Synthesis and transfecting properties of a glycosylated polycationic DNA vector

Article abstract of DOI:10.1016/S0960-894X(00)00683-1

A glycosylated analogue of dialkylamidoglycylcarboxyspermine was synthesized. The physico-chemical and transfecting properties of 11 were evaluated indicating that this cationic lipid forms stable particles at low charge ratio and is efficient for gene delivery.

Full text of DOI:10.1016/S0960-894X(00)00683-1

Bioorganic & Medicinal Chemistry Letters 11 (2001) 419±422
Synthesis and Transfecting Properties of a Glycosylated
Polycationic DNA Vector1
Christophe Jacopin,a Hans Ho¯and,b Daniel Schermana,c and Jean Herscovicia,*
aLCBBMC UMR 7001 CNRS-ENSCP-Aventis-Gencell-Gencell, ENSCP, 11 rue Pierre et Marie Curie, 75231 Paris Cedex, France
bAventis-Gencell, 3825 Bay Center Place, Hayward, CA 94545, USA
cAventis-Gencell CRVA, 13 quai Jules Guesde, B.P. 14, Vitry sur Seine Cedex, France
Received 24 August 2000; revised 13 November 2000; accepted 29 November 2000
AbstractÐA glycosylated analogue of dialkylamidoglycylcarboxyspermine was synthesized. The physico-chemical and transfecting
properties of 11 were evaluated indicating that this cationic lipid forms stable particles at low charge ratio and is ecient for gene
delivery. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
an increased toxicity.17 In designing new vectors, we
have planned to covalently connect a non-charged
hydrophilic moiety at the end of the hydrophobic, in
order to decrease the surface charge density of lipoplexes
which would harbor on their surface these non-charged
hydrophilic domains. With this concept in mind, we
were attracted by carbohydrates that o€ered, in addition,
anti-aggregation18 and targeting properties.19 Several
groups have reported that glycosylation of cationic
polymers like polylysine20,21 or polyethylenimine22 pro-
motes the transfection. In the same fashion Remy et al.23
have used dialkylamidoglycylspermine/DNA complex
targeted to hepatoma cells by adding a triantennary
galactose ligand to a lipid associated to the complex.
Logically, the next step was to connect by a covalent
bond a saccharide to the hydrophobic anchor. Recent
reports in this ®eld19 prompted us to disclose our ongoing
results. In this report, we describe the synthesis and the
biological properties of a glycosylated lipopolyamine. In
order to evaluate the di€erent properties of the carbo-
hydrate, we have synthesized a rhamnosyl derivative.
This last choice was made to evaluate the physicochemical
in¯uence of a carbohydrate linked to the lipopolyamine,
because l-rhamnose is not recognized by mammal lectins
and thus does not possess any potential targeting capacity.
The use of DNA as a drug to cure inherited and
acquired diseases will depend on the ability to achieve
ecient introduction of a therapeutic nucleic acid into
target cells. For this reason much e€ort has been
focused on the development of viral2 and non-viral gene
delivery systems which self-associate to DNA. Basically,
synthetic DNA vectors are positively charged peptides,
polymers and lipids. Among those, cationic lipids have
been extensively studied.3,4 These amphipathic compounds
are made from a hydrophobic anchor connected to a
cationic domain by a spacer. Owing to their importance,
numerous modi®cations have been introduced to improve
their transfecting properties. Thus, new structures allowing
the biodegradability,5 9 reducing their toxicity10,11 and
possessing a targeting moiety12 have been reported with
results.
Because of the net negative charge exhibited by the
membrane, it has been observed that cationic lipid±DNA
complexes must be positive for optimal transfection.13,14
However, in vivo and physicochemical studies have
shown that this positive charge induced several problems.
First, lipoplexes interact with serum proteins, and are
removed from the circulation by complement.11,15,16
Second, less charged slightly positive or neutral particles
are unstable and aggregate, while using a large excess of
cationic lipid a€ords stable small complexes but leads to
Chemical Synthesis
Recent reports have described the glycosidation of carbo-
hydrates with long chain diols using enolates,24 trichloro-
acetamidate25 or anhydro sugars26 as glycosyl donors. We
*Corresponding author. Tel.: +33-1-53-10-1295; fax +33-1-53-10-1292;
e-mail: herscovi@ext.jussieu.fr
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00683-1

420
C. Jacopin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 419±422
Scheme 1. Reagents and conditions. (i) Methyl 15-hydroxy-pentadecanoate 2, SnCl4 0 ꢀC, 30 min then rt 3 h, 60%; (ii) 2 N MeONa in MeOH
quantitative; (iii) KI, NaH, BnBr, DMF overnight 70%; (iv) 25% NaOH re¯ux 30 min 73%; (v) octadecylamine, DIPEA, BOP, CHCl3 83%; (vi)
LiAlH4 2 equiv re¯ux 10 h THF 93%; (vii) Boc protected carboxyspermine 9, BOP, DIPEA, CHCl3 rt 65%; (viii) (a) Pd/C 10% H2 MeOH, (b)
tri¯uoroacetic acid 2 steps 100%.
were attracted by another procedure, i.e., the method
described by Banoub and Bundle27 for the glycosylation
of peracetylated carbohydrate with a mono-protected
diol. The synthesis of the glycosylated vector 10 is out-
lined in Scheme 1. The synthesis was started by the
condensation of the peracetylated rhamnopyranoside 1
with the methyl 15-hydoxy-pentadecanoate 2. First, 1
was stirred at 0 ꢀC in the presence of tin(IV) chloride
(1.5 equiv) for 30 min, then 2 was added, and the reaction
mixture was stirred for 3 h at rt to yield the b-glycoside
3 (60%). Removal of the acetyls was achieved by the
Zemplen method (MeONa 2 N in MeOH, 2 equiv;
30 min) to give the triol 4. Benzylation of 4 was
conducted by adding successively to a DMF solution
of the glycoside 4 equiv of potassium iodide, sodium
hydride and benzyl bromide. After one night, the per-
benzylated glycoside 5 was isolated in 70% yield.
Saponi®cation of 5 (25% NaOH in water, re¯ux 30 min)
a€orded 6 in 73% yield. The acid 6 was condensed with
octadecylamine in the presence of BOP to gave the
amide 7 (83%). Reduction of 7 was performed with
LiAlH4 (2 equiv; re¯ux overnight) and led to the desired
amine 8 in 93% yield. With 8 in hand, the preparation
of the cationic amphiphile was straightforward. Thus 8
was condensed with the BOC protected carboxy-sper-
mine 9.12 The reaction was promoted by the BOP
reagent and gave 10 in 65% yield. Next, 10 was fully
deprotected, ®rst by removal of the benzyl groups by
catalytic hydrogenation (Pd/C 10%), then by treating
the resulting product with tri¯uoroacetic acid. The glyco-
sylated lipopolyamine 11 was recovered quantitatively.28
structural analogue RPR 120535 12. First, the size of
the cationic lipid/DNA complexes29,30 was analyzed by
measuring the hydrodynamic diameter by dynamic light
scattering.31 The samples were diluted 20-fold in order
to avoid multiple scattering (Scheme 2).
At the ratio of 3 nmoles lipid/mg DNA a mean diameter
of 130 nm was recorded for lipoplex (lipopolyamine/
plasmid complex) made from micelles.29 When the
plasmid was complexed with liposome formulations
containing either cholesterol (25 mol%) or DOPE (dio-
leyl phosphatidyl ethanolamine 50 mol%) the diameters
were 137 and 153 nm respectively (Table 1). Next, the
colloidal stability of the lipoplex was studied at various
charge ratios. Pitard et al.17 have observed three zones
Scheme 2. RPR 120535 12.
Table 1. Size of the cationic lipid/DNA complexes in function of the
cationic lipid/DNA ratio
Compounds 0.5 0.75
1
1.5
2
3
4
5
6
11a
11b
11c
12a
12b
12c
A
A
A
A
A
A
B
B
B
B
B
B
A
A
A
C
C
C
C
C
C
B
B
B
C (130d)
C
C
C
B
B
B
C
C
C
C
B
B
C
C
C
C
C (153d)
C (137d)
B
B
B
B
aLipoplex made from micelles.
Evaluation of the Rhamno Lipopolyamine 11
bCationic lipid/DNA complex formulated with cholesterol.
cCationic lipid/DNA complex formulated with DOPE.
dSize in nm.
Compound 11 was essayed in vitro and the results were
compared with those obtained for the non-glycosylated

C. Jacopin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 419±422
421
namely A, B and C. In zones A and C, the complexes
are stable. A corresponds to low cationic/DNA charge
ratio thus leads to negatively charged lipoplex.
addition we have shown that the presence of the sugar
did not a€ect the transfecting properties of the poly-
amino vector. The galacto- and the manno- derivatives
are now under investigation and will be published in due
course.
On the other hand, C corresponds to higher charge
ratios, thus to lipoplex of net positive charges. In zone
B, where the charged lipid/DNA complex is near neu-
trality, the complexes are colloidally unstable and pre-
cipitate after a few minutes.
Acknowledgements
This work was ®nancially supported by the CNRS,
Aventis-Gencell, the Association pour la Recherche sur
le Cancer (ARC), the MENRT and the Region Ile de
France (SESAME).
Comparison with the RPR 120535 12 shows that the
precipitation zone is very small, allowing the formation
of stable particles for a charge ratio of 1.5 instead of 5±6
for the non-glycosylated cationic lipid RPR 120535 12
(Table 1). The glycosylated lipopolyamine 11 was
essayed for transfecting activity in vitro on HeLa cells.31
Results presented in Figure 1 indicate that the cationic
lipid shows the highest level of transfection eciency for
a 1.5±4 charge ratio (500 mg luciferase/105 cells). An
interesting fact is that micelles are more active than
DOPE containing complexes.
References and Notes
1. Jacopin, C.; Ho¯and, H.; Scherman, D.; Herscovici, J. PCT
application WO 00/32803.
2. Miller, A. D. Nature 1992, 35, 455.
3. Behr, J.-P. Bioconjugate Chem. 1994, 5, 382.
4. Miller, A. D. Angew. Chem., Int. Ed. 1998, 37, 1768.
5. Wang, J.; Guo, X.; Xu, Y.; Barron, L.; Szoka, F. C., Jr. J.
Med. Chem. 1998, 41, 2207.
6. Tang, F.; Hughes, J. A. Biochem. Biophys. Res. Commun. C
1998, 242, 141.
7. Rui, Y.; Wang, S.; Low, P. S.; Thompson, D. H. J. Am.
Chem. Soc. 1998, 120, 11213.
Finally, the transfection eciency between the RPR
120535 12 and the RPR 204686 11 were compared using
M109 cells.31 These data indicate that 11 is slightly more
ecient than 12 (Fig. 2).
In conclusion, glycosylated polyamino cationic lipids
are prepared from carbohydrate alkylamino lipids. The
introduction of the carbohydrate led to vectors with a
very small precipitation zone at low charge ratio. In
8. Tang, F.; Hughes, J. A. Bioconjugate Chem. 1999, 10, 791.
9. Obika, S.; Yu, W.; Shimoyama, A.; Uneda, T.; Miyashita,
K.; Doi, T.; Imanishi, T. Bioorg. Med. Chem. Lett. 1999, 7,
1817.
10. Aberle, A. M.; Tablin, F.; Zhu, J.; Walker, N. J.; Grue-
nert, D. C.; Nantz, M. H. Biochemistry 1998, 37, 6533.
11. Guenin, E.; Herve, A.-C.; Floch, V.; Loisel, S.; Yaouanc,
J.-J.; Clement, J. C.; Ferec, C.; des Abbayes, H. Angew.
Chem., Int. Ed. 2000, 39, 629.
12. Byk, G.; Dubertret, C.; Escriou, V.; Frederic, M.; Jaslin,
G.; Rangara, R.; Pitard, B.; Crouzet, J.; Wils, P.; Schwartz,
B.; Scherman, D. J. Med. Chem. 1998, 41, 224.
13. Behr, J.-P.; Demeneix, B.; Loe‚er, J. P.; Perez-Mutul, J.
Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 6982.
14. Zhou, X.; Klibanov, A. L.; Huang, L. Biochim. Biophys.
Acta 1991, 8.
15. Lewis, R. B.; Lin, K.-Y.; Kothavale, A.; Flanagan, W. M.;
Matteucci, M. D.; DePrince, R. B.; Mook, R. A.; Hendren, R.
W.; Wagner, R. W. Proc. Natl. Acad. Sci. U.S.A. 1996, 3176.
16. Vitiello, L.; Chonn, A.; Wasserman, J. D.; Du€, C.; Wor-
ton, R. G. Gene Ther. 1996, 3, 396.
17. Pitard, B.; Aguerre, O.; Ariau, M.; Lachages, A. M.;
Boukni-kachvilli, T.; Byk, G.; Dubertret, C.; Daniel, J. C.;
Herviou, C.; Scherman, D.; Mayaux, J. F.; Crouzet, J. Proc.
Natl. Acad. Sci. U.S.A. 1997, 94, 14412.
Figure 1. Transfection eciency of 11/DNA lipoplex as a function of
the charge ratio. The transfection was studied on HeLa cells with and
without DOPE.
18. Fuhrop, J.-H.; Tank, H. Chem. Phys. Lipids 1987, 43, 193.
19. Ren, T.; Liu, D. Tetrahedron Lett. 1999, 40, 7621.
20. Erbacher, P.; Bousser, M.-T.; Raimoni, J.; Monsigny, M.;
Midoux, P.; Roche, A. C. Hum. Gene Ther. 1996, 7, 721.
21. Liang, W. W.; Shi, X.; Deshpande, D.; Malanga, C. J.;
Rojannasakul, Y. Biochim. Biophys. Acta 1996, 1279, 227.
22. Bettinger, P.; Remy, J.-S.; Erbacher, P. Bioconjugate
Chem. 1999, 10, 558.
23. Remy, J. S.; Kichler, A.; Mordvinov, V.; Schuber, F.;
Behr, J.-P. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 1744.
24. Chenault, K. H.; Castro, A.; Cha¯in, L. F.; Yang, J. J.
Org. Chem. 1996, 61, 5024.
25. Kitov, P. I.; Pavel, I.; Railton, C.; Bundle, D. R. Carbo-
hydr. Res. 1998, 307, 361.
Figure 2. Comparison of the transfection eciency between 11 and 12.
The transfection was studied on M109 cells with DOPE.

422
C. Jacopin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 419±422
26. Timmers, C. M.; Wigchert, S. C. M.; Leeuwenburgh,
M. A.; van der Marel, G. A.; van Boom, J. H. Eur. J. Org.
Chem. 1998, 1, 91.
of 11 in chloroform. The solvent was removed under argon
and the sample was dried for 1 h then the lipid was hydrated
with a solution made of 5% dextrose and 10 mM sodium
chloride. After one night the solution was heated at 60 ꢀC
(5 min) and sonicated (1 min). The process was repeated sev-
eral times until the size of the lipidic particles was stable.
Plasmid pXL3031 encoding the luciferase gene under control
of the P/E CMV promoter of the cytomegalovirus was diluted
at the concentration of 0.5 mg/mL or 1.0 mg/mL in a solution
made of 5% dextrose and 10 mM sodium chloride. The lipoplexes
were prepared by mixing various volumes of the plasmid DNA
and the lipid 11 solutions at rt.
27. Banoub, J.; Bundle, D. R. Can. J. Chem. 1979, 57, 2085.
28. Mp 54 ꢀC, [a]D 10.77 (c 0.9, MeOH), MS (FAB), m/z
1
884.9 (MH+). H NMR (DMSO-d6, 400 MHz): d 0.88 (t, 3H,
J=6.36 Hz, octadecyl), 1.13 (d, 3H, J=6.45 Hz, H-6), 1.27 (m,
52 H, (CH2)11 N-pentadecyl and (CH2)15 N-octadecyl), 1.5 (m,
4H, CH2CH2N amide), 1.58 (m, 2H, CH2 CH2O), 1.66 (m,
4H, CH2CH2NH+ butyl), 1.92±2 (m, 4H, CH2CH2NH+ propyl),
2.92±3 (m, 12H, CH2NH+), 3.14 (m, 1H, H-4), 3.18 (m, 4H,
CH2N amide), 3.22 (m, 1H, OCHaCH2), 3.39 (m, 1H, H-5),
3.41 (m, 1H, H-3), 3.52 (m, 1H, OCHbCH2), 3.59 (m, 1H, H-2),
4.06 (m, 2H, NCCˆO), 4.5 (br s, 1H, H-1). 13C NMR (DMSO-
30. Size determination of the lipoplex was determined using a
Coulter N4+ particle analyzer. Fifty mL of the lipid/DNA
complex suspension is diluted to 1 mL in (5% Dex-
trose+20 mM NaCl) and placed in the N4+cuvette.
d6+CD3COOD, 400 MHz):
d 14.38 (C-18 N-octadecyl-
amine), 18 (C-6), 22.5, 23 and 24.01 (CCNH+), 29 ((CH2)11 N-
pentadecyl and (CH2)15 N-octadecyl), 36 (CNH+3 ), 42 (CN
amide), 44±46 (CNH+2 ), 47.5 (NCCˆO), 66.5 (CH2O), 68.5
(C-5), 70.5 (C-2), 71 (C-3), 72 (C-4), 100 (C-1), 158.40 (CˆO).
29. The cationic micelles (particles from less than 5 nm dia-
meter) were prepared by drying the lipid from organic solvent
with argon or nitrogen to form a thin ®lm on the bottom of a
glass tube. The ®lm is vacuum-desiccated for 1 h. The lipid ®lm
is allowed to hydrate in 5% dextrose and 20 mM NaCl at 4 ꢀC
overnight. Subsequently, the ®lms are sonicated for 5 min and
heated to 50 ꢀC for 30 min. Liposomes formulation contain
`helper lipids', i.e., either 50 mol% dioleyl phosphatidyl ethanol-
amine (DOPE) or 25 mol% cholesterol. The appropriate
amount of helper lipid (CHCl3 stock) was added to a solution
31. HeLa cells are plated at 60,000 cells per well of a 24-well
plate, and allowed to grow overnight. The medium (DMEM+
10% FBS) is removed and replaced by 0.5 mL DMEM with-
out serum, containing 1 mg DNA complex. The cells are
transfected for 4 h. Subsequently, the complex containing
medium is replaced with DMEM+10% FBS. The cells are
allowed to grow for another 24 h, after which they are harvested
with 100 mL luciferase lysis bu€er per well. M109 cells are lung
carcinoma cells that are derived from Balb/C mice. The trans-
fection protocol is identical to that used for HeLa cells except
for the fact that we plate them at 40,000 cells/well because they
grow slightly faster (so on the day of transfection the number of
cells per well is approximately 100,000 for both HeLa and M109).