A columnar phase of dendritic lipid-based cationic liposome-DNA complexes for gene delivery: Hexagonally ordered cylindrical micelles embedded in a DNA honeycomb lattice

Article abstract of DOI:10.1021/ja055907h

Gene therapy holds great promise as a future approach to fighting disease and is explored in worldwide clinical trials. Cationic liposome (CL)-DNA complexes are a prevalent nonviral delivery vector, but their efficiency requires improvement and the unders

Full text of DOI:10.1021/ja055907h

Published on Web 03/08/2006
A Columnar Phase of Dendritic Lipid-Based Cationic
Liposome-DNA Complexes for Gene Delivery: Hexagonally
Ordered Cylindrical Micelles Embedded in a DNA Honeycomb
Lattice
Kai K. Ewert, Heather M. Evans,† Alexandra Zidovska, Nathan F. Bouxsein,
Ayesha Ahmad,§ and Cyrus R. Safinya*
Contribution from the Department of Materials, Department of Physics, and Molecular,
Cellular and DeVelopmental Biology Department, UniVersity of California,
Santa Barbara, California 93106
Received August 26, 2005; E-mail: safinya@mrl.ucsb.edu
Abstract: Gene therapy holds great promise as a future approach to fighting disease and is explored in
worldwide clinical trials. Cationic liposome (CL)-DNA complexes are a prevalent nonviral delivery vector,
but their efficiency requires improvement and the understanding of their mechanism of action is incomplete.
As part of our effort to investigate the structure-transfection efficiency relationships of self-assembled CL-
DNA vectors, we have synthesized a new, highly charged (16+) multivalent cationic lipid, MVLBG2, with
a dendritic headgroup. Our synthetic scheme allows facile variation of the headgroup charge and the spacer
connecting hydrophobic and headgroup moieties as well as gram-scale synthesis. Complexes of DNA
with mixtures of MVLBG2 and neutral 1,2-dioleoyl-sn-glycerophosphatidylcholine (DOPC) exhibit the well-
known lamellar phase at 90 mol % DOPC. Starting at 20 mol % dendritic lipid, however, two novel
nonlamellar phases are observed by synchrotron X-ray diffraction. The structure of one of these phases,
present in a narrow range of composition around 25 mol % MVLBG2, has been solved. In this novel dual
lattice structure, termed HIC, hexagonally arranged tubular lipid micelles are surrounded by DNA rods forming
a three-dimensionally continuous substructure with honeycomb symmetry. Complexes in the HIC phase
efficiently transfect mouse and human cells in culture. Their transfection efficiency, as well as that of the
lamellar complexes containing only 10 mol % dendritic lipid, reaches and surpasses that of commercially
available, optimized DOTAP-based complexes. In particular, complexes containing MVLBG2 are significantly
more transfectant over the entire composition range in mouse embryonic fibroblasts, a cell line empirically
known to be hard to transfect.
gene silencing.6-9 Thus, substantial research efforts are directed
toward developing and fundamentally understanding DNA
carriers (vectors). Engineered viruses are very efficient vec-
tors,10,11 but general concerns about their safety have recently
been intensified by a few severe setbacks.12 This has further
increased the interest in nonviral or synthetic vectors,13-15 in
which the negatively charged DNA is complexed with cationic
Introduction
Somatic gene therapy, i.e., the use of DNA as a therapeutic
agent, holds great promise for future medical applications.
Indeed, numerous clinical trials in this field are currently
ongoing.1,2 Cancers, inherited diseases, cardiovascular diseases,
and many others are targets for this novel medical approach.3,4
In fact, the therapeutic prospects of nucleic acid delivery are
expanding constantly5 due to recent discoveries such as that of
RNA interference (RNAi), which enables, in principle, selective
(5) Mahato, R. I., Kim, S. W., Eds. Pharmaceutical PerspectiVes of Nucleic
Acid-Based Therapeutics; Taylor and Francis: London and New York,
2002.
(6) McManus, M. T.; Sharp, P. A. Nat. ReV. Genet. 2002, 3, 737-747.
(7) Caplen, N. J. Gene Ther. 2004, 11, 1241-1248.
† Current address: Max Planck Institute for Dynamics and Self-
Organization, Bunsenstrasse 10, D-37073 Go¨ttingen, Germany.
§ Current address: Imperial College Genetic Therapies Centre, Depart-
ment of Chemistry, Flowers Building, Armstrong Road, Imperial College
London, London, SW7 2AZ, U.K.
(1) Extensive and current information on clinical trials in the field of gene
therapy can be found on the Internet at http://www.wiley.co.uk/genetherapy/
clinical/. See Ewert, K.; Ahmad, A.; Evans, H. M.; Safinya, C. R. Expert
Opin. Biol. Ther. 2005, 5, 33-53 for a compilation of open trials in nonviral
gene therapy as of July 2004.
(2) Griesenbach, U.; Geddes, D. M.; Alton, E. W. F. W. In NonViral Vectors
for Gene Therapy; Huang, L., Hung, M.-C., Wagner, E., Eds.; Academic
Press: San Diego, 1999; pp 337-356.
(3) Felgner, P. L.; Rhodes, G. Nature 1991, 349, 351-352.
(4) Smyth-Templeton, N., Lasic, D. D., Eds. Gene Therapy. Therapeutic
Mechanisms and Strategies; Marcel Dekker Inc.: New York, 2000.
(8) Milner, J. Expert Opin. Biol. Ther. 2003, 3, 459-467.
(9) Caplen, N. J. Expert Opin. Biol. Ther. 2003, 3, 575-586.
(10) Smith, A. E. Annu. ReV. Microbiol. 1995, 49, 807-838.
(11) Kay, M. A.; Glorioso, J. C.; Naldini, L. Nature Med. 2001, 7, 33-40.
(12) (a) Marshall, E. Science 2000, 288, 951-957. (b) Marshall, E. Science
2002, 298, 510-511.
(13) Huang, L., Hung, M.-C., Wagner, E., Eds. Non-Viral Vectors for Gene
Therapy, 2nd ed., Part I; Advances in Genetics, Vol. 53; Elsevier: San
Diego, 2005.
(14) Huang, L., Hung, M.-C., Wagner, E., Eds.; NonViral Vectors for Gene
Therapy; Academic Press: San Diego, 1999.
(15) Felgner, P. L., Heller, M. J., Lehn, P., Behr, J.-P., Szoka, F. C., Eds.;
Artificial Self-Assembling Systems for Gene DeliVery; American Chemical
Society: Washington, DC, 1996.
9
3998
J. AM. CHEM. SOC. 2006, 128, 3998-4006
10.1021/ja055907h CCC: $33.50 © 2006 American Chemical Society

Gene Delivery: Honeycomb Phase of Lipid−DNA Complexes
A R T I C L E S
liposomes (CLs)16-19 or cationic polyelectrolytes.20-22 Synthetic
vectors have several advantages over viral methods, such as
facile and variable preparation, absence of immunogenic protein
components, and unlimited length of the transported DNA. A
large number of cationic and other lipids for use in gene delivery
have been synthesized. The earlier part of these efforts has been
reviewed in detail,23-25 and several more recent reviews with a
smaller scope have also appeared.26-28 Several lipid vectors have
been commercialized, and many are able to efficiently transfect
cultured mammalian cells. For CL-DNA complexes to become
widely useful for gene therapeutic purposes, however, their
efficiency needs to be improved.29
The improvement of CL-DNA vectors requires a better
understanding of their mechanism of transfection, and the
chemical and physical parameters of CL-DNA complexes
influencing it. Thus, a large amount of work has been performed
in order to understand the formation and structure of CL-DNA
complexes, and to establish correlations between the structures
and physicochemical characteristics of the complexes and their
transfection efficiency and mechanism of action.30,31 Thus far,
Figure 1. Schematic depictions of the previously known equilibrium phases
of CL-DNA complexes, showing the local structure of their interior on
the nanometer scale. Neutral and cationic lipids are depicted as having white
and gray headgroups, respectively. The lamellar phase, LRC, consisting of
alternating lipid bilayers and DNA monolayers with an interlayer spacing
C
of d ) δw + δm is depicted on the left. The inverted hexagonal phase HII
,
which is comprised of DNA rods coated with a lipid monolayer and arranged
on a hexagonal lattice, is shown on the right. Adapted with permission
from ref 32.
shape and corresponding preferred curvature of a lipid deter-
mines the structure of its DNA complexes. This is evident by
two equilibrium structures of CL-DNA complexes have been
C
the tendency of DOPE to promote the HII phase, due to its
C 32
found: the inverted hexagonal (HII
lamellar (LR
) and the more abundant
cone-shaped molecular structure. Likewise, it seemed feasible
to induce a complex structure with positive lipid curvature by
using lipids with large headgroups (corresponding to an inverted
cone molecular shape). However, lipids with 3-5 cationic
charges in the headgroup still form LRC-DNA complexes,39-41
even though some of these lipids (without DNA) have been
shown to form spherical and wormlike micelles in solution.41
To further increase the headgroup charge and size and, thus,
possibly induce the formation of complexes containing such
micellar structures, we have synthesized a new lipid with a
highly charged, dendritic headgroup.
Dendrimers are monodisperse, highly branched synthetic
molecules, which are typically assembled by adding generations
of AB2 monomers to a core molecule.42-45 Dendrimers bearing
amino groups that can be protonated, such as poly(amidoamine)
(PAMAM) and poly(propylenimine) (PPI) dendrimers, have
been used successfully for gene delivery,46-50 and the structure
of their DNA complexes has been investigated.51 Conjugates
of dendritic molecules to PEG have also been used for gene
delivery.52
)
C 33,34 liquid crystal phase. These are schematically
shown in Figure 1. In addition, kinetically trapped structures
have been observed, e.g. by electron microscopy.35 The trans-
fection mechanisms of the two phases vary greatly as a
consequence of their structures, but hexagonal and optimized
lamellar complexes transfect equally well.36-38 In addition to
the composition and rigidity of the membrane,32 the molecular
(16) Felgner, P. L.; Gader, T. R.; Holm, M.; Roman, R.; Chan, H. W.; Wenz,
M.; Northrop, J. P.; Ringold, G. M.; Danielsen, M. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 7413-7417.
(17) Wheeler, C. J.; Felgner, P. L.; Tsai, Y. J.; Marshall, J.; Sukhu, L.; Doh, S.
G.; Hartikka, J.; Nietupski, J.; Manthorpe, M.; Nichols, M.; Plewe, M.;
Liang, X. Norman, J.; Smith, A.; Cheng, S. H. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 11454-11459.
(18) Byk, G.; Dubertret, C.; Escriou, V.; Frederic, M.; Jaslin, G.; Rangara, R.;
Pitard, B.; Crouzet, J.; Wils, P.; Schartz, B.; Scherman, D. J. Med. Chem.
1998, 41, 224-235.
(19) Remy, J.-S.; Sirlin, C.; Vierling, P.; Behr, J.-P. Bioconjugate Chem. 1994,
5, 647-654.
(20) Boussif, O.; Lezoualc’h, F.; Zanta, M. A.; Mergny, M. D.; Scherman, D.;
Demeniex, B.; Behr, J.-P. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 7297-
7301.
(21) Haensler, J.; Szoka, F. C. Bioconjugate Chem. 1993, 4, 372-379.
(22) Tang, M. X.; Redemann, C. T.; Szoka, F. C. Bioconjugate Chem. 1996, 7,
703-714.
(23) Miller, A. D. Angew. Chem., Int. Ed. 1998, 37, 1769-1785.
(24) Zabner, J. AdV. Drug DeliVery ReV. 1997, 27, 17-28.
(25) Behr, J.-P. Bioconjugate Chem. 1994, 5, 382-389.
(26) Byk, G.; Scherman, D. Drug DeV. Res. 2000, 50, 566-572.
(27) Chesnoy, S.; Huang, L. Annu. ReV. Biophys. Biomol. Struct. 2000, 29, 27-
47.
(39) Ewert, K.; Ahmad, A.; Evans, H. M.; Schmidt, H.-W.; Safinya, C. R. J.
Med. Chem. 2002, 45, 5023-5029.
(40) Ahmad, A.; Evans, H. M.; Ewert, K.; George, C. X.; Samuel, C. E.; Safinya,
C. R. J. Gene Med. 2005, 7, 739-748.
(41) Boukhnikachvili, T.; Aguerre-Chariol, O.; Airiau, M.; Lesieur, S.; Ollivon,
M.; Vacus, J. FEBS Lett. 1997, 409, 188-194.
(28) de Lima, M. C. P.; Simo˜es, S.; Pires, P.; Faneca, H.; Du¨zgu¨nes¸, N. AdV.
Drug DeliVery ReV. 2001, 47, 277-294.
(42) Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665-
(29) Miller, A. D. Curr. Med. Chem. 2003, 10, 1195-1211.
(30) Koltover, I.; Safinya, C. R. In NonViral Vectors for Gene Therapy; Huang,
L., Hung, M.-C., Wagner, E., Eds.; Academic Press: San Diego, 1999; pp
91-117.
1688.
(43) Fischer, M.; Vo¨gtle, F. Angew. Chem., Int. Ed. 1999, 38, 884-905.
(44) Frechet, J. M. J., Tomalia, D. A., Eds. Dendrimers and Other Dendritic
Polymers; Wiley: Chichester, 2001.
(31) Safinya, C. R.; Lin, A. J.; Slack, N. L.; Koltover, I. In Pharmaceutical
PerspectiVes of Nucleic Acid-Based Therapeutics; Mahato, R. I., Kim, S.
W., Eds.; Taylor and Francis: London and New York, 2002; pp 190-
210.
(45) Tomalia, D. A. Sci. Am. 1995, 272, 62-66.
(46) Tang, M. X.; Redemann, C. T.; Szoka, F. C. Bioconjugate Chem. 1996, 7,
703-714.
(47) Kukowaska-Latallo, J. F.; Bielinska, A. U.; Johnson, J.; Spindler, R.;
Tomalia, D. A.; Baker, J. R. Jr. Proc. Natl. Acad. Sci. U.S.A. 1996, 93,
4897-4902.
(32) Koltover, I.; Salditt, T.; Ra¨dler, J. O.; Safinya, C. R. Science 1998, 281,
78-81.
(33) Ra¨dler, J. O.; Koltover, I.; Salditt, T.; Safinya, C. R. Science 1997, 275,
810-814.
(34) Koltover, I.; Salditt, T.; Safinya, C. R. Biophys. J. 1999, 77, 915-924.
(35) Smyth Templeton, N.; Lasic, D. D.; Frederik, P. M.; Strey, H. H.; Roberts,
D. D.; Pavlakis, G. N. Nat. Biotechnol. 1997, 15, 647-652.
(36) Lin, A. J.; Slack, N. L.; Ahmad, A.; Koltover, I.; George, C. X.; Samuel,
C. E.; Safinya, C. R. J. Drug Targeting 2000, 8, 13-27.
(37) Ewert, K.; Slack, N. L.; Ahmad, A.; Evans, H. M.; Lin, A. J.; Samuel, C.
E.; Safinya, C. R. Curr. Med. Chem. 2004, 11, 133-149.
(38) Ewert, K.; Ahmad, A.; Evans, H. M.; Safinya, C. R. Expert Opin. Biol.
Ther. 2005, 5, 33-53.
(48) Kukowska-Latallo, J. F.; Raczka, E.; Quintana, A.; Chen, C. L.; Rymaszew-
ski, M.; Baker, J. R., Jr. Human Gene Ther. 2000, 11, 1385-1395.
(49) Rudolph, C.; Lausier, J.; Naundorf, S.; Mu¨ller, R. H.; Rosenecker J. J.
Gene Med. 2000, 2, 269-278.
(50) Ohsaki, M.; Okuda, T.; Wada, A.; Hirayama, T.; Niidome, T.; Aoyagi, H.
Bioconjugate Chem. 2002, 13, 510-517.
(51) Evans, H. M.; Ahmad, A.; Ewert, K.; Pfohl, T.; Martin-Herranz, A.;
Bruinsma, R. F.; Safinya, C. R. Phys. ReV. Lett. 2003, 91, 075501-1-
075501-4.
(52) Kim, T. I.; Seo, H. J.; Choi, J. S.; Jang, H. S.; Baek, J. U.; Kim, K.; Park,
J. S. Biomacromolecules 2004, 5, 2487-2492.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 3999

A R T I C L E S
Ewert et al.
Figure 2. Synthesis of the dendritic lipid MVLBG2, starting from ornithine methyl ester. The hydrophobic DOB moiety, ethylenediamine spacer, ornithine
headgroup core, and carboxyspermine endgroups are underlaid in yellow, blue, green, and red, respectively. Reaction conditions: i: Boc-L-ornithine(Boc),
TBTU, HOBt, DIEA; ii: TFA; iii: R-COOH (7), TBTU, HOBt, DIEA; iv: excess ethylenediamine (20×) in MeOH. HOBt: 1-hydroxybenzotriazole hydrate;
DIEA: N,N-diisopropylethylamine; TBTU: O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium tetrafluoroborate; TFA: trifluoroacetic acid.
In this paper, we report the synthesis of a new multivalent
cationic lipid, MVLBG2, with a dendritic headgroup. This
headgroup has a core based on ornithine that is terminated by
carboxyspermine moieties. The headgroup carries 16 positive
charges at complete protonation, expanding the range of
membrane charge density accessible by our experiments and
approaching the realm of lipid-polymer and lipid-dendrimer
hybrid materials, which have been reported occasionally.53-57
In addition, this lipid can be viewed as an extension of the
concept of multivalent lipids,18,58,59 since few lipids with more
than five charges per headgroup have been reported. Our
synthetic scheme allows for gram-scale synthesis and facile
variation of headgroup charge and the spacer connecting
hydrophobic and headgroup moieties. We have prepared DNA
complexes from mixtures of MVLBG2 with DOPC to inves-
tigate their structures using X-ray diffraction and measure their
transfection efficiency with a luciferase reporter gene assay.
X-ray diffraction shows that the complexes form the LRC phase33
at low contents of dendritic lipid in the membrane. For higher
contents of the highly charged MVLBG2, two nonlamellar
phases are observed. The structure of one of these, present in a
narrow range of composition around 25 mol % MVLBG2, has
been solved. This phase, termed HIC and consisting of a
hexagonal array of tubular micelles, constitutes a previously
unknown, third phase of CL-DNA complexes. DNA rods are
intercalated in the interstices to form a three-dimensionally
contiguous substructure with honeycomb symmetry, as opposed
to the isolated DNA rods and sheets of parallelly arranged
strands observed in the HII and LR phases (Figure 1),
respectively. Importantly, lipid-DNA complexes in the new
phase show high transfection efficiency in murine and human
cells, with a significant improvement (about 1 order of
magnitude) over the control DOTAP complexes observed in
mouse embryonic fibroblasts, which are known to be hard to
transfect.
C
C
Results and Discussion
Lipid Design and Synthesis. Figure 2 shows the structure
and synthesis of MVLBG2. The 3,4-dioleyloxybenzoic acid
(DOB) hydrophobic moiety39,60 (underlaid in yellow in Figure
2) provides stable membrane anchoring without side chain
crystallization, as well as miscibility with DOPC and DOPE.
The headgroup is based on the amino acid ornithine, which is
used as an AB2 branching building block. Terminating a core
made up of three (1+2) ornithine molecules (underlaid in green
in Figure 2) with carboxyspermine61,62 as end groups (underlaid
in red in Figure 2) yields eight primary and eight secondary
amino groups in the headgroup. An ethylendiamine linker
(underlaid in blue in Figure 2) connects headgroup and DOB.
(53) Dewa, T.; Ieda, Y.; Morita, K.; Wang, L.; MacDonald, R. C.; Iida, K.;
Yamashita, K.; Oku, N.; Nango, M. Bioconjugate Chem. 2004, 15, 824-
830.
(54) Sugiyama, M.; Matsuura, M.; Takeuchi, Y.; Kosaka, J.; Nango, M.; Oku,
N. Biochim. Biophys. Acta 2004, 1660, 24-30.
(55) Usol’tseva, N.; Bykova, V.; Smirnova, A.; Grusdev, M.; Lattermann, G.;
Facher, A. Mol. Cryst. Liquid Cryst. 2004, 409, 29-42.
(56) Grohn, F.; Bauer, B. J.; Amis, E. J. Macromolecules 2001, 34, 6701-
6707.
(57) Schenning, A. P. H. J.; Elissenroman, C.; Weener, J. W.; Baars, M. W. P.
L.; Vandergaast, S. J.; Meijer, E. W. J. Am. Chem. Soc. 1998, 120, 8199-
8208.
(58) Cooper, R. G.; Etheridge, C. J.; Stewart, L.; Marshall, J.; Rudginsky, S.;
Cheng, S. H.; Miller, A. D. Chem. Eur. J. 1998, 4, 137-151.
(59) Behr, J.-P.; Demeneix, B.; Leffler, J.-P.; Perez-Mutul J. Proc. Natl. Acad.
Sci. U.S.A. 1989, 86, 6982-6986.
(60) Schulze, U.; Schmidt, H.-W.; Safinya, C. R. Bioconjugate Chem. 1999,
10, 548-552.
(61) Behr, J.-P. J. Chem. Soc., Chem. Commun. 1989, 101-103.
(62) Huber, M.; Pelletier, J. G.; Torossian, K.; Dionne, P.; Gamache, I.; Charest-
Gaudreault, R.; Audette, M.; Poulin, R. J. Biol. Chem. 1996, 271, 27556-
27563.
9
4000 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Gene Delivery: Honeycomb Phase of Lipid−DNA Complexes
A R T I C L E S
Figure 4. Optical microscopy images of DNA complexes containing (A)
10 mol % and (B) 25 mol % MVLBG2. Complexes were imaged in water
using DIC (left), DNA fluorescence (middle), and lipid fluorescence (right)
modes. The scale bars correspond to 2 µm. The observation of colocalization
of DNA and lipid in the micrometer-scale particles provides direct evidence
of the existence of a complex phase at both compositions.
Figure 3. Data from an ethidium bromide (EtBr) displacement assay, used
to measure the charge of MVLBG2 in complexes with DNA. Intercalated
ethidium bromide is expelled from DNA upon complex formation with
cationic lipids, resulting in a reduction in fluorescence intensity due to self-
quenching of EtBr in solution. The fluorescence intensity is normalized to
that of DNA with EtBr and no lipid. The lines serve as a guide to the eye.
containing 10 and 25 mol % MVLBG2 in the membrane,
respectively, are shown. The complexes were prepared in water
and doubly labeled with fluorescent DNA and lipid stains. Each
complex is imaged in DIC mode (left), DNA fluorescence mode
(middle), and lipid fluorescence mode (right). The complex
particles, at both lipid compositions, have dimensions on the
µm scale, apparently consisting of aggregates of globular
submicrometer particles. They are highly mobile, as evident
from the drift and rotation of the particles seen in the images.
The colocalization observed in the fluorescence imaging modes
provides direct evidence of the existence of a phase containing
lipid and DNA, a key premise in the analysis of the X-ray
scattering data presented below. X-ray diffraction then shows
that the compositions imaged here have a very different internal
complex structure (see below). In this respect, it is interesting
The synthesis of the dendritic headgroup is similar to that of
multiple antigenic peptides (MAPs)63,64 and poly(ethylene
glycol) dendritic oligolysine block copolymers.65 Ornithine
methyl ester was first coupled with two (2.2) equivalents of
Boc-protected ornithine. After deprotection of the amino groups,
a similar coupling was performed using Boc-protected carboxy-
spermine.61,62 Aminolysis of the methyl ester with an excess of
ethylenediamine yielded the amine-functionalized spacer-head-
group building block 3. After coupling with DOB,60,39 the
resulting Boc-protected lipid was purified by flash chromato-
graphy and deprotected with TFA. This synthetic route has been
carried out on a gram scale and allows facile variation of the
spacer length and the headgroup charge by employing other
diamines and multivalent headgroup building blocks, respec-
tively.
C
to note that small aggregates of complex particles in the HII
phase predominantly show branched morphologies32 similar to
that seen in Figure 4B.
Ethidium Bromide Displacement Assay. To determine the
number of charges per headgroup effective in DNA complex-
ation, we used an ethidium bromide (EtBr) displacement
assay.40,66-68 Increasing amounts of lipid were added to a
constant amount of DNA and EtBr, recording the resulting
fluorescence and normalizing it to the fluorescence without
added lipid. Complex formation of DNA with lipids expels
intercalated EtBr, reducing its fluorescence since EtBr self-
quenches in solution. As shown in Figure 3, the observed
fluorescence drops steeply with increasing amount of added lipid
and then levels off to a saturation value. The isoelectric point
is obtained from analysis of these data,40,69,70 yielding the lipid
headgroup charge as Zexp ) 15.9 ( 1.0. Thus, the headgroup is
essentially fully protonated in the complexes.
X-ray Diffraction. The structure of DNA complexes of
MVLBG2 on the nanometer scale was investigated by small-
angle X-ray scattering (SAXS). All SAXS experiments were
performed in DMEM, the medium used in the transfection
experiments. Complexes were formed from mixtures of
MVLBG2 and neutral DOPC at a lipid-to-DNA charge ratio of
2.8, placing them in the transfection-relevant regime of cat-
ionically charged complexes, where complexes and excess
cationic liposomes coexist. While their structure changes
significantly with the ratio of neutral to cationic lipid (see
below), it is not affected by the lipid-to-DNA charge ratio (Fchg
)
above the isoelectric point (data not shown).
Optical Microscopy. DNA complexes of MVLBG2 were
imaged directly using differential interference contrast (DIC)
and fluorescence microscopy. In Figure 4A and B, complexes
As evident from the SAXS pattern shown in Figure 5,
complexes containing 10 mol % MVLBG2 form the lamellar
phase LRC. The peaks (solid arrows) at 0.087, 0.173, and 0.260
Å-1 correspond to the q001, q002, and q003 reflections from stacks
of the lipid bilayers, indicating a layer repeat distance δ ) 2π/
q001 of 72.2 Å. A broader peak, arising from smectic ordering
of DNA strands is also visible (qDNA, dashed arrow) and gives
their interaxial spacing dDNA ) 2π/qDNA ) 2π/(0.227 Å-1) )
27.7 Å.
At higher mole fractions of MVLBG2 (g20%), the scattering
pattern changes significantly. A novel structure, which we
describe below, is formed in a narrow range of composition of
a few mol % around 25 mol % MVLBG2. At higher contents
(63) Veprek, P.; Jezek, J. J. Pept. Sci. 1999, 5, 5-23.
(64) Veprek, P.; Jezek, J. J. Pept. Sci. 1999, 5, 203-220.
(65) Choi, J. S.; Lee, E. J.; Choi, Y. H.; Jeong, Y. J.; Park, J. S. Bioconjugate
Chem. 1999, 10, 62-65.
(66) Gershon, H.; Ghirlando, R.; Guttman, S. B.; Minsky, A. Biochemistry 1993,
32, 7143-51.
(67) Akao, T.; Fukumoto, A.; Ihara, H.; Ito, A. FEBS Lett. 1996, 391, 215-
218.
(68) Eastman, S. J.; Siegel, C.; Tousignant, J.; Smith, A. E.; Cheng, S. H.;
Scheule, R. K. Biochim. Biophys. Acta Biomembr. 1997, 1325, 41-62.
(69) Ham, Y. W.; Tse, W. C.; Boger, D. L. Bioorg. Med. Chem. Lett. 2003, 13,
3805-3807.
(70) Boger, D. L.; Fink, B. E.; Brunette, S. R.; Tse, W. C.; Hedrick, M. P. J.
Am. Chem. Soc. 2001, 123, 5878-5891.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4001

A R T I C L E S
Ewert et al.
Figure 6. Schematic of the new, HIC phase of CL-DNA complexes. The
large headgroup of MVLBG2 forces the lipid to assemble into tubular lipid
micelles that are arranged on a hexagonal lattice, with an interaxial distance
of 81.5 Å (see text). The diameter of the tubular micelles may be estimated
to around 41 Å as the typical thickness of a bilayer formed by lipids with
DOB hydrophobic tails. The DNA rods, with a diameter of 25-30 Å, are
arranged on a honeycomb lattice in the interstices of the lipid micelle
arrangement. The remaining interstitial space is filled by the headgroups,
water, and counterions. Interestingly, the DNA forms a three-dimensionally
contiguous substructure in the HIC phase, as opposed to isolated DNA rods
C
C
and sheets of parallel strands in the HII and LR phases (Figure 1),
respectively.
of DOTAP (2,3-dioleyloxy-propyl-trimethylammonium chlo-
ride)74 and MVLBG2 shown in Figure 7, the headgroup of
MVLBG2 is very large and may fill most of the void space. It
is the size of this headgroup that strongly favors a positive
curvature of the lipid-water interface and thus leads to the
formation of the novel phase.75 Further corroborating the
assignment of the HIC structure as opposed to the previously
Figure 5. Synchrotron SAXS patterns of DNA complexes of MVLBG2/
DOPC mixtures, containing 10 and 25 mol % of the highly charged dendritic
lipid. At 25 mol % MVLBG2, the peaks cleanly index to a hexagonal lattice,
indicating a change of structure into the new HIC phase. Data acquired on
SAXS beamline 4-2 at the Stanford Synchrotron Radiation Laboratory.
of MVLBG2, a different phase with lower symmetry is
observed.71 As shown in Figure 5 for 25 mol % MVLBG2,
peaks are now found at 0.089, 0.157, 0.177, 0.232, and 0.262
Å-1, corresponding to the q10, q11, q20, q21, and q30 reflections
of a hexagonal lattice, respectively. Figure 6 shows a schematic
of this new hexagonal CL-DNA complex structure, which we
term HIC. At 20 mol % MVLBG2, complexes in the HIC phase
are in coexistence with lamellar complexes (data not shown).
In the HIC phase, tubular lipid micelles are arranged on a
hexagonal lattice, with an interaxial distance of d ) 4π/(31/2q10)
) 81.5 Å (Figure 6). The typical thickness of a bilayer formed
by lipids with a DOB hydrophobic tail is around 41 Å, which
gives a first estimate of the diameter of a lipid tubule formed
by MVLBG2. The DNA rods, which have a diameter of 25-
30 Å, depending on the extent of hydration, are arranged on a
honeycomb lattice in the interstices of the lipid micelle
arrangement. It is remarkable that the DNA forms a three-
dimensionally contiguous substructure in the HIC phase, as
opposed to isolated DNA rods and sheets of parallel strands in
C
observed HII structure is the observation that lipid mixtures
containing MVLBG2 are soluble in water over a wide range of
composition, indicating the formation of micelles rather than
inverted micelles, which have poor water solubility. Structures
similar to the HIC phase have been proposed for complexes of
DNA with cationic detergents, but these structure assignments
were less unambiguous since they are based on fewer reflections
observed in X-ray scattering experiments.76-78 In addition, while
cationic detergents are highly toxic to cells, and thus their DNA
complexes have no potential for medical applications, the DNA
complexes of MVLBG2 efficiently transfect cells, as demon-
strated below.
Cell Transfection. For transfection studies, we compared
DNA complexes of the dendritic lipid MVLBG2 and the
commonly used monovalent DOTAP74 in four different cell
lines. Figure 8 shows the effect of varying the lipid to DNA
charge ratio Fchg in mouse fibroblast L-cells. For DOTAP, TE
(72) Bruinsma, R. Eur. Phys. J. B 1998, 4, 75-88.
(73) Harries, D.; May, S.; Gelbart, W. M.; Ben-Shaul, A. Biophys. J. 1998, 75,
159-173.
(74) Leventis, R.; Silvius, J. R. Biochim. Biophys. Acta 1990, 1023, 124-132.
(75) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press:
London, 1992.
C
C
the HII and LR phases (cf. Figure 1), respectively. The
remaining interstitial space in the HIC structure is filled by the
headgroups, water, and counterions which provide local charge
neutrality for the complexes at a positive lipid-DNA charge
ratio.72,73 As evident from a comparison of the molecular models
(76) Ghirlando, R.; Wachtel, E. J.; Arad, T.; Minsky, A. Biochemistry 1992,
31, 7110-7119.
(77) Mel’nikov, S. M.; Sergeyev, V. G.; Yoshikawa, K.; Takahashi, H.; Hatta,
I. J. Chem. Phys. 1997, 107, 6917-6924.
(71) Evans, H. M.; Ewert, K. K.; Zidovska, A.; Safinya, C. R. Unpublished
data.
(78) Zhou, S.; Liang, D.; Burger, C.; Yeh, F.; Chu, B. Biomacromolecules 2004,
5, 1256-1261.
9
4002 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Gene Delivery: Honeycomb Phase of Lipid−DNA Complexes
A R T I C L E S
we observe that for DOTAP, TE rises exponentially with the
fraction of cationic lipid in the membrane from negligible
transfection to an optimum at or near 100 mol % DOTAP. This
is typical for monovalent lipids in lamellar complexes, which
require high membrane charge densities for successful endo-
somal escape via activated fusion.40,80 On the other hand, the
TE of MVLBG2 complexes reaches a maximum much earlier:
20 mol % (human cell lines) or even 10 mol % (murine cell
lines) cationic lipid are sufficient for a TE at or near the highest
TE achievable with MVLBG2 complexes. This high level of
TE is maintained throughout the remainder of the composition
range to 100 mol % MVLBG2, with a slight drop in TE
observed for the human cell lines. Comparing the maximum
TE for DOTAP and MVLBG2 in the different cell lines, it is
evident that the TE achievable with MVLBG2 is as good (293
cells), slightly better (mouse L-cells, HeLa cells), or even
significantly better (about 1 order of magnitude for mouse
embryonic fibroblast (MEF) cells, 30 mol % MVLBG2 vs 100
mol % DOTAP). The large improvement in MEF cells when
using MVLBG2, observed throughout the composition range,
is particularly significant since the MEF cells are empirically
known to be a “hard to transfect” cell line. This is, however,
only one advantage of the multivalent lipid. Another is the ability
to add a significant amount of neutral lipid to the vector
composition without loss of efficiency, allowing the incorpora-
tion of neutral lipids bearing additional (e.g., targeting) func-
tionality into the complex. In addition, since cationic lipid is
the component mainly responsible for lipid vector toxicity, the
comparatively small number of highly charged cationic lipid
molecules required for efficient transfection by MVLBG2 (or
more biodegradable analogues) should allow the cationic
component of these complexes to be metabolized quickly,
resulting in a major advantage for potential applications. Very
importantly, the data shows that complexes in the new HIC phase
and at higher contents of MVLBG2 efficiently transfect murine
and human cells, with a TE as high as or higher than that of
optimized DOTAP complexes. It is interesting to speculate how
the structure of HIC complexes may play a role in their
transfection mechanism and efficiency. The existence of a
continuous DNA substructure likely facilitates release of the
DNA cargo, as in principle all DNA is accessible once a part
of it is exposed to the inside of the cell. This is in contrast to
lamellar complexes, where the enveloping lipid bilayers have
to be removed layer by layer to release all encapsulated DNA.
Figure 7. Molecular models of hexadecavalent MVLBG2 (left) and
monovalent DOTAP (right).
Figure 8. Transfection efficiencies in mouse L-cells for DOTAP/DOPC
and MVLBG2/DOPC complexes at 80 mol % cationic lipid, plotted against
the lipid/DNA charge ratio Fchg. While a charge ratio of three is sufficient
for optimum performance of DOTAP complexes, the TE of MVLBG2
complexes benefits from higher charge ratios. A typical value of TE for
DNA without lipid is 9.6 × 104, corresponding to the starting value of the
TE axis.
Experimental Section
at all mole fractions of cationic lipid improves with Fchg up to
a saturation value,79 and the same is observed for MVLBG2.
However, the onset of the saturation occurs at a higher charge
ratio for the dendritic lipid: Fchg ) 4.5 lies in the saturated
regime but Fchg ) 3 does not, despite marking saturation for
DOTAP. Accordingly, all transfection experiments discussed
below were performed at Fchg ) 4.5 except for those on HeLa
cells, where a higher Fchg of 7 was used to ensure optimum
performance of MVLBG2.
General Methods. NMR spectroscopy was carried out on a Bruker
Avance 200 MHz spectrometer. MALDI-TOF mass spectrometry was
performed on a Dynamo spectrometer from Thermal BioAnalysis Ltd.
using 2,5-dihydroxybenzoic acid as the matrix and 30 vol.-% water in
acetonitrile as the solvent. For thin-layer chromatography, silica covered
plastic sheets with fluorescence indicator (Macherey-Nagel) were used.
Detection of spots was achieved with UV light and ninhydrin reagent
(300 mg in 95 mL of 2-propanol and 5 mL of acetic acid). By addition
of methanol to a 4:2:1 (v/v/v) mixture of chloroform, methanol, and
25% ammonia until the mixture became homogeneous, solvent mixture
A was obtained. Similarly, solvent mixture B was prepared starting
Figure 9 shows TE plotted against the mole fraction of
cationic lipid for the different cell lines. As a common feature,
(79) Slack, N. Ph.D. Dissertation, University of California, Santa Barbara,
California, 2000.
(80) Lin, A. J.; Slack, N. L.; George, C. X.; Samuel, C. E.; Safinya, C. R.
Biophys. J. 2003, 84, 3307-3316.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4003

A R T I C L E S
Ewert et al.
Figure 9. Transfection efficiencies for DOTAP/DOPC and MVLBG2/DOPC complexes in four different cell lines, plotted against the mole fraction of
cationic lipid. The data points were obtained at a constant Fchg (7 for HeLa cells, 4.5 for all others), corresponding to a constant amount of DNA applied to
the cells for each data point in a plot.
with a 1:1:1 (v/v/v) mixture of the same solvents. Silica gel from Fisher
Scientific with a mesh size of 200-425 was used for flash chroma-
tography.81
Transfection. Wild-type mouse embryonic fibroblast cells were a
gift from C. Samuel (UCSB) and originally derived as described
(Schreiber, R. D. et al. Cell 1998, 93, 373-383).82 The other cell lines
were from ATCC (ATCC numbers: mouse fibroblast L-cells: CCL-
1; HeLa cells: CCL-2; human 293 cells (kidney-derived fetal epithelial
cells): CRL-1573). Cells were cultured at 37 °C in supplemented cell
medium (Dulbecco’s Modified Eagle’s Medium (DMEM) with 1%
penicillin-streptomycin and 5% fetal bovine serum, v/v; from Gibco
BRL) in an atmosphere containing 5% CO2 and were reseeded
approximately every 72 h to maintain subconfluency. The day before
the transfection experiment, cells were seeded in 24-well plates. At
the time of transfection, the cells were approximately 70% confluent.
Luciferase plasmid DNA (pGL3 Control Vector, Promega) was prepared
using a Qiagen Giga Kit (Qiagen). For each well, 0.4 µg of luciferase
plasmid DNA was used. Lipid solution (0.6 mM total lipid) was added
to the DNA and the mixture was diluted to a final volume of 0.1 mL
with nonsupplemented DMEM. The cells were incubated with this
solution for 6 h, then rinsed three times with Phosphate Buffered Saline
(PBS, Invitrogen) and incubated with supplemented cell media for an
additional 24 h. Luciferase gene expression was measured with the
Luciferase Assay System (Promega), and light output readings were
taken on a Berthold AutoLumat luminometer. Transfection efficiency,
measured as relative light units (RLU), was normalized to the weight
of total cellular protein determined using Bio-Rad Protein Assay Dye
Reagent (Bio-Rad). All experiments were performed at least in
duplicate.
Lipid Solutions. Stock solutions of MVLBG2 were prepared in
chloroform/methanol (9:1, v/v). DOPC and DOTAP were purchased
as a solution in chloroform from Avanti Lipids. Lipid solutions were
combined at the desired ratio of lipids and dried, first by a stream of
nitrogen and subsequently in a vacuum for 8 to 12 h. To the residue,
high conductivity (18.2 MΩ) water was added, and the mixture was
incubated at 37 °C for at least 12 h to give solutions of a final
concentration of 30 mM (DOTAP) or 15 mM (MVLB2) for X-ray
samples. For transfection, solutions were prepared at 0.6 mM. Lipid
mixtures containing higher mole fractions of DOPC formed opaque
suspensions, which were sonicated to clarity and filtered through 0.2
µm filters. The lipid solutions were stored at 4 °C until use.
Ethidium Bromide Replacement Assay. Lipid-DNA complexes
were prepared in triplicate by adding a DNA/EtBr mixture to the lipid
solution in a 96 well plate. Each well contained 2.4 µg HPCT DNA,
0.28 µg EtBr, and the desired amount of lipid (100% MVLBG2, no
DOPC) in a final volume of 250 µL. The emission at 605 nm was
measured on a Cary Eclipse fluorescence spectrophotometer, following
excitation at 519 nm.
Optical Microscopy. A Nikon Diaphot 300 inverted microscope
equipped for epifluorescence and DIC and a SensiCamQE high-speed
digital camera were used. For fluorescence microscopy, aqueous lipid
stock solutions were prepared at 1 mM with 1 mol % Texas Red-DHPE
dye (Molecular Probes). DNA (0.1 mg/mL) was stained using 3.3 mol
% (calculated per base pair) YOYO dye (Molecular Probes). Samples
were prepared by combining appropriate volumes of lipid and DNA
stock solutions to reach a DNA/lipid charge ratio of 2.8 and diluting
2-fold with water. Thirty minutes later, an equal volume of an oxygen
scavenging system consisting of 0.5% â-mercaptoethanol, 200 µg/mL
glucose oxidase, 35 µg/mL catalase, and 4.5 mg/mL glucose was added
to the sample. The complexes were imaged 10 min after the scavenging
system was introduced.
Small Angle X-ray Scattering. Lipid solution (see above) was added
to 0.2 mg of highly polymerized calf thymus DNA (HPCT, from USB;
prepared at 5 mg/mL in water) and mixed with an equal volume of
DMEM. Samples were centrifuged for 3 h at 19,000 rpm in a Sorvall
SS34 rotor and stored at 4 °C for at least 3 days. Typically, the CL-
DNA complexes formed a white precipitate, and these pellets were
transferred to 1.5-mm quartz capillaries and flame sealed. Scattering
data were collected at beamline 4-2 at the Stanford Synchrotron
Radiation Laboratory (SSRL).
(81) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923-2925.
(82) Schreiber, R. D. et al. Cell 1996, 84, 431-442.
9
4004 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Gene Delivery: Honeycomb Phase of Lipid−DNA Complexes
A R T I C L E S
Synthesis. Chemicals were from Fisher Scientific and at least of
analytical grade unless otherwise indicated. They were used as received.
NR,Nδ-bis(tert-Butyl-carbamoyl)ornithine was prepared by a procedure
similar to that described in ref 60, which can be found in the Supporting
Information.
hydrotrifluoroacetate form) and n mmol DIEA in DMF was added to
the resulting solution with stirring. Stirring was continued at room
temperature for 2 h, and chloroform and water were added with vigorous
stirring after evaporation of most of the DMF.84 The phases were
separated, and after extracting two more times with chloroform, the
organic phases were combined and washed twice with 2 M HCl, twice
with 2 M NaOH, and once with water. After drying (Na2SO4), the
solvents were evaporated.
NR,Nδ-bis(2-Cyanoethyl)ornithine (5). To a solution of 474 mg
(11.9 mmol) NaOH in 6 mL of methanol, were added 1.00 g (5.93
mmol) of L-ornithine hydrochloride (Aldrich) and 15 mL of methanol.
The mixture was stirred at room temperature, 1.56 mL (1.26 g, 23.7
mmol) of acrylonitrile (Fluka) was added, and stirring was continued
for 2 h. The reaction mixture was acidified slightly by the addition of
concentrated hydrochloric acid (about 1.5 mL) and filtered immediately.
The residue was dried in vacuo to yield 1.4 g (5.66 mmol, 86%) of 5
hydrochloride as colorless microcrystals: Rf ) 0.73 (solvent mixture
A); 1H NMR (D2O): δ ) 1.65-2.25 (m, 4 H, CH2-CH2-CH), 2.85-
3.1 (m, 4 H), 3.1-3.25 (“t”, 2 H), 3.25-3.7 (“t”, 4 H) (CH2-N, CH2-
CN), 3.7-3.9 (“t”, 1 H, CH); 13C NMR (D2O): δ ) 15.2, 15.4 (CH2-
CN), 21.8, 26.9 (CH2-CH2-CH), 42.7, 43.1, 47.4 (CH2-N), 62.3
(CH), 118.1 (CN), 172.7 (CO2H).
Methyl (2S)-2,5-Di[(1S)-1,4-di(tert-butyl-carbamoyl-amino)butyl]-
carboxamidopentanoate (1). As described in the general procedure
for TBTU-mediated coupling (n ) 2), 14.6 g (44.0 mmol) of NR,Nδ-
bis(tert-Butyl-carbamoyl)ornithine was reacted with 4.38 g (20 mmol)
of ornithine methyl ester dihydrochloride (Bachem) to yield 14.1 g (18.2
mmol, 91%) of 1 as a colorless solid: Rf ) 0.52 (CHCl3/MeOH, 9:1,
1
1% AcOH); H NMR (CDCl3): δ ) 0.8-2.1 (m, C-CH2-C), 1.37
(s, C(CH3)3), 2.9-3.5 (2 m, 6 H, CH2-N), 3.65 (s, 3 H, OCH3), 4.1-
4.5 (2 m, 3 H, CH), 4.8-5.1 (bm, 2 H), 5.4-5.8 (2 b, 2 H), 7.22 (b,
1 H), 7.55 (b, 1 H) (all NH);13C NMR (CDCl3): δ ) 25.4, 25.8, 26.1
(CH2-CH2-CH2), 28.3, 28.4 (C(CH3)3), 30.0 (CH2-CH2-CH), 38.7,
39.6 (CH2-NH), 52.2 (OCH3), 52.1, 53.5, 53.9 (CH), 79.1, 80.0
(C(CH3)3),156.2 (O-C(O)-N), 172.1, 172.9, 173.2 (C-C(O)).
Methyl (2S)-2,5-Di[(1S)-1,4-diaminobutyl]carboxamidopentanoate
(8). About 70 mL of TFA (Acros) was saturated with nitrogen, cooled
on ice, and added to 10.0 g (12.9 mmol) of 1. The mixture was
evaporated after stirring for 1 h at room temperature, and diethyl ether
(300 mL) was added. The ether was decanted, and the solid product
was triturated with ether and finally dried in vacuo to yield 4.54 g
(12.1 mmol, 94%) of 8 (trifluoroacetate) as a colorless solid: 1H NMR
(MeOH-d4): δ ) 1.5-2.1 (m, 12 H, C-CH2-C), 2.9-3.1 (m, 4 H,
CH2-NH3+), 3.2-3.4 (m, 2 H, CH2-NH), 3.74 (s, 3 H, OCH3), 3.92
(“t”, 1 H), 4.04 (“t”, 1 H), 4.50 (“q”, 1 H) (CH);13C NMR (MeOH-d4):
δ ) 23.8, 24.1, 26.5, 29.5 (C-CH2-C), 40.0, 40.1 (CH2-NH), 53.0,
53.6, 53.9 (CH, OCH3), 169.9, 170.1, 173.6 (C(O)).
NR,Nδ-Bis(3-aminopropyl)ornithine (6). A mixture of 29.8 g (109
mmol) of 5, 175 mL of 1.4 m NaOH in 95% ethanol, and 6.7 g of a
supension of Raney-nickel (Aldrich) in water was prepared in a Parr
hydrogenator flask and agitated for 24 h at room temperature under 60
psi (4 bar) hydrogen pressure. The Raney-nickel was removed by
filtration over Celite (diatomaceous earth, acid washed; Sigma), and
the solvent was evaporated. The resulting oil was used for the next
step without further purification. Pure 6 hydrochloride for analytical
purposes was obtained by adding concentrated hydrochloric acid to
the reaction solution and collecting and drying the precipitate: 1H NMR
(D2O): δ ) 1.65-2.25 (2 m, 8 H, C-CH2-C), 3.0-3.35 (m, 10 H,
CH2-N), 4.12 (t, 3J ) 6.0 Hz, 1 H, CH); 13C NMR (D2O): δ ) 21.8,
24.1, 24.2, 26.2 (C-CH2-C), 37.0 (CH2-NH3+), 44.2, 45.0, 47.2
(CH2-NH2+), 59.9 (CH), 170.8 (CO2H).
NR,Nδ-Bis(tert-butyl-carbamoyl)-NR,Nδ-bis(3-[tert-butyl-carbam-
oyl-amino]propyl)ornithine (7). The raw 6 (assay calculated for 100%
yield from the previous step: 108 mmol) was dissolved in water/THF
(5:1, v/v). Over the course of 1 h, 100 mL of 4 M NaOH and a solution
of 104.2 g (477 mmol) of Boc2O (Novabiochem) in 200 mL of THF
were added in four portions with stirring while the reaction mixture
was kept at room temperature by cooling with a water bath. Stirring
was continued for 6 h, and most of the THF was then evaporated. After
addition of 600 mL of diethyl ether and 200 mL of water, the mixture
was acidified using half-concentrated hydrochloric acid. The phases
were separated, and the aqueous phase was extracted two more times
with 200 mL of ether. The combined organic phases were washed twice
with water, dried (Na2SO4), and evaporated. The residue was purified
in portions of 44 g by flash chromatography on 930 g of silica gel
using chloroform/methanol (19:1) as the eluent83 to yield 26.5 g (41
mmol, 38%) of 7 as a colorless solid: Rf ) 0.72 (9:1 CHCl3/MeOH);
1H NMR (CDCl3): δ ) 1.37, 1.38 (2s, CH3), 0.8-2.2 (ms, C-CH2-
C) (44 H), 2.7-3.9 (ms, 10 H, CH2-N), 4.26 (bm, 1 H, CH), 5.1-5.5
(b/m, 2 H, NH), 8.58 (b, 1 H, COOH); 13C NMR (CDCl3): δ ) 25.3,
27.5, 28.7, 29.8 (C-CH2-C), 28.4 (C(CH3)3), 37.4, 43.8, 44.4, 45.2
(CH2-N), 59.3 (CH), 79.0, 79.7, 80.8 (CCH3), 155.5, 156.1 (O-C(O)-
N), 174.9 (CO2H).
Methyl (2S)-2,5-Di((1S)-1,4-di[((1S)-1,4-di(tert-butyl-carbamoyl)-
(3-[tert-butyl-carbamoyl-amino]propyl)aminobutyl)carboxamido]-
butylcarboxamido)pentanoate (2). As described in the general
procedure for TBTU-mediated coupling (n ) 4), 5.83 g (9.01 mmol)
of 7 was reacted with 1.70 g (2.05 mmol) of 8 to yield 5.15 g (1.78
mmol, 87%) of 2 as a colorless solid: Rf ) 0.53 (CHCl3/MeOH 9:1,
1% concentrated NH4OH); 0.80 (CHCl3/MeOH 6:1, 1% concentrated
NH4OH); For characterization purposes, a small sample was deprotected
by adding cold TFA, stirring for 1 h and subsequent evaporation in
vacuo: 1H NMR (D2O): δ ) 1.2-2.0 (m, 44 H, C-CH2-C), 2.6-
3.2 (m, 52 H, CH2-N), 3.43 (s, 3 H, OCH3), 3.55-3.85, 3.90-4.15
(2 m, 7 H, CH), 7.2-7.7, 8.2-8.5 (m, amide-H); 13C NMR (D2O): δ
) 21.2 (2), 21.4 (2), 23.9 (4), 24.0 (4), 24.6, 24.8 (2), 27.1 (4), 28.0,
28.7 (2) (C-CH2-C), 36.7 (8) (CH2NH3+), 39.0, 39.4 (2), 43.8 (4),
44.8 (4), 47.0 (4) (CH2-N), 52.9, 53.1, 54.0, 54.4, 60.0 (2), 60.4(2)
(CH, OCH3), 116.3 (q, J ) 291 Hz, CF3), 162.4 (q, J ) 36 Hz,
CF3C(O)), 167.8 (2), 167.6 (2), 172.9, 173.2, 173.9 (C(O)).
N1-(2-Aminoethyl)-(2S)-2,5-di((1S)-1,4-di[((1S)-1,4-di(tert-butyl-
carbamoyl)(3-[tert-butyl-carbamoyl-amino]propyl)aminobutyl)car-
boxamido]butylcarboxamido)pentanamide (3). To a solution of 2.00
g (0.69 mmol) of 2 in methanol, was added 927 µL (830 mg, 13.8
mmol) of ethylenediamine in one portion, and the mixture was stirred
at room temperature for 10 days. The solvent was evaporated, and water
and chloroform were added. After agitation, the phases were separated,
and the organic layer was washed two more times with water, dried
(Na2SO4), and evaporated. The residue was purified by flash chroma-
tography on 120 g of silica gel using chloroform/methanol (15:1,
changed to 6:1 once the product started to elute) as the eluent to yield
1.33 g (0.45 mmol, 66%) of 3 as a colorless solid: Rf ) 0.61 (CHCl3/
MeOH 6:1, 1% concentrated NH4OH); 1H NMR (CDCl3, trace MeOH-
General Procedure for TBTU-Mediated Coupling (Amine Com-
pound with n Amino Groups). To a mixture of 1.1n mmol acid
component, 1.1n mmol O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluro-
nium tetrafluoroborate (TBTU; Bachem), and 1.1n mmol 1-hydroxy-
benzotriazole (HOBt) hydrate (Peptides International) in dimethylfor-
mamide (DMF) (peptide synthesis grade, Applied Biosystems), was
added 2n mmol N,N-diisopropylethylamine (DIEA; Aldrich). After 10
min, a mixture of 1 mmol amine compound (hydrochloride or
(83) This compound elutes faster than estimated by thin-layer chromatography,
probably because its retention factor increases with its concentration.
(84) Due to the sensitivity of the methyl ester to hydrolysis, immediate workup
of the reaction mixture is mandatory after the addition of water.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4005

A R T I C L E S
Ewert et al.
d4): δ ) 0.8-2.1 (m), 1.42, 1.44, 1.46 (3 s) (C-CH2-C, C(CH3)3,
188 H), 2.7-3.5 (m, 50 H, CH2-N), 4.1-4.5 (bm, 7 H, CH);13C NMR
(CDCl3): δ ) 25.3, 26.1, 29.9 (C-CH2-C), 28.3 (C(CH3)3), 37.4,
38.0, 39.1, 40.7, 43.1, 43.7, 44.3, 44.7, 46.4 (CH2-N), 52.4, 55.7, 58.1,
59.0 (CH), 78.8, 79.5, 80.7, 80.8 (C(CH3)3), 155.5, 156.1 (O-C(O)-
N), 171.8, 172.5 (C-C(O)); MALDI-MS: m/z ) 2917.3 (M + H+),
2939.1 (M + Na+).
been saturated with nitrogen and cooled on ice. The mixture was left
to stand at room temperature for 25 min with occasional swirling, and
most of the TFA was then evaporated in vacuo. A mixture of diethyl
ether/pentane (1:1, 30 mL/g protected lipid) was added with stirring,
and the precipitate was isolated by decanting. This was repeated twice,
and the precipitate was dried in vacuo to yield 1.31 g (348 µmol, 95%)
1
of MVLBG2 as a colorless solid: Rf ) 0.07 (solvent mixture B); H
N-1-2-[(1S)-1,4-Di((1S)-1,4-di[(tert-butyl-carbamoyl)(3-[tert-butyl-
carbamoyl-amino]propyl)amino]butyl)carboxamido]butylcarbox-
amido)butylcarboxamido]ethyl-3,4-di[(Z)-9-octadecenyloxy]benz-
amide (9). A solution of 301 mg (460 µmol) of DOB60,39 in
dichloromethane was added to a mixture of 148 mg (460 µmol) of
TBTU and 71 mg (460 µmol) of HOBt hydrate in DMF. Following
the addition of 216 µL (160 mg, 1.24 mmol) of DIEA, the mixture
was left to stand for 10 min, and a solution of 1.22 g (418 µmol) of 3
in DMF was added. After standing for 4 h, the solvents were evaporated.
Following the addition of chloroform and 5% citric acid and agitation,
the phases were separated, and the organic phase was washed one more
time with 5% citric acid, twice with 10% NaHCO3, and once with water.
After drying (Na2SO4), the solvent was evaporated, and the residue
was purified by flash chromatography on 100 g of silica gel using
chloroform/methanol (25:1) as the eluent83 to yield 1.29 g (364 µmol,
NMR (MeOH-d4): δ ) 0.88 (“t”, CH3), 0.8-2.4 (several m, lg. peaks
at 1.29, 1.35, C-CH2-C) (106 H), 2.85-3.65 (m, 50 H, N-CH2),
3.8-4.2 (m, 8 H, O-CH2, CH), 4.2-4.6 (m, 3 H, CH), 5.2-5.45 (m,
3
4 H, dCH-), 6.98 (d, J ) 8.6 Hz, 1 H, Har m-C(O)), 7.4-7.55 (m,
2 H, Har), 8.1-9.2 (m, NH); 13C NMR (MeOH-d4): δ ) 14.6 (CH3),
22.5, 22.9, 23.8, 25.4, 25.5, 26.6, 27.5, 28.3, 28.5, 30.5, 30.6, 30.7,
30.8, 30.95, 31.01, 33.2 (C-CH2-C), 38.0, 38.5, 40.1, 40.4, 45.0, 45.9,
46.0, 48.2 (CH2-N), 54.6, 55.2, 61.0, 61.4 (CH), 70.3, 70.8 (CH2-
O), 113.9, 114.3, 122.5 (CarH), 118.2 (q, J ) 293 Hz, CF), 127.6 (Car-
C(O)N), 130.9, 131.0 (dCH-CH2), 150.2, 154.0 (Car-O), 163.3 (q, J
) 35 Hz, TFA C(O)), 168.6, 168.7, 168.9, 169.0, 170.2, 173.7, 173.8,
174.4 (C(O)); MALDI-MS: m/z ) 1951.0 (M + H+).
Acknowledgment. This work was primarily supported by
NIH GM59288, with additional funding by NSF DMR 0503347
and NSF CTS 0404444. Our work made use of MRL Central
Facilities supported by the MRSEC Program of the National
Science Foundation under Award No. DMR 05-20415. The
SAXS data was obtained at the Stanford Synchrotron Radiation
Laboratory, a national user facility operated by Stanford
University and supported by the U.S. DOE.
1
87%) of 9 as a colorless solid: Rf ) 0.28 (CHCl3/MeOH 19:1); H
NMR (CDCl3): δ ) 0.76 CH2-CH3), 0.9-2.3 (several m, lg. peaks
at 1.16, 1.21, 1.31, 1.33, 244 H, C-CH2-C, C(CH3)3), 2.7-3.7 (m,
50 H, N-CH2), 3.89 ((“t”, 4 H, O-CH2), 4.0-4.5 (bm, 7 H, CH),
4.7-5.7 (m, 12 H, dCH-, NH), 6.5-7.8 (several m and b, 11 H),
Har, NH); 13C NMR (CDCl3): δ ) 13.9 (CH3), 22.5 (CH2-CH3), 25.3
(very b), 25.9, 27.0, 29.1, 29.3, 29.6, 31.7 (C-CH2-C), 28.3 (C(CH3)3),
37.5, 38.0, 39.4, 39.9, 42.3, 43.7, 44.3, 46.4 (CH2-N), 52.1, 54.7, 58.6,
59.3 (very b, CH), 68.9, 69.1 (CH2-O), 78.7, 79.5, 80.6, 80.8
(C(CH3)3), 112.1, 112.9, 120.0 (CarH), 126.6 (Car-C(O)N), 129.6, 129.7
()CH-CH2), 148.6, 151.7 (Car-O), 155.4, 156.0 (b, O-C(O)-N),
167.5, 171.4, 172.8, 172.5 (C-C(O)).
Supporting Information Available: Complete ref 82; full
experimental details for the synthesis of NR,Nδ-bis(tert-butyl-
carbamoyl)ornithine and corresponding characterization data.
This material is available free of charge via the Internet at
http://pubs.acs.org.
N-1-2-[((1S)-1,4-Di[(1S)-1,4-di((1S)-1,4-di[(3-aminopropyl)amino]-
butylcarboxamido)butyl]carboxamidobutyl)carboxamido]ethyl-3,4-
di[(Z)-9-octadecenyloxy]benzamide (MVLBG2, 4). 9 (1.30 g (366
mmol)) was dissolved in 10 mL of 95% TFA (5% water), which had
JA055907H
9
4006 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006