Synthesis and characterization of novel zwitterionic lipids with pH-responsive biophysical properties

Article abstract of DOI:10.1039/c2cc31710a

We report the synthesis and biophysical characterization of a novel class of zwitterionic lipids (ZL) with head groups containing a 3° or 4° amine and carboxylate. ZL form stable liposomes that exhibit head group dependent, pH-responsive biophysical characteristics. These lipids may be suitable components for drug delivery applications due to their ease of synthesis and unique pH-dependent properties.

Full text of DOI:10.1039/c2cc31710a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 5575–5577
www.rsc.org/chemcomm
COMMUNICATION
Synthesis and characterization of novel zwitterionic lipids with
pH-responsive biophysical propertiesw
Colin L. Walsh,a Juliane Nguyenb and Francis C. Szoka*ab
Received 7th March 2012, Accepted 10th April 2012
DOI: 10.1039/c2cc31710a
We report the synthesis and biophysical characterization of a
novel class of zwitterionic lipids (ZL) with head groups containing
a 38 or 48 amine and carboxylate. ZL form stable liposomes
that exhibit head group dependent, pH-responsive biophysical
characteristics. These lipids may be suitable components for
drug delivery applications due to their ease of synthesis and
unique pH-dependent properties.
Fig. 1 (a) Schematic diagram of tertiary and quaternary zwitterionic
Lipid-based nanoparticles have long been used to deliver
biologically active molecules including drugs, DNA, and more
recently siRNA in vivo.1,2 There have been extensive synthetic
efforts to develop novel lipids with biophysical properties that
can be exploited to promote efficient siRNA delivery while
maintaining low cytotoxicity and immunogenicity.2–6 siRNA
delivery systems are generally cationic or ionizable to promote
efficient siRNA encapsulation and membrane fusion.2,7
lipids. (b) Theoretical pH-dependent ionization of quaternary ZL.
a critical step in the delivery of functional nucleic acids to the
cytosol.7
All ZL head groups contain a single amine (31 or 41) and
carboxylate linked by a carbon spacer of varying molecular
structure. The structures synthesized were chosen to test how
the length and hydrophobicity of the linker region affect the
biophysical behavior of ZL. We hypothesized that differences
in linker structure could impact the protonation behavior of
ZL, their ability to form stable liposomes, or their membrane
interactions. Additionally, we studied the impact of amine
substitution on lipid behavior as a function of pH.
Here, we describe the design, synthesis, and biophysical
characterization of novel zwitterionic lipids (ZL) with pH-
responsive properties. ZL consist of a 3-amino-1,2-propanediol
backbone, ester-linked oleic acid tails, and a head group
containing a carboxylic acid and amine linked through a
carbon spacer (Fig. 1a). The length and structure of the spacer
was varied to determine its effect on the biophysical properties
of ZL. We hypothesize that ZL are anionic (31 amine) or
neutral (41 amine) at physiological pH, and become cationic
with decreasing pH (Fig. 1b, S1w). This pH-dependent ioniza-
tion allows for efficient siRNA encapsulation in a low pH
buffer.8 Because ZL remain neutral or anionic at physiological
pH, reduced opsonization of serum proteins, increased circu-
lation times and lower immunoreactivity compared to cationic
amino lipids are expected.9 Once internalized, ZL liposomes
will become cationic if endocytic pH is less than lipid pKa,
allowing for membrane interactions between anionic cellular
lipids and transiently cationic ZL.3 These interactions may
promote membrane destabilization along the endocytic pathway,
The synthesized library consists of seven ZL structures with
four distinct carboxylic acid containing headgroups (Fig. 2).
These lipids are colloquially named using the acronyms shown
under their respective structures in Fig. 2. DOPAQ was
synthesized, but degraded following purification. This was
believed to be due to a retro-Michael reaction, yielding
1,2-dioleoyloxy-3-dimethylamino propane and acrylic acid (data
not shown). Therefore, further characterization of DOPAQ was
not completed. The head group in DOAAQ was not protonated
in the tested pH range, indicating that this structure was not a
a The UC Berkeley-UCSF Graduate Program in Bioengineering,
University of California, Berkeley, Berkeley, CA 94720-1762, USA
b Department of Bioengineering and Therapeutic Sciences, UCSF,
513 Parnassus Ave, San Francisco, CA 94143, USA.
E-mail: szoka@cgl.ucsf.edu; Fax: (415) 476-0688;
Tel: (415) 476-3895
w Electronic supplementary information (ESI) available: Details of
synthesis and characterization, experimental methods, figures, and
tables. See DOI: 10.1039/c2cc31710a
Fig. 2 Schematic structures and colloquial names of synthesized ZL.
Chem. Commun., 2012, 48, 5575–5577 5575
c
This journal is The Royal Society of Chemistry 2012

View Article Online
suitable candidate for further development. Therefore, an
analogous tertiary lipid was not synthesized.
potential for all formulations occurred at the lowest pH tested
(pH = 3.0), and the minimum zeta potential occurred at the
highest pH tested (pH = 8.5) (Table S1w). However, the
magnitude of the surface charge depends on the percentage
of ZL present in liposomes. Increasing the ratio of ZL
increases the number of ionizable head groups in the formula-
tion, thereby increasing the magnitude of the surface charge of
the vesicle in pH ranges where those groups are protonated.
The effect of structural differences in ZL head groups on
their biophysical behavior was determined by studying the
pH- dependent zeta potential behavior of liposomes containing
1 : 3 ZL : DOPC (Fig. 3c, d). The length and structure of the
region linking a carboxylate and amine can affect the pKa of
these moieties. Similarly, the head group structure impacts
the ionization behavior of ZL liposomes. DOAAQ, which
contains a single carbon linker between the carboxylate and
amine, is not protonated in the pH range tested, and therefore
remains neutral. This occurs because the close proximity of the
electron withdrawing quaternary amine lowers the pKa of the
carboxylate. DOAAQ shows similar behavior to membrane
phospholipids such as DOPC, and may be an interesting
alternative to PC lipids for drug delivery.10–12 As the amine is
moved farther from the carboxylate, its impact on carboxylate
protonation decreases, resulting in a higher lipid pKa (Fig. 3c).
Indeed, all other ZL liposomes show increasing zeta potential
with decreasing pH (Fig. 3c, d).
ZL generally form liposomes as pure component systems.
However, liposomes containing DOBAT and DOBAQ, which
contain a 1,4-phenylene in the linker region, were less stable
than the remaining non-aromatic ZL, and aggregated over
time (data not shown). This may be due to the presence of a
rigid, hydrophobic region in the hydrophilic head group. The
addition of helper lipids, including DOPC, cholesterol, and
PEG-DMG allowed for formulation of stable ZL liposomes.
Understanding the ionization behavior of ZL in liposomes is
critical to the development of structure–activity relationships,
which can be used to rationally design lipids with optimal
biophysical properties for siRNA delivery. In a liposome, the
ionizable molecule is located at a hydrophobic/hydrophilic
interface with zwitterionic, polyelectrolyte characteristics,
making it difficult to directly measure lipid pKa. Zeta potential
serves as a surrogate measurement to analyze the pH-dependent
surface charge of ZL containing liposomes.
To identify the impact of amine substitution on ZL behavior,
we compared the pH-responsive zeta potential of analogous
tertiary (DOMPAT) and quaternary (DOMPAQ) ZL lipo-
somes. At low pH, formulations containing equal amounts of
DOMPAT and DOMPAQ have nearly identical cationic zeta
potentials. As the pH increases, formulations containing
DOMPAT become neutral around pH = 4.5 and eventually
anionic around pH = 7.5. Similarly, DOMPAQ liposomes
become neutral around pH = 4.5, however they remain
neutral as the pH increases (Fig. 3a). This behavior is con-
sistent for all ZL except DOAAQ, which is protonated below
pH = 3.0 (Table S1w). At low pH, the tertiary amine is likely
fully protonated, and the limiting step in these vesicles becoming
cationic is the protonation of the carboxylate. The different
behavior of tertiary and quaternary ZL at physiological pH
allows for the control of vesicle properties based on the desired
behavior.
ZL containing aromatic linkers show ionization behavior
that differs from non-aromatic ZL. It is possible that the
aromatic ring in the linker region of DOBAT and DOBAQ is
altering the ionization behavior of the amine and/or carboxylate
when assembled in a lipid bilayer. The carboxylate of DOBAQ
appears to be more easily protonatable than the other ZL tested,
resulting in higher zeta potentials at low pH. Alternatively, the
31 amine in DOBAT appears to be more difficult to protonate,
thereby decreasing the pH at which these vesicles become
neutral. This could be due to head group orientation or
molecular interactions in the bilayer. Regardless, these data
indicate that lipid structure can impact the pH-dependent
ionization behavior of ZL liposomes.
Changing the mole percentage of ZL in liposomes alters the
magnitude of their zeta potentials, but not their general
ionization behavior (Fig. 3b). As expected, the maximum zeta
Membrane fusion is thought to be a critical component
of efficient siRNA delivery by lipid-based vectors. Because
siRNA cannot passively diffuse through membranes, an active
endosomal escape mechanism is required to reach its site of
action. Cationic or ionizable lipid systems promote endosomal
escape by perturbing endosomal membranes through the
formation of ion-pairs between cationic vector lipids and
anionic membrane lipids. These electrostatic interactions drive
the formation of the unstable inverted hexagonal (HII) phase,
which promotes membrane fusion and contents release.13 In
addition to electrostatics, lipid structure plays a critical role in
membrane disruption and endosomal escape by impacting
lipid phase preference and fusogenicity.14 We investigated
interactions between ZL liposomes and anionic membranes
using a lipid mixing and membrane lysis assay to determine if
ZL head group structure impacts their membrane disruption
properties.
Fig. 3 pH-dependent zeta potential data for 1 : 3 ZL : DOPC liposomes.
(a) Effect of amine substitution on zeta potential. (b) Effect of liposomal
formulation on zeta potential. (c) Effect of headgroup structure on zeta
potential. (d) Zeta potential of all ZL.
A FRET based lipid mixing assay, which measures lipid
exchange due to membrane fusion or aggregation, was used to
study how ZL liposomes interact with anionic membranes as a
c
5576 Chem. Commun., 2012, 48, 5575–5577
This journal is The Royal Society of Chemistry 2012

View Article Online
Several DOBAQ containing liposomal formulations showed
no knockdown of Factor VII after 48 h following a single
5 mg kgÀ1 injection (Fig. S3w). The inability of ZL systems to
transfect is likely because they are not effectively protonated in
the endosome. However, ZL were generally well tolerated
across multiple liposomal formulations, and did not induce
inflammatory cytokines in vivo (Fig. S4w).
We have described the synthesis and pH-dependent biophysical
behavior of a novel class of ionizable zwitterionic lipids
containing a cationic amine and anionic carboxylate. Structural
variations in the head group were shown to impact the ionization
behavior of ZL; however once the carbon spacer is greater than
one carbon, all ZL show the same general trends. ZL are capable
of forming small, monodisperse liposomes that efficiently
encapsulate siRNA when ionized, and show pH-dependent
zeta potentials ranging from anionic (31 ZL) or neutral (41 ZL)
at physiological pH to cationic at low pH. ZL promote lipid
mixing with anionic vesicles when cationic, and at pH = 4.0,
vesicles containing ZL induce lysis of biomembrane mimicking
vesicles. No membrane disruption was seen with ZL containing
liposomes at physiological pH, and no cytokine induction was
observed in vivo. Structural modifications to increase lipid pKa
may result in zwitterionic lipid systems that efficiently deliver
siRNA while still limiting immune stimulation.
Fig. 4 (a) pH-dependent lipid mixing between ZL liposomes and
anionic liposomes. (b) pH-dependent membrane lysis of anionic
membrane mimicking liposomes by ZL liposomes.
function of pH.15 This provides insight into the fusogenic
potential of ZL, as lipid mixing between the outer membrane
leaflets is a required step in membrane fusion, and may
correlate with their ability to deliver siRNA. ZL are capable
of lipid mixing when cationic (Fig. 4a). Maximum lipid mixing
for all ZL is achieved at pH = 3.5, which corresponds with the
highest zeta potentials. DOBAQ liposomes, which had the
highest zeta potential at all pH o 5.0, show the highest levels
of lipid mixing. DOAAQ, which is not protonated in the range
tested, shows baseline levels of lipid mixing across all tested
pH values. These results agree with published data indicating
that electrostatics drive lipid mixing.6 However, they do
indicate that ZL may not promote endosomal escape because
they do not become cationic at endocytic pH (pH = 5.0–7.0).
Lysis of membrane mimicking anionic vesicles is also used
as a surrogate assay to predict the ability of liposomal systems
to promote endosomal escape.6 We investigated the capacity
of ZL containing liposomes to lyse biomembrane mimicking
vesicles (BMV) encapsulating ANTS/DPX as a function of pH
(Fig. 4b). pH = 4.0 and 7.4 were chosen because previous
experiments confirmed that liposomes containing ZL were
cationic and membrane active at pH = 4.0 but neutral and
inactive at pH = 7.4. ZL liposomes lyse BMV at pH = 4.0 but
cause no lysis at pH = 7.4, indicating they may exhibit
reduced immune stimulation due to their neutral charge at
physiological pH. These results confirm that ZL liposomes
disrupt anionic membranes only when cationic. In this model
system, DOPAT liposomes showed significantly (p o 0.05)
more lysis than all other ZL, indicating that head group
structure may impact membrane lysis.
This work was supported by the NSF Graduate Research
Fellowship Program, NIH–EB003008, Deutsche Forschungs
Gemeinschaft (DGF), and Pfizer.
References
1 V. P. Torchilin, Nat. Rev. Drug Discovery, 2005, 4, 145–160.
2 M. G. Stanton and S. L. Colletti, J. Med. Chem., 2010, 53, 7887–7901.
3 S. C. Semple, A. Akinc, J. Chen, A. P. Sandhu, B. L. Mui, C. K. Cho,
D. W. Y. Sah, D. Stebbing, E. J. Crosley, E. Yaworski, I. M. Hafez,
J. R. Dorkin, J. Qin, K. Lam, K. G. Rajeev, K. F. Wong, L. B. Jeffs,
L. Nechev, M. L. Eisenhardt, M. Jayaraman, M. Kazem, M. A.
Maier, M. Srinivasulu, M. J. Weinstein, Q. Chen, R. Alvarez, S. A.
Barros, S. De, S. K. Klimuk, T. Borland, V. Kosovrasti, W. L.
Cantley, Y. K. Tam, M. Manoharan, M. A. Ciufolini, M. A. Tracy,
A. de Fougerolles, I. Maclachlan, P. R. Cullis, T. D. Madden and
M. J. Hope, Nat. Biotechnol., 2010, 28, 172–176.
4 R. C. Adami, S. Seth, P. Harvie, R. Johns, R. Fam, K. Fosnaugh,
T. Zhu, K. Farber, M. McCutcheon, T. T. Goodman, Y. Liu,
Y. Chen, E. Kwang, M. V. Templin, G. Severson, T. Brown,
N. Vaish, F. Chen, P. Charmley, B. Polisky and M. E. J. Houston,
Mol. Ther., 2011, 19, 1141–51.
5 J. Heyes, L. Palmer, K. Bremner and I. MacLachlan, J. Controlled
Release, 2005, 107, 276–287.
6 J. Zhang, H. Fan, D. A. Levorse and L. S. Crocker, Langmuir,
2011, 27, 1907–14.
7 W. Li and F. C. Szoka, Pharm. Res., 2007, 24, 438–449.
8 L. B. Jeffs, L. R. Palmer, E. G. Ambegia, C. Giesbrecht,
S. Ewanick and I. Maclachlan, Pharm. Res., 2005, 22, 362–372.
9 M. T. Abrams, M. L. Koser, J. Seitzer, S. C. Williams, M. A. DiPietro,
W. Wang, A. W. Shaw, X. Mao, V. Jadhav, J. P. Davide, P. A. Burke,
A. B. Sachs, S. M. Stirdivant and L. Sepp-Lorenzino, Mol. Ther., 2010,
18, 171–180.
Ionized ZL can encapsulate anionic molecules such as
siRNA more efficiently than zwitterionic phospholipids by
exploiting their cationic charge at low pH. DOBAQ, DOPAT,
and DOAAQ were chosen as representative ZL for siRNA
encapsulation experiments using a modified ethanol dilution
method.8 DOBAQ and DOPAT, both of which are cationic at
pH = 4.0, form small, monodisperse liposomes with over 80%
siRNA encapsulation. DOAAQ, which is not ionized at
pH = 4.0 showed less than 5% encapsulation, indicating that
ionized ZL can promote encapsulation of anionic molecules
(Table S2w). This demonstrates the potential utility of the
pH-sensitive characteristics of ZL for encapsulating multivalent
anionic drugs in small, neutral particles by modulating the
pH of the system during encapsulation. DOBAQ : DOPE
(65 : 35) liposomes showed no siRNA knockdown of luciferase
(ZL : siRNA ratio (w : w) = 15 or 30; 90 or 180 nM siRNA)
in stably transfected HeLa-Luc cells in vitro (Fig. S2w).
10 A. G. Kohli, C. L. Walsh and F. C. Szoka, Chem. Phys. Lipids,
2012, 165, 252–259.
11 E. K. Perttu, A. G. Kohli and F. C. Szoka, J. Am. Chem. Soc.,
2012, 134, 4485–4488.
12 E. K. Perttu and F. C. Szoka, Chem. Commun., 2011, 47, 12613–15.
13 O. Zelphati and F. C. Szoka, Proc. Natl. Acad. Sci. U. S. A., 1996,
93, 11493–98.
14 J. Nguyen and F. C. Szoka, Acc. Chem. Res., 2012, DOI: 10.1021/
ar3000162.
15 D. Struck, D. Hoekstra and R. Pagano, Biochemistry, 1981, 20,
4093–99.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 5575–5577 5577