Il Farmaco 60 (2005) 701–710

http://france.elsevier.com/direct/FARMAC/

Preparation and phase behaviour of surface-active pharmaceuticals:

self-assembly of DNA and surfactants with membranes.
Differential adiabatic scanning microcalorimetric study
Erhan Süleymanoğlu a,b,*
a
Biophysics Section, Department of Physical Chemistry of Drugs, Faculty of Pharmacy, J.A. Comenius University, Odbojarov 10,
83 232 Bratislava, The Slovak Republic
b
Department of Biophysics, The Slovak Academy of Sciences, Institute of Experimental Physics, Košice, The Slovak Republic
Received 18 March 2005; accepted 3 May 2005

Available online 14 July 2005

Abstract

Some energetics issues relevant to preparation and surface characterization of zwitterionic phospholipid–DNA self-assemblies, as alter-
native models of the currently used problematic lipoplexes are presented. Nucleic acid compaction capacities of Mg2+ and N-alkyl-N,N,N-
trimetylammonium ions (CnTMA, n = 12) were compared, with regard to surface interaction with unilamellar vesicles. Differential adiabatic
scanning microcalorimetric measurements of synthetic phosphatidylcholine liposomes and calf thymus DNA and their ternary complexes
with Mg2+ and C12TMA, were employed for deduction of the thermodynamic model describing their structural transitions. Small monodis-
perce and highly stable complexes are established after precompaction of DNA with detergent, followed by addition of liposomes. In contrast,
divalent metal cation-mediated aggregation of vesicles either leads to heterogeneous multilamellar DNA–lipid arrangements, or to DNA-
induced bilayer destabilization and lipid fusion. Possible dependence of the cellular internalization and gene transfection efficiency on the
structure and physicochemical properties of DNA–Mg2+–liposomes or DNA–cationic surfactant–liposome systems is emphasized by propos-
ing the structure of their molecular self-organizations with further implications in gene transfer research.
© 2005 Elsevier SAS. All rights reserved.

Keywords: Zwitterionic liposomes–DNA self-organization; Surface active pharmaceuticals; Differential adiabatic scanning microcalorimetry; Phase behavior;
Non-viral gene delivery

1. Introduction Gene transfection is commonly used in biotechnology and

The molecular associations in mixed solutions of DNA
with oppositely charged cosolutes-metal cations, cationic
amphiphiles and macromolecules, have attracted research
interest and efforts not only in terms of their physicochemi-
cal [1–3] and pharmaceutical relevance [4–6], but also due to
their potential for applications in separation, purification and
transfection of DNA [7,8].
* Corresponding author. G.Ü.E.F., The Central Laboratory, Department
of Pharmaceutical Chemistry, Gazi Mahallesi, Polatli Caddesi, No: 115/5,
Yenimahalle, 06560 Ankara, Turkey. Tel.: +90 312 211 1947; fax: +90 312
223 5018. http://erhan.boom.ru.
E-mail address: erhan@atlas.cz (E. Süleymanoğlu).
0014-827X/$ - see front matter © 2005 Elsevier SAS. All rights reserved.
doi:10.1016/j.farmac.2005.05.010
has received considerable attention in biomedicine for curing
genetic diseases [9,10]. Therapeutic gene transfer is achieved
by employing viral [11] or non-viral [12–14], synthetic
[15–17] or physical methods [18]. Research on human gene
therapy faces certain hurdles due to the lack of suitable deliv-
ery tools for therapeutic nucleic acid transfer to target cells.
Having considered the risky viral based delivery systems and
problematic physical methods, nowadays much research effort
is devoted to the design of artificial non-viral vehicles formed
by association of nucleic acids with lipids (lipoplexes) or poly-
mers (polyplexes). Whatever the approach is, in both designs,
the objective is to achieve a stable packaging of genes to be
transfected, to increase transgene expression and to improve
their bioavailability, while decreasing their cytotoxicity. In
702 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
this respect, the desired gene packaging becomes a physical
pharmaceutics problem, requiring contributions from physi-
cochemically oriented groups.
There is an urgent need for optimized gene transfection
methods capable of protecting the DNA from degradation
via its route to gene expression [9,10]. Among these, lipid-
based delivery method, has gained preference
(http://www.wiley.co.uk/genetherapy/clinical/) in the light of
the possibilities for performing simpler quality controls, as
well as easier satisfaction of pharmaceutical requirements.
Designing suitable lipoplex formulations requires systematic
characterization of the DNA–lipid complexes as gene pack-
aging systems for transfection [19]. From lipid-based experi-
mental designs, various monolayer (Langmuir–Blodgett films)
[20], black lipid membranes (BLMs) [21] and liposomal sys-
tems [10,13,14,22–25] are studied concerning their ability to
bind nucleic acids. Using a broad range of spectroscopic,
microscopic, thermodynamic, hydrodynamic, and other rel-
evant techniques [26], a useful laboratory data have been col-
lected, in search for a details regarding biophysical and col-
loidal factors determining the stability of various DNA–lipid
self-assemblies. Since mainly electrostatic and hydrophobic
interactions govern the formation of these self-assemblies,
most of the designs are concerned with colloidal or interfa-
cial electrified surface forces. Thus, mixed systems between
lipids and surfactants can be employed as packaging agents
for DNA delivery to the affected cells.
Before switching to real in vitro transfection experiments
with numerous gene reporter molecules, it is crucial first to
achieve a stable DNA–lipid formulation with controllable
properties. Measurable parameters of potential interest are
phase behavior, morphology and structural characterization
of DNA compaction with various condensing agents, such
as, surfactants, charged and neutral polymers, metal ions, as
well as mixtures of cationic and anionic macromolecules, and
thermodynamically stable lipid vesicles. Thus, the molecular
details of cationic lipid binding to DNA polyanion will be
clarified. However, despite the collected research data, cur-
rently neither the energetics of these associations nor struc-
ture of the resultant complex is well understood.
Following the interesting recent results of using cationic,
small membrane-permeant molecules [27], as well as our own
data on Mg2+ as bridging DNA with liposomes [28] we have
focused on designs involving such agents as rapidly moving
through model cellular and nuclear membranes. These could
then bind nuclear DNA with high affinity. In the light of the
latter proposal, we hypothesized that such molecules could
affect the topology of bound DNA after associating with it so
as to enable penetration of model cell membranes by this
bound nucleic acid. Usually, the positive charge of these small
molecules would complex with the negatively charged phos-
phate groups of DNA, forming a hydrophobic ion pair. The
latter would then alter the solution properties of the com-
plexed DNA by reducing its polarity, thus giving rise to
membrane-permeant agents to trigger DNA transport through
simulated cell membranes.
To explore new mechanisms for overcoming physical
membrane barriers to the intracellular delivery of therapeutic
nucleic acids (both DNA and RNA) it is worth studying the
design of non-viral delivery tools that could form thermody-
namically stable colloidal complexes with DNA, thus obvi-
ating the need for chemical conjugation of DNA to various
ligands. Interestingly, these small cationic molecules would
be able to form membrane-permeant ion pairs with the nucleic
acid. In this respect, we have already reported our results with
divalent metal cations, capable of complexing liposomes with
DNA, followed by attachment to model cell membrane, desta-
bilizing and internalizing through it, as an alternative lipoplex
entry route to target cell via problematic receptor-mediated
endocytic route. In our opinion, such membrane fluidity
effects of various small cations deserve to be studied with
respect to gene delivery vesicle designs, which potentially
could enhance their transfection efficiencies, as suggested pre-
viously [28]. Hence, it is interesting to compare our previous
results on divalent metal cations with other promising mol-
ecules, such as cationic surfactants for instance, which are
well-known gene transfection enhancers [9]. Experimental
designs of this sort focus on features such as influence of
vesicle size and dose, surface charge and properties, and sta-
bility characteristics, which are to a major extent, thermody-
namically governed processes. Unfortunately, little has been
done so far to relate these physicochemical considerations to
in vivo behavior of these dispersions.
To understand the energetics of this important system, the
thermodynamics of lipid and amphiphile-like ligand binding
to DNA was studied and some preliminary results are pre-
sented here. Both naturally occurring lipids and amphiphile
resembling model compounds were used, trying to compare
the phase behavior of various ligands and their binding modes
and their effect on the energetics of DNA–lipid complex for-
mation. Assuming that alkylammonium ion and related forms
form a relevant structure, their phase paramaters were com-
pared with 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine
(DPPC) concerning affinity for DNA binding, trying to find
more clues from the correlation between DNA compaction
and transfection efficiency obtained from previous studies of
DNA associations with lipid dispersions and polycations with
different chain length.
Deducing from both theoretical and experimental studies
will improve the current knowledge concerning molecular
interactions and general biophysical chemistry of these prom-
ising formulations in terms of designing improved gene deliv-
ery systems.
2. Experimental
2.1. Materials
DPPC, SSC (1.5 × 10–4 mol/l Na-citrate, 1.5 × 10–3 mol/l
NaCl, pH 7.2) reagents were purchased from Sigma (St. Louis,
MO, USA) and used without further purification. All other
 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
reagents were of analytical grade. Alkylamine aqueous solu-
tions were stored in tightly sealed containers, to prevent their
reactions with atmospheric CO2, as suggested [29]. Lipid and
C12TMA solutions were mixed in polyethylene vials in
desired ratios. Following solvent evaporation under nitrogen
gas flow, the samples were evacuated at room temperature
for couple of hours. Cationic detergent was dissolved in
double distilled water and was added to the dry lipid, prior to
measurement.
2.2. Methods
2.2.1. Preparation of polynucleotide solutions
and concentration determinations
Calf thymus DNA with MW of 8.6 MDa (= 13 kb), 42%
GC, Tm = 87 °C, ~20 A260 units per mg DNA was used. Poli-
nucleotide concentrations and molar ratios are based on the
average nucleotide molecular weight of 308 [28].
2.2.2. Preparation of liposomes
Chromatographic tests for purity of the lipids were not per-
formed, however, the purity of the lipid preparation was
assured from the half-widths of their main phase transitions.
1.2 mM lipid in standard SSC buffer, pH 7.2 was used in all
experiments and was stored at 4 °C.
The formation of a thin layer of lipids of a 15 ml round-
bottomed flask was achieved by a hand-shaking and hydra-
tion in particular buffer at around 70 °C. Vortexing of the
lipid with the desired aqueous solution above the gel-to-
liquid crystalline phase transition of the lipid (Tm) for around
30 min resulted in multilamellar vesicles.
Unilamellar vesicles (ULV) were obtained by extrusion of
multilamellar vesicle (MLV) suspension through two stacked
polycarbonate filters (Nucleopore, Inc.) of 100 nm pore size
at around 60 °C. Repeated extrusion (10 times) through the
extruder (Lipex Biomembranes, Inc., Vancouver, BC, Canada)
created homogeneous vesicle suspension. This allowed the
preparation of vesicles with a mean diameter of 90 nm and a
trap volume in the range of 1.5–2.0 l/mol.
2.2.3. Preparation of liposome–nucleic acid mixtures
Nucleic acid–lipid mixtures were prepared 1 h before
microcalorimetric measurements by vigorous mixing of either
phosphatidylcholine MLV or ULV dispersions and solvent,
varying nucleic acid concentration and keeping DPPC con-
centration fixed. Control experiments of DNA–lipids in the
absence of detergent or divalent cations, were performed in
parallel. Phospholipid concentration employed was
0.3 mg/ml.
The lipid samples were hydrated, as described above to
form first MLV. The preparation of phosphatidylcholine ULV–
calf thymus DNA complexes, was the same as in the case of
MLVs, i.e. by mixing DNA solution with aqueous DPPC dis-
persion in the presence of Mg2+.
The DNA concentration used throughout all experiments
was 1.8 mM based on the above-mentioned assumption. A
703
freeze–thaw protocol was followed to ensure equal distribu-
tion of solutes between lamellae and adequate hydration of
the lipids.
Comparison with the case of liposomal preparations with-
out employing freeze–thaw procedure showed no difference
in terms of homogeneity of the suspension. This was done by
placing the sample in a cryo-tube and freezing it, in liquid
nitrogen for around 30 s. The cryo-tube was subsequently
removed and was plunged into warm water (~60 °C). When
the sample was thawed, the whole cycle of freeze–thawing
was repeated six times.
2.2.4. Differential scanning calorimetry
Calorimetric measurements were performed using Privalov
type high sensitivity differential adiabatic scanning micro-
calorimeter DASM-4 (Biopribor, Pushchino, Russian Federa-
tion) with sensitivity higher than 4 × 10–6 cal/K and a noise
level less than 5 × 10–7 W. Heating runs were performed with
a scan rate of 0.5 K/min. The temperature at the maximum of
the excess heat capacity curve was taken as the transition tem-
perature Tm and the transition width DT1/2 was determined at
the transition half-height. The calorimetric enthalpy DHcal of
the transition was determined as the area under the excess
heat capacity curve [28].
3. Results and discussion
The presented work describes preliminary microcalorimet-
ric measurements on designing and exploiting of surface
active molecular assemblies of DNA, liposomes and N-alkyl-
N,N,N-trimetylammonium ions (CnTMA, n = 12) and Mg2+,
as a model supramolecular pharmaceutical formulation com-
pacted for therapeutic gene transfection, or alternatively for
use in DNA chromatography. Following the relevant pro-
posal to mimic the way of action of much superior viruses, as
compared to non-viral DNA carriers, for improvement of the
transfection efficiencies of therapeutic gene delivery vehicles,
our efforts concentrated on designing such tools, acting via
membrane destabilization and fusion pathway instead of
receptor-mediated route [30]. The employment of more natu-
rally encountered surfaces is emphasized, as described in our
recent study [28]. The advantages of using liposomes as com-
pared to other lipid-containing drug carrier systems are well
established [31]. Having considered the cytotoxicity prob-
lems of the currently employed cationic lipids, neutral phos-
pholipid vesicles deserve to be studied in more detail as prom-
ising alternatives. Our current research focuses on a nanoscale
complex formations between zwitterionic liposomes and
DNA, mediated by various inorganic cations, acting as con-
densing agents. Our motivation for such an experimental
design has come from recent reports on positive effects of
divalent cations, such as Ca2+ and Mg2+ on transfection effi-
ciency [32,33].
The phase behavior of complex formed between neutral
lipid and calf thymus DNA in the presence of Mg2+ is pre-
704 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
sented. N-alkyl-N,N,N-trimetylammonium ions were chosen
as an amphiphilic system, whose surface properties are to be
compared with Mg2+ with regard to ternary complex forma-
tion with liposomes and ability to compact DNA. Mg2+ is
tested as a compaction agent in the presence of phospholipid
vesicle suspensions, in the light of its role in numerous cel-
lular events. Its high intracellular concentrations, as well as
its well-known property of phosphate group transfer [34]
makes it preferred natural divalent inorganic cation in com-
parison with commonly used synthetic polymers, which offer
system versatility and large selection of polymer species, but
are not encountered in biointerfaces. Ion transport, for
instance, takes place by metal binding to cell membranes.
Neutral phospholipids are interesting research subject not only
in terms of the abovementioned preference of liposome gene
delivery designers, but also concerning fundamental cell biol-
ogy events. Since phosphatidylcholine moeity is a major con-
stituent of the total phospholipid bilayer content, it is useful
to study its interactions with various metals. Hence, the sug-
gested phosphatidylcholine–Mg2+ binary mixture would then
give further data on biological implications of metal ion con-
trol of cell membrane fluidity. N-alkyl-N,N,N-trimetyl-
ammonium ions (C12TMA) ions were selected due to their
interesting but yet insufficiently studied effects as membrane
destabilizing and lipid penetrating agent. Provided its inter-
actions with model membranes are characterized suffi-
ciently, a synthesis of new relevant quaternary bisammonium
Fig. 1. Thermotropic phase behavior of calf thymus DNA in complex with
lipid in the presence of C12TMA and Mg2+, respectively. (1) DPPC–MLV;
(2) equimolar mixture of DPPC–ULV with DNA: (3) DPPC–ULV–DNA
complex by Mg2+ in 1:3:1 ratio; (4) DPPC–ULV–DNA complex in the pre-
sence of C12TMA; (5) DPPC–ULV–DNA mixture in the presence of C12TMA
after heating of the sample; (6) DNA–Mg2+ binary mixture.
Table 1
Thermodynamics of DPPC–vesicle binding to calf thymus DNA in the presence of Mg2+ a
DNA–lipid mixtures Tp Tm
DPPC–MLV 36.05 41.9
DPPC–ULV–Mg2+–DNA (1:1:1 ratio) – 41.7
DPPC–ULV–Mg2+–DNA (1:3:1 ratio) – 41.75
a
Represents a mean of six independent DSC measurements. Each measurement was performed as described in Sections 2.1 and 2.2. r is estimated as:
r = DH/DHcal.
compounds and further studies on their effects on cell sur-
faces with biomedical profits will be stimulated [35].
Fig. 1 depicts DSC heating scans of DPPC vesicles and
their ternary complexes with calf thymus DNA in the pres-
ence of either Mg2+ or C12TMA. The first curve (1) is a cali-
bration mark starting with a typical DPPC multilayer phase
transitions, with a pre-transition temperature peak at around
36 °C with DHcal of 3.9 kJ/mol and the gel-to-liquid crystal
temperature at 41.9 °C (Table 1). The determined values are
in a good agreement with those reported previously in the
lipid database (LIPIDAT): http://lipidat.chemistry.ohio-
state.edu. The signal after the lipid phase peaks is a base line
with calibration mark (50 µW, DT = 4), obtained during the
filling of both calorimetric cells with solvent. The next curves
show the change in phase behavior of extruded unilamellar
lipids upon their reaction with DNA in the presence of Mg2+
(2) and (3) in various ratios or C12TMA (4) and (5), respec-
tively. Compared with pure DPPC, triple complexes of DPPC–
ULV with DNA and Mg2+ in equimolar ratios possess broader
lipid peak with decreased maximum. The pre-transition dis-
appears (Fig. 1 (2, 3) and Table 1). Similar thermogram is
obtained for this ternary complex in ratio of 1:3:1, however,
with a shift of the second lipid–DNA phase to lower tempera-
ture at around 70 °C. The last endotherm belonging to free
nucleic acid melting depicts a difference with that reported
for plasmids, which show peaks at 60 °C attributed to linear
and open-circle plasmid DNA. Another peak of such plasmid
DNA phase is usually larger, seen at 80 °C and corresponds
to supercoiled form of the plasmid [36]. This DSC scan of
ternary complexation was preceded first by DPPC–ULV peak
at 41.9 °C, by a second peak belonging to DPPC–DNA
equimolar binary complex at 51.3 °C and a third minor peak
of unbound DNA appearing at ~60 °C. This is a resolution
profile of the thermograms (2) and (3), which results only
after addition of Mg2+.
The observed peak distribution indicates that the fraction
of liposome-free DNA is less encountered than the bound
DNA in the lipoplex. The liposome–DNA association results
in the decrease of the DCp. Interestingly, no any DSC signal
was detected for Mg2+–DPPC mixture (data not shown),
which was observed previously turbidimetrically [28]. Addi-
tion of C12TMA to DPPC–ULV–DNA mixture results in
broadening of the thermograms leading to difficulties of esti-
mations of its onset point, peak top, and hence the endother-
mic effect. For these reasons, the quantitative evaluation of
the surfactant effects should be done with precautions and
are, therefore, not represented in Table 1. On the other hand,
their properties reported here with respect to DNA–liposome
DT1/2 DHcal DHvH R
0.6 31.9 5501 172
2.35 10.8 1402 129
1.65 9.7 1998 206
 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
condensation is compared with that of Mg2+. This is done
solely for testing the electrostatics and hydrophobicity rea-
soning with regard to the thermodynamics of phospholipid
binding to DNA followed by its compaction. Moreover, DSC
is a very sensitive way of studying certain changes in bilayer
packing, with the presence of units interfering with chain
packing in the bilayer causing a decrease in the temperature
of the main transition [37]. The incorporation of a substance
in a liposome bilayer has a more profound effect on the lipid
phase transitions peaks. The shift or disappearance of pre-
transition or main phase transition can thus be used as an
informative indicator of the existence and level of included
materials in the leaflets. Thus, the entrapment of molecules
in liposomes can be quantified by determining the change in
temperature of the onset of the transition, as well as by mea-
suring of the peak temperature as shifts in the transition tem-
perature, rather than presenting them as abstract values
(Table 1).
The DSC scan (4) of Fig. 2, represents a thermogram of a
complex formed between DPPC and DNA in the presence of
C12TMA ions. It is interesting to study assemblies of this type,
since surfactants are used for extraction of proteins from phos-
pholipid membranes [38], which could also appear as nucleic
acid–protein complexes, specifically or non-specifically
attached to cell membrane fractions. These interactions can
be approached by thermodynamic measurements. It is worth
investigating their binding characteristics, as drugs usually
form supramolecular complexes with proteins, membrane
phospholipids and nucleic acids (DNA and RNA), and their
subsequent release depends on the strength of binding and on
the reversibility of this interaction. The evaluation of physi-
cochemical stability of the samples with respect to tempera-
ture variations is crucial due to their thermodynamic profiles
in terms of manufacturing issues, as recommended also for
other relevant multiple emulsions [39]. Moreover, since the
enthalpy of binding is a temperature dependent event [29],
the effect of heating of the sample is emphasized (Fig. 1 (4)
Fig. 2. Phase behavior of the control samples used: (1) DNA–C12TMA binary
mixture; (2) DNA in concentration used as in Sections 2.1 and 2.2; (3) DNA
in doubled concentration; (4) DPPC–ULV–C12TMA binary mixture.
705
vs. (5). The single melting peak corresponding to the ternary
complex formed between DPPC–C12TMA–DNA appears at
approximately 33 °C. The melting behavior of the control
sample DNA–C12TMA binary mixture (Fig. 2 (1)) is com-
posed of two separate phases. The first situated at 47 °C
belongs to DNA phase, the separate melting behavior of which
is shown with two different concentrations on Fig. 2 (2) and
(3), respectively. The second phase is a structural transition
seen after 85 °C and continues until 100 °C, corresponding to
complex of DNA bound to C12TMA. This type of melting
behavior indicates the occurrence of aggregation reaction
between the latter molecules. While, large enthalpy varia-
tions with temperature are compensated by the hydrophobic
component of entropy, it is possible to estimate electrostatic
and hydrophobic contributions to the enthalpy at various tem-
peratures by applying additivity, as recently shown for alky-
lammonium binding to DNA by isothermal titration calorim-
etry [29]. Fig. 2 (2) is a DNA sample, prepared as described
in Sections 2.1 and 2.2, while (3) is the same DNA sample in
doubled concentration. These two thermograms show a typi-
cal melting behavior of DNA and its synthetic models, which
in calorimetric determinations appear as a single peak. This
highly cooperative process represents unwinding of the double
stranded structure into two polynucleotide strands, which fold
into separate chaotic globules. The equality of the heat capaci-
ties of the native and denatured DNA states is a typical fea-
ture of this process, which makes quantitative determination
of thermal effects of melting, as well as free energy DG of
stabilization of native DNA structure easier. This kind of struc-
tural transitions of DNA (Fig. 2 (2) and (3) are different from
those of lipid phase behavior (Fig. 1 (1)). Lipids undergo
another type of structural transitions, or gel-to-liquid crystal-
line transitions, which is also their main phase transition. In
this respect, these are different from the abovementioned
events in that in phospholipids the degree of cooperativity of
their main phase transitions entirely depends not on the inner
molecular but on interactions between molecules. For such
systems, the effective Van’t Hoff enthalpy is greater than calo-
rimetrically determined. In general, the degree of cooperat-
ivity of the is given as: DHvH = NDH, where N is the number
of molecules in the cooperativity unit. Testing this kind of
dependence on DNA concentration is crucial, as also shown
by another relevant thermodynamic study [29]. In agreement
with this reminding and with the mathematical model pre-
sented in the latter study, the prediction is that the position of
the binding peak would be shifted upon changing the DNA
concentration. Therefore, it is important to perform micro-
calorimetric scans at several nucleic acid concentrations at
previously prepared DNA–ligand ratios. This is also seen for
DPPC–Mg2+–DNA ternary complex in equimolar and
1:3:1 ratios, respectively (Fig. 1. (2) and (3)). DNA sample
examined in two different concentrations already shows simi-
lar melting behavior, but with various peaks (Fig. 2 (2) and
(3)), which supports the expectation of dependence of the
results on nucleic acid concentration used. The lower DNA
concentration in lipoplexes results in a little bit prolonged
706 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
second peak, as compared with that of higher DNA amount
(Fig. 1 (2) vs. (3)), as also indicated by the experimental
enthalpies. However, to relate this effect to transfection effi-
ciency of these lipoplexes still remains to be elucidated. Inter-
estingly, this effect is less seen in the case of C12TMA–DNA
binary mixtures, where the most reproducible results were
obtained for their ratio of 1–1.8. In this case, the effect of
heating and aggregation is more profound (Fig. 1 (4) and (5)
and Fig. 2 (1)). Such dependence of the electrostatic compo-
nent of binding Gibbs free energy and entropy on reactant
concentration suggests the existence of an aggregation reac-
tion, which was also observed in DNA–lipid complexes
[28,29].
The Mg2+-ions at the equimolar amounts with DNA
increases the Tm value by 33.7 °C, due to Mg2+-induced
duplex stabilization. ULVs treated with the same Mg2+ con-
centration did not produce such a shift, which is normally
detected spectroscopically [28]. Mg2+ induces the formation
of substantial amount of circular DNA, suggesting that Mg2+
cations stabilize the interaction of polynucleotide cohesive
ends, the effect being dependent on the concentration of
MgCl2 and possibly being a sequence-specific event [40]. The
formed circular molecules are stabilized by Mg2+, but are not
covalently closed. Although, Mg2+ stabilizes end-to-end inter-
actions, it is likely that a dynamic equilibrium exist between
linear and circular fragments.
C12TMA ions cause a decrease in lipid ordering by exert-
ing a perturbation effect on the lipid bilayer structure. Trim-
ethylammonium ions possess an intermediary position among
CnTMA homologues, in terms of lipid ordering, as demon-
strated by recent ESR studies [41]. This is characteristic bipha-
sic behavior of these alkylammonium series, similar to other
amphiphilic compounds, suggesting that the interaction of
CnTMA molecules with lipid membranes has important bio-
logical implications. Moreover, other relevant ions, such as
bis-quaternary ammonium ions (bis-A2+), e.g. alkane-bis-
alkylammonium ion, which is a divalent organic ion with two
positively charged sites connected by a spacer chain and hav-
ing also six side chains, has attracted research interest not
only due to its biological properties as ion channel blocker of
acetylcholine receptors, but also as a model compound for
studying the cation effects on hydration of hydrophobic cat-
ions. In this respect, it is argued that if its hydrophobicity
could be controlled precisely, it could have had a more sig-
nificant value in a two-phase system, such as a membrane,
indicating the importance of evaluating various alkylammo-
nium species with different structures concerning this fea-
ture. C12TMA ions possess a hydrophilic positively charged
ammonium group and a hydrophobic alkyl chain of 12 –CH2
groups. Due to this amphiphilic character these ions partition
between aqueous and lipid phases. However, the observed
cut-off effect of the whole series of alkylammonium ions in
decreasing lipid order cannot be the result only of partition
equilibrium between these two phases. This is also a concen-
tration dependent effect possibly caused by lateral bilayer
expansion due to positioning of such amphiphilic ions in the
region of the lipid leaflet, exerting an additional effect of
amphiphile partitioning between aqueous and lipid phases
[41]. Hydrophobically driven insertion of C12TMA ions into
the phosphatidylcholine lipid layer brings about attractive
electrostatic forces for DNA polyanion. In this respect,
C12TMA ions affect the phospholipid vesicle in a similar way
of Mg2+ adsorption onto their surface (Fig. 3) creating a posi-
tive liposomal interface, which can also be detected by their
electrophoretic mobility measurements [42].
The structure of the interacting biopolymers in lipoplex
formulations is a matter of controversy. Research evidence
indicates that both DNA [43] and lipids [44] affects each oth-
er’s structural transitions during complex formation. Most of
the interpretations for this self-assembly in terms of well-
established polyelectrolyte theories for interactions between
oppositely charged macromolecules are an oversimplified
view of the real structures.
It is likely that structures similar to cationic lipid–DNA
complexes are formed [35,45]. In our opinion, kinetic vs. ther-
modynamic stability features govern this self-association in
more complex way. Within this respect, the model considers
the lipid–DNA structures as overlaying layers of DNA
adsorbed onto lipid bilayers, after charge neutralization [28].
The process is governed by adsorbed cations (Me2+) on the
surface of the zwitterionic lipids. The presence of such mul-
tilayers of alternating lipid–DNA assemblies is due to forma-
tion of condensed DNA as parallel arrangements between the
lipid bilayers [46]. This is expected, due to the existence of
3-D correlation forces between the DNA-covered lipid lay-
ers, following DNA-driven formation of multilamellar lipo-
somes from ULVs. Based on our own polyelectrolyte data, as
well as on relevant measurements reported in the literature,
Fig. 3 represents a proposal of the possible structures of
C12TMA–DNA–liposomes and DNA–Mg2+-neutral lipo-
somes ternary complexes.
Fig. 3a suggests the possible structure of DNA–C12TMA–
liposome complexes. Under the employed conditions, the ini-
tially relaxed DNA in solution is complexed by surfactant
molecules, which adsorb on the nucleic acid surface forming
a, micelle-like domains. C12TMA molecules also partition in
lipid bilayer forming a swollen mixed bilayers. The latter can
also lead to subsequent humpbacked vesicles with surfactant
at regions of high curvature. Afterwards, the DNA in unfolded
form is apparently adsorbed on the surface of surfactant–
DPPC vesicles, as also suggested by a fluorescent study on
neutral lipids, employing cationic surfactant [47]. Since sur-
factant molecules become incorporated into the liposome
bilayer due to the partition equilibrium between bilayer and
aqueous phase, normally the binding of C12TMA–DNA com-
plexes to the vesicle surface through hydrophobic forces
results in opening of the micelle-like domains and partition-
ing of C12TMA ions in the lipid bilayer. Hence, employment
of cationic surfactant tend to form a fully relaxed DNA, which
is bound with significant stability to plasma membranes,
resulting in a difficult to internalize in a cell structure by
endocytosis. In contrast, unilamellar phosphatidylcholine
 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710 707

Fig. 3. Proposed structures of neutral liposome–detergent–DNA (a) and DPPC–Mg2+–neutral liposome (b) self-assemblies. Me2+ refers to inorganic divalent
metal cations (modified version of the structures proposed originally by Clamme et al. [47] and Kuvitchkin and Sukhomudrenko [54]).

vesicles interact with DNA via a mechanism of nucleic acid
helix-induced liposome aggregation and subsequent lipid
fusion (Fig. 3b). A liposome with higher curvature could be
formed, which is advantageous for further increasing the gene
transfection efficacy. Thermotropic phase behavior reported
in Figs. 1 and 2, probably represent several intermediary
DNA–lipid structures [22]. Usually, several MLVs aggregate
around a Mg2+-induced compacted DNA as grapes. ULVs
follow DNA-induced fusion resulting in generation of lipo-
some with preferable higher curvature. Even though the sug-
gested structure remains to be confirmed by further structural
analyses, such a local DNA unwinding is likely to occur,
because of the spectroscopic evidence for significant disrup-
tion of the planar interactions between the bases [47,48].
The process is hydrophobically and electrostatically con-
trolled, since liposomes aggregate and fuse, in the presence
of oppositely charged particles [49]. In the currently pre-
sented system, zwitterionic liposomes fuse in the presence of
Mg2+. Hence, polyanions like DNA play an active role in
adhesion, aggregation and fusion, by bridging two liposomes
in close contact, with surface adsorbed metal cation-induced
fusion. Since, a variety of possibilities exist regarding the
structure of liposome–DNA formulations, it seems that,
besides contributions due to charge neutralizations or rela-
tive lipid/DNA ratios, the absolute concentrations of the
engaged system components play an additive role in thermo-
dynamically preferred lipoplex structure formation.
The major problem with nucleic acid delivery employing
liposomes as a therapeutic gene vector is efficient escape from
the clearance by the reticuloendotelial system (RES)
[8,9,30,50,51]. The immune response is due to accumulation
of serum proteins on the liposome surface, resulting in rapid
removal primarily by liver, spleen, lymph nodes, and lung
tissue, the process being referred to as opsonization. Usually
neutral multilamellar and unilamellar vesicles follow a slower
708 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710

clearance rate than negatively and positively charged MLVs. ating a more hydrophobic DNA, then they could facilitate
ULVs are characterized with longer residence time than MLVs surpassing the membrane barriers that currently limit DNA
[9]. delivery into the nucleus by non-viral vectors, remains to be
The unavoidable hurdle until now has been the final fate tested.
of liposome, which is engulfed by macrophages of the RES, There are several potential ways, though which metal ions
via the lysosomal participation. Such fusion with lysosomes can increase gene transfection efficiency. Thus, their ability
leads to destruction of liposomes through the action of phos- to partition rapidly through cell membranes entering the
pholipases with subsequent release of vesicle’s content. nucleus may confer novel intracellular trafficking pathways
Hence, efficient strategies to overcome opsonization of lipid on complexed DNA. Fig. 4 shows the simplified possibility
vesicles have been designed, thus developing the concept of of the proposed DNA-mediated liposome fusion with target
liposome targeting to specific target cells and tissues. cell membranes. The suggested structure of the entrapped
Initially, conjugation of various polymers and other mac- DNA [28] is based on Fig. 3b, originally described by
romolecules for this purpose was highly cell type-dependent Kuvitchkin and Suchomudrenko (1987) [54]. The model
and ended with discouraging results. Therefore, besides the describes the aggregation of several vesicles resulting in fusion
required knowledge on the influence of in vivo environment of ULVs, induced by polynucleotide chain unwinding. Thus,
on both vesicle leakage and clearance rates, searching for the desired highly fusogenic vesicle with higher curvature is
improvement of the specificity of cell and tissue recognition formed. Mg2+ used, bridge the DNA polyanion to charge
becomes essential. The objective is to deduce alternative route reversed liposomes accelerating further their membrane desta-
of successful gene delivery avoiding the receptor-mediated bilizing properties.
intake. With this respect, we have worked on electrostatic and We hypothesized that based on these features of Mg2+ as a
hydrophobic control of membrane destabilization. Despite the surface active compound, similarly to cationic peptides, an
fact that often the fluidity of the phospholipid surface is inorganic cation-mediated cell membrane destabilization
described as disordered phase, a certain level of interfacial could occur, governed by electrostatic and hydrophobic forces.
organization or lateral heterogeneity is present due to multi- As shown on Fig. 4, Mg2+ bring the neutral lipid vesicles
lamellar nature of the vesicles with high degree of cooperat- with the encapsulated nucleic acid in close proximity with
ivity. Hence, knowledge of its surface structure is crucial for negatively charged target cell surface. When optimal combi-
designing efficient lipid-based gene delivery systems and sug- nation of surface factors needed for bilayer destabilization is
gesting new clues on their interactions with cellular surfaces,
as well as internalization mechanisms of the entrapped thera-
peutic DNA through membranes. Although increased sur-
face rigidity prolongs the circulatory half-time of lipid
vehicles, the commonly employed phospholipids in drug
delivery are at physiological temperature entirely or partly in
the fluid state [52]. For proper understanding of the relation-
ships between surface heterogeneity and lipid-carrier func-
tion it is important to distinguish between various levels of
lipid lateral organization differing in their lifetimes and sizes.
Lipid heterogeneity is a consequence of “solid–liquid” or “liq-
uid–liquid” phase separation, as well as by thermal density
fluctuations leading to the appearance of short lived gel-like
microdomains in a fluid (liquid-crystalline) environment
(Fig. 3). Their appearance implies function of low-ordered
boundary regions, which in spite of their short lifetime and
small size serve as sites of ion penetration and enzymatic reac-
tions. They also may serve as nucleation centers preceding
gross lipid phase separation induced by cations and are
believed to play a role in fusion of the lipid carriers with a
plasma membrane of living cells [52].
Numerous suggestions have been proposed for overcom-
ing the membraneous barriers, which significantly inhibit the
entry of therapeutic DNA to the nucleus. Certain cationic
amphiphiles, such as those presented, able of free partition-
ing through cell membranes and rapidly localizing to specific
nuclear chromosomal regions are being studied for their car-
rier potential [53]. The proposal that if such nucleic acid bind- Fig. 4. Proposed mechanism of Mg2+-induced cellular internalization of neu-
ing agents express their membrane transport properties cre- tral liposome–DNA formulation.
 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
reached, coupled with suitable amphiphilicity, the DNA-
mediated fusion between zwitterionic liposomes and cell
membrane occurs. This proposal does not, however, rule out
the possibility of existence of another cellular internalization
mechanism of neutral liposome–DNA complexes. Neverthe-
less, the described ability of free metal cations for both DNA
and lipid binding, partitioning and permeating through cell
membranes entering the nucleus [31,32], coupled with accel-
erated nuclear delivery of DNA, supported by helix-binding
molecules [27], is consistent with the hypothesis that diva-
lent metal cations can confer membrane-permeant properties
on complexed DNA. The exact mechanism of nucleic acid-
exerted phospholipid fusion in the presence of metal cations
is unknown. One possibility could be, that similarly to cat-
ionic peptides [30], the electrostatic screening of the hydra-
tion shell of inorganic cations compensates for the inter-
vesicle repulsive forces. Once the opposing membranes are
in close contact zone, the hydrophobic and electrostatic fea-
tures of metal cation form a complex between liposome and
cell phospholipids followed by lipid mixing (Fig. 4). Our pro-
posal is further supported by recent results of Sato et al. (2003)
[55] on Mg2+-induced DNA attachment to phospholipids. In
addition, closely relevant work of describes DNA–Mg2+
nanoscale mixtures as potential non-viral vector for gene
delivery [56].
Besides their membrane permeating properties, Mg2+-
ions employed could also increase the transfection efficiency
in other respects. Nuclear stability and topology of genomic
DNA seems to be one cellular mechanism for controlling gene
transcription [27]. Metal cations may target complexed DNA
via nuclear localization signals (NLS) to transcriptionally
active chromatin regions. Thus, gene carrier-mediated nuclear
targeting could represent a way of enhancing the gene expres-
sion levels. This suggestion is in agreement with recent reports
on Ca2+ and Mg2+-induced increase of expression of the trans-
fected genes [31,32]. Metal cations can govern the conforma-
tion of the transfected DNA, showing the role of nucleic acid
topology control in gene transfection. Employment of natu-
rally occurring divalent metal cations as oppose to other syn-
thetic and potentially mutagenic molecules [27], deserves to
be studied in more detail.
4. Conclusions
Neutral liposome–Mg2+–DNA formulations described here
are highly dynamic structures, frequently changing their trans-
fection characteristics, depending on particular laboratory pro-
tocol. On the other hand, these ternary complexes appear to
be a reproducible lipoplex design, but their serum stability
profiles are not yet well defined. While promising results are
reported recently on their use [57], more immunological stud-
ies are needed for their in vivo evaluation. Nevertheless, the
concept of metal-based pharmaceuticals [58], undertaken
here, will open new insights for studying whether these
supramolecular complexes follow the similar principles of
709
binding to cellular receptors and will further define the issue
of nucleic acid receptors on cell surfaces.
Acknowledgments
I thank Professor R.I. Zhdanov (V.N. Orekhovich Institute
of Biomedical Chemistry, Russian Academy of Medical Sci-
ences) for introducing me to the topic of nucleic acid–
membrane interactions and for his help as my postgraduate
supervisor. The hospitality of Professor P. Bálgavy (J.A.
Comenius University-Bratislava) and Dr. J. Bagelova (Insti-
tute of Experimental Physics, The Slovak Academy of
Sciences-Košice) is greatly acknowledged.
References
[1] R. Chang, Physical Chemistry for the Chemical and Biological Sci-
ences, Chapter 16, Intermolecular Forces, University Science Books,
Sausalito, California, 2000 (pp. 669–700).
[2] D. Leckband, J. Israelachvili, Intermolecular forces in biology, Q.
Rev. Biophys. 34 (2001) 105–267.
[3] J.B.F.N. Engberts, M.J. Blandamer, Understanding organic reactions
in water: from hydrophobic encounters to surfactant aggregates,
Chem. Commun. (2001) 1701–1708.
[4] A.T. Florence, D. Attwood, Physicochemical Principles of Pharmacy,
third ed, Palgrave, 1998.
[5] E.H. Kerns, L. Di, Physicochemical profiling: overview of the
screens, Drug Discovery Today, Technologies 1 4 (2004) 343–348.
[6] P.J. Crowley, L.G. Martini, Formulation design: new drugs from old,
Drug Discovery Today, Therapeutic Strategies 1 4 (2004) 537–542.
[7] A. Simmonds, M. Cunningham, The preparation of protein–nucleic
acid conjugates, in: M. Aslam, A. Dent (Eds.), Bioconjugation: Pro-
tein Coupling Techniques for the Biomedical Sciences, Macmillan
Reference Ltd, 1998, pp. 483–503.
[8] P.-G. de Gennes, Problems of DNA entry into a cell, Physica A
(Amsterdam) 274 (1999) 1–7.
[9] A.J. Kirby, P. Camilleri, J.B.F.N. Engberts, M.C. Feiters,
R.J.M. Nolte, O. Söderman, M. Bergsma, P.C. Bell, M.L. Fielden,
P. García Rodríguez, A. Guédat, C. Kremer, C. McGregor, G. Perrin,
M.C.P. Ronsin, van Eijk, Gemini surfactants: new synthetic vectors
for gene transfection, Angew. Chem. Int. Ed. Engl. 42 (2003) 1448–
1457.
[10] N.S. Templeton, Gene and Cell Therapy-Therapeutic Mechanisms
and Strategies, second ed, Marcel Dekker, 2003.
[11] N.A. Kootstra, I.M. Verma, Gene therapy with viral vectors, Annu.
Rev. Pharmacol. Toxicol. 43 (2003) 413–439.
[12] R. Dass, Cytotoxicity issues pertinent to lipoplex-mediated gene
therapy in-vivo, J. Pharm. Pharmacol. 54 (2002) 593–601.
[13] A.D. Miller, The problem with cationic liposome/micelle-based non-
viral vector systems for gene therapy, Curr. Med. Chem. 10 (2003)
1195–1211.
[14] A.D. Miller, Nonviral liposomes, Methods Mol. Med. 90 (2004)
107–138.
[15] A.V. Kabanov, V.A. Kabanov, DNA complexes with polycations for
the delivery of genetic material into cells, Bioconjug. Chem. 6 (1)
(1995) 7–20 (Jan–Feb).
[16] C.L. Gebhart, A.V. Kabanov, Evaluation of polyplexes as gene trans-
fer agents, J. Contr. Releas. 73 (2001) 401–416.
[17] K.B. Anwer, G. Rhee, S.K. Mendiratta, Recent progress in polymeric
gene delivery systems, Crit. Rev. Ther. Drug Carrier Syst. 20 (4)
(2003) 249–293.
710 E. Süleymanoğlu / Il Farmaco 60 (2005) 701–710
[18] S. Mehier-Humbert, R.H. Guy, Physical methods for gene transfer:
improving the kinetics of gene delivery into cells, Adv. Drug Deliv.
Rev. 57 (5) (2005) 733–753.
[19] M.G. Miguel, A.A.C.C. Pais, R.S. Dias, C. Leal, M. Rosa, B. Lind-
man, DNA–cationic amphiphile interactions, Colloids Surf. A: Physi-
cochem. Eng. Aspects 228 (2003) 43–55.
[20] T. Kunitake, Self-assemblies of biomembrane mimics, in: A. Baskin,
W. Norde (Eds.), Physical Chemistry of Biological Interfaces, Marcel
Decker, Inc., New York, Basel, 2000, pp. 283–305.
[21] T. Hianic, A. Labajova, Electrostriction of supported lipid films at
presence of cationic surfactants, surfactant–DNA and DNA–Mg2+
complexes, Bioelectrochemistry 58 (2002) 97–105.
[22] F. Liu, L. Huang, Development of non-viral vectors for systemic gene
delivery, J. Contr. Release 78 (2002) 259–266.
[23] N.S. Templeton, Developments in liposomal gene delivery systems,
Expert Opin. Biol. Ther. 1 (2001) 1–4.
[24] N.S. Templeton, Liposomal delivery of nucleic acids in vivo, DNA
Cell Biol. 12 (2002) 857–867.
[25] A.S. Ulrich, Biophysical aspects of using liposomes as delivery
vehicles, Bioscience Rep. 22 2 (2002) 129–149.
[26] in: R.I. Mahato, S.W. Kim (Eds.), Pharmaceutical Perspectives of
Nucleic Acid-Based Therapeutics, Taylor and Francis, 2002.
[27] S. Fong, Y. Liu, T. Heath, P. Fong, D. Liggitt, R.J. Debs, Membrane-
permeant, DNA-binding agents alter intracellular trafficking and
increase the transfection efficiency of complexed plasmid DNA, Mol.
Ther. 10 4 (2004) 706–718.
[28] E. Süleymanoğlu, Self-assembling polyelectrolyte nanostructures for-
mulated for therapeutic gene delivery, Mahidol University J. Pharm.
Sci. 30 (2003) 37–51.
[29] D. Matulis, I. Rouzina, V.A. Bloomfield, Thermodynamics of cationic
lipid binding to DNA and DNA condensation: roles of electrostatics
and hydrophobicity, J. Am. Chem. Soc. 124 (2002) 7331–7342.
[30] G. Fujii, To fuse or not to fuse: the effects of electrostatic interactions,
hydrophobic forces, and structural amphiphilicity on protein-
mediated membrane destabilization, Adv. Drug Deliv. Rev. 20 38 (3)
(1999) 257–277.
[31] P.H. Stachl, Drug solubilization with organic solvents, surfactants and
lipids, in: C.G. Wermuth (Ed.), The Practice of Medicinal Chemistry,
second ed, Elsevier, 2003.
[32] M.I. Lam, P.R. Cullis, Calcium enhances the transfection potency of
plasmid DNA–cationic liposome complexes, Biochim. Biophys. Acta
1463 (2000) 279–290.
[33] E.H. Chowdhury, K. Megumi, I. Harada, A.K. Kundu, A. Toshihiro,
Dramatic effect of Mg2+ on transfecting mammalian cells by
DNA/calcium phosphate precipitates, Anal. Biochem. 328 (2004)
96–97.
[34] J.J.R. Fraústo da Silva, R.J.P. Williams, The Biological Chemistry of
the Elements, The Inorganic Chemistry of Life, Chapter 9: The Bio-
logical Chemistry of Magnesium: Phosphate Metabolism, second ed,
Oxford University Press, Oxford, New York, 2001 (251–276).
[35] V. Majtán, L. Majtánova, Effect of new quaternary bisammonium
compounds on the growth and cell surface hydrophobicity of Entero-
bacter cloaceae, Cent. Eur. J. Publ. Health 8 (2000) 80–82.
[36] B.A. Lobo, S.A. Rogers, C.M. Wiethoff, S. Choosakoonkriang,
S. Bogdanovich-Knipp, C.R. Middaugh, Characterization of cationic
vector-based gene delivery vehicles using isothermal titration and
differential scanning calorimetry, in: M.A. Findeis (Ed.), Methods in
Molecular Medicine, Vol. 65, Nonviral Vectors for Gene Therapy,
Humana Press Inc., Totowa, NJ, USA, 2001, pp. 319–348.
[37] K.M.G. Taylor, D.Q.M. Craig, Physical methods of study: differential
scanning calorimetry, in: V.P. Torchilin, V. Weissig (Eds.), Liposomes,
second ed, Oxford University Press, 2003, pp. 79–101.
[38] A. Chatterjee, S.P. Moulik, P.R. Majhi, S.K. Sanyal, Studies on sur-
factant–biopolymer interaction. I. Microcalorimetric investigation on
the interaction of cetyltrimetylammonium bromide (CTAB) and
sodium dodecylsulfate (SDS) with gelatin (Gn), lysozyme (Lz) and
deoxyribonucleic acid (DNA), Biophys. Chem. 98 (2002) 313–327.
[39] A. Baillet, P. Prognon, Analytical techniques for tracer method moni-
toring, in: J.-L. Grossiord, M. Seiller (Eds.), Multiple Emulsions:
Structure, Properties and Applications, Éditions de Santé, Paris, 1998,
pp. 225.
[40] P.R. Dahlgren, Y.L. Lyubchenko, Atomic force microscopy study of
the effects of Mg2+ and other divalent cations on the end-to-end DNA
interactions, Biochemistry 41 (2002) 11372–11378.
[41] J. Gallová, F. Sersen, F. Devinski, P. Balgavy, The perturbation effect
of N-alkyl-N,N,N-trimetylammonium ions on a model membrane. A
spin-label study, Acta Facultatis Pharmaceuticae Universitatis Come-
nianae, Tommus L (2003) 1–10.
[42] A.A. Yaroslavov, O.Y. Udalykh, N.S. Melik-Nubarov, V.A. Kabanov,
Y.A. Ermakov, V.A. Azov, F.M. Menger, Conventional and gemini
surfactants embedded within bilayer membranes: contrasting behav-
iour, Chem. Eur. J. 7 (2001) 4835–4843.
[43] Y.S. Tarahovsky, R.S. Khusainova, A.V. Gorelov, T.I. Nikolaeva,
A.A. Deev, A.K. Dawson, G.R. Ivanitsky, DNA initiates polymorphic
structural transitions in lecithin, FEBS Lett. 390 (1996) 133–136.
[44] T. Akao, T. Fukumoto, H. Ihara, A. Ito, Conformational change in
DNA induced by cationic bilayer membranes, FEBS Lett. 391 (1–2)
(1996) 215–218.
[45] C.R. Safinya, Structures of lipid–DNA complexes: supramolecular
assembly and gene delivery, Curr. Opin. Struct. Biol. 11 49 (2001)
440–448.
[46] W.M. Gelbart, R.F. Bruinsma, P.A. Pincus, V.A. Parsegian, DNA-
inspired electrostatics, Phys. Today (September 2000) 38–44.
[47] J.P. Clamme, S. Bernacchi, G. Vuilleumier, G. Duportail, Y. Mély,
Gene transfer by cationic surfactant is essentially limited by the
trapping of the surfactant/DNA complexes onto the cell membrane: a
fluorescence investigation, Bichim. Biophys. Acta 1467 (2000) 347–
361.
[48] D. Llères, J.-M. Weibel, D. Heissler, G. Zuber, G. Duportail, Y. Mély,
Dependence of the cellular internalization and transfection efficiency
on the structure and physicochemical properties of cationic
detergent/DNA/liposome, J. Gene Med. 6 (2004) 415–428.
[49] S.M. Mel’nikov, R. Dias, Y.S. Mel’nikova, E.F. Marques,
M.G. Miguel, B. Lindman, DNA conformational dynamics in the
presence of catanionic mixtures, FEBS Lett. 453 (1999) 113–118.
[50] P. Opanasopit, M. Nishikawa, M. Hashida, Factors affecting drug and
gene delivery: effects of interaction with blood components, Crit. Rev.
Ther. Drug Carrier Syst. 19 (3) (2002) 191–233.
[51] H. Kamiya, H. Akita, H. Harashima, Pharmacokinetic and pharmaco-
dynamic considerations in gene therapy, Drug Discov. Today 1 8 (21)
(2003) 990–996.
[52] L. Bergelson, A. Domb, The surface structure of lipid drug carriers-
influence on carrier–cell interaction, in: R.H. Müller, W. Mehnert
(Eds.), Particle and Surface Characterization Methods, Medpharm
GmbH Scientific Publishers, Stuttgart, Germany, 1997, pp. 169–184.
[53] A. Asokan, M.J. Cho, Cytosolic delivery of macromolecules. 3. Syn-
thesis and characterization of acid-sensitive bis-detergents, Bioconj,
Chem. 15 (2004) 1166–1173.
[54] V.V. Kuvitchkin, A.G. Sukhomudrenko, Interaction of natural and
synthetic polynucleotides with liposomes in the presence of divalent
actions, Biophysics, XXXII 4 (1987) 628–633 (In Russian).
[55] Y. Sato, S.-I. Nomura, K. Yoshikawa, Enhanced uptake of giant DNA
in cell-sized liposomes, Chem. Phys. Lett. 380 (2003) 279–285.
[56] G. Bhakta, S. Mitra, A. Maitra, DNA encapsulated magnesium and
manganous phosphate nanoparticles: potential non-viral vectors for
gene delivery, Biomaterials 26 (2005) 2157–2163.
[57] P. Esponda, M. Goldstein, S.S. Witkin, In vitro transfection of the
human vas deferens using DNA–liposome and DNA–neutral lipid
complexes, Fertility and Sterility 81 1 (2004) 171–175.
[58] J. Mönkkönen, A. Urtti, Lipid fusion in oligonucleotide and gene
delivery with cationic lipids, Adv. Drug Deliv. Rev. 34 (1998) 37–49.