and, specically, the length of the spacer in the gemini lipid
play a key role in determining an optimum transfection
efficiency.17 It was found that GCLs with short spacers affor-
ded higher transfection efficiency, probably due to the pres-
ence of two coexisting lamellar structures in such lipoplexes.
The higher charge density and lower hydrophobicity of the
head groups might have also played a role.31 In addition,
lipoplexes formed from GCLs with two imidazolium cationic
head groups rearrange in cluster-type lamellar structures,
their TE being controlled by the spacer length between the
imidazolium head groups.18
In the present study, we use the lipoplex formed by pEGFP-
C3 plasmid DNA (pDNA) and mixed liposomes, containing a
GCL of the bis(hexadecyl dimethyl ammonium) oxyethylene
series, referred to as (C16Am)2(C2O)n, (where n is the oxy-
ethylene spacer length, n ¼ 1, 2 or 3, between the ammonium
heads) and DOPE as the helper zwitterionic lipid.32 It is
2.3. Characterization of lipoplexes
expected that the use of a hydrophilic spacer will provide a The electrophoretic mobility (me) and hydrodynamic diameter of
better lipid–DNA interaction and better transfection results liposomes and/or lipoplexes were determined using a Phase
compared with the GCL bis(hexadecyl dimethyl ammonium) Analysis Light Scattering technique (Zeta PALS, Brookhaven
alkane series previously reported.9 The lipoplexes have been Instrum. Corp., USA).33,35 Each electrophoretic mobility datum
analyzed by zeta potential, gel electrophoresis, SAXS, cryo- is taken as an average over 50 independent measurements. The
TEM, uorescence anisotropy, transfection efficiency, uo- zeta potential (z) has been obtained from the electrophoretic
rescence and confocal uorescence microscopy and cell mobility, using the Henry equation.33 Zeta potential values ꢁfor
viability/cytotoxicity experiments, to establish a structure– the mixed liposomes and lipoplexes were measured, at 25 C,
biological activity relationship. As it will be shown, the for each a at different L/D ratios of the lipoplexes.
results reported here for in vitro studies support the use of
Gel electrophoretic mobilities of lipoplexes, along with
the formulations proposed in this work as optimized gene uncomplexed pure plasmid DNA, were photographed under UV
therapy vectors of plasmid DNA and may be extended for in light by running on 1% agarose gel under 100 mV in 1ꢂ TAE
vivo studies too.
(Tris–HCl, acetate and EDTA) buffer for ꢃ30 min. The
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4641

View Article Online
Journal of Materials Chemistry B
Paper
complexed pDNA appeared as a uorescent band in the wells and Lipo2000* under (ꢀFBS ꢀ FBS) and (ꢀFBS + FBS) condi-
while uncomplexed pDNA moved outside the well. The uo- tions, respectively (see ESI for details†). Samples were observed
rescence intensity of each band was an indication of the pDNA under a confocal microscope aer 48 h of incubation under
amount compared to the known amount of uncomplexed pDNA ambient conditions (Zeiss LSM 510-Meta Apochromat).
running outside the well (see ESI for details†).
The cell viability efficiency of optimized lipoplexes was
Small-angle X-ray scattering (SAXS) experiments were carried determined using the 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl
out using the Beam-line 26 at European Synchrotron Radiation tetrazolium bromide reduction method (MTT assay) in various
Facility (ESRF), Grenoble (France). The energy of the incident cell lines used for the transfection studies following a literature
beam was 12.6 keV (l ¼ 0.995 A). Samples were placed in sealed procedure.40,41 Lipoplexes were incubated with cells for 6, 24
˚
glass capillaries purchased from Hilgenberg with an outside and 48 h and MTT were added to each well for 4 h before dis-
diameter of 1.5 mm and a wall thickness of 0.01 mm. SAXS solving the blue formazan crystals in DMSO. The absorbance
experiments were carried out at different CL compositions (a) of was measured using a microtiter plate reader and % cell
the mixed lipid, and at several effective charge ratios (reff) of the viability was determined and is reported in detail in the ESI.†
lipoplexes.33
Cryo-TEM experiments were prepared following the general
3. Results and discussion
3.1. Mixed liposome–DNA electrostatic interaction
procedures described by several groups.36,37 In these experi-
ments we have used perforated Quantifoil R1.2/1.3 (1.2 mm of
hole diameter) on a 400-mesh copper grid. Images were The knowledge of electroneutrality in the mixed liposome–
obtained using a Jeol JEM 2011 cryo-electron microscope pDNA complex is the most important factor in the preparation
operated at 200 kV, under low-dose conditions, and using of the lipoplexes before the transfection studies. This is because
different degrees of defocus (500–700 nm) to obtain an lipoplexes must have a net positive charge (reff > 1) to be capable
adequate phase contrast.38 Images were recorded using a Gatan of interacting with the negative membrane of cells to become a
794 Multiscan digital camera. These CCD images were pro- potentially efficient cell transfecting agent. The most widely
cessed and analyzed with a digital micrograph.
used technique for determining the electroneutrality ratio ((L/
The uorescence anisotropy of the 1,6-diphenylhexatriene D)f at r ¼ 1) of nanoparticles is the zeta potential, although
(DPH) probe was determined with a Perkin Elmer LS-50B other methods such as agarose gel electrophoresis and
Luminiscence Spectrometer39 by measuring the intensities of ethidium bromide or GelRed intercalation assays may be
the light emitted by DPH as described previously.31 DPH used.12,39,42,43 Recently, we have demonstrated that in lipoplexes
included in the liposomal membrane was excited at 360 nm and with ctDNA the available charge of GCLs does not match with its
þ þ
its uorescence emission was recorded at 430 nm, with a slit nominal one (qL ¼ þ2), and the determination of its real
width of 2.5 nm for both the excitation and the emission.
value is crucial.31 Thus, the actual available charges of the
The transfection of pEGFP-C3 plasmid DNA across HEK293T cationic lipids ((C16Am)2(C2O)n with n¼ 1, 2 or 3) studied in this
(Human embryo kidney transformed cancer), HeLa (Human work have been determined following a repoꢀrted procedure31
cervical cell carcinoma), CaCO-2 (Human epithelial colorectal based on the charge of the ctDNA (qlinear DNA ¼ ꢀ2 per bp).
adenocarcinoma), Hep3B (Human hepatoma) and MDA-MB231 Fig. 1a shows the typical sigmoidal behaviour of the zeta
(Human breast cancer) cells using (C16Am)2(C2O)n/DOPE lipid potential vs. L/D for the (C16Am)2(C2O)n/DOPE-ctDNA lip-
mixtures was performed in the absence (ꢀFBS ꢀ FBS) and oplexes. The procedure permits determination of the electro-
þ þ
presence (ꢀFBS + FBS) of serum (see ESI for details†). Trans- neutrality (L/D)f of each GCL (ESI, Table S1† where the real qL
fection efficiencies were followed by uorescence microscopy at and (L/D)f values are reported). The average charge
þ þ
different time intervals and photographed at the end of 48 h of qL ¼ 1:75 ꢄ 0:05, obtained for (C16Am)2(C2O)n/DOPE-ctDNA
þ þ
incubation (see ESI for details†). Harvested cells were used for at a ¼ 0.5, its close to its expected nominal values (qL ¼ þ2).
þ þ
the Flow Assisted Cell Sorting (FACS), and % GFP cells and Using those qL values, the next step is the determination of
mean uorescence intensity (MFI) were obtained. Data were (L/D)f for lipoplexes formed by the mixed liposome and pDNA
analyzed using public domain soware WinMDI 2.8. The results at several GCL compositions (a) (Fig. 1a and S1, ESI†). Aer-
were plotted as % GFP cells and MFI against the effective charge wards, the effective charge of pDNA (qpDNAꢀ) has been obtained
ratio (reff) using graph pad prism 5.0 soware. Lipofect- according to the procedure previously described31,35 and also
amine2000 was used as a positive control during all the trans- detailed in the ESI.† Fig. 1b shows the effective charge of pDNA
fection experiments and acronymed as Lipo2000 and Lipo2000* (qpDNAꢀ) when interacts with the mixed liposomes containing
while used to transfect pDNA in the absence (ꢀFBS ꢀ FBS) or each of the three GCLs at several a compositions of the mixed
presence of serum (ꢀFBS + FBS), respectively.31,33
liposome, these values (ESI, Table S1†) being only 15–30% of the
Confocal Fluorescence Microscopy (CFM) study was per- total charge (ꢀ2 per bp), a similar range to that reported for
formed on HEK293T cells to visualize the GFP expression in pDNA when is compacted by mixed liposomes containing other
transfected cells along with PI stained nuclei. Experiments were types of GCLs.33,35,44 For (C16Am)2(C2O)n/DOPE-pDNA lipoplexes,
ꢀ
performed for the optimized formulations in the same way as in when n ¼ 1, qpDNA shows a slight decrease when a increases,
the transfection experiments with (ꢀFBS + FBS) and without while for n ¼ 2 or 3 remains roughly constant with a. In addi-
(ꢀFBS ꢀ FBS) serum conditions. Control experiments were tion, at constant a qpDNAꢀ shows a slight decrease as the spacer
performed using a commercial transfection reagent, Lipo2000 length (n) of the GCL increases. These results demonstrate that
4642 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Fig. 2 Gel electrophoretic retardation of (a) plasmid pEGFP-C3 and (b)
ctDNA by complexing them with (C16Am)2(C2O)/DOPE mixed lipo-
somes at a ¼ 0.7. Experiment was performed at reff ¼ 0.5, 1, 1.5, 2, 2.5
and 4.
wells by producing a fully complexed lipoplex at reff < 2 whereas
linear ctDNA requires more GCL for complete gel retardation.
These results are in full accordance with those obtained from
the zeta potential study.
3.2. Structure, morphology and bilayer uidity of the
lipoplex nano-aggregates
The structure, size, and morphology of the (C16Am)2(C2O)n/
DOPE-pDNA lipoplexes have been analyzed by means of PALS,
cryo-TEM and SAXS methods. The particle size was measured
for the three mixed liposomes, (C16Am)2(C2O)n/DOPE with
n ¼ 1, 2 or 3, the obtained average sizes in nm, independent of
n, were 106 ꢄ 2; 90 ꢄ 1; and 75 ꢄ 3 at a ¼ 0.2, 0.5 and 0.7,
respectively (ESI, Table S2†). Although, the same standard
hydration and extrusion procedure45 was followed for all the
samples, the size decreases and the polydispersity increases as
the GCL content, i.e., a increases. This feature is attributed to
the different aggregation properties of the GCL with respect to
the DOPE helper lipid.23,27
Cryo-TEM experiments were run on (C16Am)2(C2O)n/DOPE
mixed liposomes and positively charged (C16Am)2(C2O)n/DOPE-
pDNA lipoplex samples (reff > 1). The analysis of the micro-
graphs of mixed liposomes (data not shown) indicates that they
consist mostly of unilamellar and spherical nanostructures.
From all the micrographs three different liposome sizes have
been distinguished. Table S3† of ESI reports the three size
groups and the percentage contribution of each one to the
whole average liposome size at different GCL compositions (a).
Thus, at a ¼ 0.2, the percentage of the size distribution is higher
(55%) for bigger liposomes (ꢃ110 nm), while liposomes with
medium size (ꢃ80 nm) contribute 39% and those with a lower
size (ꢃ50 nm) only contribute 6%. On the other hand, at a ¼ 0.5,
the percentages of the size distribution for the three sizes are
comparable (36%, 43% and 21%, respectively) indicating an
increase of the presence of smaller liposomes vs. a, and thus
justifying the polydispersity increase (ESI, Table S2†). These
results are in agreement with the data found using the PALS
method where a decrease of the whole average size in the mixed
liposomes was also observed as a increases. Using an ImageJ
program, cryo-TEM micrographs (not shown) have permitted
estimation of the width of the lipid bilayer (dm) of the uni-
lamellar mixed liposomes; the average values decrease with a in
the 5.5–3.5 nm range. Fig. 3a–f and S2† of the ESI show a
selection of micrographs of (C16Am)2(C2O)n/DOPE-pDNA
Fig. 1 Plots of: (a) zeta potential (z) against the lipoplex nano-aggre-
gate composition (L/D) of (C16Am)2(C2O)n/DOPE plus ctDNA (solid
symbols) and pDNA (open symbols) for n ¼ 1 (squares), 2 (circles) or 3
(triangles) at
a
¼
0.5, and (b) effective pDNA charge for
(C16Am)2(C2O)n/DOPE-pDNA at several a. The dashed line corre-
ꢀ
sponds to a value of qlinear DNA ¼ ꢀ2.
the negative available charge of pDNA is dependent on the
surface charge of the mixed liposome, so that, when a decreases
and/or spacer increases,29 results in a greater handicap for the
mixed liposome to access to the negative charges of the pDNA
(Scheme 1). This feature is in agreement with previous results
and conrms that a less amount of GCL is necessary to compact
the pDNA according to its less negative available charge,31,33,35
resulting in a reduction of the lipoplex cytotoxicity during the
pDNA transfection process.
Gel electrophoretic retardation of pDNA was performed by
complexing it with (C16Am)2(C2O)n/DOPE mixed liposomes at
several a and reff values. A representative gel electrophoresis
image (Fig. 2 for n ¼ 1) shows the high efficiency of
(C16Am)2(C2O)/DOPE, which retards the pDNA progress to the
Scheme 1 (C16Am)2(C2O)n/DOPE-pDNA lipoplexes with n ¼ 1, 2 or 3
at constant a.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4643

View Article Online
Journal of Materials Chemistry B
Paper
Fig. 3 A selection of cryo-TEM micrographs showing a general view of the (C16Am)2(C2O)n/DOPE-pDNA lipoplex nano-aggregates at reff ¼ 1.5
and at several a values. (Panels a, c and e) (C16Am)2(C2O)1/DOPE-pDNA for GCL composition a ¼ 0.2, a ¼ 0.5 and a ¼ 0.7 and, (panels b, d and f)
(C16Am)2(C2O)3/DOPE-pDNA for a ¼ 0.2, 0.5 and 0.7.
lipoplexes at GCL composition, a ¼ 0.2, 0.5 or 0.7. These or 3, at several values of a, and reff > 1. At all a compositions,
micrographs (and others not shown) reveal that the resulting diffractograms show three Bragg peaks corresponding to a
lipoplexes are characterized by a multilamellar compaction lamellar structure (La) with the interlayer distance (d) being
pattern, consisting of a series of cationic lipidic bilayers with directly related to the q factor (d ¼ 2p/q100). This structure may
pDNA sandwiched between them, similarly to other lipoplexes be represented as alternating layers of the mixed lipids and the
previously reported.33,34,46,47 However and irrespective of the supercoiled pDNA, where d is the sum of the thicknesses of the
spacer length (n) of the GCL, two different multilamellar types lipid bilayer (dm) and the pDNA aqueous monolayer (dw), i.e.,
may be distinguished: cluster-type (CT) and ngerprint-type d ¼ dm + dw. The middle broad and smoothed peak that appears
(FP) nanostructures.
in all diffractograms arise from the DNA–DNA correlation (black
Micrographs shown in Fig. 3a and b for lipoplex nano- arrow), and its qpDNA factor permits the determination of the
aggregates with a DOPE rich content (a ¼ 0.2) indicate that the separation between pDNA in the monolayer (dpDNA ¼ 2p/qpDNA).
presence of pDNA induces aggregation leading to the formation Extra Bragg peaks with respect to those corresponding to a
of CT-type nano-aggregates, where the lipoplexes are deformed lamellar structure were observed in the diffractograms of
upon contact with other adjacent lipoplexes. Fig. 3c and d show (C16Am)2(C2O)/DOPE-pDNA lipoplexes (n ¼ 1) at a ¼ 0.2, 0.4,
that, when the GCL content in the lipoplexes increases (a ¼ 0.5), and 0.7, and for the spacers n ¼ 2 and 3 at a ¼ 0.2 and 0.7. These
in addition to the CT-type nano-aggregates, one also can see a results imply that there are at least two lamellar structures
continuum arrangement of multilamellae of mixed lipids with coexisting at these molar fractions. Thus, (a) at all the lipid
pDNA sandwiched among them following a FP-type structure. compositions (a) a lamellar structure (La, main) is present, (b) in
Finally, (Fig. 3e and f) at higher GCL content (a ¼ 0.7), only the DOPE rich region, where the GCL content is low (a ¼ 0.2 and
FP-type nano-aggregates can be found. It can be concluded that 0.4), a second structure (La, DOPE rich) has been found, and (c) at
when the DOPE content is high (low a), mixed liposomes higher GCL compositions (a ¼ 0.7) a third structure (La, GCL rich
)
essentially maintain their shape while supercially compacting is also present. It may be remarked that the phenomena of de-
pDNA (CT-type), their lipid bilayers becoming thickened and mixing,48 i.e., the presence of several coexisting structures, has
twisted or even looped. Note that many of the mixed liposomes been already found in other lipoplexes containing other GCLs
are not disrupted. However, when the GCL content increases, mainly at low or high GCL compositions.31,33,35,44 Also noticeable
the liposome–pDNA interaction is clearly favoured and lipo- is that both cryo-TEM and SAXS agree on showing this coexis-
somes tend to disrupt in order to form FP-type nano-aggregates, tence of lamellar structures which, as will be described later,
a feature already found in lipoplexes containing other GCL may improve pDNA transfection. Diffractograms of Fig. 4 also
series.31
show that the Bragg peaks corresponding to the La, main struc-
Fig. 4a–c and S3–S5† of ESI show SAXS diffraction data ture are shied to the right, which results in higher q values as a
collected for (C16Am)2(C2O)n/DOPE-pDNA lipoplexes for n ¼ 1, 2 increases. According to this, the interlamellar distance (d)
4644 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Fig. 5 (a) Plots of the periodic distance of the lamellar structure (d) for
(C16Am)2(C2O)/DOPE-pDNA lipoplex nano-aggregates as a function
of GCL composition (a) at several effective charge ratios (reff): red
symbols, La, DOPE rich structure; black symbols, La, main structure; and
blue symbols, La, GCL rich structure. In all the structures: squares, reff
¼
1.5; circles, reff ¼ 2.0; up triangles, reff ¼ 2.5; down triangles, reff ¼ 3.0;
and left triangles, reff ¼ 5.0. (b) Plots of the distance dpDNA for
(C16Am)2(C2O)/DOPE-pDNA nano-aggregates as a function of GCL
composition (a) at several effective charge ratios (reff).
factors: a thinner bilayer (dm) as a increases because the length
of the cationic gemini surfactants is slightly shorter than that of
DOPE, and/or a decrease of the monolayer (dw) at higher a
because the surface charge area in the liposomes is increased
resulting in a higher compaction of pDNA. The other two found
structures, the one rich in DOPE (La, DOPE rich) has a higher inter-
lamellar distance (dLa, DOPE rich z 8.7–7.3 nm), while the one
rich in gemini (La, CL rich) has a lower distance (dLa, GCL rich z 5.5
nm), both compared to the La,
structure. Structural
main
parameters of all the lamellar structures found at any a and reff
are reported in Tables S4–S6 of ESI.† In any case, the calculated
d values together with those obtained from cryo-TEM for the
lipid bilayer (dm) permit estimation of the thickness of the
monolayer (dw ꢃ 2.0–2.5 nm) appropriate to host the plasmid
DNA. Because the peak related to the DNA–DNA correlation is
low and very broad in all diffractograms, it has been assigned
mainly to the two lamellar structures formed at the medium-to-
Fig.
4 SAXS diffractograms of (C16Am)2(C2O)n/DOPE-pDNA lip-
oplexes at various GCL compositions (a) and effective charge ratio
(reff) ¼ 1.5. (a) (C16Am)2(C2O)/DOPE-pDNA, (b) (C16Am)2(C2O)2/DOPE-
pDNA, and (c) (C16Am)2(C2O)3/DOPE-pDNA. In all the structures: black
lines, a ¼ 0.2; red lines, a ¼ 0.4; blue lines, a ¼ 0.5; and purple lines,
a ¼ 0.7. Miller indices (hkl) correspond to the lamellar structures: red,
L
a, DOPE rich; black, La, main; and blue, La, GCL rich. Arrows indicate the
pDNA–pDNA correlation peak.
high GCL composition, i.e., La, main and La, GCL rich
.
From the qpDNA of this peak, the distance DNA–DNA (dpDNA
)
plotted vs. a shown in Fig. 5a and S6(a and c) of ESI† for has been determined in the three (C16Am)2(C2O)n/DOPE-pDNA
(C16Am)2(C2O)n/DOPE-pDNA lipoplexes (at all reff) slightly lipoplexes at all the studied GCL compositions (a) and effective
decreases with a in the La, main structure (dLa, main z 7.0–6.0 charge ratios (reff) (see ESI, Tables S4–S6†), the values are
nm). This feature may be attributed to a combination of two plotted vs. a in Fig. 5b and S6(b and d) of the ESI.† The dpDNA
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4645

View Article Online
Journal of Materials Chemistry B
Paper
plots show a slight decrease (5–3 nm) at all reff, which indicates 0.5 showed lower % GFP and MFI with not much variation in
that DNA strands reduce the dpDNA as a increases in the lip- GFP expression (ESI, Fig. S8†). Under similar conditions,
oplex, similar to what was reported for lipoplexes with other (C16Am)2(C2O)2/DOPE could show bar diagrams at all the
GCLs with different head groups or the type of spacer.31,33,48,49
compositions (a) with a maximum of % GFP cells ꢃ85 and MFI
The lipid bilayer uidity is a biophysical parameter known to ꢃ140 at a ¼ 0.4 and reff ¼ 1.5 while lipoplexes at a ¼ 0.2, 0.5 and
affect the viability of the liposomes as DNA carriers and their 0.7 showed not much variation in GFP expression (ESI, Fig. S9†).
capability as transfecting agents.50–52 In this work, uorescence Similarly, (C16Am)2(C2O)3/DOPE could transfect the cells in
anisotropy (r) was used to determine the uidity of the bilayer of such a way that the (C16Am)2(C2O)3 and DOPE composition at a
the (C16Am)2(C2O)n/DOPE liposomes and (C16Am)2(C2O)n/ ¼ 0.7 and reff ¼ 2 could express the % GFP cells to ꢃ50 with a
DOPE-pDNA lipoplexes. Fig. S7(a–f)† of ESI shows the anisot- MFI of ꢃ290 leaving other compositions not so good in trans-
ropy values of the DPH uorophore located in the bilayer of the fection efficiency in the order of compositions 0.7 > 0.5 [ 0.4 >
(C16Am)2(C2O)n/DOPE liposomes, in the absence (reff ¼ N) and 0.2 (ESI, Fig. S10†). Only (C16Am)2(C2O)3/DOPE at a ¼ 0.7 at reff
in the presence of pDNA, as a function of temperature. In all the ¼ 2 with % GFP cells ꢃ50 and MFI ꢃ290 was found to be better
cases, the lipid bilayer uidity increases with T, in agreement than the Lipo2000 a positive control used under similar
with previous reports for a wide variety of lipoplexes.42,46,47,50,53–55 conditions.
As expected, the uidity increases slightly at higher DOPE
HEK293T cells were further screened for GFP gene expres-
content (low a).46,50 According to the anisotropy results, the sion followed by transfection of pDNA using the same lipid
transition temperatures (Tm) of the (C16Am)2(C2O)n/DOPE lipo- mixtures but under ꢀFBS + FBS conditions having 10% FBS
somes are below the experimental temperature range studied, with lipoplexes while incubated with cells during transfection
which conrms that liposome and lipoplex bilayers are in the experiments for 6 h. It was noticed that addition of serum
uid state at physiological temperature, even in the presence of decreases the MFI values to a considerable extent for all the
pDNA. The low anisotropy levels for all the (C16Am)2(C2O)n/ used a and reff values across all the three lipid mixtures
DOPE-pDNA lipoplexes (0.1–0.22) point to a high enough (C16Am)2(C2O)n/DOPE. The lipoplex of the (C16Am)2(C2O)/DOPE
uidity to consider these lipoplex nano-aggregates as poten- lipid mixture could transfect a maximum of ꢃ40% cells with a
tially efficient transfecting agents of pDNA.
MFI of ꢃ110 at a ¼ 0.7 and reff ¼ 1.5 while the trend of
decreasing order of TE was followed by lipoplexes at a ¼ 0.4 and
reff ¼ 2, at a ¼ 0.5 and reff ¼ 4, and at a ¼ 0.2 and reff ¼ 4, with
% GFP positive cells ꢃ40, 25, 30 with MFI ꢃ80, 75, 80, respec-
3.3. Transfection efficiency of lipoplexes
The transfection efficiency (TE) of various lipid mixtures has to tively and comparable or better than the positive control
be ascertained for generating a correlation between biophysical Lipo2000* (50% GFP with MFI ꢃ70) (ESI, Fig. S11†). The
characteristics of lipoplexes and their TE. There is an urgent (C16Am)2(C2O)2/DOPE-pDNA lipoplexes showed poor TE and
need to understand the biasness of lipoplexes against some of lower than Lipo2000* for all a and reff values, with a maximum
the cell lines where they do not perform well in terms of TE, of ꢃ35% transfected cells having a MFI of ꢃ50 at a ¼ 0.4 and
while being better in other cells. It has to be established that the reff ¼ 1.5 (ESI, Fig. S12†). On the other hand, lipoplex
combination of physical factors, nature and origin of cell lines (C16Am)2(C2O)3/DOPE-pDNA was much better at a ¼ 0.5 and 0.7
determines the nal outcome of the transfection procedure.
than at 0.4 or 0.2 which were very poor in terms of % GFP while
To investigate the cell dependent transfection efficiency of MFI (ꢃ80) was a little better at a ¼ 0.2 and reff ¼ 1. Thus, TE
various used lipoplexes of GCLs mixed with DOPE at various showed a maximum at ꢃ25% GFP with MFI ꢃ85 at a ¼ 0.5 and
GCL compositions and effective charge ratios with pDNA, FACS reff ¼ 1.5, and ꢃ45% GFP with MFI ꢃ135 at a ¼ 0.7 and reff ¼ 2
(ow assisted cell sorting) analysis was performed on trans- last formulation much better than the control Lipo2000* (ESI,
fected cells. FACS analysis gives information about the Fig. S13†).
comparative expression of green uorescent protein and
To conclude the FACS analysis for the TE of different lipid
transfection efficiency in terms of % GFP cells and the mean mixtures (see Fig. 6), in the absence of serum (ꢀFBS ꢀ FBS)
uorescence intensity (MFI) into the cells. Here the MFI value conditions, lipid mixture (C16Am)2(C2O)3/DOPE, at a ¼ 0.7 and
indicates the level of GFP protein expression per cell which reff ¼ 2, was found to be the best which was at least ꢃ2 times
occurred because of the high copy number of pDNA expressed better in terms of MFI compared to control Lipo2000 (Fig. 6c).
as protein in each cell while % GFP cells indicates the Although in the presence of 10% serum conditions (ꢀFBS +
percentage of cell population expressing GFP protein over FBS) the TE of all the lipid compositions was brought down
residual uorescence of untransfected cells or transfected with signicantly, the lipid mixture (C16Am)2(C2O)3/DOPE was again
only pDNA and lipid mixture alone.
found to be the best with ꢃ40% transfected cells with ꢃ135 of
Initially, HEK293T cells were used to transfect with MFI which was, in terms of MFI, at least ꢃ2 times better than
(C16Am)2(C2O)n/DOPE at several a and reff values, in the absence Lipo2000* (Fig. 6f). The overall trend of TE, in the absence of
(ꢀFBS ꢀ FBS) and presence (ꢀFBS + FBS) of 10% serum serum, among all the lipid mixtures was such that
condition. In the absence of serum (ꢀFBS
ꢀ
FBS), (C16Am)2(C2O)3/DOPE (Fig. 6c) was found to be better than
(C16Am)2(C2O)/DOPE could show bar diagrams vs. reff at all the (C16Am)2(C2O)/DOPE (Fig. 6a) which is comparable to
compositions (a) with a maximum of % GFP cells ꢃ40 and MFI (C16Am)2(C2O)2/DOPE (Fig. 6b) stating that the spacer length
ꢃ150 at a ¼ 0.7 and reff ¼ 1.5 while lipoplexes at a ¼ 0.2, 0.4 and did not inuence the TE of the lipid mixture too much. But in
4646 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Fig. 6 Transfection of pEGFP-C3 in HEK293T cells in the absence (ꢀFBS ꢀ FBS) [a–c] and presence of serum (ꢀFBS + FBS) [d–f] using lipoplexes
of (C16Am)2(C2O)/DOPE (a and d), (C16Am)2(C2O)2/DOPE (b and e) and (C16Am)2(C2O)3/DOPE (c and f) at a ¼ 0.2, 0.4, 0.5 and 0.7, and reff ¼ 1, 1.5,
2, 2.5 and 4 as mentioned in bar diagrams. Experiments were performed using 0.8 mg pDNA per well. Lipo2000 and Lipo2000* were used as
positive controls for (ꢀFBS ꢀ FBS) and (ꢀFBS + FBS) transfection conditions, respectively.
the presence of serum, it seems that the spacer length highly moderate % GFP and MFI values not showing any particular
inuences the TE of lipid mixtures and a maximum drop in TE pattern. But in any case, the optimized formulations using lipid
was noticed in the case of (C16Am)2(C2O)2/DOPE (Fig. 6e) fol- mixtures (C16Am)2(C2O)n/DOPE with n ¼ 1 or 3, the TE of the
lowed by (C16Am)2(C2O)/DOPE (Fig. 6d) and lower inuence on lipid mixture in all cell lines is comparable or much better than
(C16Am)2(C2O)3/DOPE (Fig. 6f). Additionally, the other point to that of the positive control Lipo2000*. This comparative anal-
be noted here is that optimized formulation of the lipid ysis states that: (a) two units of oxyethylene were good for none
mixtures with higher TE is associated with low DOPE content of the cell lines and, (b) three oxyethylene units in the spacer
(a
¼
0.7), especially for those with better TE, viz. region between two positive head groups improves the TE in
(C16Am)2(C2O)n/DOPE with n ¼ 3 and 1, than the control HEK293T cells whereas in all other cells only a single unit of
Lipo2000*. This behaviour will be deeply discussed further on oxyethylene is required. Probably, the combination of a long
the basis of the relationship between the lipoplex structure and poly-oxyethylene spacer and two ammonium cationic head-
its biological activity.
groups is better suited for cancer cells of transformed nature
3.3.1. Cell dependent transfection efficiency of lipoplexes. (HEK293T) compared to a shorter oxyethylene spacer and two
The TE of the prepared lipoplexes, at previously optimized a ammonium cationic head-groups which were found better for
and reff values, was investigated in four more human cell lines cancer cells of human origin viz., CaCO-2, HeLa, Hep3B and
of different origins but possessing cancer characteristics. MDA-MB231 cells because of unknown reasons. This exciting
Optimized lipoplexes [(C16Am)2(C2O)/DOPE at a ¼ 0.7 with information needs further investigation to reveal the biasness
reff ¼ 2; (C16Am)2(C2O)2/DOPE at a ¼ 0.7 with reff ¼ 1.5 and in the interaction of the lipid mixture against different cancer
(C16Am)2(C2O)3/DOPE at a ¼ 0.7 with reff ¼ 2] were used to cells. Experiments are underway to nd out the actual reason
transfect CaCO-2, HeLa, Hep3B and MDA-MB231 cells in the behind this discrepancy when no targeting moiety has been
presence of serum (ꢀFBS + FBS) condition and the obtained introduced into the lipid structure or targeting molecule in the
FACS data were compared with the TE of Lipo2000* positive lipid mixtures.
control (ESI, Fig. S14†). It was found that the lipoplex
3.3.2. Transfection efficiency by uorescence microscopy.
(C16Am)2(C2O)/DOPE was the best among all three optimized Fluorescence microscopy could show expression of green uo-
lipoplexes in all other cell lines except HEK293T where the rescent protein in all the transfected HEK293T cells whereas
lipoplex (C16Am)2(C2O)3/DOPE was the best transfecting agent untransfected cells were devoid of any specic and signicant
(ESI, Fig. S14†). In all the cases the lipid mixture green uorescence as shown in representative uorescence
(C16Am)2(C2O)2/DOPE was reported poor irrespective of the cell images in the presence (Fig. 7) and absence (ESI, Fig. S15†) of
lines used for the transfection experiment. The pattern of cell serum. In the absence of serum (ꢀFBS ꢀ FBS), cells transfected
dependent TE was also found true for positive control with the lipoplex formulation (C16Am)2(C2O)3/DOPE-pDNA at a
Lipo2000* in CaCO-2, HeLa, Hep3B and MDA-MB231 cells, with ¼ 0.7 with reff ¼ 2 showed maximum green uorescence (ESI,
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4647

View Article Online
Journal of Materials Chemistry B
Paper
Fig. S15c2†) followed by cells transfected with (C16Am)2(C2O)/ visual level were also reported (Fig. 8(a1–a4)). A signicant GFP
DOPE-pDNA at optimized a ¼ 0.7 with reff ¼ 1.5 (ESI, expression was seen in cells transfected with (C16Am)2(C2O)3/
Fig. S15a2†) and (C16Am)2(C2O)2/DOPE-pDNA at a ¼ 0.4 and reff DOPE-pDNA while the GFP protein was distributed to the whole
¼ 1 (ESI, Fig. S15b2†). In the presence of serum (ꢀFBS + FBS) cytoplasm (Fig. 8c4) as in the case using the positive control
cells transfected with the lipoplex formulation (C16Am)2(C2O)3/ Lipo2000* (Fig. 8b4). This feature indicates the existence of GFP
DOPE-pDNA at a ¼ 0.7 with reff ¼ 2 (Fig. 7c2) were found at least protein inside the cell which was produced aer successful
2 times more green uorescent compare to cells transfected transfection of much higher pDNA when using the optimum
with Lipo2000* (Fig. 7b2). These results were very much in formulation of the (C16Am)2(C2O)3/DOPE-pDNA lipoplex and
agreement with the FACS data acquired from cell transfected conrms its better suitability as the transfecting agent
under similar conditions, and conrm the formulation compared to the control Lipo2000*.
(C16Am)2(C2O)3/DOPE-pDNA at a ¼ 0.7 with reff ¼ 2 as the
signicantly better transfecting agent compared to Lipo2000*
under 10% FBS conditions.
3.4. Cell viability assay
The GFP expression and its localization in transfected
HEK293T cells were visualized using confocal uorescence
microscopy study 48 h post transfection (Fig. 8 and S16, ESI†).
The nucleus of the cells has been stained (red color) with pro-
pidium iodide (PI) due to the presence of endogenous DNA in
cells. Highly uorescent GFP positive HEK293T cells were seen
aer transfection with pDNA in the presence (Fig. 8) and
absence (ESI, Fig. S16†) of serum indicating that % GFP positive
cell data obtained from FACS analysis were not a false positive.
Representative confocal images of HEK293T cells transfected in
the absence of serum with lipoplexes (C16Am)2(C2O)/DOPE-
pDNA (ESI, Fig. S16(a1–a4)†), (C16Am)2(C2O)2/DOPE-pDNA (ESI,
Fig. S16(b1–b4)†) and (C16Am)2(C2O)3/DOPE-pDNA (ESI,
Fig. S16(c1–c4)†) showed the existence of integral GFP expres-
sion. Similarly, representative confocal images of HEK293T
cells transfected in the presence of serum were reported with:
(a) the optimum formulation for the lipoplex (C16Am)2(C2O)3/
DOPE-pDNA at a ¼ 0.7 with reff ¼ 2 (Fig. 8(c1–c4)), and (b)
Lipo2000* as positive control (Fig. 8(b1–b4)). Untransfected
cells showing no residual green uorescence in the cells at the
MTT assay experiments were performed on the ve cell lines
used for transfection studies to nd the biocompatibility of all
the (C16Am)2(C2O)n/DOPE-pDNA lipoplexes at the optimized
GCL composition (a) and effective charge ratios (reff) (Fig. 9 and
S17, ESI†). Representative data using the HEK293T cell line
showed that a considerably high percentage of cells remained
viable to grow and divide aer the lipoplex incubation with cells
for 6 and 12 h while 24 h incubation causes a certain decrease in
cell viability (Fig. 9). Irrespective of the spacer length, all the
GCLs were found much better than the Lipo2000* which causes
considerable cell death even at smaller periods of incubation
which was further enhanced at 24 and 48 h incubation over all
the used cell lines (Fig. 9 and S17, ESI†). This observation
potentiates very high feasibility of using such GCLs for better
transfecting efficiency. The cell viability obtained using
(C16Am)2(C2O)n/DOPE-pDNA lipoplexes of this work is higher
than that was found with (C16Am)2Cn/DOPE-pDNA lipoplexes
reported recently.33 We believe that this high cell viability (lower
cytotoxicity) is a result of two factors: (a) the lower toxicity of the
polyoxyethylene spacer of the (C16Am)2(C2O)n lipid series
Fig. 7 Representative bright field (a1–c1) and fluorescence images (a2–c2) of untransfected HEK293T cells (a1 and a2); transfected with Lipo2000*
(b1 and b2) and with [(C16Am)2(C2O)3/DOPE-pDNA] (c1 and c2) prepared at the optimized formulation (a ¼ 0.7 and reff ¼ 2) in the presence of serum
(ꢀFBS + FBS). Experiments were performed using 0.8 mg pDNA per well. Lipo2000* was used as positive control. Scale bar is 50 mm.
4648 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Fig. 8 Representative confocal fluorescence images of untransfected HEK293T cells (a1–a4); transfected with Lipo2000* (b1–b4) and with
[(C16Am)2(C2O)3/DOPE-pDNA] (c1–c4) prepared at the optimized formulation (a ¼ 0.7 and reff ¼ 2) in the presence of serum (ꢀFBS + FBS). Panel
(a1–c1), (a2–c2), (a3–c3) and (a4–c4) represent GFP expression in cells, bright field, PI staining and overlapped images, respectively. Experiments
were performed using 0.8 mg pDNA per well. Lipo2000* was used as a positive control. Scale bar is 50 mm.
compared to the alkyl spacer of the (C16Am)2Cn lipid series, and therapy. As detailed in a previous section, SAXS and cryo-TEM
(b) the small quantity of the GCL used in the formulations of the methods, used to determine the structures of the compacted
lipoplexes at the optimum formulations (reff ¼ 2) in the pres- lipoplexes studied in this work, have concluded that, for any of
ence of the (C16Am)2(C2O)n/DOPE mixed lipids of this work (C16Am)2(C2O)n/DOPE-pDNA lipoplexes analyzed at several a
compared to the (C16Am)2Cn/DOPE ones (reff ¼ 4) recently and reff values, no HcII structure has been found but three La
studied.33
structures (La, main, La, DOPE rich and La, CL rich) are present, two of
them (La, main and La, CL rich) coexisting at high GCL composi-
tion (a ¼ 0.7) where the highest transfection efficiency has been
found for the three (C16Am)2(C2O)n/DOPE-pDNA lipoplexes.
Nevertheless, a recent study33 of the (C16Am)2Cn/DOPE-pDNA
lipoplexes with an alkyl spacer of n ¼ 2, 3, 5 or 12 indicated the
presence of two coexisting La structures in the whole GCL
content, and the best TE was found in lipoplexes containing
a shorter alkyl spacer (n ¼ 2 or 3) and at low GCL content
(a ¼ 0.2). Thus, the synergistic effect produced by the presence
of several lamellar structures may be a reason to explain why the
3.5. Lipoplex structure–biological activity relationship
When DNA is initially compacted by the mixed liposomes, it is
expected that the resulting structure will depend on the
molecular characteristics of the three compounds (GCL, helper
lipid and DNA). It is obvious that the lipoplex structure may
inuence the transfection efficiency shown by the lipoplex when
it delivers the DNA to the cells. In general, it was initially
assumed that lipoplexes with an inverted hexagonal structure
(HcII) are more efficient than lipoplexes with a lamellar one (La).
This feature was attributed to that the HcII structure is more
fusogenic than La, and, aer endocytosis, lipoplexes with the
HcII structure interact better with the cell membrane and release
DNA, escaping from the endosome to the cytosol. However, this
analysis was questioned from a deeper study of the lamellar
structures54,56 which indicated that the membrane charge
density of the mixed liposome is a universal parameter that
controls the transfection efficiency in lamellar lipoplexes.
Nevertheless, this conclusion has been subjected to a high
controversy and, in fact, several studies have concluded that the
membrane charge density is not
a universal parameter
controlling the TE of lipoplexes57,58 and that highly efficient
lipoplexes can be lamellar in nature, so that the correlation
hexagonal phase-high TE does not exist.59 On the other hand,
systematic studies of the structure and transfection efficiency of
several lipoplexes containing GCLs have made it possible to
establish a correlation between the coexistence of several lip- Fig. 9 Cell viability assay of HEK293T cells transfected in the presence
oplex structures and higher transfection efficiency.31,33,60 Those
results conclude that the presence of a structural poly-
morphism, i.e., the coexistence of La and inverted HcII struc-
tures,31 or even two La ones33 synergizes the transfection and
of serum (ꢀFBS + FBS) with lipoplexes prepared at the optimized
formulations: [(C16Am)2(C2O)/DOPE-pDNA] at a ¼ 0.7 and reff ¼ 1.5;
[(C16Am)2(C2O)2/DOPE-pDNA] at
[(C16Am)2(C2O)3/DOPE-pDNA] at a ¼ 0.7 and reff ¼ 2. Experiments
a
¼
0.4 and reff
¼
1.5; and
were performed by incubating lipoplexes at different time points of 6,
points to the GCLs among the best CLs to be used in gene 24 and 48 h.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4649

View Article Online
Journal of Materials Chemistry B
Paper
(C16Am)2(C2O)n/DOPE-pDNA lipoplexes with
a
high GCL compact and transfect pEGFP-C3 plasmid DNA. The spacer
content are the best lipid mixture for TE, but other effects must length between the two head groups, the mixed lipid compo-
be considered to explain why: (a) in the (C16Am)2Cn/DOPE- sitions and the lipoplex effective charge ratio hold the key that
pDNA lipoplexes the best TE was found when the alkyl spacer is determine the properties, structure and transfection efficiency
shorter33 while in the (C16Am)2(C2O)n/DOPE-pDNA lipoplexes of of the resulting nano-aggregates. The three gemini lipids
this work the best TE has been found with the shorter and (C16Am)2(C2O)n with n ¼ 1, 2 or 3 have a lower effective charge
þ þ
longer oxyethylene spacer (n ¼ 1 or 3), and (b) the best lipid qL ¼ ð1:75 ꢄ 0:05Þ than its nominal one (+2), while pDNA is
mixture composition (a) for the (C16Am)2Cn/DOPE-pDNA lip- compacted yielding only a 10–30% effective negative charge
oplexes with an alkyl spacer was reported at high DOPE content (qpDNAꢀ) than that of the linear DNA (ꢀ2 per bp). Cryo-TEM
(a ¼ 0.2), while in the case of the (C16Am)2(C2O)n/DOPE-pDNA micrographs show two types (cluster-type and ngerprint-type)
lipoplexes with a polyoxyethylene spacer of this work a best lipid of lamellar packing. The cluster-type is observed at low a (¼ 0.2),
mixture has been found at high GCL content (a ¼ 0.7). In this both, cluster and ngerprint types, coexist at a ¼ 0.5, and only
sense, one notable difference between the (C16Am)2Cn and the ngerprint-type is present at high a (¼ 0.7). SAXS dif-
(C16Am)2(C2O)n lipid series is that the alkyl spacer (Cn) is a fractograms of the (C16Am)2(C2O)n/DOPE-pDNA lipoplexes
hydrophobic moiety that has to be allowed as deep as possible reveal the presence of three lamellar structures, two of them
in the lipid bilayer of the mixed liposome, while the polyoxy- coexisting at low and high GCL compositions (a). At all the
ethylene spacer [(C2O)n] is a more polar moiety that can be lipid compositions, a lamellar structure (La, main) is present; in
allocated on the surface of the lipid bilayer of the mixed lipo- the DOPE rich region (a ¼ 0.2 and 0.4) a second structure
some, and thus, can also interact with DNA. Further, the higher (La,
rich) coexists with the La, main; and at higher GCL
DOPE
TE of lipid mixtures (C16Am)2(C2O)/DOPE and (C16Am)2(C2O)3/ compositions (a ¼ 0.7) a third structure (La, CL rich) coexisting
DOPE over (C16Am)2(C2O)2/DOPE may reveal that the orienta- with the La, has been also found. The optimized trans-
main
tion of the polyoxyethylene region in (C16Am)2(C2O)3 probably fection efficiency (TE) of pDNA was higher for lipoplexes con-
occurs in such a way that the distance between two cationic taining GCLs with a longer (n ¼ 3) or shorter (n ¼ 1)
heads reduces signicantly and becomes smaller than that one polyoxyethylene spacer, at high GCL composition (a ¼ 0.7) and
in (C16Am)2(C2O)2 which is poor in terms of TE. In this regard low charge ratio (reff ¼ 2), and in all cancer cell lines studied,
we have already reported that a smaller distance between two much better than both Lipo2000* and lipoplexes with GCLs of
cationic head groups on GCLs was benecial for higher TE.35 On the bis(hexadecyl dimethyl ammonium) alkane series recently
the other hand, it has been also demonstrated recently61 that reported. Probably, (a) the coexistence of two lamellar struc-
the different composition in the mixed liposomes plays an tures at high GCL composition synergizes the TE of these lipid
important role in the preferential adsorption of serum proteins. vectors, (b) the orientation of the polyoxyethylene region in
Additionally, the liposome–protein corona interactions have (C16Am)2(C2O)3/DOPE may occur in such a way that the distance
been deeply studied recently.62–64 In fact, a structural analysis between two cationic heads becomes smaller than that in
has revealed61 that DOPE (helper lipid) interacts with human (C16Am)2(C2O)2/DOPE which is poor in terms of TE, and (c)
serum albumin (HSA) polypeptide polar groups involving the synergistic interactions between serum proteins and
changes in hydrogen bonding, while hydrophobic interactions (C16Am)2(C2O)n/DOPE-pDNA lipoplexes containing a polyoxy-
between CLs of one head group (as DOTAP or DDAB) with HSA ethylene spacer improve TE, especially at high GCL content.
are stronger and may cause major changes in the protein Lipoplexes studied here show very low levels of toxicity in vitro,
conformation caused by a reorganization of intra- and inter- conrming them as improved vectors of pDNA and in turn
molecular hydrogen bonding similar to drug/HSA complexes.61 potentially adequate for gene therapy studies in vivo too.
Additionally, it must be noted that GCLs with two head groups
linked by a polyoxyethylene spacer as those used in this work
may also affect the GCL–serum protein interaction. In fact, a
Acknowledgements
recent study reports that similar GCLs with oxyethylene spacers
The authors thank MICINN of Spain (projects no. ACI2009-
show a synergistic interaction with serum proteins which led to
0867, CTQ2009-10002BQU and CTQ2012-30821), DST of India
enhanced DNA binding, DNA release, stability against DNase I
(project no. DST/INT/SPAIN/P-8/2009), and Univ. Complutense
and in turn TE of lipoplexes.65 Probably, similar interactions
of Madrid (Spain) (project no. UCMA05-33-010). The authors
govern the TE of (C16Am)2(C2O)n/DOPE-pDNA lipoplexes while
also thank BM26 at ESRF, Grenoble (France) for SAXS experi-
the number of oxyethylene units decides the fate of the inter-
´
ments, and Servei de Microscopia of Univ. Autonoma of Bar-
celona (Spain) for cryo-TEM experiments.
action. In any case and in order to conrm this feature, further
studies in vitro and in vivo are welcomed.
4. Conclusions
Notes and references
The biophysical and biological experiments reported conrm
that lipoplex nano-aggregates of DOPE and a gemini cationic
lipid (GCL) with two ammonium head groups separated by a
polyoxyethylene spacer ((C16Am)2(C2O)n/DOPE) efficiently
1 I. M. Verma, Science, 2013, 341, 853–855.
2 I. M. Verma and N. Somia, Nature, 1997, 389, 239–242.
3 S. Razi Sooyani, B. Baradaran, F. Lotpour, T. Kazemi and
L. Mohammadnejad, Adv. Pharm. Bull., 2013, 3, 249–255.
4650 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
4 P. L. Sinn, R. M. Anthony and P. B. McCray Jr, Hum. Mol. 28 L. Karlsson, M. C. van Eijk and O. Soderman, J. Colloid
Genet., 2011, 20, R79–R86. Interface Sci., 2002, 252, 290–296.
5 R. I. Zhdanov, O. V. Podobed and V. V. Vlassov, 29 S. De, V. K. Aswal, P. S. Goyal and S. Bhattacharya, J. Phys.
Bioelectrochemistry, 2002, 58, 53–64. Chem. B, 1998, 102, 6152–6160.
6 J. H. Felgner, R. Kumar, C. N. Sridhar, C. J. Wheeler, 30 S. Bhattacharya and A. Bajaj, J. Phys. Chem. B, 2007, 111,
Y. T. Tsai, R. Border, P. Ramsey, Martin and P. L. Felgner,
J. Biol. Chem., 1993, 269, 2550–2561.
7 C. R. Sanya, K. Ewert, A. Ahmad, H. M. Evans, U. Raviv,
2463–2472.
31 S. Asayama, A. Nohara, Y. Negishi and H. Kawakami,
Biomacromolecules, 2014, 15, 997–1101.
D. J. Needleman, A. J. Lin, N. L. Slack, C. George and 32 G. Caracciolo, D. Pozzi, H. Amenitsch and R. Caminiti,
C. E. Samuel, Philos. Trans. R. Soc., A, 2006, 364, 2573–2596. Langmuir, 2005, 21, 11582–11587.
8 G. Caracciolo and H. Amenitsch, Eur. Biophys. J., 2012, 41, 33 M. Munoz-Ubeda, S. K. Misra, A. L. Barran-Berdon, S. Data,
˜
815–829.
C. Aicart-Ramos, P. Castro-Hartmann, P. Kondaiah,
E. Junquera, S. Bhattacharya and E. Aicart,
Biomacromolecules, 2012, 13, 3926–3937.
9 S. Bhattacharya and A. Bajaj, Chem. Commun., 2009, 4632–
4656.
10 D. D. Lasic, Liposomes in Gene Delivery, CRC Press, Boca 34 A. Rodriguez-Pulido, F. Ortega, O. Llorca, E. Aicart and
Raton, FL, 1997. E. Junquera, J. Phys. Chem. B, 2008, 112, 12555–12565.
11 R. Koynova and B. Tenchov, So Matter, 2009, 5, 3187–3200. 35 M. Munoz-Ubeda, S. K. Misra, A. L. Barran-Berdon, C. Aicart-
˜
12 H. Amenitsch, G. Caracciolo, P. Foglia, V. Fuscoletti,
P. Giansanti, C. Marianecci, D. Pozzi and A. Lagana,
Colloids Surf., B, 2011, 82, 141–146.
Ramos, M. B. Sierra, J. Biswas, P. Kondaiah, E. Junquera,
S. Bhattacharya and E. Aicart, J. Am. Chem. Soc., 2011, 133,
18014–18017.
13 C. Marchini, D. Pozzi, M. Montani, C. Alfonsi, A. Amici, 36 J. Dubochet, M. Adrian, J. J. Chang, J. C. Homo, J. Lepault,
H. Amenitsch, S. C. De Sanctis and G. Caracciolo,
Langmuir, 2010, 26, 13867–13873.
A. W. McDowall and P. Schultz, Q. Rev. Biophys., 1988, 21,
129–228.
14 A. Kulkarni, R. VerHeul, K. DeFrees, C. J. Collins, 37 J. Bednar and C. L. Woodcock, in Chromatin, Academic Press
R. A. Schuldt, A. Vlahu and D. H. Thompson, Biomater.
Sci., 2013, 1, 1029–1033.
15 M. Donkuru, I. Badea, S. Wettig, R. Verrall, M. Elsabahy and
M. Foldvari, Nanomedicine, 2010, 5, 1103–1127.
Inc., San Diego, 1999.
38 J. Dubochet, B. Zuber, M. Eltsov, C. Bouchet-Marquis, A. Al-
Amoudi and F. Livolant, in Cellular Electron Microscopy,
2007, pp. 385–406.
16 M. Kapoor and D. J. Burgess, Int. J. Pharm., 2012, 432, 80–90. 39 A. Rodriguez-Pulido, E. Aicart and E. Junquera, Langmuir,
17 R. Zana, Adv. Colloid Interface Sci., 2002, 97, 205–253. 2009, 25, 4402–4411.
18 A. J. Kirby, P. Camilleri, J. Engberts, M. C. Feiters, 40 M. B. Hansen, S. E. Nielsen and K. Berg, J. Immunol. Methods,
R. J. M. Nolte, O. Soderman, M. Bergsma, P. C. Bell, 1989, 119, 203–210.
M. L. Fielden, C. L. G. Rodriguez, P. Guedat, A. Kremer, 41 A. Bajaj, P. Kondiah and S. Bhattacharya, J. Med. Chem.,
C. McGregor, C. Perrin, G. Ronsin and M. C. P. van Eijk,
Angew. Chem., Int. Ed., 2003, 42, 1448–1457.
19 F. M. Menger and J. S. Keiper, Angew. Chem., Int. Ed., 2000,
39, 1906–1920.
2007, 50, 2432–2442.
42 K. K. Ewert, H. M. Evans, N. F. Bouxsein and C. R. Sanya,
Bioconjugate Chem., 2006, 17, 877–888.
43 M. D. Kearns, Y. N. Patel and M. Savva, Chem. Phys. Lipids,
2010, 163, 755–764.
20 G. Caracciolo, D. Pozzi, R. Caminiti, G. Mancini, P. Luciani
˜
and H. Amenitsch, J. Am. Chem. Soc., 2007, 129, 10092– 44 A. L. Barran-Berdon, M. Munoz-Ubeda, C. Aicart-Ramos,
10093.
L. Perez, M. R. Infante, P. Castro-Hartmann, A. Martin-
Molina, E. Aicart and E. Junquera, So Matter, 2012, 8,
7368–7380.
21 P. Luciani, C. Bombelli, M. Colone, L. Giansanti,
S. J. Ryhanen, V. M. J. Saily, G. Mancini and
P. K. J. Kinnunen, Biomacromolecules, 2007, 8, 1999–2003.
22 S. D. Wettig, R. E. Verrall and M. Foldvari, Curr. Gene Ther.,
2008, 8, 9–23.
45 A. Rodriguez-Pulido, E. Aicart, O. Llorca and E. Junquera, J.
Phys. Chem. B, 2008, 112, 2187–2197.
46 A. Rodriguez-Pulido, A. Martin-Molina, C. Rodriguez-Beas,
O. Llorca, E. Aicart and E. Junquera, J. Phys. Chem. B, 2009,
113, 15648–15661.
23 Y. Han and Y. Wang, Phys. Chem. Chem. Phys., 2011, 13,
1939–1956.
˜
24 X. Chen, J. Wang, N. Shen, Y. Luo, L. Li, M. Liu and 47 M. Munoz-Ubeda, A. Rodriguez-Pulido, A. Nogales,
R. K. Thomas, Langmuir, 2002, 18, 6222–6228.
O. Llorca, M. Quesada-Perez, A. Martin-Molina, E. Aicart
25 A. Bendjeriou-Sedjerari, G. Derrien, C. Charnay, J. Zajac,
and E. Junquera, So Matter, 2011, 7, 5991–6004.
L. C. De Menorval and M. Lindheimer, J. Colloid Interface 48 I. Koltover, T. Salditt and C. R. Sanya, Biophys. J., 1999, 77,
Sci., 2009, 331, 281–287.
26 A. Bajaj, P. Kondaiah
Biomacromolecules, 2008, 9, 991–999.
915–924.
and
S.
Bhattacharya, 49 G. Caracciolo, D. Pozzi, A. Amici and H. Amenitsch, J. Phys.
Chem. B, 2010, 114, 2028–2032.
27 X. Wang, J. Wang, Y. Wang, H. Yan, P. Li and R. K. Thomas, 50 S. Bhattacharya and S. S. Mandal, Biochemistry, 1998, 37,
Langmuir, 2004, 20, 53–56.
7764–7777.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 4640–4652 | 4651

View Article Online
Journal of Materials Chemistry B
Paper
51 R. Zantl, L. Baicu, F. Artzner, I. Sprenger, G. Rapp and 59 G. Caracciolo, D. Pozzi, R. Caminiti and A. Congiu
¨
J. O. Radler, J. Phys. Chem. B, 1999, 103, 10300–10310.
Castellano, Eur. Phys. J. E: So Matter Biol. Phys., 2003, 10,
331–336.
60 M. Foldvari, I. Badea, S. Wettig, R. Verrall and M. Bagonluri,
J. Exp. Nanosci., 2006, 1, 165–176.
52 A. Aljaberi, M. Spelios, M. Kearns, B. Selvi and M. Savva,
Colloids Surf., B, 2007, 57, 108–117.
53 C. R. Sanya, Curr. Opin. Struct. Biol., 2001, 11, 440–448.
54 A. J. Lin, N. L. Slack, A. Ahmad, C. X. George, C. E. Samuel 61 D. Charbonneau, M. Beauregard and H. A. Tajmir-Riahi,
and C. R. Sanya, Biophys. J., 2003, 84, 3307–3316. J. Phys. Chem. B, 2009, 113, 1777–1784.
55 M. Munoz-Ubeda, A. Rodriguez-Pulido, A. Nogales, 62 D. Pozzi, V. Colapicchioni, G. Caracciolo, S. Piovesana,
˜
A. Martin-Molina, E. Aicart and E. Junquera,
Biomacromolecules, 2010, 11, 3332–3340.
A. L. Capriotti, S. Palchetti, S. De Grossi, A. Riccioli,
H. Amenitsch and A. Lagana, Nanoscale, 2014, 6, 2782–
2792.
56 A. Ahmad, H. M. Evans, K. Ewert, C. X. George, C. E. Samuel
and C. R. Sanya, J. Gene Med., 2005, 7, 739–748.
57 G. Caracciolo, C. Marchini, D. Pozzi, R. Caminiti,
H. Amenitsch, M. Montani and A. Amici, Langmuir, 2007,
23, 4498–4508.
63 A. L. Barran-Berdon, D. Pozzi, G. Caracciolo, A. L. Capriotti,
G. Caruso, C. Cavaliere, A. Riccioli, S. Palchetti and
A. Lagana, Langmuir, 2013, 29, 6485–6494.
64 G. Caracciolo, D. Pozzi, A. Capriotti, C. Cavaliere and
`
A. Lagana, J. Nanopart. Res., 2013, 15, 1–11.
58 G. Caracciolo, D. Pozzi, R. Caminiti, C. Marchini,
M. Montani, A. Amici and H. Amenitsch, J. Phys. Chem. B, 65 S. K. Misra, J. Biswas, P. Kondaiah and S. Bhattacharya, PLoS
2008, 112, 11298–11304.
One, 2013, 8, e68305.
4652 | J. Mater. Chem. B, 2014, 2, 4640–4652
This journal is © The Royal Society of Chemistry 2014