Solid-Phase Synthesis of 89 Polyamine-Based Cationic Lipids for DNA Delivery to Mammalian Cells

Article abstract of DOI:10.1002/chem.200305232

The ability of non-viral gene delivery systems to overcome extracellular and intracellular barriers is a critical issue for future clinical applications of gene therapy. In recent years much effort has been focused on the development of a variety of DNA c

Full text of DOI:10.1002/chem.200305232

FULL PAPER
Solid-Phase Synthesis of 89 Polyamine-Based Cationic Lipids for DNA
Delivery to Mammalian Cells
Boon-ek Yingyongnarongkul,[a] Mark Howarth,[b] Tim Elliott,[b] and Mark Bradley*[a]
Abstract: The ability of non-viral gene
delivery systems to overcome extracel-
lular and intracellular barriers is a criti-
cal issue for future clinical applications
of gene therapy. In recent years much
effort has been focused on the develop-
ment of a variety of DNA carriers, and
cationic liposomes have become the
most common non-viral gene delivery
system. Solid-phase synthesis was used
to produce three libraries of poly-
amine-based cationic lipids with diverse
hydrophobic tails. These were charac-
terised, and structure-activity relation-
ships were determined for DNA bind-
ing and transfection ability of these
compounds when formulated as cation-
ic liposomes. Two of the cationic lipids
produced high-efficiency transfection
of human cells. Surprisingly, these two
compounds were from the library with
two headgroups and one aliphatic tail,
a compound class regarded as deter-
gent-like and little investigated for
transfection. These cationic lipids are
promising reagents for gene delivery
and illustrate the potential of solid-
phase synthesis methods for lipoplex
discovery.
Keywords: cationic lipids ¥ combi-
natorial chemistry ¥ gene therapy ¥
non-viral vector ¥ solid-phase syn-
thesis
Introduction
DEAE-dextran[7] and cationic lipid-mediated transfection.[8]
Of all the non-viral vectors, cationic lipids have shown the
most promise for in vivo applications, based on a combina-
tion of efficiency, stability and lack of toxicity. Since the first
application of cationic lipids in DNA delivery,[9] numerous
cationic lipids have been synthesised. Some of these have
been used in gene therapy clinical trials,[10,11] while many
others have established cationic lipids as the most common
method for the transfection of cell lines in the laboratory.
Transfection of DNA into mammalian cells involves nu-
merous steps.[12] The DNA is first compacted by the cationic
transfection agent; the positively charged transfection com-
plex may then undergo electrostatic interactions with the
negatively charged cell surface. Uptake into the cell seems
to be principally by endocytosis, and a small fraction of the
DNA taken up into endosomes then escapes into the cyto-
sol,[12] where a small fraction enters the nucleus. It is in the
nucleus that the DNA can be transcribed to produce transi-
ent gene expression insert into host chromosomes to pro-
duce stable gene expression or to switch off gene expres-
sion/transcription. The presence of so many steps in the
transfection pathway means that it has been very difficult to
generate strong structure-activity relationships for transfec-
tion compounds, necessitating an empirical approach, which
makes the use of the combinatorial library approach em-
ployed here highly advantageous. One key feature of trans-
fection compounds, however, seems to be the ability to
change the structure of the liposomes from a lamellar phase,
with DNA between lipid bilayers, to an inverted hexagonal
Gene therapy represents an important advance in the alter-
native treatment of a variety of diseases of both genetic and
acquired origin. In this process, the corrected exogeneous
genes or portions of a gene are introduced into target cells
to replace defective DNA sequences. The most efficient
method for transfer of DNA into cells is the use of viral vec-
tors.[1] However, there are growing concerns about both the
short- and the long-term risks of viral vectors, the key prob-
lems being the immune response they provoke, which can
even be lethal,[2] the limit on the size of the DNA that can
be introduced, and the difficulty of large-scale production of
virus. There have hence been great efforts to develop a
range of non-viral vectors for gene therapy applications.[3]
broad range of non-viral delivery systems have been descri-
bed to date, including microinjection,[4] electroporation,[5]
and chemical-based systems such as calcium phosphate,[6]
A
[a] B.-e. Yingyongnarongkul, Prof. Dr. M. Bradley
Department of Chemistry
University of Southampton
Southampton SO17 1BJ (UK)
Fax : (+44)23-80583789
E-mail: mb14@soton.ac.uk
[b] M. Howarth, T. Elliott
Cancer Sciences Division
University of Southampton School of Medicine
Southampton General Hospital, SO16 6YD (UK)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org or from the author.
463
Chem. Eur. J. 2004, 10, 463 473
DOI: 10.1002/chem.200305232
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER
phase, with DNA at the centre of hexagonal arrays of lipid
tubes,[13] and this phase transition is believed to promote the
escape of the DNA from the endosomes into the cytosol.
Dioleoyl-l-a-phosphatidylethanolamine (DOPE) is included
as a co-lipid with the cationic lipids used in this study be-
cause it helps to promote this phase transition.[13]
philic, positively charged section of the transfection agent,
since it has been shown that compounds containing the gua-
nidinium group are highly efficient for transfection of a vari-
ety of mammalian cell lines.[17] In order that the potential of
these guanidinium-containing headgroups could be explored
fully, they were conjugated to a wide range of hydrophobic
tails: straight-chain and branched-chain tails of 5 18 carbon
units, as well as steroidal bile acids (Table 1).
While there are no absolute features for the structure of a
cationic lipid, the cationic lipid generally contains one or
more cationic headgroup(s) attached to one or more hydro-
phobic tail(s), the headgroup often consisting of a guanidini-
um functionality or an amine (primary, secondary, tertiary
or quaternary). The hydrophobic domain is generally a
single- or double-chain hydrocarbon of 12 18 carbon units
in length and can be a straight chain, saturated or unsaturat-
ed, or a steroidal derivative. Combinatorial solid-phase syn-
thesis allows highly efficient production and purification of
diverse libraries of compounds in the search for new thera-
peutic agents.[14] The first reported use of a solid-phase syn-
thesis approach to generate[15] and study structure activity
relationships[16] of DNA delivery compounds was by Byk.
However, the protected polycationic building blocks had to
be cleaved from the solid support prior to attachment of the
hydrophobic tails. Our work reported here has an advantage
that polyamines are directly attached to–and the cationic
head groups and hydrophobic tails coupled–on the solid
phase. Here we report the synthesis and testing of 89 poly-
amine-based cationic lipids and the isolation of two com-
pounds with high transfection activity and low toxicity. The
two most effective compounds in the library had transfec-
tion activity superior to that of the commercial lipid trans-
fection reagent effectene and merit further investigation.
Synthesis of the polyamine scaffold: The first step was the
construction of a polyamine scaffold, which was used in the
synthesis of all three libraries. An unsymmetrical polyamine
(5) was synthesised, with trifluoroacetyl (Tfa) and (4,4-di-
methyl-2,6-dioxocyclohexylidene)ethyl (Dde) protecting
groups, because of their relative orthogonality (Scheme 1).
The Tfa group has been used specifically to protect primary
over secondary amines[18] and can be removed with aqueous
ammonia[19] or aqueous methanolic K2CO3 solution.[20] The
Dde group was chosen for its ability to react selectively with
primary amines even when used in large excess, while it is
stable to both acidic and basic conditions and can be re-
moved under mild conditions.[21] The initial synthetic strat-
egy was to protect one of the primary amino groups of poly-
amine 4 selectively by treatment with ethyl trifluoroacetate
(1 equiv) (Scheme 1). The crude product was subjected to
treatment with DdeOH (1.2 equiv) to afford, after careful
column chromatography, polyamine 5 (61% yield). This or-
thogonally protected polyamine 5 was treated with linker 6
to afford 7 in high yield. To attach the polyamine scaffold 7
onto aminomethyl resin, the Aloc protecting group was
cleaved by treatment with [Pd(PPh3)4], with thiosalicylic
acid as the nucleophilic scavenger.[22] This afforded the cor-
responding acid 8, which was attached onto aminomethyl
resin by a standard DIC coupling strategy to give 9.
Results and Discussion
In the search for new efficient cationic liposomes as DNA
carriers, three libraries of transfection agents 1 3 were syn-
thesised. The guanidinium group was selected as the hydro-
Synthesis of guanidinium-containing transfection agents
Library 1: The first library contained compounds with one
guanidinium polar headgroup and one hydrophobic tail (1)
and was synthesised as shown in Scheme 2. Starting from
polyamine scaffold-bound resin 9, the Dde protecting group
was removed by treatment with hydrazine in DMF (5%).
Several methods and reagents for the conversion of an
amine to a guanidine to generate the guanidine functionality
were investigated. The most commonly used reagents in-
clude S-alkylisothiouronium salts,[23] N,N’-bis(alkoxycarbon-
yl)-protected S-alkylisothiourea derivatives,[24] pyrazole-1-
carboxamidine hydrochloride,[25] di(benzotriazol-1-yl)metha-
nimine[26] and di(imidazol-1-yl)methanimine.[27] In this case,
the S-alkylisothiouronium salt was observed to be the most
reactive. Resin 9 was thus deprotected before treatment
with
N,N’-bis(tert-butoxycarbonyl)-S-methylisothiourea
(4 equiv) to afford 10 (a ninhydrin test was negative, show-
ing that there was no primary amine group present). The
Tfa group was removed with 1m KOH/THF/MeOH (4:3:1),
and various aliphatic carboxylic acids were coupled to the
free amine, to give compounds 11. The products were ob-
tained in yields of 76 100% after cleavage from the resin
with a cocktail of TFA/CH2Cl2/H2O/thioanisole (8:0.5:0.5:1).
464
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 463 473

Polyamine-Based Cationic Lipids for DNA Delivery
463 473
Table 1. Structures of the hydrophobic tails used in the library of cationic lipids.
Library 2: The second library
contained compounds with two
guanidinium polar headgroups
and one hydrophobic tail (2)
and was synthesised as shown
in Scheme 3 and Scheme 4.
Acid 13 was synthesised in high
yield (95%) by treatment of
N1,N9-bis-(Dde)-norspermidine
12 with succinic anhydride
(Scheme 3). Compound 12 was
a by-product from the prepara-
tion of the asymmetric protect-
ed polyamines 5 (Scheme 1) or
was obtained simply by treat-
ment of norspermidine 4 with
DdeOH (2.2 equiv). Resin 14,
obtained by treatment of resin
Compound
R group
Compound
R group
x.1
x.2
x.3
x.5
x.4
x.6
x.7
x.8
x.9
x.10
x.12
x.14
x.11
x.13
x.15
x.16
x.17
x.19
x.18
x.20
9
with hydrazine in DMF
(5%), was treated with acid 13
(2 equiv) in the present of DIC
(2 equiv), HOBt (2 equiv) and
DMAP to give polyamine scaf-
fold 15 (Scheme 4). The Dde
protecting groups were selec-
tively removed as described
above. The two free amine
groups were then converted
into the guanidinium function-
ality as described for library 1.
After removal of the trifluoroa-
cetyl protecting group with
x.21
x.22
x.23
x.25
x.24
x.26
x.27
x.28
x.31
x.28
x.30
aqueous methanolic base,
a
range of acids were coupled to
the free amino group. The
products were obtained in good
yield (54 98% pure) after
cleavage from the solid support
(Scheme 4).
Structure elucidation of
2
was achieved as follows
(Figure 2). The methylene pro-
tons (H-8’’) correlated with the
guanidine carbons by HMBC.
The protons H-8’’, H-7’’ and H-
Spectroscopic characterisation of these groups of com-
6’’ were assigned by 1H 1H COSY. The H-2’ and H-3’ proton
signals exhibited cross-peaks to the carbonyl signal (C-1’),
which also correlated with H-2 in the HMBC experiment.
Protons H-2, H-3 and H-4 were confirmed by a 1H 1H
pounds was achieved by NMR (1H, 13C, H 1H COSY, H
13C COSY and HMBC), to provide the connections of the
polyamines to the guanidine and carboxylic acids. Thus, H-8
showed a cross-peak with the guanidine carbon, whereas H-
2 showed a cross-peak with the carbonyl carbon, which also
correlated to H-2’ and H-3’ (Figure 1).
1
1
1
Figure 1. Arrows showing observed H 13C correlation from HMBC spec-
tra of compounds from library 1.
Figure 2. The arrows show 1H 13C correlation from HMBC spectra from
compound library 2.
465
Chem. Eur. J. 2004, 10, 463 473
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

M. Bradley et al.
FULL PAPER
COSY experiment. Proton H-8
showed a cross-peak to C-1’’,
while the H-2’’ and H-3’’ signals
also correlated with this car-
bonyl carbon.
Library 3: The third library
contained compounds with one
guanidinium polar head group
and two hydrophobic tails (3)
and was synthesised as shown
in Scheme 5. The trifluoroacetyl
group on compound 10 (see
Scheme 2) was removed (1m
KOH/THF/MeOH 4:3:1) and
the resulting resin was treated
with acid 13 to give resin 17.
The two Dde protecting groups
were removed (N2H4/DMF,
5%) and the resulting product
was coupled with a range of
acids to afford compounds 18.
Compound 3 was obtained after
treatment of 18 with a solution
of TFA/CH2Cl2/H2O/thioanisole
(8:0.5:0.5:1). The structure of 3
was again confirmed by 1D and
2D NMR techniques. The hy-
drophobic tails of all the trans-
fection agents prepared are
shown in Table 1.
Scheme 1. Reagents and conditions: a) Ethyl trifluoroacetate (1 equiv), MeOH, À788C, 1 h, 0 8C, 4 h; b) 2-ace-
tyldimedone (1.2 equiv), CH2Cl2, 12 h, two steps 61%; c) linker 6 (1.2 equiv), DMF, 12 h, 88%; d) [Pd(PPh3)4]
(0.1 equiv), thiosalicylic acid (4 equiv), CH2Cl2/THF (1:1), 2 h, 72%; e) aminomethyl polystyrene resin, DIC
(1.5 equiv), HOBt (1.5 equiv), CH2Cl2, 12 h.
Derivatives with tails 1 to 31
were prepared for Libraries 1
and 2. However, for Library 3
only derivatives with tails 1 to
27 were synthesised, as treat-
ment with the steroidal groups
produced multiple products.
Data for all library members
is given in the Supporting Infor-
mation.
DNA bindingaffinities
: The
relative binding affinities of the
transfection agents for DNA
were evaluated to determine
whether transfection activities
correlated with DNA binding.
Relative binding affinities were
assessed in two ways: i) a gel re-
tardation assay[28] (Figure 3)
and ii) an ethidium bromide dis-
placement assay[29] (Table 2).
Scheme 2. Reagents and conditions: a) Hydrazine/DMF (5%), 2 h; b) N,N’-bis(tert-butoxycarbonyl)-S-methyli-
sothiourea (4 equiv), pyridine, THF, 12 h; c) 1m KOH/THF/MeOH (4:3:1), 2 h; d) carboxylic acids (4 equiv),
DIC (4 equiv), HOBt (4 equiv), CH2Cl2, 2 h; e) TFA/CH2Cl2/H2O/thioanisole (8:0.5:0.5:1), 2 h.
For the gel retardation assay,
compounds were mixed with
plasmid DNA at ratios of 1:5
and 1:20 (DNA/sample, w/w)
and loaded onto an agarose gel
Scheme 3. Reagents and conditions: a) Succinic anhydride (1.1 equiv), pyridine (1.1 equiv), CH2Cl2, 1 h, 95%. (Figure 4). All compounds con-
466
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 463 473

Polyamine-Based Cationic Lipids for DNA Delivery
463 473
Scheme 4. Reagents and conditions: a) Acid 13 (2 equiv), DIC (2 equiv), HOBt (2 equiv), CH2Cl2, DMAP (0.1 equiv), 2 h; b) hydrazine/DMF (5%), 2 h;
c) N,N’-bis(tert-butoxycarbonyl)-S-methylisothiourea (4 equiv), pyridine, THF, 2 h; d) 1m KOH/THF/MeOH (4:3:1), 2 h; e) carboxylic acids (4 equiv),
DIC (4 equiv), HOBt (4 equiv), CH2Cl2, 2 h; f) TFA/CH2Cl2/H2O/thioanisole (8:0.5:0.5:1), 2 h.
than that of library 1 (14 18
carbons). In library 2, in con-
trast with library 1, the steroids
were able to bind DNA. The
steroids bearing hydroxy groups
in library 2 (2.28, 2.29 and 2.31)
bound DNA more efficiently
than the steroid bearing car-
bonyl groups (2.30).
The ethidium bromide dis-
placement assay was based on
the displacement of ethidium
bromide (0.125 mg) from its in-
tercalation site in DNA (0.5 mg)
by the cationic lipids, and the
decrease in fluorescence meas-
ured (lexcit = 485 nm, lemiss
=
590 nm) after 1 minute of equi-
libration, and the weight ratio
(lipid/DNA) at which 50% of
the ethidium bromide was dis-
placed from the DNA (WR50)
determined (Table 2). WR50
values greater than 3 indicate
weak binding to DNA and
these compounds were exclud-
ed from the table. There was
good correspondence between
DNA binding affinity from
Scheme 5. Reagents and conditions: a) 1m KOH/THF/MeOH (4:3:1), 2 h; b) acid 13 (1.7 equiv), DIC
(2 equiv), HOBt (2 equiv), CH2Cl2, 2 h; c) hydrazine/DMF (5%), d) carboxylic acids (4 equiv), DIC (4 equiv),
HOBt (4 equiv), CH2Cl2, 2 h; e) TFA/CH2Cl2/H2O/thioanisole (8:0.5:0.5:1), 2 h.
taining a tail of fewer than nine carbons bound poorly to
DNA. In libraries 2 and 3, the chain length of the hydropho-
bic tail required for DNA binding (9 18 carbons) was less
ethidium bromide displacement and from the gel electro-
phoresis assay (Figure 3). Again, compounds with a hydro-
phobic tail of fewer than nine carbons bound weakly to
467
Chem. Eur. J. 2004, 10, 463 473
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

M. Bradley et al.
FULL PAPER
Figure 3. Gel electrophoresis assay of the mixtures of plasmid DNA and cationic lipids at 1:5 (a) and 1:20 (b) weight ratios. Lanes marked C contained
DNA alone and were used as a control. The presence of a lower band shows that the DNA has migrated and so has not been bound by the transfection
compound.
DNA. Library 2 generally had higher binding affinity than
libraries 1 and 3. Nearly all the steroid-containing samples
Table 2. WR50 values from the ethidium bromide displacement assay.
Values for all other compounds were greater than 3, indicating poor
DNA binding.
showed poor DNA binding. Compounds containing
a
straight chain were more effective at binding DNA than
compounds containing a branched chain of the same length
(3.17 versus 3.18, 2.19 versus 2.20 or 2.21). Conflicting data
about the ideal lipid chain length have been obtained previ-
Compound WR50
(lipid/
Compound WR50
(lipid/
Compound WR50
(lipid/
DNA)
DNA)
DNA)
1.24
1.25
1.26
1.27
1.31
2.16
2.17
2.18
2.19
1.08
0.67
1.69
0.95
1.70
1.91
1.71
2.59
2.54
2.20
2.21
2.22
2.23
2.24
2.25
2.26
2.27
2.31
2.16
1.14
1.25
0.81
0.63
0.77
0.78
0.51
1.32
3.17
3.20
3.21
3.22
3.23
3.24
3.25
3.27
0.73
2.36
2.43
2.10
1.88
2.45
2.41
1.98
32]
ously,[30 but the consensus is that tails shorter than 12 car-
bons are unable to form lipid bilayers and thus do not inter-
act well with DNA.
Transfection results: The cationic lipids that bound DNA
(from the binding affinity assays) were used to study trans-
fection of mammalian cell lines. Typically, cationic liposomes
468
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 463 473

Polyamine-Based Cationic Lipids for DNA Delivery
463 473
50 mL (20 mgmLÀ1 in ethanol) and DOPE 50 mL (20 mgmLÀ1
in chloroform) were mixed and the organic solvent was re-
moved under a stream of nitrogen. The resulting thin film
was dried under high vacuum (>2 h). The thin film was hy-
drated with 100 mL phosphate-buffered saline (PBS). The
soution was vortexed (1 min) and sonicated (2î15 min) in a
bath-type sonicator to form the liposomes. The resulting so-
lution was stored at 48C for 24 h before use. The transfec-
tion activities of cationic liposomes were evaluated in com-
parison with the commercially available transfection reagent
effectene by use of b-galactosidase as a reporter gene, the
transfection activity of each cationic liposome being report-
headgroups, caused much higher levels of protein expres-
sion. Compounds 2.24 and 2.27 gave transfection levels of
117% and 105%, respectively. Compounds 2.23, 2.25, 2.26,
3.23 and 3.27 were also identified as active transfection
agents, but the levels of gene expression were lower than
those for 2.24 and 2.27. DNA liposome complexes need a
net positive charge for high transfection efficiencies,[33] al-
though the activity decreases when the lipid ratio is too
high, due to cytotoxicity[9,33] Most of the compounds in li-
brary 2 had charge ratios lower than 15. DNA/cationic lipid
of lower charge ratios (<15) gave better transfections than
those of higher charge ratios (Figure 5). The steroidal com-
pound 2.30 was an exception, with transfection activity only
at the higher charge ratios (>20). These compound libraries
did not show significant transfection activity when DOPE
was omitted (data not shown).
The most active compounds in transfection, 2.24 and 2.27,
were two of the compounds with the highest affinity for
DNA as measured by the ethidium bromide displacement
assay. No correlation was found between the measured
transfection activity or DNA binding and the hydrophobici-
ty of the compound (cLogP) (as estimated from ChemDraw
Ultra, data not shown).
Transfection compound toxicity: To assess the relationship
between cytotoxicity and gene expression efficiency, the tox-
icity of the two cationic lipids (2.24 and 2.27) most active in
transfection was examined by measuring changes in cell
metabolic activity (MTT assay)[34] and by induction of cell
death by trypan blue exclusion[35] after transfection
(Figure 6). The IC50 values of most complexes were 30 mm,
two times higher than the concentrations used for transfec-
tion. The exception was compound 2.27 which had an IC50
of approximately 8mm in the presence of DNA (but in the
absence of DOPE). Effectene was slightly more toxic than
compounds 2.24 and 2.27 in the MTT assay (data not
shown). The effect of the transfection compounds on cell
death, as assessed by trypan blue assay, was much less than
their effect on metabolic activity (Figure 6b). The com-
plexes from 2.24 and 2.27 exhibited minimal toxicity and
this was not greatly changed by inclusion of DOPE as a co-
lipid. Thus, overall, compounds 2.24 and 2.27 can produce
high levels of transfection without inducing significant cell
death, although there is a reduction in metabolic activity.
This reduction in metabolic activity is also observed in other
commercial transfection reagents and should not limit the
application of these compounds.
Figure 4. Relative transfection activities (%) of each library of cationic
lipids relative to effectene. Cationic lipids were mixed with plasmid DNA
(b-Gal, 0.1 mgmLÀ1 per well) at weight ratios (DNA/cationic lipid) of 1:10
and 1:20 and used for transfection of 293T cells.
Conclusions
It is surprising that the compounds in library 2, with two
headgroups and one tail, showed the highest transfection ac-
tivities, since most cationic lipid transfection compounds
(e.g., N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammoni-
um methylsulfate (DOTAP), N-[1-(2,3-dioleyloxy)propyl]-
N,N,N-trimethylammonium chloride (DOTMA), (+)-N-(2-
hydroxyethyl)-N,N-dimethyl-2,3-bis(tetradecyloxy)-1-propa-
naminium bromide (DMRIE)) have two aliphatic tails.[8]
ed as a percentage of that of the effectene control. Figure 4
displays the data generated by employment of a plasmid en-
coding b-galactosidase (100 ng/well) at DNA/cationic lipid
ratios (w/w) of 1:10 and 1:20. Most of the compounds in li-
brary 1 and 3 bearing one guanidinium group did not medi-
ate transfection under these conditions. However, several
compounds from library 2, containing two guanidinium
469
Chem. Eur. J. 2004, 10, 463 473
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

M. Bradley et al.
FULL PAPER
Cationic detergents have also
been shown to condense
DNA.[36] Detergents have
a
much higher water solubility
than cationic lipids, release
their DNA rapidly and kill cells
and consequently cannot be
used as transfection agents.[37]
A
previous study with lipo-
somes made of DOPE com-
bined with a compound con-
taining a single aliphatic tail
produced a high degree of tox-
icity even after three hours of
exposure,[38] which was not ob-
served with any of the com-
pounds in library 2, suggesting
that investigation of further
compounds containing one ali-
phatic tail for their transfection
potency is warranted.
The generation of com-
pounds with transfection abili-
ties similar to or greater than
those of a widely used commer-
cial reagent suggests that this
Figure 5. Relative transfection activities (%) of cationic lipids (selected from Figure 4) at lower (grey bar) and
higher (black bar) charge ratios.
class of compound has significant potential for overcoming
gene transfer difficulties in vitro and possibly in vivo. Fur-
ther work will investigate the effect of compound formula-
tion and examine the varying susceptibilities of different cell
types (e.g., B cells, dendritic cells, fibroblasts) to transfection
by different members of these libraries.
Experimental Section
General information: NMR spectra were recorded on a Bruker AC 400
1
spectrometer operating at 400 MHz for H and 100 MHz for 13C. All cou-
pling constants (J values) were measured in Hz. ES mass spectra were re-
corded with a VG Platform Quadrupole Electrospray Ionisation mass
spectrometer. High-resolution electrospray mass spectra were recorded
on a Bruker Apex III FT-ICR mass spectrometer. Infra-red spectra were
recorded on a BioRad Golden Gate FTS 135 Paragon 1000. Reversed-
phase analytical HPLC (RP-HPLC) was performed on a Hewlett Pack-
ard HP1100 Chemstation, with a Phenomenex C18 prodigy 5m (150 mmî
3.0 mm i.d.) column. Thin layer chromatography (TLC) was performed
on Alugram SIL G/UV/254 precoated plates. Bands were visualised
under UV light. Column chromatography was carried out on SiO2. All re-
actions were carried out at room temperature unless otherwise stated.
N1-(Trifluoroacetyl)-N9-1-(4,4-dimethyl-2,6-dioxocyclohexylidene)ethyl-
norspermidine (5) and N1,N9-bis-1-(4,4-dimethyl-2,6-dioxocyclohexylide-
ne)ethyl-norspermidine (12): A solution of ethyl trifluoroacetate (1 equiv,
5.30 mL, 44.54 mmol) in MeOH (20 mL) was added dropwise at À788C
over 1 h to a stirred solution of norspermidine (4) (1 equiv, 5.85 g,
44.58 mmol) in MeOH (30 mL) and the system was left to stir at 08C for
4 h. The solvent was removed under reduced pressure and the residue
was co-evaporated with CH2Cl2 (2î50 mL) to remove trace amounts of
MeOH. The crude product was dissolved in CH2Cl2 (50 mL), 2-acetyldi-
medone (1.2 equiv, 9.75 g, 53.5 mmol) was added, and the reaction mix-
ture was left to stir overnight. The solvent was evaporated to dryness and
the crude product was purified by chromatography on silica gel, with elu-
tion with CH2Cl2/MeOH (95:5 to 93:7), to afford 12 (3.89 g, 19%) and 5
(10.68 g, 61%) as viscous oils.
Figure 6. Toxicity of the two most active compounds used for transfection
of 293T cells. These two compounds were tested for reduction in meta-
bolic activity (MTT assay) and induction of cell death (trypan blue). De-
terminations were performed after 24 h incubation with medium contain-
ing the complexes (0.1 mg DNA/100 mL).
470
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 463 473

Polyamine-Based Cationic Lipids for DNA Delivery
463 473
Compound 12: Rf
=
0.30 (CH2Cl2/MeOH 9:1); 1H NMR (400 MHz,
196.9 (C=OÀDde) ppm; IR (CH2Cl2): n˜ = 1691, 1570 cmÀ1 (C=O); MS
(ES+): m/z (%): 600.3 (20) [M+H]+, 622.3 (5) [M+Na]+; HRMS (ES+):
m/z: calcd for C28H37F3N3O8 [M+H]+: 600.2527; found: 600.2539.
CDCl3): d = 1.05 (s, 12H; 2î(C(CH3)2), 1.87 (tt, 3J(H,H) = 7, 7 Hz,
4H; CH2CH2NHCH2CH2), 2.37 (s, 8H; COCH2), 2.57 (s, 6H; C=CCH3),
2.76 (t, 3J(H,H) = 7 Hz, 4H; CH2NHCH2), 3.51 (m, 4H; 2îCH2NHÀ
Dde), 13.44 (brs, 2H; NHÀDde) ppm; 13C NMR (CDCl3, 100 MHz): d =
18.3 (C=CCH3), 28.6 (C(CH3)2), 29.7 (CH2CH2NHCH2CH2), 30.5
(C(CH3)2), 41.7 (CH2NHÀDde), 47.2 (CH2NHCH2), 53.2 (COCH2), 108.3
Synthesis of polyamine scaffold resin 9: Compound
8 (3 equiv,
6.43 mmol) was dissolved in CH2Cl2 (15 mL), and then DIC (3.5 equiv,
7.63 mmol) and HOBt (3.5 equiv, 7.63 mmol) were added. After 10 min,
this solution was added to aminomethyl resin (1 equiv, 2.18 mmol) and
the suspension was shaken overnight. The resulting resin was washed
with CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (3î15 mL each) and
dried under high vacuum. The resin gave a negative ninhydrin test.
(C=CCH3), 174.0 (C=CCH3), 198.4 (C=O) ppm; IR (CH2Cl2): n˜
=
1628 cmÀ1 (C=O); MS (ES+): m/z (%): 460.4 (100) [M+H]+; HRMS
(ES+): m/z: calcd for C26H42N3O4 [M+H]+: 460.3170; found: 460.3172.
Compound 5: Rf
CDCl3):
=
0.26 (CH2Cl2/MeOH 9:1); 1H NMR (400 MHz,
1.05 (s, 6H; C(CH3)2, 1.75 1.92 (m, 4H;
Solid-phase synthesis of transfection agent library: synthesis of transfec-
tion library 1
d
=
CH2CH2NHCH2CH2), 2.39 (s, 4H; COCH2), 2.59 (s, 3H; C=CCH3), 2.77
(t, 3J(H,H) = 7 Hz, 2H; CH2CH2CH2NHDde), 2.84 (t, 3J(H,H) = 6 Hz,
2H; TfaNHCH2CH2CH2), 3.52 (m, 4H; NH(CH2CH2CH2)2), 5.03 (brs,
2H; NH), 13.45 (brs, 1H; NHÀDde) ppm; 13C NMR (CDCl3, 100 MHz):
Full experimental data for all library compounds 1,2, and 3 is given in
the Supporting Information.
General procedures
Dde deprotection: Resin (1 equiv) was pre-swollen in DMF for 30 min
and filtered. Hydrazine in DMF (5% v/v) was added to this resin and the
suspension was shaken for 2 h. The resin was filtered and washed succes-
sively with CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (positive ninhydrin
test).
d
= 18.2 (C=CCH3), 27.8 (TfaNHCH2CH2), 28.6 (C(CH3)2); 29.4
(CH2CH2NHDde);
30.5
(C(CH3)2);
39.9
(TfaNHCH2);
41.6
(CH2NHDde), 47.1 (CH2CH2CH2NHDde); 48.6 (TfaNHCH2CH2CH2);
53.2 (COCH2); 108.4 (C=CCH3); 116.5 (q, 1J = 288 Hz; CF3ÀCONH),
157.6 (q, 2J
=
36 Hz, CF3ÀCOÀNH), 174.0 (C=CCH3), 198.5 (COÀ
Dde) ppm; IR (CH2Cl2): n˜ = 1717 cmÀ1 (C=O); MS (ES+): m/z (%):
392.3 (16) [M+H]+; HRMS (ES+): m/z: calcd for C18H29F3N3O3 [M+H]+:
392.2156; found: 392.2145.
Guanidation: The resulting resin was pre-swollen in THF for 30 min and
filtered. A solution of N,N’-bis(tert-butoxycarbonyl)-S-methylisothiourea
(4 equiv, 1m) in THF and a few drops of pyridine were added and the
suspension was shaken overnight. The solution was collected and the
resin was washed as above (negative ninhydrin test).
N1-1-(4,4-Dimethyl-2,6-dioxocyclohexylidene)ethyl-N5-(4-(allyloxycarbo-
nylmethoxy)phenyl-methoxycarbonyl)-(N9-trifluoroacetyl)-norspermidine
(7): Compound 5 (1 equiv, 5.42 g, 13.85 mmol) was dissolved in DMF
(10 mL), linker 6 (1.2 equiv, 6.43 g, 16.62 mmol) was added, and the reac-
tion mixture was stirred overnight. It was then poured into KHSO4 (1m,
100 mL) and extracted with EtOAc (2î100 mL). The combined organic
layers were washed with water, dried over Na2SO4 and evaporated to
dryness. The crude product was purified on silica gel, with elution with
CH2Cl2 to CH2Cl2/MeOH (97:3), to afford the desired product 7 (7.77 g,
88%). Rf = 0.37 (CH2Cl2/MeOH 20:1); 1H NMR (400 MHz, CDCl3): d
= 0.96 (s, 6H; C(CH3)2), 1.69 (brs, 2H; TfaHNCH2CH2), 1.83 (brs, 2H;
DdeHNCH2CH2), 2.29 (s, 4H; 2îCOCH2ÀDde), 2.42 (s, 3H; C=CCH3),
3.18 3.35 (m, 8H; N(CH2CH2CH2)2), 4.59 (s, 2H; OCH2C6H4), 4.63 (m,
2H; OCH2CH=CH2), 5.02 (s, 2H; C6H4OCH2CO), 5.18 5.30 (m, 2H;
Tfa deprotection: A mixture of 1m KOH/THF/MeOH (4:3:1) was added
to the resin (pre-swollen in THF and filtered). This suspension was
shaken for 2 h. The resin was washed successively with CH2Cl2, MeOH,
DMF, MeOH and CH2Cl2 (positive ninhydrin test).
Couplingof carboxylic acid : The resin was pre-swollen for 30 min in
CH2Cl2 and filtered. A solution of carboxylic acid (4 equiv, 0.5m), DIC
(4 equiv, 0.5m) and HOBt (4 equiv, 0.5m) in CH2Cl2 was added to this
resin. The suspension was shaken for 2 h and the resin was washed as de-
scribed above (negative ninhydrin test).
Cleavage of the product from the resin: The resin was pre-swollen in
CH2Cl2 for 30 min and filtered. A cocktail of TFA/CH2Cl2/H2O/thioani-
sole (8:0.5:0.5:1) was added and the suspension was shaken for 2 h. The
resin was filtered and the solution was collected. The solvent was re-
moved under vacuum, and the crude product was re-dissolved in H2O
and extracted with CH2Cl2. The aqueous layer was evaporated under re-
duced pressure to afford the pure products in 76 100% yield (% yield
based on the loading of the aminomethyl resin and product as 4¥TFA
salts).
Synthesis of N1,N9-bis-1,1-(4,4-dimethyl-2,6-dioxocyclohexylidene)ethyl-
N5-(3-carboxypropanoyl)-norspermidine (13): Succinic anhydride (0.48 g,
4.74 mmol) was added to a solution of 12 (1.98 g, 4.31 mmol) in CH2Cl2
(5 mL) and the mixture was stirred for 1 h. The reaction mixture was
poured into water (200 mL) and extracted with EtOAc (2î100 mL). The
combined organic layers were dried over anhydrous Na2SO4 and evapo-
rated under reduced pressure to afford the title compound as an amor-
phous solid (2.22 g, 95%).
3
OCH2CH=CH2), 5.85 (m, 1H; OCH2CH=CH2), 6.81 (d, J(H,H) = 9 Hz,
3
2H; ArH), 7.22 (d, J(H,H) = 9 Hz, 2H; ArH), 7.93 (brs, 1H; NHÀTfa),
13.44 (brs, 1H; NHÀDde) ppm; 13C NMR (CDCl3, 100 MHz): d = 18.3
(C=CCH3), 27.3 (TfaHNCH2CH2), 28.3 (DdeHNCH2CH2), 28.6
(C(CH3)2), 30.5 (C(CH3)2), 36.4 (TfaHNCH2), 41.1 (TfaHNCH2CH2CH2),
44.3 (DdeHNCH2), 44.5 (DdeHNCH2CH2CH2), 53.2 (COCH2ÀDde), 65.7
(OCH2C6H4), 66.3 (OCH2CH=CH2), 67.8 (C6H4OCH2CO), 108.4 (C=
CCH3), 115.2 (CHarom), 115.8 (OCH2CH=CH2), 119.6 (OCH2CH=CH2),
129.8 (Carom), 130.6 (CHarom), 131.8 (Carom), 158.3 (C=O amide), 168.8 (C=
O ester), 174.1 (C=CCH3), 198.4 (C=OÀDde) ppm; IR (CH2Cl2): n˜
=
1757, 1718 cmÀ1 (C=O); MS (ES+): m/z (%): 639.9 (100) [M+H]+, 661.9
(70) [M+Na]+; HRMS (ES+): m/z: calcd for C31H40F3N3O8Na [M+Na]+:
662.2659; found: 662.2641.
N1-1-(4,4-Dimethyl-2,6-dioxocyclohexylidene)ethyl-N5-(4-(carboxyme-
thyl)phenyl-methoxy-carbonyl))-N9-(trifluoroacetyl)-norspermidine (8):
Compound 7 (1 equiv, 6.34 g, 9.91 mmol) in CH2Cl2/THF (1:1, 30 mL
each) was purged with N2 (g) for 1 h. [Pd(PPh3)4] (0.1 equiv, 1.14 g,
0.99 mmol) and thiosalicylic acid (2 equiv, 3.06 g, 19.82 mmol) were
added in one portion, and the reaction mixture was left to stir for 2 h.
The reaction mixture was evaporated to dryness and the crude product
was purified on silica gel, with elution with EtOAc to 50% MeOH/
Rf = 0.48 (CH2Cl2/MeOH 9:1); 1H NMR (400 MHz, [D6]DMSO): d =
1.08 (s, 12H; 2îC(CH3)2ÀDde); 1.87 (m, 2H; NCH2CH2CH2NH), 1.99
(m, 2H; NHCH2CH2CH2N), 2.39 (s, 8H; COCH2), 2.50 (m, 2H;
NCOCH2), 2.56 (s, 3H; C=CÀCH3ÀDde), 2.62 (m, 5H; m, C=CÀ
CH3+CH2CO2H), 3.47 (m, 6H; CH2NCH2CH2CH2NH), 3.58 (m, 2H;
NHCH2CH2CH2N), 13.34 (brs, 1H; NH), 13.42 (brs, 1H; NH) ppm; 13C
=
0.23 (CH2Cl2/MeOH 2:1); 1H
NMR (100 MHz, [D6]DMSO):
(NCH2CH2CH2NH), 30.2
(C(CH3)2+NHCH2CH2CH2N), 32.4 (NCH2), 43.0 (NHCH2CH2CH2N),
43.4 (NCH2CH2CH2NH), 45.1 (NCH2CH2CH2NH), 46.9
d
=
20.0 (C=CÀCH3), 29.8
EtOAc, to give 8 (4.29 g, 72%). Rf
(CH2CO2H), 33.6
NMR (400 MHz, [D6]DMSO): d = 0.96 (s, 6H; C(CH3)2), 1.75 1.90 (m,
4H; CH2CH2NHCH2CH2), 2.36 (s, 4H; COCH2ÀDde), 2.55 (s, 3H; C=
CCH3); 3.20 3.40 (m, 8H; HN(CH2CH2CH2)2), 4.33 (s, 2H; OCH2C6H4),
5.07 (s, 2H; OCH2CO2H), 6.90 (d, 3J(H,H) = 9 Hz, 2H; ArH), 7.32 (d,
3J(H,H) = 9 Hz, 2H; ArH), 9.57 (brs, 1H; NHÀTfa), 13.38 (brs, 1H;
NHÀDde) ppm; 13C NMR (100 MHz, [D6]DMSO): d = 17.7 (C=CCH3),
27.3 (TfaHNCH2CH2), 28.3 (DdeHNCH2CH2+C(CH3)2), 30.2 (C(CH3)2),
37.5 (TfaHNCH2), 41.5 (TfaHNCH2CH2CH2), 45.0 (DdeHNCH2), 49.0
(DdeHNCH2CH2CH2), 52.8 (COCH2ÀDde), 66.6 (OCH2C6H4), 67.7
(OCH2CO2H), 107.4 (C=CCH3), 114.8 (CHarom), 128.5 (Carom), 129.6
(CHarom), 131.8 (Carom), 159.1 (C=O amide), 173.3 (C=CCH3+CO2H),
(NHCH2CH2CH2N), 55.1 (2îCH2ÀDde), 109.7 (C=CÀCH3), 174.0 (C=
CÀCH3), 175.5, 175.7 (NCOCH2CH2CO2H), 199.1, 199.2 (C=O) ppm; IR
(film): n˜ = 1725, 1635 cmÀ1 (C=O); MS (ES+): m/z (%): 528.4 (100)
[M+H]+; HRMS (ES+): m/z: calcd for C30H46N3O7 [M+H]+: 560.3330;
found: 560.3341.
Synthesis of compound library 2: Hydrazine in DMF (5%, v/v) (10 mL)
was added to the resin 9 (1.18 g, 0.76 mmol, pre-swollen in DMF and fil-
tered) and the suspension was shaken for 2 h. The resin was washed with
471
Chem. Eur. J. 2004, 10, 463 473
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

M. Bradley et al.
FULL PAPER
CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (3î10 mL each). The result-
ing resin (positive ninhydrin test) was pre-swollen in CH2Cl2 and filtered.
MeOH, DMF, MeOH and CH2Cl2 (3î3mL each). The resin gave a nega-
tive ninhydrin test. A cleavage cocktail of TFA/CH2Cl2/H2O/thioanisole
(8:0.5:0.5:1, 2 mL) was added to this resin (pre-swollen in CH2Cl2 and fil-
tered) and the suspension was shaken for 2 h. The resin was filtered and
washed with CH2Cl2 (1 mL) and the combined solution was collected.
The solvent was removed under vacuum, and the crude product was re-
dissolved in H2O (20 mL) and extracted with CH2Cl2 (20 mL). The aque-
ous layer was evaporated under reduced pressure to afford the final
products 3 as 4¥TFA salts (55 100%).
A
solution of acid 13 (0.89 g, 1.65 mmol) in CH2Cl2 (8 mL), DIC
(0.26 mL, 1.65 mmol) and HOBt (0.22 g, 1.65 mmol) were added to this
resin. The suspension was shaken for 2 h and the resin was filtered and
washed successively with CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (3î
10 mL each). Resin 15 was dried under vacuum and gave a negative nin-
hydrin test. The Dde protecting groups were removed by treatment with
N2H4 in DMF (5% v/v) for 2 h. The resulting resin was washed with
CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (3î10 mL each). The resin
gave a positive ninhydrin test. A solution of the N,N’-bis(tert-butoxycar-
bonyl)-S-methylisothiourea (0.99 g, 3.41 mmol) in THF (5 mL) and pyri-
dine (5 drops) were added to this resin. The resulting suspension was
shaken overnight. The solution was filtered to recover the guanylating re-
agent and the resin was washed as above. The resin gave a negative nin-
hydrin test. A solution of 1m KOH/THF/MeOH (4:3:1, 10 mL) was
added to this resin (pre-swollen in THF and filtered) and the suspension
was shaken for 2 h. The resin was washed with MeOH/H2O (1:1),
MeOH, DMF and CH2Cl2 (3î10 mL each) and the resin gave a positive
ninhydrin test. The resulting resin (35 50 mg per carboxylic acid) was
pre-swollen in CH2Cl2 for 30 min and the resin was filtered. A solution
of commercially available carboxylic acid (4 equiv) in CH2Cl2 (2 mL) and
3 5 drops of DMF, DIC (4 equiv) and HOBt (4 equiv) were added. The
suspension was shaken for 2 h. The resin was washed with CH2Cl2,
MeOH, DMF, MeOH and CH2Cl2 (3î10 mL each). The resulting resin
16 gave a negative ninhydrin test. A cleavage cocktail of TFA/CH2Cl2/
H2O/thioanisole (8:0.5:0.5:1, 2 mL) was added to this resin and the sus-
pension was shaken for 2 h. The resin was filtered and washed with
CH2Cl2 (1 mL) and the combined solutions were collected. The solvent
was removed in vacuum and the crude product was re-dissolved in H2O
(20 mL) and extracted with CH2Cl2 (20 mL). The aqueous layer was
evaporated under reduced pressure to afford the final products 2 (34
77% as 7¥TFA salts).
DNA bindingaffinities
Electrophoresis assay: DNA binding affinities of all samples were meas-
ured at two DNA/sample ratios (w/w), 1:5 and 1:20, by electrophoresis.
DNA/sample complexes were formed at a ratio of 1:5 (w/w) by transferal
of 12.5 mL (30 mgmLÀ1) of sample into an Eppendorf tube. Each sample
was further diluted with 37.5 mL of acetate buffer (20 mm, pH 7.4). To
this solution, an aqueous solution of plasmid DNA (50 mL, 3 mg/25 mL),
purified from E. coli HMS174 by a Qiagen Maxiprep, was added to each
sample, and the solutions were successively mixed. The DNA/sample
complex 1:20 (w/w) was prepared as above except that the 50 mL of stock
sample (30 mgmLÀ1) was used without dilution. DNA complexes were in-
cubated at room temperature for 30 min. Bromophenol blue-free gel-
loading buffer (100 mL, 13.3% w/v sucrose in water) was added to 100 mL
of this complex. The solution was inverted to mix, and each sample
(10 mL) was loaded onto a 1% agarose gel (0.5îTBE buffer). The gel
was run at 200 V, 400 mA for 2 h. DNA bands were viewed under UV
light by ethidium bromide staining.
Ethidium bromide displacement assay: A fluorescence microplate reader
was used to measure the fluorescence intensity of eight samples with dif-
ferent charge ratios in a 96-well PS plate. A stock solution of DNA of
0.5 mg/5 mL was prepared in low salt buffer (20 mm NaCl, 2 mm HEPES,
10 mm EDTA, pH7.4). Ethidium bromide and synthetic compounds were
weighed, and stock solutions were prepared with a final concentration of
0.125 mgmLÀ1 in water and 0.1 mg/5 mL in ethanol, respectively. Each
sample was prepared as shown in Table 3. Fluorescence was measured
(ex filter = 485 nm, em filter = 590 nm) after 1 minute of equilibration.
The fluorescence was calculated as a percentage of the maximum fluores-
cence intensity when ethidium bromide was bound to DNA in the ab-
sence of the transfection compound. The WR50 value is the lipid/DNA
weight ratio that gives a 50% reduction in the fluorescence intensity of
the solution containing DNA (0.5 mg) and ethidium bromide (0.125 mg).
Synthesis of compound library 3: Resin
9
(1.19 g, 0.76 mmol,
0.64 mmolgÀ1) was pre-swollen in DMF for 30 min and filtered. Hydra-
zine in DMF (5% v/v, 10 mL) was added and the suspension was shaken
for 2 h. The resin was washed successively with CH2Cl2, MeOH, DMF,
MeOH and CH2Cl2 (3î10 mL each). The resulting resin (positive ninhy-
drin test) was pre-swollen in THF and filtered. A solution of the N,N’-
bis(tert-butoxycarbonyl)-S-methylisothiourea (5 equiv, 1.14 g, 3.93 mmol)
in THF (5 mL) and pyridine were added to this resin. The suspension
was shaken overnight. The solution was collected and the resin was
washed as described above. The resin gave a negative ninhydrin test. A
solution of 1m KOH/THF/MeOH (4:3:1, 10 mL) was added to resin 10
(pre-swollen in THF and filtered) and the suspension was shaken for 2 h.
The resin was washed as described above and gave a positive ninhydrin
test. The resulting resin was pre-swollen in CH2Cl2 and filtered. A solu-
tion of acid 13 (0.75 g, 1.34 mmol) in CH2Cl2 (8 mL), DIC (0.26 mL,
1.65 mmol) and HOBt (0.22 g, 1.65 mmol) were added to this resin. The
suspension was shaken for 2 h, and the resin was filtered and washed suc-
cessively with CH2Cl2, MeOH, DMF, MeOH and CH2Cl2 (3î10 mL
each). Resin 17 was dried under vacuum and gave a negative ninhydrin
test.
Liposome preparation: Dioleoyl-l-a-phosphatidylethanolamine (DOPE)
(Sigma) 50 mL (20 mgmLÀ1 in chloroform) and cationic lipid 50 mL
(20 mgmLÀ1 in ethanol) were mixed (weight ratio 1:1). The organic sol-
vents were evaporated under a stream of nitrogen gas and further dried
under high vacuum (>2 h.). The resulting thin film was hydrated with
phosphate buffered saline (PBS, pH 7.4) 100 mL and hydrated at room
temperature for one hour. The mixture was vortexed for one minute and
sonicated (2î15 min) with one hour rests between sonications in a bath-
type sonicator, producing small unilamellar vesicles.[29] The liposomes
were stored at 48C for 24 h prior to use.
Transfection procedure: 293T (human embryonic kidney) cells were
grown in DMEM supplemented with 10% heat inactivated foetal calf
serum, penicillin (100 unitsmLÀ1), streptomycin (100 mgmLÀ1) and l-glu-
tamine (4 mm) at 378C under 5% CO2. For transfection, 1.8î104 cells/
well were seeded in medium (140 mL) in a 96-well culture plate, to give
50 70% confluence for use the next day. The growth medium was re-
moved and replaced with 100 mL of serum-free medium (AIM-V, Sigma).
DNA/cationic liposome complexes were prepared as follows. An appro-
priate volume of each cationic liposome (1 mgmLÀ1) was added to the
plasmid DNA 0.4 mL (0.5 mgmLÀ1) and the system was incubated at room
temperature for 30 min before being diluted with phosphate-buffered
Hydrazine in DMF (5% v/v, 10 mL) was added to resin 17 (pre-swollen
in DMF and filtered). This suspension was shaken for 2 h. The resin was
washed successively with CH2Cl2, MeOH, DMF, MeOH, CH2Cl2 and
Et2O (3î10 mL each) and dried under vacuum. The resin gave a positive
ninhydrin test. The resulting resin (35 50 mg per carboxylic acid) was
pre-swollen in CH2Cl2 for 30 min and filtered. A solution of carboxylic
acid (4 equiv) in CH2Cl2 (2 mL) and 3 5 drops of DMF, DIC (4 equiv)
and HOBt (4 equiv) were added to this resin. The suspension was shaken
for 2 h, and then the resulting resin was washed successively with CH2Cl2,
Table 3. Amounts of compounds used in the ethidium bromide displacement assays in a 96-well plate format.
Well
1
2
3
4
5
6
7
8
9
10
11
12
buffer [mL]
DNA [mL]
Et Br [mL]
lipid [mL]
190
5
5
185
5
5
180
5
5
175
5
5
170
5
5
165
5
5
160
5
5
155
5
5
150
5
5
145
5
5
140
5
5
115
5
5
0
5
10
15
20
25
30
35
40
45
50
75
472
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 463 473

Polyamine-Based Cationic Lipids for DNA Delivery
463 473
saline to make a final DNA concentration of 0.1 mg/10 mL. DNA/cationic
liposome complexes (10 mL) were added to the cells and left to incubate
at 378C under 5% CO2 for 48 h. For effectene (Qiagen) transfections,
enhancer (1.6 mL) was added to plasmid DNA (0.2 mg) in buffer EC
(60 mL). The mixture was vortexed for 2 s and left on bench for 3 min.
Effectene was added to this mixture (5 mL), which was vortexed for 10
seconds and left for 7 min. Serum-free medium (350 mL) was added and
mixed by pipetting up and down twice. DNA/effectene complex (50 mL)
was added to the cells. b-Galactosidase expression was measured with
the FluoReporter LacZ/Galactosidase Quantitation kit (Molecular
Probes) according to the manufacturer×s instructions, with the reaction
developed for 10 min at room temperature. Transfection efficiency was
calculated as a percentage relative to effectene transfection, after sub-
traction of the value of untransfected cells.
[11] N. J. Caplen, E. W. Alton, P. G. Middleton, J. R. Dorin, B. J. Steven-
son, X. Gao, S. R. Durham, P. K. Jeffery, M. E. Hodson, C. Coutelle,
L. Huang, D. J. Porteous, R. Williamson, D. M. Geddes, Nat. Med.
1995, 1, 39 46.
[12] J. Zabner, A. J. Fasbender, T. Moninger, K. A. Poellinger, M. J.
Welsh, J. Biol. Chem. 1995, 270, 18997 19007.
[13] I. Koltover, T. Salditt, J. O. Raedler, C. R. Safinya, Science 1998,
281, 78 81.
[14] G. Jung, Combinatorial Chemistry Synthesis, Analysis, Screening,
Wiley-VCH, Weinheim, Germany, 1999.
[15] G. Byk, M. Ferderic, D. Scherman, Tetrahedron Lett. 1997, 38,
3219 3222.
[16] G. Byk, C. Dubertret, V. Escriou, M. Frederic, G. Jaslin, R. Rangara,
B. Pitard, J. Crouzet, P. Wils, B. Schwartz, D. Scherman, J. Med.
Chem. 1998, 41, 224 235.
Cytotoxicity assay: Cytotoxicity was evaluated by use of 3-(4,5-dime-
thylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reagent (MTT test)
and by trypan blue exclusion. Cells were seeded in 96-well plates at 1î
104 cells per well. The growth medium was removed next day and re-
placed with 100 mL of serum-free medium (AIM-V, Sigma). Cationic lipo-
some/DNA complexes and effectene were prepared as described in the
transfection procedure. For the MTT assay, an increasing amount of cati-
onic lipid/DNA (1:10 wt/wt) and cationic liposome/DNA (1:10 wt/wt)
was added to triplicate wells and the system was incubated at 378C under
5% CO2 for 24 h. The medium was then removed and replaced with a
phenol red-free medium (90 mL). MTT (3 mgmLÀ1) was added (10 mL/
well) to the cells and formazan crystals formed in the incubator over 3 h.
MTT solubilization solution (Sigma, 100 mL) was added to dissolve the
resulting crystals. Absorbance was measured at 570 nm on a microplate
reader (Bio-Rad). Reduction in metabolic activity was calculated as (A570
with compound/A570 untreated). For trypan blue exclusion, 10 mL (1 mg/
10 mL) of cationic lipid/DNA (1:10 wt/wt) and 10 mL cationic liposome/
DNA (1:10 wt/wt) were added to cells in duplicate wells, and the system
was incubated at 378C under 5% CO2 for 24 h. The cells were detached
from the well by pipetting, and the suspension (10 mL) was mixed with
10 mL of 0.4% trypan blue (Sigma). Live and dead cells were counted
within four microscopic fields of a haemocytometer. The viability was
calculated as a percentage relative to untreated cells.
[17] J.-P. Vigneron, N. Oudrhiri, M. Fauquet, L. Vergely, J.-C. Bradley,
M. Basseville, P. Lehn, J.-M. Lehn, Proc. Natl. Acad. Sci. USA 1996,
93, 9682 9686.
[18] a) M. C. O×Sullivan, D. M. Dalrymple, Tetrahedron Lett. 1995, 36,
3451 3452; b) D. Q. Xu, K. Prasad, O. Repic, T. J. Blacklock, Tetra-
hedron Lett. 1995, 36, 7357 7360; c) I. S. Blagbrough, A. J. Geall,
Tetrahedron Lett. 1998, 39, 439 442.
[19] M. Imazawa, F. Eckstein, J. Org. Chem. 1979, 44, 2039 2041.
[20] R. J. Bergeron, J. S. McManis, J. Org. Chem. 1988, 53, 3108 3111.
[21] a) B. W. Bycroft, W. C. Chan, S. R. Chhabra, P. H. T. Spittle, P. M.
Hardy, J. Chem. Soc. Chem. Commun. 1993, 776 77; b) S. R. Chha-
bra, A. N. Khan, B. W. Bycroft, Tetrahedron Lett. 1998, 39, 3585
3588.
[22] S. Lemaire-Audoire, M. Savignac, J. P. Genet, Tetrahedron Lett.
1995, 36, 1267 1269.
[23] J. Wityak, S. J. Gould, S. J. Hein, P. A. Keszler, J. Org. Chem. 1987,
52, 2179 2183.
[24] a) R. J. Bergeron, J. S. McManis, J. Org. Chem. 1987, 52, 1700
1703.; b) M. Moroni, B. Koksch, S. N. Osipov, M. Crucianelli, M. Fri-
gerio, P. Bravo, K. Burger, J. Org. Chem. 2001, 66, 130 133.
[25] M. A. Exposito, B. Lopez, R. Fernandez, M. Vazquez, C. Debitus, P.
Bravo, K. Burger, Tetrahedron 1998, 54, 7539 7550.
[26] A. R. Katritzky, B. V. Rogovoy, C. Chassaing, V. Vvedensky, J. Org.
Chem. 2000, 65, 8080 8082.
[27] Y.-Q. Wu, S. K. Hamilton, D. E. Wilkinson, G. S. Hamilton, J. Org.
Chem. 2002, 67, 7553 7556.
[28] D. M. Lynn, D. G. Anderson, D. Putnam, R. Langer, J. Am. Chem.
Soc. 2001, 123, 8155 8156.
[29] a) H. Gershon, R. Ghirlando, S. B. Guttman, A. Minsky, Biochemis-
try 1993, 32, 7143 7151; b) Y. Xu, F. C. Szoka, Biochemistry, 1996,
35, 5616 5623.
Acknowledgments
B.Y. would like to thank Ramkhamhaeng University for a studentship.
M.B. would like to thank the BBSRC. M.H. is supported by the Medical
Research Council and T.E. by Cancer Research UK.
[30] J. H. Felgner, R. Kumar, C. N. Sridhar, C. J. Wheeler, Y. J. Tsai, R.
Border, P. Ramsey, M. Martin, P. L. Felgner, J. Biol. Chem. 1994,
269, 2550 2561.
[31] R. P. Balasubramaniam, M. J. Bennett, A. M. Aberle, J. G. Malone,
M. H. Nantz, R. W. Malone, Gene Ther. 1996, 3, 163 172.
[32] J. A. Heyes, D. Niculescu-Duvaz, R. G. Cooper, C. J. Springer, J.
Med. Chem. 2002, 45, 99 114.
[1] R. G. Amado, I. S. Y. Chen, Biomedicine 1999, 285, 674 676.
[2] N. Somia, I. M. Verma, Nat. Rev. Genet. 2000, 1, 91 99.
[3] P. L. Felgner, Adv. Drug Delivery Rev. 1990, 5, 163 187.
[4] R. Harland, H. Weintraub, J. Cell Biol. 1985, 101, 1094 1099.
[5] W. C. Heiser, Methods Mol. Biol. 2000, 130, 117 134.
[6] E. T. Schenborn, V. Goiffon, Methods Mol. Biol. 2000, 130, 135
145.
[33] J.-P. Behr, B. Demeneix, J.-P. Loefeler, J. Perez-Mutul, Proc. Natl.
Acad. Sci. USA 1989, 86, 6982 6986.
[7] J. S. Pagano, Prog. Med. Virol. 1970, 12, 1 48.
[8] a) A. D. Miller, Angew. Chem. 1998, 110, 1862 1880; Angew. Chem.
Int. Ed. 1998, 37, 1768 1785.
[34] F. Denizot, . R. Lang, J. Immunol. Methods 1986, 89, 271 277.
[35] H. Niu, R. D. Simari, E. M. Zimmermann, G. M. Christman, Gene
Ther. 1998, 91, 735 740.
[9] P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan, M.
Wenz, J. P. Northrop, G. M. Ringold, M. Danielsen, Proc. Natl.
Acad. Sci. USA 1987, 84, 7413 7417.
[10] G. J. Nabel, E. G. Nabel, Z. Y. Yang, B. A. Fox, G. E. Plautz, X.
Gao, L. Huang, S. Shu, D. Gordon, A. E. Chang, Proc. Natl. Acad.
Sci. USA 1993, 90, 11307 11311.
[36] J.-P. Behr, Tetrahedron Lett. 1986, 27, 5861 5864.
[37] J.-P. Behr, Acc. Chem. Res. 1993, 26, 274 278.
[38] P. Pinnaduwage, L. Schmitt, L. Huang, Biochim. Biophys. Acta 1989,
985, 33 37.
Received: June 13, 2003
Revised: September 5, 2003 [F5232]
473
Chem. Eur. J. 2004, 10, 463 473
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim