Folate-equipped pegylated archaeal lipid derivatives: Synthesis and transfection properties

Article abstract of DOI:10.1002/chem.200800950

We have previously shown that synthetic archaeal lipid analogues are useful vectors for drug/gene delivery. We report herein the synthesis and gene transfer properties of a series of novel di- and tetraether-type archaeal derivatives with a poly(ethylene glycol) (PEG) chain and further equipped with a folic acid (FA) group. The synthetic strategy and the purification by dialysis ensured complete removal of free FA. The lipids were mixed with a conventional glycine betaine-based cationic lipid and the resulting formulations were tested in transfection assays after complexation with plasmid DNA. All four novel co-lipids afforded efficient in vitro gene transfection. Moreover, the FA-equipped derivatives permitted ligand/receptor-based targeted transfection; their activity was inhibited when free FA was added to the transfection medium. These novel archaeal derivatives equipped with FA-PEG moieties may thus be of great interest for targeted in vivo transfection.

Full text of DOI:10.1002/chem.200800950

DOI: 10.1002/chem.200800950
Folate-Equipped Pegylated Archaeal Lipid Derivatives: Synthesis and
Transfection Properties
CØline LainØ,[a] Emmanuel Mornet,[b] Loïc Lemi›gre,[a] Tristan Montier,[b]
Sandrine Cammas-Marion,[a] CØcile Neveu,[a] Nathalie Carmoy,[b] Pierre Lehn,[b] and
Thierry Benvegnu*[a]
Abstract: We have previously shown
that synthetic archaeal lipid analogues
are useful vectors for drug/gene deliv-
ery. We report herein the synthesis and
gene transfer properties of a series of
novel di- and tetraether-type archaeal
derivatives with a poly(ethylene glycol)
(PEG)chain and further equipped
with a folic acid (FA)group. The syn-
thetic strategy and the purification by
dialysis ensured complete removal of
free FA. The lipids were mixed with a
conventional glycine betaine-based cat-
ionic lipid and the resulting formula-
tions were tested in transfection assays
after complexation with plasmid DNA.
All four novel co-lipids afforded effi-
cient in vitro gene transfection. More-
over, the FA-equipped derivatives per-
mitted ligand/receptor-based targeted
transfection; their activity was inhibit-
ed when free FA was added to the
transfection medium. These novel arch-
aeal derivatives equipped with FA-
PEG moieties may thus be of great in-
terest for targeted in vivo transfection.
Keywords: archael lipids
·
cell
targeting · folic acid · gene delivery
Introduction
Attempts to optimise the lipoplex stability 1)by the incor-
poration of high amounts of cholesterol (to increase the lip-
oplex membrane rigidity)or 2)by coating the lipoplex sur-
face with hydrophilic poly(ethylene glycol)(PEG)(to form
sterically stabilised liposomes, called Stealth liposomes, with
reduced interactions with blood proteins)led to an in-
creased residency time in blood in vivo, and therefore, have
permitted some passive targeting. Unfortunately, such modi-
fications also decrease the electrostatic interactions between
negatively charged cell surface residues and the positively
charged lipoplexes, which is required for their internalisa-
tion by endocytosis.[3] It is thus expected that the next gener-
ation of lipoplexes should be capable of active cell targeting
through ligand–receptor interactions. Indeed, equipping the
lipoplexes with an appropriate ligand should have the fol-
lowing two advantages: 1)provide selective entry into the
cells that express the corresponding receptor and 2)as the
internalisation process no longer involves electrostatic inter-
actions, allow the use of neutral lipoplexes, thereby avoiding
unwanted interactions with anionic plasma or extracellular
proteins and reducing the toxicity of the lipoplexes. Among
the numerous ligands available (antibodies, proteins, pep-
tides, carbohydrates)folic acid (FA)is particularly attractive
because it is well known in specific drug/gene delivery.[4–8]
FA is a small molecule that is readily available and non-im-
munogenic. It binds to specific folate receptors (FRs), in
particular to those of the a subtype (FRa, Kd ꢀ10À10 m).[6]
Over the last several years, both recombinant viruses and
non-viral vectors have been developed for gene therapy ap-
plications. However, neither viral vectors nor current syn-
thetic vectors fulfil all the requirements of an ideal DNA
delivery system. Although current non-viral vectors are less
efficient than their viral counterparts (especially in vivo),
non-viral systems are nevertheless of particular interest be-
cause they offer a series of advantages (including the ab-
sence of safety and large-scale production issues). Among
the synthetic carriers, cationic lipids, which form complexes,
called lipoplexes, with the DNA to be transferred, are espe-
cially promising. It is however generally recognised that effi-
cient in vivo transfection will require the development of
vectors that provide lipoplexes with favourable properties
such as increased stability and cell-targeting capabilities.[1,2]
[a] Dr. C. LainØ, Dr. L. Lemi›gre, Dr. S. Cammas-Marion, C. Neveu,
Prof. T. Benvegnu
Ecole Nationale SupØrieure de Chimie de Rennes, UMR CNRS 6226
Equipe “Chimie Organique et SupramolØculaire”
Av. GØnØral Leclerc, 35700 Rennes (France)
Fax : (+33)022-323-8046
E-mail: Thierry.Benvegnu@ensc-rennes.fr
[b] E. Mornet, Dr. T. Montier, N. Carmoy, Prof. P. Lehn
Inserm, U613, Brest, UniversitØ de Bretagne Occidentale
Brest, 29200 (France)
8330
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340

FULL PAPER
Because FRs are usually overexpressed at the surface of
human cancer cells, folate-based delivery systems have been
mostly used in cancer chemotherapy and medical imaging.
However, as FRs are also expressed at the surface of other
cell types (such as activated macrophages and epithelial res-
piratory cells), the therapeutic scope of cell targeting by the
FA–FR interaction extends beyond cancer and includes
other human diseases, for example, inflammatory diseases
and cystic fibrosis.[9,10]
In previous studies we have synthesised and evaluated the
properties of cationic and neutral “archaeosomes” with a
view to developing gene/drug delivery systems with en-
hanced stability.[11,12] Archaeosomes are liposomes composed
of natural lipids found in Archaea or of archaea-derived
synthetic lipids that have the unique structural characteris-
tics of archaeal membrane lipids.[13–15] Lipid components of
Archaea are indeed strikingly different from those found in
other forms of life and they are considered to be specific
and useful markers of this evolutionary line. Apart from the
archaea domain, all forms of life usually have lipids that
contain ester linking groups. In contrast, archaeal lipids con-
tain ether linkages and saturated isoprenyl units (Scheme 1).
they are also good “helper” lipids for gene delivery applica-
tions (when compared with the commonly used dioleoyl-
phosphatidylethanolamine (DOPE)or cholestero.l)
[11,12,18]
We thus reasoned that it might be interesting to combine
these stabilising effects with other properties desirable for in
vivo transfection. Indeed, as our archaeal lipid analogues
permit relatively easy chemical functionalisation,[18] they can
be regarded as good anchoring points for PEG chains,
equipped or not with targeting ligands at their end, to pro-
vide lipoplexes with specific targeting or stealth properties,
respectively.
We thus describe herein the preparation of a series of
archaeal lipid analogues with a PEG chain and their further
functionalisation with a folate group. Two lipid structures
(di- and tetraether)and one length of PEG chain (10 ethyl-
ene oxide units)were envisaged. The length of the PEG
chain was chosen as a “not too short, not too long” compro-
mise, the idea being 1)to facilitate the FA–FR interaction
by avoiding steric hindrance problems due to the lipoplex
itself and 2)to keep the preparation of lipoplexes easy de-
spite the presence of the PEG arms. Size measurements by
using dynamic light scattering were performed on both the
archaeosome and lipoplex for-
mulations. By using HeLa
cells, which express high levels
of FRs, we studied the in vitro
transfection activity of various
formulations in which the new
amphiphiles synthesised herein
were combined with a conven-
tional cationic lipid derived
from glycine betaine[19] and
previously found to mediate
significant gene transfection.
In particular, the targeting
ability of the formulations was
Scheme 1. A)Natural diether and B)tetraether structures of the archaeal lipids.
evaluated by comparing the
The ether bonds are chemically and enzymatically more
stable than esters located in comparable positions within the
structures of lipids, whereas the branching methyl groups in
isoprenyl units help to keep the membrane in a fluid state.
Other typical lipids from the archaea domain (Scheme 1,
type B)are characterised by aliphatic phytanyl chains linked
through ether units to two sn-2,3-glycerol units. Note that
such lipids span the membrane from the inner to the outer
side, thus forming monolayered membranes with increased
physical stability. Archaeal membranes are therefore rela-
tively more stable than those of other forms of life, thereby
enabling these microorganisms to survive under extreme
conditions of pH, temperature, pressure and salt concentra-
tion. It is thus not surprising that natural archaeal lipids
have been used as innovative materials in liposome formula-
tions with a view to exploiting their ability to increase lipid
membrane stability.[16,17]
transfection activity of lipids with a folate group and lipids
devoid of a targeting residue and, most importantly, by com-
petitive inhibition assays with free FA.
Results and Discussion
Synthesis: Two types of lipids were envisaged in the present
study: di- and tetraether-type structures, both of which are
analogues of archaeal membrane lipids. Preparation of the
diether backbone was based on the functionalisation of the
readily available glycerol derivative 3 (Scheme 2).[18] Reac-
tion of hexadecyl triflate with primary alcohol 3 gave, in the
presence of proton sponge, benzyloxy diether 4 in a yield of
83%. Hydrogenolysis of the benzyloxy group provided an
efficient access to alcohol 5, which was easily converted into
its acid and amine counterparts. Amine 7 was obtained in an
overall yield of 45% following a three step procedure from
alcohol 5: 1)mesylation of 5 with mesyl chloride in the pres-
ence of triethylamine, 2)substitution of the corresponding
We have previously shown that synthetic archaeal lipids
could significantly increase the stability of drug delivery sys-
tems even under oral administration conditions and that
Chem. Eur. J. 2008, 14, 8330 – 8340
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8331

T. Benvegnu et al.
(Table 1, entry 2). The amount
of aqueous phase in the reac-
tion mixture was also found to
be an important parameter as
the yield dropped to 19% for a
2:1 ratio of THF/ NaHCO3
(aq), even if the quantity of
catalyst was increased to
0.6 equiv (Table 1, entry 3).
This last result underlines the
importance of favouring the
hydrate intermediate for its ox-
mesylate with sodium azide in DMF and 3)reduction of
azide 6 by PPh3 in THF/H2O. Acid 8 was obtained by oxida-
tion of primary alcohol 5 using a 2,2,6,6-tetramethylpiperi-
dine N-oxide (TEMPO)-benzoate-catalysed oxidation reac-
tion in aqueous and basic media (THF/ NaHCO3 (aq)). Cal-
cium hypochlorite was found to be the best oxidant and af-
forded carboxylic acid 8 in a satisfactory yield (70%;
Table 1, entry 1).
idation to the carboxylic acid.
Interestingly, reduction of azide 6 under usual hydrogena-
tion conditions (H2, Pd/C), but in chlorinated solvents such
as chloroform led to an unexpected product, which was un-
ambiguously characterised as chlorinated diether
9
(Scheme 3). Note that we have already reported such a reac-
tion with a tetraether-type azide.[18] This unexpected reaction
and its mechanism are still under investigation.
Note that a homogeneous medium had a dramatic effect
on the oxidation efficiency because the use of CH2Cl2 in-
stead of THF led to a lower yield of 45% after 48 h
We described the tetraether backbone (diol 10)in our
previous work.[18] In this work, we envisaged its functionali-
sation at one end by a PEG derivative, a step that first re-
quired desymmetrisation of starting diol 10 (Scheme 4). This
crucial step was carried out by an easy monobenzylation
with benzyl bromide (0.9 equiv)and potassium hydride
(1.0 equiv). After treatment, 40% of monobenzylated 11
was isolated in addition to recyclable starting diol 10 and
the corresponding dibenzylated derivative in yields of 35
and 25%, respectively. Alcohol 11 was then oxidised in a
one-pot two-step procedure under TEMPO catalysis condi-
tions with NaOCl and NaClO2 as the oxidising agents. Fine
Table 1. Oxidation conditions for the preparation of acid 8.
Entry 4-BzO-TEMPO
[equiv]
Solvent 5% NaHCO3/sol-
vent (v/v)
Time Yield
[h]
[%]
1
2
3
0.1
0.1
0.6
THF
CH2Cl2 1:1
THF 2:1
1:1
24
48
24
70
45
19
Scheme 2. Preparation of the amine and carboxylic acid diether lipid backbones.
Scheme 3. Chlorination of the azide 6.
8332
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340

Archaeal Lipid Derivatives
FULL PAPER
(HOAt). HOBt led to a very
low yield (5%; Table 2,
entry 1), whereas HOAt af-
forded a 50% yield of isolated
14 (Table 2, entry 2). Moving
from DCC/HOAt or HOBt to
a uronium salt (O-(benzotria-
zol-1-yl)1,1,3,3-tetramethyluro-
nium
tetrafluoroborate,
TBTU)increased the yield to
70% for the same reaction
time (24 h)(Table 2, entry 3.)
Stirring the reaction mixture
for four days at room tempera-
ture allowed
a good yield
(93%)to be obtained (Table 2,
entry 4).
Reduction of azide 14 was
envisaged by two different pro-
Scheme 4. Preparation of the dissymmetrical tetraether backbones.
cedures: 1)Pd/C-catalysed hy-
tuning of the pH during the reaction gave a clean oxidation
of 11 to carboxylic acid 12 in a yield of 73%.
drogenation in THF/MeOH gave H2N-PEG570-diether 15 in
90% yield, removal of the Pd/C catalyst requiring several
filtrations on a silica gel pad, and 2)reduction by PPh 3 in
THF/H2O gave 15 in a moderate yield of 72%, but with the
advantage of a straightforward purification step (Scheme 5).
Application of the above coupling conditions to acid 12
did not provide satisfactory yields. We could however
With acids 8 and 12 in hand, we introduced a 10-unit
PEG chain into both compounds. The commercially avail-
able dissymmetrical H2N-PEG570-N3 chain (13)was selected
due to its ease of transformation into the primary amine by
reduction. We first optimised the conditions for coupling
with compound 8 and then applied the best conditions to
acid 12 (Scheme 5, Table 2). The first attempts were carried
out with dicyclohexylcarbodiimide (DCC)and 1-hydroxy-
benzotriazole (HOBt)or 1-hydroxy-7-azabenzotriazole
A
heated at reflux in chloroform, which gave the coupling
product 16 (N3-PEG570-tetraether)in a good yield (90%;
Scheme 6). Hydrogenation (H2, Pd/C)in THF/MeOH per-
mitted both the reduction of the azide group to the corre-
sponding primary amine and the removal of the benzyl
group to give the H2N-PEG570-tetraether 17 in a yield of
51%.
Table 2. Coupling conditions for the synthesis of the pegylated diether
14.
Entry
Coupling system
Time
Yield [%]
The introduction of FA into the pegylated lipid amines
can be expected to be associated with two main problems:
1)the poor solubility of FA and folate conjugates requires
the reaction to be carried out in DMSO, which complicates
the purification steps, and 2)the presence of two carboxylic
acids on the glutamate moiety
1
2
3
4
HOBt (1.2 equiv), DCC (1.2 equiv)
HOAt (1.2 equiv), DCC (1.2 equiv)
TBTU (1.3 equiv), DIEA (1.2 equiv)
TBTU (1.3 equiv), DIEA (1.2 equiv)
24 h
24 h
24 h
4 d
5
50
70
93
cannot guarantee a good regio-
selectivity of the coupling reac-
tion at the a or g position. It
should however be emphasised
here that coupling through the
carboxylic acid groups does
not affect the pteroate moiety
of FA, which is crucial for the
specific binding of FA to the
FR.
For H2N-PEG570-diether 15,
we could promote the regiose-
lectivity in clear favour of 19g
(19a/19g, 8:92)[20] because the
g position is more accessible
Scheme 5. Preparation of the pegylated diether 15.
(Scheme 7). However, a slight
Chem. Eur. J. 2008, 14, 8330 – 8340
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8333

T. Benvegnu et al.
a yield of 31% and as a mix-
ture of a-/g-FA-PEG570-dieth-
er/H2N-PEG570-diether in
a
40:48:12 molar ratio (19b).[20]
Note that these latter condi-
tions led to a poor a/g regiose-
lectivity (a/g=45:55).
H2N-PEG570-tetraether
17
was functionalised in the pres-
ence of an excess of TBTU
and DIEA (both 5.0 equiv;
Scheme 8). Under these condi-
tions and after dialysis against
DMSO, H2N-PEG570-tetraether
17 was quasi-quantitatively
converted into FA-PEG570-tet-
raether 20 (a/g=52:48), which
was obtained in a yield of 56%
as a 47:44:9 mixture of the two
regioisomers 20a and 20g and
Scheme 6. Preparation of the pegylated tetraether 17.
H2N-PEG570-tetraether 17, re-
spectively.[20]
excess of TBTU (1.3 equiv)did not permit more than 25%
conversion. Purification of FA-PEG570-diether 19a by dialy-
sis against DMSO allowed complete removal of the free FA
molecules and furnished a mixture of a-/g-FA-PEG570-dieth-
er/H2N-PEG570-diether in a 2:23:75 molar ratio (19a).[20]
This low yield (4%)was mainly due to a loss of material
during the dialysis process, the 1000 D cell-membrane cutoff
was not quite adapted to the retention of all pegylated com-
pounds. Because the positive results obtained with 19a-
based formulations in in vitro transfection experiments (see
below)strongly invite us to perform, in the near future, ex-
tensive in vivo transfection studies that will require large
amounts of reagent, we decided to further investigate this
synthesis step to improve both the reaction yield and con-
version. Thus, it should be stressed here that the use of
three equivalents of TBTU/FA/DIEA (DIEA=diisopropy-
lethylamine)significantly improved the conversion to the
FA-PEG570-diether. Indeed, although the dialysis purifica-
tion step remained unchanged, the product was obtained in
In summary, we have designed and synthesised four origi-
nal pegylated di- (15 and 19)and tetraether-type ( 17 and
20)archaeal lipid analogues, two of which ( 19 and 20)were
further equipped with a FA group at the end of the PEG
chain. Note that the absence of free FA was confirmed for
all samples; this is indeed critical for the evaluation of their
cell-targeting capability. As already stated above, these four
compounds were then used, in combination with the con-
ventional glycine betaine-based cationic lipid 21[19] for gene
transfection experiments.
Liposome/lipoplex preparation
and size determination: First,
liposomes and archaeosomes
were prepared by hydrating
lipid films composed of bilay-
er-forming cationic lipid 21
alone (liposomes)or combined
with one bi- or monolayer-
forming co-lipid 7, 15, 17, 19a
or 20 (archaeosomes)with
water for 12 h at 48C followed
by sonication (2”5 min). The
sizes of the vesicles obtained
were determined by dynamic
light scattering. As shown in
Scheme 7. Synthesis of the targeting folate derivative 19.
Table 3, incorporation of
a
8334
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340

Archaeal Lipid Derivatives
FULL PAPER
Scheme 8. Synthesis of the targeting folate derivative 20.
Table 3. Sizes of the liposomes and archaeosomes.
DNA phosphate groups). The lipid–DNA complexes formed
were analysed after 30 min of incubation at room tempera-
ture. Size determinations led to the same conclusions as
those for the corresponding liposomes. However, positively-
charged lipoplexes (charge ratio >1)were generally larger
(by 30–300 nm)than the corresponding liposomes (data not
shown). Of note, the smallest size differences were observed
with H2N-PEG570-diether 15 and H2N-PEG570-tetraether 17,
whereas the largest lipoplexes were obtained with the tar-
geting co-lipids (19a and 20).
Co-lipid
21/co-lipid Average
Polydispersity
AHCTREUNG
none –
H2N-diether 7
100:0
95:5
85:15
70:30
95:5
85:15
70:30
50:50
95:5
106
373
510
827
182
308
110
214
89
0.4
0.8
1
1
H2N-PEG570
H2N-PEG570
-
-
G
0.7
0.7
0.4
0.2
0.6
0.5
0.3
0.3
0.4
0.6
0.5
0.4
0.5
0.3
85:15
70:30
50:50
98:2
95:5
90:10
98:2
107
129
139
111
232
130
154
171
138
Transfection experiments and targeting properties: The
transfection activities of the various formulations were eval-
uated by transfecting the pCMV-luc plasmid (see above)
into HeLa cells. The commercially available cationic lipid
Lipofectamine was used as a positive control (at a charge
ratio of 2), whereas unreacted (“naked”) plasmid DNA and
untreated cells were used as negative controls. Luciferase
expression in the transfected cells was quantified by lumin-
ometry (MLX Dynex)and the results were expressed as
FA-PEG570
FA-PEG570
-
-
N
95:5
90:10
non-pegylated archaea-derived co-lipid, such as amine 7, led
to larger vesicles with a high polydispersity index relative to
the small liposomes made solely of 21. Note also that formu-
lations prepared from lipid 21 and a pegylated co-lipid (15,
17, 19a or 20)yielded vesicles of moderate size that range
from 89 to 308 nm with acceptable-to-good polydispersity
indexes. This may be due to a reduced aggregation of the ar-
chaeosomes coated with PEG chains. Interestingly, mono-
layer-forming H2N-PEG570-tetraether 17 generally yielded
smaller vesicles and better polydispersities than its bilayer-
forming diether counterpart 15.
Next, lipid–DNA complexes were prepared by mixing ap-
propriate amounts of aqueous liposome or archaeosome sus-
pensions with plasmid DNA expressing the luciferase re-
porter gene (pCMV-luc, 10.6 kb). We added increasing
amounts of cationic lipid 21, lipids 21/7, lipids 21/15, lipids
21/17, lipids 21/19a or lipids 21/20 to a fixed amount of
DNA (4 mg)to form lipoplexes with increasing ( +/À)
charge ratios that range from 0.5 to 8 (mean theoretical
ratio of positive charges due to 21 to negative charges of the
ACHTREUNG
AHCTREUNG
evaluated. Figure 1 shows the transfection efficiencies of cat-
ionic lipid 21 alone and combined with diether derivatives 7
and 15. It clearly indicates that 1)the introduction of small
amounts of an archaea derivative increases the transfection
efficiency of cationic lipid 21 and 2)the optimal ( +/À)
charge ratio for these formulations lies between 2 and 4.
Indeed, the introduction of more than 30% of 7 causes a de-
crease in the beneficial effect brought about by this co-lipid
in low proportions and the presence of the PEG570 chain
(15)leads to a decrease in the transfection efficiency, espe-
cially for amounts above 15%. This last observation may be
due to a shielding effect by PEG. This results in a reduced
accessibility of the positive charges, which may limit the
electrostatic interactions required for internalisation of the
lipoplexes into the cell.
The H2N-PEG570-tetraether 17 and H2N-PEG570-diether
15 formulations yielded similar transfection efficiencies
Chem. Eur. J. 2008, 14, 8330 – 8340
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8335

T. Benvegnu et al.
case, only negative or slightly positive lipoplexes were tested
(0.5, 1 and 2 (+/À)charge ratio)to limit as much as possible
the cellular uptake of the lipoplexes by unspecific electro-
static interactions and to favour targeted folate-mediated
endocytosis.[2] Indeed, as mentioned in the Introduction, effi-
cient targeted gene transfer should have two main advantag-
es: 1)specific targeting and 2)the possibility of using neu-
tral lipoplexes.
Figures 3 and 4 show the transfection results obtained
with the most efficient ratios of 21/19a and 21/20, which
were 90:10 and 95:5, respectively. Interestingly, these lipid/
Figure 1. Transfection efficiencies for formulations of lipids 21/7 and 21/
15 (H2N-diether 7 and H2N-PEG570-diether 15) . (+/À)charge ratio from
0.5 to 8.
(both were just a bit less efficient than Lipofectamine). With
regards to the beneficial effect of the addition of 17, a limit
of 15% was also observed as in the case of H2N-PEG570-di-
ether 15 and the optimal (+/À)charge ratios were also be-
tween 2 and 4 (Figure 2). Altogether, these results confirm
Figure 3. Transfection efficiencies and competitive inhibitions for the
90:10 formulation of 21/19a. (+/À)charge ratio from 0.5 to 2 and con-
centration of FA from 0 to 50 nm.
co-lipid ratios correspond to similar molar ratios of folate-
equipped lipid (di- or tetraether)in the combined system
(1.26 mol% for a ratio of 90:10 of 21/19a and 1.45 mol%
for a 95:5 ratio of 21/20). This is indicative of the proportion
of ligand (FA)that is required on the surface of the lipo-
plexes for efficient cellular internalisation and transfection.
As shown in Figure 3, FA-PEG570-diether 19a combined
with cationic lipid 21 exhibited a very high transfection effi-
ciency for a neutral charge ratio (compared with H2N-
PEG570-diether 15, Figure 1); this activity was actually much
higher than that of the Lipofectamine positive control. Most
importantly, Figure 3 also shows that competitive inhibition
with free FA led to highly reduced transfection on addition
of FA (at concentrations equal to or higher than 25 nm)to
neutral lipoplexes (charge ratio of 1). This may be due to
the saturation of the FRs by free FA and suggests that the
internalisation of these neutral FA-equipped lipoplexes was
mainly mediated by FA–FR interactions. Moreover, note
that the transfection efficiency of Lipofectamine, which is
devoid of FA groups, was not affected by the addition of
free FA (even at 100 nm), the internalisation process is thus
being mediated here only by electrostatic interactions. In ad-
dition, Figure 3 also shows that the effect of competitive in-
hibition was clearly less pronounced for FA-equipped lipo-
plexes with a charge ratio of two (versus the neutral ones), a
Figure 2. Transfection efficiencies for formulations of 21/17 (H2N-PEG570
tetraether). (+/À)charge ratio from 0.5 to 8.
-
that the addition of archaea derivatives can be beneficial for
gene transfection by conventional cationic lipids. They also
show that there is a limit above which the incorporation of
pegylated derivatives becomes detrimental for transfection,
probably because of the PEG shielding effect.
Finally, we evaluated the transfection efficiencies and tar-
geting capabilities of the two folate co-lipids 19a and 20
combined with standard cationic lipid 21, the 21/co-lipid
ratio ranging from 98:2 to 90:10. Here, Hela cells were
again chosen because this cell line expresses high levels of
high-affinity a-folate receptors (FRa). The transfection effi-
ciencies were determined in the same way as the non-target-
ing co-lipids (see above)and competitive inhibition assays
were performed by adding increasing amounts of free FA
(0, 1, 10, 25 and 50 nm)into the cell culture medium. In this
8336
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340

Archaeal Lipid Derivatives
FULL PAPER
fact that supports the presence of an electrostatic contribu-
tion to the transfection process under such conditions.
Although the transfection efficiencies were lower than
with FA-PEG570-diether formulations, the FA-PEG570-tet-
raether formulations also displayed significant transfection
activity, which was strongly inhibited by the presence of free
FA (25 nm)in the cell culture medium (Figure 4 and by
significant increase in the efficiency of gene transfection in
vitro. Most importantly, neutral lipoplexes incorporating
FA-equipped di- and tetraether derivatives permitted
ligand/receptor-based targeted gene transfection because
their activity was inhibited when free FA was added to the
transfection medium. These novel lipids equipped with FA-
PEG moieties may thus be of great interest for targeted
gene transfection in vivo.
Experimental Section
General: Commercially available chemicals were used without further
purification and solvents were carefully dried and distilled prior to use.
Unless otherwise noted, non-aqueous reactions were carried out under a
nitrogen atmosphere. Analytical TLC was performed on Merck 60 F254
silica gel non-activated plates. A solution of 5% H2SO4 in EtOH was
used to develop the plates. Merck 60 H (5–40 mm)silica gel was used for
column chromatography. 1H and 13C NMR spectra were recorded at 400
and 100 MHz, respectively, on a Bruker Avance III. MS spectra were re-
corded on a Waters Micromass Q-TOF equipped with a Z-spray ion
source, IR spectra were recorded on a Thermo Nicolet 320 FTIR and op-
tical rotation were recorded on a Perkin–Elmer 341 polarimeter.
1-O-Benzyl-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexa-
decyl-sn-glycerol (4)
Figure 4. Transfection efficiencies and competitive inhibition for the 95:5
formulation of 21/20. (+/À)charge ratio from 0.5 to 2 and concentration
of FA from 0 to 50 nm.
Hexadecyl triflate: Triflic anhydride (9.0 mL, 54 mmol, 2 equiv)was
added dropwise to a stirred solution of 2,6-lutidine (6.3 mL, 54 mmol,
2 equiv)in dry CH Cl2 (60 mL)at 0 8C under a nitrogen atmosphere. The
2
reaction mixture was stirred at 08C for 10 min and hexadecan-1-ol (6.6 g,
27 mmol, 1 equiv)was added. After 1 h at room temperature, the reac-
tion was quenched by the addition of water and the aqueous phase was
extracted with CH2Cl2. The combined organic phases were successively
washed with 5% aqueous HCl, 5% aqueous NaHCO3 and brine, dried
(MgSO4)and concentrated under reduced pressure. Flash column chro-
matography on silica gel (petroleum ether (40–608C)(PE/)EtOAc, 9:1)
yielded hexadecyl triflate as a brown solid (9.95 g, 26.6 mmol, 98%).
comparison with H2N-PEG570-tetraether 17, Figure 2). This
again supports a targeted transfection that relies on ligand/
receptor-mediated internalisation. In these experiments, the
transfection efficiency of Lipofectamine was again not sig-
nificantly affected by the addition of free FA (100 nm)as
previously found with FA-PEG570-diether formulations (see
above and Figure 3). Finally, the lower transfection efficien-
cy of the FA-equipped tetraether 20 formulation (compared
with 19a)may be explained by the rigidity of the tetraether
structure[21–23] (compared with the fluidity/fusogenicity of the
diether derivatives[23,24]), which may increase too much the
stability of the lipoplexes and thus hinder their endosomal
escape, a critical transfection step known to require some
fluidity of the lipoplex membranes. Such rigidity may how-
ever be quite favourable for in vivo gene transfection, in
particular by the intravenous route (by increasing the lipo-
plex circulation time)and by aerosol delivery (by providing
resistance to the shear forces of the aerosolisation process).
Compound 4: A solution of the hexadecyl triflate (9.2 g, 24.6 mmol,
3 equiv)in dry CH 2Cl2 (10 mL)was added under a nitrogen atmosphere
to a mixture of 3 (3.8 g, 8.2 mmol, 1 equiv)and 1,8-bis(dimethylamino-)
naphthalene (proton sponge; 5.3 g, 24.6 mmol, 3 equiv)in dry CH 2Cl2
(40 mL). After 4 d at reflux, 5% aqueous HCl and CH2Cl2 were added.
The aqueous phase was extracted with CH2Cl2. The combined organic
phases were washed with water, dried (MgSO4)and concentrated under
reduced pressure. Flash chromatography on silica gel (PE/Et2O, 95:5)
gave compound 4 as a red oil (4.8 g, 83% in two steps). Rf =0.4 (PE/
1
EtOAc, 9:1); [a]2D0 =+3.7 (c=1 in CH2Cl2); H NMR (400 MHz, CDCl3):
d=0.75–0.83 (m, 18H), 1.00–1.52 (m, 52H), 3.55–3.62 (m, 6H), 3.63–3.70
(m, 2H), 3.74–3.79 (m, 1H), 4.48 (s, 2H), 7.19–7.27 ppm (m, 5H);
13C NMR (100 MHz, CDCl3): d=14.14, 19.63, 19.68, 19.75, 22.63, 22.73,
22.73, 24.37, 24.50, 24.81, 26.13, 27.98, 29.80, 32.77, 32.79, 29.38, 29.51,
29.65, 29.71, 31.93, 37.08, 37.16, 37.28, 37.35, 37.39, 37.46, 37.51, 39.36,
68.87, 70.25, 70.72, 71.65, 73.33, 77.90, 127.49, 127.58, 128.29, 138.98 ppm;
HRMS (ESI): m/z: calcd for C46H86O3 [M+Na]+: 709.6475; found:
709.6472; calcd for [M+K]+: 725.6214; found: 725.6224; elemental analy-
sis calcd (%)for C 46H86O3: C 80.40, H 12.61; found: C 80.65, H 12.65.
Conclusion
1-O-Hydroxy-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexa-
decyl-sn-glycerol (5): Palladium on activated carbon was added to a solu-
tion of 4 (1.6 g, 2.33 mmol, 1 equiv)in THF/MeOH (20 mL, 1:1.) After
3 h at room temperature under a hydrogen atmosphere, the suspension
was filtered through Celite and the solvent removed under reduced pres-
sure. Flash chromatography on silica gel (PE/EtOAc, 95:5)gave com-
pound 5 as a colourless oil (1.25 g, 90%). Rf =0.1 (PE/EtOAc, 95:5);
[a]2D0 =+7.3 (c=1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–0.89
(m, 18H), 1.07–1.57 (m, 52H), 2.17–2.20 (t, J=2.0 Hz, 1H), 3.41–3.45 (m,
2H), 3.46–3.55 (m, 4H), 3.62–3.64 (m, 2H), 3.71–3.72 ppm (m, 1H);
13C NMR (100 MHz, CDCl3): d=14.13, 19.61, 19.68, 19.75, 22.63, 22.72,
We have developed a series of novel archaeal lipid deriva-
tives that can be incorporated as co-lipids into cationic lipid
formulations for gene transfection. Both di- and tetraether
lipids were synthesised and functionalised with a PEG570
chain that was further equipped with a folate group with a
view to targeted transfection. When added in relatively
small proportions to a conventional glycine betaine-based
cationic lipid, the non-targeting pegylated co-lipids led to a
Chem. Eur. J. 2008, 14, 8330 – 8340
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8337

T. Benvegnu et al.
22.69, 24.35, 24.47, 24.80, 24.10, 27.97, 29.81, 32.77, 32.79, 29.36, 29.47,
29.61, 29.70, 31.92, 37.05, 37.13, 37.28, 37.31, 37.34, 37.38, 37.44, 37.49,
39.36, 63.10, 68.63, 70.92, 71.85, 78.23 ppm; IR (neat): n˜ =3446, 2924–
2854, 1465, 1377, 1371, 1118, 715–666 cmÀ1; HRMS (ESI): m/z: calcd for
C39H80O3 [M+Na]+: 619.6005; found: 619.5999; calcd for [M+K]+:
635.5744; found: 635.5752; elemental analysis calcd (%)for C 39H80O3: C
78.46, H 13.51; found: C 78.52, H 13.56.
1-O-Carboxyl-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexa-
decane-sn-glycerol (8): A 5% aqueous solution of NaHCO3 (6 mL)and
calcium hypochlorite (204.9 mg, 1.43 mmol, 8 equiv)were added portion-
wise to a mixture of 5 (106.9 mg, 0.18 mmol, 1 equiv)and TEMPO-ben-
zoate (4.5 mg, 0.018 mmol, 0.1 equiv)in THF (6 mL)at 0 8C. The reaction
mixture was stirred at 08C for 15 min and then at room temperature for
20 h. The white precipitate thus formed was dissolved with aq. 10% HCl
and the aqueous phase was extracted with EtOAc. The combined organic
phases were dried (Na2SO4)and concentrated under reduced pressure.
Column chromatography on silica gel (PE/EtOAc, 7:3)gave 8 as a col-
ourless oil (77.0 mg, 70%). Rf =0.2 (PE/EtOAc, 7:3); [a]2D0 =+1.7 (c=1
in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–0.89 (m, 18H), 1.02–
1.57 (m, 52H), 3.44–3.52 (m, 2H), 3.65–3.71 (m, 3H), 3.78–3.82 (dd, J=
15.1, 4.2 Hz, 1H), 4.03–4.05 (m, 1H), 11.43 ppm (s, 1H); 13C NMR
(100 MHz, CDCl3): d=14.13, 19.56, 19.69, 19.74, 22.62, 22.72, 22.72,
24.38, 24.49, 24.80, 25.95, 27.97, 29.82, 32.80, 32.81, 29.33, 29.37, 29.48,
29.66, 29.72, 31.93, 36.51, 36.59, 36.63, 37.28, 37.36, 37.41, 37.46, 69.68,
69.72, 72.05, 78.53, 172.20 ppm; IR (neat): n˜ =2925–2854, 1724, 1462,
1378, 1357, 1242, 1127, 937, 713–666 cmÀ1; HRMS (ESI): m/z: calcd for
C39H78O4 [MÀH]À: 609.5822; found: 609.5830; elemental analysis calcd
(%)for C 39H78O4: C 76.66, H 12.87; found: C 76.20, H 12.94.
1-O-Azide-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexade-
cane-sn-glycerol (6)
Mesylate: Mesyl chloride (245 mL, 3.15 mmol, 1.5 equiv)was added drop-
wise to a stirred solution of alcohol 5 (1.25 g, 2.1 mmol, 1 equiv)and tri-
A
After a yellow colour had appeared, water was added and the aqueous
phase was extracted with CH2Cl2. The combined organic phases were
successively washed with a saturated aqueous solution of NaHCO3 and
brine, dried (MgSO4)and concentrated under reduced pressure. Flash
chromatography on silica gel (PE/EtOAc, 9:1)gave the mesylate of 5 as
a colourless oil (1.20 g, 85%). Rf =0.4 (PE/EtOAc 9:1); [a]2D0 =+6.4 (c=
1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–089 (m, 18H), 1.04–
1.56 (m, 52H), 3.04 (s, 3H), 3.41–3.45 (m, 2H), 3.46–3.54 (m, 2H), 3.58–
3.52 (m, 2H), 3.65–3.68 (m, 1H), 4.42–4.26 (dd, J=3.6, 10.9 Hz, 1H),
4.36–4.40 ppm (dd, J=3.6, 10.9 Hz, 1H); 13C NMR (100 MHz, CDCl3):
d=14.12, 19.64, 19.67, 19.74, 22.62, 22.71, 22.67, 24.34, 24.47, 24.78, 26.05,
27.97, 31.92, 32.77, 32.78, 27.97, 31.92, 32.77, 32.78, 29.34, 29.45, 29.56,
29.59, 29.68, 31.90, 37.26, 37.32, 37.38, 37.42, 37.47, 37.50, 39.33, 37.44,
1-Chloro-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexade-
cane-sn-glycerol (9): Palladium on activated carbon was added to a solu-
tion of 6 (52.2 mg, 0.084 mmol, 1 equiv)in CHCl 3/MeOH (4 mL, 1:1).
After 4 h at room temperature under a hydrogen atmosphere, the suspen-
sion was filtered through Celite and the solvent was removed under re-
duced pressure. Flash chromatography on silica gel (PE/EtOAc, 95:5)
gave compound 9 as a colourless oil (43.9 mg, 90%). Rf =0.5 (PE/EtOAc,
98:2); [a]2D0 =+2.3 (c=1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=
0.83–0.89 (m, 18H), 1.07–1.57 (m, 52H), 3.43–3.46 (t, J=6.65 Hz, 2H),
3.52–3.53 (m, 2H), 3.55–3.67 ppm (m, 5H); 13C NMR (100 MHz, CDCl3):
d=14.13, 19.61, 19.68, 19.75, 22.63, 22.72, 22.69, 24.34, 24.50, 24.80, 26.08,
27.98, 29.70, 32.77, 32.79, 29.36, 29.46, 29.59, 29.62, 29.66, 31.92, 36.91,
36.99, 37.28, 37.34, 37.38, 37.40, 37.45, 37.47, 37.49, 39.36, 43.97, 68.90,
69.99, 71.77, 78.32 ppm; IR (neat): n˜ =2924–2854, 1464, 1377, 1366, 1123,
69.03, 69.07, 69.71, 71.88, 76.37 ppm; IR (Nujol): n˜ =1180, 1376 cmÀ1
;
HRMS (ESI): m/z: calcd for C40H82O5S [M+Na]+: 697.5781; found:
697.5783; elemental analysis calcd (%)for C 40H82O5S: C 71.16, H 12.24,
S 4.75; found: C 70.65, H 12.35, S 4.25.
Azide 6: Sodium azide (492 mg, 7.56 mmol, 1.5 equiv)and tetrabutylam-
monium bromide (812 mg, 2.52 mmol, 0.5 equiv)were added under a ni-
trogen atmosphere to a stirred solution of the mesylate of 5 (3.4 g,
5.04 mmol, 1 equiv)in dry DMF (10 mL.) The reaction mixture was
heated at reflux for 18 h, water was added and the aqueous phase was ex-
tracted with CH2Cl2. The combined organic phases were dried (MgSO4)
and concentrated under reduced pressure. The residue was purified by
721–666 cmÀ1
;
HRMS (ESI): m/z: calcd for C39H79O2Cl [M+Na]+:
637.5667; found: 637.5656; calcd for [M+K]+: 653.5406; found: 653.5413;
elemental analysis calcd (%)for C 39H79O2Cl: C 76.10, H 12.94; found: C
75.59, H 13.03.
flash chromatography on silica gel (PE/EtOAc, 98:2)to give the azide
6
as a colourless oil (2.97 g, 95%). Rf =0.8 (PE/EtOAc, 98:2); [a]2D0 =+3.2
(c=1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–0.89 (m, 18H),
1.04–1.66 (m, 52H), 3.32–3.35 (m, 2H), 3.41–3.45 (m, 3H), 3.49–3.53 (m,
1H), 3.55–3.67 ppm (m, 3H); 13C NMR (100 MHz, CDCl3): d=14.13,
19.59, 19.65, 19.69, 22.63, 22.72, 22.73, 24.33, 24.47, 24.81, 26.10, 27.97,
31.92, 32.77, 32.79, 29.37, 29.47, 29.60, 29.71, 31.93, 36.98, 37.00, 37.06,
37.09, 37.29, 37.35, 37.41, 37.46, 39.37, 52.05, 68.89, 70.09, 71.78,
77.86 ppm; IR (neat): n˜ =2925–2854, 2099, 1463, 1377, 1374, 1278, 1120,
cis-1,3-Bis(dodecan-1-O-{[(R)-tetramethylhexadecyl]-sn-glycerol}oxy-
methyl)-3-O-({[(R)-tetramethylhexadecyl]-sn-glycerol}oxybenzyl)cyclo-
pentane (11): Under a nitrogen atmosphere, a solution of diol 10 (97 mg,
0.080 mmol, 1 equiv)in dry THF (4 mL)and benzyl bromide (9 mL,
0.072 mmol, 0.9 equiv)were added dropwise at 0 8C to a suspension of
30% potassium hydride in oil (11 mg, 0.080 mmol, 1 equiv; previously
washed twice with THF). After 2 h at room temperature, water was
added slowly at 08C and the aqueous phase was extracted with EtOAc.
The combined organic phases were dried (MgSO4)and concentrated
under reduced pressure. Flash chromatography on silica gel (PE/EtOAc,
9:1)gave dibenzylated compound (26 mg, 25%,) diol 10 (26 mg, 35%)
and monobenzylated 11 (45 mg, 40%)as colourless oils. Rf =0.22 (PE/
714–666 cmÀ1
;
HRMS (ESI): m/z: calcd for C39H79O2N3 [M+Na]+:
644.6070; found: 644.6061; calcd for [M+K]+: 660.5809; found: 660.5819;
calcd for [MÀN2+Na]+: 616.6008; found: 616.6005; elemental analysis
calcd (%)for C 39H79N3O2: C 75.30, H 12.80, N 6.75; found: C 75.20, H
12.94, N 6.80.
1-O-Amine-2-O-[(7R,11R)-3,7,11,15-tetramethylhexadecyl]-3-O-hexade-
cane-sn-glycerol (7): PPh3 (48.1 mg, 0.179 mmol, 1.5 equiv)was added
portionwise to a stirred solution of 6 (74 mg, 0.119 mmol, 1 equiv)in
THF/H2O (6 mL, 1:1). After 4 h at room temperature the solvent was re-
moved under reduced pressure. Flash chromatography on silica gel
(CH2Cl2/MeOH, 96:4)yielded the amine 7 as a colourless oil (38.6 mg,
55%). Rf =0.04 (CH2Cl2/MeOH, 96:4); [a]2D0 =+5.7 (c=1 in CH2Cl2);
1H NMR (400 MHz, CDCl3): d=0.83–0.89 (m, 18H), 1.04–1.66 (m, 52H),
1.84 (brs, 2H), 2.77 (m, 1H), 2.87 (m, 1H), 3.38–3.45 (m, 3H), 3.47–3.55
(m, 3H), 3.60–3.70 ppm (m, 1H); 13C NMR (100 MHz, CDCl3): d=14.10,
19.60, 19.67, 19.72, 22.60, 22.70, 22.67, 24.34, 24.46, 24.78, 26.10, 27.94,
29.81, 32.76, 32.77, 29.34, 29.47, 29.63, 29.68, 31.90, 37.12, 37.19, 37.26,
37.32, 37.36, 39.43, 37.47, 39.33, 43.31, 68.55, 71.14, 71.72, 79.45 ppm; IR
(neat): n˜ =3372, 2924–2853, 1463, 1378, 1357, 1117, 729–666 cmÀ1; HRMS
(ESI): m/z: calcd for C39H81NO2 [M+Na]+: 618.6165; found: 618.6194;
calcd for [M+H]+: 596.6346; found: 596.6339; elemental analysis calcd
(%)for C 39H81NO2: C 78.58, H 13.70, N 2.35; found: C 78.35, H 13.77, N
2.50.
1
EtOAc, 9:1); [a]2D0 =+2.7 (c=1 in CH2Cl2); H NMR (400 MHz, CDCl3):
d=0.80–0.90 (m, 31H), 1.03–1.41 (m, 81H), 1.41–1.63 (m, 8H), 1.70–1.77
(m, 3H), 1.92–1.99 (dt, J=12.41 Hz, 8.0 Hz, 1H), 2.14–2.22 (m, 2H),
3.27–3.29 (d, J=6.8 Hz, 4H), 3.67–3.40 (t, J=6.8 Hz, 4H), 3.41–3.43 (t,
J=6.8 Hz, 4H), 3.45–3.55 (m, 7H), 3.52–3.53 (d, J=1.2 Hz, 1H), 3.57–
3.58 (d, J=1.2 Hz, 1H), 3.60–3.62 (m, 4H), 3.65–3.68 (m, 1H), 3.69–3.74
(m, 1H), 4.55 (s, 2H), 7.27–7.34 ppm (m, 5H); 13C NMR (100 MHz,
CDCl3): d=19.61, 19.68, 19.75, 22.63, 22.73, 22.63, 22.73, 24.36, 26.10–
26.18, 29.47–29.73, 28.85, 27.97, 29.80, 32.77, 32.79, 33.93, 37.05–37.48,
39.36, 39.71, 63.10, 68.62, 68.87, 70.27, 70.71, 70.92, 71.14, 71.66, 71.85,
73.34, 75.62, 77.91, 78.20, 127.48, 127.57, 128.29, 138.40 ppm; IR (neat):
n˜ =3473, 2925–2854, 1496, 1464, 1377, 1366, 1114, 828, 733–666 cmÀ1
;
HRMS (ESI): m/z: calcd for C84H160O8 [M+Na]+: 1320.2011; found:
1320.2016; elemental analysis calcd (%)for C 83H158O8: C 77.33, H 12.40;
found: C 77.46, H 12.47.
cis-1,3-Bis[dodecan-1-O-{[(R)-tetramethylhexadecyl]-sn-glycerol}carbox-
yl)-3-O-({[(R)-tetramethylhexadecyl]-sn-glycerol}oxybenzyl)cyclopentane
8338
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340

Archaeal Lipid Derivatives
FULL PAPER
(12): An aqueous solution of KBr (670 mL, 0.5m)and TEMPO (1 mg,
added and the organic phase was washed with water. The combined or-
ganic phases were dried (MgSO4)and concentrated under reduced pres-
sure. Flash chromatography on silica gel (CH2Cl2/MeOH, 9:1)gave 16 as
a yellow oil (60 mg, 90%). Rf =0.2 (CH2Cl2/MeOH, 9:1); [a]2D0 =À2.6
(c=1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–0.89 (m, 31H),
1.04–1.45 (m, 81H), 1.48–1.56 (m, 8H), 1.71–1.76 (m, 3H), 1.91–1.98 (dt,
J=12.4, 8.0 Hz, 1H), 2.13–2.22 (m, 2H), 3.27–3.29 (d, J=6.8 Hz, 4H),
3.37–3.39 (t, J=6.8 Hz, 4H), 3.40–3.44 (t, J=6.8 Hz, 2H), 3.44–3.57 (m,
10H), 3.57–3.60 (m, 7H), 3.61–3.69 (m, 37H), 3.75–3.78 (dd, J=2.4,
10.4 Hz, 1H), 3.88–3.90 (dd, J=2.4, 6.0 Hz, 1H), 4.55 (s, 2H), 7.04 (brs,
1H), 7.28–7.34 ppm (m, 5H); 13C NMR (100 MHz, CDCl3): d=19.62,
19.68, 19.75, 22.63, 22.72, 24.49, 24.81, 26.06, 26.13, 26.18, 26.19, 29.52,
29.53, 29.56, 29.64, 29.65, 28.86, 27.97, 29.74, 32.79, 33.93, 37.29–37.49,
39.36, 38.71, 39.73, 50.71, 68.87, 69.86, 70.06, 70.30, 70.59–70.73, 71.10,
71.66, 71.73, 73.34, 75.63, 77.93, 80.53, 127.48, 127.56, 128.29, 138.43,
170.60 ppm; IR (neat): n˜ =3583, 3473, 2925–2855, 2103, 1676, 1523, 1463,
0.006 mmol, 0.02 equiv)were added to
a
solution of
11 (400 mg,
0.31 mmol, 1 equiv)in EtOAc (5 mL.) A 12% aqueous solution of
NaOCl (200 mL)was added slowly at 0 8C. After 30 min at room tempera-
ture, the reaction mixture was acidified with 5% aqueous HCl to pH<3–
4. Then 12% aqueous NaClO2 (140 mL)was added slowly. The reaction
mixture was stirred at room temperature for 2 h and the aqueous phase
was extracted with EtOAc. The combined organic phases were washed
with brine and dried (MgSO4). Flash chromatography on silica gel (PE/
EtOAc, 9:1 then 7:3)gave 12 as a colourless oil (297 mg, 73%). Rf =0.2
(PE/EtOAc, 9:1); [a]D20 =À1.0 (c=1 in CH2Cl2); 1H NMR (400 MHz,
CDCl3): d=0.81–0.90 (m, 31H), 1.03–1.38 (m, 81H), 1.49–1.59 (m, 8H),
1.69–1.77 (m, 3H), 1.91–1.98 (dt, J=12.4, 8.0 Hz, 1H), 2.16–2.20 (m,
2H), 3.28–3.30 (d, J=6.8 Hz, 4H), 3.37–3.41 (t, J=6.8 Hz, 2H), 3.43–3.44
(t, J=6.8 Hz, 2H), 3.49–3.54 (m, 6H), 3.56–3.61 (m, 5H), 3.65–3.72 (m,
3H), 3.77–3.80 (dd, J=2.4, 10.4 Hz, 1H), 4.02–4.06 (q, J=3.2 Hz, 1H),
4.55 (s, 2H), 7.26–7.33 ppm (m, 5H); 13C NMR (100 MHz, CDCl3): d=
19.62, 19.63, 19.68, 19.75, 22.63, 22.71, 24.37, 24.47, 24.49, 24.80, 25.95,
26.14, 26.15, 26.18, 29.38, 29.42, 29.45, 29.52, 29.63, 29.67, 29.68, 29.71,
29.76, 28.85, 27.97, 29.82, 32.78, 32.80, 33.93, 37.11–37.52, 39.37, 39.73,
68.88, 69.94, 70.32, 70.52, 70.75, 71.14, 71.67, 72.05, 73.35, 75.61, 77.96,
78.61, 127.48, 127.57, 128.28, 138.43, 171.72 ppm; IR (neat): n˜ =3582,
2925–2854, 1733, 1496, 1464, 1377, 1366, 1260, 1114, 1029, 937, 733–
666 cmÀ1; elemental analysis calcd (%)for C 84H158O9: C 76.89, H 12.14;
found: C 76.53, H 12.37.
1376, 1350, 1257, 1114, 1029, 734–666 cmÀ1
.
H2N-PEG570-tetraether 17: Palladium on activated carbon was added to a
solution of 16 (145 mg, 0.078 mmol, 1 equiv)in CHCl 3/MeOH (12 mL,
1:1). After 18 h at room temperature under a hydrogen atmosphere, the
suspension was filtered through Celite and the solvent was removed
under reduced pressure. Flash chromatography on silica gel (CH2Cl2/
MeOH, 95:5)gave 17 as a colourless oil (70 mg, 51%). Rf =0.1 (CH2Cl2/
MeOH, 95:5); [a]2D0 =+4.4 (c=1 in CH2Cl2); 1H NMR (400 MHz,
CDCl3): d=0.83–0.88 (m, 31H), 1.02–1.31 (m, 81H), 1.42–1.58 (m, 8H),
1.69–1.76 (m, 3H), 1.93–1.98 (dt, J=12.4, 8.0 Hz, 1H), 2.14–2.22 (m,
2H), 3.27–3.29 (d, J=6.8 Hz, 4H), 3.37–3.40 (t, J=6.8 Hz, 2H), 3.42–3.44
(t, J=6.8 Hz, 2H), 3.45–3.57 (m, 12H), 3.57–3.60 (m, 7H), 3.63–3.69 (m,
37H), 3.75–3.78 (m, 1H), 3.89–3.90 (m, 1H), 3.93–3.94 (m, 2H), 7.04 ppm
(brs, 1H); 13C NMR (100 MHz, CDCl3): d=19.61, 19.68, 19.75, 22.62,
22.71, 24.35, 24.46, 24.79, 26.06, 26.10, 26.18, 26.19, 29.47–29.73, 28.85,
27.96, 29.82, 29.88, 32.78, 32.81, 33.93, 37.05–37.48, 39.36, 38.72, 39.72,
63.09, 66.94, 68.62, 69.85, 70.09, 70.34, 69.70, 69.77, 69.88, 70.00, 70.17,
70.21, 70.36–70.61, 71.14, 71.72, 70.92, 71.54, 71.85, 75.62, 78.24,
80.51 ppm; IR (neat): n˜ =3420, 2923–2854, 1661, 1530, 1463, 1377, 1350,
1250, 1110, 734–666 cmÀ1; HRMS (ESI): m/z: calcd for C101H202O19N2
[M+H]+: 1748.4980; found: 1748.4943.
N3-PEG570-diether 14: DIEA (149 mL, 0.9 mmol, 1.3 equiv)was added to
a mixture of 6 (460 mg, 0.9 mmol, 1 equiv)and TBTU (321 mg, 1 mmol,
1.3 equiv)in dry CH 2Cl2 (15 mL)under a nitrogen atmosphere. After
20 min at room temperature, a solution of N3-PEG570-NH2 13 (400 mg,
0.7 mmol, 1 equiv)in dry CH Cl2 (3 mL)was added and the reaction mix-
2
ture was stirred for 4 d. An aqueous solution of 1N HCl was added and
the organic phase was washed with water. The combined organic phases
were dried (MgSO4)and concentrated under reduced pressure. Flash
chromatography on silica gel (EtOAc/MeOH, 9:1)gave 14 as a viscous
yellow oil (941 mg, 93%). Rf =0.3 (EtOAc/MeOH, 9:1); [a]2D0 =À4.9 (c=
1 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=0.83–0.87 (m, 18H), 1.04–
1.78 (m, 52H), 3.37–3.40 (m, 2H), 3.41–3.50 (m, 4H), 3.53–3.57 (m, 4H),
3.58–3.69 (m, 43H), 3.75–3.78 (m, 1H), 3.88–3.90 (dd, J=2.51, 5.92 Hz,
1H), 7.02–7.05 ppm (m, 1H); 13C NMR (100 MHz, CDCl3): d=14.10,
19.58, 19.65, 19.72, 22.60, 22.69, 22.66, 24.36, 24.47, 24.78, 26.03, 27.94,
29.84, 32.76, 32.78, 29.34, 29.46, 29.52, 29.63, 29.68, 31.89, 37.26, 37.36,
37.39, 37.45, 37.50, 39.32, 38.67, 50.63, 69.72, 69.83, 70.3–70.7, 71.47,
71.68, 80.48, 170.57 ppm; IR (neat): n˜ =2025–2856, 2103, 1678, 1524,
1463, 1300, 1250, 1117, 720–666 cmÀ1; elemental analysis calcd (%)for
C63H126O14N4: C 65.02, H 10.91, N 4.81; found: C 64.79, H 10.95, N 4.79.
FA-PEG570-diether 19:
A solution of FA (18; 13.7 mg, 0.032 mmol,
1.2 equiv)in dry DMSO (5 mL)was heated until all material was dis-
solved. TBTU (16.7 mg, 0.052 mmol, 2 equiv), DIEA (10.7 mL,
0.065 mmol, 2.5 equiv)and a solution of 15 (29.6 mg, 0.026, 1 equiv)in
dry DMSO (2 mL)were added at room temperature under a nitrogen at-
mosphere. After 48 h at 308C in the dark, the reaction mixture was
lyophilised. The residue was purified by dialysis (molecular weight cut-
off (MWCO)=1000 D)against DMSO to afford 19 as a yellow solid mix-
ture of 19a, 19g and 15 (1.4 mg, 4 mol%, 11:14:75). 1H NMR (400 MHz,
[D6]DMSO): d=0.80–0.83 (m, 18H), 1.05–1.71 (m, 52H), 1.84–1.99 (m,
1H), 2.19–2.33 (m, 3H), 3.17–3.49 (m, 53H), 3.66–3.70 (m, 1H), 3.80–
3.82 (m, 1H), 4.25–4.26 (m, 1H), 4.47–4.48 (d, J=5.6 Hz, 2H), 5.63 (s,
1H), 6.53–6.54 (d, J=8.8 Hz, 2H), 6.94 (t, J=6.4 Hz, 1H), 7.58–7.59 (d,
J=8.4 Hz, 2H), 9.24 ppm (s, 1H); 13C NMR (100 MHz, [D6]DMSO): d=
14.10, 19.58, 19.65, 19.72, 22.61, 22.71, 22.66, 24.34, 24.45, 24.76, 26.00,
27.95, 29.81, 29.85, 32.79, 29.34, 29.46, 29.52, 29.63, 29.68, 31.89, 37.26,
37.36, 37.39, 37.45, 37.50, 39.32, 26.25, 30.63, 38.70, 39.38, 46.25, 52.50,
65.63, 80.63, 111.88, 129.38, 166.96, 170.63 ppm; HRMS (ESI): m/z: calcd
for C82H145O19N9 [MÀ2H+3Na]+: 1627.0193; found: 1627.0189;
[MÀH+2Na]+: 1605.0374; found: 1605.0337.
FA-PEG570-tetraether 20: A solution of FA (18; 11 mg, 0.025 mmol,
2.5 equiv)in dry DMSO (3 mL)was heated until all material was dis-
solved. TBTU (16 mg, 0.05 mmol, 5 equiv), DIEA (8 mL, 0.05 mmol,
5 equiv)and a solution of 17 (17 mg, 0.01, 1 equiv)in dry DMSO (4 mL)
were added at room temperature under a nitrogen atmosphere. After
24 h at 408C in the dark, the reaction mixture was lyophilised. The resi-
due was purified by dialysis (MWCO=1000)against DMSO to give 20
as yellow solid mixture of 20a, 20g and 17 (11.6 mg, 56 mol%, 47:44:9).
1H NMR (400 MHz, [D6]DMSO): d=0.74–0.84 (m, 31H), 0.98–1.58 (m,
89H), 1.69–1.86 (m, 3H), 1.92–2.06 (m, 9H), 2.37–2.40 (m, 4H), 3.11–
3.68 (m, 64H), 4.48–4.51 (m, 3H), 4.55–4.57 (m, 2H), 4.58–4.63 (m, 1H),
H2N-PEG570-diether 15: PPh3 (157.37 mg, 0.6 mmol, 1.5 equiv)was added
portionwise to a stirred solution of 14 (591 mg, 0.4 mmol, 1 equiv)in
THF/H2O (6 mL, 1:1). After 18 h at room temperature, the solvent was
removed under reduced pressure. Flash chromatography on silica gel
(CH2Cl2/MeOH, 95:5)gave amine 15 as a viscous yellow oil (327 mg,
72%). Rf =0.3 (CH2Cl2/MeOH, 9:1); [a]2D0 =À6.6 (c=1 in CH2Cl2);
1H NMR (400 MHz, CDCl3): d=0.82–0.88 (m, 18H), 0.99–1.69 (m, 52H),
3.16–3.18 (m, 2H), 3.39–3.49 (m, 4H), 3.54–3.57 (m, 4H), 3.59–3.70 (m,
38H), 3.72–3.75 (m, 3H), 3.76–3.77 (m, 1H), 3.87–3.90 (m, 3H), 7.04–
7.06 ppm (m, 1H); 13C NMR (100 MHz, CDCl3): d=14.10, 19.58, 19.65,
19.72, 22.61, 22.71, 22.66, 24.34, 24.45, 24.76, 26.00, 27.95, 29.77, 29.85,
32.79, 29.32, 29.45, 29.54, 20.61, 29.66, 31.88, 37.23, 37.34, 37.37, 37.43,
39.31, 38.70, 40.45, 66.89, 69.72–70.53, 71.45, 71.70, 80.47, 170.63 ppm; IR
(neat): n˜ =3426, 2024–2855, 1674, 1524, 1463, 1257, 1114, 714–666 cmÀ1
;
HRMS (ESI): m/z: calcd for C63H128O14N2 [M+H]+: 1137.9444; found:
1137.9438; elemental analysis calcd (%)for C 63H128O14N2·2H2O: C 64.47,
H 11.34, N 2.39; found: C 64.56, H 11.18, N 2.35.
N3-PEG570-tetraether 16: DIEA (8 mL, 0.046 mmol, 1.3 equiv)was added
to a mixture of 12 (55 mg, 0.042 mmol, 1.2 equiv)and TBTU (14 mg,
0.046 mmol, 1.3 equiv)in dry CH 2Cl2 (2 mL)under a nitrogen atmos-
phere. After 20 min at room temperature, a solution of N3-PEG570-NH2
13 (20 mg, 0.035 mmol, 1 equiv)in dry CH Cl2 (2 mL)was added and the
2
reaction mixture was stirred for 6 d. An aqueous solution of 1N HCl was
Chem. Eur. J. 2008, 14, 8330 – 8340
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8339

T. Benvegnu et al.
6.60–6.62 (m, 2H), 6.69–6.71 (m, 2H), 6.94 (m, 1H), 7.25 (m, 2H), 7.43–
7.45 (m, 2H), 7.71–7.80 (m, 2H), 7.80 (m, 1H), 8.67 (s, 1H), 8.68 ppm (s,
1H);13C NMR (100 MHz, [D6]DMSO): d=17.23, 18.54, 19.62, 19.92,
22.00, 22.77, 24.54, 27.85, 29.46, 32.62, 37.31, 39.31, 47.85, 48.15, 59.69,
70.15, 115.42, 115.97, 136.32, 137.01, 144.51 ppm. LRMS (ESI): m/z: 2170
[M]À.
[1] M. D. Brown, A. G. Schatzlein, I. F. Uchegbu, Int. J. Pharm. 2001,
229, 1.
[2] L. Wasungu, D. Hoekstra, J. Controlled Release 2006, 116, 255.
[3] M. C. Deshpande, M. C. Davies, M. C. Garnett, P. M. Williams, D.
Armitage, L. Bailey, M. Vamvakaki, S. P. Armes, S. Stolnik, J. Con-
trolled Release 2004, 97, 143.
Transfection assays: Adherent HeLa cells were seeded onto a 24-well
plate in RPMI-FF (folate free; 500 mL)and with 10% foetal calf serum
(FCS)at 100000 cells per well 24 h before transfection and incubated
overnight in a humidified 5% CO2 atmosphere at 378C. For transfection,
plasmid DNA (4 mg)encoding the luciferase reporter gene was mixed
with the appropriate amounts of cationic liposomes depending on the
charge ratio tested in a RPMI-FF medium without FCS. During the com-
plexation time, cells were washed twice with PBS 1X and then preserved
in the RPMI-FF medium (450 mL)not supplemented with FCS. The lipo-
plexes were added to the cell cultures 30 min later for competitive assays,
various amounts of folate (1, 10, 25 and 50 nm)were also rapidly added
to the medium. After 4 h, standard DMEM medium (2 mL)supplement-
ed with FCS and folate was added to each well. Luciferase activity of the
transfected cells was measured 48 h after transfection by using a chemilu-
minescent assay (Promega, Charbonni›re, France). Measurements were
carried out as described by the manufacturer. After rinsing twice in PBS
1X, cells were lysed with Lysis Buffer (200 mL, Promega)for 15 min. The
supernatant was then distributed into the wells of a 96-well opaque plate.
Luciferase activity in the supernatant was quantified with a luminometer
(MLX Microtiter Plate Luminometer, Dynex, Gyancourt, France)to in-
tegrate light emission over a 15 s reaction period. In parallel, the total
protein concentration was determined by using the BC Uptima protein
assay (Interchim, MontluÅon, France). The luciferase activity of each
sample was normalised to relative light units (RLU)per mg of extracted
protein. The total RLU mgÀ1 of protein obtained by summing up the
RLU values of eight wells of microtitre plates, were plotted on the y axis.
Each data point indicates the mean value of the total RLUmgÀ1 of pro-
tein from at least three experiments and the bars indicate the standard
deviation of this mean.
[4] J. Sudimack, R. J. Lee, Adv. Drug Delivery Rev. 2000, 41, 147.
[5] S. Wang, P. S. Low, J. Controlled Release 1998, 53, 39.
[6] P. S. Low, W. A. Henne, D. D. Doorneweerd, Acc. Chem. Res. 2008,
41, 120.
[7] J. A. Reddy, P. S. Low, J. Controlled Release 2000, 64, 27.
[8] J. A. Reddy, D. Dean, M. D. Kennedy, P. S. Low, J. Pharm. Sci. 1999,
88, 1112.
[9] P. S. Low, A. C. Antony, Adv. Drug Delivery Rev. 2004, 56, 1055.
[10] P. L. Sinn, M. A. Hickey, P. D. Staber, D. E. Dylla, S. A. Jeffers, B. L.
Davidson, D. A. Sanders, P. B. McCray, Jr., J. Virol. 2003, 77, 5902.
[11] T. Benvegnu, G. RØthorØ, M. Brard, W. Richter, D. Plusquellec,
Chem. Commun. 2005, 5536.
[12] G. RØthorØ, T. Montier, T. Le Gall, P. DelØpine, S. Cammas-Marion,
L. Lemi›gre, P. Lehn, T. Benvegnu, Chem. Commun. 2007, 2054.
[13] M. De Rosa, A. Gambacorta, A. Gliozzi, Microbiol. Rev. 1986, 70.
[14] G. D. Sprott, J. Bioenerg. Biomembr. 1992, 24, 555.
[15] T. Benvegnu, M. Brard, D. Plusquellec, Curr. Opin. Colloid Interface
Sci. 2004, 8, 469.
[16] G. B. Patel, G. D. Sprott, Crit. Rev. Biotechnol. 1999, 19, 317.
[17] G. B. Patel, W. Chen, Curr. Drug Delivery 2005, 2, 407.
[18] M. Brard, C. LainØ, G. RØthorØ, I. Laurent, C. Neveu, L. Lemi›gre,
T. Benvegnu, J. Org. Chem. 2007, 72, 8267.
[19] D. Gilot, M.-L. Miramon, T. Benvegnu, V. Ferrieres, O. Loreal, C.
Guguen-Guillouzo, D. Plusquellec, P. Loyer, J. Gene Med. 2002, 4,
415.
1
[20] Determined by H NMR analysis.
[21] A. P. Patwardhan, D. H. Thompson, Langmuir 2000, 16, 10340.
[22] J. L. C. M. van de Vossenberg, A. J. M. Driessen, W. N. Konings, Ex-
tremophiles 1998, 2, 163.
[23] K. Arakawa, T. Eguchi, K. Kakinuma, Bull. Chem. Soc. Jpn. 2001,
74, 347.
[24] R. Auzely-Velty, T. Benvegnu, G. Mackenzie, J. A. Haley, J. W.
Goodby, D. Plusquellec, Carbohydr. Res. 1998, 314, 65.
Acknowledgements
This work was supported by the Association “Vaincre La Mucoviscidose”
(Paris, France)(grant to C.L.), by the “RØgion Bretagne” and by “Ouest-
Genopole”. We are thankful to S. Pilard (UniversitØ de Picardie Jules
Verne)for recording the mass spectra.
Received: May 19, 2008
Published online: July 30, 2008
8340
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8330 – 8340