Cationic nucleoside lipids derived from universal bases: A rational approach for siRNA transfection

Article abstract of DOI:10.1021/bc100005k

Cationic nucleoside lipids (CNLs) derived from 5-nitroindole and 4-nitroimidazole bases were prepared from d-ribose by using a straightforward chemical synthesis. TEM experiments indicate that these amphiphilic molecules self-assemble to form supramolecul

Full text of DOI:10.1021/bc100005k

1062
Bioconjugate Chem. 2010, 21, 1062–1069
Cationic Nucleoside Lipids Derived from Universal Bases: A Rational
Approach for siRNA Transfection
Claire Ceballos,† Salim Khiati,‡ Carla A. H. Prata,§ Xiao-Xiang Zhang,§ Suzanne Giorgio,† Philippe Marsal,†
Mark W. Grinstaff,§ Philippe Barthe´le´my,*,‡ and Michel Camplo*,†
Centre Interdisciplinaire de Nanoscience de Marseille CINaM, Upr-Cnrs 3118, Universite´ Aix-Marseille II, Luminy, case
913, 13288 Marseille cedex 09, France, Inserm, U869, Universite´ de Bordeaux, F-33076, Bordeaux, France, and
Departments of Chemistry and Biomedical Engineering, Metcalf Center for Science and Engineering, Boston University,
Boston, Massachusetts 02215. Received January 6, 2010; Revised Manuscript Received April 27, 2010
Cationic nucleoside lipids (CNLs) derived from 5-nitroindole and 4-nitroimidazole bases were prepared from
D-ribose by using a straightforward chemical synthesis. TEM experiments indicate that these amphiphilic molecules
self-assemble to form supramolecular organizations in aqueous solutions. Electrophoresis and standard ethidium
bromide (EB) fluorescence displacement assay shows that CNLs are able to bind siRNA. We demonstrated that
both the nature of the universal bases and the stereochemistry of the anomeric position (R, ꢀ) have an impact on
the CNLs-siRNA complex formation. Correlations among chemical structure, stereochemistry, siRNA knockdown
effect, and binding affinities for all the compounds were shown and analyzed with a simple molecular modeling
study. The best binding affinities for siRNA were found for the ꢀ anomer of the 5-nitroindole CNL which exhibits
protein knockdown activity similar to the standard siPORT NeoFX positive control. It is noteworthy that no
significant cytotoxicity at the tested concentration was observed for the novel CNLs.
INTRODUCTION
Currently, there is substantial interest in the design of
bioinspired molecules having a double functionality based on
the combination of nucleic acids and lipid characteristics (1-4).
Nucleoside lipids are emerging as tools for constructing both
supramolecular assemblies and nucleic acid delivery applica-
tions. The 2′-deoxyribosyl derivatives of 3-nitropyrrole and
5-nitroindole represent a new class of compounds known as
universal bases. When incorporated into oligonucleotides, they
show little discrimination in their base-pairing properties, as
they have no hydrogen bonding capability. Recently (5), we
Figure 1. Chemical structures of the (ꢀ) universal nucleobase (left)
and both CNLs (tosylate salts of 1′-(2′,3′-dioleyl-5′-trimethylammo-
have described the synthesis and siRNA transfection activities
of cationic nucleoside lipids (CNLs) A and B based on a
3-nitropyrrole universal base (Figure 1). The encouraging
biological results with such derivatives prompt us to further
investigate the potential of new CNL analogues containing the
5-nitroindole and 4-nitroimidazole universal bases (Figure 2).
We hypothesized that the chemical structure of CNLs could
have an impact on the siRNA/CNL stability and siRNA delivery.
Indeed, puzzling questions are emerging from the siRNA/CNL
complexes, for example, how can the structure and/or the stere-
ochemistry of the universal base affect the formation siRNA/CNL
or the transfection properties? To address these issues, the following
two structural parameters were modulated: (i) the nature of the
universal bases (5-nitroindole and 4-nitroimidazole) and (ii) the
stereochemistry of the anomeric position.
The target CNLs shown in Figure 2 possess three key
components: a quaternized amine at the 5 position, two oleic
acid chains linked at the 2 and 3 positions, and a 5-nitroindole
or a 4-nitroimidazole attached at the anomeric position. As for
the 3-nitropyrrole derivatives A and B shown in Figure 1, the
cationic charge of these new analogues is responsible for electro-
static binding to the nucleic acids, whereas the oleyl chains
nium-R-D-ribofuranosyl)-3-nitropyrrole A and 1′-(2′,3′-dioleyl-5′-tri-
methylammonium-ꢀ-D-ribofuranosyl)-3-nitropyrrole B (2).
contribute to stabilization of the resulting nucleic acid-nucleolipid
assembly through hydrophobic chain-chain interactions.
The 5-nitroindole universal base was selected because 1-(2′-
deoxyribofuranosyl)-5-nitroindole has been shown to be gener-
ally less destabilizing than 1-(2′-deoxyribofuranosyl)-3-nitro-
pyrrole analogue, with only a 2 °C decrease in melting
temperature (Tm) value when incorporated toward the end of a
17mer duplex and 5 °C in the middle (6). It was suggested that
this stabilization occurs because this nucleoside is sufficiently
large to be able to intercalate into the opposite strand. The crystal
structure of 1-(2′-deoxyribofuranosyl)-5-nitroindole shows that
the base residues stack in columns, with overlap between each
aromatic base (7). Also, 1-(2′-deoxyribofuranosyl)-5-nitroindole
is nondiscriminating toward the natural DNA bases, with a Tm
range of 3 °C, similar to that observed with the 3-nitropyrrole
analogue.
Herein, we report the synthesis, physicochemical properties,
and siRNA transfection efficiency of the universal-base nucleo-
lipids, 2′,3′-dioleyl-5′-O-(R,ꢀ)-D-ribofuranosyl-5-nitroindole and
2′,3′-dioleyl-5′-O-(R-D-ribofuranosyl-4-nitroimidazole.
* Corresponding author. camplo@univmed.fr.
† Upr-Cnrs 3118.
MATERIAL AND METHODS
‡ Inserm, U869.
Unless noted otherwise, all starting materials were purchased
from Sigma-Aldrich, Acros, or Alfa Aesar and used as received.
§ Boston University.
10.1021/bc100005k  2010 American Chemical Society
Published on Web 05/19/2010

Cationic Nucleoside Lipids from Universal Bases
Bioconjugate Chem., Vol. 21, No. 6, 2010 1063
Figure 2. Chemical structures of the universal cationic nucleoside lipids (CNLs) used in this study. Tosylate salts of 1-(2,3-dioleyl-5-
trimethylammonium-R-D-ribofuranosyl)-5-nitroindole and 1-(2,3-dioleyl-5-trimethylammonium-ꢀ-D-ribofuranosyl)-5-nitroindole, 7a and 7b, re-
spectively. 1-(2,3-dioleyl-5-trimethylammonium-ꢀ-D-ribofuranosyl)-4-nitroimidazole) 12a.
Some solvents were distilled under argon over appropriate
drying agent: tetrahydrofuran and diethylether over sodium/
benzophenone, dichloromethane over CaCl2. Unless otherwise
stated, the other solvents were used as commercially available.
All compounds were characterized by 1H NMR, 13C NMR, and
mass spectroscopy. 250 and 62.5 MHz 13C nuclear magnetic
resonance (NMR) spectra were recorded on a Bruker AC 250.
The spectra were taken at 20 °C and referenced against residual
proton-solvent signals (1H, CDCl3, 7.26 ppm; and 13C, 77 ppm).
The 1H NMR coupling constants (J) are reported in hertz. ESI-
MS and elemental analyses were performed at the spectropole
of the Faculte´ des Sciences et Techniques de Saint Jeroˆme,
Marseille, France. Electrospray ionization mass spectra (ESI-
MS) were recorded with an API///Plus sciex triple quadripole
spectrophotometer in the positive mode. Elemental analyses
were recorded with a Thermo Finnigan EA 1112 elemental
analysis apparatus driven by the Eager 300 software. TEM
microscopy experiments were performed on a Philipps CM10
(negative staining with ammonium molybdate 1% in water, Cu/
Pd carbon coated grids). Analytical thin layer chromatography
(TLC) was performed on precoated glass plates with Kieselgel
60F254 neutral with aluminum support plate (0.25 mm layer
thickness), SDS. The developed plates were air-dried and
exposed to UV light and/or sprayed with a solution of cerium
sulfate (1%) and molybdic acid (1.5%) in 10% aqueous sulfuric
acid and heated at 150 °C. Preparative thin layer chromatography
was performed on glass plates with Merck silica gel 60F-254
as the adsorbent (layer thickness 2 mm or 1 mm). Column
chromatography was performed on Merck silica gel 60
(0.063-0.200 mm). Melting point (mp) were measured with
an Electrothermal 9100. The synthesis of compounds 1, 2, and
2′ is described in the Supporting Information.
1-(5-O-tert-Butyldimethylsilyl-2,3-O-isopropylidene-D-ri-
bofuranosyl)-5-nitroindole, 3. Step 1. Carbon tetrachloride (433
µL, 4.49 mmol) was added to a stirring solution of 2 (872 mg;
2.87 mmol) in anhydrous THF (10 mL). The solution was cooled
to -70 °C and hexamethylphosphorous triamide (700 µL, 3.82
mmol) was added dropwise for 5 min, and the mixture was
allowed to stir at the same temperature for 5 min. The resulting
chewy compound was allowed to warm to room temperature
and further stirred for 2 h. The solvent was evaporated and the
residue dissolved in diethyl ether. The resulting insoluble
material was removed by decantation and the organic layer was
evaporated to give the crude compound 3 as an orange oil.
Step 2. To a cooled solution (0 °C) of 5-nitroindole (906 mg,
5.59 mmol) dissolved in anhydrous acetonitrile (9 mL) was
added NaH (60%, 223 mg, 5.61 mmol); the solution was stirred
at room temperature for 30 min. To the resulting bright yellow
solution, the chlorosugar obtained from the first step was
dissolved in anhydrous acetonitrile and transferred via canula.
The reaction mixture was then warmed to room temperature
and stirred for 40 h. The mixture was poured into H2O (50 mL)
and extracted with CH2Cl2 (3 × 50 mL). The organic layers
were combined and successively washed with saturated aqueous
NH4Cl and saturated aqueous NaCl. The organic layer was dried
over MgSO4 and concentrated in vacuo. The residue was purified
by flash column chromatography (90/10: cyclohexane/EtOAc)
to give 3 (502 mg, 39%) as a yellow oil (anomeric mixture).
1H NMR (CDCl3): δ 8.57 (1H, d, J ) 2.14 Hz, H4); 8.11 (1H,
dd, J ) 2.14 Hz, J ) 9.12 Hz, H6); 7.60 (1H, d, J ) 3.43 Hz,
H3); 7.30 (1H, d, J ) 9.12 Hz, H9); 6.71 (1H, d, J ) 3.43 Hz,
H8); 6.48 (1HR, d, J ) 4.11 Hz, H1′R); 6.03 (1Hꢀ, d, J ) 3.38
Hz, H1′ꢀ); 4.96 (1H, d, J ) 6.06 Hz, H2′); 4.88 (1H, dd, J )
4.15 Hz, J ) 6.06 Hz, H3′); 4.45 (1H, m, H4′); 3.95 (1H, dd, J
) 2.14 Hz, J ) 11.05 Hz, H5′); 3.83 (1H, dd, J ) 2.14 Hz, J
) 11.05 Hz, H5′′); 1.43 (3H, s, CH3-C); 1.30 (3H, s, CH3-C);
1.00 (9H, s, (CH3)3-C); 0.15 (6H, s, Si(CH3)2). 13C NMR
(CDCl3) δ: 138.34 (C2); 129.81 (C5); 128.14 (C7); 118.08 (C9);
117.30 (C6); 113.29 (C4); 109.24 (C3); 104.16 (C-(CH3)2);
100.57 (C8); 88.33 (C1′); 82.76 (C2′); 82.74 (C4′); 80.53 (C3′);
66.06 (C5′); 29.69 (CH3-C); 25.89 (CH3-C); 25.71 ((CH3)3-
C); 18.13 ((CH3)3-C); -5.46 (Si(CH3)2). MS (GT, FAB+): 449
[M+H]+. Rf (C6H12/EtOAc: 90/10): 0.54.
1-(2,3-O-Isopropylidene-D-ribofuranosyl)-5-nitroindole, 4a
and 4b. To a solution of 3 (596 mg, 1.49 mmol) in dry THF (5
mL) was added TBAF (1 M in THF) (3 mL, 2.98 mmol) at
room temperature, and the mixture was stirred for 30 min. The
solution was evaporated under reduced pressure, diluted with
CH2Cl2, and washed with brine. The organic layer was dried
on MgSO4 and evaporated under reduced pressure. The crude
product was purified by flash column chromatography (cyclo-
hexane/EtOAc: 70/30) to give the anomer R 4a (198 mg, 54%)
as a yellow oil and the anomer ꢀ 4b (43 mg, 11%) as a yellow
oil.
1-(2,3-O-Isopropylidene-r-D-ribofuranosyl)-5-nitroin-
1
dole, 4a. H NMR (CDCl3): δ 8.57 (1H, d, J ) 2.16 Hz, H4);
8.10 (1H, dd, J ) 2.16 Hz, J ) 9.10 Hz, H6); 7.60 (1H, d, J )
3.43 Hz, H3); 7.41 (1H, d, J ) 9.10 Hz, H9); 6.72 (1H, d, J )
3.43 Hz, H8); 6.45 (1H, d, J ) 3.93 Hz, H1′); 4.97 (2H, m, H3′
+ H2′); 4.50 (1H, t, J ) 2.75 Hz, H4′); 3.99 (1H, dd, J ) 2.75
Hz, J ) 11.33 Hz, H5′); 3.88 (1H, dd, J ) 2.75 Hz, J ) 11.33
Hz, H5′′); 1.66 (1H, sl, OH); 1.40 (3H, s, CH3-C); 1.30 (3H, s,
CH3-C). 13C NMR (CDCl3) δ: 141.79 (C2); 138.39 (C5); 129.78
(C7); 128.08 (C9); 118.04 (C6); 117.23 (C4); 113.59 (C3); 109.59
((CH3)2-C); 104.27 (C8); 87.92 (C1′); 82.75 (C2′); 82.00 (C4′);
80.45 (C3′); 64.57 (C5′); 25.67 (CH3-C); 24.13 (CH3-C). MS
(GT, FAB+): 335 [M+H]+. HRMS (FAB+): calculated for
[M+H]+ 335.1237, found 335.1243. Rf (C6H12/EtOAc: 50/50):
0.22.
1-(2,3-O-Isopropylidene-ꢀ-D-ribofuranosyl)-5-nitroin-
1
dole, 4b. H NMR (CDCl3): δ 8.53 (1H, d, J ) 2.13 Hz, H4);
8.10 (1H, dd, J ) 2.13 Hz, J ) 9.10 Hz, H6); 7.52 (1H, d, J )
9.10 Hz, H3); 7.47 (1H, d, J ) 3.36 Hz, H9); 6.70 (1H, d, J )
3.36 Hz, H8); 6.05 (1H, d, J ) 3.13 Hz, H1′); 4.94 (2H, m, H3′
+ H2′); 4.31 (1H, q, J ) 3.29 Hz, H4′); 3.92 (1H, dd, J ) 3.29
Hz, J ) 12.00 Hz, H5′); 3.81 (1H, dd, J ) 3.29 Hz, J ) 12.00
Hz, H5′′); 2.32 (1H, sl, OH); 1.64 (3H, s, CH3-C); 1.37 (3H, s,
CH3-C). 13C NMR (CDCl3) δ: 142.09 (C2); 138.32 (C5); 128.43
(C7); 127.76 (C9); 118.05 (C6); 117.71 (C4); 114.96 (C3); 110.11
((CH3)2-C); 105.42 (C8); 91.76 (C1′); 85.03 (C2′); 84.42 (C4′);
80.29 (C3′); 62.29 (C5′); 27.23 (CH3-C); 25.22 (CH3-C). SM

1064 Bioconjugate Chem., Vol. 21, No. 6, 2010
Ceballos et al.
(GT, FAB+): 335 [M+H]+. HRMS (FAB+): calculated for
[M+H]+ 335.1237, found 335.1244; Rf (C6H12/EtOAc: 50/50):
0.38.
Hz, H8); 5.62 (1H, t, J ) 5.69 Hz, H2′); 5.46 (1H, m, H3′); 5.33
(4H, m, 2 × CHdCH); 4.56 (1H, q, J ) 2.84 Hz, H4′); 4.30
(2H, dd, J ) 2.84 Hz, J ) 11.21 Hz, H5′ + H5′′); 2.43 (3H, s,
CH3-CAr); 2.34 (2H, t, J ) 7.66 Hz, CH2-CO); 2.25 (2H, t, J
) 7.33 Hz, CH2-CO); 2.00 (8H, m, 2 × CH2-CHdCH-CH2);
1.60 (4H, m, 2 × CH2-CH2-CO); 1.26 (40H, m, 2 × 10 CH2);
0.87 (6H, t, J ) 6.48 Hz, 2 × CH3). 13C NMR (CDCl3) δ:
172.15 (CO); 171.53 (CO); 145.50 (CAr); 142.12 (C2); 138.70
(C5); 132.06 (CAr); 130.01 (2 × CHdCH); 129.53 (2 × CHAr);
128.96 (C7); 128.49 (2 × CHAr); 127.97 (C9); 125.85 (C6);
125.19 (C4); 104.31 (C3); 100.78 (C8); 79.82 (C1′); 74.43 (C4′);
73.80 (C3′); 73.59 (C2′); 65.57 (C5′); 31.87 (2 × CH2-CO);
29.78-28.67 (2 × 9 CH2); 27.18 (2 × CH2-CHdCH-CH2);
24.39 (2 × CH2-CH2-CO); 22.65 (CH3-CAr); 21.66 (2 ×
CH3-CH2); 14.09 (2 × CH3). MS (GT, FAB+): 978 [M+H]+;
1000 [M+23]+. Rf (C6H12/EtOAc: 80/20): 0.39.
1-(2,3-O-Isopropylidene-5-O-tosyl-r-D-ribofuranosyl)-5-ni-
troindole, 5a. Compound 4a (100 mg, 0.30 mmol) was
dissolved in dry dichloromethane (4 mL) and the solution was
cooled to 0 °C. NEt3 (168 µL, 1.20 mmol) was added, as well
as tosyl chloride (171 mg, 0.90 mmol), in small portions. The
solution was then allowed to warm to room temperature and
stirred overnight. Dichloromethane was added and the solution
was successively washed with brine and aqueous NaHCO3 5%
w/v. The organic layer was dried on MgSO4, filtered, and
concentrated under reduced pressure. The residue was purified
by preparative layer chromatography (cyclohexane/EtOAc: 60/
40) to give 5a (137 mg, 94%) as a yellow oil. 1H NMR (CDCl3)
δ: 8.55 (1H, d, J ) 2.21 Hz, H4); 8.10 (1H, dd, J ) 2.21 Hz,
J ) 9.00 Hz, H6); 7.83 (2H, d, J ) 9.00 Hz, HAr); 7.54 (1H, d,
J ) 3.47 Hz, H3); 7.35 (3H, m, HAr + H9); 6.70 (1H, d, J )
3.47 Hz, H8); 6.26 (1H, d, J ) 2.37 Hz, H1′); 4.93 (2H, m, H2′
+ H3′); 4.55 (1H, t, J ) 2.37 Hz, H4′); 4.27 (2H, d, J ) 2.64
Hz, H5′ + H5′′); 2.44 (3H, s, CH3-CAr); 1.41 (3H, s, CH3-C);
1.25 (3H, s, CH3-C). 13C NMR (CDCl3) δ: 145.76 (CAr); 141.97
(C2); 138.41 (C5); 132.06 (CAr); 130.20 (2 × CHAr); 129.64 (C7);
128.06 (2 × CHAr); 127.86 (C9); 117.98(C6); 117.36(C4); 114.04
(C3); 109.54 ((CH3)2-C); 104.52 (C8); 87.84 (C1′); 81.67 (C2′);
80.09 (C3′); 79.81(C4′); 71.22 (C5′); 25.65 (CH3-C); 24.09
(CH3-C); 21.67 (CH3-CAr). MS (GT, FAB+): 489 [M+H]+;
506 [M+NH4]+. Rf (C6H12/EtOAc: 70/30): 0.22.
1-(2,3-Dioleyl-5-O-tosyl-ꢀ-D-ribofuranosyl)-5-nitroin-
dole, 6b. Compound 6b was obtained from 5b with the same
procedure described above. Purification by flash column chro-
matography (cyclohexane/EtOAc: 60/40) gave 6b as a yellow
1
oil in 73% yield. H NMR (CDCl3): δ 8.56 (1H, d, J ) 2.21
Hz, H4); 8.08 (1H, dd, J ) 2.21 Hz, J ) 9.11 Hz, H6); 7.81
(2H, d, J ) 8.37 Hz, HAr); 7.46 (2H, m, H3 + H9); 7.32 (2H, d,
J ) 8.37 Hz, HAr); 6.74 (1H, d, J ) 3.47 Hz, H1′); 6.13 (1H, d,
J ) 6.63 Hz, H8); 5.45 (1H, t, J ) 6.00 Hz, H2′); 5.34 (6H, m,
2 × CHdCH + H3′ + H4′); 4.34 (2H, m, H5′ + H5′′); 2.43 (3H,
s, CH3-CAr); 2.37 (2H, t, J ) 7.26 Hz, CH2-CO); 2.25 (2H,
t, J ) 7.26 Hz, CH2-CO); 1.99 (8H, m, 2 × CH2-CHd
CH-CH2); 1.63 (4H, m, 2 × CH2-CH2-CO); 1.26 (40H, m,
2 × 10 CH2); 0.87 (6H, t, J ) 6.27 Hz, 2 × CH3). 13C NMR
(CDCl3) δ: 172.44 (CO); 171.86 (CO); 145.50 (CAr); 142.34
(C2); 138.91 (C5); 132.04 (CAr); 130.07 (2 × CHdCH); 130.08
(2 × CHAr); 129.53 (C7); 127.99 (2 × CHAr); 127.25 (C9);
118.12 (C6); 117.82 (C4); 109.62 (C3); 99.69 (C8); 86.90 (C1′);
73.63 (C4′); 73.02 (C3′); 70.46 (C2′); 64.07 (C5′); 31.87 (2 ×
CH2-CO); 29.72-28.89 (2 × 9 CH2); 27.19 (CH2-CHd
CH-CH2); 27.12 (CH2-CHdCH-CH2); 24.76 (CH2-CH2-
CO); 24.62 (CH2-CH2-CO); 22.65 (CH3-CAr); 21.65 (2 ×
CH3-CH2); 14.09 (2 × CH3). MS (GT, FAB+): 977 [M+H]+;
999 [M+23]+. Rf (C6H12/EtOAc: 60/40): 0.45.
1-(2,3-O-Isopropylidene-5-O-tosyl-ꢀ-D-ribofuranosyl)-3-ni-
troindole, 5b. Compound 5b was obtained in 53% yield from
compound 4b by the procedure described above. Purification
by preparative layer chromatography (cyclohexane/EtOAc: 60/
1
40) gave a yellow oil. H NMR (CDCl3): δ 8.56 (1H, d, J )
2.13 Hz, H4); 8.05 (1H, dd, J ) 2.13 Hz, J ) 9.10 Hz, H6);
7.52 (2H, d, J ) 7.96 Hz, HAr); 7.44 (1H, d, J ) 9.10 Hz, H3);
7.37 (1H, d, J ) 3.45 Hz, H9); 7.27 (2H, d, J ) 7.96 Hz, HAr);
6.71 (1H, d, J ) 3.45 Hz, H8); 6.02 (1H, d, J ) 2.84 Hz, H1′);
4.91 (2H, m, H3′ + H2′); 4.41 (1H, q, J ) 2.84 Hz, H4′); 4.30
(1H, dd, J ) 2.84 Hz, J ) 10.90 Hz, H5′); 4.19 (1H, dd, J )
2.84 Hz, J ) 10.90 Hz, H5′′); 2.41 (3H, s, CH3-CAr); 1.63 (3H,
s, CH3-C); 1.36 (3H, s, CH3-C). 13C NMR (CDCl3) δ: 154.49
(CAr); 142.22 C2); 138.24(C5); 131.97 (CAr); 129.98 (2 × CHAr);
128.51 (C7); 127.89 (2× CHAr); 127.72 (C9); 118.04 (C6); 117.74
(C4); 115.30 (C3); 110.15 ((CH3)2-C); 105.54 (C8); 91.99 (C1′);
84.08 (C2′); 81.96 (C3′); 80.12 (C4′); 68.42 (C5′); 27.17 (CH3-C);
25.19 (CH3-C); 21.62 (CH3-CAr). MS (GT, FAB+): 489
[M+H]+, 511 [M+Na]+. Rf (C6H12/EtOAc: 70/30): 0.60.
Tosylate Salt of 1-(2,3-Dioleyl-5-trimethylammonium-r-
D-ribofuranosyl)-5-nitroindole, 7a. Anhydrous triethylamine
(1 mL) was transferred to a pressure tube cooled at -50 °C via
canula. Next, anhydrous acetonitrile (2 mL) and a solution of
6a (51 mg, 0.05 mmol) in dry THF (1 mL) were added. The
tube was sealed and heated in an oil bath at 50 °C during 48 h,
then cooled to -20 °C and opened. The solvents were
evaporated under reduced pressure to give 7a (54 mg, quantita-
1-(2,3-dioleyl-5-O-tosyl-r-D-ribofuranosyl)-5-nitroin-
dole, 6a. At 0 °C, compound 5a (40 mg, 0.08 mmol) was
dissolved in 1 mL of an aqueous trifluoroacetic acid/H2O (6/1)
solution and stirred for 1 h at 0 °C. The solvents were
coevaporated with methanol under reduced pressure, and the
residue was dried in vacuo. The crude product was dissolved
in dry CH2Cl2 (4 mL) and the solution was cooled to 0 °C.
Dicyclohexylcarbodiimide (68 mg, 0.33 mmol), (dimethylami-
no)pyridine (40 mg, 0.33 mmol), and oleic acid (93 mg, 0.33
mmol) were added, and the mixture was stirred at room
temperature overnight. DCU was then filtered, and the filtrate
was successively washed with aqueous 2 N HCl and aqueous
saturated NaHCO3 solutions. The organic layer was dried on
MgSO4, filtered, and evaporated under reduced pressure. The
crude product purified by flash column chromatography (cy-
clohexane/EtOAc: 70/30) gave 6a (51 mg, 65%) as a yellow
1
tive) as a yellow oil. H NMR (CDCl3): δ 8.48 (1H, d, J )
2.05 Hz, H4); 7.88 (1H, dd, J ) 2.05 Hz, J ) 9.07 Hz, H6);
7.74 (2H, d, J ) 8.05 Hz, HAr); 7.49 (1H, d, J ) 2.05 Hz, H3);
7.35 (1H, d, J ) 6.48 Hz, H9); 7.11 (2H, d, J ) 8.05 Hz, HAr);
6.66 (1H, d, J ) 3.50 Hz, H1′); 6.21 (1H, d, J ) 6.48 Hz, H8);
5.33 (4H, m, 2 × CHdCH); 4.70 (2H, m, H2′ + H3′); 4.28 (1H,
m, H4′); 3.41 (9H, s, NCH3+); 3.33 (2H, m, H5′ + H5′′); 2.31
(3H, s, CH3-CAr); 2.24 (4H, t, J ) 7.26 Hz, 2 × CH2-CO);
1.98 (8H, m, 2 × CH2-CHdCH-CH2); 1.53 (4H, m, 2 ×
CH2-CH2-CO); 1.25 (40H, m, 2 × 10 CH2); 0.87 (6H, t, J )
6.65 Hz, 2 × CH3). 13C NMR (CDCl3) δ: 172.65 (CO); 171.82
(CO); 145.50 (C2); 143.39 (CAr); 142.29 (C5); 139.63 (CAr);
130.03 (2 × CHdCH); 130.00 (2 × CHAr); 129.49 (C7); 128.76
(2 × CHAr); 127.16 (C9); 125.91 (C6); 125.73 (C4); 101.20 (C3);
96.05 (C8); 83.17 (C1′); 71.98 (C3′); 66.55 (C4′); 59.44 (C2′);
66.93 (C5′); 54.34 (NCH3+); 31.84 (2 × CH2-CO); 30.23-28.90
(2 × 10 CH2); 27.07 (2 × CH2-CHdCH-CH2); 22.62 (2 ×
CH2-CH2-CO); 21.22 (CH3-CAr); 21.22 (2 × CH3-CH2);
1
oil. H NMR (CDCl3): δ 8.55 (1H, d, J ) 2.21 Hz, H4); 8.11
(1H, dd, J ) 2.21 Hz and J ) 9.16 Hz, H6); 7.84 (2H, d, J )
8.27 Hz, HAr); 7.46 (1H, d, J ) 2.21 Hz, H3); 7.43 (3H, m, HAr
+ H9); 6.68 (1H, d, J ) 2.84 Hz, H1′); 6.33 (1H, d, J ) 5.37

Cationic Nucleoside Lipids from Universal Bases
Bioconjugate Chem., Vol. 21, No. 6, 2010 1065
14.09 (2 × CH3). MS (GT, FAB+): 864 [M-TsO-]+. Anal.
Calcd. C, 68.12; H, 9.00; N, 4.00; found C, 68.13; H, 8.97; N,
4.11.
7.99 (1H, d, J ) 1.13 Hz, H5); 7.93 (1H, m, H2); 6.62 (1H, d,
J ) 2.78 Hz, H1′); 5.15 (1H, t, J ) 5.93 Hz, H3′); 4.87 (1H, dd,
J ) 2.78 Hz, J ) 5.93 Hz, H2′); 4.63 (1H, t, J ) 2.78 Hz, H4′);
3.95 (1H, dd, J ) 2.78 Hz, J ) 11.59 Hz, H5′); 3.81 (1H, dd,
J ) 2.78 Hz, J ) 11.59 Hz, H5′′); 2.41 (1H, sl, OH); 1.27 (3H,
s, CH3-C); 0.98 (3H, s, CH3-C). 13C NMR (CDCl3) δ: 139.31
(C4); 135.00 (C2); 118.72 (C5); 113.96 ((CH3)2-C); 89.30 (C1′);
83.73 (C2′); 82.13 (C4′); 80.35 (C3′); 64.39 (C5′); 25.50 (CH3-C);
24.15 (CH3-C). MS (GT, FAB+): 286 [M+H]+; 308 [M+Na]+.
Rf (EtOAc): 0.43.
1-(2,3-O-Isopropylidene-ꢀ-D-ribofuranosyl)-4-nitroimida-
zole, 9b. Compound 8b (166 mg, 0.32 mmol) was converted
to 9b according to the procedure described for 4b and purified
by flash column chromatography (cyclohexane/EtOAc: 70/30)
to give 9b (58 mg, 64%) as a white solid. mp ) 142-144 °C.
1H NMR (CDCl3) δ: 7.91 (1H, d, J ) 1.37 Hz, H5); 7.61 (1H,
d, J ) 1.37 Hz, H2); 6.11 (1H, d, J ) 2.51 Hz, H1′); 4.97 (1H,
dd, J ) 2.51 Hz, J ) 4.48 Hz, H3′); 4.90 (1H, dd, J ) 1.58 Hz,
J ) 4.48 Hz, H2′); 4.49 (1H, t, J ) 2.51 Hz, H4′); 3.98 (1H,
dd, J ) 2.51 Hz, J ) 11.47 Hz, H5′); 3.85 (1H, dd, J ) 2.51
Hz, J ) 11.47 Hz, H5′′); 1.38 (3H, s, CH3-C); 1.33 (3H, s,
CH3-C). 13C NMR (CDCl3) δ: 138.32 (C4); 137.85 (C2); 132.85
(C5); 113.49 ((CH3)2-C); 90.93 (C1′); 84.28 (C2′); 81.73 (C4′);
81.12 (C3′); 63.74 (C5′); 25.03 (CH3-C); 24.40 (CH3-C). MS
(GT, FAB+): 286 [M+H]+; 308 [M+Na]+. Rf (EtOAc): 0.72.
Tosylate Salt of 1-(2,3-Dioleyl-5-trimethylammonium-ꢀ-
D-ribofuranosyl)-5-nitroindole, 7b. This compound was pre-
pared from 6b with the same procedure described above to yield
7b as a brown oil with quantitative yield. 1H NMR (CDCl3): δ
8.48 (1H, d, J ) 2.21 Hz, H6); 8.05 (1H, dd, J ) 2.21 Hz, J )
9.16 Hz, H4); 7.73 (3H, m, H3 + HAr); 7.47 (1H, d, J ) 6.16
Hz, H9); 7.12 (2H, d, J ) 7.90 Hz, HAr); 6.57 (1H, d, J ) 3.43
Hz, H1′); 6.21 (1H, d, J ) 6.16 Hz, H8); 5.32 (4H, m, 2 ×
CHdCH); 4.74 (1H, m, H3′); 4.64 (1H, m, H2′); 4.22 (1H, m,
H4′); 3.34 (9H, s, N(CH3)3+); 3.27 (2H, m, H5′ + H5′′); 2.32
(3H, s, CH3-CAr); 2.26 (4H, t, J ) 7.42 Hz, 2 × CH2-CO);
1.98 (8H, m, 2 × CH2-CHdCH-CH2); 1.54 (4H, m, 2 ×
CH2-CH2-CO); 1.24 (40H, m, 2 × 10 CH2); 0.86 (6H, t, J )
6.32 Hz, 2 × CH3). 13C NMR (CDCl3) δ: 172.15 (CO); 171.86
(CO); 145.50 (C2); 142.12 (2 × CAr); 138.70 (C5); 132.06 (2 ×
CAr); 130.01 (2 × CHdCH); 129.53 (2 × CHAr); 128.96 (C7);
128.49 (2 × CHAr); 127.97 (C9); 125.85 (C6); 125.19 (C4);
104.31 (C3); 100.78 (C8); 79.82 (C1′); 74.43 (C3′); 73.80 (C4′);
73.59 (C2′); 65.57 (C5′); 55.45 (N(CH3)3+); 31.87 (2 ×
CH2-CO); 29.78-28.67 (2 × 9 CH2); 27.18 (2 × CH2-
CHdCH-CH2); 22.65 (2 × CH2-CH2-CO); 21.66 (CH3-
CAr); 21.65 (2 × CH3-CH2); 14.09 (2 × CH3). MS (GT, FAB+):
864 [M-TsO-]+. HMRS (FAB+): calculated for [M-TsO-]+
864.6460; found 864.6443. Anal. Calcd. C, 68.13; H, 9.01; N,
3.80; found C, 68.13; H, 8.97; N, 4.11.
1-(5-O-tert-Butyldiphenylsilyl-2,3-O-isopropylidene-D-ribo-
furanosyl)-4-nitroimidazole, 8. Compound 2′ (1.22 g, 2.85
mmol) was converted to 8 according to the procedure of
glycosylation described for 3 and purified by flash column
chromatography (cyclohexane/EtOAc: 90/10f80/20) to give 8a
(324 mg, 22%) and 8b (170 mg, 11%).
1-(5-O-tert-Butyldiphenylsilyl-2,3-O-isopropylidene-r-D-ri-
bofuranosyl)-4-nitroimidazole, 8a. 1H NMR (CDCl3) δ: 8.00
(1H, d, J ) 1.13 Hz, H5); 7.96 (1H, m, H2); 7.66 (4H, m, HAr);
7.43 (6H, m, HAr); 6.69 (1H, d, J ) 4.99 Hz, H1′); 5.22 (1H, t,
J ) 5.53 Hz, H2′); 4.78 (1H, d, J ) 5.53 Hz, H3′); 4.56 (1H, m,
H4′); 3.95 (1H, dd, J ) 2.10 Hz, J ) 11.39 Hz, H5′); 3.73 (1H,
dd, J ) 2.10 Hz, J ) 11.39 Hz, H5′′); 1.25 (3H, s, CH3-C);
1.12 (9H, s, (CH3)3-C); 1.01 (3H, s, CH3-C). 13C NMR
(CDCl3): δ 138.55 (C4); 135.52 (C2); 133.06 (2 × CAr); 132.31
(2 × CHAr); 132.04 (2 × CHAr); 130.17 (2 × CHAr); 130.09 (4
× CHAr); 127.99 (C5); 113.19 ((CH3)2-C); 91.12 (C1′); 84.11
(C2′); 82.03 (C4′); 81.11 (C3′); 65.60 (C5′); 26.81 (CH3-C); 25.15
(CH3-C); 24.37 ((CH3)3-C); 19.04 ((CH3)3-C). MS (GT,
FAB+): 524 [M+H]+. Rf (C6H12/EtOAc: 90/10): 0.23.
1-(2,3-O-Isopropylidene-5-O-tosyl-r-D-ribofuranosyl)-4-ni-
troimidazole, 10a. Compound 9a (80 mg, 0.28 mmol) was
converted to 10a according to the procedure described for 5a
and purified by flash column chromatography (cyclohexane/
1
EtOAc: 70/30) to give 10a (50 mg, 46%) as an orange oil. H
NMR (CDCl3) δ: 7.86 (1H, d, J ) 1.46 Hz, H5); 7.80 (2H, d,
J ) 7.90 Hz, HAr); 7.56 (1H, d, J ) 1.46 Hz, H2); 7.40 (2H, d,
J ) 7.90 Hz, HAr); 5.93 (1H, d, J ) 4.11 Hz, H1′); 4.89 (2H, m,
H4′ + H2′); 4.56 (1H, t, J ) 1.74 Hz, H3′); 4.22 (2H, d, J )
2.37 Hz, H5′ + H5′′); 2.47 (3H, s, CH3-CAr); 1.36 (3H, s,
CH3-C); 1.30 (3H, s, CH3-C). 13C NMR (CDCl3) δ: 145.93
(C4); 138.81 (C2); 134.85 (CAr); 131.83 (CAr); 130.25 (2 ×
CHAr); 127.87 (2 × CHAr); 118.43 (C5); 114.59 ((CH3)2-C);
89.12 (C1′); 81.66 (C2′); 81.07 (C3′); 79.96 (C4′); 70.90 (C5′);
25.45 (CH3-C); 24.15 (CH3-C); 21.69 (CH3-CAr). MS (GT,
FAB+): 440 [M+H]+. Rf (C6H12/EtOAc: 70/30): 0.36.
1-(2,3-O-isopropylidene-5-O-tosyl-ꢀ-D-ribofuranosyl)-4-ni-
troimidazole, 10b. Compound 9b (140 mg, 0.49 mmol) was
converted to 10b according to the procedure described for 5b
and purified by preparative layer chromatography (cyclohexane/
EtOAc: 30/70) to give 10a (90 mg, 42%) as a pale yellow oil.
1H NMR (CDCl3) δ: 7.98 (1H, d, J ) 1.43 Hz, H5); 7.82 (3H,
d + m, J ) 8.53 Hz, HAr + H2); 7.38 (2H, d, J ) 8.53 Hz,
HAr); 6.34 (1H, d, J ) 2.84 Hz, H1′); 5.05 (1H, t, J ) 4.84
Hz, H4′); 4.77 (1H, d, J ) 2.84 Hz, H2′); 4.70 (1H, t, J ) 2.84
Hz, H3′); 4.26 (2H, m, H5′ + H5′′); 2.46 (3H, s, CH3-CAr); 1.23
(3H, s, CH3-C); 0.96 (3H, s, CH3-C). 13C NMR (CDCl3) δ:
145.77 (C4); 134.02 (C2); 132.37 (CAr); 131.31 (CAr); 130.19
(2 × CHAr); 127.84 (2 × CHAr); 121.95 (C5); 114.07
((CH3)2-C); 81.35 (C1′); 81.22 (C2′); 80.73 (C3′); 70.75 (C4′);
69.87 (C5′); 22.98 (CH3-C); 24.36 (CH3-C); 21.70 (CH3-CAr).
MS (GT, FAB+): 440 [M+H]+. Rf (C6H12/EtOAc: 90/10): 0.74.
1-(5-O-tert-Butyldiphenylsilyl-2,3-O-isopropylidene-ꢀ-D-ri-
bofuranosyl)-4-nitroimidazole, 8b. 1H NMR (CDCl3) δ: 7.85
(1H, d, J ) 1.44 Hz, H5); 7.64 (4H, m, HAr); 7.52 (1H, d, J )
1.46 Hz, H2); 7.44 (6H, m, HAr); 6.10 (1H, d, J ) 2.91 Hz,
H1′); 4.91 (2H, m, H2′ + H3′); 4.44 (1H, t, J ) 2.10 Hz, H4′);
3.99 (1H, dd, J ) 2.10 Hz, J ) 11.44 Hz, H5′); 3.79 (1H, dd,
J ) 2.10 Hz, J ) 11.44 Hz, H5′′); 1.37 (3H, s, CH3-C); 1.32
(3H, s, CH3-C); 1.11 (9H, s, (CH3)3-C). 13C NMR (CDCl3)
δ: 135.49 (C4); 135.33 (C2); 134.85 (2 × CAr); 132.11 (4 ×
CHAr); 130.34 (2 × CHAr); 130.26 (4 × CHAr); 128.07 (C5);
113.98 ((CH3)2-C); 89.41 (C1′); 83.87 (C2′); 82.32 (C4′); 80.41
(C3′); 66.13 (C5′); 26.89 (2 × CH3-C); 25.51 ((CH3)3-C); 19.09
((CH3)3-C). MS (GT, FAB+): 524 [M+H]+. mp ) 40 °C. Rf
(C6H12/EtOAc: 90/10): 0.18.
1-(2,3-dioleyl-5-O-tosyl-r-D-ribofuranosyl)-4-nitroimida-
zole, 11a. Compound 10a (80 mg, 0.28 mmol) was converted
to 11a according to the procedure described for 6a and purified
by flash column chromatography (cyclohexane/EtOAc: 70/30)
to give 11a (41 mg, 32%) as a yellow oil. 1H NMR (CDCl3): δ
7.88 (1H, d, J ) 1.46 Hz, H5); 7.82 (2H, d, J ) 8.06 Hz, HAr);
7.51 (1H, d, J ) 1.46 Hz, H2); 7.39 (2H, d, J ) 8.06 Hz, HAr);
6.09 (1H, d, J ) 5.49 Hz, H1′); 5.48 (2H, m, H2′ + H3′); 5.34
(4H, m, 2 × CH2-CHdCH-CH2); 4.58 (1H, m, H4′); 4.34 (1H,
dd, J ) 2.21 Hz, J ) 11.26 Hz, H5′); 4.21 (1H, dd, J ) 2.21
1-(2,3-O-Isopropylidene-r-D-ribofuranosyl)-4-nitroimida-
zole, 9a. Compound 8a (325 mg, 0.62 mmol) was converted to
9a according to the procedure described for 4a and purified by
flash column chromatography (cyclohexane/EtOAc: 90/
1
10fEtOAc) to give 9a (80 mg, 46%). H NMR (CDCl3) δ:

1066 Bioconjugate Chem., Vol. 21, No. 6, 2010
Ceballos et al.
Hz, J ) 11.26 Hz, H5′′); 2.47 (3H, s, CH3-CAr); 2.32 (4H, m,
2 × CH2-CO); 1.99 (8H, m, 2 × CH2-CHdCH-CH2); 1.56
(4H, m, 2 × CH2-CH2-CO); 1.26 (40H, m, 2 × 10 CH2);
0.87 (6H, t, J ) 6.64 Hz, 2 × CH3). 13C NMR (CDCl3) δ:
173.01 (CO); 172.59 (CO); 149.54 (CAr); 147.69 (C2); 143.42
(CAr); 140.11 (CAr); 133.65 (2 × CHdCH); 131.27 (2 × CHAr);
130.96 (2 × CHAr); 118.88 (C5); 87.71 (C1′); 78.45 (C4′); 76.23
(C3′); 71.36 (C2′); 65.67 (C5′); 32.36 (2 × CH2-CO); 30.78
(CH2-CHdCH-CH2); 30.66 (CH2-CHdCH-CH2); 29.88-
29.19 (2 × 9 CH2); 26.76 (CH2-CH2-CO); 25.84 (CH2-CH2-
CO); 23.05 (CH3-CAr); 21.52 (2 × CH3-CH2); 14.09 (2 ×
CH3). MS (GT, FAB+): 929 [M+H]+.
anomer nucleolipids formed small vesicles with diameters
between 40 and 60 nm, respectively.
siRNA Transfection. Trypsine adherent cells were diluted
in normal growth medium to 1 × 105 cells per milliliter. The
transfectant reagent was diluted in serum free medium and
incubated at room temperature for 10 min before use. Next,
siRNA (GAPDH siRNA or negative control siRNA, Ambion)
was diluted in serum free medium. The siRNA and the
tranfection reagent were mixed, incubated for 10 min at room
temperature, and dispensed into a culture plate. Depending
on the experimental design, the ratio of lipid, siRNA, and
incubation time was varied. Cell suspension was overlayed
onto the transfection mixture, then incubated at 37 °C and 5%
CO2. Assays for gene knockdown were assessed after 48 h
depending on the experimental protocol.
The gene knockdown assay performed was KDalert GAPDH
Assay (Ambion) following the manufacturer’s protocol. Briefly,
48 h after siRNA transfection, the culture medium was aspirated
from transfected cells; 200 µL KDalert Lysis Buffer was added
to each sample well and the mixture incubated at 4 °C for 20
min to lyse the cells. The cell lysate was pipeted up and down
4-5 times to homogenize the lysate. 10 µL aliquots of each
lysate or GAPDH enzyme dilution (including the GAPDH
working stock) were transfered to the wells of a clean 96 well
plate. 90 µL aliquots of KDalert master mix were added to each
sample using a multichannel pipettor to dispense the KDalert
master mix quickly. The increase in fluorescence of the samples
was measured at room temperature.
Cytotoxicity. Cytotoxicity was assessed using both a forma-
zan-based proliferation assay (CellTiter 96 AQueous One
Solution Cell Proliferation Assay kit, Promega) and a total
protein-based assay (Pierce). Briefly, CHO cells were seeded
onto a 24 well plate with an appropriate density of 1 × 106
cells per well. After 48 h, the MTS substrate was added to each
well, and the plate was incubated for 4 h at 37 °C in a
humidified, 5% CO2 incubator. The amount of soluble formazan
produced by cellular reduction of the substrate MTS was
recorded at 490 nm using a multiwell plate reader. For the total
protein based proliferation assay, cells were lysed at the same
time as transfection efficiency was assayed. 10 µL aliquots of
lysates were transferred to a separate multiwell plate. Total
protein contents were assessed using the Coomassie Blue protein
kit (Pierce, Rockford, IL) following the manufacturer protocol.
Negative and positive controls were nontreated cells and
commercial lipids treated cells, respectively. The proliferation
results were expressed as percentages of nontreated cells.
Molecular Modeling. In order to obtain some information
about the structure-activity of these universal base analogue
nucleolipids, modelization of the five tested compounds was
performed at the density functional theory (DFT) level (B3LYP)
with a 6-31G* split-valence basis set (8, 9). Cavity surface area
(CSA) and (polarized solute)-solvent (PSS) are determined
within the same methodology in a solvent, water, using a PCM
model (10). Briefly, cavity surface area (CSA) and (polarized
solute)-solvent (PSS) were calculated for all the compounds as
well as the quotient MC ) |PSS/CSA|.
Tosylate Salt of 1-(2,3-Dioleyl-5-trimethylammonium-r-
D-ribofuranosyl)-4-nitroimidazole, 12a. Compound 11a (80
mg, 0.28 mmol) was converted to 12a according to the
procedure described for 7a to give 11a as a yellow oil with
1
quantitative yield. H NMR (CDCl3): δ 7.95 (1H, d, J ) 1.46
Hz, H5); 7.68 (3H, d+m, J ) 7.90 Hz, HAr + H2); 7.15 (2H, d,
J ) 7.90 Hz, HAr); 5.76 (1H, d, J ) 5.50 Hz, H1′); 5.33 (4H, m,
2 × CH2-CHdCH-CH2); 4.91 (3H, m, H2′ + H3′ + H4′); 3.37
(9H, s, NCH3+); 3.30 (2H, m, H5′ + H5′′); 2.33 (3H, s,
CH3-CAr); 2.15 (4H, m, CH2-CO); 1.99 (8H, m, 2 × CH2-
CHdCH-CH2); 1.50 (4H, m, 2 × CH2-CH2-CO); 1.26 (40H,
m, 2 × 10 CH2); 0.87 (6H, t, J ) 6.16 Hz, 2 × CH3). MS (GT,
FAB+): 815 [M-TsO-]+. Anal. Calcd. 65.63; H, 9.11; N, 5.62;
Found C, 65.69; H, 9.19; N, 5.67.
Vesicle Preparation. 1 mg/mL of compounds 7a or 7b was
dissolved in chloroform. The solution was evaporated and dried
under vacuum for 2 h. 1 mL of DI water was added, and the
solution was hydrated overnight at 4 °C. The resulting solution
was sonicated for 20 min prior to any measurements.
Electrophoresis Studies. Electrophoresis studies were con-
ducted in 2% agarose gels containing ethidium bromide in a
0.5 Tris-Borate-EDTA buffer. Cationic lipoplexes were prepared
10 min before use. 250 ng of siRNA (21 bp, Eurogentec SR-
CL010-005) was mixed with lipids at different ratios (final
volume is 20 µL). Then, 20 µL of each lipoplexes (samples)
were mixed with 4 µL of loading buffer (glycerol 30%, v/v),
bromophenol blue 0.25% (w/v), and xylene cyanol 0.25 (w/v)
and subjected to agarose gel electrophoresis for 20 min at 100
V. The electrophoresis gel was visualized and digitally photo-
graphed using a G.BOX camera.
Exclusion Assay. Ten milliliters of EB (10 mg/mL) was
diluted in 580 mL in deionized water. Ten microliters of 90
mg/mL solution of siRNA was added. A varying amount of
lipid (depending on the siRNA/lipid ratio required) was added
to the EB solution finally. The solutions were mixed on a
benchtop vortex, and the fluorescence was measured (λexc
)
470 nm, λem ) 600 nm; 0.5 cm path length glass cuvette). The
fluorescence was expressed as the percentage of the maximum
fluorescence signal when EB was bound to the siRNA in the
absence of lipid.
ꢁ Potential and Size Distribution of 7b CNLs (with and
without siRNA). Experiments were realized with N/P ratio of
10 (CNLs/siRNA).
Transmission Electronic Microscopy (TEM). TEM experi-
ments were performed on vesicle samples 7a, 7b, and 12a using
a Philips CM 10 (negative staining with ammonium molybdate
1% water, Cu/Pd carbon coated grids). Both (R) and (ꢀ) anomer
nucleolipids formed small vesicles with diameters between 40
and 60 nm, respectively. Cationic lipoplexes were prepared as
described above and were visualized by negative staining
microscopy. Ten microliter aliquots of samples (either liposomes
or lipoplexes) were transferred to a carbon-coated grid for 10
min. Then, the samples were dried and stained with 2.5% (w/
w) of uranyl acetate for 30 s. The specimens were observed
with a Hitachi H 7650 electron microscope. Both (R) and (ꢀ)
RESULTS AND DISCUSSION
Synthesis. The (R) and (ꢀ) anomers of 2′,3′-dioleyl-5′-O-D-
ribofuranosyl-3-nitroindole, 7a and 7b, respectively, and the (R)
anomer 2′,3′-dioleyl-5′-O-D-ribofuranosyl-4-nitroimidazole, 12a,
were synthesized from D-ribose as shown in Scheme 1.
Accordingly, 5-O-tert-butyldimethylsilyl-2,3-O-isopropylidene-
D-ribose 2 and 5-O-tert-butyldimethylsilyl-2,3-O-isopropylidene-
D-ribose 2′ were prepared in two steps (see Supporting
Information) from D-ribose following a published procedure
(11). Next, the stereoselective chlorination of the protected

Cationic Nucleoside Lipids from Universal Bases
Bioconjugate Chem., Vol. 21, No. 6, 2010 1067
Scheme 1a
a Reaction conditions: (a) (i) acetone, CuSO4, 1 (94%), (ii) TBDMSCl or TBDPSCl, DMAP, TEA; 2 (35%) and 2′ (46%); (b) (i) CCl4, HMPT,
(ii) 5-nitroindole, NaH, CH3CN; 3 (40%); (c) (i) CCl4, HMPT, (ii) 4-nitroimidazole, NaH, CH3CN; 8a (22%) and 8b (11%); (d) TBAF dry THF;
4a/4b (5/1; 65%); 9a (46%), 9b (64%); (e) (i) TsCl, TEA, DCM; 5a (94%), 5b (56%) and 10a (42%), 10b (46%), (ii) TFA/H2O, quant., (iii) Oleic
acid, DCC/DMAP; 6a (65%), 6b (73%), and 11a (32%); (f) NMe3, CH3CN 60°C, 40 h, quantitative.
sugars 2 and 2′ with CCl4 and hexamethylphosphorus triamide
(HMPT) (12) was followed by coupling with the anions of
3-nitroindole pyrrole or 4-nitroimidazole. Deprotection of the
resulting derivative 3 and 8 afforded the anomeric mixture of
(R) 4a (53%), (ꢀ) 4b (18%), and (R) 9a (53%), (ꢀ) 9b (18%).
As described previously for nitropyrrole derivatives, the (R)
forms were found to be the major isolated polar isomers. The
relative stereochemistry of these compounds was assigned by
1H NMR when compared to the chemical shifts of the acetonide
methyl group as well as those of the H-1′ (13). The anomeric
pairs of compounds 4 and 9 were then separately tosylated with
TsCl/TEA, deprotected with TFA/H2O, and coupled with oleic
acid using DCC/DMAP to afford the nucleolipids 6a, 6b, 11a,
and the side product 11b. The latter was isolated and character-
ized by mass spectra. Unfortunately, we were unable to isolate
the ꢀ anomer of 11a. Indeed, the instability of the sulfonated
intermediates has been demonstrated for adenosine analogues
with intramolecular cyclization of the 5′-carbon and N3-position
of the adenine ring (14, 15). It is likely that such an intermo-
lecular cyclization occurs in the case of the nitroindole
intermediate 11a.
Figure 3. Electrophoretic mobility of 7b (ꢀ) and 7a (R). The lines
represent the following (1): DNA stepladder 25 pb (2-6), siRNA with
7b at different ratios N/P (2, 4, 6, 8, 10 respectively (7)), siRNA (8-12),
siRNA with 7a at different ratios N/P (2, 4, 6, 8, 10, respectively).
bp) as a model system (16). The siRNA gel electrophoresis was
performed on compounds 7a and 7b; as shown in Figure 3, a
full complexation occurs for a charge ratio 7b/siRNA (ꢀ anomer)
of 10/1 (Figure 3 column 6), whereas in the case of the 7a/
siRNA (R anomer), free siRNA is observed for all the ratios
investigated (Figure 3 columns 8-12).
Finally, compounds 6a,b, and 11a were reacted with trim-
ethylamine for two days in a sealed tube to afford the target
nucleolipids 7a, 7b, and 12a.
Complexation Studies. In order to determine whether the
CNLs 7a, 7b, and 12a bind nucleic acids, we performed (i)
siRNA (21 bp) gel electrophoresis and (ii) standard ethidium
bromide (EB) fluorescence displacement assays with siRNA (21
The binding affinity of siRNA by CNLs was also investigated
by the fluorescence titration of EB with a premixed solution of
siRNA and various concentrations of CNLs. In the presence of
siRNA, the fluorescence emission of EB is enhanced relative
to that in water as a result of EB intercalation between the RNA
base pairs. The subsequent addition of CNLs to the siRNA-EB
solution results in CNL binding to RNA and displacement of

1068 Bioconjugate Chem., Vol. 21, No. 6, 2010
Ceballos et al.
Figure 5. TEM image of CNL-siRNA systems obtained with 7b (bar
) 50 nm).
Figure 4. EB exclusion assays with A, B, 7a, 7b, and 12a in the
presence of siRNA (fluorescence intensity arbitrary units versus N/P
charge ratio).
EB from the helix to water, which induces a decrease in
fluorescence intensity. As shown in Figure 4, the fluorescence
intensity decreases rapidly for all the CNLs investigated,
including both 5-nitroindole derivatives 7b (ꢀ), 7a (R), the
4-nitroimidazole 12a, and nitropyrrole CNLs previously reported
(R,ꢀ anomers, compounds A and B, respectively). Interestingly,
with a total decrease in fluorescence for N/P ratios higher than
8, the ꢀ anomer 5-nitroindole derivative 7b exhibits the best
binding affinity of the CNL series. The R anomer 7a, which
forms 7a/siRNA complex at higher N/P ratios (higher than 10)
compared to ꢀ analogue 7b, shows a lower binding affinity
indicating that stereochemistry of the anomeric position has an
impact on the complexation with siRNA double helix. Likewise,
a higher binding affinity is observed for nitropyrrole CNL B
(ꢀ) compared to CNL A (R) confirming the importance of
anomeric configuration in the complex formation. As expected,
the nature of the CNL universal base has an impact on the
formation of CNLs/siRNA complexes. With no residual EB
fluorescence for ratio of 10/1, the 5-nitroindole derivatives 7b
(ꢀ) and 7a (R) feature the best binding affinity, whereas
4-nitroimidazole 12a is the least efficient of the CNLs. It is
well-known that intermolecular overlapping of p-orbitals in
π-conjugated systems causes π-π interactions. Hence, these
interactions become stronger as the number of π-electrons
increases. Accordingly, the best binding affinity observed for
the 5-nitroindole universal base can be explained by the 10 π
electrons, whereas the 6 π electrons of both nitroindole and
nitropyrrole induce fewer π-π interactions.
Electronic Microscopy Studies, ꢁ Potential, and Size
Distribution. TEM experiments were performed on vesicle
samples 7a, 7b, and 12a using a Hitachi H 7650 electron
microscope (negative staining with 2.5% (w/w) of uranyl
acetate). Both (R) and (ꢀ) anomer nucleolipids formed small
vesicles with diameters between 40 and 60 nm, respectively
(Supporting Information Figure S1). TEM experiments were also
performed on samples prepared with siRNA. For these experi-
ments, the CNL vesicles were incubated for 15 min in the
presence of siRNA (21 bp). TEM micrographs (Figure 5) of
the CNL-siRNA systems exhibited multilamellar vesicles (which
are not observed in the case of CNL vesicles; see Supporting
Information Figure S2) suggesting that the siRNA molecules
are entrapped within the vesicular structures. ꢁ-potential and
size distribution were determined for the CNL vesicles of 7b
with and without siRNA. As expected, the results show a
significative decrease of ꢁ-potential (from 74 to 36.7 mV), as
well as an increase of size distribution (from 108 to 360 nm) in
Figure 6. siRNA transfection results after 48 h in HepG2 as a function
of nucleolipid (A, B, 7a,7b, and 12a) and N/P mole ratio. n ) 3.
the presence of siRNA suggesting its complexation with the
CNL vesicles of 7b.
Transfection. The transfection of siRNA with A, B, 7a, 7b,
and 12a was next performed to evaluate the reduction in protein
synthesis. Experiments used glyceraldehyde-3-phosphate dehy-
drogenase (GAPDH) siRNA (Ambion) on human liver (HepG2)
cell line, and the complete details can be found in the Supporting
Information. siRNA transfection results were accessed after 48 h
as a function of both anomer and cation/anion (N/P) mole ratio.
In each experiment, controls using GAPDH siRNA and negative
control siRNA were used at 30 mM. siPORT NeoFX was used
as the positive control, and the previously evaluated 3-nitropy-
rrole nucleolipids A and B (2) (Figure 1) were tested again in
the experiment. As shown in Figure 6, siRNA knockdown was
observed for most nucleolipids, with values similar to the
positive control for compounds 7a and 7b. The transfection
activity of 7a was significant vs siRNA with N/P ratios of 5
and 10, whereas for 7b, it was significant with N/P ratios of 5,
10, 15, and 20. Interestingly, the 5-nitroindole CNL 7b (ꢀ),
which features the best binding affinity for siRNA, is also
globally the most efficient in term of transfection. In addition,
none of the CNLs showed a significant cytotoxicity at these
concentrations, similar to that to the negative control (Supporting
Information Figure S3).
Molecular Modeling. A basic molecular modeling method
described in the section above was used to calculate the MC
quotient (MC ) |PSS/CSA|). This quotient represents the
stabilization in kilojoules per mole per surface unit; it reflects
the system’s capacity to maintain its cohesion, i.e., the smaller
MC values reflect lower membrane cohesion. As shown in Table
1 with N/P ratios of >5, it can be seen that low MC values are
correlated both to better transfection activity (Table 1, line 4)

Cationic Nucleoside Lipids from Universal Bases
Bioconjugate Chem., Vol. 21, No. 6, 2010 1069
Table 1. Relationship among MC Quotient, % Knockdown, and
(4) Barthe´le´my, P. (2009) Nucleoside-based lipids at work: From
supramolecular assemblies to biological applications. C. R. Chim.
12, 171–179.
(5) Ceballos, C., Prata, C. H., Giorgio, S., Payet, D., Barthe´le´my,
P., Grinstaff, M. W., and Camplo, M. (2009) Cationic nucleoside
lipids possessing a 3-nitropyrrole universal base for siRNA
delivery. Bioconjugate Chem. 20, 193–196.
(6) Loakes, D. (2001) The applications of universal DNA base
analogues. Nucleic Acids Res. 29, 2437–2447.
(7) Loakes, D., Hill, F., Brown, D. M., and Salisbury, S. A. (1997)
Stability and structure of DNA oligonucleotides containing non-
specific base analogues. J. Mol. Biol. 270, 426–435.
Fluorescence Intensity for Compounds A, B, 7a-b, and 12a-ba
compounds
PSS (kJ.mol-1
CSA (Å2)
12a
A
B
7a
7b
)
309.57 287.79 270.78 279.81 269.07
389.02 396.10 390.84 438.56 432.64
MC ) |PSS/CSA|
0.80
0.73
0.69
15
20
0.64
0.62
(kJ.mol-1 Å-2
% knockdown
)
3
8
18
50
(N/P ratios of 15)
fluorescence intensity
(N/P ratios of 10, Figure 4)
70
30
5
0
a CSA: Cavity Surface Area. PSS: (Polarized Solute)-Solvent.
(8) Lee, C. T., Wang, W. T., and Parr, R. G. (1988) Development
of the the colle-salvettu correlation-energy formula into a
functional of the electron density. Phys. ReV. B 37, 785–789.
(9) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E.,
Robb, M. A., Cheeseman, J. R., Montgomery, Jr., J. A. ,Vreven,
T., Kudin, K. N., Burant, J. C., Millam, J. M., Iyengar, S. S.,
Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G.,
Rega, N.,Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M.,
Toyota, K., Fukuda,R., Hasegawa, J., Ishida, M., Nakajima, T.,
Honda, Y., Kitao, O., Nakai, H., Klene, M., Li, X., Knox, J. E.,
Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C., Jaramillo,
J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J.,
Cammi, R., Pomelli, C., Ochterski, J. W., Ayala, P. Y.,
Morokuma, K., Voth, G. A., Salvador, P., Dannenberg, J. J.,
Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C.,
Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K.,
Foresman, J. B., Ortiz, J. V., Cui, Q., Baboul, A. G., Clifford,
S., Cioslowski, J., Stefanov, B. B., Liu, G., Liashenko, A.,
Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T.,
Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe,
M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W.,
Gonzalez, C., Pople, J. A. (2004) Gaussian 03, reV C.02,
Gaussian, Inc., Wallingford, CT.
(10) Cossi, M., Scalmani, G., Rega, N., and Barone, V. (2002)
New developments in the polarizable continuum model for
quantum mechanical and classical calculations on molecules in
solution. J. Chem. Phys. 117, 43–54.
(11) Stewart, A. O., and Williams, R. M. (1985) C-glycosidation
of pyridyl thioglycosides. J. Am. Chem. Soc. 107, 4289–4296.
(12) Harusawa, S., Imazu, T., Takashima, S., Araki, L., Ohishi,
H., Kurihara, T., Sakamoto, Y., Yamamoto, Y., and Yamatodani,
A. J. (1999) Synthesis of 4(5)-[5-(aminomethyl)tetrahydrofuran-
2-yl- or 5-(aminomethyl)-2,5-dihydrofuran-2-yl]imidazoles by
efficient use of a PHSe group: Application to novel histamine
H3-ligands. J. Org. Chem. 64, 8608–8615.
and to binding affinity (Figure 4). These results confirm that to
a certain extent (N/P ratio >5) the lowest membrane cohesion
of these series of CNLs vesicles are beneficial for the release
and the activities of siRNA in the cytoplasm.
It is known that subtle change such as headgroup nature,
headgroup charges, counterion nature, and so forth can dramati-
cally affect the transfection properties (17, 18). However, for
the first time to our knowledge, the results highlight the impact
of both stereochemistry and base nature of these nucleoside
amphiphiles, i.e., ꢀ anomeric configuration, as well as highly
π conjugated bases, is likely to improve the transfection efficacy.
These findings could provide a rationale for the design of
universal base nucleoside amphiphiles for siRNA delivery.
CONCLUSION
In summary, we have synthesized and characterized several
new universal base CNLs for siRNA delivery. The results
obtained by gel electrophoresis and ethidium bromide exclusion
assays clearly indicate that (i) the nature of the universal base
(5-nitroindole and 4-nitroimidazole) and (ii) the stereochemistry
of the anomeric position (R, ꢀ) have an impact on the CNLs-
siRNA complex formation. We observed that the CNL featuring
the nitroindole as universal base exhibits the best binding affinity
for siRNA. We also demonstrated that higher binding affinities
for siRNA are observed for ꢀ anomers of both nitroindole and
nitropyrrole CNLs. Molecular modeling has shown that calcula-
tion of simple values (MC) reflecting the system’s capacity to
maintain its cohesion can be correlated to both binding affinities
and transfection efficiencies. On the basis of this rational
approach, work is in progress to synthesize and improve the
efficacy of new, original cationic nucleoside amphiphiles for
siRNA delivery.
(13) Imbach, J. L., Barascut, J. L., Kam, B. L., Rayner, B., Tamby,
C., and Tapiero, C. (1973) Synthesis of nucleosides. III. New
criterion determination of the anomer configuration of ribosidic
nucleosides. J. Heterocycl. Chem. 10 (6), 1069–70.
(14) Bakthavachalam, V., Lin, L. G., Cherian, X. M., and Czarnik,
A. W. (1987) Synthesis, stereochemistry, intramolecular cycliza-
tion, and rates of hydrolysis of adenosine 2′,3′-acetals. Carbo-
hydr. Res. 170, 124–135.
(15) Clark, V. M., Todd, A. R., and Zussman, J. (1951) Nucleotides.
Part VIII cycloNucleoside salts. A novel rearrangement of some
toluene-p-sulphonylnucleosides. J. Chem. Soc. 2952–2958.
(16) Geall, A. J., and Blasgbrough, I. S. (2000) Rapid and sensitive
ethidium bromide fluorescence quenching assay of polyamine
conjugate-DNA interactions for the analysis of lipoplex formation
in gene therapy. J. Pharm. Biomed. Anal. 22, 849–859.
(17) Karmali, P. P., and Chaudhuri, A. (2006) Cationic liposomes as
non-viral carriers of gene medicines: resolved issues, open questions,
and future promises. Med. Chem. Res. 27 (5), 696–722.
(18) Alberle, A. M., Bennet, M. J., Malone, R. W., and Nantz,
M. H. (1996) The counterion influence of cationic lipid-mediated
transfection of plasmid DNA. Biochim. Biophys. Acta 1299,
281–283.
ACKNOWLEDGMENT
The authors thank the CNRS and the “Ministe`re de l’enseignement
supe´rieur et de la recherche” for financial support.
Supporting Information Available: Synthesis of compounds
1, 2, and 2′; ꢁ potential and size distribution of 7b CNLs; TEM
images of 7a and 12a; cytotoxicities of CNLs on HepG2 cell;
EB exclusion assays with 7a and 7b. This material is available
free of charge via the Internet at http://pubs.acs.org.
LITERATURE CITED
(1) Chabaud, P., Camplo, M., Payet, D., Serin, G., Moreau, L.,
Barthe´le´my, P., and Grinstaff, M. (2006) Cationic nucleoside lipids
derived for gene delivery. Bioconjugate Chem. 17, 466–472.
(2) Gissot, A., Camplo, M., Grinstaff, M. W., and Barthe´le´my, P.
(2008) Nucleoside based amphiphiles: A Successful Marriage of
nucleic acids with lipids. Org. Biomol. Chem. ReV. 6, 1324–1333.
(3) Moreau, L., Camplo, M., Wathier, M., Taib, N., Laguerre, M.,
Bestel, I., Grinstaff, M. W., and Barthe´le´my, P. (2008) Real time
imaging of supramolecular assembly formation via programmed
nucleolipid recognition. J. Am. Chem. Soc. 130, 14454–14455.
BC100005K