Asymmetric phospholipid: lipopolysaccharide bilayers; A Gram-negative bacterial outer membrane mimic

Abstract

The Gram-negative bacterial outer membrane (OM) is a complex and highly asymmetric biological barrier but the small size of bacteria has hindered advances in in vivo examination of membrane dynamics. Thus, model OMs, amenable to physical study, are important sources of data. Here, we present data from asymmetric bilayers which emulate the OM and are formed by a simple two-step approach. The bilayers were deposited on an SiO2 surface by Langmuir-Blodgett deposition of phosphatidylcholine as the inner leaflet and, via Langmuir-Schaefer deposition, an outer leaflet of either Lipid A or Escherichia coli rough lipopolysaccharides (LPS). The membranes were examined using neutron reflectometry (NR) to examine the coverage and mixing of lipids between the bilayer leaflets. NR data showed that in all cases, the initial deposition asymmetry was mostly maintained for more than 16 h. This stability enabled the sizes of the headgroups and bilayer roughness of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine and Lipid A, Rc-LPS and Ra-LPS to be clearly resolved. The results show that rough LPS can be manipulated like phospholipids and used to fabricate advanced asymmetric bacterial membrane models using well-known bilayer deposition techniques. Such models will enable OM dynamics and interactions to be studied under in vivo-like conditions.

Full text

rsif.royalsocietypublishing.org
Research
Cite this article: Clifton LA, Skoda MWA,
Daulton EL, Hughes AV, Le Brun AP, Lakey JH,
Holt SA. 2013 Asymmetric phospholipid:
lipopolysaccharide bilayers; a Gram-negative
bacterial outer membrane mimic. J R Soc
Interface 10: 20130810.
http://dx.doi.org/10.1098/rsif.2013.0810
Received: 2 September 2013
Accepted: 25 September 2013
Subject Areas:
biophysics, synthetic biology
Keywords:
lipopolysaccharide, Gram-negative bacterial
outer membrane, neutron reflection, rough
mutant lipopolysaccharides, isotopic labelling
Author for correspondence:
Luke A. Clifton
e-mail: luke.clifton@stfc.ac.uk
Electronic supplementary material is available
at http://dx.doi.org/10.1098/rsif.2013.0810 or
via http://rsif.royalsocietypublishing.org.
Asymmetric phospholipid:
lipopolysaccharide bilayers; a
Gram-negative bacterial outer
membrane mimic
Luke A. Clifton1, Maximilian W. A. Skoda1, Emma L. Daulton1,2, Arwel
V. Hughes1, Anton P. Le Brun3, Jeremy H. Lakey4and Stephen A. Holt3
1
ISIS Pulsed Neutron and Muon Source, Science and Technology Facilities Council, Rutherford Appleton
Laboratory, Harwell, Oxfordshire OX11 OQX, UK
2
Department of Chemistry, University of Bath, Bath BA2 7AY, UK
3
Bragg Institute, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC,
New South Wales 2232, Australia
4
Institute for Cell and Molecular Biosciences, Newcastle University, Framlington Place, Newcastle upon Tyne
NE2 4HH, UK
The Gram-negative bacterial outer membrane (OM) is a complex and highly
asymmetric biological barrier but the small size of bacteria has hindered
advances in in vivo examination of membrane dynamics. Thus, model OMs,
amenable to physical study, are important sources of data. Here, we present
data from asymmetric bilayers which emulate the OM and are formed by a
simple two-step approach. The bilayers were deposited on an SiO
2
surface
by Langmuir– Blodgett deposition of phosphatidylcholine as the inner leaflet
and, via Langmuir–Schaefer deposition, an outer leaflet of either Lipid A
or Escherichia coli rough lipopolysaccharides (LPS). The membranes were
examined using neutron reflectometry (NR) to examine the coverage and
mixing of lipids between the bilayer leaflets. NR data showed that in all
cases, the initial deposition asymmetry was mostly maintained for more
than 16 h. This stability enabled the sizes of the headgroups and bilayer rough-
ness of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine and Lipid A, Rc-LPS
and Ra-LPS to be clearly resolved. The results show that rough LPS can be
manipulated like phospholipids and used to fabricate advanced asymmetric
bacterial membrane models using well-known bilayer deposition techniques.
Such models will enable OM dynamics and interactions to be studied under
in vivo-like conditions.
1. Introduction
Bacteria are differentiated into two main groups, Gram-positive or Gram-
negative, based on a technique which detects the thick peptidoglycan cell
wall characteristic of Gram-positive bacteria. Gram-negative bacteria are of
particular biomedical, technological interest owing to their increasing antibiotic
resistance and their utility in many biotechnological processes. The most com-
monly known example is Escherichia coli, found naturally in our digestive
system and extensively used in biomedical research and industry. However,
some strains may cause food poisoning, septicaemia or meningitis while, in
developing countries, it remains a major cause of infant mortality. Furthermore,
this group also includes Klebsiella (hospital-acquired infections), Legionella
(Legionnaires’ disease), Neisseria (meningitis and gonorrhoea), Pseudomonas
(lung infections in cystic fibrosis patients) and even Yersinia pestis (bubonic
plague). However, just as Legionella was unknown until recently, previously
unnoticed Gram-negative bacteria such as Acinetobacter are now a significant
&2013 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution
License http://creativecommons.org/licenses/by/3.0/, which permits unrestricted use, provided the original
author and source are credited.

threat to hospital patients and are rapidly acquiring multiple
antibiotic resistances. As the outer membrane (OM) presents
an additional barrier to antibiotics entering Gram-negative
bacteria, biophysical and structural studies are of significant
interest [1].
The Gram-negative OM resembles most biological membra-
nes being a lipid bilayer with embedded membrane proteins,
however it is extremely asymmetric [2]. In lipid terms, the
inner, cytoplasmic, membrane of Gram-negative bacteria is
composed predominately of phospholipids, in particular
phosphatidylethanolamine and phosphatidylglycerol, as well
as cardiolipin [3]. The OM has a phospholipid-rich inner leaflet,
with a similar composition to the cytoplasmic membrane how-
ever, the outer leaflet which faces the extracellular environment,
is predominantly composed of lipopolysaccharides (LPSs) [2].
LPSs are complex molecules which can be considered to
consist of three parts. Lipid A, which is a phosphorylated diglu-
cosamine (di-GlcN) molecule with covalently attached acyl
chains, which anchors the LPS molecule to the hydrophobic
interior of the OM. Attached to the glucosamine (GLcN) head-
group of Lipid A and facing the outer surface is the core
oligosaccharide region, which can be further broken down into
the inner and outer core. The inner core is composed of the
sugars 3-deoxy-D-manno-octulsonic acid (Kdo) and L-glycero-D-
manno-heptose (Hep) and the outer core region is composed of
sugars such as hexoses and hexosamines. Attached to the core
is the O-antigen region, the largest part of LPS and composed
of a repeating chain of oligosaccharides with high variability
across bacterial strains [4,5]. The charge on the Gram-negative
bacterial OM (GNB-OM) surface is negative owing to the high
levels of phosphorylation of both the GLcNs on Lipid A and
the Kdo and Hep groups in the inner core [5].
In Gram-negative bacteria, LPS may be present in the
smooth or rough form. Smooth LPS contains the complete
core oligosaccharide and O-antigen regions. Bacterial colo-
nies which possess these LPS types form visibly smooth
colonies on agar plates, hence the name. Colonies of bacteria
expressing types of LPS which do not contain the O-antigen
region with either complete or truncated core oligosaccharide
regions appear roughened and are termed rough mutants [6].
Rough mutant LPSs are obtained from mutated bacteria
which are, in general, not found in nature but are viable,
with the genes which encode for LPS formation altered to
produce a truncated LPS in the OM outer leaflet [7].
Previous studies have examined the structure of model
Gram-negative bacterial membranes composed of deep rough,
rough and smooth LPS in bilayer structures composed of LPS
only or LPS/phospholipid mixtures. Studies have ranged
from examining the formation, structure and physiochemical
properties of LPS-containing vesicles in solution [8,9] and in
dry and hydrated powders [10]. Neutron and X-ray diffraction
studies have been used to examine suspensions and stacked
bilayers of both smooth and rough LPS types [11,12]. Studies
on deep rough LPSs in monolayers at the air–liquid interface
have revealed the effect of divalent cations on the packing and
interaction of antimicrobial peptides with these interfacial
films [13], and recently it was shown that Rc-LPS which pos-
sesses a significant portion of the LPS core oligosaccharide
region could be deposited at the air–liquid interface as stable
monolayers [14]. Schneck et al. [15] were able to deposit
smooth LPS monolayers onto a hydrophobically modified sili-
con surface, using these monolayers to examine the effect of
Ca
2þ
on the conformation of the O-antigen.
Here, we have created and examined model Gram-
negative bacterial membranes similar to the GNB-OM in
both lipid components and asymmetry. These GNB-OM
mimics were single bilayers deposited on the surface of silicon
crystals. The GNB-OM has a phosphatidylethanolamine-rich
inner leaflet and to mimic this zwitterionic phospholipid
1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) was
deposited as the first layer [2,3]. The LPS outer leaflet was
composed of either E. coli Lipid A, or the rough mutant
LPSs, Rc-LPS or Ra-LPS.
Escherichia coli Lipid A is the smallest LPS used in these
studies and contains six saturated acyl chains attached to a
GlcN headgroup [5]. In addition to Lipid A, E. coli Rc-LPS
contains a significant proportion of the Hep, glucose (Glc),
galactose (Gal) and Kdo of the LPS core oligosaccharide
region, whereas Ra-LPS contains the complete inner core
region [10] (figure 1).
Neutron reflectometry (NR) was used to examine the
structure normal to the interface of asymmetric bilayers
core oligosaccharide region
Ra-LPS
Hep
Hep
Hep
PO4
–2
PO4
–2
PO4
–2
PO3
–
PO3
–
NH3
+
Hep Kdo
Kdo GlcN
GlcN
Gal
Glc Glc Glc
O-antigen
o
o
oo
oo
oo
o
oo
o
o
Lipid-ARc-LPS
Figure 1. A cartoon representation of the structure of E. coli lipopolysaccharide, showing the lipid tails and sugar groups. The limiting regions corresponding to Lipid A
and Rc and Ra-LPS from rough mutant bacterial strains are shown. Within this general model there are small variations of the core region sugars and phosphates [5].
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
2

deposited on the silicon surface. This study examines the
structural asymmetry and stability of three differing model
OMs with increasing core oligosaccharide size, and therefore
increasing compositional similarity to smooth GNB-OM.
The present systems biology investigation provides a hol-
istic approach for the detailed study of realistic models of the
GNB-OM ranging from the synthesis and assembly on a well-
defined surface to the precise quantitative determination of
structure and composition via multi-contrast fitting of NR
data. Our realistic model was relatively easily obtained and
owing to this it can serve as a platform for more advanced
studies of the GNB-OM at the molecular level, such as
interaction/binding studies, transport, complexation, kinetic
studies to name a few.
2. Material and methods
2.1. Materials
Lipid A (diphosphoryl from E. coli F583), Rc mutant rough strain
LPS (Rc-LPS, from J5 E. coli) and Ra mutant rough strain LPS
(Ra-LPS, from EH100 E. coli) were obtained from Sigma-Aldrich
(Dorset, UK). DPPC and tail-deuterated DPPC (d-DPPC, 1,2-
dipalmitoyl(d62)-sn-glycero-3-phosphocholine) were obtained
from Avanti polar lipids (Alabaster, AL, USA). All phospholipid
and LPS samples were used without further purification. All
other chemicals were sourced from Sigma-Aldrich.
2.2. Asymmetric bilayer deposition
Model Gram-negative bacterial membranes were deposited on
the Piranha-cleaned (SiO
2
) surface of single silicon crystals
using a purpose-built Langmuir–Blodgett (LB) trough (KSV-
Nima, Biolin Scientific, Finland) [16]. LB deposition was used
to deposit the inner leaflet of the membrane on the silicon surface
and Langmuir–Schaefer (LS) deposition used for the outer leaflet
[17] (for a pictorial description, see the electronic supplementary
material, figure S1). For the LB deposition of the inner bilayer
leaflet, tail-hydrogenated DPPC (h-DPPC) or d-DPPC was
deposited from chloroform onto a clean air – liquid interface of
non-buffered water and compressed to a surface pressure of
27 mN m
21
. A submerged silicon crystal was then lifted through
the air–water interface at a speed of 3 mm min
21
while surface
pressure was kept constant. The LB trough was then cleaned
and an air/liquid interfacial monolayer of Lipid A, Rc-LPS or
Ra-LPS was deposited on the water surface ( from 60% CH
3
Cl,
39% MeOH and 1% H
2
O v/v) [14] and compressed to
27 mN m
21
. The LS deposition of the bilayer outer leaflet was
achieved by placing the silicon crystal containing the LB-
deposited DPPC monolayer in a holder directly above the
air– liquid interface of the LB trough. The angle of crystal adjusted
using a purpose-built levelling device to make crystal face parallel
to the water surface. The silicon crystal (and LB film) was then
dipped through the interface at a constant speed of 3 mm min
21
and lowered into a purpose-built sample cell in the well of
the trough. All bilayer deposition took place under ambient
conditions and without subphase buffering until NR analysis.
Initially, 27 mN m
21
was chosen as the monolayer deposi-
tion pressure for the fabrication of the bilayers as DPPC is in the
condense phase at this surface pressure under the ambient con-
ditions [18,19]. It was discovered that high coverage bilayers of
asymmetrically deposited bilayer of DPPC (inner leaflet) and
Lipid A (outer leaflet) (DPPC :Lipid A) could be deposited with
both the inner and outer bilayer leaflets deposited at 27 mN m
21
(see results section). Therefore, this pressure was then used for
the deposition of all bilayer samples described here.
2.3. Neutron reflectometry measurements
Specular NR measurements were carried out using the INTER
[20], SURF [21] and CRISP [22] time-of-flight reflectometers
at the Rutherford Appleton Laboratory (Oxfordshire, UK),
using neutron wavelengths from 0.5 to 6.5 A
˚
for CRISP, 0.5
to 6.8 A
˚
for SURF and 1 to 16 A
˚
for INTER. The reflected inten-
sity is measured as a function of the momentum transfer, Q
z
(Q
z
¼(4
p
sin
u
)/
l
, where
l
is wavelength and
u
is the incident
angle). The collimated neutron beam was reflected from the
silicon–liquid interface at different glancing angles of incidence,
being 0.358, 0.88and 1.88(for CRISP), 0.358, 0.658and 1.58( for
SURF) and 0.78and 2.38(for INTER).
Purpose-built liquid flow cells for analysis of the silicon– liquid
interface were placed on a variable angle sample stage in the
NR instrument and the inlet to the liquid cell was connected to
a liquid chromatography pump (L7100 HPLC pump, Merck,
Hitachi), which allowed for easy exchange of the solution isotopic
contrast within the (3 ml volume) solid–liquid sample cell. For
each solution isotopic contrast change, a total of 22.5 ml of 20 mM
pH/D 7.0 sodium phosphate buffer solution was pumped through
the cell (7.5 cell volumes) at a speed of 1.5 ml min
21
. This was found
by examination of the NR data to completely exchange the solution
in the cell from one isotopic contrast to another. Each solution con-
trast was run in duplicate with the repeat analysis taken at 16 h
intervals. This was conducted to check the stability of the bilayer
over time and under periodic flow (due to changing the solution
contrast within the solid–liquid flow cell).
2.4. Neutron reflectometry data analysis
Reflectivity profiles were obtained from series of samples, which
were chemically similar but differed in the isotopic (deuterium)
composition of either aqueous or lipid contents. Specifically, the
isotopic contrast series contained data from two bilayers which
differed in phospholipid isotopic contrast labelling (one h-DPPC
labelled and another d-DPPC labelled) which were measured
under three-solution isotopic contrasts yielding a total of six
different reflectivity profiles for each model membrane.
As there is no isotopic contrast between the tails of the hydro-
genated phospholipid and the hydrogenated LPS (table 1), there
is no way of determining the contribution of each individual
component to the bilayer structure if only hydrogenous com-
ponents are examined. However, owing to the large difference
in neutron scattering length density (SLD,
r
) between hydrogen-
ated and deuterated alkyl chains the use of deuterated and
hydrogenated lipids within the same bilayer can highlight asym-
metry in the inner and outer leaflet composition and allow for
the structural parameters from the lipid tails in individual bilayer
leaflets to be determined [26]. As DPPC is relatively easily
obtained in its deuterium labelled form from a commercial sup-
plier, this was used as the deuterated lipid component in the
work described here. The LPS was hydrogenous (i.e. natural
abundance) material. Table 1 gives a list of the neutron SLD of
the components used in this study.
The DPPC : LPS bilayers were examined under three-solution
isotopic contrast conditions which were used to highlight the
different components of the bilayer structure. Reflectivity profiles
were obtained with a solution subphase of D
2
O (99.9%,
r
of
6.35 10
26
A
˚
22
), silicon scattering length density matched
water (SMW, 38% D
2
O: 62% H
2
O v/v;
r
¼2.07 10
26
A
˚
22
)
and water (
r
¼20.56 10
26
A
˚
22
).
Neutron reflectivity profiles were simultaneously analysed
using RasCal [27], which employs an optical matrix formalism
(described in detail by Born & Wolf [28]) to fit layer models to
the interfacial structure. In this approach, the interface is
described as a series of slabs, each of which is characterized by
its SLD, thickness and roughness. The reflectivity for the model
starting point is then calculated and compared with the
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
3

experimental data. A least-squares minimization is used to adjust
the fit parameters to reduce the differences between the model
reflectivity and the data. In all cases, the simplest possible
model (i.e. least number of parameters (layers)), which ade-
quately described the data, was selected. NR profiles obtained
from samples under differing solution isotopic conditions were
constrained to fit to the same layer and thickness profile with
SLD varied between datasets as required.
The fitted results of reflectivity data obtained from d-DPPC-
labelled bilayers (in particular the tail layer SLDs) at three differing
solution H
2
O/D
2
O mixtures (100% H
2
O, 38% D
2
O and 100% D
2
O)
were used to determine the relative contribution of the three mem-
brane components, DPPC, LPS and water, to the inner leaflet
(closest to the silicon surface) and outer leaflet (furthest from the
Si surface) tails of the bilayer using a set of linear equations. The
three individual components of a fitted layer within the bilayer
will contribute to the SLD of this layer as shown in equation (2.1)
r
¼ð
r
ðDPPCÞ
w
ðDPPCÞÞþð
r
ðLPSÞ
w
ðLPSÞÞþð
r
ðwaterÞ
w
ðwaterÞÞ;ð2:1Þ
where
r
is the SLD of a given layer.
r
(DPPC)
,
r
(LPS)
and
r
(water)
are the individual SLDs of the DPPC, LPS and solvent, respec-
tively (values are given for these in table 1) and
w
(DPPC)
,
w
(LPS)
and
w
(water)
are the volume fractions of these components within
a particular layer. In the tail regions of the bilayers,
w
(water)
can
be determined by the difference in
r
of the lipid tail layers in
H
2
O, D
2
O and SMW solvent contrasts, which will be owing to
the water contribution to this region of the bilayer only as the
DPPC and LPS lipid tails do not possess labile hydrogens, and
therefore will not undergo solvent-contrast-related changes in
SLD [16].
w
(water)
was determined by
w
ðwaterÞ¼ð
r
ðwater contrast1Þ
r
ðwater contrast2ÞÞ
ð
r
ðwater 1Þ
r
ðwater 2ÞÞ;ð2:2Þ
where
r
water contrast 1
and
r
water contrast 2
are the SLDs of the same
tail layer in two different water isotopic contrasts (in this case
either D
2
O, SMW or H
2
O) and
r
water 1
and
r
water 2
are the
SLDs of each H
2
O/D
2
O mix, respectively. In this way,
w
(water)
and
r
2(
r
(water)
w
(water)
) are obtained, these values relate the
relative contributions of DPPC and LPS tails to this layer by
r
ð
r
ðwaterÞ
w
ðwaterÞÞ¼ð
r
ðDPPC tailsÞ
w
ðDPPC tailÞÞ
þð
r
ðLPS tailsÞ
w
ðLPS tailsÞÞ:ð2:3Þ
Therefore, once
w
(water)
and
r
2(
r
(water)
w
(water)
) were known,
these values were used to determine the relative mixing of
the hydrogenated LPS and the deuterated phospholipid in the
bilayer leaflets. The
w
(DPPC tails)
in the tail layers of the bilayer
was determined by
w
ðDPPC tailsÞ¼ð
r
ð
r
ðwaterÞ
w
waterÞð
r
LPS tailsð1
w
waterÞÞÞ
ð
r
ðdDPPC tailsÞ
r
ðLPS tailsÞÞ
!
:
ð2:4Þ
Once
w
(DPPC tails)
was determined, the contribution of
w
(LPS tails)
to these layers was deduced by
w
ðLPS tailsÞ¼1ð
w
ðDPPC tailsÞþ
w
ðwaterÞÞ:ð2:5Þ
The relative volume fractions of the LPS and DPPC in the
headgroup layers of the bilayer structures were not able to be
determined owing to the minimal isotopic contrast between the
DPPC headgroups and the LPS core oligosaccharide region
(table 1). Therefore, all volume fractions of DPPC, LPS and
water quoted in this article are describing the lipid tail regions
of each leaflet within the bilayer.
2.5. Model to experimental data fitting error analysis
Model to experimental data fitting errors were obtained using
Rascals ‘bootstrap’ error analysis function, in which the original
dataset is resampled and these new datasets fitted via the same
methods as described earlier. The parameter value distributions
obtained across these fits were used to estimate errors which
were then propagated through the calculations of the derived
parameters according to standard error treatment methods [29].
3. Results
3.1. Asymmetric DPPC :Lipid A bilayer
Figure 2 shows the NR profiles, model datafits and the resulting
SLD profiles for an asymmetrically deposited DPPC :Lipid A
bilayer deposited on a silicon surface. As mentioned previously,
reflectivity data for two individual bilayers, an h-DPPC- and a
d-DPPC-labelled bilayer, were examined under three-solution
contrasts (H
2
O, SMW and D
2
O) producing six reflectivity
profiles. During fitting of the data, the layer thicknesses and
roughness of both bilayers were constrained to fit a single pro-
file, however the hydration and the SLD of the layers was fitted
individually for each bilayer.
The NR obtained from the DPPC :Lipid A bilayer was
fitted to a five-layer model of the interfacial structure. This
model represents the minimal number of layers with which
the reflectivity data could be fitted. The layers in this struc-
tural model describe (moving from silicon to the bulk
Table 1. Summary of scattering length densities of the lipid components
studied and the solution subphases.
lipid/solvent
a
neutron scattering
length density (
r
)
(310
26
A
˚
22
)
20 mM pD 7.0 D
2
O phosphate buffer 6.35
20 mM pH/D 7.0 SMW phosphate buffer 2.07
20 mM pH 7.0 H
2
O phosphate buffer 20.56
silicon 2.07
silicon oxide (SiO
2
) 3.41
DPPC headgroup 1.98
h-DPPC tails 20.39
d-DPPC tails 7.45
Lipid A tails 20.39
Lipid A GlcN (headgroup) in D
2
O 3.39
Lipid A GlcN (headgroup) in H
2
O 2.58
Rc-LPS hydrophilic core oligosaccharide
(headgroup) region in D
2
O
4.2
Rc-LPS hydrophilic core oligosaccharide
(headgroup) region in H
2
O
2.04
Ra-LPS hydrophilic core oligosaccharide
(headgroup) region in D
2
O
4.28
Ra-LPS hydrophilic core oligosaccharide
(headgroup) in H
2
O
2.01
a
The volumes used to calculate SLD for Lipid A and LPS headgroups are
based on volumes from the crystal structures of sugars [14,23]. Values
from H
2
O, D
2
O, Si and SiO
2
have been reported previously [24,25].
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
4

solution) a silicon oxide layer (1st layer), the inner bilayer
leaflet headgroups (2nd layer), the inner bilayer leaflet acyl
chains (3rd layer), the outer leaflet acyl chains (4th layer)
and the outer leaflet headgroups (5th layer). Table 2 describes
the structural parameters obtained from fitting of the NR data
obtained from the d-DPPC-labelled DPPC :Lipid A bilayer
sample. It should be noted that fitting of the asymmetrical
DPPC :Lipid A bilayer to a simple five-layer description
of the interfacial structure produced fits that were less
complete than those obtained from fitting the reflectivity
profiles obtained from the Rc- or Ra-LPS-containing bilayer
samples using the same model.
Analysis of the reflectivity data revealed high coverage
for both bilayers examined (h-DPPC and d-DPPC labelled).
Based on the hydration of the lipid tail regions of the bilayer,
the fully hydrogenated bilayer (h-DPPC :Lipid A) was found
to have an average surface coverage (determined by the
addition of
w
LPS
and
w
DPPC
of the inner and outer leaflets
combined) of 99 +5%, whereas the d-DPPC-labelled bilayer
was found to have an average coverage of 91+5%. Although
similar, it is clear that repeated bilayer production produces
bilayers with minor differences in coverage probably owing
to random error during the bilayer fabrication process.
As previously mentioned, coverages were determined from
the combined volume fractions of the DPPC and LPS in the
lipid tail regions of the bilayer as calculation of the head-
group volume fractions could not be accurately determined
for reasons described previously. However based on the scat-
tering length densities obtained from the headgroup layers
(see the electronic supplementary material), the hydration
of the lipid DPPC :Lipid A headgroup is likely to be
significantly higher than that determined for the tail regions,
which is expected owing to the hydrophilic nature of this
moiety of the bilayer.
The d-DPPC-labelled DPPC :LPS bilayers were used to
examine the asymmetry of the two lipid component bilayers.
Analysis of the scattering length densities of the inner and
outer bilayer tails of the d-DPPC :Lipid A bilayer reveals
that although an asymmetrical structure had been produced
there was mixing of the DPPC and Lipid A. Indeed, the
outer leaflet of the bilayer was found to be composed of
65% Lipid A (
w
Lipid A
¼0.65) and 26% DPPC, whereas con-
versely the inner bilayer leaflet showed almost the reverse
mixing with 36% Lipid A and 55% DPPC found. As the
inner leaflet of the bilayer was deposited as a pure DPPC
layer and the outer leaflet was deposited as a pure Lipid A
layer, this implies mixing of the two leaflets. However, the
collection of repeat reflectivity data at 16 h intervals revealed
that although significant mixing had occurred between the
leaflets prior to initial NR analysis no further mixing between
the layers occurred overtime under periodic flow (see the
electronic supplementary material, figure S3 for a comparison
of NR data).
Gerelli et al. [30] have recently examined the mixing of
asymmetrically deposited phospholipid bilayers and have
found that significant flipping between the inner and outer
bilayer leaflets would only be expected when the bilayer com-
ponents are in the liquid phase. As the bilayer structures
described here were both deposited and examined at room
temperature (208C) where both the DPPC and the Lipid A
components of the bilayer would be expected to be in the
gel or subgel phases [31,32], flipping between the layers
1
(a)(d)
(b)
(c)
0.01 0.1
0.01 0.1
0.01
Q (Å–1)
distance (Å)
0.1
0
–1
0
1
2
3
4
5
6
7
20 40 60 80 100 120
10–1
10–2
10–2
reflectivity
scattering length density (×10–6 Å–2)
Si
SiO2
inner head group
outer head group
solution
inner tails
outer tails
reflectivityreflectivity
10–3
10–3
10–4
10–4
10–5
10–6
0–6
10–5
10–2
10–1
10–3
10–4
10–6
10–5
Figure 2. Neutron reflectometry profile and model data fits (a–c) and the scattering length density profiles these fits describe (d) for asymmetrically deposited
DPPC (inner leaflet) : Lipid A (outer leaflet) bilayer. The six simultaneously fitted isotopic contrasts shown are (a)d-DPPC :Lipid A in D
2
O (red line), h-DPPC :Lipid A
in D
2
O (blue line); (b)d-DPPC :Lipid A in SMW (black line), h-DPPC :Lipid A in SMW (grey line); (c)d-DPPC :Lipid A in H
2
O (green line), h-DPPC :Lipid A in H
2
O
(purple line). (Online version in colour.)
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
5

would not be expected. Therefore, the mixing between the
layers observed here is only likely to have occurred during
the LS deposition of the outer leaflet of the bilayer, in agree-
ment with previously observed results for phospholipids [30].
3.2. Asymmetric DPPC :Rc-LPS bilayer
To examine whether more realistic mimics of the structure of
the GNB-OM could be achieved, we formed bilayers with an
inner leaflet of DPPC and an outer leaflet of Rc-LPS or
Ra-LPS. Rc-LPS is preferable to Lipid A for use inouter bacterial
membrane mimics as this rough strain LPS possesses a signifi-
cant portion of the core oligosaccharide region, thus providing
a better mimic of the surface structure of the GNB-OM. The use
of Ra-LPS in these bilayers is better still as this rough mutant
LPS contains the complete core oligosaccharide region of a
full-length LPS molecule. Figure 3 shows the NR profiles,
model data fits and the SLD profiles, and these fits describe
for an asymmetrically deposited bilayer of DPPC (inner leaflet)
and Rc-LPS (outer leaflet) (DPPC :Rc-LPS) with table 3
showing the parameters.
As with the DPPC : Lipid A bilayer, a five-layer model of
the interfacial structure was suitable for fitting reflectometry
profiles obtained from the asymmetrically deposited DPPC :
Rc-LPS bilayer. As with the DPPC : Lipid A bilayer, compari-
son of the coverage of the h-DPPC and d-DPPC labelled
bilayers revealed minor differences in coverage between the
two bilayers, with average coverages of 90 +10% and 84 +
5% for the h-DPPC and d-DPPC-labelled bilayers, respect-
ively. It should be noted however that the surface coverage
for each bilayer are within error of each other.
The bilayer roughness was somewhat higher than that
found for the DPPC :Lipid A membrane, fitted at 5.4 +
3.1 A
˚
for the Rc-LPS-containing membrane compared with
2.4 +1.5 A
˚
for the Lipid A-containing membrane. The fits
of the NR for d-DPPC-labelled DPPC :Rc-LPS bilayer demon-
strated that the inner leaflet, which was deposited as DPPC
only, contained 58 +4% DPPC and 25 +8% Rc-LPS, whereas
the outer leaflet consisted of 28 +1% DPPC and 57 +2%
Rc-LPS. As with the Lipid A bilayer, no change in the
asymmetry was noted over time (see electronic supplemen-
tary material, figure S4).
Notably, the outer headgroup region of the DPPC :
Rc-LPS bilayer was found to be significantly thicker than
the equivalent region of the DPPC : Lipid A outer bilayer,
being 20.9 +2.0 A
˚
in thickness (compared with 8.0 +5.0 A
˚
for the Lipid A-containing membrane). This difference in
thickness between the two bilayers is likely owing to the
presence of a significant proportion of the LPS core oligosac-
charide region on the hydrophilic moiety of Rc-LPS (figure 1)
compared with Lipid A.
3.3. Asymmetric DPPC :Ra-LPS bilayer
Ra-LPS, despite its large size and being highly water soluble,
formed stable, reproducible monolayers at the air–liquid inter-
face so that LS deposition could occur effectively (see electronic
supplementary material, figure S2). The collapse pressure of
the monolayer was 52 mN m
21
, well above the 27 mN m
21
used in bilayer fabrication. Figure 4 shows the NR profiles,
model data fits and the SLD profiles from these fits describing
an asymmetrically deposited asymmetrically deposited bilayer
of DPPC (inner leaflet) and Ra-LPS (outer leaflet) (DPPC :
Ra-LPS) bilayer fitted to the same five-layer model structure
which was previously found to be optimal for the Lipid A
and Rc-LPS-containing bilayers. Table 4 lists the parameters
obtained from these fits of the experimental data.
The total surface coverage of lipid (
w
DPPC
þ
w
Ra-LPS
) in the
DPPC :Ra-LPS bilayer was found to be approximately 85%
(table 4), and therefore significantly lower than that found
for the DPPC : Lipid A bilayer. The structure and structural
asymmetry observed across the membrane was similar to
that found for the DPPC :Lipid A and DPPC :Rc-LPS bilayers,
with a DPPC-rich inner leaflet (
w
DPPC
¼0.66) and an LPS-
rich outer leaflet (
w
Ra-LPS
¼0.67), which as with the other
DPPC :LPS bilayers suggested that partial asymmetry had
been maintained in the interfacial film.
The most notable feature of the DPPC :Ra-LPS bilayer was
the outer headgroup region of the bilayer, notably thicker
(31 +1.2 A
˚
) than that for Lipid A or Rc-LPS-containing mem-
branes. This region of the bilayer structure is likely to be
dominated by the contribution of the Ra-LPS headgroup
region owing to the significantly larger size of the Ra-LPS
hydrophillic inner core region compared with the DPPC
Table 2. Structural parameters obtained for an asymmetrically deposited d-DPPC (inner leaflet) E. coli Lipid A (outer leaflet) bilayer deposited on a silicon
surface at 27 mN m
21
monolayer pressure.
layer thickness (A
˚
)
w
DPPC
w
Lipid A
w
water
roughness (A
˚
)
layer 1
silicon oxide
14.3 +2.9 n.a. n.a. 0.07 +0.06 2.9 +1.3
layer 2
inner headgroup
8.5 +1.0 0.54 +0.03 0.36 +0.05 0.092 +0.050
a
bilayer
roughness ¼2.4 +1.5
layer 3
inner tails
19.8 +2.0
layer 4
outer tails
17.6 +3.4 0.26 +0.03 0.65 +0.06 0.15 +0.03
a
as above
layer 5
outer headgroup
8+5
a
w
water
for the headgroups does not include water of hydrations as this is accounted for in the headgroup volume fraction.
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
6

headgroup and the higher volume fraction of Ra-LPS found
in the outer bilayer leaflet tails compared with DPPC.
As previously mentioned, in all cases the bilayer structure
was examined with NR over a 16 h period after initial NR
measurement to asses whether any changes to the bilayer
structure took place over time and after periodic flow in the
solid–liquid cell. In the cases of the DPPC :Lipid A,DPPC :
Rc-LPS and the DPPC :Ra-LPS, bilayer showed no significant
changes to the interfacial structure over this time period (see
the electronic supplementary material, figure S5).
4. Discussion
Here, we have examined whether it is possible to create asym-
metrical GNB-OM models using Lipid A and rough mutant
LPSs. The asymmetry of these GNB-OM models was intended
to mimic that of the GNB-OM where a phosphatidylethanola-
mine-rich inner leaflet and an LPS-rich outer leaflet are found
[2,5]. By varying the LPS used in the outer leaflet from Lipid
A to Ra-LPS, the size of the LPS inner core region has been
increased in the bilayer outer leaflet from the minimal size
(GlcN only in Lipid A, figure 1) to having a complete core
1
(a)(d)
(b)
(c)
0.01 0.1
0.01 0.1
0.01
Q (Å–1)
distance (Å)
0.1
0
–1
0
1
2
3
4
5
6
7
20 40 60 80 100 120
10–1
10–2
10–2
reflectivity
scattering length density (×10–6 Å–2)
Si
SiO2
inner head group
outer head group
solution
inner tails
outer tails
reflectivityreflectivity
10–3
10–3
10–4
10–4
10–5
10–6
10–6
10–5
10–2
10–1
10–3
10–4
10–6
10–5
Figure 3. Neutron reflectometry profile and model data fits (a–c) and the scattering length density profiles these fits describe (d) for asymmetrically deposited
DPPC (inner leaflet) : Rc-LPS (outer leaflet) bilayer. The six simultaneously fitted isotopic contrasts shown are (a)d-DPPC :Rc-LPS in D
2
O (red line), h-DPPC :Rc-LPS in
D
2
O (blue line); (b)d-DPPC :Rc-LPS in SMW (black line), h-Rc-LPS :Lipid A in SMW (grey line); (c)d-DPPC :Rc-LPS in H
2
O (green line), h-DPPC :Rc-LPS in H
2
O
(purple line). (Online version in colour.)
Table 3. Fitting parameters obtained for an asymmetrically deposited d-DPPC (inner leaflet) Rc-LPS (outer leaflet) bilayer deposited on a silicon surface at
27 mN m
21
monolayer pressure.
layer thickness (A
˚
)
w
DPPC
w
Rc-LPS
w
water
roughness (A
˚
)
layer 1
silicon oxide
11.1 +1.9 n.a. n.a. 0.15 +0.10 3 +2
layer 2
inner headgroup
8.4 +11.2 0.58 +0.04 0.25 +0.08 0.16 +0.06
a
bilayer
roughness ¼5.4 +3.1
layer 3
inner tails
18.2 +2.5
layer 4
outer tails
15.3 +3.0 0.28 +0.01 0.57 +0.02 0.15 +0.03
a
as above
layer 5
outer headgroup
20.9 +2.0
a
w
Water
for the headgroups does not include water of hydrations as this is accounted for in the headgroup volume fraction.
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
7

oligosaccharide region (Ra-LPS). In doing so, the accuracy of
the membrane model moves from a bilayer that represents
only the GNB-OM core hydrophobic region (DPPC :Lipid A)
to a bilayer that is more structurally similar to the GNB-OM,
with the Lipid A and core oligosaccharide present on the LPS
in the outer leaflet of the membrane present (DPPC :Ra-LPS),
and only the O-antigen missing.
Results revealed that complex asymmetrical bilayers
could indeed be fabricated using both a combination of syn-
thetic phospholipids and E. coli rough mutant LPSs. All the
bilayers examined were found to be asymmetric in nature
with a DPPC-rich inner leaflet and an LPS-rich outer leaflet.
The discovery that complex amphiphillic natively extracted
molecules, such as Rc- and Ra-LPS, can be incorporated
into complex structures that are amenable to molecular
level structural studies shows the potential of complex
wild-type lipids and amphiphiles for use in the fabrication
of model biological surfaces. Indeed, the GNB-OM mimics
described in this study may allow for molecular level
examinations of the dynamics and interactions of this mem-
brane to be conducted under conditions close to those
found in vivo.
1
(a)
(d)
(b)
(c)
0.01 0.1
0.01 0.1
0.01
Q (Å–1)
distance (Å)
0.1
0
–1
0
1
2
3
4
5
6
7
20 40 60 80 100 120
10–1
10–2
10–2
10–1
reflectivity
scattering length density (×10–6 Å–2)
Si
SiO2
inner head group
outer head group
solution
inner tails
outer tails
reflectivityreflectivity
10–3
10–3
10–4
10–4
10–5
10–6
10–6
10–5
10–2
101
10–3
10–4
10–6
10–5
Figure 4. Neutron reflectometry profile and model data fits (a–c) and the scattering length density profiles these fits describe (d) for asymmetrically deposited
DPPC (inner leaflet) : Ra-LPS (outer leaflet) bilayer. The six simultaneously fitted isotopic contrasts shown are (a)d-DPPC :Ra-LPS in D
2
O (red line), h-DPPC :Ra-LPS
in D
2
O (blue line); (b)d-DPPC :Ra-LPS in SMW (black line), Ra-LPS :Lipid A in SMW (grey line); (c)d-DPPC :Ra-LPS in H
2
O (green line), h-DPPC :Ra-LPS in H
2
O
(purple line). (Online version in colour.)
Table 4. Fitting parameters obtained for an asymmetrically deposited d-DPPC (inner leaflet) Ra-LPS (outer leaflet) bilayer deposited on a silicon surface at
27 mN m
21
monolayer pressure.
layer thickness (A
˚
)
w
DPPC
w
Ra-LPS
w
water
roughness (A
˚
)
layer 1
silicon oxide
13.4 +2.0 n.a. n.a. 0.104 +0.040 3.0 +1.0
layer 2
inner headgroup
14.8 +2.0 0.66 +0.05 0.19 +0.09 0.16 +0.08
a
bilayer
roughness ¼7.90 +0.55
layer 3
inner tails
15.6 +0.6
layer 4
outer tails
16.0 +4.8 0.22 +0.05 0.67 +0.07 0.11 +0.07
a
as above
layer 5
outer headgroup
31.0 +1.2
a
w
water
for the headgroups does not include water of hydrations as this is accounted for in the headgroup volume fraction.
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
8

Controlled deposition of LPS monolayers in multi-layered
LPS only bilayer samples by LB deposition has previously
been reported as intractable with LPS types possessing a
core oligosaccharide region larger in size than that found in
Re-LPS [10]. Here, we have been able to deposit solid sup-
ported single bilayers containing an LPS-rich outer leaflet
with LPS types which contain a significant proportion of
(Rc-LPS) or all of (Ra-LPS) the core oligosaccharide region.
The stability of the single bilayers containing the relatively
hydrophilic rough mutant LPSs may be owing to the DPPC
anchoring the LPS within the bilayer. DPPC is known to form
stable bilayers on an oxidized silicon crystal surface [17], with
the bilayers held in place owing to a combination of electrostatic
attraction between the cationic choline group on the inner
bilayer leaflet and surface oxide (in this case SiO
2
)andvan
der Waals forces [33,34]. Hydrophobic interactions with the
DPPC tails likely keep the LPS tails, and therefore the whole
molecule anchored to the bilayer, which results in the stable
silicon-surface-bound bilayers described here. Indeed, it has
been shown previously that monolayers of smooth LPS
can be deposited on to silicon surfaces hydrophobized by a
covalently attached alkyl silane monolayer [15].
The structural parameters obtained for the Lipid A and
LPS-rich outer leaflets of the asymmetric DPPC :Lipid A
and DPPC :Rc-LPS compare well with those determined for
monolayers of these lipids by Le Brun et al. [14]. Lipid A
monolayers were found to have a headgroup thickness of
8+1A
˚
, which compares well with the 8 +5A
˚
found for
the Lipid A-rich outer leaflet of the DPPC :Lipid A bilayer.
Rc-LPS monolayers at the air–liquid interface were found
to have headgroup regions that had a total thickness of
29 +7A
˚
at high surface pressures, and therefore high LPS
densities. We found the LPS-rich leaflet of the Rc-LPS bilayer
to have a headgroup thickness of 20.9 +2A
˚
, this slightly
thinner layer may suggest a tilted orientation of the Rc-LPS
headgroup in the outer leaflet of the bilayer. The outer leaflet
acyl chain regions of the DPPC :Lipid A and DPPC :Rc-LPS
bilayers were found to be slightly thicker than found for
Lipid A and Rc-LPS monolayers, which is probably owing
to the presence of the palmitic acid chains of the phospho-
lipid within this region. Previously, Snyder et al. [11] were
able to resolve the trend of increasing core oligosaccharide
region size where Re ,Rd ,Rc ,Ra-LPS, when LPS only
stacked bilayer samples were examined. Here, the same
trend has been observed (Lipid A ,Rc-LPS ,Ra-LPS) in
the complex asymmetrical single bilayer structures examined
with the core oligosaccharide thicknesses (obtained from the
outer leaflet headgroup thicknesses) being in general in good
agreement with the aforementioned studies.
In all the bilayers examined, some mixing of the lipids
between the leaflets was observed. This is likely to have
occurred owing to the mechanical shock of the LS dipping
phase of bilayer fabrication [30], as mixing between the
leaflets was not observed over time under the ambient con-
ditions used in these studies. The general asymmetry of the
phospholipid : LPS bilayers described here showed approxi-
mately 25% LPS and approximately 65% DPPC in the inner
leaflet (þ10% water) and approximately 65% LPS and
approximately 25% DPPC (þ10% water) in the outer leaflet
(figure 5 shows a cartoon representation of the interfacial
structure). The GNB-OM outer leaflet is known to possess
LPS predominantly as its lipid component [2], approximately
25% phospholipid found in the outer leaflet here could be
considered to make these membrane models less biologically
relevant. However, it should be noted that the phospholipid
is a minor component of the outer leaflet. The significant
asymmetry observed in these easily formed, analysed (by
NR) and stable bilayer models could be considered as a
reasonable representative of the GNB-OM for future bio-
logical interaction studies. Gerelli et al. [30] have improved
phospholipid bilayer asymmetry by preparing the samples
below the phase transition temperature of both lipids depos-
ited which, for the lipids used was below room temperature.
Despite the improved asymmetry when conducting the LB–LS
depositions below the lipid phase transition temperatures,
there was still a 10% mixing of phospholipids. The bilayers
described here were all prepared at room temperature,
which is well below the phase transition temperature of
both the DPPC and the LPS used. Another potential way of
reducing inner/outer leaflet mixing during LS deposition of
the outer bilayer leaflet maybe to introduce divalent cations
to the solution subphase below the LPS monolayer, as the
interaction of the cations with the LPS may increase the rigid-
ity of the monolayer causing less mixing to occur during
bilayer production [13].
Total interfacial coverages of the bilayers ranged from
greater than 90% for the Lipid A-containing bilayers to
approximately 85% for the Rc- and Ra-LPS-containing mem-
branes. The hydration of the tail region of the bilayers (which
we use here to measure coverage) is likely owing to defects in
the bilayer film, that is relatively small regions of the silicon
surface with low or no lipid coverage [17], which are likely
to be formed during the LS stage of the asymmetrical bilayer
fabrication process. The possible reason for the slightly
higher coverage for the Lipid A-containing bilayers compared
with the rough LPS-containing membranes maybe owing to
the higher phosphorylation found on the rough mutant LPS
types. As electrostatic repulsion between neighbouring anio-
nic LPS molecules may have caused some loss of this material
during the LS deposition of the outer bilayer leaflet, the
increased hydrophilicity of the rough mutant LPSs compared
to Lipid A may also be causing a higher loss of the LPS to the
bulk solution compared with the more hydrophobic Lipid A
during the bilayer fabrication process, this may also partly
account for the slightly lower coverage of the Rc- and
Ra-LPS-containing bilayers.
The rough mutant LPS-containing bilayers were found to
be rougher than the DPPC :Lipid A bilayer, with both the
Rc- and Ra-LPS-containing films found to be 5 and 8 A
˚
in
roughness, respectively, compared with 3 A
˚
for the DPPC :
Lipid A bilayer. The size of the DPPC and Lipid A headgroup
regions have been found to be the same at approximately 9 A
˚
(table 2; [14,16]). Conversely, the headgroups of Rc- and Ra-
LPS were found here to be 20.9 and 31 A
˚
in thickness, respect-
ively (based on the thicknesses of the LPS-rich outer leaflet
headgroups; tables 3 and 4) and are thus significantly larger
in size than the DPPC headgroup region. The size mismatch
between the PC and LPS in the inner leaflet of the bilayer,
which is next in close proximity to the relatively flat silicon
oxide surface, would induce an increased roughness across
the entire bilayer compared with the DPPC :Lipid A bilayer,
where no significant size mismatch was present (a cartoon rep-
resentation of this is shown in figure 5). Thus, the increases in
roughness (and headgroup thickness in the case of the Ra-LPS-
containing bilayer) are likely owing to the presence of the
rough E. coli mutant LPS within the inner leaflet of the bilayer.
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
9

The increased roughness may also be in part owing to the flexi-
bility of the core oligosaccharide headgroup regions of Rc- and
Ra-LPS [35] in the outer headgroup region making this region
less defined using layer models. The interfacial roughness can
be viewed as either representing a sharp interface with special
variations, or as is more likely here, represents a gradual
change in the neutron SLD as a function of distance. However,
the interfacial roughnesses of all the bilayers studied were
relatively low and were only approximately 10% of the total
thickness of the membranes.
Owing to the asymmetric nature of the bilayers under
study and the presence of large hydrophilic regions on
Rc- and Ra-LPS-containing bilayers, the structure of the
bilayers was studied over time to observe potential changes
to the structure, i.e. lipid flip flop between bilayer leaflets
and/or loss of material from the interface. All three bilayer
types showed no change over a 16 h period in the presence
of periodic fluid flows used to change isotopic contrast (see
electronic supplementary material, figures S3– S5). The lack
of membrane leaflet mixing observed here would be expected
as both the phospholipid and LPS used in this study were
examined in the gel or subgel phase [31]. It is possible that
if we examined the bilayer structure above the phase tran-
sition of both DPPC and LPS which is both cases would
be greater than the ambient temperature under which this
study was conducted. Indeed, a relatively high temperature
(more than 408C) would be required to have both the LPS
and phospholipid components in the fluid phase [31,32].
5. Conclusion
Bilayers which mimic lipid content and asymmetry of the
GNB-OM have been successfully fabricated. The asym-
metrical bilayer structures described here are composed of a
mixture of synthetic phospholipids and naturally extracted
LPSs making these biological membrane models accurate in
the composition and asymmetry of the Gram-negative bac-
terial membrane which they intend to mimic. Future work
will both examine interactions of the membrane models
described here with antimicrobial proteins and peptides
and continue to develop the complexity of the GNB-OM
models; this could include embedding integral membrane
proteins [36,37] and increasing the fluidity of the structures
by preventing the bilayer from being in direct contact with
the solid–liquid interface. However, for many methods to
study Gram-negative membrane interactions, the simple,
robust and long-lived model structures presented here may
provide a useful tool.
Acknowledgements. We thank Christian Kinane for assistance with
running the CRISP reflectometer.
Funding statement. This work was supported by ISIS beam-time award
no. 1310101. J.H.L. wishes to thank the Wellcome Trust for support
(grant nos. 080342 and 093581).
References
1. Tenover FC. 2006 Mechanisms of antimicrobial
resistance in bacteria. Am. J. Med.119, S3– S10;
discussion S62–S70. (doi:10.1016/j.amjmed.
2006.03.011)
2. Nikaido H. 2003 Molecular basis of bacterial outer
membrane permeability revisited. Microbiol. Mol.
Biol. Rev. 67, 593–656. (doi:10.1128/mmbr.67.4.
593-656.2003)
3. Delcour AH. 2009 Outer membrane permeability and
antibiotic resistance. Biochem. Biophys. Acta 1794,
808– 816. (doi:10.1016/j.bbapap.2008.11.005)
4. Nilsson C, Skoglund A, Moran AP, Annuk H, Engstrand
L, Normark S. 2008 Lipopolysaccharide diversity
evolving in Helicobacter pylori communities through
genetic modifications in fucosyltransferases. PLoS ONE
3, e3811. (doi:10.137/journal.pone.0003811)
5. Erridge C, Bennett-Guerrero E, Poxton IR. 2002
Structure and function of lipopolysaccharides.
Microbes Infect. 4, 837–851. (doi:10.1016/s1286-
4579(02)01604-0)
6. Mayer H, Bhat UR, Masoud H, Radziejewska-
Lebrecht J, Widemann C, Krauss JH. 1989 Bacterial
lipopolysaccharides. Pure Appl. Chem. 61,
1271–1282. (doi:10.1351/pac198961071271)
7. Raetz CRH. 1990 Biochemistry of endotoxins. Annu.
Rev. Biochem. 59, 129–170. (doi:10.1146/annurev.
bi.59.070190.001021)
8. D’Errico G, Silipo A, Mangiapia G, Molinaro A,
Paduano L, Lanzetta R. 2009 Mesoscopic and
microstructural characterization of liposomes formed
by the lipooligosaccharide from Salmonella minnesota
strain 595 (Re mutant). Phys. Chem. Chem. Phys. 11,
2314– 2322. (doi:10.1039/B816248D)
(a)
(b)
Si SiO2
(c)
Figure 5. A cartoon representation of the structure across the interface found
for the DPPC : LPS bilayers described here, showing the structure of (a)DPPC :
Lipid A,(b)DPPC :Rc-LPS and (c)DPPC :Ra-LPS examined here. (Online
version in colour.)
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
10

9. Kubiak J, Brewer J, Hansen S, Bagatolli LA. 2011
Lipid lateral organization on giant unilamellar
vesicles containing lipopolysaccharides. Biophys. J.
100, 978–986. (doi:10.1016/j.bpj.2011.01.012)
10. Kastowsky M, Gutberlet T, Bradaczek H. 1993
Comparison of X-ray powder-diffraction data of various
bacterial lipopolysaccharide structures with theoretical
model conformations. FEBS J. 217, 771–779. (doi:10.
1111/j.1432-1033.1993.tb18305.x)
11. Snyder S, Kim D, Mcintosh TJ. 1999
Lipopolysaccharide bilayer structure: effect of
chemotype, core mutations , divalent cations and
temperature. Biochemistry 38, 10 758 –10 767.
(doi:10.1021/bi990867d)
12. Abraham T, Schooling SR, Nieh M-P, Kucerka N,
Beveridge TJ, Katsaras J. 2007 Neutron diffraction
study of Pseudomonas aeruginosa lipopolysaccharide
bilayers. J. Phys. Chem. B 111, 2477– 2483. (doi:10.
1021/jp066012+)
13. Jeworrek C, Evers F, Howe J, Brandenburg K, Tolan
M, Winter R. 2011 Effects of specific versus
nonspecific ionic interactions on the structure
and lateral organization of lipopolysaccharides.
Biophys. J. 100, 2169– 2177. (doi:10.1016/j.bpj.
2011.03.019)
14. Le Brun AP, Clifton LA, Halbert CE, Lin B, Meron M,
Holden PJ, Lakey JH, Holt SA. 2013 Structural
characterization of a model Gram-negative bacterial
surface using lipopolysaccharides from rough
strains of Escherichia coli.Biomacromolecules 14,
2014–2022. (doi:10.1021/bm400356m)
15. Schneck E, Papp-Szabo E, Quinn BE, Konovalov OV,
Beveridge TJ, Pink DA, Tanaka M. 2009 Calcium ions
induce collapse of charged O-side chains of
lipopolysaccharides from Pseudomonas aeruginosa.
J. R. Soc. Interface 6, S671– S677. (doi:10.1098/rsif.2009.
0190.focus)
16. Hughes AV, Howse JR, Dabkowska A, Jones RAL,
Lawrence MJ, Roser SJ. 2008 Floating lipid bilayers
deposited on chemically grafted phosphatidylcholine
surfaces. Langmuir 24, 1989–1999. (doi:10.1021/
la702050b)
17. Tamm LK, McConnell HM. 1985 Supported
phospholipid bilayers. Biophys. J. 47, 105– 113.
(doi:10.1016/s0006-3495(85)83882-0)
18. Ma G, Allen HC. 2006 DPPC Langmuir monolayer at the
air– water interface: probing the tail and head groups
by vibrational sum frequency generation spectroscopy.
Langmuir 22, 5341–5349. (doi:10.1021/la535227)
19. McConlogue CW, Vanderlick TK. 1997 A close look at
domain formation in DPPC monolayers. Langmuir
13, 7158–7164. (doi:10.1021/la970898e)
20. Webster JRP, Holt SA, Dalgliesh R. 2006 INTER the
chemical interfaces reflectometer on target station 2
at ISIS. Phys. B 385– 386, 1164 – 1166. (doi:10.10.
6/j.physb.2006.05.400)
21. Penfold J et al. 1997 Recent advances in the study
of chemical surfaces and interfaces by specular
neutron reflection. Trans. Faraday Soc. J. Chem. Soc.
93, 3899–3917. (doi:10.1039/a702836i)
22. Penfold J, Ward RC, Williams WG. 1987 A time-of-
flight neutron reflectometer for surface and
interfacial studies. J. Phys. E: Sci. Instrum. 20,
1411–1417. (doi:10.1088/0022-3735/20/11/024)
23. Ferrier WG. 1960 The crystal structure of b-D-
glucose. Acta Crystallogr. 13, 678– 679. (doi:10.
1107/s0365110X60001588)
24. Stidder B, Fragneto G, Roser SJ. 2005 Effect of low
amounts of cholesterol on the swelling behavior of
floating bilayers. Langmuir 21, 9187– 9193. (doi:10.
1021/la50962p)
25. Fragneto G, Charitat T, Daillant J. 2012 Floating lipid
bilayers: models for physics and biology. Eur.
Biophys. J. Biophy. 41, 863– 874. (doi:10.1007/
s00249-0834-4)
26. Wacklin HP. 2011 Composition and asymmetry in
supported membranes formed by vesicle fusion.
Langmuir 27, 7698–7707. (doi:10.1021/la200683e)
27. Hughes AV. 2013 RasCAL. Sourceforge. See http://
sourceforge.net/projects/rscl/.
28. Born M, Wolf E. 1970 Principles of optics. Oxford,
UK: Pergamon Press.
29. Clifton LA, Sanders M, Kinane C, Arnold T, Edler KJ,
Neylon C, Green RJ, Frazier RA. 2012 The role of
protein hydrophobicity in thionin– phospholipid
interactions: a comparison of a1 and a2-
purothionin adsorbed anionic phospholipid
monolayers. Phys. Chem. Chem. Phys. 14, 13 569–
13 579. (doi:10.1039/c2cp42029e)
30. Gerelli Y, Porcar L, Fragneto G. 2012 Lipid
rearrangement in DSPC/DMPC bilayers: a neutron
reflectometry study. Langmuir 28, 15 922 –15 928.
(doi:10.1021/la303662e)
31. Brandenburg K, Seydul U. 1990 Investigation into
the fluidity of lipopolysaccharide and free lipid A
membrane systems by Fourier transform infrared
spectroscopy and differential scanning calorimetry.
Biophys. J. 64, 1215– 1230. (doi:10.1016/s0006-
3495(93)81488-7)
32. Yarrow F, Vlugt TJH, van der Eerden JPJM, Snel
MME. 2005 Melting of a DPPC lipid bilayer observed
with atomic force microscopy and computer
simulation. J. Cryst. Growth 275, e1417–e1421.
(doi:10.1016/j.jcrysgro.2004.11.237)
33. Cremer PS, Boxer SG. 1999 Formation and spreading
of lipid bilayers on planar glass supports. J. Phys.
Chem. B 291, 2554–2559. (doi:10.1021/jp983996x)
34. Castellana ET, Cremer PS. 2006 Solid supported lipid
bilayers: from biophysical studies to sensor design.
Surf. Sci. Rep. 61, 429– 444. (doi:10.1016/j.surfrep.
2006.06.001)
35. Kastowsky M, Gutberlet T, Bradaczek H. 1992
Molecular modelling of the three-dimensional
structure and conformational flexibility of
bacterial lipopolysaccharide. J. Bacteriol. 174,
4798–4806.
36. Holt SA, Le Brun AP, Majkrzak CF, McGillivray DJ,
Heinrich F, Loesche M, Lakey JH. 2009 An ion-
channel-containing model membrane: structural
determination by magnetic contrast neutron
reflectometry. Soft Matter 5, 2576– 2586.
(doi:10.1039/b822411k)
37. Clifton LA et al. 2012 Low resolution structure and
dynamics of a colicin –receptor complex determined
by neutron scattering. J. Biol. Chem. 287, 337– 346.
(doi:10.1074/jbc.m111.302901)
rsif.royalsocietypublishing.org J R Soc Interface 10: 20130810
11