13574 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
Characterization of liposomes formed by lipopolysaccharides from
Burkholderia cenocepacia, Burkholderia multivorans and Agrobacteriumtumefaciens: from the molecular structure to the aggregate architecture
Gerardino D’Errico,ab
Alba Silipo,cGaetano Mangiapia,
abGiuseppe Vitiello,
ab
Aurel Radulescu,dAntonio Molinaro,
cRosa Lanzetta
cand Luigi Paduano*
ab
Received 30th March 2010, Accepted 12th July 2010
DOI: 10.1039/c0cp00066c
The microstructure of liposomes formed by the lipopolysaccharides (LPS) derived from
Burkholderia cenocepacia ET-12 type strain LMG 16656, Burkholderia multivorans strain
C1576 and Agrobacterium tumefaciens strain TT111 has been investigated by a combined
experimental strategy, including dynamic light scattering (DLS), small-angle neutron scattering
(SANS) and electron paramagnetic resonance (EPR). The results highlight that the LPS molecular
structure determines, through a complex interplay of hydrophobic, steric and electrostatic
interactions, the morphology of the aggregates formed in aqueous medium. All the considered
LPS form liposomes that in most cases present a multilamellar arrangement. The thickness of the
hydrophobic domain of each bilayer and the local ordering of the acyl chains are determined not
only by the molecular structure of the LPS glycolipid portion (lipid A), but also, indirectly, by the
bulkiness of the saccharidic portion. In the case of a long polysaccharidic chain, such as that of
the LPS derived from Burkholderia multivorans, liposomes coexist with elongated micellar
aggregates, whose population decreases if a typical phospholipid, such as dioleoyl
phosphatidylethanolamine (DOPE) is introduced in the liposome formulation. The effect of
temperature has also been considered: for all the considered LPS an extremely smooth transition
of the acyl chain self-organization from a gel to a liquid crystalline phase is detected around
30–35 1C. In the biological context, our results suggest that the rich biodiversity of LPS molecular
structure could be fundamental to finely tune the structure and functional properties of the outer
membrane of Gram negative bacteria.
1. Introduction
Bacteria are unicellular organisms devoid of a membrane-
delimited nucleus and other organelles like mitochondria and
chloroplasts. Depending on their answer to the Gram staining,
they can be divided in two main families: Gram negative
bacteria and Gram positive bacteria. Gram-positive bacteria
are those that are stained dark blue or violet by Gram staining.
This is in contrast to Gram-negative bacteria, which cannot
retain the crystal violet stain, instead taking up the counter-
stain (safranin or fuchsin) and appearing red or pink. The
different responsiveness of the bacteria to the Gram staining
test is due to the differences existing between their cell walls.
In particular, Gram-negative bacteria possess a multilayer
architecture of their cell membrane.1 The internal layer is
quite thin (B3 nm) and is constituted by peptidoglycan, a
polymer consisting of saccharide backbones cross-linked by
peptide side chains, forming a mesh-like local microstructure.
In contrast, the most external shell of these bacteria is formed
by a thicker wall (B7 nm) whose composition is different
among the two sides. The internal side is mainly characterized
by the presence of glycerol-based phospholipids, whereas the
external side is composed of lipopolysaccharides (LPS), large
molecules consisting of a glycolipid moiety and a saccharide
portion covalently linked.2
LPS are also called endotoxins because, once released from the
membrane, they play a key role in the pathogenesis of Gram-
negative infections due to their immunostimulatory properties.
They are indispensable for the bacterial growth and viability and
are responsible for the correct assembly of the membrane and its
structural integrity, being able to protect the bacterial cell from
physical and chemical attacks. From a chemical and bio-
synthetical viewpoint, they are composed of three parts:
—a glycolipid portion known as lipid A. This part is
composed of a disaccharide with multiple fatty acid tails,
embedded in the lipid bilayer while the rest of the LPS projects
toward the external environment;
—a core oligosaccharide portion, covalently attached to the
lipid A. In the core region the 3-deoxy-D-manno-octulosonic
acid residue (Kdo) is always present as well as, in most cases,
heptose residues;
—an external O-specific long polysaccharide chain, also
known as O-chain, that extends from the core oligosaccharide.
Although from a chemical point of view, the O-chain is
essentially similar to the core oligosaccharide, they are
aDepartment of Chemistry, University of Naples ‘‘Federico II’’,Naples, Italy
bCSGI (Consorzio per lo Sviluppo dei Sistemi a Grande Interfase),Florence, Italy
cDepartment of Organic Chemistry and Biochemistry, University ofNaples ‘‘Federico II’’, Naples, Italy
d Institut fur Festkorperforsching Forschungszentrum Julich and JulichCentre for Neutron Science, Garching bei Munchen, Germany
PAPER www.rsc.org/pccp | Physical Chemistry Chemical Physics
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13575
biologically distinct. Indeed, the O-chain is a proper poly-
saccharide with a repeating oligosaccharide unit.
All three parts of LPS can vary among different Gram-
negative bacterial strains; the largest variability is located in
the O-chain portion (that can also be missing) which is more
exposed to selective pressures of the outer environment and to
modifications induced by external stimuli. O-chains are
characterized by high specificity within a species and makes
LPS immunogenic being the antigenic determinant recognized
by the host specific immune system; often even slight
modifications to the primary structure can be carried out to
avoid host detection. Usually, the presence or the absence of
the O-chain determines whether the LPS are considered
smooth or rough type, respectively, depending on the appearance
of the bacterial colonies on agar plates. Full length O-chains
make the LPS smooth type (S-LPS) whereas their absence
makes the LPS rough type (R-LPS, also known as LOS,
lipooligosaccharide).3
The lipid A portion of LPS is the most conserved part,
despite its composition may also vary. Indeed, the lipid A
component is recognized as the primary immunostimulator. It
acts as strong elicitor of the innate immune system by the
induction of inflammatory cytokines in mammalian cells.4
The recognition of bacterial lipid A allows the host to
unequivocally signal the presence of infection. The protein
toll-like receptor 4 (TLR-4) is the signal transducing receptor
of LPS; lipid A is the real effector required to activate TLR-4
signalling pathway in conjunction with a soluble co-receptor
protein lymphocyte antigen 96 (MD-2), which directly and
physically binds to LPS. The low and balanced concentrations
of these mediators and soluble immune response modulators
lead to a resulting inflammation that is one of the most
important and ubiquitous aspects of the immune host defence
against invading microorganisms. Beside these positive effects,
an uncontrolled and massive immune response, due to
the circulation of a large amount of endotoxins, leads to
symptoms of sepsis and septic shock.4
The LPS molecules present on the external surface of the
bacteria are stabilized by the electrostatic interactions existing
between the negative charges located in the lipid A part (1- and
40-phosphates) and in the adjacent Kdo moiety (carboxylate
groups), and the divalent cations present in the aqueous
environment (mainly Ca2+ and Mg2+). These interactions
make LPS layer semirigid with a highly ordered structure
and low permeability towards hydrophobic solutes, thus
explaining the high resistance of Gram-negative bacteria
towards hydrophobic host-released antimicrobial peptides
and negatively charged antibiotics.5
From the arguments summarized above, it appears evident
that many of the relevant aspects of Gram-negative bacteria
behavior can be ascribed to the peculiar physico-chemical
properties of the external layer of their cell wall, whose
architecture is in turn determined by the molecular structure
of the constituting LPS. Consequently, structural and
functional investigations on liposomes or membranes formed
by LPS molecules can greatly help to increase the knowledge
on the stability, low permeability and high resistance of the
outer membrane of Gram negative bacteria and also on their
biological activity. This information may, in turn, lead to new
strategies for the design and synthesis of new molecules
capable of penetrating into the cell wall and causing the
bacterium death. To this aim, several studies have been
performed using several techniques (fluorescence spectro-
scopy, light and electron microscopy, X-ray and neutron
scattering, differential scanning calorimetry, etc.), mainly on
LPS derived by Escherichia coli and Salmonella enterica.6–14
In this scenario, here we present a detailed investigation on
liposomes constituted of three different LPS: (i) R-LPS from
Burkholderia cenocepacia ET-12 type strain LMG 16656;
(ii) S-LPS from Burkholderia multivorans strain C1576;
(iii) R-LPS from Agrobacterium tumefaciens strain TT111.
The results are also compared with those obtained for a
R-LPS derived from S. enterica serotype minnesota strain
595 (Re mutant), presented in a previous work.15 Selection
of LPS is detailed in the Discussion section. Particularly, we
try to connect their self-aggregation behavior to the molecular
structure (shown in Fig. 1). In some samples, a diacylglycero-
phosphoethanolamine has been included in liposomes
formulation. Phosphatidylethanolamines represent, besides
LPS, the most abundant lipid component in the outer
membrane of Gram-negative bacteria.16
The investigation has been performed using an experimental
strategy which has been proved to be extremely informative on
liposome aqueous suspensions15 and combines dynamic light
scattering (DLS) to estimate liposome dimension, small angle
neutron scattering (SANS) to analyze the aggregate morphology
and to estimate the thickness of the lipid bilayer and electron
paramagnetic resonance (EPR) to investigate the dynamics of the
lipid hydrophobic tail in the bilayer. The report is organized as
follows: in the Results (section 3) the results obtained by different
techniques are presented separately; in the Discussion (section 4),
the findings are discussed and compared, in an attempt to achieve
general comprehension of the system behavior.
2. Materials and methods
2.1 Materials and sample preparation
LPS compounds have been extracted and purified as reported
elsewhere.17–19 Liposomes have been prepared by dissolving
an appropriate amount of the molecules in chloroform
(HPLC-grade, obtained from Sigma, St. Louis, MO, USA),
in order to have a final concentration of B1 mg ml�1.
The dissolution has been favored through a slight warming
(B40 1C) and a short sonication treatment (B5 min at 59 kHz).
Subsequently, appropriate amounts of the obtained solution
have been transferred to a round-bottom glass tube. A thin
film of the lipid has been obtained through evaporation of the
solvent with dry nitrogen gas and vacuum desiccation.
For preparing samples including 1,2-dioleyl-sn-glycero-3-
phosphoethanolamine (DOPE, obtained from Avanti Polar
Lipids, Birmingham, AL, USA), a suitable amount of phospho-
lipid dissolved in chloroform (concentration B10 mg ml�1)
has been mixed to the LPS solution in the round-bottom glass
tubes before desiccation. The DOPE/LPS molar ratio has been
set to 25/75, which is close to the LPS composition of the
external leaflet of the outer membrane of the Gram-negative
bacteria.20
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
13576 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
Fig. 1 Molecular structure of LPS from Burkholderia cenocepacia (A, mw = 3500 amu), Burkholderia multivorans (B, mw E 10000 amu),
Agrobacterium tumefaciens (C, mw = 3671 amu) and Salmonella minnesota (D, mw = 2400 amu).
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13577
Samples to be analyzed by EPR also included 1% (w/w) of
spin-labeled phosphatidylcholine (1-palmitoyl-2-[n-(4,4-di-
methyloxazolidine-N-oxyl)]stearoyl-sn-glycero-3-phosphocholine,
n-PCSL, n = 5,14), purchased from Avanti Polar Lipids and
stored at �20 1C in ethanol solutions.
The samples have then been hydrated with different buffer
solutions prepared in D2O (isotopic enrichment >99.8%,
purchased from Sigma), at pH 7.4 (i.e. the physiological pH)
and at pH 9.1 (the pKa of the ammonium groups present onto
the polar head of DOPE molecules) and vortexed. Afterwards,
the obtained suspensions have been sonicated and repeatedly
extruded through polycarbonate membranes of 100 nm pore
size, at least 11 times. The final concentration of LPS is
0.5 mmol dm�3.
Buffer at pH 7.4 has been prepared by dissolving sodium
dihydrogenphosphate NaH2PO4 and sodium hydrogen-
phosphate Na2HPO4 in D2O at concentrations equal to 0.0773
and 0.123 mol dm�3, respectively, whereas buffer at pH 9.1 has
been prepared by dissolving sodium carbonate Na2CO3 and
sodium hydrogencarbonate NaHCO3 in D2O at concentrations
equal to 0.0132 and 0.187 mol dm�3, respectively. The
pH values has been then verified to be within 0.1 pH units
by means of a Radiometer pHM220 pH-meter, equipped with
a saturated calomel electrode and a glass electrode previously
calibrated with the IUPAC standard buffer solutions.21
2.2 DLS measurements
The light scattering setup was an ALV/CGS-3 based compact
goniometer system from ALV-GmbH (Langen, Germany).
The light source was constituted by a He–Ne laser operating
at 632.8 nm with a fixed output power of 22 mW.
In DLS, the intensity autocorrelation function g(2)(t) is
measured and related to the electric field autocorrelation
g(1)(t) by the Siegert relation. The parameter g(1)(t) can be
written as the Laplace transform of the distribution of the
relaxation rate G. Laplace transform was performed using the
RILT algorithm incorporated in the commercial ALV/CGS-3
software.22 From the relaxation rates, the translational
diffusion coefficient D may be obtained as23
D ¼ limq!0
Gq2
ð1Þ
where q = 4pn0/l sin(y/2) is the modulus of the scattering
vector, n0 is the refractive index of the solution, l is the
incident wavelength and y represents the scattering angle.
Thus D is obtained from the limit slope of G as a function
of q2, where G is measured at different scattering angles.
All measurements were performed at (25.00 � 0.05) 1C by
using a thermostat bath.
2.3 SANS measurements
SANS measurements were performed at 25 1C with the KWS2
instrument located at the Heinz Meier Leibnitz Source,
Garching Forschungszentrum (Germany). Neutrons with a
wavelength spread Dl/l r 0.2 were used. A two-dimensional
array detector at three different wavelengths (W)/collimation
(C)/sample-to-detector(D) distance combinations (W7AC8mD2m,
W7AC8mD8m and W19AC8mD8m), measured neutrons scattered
from the samples. These configurations allowed collecting data
in a range of the scattering vector modulus between 0.0019 and
0.179 A�1. The investigated samples were contained in a
closed quartz cell, in order to prevent the solvent evaporation,
and kept under measurements for a period such as to have
B2 million counts of neutrons. The obtained raw data were
then corrected for background and empty cell scattering.
Detector efficiency corrections, radial average and transformation
to absolute scattering cross sections dS/dO were made with a
secondary plexiglass standard.24,25
2.4 EPR measurements
EPR spectra of 5-PCSL or 14-PCSL in LPS samples were
recorded on a Elexys E-500 EPR spectrometer from Bruker
(Rheinstetten, Germany) operating in the X band. Capillaries
containing the samples were placed in a standard 4 mm quartz
sample tube. The temperature of the sample was regulated and
maintained constant during the measurement by blowing
thermostatted nitrogen gas through a quartz Dewar. The
samples were investigated in a temperature range spanning
from 5 to 45 1C. The instrumental settings were as follows:
sweep width, 120 G; resolution, 1024 points; modulation
frequency, 100 kHz; modulation amplitude, 1.0 G; time
constant, 20.5 ms; sweep time, 42 s; incident power, 5.0 mW.
Several scans, typically 16, were accumulated to improve the
signal-to-noise ratio.
3. Results
3.1 DLS measurements: aggregate dimension
The size of aggregates formed by LPS in buffer at pH 7.4, and
LPS-DOPE mixtures at pH 7.4 and 9 have been obtained
through DLS measurements, carried out in the angular range
30–1201. The relaxation time distribution, obtained by regularized
inverse Laplace transformation of the correlation functions,
are shown in Fig. 2–4. For pure LPS (solid lines), they are
monomodal for all the systems except for that constituted by
S-LPS from B. multivorans that showed a bimodal distribution,
indicating the presence of two well-distinct species in solution.
The experimental translational diffusion coefficients, D,
obtained from the relaxation rates as described in the
Experimental section, are reported in Table 1. In the same
table the literature data relative to S. minnesota are also
reported.15
In the approximation of very diluted solutions, D can be
directly related to the hydrodynamic radius of the aggregates,
RH, through the Stokes–Einstein equation, which holds for
infinitely diluted hard spheres diffusing in a continuous medium26
RH ¼kBT
6pZDð2Þ
where kB is the Boltzmann constant, T is the absolute
temperature, and Z is the viscosity of the medium (i.e. the solvent).
RH values of LPS aggregates are also reported in Table 1.
From inspection of the table, it is possible to infer the presence
of large diffusing aggregates with radii ranging between B100
and B180 nm, compatible with the presence of liposomial
aggregates. For the system composed of B. multivorans S-LPS,
smaller aggregates with RH E 25 nm are also present.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
13578 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
The presence of DOPE in the liposome formulation does
not change the kind of aggregates present in each system at
both pH values investigated, see Fig. 2–4. In the case of S-LPS
derived from B. multivorans, DOPE alters the relative
populations of the two kinds of aggregates, in that the relative
volume fraction of larger aggregates increases.
3.2 SANS measurements: aggregate structural parameters
The structural parameters of the aggregates have been
obtained by applying the appropriate model to the experimental
scattering cross sections.27 Measurements have been performed
for all the systems in the absence and in the presence of DOPE,
at pH 7.4 and 9.1, see Fig. 5–7. In particular, the scattering
cross section pattern dS/dO for the aqueous dispersion of
R-LPS derived from B. cenocepacia and A. tumefaciens are
reported in Fig. 5 and 7, respectively (open circles). Qualitative
analysis of the cross sections allows revealing the presence of a
power law dS/dOp q�a in the small q region (qr 0.005 A�1),
with the exponent a > 2.28 Furthermore, in the intermediate
q region (q D 0.05 A�1) a shoulder is present. All these
features, consistent with the molecular architecture of the
molecules, are typical characteristics of the presence of multi-
lamellar vesicles/liposomes. Indeed, the value of the exponent
a depends on the average lamellarity of the vesicles29 whereas
the shoulder should arise from the interference occurring
among the concentric vesicular layers. Despite the presence
in literature of more complex models for multilamellar
vesicles,29 in order to reduce the number of fitting parameters,
we have used the simplest model30 available for structural
characterization of such aggregates. Accordingly, the scattering
function has been modeled considering the liposomes as a
collection of one-dimensional paracrystalline stack. The
theoretical expression of dS/dO is:
dSdOðqÞ ¼ 1
q2hjf ðqÞj2i 1þ hjf ðqÞj
2ihjf ðqÞj2i
ðSðqÞ � 1Þ !
þ dSdO
� �incoh
ð3Þ
where f(q) is the form factor of a bilayer, whereas S(q) is the
structure factor that takes into account the interferences
occurring among the bilayers belonging to a single stack.
Finally, (dS/dO)incoh represents the incoherent contribution
to the cross section measured, mainly due to the presence of
hydrogenated molecules. Both form and structure factors
depend in a complex way on the geometrical parameters of
the stack, namely the number of layers in the stack, N, the
mean layer thickness, d, and the distance between the centers
of two consecutive layers dl. Thus, optimized values of these
parameters have been obtained from data fitting. Actually, the
Fig. 2 Relaxation time distributions at 25 1C and y = 901 for the
following aqueous systems: LPS from Burkholderia cenocepacia at
pH 7.4 (continuous line); LPS from Burkholderia cenocepacia—DOPE
(3 : 1) at pH 7.4 (dashed line); LPS from Burkholderia cenocepacia—
DOPE (3 : 1) at pH 9 (dotted line).
Fig. 3 Relaxation time distributions at 25 1C and y = 901 for the
following aqueous systems: LPS from Burkholderia multivorans at
pH 7.4 (continuous line); LPS from Burkholderia multivorans—DOPE
(3 : 1) at pH 7.4 (dashed line); LPS from Burkholderia multivorans—
DOPE (3 : 1) at pH 9 (dotted line).
Fig. 4 Relaxation time distributions at 25 1C and y = 901 for the
following aqueous systems: LPS from Agrobacterium tumefaciens at
pH 7.4 (continuous line); LPS from Agrobacterium tumefaciens—
DOPE (3 : 1) at pH 7.4 (dashed line); LPS from Agrobacterium
tumefaciens—DOPE (3 : 1) at pH 9 (dotted line).
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13579
number of layers N is generally only approximately determined;
it determines the upturn in scattering at low q values where the
total thickness of the stack is seen. Furthermore, because of the
low scattering contrast existing between the highly hydrated LPS
hydrophilic moieties and D2O bulk, the obtained d values
represent the thickness of the hydrophobic layer constituted of
the acyl chains rather than the whole thickness of LPS bilayer.
Both d and dl have been allowed to be polydisperse: the former
with a Schulz–Zimm distribution function related to the Zimm
polydispersity indexZ, the latter with a Gaussian distribution with
a standard deviation sdl. In particular, the ratio sdl/dl, named
Hosemann factor, has been estimated. The values of all the
parameters are collected in Table 2. It is worth pointing out that,
because of the limited instrument resolution, the polydispersity
parameters, Z and sdl/dl tend to be overestimated, since both the
resolution and the polydispersity smear out the oscillations
contained in the form factor. Inter-particle structure factor may
be approximated to the unity because of the low amount of LPS
(0.5 mmol dm�3, corresponding to a volume fraction of D0.07%
of solute volume fraction).
Different experimental evidences have been obtained for the
S-LPS derived from B. multivorans. In this case the scattering
profile obtained for the aqueous dispersion (open circles in
Fig. 6) is typical of two different size scattering objects:31,32 in
the region at high q (q > 0.006 A�1) the cross-section shows a
power law decrease, (dS/dO) p q�1, characteristic of
elongated cylindrical micelles, whose theoretical expression,
in the region where the power law holds, can be written as33
dSdOðqÞ ¼
fcylð1� fcylÞp2R2ðrc � r0Þ2
qexp � q2R2
4
� �
þ dSdO
� �incoh
ð4Þ
Table 1 Translational diffusion coefficients, D, and hydrodynamic radii, RH, obtained through DLS measurements at 25 1C. In the table, LPS areindicated by the name of the bacterium from which they are obtained
System 1012D=m2 s�1 RH=nm
Salmonella minnesotaa 2.7 � 0.9 80 � 20Burkholderia cenocepacia pH 7.4 1.66 � 0.15 146 � 13Burkholderia cenocepacia—DOPE pH 7.4 1.39 � 0.06 175 � 7Burkholderia cenocepacia—DOPE pH 9.1 2.14 � 0.05 113 � 3Burkholderia multivorans pH 7.4 9.59 � 0.15 25.3 � 0.8
1.78 � 0.11 136 � 8Burkholderia multivorans—DOPE pH 7.4 8.75 � 0.18 27.8 � 0.6
1.33 � 0.06 183 � 8Burkholderia multivorans—DOPE pH 9.1 8.8 � 0.2 27.6 � 0.6
1.32 � 0.07 184 � 10Agrobacterium tumefaciens pH 7.4 1.64 � 0.03 148 � 3Agrobacterium tumefaciens—DOPE pH 7.4 1.86 � 0.02 130 � 2Agrobacterium tumefaciens—DOPE pH 9.1 1.64 � 0.04 148 � 4
a Data from ref. 15.
Fig. 5 Scattering cross sections at 25 1C for the following aqueous
systems: LPS from Burkholderia cenocepacia at pH 7.4 (open circles);
LPS from Burkholderia cenocepacia—DOPE (3 : 1) at pH 7.4
(full triangles); LPS from Burkholderia cenocepacia—DOPE (3 : 1) at
pH 9 (open squares). For a better comparison, data have been multiplied
by a scale factor, as indicated. Fitting curves to the experimental data
through the models reported in the text are also shown.
Fig. 6 Scattering cross sections at 25 1C for the following aqueous
systems: LPS from Burkholderia multivorans at pH 7.4 (open circles);
LPS from Burkholderia multivorans—DOPE (3 : 1) at pH 7.4
(full triangles); LPS from Burkholderia multivorans—DOPE (3 : 1) at
pH 9 (open squares). For a better comparison, data have been multiplied
by a scale factor, as indicated. Fitting curves to the experimental data
through the models reported in the text have are also shown.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
13580 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
where R is the radius of the base, fcyl the cylinder volume
fraction, and rc � r0 the scattering length density difference
between the cylinders and the solvent. In the low q region
(q o 0.006 A�1) the trend is similar to what was found for the
B. cenocepacia and A. tumefaciens, denoting the presence of
multilamellar liposomes. In this case, scattering cross sections
have been analyzed assuming each kind of aggregate to scatter
independently from the other and expressing (dS/dO) as the
sum of the eqn (3) and (4), taking into account for the relative
number density of the objects, strictly connected to the volume
fractions fcyl and flip. This procedure allows the estimation of
the micelle radius R that results to be (4.2 � 0.2) nm, while the
liposome structural parameters are reported in Table 2.
Although the lack of the Guinier regime for the micelles does
not allow obtaining a value of the length of cylindrical
micelles, l, nonetheless an estimation for such value can be
obtained from the knowledge of the DLS hydrodynamic
radius RH that for a cylinder is given by34
RH ¼ffiffiffiffiffiffiffiffiffiffiffi3
4R2l
3
rð1:0304þ 0:0193xþ 0:06229x2
þ 0:00476x3 þ 0:00166x4 þ 2:66� 10�6x7Þð5Þ
where R is the cylinder radius and x= ln[l/(2R)]. From eqn (5)
the length of the micelles formed by B. multivorans S-LPS has
been estimated to be (68 � 2) nm.
In Table 2 we also report the structural parameters obtained
for S. minnesota R-LPS.15 In this case only unilamellar
liposomes were detected.
In the presence of DOPE in aggregate formulation, at both
investigated pH values, the scattering profile is quite similar
for all three LPS considered in the present work. In all cases,
dS/dO shows a power law decrease with a o 2 at low q and a
shoulder at higher q (q D 0.02 A�1 for both B. cenocepacia
R-LPS and B. multivorans S-LPS and q D 0.07 A�1 for
A. tumefaciens R-LPS). Consequently the experimental data
were fitted modeling the systems as constituted by multi-
lamellar vesicles, using the model developed by Kotarchyk
and Ritzau.30 The fitting parameters are reported in Table 2.
Interestingly, in the presence of DOPE, also in the case of
S-LPS derived from B. multivorans no evidence of smaller
aggregate was detected.
3.3 EPR measurements: bilayer structuring
EPR spectroscopy, by using spin-labeled lipids has been
proved to give substantial information on the acyl chains
structuring in the lipid bilayers.35–38 In the present study, the
samples investigated were phosphatidylcholine spin-labeled on
the 5 or 14 C-atom of the sn-2 chain (5-PCSL and 14-PCSL,
respectively) incorporated in LPS liposomes. Incorporation,
and molecular dispersion, of the spin-label into the LPS
bilayer are highlighted by the absence of any evidence of
spin-exchange in the registered spectra, as would be expected
in the case of spin-label self-aggregation. 5-PCSL bears
the radical label close to the molecule head group, and
consequently allows monitoring the behavior of the region
Fig. 7 Scattering cross sections at 25 1C for the following aqueous
systems: LPS from Agrobacterium tumefaciens at pH 7.4 (open circles);
LPS from Agrobacterium tumefaciens—DOPE (3 : 1) at pH 7.4
(full triangles); LPS from Agrobacterium tumefaciens—DOPE (3 : 1) at
pH 9 (open squares). For a better comparison, data have been multiplied
by a scale factor, as indicated. Fitting curves to the experimental data
through the models reported in the text are also shown.
Table 2 Structural parameters at 25 1C for the liposomes formed by LPS, obtained through the fitting of the model described in the text to SANSdata. The table reports the number of lamellae, N, the average lamellar thickness, d, the Zimm index of the thickness distribution function, Z, themean distance between two consecutive lamellae, dl, and the Hosemann factor, sdl/dl. Finally, for samples containing both cylindrical micelles andliposomes, the relative volume fractions of both kinds of aggregates fcyl and flip are also reported. In the table, LPS are indicated by the name ofthe bacterium from which they are obtained
System N d=nm Z dl=nm sdl=dl fcyl flip
Salmonella minnesotaa 1 4.3 � 0.2Burkholderia cenocepacia pH 7.4 4 � 1 2.6 � 0.2 10.5 � 0.1 17.0 � 0.3 0.59 � 0.01Burkholderia cenocepacia—DOPE pH 7.4 3 � 1 2.4 � 0.6 10.3 � 0.1 26.4 � 0.4 0.55 � 0.01Burkholderia cenocepacia—DOPE pH 9.1 4 � 1 2.8 � 0.2 10.3 � 0.1 23.4 � 0.4 0.81 � 0.02Burkholderia multivorans pH 7.4b 2 � 1 4.9 � 0.2 10.4 � 0.1 30.0 � 0.5 0.60 � 0.01 0.0018 � 0.0002 0.0049 � 0.0005Burkholderia multivorans—DOPE pH 7.4 4 � 1 5.0 � 0.1 10.3 � 0.1 32.0 � 0.5 0.60 � 0.01 0.0019 + 0.0002 0.0048 � 0.0005Burkholderia multivorans—DOPE pH 9.1 5 � 1 4.2 � 0.1 10.5 � 0.1 27.8 � 0.5 0.62 � 0.01 0.0020 � 0.0002 0.0047 � 0.0005Agrobacterium tumefaciens pH 7.4 3 � 1 4.7 � 0.4 10.3 � 0.1 13.7 � 0.5 0.60 � 0.01Agrobacterium tumefaciens—DOPE pH 7.4 5 � 1 4.8 � 0.4 10.3 � 0.1 11.5 � 0.4 0.60 � 0.01Agrobacterium tumefaciens—DOPE pH 9.1 4 � 1 4.5 � 0.2 10.3 � 0.1 16.9 � 0.3 0.60 � 0.01
a Data from ref. 15. b In this sample cylindrical micelles were also present, with radius R = (4.2 � 0.2) nm and length (68 � 2) nm, see text.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13581
of the membrane inner core closer to the polar external layers.
In contrast, 14-PCSL bears the radical label close to the
terminal methyl group of the acyl chain, thus allowing
monitoring the behavior of the more internal region of the
membrane hydrophobic core.
In all the systems analyzed in the present work, the 5-PCSL
spectrum presents a clearly defined axial anisotropy, see Fig. 8,
thus indicating that, in all cases, the mobility of the label in the
region of the bilayer just below the hydrophilic external
surface is strongly reduced. In contrast, all the 14-PCSL
spectra show an almost isotropic three-line signal, indicative
of a rather free motion of the radical label.
In an attempt to quantitatively analyze the spectra, we
determined the order parameter, S. This parameter is related
to the angular amplitudes of motion of the label, which in turn
reflects the motion of the acyl chain segment to which the label
is bound. In particular, S was calculated according to the
relation
S ¼ðTk � T?ÞðTzz � TxxÞ
aN
a0Nð6Þ
where TJ and T> are two phenomenological hyperfine
splitting parameters which can be experimentally determined
for each spin-labeled phospholipid as shown in Fig. 8 (note
that 2T0? ¼ 2T? � 1:6, see ref. 39). Txx and Tzz are the
principal elements of the real hyperfine splitting tensor in the
spin Hamiltonian of the spin-label, which can be measured
from the corresponding single-crystal EPR spectrum and are
reported in the literature (Txx = 6.1 G and Tzz = 32.4 G,
ref. 39). aN and a0N are the isotropic hyperfine coupling
constants for the spin-label in crystal state and in the
membrane, respectively, given by:
aN =1
3(Tzz+2Txx) (7)
a0N ¼
1
3ðTk þ 2T?Þ ð8Þ
The isotropic hyperfine coupling constant is an index of the
micropolarity experienced by the nitroxide; in particular, it
increases with the environmental polarity. The aN=a0N ratio in
eqn (6) corrects the order parameter for polarity differences
between the crystal state and the membrane.
The S and a0N values obtained from the spectra registered at
25 1C are collected in Table 3. In the case of S. minnesota
R-LPS we report the value determined from the 5-PCSL
spectrum reported in literature.15 By comparing the results
obtained for 5-PCSL and 14-PCSL it is possible to observe
that both parameters decrease with increasing the depth at
which the label is inserted in the bilayer, indicating that both
the acyl chains ordering and the environmental polarity
decrease. Concerning variations due to the peculiar LPS, it is
to be observed that the values obtained for 5-PCSL show that
liposomes formed by R-LPS from S. minnesota present values
of order parameter substantially lower than those observed in
the case of B. cenocepacia and A. tumefaciens R-LPS. For
S-LPS from B. multivorans an intermediate value is obtained.
It is interesting to observe that ordering of acyl chains in the
LPS bilayer is higher than that observed for typical phospho-
lipids (S E 0.6 for 5-PCSL39).
The lower S values obtained for 14-PCSL reflect the higher
fluidity of the more internal region of the bilayer hydrophobic
core. Among the various LPS, the same trend found for
5-PCSL is observed. For both spin-labels, Table 3 shows that
the a0N values do not depend on the considered LPS.
In the case of 5-PCSL, we also studied the S and a0N
variation with temperature, see Fig. 9. In all cases, a continuous
decrease was registered, thus indicating that the dynamics of
the lipid chain significantly increases with temperature. In the
case of R-LPS from S. minnesota, perusal of the figure reveals
a very broad sigmoidal trend, with the inflection centered
Fig. 8 EPR spectra of 5-PCSL and 14-PCSL in liposomes formed by
LPS from Burkholderia cenocepacia (A and A0 for 5- and 14-PCSL,
respectively),Burkholderia multivorans (B, B0),Agrobacterium tumefaciens
(C, C0) at pH 7.4 and 25 1C. The magnetic field scan range is 90 G.
Table 3 Microstructural parameters at 25 1C for the liposomesformed by LPS, through the quantitative analysis of EPR datadescribed in the text. The table reports the order parameter, S, andthe isotropic hyperfine coupling constant, a
0N . In the table, LPS are
indicated by the name of the bacterium from which they are obtained
System S a0N=G
5-PCSLSalmonella minnesotaa 0.69 � 0.01 15.4 � 0.1Burkholderia cenocepacia pH 7.4 0.77 � 0.02 15.6 � 0.2Burkholderia multivorans pH 7.4 0.73 � 0.01 15.5 � 0.2Agrobacterium tumefaciens pH 7.4 0.76 � 0.02 15.6 � 0.214-PCSLBurkholderia cenocepacia pH 7.4 0.24 � 0.02 13.6 � 0.2Burkholderia multivorans pH 7.4 0.20 � 0.03 13.5 � 0.2Agrobacterium tumefaciens pH 7.4 0.27 � 0.02 13.6 � 0.2
a Data computed from the spectrum reported in ref. 15.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
13582 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
between 30 and 35 1C; this evidence is related to a transition
from gel- to fluid-phase of the LPS acyl chain arrangement in
the bilayer. For all the other LPS considered, only a slight
slope change is observed in the same temperature range.
Even in this case, no significant a0N variation was observed
(mean value 15.6 � 0.3 G).
4. Discussion
The focus of the present work relies on the relationship
between the LPS molecular structure and the morphology
and behavior of the aggregates they form in aqueous medium.
Indeed, the molecular structure determines the local self-
organization of the molecules, which in turn influences the
aggregate architecture and dynamics. For this reason it is
worthy to analyse in details the main analogies/differences
among the molecular structures of the LPS used, shown in
Fig. 1, before discussing the experimental results.
Two of the LPS considered in the present work, specifically
those derived from B. cenocepacia and A. tumefaciens, lack the
O-chain and therefore are R-LPS. In contrast, B. multivorans
LPS presents an extended O-chain (i.e., it is a S-LPS), whose
primary structure consists of two O-polysaccharide chains
present in different amounts and made up of repeating units
both containing deoxy sugars. In this case, Fig. 1 does not
show the entire S-LPS molecule, but only the lipid A and the
repeating units structures. B. multivorans S-LPS presents the
same lipid A of B. cenocepacia R-LPS, and a similar core
oligosaccharide. Thus, the comparison between the results
obtained for B. cenocepacia R-LPS and B. multivorans
S-LPS allows to discriminate the effect of the O-chain on the
molecules self-aggregation.
The R-LPS derived from A. tumefaciens presents a peculiar
lipid A, because of the presence of an unusually long chain
fatty acid (C28 : 0 (27-OH)), not stoichiometrically esterified
by a 3-hydroxy-butyroyl residue at its hydroxy group.40 Thus,
the analysis of the results obtained for A. tumefaciens R-LPS
highlights the extent to which LPS self-aggregation is
influenced by the lipid A primary structure.
Last, we compare the results obtained for these three LPS
with those already published by us relative to the R-LPS
derived from S. enterica serotype minnesota strain 595
(Re mutant).15 The R-LPS from the Re mutants represents
the deep-rough chemotype with the shortest saccharidic chain
length formed by the lipid A and Kdo as the sole constituent of
the core. Thus, in the present work, the comparison between
the properties of S. minnesota R-LPS aggregates with those
presented by the other LPS gives an indications of the
relevance of the glycolipid portion with respect to the
saccharidic component in determining the molecule self-
aggregation.
However, before starting to discuss the experimental results,
it is to be stressed that all the comparisons presented below
have to be handled with caution, since a single-variable
analysis is not feasible, the LPS molecules differentiating for
more than one feature.
The data collected in the present work indicate that all the
considered LPS form liposomes, whose dimension is coarsely
regulated by pore dimension of the membrane used for the
extrusion procedure. These aggregates, similar to phospholipid
liposomes, are likely to be metastable. However, their
morphological features are constant over long periods
(weeks, at least). The main driving force of LPS self-aggregation
is the hydrophobic interaction among the lipid A acyl chains.
This interaction forces the molecules to self-organize in a
bilayered structure with the acyl chains forming an inner
hydrophobic domain and the saccharidic portion protruding
in the external aqueous medium. The bilayer thickness, d, can
be evaluated by SANS measurements. As discussed in the
Results section, the d values represent the thickness of the
hydrophobic layer constituted by the acyl chains rather than
the whole thickness of LPS bilayer. B. cenocepacia presents the
lower d value with respect to the other LPS (d = 2.6 nm). The
linear extension of the lipophilic part of a single molecule of
lipid A is approximately 1.9 nm, so that a bilayer composed of
two opposing leaflets of LPS molecules with completely
extended acyl chains perpendicular to the disaccharidic back-
bone is expected to have a thickness of at leastB3.8 nm. Thus,
our experimental evidence suggests either a certain degree of
interdigitation between the acyl chains of the two opposing
lipid A leaflets, or a tilting of the acyl chains with respect
to the disaccharidic backbone. EPR measurements allow to
discriminate between these two possibilities; in fact, inter-
digitation reduces the mobility of the more internal segments
of the acyl chains, so that the spectrum of the spin-probe
14-PCSL is expected to present a slower component.41 The
14-PCSL spectrum in liposomes formed by B. cenocepacia
R-LPS presents an almost isotropic lineshape indicating a
relatively fast motion of the chain segments, an evidence of
no interdigitation to occur. Consequently, we conclude that
the low d value obtained by SANS is likely to be due to a
tilting of the acyl chains relative to the disaccharide units
delimiting the bilayer inner hydrophobic domain. A similar
chain arrangement in LPS bilayers was suggested by the
analysis of FTIR and AFM experiments.42,43
Interesting is the comparison of the self-association
behavior of B. cenocepacia R-LPS with the A. tumefaciens
and S. minnesota ones. In the last two cases, the higher bilayer
Fig. 9 Order parameter, S, of 5-PCSL in liposomes formed by LPS
from Burkholderia cenocepacia (full squares), Burkholderia multivorans
(open circles), Agrobacterium tumefaciens (open squares) and
Salmonella minnesota (full circles), at pH 7.4 as a function of the
temperature.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13583
thickness could be related to a different arrangement of the
acyl chains, which probably tend to be perpendicular to
the disaccharide backbone. This change could be induced
by the higher number of acyl chains per disaccharide unit
(6 for S. minnesota R-LPS vs. 5 for B. cenocepacia R-LPS) or
to the presence of the unusually longer fatty acid, in the case of
A. tumefaciens R-LPS. For S. minnesota R-LPS, the different
orientation significantly reduces the acyl chains ordering, as
shown by the S value obtained for 5-PCSL by EPRmeasurements.
In the case of A. tumefaciens R-LPS, the relatively high
ordering of the acyl chains in both 5- and 14- position could
support the hypothesis that the acyl chain of the (C28:0 (27-OH))
fatty acid passes through both leaflets constituting the
bilayer,40 with a consequent mobility reduction. Actually,
the presence of the long acyl chain could be an evolutive
strategy for stabilizing liposomes formed by A. tumefaciens
R-LPS, which are more negatively charged than those derived
from B. cenocepacia and B. multivorans LPS, whose phosphate
groups are partially substituted by a 4-deoxy-4-amino-arabinose
moiety, see Fig. 1.
Even more stimulating is the comparison between the
d values obtained for liposomes formed by B. multivorans
S-LPS and B. cenocepaciaR-LPS, which present the same lipid
A and a similar saccharidic core, and differ only in the
O-chain, which is absent in the latter case. As discussed above,
SANS measurements are almost insensitive to the saccharidic
portions of the bilayer, and consequently one could expect
similar SANS results for the two LPS. However, this is
absolutely not confirmed by the experimental results, showing
a d value much higher for liposomes formed by B. multivorans
S-LPS. This evidence has to be related to an indirect effect of
the O-chain. In consideration of the bulkiness of the saccharidic
portion of B. multivorans S-LPS, strong steric repulsion occurs
among the hydrophilic moieties of adjacent molecules in the
bilayer, thus disturbing the molecules packing. This distortion
propagates to the acyl chains, affecting their arrangement in
the hydrophobic inner layer, reducing the tightness of their
packing and their tilting with respect the bilayer surface,
and finally resulting in a d increase. This interpretation is
supported by the reduction of the S parameter derived from
EPR. Thus, the comparison between the data collected for
B. cenocepacia R-LPS and B. multivorans S-LPS indicates
that the self-aggregation of the glycolipid portions of LPS
molecules is strongly affected by the saccharidic portion, in a
complex interplay of hydrophobic, steric and electrostatic
interactions.
In this context, here we comment on the peculiarity of a
S-LPS (i.e., the LPS derived from B. multivorans) with respect
to R-LPS (all the others considered LPS), in that only in the
first case, according to the DLS measurements, liposomes
coexist with elongated micelles. Coexistence of micelles and
vesicles is well documented in the recent literature. Interestingly,
it is often found in aqueous mixtures of amphiphiles presenting
an extended hydrophilic moiety, such as a poly(ethylene oxide
chain).44–46 Conformation of the PEG coil is considered to be
decisive in establishing the morphology of the supramolecular
aggregates. We propose the polysaccharidic chain of S-LPS to
play a similar role. For steric reasons, a ‘‘globular’’ coil would
favor formation of spherical micelles, while a more extended
chain conformation is required in bilayered structures. Thus our
evidence has to be ascribed to the bulkiness of the hydrophilic
portion of the B. multivorans S-LPS molecule, which leads to an
increased curvature of the aggregate surface, finally resulting in
the formation of rodlike micellar aggregates.
Interestingly, the 5-PCSL EPR spectrum in aggregates
formed by S-LPS derived from B. multivorans presents the
same line shape observed for the other LPS, with no super-
position of signals deriving from micellar aggregates. This
evidence, already observed for other lipids,47 suggests that
the dynamic state of the lipid acyl chains is preserved, despite
the change in morphology.
Coming back to a general discussion on liposomes formed
by LPS molecules, local ordering of the acyl chains is expected
to be affected by temperature. In the present work this aspect
was analyzed by EPR. With increasing temperature, the
fluidity of the bilayer increases, as highlighted by the
S decrease. In the case of the R-LPS deriving from S. minnesota,
the S values decrease according to a broad sigmoidal trend,
with an inflection at 30–35 1C. This trend could be related to a
transition of the acyl chains from a gel to a liquid crystalline
phase (a - b), which has been largely reported in the
literature.5 This transition is much less steep than that
observed for typical phospholipids.13,47 Furthermore, our data
indicate that it becomes even smoother in the case of LPS
derived from B. cenocepacia, B. multivorans and A. tumefaciens,
for which only a slight slope change is observed. The reason
why in the case of LPS the transition is so smooth, could be the
polydispersity of acyl chain length, and their different
positioning with respect to the saccharidic headgroup.
For all the considered LPS, we have observed the tendency
of the liposomes to re-arrange forming multilayered
structures. This is unequivocally shown by SANS data, even
though, because of the quality of the data and of limitations of
the adopted fitting model, the number of concentric lamellae
can be only qualitatively estimated. The extrusion technique
we used in the present work is considered a reliable method to
produce unilamellar vesicles,48 even though in a limited number
of cases it has also been found to lead to formation of
multilamellar aggregates. This could depend on lipid
composition49,50 or on external conditions, such as solvent
composition.51 Concerning LPS, we tentatively propose liposome
multilamellarity to be related to the favorable interaction
among saccharidic surfaces of concentric bilayers, probably
connected to the formation of an extended network of
H-bonds. Interestingly, a similar interpretation has been
recently proposed to explain why glycolipids tend to form
multi- and not unilamellar structures.52
Last, we comment on the effect of inclusion of DOPE in
liposomes formulation. Generally speaking, the effects are not
dramatic. The only relevant DOPE effect was registered in the
case of S-LPS derived from B. multivorans. In this case, the
addition of DOPE leads to an increase of the vesicles
population (see Fig. 3), as put in evidence by DLS results.
Furthermore, the reduction of the micelles population is such
that they are not detected by SANS measurements. These
experimental evidences indicate that the bilayer is stabilized by
the insertion of DOPE, which causes a reduction of the steric
repulsion among the LPS headgroups.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
13584 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 This journal is c the Owner Societies 2010
In conclusion, the results presented in this work have high-
lighted the relationship between the molecular structure of
some LPS molecules and the structure of their aggregates.
Interestingly, the relationship is often not obvious, in that the
aggregate architecture and dynamics are the final result of a
complex interplay of all the possible interactions (hydrophobic,
electrostatic, steric) set up among the LPS molecules. In
biological context, our results suggest that the rich biodiversity
of LPS molecular structure could be fundamental to finely
tuning the structure and functional properties of the outer
membrane of Gram negative bacteria and, consequently, their
biological behavior.
Acknowledgements
The authors thank CSGI (Consorzio Interuniversitario per lo
sviluppo dei Sistemi a Grande Interfase) and MIUR (PRIN
2007) for financial support. B. multivorans cells were kindly
furnished by Dr Paola Cescutti (Universita di Trieste),
B. cenocepacia cells were kindly furnished by Dr Anthony
De-Soyza (University of Newcastle) and R-LPS from
A. tumefaciens was a kind gift from Dr Cristina De Castro
(Universita di Napoli). Forschungszentrum Julich is
acknowledged for provision of beam time. SANS experiments
were supported by the European Commission, NMI3 contract
RII3-CT-2003-505925. The authors thank Prof. Lucia
Costantino for her helpful comments. Finally, we thank the
referees whose comments helped improve the manuscript.
References
1 N. Ruiz, D. Kahne and J. Silhavy Thomas, Advances inunderstanding bacterial outer-membrane biogenesis, Nat. Rev.Microbiol., 2006, 4(1), 57–66.
2 A. Silipo, C. De Castro, R. Lanzetta, M. Parrilli and A. Molinaro,in Prokaryotic Cell Wall Compounds: Structure and Biochemistry,ed. H. Konig, H. Claus and A. Varma, Springer, Heidelberg, 2010.
3 O. Holst and A. Molinaro, in Microbial Glycobiology,ed. A. P. Moran, Elsevier, London, 2009.
4 U. Seydel, A. J. Ulmer, S. Uhlig and E. T. Rietschel, in MemnrnStructure in Disease and Drug Therapy, ed. G. Zimmer, MarcelDekker Inc., New York, 1999.
5 K. Brandenburg and U. Seydel, A comment on the preparation ofliposomes from and on the beta .tautm. alpha acyl chain meltingbehavior of rough mutant lipopolysaccharide, Biochim. Biophys.Acta, Biomembr., 1991, 1069(1), 1–4.
6 K. Nixdorff, J. Gmeiner and H. H. Martin, Interaction oflipopolysaccharide with detergents and its possible role in thedetergent resistance of the outer membrane of Gram-negativebacteria, Biochim. Biophys. Acta, Biomembr., 1978, 510(1), 87–98.
7 T. Gutsmann, A. B. Schromm, M. H. J. Koch, S. Kusumoto,K. Fukase, M. Oikawa, U. Seydel and K. Brandenburg,Lipopolysaccharide-binding protein-mediated interaction oflipid A from different origin with phospholipid membranes,Phys. Chem. Chem. Phys., 2000, 2(20), 4521–4528.
8 K. Brandenburg, M. Matsuura, H. Heine, M. Muller, M. Kiso,H. Ishida, M. H. J. Koch and U. Seydel, Biophysical characterizationof triacyl monosaccharide lipid A partial structures in relation tobioactivity, Biophys. J., 2002, 83(1), 322–333.
9 H. Labischinski, E. Vorgel, W. Uebach, R. P. May andH. Bradaczek, Architecture of bacterial lipid A in solution.A neutron small-angle scattering study, Eur. J. Biochem., 1990,190(2), 359–63.
10 E. Urban, A. Bota and B. Kocsis, Non-bilayer formation in theDPPE-DPPG vesicle system induced by deep rough mutant ofSalmonella minnesota R595 lipopolysaccharide, Colloids Surf., B,2006, 48(2), 106–11.
11 K. Nomura, T. Inaba, K. Morigaki, K. Brandenburg, U. Seydeland S. Kusumoto, Interaction of lipopolysaccharide and phospho-lipid in mixed membranes: solid-state 31P-NMR spectroscopic andmicroscopic investigations, Biophys. J., 2008, 95(3), 1226–1238.
12 N. C. Santos, A. C. Silva, M. A. R. B. Castanho, J. Martins-Silvaand C. Saldanha, Evaluation of lipopolysaccharide aggregation bylight scattering spectroscopy, ChemBioChem, 2003, 4(1), 96–100.
13 E. Urban, A. Bota, B. Kocsis and K. Lohner, Distortion of thelamellar arrangement of phospholipids by deep rough mutantlipopolysaccharide from Salmonella minnesota, J. Therm. Anal.Calorim., 2005, 82(2), 463–469.
14 M. F. Henning, S. Sanchez and L. Bakas, Visualization andanalysis of lipopolysaccharide distribution in binary phospholipidbilayers, Biochem. Biophys. Res. Commun., 2009, 383(1), 22–26.
15 G. D’Errico, A. Silipo, G. Mangiapia, A. Molinaro, L. Paduanoand R. Lanzetta, Mesoscopic and microstructural characterizationof liposomes formed by the lipooligosaccharide from Salmonellaminnesota strain 595 (Re mutant), Phys. Chem. Chem. Phys., 2009,11(13), 2314–2322.
16 R. F. Epand, P. B. Savage and R. M. Epand, Bacterial lipidcomposition and the antimicrobial efficacy of cationic steroidcompounds (Ceragenins), Biochim. Biophys. Acta, Biomembr.,2007, 1768(10), 2500–2509.
17 T. Ierano, A. Silipo, P. Cescutti, M. R. Leone, R. Rizzo,R. Lanzetta, M. Parrilli and A. Molinaro, Structural study andconformational behavior of the two different lipopolysaccharideO-antigens produced by the cystic fibrosis pathogen Burkholderiamultivorans, Chem.–Eur. J., 2009, 15(29), 7156–7166, S7156/1–S7156/4.
18 A. Silipo, A. Molinaro, T. Ierano, A. De Soyza, L. Sturiale,D. Garozzo, C. Aldridge, P. A. Corris, C. M. A. Khan,R. Lanzetta and M. Parrilli, The complete structure and pro-inflammatory activity of the lipooligosaccharide of the highlyepidemic and virulent gram-negative bacterium Burkholderiacenocepacia ET-12 (strain J2315), Chem.–Eur. J., 2007, 13(12),3501–3511, S3401/1–23501/9.
19 C. De Castro, A. Carannante, R. Lanzetta, V. Liparoti,A. Molinaro and M. Parrilli, Core oligosaccharide structure fromthe highly phytopathogenic Agrobacterium tumefaciens TT111and conformational analysis of the putative rhamnan epitope,Glycobiology, 2006, 16(12), 1272–1280.
20 M. Ishinaga, R. Kanamoto and M. Kito, Distribution of phospho-lipid molecular species in outer and cytoplasmic membranes ofEscherichia coli, Journal of Biochemistry, 1979, 86(1), 161–5.
21 A. K. Covington, F. S. Bates and R. A. Durst, Definition ofpH scales, standard reference values, measurement of pH andrelated terminology, Pure Appl. Chem., 1985, 57(3), 531–42.
22 R. Peters, ALV-5000/E/EPP% ALV-60X0 for WINDOWS-95/NT4.0 Software, 3.0.1.12, 2003.
23 G. A. Brehm and V. A. Bloomfield, Analysis of polydispersity inpolymer solutions by inelastic laser light scattering, Macro-molecules, 1975, 8(5), 663–5.
24 G. D. Wignall and F. S. Bates, Absolute calibration of small-angleneutron scattering data, J. Appl. Crystallogr., 1987, 20(1), 28–40.
25 T. P. Russell, J. S. Lin, S. Spooner and G. D. Wignall, Inter-calibration of small-angle X-ray and neutron scattering data,J. Appl. Crystallogr., 1988, 21(6), 629–38.
26 H. J. V. Tyrrell and K. R. Harris, Diffusion in Liquids:A Theoretical and Experimental Study, Butterworths, London, 1984.
27 M. Kotlarchyk and S. H. Chen, Analysis of small angle neutronscattering spectra from polydisperse interacting colloids, J. Chem.Phys., 1983, 79(5), 2461–9.
28 M. Vaccaro, R. Del Litto, G. Mangiapia, A. M. Carnerup,G. D’Errico, F. Ruffo and L. Paduano, Lipid based nanovectorscontaining ruthenium complexes: a potential route in cancertherapy, Chem. Commun., 2009, (11), 1404–1406.
29 H. Frielinghaus, Small-angle scattering model for multilamellarvesicles, Physical Review E: Statistical, Nonlinear, and Soft MatterPhysics, 2007, 76(5–1), 051603/1–051603/8.
30 M. Kotlarchyk and S. M. Ritzau, Paracrystal model of the high-temperature lamellar phase of a ternary microemulsion system,J. Appl. Crystallogr., 1991, 24(5), 753–8.
31 M. Vaccaro, G. Mangiapia, L. Paduano, E. Gianolio, A.Accardo, D. Tesauro and G. Morelli, Structural and relaxometriccharacterization of peptide aggregates containing gadolinium
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online
This journal is c the Owner Societies 2010 Phys. Chem. Chem. Phys., 2010, 12, 13574–13585 13585
complexes as potential selective contrast agents in MRI,ChemPhysChem, 2007, 8(17), 2526–2538.
32 M. Vaccaro, A. Accardo, D. Tesauro, G. Mangiapia, D. Loef,K. Schillen, O. Soederman, G. Morelli and L. Paduano,Supramolecular Aggregates of Amphiphilic GadoliniumComplexes as Blood Pool MRI/MRA Contrast Agents: Physico-chemical Characterization, Langmuir, 2006, 22(15), 6635–6643.
33 A. Radulescu, R. T. Mathers, G. W. Coates, D. Richter andL. J. Fetters, A SANS Study of the Self-Assembly in Solution ofSyndiotactic Polypropylene Homopolymers, Syndiotactic Poly-propylene-block-poly(ethylene-co-propylene) Diblock Copolymers,and an Alternating Atactic-Isotactic Multisegment Polypropylene,Macromolecules, 2004, 37(18), 6962–6971.
34 S. Hansen, Translational friction coefficients for cylinders ofarbitrary axial ratios estimated by Monte Carlo simulation,J. Chem. Phys., 2004, 121(18), 9111–9115.
35 D. Marsh, Structural and thermodynamic determinants of chain-melting transition temperatures for phospholipid and glycolipidsmembranes, Biochim. Biophys. Acta, Biomembr., 2010, 1798(1),40–51.
36 A. Lange, D. Marsh, K. H. Wassmer, P. Meier and G. Kothe,Electron spin resonance study of phospholipid membranesemploying a comprehensive line-shape model, Biochemistry,1985, 24(16), 4383–92.
37 M. Moser, D. Marsh, P. Meier, K. H. Wassmer and G. Kothe,Chain configuration and flexibility gradient in phospholipidmembranes. Comparison between spin-label electron spin resonanceand deuteron nuclear magnetic resonance, and identification ofnew conformations, Biophys. J., 1989, 55(1), 111–23.
38 W. L. Hubbell and H. M. McConnell, Molecular motion in spin-labeled phospholipids and membranes, J. Am. Chem. Soc., 1971,93(2), 314–26.
39 L. M. Gordon and R. D. Sauerheber, Studies on spin-labeled egglecithin dispersions, Biochim. Biophys. Acta, Biomembr., 1977,466(1), 34–43.
40 A. Silipo, C. de Castro, R. Lanzetta, A. Molinaro and M. Parrilli,Full structural characterization of the lipid A components fromthe Agrobacterium tumefaciens strain C58 lipopolysaccharidefraction, Glycobiology, 2004, 14(9), 805–815.
41 I. Plasencia, F. Baumgart, D. Andreu, D. Marsh and J. Perez-Gil,Effect of acylation on the interaction of the N-Terminal segment of
pulmonary surfactant protein SP-C with phospholipid membranes,Biochim. Biophys. Acta, Biomembr., 2008, 1778(5), 1274–1282.
42 U. Seydel, M. Oikawa, K. Fukase, S. Kusumoto andK. Brandenburg, Intrinsic conformation of lipid A is responsiblefor agonistic and antagonistic activity, Eur. J. Biochem., 2000,267(10), 3032–3039.
43 S. Roes, U. Seydel and T. Gutsmann, Probing the Properties ofLipopolysaccharide Monolayers and Their Interaction with theAntimicrobial Peptide Polymyxin B by Atomic Force Microscopy,Langmuir, 2005, 21(15), 6970–6978.
44 F. Li, S. Prevost, R. Schweins, A. T. M. Marcelis, F. A. M.Leermakers, M. A. Cohen Stuart and E. J. R. Sudhoelter,Small monodisperse unilamellar vesicles from binary copolymermixtures, Soft Matter, 2009, 5(21), 4169–4172.
45 C. Leal, S. Roegnvaldsson, S. Fossheim, E. A. Nilssen andD. Topgaard, Dynamic and structural aspects of PEGylatedliposomes monitored by NMR, J. Colloid Interface Sci., 2008,325(2), 485–493.
46 M. Vaccaro, G. Mangiapia, A. Radulescu, K. Schillen,G. D’Errico, G. Morelli and L. Paduano, Colloidal particlescomposed of amphiphilic molecules binding gadolinium complexesand peptides as tumor-specific contrast agents in MRI: physico-chemical characterization, Soft Matter, 2009, 5(13), 2504–2512.
47 G. D’Errico, A. M. D’Ursi and D. Marsh, Interaction of a PeptideDerived from Glycoprotein gp36 of Feline ImmunodeficiencyVirus and Its Lipoylated Analogue with Phospholipid Membranes,Biochemistry, 2008, 47(19), 5317–5327.
48 B. Mui, L. Chow and M. J. Hope, Extrusion technique togenerate liposomes of defined size, Methods Enzymol., 2003,367(Liposomes, Part A), 3–14.
49 P. Fromherz, Lipid-vesicle structure: size control by edge-activeagents, Chem. Phys. Lett., 1983, 94(3), 259–66.
50 N. E. Gabriel and M. F. Roberts, Spontaneous formation of stableunilamellar vesicles, Biochemistry, 1984, 23(18), 4011–15.
51 M. A. Kiselev, P. Lesieur, A. M. Kisselev, D. Lombardo,M. Killany and S. Lesieur, Sucrose solutions as prospectivemedium to study the vesicle structure: SAXS and SANS study,J. Alloys Compd., 2001, 328(1–2), 71–76.
52 J. Howe, M. Von Minden, T. Gutsmann, M. H. J. Koch, M. Wulf,S. Gerber, G. Milkereit, V. Vill and K. Brandenburg, Structuralpreferences of dioleoyl glycolipids with mono- and disaccharidehead groups, Chem. Phys. Lipids, 2007, 149(1–2), 52–58.
Dow
nloa
ded
by J
OIN
T I
LL
- E
SRF
LIB
RA
RY
on
18 O
ctob
er 2
010
Publ
ishe
d on
20
Sept
embe
r 20
10 o
n ht
tp://
pubs
.rsc
.org
| do
i:10.
1039
/C0C
P000
66C
View Online