Novel macrocyclic and acyclic cationic lipids for gene transfer: Synthesis and in vitro evaluation

Article abstract of DOI:10.1016/j.bmcl.2012.05.080

The synthesis and in vitro evaluation of four cationic lipid gene delivery vectors, characterized by acyclic or macrocyclic, and saturated or unsaturated hydrophobic regions, is described. The synthesis employed standard protocols, including ring-closing metathesis for macrocyclic lipid construction. All lipoplexes studied, formulated from plasmid DNA and a liposome composed of a synthesized lipid, 1,2-dimyristoyl-sn-glycero-3-ethylphosphocholine (EPC), and either 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) or cholesterol as co-lipid, exhibited plasmid DNA binding and protection from DNase I degradation, and concentration dependent cytotoxicity using Chinese hamster ovary-K1 cells. The transfection efficiency of formulations with cholesterol outperformed those with DOPE, and in many cases the EPC/cholesterol control, and formulations with a macrocyclic lipid (+/- 10:1) outperformed their acyclic counterparts (+/- 3:1).

Full text of DOI:10.1016/j.bmcl.2012.05.080

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4686–4692
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel macrocyclic and acyclic cationic lipids for gene transfer: Synthesis
and in vitro evaluation
⇑
,
William P. D. Goldring a, , Emile Jubeli b, , Rachael A. Downs a, , Adam J. S. Johnston a,
Nada Abdul Khalique b, Liji Raju b, Deena Wafadari c, Michael D. Pungente c, ,
⇑
a School of Chemistry and Chemical Engineering, Queen’s University Belfast, David Keir Building, Stranmillis Road, Belfast, Northern Ireland, BT9 5AG, United Kingdom
b Research Division, Weill Cornell Medical College in Qatar, Education City, PO Box 24144, Doha, Qatar
c Premedical Unit, Weill Cornell Medical College in Qatar, Education City, PO Box 24144, Doha, Qatar
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis and in vitro evaluation of four cationic lipid gene delivery vectors, characterized by acyclic
or macrocyclic, and saturated or unsaturated hydrophobic regions, is described. The synthesis employed
standard protocols, including ring-closing metathesis for macrocyclic lipid construction. All lipoplexes
studied, formulated from plasmid DNA and a liposome composed of a synthesized lipid, 1,2-dimyri-
stoyl-sn-glycero-3-ethylphosphocholine (EPC), and either 1,2-dioleoyl-sn-glycero-3-phosphoethanol-
amine (DOPE) or cholesterol as co-lipid, exhibited plasmid DNA binding and protection from DNase I
degradation, and concentration dependent cytotoxicity using Chinese hamster ovary-K1 cells. The trans-
fection efficiency of formulations with cholesterol outperformed those with DOPE, and in many cases the
EPC/cholesterol control, and formulations with a macrocyclic lipid (+/ꢀ 10:1) outperformed their acyclic
counterparts (+/ꢀ 3:1).
Received 19 April 2012
Revised 17 May 2012
Accepted 21 May 2012
Available online 26 May 2012
Keywords:
Non-viral gene therapy
Ring-closing metathesis
Macrocyclic lipids
DNA transfection
Ó 2012 Elsevier Ltd. All rights reserved.
Gene therapy has significant potential for the treatment of
inherited and acquired life threatening diseases, such as cancer,
AIDS, cardiovascular diseases, and certain autoimmune disorders.1
Viral vectors are the most widely used gene delivery vehicle due to
their high transfection efficiency. However, the use of viral based
therapeutics is restricted by drawbacks including immunogenicity,
biological safety, and size restrictions on the plasmid DNA (pDNA)
cargo. For these reasons, non-viral synthetic gene delivery reagents
represent an attractive, alternative approach to gene delivery that
offers many advantages over viral vectors. Liposomes based on cat-
ionic lipids have been the subject of considerable interest as non-
viral delivery vectors,2,3 which, in general, exhibit excellent bio-
compatibility, low immunogenicity,4 low toxicity,5,6 well defined
physical and chemical composition, large nucleic acid payload
capacity, and the potential to transfect numerous tissue and cell
types.7 Furthermore, cationic lipids are characterised by their sim-
plicity of production and ease of application.
phobic domain architectural constructs are composed of either; (i)
two flexible (non-rigid) hydrophobic tails, mainly as saturated or
mono-unsaturated fatty acid chains containing typically 14 to 18
carbon atoms, where 1,2-dimyristoyl-sn-glycero-3-ethylphosph-
ocholine (EPC) in Figure 1 is a representative example, or (ii) a rigid
cholesteryl moiety, exemplified by 3b-[N-(N0,N0-dimethylamino-
ethyl)-carbamoyl]cholesterol (DC-Chol).
A
common empirical
H
Cl
H
H
N
H
Rigid
O
O
H
N
H
O
DC-Chol
Rigidity-
range
O
The typical structure of cationic lipid vectors includes a posi-
tively charged polar (hydrophilic) headgroup connected through
to a hydrophobic domain via a linker unit (e.g., an ester) and a
backbone, typically glycerol (Fig. 1). The two most common hydro-
O
Cl
H
O
P
O
O
Non-rigid
N
OEt
O
EPC
⇑
Corresponding authors. Tel.: +44 (0) 28 9097 4414; fax: +44 (0) 28 9097 6524
(W.P.D.G.); tel.: +974 4492 8216; fax: +974 4492 8222 (M.D.P.).
E-mail addresses: w.goldring@qub.ac.uk (W.P.D. Goldring), mdp2001@qatar-
med.cornell.edu (M.D. Pungente).
hydrophobic domain backbone headgroup
Figure 1. Lipid structural features and the rigidity range associated with the
These authors have contributed equally to this work.
hydrophobic domain of common cationic lipids DC-Chol and EPC.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.080

W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
4687
approach to improving transfection efficiency and reducing cellu-
lar toxicity has focused on synthetic modifications to the positively
charged headgroup and/or linker functionality, while less attention
has been given to the role of the hydrophobic domain. Our recent
research efforts have focused on structural modifications to the li-
pid headgroup, but also an investigation of the rigidity-range
(Fig. 1) associated with the hydrophobic core of lipid vectors to-
wards enhancing nucleic acid delivery,8,9 where the lipids EPC
and DC-Chol represent two extremes along the rigidity-range.
Macrocyclic archaebacteria lipids10 have attracted much atten-
tion because they form liposomes that demonstrate enhanced sta-
bility to oxidative stress, high temperature, alkaline pH and the
action of serum proteins.11 It has been previously reported that
these lipids form more tightly packed vesicles than their acyclic
counterparts.12 Furthermore, Benvegnu and co-workers13 have re-
ported the use of novel, archaeobacterial-like cationic lipids for
in vitro gene transfection for the purpose of modulating the lipidic
membrane fluidity of the complexes they form with DNA. Such re-
ports have led us to hypothesize that macrocyclic cationic lipids
may package pDNA sufficiently different than their acyclic lipid
analogues, possibly giving rise to enhanced protection of the pDNA
cargo from degradative enzymes, and ultimately enhanced gene
transfer. To the best of our knowledge there are no reports on
the use of cationic lipids containing a macrocyclic hydrophobic do-
main as gene carriers. The synthesis of a macrocyclic lipid as a ‘per-
spective’ molecule for gene transfer was previously described in
the literature.14 Reports on the use of cyclic structures in the head-
group have appeared in the literature, including applications of;
liposomes containing varying amounts of surfactin, a biosurfactant
containing a cyclic peptide in the headgroup,15 a novel, single-
chain lipid containing a macrocyclic polyamine (Cyclen) head-
group,16 and a cationic glycolipid containing a cyclic galactose-
based headgroup.17 Finally, examples of gene carriers which incor-
porate a macrocyclic structure in the backbone include structures
which possess a calix[4]arene18 or a cyclodextrin.19
In the present work, four novel cationic lipids, 1–4, were syn-
thesized as gene delivery reagents (Fig. 2a), and liposomes were
prepared from the synthetic lipids mixed with the commercial cat-
ionic lipid EPC together with a neutral co-lipid, either 1,2-dioleoyl-
sn-glycero-3-phosphoethanolamine (DOPE) or cholesterol (Chol)
(Fig. 2b). An equal molar proportion of the cationic lipid EPC was
included in each lipoplex formulation to aid in the solubility of
the synthetic lipids 1–4. Furthermore, the formulations contained
either the neutral lipid DOPE or Chol to assess the impact these
co-lipids had on gene delivery. Lipid/pDNA complexes (lipoplexes)
were formulated by associating varying amounts of liposomes with
a constant amount of pDNA. The interaction of liposomes with
pDNA was investigated via gel electrophoresis and the protective
character of the cationic lipids within these structures was as-
sessed via a degradation assay. The set of four related cationic lip-
ids possess identical amine headgroups, and comprise two pairs;
each of which contains, in the hydrophobic domain, an acyclic
and a cyclic counterpart. One pair, 1 and 3, is composed of lipids
containing one or two alkene units in the hydrophobic region,
and the other pair, 2 and 4, is composed of the corresponding sat-
urated lipids. An evaluation of the in vitro cytotoxicity and gene
transfer structure-activity relationships were conducted with Chi-
nese hamster ovary-K1 (CHO-K1) cells.
Our synthesis of a cationic lipid library began with the protec-
tion of commercially available solketal (5), using 4-methoxybenzyl
chloride and potassium hydroxide in dimethyl sulfoxide (DMSO),
followed by treatment of the cyclic acetal with 10% aqueous acetic
acid to give the diol 6 in 50% yield over two steps (Scheme 1a).
Esterification of the diol 6 with two equivalents of 6-heptenoic acid
using N,N0-dicyclohexylcarbodiimide (DCC) and 4-(dimethyl-
amino)pyridine (DMAP) gave the diene 7 in 58% yield. The diene
7 was used as a common building block from which all of the re-
quired cationic lipids were elaborated. For example, construction
of the unsaturated acyclic cationic lipid 1 began with deprotection
of the benzyl ether in 7 using 2,3-dichloro-5,6-dicyano-p-benzo-
quinone (DDQ) to give the crude alcohol 8, which was used in
the subsequent step without further purification. Our attempts to
purify the alcohol 8 by flash column chromatography on silica,
and other related structures such as the alcohol 12 (Scheme 2),
led to an inseparable mixture of acyl shift products, which included
the corresponding symmetrical constitutional isomer.20 Next,
installation of the cationic headgroup began with esterification of
the crude alcohol 8 using 5-bromovaleric acid, in the presence of
DCC and DMAP, to give the tri-ester 9 in 17% yield over two steps.
We were disappointed to discover this reaction, and others such as
the construction of 10, suffered from the formation of a symmetri-
cal tri-ester, which made up the remainder of the material balance,
resulting from secondary to primary hydroxyl acyl shift, followed
by esterification of the resulting secondary alcohol with 5-brom-
ovaleric acid. Finally, the bromide 9 was displaced with a 10-fold
excess of dimethylamine (2 M in tetrahydrofuran (THF)), followed
by treatment of the resulting amine with a 2 M solution of hydro-
chloric acid in diethyl ether to give the cationic unsaturated acyclic
lipid 1 in 99% yield over two steps.21
O
a
O
Cl
O
O
O
N
N
H
H
O
1
O
O
O
Cl
Cl
O
O
O
2
O
O
O
O
O
O
N
H
O
3
O
O
O
Cl
O
O
N
H
4
H
b
H
H
OH
H
Construction of the cationic saturated acyclic lipid 2, which fol-
lowed a similar sequence of steps to those described above for 1,
started from the common building block 7. Hydrogenation and
hydrogenolysis of 7 using hydrogen gas and 10% palladium on car-
bon, followed by esterification of the resulting crude alcohol gave
the tri-ester 10 in 67% yield over two steps (Scheme 1b). In a
two-step sequence, involving amination and subsequent treatment
with acid, the bromide 10 was converted into the cationic lipid 2 in
99% yield over two steps.
cholesterol
O
O
H
O
P
O
5
5
5
5
O
O
O
NH3
O
DOPE
Figure 2. The structures of lipids employed in this study: (a) the acyclic and
macrocyclic cationic lipids, 1–4; and (b) the neutral co-lipids cholesterol and DOPE.

4688
W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
O
O
O
a
O
O
i,ii
HO
HO
iii
iv
O
O
O
O
OH
OPMB
O
OPMB
OH
5
6
7
8
O
O
O
O
O
O
Cl
v
O
vi
vi
O
O
O
O
Br
Br
O
O
N
N
H
O
O
9
1
O
O
O
O
b
7
O
Cl
vii,v
O
O
O
O
H
10
2
Scheme 1. Reagents and conditions: (i) PMBCl, KOH, DMSO, rt; (ii) 10% AcOH, water, 90 °C, 50% (2 steps); (iii) 6-heptenoic acid (2 equiv.), DCC, DMAP, CH2Cl2, rt, 58%;
(iv) DDQ, water/CH2Cl2 (1:10), rt; (v) 5-bromovaleric acid, DCC, DMAP, CH2Cl2, rt, 17% (9) or 67% (10) (2 steps); (vi) a. 2 M (CH3)2NH in THF, THF, rt; then b. 2 M HCl in diethyl
ether, rt, 99% (2 steps); (vii) H2, 10% Pd-C, AcOH/EtOH (1:10), rt. PMB = p-methoxybenzyl.
O
O
O
O
O
Cl
Cl
i
ii-iv
O
O
O
O
7
OPMB
O
O
N
N
H
H
11
3
v
O
O
O
O
O
O
O
iii,iv
O
O
OH
12
4
Scheme 2. Reagents and conditions: (i) Grubbs’ 2nd generation pre-catalyst, CH2Cl2, 40 °C, 73% (3:1, E:Z); (ii) DDQ, water/CH2Cl2 (1:10), rt; (iii) 5-bromovaleric acid, DCC,
DMAP, CH2Cl2, rt, 44% (to 3) or 52% (to 4); (iv) a. 2 M (CH3)2NH in THF, THF, rt; then, b. 2 M HCl in diethyl ether, rt, 99% (3) or 93% (4); (v) H2, 10% Pd-C, AcOH/EtOH (1:10), rt.
The common building block 7 was employed to elaborate the
macrocyclic cationic lipids 3 and 4, corresponding to the cationic
the combined lipids formed upon evaporation of ethanol under re-
duced pressure. Liposome particle sizing was obtained for all lipo-
some formulations (with DOPE or Chol as co-lipid) using dynamic
light scattering and revealed a range in diameter between 150 and
530 nm. It is interesting to note, all of the formulations using the
unsaturated lipids, 1 or 3, with the exception of 1/EPC/Chol, were
approximately three times larger than their saturated counter-
parts. Subsequently, combining the various cationic liposomes
with negatively charged pDNA resulted in lipoplex formation, ini-
tially mediated by electrostatic interactions and subsequently by
hydrophobic effects.23–25 The lipoplexes assembled into nano-
sized particles with diameters in the range 290–900 nm.
Lipoplex formation upon association of liposomes with pDNA
was investigated over a range of cationic lipid:DNA (+/ꢀ, or
nitrogen/phosphorous (N/P)) molar charge ratios using a gel
retardation assay (Fig. 3). Similar retention trends were observed
in formulations with either DOPE or cholesterol as the co-lipid.
Liposome formulations with high N:P molar charge ratios re-
vealed the highest level of pDNA association, evident from Fig-
ure 3. Lipoplexes containing 2/EPC/DOPE, 4/EPC/DOPE, 2/EPC/
Chol and EPC/Chol revealed complete retention at a N:P molar
charge ratio of 10:1. All other formulations showed incomplete
retention, even at the higher molar charge ratios, however most
lipoplexes exhibited some degree of retention from a N:P molar
charge ratio of 3:1. The minor differences observed for pDNA
association between lipoplexes using the same co-lipid and at
the same molar charge ratio, were likely due to the differences
in the hydrophobic domain of the cationic lipids, 1–4, each of
which possess the same headgroup.
acyclic lipids
1 and 2, respectively. Ring-closing metathesis
(RCM) of the diene 7 using Grubbs’ second generation pre-catalyst,
under high dilution conditions, gave an inseparable 3:1 mixture of
the macrocyclic alkene isomers 11 in 73% combined yield
(Scheme 2). Using proton NMR we were unable to conclusively
determine which of the alkene isomers of 11 was the major prod-
uct. Molecular mechanics calculations, together with a Monte Carlo
conformational search, were separately performed on the individ-
ual isomeric macrocyclic products 11, each of which led to a num-
ber of low energy conformations. Using the energy difference
between the global minimum conformation of each geometric iso-
mer a 3:1 (E)-11 to (Z)-11 ratio was calculated.22 Next, the macro-
cyclic lipid 11 was converted into the unsaturated macrocyclic
cationic lipid 3 in 44% yield over three steps, which included:
DDQ deprotection, esterification with 5-bromovaleric acid, and
then conversion of the alkyl bromide into the ammonium salt 3.
During the conversion of 11 into 3, all synthetic intermediates
and the cationic lipid 3 were formed as inseparable 3:1 mixtures
of geometric isomers. Finally, using a similar sequence of steps
the saturated macrocyclic cationic lipid 4 was constructed from
11, via the crude alcohol 12, in 48% yield over three steps.
Lipoplexes were prepared from liposomes that contained one of
the four synthetic cationic lipids, 1–4, combined with pDNA. The
liposomes were formulated first by combining ethanolic stock
solutions of a cationic lipid, 1–4, along with EPC and a neutral
co-lipid, DOPE or cholesterol, in a 1:1:2 molar ratio. The cationic
liposomes were prepared by sonication of hydrated thin films of

W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
4689
Figure 3. Gel retardation assays of liposomes formulated with a cationic lipid, 1–4, EPC, and either DOPE (A) or cholesterol (B) as co-lipid, combined with pDNA at various N:P
(+/ꢀ) molar charge ratios, ranging from 0.5:1 to 10:1, and run through a 1% agarose gel impregnated with the pDNA gel stain, ethidium bromide. Lanes L and D denote the 1 kb
DNA ladder and pDNA, respectively.
A
B
Figure 4. DNase I degradation assays of liposomes formulated with a cationic lipid, 1–4, EPC, and either DOPE (A) or cholesterol (B) as co-lipid, combined with pDNA at
various N:P (+/ꢀ) molar charge ratios, ranging from 0.5:1 to 10:1, and run through a 1% agarose gel impregnated with the pDNA gel stain, ethidium bromide. Lanes L and D
denote the 1 kb DNA ladder and pDNA, respectively.

4690
W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
A set of DNase I degradation assays were performed to assess
plexes containing only EPC and DOPE at all but the lowest molar
charge ratio of 0.5:1 (Fig. 5A). With the exception of 3/EPC/DOPE
and EPC/DOPE, less than 70% cell viability was observed in the
DOPE-containing formulations with a N:P molar charge ratio high-
er than 1:1. However, formulations of the cationic lipids 1–4 with
EPC and cholesterol as co-lipid showed comparable or better cell
tolerance than the control lipoplexes containing only EPC and cho-
lesterol (Fig. 5B). The formulations with a N:P molar charge ratio
up to and including 3:1 revealed approximately >80% cell viability.
Moreover, lipoplexes containing the macrocyclic cationic lipids (3
and 4) with cholesterol as co-lipid revealed cell viability approxi-
mately >60% for all N:P molar charge ratios examined in this study.
In order to evaluate the pDNA transfection efficiency of lipo-
plexes formulated with the cationic lipids 1–4, cells were transfec-
ted with lipoplexes for 4 h, then replaced with complete growth
media, followed by incubation for a further 48 h at 37 °C in the
presence of 5% CO2 before assaying b-galactosidase expression
using the Beta-GloÒ Assay System (Promega).
All lipoplex formulations containing compounds 1–4 formu-
lated with EPC and either DOPE or cholesterol as co-lipids demon-
strated some level of gene induction evidenced by the luciferin
luminescence as a result of b-galactosidase gene transfer to cells
(Fig. 6). Cells incubated with the free soluble pDNA did not show
any detectable transfection (data not shown). The transfection effi-
ciency of lipoplexes containing cholesterol as co-lipid was gener-
ally found to be greater than those formulated with DOPE, with
the exception of the lower N:P (+/ꢀ) molar charge ratios assayed
(e.g., 0.5:1 and 1:1).
the protective character the novel cationic vectors offer the nucleic
acid cargo from degradative enzymes. This assay involved incuba-
tion of each lipoplex with DNase I, which cleaves unbound and/or
unprotected pDNA into linear fragments, followed by quenching of
the DNase and incubation of the lipoplexes with the surfactant so-
dium dodecyl sulfate (SDS) to release any residual, intact DNA from
the lipoplex which was subsequently visualized by gel
electrophoresis.
DNase I protection experiments were carried out by complexing
pDNA with liposomes containing a cationic lipid, 1–4, at N:P (+/ꢀ)
molar charge ratios ranging from 0.5:1 to 10:1 and then incubating
with DNase I for 1 h at 37 °C. Figure 4 shows the naked pDNA (lane
D) followed by the naked pDNA incubated with DNase I, resulting
in complete digestion (lane DNase) as indicated by the absence of
the intense band for naked pDNA. Lipoplexes with N:P molar
charge ratios P3:1 generally revealed greater protection of the nu-
cleic acid cargo. A comparison between the lipoplexes formulated
with DOPE (Fig. 4A) versus those with cholesterol (Fig. 4B) as co-li-
pid indicated that formulations using cholesterol led to a modest
improvement in the protective character of the lipoplex. However,
it should be noted that a comparison between the various lipoplex
formulations within either co-lipid series did not indicate a signif-
icant difference in the ability to protect the pDNA cargo.
The cytotoxicity associated with lipoplex formulations at N:P
(+/ꢀ) molar charge ratios ranging from 0.5:1 to 10:1 was deter-
mined using the MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxy-
methoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay. CHO-
K1 cells were incubated with different formulations for 48 h at
37 °C in the presence of 5% CO2. The results illustrated in Figure 5
show a concentration-dependent cytotoxicity associated with all
lipoplex formulations.
Aside from the lipoplex 3/EPC/DOPE at a N:P molar charge ratio
of 10:1, all lipoplex formulations with DOPE as co-lipid showed
lower transfection efficiency when compared with the control lip-
oplex EPC/DOPE (Fig. 6A). Within this series of lipoplexes, those
composed of unsaturated lipids (1 or 3) revealed greater transfec-
tion than their saturated counterparts (2 or 4) regardless of the
Lipoplexes formulated with lipids 1–4, EPC and DOPE as co-lipid
revealed a general trend of higher toxicity than the control lipo-
A
B
Figure 5. Cytotoxicity of liposomes formulated with a cationic lipid, 1–4, EPC, and either DOPE (A) or cholesterol (B) as co-lipid, complexed with pDNA at various N:P (+/ꢀ)
molar charge ratios, ranging from 0.5:1 to 10:1, 48 h after transfection of CHO-K1 cells. The percentage of viable cells was calculated as the absorbance ratio of treated to
untreated cells. The data reported was the average of two experiments, each performed in triplicate, and expressed as mean S.E.

W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
4691
A
B
Figure 6. Transfection efficiency of liposomes formulated with a cationic lipid, 1–4, EPC, and either DOPE (A) or cholesterol (B) as co-lipid, complexed with pDNA at various
N:P (+/ꢀ) molar charge ratios, ranging from 0.5:1 to 10:1, 48 h after transfection of CHO-K1 cells. Luciferin luminescence for the pDNA treated cells was normalized to total
protein. The data reported was the average of two experiments, each performed in triplicate, and expressed as mean S.E. ⁄p <0.05, ⁄⁄p <0.01 (Student’s t test).
acyclic or cyclic nature of the hydrophobic domain. This result was
consistent with similar reports found in the literature.26–28 In gen-
eral, optimum efficiency was observed at a N:P molar charge ratio
of 1:1 for all lipids 1-4 formulated with DOPE as co-lipid.
Lipoplex formulations containing lipids 1–4 formulated with
cholesterol as co-lipid revealed, in most cases, a similar level of
transfection when compared with the control lipoplex EPC/Chol
(Fig. 6B). Three of the lipoplex formulations composed of the mac-
rocyclic cationic lipids 3 and 4 performed significantly better than
the control, as illustrated by the red correlation lines. Interestingly,
formulations containing the cationic acyclic lipids 1 and 2 showed
optimal transfection efficiency at a N:P molar charge ratio of 3:1,
while those containing the macrocyclic lipids 3 and 4 exhibited a
general concentration dependence up to a N:P molar charge ratio
of 10:1. In general, the lipoplexes formulated with the lipids 1–4,
EPC and cholesterol as co-lipid exhibited higher transfection effi-
ciency than the corresponding formulations containing DOPE as
co-lipid.
For N:P (+/ꢀ) molar charge ratios of 5:1 and 10:1, lipoplexes
formulated with the macrocyclic compounds 3 or 4, EPC, and cho-
lesterol as co-lipid revealed a trend towards better transfection
than lipoplexes composed of their acyclic counterparts 1 or 2.
Among the formulations included in the present study, the lipoplex
composed of the unsaturated macrocyclic cationic lipid 3 exhibited
the highest transfection efficiency at the N:P molar charge ratio of
10:1.
this enhancement of transfection may be due to an improvement
of pDNA delivery inside the cell. Indeed, it is well documented that
transfection using mixed cationic lipids resulting in the liposomal
incorporation of aliphatic chains of varying lengths, as is the case
with a mixture of EPC and cationic lipids 1–4, can improve trans-
fection efficiency potentially by promoting endosomal escape.29–
MacDonald and co-workers32 have demonstrated that when
31
mixtures of cationic lipids, particularly with significantly different
structural properties, are used together, the system is no longer in
a single fully miscible phase. That is, the lipids pack poorly and
thus constitute a defect, resulting in fusion with bilayers of the
endosome membrane, a key step for DNA release from the lipo-
plexes.33,34 This may be the case in the current study where the
length of the hydrophobic domain of our novel lipids, 1–4, differs
from that of EPC. This effect is pronounced with cholesterol and
particularly with the cyclic compounds that have a more rigid
hydrophobic domain enabling the formation of such structural de-
fects, and may result in a more effective release of the DNA, and
ultimately better transfection efficiency. This hypothesis will be
evaluated through further studies on the next generation of cyclic
compounds whose synthesis is already in progress.
We report for the first time the successful application of cat-
ionic lipids containing a macrocyclic hydrophobic domain as gene
delivery vectors. Four novel cationic lipids with short hydrophobic
domains, including two acyclic (1 and 2) and two macrocyclic com-
pounds (3 and 4), were designed and synthesized. In association
with EPC and either DOPE or cholesterol as co-lipid the synthesized
lipids 1–4 achieved efficient plasmid DNA complexation and pro-
tection from DNase I degradation, with no significant difference
Taken into consideration that the lipoplexes composed of 3 or 4
did not show better association with pDNA or better protection
against DNase I than their acyclic counterparts 1 or 2, it is possible

4692
W. P. D. Goldring et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4686–4692
14. Miramon, M.-L.; Mignet, N.; Herscovici, J. J. Org. Chem. 2004, 69, 6949.
15. Shim, G. Y.; Kim, S. H.; Han, S.-E.; Kim, Y. B.; Oh, Y.-K. Asian J. Pharm. Sci. 2009, 4,
207.
16. Huang, Q.-D.; Chen, H.; Zhou, L.-H.; Huang, J.; Wu, J.; Yu, X.-Q. Chem. Biol. Drug
Des. 2008, 71, 224.
17. Mukthavaram, R.; Marepally, S.; Venkata, M. Y.; Vegi, G. N.; Sistla, R.;
Chaudhuri, A. Biomaterials 2009, 30, 2369.
18. (a) Horiuchi, S.; Aoyama, Y. J. Control. Release 2006, 116, 107; (b) Sansone, F.;
observed between the cationic lipids. All lipoplex formulations
generally exhibited molar charge ratio dependent cytotoxicity.
The formulations which contained cholesterol generally exhibited
marginally to significantly higher cell viability and significantly
higher in vitro transfection into CHO-K1 cells than those formu-
lated with DOPE as co-lipid. In some cases the observed transfec-
ˇ
Dudic, M.; Donofrio, G.; Rivetti, C.; Baldini, L.; Casnati, A.; Cellai, S.; Ungaro, R. J.
tion efficiency was greater than the EPC/Chol control,
a
Am. Chem. Soc. 2006, 128, 14528; (c) Lalor, R.; DiGesso, J. L.; Mueller, A.;
Matthews, S. E. Chem. Commun. 2007, 4907; (d) Bagnacani, V.; Sansone, F.;
Donofrio, G.; Baldini, L.; Casnati, A.; Ungaro, R. Org. Lett. 2008, 10, 3953.
19. Srinivasachari, S.; Fichter, K. M.; Reineke, T. M. J. Am. Chem. Soc. 2008, 130,
4618.
formulation widely used for gene transfer. Within the cholesterol
formulation series, the highest transfection efficiency was achieved
at N:P 3:1 for lipoplexes composed of the acyclic lipids, 1 or 2, and
at N:P 10:1 for the macrocyclic lipids, 3 or 4. However, were the
macrocyclic lipids 3 or 4 to be employed for gene therapy at this
molar charge ratio (N:P 10:1) it would be at the expense of low cell
viability. In this context, the matched characteristics of high cell
viability (approximately >80%) and high transfection efficiency
was achieved at a N:P molar charge ratio of 3:1 for all lipoplexes
formulated with a cationic lipid (1-4), EPC and cholesterol.
The results of these studies are very encouraging and support
our hypothesis that macrocyclic lipids possess interesting proper-
ties as non-viral gene delivery vectors, including enhanced trans-
fection efficiency. Further work, including the optimization of the
lipid synthetic route, an investigation of different hydrophobic do-
mains and headgroups, as well as an investigation of the intracel-
lular trafficking of these lipoplexes are in progress and will be
disclosed in due course.
20. Liu, H.; Olsen, C. E.; Christensen, S. B. J. Nat. Prod. 2004, 67, 1439.
21. Selected data recorded for the amine lipids 1–4:
3-((5-(Dimethylamino)pentanoyl)oxy)propane-1,2-diyl bis(hept-6-enoate) (1):
mmax (film)/cmꢀ1 3072, 2963, 2919, 2855, 1739, 1606, 1462, 1415, 1261,
1072 and 802; dH (300 MHz, CDCl3) 5.85–5.72 (2H, m, H2C@CHCH2), 5.30–5.23
(1H, m, OCH), 5.04–4.93 (4H, m, H2C@CHCH2), 4.33–4.27 (2H, m, OCH2), 4.14
(1H, dd, J 11.9 and 6.0, OCHH), 4.13 (1H, dd, J 11.9 and 6.0, OCHH), 2.40–2.27
(8H, m, C(O)CH2 and CH2N(CH3)2), 2.30 (6H, s, CH2N(CH3)2), 2.10-2.03 (4H, m,
H2C@CHCH2), 1.68–1.58 (2H, m, CH2), 1.47–1.39 (2H, m, CH2), 1.32–1.20 (8H,
m, CH2); dC (100 MHz, CDCl3) 173.13, 173.09, 172.7, 138.3, 114.8, 68.9, 62.2,
62.1, 59.0, 45.3, 34.0, 33.8, 32.8, 30.1, 30.0, 29.3, 28.2, 27.0, 24.3, 24.2, 22.7; m/z
(ES) 440.3013 (M++H, 100%, C24H42O6N requires 440.3012).
3-((5-(Dimethylamino)pentanoyl)oxy)propane-1,2-diyl diheptanoate (2): mmax
(film)/cmꢀ1 2961, 2927, 2856, 1734, 1456, 1415, 1261, 1098 and 800; dH
(300 MHz, CDCl3) 5.30–5.23 (1H, m, OCH), 4.30 (1H, dd, J 11.9 and 4.3, OCHH),
4.29 (1H, dd, J 11.9 and 4.3, OCHH), 4.15 (1H, d, J 11.9, OCHH), 4.13 (1H, d, J
11.9, OCHH), 2.39-2.26 (8H, m, C(O)CH2 and CH2N(CH3)2), 2.27 (6H, s,
CH2N(CH3)2), 1.67–1.50 (8H, m, CH2), 1.34–1.21 (12H, m, CH2), 0.90–0.83
(6H, m, CH3); dC (100 MHz, CDCl3) 173.3, 173.0, 172.9, 68.8, 62.2, 62.1, 59.2,
45.3, 34.2, 34.0, 33.8, 32.8, 32.7, 30.0, 29.4, 28.75, 28.70, 27.1, 25.6, 22.7, 22.4,
14.1, 14.0; m/z (ES) 444.3336 (M++H, 100%, C24H46O6N requires 444.3325).
Acknowledgments
(E)-
and
(Z)-(5,16-Dioxo-1,4-dioxacyclohexadec-10-en-2-yl)methyl
5-
(dimethylamino)-pentanoate (3): Data for the major (E)-isomer; mmax (film)/
cmꢀ1 2961, 2928, 2857, 1732, 1635, 1461, 1262, 1099, 1016 and 808; dH
(300 MHz, CDCl3) 5.31-5.21 (3H, m, OCH and CH@CH), 4.31–4.11 (4H, m,
OCH2), 2.37–2.26 (8H, m, C(O)CH2 and CH2N(CH3)2), 2.25 (6H, s, CH2N(CH3)2),
2.10–2.02 (4H, m, C@CHCH2), 1.71–1.38 (8H, m, CH2), 1.36–1.21 (4H, m, CH2);
dC (100 MHz, CDCl3) 173.2, 172.9, 172.8, 131.0, 130.9, 68.9, 62.6, 62.1, 59.1,
45.2, 33.8, 32.0, 31.9, 30.3, 30.1, 29.4, 28.04, 27.99, 26.8, 24.8, 22.7; m/z (ES)
412.2701 (M++H, 100%, C22H38O6N requires 412.2699).
The authors thank the British Council (PMI2 Gulf States Cooper-
ation Grant No. RCGS206) and the Biomedical Research Program
intramural funding at WCMC-Q for support of this research.
Supplementary data
(5,16-Dioxo-1,4-dioxacyclohexadecan-2-yl)methyl 5-(dimethylamino)pentanoate
(4): mmax (film)/cmꢀ1 2927, 2856, 1741, 1457, 1260, 1129 and 735; dH
(300 MHz, CDCl3) 5.32–5.26 (1H, m, OCH), 4.39–4.26 (2H, m, OCH2), 4.21–
4.12 (2H, m, OCH2), 2.39–2.31 (8H, m, C(O)CH2 and CH2N(CH3)2), 2.29 (6H, s,
CH2N(CH3)2), 1.94–1.60 (8H, m, CH2), 1.35–1.29 (12H, m, CH2); dC (100 MHz,
CDCl3) 173.5, 173.05, 173.00, 69.1, 62.8, 62.1, 59.2, 45.4, 33.8, 33.72, 33.69,
31.9, 30.0, 29.4, 27.1, 27.0, 26.1, 26.02, 25.95, 25.90, 22.7; m/z (ES) 414.2853
(M++H, 100%, C22H40O6N requires 414.2856).
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.05.
080.
References and notes
22. For a related application of molecular mechanics for the calculation of E/Z
ratios of RCM alkene isomer products, see: Goldring, W. P. D.; Hodder, A. S.;
Weiler, L. Tetrahedron Lett. 1998, 39, 4955.
23. Hirko, A.; Tang, F.; Hughes, J. A. Curr. Med. Chem. 2003, 10, 1185.
24. Zhang, Y.-P.; Reimer, D. L.; Zhang, G.; Lee, P. H.; Bally, M. B. Pharm. Res. 1997,
14, 190.
25. Gaucheron, J.; Wong, T.; Wong, K. F.; Maurer, N.; Cullis, P. R. Bioconjugate Chem.
2002, 13, 671.
26. Ferrari, M. E.; Rusalov, D.; Enas, J.; Wheeler, C. J. Nucleic Acids Res. 2002, 30,
1808.
1. Both, G.; Alexander, I.; Fletcher, S.; Nicolson, T. J.; Rasko, J. E. J.; Wilton, S. D.;
Symonds, G. Pathology 2011, 43, 642.
2. Chesnoy, S.; Huang, L. Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 27.
3. Zhi, D.; Zhang, S.; Wang, B.; Zhao, Y.; Yang, B.; Yu, S. Bioconjugate Chem. 2010,
21, 563.
4. Liu, Y.; Liggitt, D.; Zhong, W.; Tu, G.; Gaensler, K.; Debs, R. J. Biol. Chem. 1995,
270, 24864.
5. Stewart, M. J.; Plautz, G. E.; Del Buono, L.; Yang, Z. Y.; Xu, L.; Gao, X.; Huang, L.;
Nabel, E. G.; Nabel, G. J. Hum. Gene Ther. 1992, 3, 267.
6. Zhu, N.; Liggitt, D.; Liu, Y.; Debs, R. Science 1993, 261, 209.
7. Thierry, A. R.; Lunardi-Iskandar, Y.; Bryant, J. L.; Rabinovich, P.; Gallo, R. C.;
Mahan, L. C. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 9742.
8. Popplewell, L. J.; Abu-Dayya, A.; Khanna, T.; Flinterman, M.; Abdul Khalique, N.;
Raju, L.; Øpstad, C. L.; Sliwka, H.-R.; Partali, V.; Leopold, P. L.; Sliwka, H.-R.;
Partali, V.; Dickson, G.; Pungente, M. D. Molecules 2012, 17, 1138.
9. Pungente, M. D.; Jubeli, E.; Øpstad, C. L.; Al-Kawaz, M.; Barakat, N.; Ibrahim, T.;
Abdul Khalique, N.; Raju, L.; Jones, R.; Leopold, P. L.; Sliwka, H.-R.; Partali, V.
Molecules 2012, 17, 3484.
10. Gliozzi, A.; Relini, A.; Chong, P. L.-G. J. Membr. Sci. 2002, 206, 131.
11. Patel, G. B.; Sprott, G. D. Crit. Rev. Biotechnol. 1999, 19, 317.
12. Dannenmuller, O.; Arakawa, K.; Eguchi, T.; Kakinuma, K.; Blanc, S.; Albrecht, A.-
M.; Schmutz, M.; Nakatani, Y.; Ourisson, G. Chem. Eur. J. 2000, 6, 645.
13. Réthoré, G.; Montier, T.; Le Gall, T.; Delépine, P.; Cammas-Marion, S.; Lemiègre,
L.; Lehn, P.; Benvegnu, T. Chem. Commun. 2007, 2054.
27. Zuhorn, I. S.; Oberle, V.; Visser, W. H.; Engberts, J. B. F. N.; Bakowsky, U.;
Polushkin, E.; Hoekstra, D. Biophys. J. 2002, 83, 2096.
28. Koynova, R.; Wang, L.; MacDonald, R. C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
14373.
29. Felgner, P. L.; Ringold, G. M. Nature 1989, 337, 387.
30. Tang, F.; Hughes, J. A. J. Control. Release 1999, 62, 345.
31. Rosenzweig, H. S.; Rakhmanova, V. A.; McIntosh, T. J.; MacDonald, R. C.
Bioconjugate Chem. 2000, 11, 306.
32. Wang, L.; Koynova, R.; Parikh, H.; MacDonald, R. C. Biophys. J. 2006, 91, 3692.
33. Zhou, X.; Huang, L. Biochim. Biophys. Acta 1994, 1189, 195.
34. Friend, D. S.; Papahadjopoulos, D.; Debs, R. J. Biochim. Biophys. Acta 1996, 1278,
41.