Glycine-terminated dendritic amphiphiles for nonviral gene delivery

Article abstract of DOI:10.1021/bm300892v

Development of nonviral vectors for the successful application of gene therapy through siRNA/DNA transfection of cells is still a great challenge in current research.(1, 2) In the present study, we have developed multivalent polyglycerol dendron based amphiphiles with well-defined molecular structures that express controlled glycine arrays on their surfaces. The structure-activity relationships with respect to the siRNA complexation, toxicity, and transfection profiles were studied with synthesized amphiphilic polycations. Our findings revealed that a second-generation amphiphilic dendrimer (G2-octaamine, 4) that has eight amine groups on its surface and a hydrophobic C-18 alkyl chain at the core of the dendron, acts as an efficient vector to deliver siRNA and achieve potent gene silencing by investigating the knockdown of luciferase and GAPDH gene activity in HeLa cells. Interestingly, the amphiphilic vector is nontoxic even at higher ratio of N/P 100. To the best of our knowledge this is the first example of successful in vitro siRNA transfection using dendritic amphiphiles. We believe that this supramolecular complex may serve as a new promising alternative for nonviral siRNA delivery systems and will be investigated for in vivo siRNA delivery in the future.

Full text of DOI:10.1021/bm300892v

Article
pubs.acs.org/Biomac
Glycine-Terminated Dendritic Amphiphiles for Nonviral Gene
Delivery
Shashwat Malhotra,† Hannah Bauer,‡ Ariane Tschiche,† Anna Maria Staedtler,† Andreas Mohr,†
Marcelo Calderon,† Virinder S. Parmar,§ Lena Hoeke,‡ Soroush Sharbati,‡ Ralf Einspanier,*,‡
́
and Rainer Haag*,†
†Institut fur Chemie und Biochemie, Freie Universitat Berlin, Takustrasse 3, Berlin 14195, Germany
̈
̈
‡Institute of Veterinary Biochemistry, Freie Universitat Berlin, Oertzenweg 19b, Berlin 14163, Germany
̈
§Bioorganic Laboratory, Department of Chemistry, University of Delhi, Delhi 110 007, India
S
* Supporting Information
ABSTRACT: Development of nonviral vectors for the
successful application of gene therapy through siRNA/DNA
transfection of cells is still a great challenge in current
research.1,2 In the present study, we have developed
multivalent polyglycerol dendron based amphiphiles with
well-defined molecular structures that express controlled
glycine arrays on their surfaces. The structure−activity
relationships with respect to the siRNA complexation, toxicity,
and transfection profiles were studied with synthesized
amphiphilic polycations. Our findings revealed that a second-
generation amphiphilic dendrimer (G2-octaamine, 4) that has
eight amine groups on its surface and a hydrophobic C-18
alkyl chain at the core of the dendron, acts as an efficient vector to deliver siRNA and achieve potent gene silencing by
investigating the knockdown of luciferase and GAPDH gene activity in HeLa cells. Interestingly, the amphiphilic vector is
nontoxic even at higher ratio of N/P 100. To the best of our knowledge this is the first example of successful in vitro siRNA
transfection using dendritic amphiphiles. We believe that this supramolecular complex may serve as a new promising alternative
for nonviral siRNA delivery systems and will be investigated for in vivo siRNA delivery in the future.
nonviral gene carriers often exhibit significantly reduced
transfection efficiencies as they are stalled by numerous extra-
and intracellular obstacles. However, biocompatibility, ease, and
potential for large-scale production make these compounds
increasingly attractive for gene therapy.11 As a result, a
significant amount of research in the past decade has been
focused on designing cationic compounds that can form
complexes with DNA and overcome in vitro and in vivo
barriers for gene delivery. Among them, dendrons have been of
particular interest as a result of their ability to deliver genetic
material into the cells.12 Because of the multivalency effect,
dendrons can have multiple densely packed surface groups that
offer multiple simultaneous interactions that lead to enhanced
binding.13,14 Ever since the ground-breaking work from the
groups of Tomalia and Szoka using poly(amidoamine)
spherical dendrimers, a wide range of different dendritic
architectures have been explored for their gene delivery
potential.15,16 In most cases, large polycationic dendrimers
based on polyamines such as dendritic poly-L-lysine and
INTRODUCTION
■
Gene therapy has gained significant attention over the past two
decades as a potential method for treating chronic diseases and
genetic disorders as well as an alternative method to traditional
chemotherapy for cancer treatment.1−3 A considerable amount
of research effort is currently being focused on designing
effective gene vectors that can condense and protect
oligonucleotides for gene transfection, whereby free oligonu-
cleotides and DNA are rapidly degraded by serum nucleases in
the blood when injected intravenously and therefore cannot
easily pass through the negatively charged cell surface barrier.4
Earlier research concentrated on using viral gene vectors,
including both retro- and adenoviruses, as these vectors have
exhibited high gene transfection efficiency by delivering both
DNA and RNA to numerous cell lines.5 Elementary problems
including toxicity, inflammatory, immunogenic, and mutagenic
effects, however, make them a safety risk and underline the
urgent need for nonviral alternatives.6
Nonviral gene delivery vector systems, including cationic
lipids,7 polymers,8 dendrimers,9 and peptides,10 are frequently
regarded as a potentially safer alternative to viral gene delivery.
However, unlike viral analogues that have developed means to
overcome cellular barriers and immune defense mechanisms,
Received: June 12, 2012
Revised: July 30, 2012
Published: August 10, 2012
© 2012 American Chemical Society
3087
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
NMR spectra were recorded on a Bruker ECX 400 (1H: 400 MHz,
13C: 100.5 MHz), a Jeol Eclipse (1H: 500 MHz, 13C: 125.8 MHz), or
on a Bruker Biospin (1H: 700 MHz, 13C: 176.1 MHz) spectrometer at
25 °C and calibrated against residual solvent peaks as internal
polyglycerol amines have been employed.17−26 Such systems
can be considered to belong to the cationic polymer class of
gene delivery vehicle. Although such systems can perform
highly effective gene delivery, they also exhibit challenging
toxicity profiles and can cause problems as a result of their
persistence inside the cells after gene delivery has taken
place.27,28 A different approach has therefore been to use
dendron based architectures in which a hydrophobic group at
the focal point (amphiphiles) encourages self-assembly of the
dendrons into a larger aggregate, “pseudodendrimer”. Such
systems, as reported by different research groups, are not only
capable of condensing the genetic materials but also for
delivering and silencing the cells.29−33 Based on their
amphiphilic nature, such dendrons are considered to belong
to the cationic lipid class for gene delivery vehicle.34,35 In
contrast to most phospholipid based vesicles, these dendritic
glycerol amphiphiles form defined and shape persistant
micelles.36
Our interest is in optimizing siRNA binding and delivery into
cells by developing low-molecular-weight oligoglycerol based
amphiphiles with low cellular toxicity. We therefore decided to
develop multivalent amphiphiles with well-defined molecular
structures that express controlled multivalent glycine arrays on
their surfaces. These amino acids are naturally occurring
building blocks and can be used as biodegradable amine-
functionalities. Our endeavor was to control the loading of
amine content on the surface of the dendritic head-groups of
the synthesized amphiphiles using chemo-enzymatic and
classical chemical synthetic routes. Such systems enable an
understanding of structure−activity relationships with respect
to the siRNA complexation, toxicity, and transfection profiles
often associated with amine bearing polycations. Here, we
present a second-generation amphiphilic dendrimer (G2-
octaamine, 4) with eight amine groups on its surface and a
C-18 hydrophobic alkyl chain at the core that acts as an
efficient vector to deliver siRNA and achieve potent gene
silencing by investigating the knockdown of a luciferase gene in
stably transfected HeLa cells as well as knockdown of the
intrinsic expression of GAPDH in HeLa cells. To the best of
our knowledge, this is the first example of successful in vitro
siRNA transfection using dendritic amphiphiles. It may
represent a promising nonviral carrier system for siRNA
delivery with a clear in vivo application potential.
1
standard. All 13C NMR spectra were recorded with H broadband
decoupling. Chemical shifts δ are given in ppm according to calibration
to the corresponding solvents CDCl3 (1H: 7.26 ppm, 13C: 77.00 ppm)
and CD3OD (1H: 3.31 ppm, 13C: 49.05 ppm). Electrospray-ionization
time-of-flight mass spectrometry (ESI-TOF-MS) experiments were
carried out on an Agilent 6210 ESI-TOF, Agilent Technologies.
Critical Micelle Concentration (CMC). Fluorescence emission
spectra were taken with a Jasco FP-6500 spectrofluorimeter equipped
with a thermostatted cell holder, a DC-powered 150 W xenon lamp, a
Hamamatsu R928 photomultiplier, and a variable slit system. Both
excitation and emission slits were set at 5 nm. In the present study,
diphenyl-1,3,5-hexatriene (DPH) was used as a hydrophobic probe to
determine the critical micelle concentrations (CMCs) of the
amphiphiles in buffered aqueous solution (HEPES saline buffer, 9.4
mM NaCl, pH 7.2). Fluorescence of DPH was recorded from 360 to
600 nm after excitation at 345 nm. In order to determine the CMCs,
the fluorescence intensity of the most intensive peak at 430 nm was
plotted against the amphiphile concentration. Independent linear
regressions were performed on the data points above and below the
putative CMC.37−40 The CMCs were derived from the points of
intersection of the independent linear regressions.
Prior to measurements, DPH stock solution of 0.5 μM (in HEPES
saline buffer, 9.4 mM NaCl, pH 7.2) was freshly prepared by dissolving
the probe in THF or acetonitrile. Increasing concentrations of
amphiphiles were then added to DPH (0.5 μM) from a stock solution
and fluorescence emission spectra were measured at different
amphiphile concentrations. To ensure proper mixing and dissolution
of the compounds all samples were stirred thoroughly by using a
laboratory vortex shaker. The samples were then incubated for at least
12 h at room temperature. All measurements were carried out at 22
2 °C and taken in triplicate and averaged. Data analysis was performed
using SigmaPlot 8.0 software.
Dynamic Light Scattering (DLS) and Zeta Potential. DLS and zeta
potential measurements were conducted at 25 °C using a Zetasizer
Nano ZS analyzer with integrated 4 mW He−Ne laser, λ = 633 nm
(Malvern Instruments Ltd., U.K.). The amphiphiles were measured in
HEPES saline buffer (2 mM, EDTA 10 μM, NaCl 9.4 mM) at a pH of
7.2. Lipopolyplex solutions were obtained as follows: first a 200 μM
DNA (21-mer oligonucleotides) solution was freshly prepared in the
relevant buffer solution. With respect to a certain N/P ration, the
appropriate amount of amphiphile was added. Here, it should be noted
that the final concentrations of the respective amphiphiles were above
each CMC. After mixing and incubation for 30 min, the sample
solutions were directly measured. All measurements were carried out
using folded capillary cells (DTS 1060) in three replicate measure-
ments.
Ethidium Bromide Assay. The experimental protocol for the
ethidium bromide (EthBr) displacement assay was based on a
previously reported study.41 Fluorescence spectroscopy measurements
were performed with a JASCO FP-6500 spectrofluorometer. Excitation
of the sample was done with 546 nm excitation light, and emission was
measured from 560 to 700 nm. We studied the DNA-binding
properties at low salt concentration in HEPES saline buffer (pH 7.2, 2
mM HEPES, 9.4 mM NaCl). A total of 0.1242 mg of EthBr was
dissolved in 50 mL of buffer to provide 1.26 μM concentration. In
order to prepare 100 mL of a 5 μM solution of DNA (21-mer
oligonucleotides), 0.0762 mg of each strand of DNA was measured in
an accurate scale. For preparation of double helix, incubation was done
by keeping strands at 90 °C for 1 min and then 2 h at 37 °C. DNA was
provided from Operon Biotechnologies GmbH. Mother solutions
(1000 μM based on nitrogen content) for each tested amphiphile were
prepared in buffer solution. Solutions containing DNA and EthBr in
buffer were first incubated at room temperature for 5 min to ensure
interaction. The fluorescence of the DNA solution with EthBr was set
at 100%. Consequently, an appropriate quantity of the corresponding
amphiphile was added in order to reach the desired N/P ratio followed
EXPERIMENTAL SECTION
■
Materials. Air and moisture sensitive reactions were carried out in
flame-dried glass ware under argon atmosphere. Anhydrous solvents
were either commercially purchased from Acros OrganicsTM in
septum-sealed bottles or chemically dried using a MBRAUN SPS
800 solvent purification system. Shell acetal-protected PG dendrons
[G1.0]−OH (5) and [G2.0]−OH (10) were provided by our group.
Mono-Boc-protected glycine was supplied by the Acros Organics. All
other chemicals were of reagent grade quality and used without further
purification from the suppliers Acros Organics, Fluka, Sigma-Aldrich,
Roth, Invitrogen, and Merck.
Measurements. Chromatography and Spectroscopy. Thin layer
chromatography (TLC) analysis was carried out on silica coated
aluminum plates from Merck either using silica gel 60, F254, or silica gel
60 RP-18 F254s when performing reversed phase (RP) analysis.
Preparative column chromatography was conducted on silica gel 60
(0.040−0.063 mm, 230−400 mesh ASTM). Detection was accom-
plished by UV irradiation (254 nm; 366 nm) and different staining
solutions such as potassium permanganate, cerium molybdate,
ninhydrin, bromocresol green, and Dragendorff reagent.
3088
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
by further incubation for 5 min, after each addition. Here, it should be
noted that the final concentrations of the respective amphiphiles were
above each CMC. Control experiments were also conducted by
measuring solutions which only contained EthBr and the amphiphiles.
Results were normalized against the set 100% value and expressed as a
percentaged reduction of the relative fluorescence intensity.
Agarose Gel Electrophoresis. One day prior to electrophoresis the
glycine amphiphiles were dissolved in ddH2O. Complexes of glycine
amphiphiles and fluorescently labeled siRNA (FAM-siRNA, GU-
CAACGGAUUUGGUCGUA, Eurogentec) were created by incuba-
tion for 30 min at room temperature. They were loaded on 4% high-
resolution agarose gels (MetaPhor Agarose, Lonza) and subjected to
electrophoresis at 70 V. Polymer-FAM-siRNA-complexes were
visualized after excitation at 495 nm and acquiring the emitted
fluorescence signal at 520 nm using the Fusion SL imager (Vilber
Lourmat).
Cell Culture and Treatment. Two human cell lines HeLa (ATCC
No. CCL-2) and HeLa-Luc (provided by Biontex Laboratories GmbH,
constitutively expressing the firefly luciferase) were cultured at 37 °C
and 5% CO2 either in RPMI 1640 or DMEM (Biochrom AG) with 1.5
g/L NaHCO3 both supplemented with 10% fetal bovine serum. Cells
were grown in 96-well plates (6 × 104 cells per well) for 24 h and
transfected using the glycine amphiphiles and Lipofectamine 2000
transfection reagent (Invitrogen) as control. One day before
transfection the glycine amphiphiles were dissolved in ddH2O.
Depending on the following experiments different double-stranded
siRNA molecules were used.
Intracellular Uptake. For the determination of transfection
efficiency of each glycine amphiphile the intracellular siRNA-uptake
was measured using fluorescently labeled nonsense siRNA (Cy3 dye-
Labeled Pre-miR Negative Control #1 - Life Technologies). 48 h after
transfection, efficiency of delivery was analyzed by fluorescence
microscopy.
Cell Viability Assays. Cytotoxicity and cell viability were measured
using a colorimetric WST-1 assay (Roche) and the xCELLigence
system (Roche) to monitor cell index profiles and determining cell
counts, proliferation and cytotoxicity. Both assays were performed in
96-well plates (6 × 104 cells per well). Each plate contained blanks,
controls (negative and positive), and substance dilution series with
four replicates. Nontargeting siRNA (ON-TARGETplus nontargeting
pool, D-001810-10, Thermo Fisher Scientific) transfection was
accomplished and WST-1 assay was performed according to the
manufacturer’s instructions.
RNA Interference and Transient Transfections. siRNAs targeting
GAPDH (ON-TARGETplus GAPD Control siRNA, D-001830-01,
Thermo Fisher Scientific) or Luciferase (CUUACGCUGAGUA-
CUUCGA, Eurofins MWG Operon) and nontargeting siRNA
(Thermo Fisher Scientific) as a control were used for performing
the transient transfections. As positive control for transfection, the
reagent Lipofectamine 2000 was used following the manufacturer’s
protocol. siRNA-transfection-reagent-complexes were prepared by
incubating 5 pmol siRNA with various amounts of aqueous glycine
amphiphiles depending on N/P ratios in OptiMEM (Gibco) for 30
min. Thereafter, the siRNA-transfection-reagent-complexes were
added to the cells. Both were incubated for at least 48 h at 37 °C
and 5% CO2. Samples were taken after transfection for RNA extraction
or luciferase assays.
RNA isolation was performed with the NucleoSpin RNA XS Kit
(Macherey-Nagel). Determination of RNA quality and RT-qPCR
assays were performed as described previously42 and the normalization
of the gene of interest (GAPDH; Gene ID: 2597; forward: 5′-
CCATCTTCCAGGAGCGAGAT-3′; reverse: 5′-CTAAG-
CAGTTGGTGGTGCAG-3′) was performed using geNorm.43 In
t h i s s t u d y B 2 M ( G e n e I D : 5 6 7 ; f o r w a r d : 5 ′ -
TGCTGTCTCCATGTTTGATGTATCT-3′; reverse: 5′-
TCTCTGCTCCCCACCTCTAAGT-3′), SDHA (Gene ID: 6389;
forward: 5′-TGGGAACAAGAGGGCATCTG-3′; reverse: 5′-CCAC-
CACTGCATCAAATTCATG-3′), and ACTB (Gene ID: 60; forward:
5′-GGACTTCGAGCAAGAGATGG′3; reverse: 5′-AGCACTGTG
TTGGCGTACAG-3′) were used as housekeeping genes.
For normalization of luciferase assays, cells were treated with
Calcein AM (Sigma Aldrich) to quantify the amount of viable cells.
After uptake of Calcein AM by living cells and acetoxymethyl ester
hydrolysis by cellular esterases, fluorescence is detected only in viable
cells, which enables the estimation of cell viability in a population. For
this purpose, the cell culture medium was removed and cells were
washed with PBS (PAA) and incubated with fresh medium containing
4 mM Calcein AM for 30 min at 37 °C and 5% CO2. After removal of
medium 100 μL PBS per well were added and fluorescence was
measured by excitation at 485 nm and acquiring the emission at 520
nm using the FLUOStar OPTIMA (BMG Labtech). An additional
washing step with PBS was done and 20 μL of lysis-juice (PJK GmbH)
per well was added. Measuring of luciferase activity was performed
according to the manufacturer’s manual.
In Vivo Toxicity. Three BALB/c mice per group were treated
intravenously (i.v.) with 8, 20, and 40 mg/kg G2-octaamine (4)
complexed with nontargeting siRNA (ON-TARGETplus Nontargeting
siRNA #1, Dharmacon) at N/P ratios of 50, 70, and 100, respectively.
HiPure water was administered i.v. as control. Retrobulbar blood was
taken 1 h after administration. Cytokine levels in the blood were
evaluated using the Meso Scale Discovery Multi-Spot Assay System,
Mouse ProInflammatory 7-Plex Assay Ultra-Sensitive Kit.
Synthesis of Compounds 5−8. The synthesis of compounds 5,
6, 7, and 8 were performed according to our previous report.44
General Procedure for the Esterification of Compounds 8
and 13. N-boc glycine (2.66 mmol), 4-(dimethylamino)pyridine
(20.0 mg), and the relavent compounds 8 (0.332 mmol) and 13
(0.166 mmol) were dissolved in 15 mL of DMF and cooled to 0 °C
under ice bath. l-Ethyl-3-[3-(dimethylamino) propyl]carbodiimide
hydrochloride (EDCI, 2.66 mmol) along with a few drops of triethyl
amine were added and the reaction mixture was stirred first at 0 °C for
2 h and then at room temperature, overnight. The solution was
concentrated to dryness in vacuum, and the residue was taken up in
chloroform (25 mL) and water (5 mL). The organic layer was
separated, washed with saturated sodium bicarbonate (2 × 15 mL) and
water (2 × 15 mL), and dried over MgSO4. The solvent was removed
in vacuum, and the products 9 and 14 were purified by column
chromatography (chloroform/methanol, 95:5) as viscous oils.
Compound 9. Obtained as light yellow viscous oil (318 mg, 78%
yield). 1H NMR (400 MHz, methanol-d4) δ 0.83−0.88 (3H, t, J = 8.0
Hz, −CH3), 1.24 (30H, s, methylene protons), 1.40 (36H, s,
−C(CH3)3), 1.51−1.59 (2H, m, −O−CH2CH2-), 3.24−3.28 (2H,
m), 3.45−3.50 (2H, m, PG dendron), 3.56−3.63 (3H, m, PG
dendron), 3.73 (8H, s, −CO−CH2-NHBoc), 3.86−3.93 (4H, m, PG
dendron), 4.12−4.25 (4H, m, PG dendron), 4.54 (2H, s), 5.14−5.18
(2H, m, PG dendron), 7.97−8.00 (1H, m, Triazole-H); 13C NMR
(100.5 MHz, methanol-d4) δ 15.61, 24.87, 26.39, 28.36, 29.90, 31.61,
31.78, 31.92, 34.20, 44.07, 44.19, 50.97, 63.20, 65.01, 65.77, 71.45,
72.20, 72.95, 73.17, 81.72, 126.26, 146.90, 159.46, 172.59, 172.85.
HRMS: m/z Calcd for C58H103N7O19Na: 1224.7223 [M+Na]+.
Found: 1224.7373. (For 1H and 13C NMR spectra, see the Supporting
Information.)
General Procedure for the Boc-Deprotection of Compounds
9, 14, and 15. Compounds 9, 14, and 15 (100 mg) were treated
overnight with a mixture of trifluoroacetic acid/dichloromethane
(1:3). The reaction mixture was evaporated under reduced pressure to
afford the tetra, octa, and diamino compounds 3, 4, and 16,
respectively, as viscous oils.
Compound 3. Obtained as colorless viscous oil (104 mg, quant.
1
yield). H NMR (400 MHz, methanol-d4) δ 0.87 (3H, t, J = 6.9 Hz),
1.26 (30H, s, methylene protons), 1.56−1.62 (2H, m, −O−CH2CH2-
), 3.52 (2H, t, J = 6.7 Hz), 3.58−3.65 (2H, m, PG dendron), 3.67−
3.76 (2H, m, PG dendron), 3.85 (8H, s, −CO−CH2-NHBoc), 3.87−
4.01 (5H, m, PG dendron), 4.23−4.27 (2H, m, PG dendron), 4.39−
4.41 (2H, m, PG dendron), 4.57 (2H,s), 5.23−5.31 (2H, m, PG
dendron), 7.95 (1H, s, Triazole-H); 13C NMR (100.5 MHz, methanol-
d4) δ 13.12, 22.40, 25.86, 26.40, 29.14, 29.29, 29.46, 30.37, 31.74,
35.69, 39.59, 39.71, 63.22, 63.46, 68.52, 69.85, 70.68, 71.84, 113.28
(TFA), 116.21 (TFA), 123.73, 144.72, 160.24, 160.61, 166.88, 167.08.
HRMS: m/z Calcd for C38H72N7O11: 802.5284 [M+H]+. Found:
3089
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Figure 1. Structures of the PG-amphiphiles investigated in this paper.
802.5295. (For 1H and 13C NMR spectra, see the Supporting
Information.)
Procedure for the Synthesis of Compounds 10 and 11. The
synthesis of compounds 10 and 11 were performed according to our
previous report.45
0.388 mmol) in methanol (5 mL) and stirred overnight at refluxing
temperature. After TLC analysis, dowex was filtered off and the solvent
was evaporated in vacuum. After column chromatography of the
obtained residue, pure product was obtained as a white sticky solid
(280 mg, 83%).
Compound 12. Compound 11 (1.00 g, 1.38 mmol, 1 equiv), 1-
prop-2-ynyloxy-octadecane (0.47 g, 1.52 mmol, 1.1 equiv), DIPEA
(0.14 mL, 0.83 mmol, 0.6 equiv), and bromotris (triphenylphosphine)-
copper(I) (0.26 g, 0.28 mmol, 0.2 equiv) were dissolved in dry THF
and the reaction mixture was left for stirring at 40 °C for 16 h. Since
TLC analysis did not indicate full consumption of the starting material
(generally at any point of time), the reaction mixture was treated with
saturated EDTA solution. Before doing so, the solvent was evaporated
and the residue dissolved in DCM. After aqueous workup the solvent
was removed under reduced pressure. The resulting residue was then
subjected to the same reaction conditions as initially employed
including also the same amounts of reactants referred to DIPEA, Cu
catalyst and THF. After completion of reaction, the mixture was
evaporated and again extracted with saturated EDTA solution in the
same manner as above-mentioned in order to remove any traces of
copper ions. The resulting crude product was purified by column
chromatography (CHCl3/MeOH, 99:1) yielding compound 12 as
yellowish viscous oil (0.91 g, 64%).
1H NMR (700 MHz, CD3OD): δ 0.97 (t, J = 7.1 Hz, 3 H, CH3),
1.35 (br s, 30 H, methylene CH2), 1.66 (quint, J = 7.8 Hz, 2 H, CH2-
CH2−O), 3.65−3.46 (m, 24 H, PG dendron), 3.68 (mc, 2 H, CH2−
CH2-O), 3.84−3.73 (m, 5 H, PG dendron), 4.22−3.97 (m, 5 H, PG
dendron), 4.65 (s, 2 H, O−CH2-trz), 5.04 (mc, 1 H, PG dendron),
8.18 (mc, 1 H, trz); 13C NMR (176 MHz, CD3OD): δ 14.4, 23.7, 27.2,
30.5, 30.6, 30.7, 30.8, 33.1, 62.8, 64.4, 79.7, 64.5, 64.7, 70.4, 70.5, 71.4,
71.8, 72.1, 72.2, 72.4, 72.5, 72.8, 72.9, 74.0, 80.0, 125.3, 145.7. HRMS:
m/z Calcd for C42H83N3O15Na: 892.5722 [M+Na]+. Found: 892.5742.
Compound 14. Obtained as colorless viscous oil (246 mg, 70%
1
yield). H NMR (400 MHz, CDCl3) δ 0.82−0.91 (3H, m, −CH3),
1.24 (30H, s, methylene protons), 1.43 (72H, s, −C(CH3)3), 1.54−
1.62 (2H, m, −O−CH2CH2-), 3.38−3.67 (19H, m, PG dendron and
−O-CH2CH2−), 3.87 (16H, s, −CO−CH2-NHBoc), 3.96−4.01 (2H,
m, PG dendron), 4.17−4.39 (8H, m, PG dendron), 4.47 (4H, m, PG
dendron), 4.59−4.61 (2H, m), 5.20−5.28 (4H, m, PG dendron), 7.74
− 7.79 (1H, m, Triazole-H); 13C NMR (100.5 MHz, CDCl3) δ 14.01,
22.60, 26.07, 27.80, 28.29, 29.27, 29.62, 31.85, 42.30, 63.07, 63.86,
68.34, 69.37, 70.83, 71.19, 78.51, 79.87, 84.63, 122.63, 144.76, 155.95,
169.94, 170.21. HRMS: m/z Calcd for C98H171N11NaO39: 2150.1667
1H NMR (700 MHz, CDCl3): δ 0.87 (t, J = 7.1 Hz, 3 H, alkyl CH3),
1.24 (br s, 30 H, methylene CH2), 1.34 (br s, 12 H, acetal CH3), 1.39
(br s, 12 H, acetal CH3), 1.58 (mc, 2 H, CH2-CH2−O), 3.65−3.38 (m,
20 H, PG dendron), 3.68 (mc, 4 H, PG dendron), 3.88 (mc, 2 H,
CH2−CH2-O), 4.02 (mc, 6 H, PG dendron), 4.24−4.16 (m, 4 H, PG
dendron), 4.59 (s, 2 H, O−CH2-trz), 4.85 (mc, 1 H, PG dendron),
7.73 (mc, 1 H, trz); 13C NMR (176 MHz, CDCl3): δ 14.2, 22.8, 25.5,
26.3, 26.9, 29.5, 29.7, 29.8, 32.1, 61.0, 64.5, 66.8, 66.9, 69.5, 70.4, 71.1,
71.3, 71.4, 71.6, 71.7, 71.8, 72.7, 74.8, 74.9, 78.6, 78.9, 109.5, 122.9,
145.1. HRMS: m/z Calcd for C54H99N3O15Na: 1052.6974 [M+Na]+.
Found: 1052.7005.
1
[M+Na]+. Found: 2150.1705. (For H and 13C NMR spectra, see the
Supporting Information.)
Compound 4. Obtained as colorless viscous oil (105 mg, quant.
1
yield). H NMR (500 MHz, CDCl3) δ 0.85−0.92 (3H, m, −CH3),
1.28 (30H, s, methylene protons), 1.56−1.64 (2H, m, −O−CH2CH2-
), 3.42−3.76 (22H, m, PG dendron and −O-CH2CH2−), 3.81−3.84
(1H, m, PG dendron), 3.89 and 3.91 (16H, 2s, −CO−CH2-NHBoc),
3.99−4.10 (2H, m, PG dendron), 4.30−4.42 (4H, m, PG dendron),
4.47−4.56 (4H, m, PG dendron), 4.55−4.61 (2H, m), 5.24−5.37 (4H,
m, PG dendron), 8.04 (1H, s, Triazole-H); 13C NMR (176 MHz,
MeOD) δ 12.77, 22.47, 25.10, 25.87, 26.41, 29.02, 29.30, 31.64, 39.53,
Compound 13. Acidic ion-exchange resin Dowex (200 mg) and a
few drops of water were added to a solution of compound 12 (400 mg,
3090
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Scheme 1. Synthesis of G1-Tetraamine (3)
Scheme 2. Synthesis of G2-Octaamine (4)
63.09, 63.60, 67.57, 68.63, 70.72, 72.00, 78.55, 115.52 (TFA), 117.17
Compound 15. Compound 15 was synthesized according to our
previously reported procedure.44
Compound 16. Obtained as light yellow viscous oil (102 mg, quant.
yield). 1H NMR (500 MHz, MeOD) δ 3.53−3.60 (6H, m), 3.67−3.77
(4H, m), 3.87 (4H, s), 3.99−4.01 (1H, m), 4.24−4.28 (2H, m), 4.34
(TFA), 124.74, 144.38, 160.89, 167.01. HRMS: m/z Calcd for
1
C58H107N11NaO23: 1348.7439 [M+Na]+. Found: 1348.7443. (For H
and 13C NMR spectra, see the Supporting Information.)
3091
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
(2H, dd, J = 11.3, 3.9 Hz); 13C NMR (126 MHz, MeOD) δ 39.73,
60.17, 60.69, 66.91, 67.87, 70.79, 71.86, 114.78 (TFA), 117.08 (TFA),
159.85 (TFA), 167.14. HRMS: m/z Calcd for C13H26N5O8: 380.1770
[M+H]+. Found: 380.1777.
trifluoroacetic acid/dichloromethane (1:3) to afford octaami-
nated amphiphile (G2-octaamine) 4, in quantitative yield as
trifluoroacetate salt (Scheme 2).
In order to investigate that the siRNA complexation ability of
the glycine functionalized amphiphiles 1−4 was indeed due to
the combined effect of the hydrophobic alkyl chain and the
hydrophilic glycine functionalized dendrons, we also synthe-
sized a reference dendron 16 having two surface amino groups
and without having a C-18 alkyl chain at the core (Scheme 3).
Compound 16 was obtained in quantitative yield by the
deprotection of boc-groups of compound 1544 using a mixture
of trifluoroacetic acid/dichloromethane (1:3) (Scheme 3).
RESULTS AND DISCUSSION
■
Synthesis. In order to correlate the number of amine
functionalities on the surface of the PG dendron based
amphiphiles with respect to siRNA complexation, delivery,
toxicity, and transfection, we decorated the surface of
bifunctional dendrons with different degrees of glycine loadings
in the shell and a C-18 hydrophobic alkyl chain at the focal
point. Thus, by controlling the number of amine functionaliza-
tion on the surface of PG-amphiphiles of generation 1 and 2,
viz. G1-monoamine (1), G1-diamine (2), G1-tetraamine (3),
and G2-octaamine (4) were synthesized (Figure 1). The
aminated compounds 1−4 were synthesized using facile
chemoenzymatic and classical organic synthesis. G1-mono-
amine (1) and G1-diamine (2) were synthesized according to
our previously reported procedure.44
Scheme 3. Synthesis of Dendron 16
For the synthesis of G1-tetraamine (3), we had first prepared
the deprotected amphiphile 8 according to our previous
report.44 Starting from [G1.0]−OH (5), the free secondary
hydroxyl group of [G1.0]−OH (5) was converted to the
corresponding mesylate and without further purification treated
with sodium azide to give [G1.0]−N3 (6) in 96% yield over
two steps (Scheme 1). The acetal protecting groups of
compound 6 were removed in 90% yield by refluxing with an
acidic ion-exchange resin in methanol to obtain compound 7.
We then performed the click reaction on compound 7 with
octadecyl-propargylether using copper triphenylphosphine
bromide, diisopropylethylamine and N,N-dimethylformamide
as a solvent at 50 °C for 24 h to afford G1 click amphiphile 8 in
83% yield. The hydroxyl groups of G1 click amphiphile 8 were
then decorated via esterification, using N-Boc-glycine as the
acylating agent in the presence of condensation reagent EDCI
and DMAP as a base at room temperature in DMF solvent to
furnish tetra boc-protected compound 9 in 78% isolated yield
(Scheme 1). Tetra boc-protected compound 9 was then treated
overnight with a mixture of trifluoroacetic acid/dichloro-
methane (1:3) to afford G1-tetraamine (3), in quantitative
yield as trifluoroacetate salt (Scheme 1).
Regarding the synthesis of G2-octaamine (4) (Scheme 2), we
had first prepared the compound [G2.0]−OH (10) according
to our previous report.45 As described above, the free secondary
hydroxyl group of [G2.0]−OH (10) was also converted to the
corresponding mesylate and without further purification treated
with sodium azide to give [G2.0]−N3 (11) in 85% yield over
two steps. We then carried out the click reaction on compound
11 with octadecyl-propargylether (as mentioned above for
compound 7) using copper triphenylphosphine bromide,
diisopropylethylamine and N,N-dimethylformamide as a
solvent at 50 °C for 16 h to afford G2 click product 12 in
64% yield.
Physico-Chemical Characterization of Amphiphiles.
Amphiphilic structures with long alkyl groups form stable
aggregates and are therefore useful for transporting e.g. genetic
material, particularly if the formed supramolecular architectures
exhibit a net positive charge on the surface to complex the
negatively charged siRNA. To examine the characteristics of the
dendritic amphiphiles to be compared, the aggregation behavior
as well as the size and zeta potential were studied using
different physical characterization methods such as fluorescence
spectroscopy, DLS, and zeta potential measurements.
Initially, the aggregation behavior of the four amphiphiles
was studied by means of fluorescence spectroscopy employing
hydrophobic fluorescent dye DPH in HEPES saline buffer
solution.37−40 It is known that above a particular concentration,
called the critical micelle concentration (CMC), amphiphilic
molecules self-associate to form thermodynamically stable
micellar aggregates.37−40 We have chosen a neutral fluorescent
probe (DPH) for the determination of CMC’s so as to avoid
undesirable ionic interactions with our positively charged
amphiphiles. At concentrations below CMC, DPH exists
predominantly in an aqueous environment and exhibits low
fluorescence. Formation of micelles results in the preferential
partitioning of the DPH into the hydrophobic interior of the
micelles with a concomitant sharp increase in fluorescence.
Figure 2 shows that as the concentration of the amphiphile is
increased, fluorescence is weak at the lowest concentrations,
then rises (first end-point), and finally increases sharply
(second end-point). This is interpreted as follows: The first
rise in fluorescence (first end-point) occurs at the premicellar
concentration by the formation of small aggregates between the
probe molecule and the surfactant, and the second rise in
fluorescence (second end-point) occurs at and above the CMC
of the amphiphile. As the amount of amphiphile is increased,
the number of micelles and the amount of bound DPH
increase, so that fluorescence also increases. The CMC is given
by the intersection of the straight line through the fluorescence
at low detergent concentrations with a straight line through the
fluorescence values in the region of rapid intensity increase.
The determined CMC values, given in Table 1, showed that all
four amphiphiles aggregate at a micromolar level ranging from
10 to 60 μM. The CMC values are comparable to our
The acetal protecting groups were removed in 83% yield by
refluxing with an acidic ion-exchange resin dowex in methanol
to obtain the water-soluble core−shell architecture 13. The
eight hydroxyl groups of the G2 click amphiphile 13 were then
completely esterified, again by using N-Boc-glycine as the
acylating agent in the presence of EDCI and DMAP at room
temperature in DMF solvent to furnish octa boc-protected
compound 14 in 70% isolated yield. Octa boc-protected
compound 14 was then treated overnight with a mixture of
3092
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Figure 2. Determination of CMC of glycine amphiphiles in 0.5 μM DPH (aqueous HEPES saline (pH 7.2, 9.4 mM NaCl).
However, this effect appears to be of minor importance with
regard to the concentration of DPH employed here.38−40
The size (hydrodynamic diameter) of the aggregates and
their relative size distribution (PDI: polydispersity index) were
determined by means of dynamic light scattering (DLS)
measurements in HEPES saline buffer (pH 7.2, 2 mM HEPES,
9.4 mM NaCl). As shown in Table 1, all four amphiphiles
formed small micelles in the range of 7−9 nm. The
hydrodynamic diameters of all four micellar aggregates
measured by DLS (7−9 nm) correlate very well with the
theoretical value of around 6−8 nm, since the Z-average value
displays the size of the micelle including its hydrodynamic
water shell, which is slightly larger than the pure amphiphile
size by definition (3−4 nm). The corresponding graphs are
shown in the Supporting Information as Figure S3.
According to the conducted zeta potential measurements
(Table 1), all four micellar constructs showed positive charges
within a range of 40−58 mV. This can be attributed to the large
number of cationic amine groups, which are most probably
located at the surface of the micellar aggregates. Interestingly,
G1-monoamine (1) has shown maximum positive charge on
the surface which could be due to less repulsion of positive
charge on the surface of G1-monoamine (1).
Table 1. CMC, Size, and Zeta-Potential of the Synthesized
Glycine Amphiphiles
a
a,
b
a
CMC
size (solo)
ζ-potential (solo)
amphiphile
[μM]
dH [nm]
[mV]
G1-monoamine (1)
G1-diamine (2)
G1-tetraamine (3)
G2-octaamine (4)
DNA
10
26
60
58
8.4 ( 0.3)
7.3 ( 0.4)
6.4 ( 0.2)
8.2 ( 0.9)
57.3 ( 0.4)
40.1 ( 3.2)
47.5 ( 2.7)
58.1 ( 1.3)
−41
a
Micellar dispersions in HEPES saline buffer (pH 7.2, 9.4 mM NaCl).
b
T = 25 °C. Size-distribution by DLS (by volume), PDI = ∼0.3−0.5.
previously reported nonionic derivatives36 using also PG-
derived, hydrophobically modified amphiphiles.
The observed variations of up to 1 order of magnitude can be
attributed to the fact that cationic amphiphiles typically exhibit
higher CMC values due to repulsion forces induced by the
incorporated charged moieties in the headgroup of the
amphiphiles. Also, the CMC values increased from G1-
monoamine (1) to G1-tetramine (3) which is in accordance
considering the number of positive charges which induces more
repulsion in the head groups. In order to achieve a good signal-
to-noise ratio the concentration of DPH was kept constant (0.5
μM) and adjusted according to the CMCs of the amphiphiles.
In this context, it of interest to note that the critical micelle
concentration for probe molecules that are incorporated in the
micelles may be dependent on the concentration and the size of
the probe molecules. Hence, the presence of DPH might affect
the aggregation process and disturb the micellar aggregates.
DNA Binding and Condensation. The process of binding
and condensation of DNA represents a prerequisite of nonviral
gene therapy, since it facilitates the cellular uptake of the
genetic material.
Complexation through cationic carriers is mainly based on
electrostatic interactions between the opposite charges of the
DNA and the gene vehicle. Cationic complexes are favored by
cell membranes owing to negatively charged membrane
3093
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
components, thus encouraging cell surface binding through
electrostatic interactions. We used double-stranded DNA with
21 base pairs instead of siRNA as a model compound to
evaluate the gene binding affinities of the dendritic structures.
The DNA binding and complexation properties of the
glycine loaded amphiphiles 1−4 were initially studied with an
EthBr displacement assay. This assay utilizes the competition
between the DNA binding ligands and EthBr for binding to
DNA. EthBr exhibits intense fluorescence when bound to
DNA, but when it is displaced from the DNA by a DNA-
binding ligand, its fluorescence is quenched. Notably, the
fluorescence of EthBr can increase again if reintercalation in
DNA becomes possible. Binding values are best described by
CE50 values, which report the relevant N/P ratio indicating the
nominal dendron charge excess causing a 50% decrease in
fluorescence intensity. The concentration of dendritic amphi-
phile required for effective DNA binding can also be calculated
(C50).
μM), even though the former has eight surface positive charges
and the latter has only four.
Thus, it can be concluded that G1-tetraamine (3) is the most
effective DNA binder in terms of the CE50 parameter, which
reflects the relative ability of the cationic amine groups to bind
anionic DNA. The corresponding titration graphs are shown in
the Supporting Information in Figure S2.
Furthermore, we confirmed that the DNA complexation
ability of the glycine functionalized amphiphiles 1−4 was
indeed due to the combined effect of the hydrophobic alkyl
chain and the hydrophilic glycine functionalized dendrons,
since the dendron 16 alone did not led to any DNA binding
effect. As revealed from the Ethbr displacement assay (data not
shown), dendron 16 was not able to displace even 20% of the
Ethbr from the DNA even at an N/P ratio of 5, thus revealing
no complexation ability. The combined effect of the hydro-
phobic alkyl chain and the cationic hydrophilic glycine
functionalized dendron was able to impart 1−4 with a strong
capacity to form stable self-assembled complexes with DNA
and effectively displace Ethbr from the DNA, thus giving good
binding values (CE50) in the range of 0.8−1.3. This indicates
that the alkyl chain considerably favors the DNA/vector
assembly and increases their stability via hydrophobic
interactions.
The surface charge of the dendron−DNA polyplexes was
then evaluated by zeta-potential (ζ) measurements (see Table
3), which were carried out at 9.4 mm NaCl concentration to
allow accurate measurements. We have used double-stranded
DNA with 21 base pairs for zeta-potential (ζ) measurements as
well. Negative ζ potentials close to −41 mV were observed for
free DNA (see Table 1).
We studied the DNA-binding properties at low salt
concentration in HEPES saline buffer (pH 7.2, 2 mM
HEPES, 9.4 mM NaCl). The data shown in Table 2
Table 2. DNA Binding Data Extracted from EthBr
Displacement Assays Using HEPES Saline Buffer (pH 7.2,
9.4 mM NaCl)
a
b
amphiphile
CE50 value
C50 value [μM]
G1-monoamine (1)
G1-diamine (2)
1.3
1.0
0.8
1.3
3.3
1.3
0.5
0.8
G1-tetraamine (3)
G2-octaamine (4)
a
CE50 represents the charge excess (N/P ratio) required to decrease
b
The obtained ζ potentials of the polyplexes showed the
general trend that, along with rising N/P ratios, the ζ potential
increases, which is in complete agreement with theoretical
expectations. All four amphiphilic constructs showed positive ζ
potentials in the range of 20−43 mV at N/P 10. However, only
G1-monoamine (1) and G1-diamine (2) derived micelles
exhibited positive net charges at the N/P ratio of 5 as well.
In addition, the size of amphiphile-DNA polyplexes of
different N/P ratios were analyzed by dynamic light scattering
(DLS) measurements in order to provide insight into the
dimensions of the formed polyplexes which later on should be
internalized by the cells, thereby facilitating nucleic acid
delivery. We have used double-stranded DNA with 21 base
pairs for DLS studies as well. Generally, two different N/P
ratios were tested: N/P 5 and 10 (see Table 3). The
hydrodynamic diameters, ranging from approximately 69−306
nm at N/P ratio 10, were obtained for all four amphiphiles
tested, whereas large aggregates in the range of 212−1383 nm
were observed at N/P ratio 5. The corresponding DLS graphs
are shown in Supporting Information as Figures S4 and S5.
EthBr fluorescence by 50%. C50 represents the concentration of
amphiphile required to displace 50% of EthBr.
demonstrate that all of our four glycine loaded amphiphiles
effectively bind to DNA although G1-tetraamine (3) had the
highest DNA binding affinity in comparison to the CE50 as well
as the C50 values of all tested amphiphiles, since already at a N/
P ratio of 0.8 (CE50) fifty percent of the ethidium bromide
intercalated in the DNA could be replaced. In addition, a
positive trend within the G1 derivatives from the relevant
monoamine to the fully substituted tetraamine is noticeable,
which clearly can be attributed to multivalency effects. Hence,
within this testing series G1-monoamine (1) has the lowest
ability to bind DNA (CE50 = 1.3). Interestingly, G2-octaamine
(4) did not show better results (CE50 = 1.3) than their G1
analogues, even though this would have been expected when
considering multivalent aspects. In fact, only in concentration
terms, the performance of G2-octaamine (4) (C50 = 0.8 μM) is
relatively comparable to that of G1-tetraamine (3) (C50 = 0.5
Table 3. Size and Zeta-Potential of the Glycine Amphiphiles−DNA Polyplexes at Different N/P Ratios
a
size (dH) [nm]
ζ-potential [mV]
amphiphile
N/P 5
N/P 10
N/P 5
N/P 10
G1-monoamine (1)
G1-diamine (2)
211.8 ( 10.4)
1249 ( 36.9)
1383 ( 51.3)
1181 ( 12.0)
141.1 ( 2.8)
80.7 ( 0.5)
68.9 ( 0.8)
305.9 ( 2.8)
+26.5 ( 0.9)
+4.9 ( 0.1)
−12.0 ( 1.5)
−10.2 ( 2.6)
+43.1 ( 1.3)
+37.9 ( 1.1)
+42.3 ( 0.6)
+20.6 ( 3.9)
G1-tetraamine (3)
G2-octaamine (4)
a
Amphiphile−DNA polyplexes in HEPES saline buffer (pH 7.2, 9.4 mM NaCl). T = 25 °C. Size-distribution by DLS (by volume), PDI = ∼0.1−0.3.
3094
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Figure 3. Binding capacity and migration assay of glycine amphiphiles-FAM-siRNA complexes by agarose gel electrophoresis. G1-monoamine (1),
G1-diamine (2), G1-tetraamine (3), and G2- octaamine (4) are shown with increased N/P ratios (20, 50, 70, 90, and 100) complexed with FAM-
labeled siRNA. Complex-free FAM-siRNA is illustrated in the last slot on the right. Images were acquired by detecting the green fluorescence signal.
The binding capacity of all glycine amphiphiles was further
demonstrated using agarose gel electrophoresis retardation
assay, which directly measures siRNA-dendron interactions. For
this purpose, we used FAM-labeled siRNA. As a sign of
successful siRNA neutralization and compaction, the constructs
(amphiphiles) should either reduce or completely retard the
electrophoretic mobility of siRNA. As shown in Figure 3, all
glycine amphiphiles exhibited enhanced binding capacity with
increased N/P ratios while monoamine was the most
inefficient. The complexes of FAM-siRNA with tetraamine
mostly remained in the slots indicating proper binding of
siRNA molecules. Di- and octaamine derivatives possessed
similar and moderate binding capacities as shown by migration
assays. Complex-free FAM-siRNA was loaded as migration
control. The results were visualized by detecting the green
fluorescence signal.
Thus, from the nucleic acids complexation studies it could be
concluded that all four glycine functionalized amphiphiles (1−
4) were successfully able to complex nucleic acids with high
affinities. Tetraamine was found to be the best DNA binder.
Cytotoxicity and siRNA Transfection Results. The
cytotoxic effect of each glycine amphiphile for N/P ratios 20,
50, and 100 was tested by WST-1 assays and xCELLigence
(Figure 4) using HeLa cells. The results of WST-1 assays
pointed out high toxicity of (A) G1-monoamine (1) with N/P
ratios higher than 50 and of (B) G1-diamine (2) using N/P
100. (C) G1-tetraamine (3) and (D) G2-octaamine (4)
indicated no lethal effects. These findings were confirmed by
real-time determination of cell index profiles (E-H) obtained
from xCELLigence. The graphs demonstrate the baseline delta
cell indexes related to the control cells (control+siRNA, Figure
4). Control cells were incubated with siRNA but without
transfection reagent. Incubation with lipofectamine (Figure 4)
in comparison with the control cells displayed no difference in
cell proliferation over the observed time period.
transfer of siRNA with G1-tetraamine (3) was effective at all
investigated N/P ratios.
G2-octaamine (4) also turned out to be an efficient
nanotransporter for fluorescently labeled siRNA at ratios higher
than N/P 20 (Table 4). Lipofectamine was used as a control
transfection reagent and showed efficient uptake in all
experiments.
Further functional analysis of all glycine amphiphiles was
performed by investigating the knockdown of normalized
luciferase activity in HeLa cells constitutively expressing the
luciferase gene. Figure 5 illustrates normalized luciferase
knockdown using (A) G1-monoamine (1), (B) G1-diamine
(2), (C) G1-tetraamine (3), and (D) G2-octaamine (4) at N/P
ratios of 20, 50, 70, 90, and 100 for luciferase specific siRNA
delivery after 48 h.
The relative luminescence was determined by measuring the
luciferase activity per living cell relative to the Calcein AM
signal. This approach provided a simultaneous measure for
transfection efficiency and cytotoxicity of investigated nano-
transporters within the same assay. This experiment revealed
that both G1-mono-(1) and diamine (2) are both toxic
independently of investigated N/P ratio compared with control
cells. Both the control+siRNA and control were without
transfection reagent. G1-monoamine (1) at a ratio of N/P 20
exhibited low knockdown of luciferase in comparison to the
transfection data of nontargeting (nT) siRNA while diamine
showed no obvious downregulation. Transfections with G1-
tetraamine (3) up to N/P 70 led to luciferase knockdown while
increased N/P ratios at 90 and 100 caused toxic effects. G2-
octaamine (4) had no toxic impact on cells and exhibited the
most efficient combination of low toxicity and high knockdown
efficiency regarding the entire spectrum of investigated N/P
ratio.
Furthermore, the toxicity of G2-octaamine (4) complexed
with siRNA was tested in vivo by intravenous administration.
No significant increase in the tested proinflammatory cytokine
levels compared to the control was detected in the serum
reflecting the nontoxic effect of G2-octaamine (4) detected in
vitro (see the Supporting Information, Figure S6).
The knockdown of GAPDH after siRNA delivery in HeLa
cells was proven by RT-qPCR (Figure 6). Due to the cytotoxic
effects caused by G1-monoamine (1), G1-diamine (2) and G1-
Table 4 summarizes the outcome of the Cy3-labeled siRNA
delivery using the glycine amphiphiles. HeLa cells were
transfected with fluorescently labeled siRNA and efficiency of
delivery was analyzed after 48 h by fluorescence microscopy.
G1-monoamine (1) and G1-diamine (2) showed positive
uptake of labeled siRNA using a ratio of N/P 20. Higher N/P
ratios at 50 and 100 caused cytotoxic effects. Intracellular
3095
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Figure 4. Cytotoxicity by WST-1 assay and xCELLigence (Roche). (A−D) Results of WST-1 assay and (E−H) of xCELLigence with nontargeting
siRNA transfected HeLa cells. G1-monoamine (1) is shown in panels A and E, G1-diamine (2) in B and F, G1-tetraamine (3) in C and G, and G2-
octaamine (4) in D and H. The N/P ratios 20, 50, and 100 were tested. Cells treated with siRNA and without transfection reagent were used as
reference. Lipofectamine was used as a control transfection reagent.
experiments using G2-octaamine (4) at a ratio of N/P 50.
Knockdown efficiency of Lipofectamine was around 50%.
Increased nonspecific knockdown was observed for G2-
octaamine (4) with N/P 20, 50, and 100, which is evidence
of an increase in the cytotoxicity from the polyplexes.
After observing the successful luciferase and GAPDH gene
silencing using G2-octaamine (4), we recorded the size and
zeta-potential of the polyplexes of G2-octaamine (4) and
siRNA (ON-TARGETplus GAPD Control siRNA, D-001830-
01, Thermo Fisher Scientific) at different N/P ratios in aqueous
HEPES saline buffer (9.4 mM NaCl, pH 7.4). We observed the
size of the G2-octaamine (4)-siRNA polyplex to be around
Table 4. Cy3-Labeled siRNA Delivery with Glycine
Amphiphiles (+ = Successful, − = Failed)
amphiphile
N/P 20
N/P 50
N/P 100
G1-monoamine (1)
G1-diamine (2)
+
+
+
−
toxic
toxic
+
toxic
toxic
+
G1-tetraamine (3)
G2-octaamine (4)
+
+
tetraamine (3) at high N/P ratios, the results of respective
knockdown experiments are shown for N/P ratio 20 and for
G2-octaamine (4) N/P 20, 50, and 100. No significant
downregulations were observed by N/P 20 for all glycine
amphiphiles. GAPDH possessed 60% downregulation in
118.4
1.7 at an N/P ratio of 50 (see Table S1 in the
3096
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
Figure 5. Normalized luciferase assay of HeLa-Luc cells. HeLa-Luc cells were transfected with luciferase specific and nontargeting (nT) siRNA for 48
h. The results pointed out the relative luminescence for (A) G1-monoamine (1), (B) G1-diamine (2), (C) G1-tetraamine (3), and (D) G2-
octaamine (4) at N/P ratios of 20, 50, 70, 90, and 100. Outcomes of controls (cells with and without siRNA) and transfections with Lipofectamine
are represented in each figure. The normalization was performed using data of simultaneous Calcein AM measurement. (Statistical data by paired t
test; * p < 0.05, ** p < 0.01.)
(glycine loading) on the surface of dendritic headgroup make
the vector more efficient in terms of siRNA transfection and
cytotoxicity. Furthermore, we found that a second-generation
amphiphilic dendrimer (G2-octaamine, 4) with eight amine
groups on its surface and a hydrophobic C-18 alkyl chain at the
core, acts as an efficient vector to deliver siRNA inside the cell
and achieved potent gene silencing as demonstrated by the
knockdown of normalized luciferase activity and also for
GAPDH in HeLa cells. The amphiphilic vector is nontoxic even
at higher ratio of N/P 100 both in vitro and in vivo. To the best
of our knowledge, this is the first example of successful in vitro
siRNA transfection using dendritic amphiphiles. It may serve as
Figure 6. Quantification of GAPDH mRNA by RT-qPCR. Values of
a new promising alternative for nonviral siRNA delivery system
relative knockdown of GAPDH transcripts are shown after 48 h of
and be useful for further applications in nanobiotechnology.
transfection with GAPDH specific and nontargeting (nT) siRNA in
HeLa cells. The data show the results of G1-monoamine (1), G1-
ASSOCIATED CONTENT
diamine (2), and G1-tetraamine (3) with N/P 20 and of G2-
octaamine (4) with N/P 20, 50, and 100 as well as Lipofectamine as
control transfection. (Statistical data by paired t test; * p < 0.05, ** p <
0.01.)
■
S
* Supporting Information
1H and 13C NMR spectra of compounds 9, 3, 14, and 4. Graphs
and data for the physicochemical characterization of com-
pounds 1−4 and also for their polyplexes with DNA and
siRNA, and the in vivo toxicity graphs for compound 4-siRNA
polyplex. This material is available free of charge via the
Internet at http://pubs.acs.org.
Supporting Information), which suggests that the polyplex at
this N/P ratio can be easily internalized by the cell.
CONCLUSIONS
■
In conclusion, we have developed multivalent oligoglycerol
dendron based low molecular weight amphiphiles with well-
defined molecular structures that express controlled glycine
arrays on their surfaces. In this study, we have controlled the
loading of amine content on the surface of the dendritic head-
groups, using facile chemo-enzymatic and chemical synthetic
routes. We studied the structure−activity relationships with
respect to the siRNA/DNA complexation, toxicity, and
transfection profiles with the synthesized polycations. Our
findings disclose that a higher number of amine functionalities
AUTHOR INFORMATION
■
Corresponding Author
*(R.H.) Phone: +49-30-838-52633. Fax: +49-30-838-53357. E-
mail: haag@chemie.fu-berlin.de. Website: http://www.polytree.
de. (R.E.) Phone: +49-30-838-62575. Fax: +49-30-838-62584.
E-mail: einspani@zedat.fu-berlin.de.
Notes
The authors declare no competing financial interest.
3097
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098

Biomacromolecules
Article
(31) Bayele, H. K.; Sakthivel, T.; O’Donell, M.; Pasi, K. J.;
Wilderspin, A. F.; Lee, C. A.; Toth, I.; Florence, A. T. J. Pharm. Sci.
2005, 94, 446−457.
(32) Jones, S. P.; Gabrielson, N. P.; Wong, C.-H.; Chow, H.-F.; Pack,
D. W.; Posocco, P.; Fermeglia, M.; Pricl, S.; Smith., D. K. Mol. Pharm.
2011, 8, 416−429.
(33) Poon, Z.; Lee, J. A.; Huang, S.; Prevost, R. J.; Hammond, P. T.
Nanomedicine 2011, 7, 201−209.
(34) Karmali, P. P.; Chaudhuri, A. Med. Res. Rev. 2007, 27, 696−722.
(35) Bhattacharya, S.; Bajaj, A. Chem. Commun. 2009, 4632−4656.
(36) Trappmann, B.; Ludwig, K.; Radowski, M. R.; Shukla, A.; Mohr,
ACKNOWLEDGMENTS
■
The authors acknowledge the support from the Biotransporter
project of the Bundesministerium fur Bildung und Forschung
(BMBF).
̈
REFERENCES
■
(1) Zhang, X.; McIntosh, T. J.; Grinstaff, M. W. Biochimie 2012, 94,
42−58.
(2) Cavazzana-Calvo, M.; Hacein-Bey, S.; De.Saint Basile, G.; Gross,
F.; Yvon, E.; Nusbaum, P.; Selz, F.; Hue, C.; Certain, S.; Casanova, J.
L.; Bousso, P.; Le Deist, F.; Fischer, A. Science 2000, 288, 669−672.
(3) Yang, Z. R.; Wang, H. F.; Zhao, J.; Peng, Y. Y.; Wang, J.; Guinn,
B.-A.; Huang, L. Q. Cancer Gene Ther. 2007, 14, 599−615.
(4) Niven, R.; Pearlman, R.; Wedeking, T.; Mackeigan, J.; Noker, P.;
Simpson-Herren, L.; Smith, J. G. J. Pharm. Sci. 1998, 87, 1292−1299.
(5) Anderson, W. F. Nature 1998, 392, 25−30.
A.; Rehage, H.; Bottcher, C.; Haag, R. J. Am. Chem. Soc. 2010, 132,
̈
11119−11124.
(37) Purohit, V. S.; Balasubramanian, S. V. AAPS J. 2006, 8, E362−
370.
(38) Chattopadhyay, A.; London, E. Anal. Biochem. 1984, 139, 408−
412.
(39) Nakahara, Y.; Kida, T.; Nakatsuji, Y.; Akashi, M. Langmuir 2005,
21, 6688−6695.
(40) Mohr, A.; Talbiersky, P.; Korth, H.-G.; Sustmann, R.; Boese, R.;
(6) Verma, I. M.; Somia, N. Nature 1997, 389, 239−242.
(7) Felgner, P. L.; Gadek, T. R.; Holm, M.; Roman, R.; Chan, H. W.;
Wenz, M.; Northrop, J. P.; Ringold, G. M.; Danielsen, M. Proc. Natl.
Acad. Sci. U.S.A. 1987, 84, 7413−7417.
(8) Fukushima, S.; Miyata, K.; Nishiyama, N.; Kanayama, N.;
Yamasaki, Y.; Kataoka, K. J. Am. Chem. Soc. 2005, 127, 2810−2811.
(9) Harada, A.; Kawamura, M.; Matsuo, T.; Takahashi, T.; Kono, K.
Bioconjugate Chem. 2006, 17, 3−5.
(10) Gupta, B.; Levchenko, T. S.; Torchilin, V. P. Adv. Drug Delivery
Rev. 2005, 57, 637−651.
(11) Behr, J. P. Acc. Chem. Res. 1993, 26, 274−278.
(12) Smith, D. K. Curr. Top. Med. Chem. 2008, 8, 1187−1203.
(13) Badjic, J. D.; Nelson, A.; Cantrill, S. J.; Turnbull, W. B.;
Stoddart, J. F. Acc. Chem. Res. 2005, 38, 723−732.
(14) Kitov, P. I.; Bundle, D. R. J. Am. Chem. Soc. 2003, 125, 16271−
16284.
Blaser, D.; Rehage, H. J. Phys. Chem. B 2007, 111, 12985−12992.
̈
(41) Kostiainen, M. A.; Smith, D. K.; Ikkala, O. Angew. Chem. 2007,
119, 7744−7748.
(42) Sharbati, J.; Lewin, A.; Kutz-Lohroff, B.; Kamal, E.; Einspanier,
R.; Sharbati, S. PLoS ONE 2011, 6, e20258.
(43) Vandesompele, J; De Preter, K; Pattyn, F; Poppe, B; Van Roy,
N; De Paepe, A; Speleman, F. Genome Biol. 2002, 3, r0034.1−0034.11.
́
(44) Malhotra, S.; Calderon, M.; Prasad, A. K.; Parmar, V. S.; Haag,
R. Org. Biomol. Chem. 2010, 8, 2228−2237.
(45) Wyszogrodzka, M.; Haag, R. Chem.Eur. J. 2008, 14, 9202−
9214.
(15) Kukowska-Latallo, J.-F.; Bielinska, A. U.; Johnson, J.; Spindler,
R.; Tomalia, D. A.; Baker, J. R. Proc. Natl. Acad. Sci. 1996, 93, 4897−
4902.
(16) Turunen, M. P.; Hiltunen, M. O.; Ruponen, M.; Virkamaki, L.;
Szoka, F. C.; Urtti, A.; Yla-Herttuala, S. Gene Ther. 1999, 6, 6−11.
(17) Ohsaki, M.; Okuda, T.; Wada, A.; Hirayama, T.; Niidome, T.;
Aoyagi, H. Bioconjugate Chem. 2002, 13, 510−517.
́
(18) Fischer, W.; Calderon, M.; Schulz, A.; Andreou, I.; Weber, M.;
Haag, R. Bioconjugate Chem. 2010, 21, 1744−1752.
(19) Ofek, P.; Fischer, W.; Calderon, M.; Haag, R.; Satchi-Fainaro, R.
FASEB J. 2010, 3122−3134.
(20) Germershaus, O.; Pickaert, G.; Konrad, J.; Kruger, U.; Kissel, T.;
̈
Haag, R. Macromol. Biosci. 2010, 10, 1055−1062.
(21) Fischer, W.; Calderon, M.; Haag, R. Hyperbranched Polyamines
for Transfection. In Nucleic Acid Transfection; Springer-Verlag: Berlin,
2010; Vol. 296, pp 95−129.
(22) Fischer, W.; Quadir, M. A.; Barnard, A.; Smith, D. K.; Haag, R.
Macromol. Biosci. 2011, 11, 1736−1746.
(23) Wan, K.; Ebert, B.; Voigt, J.; Haag, R.; Kemmner, W. Nanomed.
Nanotech. Biol. Med. 2012, 8, 393−398.
(24) Khandare, J.; Calderon, M.; Dagia, N.; Haag, R. Chem. Soc. Rev.
2012, 41, 2824−2848.
(25) Calderon, M.; Quadir, M. A.; Sharma, S. K.; Haag, R. Adv.
Mater. 2010, 22, 190−218.
́
(26) Quadir, M. A.; Calderon, M.; Haag, R. Cancer Drug Delivery;
Wiley-VCH Books: Weinheim, Germany, 2011; pp 513−553.
(27) Lv, H.; Zhang, S.; Wang, B.; Cui, S.; Yan, J. J. Controlled Release
2006, 114, 100−109.
(28) Kuo, J.-H. S.; Liou, M.-J.; Chiu, H.-C. Mol. Pharm. 2010, 7,
805−814.
(29) Guillot, M.; Eisler, S.; Weller, K.; Merkle, H. P.; Gallani, J. L.;
Diederich, F. Org. Biomol. Chem. 2006, 4, 766−769.
(30) Barnard, A.; Posocco, P.; Pricl, S.; Calderon, M.; Haag, R.;
Hwang, M. E.; Shum, V. W. T.; Pack, D. W.; Smith, D. K. J. Am. Chem.
Soc. 2011, 133, 20288−20300.
3098
dx.doi.org/10.1021/bm300892v | Biomacromolecules 2012, 13, 3087−3098