Novel Biodegradable Pyridinium Amphiphiles for Gene Delivery

Article abstract of DOI:10.1002/ejoc.200300361

Biodegradable synthetic cationic pyridinium-based amphiphiles (SAINTs) prove to be promising non-viral carrier systems for delivery of DNA into eukaryotic cells. Six novel SAINTs were synthesised from 3,5-pyridinedicarboxylic acid as starting material, with two ester groups as linkers between the cationic headgroup and the hydrophobic tails. The vesicle-forming properties of the amphiphiles were studied by differential scanning calorimetry and transmission electron microscopy, whereas the hydrolysis of the diesters in water was investigated by NMR spectroscopy. Finally, the transfection potential and cytotoxicity were determined on COS-7 and HepG-2 cells in culture. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejoc.200300361

FULL PAPER
Novel Biodegradable Pyridinium Amphiphiles for Gene Delivery
[a]
[c]
[c]
[a]
[d]
ˇ
´
Dirk Pijper, Erna Bulten, Jarmila Smisterova, Anno Wagenaar, Dick Hoekstra,
Jan B. F. N. Engberts,*[a] and Ron Hulst*[a,b]
Keywords: Amphiphiles / Biodegradable / Gene technology / Vesicles
Biodegradable synthetic cationic pyridinium-based amphi-
philes (SAINTs) prove to be promising non-viral carrier sys-
tems for delivery of DNA into eukaryotic cells. Six novel
SAINTs were synthesised from 3,5-pyridinedicarboxylic acid
differential scanning calorimetry and transmission electron
microscopy, whereas the hydrolysis of the diesters in water
was investigated by NMR spectroscopy. Finally, the transfec-
tion potential and cytotoxicity were determined on COS-7
as starting material, with two ester groups as linkers between and HepG-2 cells in culture.
the cationic headgroup and the hydrophobic tails. The ves- ( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
icle-forming properties of the amphiphiles were studied by
Germany, 2003)
Introduction
DNA in eukaryotic cells.[8Ϫ14] Although promising, great
improvements in the transfection efficiency are still neces-
sary and their use is often accompanied by undesired cyto-
toxic complications.
Our group has previously reported on a class of cationic
pyridinium amphiphiles (SAINTs ϭ Synthetic Amphiphiles
Interdisciplinary) showing high transfection efficiency com-
bined with low cytotoxicity.[15Ϫ17] Thanks to the synthetic
ease with which structural modifications can be introduced
in the original SAINT concept, this class is very suitable
for studies regarding the relationship between chemical
structure, transfection efficiency and cytotoxicity.
Gene therapy is widely regarded as a promising method
in the treatment of diseases caused by genetic disorders
(e.g., cystic fibrosis) and, more importantly, can also be ap-
plied for the delivery of any type of therapeutic protein,
opening up possibilities for treatment of a wide range of
diseases, including infectious (e.g., AIDS), acquired (e.g.,
Parkinson’s, arthritis) or cardiovascular (e.g., arterioscle-
rosis) diseases, besides several types of cancer.[1Ϫ3] The con-
cept is straightforward, and involves the induction of ex-
pression of certain proteins in the diseased target cells
through the introduction of foreign DNA.
The most efficient carrier systems for the transfer of
DNA into cells (transfection) at present are viral vectors.[4]
These, however, are associated with risks regarding uncon-
trolled recombination of viral DNA with host DNA and,
in particular, immunogenicity.[5Ϫ7] These problems can
(partly) be circumvented by the use of nonviral synthetic
vector systems. Cationic amphiphiles have established a
promising position as reliable transfection agents, and nu-
merous studies have reported a number of cationic amphi-
philes promoting the delivery and expression of foreign
[a]
Physical Organic Chemistry Unit, Stratingh Institute,
University of Groningen,
Nijenborgh 4, 9747 AG Groningen, The Netherlands
Fax: (internat.) ϩ31-(0)50-3634269
E-mail: j.b.f.n.engberts@chem.rug.nl
[b]
Kiadis BV,
Niels Bohrweg 11Ϫ13, 2333 CA Leiden, The Netherlands
[c]
BiOMaDe Technology Foundation,
Nijenborgh 4, 9747 AG Groningen, The Netherlands.
[d]
Department of Membrane Cell Biology, University of
Figure 1. Generalised structures 1 and 2 of SAINT amphiphiles;
diester SAINTs 3 represent the amphiphiles used in the current
study; diesters 4 represent possible structural variations[18]
Groningen,
Antonius Deusinglaan 1, 9713 AV Groningen, The Netherlands
4406
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200300361
Eur. J. Org. Chem. 2003, 4406Ϫ4412

Novel Biodegradable Pyridinium Amphiphiles for Gene Delivery
FULL PAPER
In a recent study,[18] an ester moiety was introduced as a
linker between the cationic headgroup and the hydrophobic
tails in order to increase the transfection potential of the
cytofectin for enhanced (controlled) release.[19,20] Recent
findings clearly point in the direction of an endocytotic
pathway by which the lipoplexes are taken up by the cell.[21]
It was initially anticipated that, at the level of endosomal
escape, the lipoplex would be destabilised through acid-in-
duced hydrolysis of the amphiphiles, facilitating the escape
of DNA.[22,23] Incorporation of biodegradable ester groups
into the amphiphiles might also enhance intracellular
metabolism, thus avoiding long-term persistence of the un-
natural cationic lipids in cellular membranes and corre-
sponding
undesirable
consequences
for
cellular
functions.[24Ϫ26]
Here we describe the synthesis and some characteristics
of six new pyridinium amphiphiles in which the two hydro-
carbon tails are connected to the headgroup by two ester
linkages. In these diester amphiphiles, hydrolysis of the first
ester is significantly enhanced by the electron-withdrawing
cationic pyridine unit, in combination with the second ester
group. The hydrolysis of the second ester moiety is strongly
retarded through deactivation by the zwitterionic nature of
the headgroup after hydrolysis of the first ester moiety.
The selective hydrolysis of only one of the two esters of-
fers unprecedented potential to direct the morphology as a
function of the conditions of the surroundings or the endo-
somal stage. Hypothetically, selective hydrolysis of the lipo-
plex in the endosome could result in a transition from the
‘‘relatively stable’’ lamellar bilayer vesicular to an ‘‘aggres-
sive and fusogenic’’ micellar/hexagonal HII phase.[27] The
latter process might give rise to destabilisation and rupture
of the endosomal membrane, enhancing the release of DNA
into the cytoplasm and thus increasing the transfection ef-
ficiency.[28,29] Some of the physiochemical properties of the
new diester amphiphiles and the pH-dependent hydrolytic
behaviour of the esters are reported, together with the
transfection potentials.
Scheme 1. Synthesis of diester SAINTs 3; see text for expla-
nation[18]
For this sequence, the alcohols 8a/b/d/f were converted
into the corresponding mesylates 9a/b/d/f by straightfor-
ward treatment with methanesulfonyl chloride in the pres-
ence of triethylamine as a base (Scheme 2). Treatment of
mesylate 9a with NaI in dry acetone afforded iodide 10a.
Scheme 2. Synthesis of mesylates 9 and iodide 10a[18]
Subsequently, the diesters 7aϪg were methylated on the
nitrogen atom by treatment with a 20-fold excess of methyl
iodide in dry acetone, resulting in complete conversion into
the N-methylated products as the relatively unstable iodide
salts (for a discussion on the degradation pathways of the
iodide products, see ref.[36]). Ion exchange carried out with
the very mild ion-exchange material Sephadex (DEAE A25,
chloride form) yielded the desired amphiphiles 3aϪg as the
more stable chloride salts.
In addition to the long-tailed pyridinium amphiphiles,
the methyl ester 3g was prepared as a model compound in
order to investigate the hydrolytic behaviour of the diester
system without possible interference from the long alkyl
tails.
Results and Discussion
Synthesis
The diesters 7c and 7e (Scheme 1) were prepared by treat-
ment of 3,5-pyridinedicarboxylic acid (5) with excess
thionyl chloride to yield the acid chloride 6 (which was not
isolated), followed by addition of 2.2 equivalents of hexa-
decyl alcohol or octadecyl alcohol and 6 equivalents of tri-
ethylamine. The crude products were purified by column
chromatography, affording 7c and 7e as white solid materi-
als in reasonable yields.
Unfortunately, the esterification of tetradecyl alcohol and
the unsaturated alcohols by this methodology gave very low
yields. The diesters 7a, 7b, 7d and 7f were therefore pre-
pared by treatment of the caesium salt of 3,5-pyridinedicar-
Characteristics of the Amphiphilic Aggregates
All the amphiphiles used in this study form bilayer ves-
boxylic acid (5) with 2.2 equivalents of the appropriate icles after hydration and sonication, as verified by trans-
iodide or, alternatively, the mesylate, yielding the diester mission electron microscopy (TEM). Figure 2 shows a
products in excellent yields[30,31] (Scheme 1).
transmission electron micrograph of 3f, negatively stained
Eur. J. Org. Chem. 2003, 4406Ϫ4412
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4407

J. B. F. N. Engberts, R. Hulst et al.
FULL PAPER
ing a transfection experiment carried out at 37 °C, these
bilayers adopt a fluid liquid-crystalline phase, whereas bi-
layers of the ‘‘saturated’’ amphiphiles adopt an ordered,
gel-like phase. From these results it may be expected that
the amphiphiles with the unsaturated tails should show
higher transfection efficiencies than their counterparts with
saturated tails.[32,33] As can be seen below, this expectation
was partly fulfilled.
Hydrolysis
To investigate the hydrolytic behaviour of the diester am-
phiphiles, time-resolved pH-dependent 1H NMR spec-
troscopy was used to follow the hydrolysis of model com-
1
pound 3g. By recording of H NMR spectra at subsequent
time intervals and by measurement of the integrals of the
three protons on the pyridine ring, the concentrations of
the starting diester, the singly hydrolysed monoester and the
doubly hydrolysed diacid were monitored. The concen-
trations of the diester and the monoester are plotted as a
function of time and pH[34] in Figure 3.
Figure 2. Transmission electron micrograph (TEM) of vesicles of
3f, negatively stained with UAc; vesicle size varies between 40 and
200 nm; bar represents 200 nm
with uranyl acetate (UAc). The vesicle size typically varied
between 40 and 200 nm.
The fluidity of the lipid bilayer is generally believed to
play an important role in the interactions between the lipids
and the DNA in the lipid-DNA complex (lipoplex).[32,33] It
also appears that a high fluidity has a beneficial effect on
the transfection efficiency. The main phase transition tem-
peratures for the bilayer vesicles formed by amphiphiles
3aϪf were therefore determined by differential scanning
calorimetry (DSC); the results are summarized in Table 1.
Table 1. Phase transition temperatures of diester amphiphiles 3 as
determined by DSC
Compound (R)
PTT (°C)
3a (C14:0
3b (C14:1
)
)
)
46
Ͻ0
64
Ͻ0
72
3c (C16:0
3d (C16:1
3e (C18:0
3f (C18:1
)
)
)
Ͻ0
The amphiphiles with saturated hydrocarbon chains all
show clear phase transitions, with an increase of about 15°
for each two additional CH2 groups, consistently with ear-
lier findings.[17] The signal shifted slightly towards lower
temperatures in successive temperature scans, probably as a
result of partial hydrolysis of the amphiphiles due to the
elevated temperatures during the measurements. This hy-
Figure 3. Hydrolysis of the methyl ester 3g is monitored by follow-
ing the concentration changes of 3g (A) and singly hydrolysed
monoester (B) with time at different pHs; measurements were per-
formed at 37 °C in HEPES buffer (1 ), initial concentration of 3g
ca. 10 m
1
pothesis was confirmed by H NMR (results not shown),
which clearly showed that approximately 30% of the ester
Pseudo-first order rate constants (kobsd.) were determined
moieties in the amphiphiles had been hydrolysed after the from a fit to the curve in Figure 3 (A) and the second half
DSC experiment. The amphiphiles with unsaturated tails of the curve in Figure 3 (B) (starting from the moment that
showed no phase transitions in the temperature range meas- the concentration 3g is zero) for the first and the second
ured (0Ϫ100 °C); it is assumed that the bilayers formed by ester hydrolysis, respectively. The results are listed in
these amphiphiles have phase transitions below 0 °C. Dur- Table 2.
4408
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4406Ϫ4412

Novel Biodegradable Pyridinium Amphiphiles for Gene Delivery
FULL PAPER
10Ϫ15 min mixing at room temp. the resulting complexes
were added to culture cells (COS-7 or HepG-2) and incu-
bated at 37 °C for 48 h. The percentage of GFP-positive
cells was then identified by Fluorescent Activated Cell
Sorting (FACS). The transfection results are summarized in
Figure 4. Transfection efficiencies are compared with that
of SAINT-2, an amphiphile showing excellent transfection
efficiency and a generally remarkable low toxicity towards
different cell types.[15]
Table 2. Pseudo-first-order (kobsd.) rate constants for the hydrolysis
of the methyl ester 3g at different pHs
pH
kobsd. (minϪ1
1st hydrolysis
)
kobsd. (minϪ1
)
2nd hydrolysis
7.1
7.4
7.7
0.017
0.023
0.038
0.0038
0.004
0.0098
Both hydrolytic reactions are significantly accelerated at
higher pH, indicating that the observed rate is mainly being
determined by hydroxide ion-catalysed hydrolysis, whereas
hydrolysis at much lower pH values did not proceed at sig-
nificant rates. From plots of kobsd. against the hydroxide ion
Ϫ
concentration,
kOH
could
be
determined
as
8.2 l·molϪ1·minϪ1 for the first ester hydrolysis and
1.9 l·molϪ1·minϪ1 for the second. At endosomal pH (ഠ
5.0), kobsd. would be 8.19·10Ϫ5 minϪ1 for the first hydrolysis,
corresponding with a half-life of 141 h, while at physiologi-
cal pH (ഠ 7.4) kobsd. would be 2.3·10Ϫ3 minϪ1, with an ex-
pected half-life of 30 min. The endosomal pH drop strongly
retards the hydrolytic efficiency, instead of the expected ac-
celeration (vide supra).
It is also clear that the hydrolysis of the first ester moiety
is faster than the hydrolysis of the second ester group. At
physiological pH, kobsd. of the second hydrolysis is esti-
mated as 4.8·10Ϫ3 minϪ1, corresponding to a half-life of
143 h, in comparison with 30 min for the first ester group
hydrolysis. This can be explained by the fact that the first
ester is activated towards hydrolysis by the neighbouring
electron-withdrawing cationic headgroup as well as by the
second ester moiety, whereas after hydrolysis of the first es-
ter the remaining ester is adjacent only to a less strongly
electron-withdrawing zwitterionic headgroup. In this sys-
tem, controlled hydrolysis of one of the two ester linkages
is possible, offering the potential to direct the morphology
as a function of the pH conditions of the surroundings (or
stage of the transfection event).
Unfortunately it was not possible to use NMR to per-
form hydrolysis measurements on the amphiphiles in the
vesicular state, due to the high concentrations of substrate
and, in particular, the buffers needed, which caused the ves-
icles to aggregate and precipitate from the solution. It also
proved impossible to monitor the hydrolysis by other tech-
niques, such as by mapping the UV absorptions. We expect
the amphiphiles to show the same hydrolytic behaviour as
the model compound 3g; that is, base-catalysed hydrolysis
and faster hydrolysis of the first ester moiety than of the
second group.
Figure 4. Transfection efficiencies of SAINT diesters 3; transfection
efficiencies were evaluated by use of FACS, GFP was used as re-
porter protein; SAINT-2 transfection efficiency is arbitrarily set at
100%; cell survival is expressed as the percentage of surviving cells
after the transfection event
The transfection efficiencies of the SAINT diesters with
COS-7 cells ranged from 47 to 69% relative to SAINT-2,
in combination with a generally somewhat lower toxicity.
Previous work on non-ester SAINT amphiphiles had shown
that the introduction of a double bond in the hydrophobic
tail resulted in a significant increase in transfection ef-
ficiency,[17] whereas in monoester SAINTs the transfection
efficiency was considerably reduced.[18] For the SAINT di-
esters, a trend similar to that of the monoester adducts,
though less pronounced, is observed. It can be speculated
that this behaviour may arise from a different morphology
of the lipid bilayer formed by the diesters relative to the
non-ester SAINTs, caused by the significant differences in
headgroup structure and the fact that the hydrophobic tails
are connected to the pyridinium headgroup in a 3,5 orien-
tation and not to the same carbon atom as in the non-
Transfection of Cells in Culture
Information on the number of transfected cells was ob- ester SAINTs.
tained by the use of plasmid DNA containing Green Fluor-
Generally, lower transfection efficiencies were obtained
escent Protein (GFP). Vesicles composed of cationic lipid/ with HepG2 cells than with the COS-7 cell line. The clearest
DOPE in 1:1 ratio were mixed with an equal volume of differences between the two cell lines were observed for 3b
plasmid DNA in a molar charge ratio of 2.5:1. After (C14:0), 3c (C16:0) and 3d (C16:1). On HepG2 cells, 3b
Eur. J. Org. Chem. 2003, 4406Ϫ4412
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4409

J. B. F. N. Engberts, R. Hulst et al.
FULL PAPER
MgSO4. Evaporation of the solvents yielded slightly viscous, yel-
lowish oils.[18,31,35]
reached approximately 60% of the transfection efficiency
with COS-7 and, surprisingly, a two times higher efficiency
with the HepG2 cell line than seen with SAINT-2, whereas
3d showed no transfection at all. In relation to SAINT-2,
toxicity was generally somewhat higher, although on aver-
age high survival rates were observed.
Tetradecyl Methanesulfonate (9a):[18,35] Yield 5.42 g, 98%. 1H NMR
(CDCl3): δ ϭ 0.86 (t, 3 H), 1.23 (m, 22 H), 1.79 (m, 2 H), 2.98 (s,
3 H), 4.19 (t, 2 H) ppm.
(Z)-Tetradec-11-enyl Methanesulfonate (9b):[18,35] Yield 5.33 g, 97%.
1H NMR (CDCl3): δ ϭ 0.83 (t, 3 H), 1.30 (m, 14 H), 1.72 (m, 2
H), 2.00 (m, 4 H), 2.99 (s, 3 H), 4.20 (t, 2 H), 5.31 (m, 2 H) ppm.
Conclusions
(Z)-Hexadec-11-enyl Methanesulfonate (9d):[18,35] Yield 5.45 g, 90%.
1H NMR (CDCl3): δ ϭ 0.82 (t, 3 H), 1.24 (m, 18 H), 1.71 (m, 2
H), 1.98 (m, 4 H), 2.98 (s, 3 H), 4.19 (t, 2 H), 5.31 (m, 2 H) ppm.
Six novel cationic pyridinium-based amphiphiles were
prepared by employment of two hydrolytically active ester
moieties as linking entities between the headgroup and the
hydrocarbon tails. Although it was initially envisaged that
a drop in pH (from physiological to endosomal pH) should
trigger localised rapid hydrolysis of the esters, resulting in
destabilisation of the lipoplex in the endosome, the esters
appeared to hydrolyse by base catalysis, producing fast hy-
drolysis at physiological pH and higher stability in the en-
dosomal pH range. In this respect, the use of an acid-labile
or ‘‘activated’’ ketal functional moiety, which does hydro-
lyse by acid catalysis, would be an interesting option for
further research. Furthermore, it should be emphasised that
under physiological conditions hydrolysis might be depen-
dent not only on pH, but also, and probably more impor-
tantly, on enzymatic pathways, so an amide linker that
would hydrolyse solely by enzyme catalysis might also be
employed.
(Z)-Octadec-11-enyl Methanesulfonate (9f):[18,35] Yield 6.28 g, 95%.
1H NMR (CDCl3): δ ϭ 0.82 (t, 3 H), 1.24 (m, 22 H), 1.71 (m, 4
H), 1.98 (m, 4 H), 2.98 (s, 3 H), 4.19 (t, 2 H), 5.31 (m, 2 H) ppm.
1-Iodotetradecane (10a):[18,36] NaI (27.3 mmol) was added to a solu-
tion of tetradecyl mesylate (17.5 mmol) in acetone (50 mL), and
the mixture was heated at reflux for 2 h. After the mixture had
been cooled, the salts were filtered off. Low-temperature crystalli-
sation from ethanol yielded colourless crystals, which turned into
colourless oil at room temperature; yield 3.97 g, 70%. 1H NMR
(CDCl3): δ ϭ 0.84 (t, 3 H), 1.24 (m, 22 H), 1.81 (m, 2 H), 3.19 (t,
2 H) ppm.
General Procedure for the Synthesis of the Pyridinium Diesters 7c/
e:18,31 Thionyl chloride (30 mmol) was added dropwise to a suspen-
sion of 3,5-pyridinedicarboxylic acid (6 mmol) in toluene (20 mL).
After the mixture had been heated at reflux for one hour, solvents
and excess thionyl chloride were evaporated off. This yielded the
carboxylic chloride as a brown solid, which was used in the esterifi-
cation reaction without further purification. The diacid dichloride
(6 mmol) was added in one portion to a solution of alcohol
(13.2 mmol) in dichloromethane (50 mL). Triethylamine (36 mmol)
was added dropwise to the refluxing solution, and the reaction mix-
ture was heated at reflux overnight. After evaporation of the sol-
vents, the yellow mixture was suspended in n-hexane and the salts
were filtered off. The diester was obtained pure after column chro-
matography over Al2O3 (act. III), with dichloromethane/n-hexane
(1:1) as eluent. The products were obtained as colourless solids.
Despite the observed pH behaviour, diester amphiphiles
3 showed moderate to good transfection efficiencies in re-
lation to SAINT-2 on the two cell lines studied. Moreover,
the observed toxicity was generally very low, though
strongly dependent upon the cell lines under study.
Experimental Section
General: All reactions were carried out in anhydrous solvents under
nitrogen in oven-dried glassware. For column chromatography,
Al2O3 (activity II-III) was used, prepared by addition of the indi-
cated amount of water to Merck Al2O3 90 active neutral (activity
I). Melting points were determined with a Kofler melting point
microscope. Several compounds displayed liquid crystalline behav-
iour; in these cases no melting points are reported. 1H NMR spec-
tra were recorded with Varian Gemini 200 and Varian VXR 300
spectrometers operating at 200 and 300 MHz for the proton chan-
nels, respectively. Elemental analyses were carried out at the Ana-
lytical Department of the University of Groningen. (Z)-11-Tetrade-
cen-1-ol, (Z)-11-hexadecen-1-ol, oleyl alcohol, 1-tetradecanol, 1-
hexadecanol and 1-octadecanol were obtained from Aldrich,
whereas 3,5-pyridinedicarboxylic acid was purchased from Acros
Organics.
Dihexadecyl Pyridine-3,5-dicarboxylate (7c):[18,31] Yield 1.48 g,
40%. 1H NMR (CDCl3): δ ϭ 0.87 (t, 6 H), 1.22 (m, 52 H), 1.79
(m, 4 H), 4.38 (t, 4 H), 8.82 (s, 1 H), 9.38 (s, 2 H) ppm.
Dioctadecyl Pyridine-3,5-dicarboxylate (7e):[18,31] Yield 1.61 g, 40%.
1H NMR (CDCl3): δ ϭ 0.83 (t, 6 H), 1.22 (m, 60 H), 1.74 (m, 4
H), 4.37 (t, 4 H), 8.81 (s, 1 H), 9.33 (s, 2 H) ppm.
Procedure for the Synthesis of the Pyridinium Diesters 7b/d/f:[31]
Caesium carbonate (6 mmol) was added to a suspension of 3,5-
pyridinedicarboxylic acid (6 mmol) in DMF (40 mL). The mixture
was stirred for 1 h. Mesylate (13.2 mmol) was added dropwise, and
the mixture was stirred overnight at 75 °C. After evaporation of
the solvents, the yellow mixture was suspended in n-hexane and the
salts were filtered off. The pure diester was obtained as a colourless
oil after column chromatography over Al2O3 (act. III), with di-
chloromethane/n-hexane (1:1) as eluent.
General Procedure for the Synthesis of Mesylates 9a/b/d/f: Triethyl-
amine (30 mmol) and then mesyl chloride (30 mmol) were added
dropwise to a solution of alcohol (20 mmol) in dichloromethane
(50 mL), cooled to 0 °C. After stirring at 0 °C for 40 min, the reac-
tion mixture was quenched with water (15 mL). The organic layer
was washed with 1  HCl, a saturated Na2CO3 solution and brine,
after which the combined aqueous layers were extracted with di-
chloromethane. The combined organic layers were dried with
(Z)-Di(tetradec-11-enyl) Pyridine-3,5-dicarboxylate (7b):[31] Colour-
less oil, yield 0.67 g, 20%. 1H NMR (CDCl3): δ ϭ 0.93 (t, 6 H),
1.23 (m, 28 H), 1.75 (m, 4 H), 1.95 (m, 8 H), 4.35 (t, 4 H), 5.30
(m, 4 H), 8.81 (s, 1 H), 9.35 (s, 2 H) ppm.
4410
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4406Ϫ4412

Novel Biodegradable Pyridinium Amphiphiles for Gene Delivery
FULL PAPER
(Z)-Di(hexadec-11-enyl) Pyridine-3,5-dicarboxylate (7d):[31] Colour-
12.88, 18.98, 24.27, 25.57, 26.90, 27.70, 27.76, 28.00, 28.25, 48.89,
less oil, yield 1.84 g, 50%. 1H NMR (CDCl3): δ ϭ 0.85 (t, 6 H), 66.18, 127.76, 129.00, 129.99, 142.79, 148.47, 159.23 ppm.
1.22 (m, 36 H), 1.72 (m, 4 H), 1.95 (m, 8 H), 4.33 (t, 4 H), 5.28 C36H60ClNO4·1H2O: calcd. C 69.26, H 10.01, N 2.24; found C
(m, 4 H), 8.79 (s, 1 H), 9.30 (s, 2 H).
69.34, H 10.25, N 2.26. Mass calcd. for C36H60ClNO4: 606.330;
found 555.4 [ Ϫ HCl and CH3].
(Z)-(Dioctadec-11-enyl) Pyridine-3,5-dicarboxylate (7f):[31] Colour-
less oil, yield 2.32 g, 58%. 1H NMR (CDCl3): δ ϭ 0.85 (t, 6 H), 1-Methyl-3,5-bis(hexadecyloxycarbonyl)pyridinium Chloride (3c):[30]
1.23 (m, 44 H), 1.78 (m, 4 H), 1.99 (m, 8 H), 4.38 (t, 4 H), 5.37 Colourless solid, m.p. 113Ϫ115 °C, yield 0.43 g, 65%.
(m, 4 H), 8.82 (s, 1 H), 9.38 (s, 2 H) ppm.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.82 (t, 6 H), 1.25 (m, 52 H),
Ditetradecyl Pyridine-3,5-dicarboxylate (7a):[31] Caesium carbonate
1.80 (m, 4 H), 4.42 (t, 4 H), 4.90 (s, 3 H), 9.25 (s, 1 H), 9.72 (s, 2
H) ppm.
(6 mmol) was added to a suspension of 3,5-pyridinedicarboxylic
acid (6 mmol) in DMF (40 mL). The mixture was stirred for 1 h.
1-Iodotetradecane (13.2 mmol) was added and the mixture was
stirred overnight at 75 °C. After evaporation of the solvents, the
yellow mixture was suspended in n-hexane and the salts were fil-
tered off. The pure diester was obtained as a white solid after col-
umn chromatography over Al2O3 (act. III), with dichloromethane/
n-hexane (1:1) as eluent. Colourless solid, yield 3.17 g, 88%. 1H
NMR (CDCl3): δ ϭ 0.87 (t, 6 H), 1.22 (m, 44 H), 1.78 (m, 4 H),
4.38 (t, 4 H), 8.82 (s, 1 H), 9.36 (s, 2 H) ppm.
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.82 (t, 6 H), 1.27 (m, 52
H), 1.81 (m, 4 H), 4.43 (t, 4 H), 5.05 (s, 3 H), 9.35 (s, 1 H), 9.91
(s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ 12.62, 21.18, 24.29, 26.91,
27.72, 27.86, 28.03, 28.20, 30.42, 48.92, 66.24, 129.08, 142.88,
148.40, 159.20 ppm. C40H72ClNO4: calcd. C 72.09, H 10.89, N
2.10; found
C 71.83, H 10.85, N 2.10. Mass calcd. for
C40H72ClNO4: 666.469; found 615.5 [Ϫ HCl and CH3].
(Z)-1-Methyl-3,5-bis(hexadec-11-enyloxycarbonyl)pyridinium Chlo-
ride (3d): Colourless, waxy solid, yield 0.32 g, 48%.
Dimethyl Pyridine-3,5-dicarboxylate (7g):[37] Sulfuric acid (15 mL)
was added dropwise to a suspension of 3,5-pyridinedicarboxylic
acid (6 mmol) in methanol (150 mL), while the mixture was cooled
with ice, after which the mixture was heated at reflux overnight.
The pH of the mixture was neutralised by addition of solid
Na2CO3, and the mixture was extracted several times with diethyl
ether. After drying with MgSO4, evaporation of the diethyl ether
yielded a colourless solid, yield 0.82 g, 70%. 1H NMR (CDCl3):
δ ϭ 3.98 (s, 6 H), 8.87 (t, 1 H), 9.37 (d, 2 H) ppm.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.86 (t, 6 H), 1.25 (m, 36 H),
1.81 (m, 4 H), 1.99 (m, 8 H), 4.41 (t, 4 H), 4.90 (s, 3 H), 5.33 (m,
4 H), 9.23 (s, 1 H), 9.73 (s, 2 H) ppm.
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.91 (t, 6 H), 1.25 (m, 36
H), 1.80 (m, 4 H), 1.99 (m, 8 H), 4.41 (t, 4 H), 4.99 (s, 3 H), 5.33
(s, 3 H), 9.30 (s, 1 H), 9.85 (s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ
12.49, 20.82, 24.27, 25.39, 25.67, 26.90, 27.72, 27.78, 28.01, 28.25,
30.44, 48.90, 66.18, 128.30, 128.32, 129.02, 142.80, 148.46, 159.23
ppm. C40H68ClNO4·1H2O: calcd. C 70.61, H 10.37, N 2.06; found
C 70.26, H 10.50, N 2.06. Mass calcd. for C40H68ClNO4: 662.437;
found 611.4 [Ϫ HCl and CH3].
General Procedure for the Synthesis of the Methylated Pyridinium
Diesters 3aϪg: A solution of pyridine diester 7 (1 mmol) and
methyl iodide (20 mmol) in dry acetone (30 mL) was heated at re-
flux overnight. After evaporation of the solvent and excess methyl
iodide, the crude iodide salts were used in ion-exchange chromatog-
raphy without further purification. Ion-exchange chromatography
over a Sephadex column (chloride form, DEAE, A25), with meth-
anol as eluent, afforded the products as colourless/waxy solids.
These were crystallised from acetonitrile (the waxy solids at low
temperature).
1-Methyl-3,5-bis(octadecyloxycarbonyl)pyridinium Chloride (3e):
Colourless solid, m.p. 116Ϫ118 °C, yield 0.33 g, 46%.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.82 (t, 6 H), 1.25 (m, 60 H),
1.78 (m, 4 H), 4.40 (t, 4 H), 4.81 (s, 3 H), 9.24 (s, 1 H), 9.63 (s, 2
H) ppm.
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.83 (t, 6 H), 1.26 (m, 60
H), 1.81 (m, 4 H), 4.41 (t, 4 H), 5.00 (s, 3 H), 9.35 (s, 1 H), 9.83
(s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ 12.63, 21.19, 24.29, 26.92,
27.72, 27.87, 28.03, 28.21, 30.42, 48.90, 66.25, 129.07, 142.90,
148.37, 159.16 ppm. C44H80ClNO4: calcd. C 73.14, H 11.16, N
1-Methyl-3,5-bis(tetradecyloxycarbonyl)pyridinium Chloride (3a):
Colourless solid, m.p. 115Ϫ117 °C, yield 0.30 g, 50%.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.80 (t, 6 H), 1.22 (m, 44 H),
1.78 (m, 4 H), 4.39 (t, 4 H), 4.86 (s, 3 H), 9.21 (s, 1 H), 9.69 (s, 2
H) ppm.
1.94; found
C 72.86, H 11.08, N 1.95. Mass calcd. for
C44H80ClNO4: 722.576; found 671.5 [Ϫ HCl and CH3].
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.82 (t, 6 H), 1.23 (m, 44
H), 1.80 (m, 4 H), 4.41 (t, 4 H), 5.05 (s, 3 H), 9.33 (s, 1 H), 9.91
(s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ 12.61, 21.18, 24.29, 26.90,
27.72, 27.85, 28.01, 28.15, 30.41, 48.87, 66.20, 129.00, 142.81,
148.45, 159.21 ppm. C36H64ClNO4: calcd. C 70.84, H 10.57, N
(Z)-1-Methyl-3,5-bis(octadec-11-enyloxycarbonyl)pyridinium Chlo-
ride (3f): Colourless waxy solid, yield 0.36 g, 50%.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.81 (t, 6 H), 1.25 (m, 44 H),
1.77 (m, 4 H), 1.95 (m, 8 H), 4.38 (t, 4 H), 4.82 (s, 3 H), 5.29 (m,
4 H), 9.19 (s, 1 H), 9.70 (s, 2 H) ppm.
2.29; found
C 70.83, H 10.58, N 2.30. Mass calcd. for
C36H64ClNO4 610.361; found 559.4 [ Ϫ HCl and CH3].
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.81 (t, 6 H), 1.25 (m, 44
H), 1.77 (m, 4 H), 1.97 (m, 8 H), 4.40 (t, 4 H), 4.98 (s, 3 H), 5.30
(s, 3 H), 9.25 (s, 1 H), 9.83 (s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ
12.58, 21.13, 24.25, 25.66, 26.87, 27.70, 27.77, 27.89, 27.97, 28.11,
28.21, 30.36, 31.06, 48.78, 66.08, 128.16, 128.42, 128.82, 142.66,
148.56, 159.21 ppm. No conclusive elemental analysis could be ob-
tained, probably due to small remainders of solvents.
(Z)-1-Methyl-3,5-bis(tetradec-11-enyloxycarbonyl)pyridinium Chlo-
ride (3b): Colourless, waxy solid, yield 0.27 g, 45%.
1
Iodide Salt: H NMR (CDCl3): δ ϭ 0.90 (t, 6 H), 1.27 (m, 28 H),
1.79 (m, 4 H), 1.98 (m, 8 H), 4.41 (t, 4 H), 4.82 (s, 3 H), 5.30 (m,
4 H), 9.25 (s, 1 H), 9.67 (s, 2 H) ppm.
Chloride Salt: 1H NMR (CDCl3): δ ϭ 0.91 (t, 6 H), 1.27 (m, 28 C44H76ClNO4: calcd. C 73.55, H 10.66, N 1.95; found C 72.76, H
H), 1.79 (m, 4 H), 1.99 (m, 8 H), 4.41 (t, 4 H), 4.95 (s, 3 H), 5.30 11.34, N 1.95. Mass calcd. for C44H76ClNO4: 718.545; found 667.4
(s, 3 H), 9.26 (s, 1 H), 9.80 (s, 2 H) ppm. 13C NMR (CDCl3): δ ϭ [Ϫ HCl and CH3].
Eur. J. Org. Chem. 2003, 4406Ϫ4412
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4411

J. B. F. N. Engberts, R. Hulst et al.
FULL PAPER
1-Methyl-3,5-bis(methyloxycarbonyl)pyridinium
[11]
[12]
B. Abdallah, L. Sachs, B. A. Demeneix, Biol. Cell 1995, 85,
1Ϫ7.
G. Byk, C. Dubertret, V. Escriou, M. Frederic, G. Jaslin, R.
Rangara, B. Pitard, J. Crouzet, P. Wils, B. Schwartz, D. Scher-
man, J. Med. Chem. 1998, 41, 229Ϫ235.
Chloride
(3g):
Colourless solid, decomposes above 150 °C, yield 0.20 g, 80%.
1
Iodide Salt: H NMR (D2O): δ ϭ 4.12 (s, 6 H), 4.59 (s, 3 H), 9.51
(s, 1 H), 9.66 (s, 2 H) ppm.
[13]
[14]
[15]
[16]
[17]
J. Marshall, J. B. Nietupski, E. R. Lee, C. S. Siegel, P. W. Rafter,
S. A. Rudginsky, C. D. Chang, S. J. Eastman, D. J. Harris, R.
K. Scheule, S. H. Cheng, J. Drug Target. 2000, 7, 453Ϫ469.
M. J. Bennett, A. M. Aberle, R. P. Balasubramaniam, J. G.
Malone, R. W. Malone, M. H. Nantz, J. Med. Chem. 1997,
40, 4069Ϫ4078.
I. van der Woude, H. W. Visser, M. B. A. ter Beest, A. Wag-
enaar, M. H. J. Ruiters, J. B. F. N. Engberts, D. Hoekstra, Bi-
ochim. Biophys. Acta 1995, 1240, 34Ϫ40.
Chloride Salt: 1H NMR (D2O): δ ϭ 4.12 (s, 6 H), 4.59 (s, 3 H),
9.54 (s, 1 H), 9.69 (s, 2 H) ppm. 13C NMR (D2O): δ ϭ 47.52, 52.56,
129.25, 143.66, 148.00, 161.07 ppm. C10H12ClNO4·1/2H2O: calcd.
C 47.16, H 5.15, N 5.50; found C 47.22, H 5.10, N 5.56. Mass
calcd. for C10H12ClNO4: 245.663; found 210.0 [Ϫ HCl].
Preparation of the Vesicle Solutions: A solution of the amphiphile
in a small amount of dichloromethane was concentrated under a
stream of nitrogen. The resulting lipid film was subjected to high
vacuum for several hours in order to remove all the residual sol-
vent. The lipid was then hydrated in water to a final lipid concen-
tration of 2 m and sonicated for several minutes at a temperature
above the phase transition temperature of the particular cationic
amphiphile by use of a Branson B15 sonication tip.
I. van der Woude, A. Wagenaar, A. A. P. Meekel, M. B. A. ter
Beest, M. H. J. Ruiters, J. B. F. N. Engberts, D. Hoekstra, Proc.
Natl. Acad. Sci. USA 1997, 94, 1160Ϫ1165.
ˇ
´
A. A. P. Meekel, A. Wagenaar, J. Smistrova, J. E. Kroeze, P.
Haadsma, B. Bosgraaf, M. C. A. Stuart, A. Brisson, M. H. J.
Ruiters, J. B. F. N. Engberts, D. Hoekstra, Eur. J. Org. Chem.
2000, 4, 665Ϫ673.
[18]
ˇ
EM, DSC, and NMR Measurements: EM experiments (negative
staining) were carried out with a Jeol JEM 1200 EX electron micro-
scope operating at 80 kV. Samples were prepared on carbon-coated
collodion grids and stained with uranyl acetate (UAc). Differential
scanning calorimetry measurements were performed with a Micro-
cal VP-DSC apparatus. Samples used were 2 m, and measure-
ments were carried out with heating rates of 1 °C/min. Time-
resolved 1H NMR experiments were performed on a Varian
UnityPlus 500 MHz spectrometer. Stock solutions of 3g (20 m in
D2O at pH 5.5) were mixed 1:1 with HEPES buffer (2  in D2O at
pH 7.1/7.4/7.7) prior to measurement.
´
A. Roosjen, J. Smistrova, C. Driessen, J. T. Anders, A. Wag-
enaar, D. Hoekstra, R. Hulst, J. B. F. N. Engberts, Eur. J. Org.
Chem. 2002, 7, 1271Ϫ1277.
[19]
[20]
J. A. Boomer, D. H. Thompson, S. M. Sullivan, Pharm. Res.
2002, 19, 1292Ϫ1301.
S. Obika, W. Yu, A. Shimoyama, T. Uneda, T. Minami, K.
Miyashita, T. Doi, T. Imanishi, Biol. Pharm. Bull. 1999, 22,
187Ϫ190.
I. S. Zuhorn, R. Kalicharan, D. Hoekstra, J. Biol. Chem. 2002,
277, 18021Ϫ18028.
B. Tycko, F. R. Maxfield, Cell 1982, 28, 643Ϫ651.
I. Mellman, R. Fuchs, A. Helenius, Annu. Rev. Biochem. 1986,
55, 663Ϫ700.
[21]
[22]
[23]
Transfection Experiments: Cells (COS-7 and HepG-2) were cul-
tured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, The
Netherlands) containing 7% fetal calf serum, 2 m -glutamine,
100 units/mL penicillin and 100 mg/mL streptomycin at 37 °C and
under 5% CO2. EGFP-N1 plasmid DNA (Clontech, Palo Alto,
CA) was isolated from Escherichia coli by use of a Qiagen plasmid
kit (Qiagen, Chatsworth, CA) by the manufacturerЈs instructions.
Cells (COS-7 and HepG-2) were seeded into twelve-well plates at
1.5·105 (HepG2)/0.8·105 (COS-7) cells per well and allowed to grow
overnight. Lipoplexes were prepared in cell culture medium
(DMEM): 0.5 mL of medium, containing 1 µg of pEGFP-N1
(Clontech) was added to an equal volume of medium containing
cationic lipid, at a charge ratio (cationic lipid/DNA) of 2.5:1. After
the incubation with lipoplexes for 4 h at 37 °C, the transfection
medium was replaced by complete cell culture medium. After 24 h,
the medium was refreshed, and after another 24 h the cells express-
ing reporter protein (GFP) were quantified by FACS analysis (Elite,
Coulter, Miami, FL; λex ϭ 488 nm, λem ϭ 530 nm; 5000 events).
[24]
[25]
[26]
[27]
H. Farhood, R. Bottega, R. M. Epand, L. Huang, Biochim.
Biophys. Acta 1992, 1111, 239Ϫ346.
R. Leventis, J. R. Silvius, Biochim. Biophys. Acta 1990, 1023,
124Ϫ132.
A. M. Aberle, F. Tablin, J. Zhu, N. J. Walker, D. C. Gruenert,
M. H. Nantz, Biochemistry 1998, 37, 6533Ϫ6540.
For an example of the morphological behaviour of SAINT ma-
ˇ
´
terials see: J. Smisterova, A. Wagenaar, M. C. A. Stuart, E.
Polushkin, E. ten Brinke, R. Hulst, J. B. F. N. Engberts, D.
Hoekstra, J. Biol. Chem. 2001, 276, 47615Ϫ47622.
S. Audouy, D. Hoekstra, Mol. Membr. Biol. 2001, 18, 129Ϫ143.
P. C. Bell, M. Bergsma, I. P. Dolbnya, W. Bras, M. C. A. Stuart,
A. E. Rowan, M. C. Feiters, J. B. F. N. Engberts, J. Am. Chem.
Soc. 2003, 125, 1551Ϫ1558.
[28]
[29]
[30]
[31]
[32]
E. J. R. Südholter, J. B. F. N. Engberts, J. Am. Chem. Soc. 1980,
102, 2467.
E. J. R. Südholter, J. B. F. N. Engberts, W. H. de Jeu, J. Phys.
Chem. 1982, 86, 1908Ϫ1913.
J. H. Felgner, R. Kumar, C. N. Sridhar, C. J. Wheeler, Y. J. Tjai,
R. Border, P. Ramsey, M. Martin, P. L. Felgner, J. Biol. Chem.
1994, 269, 2550Ϫ2561.
I. S. Zuhorn, V. Oberle, W. H. Visser, J. B. F. N. Engberts, U.
Bakowsky, E. Polushkin, D. Hoekstra, Biophys. J. 2002, 83,
2096Ϫ2108.
[1]
R. C. Crystal, Science 1995, 270, 404Ϫ410.
[33]
[2]
W. F. Anderson, Nature 1998, 392 (Suppl.), 25Ϫ30.
[3]
K. K. Hunt, S. A. Vorburger, Science 2002, 297, 415Ϫ416.
I. MacLachlan, P. Cullis, R. W. Graham, Curr. Opin. Mol.
[4]
[34]
[35]
[36]
Formally the term pD instead of pH should be used, as the
measurements were performed in D2O.
Ther. 1999, 1, 252Ϫ259.
Y. Ilan, G. Droguett, N. R. Chowdhury, Y. Li, K. Sengupta,
[5]
W. J. Baumann, H. K. Mangold, J. Org. Chem. 1964, 29,
3055Ϫ3069.
N. R. Thummala, A. Davidson, J. R. Chowdhury, MS Horwitz,
Proc. Natl. Acad. Sci. USA 1997, 94, 2587Ϫ2592.
[6]
ˇ
I. Muizebelt, P. Oosting, C. van der Pol, A. Wagenaar, J. Smis-
G. M. Rubanyi, Mol. Asp. Med. 2001, 22, 113Ϫ142.
[7]
´
terova, E. Bulten, C. Driessen, I. S. Zuhorn, D. Hoekstra, R.
D. Ferber, Science 2001, 294, 1638Ϫ1642.
[8]
Hulst., J. B. F. N. Engberts, Eur. J. Org. Chem., manuscript
accepted.
M. Kristiansen, J. Eriksen, K. A. Joergensen, J. Chem. Soc.,
Perkin Trans. 1 1990, 1, 101Ϫ103.
A. D. Miller, Angew. Chem. Int. Ed. 1998, 37, 1768Ϫ1785.
[9]
J. P. Behr, Bioconj. Chem. 1994, 5, 382Ϫ389.
[37]
[10]
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, 7413Ϫ7417.
Received June 18, 2003
4412
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4406Ϫ4412