Designing pH-sensitive gemini nanoparticles for non-viral gene delivery into keratinocytes

Article abstract of DOI:10.1039/c2jm15719e

This study aimed to develop a more effective non-viral gene delivery system for keratinocyte transfection. To this end, gemini nanoparticles were formulated from plasmid DNA, the lipid DOPE (dioleoylphosphatidylethanolamine) and surfactants, where the sur

Full text of DOI:10.1039/c2jm15719e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 6232
www.rsc.org/materials
PAPER
Designing pH-sensitive gemini nanoparticles for non-viral gene delivery into
keratinocytes
a
b
c
a
ab
McDonald Donkuru, Shawn D. Wettig, Ronald E. Verrall, Ildiko Badea and Marianna Foldvari*
Received 7th November 2011, Accepted 6th February 2012
DOI: 10.1039/c2jm15719e
This study aimed to develop a more effective non-viral gene delivery system for keratinocyte
transfection. To this end, gemini nanoparticles were formulated from plasmid DNA, the lipid DOPE
(
dioleoylphosphatidylethanolamine) and surfactants, where the surfactant components are novel pH-
sensitive gemini surfactant derivatives based on the m-7-m (alkyl chain–spacer–alkyl chain, m-s-m)
unsubstituted base structure. The resultant 1,9-bis(alkyl)-1,1,9,9-tetramethyl-5-amino-1,9-
nonanediammonium dibromide surfactants, m-7NH-m, where m ¼ 12, 16, 18 and 18:1, are
hypothesized to stage endosomal release of DNA. The m ¼ 18:1 chain, i.e., mono-unsaturated oleyl
chain, is an alkenyl chain analogue for comparison with the saturated m ¼ 18 alkyl chain based on the
m-7NH-m frame. Analytical and physicochemical characterization of the gemini surfactants included
purity, aggregation properties under pH 2.5–10.5, critical micellar concentration and pK . Gemini
a
nanoparticles were characterized by dynamic light scattering, zeta potential, small-angle X-ray
scattering and transmission electron microscopic studies. Keratinocyte transfection efficiency and
cytotoxicity were evaluated using the luciferase reporter assay and luminescent cell viability assay,
respectively. Gemini nanoparticles formulated from the m-7NH-m gemini surfactants at
a surfactant:DNA charge ratio (r ) 10:1 showed higher transfection efficiency compared to the
ꢀ
unsubstituted compounds (m-7-m, m-3-m series) (p < 0.01). Transfection efficiency increased with
increasing tail length, i.e., m ¼ 12 < m ¼ 16 < m ¼ 18 (p > 0.01), although the difference between m ¼ 16
and m ¼ 18 was insignificant. Morphological studies of the nanoparticles by transmission electron
microscopy showed fusogenic changes at pH ¼ 5. The incorporation of a pH-active amine group within
the spacer of the gemini surfactants significantly enhanced transfection efficiency in keratinocytes. This
may be attributed to optimal interactions between DNA phosphate groups and the m-7NH-m gemini
surfactants owing to their –NH– groups, trimethylene spacing between nitrogen centers and the acidic
pH-induced polymorphic changes, leading to endosomal release of plasmid. Such results highlight the
amino-substituted gemini surfactants as potential components for developing non-viral nanoparticles
with enhanced gene delivery for targeting diseases affecting the skin.
3
hypothesis of Behr et al. There has also been a rapid increase in
Introduction
the types of lipids, polymers, and gemini surfactants used for
4,5
non-viral gene delivery in vitro (see reviews ). Notable also has
In 1987 the first report of lipid-based non-viral gene delivery was
1
published, describing the delivery and expression of DNA
been the application of colloid stability principles for augmenting
the stability of preparations made of DNA and a non-viral
in vitro using a DOTMA/DOPE lipid-based system. Many
significant advances have occurred since then; including the
2
demonstration by Zabner et al. that endocytosis is the primary
6
delivery agent for both in vitro and in vivo use.
Recent advances have focused on rational design and
synthesis of simple as well as structurally complex molecules
with improved properties for gene delivery and the evaluation
route for cellular uptake of DNA bound within a non-viral,
cationic molecular delivery system, supporting the earlier
7,8
of structure–activity relationships. In this regard, the role of
organic synthesis in generating molecules with useful organic
functional groups or structural parts to enhance the efficiency
of non-viral gene delivery takes on major importance. Some
early developments have led to a transition of some molecular
delivery systems from experimental in vitro studies to successful
a
College of Pharmacy & Nutrition, University of Saskatchewan, 110
Science Place, Saskatoon, SK, S7N 5C9, Canada
b
School of Pharmacy, University of Waterloo, 200 University Avenue
West, Waterloo, ON, N2L 3G1, Canada. E-mail: foldvari@uwaterloo.ca;
Fax: +1 519 888 7910; Tel: +1 519 888 4567, ext. 21306
c
Department of Chemistry, University of Saskatchewan, 110 Science Place,
Saskatoon, SK, S7N 5C9, Canada
9,10
in vivo delivery.
More importantly, molecular delivery
6
232 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012

View Article Online
systems (generally termed lipofection agents) have been used,
increasingly, in gene therapy clinical trials, accounting for 6.9%
of all current clinical trials (www.wiley.co.uk/genetherapy/
clinical).
Results
Gemini surfactants
Structure and physicochemical properties. The gemini
compounds studied in this paper can be grouped into three series;
namely, m-3-m, m-7-m and m-7NH-m surfactants (Table 1 and
Scheme 1). The compounds share the m-s-m general structure,
but, differ in the chemical composition of the spacer group and
hydrocarbon tail length. As expected from classical surfactant
behaviour and properties previously observed for m-s-m
In general, attempts to find suitable and effective means of
gene delivery for gene therapy have focused mainly on the use of
viruses, which require attenuation to render them non-patho-
genic. Viruses are currently the most efficient vectors among all
candidates for gene delivery and account for more than 65% of
ongoing trials (www.wiley.co.uk/genetherapy/clinical). At
present, two viral-based gene therapy medications have been
18
surfactants, CMC values for surfactants in each series
11
commercialized, both in China; however, the use of viral
decreased with increasing alkyl tail length, m (Table 1). For equal
tail lengths, the CMC value increased slightly for gemini
surfactants with an –NH– substituent in the spacer due to the
increased aqueous solubility of these compounds, whereas the
CMC changed little using either a trimethylene or hepta-
methylene spacer (Table 1).
vectors has been fraught with problems including dangerous side
12
effects, in the worst cases causing death. Furthermore, viral
vectors are limited in the number of base pairs of DNA that can
fit into the viral capsid, as well as by the possibility of large-scale
contamination of batches, high cost of production and undesir-
13,14
able changes during long-term storage.
pK
surfactants (Table 1) fall in the range of 4.99–5.06. Such an acidic
pK range is known to actively help engineer release of DNA
a
values determined for the amine-substituted m-7NH-m
Generally, the design of non-viral DNA nanoparticles from
bottom up has followed the approximation of viral particles,
with the aim of constructing particles with ‘intelligence’ similar to
viral vectors but without the safety problems associated with
a
from endosomal/lysosomal compartments which are character-
19,20
ized by a routine span of pH change in the range of pH 5–7.4.
13,14
these vectors.
Non-viral delivery systems offer many other
We have previously reported two gemini surfactants, 12-8N-12
advantages, including a customizable and modular design, where
DNA-compacting, targeting, fusogenic and nuclear transport
components can be combined in just the right configuration.
Nanoparticles bearing these modular features can be taken up by
cells via routes governed by cell-type and nanoparticle properties,
such as macropinocytosis, lipid raft-mediated uptake, clathrin-
and 12-7N-12, for DNA transfection having lower pK values in
a
21
the range of 4–4.5. pK considerations alone would suggest that
a
these two surfactants would be better transfection agents.
However, other factors such as the effectiveness of surfactant–
DNA interactions and structural properties of surfactant–DNA
nanoparticles also contribute. Hence, the transfection ability
would result from an optimal integration of these factors than
any of these factors in isolation.
15,16
and caveolae-dependent endocytosis, among others.
From
this point onward, the leading model regards the internalized
nanoparticles as being entrapped by endosomal compartments
that progressively develop degradative capability at low pH (pH
Characterization of surfactant aggregates: particle size. Fig. 1
5
–6) which can trigger ‘intelligent’ nanoparticles to respond by
(
a–d) shows changes in size of surfactant aggregates as a function
of pH. In comparison with the non-substituted spacer series, m-
NH-m surfactants showed the greatest change in aggregate size
adopting changing morphologies, thereby disrupting the endo-
somal membrane, releasing the entrapped DNA and allowing its
7
2
timely entry into the nucleus.
as pH varied from neutral to acidic state, as a result of proton-
ation of the amino groups. Within this group, aggregates of the
This study aimed to develop a more effective non-viral gene
delivery system for keratinocyte transfection that will allow
1
2-7NH-12 surfactant had the smallest transition in size (from
2.4 to 3.7 nm radius, Fig. 1a). This narrow range of structural
17
potential treatment of conditions like scleroderma via topical
gene delivery. As such, we report a novel class of pH-sensitive
m-7NH-m gemini surfactants which were tested to determine
both their transfection efficiencies and cytotoxicities and their
correlations to both the gemini surfactant molecular structures
and the morphological properties of nanoparticles derived from
these gemini surfactants. Findings from our previous work
highlighted the ability of trimethylene spacer (s ¼ 3) to effect
superior transfections when incorporated into the 12-s-12
ꢁ
transition may appear less significant when compared with the
vesicle-to-micelle transition observed for sugar-based gemini
19,22
surfactants;
however, this pH-induced structural change for
the 12-7NH-12 surfactant aggregates holds the potential to aid
the release of DNA once gemini surfactant-DNA complexes
have been incorporated within the cell. The aggregate sizes and
their expansions were more pronounced for the three higher m-
7
NH-m surfactants (m ¼ 16, 18, 18:1). Micelle diameters (at pH 7
or above) were in the range of 3–5 nm, and were in the order of
8:1-7NH-18:1 < 16-7NH-16 < 18-7NH-18. The micelle diame-
9
gemini surfactants and the ability of an amino group (–NH–)
to effect superior transfections when substituted within the
spacer of 12-spacer-12 gemini surfactants at intervals of 3C
8
from adjacent nitrogen centers. Thus the reported m-7NH-m
1
ters expanded (by ꢁ2.5–6 nm, Fig. 1a, b) as pH was decreased
from neutral into an acidic pH range ꢁ6.8–#3, which is the range
gemini surfactants were designed to combine the effects of the
most active previous compounds in single molecules relative to
m-7-m and m-3-m compounds used as structural controls. The
contrasting of gemini surfactants based on such critical prop-
erties as the presence or lack of pH-sensitivity or other designed
a
in which the pK values of these surfactants fall. As expected, the
aggregate size for the m-3-m (Fig. 1c) and m-7-m (Fig. 1d)
surfactants remained essentially unchanged as pH was varied,
with micelle diameters ranging from approximately 2–6 nm and
increasing in the order of 12-3-12 z 18:1-3-18:1 < 16-3-16 < 18-
‘
intelligence’ in the structures provides useful structure–activity
3
-18 and 12-7-12 < 16-7-16 < 18-7-18.
clues.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6233

View Article Online
Table 1 Gemini surfactant structures, CMC and pK
nitrogen distances
a
values, and illustration of surfactant–DNA close-match interaction models for the different inter-
e
Surfactant series and representative
structures
CMC/ꢂ 10ꢃ6
M
pK
Model
a
a
a
C
12: 1290 ꢀ 40
5.0 ꢀ 0.4
d
d
d
d
C
C
C
16: 150 ꢀ 50
5.04 ꢀ 0.3
5.06 ꢀ 0.3
4.99 ꢀ 0.2
d
18: 13 ꢀ 1
d
cis-oleoyl: 44 ꢀ 20
b
C
C
C
C
12: 980 ꢀ 40
N/A
N/A
N/A
N/A
b
16: 30 ꢀ 1
d
18: 13 ꢀ 1
c
cis-oleoyl: 15 ꢀ 0.4
d
C
C
C
12: 850 ꢀ 70
N/A
N/A
N/A
d
16: 16 ꢀ 9
d
18: 11 ꢀ 1
a
b
c
d
e
Ref. 21. Ref. 46. and ref. # 24 therein. Ref. 44. This work. Calculated from Energy Minimized Model of Molecules using ChemOffice Chem 3D
Pro Software. Molecules were energy minimized (MM2 force field) in the gas phase to a minimum root-mean-square gradient of 0.05, ref. 8. In columns 2
and 3, values ¼ mean ꢀ SD.
Gemini nanoparticles constructed from m-7NH-m compounds
from 12-7NH-12, 16-7NH-16, 18-7NH-18 and 18:1-7NH-18:1
were 90 ꢀ 6, 185 ꢀ 21, 165 ꢀ 8, 205 ꢀ 33 nm, respectively
Particle size and zeta-potential of gemini nanoparticles. The
(
intrinsic pH z 7.6 ꢀ 0.4, all cases). The pH dependence of both
mean particle size for the final gemini nanoparticles prepared
particle size and zeta potential (Fig. 2a, b) was determined for
Scheme 1 Synthesis of m-7NH-m, m-3-m and m-7-m gemini surfactant. Reagents and conditions: (i) CH
3
CN, reflux, 12–24 h.
6
234 | J. Mater. Chem., 2012, 22, 6232–6244 This journal is ª The Royal Society of Chemistry 2012

View Article Online
Fig. 1 pH dependence of the size of gemini surfactant aggregates formed in aqueous solution. Initial concentration of aqueous gemini surfactants in
each case was approximately 2 ꢂ CMC. The pH was adjusted using standard acid (HCl) and base (NaOH) solutions. The results are presented for all
gemini surfactants as follows: (a) 12-7NH-12, 18:1-7NH-18:1, (b) 16-7NH-16, 18-7NH-18, (c) 12-3-12, 16-3-16, 18-3-18, 18:1-3-18:1, (d) 12-7-12, 16-7-16,
18-7-18.
Fig. 2 pH dependence of the size and zeta potential of transfection complexes formed from gemini surfactant–plasmid–DOPE (PGL) mixtures. Results
for complexes derived from the four m-7NH-m surfactants, separately, are presented as follows: (a) size analysis – R vs. pH, (b) zeta potential analysis –
¼ 10:1). Standard HCl and NaOH solutions
H
z vs. pH. The titrations were carried out for transfection systems with the optimal ꢀ charge ratio (i.e., r
ꢀ
were used as titrants for pH adjustment.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6235

View Article Online
nanoparticles prepared from the m-7NH-m surfactants, plasmid
and DOPE. For nanoparticles featuring 12-7NH-12, 16-7NH-16
and 18-7NH-18, a pH-dependent increase in the particle size
occurred as pH was decreased below pH 7.5, with the complexes
showing somewhat rapid increase in size until about pH 4.5 was
reached (Fig. 2a). Nanoparticles featuring the 18:1-7NH-18:1
surfactant also increased in size in response to pH reduction;
however, the expansion occurred over the entire pH range
investigated. In all cases, the expansion resulted in particle
enlargement by 30 nm or more.
DNA systems (Fig. 4a and b, respectively; Table 2 shows a listing
of all observed peaks). From the general shapes of the curves in
Fig. 4, a broad peak extending over a large q range on which
sharper features are superimposed is consistent with what one
would expect for a liposomal system containing DNA. The
broad feature arises from variations in bilayer thickness and may
also indicate the presence of a small proportion of multilamellar
24
vesicles in the system. Superimposed upon this broad back-
ground feature are a series of sharper peaks (or in the case of 12-
7-12, shoulders) that are indicative of more strongly ordered,
multiple bilayer arrangements induced by DNA within the
complexes. Peaks observed with an even spacing between them
are indicative of a lamellar structure, while those appearing in
various geometric progressions are representative of lipid bila-
yers packed in inverse hexagonal or cubic packing
Zeta potential of the nanoparticles, which also showed pH-
dependence (Fig. 2b), increased as a function of decreasing pH,
consistent with the expected decrease in counterion (i.e.,
hydroxide) binding to the complex and with previous reports for
20
other transfection nanoparticles. The zeta-potential curve for
the 12-7NH-12 system revealed a unique pattern characterized
by an initial rise from basic pH, reaching a plateau around pH 7.
This was followed by a sharp decrease in zeta-potential at around
pH 5.5, reaching another plateau at lower pH (Fig. 2b). Such
behaviour is consistent with the pH-induced structural change
observed in Fig. 1, 2. The increase in zeta-potential for the 16-
25
arrangements.
The m-7NH-m surfactants (m ¼ 12, 16) both show evidence of
a more ordered lamellar phase than is observed for the m-7-m
systems (Fig. 4a, b); the lamellar phase is not detected,
convincingly, for the 12-7-12 system. In addition, a second phase
is also observed, albeit weakly, that does not appear to corre-
spond to an inverse hexagonal phase. Some gemini surfactants
have already been observed to induce a variety of cubic packing
7
NH-16, 18-7NH-18 and 18:1-7NH-18:1 surfactants generally
reached a plateau or started to form a plateau as pH was
decreased below pH 7 (Fig. 2b). In all cases, the nanoparticles
showed a positive charge in the range of +40–+45 mV (at neutral
pH, Fig. 2b), giving an indication of the existence of stable
particles.
26,27
arrangements in lipid bilayer systems.
This is most likely the
case for the m-7NH-m surfactants; however, due to the weakness
of the scattering for the second phase conclusive identification is
not possible. Nevertheless, the results obtained here are consis-
tent with our previous observations for the gemini surfactant/
26,27
Structural characteristics of particles. To probe and relate the
nanoparticle morphology to transfection efficiency, complexes
prepared from the m-7NH-m surfactants were subjected to
transmission electron microscopic (TEM) analysis. The electron
micrographs (not shown) revealed discrete particles ranging from
DOPE/plasmid systems.
Evaluation of gene delivery efficiency
Transfection results from this work were examined for trends
including the permanence or transience of transgene expression
after its successful transfer into cells. Transfection results for
1
50–250 nm, consistent with particle sizes observed from light
scattering measurements (Fig. 2a). No apparent dependence on
tail length was observed. In some cases, both individual and
fused/aggregated particles were observed, consistent with
plasmid-gemini-DOPE (PGL) complexes representing r ¼ 10:1
ꢀ
for the m-3-m surfactants are shown in Fig. 5. Luciferase
2,23
previous findings. The particle aggregation showed no corre-
lation to tail length.
expression peaked at 24 h post-transfection, and then progres-
sively declined at 48 h and 72 h post-transfection. Almost no
luciferase was expressed in cells transfected with two gemini
surfactant-free controls, i.e., preparations without the cationic
gemini surfactant component; namely, plain plasmid solution
(naked DNA) and plasmid combined with DOPE. Since the
highest protein expression was detected at 24 h post-transfection,
quantitation of luciferase reporter gene expression in subsequent
transfections was measured 24 h after transfection.
More extensive TEM investigations were focused on
complexes derived from 12-3-12 and the pH-sensitive 12-7NH-12
(
Fig. 3) and interesting differences were revealed. For systems
derived from 12-3-12, the vesicle bilayer membranes were
observed to fuse at an acidic pH of 5, resulting in the collapse of
the vesicle membranes (Fig. 3, rows a, b). But for systems derived
from 12-7NH-12, the vesicle membranes were observed to reform
after a more prominent membrane fusion/collapse phase (Fig. 3,
rows c, d). It might also be that the membranes did not
completely fuse/collapse. If the fusion/collapse and vesicle
reformation is predominant, and potentially undergoes a number
of repetitive cycles, the resultant perturbation effect could
explain why 12-7NH-12 emerged more transfection-efficient than
In addition to the influence of time on the level of transgene
expression, the effect of other factors such as the r values was
ꢀ
also of interest. Thus, to determine the r at which surfactants
ꢀ
most effectively transfer DNA into cells (most effective DNA
transfer z highest level of gene expression), transfections were
carried out at 10:1 as well as lower r values, namely, 5:1, 2.5:1,
ꢀ
1:1, 0.5:1. Fig. 6 shows that transfection efficiencies (i.e., lucif-
1
2-3-12. The perturbation effect exerted by such fusions/collapse
and reformation cycles, particularly on nanoparticle–endosome
entities, is crucial for the release of DNA trapped within these
entities.
erase expression, 24 h post-transfection) were highest in cells
transfected using PGL complexes with r ¼ 10:1. Complexes
ꢀ
with lower r values had lower efficiencies due to reduced
ꢀ
Small-angle X-ray scattering (SAXS) measurements were
carried out in order to extend our comparisons to the 12-7-12/
DNA vs. 12-7NH-12/DNA and 16-7-16/DNA vs. 16-7NH-16/
surface binding between cells and the less positive nanoparticles.
In previous investigations by our group, transfection complexes
with r ¼ 10:1 also transfected cells more efficiently than
ꢀ
6
236 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Fig. 3 Transmission electron micrographs of vesicles and vesicle-encapsulated DNA systems. Images are for complex particles prepared from plasmid–
gemini surfactant or plasmid–surfactant–DOPE mixtures at pH 5.0. The nanoparticles (NP) are composed of materials as follows: row (a) 12-3-12 and
DOPE vesicles, row (b) 12-3-12, DOPE vesicles and DNA, row (c) 12-7NH-12 and DOPE vesicles, row (d) 12-7NH-12, DOPE vesicles and DNA. The
gemini surfactant:DNA charge ratio for the ternary mixtures was 10:1. Column 1 shows low magnification images, while columns 2, 3 and 4 show
progressive high magnification images that probe further into the detailed structures. Scale bar ¼ 200 nm, column 1; scale bar ¼ 500 nm, column 2; scale
bar ¼ 100 nm, column 3; scale bar ¼ 50 nm, column 4.
9
complexes with higher r
ꢀ
values, including r
ꢀ
¼ 20:1 and 40:1.
efficient (p < 0.01) than only the m ¼ 16 and 18 analogues. For
the m-3-m surfactants, transfection efficiency increased in the
order of: m ¼ 12 < m ¼ 16 < m ¼ 18 < m ¼ 18:1, but the
difference between any two of them was insignificant (p > 0.01).
A similar observation was made for the m-7-m surfactants; m ¼
12 < m ¼ 16 < m ¼ 18. Thus, increment of surfactant tail length
led to enhancement of transfection efficiency, although the
increases were insignificant in most cases.
Also, as observed in PAM 212 cells (Fig. 6), the results of parallel
-dependent transfections in COS-7 cells (data not shown) were
the highest for PGL complexes with r
¼ 10:1, and decreased as
decreased. Thus, the optimal conditions for these trans-
r
ꢀ
ꢀ
r
ꢀ
fections encompass a r value of 10:1 and a post-transfection
ꢀ
duration of 24 h.
To examine the effect of variation in gemini surfactant struc-
ture, transfection efficiencies of the m-7NH-m, m-7-m and m-3-
m surfactants were compared using optimized conditions. Levels
of protein expression were highest for m-7NH-m surfactants,
indicative of their greater transfection ability (p < 0.01), than
both the m-3-m and m-7-m surfactants, which decrease in that
order (Fig. 7a). The superior ability of the m-7NH-m surfactants
series was at a level comparable to the commercial reagent Lip-
ofectamine Plus (p < 0.01).
One of the primary goals of our research is the development of
a gene therapeutic agent for localized scleroderma; thus, the
ability of gemini surfactants to transfect keratinocytes is partic-
ularly important. As such these transfection studies have been
carried out in PAM 212 cells. The expression of a transgene in
eukaryotic cells depends to a large extent on the cell type as much
as other factors. So parallel transfection experiments were run in
COS-7 cells, which due to their high transfectability, are often
used as a standard transfection cell line for evaluating non-viral
transfection agents. The PAM 212 cells recorded an overall
transfectability that is moderately lower than that recorded for
the COS-7 cells (Fig. 7a), which are said to have the ability to
multiply plasmid DNA. This is consistent with other trans-
fectability comparisons in which COS-7 cells were easier to
Of the three surfactant groups (m-7NH-m, 4 members; m-3-m,
4
members; m-7-m, 3 members), results from optimized condi-
tions (Fig. 7a) indicated that for the m-7NH-m surfactants,
transfection efficiency increased with increasing tail length, i.e.,
m ¼ 12 < m ¼ 16 < m ¼ 18 (p > 0.01), although the difference
between m ¼ 16 and m ¼ 18 was insignificant. The remaining
surfactant from this group, m ¼ 18:1, had the lowest efficiency
10
transfect than other cell lines. But, more importantly, the m-
7NH-m surfactants effectively transfected the reporter luciferase
(
though the reverse was expected), but was significantly less
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6237

View Article Online
Table 2 Structural parameters for the DNA/gemini surfactant/DOPE
(PGL) systems determined from small-angle X-ray scattering measure-
ments. Main peaks and corresponding d-spacings are bold (see Fig. 4)
ꢀ
ꢃ1
q (A
d ¼ 2p/q ( Aꢀ )
a
PGL systems
)
Phase
1
2-7-12
0.066
95.2
75.7
58.2
55.6
40.8
77.6
64.8
61.0
59.3
38.3
31.4
71.4
59.3
56.6
52.4
35.3
29.9
77.6
64.8
61.6
60.4
38.3
31.1
0
0
0
0
.083
.108
.113
.154
1
2-7NH-12
0.081
0.097
0.103
0.106
0.164
0.200
L
L
L
1
1
6-7-16
0.088
0.106
0.111
0.120
0.178
0.210
L
L
L
6-7NH-16
0.081
0.097
0.102
0.104
0.164
0.202
a
The letters indicate the phase associated with the observed scattering
peak, L: lamellar.
ꢀ
Fig. 4 SAXS profiles for the PGL nanoparticles (r ¼ 10:1). Scattering
profiles for particles formed from 12-7-12 and 12-7NH-12 are given in
part (a), while the profiles for 16-7-16 and 16-7NH-16 are given in part
(b), with each pair to allow for contrast and comparison.
gene into the PAM 212 cells, clearly highlighting the opportunity
to develop topical therapeutic gene delivery systems based on
these surfactants.
Cytotoxicity due to gemini surfactants in PAM 212 and COS-7
cell cultures
Fig. 5 Luciferase reporter gene expression in murine PAM 212 kerati-
nocytes as a function of post-transfection duration. PGL complexes were
Cytotoxicity in both COS-7 and PAM 212 cells, recorded under
conditions of maximum transfection efficiency, is shown in
Fig. 7b. In both cell lines, treatments with the individual gemini
surfactants resulted in cell viabilities within the range of 78–84%,
which is similar to the 80% viability seen after treatment with
Lipofectamine Plus. Cell viability after treatment with control
surfactant-free formulations (untreated positive control cells,
plasmid alone, DOPE/plasmid) remained at 100%. This marks
the absence of toxicity for the surfactant-free treatments. Cyto-
toxicity of the surfactants generally increased in the following
order: m-3-m < m-7-m < m-7NH-m z Lipofectamine Plus.
prepared with r
ꢀ
¼ 10:1 and featured the m-3-m surfactants. Trans-
fections also involved three controls: naked DNA (plasmid), plasmid/
DOPE combination and Lipofectamine Plus. The resulting protein
produced within the cells was quantified after transfection at different
times: 12 h, 24 h, 48 h, 72 h. Bar charts represent mean + SD from 4
replicates.
However, neither cell line was more susceptible to the toxic
effects.
To evaluate the effect of different doses of gemini surfactants
on the level of transfection-related toxicities, as a complementary
6
238 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Discussion
Direction of surfactant design and synthesis
We have previously investigated the effect of various spacer types
8
on the transfection efficiency of five 12–spacer–12 surfactants
and found a significant difference that separated the surfactants
into two kinds. First, the 12-7NH-12 surfactant had the highest
transfection efficiency and so was as efficient as the commercial
reagent Lipofectamine Plus. Second, the other surfactants had
significantly lower efficiencies. The present study investigated
gemini surfactants that can be grouped into three series: m-3-m,
m-7-m and m-7NH-m surfactants. They vary in both their
hydrocarbon tail length and the chemistry and architecture of
their headgroup–spacer regions (Table 1). Together, the trans-
fection results for these surfactants help determine structure–
activity correlations important to nucleic acid delivery.
Structure–activity relationship
The m-7NH-m surfactants emerged herein with superior trans-
fection efficiencies over the m-3-m and m-7-m surfactants (where
m ¼ constant, Fig. 6). This finding is consistent with the ability of
the 12-7NH-12 surfactant member to engage in close-matching
interactions with DNA phosphate groups (as in our previous
8
report ) and the ability to induce endosomal release of com-
plexed DNA necessary for nuclear entry and protein expression
herein. This integrated ability is attributable to the trimethylene
spacing between nitrogen centers and the pH-sensitive –NH–
8,28
group (Table 1), both of which are important for transfection.
Thus, as expected, the m-7NH-m surfactants have yielded the
most efficient transfections in both COS-7 and PAM 212 cells.
The effect of the –NH– group is evidenced by the fact that the
m-7-m surfactants, which lack only the –NH– group, yielded
reduced transfection. Moreover, even though the m-3-m
surfactants also have trimethylene spacing between their
nitrogen centers for close-matching interactions with DNA
phosphates, they nonetheless gave reduced transfection due to
their lack of an –NH– group. Between the m-3-m and m-7-m
surfactants, the former had slightly higher (though statistically
insignificant) transfection efficiency than the latter (where m ¼
constant, Fig. 7a). This emphasizes the beneficial effect of tri-
methylene spacing, unlike the lengthy heptamethylene spacing,
for transfection via promotion of optimal gemini surfactant
Fig. 6 Transfection efficiencies in PAM 212 cells transfected with PGL
complexes prepared using gemini surfactants belonging to three series.
Transfections with PGL complexes featuring: (a) m-7NH-m surfactants;
(
b) m-3-m surfactants; (c) m-7-m surfactants. The magnitude (in
percentage) by which transfection efficiencies for complexes with r
0:1 exceeded transfection efficiencies for complexes with lower r values
excluding 0.5:1) is indicated above the respective columns (taken as
a group) for lower r values. In all cases luciferase expression was
ꢀ
¼
1
ꢀ
(
ꢀ
quantified after 24 h of transfection. Bar charts represent mean + SD
from 4 replicates.
9,26
interaction with DNA.
investigation to the r -dependent gene expression studies
(
With general regards to cationic liposome–DNA interactions,
while some weak or non-optimal interactions as shown by the use
of dimethylene spacings (e.g., 12-5N-12, 12-8N-12) have been
ꢀ
Fig. 6), cytotoxicities in both COS-7 and PAM 212 cells were
determined after cell transfection. The cytotoxicities were
determined for various doses of gemini surfactants, with these
8,29
reported to yield decreased transfections,
others have been
30
doses corresponding to the values of r from previous trans-
ꢀ
fection (Fig. 6). The results obtained for the m-7NH-m surfac-
reported to yield increased transfections. To shed light on this
apparent inconsistency, this report extended its focus beyond
cationic liposome–DNA interactions by incorporating an –NH–
group into the gemini spacer with the view to introducing pH
tants, are consistent with the dose-dependent cytotoxicity trends
9
which have frequently been reported. The dose-dependent
2,31
results for the m-7NH-m surfactants (Fig. 8) are representative of
the trends also recorded for the m-3-m and m-7-m surfactants
buffering capacity for engineering DNA release within a cell.
These surfactants recorded the highest transfection and empha-
size the benefit of having both a trimethylene spacing and an
–NH– group in the spacer moiety of the gemini compound. In
(
data not shown). Cell viability experiments were also conducted
in COS-7 cells with the three gemini surfactant groups. The
results (data not shown) were consistent with the surfactant
concentration–toxicity proportionality trend recorded for
transfections in PAM 212 cells (Fig. 8).
a a
addition, the low pK values for the m-7NH-m surfactants (pK z
5 for all, Table 1) coupled with the high intrinsic pH of the m-
7NH-m surfactant-based transfection complexes (pH ¼ 7.6 ꢀ 0.4
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6239

View Article Online
Fig. 7 Gemini surfactant-mediated transfection in COS-7 and PAM 212 cells for m-3-m, m-7-m and m-7NH-m surfactants. The results indicate: (a)
protein expression and (b) cell viabilities following gemini surfactant-mediated transfer of a luciferase plasmid into the cells. Transfections were carried
ꢀ
out using optimized conditions, i.e., PGL complexes with r ¼ 10:1 were used, while the luciferase expression was quantified after 24 h of transfection.
Bar charts represent mean + SD from 4 replicates. The labels on the horizontal axis, from left to right, represent the bars from left to right as shown in
both (a) and (b) of this figure.
for r
ꢀ
¼ 10:1) imply the ability of the surfactants to buffer a pH
However, there is an apparent debate in the literature as to what
constitutes an optimal alkyl tail length (saturated or unsaturated,
change in the 5–7.5 pH range. Similar pH buffering capacity noted
for polyethyleneimine (due to its –NH– groups) has been sug-
gested to promote osmotic shock and release of DNA from
34,35
asymmetric or not) within the 12 to 18 C atom range.
For
instance, a study of phospholipids has indicated that shorter
10,30
endosomes into the cytosol.
surfactants recorded better transfection (Fig. 6) than both the m-
-m and m-7-m surfactants due to their low pK -conferred
In other words, the m-7NH-m
alkyl tails allowed for better transfection in vitro, whereas longer
36
tails were needed for better transfection in vivo. For a number
3
of other studies which focused on alkyl tails in the range C –C /
18
a
12
7,33,41
37,38
39,40
capacity to induce DNA release from endosomes after cellular
uptake.
C
C
18:1, each of the tails, namely, C12
,
C
14
,
C
18
,
or
7,22,39
18:1
,
has emerged the best compared to the others
Variation of the alkyl tails alone produced minor enhancement
in transfection efficiency unlike the strong enhancement seen
from modification of the headgroup–spacer region. For both the
m-3-m and m-7-m surfactants, the increase in transfection in
response to an increase in alkyl tail length, though evident and
depending upon the report. Due to the generally moderate
differences between alkyl tail lengths within the range C12–C18
,
advanced rational design efforts are concentrated around the
headgroup region of gemini surfactants targeted for gene
delivery.
32
consistent with some reports, was statistically insignificant.
Statistically significant increases, as a result of alkyl tail varia-
tion, occurred only within the m-7NH-m group. The tails m ¼
With regards to unsaturated alkyl tails, the 18:1-3-18:1
surfactant, made by incorporating mono-unsaturated oleyl tails
in the m-3-m template, yielded the highest transfection for that
series, whereas 18:1-7NH-18:1 created from the m-7NH-m
template yielded the lowest transfection efficiency in its group
(Fig. 7a). This apparent paradox interestingly fits within the
context of recent literature as C18 tails have been reported to give
1
6, 18 yielded better results than m ¼ 12, 18:1 (p < 0.01).
Hydrocarbon tails are generally known to yield efficient trans-
fections when the tail length falls within the range of 12 to 18 C
33
atoms but not within the range of 10 C atoms or fewer.
6
240 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012

View Article Online
through which cells conduct their internal natural recycling and
42
foreign particle destruction. From the zeta potential analysis,
a clear transition occurred for the 12-7NH-12 surfactant in
response to acidification while the three longer alkyl tail
analogues of the m-7NH-m series lacked a clear transition in zeta
potential and instead showed a monotonic decrease in zeta
potential as a function of decreasing pH, consistent with the
expected decrease in hydroxide ion concentration (Fig. 2b). This,
nonetheless still allows the enhanced transfection efficiency for
the m-7NH-m surfactants to be explained by pH-induced
changes of the structure of the particles (Fig. 1, 2).
The efficient transfer of DNA from outside a cell to its nucleus
and the subsequent expression of the transgene is a complex
process, depending on many factors other than the structure of
delivery molecules. This was confirmed in this study by the fact
that expression of the luciferase gene was time-dependent, rising
to a peak at 24 h after transfection and declining thereafter
(
Fig. 5). Transgene expression was also affected by cell type and
dose of gemini surfactant. As a result, simple relationships
between transfection efficiency and any of these factors, partic-
ularly structure of delivery molecules, are not expected. But the
implications of any set of variables, whether they arise regularly,
circumstantially or both, can inform the development process of
non-viral gene medications. With regard to the time-dependent
decline of luciferase expression, future genetic medications
working in a similar fashion could be administered repeatedly, as
already is standard practice for an overwhelming number of non-
genetically-based medications.
Fig. 8 Dose-dependent cytotoxicities in PAM 212 cells resulting from
gemini surfactant-mediated DNA transfection. The cell transfections
were carried out using PGL complexes that contained a uniform plasmid
dose of 0.2 mg but varying concentration of the m-7NH-m surfactants.
The numbers in parentheses next to r
amount of gemini surfactant used for preparing complexes based on the
uniform plasmid dose above. That is, the complexes had different r
values ranging from 10:1 to 0.5:1. Cell viabilities were quantified after
ꢀ
values represent the molar
ꢀ
2
4 h of transfection. Bar charts represent mean + SD from 4 replicates.
The fact that higher transfection was observed in COS-7 as
compared to PAM 212 cells (Fig. 7a) suggests that non-viral gene
medicines may need to be specialized for use in specific cell types
or a range of cell types. The need to specialize medications could
also result from the need to deliver nucleic acids to the cell
nucleus as opposed to those meant for cytosolic delivery. Anti-
sense oligonucleotides usually destined for cell cytoplasm
complete their journey upon crossing the cell membrane into the
cytoplasm. But transfer of plasmid DNA to the nucleus calls for
additional transport process(es) through the cytoplasm and
across the nuclear membrane into the nucleus.
33,41
higher transfection over oleyl tails
and yet other reports find
7,22,39
oleyl tails to be superior to their saturated analogues.
Oleyl
tails, when they result in higher transfection, are believed to do so
through their enhancement of the membrane fluidity of trans-
38
fection complexes, an effect correlated to endosomal escape.
On the other hand, a loss of the double bond via oxidation, either
during transfection complex preparation or during storage, may
35
reduce transfection yield.
Nanoparticle characteristics and efficacy
Evaluation of cytotoxicity
The enhanced transfection efficiency of the amine-substituted m-
The toxic effect on the cell lines, attributable to the gemini
surfactants, was fairly moderate. The cell viability, which was
unaffected by the plasmid alone or plasmid mixed with DOPE,
was reduced to ꢁ84% with the use of the m-3-m and m-7-m
surfactants, which was better (but insignificant, p < 0.05) than an
ꢁ80% cell viability recorded for the m-7NH-m surfactants and
Lipofectamine Plus (Fig. 7b). No clear distinction was observed
in the levels of toxicity between the m-3-m and m-7-m surfac-
tants. On the whole, even though the m-7NH-m surfactants
emerged as slightly more toxic, these surfactants were able to
transfect cells at much higher efficiencies without corresponding
7
7
NH-m surfactants, relative to the unsubstituted m-3-m and m-
-m surfactants with no observed structural changes (Fig. 1, 2), is
partly due to their ability to undergo pH-induced structural
transition(s) in response to adjustment of pH from neutral to
acidic state for both the gemini surfactant-only aggregates and
the gemini surfactant/DNA complexes. Such structural changes
and the enhanced transfection observed for the m-7NH-m
surfactants are consistent with reports in which either vesicle-to-
19,22
31
micelle
or lamellar-to-hexagonal structural changes have
been correlated to increased transfection. These structural
changes are believed to mediate the release of DNA from
endosomes through membrane perturbation or fusion that
increases in cytotoxicity. The dose corresponding to r ¼ 10:1
ꢀ
yielded the best transfection for all surfactants. Transfections
19,22
results from the change of structural conformation.
These
involving lower r values were at least 50% lower in luciferase
ꢀ
expression in this study, and were also significantly lower in
changes reflect the ability of certain types of gemini surfactants to
interface with and take advantage of the progressive acidification
process associated with lysosomal/endosomal degradation
9
previous studies for r
ꢀ
values exceeding 10:1. This result, when
combined with the fact that cytotoxicities in cells transfected with
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6241

View Article Online
r ¼ 10:1 nanocomplexes were very similar with cytotoxicities of
1,3-dibromopropane with 2.1 equivalents of N,N-dimethyldo-
ꢀ
nanocomplexes with lower r values (Fig. 8), demonstrates that
decylamine (Aldrich).
ꢀ
the r ¼ 10:1 nanocomplex preparation formula as optimal.
ꢀ
m-7-m series: 12-7-12, 16-7-16, 18-7-18. The m-7-m surfactants
were synthesized by reacting 1 equivalent of 1,7-dibromoheptane
Conclusion
(
Aldrich) with 2.1 equivalents of either N,N-dimethyl-n-octade-
cylamine (Akzo Nobel) or N,N-dimethyl-n-hexadecylamine (TCI
America) or N,N-dimethyl-n-dodecylamine (Aldrich).
A novel series of gemini surfactant compounds was synthesized
by the incorporation of a protonatable amino group into the
spacer of the m-s-m series of compounds. The resultant m-7NH-
m surfactants were used to construct DNA nanoparticles with
pH-sensitive function that produced significantly enhanced
transfection efficiency compared to the parent compound series.
The enhanced transfection may be explained by the pH-induced
structural transition and membrane fusion of nanoparticles with
endosomes leading to disruption of the endosomal membrane
and release of entrapped DNA. Such results demonstrate the
feasibility of using these amino-substituted gemini surfactants,
along with their simple but effective designs, to develop non-viral
nanoparticle gene carriers for targeting diseases affecting the
skin.
General method of purification
The above reactions were carried out under reflux between 24–72
h in acetonitrile. The resulting surfactants crystallized as white
solids upon cooling and were recovered by filtration, followed by
purification by several recrystallizations from acetonitrile or
a mixture of acetonitrile/diethyl ether (both anhydrous, in the
case of unsaturated surfactants). Additional purification of the
unsaturated surfactant 18:1-7NH-18:1 was done by Soxhlet
extraction in anhydrous diethyl ether for 72 h to remove trace
impurities. Structures of purified surfactants were confirmed by
1
H NMR spectroscopy (CDCl , 500 MHz Bruker) and electro-
3
spray ionization mass spectrometry (ESI-MS), while purities
were verified by aqueous surface tension measurements and
elemental analysis. The surface tension analysis (g vs. log C plots)
Materials and methods
Synthesis of gemini surfactants
(
not shown) for all surfactants showed the absence of a minimum
m-7NH-m series: 12-7NH-12, 16-7NH-16, 18-7NH-18, 18:1-
in the post-CMC region, indicating all surfactant products were
free of any surface active impurities. Micellar properties of
surfactants, most importantly their critical micelle concentra-
tions (CMCs, Table 1), were determined using both specific
conductivity and surface tension analyses.
7
NH-18:1. Synthesis of the 12-7NH-12 gemini surfactant,
21
belonging to the m-7NH-m series, was previously carried out,
providing the foundation for synthesis of members with longer
0
tails; i.e., m ¼ 16, 18, 18:1. In brief, 1 equivalent of 3,3 -imino-
bis(N,N-dimethylpropylamine) (Aldrich) was reacted with 2.1
equivalents of 1-bromo-hexadecane, -octadecane or -cis-9-octa-
decene (Aldrich) in acetonitrile or anhydrous acetonitrile (in the
case of 1-bromo-cis-9-octadecene). Before its use, 1-bromo-cis-9-
octadecene was obtained from the conversion of the corre-
Gemini surfactant, lipid vesicles and plasmid solutions
Aqueous solutions (1.5 mM, which is well above all CMC values)
of each gemini surfactant were prepared and filtered through 0.2
mm Acrodiscꢀ filters (Pall Gelman). A lipid vesicular solution
43
sponding cis-9-octaden-1-ol into the bromide compound as
follows: to a solution of cis-9-octadecen-ol (5.74 g, 18.62 mmol,
was
prepared
as
follows:
1,2-dioleoyl-sn-glycer-
8
5% purity; Spectrum) and pyridine (2.33 g, 29.78 mmol, 99.8%
ophosphatidylethanolamine (DOPE) (Avanti Polar Lipids) and
a-tocopherol (Spectrum) in 1:0.2 weight ratios were dissolved in
100% ethanol (Commercial Alcohols Inc., Canada) and depos-
ited as a thin film on a round-bottomed flask. The lipid film was
vacuum dried (12 h) to remove traces of solvent, followed by
resuspension in 9.25% w/v isotonic sucrose (Spectrum) solution
(pH 9) via sonication to yield a 1 mM solution. The final solution
was filtered through 0.45 mm Acrodiscꢀ filters giving DOPE
vesicles with average size of 124 ꢀ 5 nm (n ¼ 3). A pMASIA.Luc
plasmid stock, obtained from previous plasmid propagation and
purity; Aldrich) in anhydrous acetonitrile (36 mL) was added
1
0
3 2
0.62 g of solid Ph PBr (24.18 mmol, 96% purity; Aldrich) at
ꢄ
C over 10 min while stirring. The complete mixture was stirred
for 1 h (conversion of the alcohol to the bromide was checked by
TLC) and filtrate collected through a short pad of silica gel,
which was rinsed with l:10 ether/pentane (200 mL). A final mass
of 4.64 g (14.00 mmol, 75% yield) of pure 1-bromo-cis-9-octa-
decene was obtained upon solvent removal.
45
m-3-m series: 12-3-12, 16-3-16, 18-3-18, 18:1-3-18:1. Of the m-
purification work, following established methods, was prepared
ꢃ1
3
-m surfactants, the 18:1-3-18:1 surfactant used in this study was
and used as 0.1 mg mL solution.
44
previously synthesized using a bis-Menshutkin reaction of
tetramethyl-1,3-propanediamine with 2.1 equivalents of
a
Determination of pK Values
1
-bromo-cis-9-octadecene. The other surfactants in the series
were subsequently synthesized via the same reaction method as
pK values for gemini surfactants having an –NH– substituent in
a
follows: 18-3-18 resulted from a reaction of 1 equivalent of
their spacer were measured by using a Beckman pH meter
(Model 350) with a glass calomel electrode (Model 511083,
1
,3-dibromopropane (Aldrich) with 2.1 equivalents of N,N-
21
dimethyloctadecylamine (Akzo Nobel); 16-3-16 resulted from
a reaction of 1 equivalent of tetramethyl-1,3-propanediamine
Beckman), as previously reported. Aqueous solutions (2 mM)
of gemini surfactants were prepared in 1 mM NaOH to ensure
complete deprotonation of nitrogen, then titrated with 0.5 mM
HCl at room temperature.
(
Aldrich) with 2.1 equivalents 1-bromohexadecane (Aldrich);
and, 12-3-12 resulted from a reaction of 1 equivalent of
6
242 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Preparation of transfection complexes
using CellTitre-Gloꢀ luminescent cell viability assay (Promega).
The assay was prepared following manufacturer’s instructions
and luminescence was monitored using a Reporter microplate
luminometer (Turner Biosystems).
To prepare non-viral transfection complexes, it was estimated
that 1 mg of plasmid DNA (i.e., 3 nmol of phosphate) combines
with 1.5 nmol of dicationic gemini surfactants (i.e., 1.5 nmol ꢂ 2
ammonium nitrogens) to give a gemini surfactant:plasmid DNA
Physical characterization of nanoparticles
charge ratio (r
ꢀ
mixtures representing r
) of 1:1. Using this guide, different transfection
ꢀ
¼ 0.5:1, 1:1, 2.5:1, 5:1 and 10:1 were
routinely prepared and used for cell transfections. For each
The analyses of particle size and zeta potential were performed
on surfactant aggregates and those complexes prepared at an
preparation (representing a r value), 0.2 mg of plasmid was
ꢀ
optimal transfection charge ratio (r ¼ 10:1). Determinations
ꢀ
mixed with the appropriate aliquot (in nmol) of gemini surfac-
tant solution (1.5 mM) and the volume was brought to 7 mL in
each case. The binary mixtures were incubated at room
temperature for 15 min, then 50 mL of the DOPE vesicles (1 mM)
were added for a final incubation of 30 min at room temperature.
Based on our previous studies DOPE is used as a co-lipid in each
gemini nanoparticle system since it was found that it provides
synergistic transfection enhancement to the m-s-m unsubstituted
series of gemini surfactant. Without DOPE the m-s-m gemini-
plasmid complexes show very low transfection rates. Second-
generation substituted gemini surfactants show increased trans-
fection even in the absence of DOPE, however, in this study, for
appropriate comparison, all nanoparticles contained DOPE.
H
of hydrodynamic radius (R ) and zeta potential (z) as a func-
tion of pH were carried out using a Zetasizer NanoZS with
MPT-2 auto-titrator attachment (Malvern Instruments). Ten
mL of the desired sample was placed in the titration cell of the
auto-titrator and samples were titrated (in 0.2–0.5 pH unit
increments) with 0.1 M NaOH and 0.1 M HCl over the desired
pH range. Size and zeta potential measurements were made in
triplicate at each pH point, and the reported graphs were
plotted using the average values. pH titrations for the aqueous
gemini surfactant solutions were carried out using initial
surfactant concentration of ꢁ2 ꢂ CMC.
Small-angle X-ray scattering (SAXS)
SAXS measurements were made using beamline X21 at the
National Synchrotron Light Source at Brookhaven National
Laboratory, Long Island, NY. The measurements were per-
formed with 12 keV X-rays and the data covered a q-range from
Cell transfection
PAM 212 keratinocyte cells, derived from BALB/C mouse
epidermis, were cultured in Minimal Eagle’s Medium (MEM),
supplemented with 10% fetal bovine serum (FBS) and antibi-
otic-antimycotic (Ab) at 1 unit/mL. COS-7 African Green
monkey kidney fibroblast cells were cultured in Dulbecco’s
Modified Eagle Medium (DMEM), supplemented similarly as
for PAM 212 cells in MEM. MEM and DMEM as well as FBS
and Ab were obtained from Invitrogen/Gibco. In each case,
cells were harvested from substrate after 85% confluence to
initiate a transfection experiment. On the day before each
ꢀ
ꢃ1
0
.008 to 0.5 A . Samples were loaded into 1.5 mm capillaries
Charles Supper #15-BG) and the scattering pattern was recor-
ded using a 13 cm Mar CCD detector (Mar USA), at 1.26 m
calibrated with the scattering pattern of silver behenate)
(
(
downstream of the sample. Samples for the SAXS study were
prepared as described for transfection complexes, and then
concentrated 5ꢂ by evaporation. All spectra were processed to
remove background contributions by subtracting the scattering
profile obtained for a water-filled capillary. Peak fitting was
carried out using the software package fityk (http://
www.unipress.waw.pl/fityk/) using multiple Gaussian functions.
4
transfection, 2 ꢂ 10 cells/well were seeded in 96-well plates
(
Corning Inc.), keeping the two cell types in their respective
media. Volumes were kept to 100 mL per well (for 96-well
plates). One hour before transfection, cells were rinsed and
covered with fresh serum-free culture medium. Cells were
transfected with transfection mixtures representing a plasmid
dose of 0.2 mg/well by adding mixtures dropwise. After 5 h
incubation, the supernatant medium on the cells was replaced
with fresh medium made using the stated supplementing
formula. Parallel transfections were run using the non-viral
commercial transfection standard Lipofectamine Plusꢁ (Invi-
trogen Life Technologies). The Lipofectamine Plusꢁ prepara-
tion for a plasmid dose of 0.2 mg/well followed manufacturer’s
protocol. Each transfection experiment was done in quadru-
plicate. The resulting luciferase expression was quantified 12–72
h later, using Steady-Gloꢁ luciferase assay (Promega) and
a Reporter microplate luminometer (Turner Biosystems) to
measure the luminescence signal from the assayed cells.
Transmission electron microscopy (TEM)
Both the gemini surfactant vesicles and the final transfection
complexes, prepared as described earlier, were studied using
transmission electron microscopy (TEM). Each sample was
prepared by placing a small drop of ca. 10 mL of transfection
mixture on a formvar and carbon-coated 200 mesh copper grid.
After blotting off excess liquid, a thin layer of the sample was
allowed to dry and absorb onto the grids. Negative staining of
the samples was performed by using 1% w/v phosphotungstic
acid. Images of surfactant aggregates and the transfection
nanocomplexes were obtained using a JEOL 2010F field-emis-
sion transmission electron microscope operating at 200 keV.
Statistical analysis
Viability of transfected cells
Statistical analysis was performed by one-way ANOVA (with
either SPSS 13.0 software, or using MS Excel statistical func-
tions). Comparisons were made using Tukey’s multiple
comparison test and non-parametric analysis (Kruskal-Wallis
PAM 212 and COS-7 cells were prepared and transfected as
described above. However, to determine cell viabilities (and
transfection-related cytotoxicity), transfected cells were assayed
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 6232–6244 | 6243

View Article Online
test). Significant differences between the experimental groups
14 Y. Hong, K. Lee, S. S. Yu, S. Kim, J.-G. Kim, H. Y. Shin and S. Kim,
J. Gene Med., 2004, 6, 724–733.
were established at the p < 0.05 level of significance, and also at p
1
5 L. K. Medina-Kauwe, J. Xie and S. Hamm-Alvarez, Gene Ther., 2005,
2, 1734–1751.
16 Z. J. Cheng, R. D. Singh, D. K. Sharma, E. L. Holicky, K. Hanada,
<
0.01 for a number of cases where significant difference exists at
1
this lower probability level.
D. L. Marks and R. E. Pagano, Mol. Biol. Cell, 2006, 17, 3197–3210.
1
7 B. Freundlich, S. A. Jimenez, V. D. Steen, T. A. Medsger, Jr.,
M. Szkolnicki and H. S. Jaffe, Arthritis Rheum., 1992, 35, 1134–1142.
8 R. Zana, M. Benrraou and R. Rueff, Langmuir, 1991, 7, 1072–1075.
9 M. Johnsson, A. Wagenaar and J. B. F. N. Engberts, J. Am. Chem.
Soc., 2003, 125, 757–760.
Abbreviations
1
1
CMC
Critical micelle concentration
DMEM
DOPE
Dulbecco’s Modified Eagle Medium
1,2-Dioleoyl-sn-glycero-3-
20 M. Johnsson, A. Wagenaar, M. C. A. Stuart and J. B. F. N. Engberts,
Langmuir, 2003, 19, 4609–4618.
2
1 S. D. Wettig, C. Wang, R. E. Verrall and M. Foldvari, Phys. Chem.
Chem. Phys., 2007, 9, 871–877.
22 M. L. Fielden, C. Perrin, A. Kremer, M. Bergsma, M. C. Stuart,
phosphatidylethanolamine
DOTMA
N-[1-(2,3-dioleyloxy)propyl]-N,N,N-
trimethylammonium chloride
Modified Eagle Medium
P. Camilleri and J. B. F. N. Engberts, Eur. J. Biochem., 2001, 268,
1
269–1279.
3 J. Gustafsson, G. Arvidson, G. Karlsson and M. Almgren, Biochim.
Biophys. Acta, Biomembr., 1995, 1235, 305–312.
MEM
SAXS
TEM
TLC
2
Small-angle X-ray scattering
Transmission electron microscopy
Thin layer chromatography
24 J. A. Bouwstra, G. S. Gooris, W. Bras and H. Talsma, Chem. Phys.
Lipids, 1993, 64, 83–98.
2
2
2
2
2
5 R. Winter and R. Kohling, J. Phys.: Condens. Matter, 2004, 16, S327–
S352.
6 M. Foldvari, I. Badea, S. Wettig, R. Verrall and M. Bagonluri, J. Exp.
Nanosci., 2006, 1, 165–176.
7 I. Badea, S. Wettig, R. Verrall and M. Foldvari, Eur. J. Pharm.
Biopharm., 2007, 65, 414–422.
8 V. Vijayanathan, T. Thomas, A. Shirahata and T. J. Thomas,
Biochemistry, 2001, 40, 13644–13651.
Acknowledgements
The authors thank Mr. Fred Pearson for his excellent technical
assistance in obtaining high resolution micrographs and Dr
Gianluigi Botton for his technical advice in analyzing nano-
particles at the Canadian Centre for Electron Microscopy,
McMaster University. Research involving SAXS was conducted
at the National Synchrotron Light Source, Brookhaven National
Laboratory, Upton, NY, which is supported by the U.S.
Department of Energy, Division of Materials Sciences and
Division of Chemical Sciences, under Contract No. DE-AC02-
9 A. J. Geall, M. A. W. Eaton, T. Baker, C. Catterall and
I. S. Blagbrough, FEBS Lett., 1999, 459, 337–342.
30 L. Stewart, M. Manvell, E. Hillery, C. J. Etheridge, R. G. Cooper,
H. Stark, M. van Heel, M. Preuss, E. W. F. W. Alton and
A. D. Miller, J. Chem. Soc., Perkin Trans. 2, 2001, 624–632.
3
1 P. C. Bell, M. Bergsma, I. P. Dolbnya, W. Bras, M. C. Stuart,
A. E. Rowan, M. C. Feiters and J. B. Engberts, J. Am. Chem. Soc.,
2003, 125, 1551–1558.
3
2 G. Byk, C. Dubertret, V. Escriou, M. Frederic, G. Jaslin, R. Rangara,
B. Pitard, J. Crouzet, P. Wils, B. Schwartz and D. Scherman, J. Med.
Chem., 1998, 41, 224–235.
9
8CH10886. The research in this paper was supported by the
3
3
3
3
3 M. Castro, D. Griffiths, A. Patel, N. Pattrick, C. Kitson and
M. Ladlow, Org. Biomol. Chem., 2004, 2, 2814–2820.
4 R. P. Balasubramaniam, M. J. Bennett, A. M. Aberle, J. G. Malone,
M. H. Nantz and R. W. Malone, Gene Ther, 1996, 3, 163–172.
5 D. Niculescu-Duvaz, J. Heyes and C. J. Springer, Curr. Med. Chem.,
Canadian Institutes of Health Research and the Natural Sciences
and Engineering Research Council of Canada. The authors
declare no conflict of interest related to this manuscript.
2
003, 10, 1233–1261.
6 V. Floch, S. Loisel, E. Guenin, A. C. Herve, J. C. Clement,
J. J. Yaouanc, H. des Abbayes and C. Ferec, J. Med. Chem., 2000,
43, 4617–4628.
References
1
P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan,
M. Wenz, J. P. Northrop, G. M. Ringold and M. Danielsen, Proc.
Natl. Acad. Sci. U. S. A., 1987, 84, 7413–7417.
37 J. A. Heyes, D. Niculescu-Duvaz, R. G. Cooper and C. J. Springer, J.
Med. Chem., 2002, 45, 99–114.
2
3
4
5
6
7
8
9
J. Zabner, A. J. Fasbender, T. Moninger, K. A. Poellinger and
M. J. Welsh, J. Biol. Chem., 1995, 270, 18997–19007.
J. P. Behr, B. Demeneix, J. P. Loeffler and J. Perez-Mutul, Proc. Natl.
Acad. Sci. U. S. A., 1989, 86, 6982–6986.
S. D. Wettig, R. E. Verrall and M. Foldvari, Curr. Gene Ther., 2008, 8,
38 Z. Hyvonen, M. Ruponen, S. Ronkko, P. Suhonen and A. Urtti, Eur.
J. Pharm. Sci., 2002, 15, 449–460.
39 P. Delepine, C. Guillaume, V. Floch, S. Loisel, J. J. Yaouanc,
J. C. Clement, H. Des Abbayes and C. Ferec, J. Pharm. Sci., 2000,
89, 629–638.
40 T. Ren, Y. K. Song, G. Zhang and D. Liu, Gene Ther., 2000, 7, 764–
768.
41 A. Roosjen, J. Smisterova, C. Driessen, J. T. Anders, A. Wagenaar,
D. Hoekstra, R. Hulst and J. B. F. N. Engberts, Eur. J. Org.
Chem., 2002, 1271–1277.
42 A. J. Kirby, P. Camilleri, J. B. F. N. Engberts, M. C. Feiters,
R. J. M. Nolte, O. Soderman, M. Bergsma, P. C. Bell,
M. L. Fielden, C. L. Garcia Rodriguez, P. Guedat, A. Kremer,
C. McGregor, C. Perrin, G. Ronsin and M. C. P. van Eijk, Angew.
Chem., Int. Ed., 2003, 42, 1448–1457.
43 J. Sandri and J. Viala, Synth. Commun., 1992, 22, 2945–2948.
44 X. Li, S. D. Wettig, C. Wang, M. Foldvari and R. E. Verrall, Phys.
Chem. Chem. Phys., 2005, 7, 3172–3178.
45 T. Maniatis, E. F. Fritsch and J. Sambroook, In Molecular Cloning: A
Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY, USA, 1982, pp. 545.
46 S. D. Wettig and R. E. Verrall, J. Colloid Interface Sci., 2001, 235,
310–316.
9–23.
D. W. Pack, A. S. Hoffman, S. Pun and P. S. Stayton, Nat. Rev. Drug
Discovery, 2005, 4, 581–593.
A. Martin-Herranz, A. Ahmad, H. M. Evans, K. Ewert, U. Schulze
and C. R. Safinya, Biophys. J., 2004, 86, 1160–1168.
C. McGregor, C. Perrin, M. Monck, P. Camilleri and A. J. Kirby, J.
Am. Chem. Soc., 2001, 123, 6215–6220.
S. D. Wettig, I. Badea, M. Donkuru, R. E. Verrall and M. Foldvari, J.
Gene Med., 2007, 9, 649–658.
I. Badea, R. Verrall, M. Baca-Estrada, S. Tikoo, A. Rosenberg,
P. Kumar and M. Foldvari, J. Gene Med., 2005, 7, 1200–1214.
0 O. Boussif, F. Lezoualc’h, M. A. Zanta, M. D. Mergny, D. Scherman,
1
B. Demeneix and J. P. Behr, Proc. Natl. Acad. Sci. U. S. A., 1995, 92,
7297–7301.
1 H. Jia, Nat. Med. (New York, NY, United States), 2006, 12, 263–264.
2 A. Branca Malorye, Nat. Biotechnol., 2005, 23, 519–521.
3 C. L. Cepko, B. E. Roberts and R. C. Mulligan, Cell, 1984, 37, 1053–
1
1
1
1062.
6
244 | J. Mater. Chem., 2012, 22, 6232–6244
This journal is ª The Royal Society of Chemistry 2012