Membrane-based sensing approaches

Article abstract of DOI:10.1071/CH10347

Tethered bilayer lipid membranes can be used as model platforms to host membrane proteins or membrane-active peptides, which can act as transducers in sensing applications. Here we present the synthesis and characterization of a valinomycin derivative, a depsipeptide that has been functionalized to serve as a redox probe in a lipid bilayer. In addition, we discuss the influence of the molecular structure of the lipid bilayer on its ability to host proteins. By using electrical impedance techniques as well as neutron scattering experiments, a clear correlation between the packing density of the lipids forming the membrane and its ability to host membrane proteins could be shown. CSIRO 2011.

Full text of DOI:10.1071/CH10347

RESEARCH FRONT
Full Paper
CSIRO PUBLISHING
Aust. J. Chem. 2011, 64, 54–61
www.publish.csiro.au/journals/ajc
Membrane-Based Sensing Approaches
A
A
A B C
, ,
Julia Braunagel, Ann Junghans, and Ingo Koper
¨
A
Max Planck Institute for Polymer Research, 55128 Mainz, Germany.
School of Chemical and Physical Sciences, Flinders University,
Adelaide, SA 5001, Australia.
B
C
Corresponding author. Email: ingo.koeper@flinders.edu.au
Tethered bilayer lipid membranes can be used as model platforms to host membrane proteins or membrane-active
peptides, which can act as transducers in sensing applications. Here we present the synthesis and characterization of a
valinomycin derivative, a depsipeptide that has been functionalized to serve as a redox probe in a lipid bilayer. In addition,
we discuss the influence of the molecular structure of the lipid bilayer on its ability to host proteins. By using electrical
impedance techniques as well as neutron scattering experiments, a clear correlation between the packing density of the
lipids forming the membrane and its ability to host membrane proteins could be shown.
Manuscript received: 20 September 2010.
Manuscript accepted: 19 October 2010.
Introduction
thick polymer layer, however, are often decreased electrical
properties of the bilayer, when compared with natural mem-
branes. Yet, a high resistance of the membrane itself and a low
capacitance are essential when the membrane is to be used to
host ion channel proteins and to allow for the recording of
transport processes through those proteins. Typical ion channel
recordings can be made by either using patch clamp systems on
whole cells or by using planar bilayer lipid membrane (BLM)
systems.[12] These approaches have been significantly improved
in the past years in order to enhance the stability of the fragile
bilayer membrane, for example, by the construction of auto-
mated patch clamp systems,[13] or by embedding planar BLM
in supporting hydrogels.[14] Yet, tBLM have been shown to
provide the most stable architectures.[15] A tBLM consists in
principle of a lipid bilayer (Fig. 1). The inner leaflet is covalently
grafted to a solid support through a spacer moiety.[16] These
anchor lipids thus consist of three parts, a lipid group, a spacer
unit, and an anchor group, to allow for the assembly on a solid
substrate. Typically, short oligomeric poly(ethylene glycol)
units have been used. This spacer part separates the bilayer from
the substrate and provides a reservoir underneath the membrane,
where electrolytes can diffuse in. Several systems have been
proposed, with branched hydrocarbon chains in the lipid part;
typical lipids have branched hydrocarbon units, which are in a
fluid phase at ambient conditions.[1,17,18] The anchor unit pro-
vides a covalent linkage of the membrane to the substrate. A
gold surface is commonly used as substrate, allowing for elec-
trical and optical characterization of the system.[19–21] For such
surfaces, thiol moieties are ideal binding groups. In addition,
silanes have been employed to anchor membranes to oxide
surfaces, or phosphonate groups for alumina substrates.[22] The
membrane architecture has been studied in detail using various
surface analytical techniques, including surface plasmon
resonance (SPR) spectroscopy,[23] electrochemical impedance
spectroscopy (EIS),[24] quartz crystal microbalance, atomic
Applications in pharmacy, environmental control, or security
require more and more sensitive detection methods to probe
minute amounts of analyte. Therefore, a wide range of different
sensing approaches has been developed. One possibility is to
mimic concepts developed in nature. Bilayer membrane-based
sensing approaches have the advantage of ideally providing a
highly sensitive detection system.[1–4] In such a system, a
membrane protein is embedded in a lipid bilayer and serves
as the actual sensing unit. The fragile nature of a membrane
protein, however, requires a sophisticated architecture to ensure
its proper function. Typically, transmembrane proteins are
embedded in a lipid bilayer, often composed of a variety of
different lipids. Such lipid mixtures are generally too complex to
be used in practical applications. Therefore, several biomimetic
model systems have been developed, including vesicular sys-
tems, lipid monolayers at the air–water interface, or other model
membrane systems. Among these, solid supported membranes
and in particular tethered bilayer lipid membranes (tBLM) have
been established as useful architectures.[5–8] They provide a
stable and robust sensing platform, where membrane proteins
can be incorporated. The general concept of a solid supported
membrane consists of a lipid bilayer that is linked to a solid
support.[9] This linkage can be either physical, for example,
through electrostatic interactions, or chemical through covalent
bonds. Interactions between the lipid bilayer and the solid sup-
port can disturb the natural behaviour of the lipid layer, for
example, hinder diffusion of lipids and embedded proteins or
even lead to denaturation of such proteins. In order to minimize
these interactions, various models have been proposed to lift the
bilayer off the substrate. One possibility is to insert a polymer
cushion between the inner leaflet of the membrane and the
solid support.[10,11] Such systems have been shown to provide
excellent membrane fluidity and allow for the functional
incorporation of membrane proteins. The disadvantages of a
Ó CSIRO 2011
10.1071/CH10347
0004-9425/11/010054

RESEARCH FRONT
Membrane-Based Sensing Approaches
55
Here, ferrocene has been coupled to valinomycin and the
modified ion carrier has been investigated in a tBLM architec-
ture. The ferrocene/ferrocinium system is a highly reversible
one electron redox pair that creates a positive charge by oxida-
tion, even beyond a water–oil interface.[38] The thus created
positive charge, which is exposed to the solvent, hinders the
membrane permeation of the ferrocene–valinomycin. This pro-
cess is reversible when a reducing agent is added. The system
might be used as membrane-based redox sensor.
RE
RM
CM
Experimental
Synthesis of Ferrocene–Valinomycin
The synthesis of the ferrocene–valinomycin (Fig. 2) required
various reactions, which will be described in terms of standard
operation procedures (SOP). If not stated otherwise, the reaction
progress was followed by TLC, and column chromatography
was performed after each step. The purity of the products was
verified by TLC, HPLC, and NMR spectroscopy.
CS
SOP 1: Didepsides were obtained by reaction of a Boc-
protected amino acid (1 equiv.) and a benzyl- or allyl-protected
hydroxy acid (1 equiv.) with 4-dimethylaminopyridine (DMAP,
0.1equiv.) and N,N0-dicyclohexylcarbodiimide (DCC, 1.1equiv.)
in dichloromethane (DCM) at ꢀ408C for 12 h.[39]
Fig. 1. Schematic depiction of the tethered membrane architecture. The
inner leaflet is covalently bound to a substrate by a spacer unit; the distal
layer is typically formed by vesicle fusion.[25] The electrical properties of the
membrane can be modelled by a simple equivalent circuit.
SOP 2: Depsipeptides were obtained by reaction of a frag-
ment with a free carboxylic acid (1 equiv.) and a fragment with
the free amino acid (1 equiv.) with DMAP (0.1 equiv.) and DCC
(1.1 equiv.) in DCM at room temperature (RT) for 12 h.
SOP 3 (Boc-deprotection): The Boc-protected amino frag-
ment was stirred for 2 h in DCM/trifluoroacetic acid (TFA) (4/1)
and washed with NaHCO3 solution. Solvents were removed
under vacuum and the crude product was used in the next step
without further purification.
force microscopy (AFM),[25] infrared spectroscopy,[26] and
neutron reflectivity (NR).[27]
The membrane architecture has been shown to be stable
over extended periods of time up to several months.[28] At the
same time, the system provides a biomimetic layer with a
high electrical resistance, close to values of the natural analo-
gue.[23] Various peptides and proteins, especially ion channels,
have already been successfully incorporated in a tBLM
platform.[1,5,29]
Here, we will concentrate on two different aspects of the
membrane. First, the synthesis of a novel ionophore, which,
embedded in a tBLM, can be used as a redox sensor, will be
discussed. In the second part, we will discuss the influence of the
membrane structure on its ability to host membrane proteins.
SOP 4 (benzyl-deprotection): The benzyl-protected carbonic
acid fragment was stirred for 3 h in methanol with Pd/C and
hydrogen at RT. The catalyst was removed by filtration, solvents
were removed under vacuum, and the crude product was used in
the next step without further purification.
Compounds 2 and 6 were obtained with yields of 85% and
67% by reaction of hydroxyisovaleric acid (1) with benzyl and
allyl bromide (1.1 equiv.), respectively, potassium carbonate
(1 equiv.) and potassium iodide (0.1 equiv.) in N,N-dimethyl-
formamide (DMF) at RT for 24 h. The didepsides 4, 5, and
15 were prepared following SOP 1 with high yields (91–97%).
The following deprotection steps were carried out quantitatively
according to SOP 3 and 4. The tetradepsipeptide 9 was synthe-
sized according SOP 2 with a yield of 70%. Fifty percent of the
product was benzyl-deprotected (10, SOP 4) and the other half
was Boc-deprotected (11, SOP 3). Both products were coupled
with a 76% yield to the octadepsipeptide 12. Compound 12 was
Boc-deprotected and bound to benzyl-deprotected 4 according
to SOP 2. The yield for the decadepsipeptide 13 was 60%.
Ferrocene carbonic acid (16, 1 equiv.) was activated with
oxalyl chloride (2.5 equiv.) in tetrachloromethane at RT for
2 h to give ferrocene acid chloride (17). Compound 17 was
directly added to a cooled solution of 18 (1 equiv.) in DCM
with N,N-diisopropylethylamine (DIPEA, 2.5 equiv.) and DMAP
(0.1 equiv.) and stirred for 2 h at RT. The yield for 19 was
16%. Compound 19 was stirred in 20% piperidine (DMF) for
2 h, the Fmoc-deprotected didepside 20 was then coupled to the
benzyl-deprotected 14 (27% yield). The linear dodecadepsipep-
tide 21 was allyl-deprotected with Pd(PPh3)4 (0.1 equiv.) and
phenyl silane (2 equiv.) in dry DCM under argon for 2 h at RT.
Ferrocene–Valinomycin as a Redox Sensor
Valinomycin is an ion carrier that has been widely studied in
model membrane systems,[30–32] however, it has not been used
for explicit sensing purposes. The cyclic depsipeptide transports
potassium ions highly selectively across a membrane by com-
plexation with its six ester carbonyl groups. The cation is buried
inside the cavity of the ionophore, which can diffuse through
lipid membranes by exposing multiple hydrophobic side-chains
and release the ion at the other side. Natural valinomycin con-
sists of two amino acids (L- and D-valine) and two a-hydroxy
acids (L-lactic acid and D-hydroxyisovaleric acid). It has a cyclic
structure (c(L-Val-D-Hyiv-D-Val-L-Lac)3) and is usually isolated
from Streptomyces fulvissimus.[33]
The first synthesis of valinomycin was described by
Shemyakin et al.[34] Furthermore, the exchange of one or two
amino or hydroxy acids did not significantly alter the transport
properties of the molecule.[35–37] By exchange of one L-valine
by an L-lysine a free amino group can be introduced to
which different ligands can be coupled through a peptide
bond. Through this ligand, an additional functionality can be
incorporated into the molecule, for example, a control over the
membrane permeability.

RESEARCH FRONT
56
J. Braunagel, A. Junghans, and I. Ko¨per
Br
COOAll
HO
HO COOBn
COOH
HO
Br
HO COOBn
Results and Discussion
KCO3, KI
KCO3, KI
[6]
[1]
[3]
[2]
Incorporation of Ferrocene–Valinomycin in tBLM
In order to probe the transport properties of the new valinomycin
compound, a tBLM consisting of DPhyTL and DPhyPC in 0.1 M
KCl was used. Details about the system have been described
previously.[24] Membranes were assembled on gold electrodes
and characterized by impedance spectroscopy. The obtained
data can typically be analyzed in terms of an equivalent circuit of
resistors (R) and capacitors (C) (Fig. 1). The circuit consists of
an electrolyte resistance (RE), followed by a resistor (RM) and
capacitor (CM) in parallel, as it is typically used to describe pure
lipid bilayers. Finally, the spacer unit and the gold interface are
modelled by one additional capacitor (CS). Such an equivalent
circuit can also be written as R(RC)C. The obtained parameters
for the membrane resistance often vary in a range between
1 MO cm2 and up to 20 MO cm2.
The ferrocene–valinomycin was added to a membrane and
EIS spectra were recorded that showed a similar behaviour
to natural valinomycin (Fig. 3).[40] The obtained data could
be modelled using simple equivalent circuits described above.
The capacitance of the membrane as well as the parameters
for the spacer region remained relatively constant and close to
literature values throughout the experiment. For the bare bilayer,
the resistance of the spacer region was not observable within
the measured frequency range. The corresponding dataset
was thus modelled using a single capacitance to describe the
spacer region. Similar effects have been seen previously.[19]
The behaviour of the systems is actually more complex than
described by the simple equivalent circuits. However, most of
the physical principles of the architecture are well described
by using simple resistors and capacitors. In fact, using more
advanced equivalent circuits containing constant phase ele-
ments[41] did not alter either the absolute values or the quality
of the fits very much.
Boc-L-Valine,
DCC, DMAP
Boc-D-Valine,
DCC, DMAP
Fmoc-L-Lys(Boc),
DCC, DMAP
NHBoc
O
O
COOH
COOBn
NHFmoc
BocHN
O
COOAll
O
O
COOH
[4]
[5]
O
Fe
[16]
Cl
[15]
BocHN
H2
Pd/C
TFA
O
TFA
NH
2 O
O
Cl
O
COOH
COOBn
BocHN
O
O
[8]
[7] ꢀ [8]
DCC, DMAP
NHFmoc
O
[7]
COCl
COOAll
Fe
O
[17]
[18]
H2N
COOBn
O
O
O
H
N
[17] ꢀ [18]
DIPEA, DMAP
O
O
NHBoc
[9]
H2
Pd/C
TFA
O
NHFmoc
O
COOAll
COOH
O
O
COOBn
O
O
O
C
H
N
H
N
O
O
O
O
[19]
N
NHBoc
O
NH2
O
[11]
H
Fe
[10]
1. [10] ꢀ [11]
DCC, DMAP
2. TFA
Piperidine
NH
O
O
O
O
COOBn
Pd/C
H
H
H
N
N
N
2O COOAII
O
O
O
O
O
NH2
O
O
O
[12]
C
NHBoc
O
[20]
N
H
COOH
Fe
DCC, DMAP
[4]
H2
O
O
O
O
O
H
O
O
COOBn
H
H
H
N
N
N
N
O
O
O
O
NHBoc
O
O
O
O
[13]
Pd/C H2
O
O
O
O
O
O
COOH
H
H
H
H
[20] ꢀ [14]
DCC, DMAP
N
N
N
N
O
O
O
O
NHBoc
O
O
O
O
[14]
The functionality of the ionophore could be seen as changes
in the membrane resistance. The incorporation into the mem-
brane resulted in a decrease in the membrane resistance from
,2 MO cm2 to 8 ꢁ 10ꢀ3 MO cm2, when measured in 0.1 M KCl
solution. Rinsing with 0.1 M NaCl led to a re-increase in
resistance to 1.8 MO cm2, because of the high potassium selec-
tivity of valinomycin. Ideally, the resistance should reach the
initial value, however, it was very difficult to eliminate all
potassium ions from the system.
O
O
H
O
O
O
O
O
COOAII
H
H
H
H
N
N
N
N
N
O
O
O
O
C
O
O
NHBoc
O
O
O
O
1. Pd(PPh ) ,
[21]
3
4
phenylsilane
2. TFA
NHFc
O
O
NH2
O
O
O
O
O
COOH
H
N
H
N
H
N
H
N
H
N
O
O
O
O
C
O
O
O
O
O
O
[22]
DPPA,
KHCO3
NHFc
H
N
Ferrocene–Valinomycin as a Redox Sensor
Fe
O
O
O
In order to probe the sensor functionalities of the modified
valinomycin, the ferrocene unit had to be oxidised. In principle,
this can be achieved electrochemically or chemically. The
electrochemical oxidation was not successful, probably because
of the distance between the electrode and membrane, created
by the spacer part of the DPhyTL lipids. Cyclic voltammetry
experiments (0.1–0.8 V) and measurements at fixed potentials
up to 1.0 V did not show any effect.
NH
O
O
O
O
HN
O
O
NH
O
Kꢀ
O
O
HN
O
NH
O
O
H
N
O
O
O
O
Ferrocene–valinomycin
For a successful chemical oxidation, the redox agent should
be water soluble and should not interfere with the membrane
itself. Reducing agents such as ascorbic acid and sodium
thiosulfate had a significant effect on the electrical properties
of the bare membrane. Hydroquinone, however, did not seem
to perturb the membrane over a period of 24 h. Similarly, the
oxidizing agents K3[Fe(CN)6] and iron(III) chloride showed
almost no influence on the membrane, yet the oxidizing effect
of FeCl3 was more pronounced.
Fig. 2. Synthetic pathway of ferrocene–valinomycin.
The yield was 90%. After the Boc-deprotection according to
SOP 3, 22 was cyclizated under high dilution (c 0.005 M) in
DMF with diphenylphosphoryl azide (DPPA, 3 equiv.) and
KHCO3 (5 equiv.) at RT for 24 h. Ferrocene–valinomycin was
obtained in 95% yield.

RESEARCH FRONT
Membrane-Based Sensing Approaches
57
109
108
107
106
105
104
103
102
ꢂ90
ꢂ60
ꢂ30
Membrane resistance [MΩ cm2]
Bilayer in KCl
2.0 ꢁ 0.07
ꢀ Fc–val in KCl
0.008 ꢁ 0.002
ꢀ Fc–val in NaCl
1.8 ꢁ 0.08
101
0
10ꢂ3
10ꢂ2
10ꢂ1
100
101
102
103
104
105
Frequency [Hz]
Fig. 3. Bode plot showing a bilayer with incorporated ferrocene–valinomycin. Solid symbols show the impedance data
(left axis) and open symbols the phase angle (right axis). Solid lines represent fits of the experimental data using an
R(RC)(RC) equivalent circuit, the obtained fitting parameters including fitting errors are listed. After incorporation of
valinomycin, a decrease in resistance because of Kþ transport was observed. Transport could be suppressed by rinsing
with NaCl, and recovery of the original resistance was measured.
109
ꢂ90
108
107
106
105
104
103
102
101
ꢂ60
ꢂ30
Membrane resistance [MΩcm2]
Bilayer
2.0 ꢁ 0.07
ꢀ Fc–val
After oxidation
After reduction
0.008 ꢁ 0.002
0.14 ꢁ 0.02
0.05 ꢁ 0.01
0
10ꢂ3
10ꢂ2
10ꢂ1
100
101
102
103
104
105
Frequency [Hz]
Fig. 4. Bode plot showing a bilayer with incorporated ferrocene–valinomycin. After incorporation of valinomycin
a decrease in resistance attributable to Kþ transport was observed. Transport could be hindered by oxidation of the
ferrocene ligand with 0.1 M FeCl3. A reduction of the ferrocinium ion was also possible and led to a higher ion transport
again.
In order to stop the ion transport through a peptide-
functionalized membrane, the ferrocene was oxidized with
0.1 M FeCl3 for 24 h, resulting in an increase in resistance to
0.14 MO cm2 (Fig. 4). The positive charge on the ion carrier
seemed to hinder the transmembrane movement. There might
be the possibility that the charged depsipeptide could leave
the bilayer, which would also lead to the observed changes in
electrical properties. However, in that case we would expect
a more pronounced effect. Furthermore, the process could be
reversed by the addition of 0.1 M hydroquinone (þ0.1 M KCl)
for 24 h to the system. The resistance then again decreased to
0.05 MO cm2.
The novel compound might be used in a redox sensor
concept, where the ferrocene unit is used as a switch that allows
or permits ion transport. The system would thus be a logic gate
(AND operation) with the redox status of ferrocene and the
presence of potassium ions as input channels, while the output
would be the ion flow. Yet, a reaction time of 24 h is definitely

RESEARCH FRONT
58
J. Braunagel, A. Junghans, and I. Ko¨per
DPhyTL
DPhyHDL
Fig. 5. Structures of the anchor lipids DPhyTL and DPhyHDL.
not suitable for a viable sensor. This time span is attributable to
the fact that the molecules have to cross a water–membrane
barrier. However, the aim of this research was to show a proof of
concept for the system. Further optimization and modification
would be needed to enhance the reaction time of the molecules.
defects in the membrane. These electrical properties can be
easily analyzed using EIS. By comparing different molecular
structures of the anchor lipid their influence on the membrane
properties can be probed.
We have studied the molecular structure and the resulting
membrane properties of two distinctive anchor lipids (Fig. 5).[44]
Both DPhyTL and DPhyHDL consist of two phytanyl chains as
the hydrophobic part. The spacer part consists of a tetra- or hexa-
ethylene glycol unit, respectively. While DPhyTL is anchored to
the substrate by a single lipoic acid group, two units are used for
DPhyHDL. The rationale for the structures is that DPhyTL has
been shown to form very densely packed layers,[23] which might
obstruct the incorporation of large membrane proteins. The
two anchor groups in DPhyHDL should lead to a more diluted
packing in the inner leaflet of the bilayer and underneath the
membrane, thus generating more space for proteins to integrate
and function. This strategy is an alternative to the use of small
molecules as diluting units in the lipid monolayer.[45,46]
The more branched structure of DPhyHDL should lead to
a looser packing of the lipid monolayer. This should have a
significant effect on the structural properties and on the ability to
host membrane proteins.[47,48] This hypothesis has been
addressed using EIS and NR studies.[27,44]
NR is an ideal tool to study soft interfaces such as the tBLM
architectures. The technique offers the unique opportunity to
analyze individually the structures of the different layers of a
tBLM system, i.e., the spacer region and the lipid layer. The
properties of the spacer part are of special interest. This part
should not only lift off the lipid bilayer from the underlying
substrate, but also provide an ion reservoir to enable the function
of membrane proteins and also accommodate extra-membrane
parts of the proteins. Thus it is interesting to analyze how thick
the ethylene glycol part of the layer effectively is and how much
water can be incorporated. Furthermore, NR allows differences
in the tBLM structures to be probed, when the molecular
structure of the anchor lipids is modified.
Structure–Function Relationship of tBLM
The structure of a tBLM is composed of three distinct parts, the
hydrophobic lipid layer, the spacer part, and the anchor moiety.
All parts have a certain influence on the final structure and
properties of the bilayer architecture.
The assembly of a tBLM is typically a two-step process.
First, the anchor lipids bind to a suitable substrate and form a
monolayer. For gold substrates, this is typically a self-assembly
process, which can be easily followed by various surface
analytical tools. Alternatively, and especially for oxidic sub-
strates using silane anchors, the transfer of a lipid film which has
been pre-arranged at the air–water interface can be a suitable
deposition method.[42,43]
For the self-assembly processes, different timescales have
been observed, depending on the technique utilized. For exam-
ple, the formation of the monolayer takes about 3 h, when
monitored optically using SPR. When studied in more detail
using EIS or AFM, it has been shown that it takes up to 24 h
until the monolayer is completely formed and any defects are
filled.[25] Similarly, the completion of the monolayer to a full
membrane in the second step, which is typically done by fusion
with small unilamellar vesicles, is finished optically much faster
than electrically. This is probably a result of rearrangements
happening within the membrane, even after the fusion of the
vesicles.
In particular, the structure of the anchor lipid itself also has a
significant influence on the properties of the formed tBLM. In
most cases, phytanoyl moieties have been used. These archae-
analogue lipid structures have low transition temperatures,
providing fluid membrane architectures at ambient conditions.
Furthermore, the branched chains lead to a strong interaction
between the hydrophobic chains, resulting in a relatively dense
packing.[42] The spacer group is supposed to lift off the bilayer
from the substrate and to provide an electrolyte reservoir under-
neath the membrane. The lateral structure of the membrane,
i.e., the molecular details of the lipid assembly, will have a
significant influence on the properties of the bilayer itself.
tBLM have been used to incorporate ion channel proteins and
study their function. Therefore, the lipid bilayer should have a
high electrical resistance, to ensure that observed transport
processes are only attributable to protein function and not to
NR experiments have been performed using DPhyTL and
DPhyHDL-based membranes (Fig. 6).[15,27] In order to create a
signal with a sufficient intensity, membranes had to be prepared
on 3 inch (7.62 cm) silicon wafers as substrates, which have been
coated with a thin gold film.
The obtained experimental data could be analyzed in terms
of a layer model, with distinctive differences in the scattering
length density for the spacer part, and the inner and the outer
bilayer leaflet. The results of NR scattering experiments are
often presented as scattering length density (SLD) profiles,
where the SLD corresponds to the isotopic composition at a
certain distance from the substrate surface. In the present case,

RESEARCH FRONT
Membrane-Based Sensing Approaches
59
(a) 7 ꢃ 10ꢂ6
6 ꢃ 10ꢂ6
5 ꢃ 10ꢂ6
4 ꢃ 10ꢂ6
3 ꢃ 10ꢂ6
2 ꢃ 10ꢂ6
1 ꢃ 10ꢂ6
0
DPhyTL (D2O)
DPhyTL ꢀ DPhyPC (D2O)
0
50
Distance from Au interface [Å]
(b) 7 ꢃ 10ꢂ6
6 ꢃ 10ꢂ6
5 ꢃ 10ꢂ6
4 ꢃ 10ꢂ6
3 ꢃ 10ꢂ6
2 ꢃ 10ꢂ6
1 ꢃ 10ꢂ6
0
DPhyHDL (D2O)
DPhyHDL ꢀ DPhyPC (D2O)
200
250
300
350
Distance from Si interface [Å]
Fig. 6. Neutron reflectivity profiles for (a) DPhyTL and (b) DPhyHDL based membranes. Measurements have
been conducted in heavy water for mono- and bilayer assemblies. SLD, scattering length density.
the substrate was a gold film sputtered onto a silicon wafer. Thus
addition, a relatively high amount of water was incorporated
in the spacer region (,65%) and even in the alkyl part of
the membrane. The anchor lipid containing the two lipoic acid
groups as anchor moieties thus led to a more diluted membrane
structure. This had a significant influence on the electrical
properties of the membranes as well as on the ability to host
membrane proteins, here the ion channel a-hemolysin.
The electrical properties of a DPhyHDL-based membrane
showed significantly different values from those obtained for a
DPhyTL-based bilayer (Fig. 7). While resistance values up to
20 MO cm2 have been reported for DPhyTL, the anchor lipid
with the larger anchor group led to maximum values of
,1 MO cm2, as could be expected from the presumably looser
packing. At the same time, both membrane architectures
showed typical capacitance values below 1 mF cmꢀ2, indicative
of the formation of a complete lipid bilayer. The lipid
bilayers have then been exposed to a solution of a-hemolysin.
This protein is known to form heptameric pores in lipid
membranes, allowing for the flux of molecules across the
the profile plots the composition of the membrane perpendicular
to the gold interface (Fig. 6). By using the difference in the SLD
of protons and deuterons, different contrasts can be created
and specific parts of the molecular architecture analyzed. For
example, it was interesting to evaluate the amount of water
incorporated into the spacer part of the anchor lipids. Further-
more, the packing density of the lipids was of interest, since
these parameters should have a direct influence on the ability of
the created tBLM to host membrane proteins. Here, the profiles
for DPhyTL show a very high packing density in the lipid
region, both for the anchor lipids as well as for the lipids
completing the bilayer structure. Similarly, the SLD of the
spacer region is relatively high for a polymer layer and the
difference in SLD for measurements in D2O and a mixture of
D2O and H2O showed that the spacer part contained only ,5%
water. In contrast to these values, the DPhyHDL membrane
showed a much looser structure. The lipid density was increased,
which indicated a more diluted packing of the lipid chains. In

RESEARCH FRONT
60
J. Braunagel, A. Junghans, and I. Ko¨per
(a) 108
107
ꢂ90
ꢂ80
ꢂ70
ꢂ60
ꢂ50
ꢂ40
ꢂ30
ꢂ20
106
105
104
103
DPhyTL bilayer
+ α-Hemolysin
102
ꢂ10
10ꢂ3
10ꢂ2
10ꢂ1
100
101
102
103
104
105
Frequency [Hz]
(b) 108
107
ꢂ90
ꢂ80
ꢂ70
ꢂ60
ꢂ50
ꢂ40
ꢂ30
ꢂ20
ꢂ10
105
106
105
104
103
DPhyHDL bilayer
+ α-Hemolysin
102
10ꢂ3
10ꢂ2
10ꢂ1
100
101
102
103
104
Frequency [Hz]
Fig. 7. Bode plots showing the incorporation of a-hemolysin into two different tBLM architectures. Solid lines
represent fits to an R(RC)C equivalent circuit. The model has been kept consistent for all experiments in order to be able to
compare the data, even though the deviations of the experimental data from the model increased, especially after protein
incorporation.
bilayer.[49,50] In electric measurements, this incorporation
can be seen as a decrease in the membrane resistance. Indeed,
in both the DPhyTL and the DPhyHDL-based tBLM systems,
the addition of the protein (180 nM) resulted in a significant
drop in the membrane resistance. The difference in the local
structures of the two membrane architectures could be seen in
the amplitude of this reduction. While the resistance for the
DPhyTL-based membrane was reduced by about one order
of magnitude to values of 1.5 MO cm2, the resistance for
DPhyHDL-based membranes decrease by a factor of ,2700
to values of 0.8 kO cm2. Simultaneously, the capacitive values
of the bilayers still showed the presence of a lipid bilayer leaflet.
In order to be able to compare the obtained results, the electrical
data has been analyzed using a simple equivalent circuit model,
as discussed above. However, the experimental data deviated
significantly from the theoretical model once the protein was
incorporated, especially in the case of DPhyHDL. This is
probably because of the more inhomogeneous membrane archi-
tecture after the protein is embedded into the bilayer, which
renders the whole system more complex. In fact, fitting the data
using constant phase elements improves the quality of the fit
results, however, it no longer allows for a quantified comparison
of the data.
Conclusion
Tethered bilayer lipid membranes are useful model archi-
tectures that can be employed as a generic platform to study

RESEARCH FRONT
Membrane-Based Sensing Approaches
61
membrane-related processes such as the functional incorpora-
tion of membrane proteins. Furthermore, they can be used as a
sensing platform that can host engineered sensing peptides such
as the here presented valinomycin derivatives. By being able
to attach a selected binding group to the still functional depsi-
peptide, a wide range of sensing possibilities can be envisioned.
Here, as a proof of concept, the synthesis of a redox sensor has
been shown.
The bare membrane architectures can also be used for
sensing purposes. For example, the presence of membrane
active peptides or proteins can be detected by monitoring
changes in the electrical properties of the membrane. However,
the intrinsic membrane architecture, i.e., the molecular structure
of the anchor lipids used in the formation of the system, has a
significant influence both on the properties of the membrane as
well as on its ability to host membrane proteins.
[21] S. Terrettaz, H. Vogel, MRS Bull. 2005, 30, 207.
[22] R. F. Roskamp, I. K. Vockenroth, N. Eisenmenger, J. Braunagel,
I. Ko¨per, ChemPhysChem 2008, 9, 1920. doi:10.1002/CPHC.
200800248
[23] R. Naumann, S. M. Schiller, F. Giess, B. Grohe, K. B. Hartman,
I. Karcher, I. Ko¨per, J. Lubben, K. Vasilev, W. Knoll, Langmuir 2003,
19, 5435. doi:10.1021/LA0342060
[24] I. Ko¨per, Mol. Biosyst. 2007, 3, 651. doi:10.1039/B707168J
[25] I. K. Vockenroth, C. Rossi, M. R. Shah, I. Ko¨per, Biointerphases 2009,
4, 19. doi:10.1116/1.3122019
[26] J. Kunze, J. Leitch, A. L. Schwan, R. J. Faragher, R. Naumann,
S. Schiller, W. Knoll, J. R. Dutcher, J. Lipkowski, Langmuir 2006,
22, 5509. doi:10.1021/LA0535274
[27] A. Junghans, I. Ko¨per, Langmuir 2010, 26, 11035. doi:10.1021/
LA100342K
[28] I. K. Vockenroth, C. Ohm, J. W. F. Robertson, D. J. McGillivray,
M. Lo¨sche, I. Ko¨per, Biointerphases 2008, 3, FA68. doi:10.1116/
1.2912097
[29] H. Hillebrandt, M. Tanaka, E. Sackmann, J. Phys. Chem. B 2002, 106,
Accessory Publication
477. doi:10.1021/JP011693O
[30] G. Stark, R. Benz, J. Membr. Biol. 1971, 5, 133. doi:10.1007/
BF02107720
Analytical data for ferrocene–valinomycin and EIS measure-
ments are available on the Journal’s website.
[31] G. Favero, L. Campanella, A. D’Annibale, R. Santucci, T. Ferri,
Microchem. J. 2003, 74, 141. doi:10.1016/S0026-265X(02)00179-0
[32] B. Raguse, V. L. B. Braach-Maksvytis, B. A. Cornell, L. G. King,
P. D. J. Osman, R. J. Pace, L. Wieczorek, Langmuir 1998, 14, 648.
doi:10.1021/LA9711239
References
[1] B. A. Cornell, V. L. B. Braach-Maksvytis, L. G. King, P. D. J. Osman,
B. Raguse, L. Wieczorek, R. J. Pace, Nature 1997, 387, 580.
doi:10.1038/42432
[33] H. Brockmann, G. Schmidt-Kastner, I. Valinomycin, Chem. Ber.
[2] M. Winterhalter, Curr. Opin. Colloid Interface Sci. 2000, 5, 250.
doi:10.1016/S1359-0294(00)00063-7
[3] G. M. Bell, L. L. Combs, L. J. Dunne, Chem. Rev. 1981, 81, 15.
doi:10.1021/CR00041A002
[4] A. Ottova, V. Tvarozek, J. Racek, J. Sabo, W. Ziegler, T. Hianik,
H. T. Tien, Supramol. Sci. 1997, 4, 101. doi:10.1016/S0968-5677(96)
00054-5
1955, 88, 57. doi:10.1002/CBER.19550880111
[34] M. M. Schemjakin, Angew. Chem. 1960, 72, 342. doi:10.1002/ANGE.
19600721003
[35] M. Rothe, W. Kreiss, Angew. Chem. Int. Ed. Engl. 1973, 12, 1012.
doi:10.1002/ANIE.197310121
[36] C. Gilon, Y. Klausner, A. Hassner, Tetrahedron Lett. 1979, 20, 3811.
doi:10.1016/S0040-4039(01)95531-5
[5] E. Sackmann, Science 1996, 271, 43. doi:10.1126/SCIENCE.271.
5245.43
[37] B. F. Gisin, A. R. Dhundale, Int. J. Pept. Protein Res. 1979, 14, 356.
doi:10.1111/J.1399-3011.1979.TB01944.X
[38] J. Hanzl´ık, Z. Samec, J. Hovorka, J. Electroanal. Chem. 1987, 216,
[6] C. Steinem, A. Janshoff, W.-P. Ulrich, M. Sieber, H.-J. Galla, Biochim.
Biophys. Acta 1996, 1279, 169. doi:10.1016/0005-2736(95)00274-X
[7] W. Knoll, C. W. Frank, C. Heibel, R. Naumann, A. Offenha¨user,
J. Ru¨he, E. K. Schmidt, W. W. Shen, A. Sinner, Rev. Mol. Biotechnol.
2000, 74, 137. doi:10.1016/S1389-0352(00)00012-X
[8] A. E. Vallejo, C. A. Gervasi, Bioelectrochem 2002, 57, 1.
doi:10.1016/S1567-5394(01)00127-X
303. doi:10.1016/0022-0728(87)80217-6
[39] B. Neises, W. Steglich, Angew. Chem. Int. Ed. Engl. 1978, 17, 522.
doi:10.1002/ANIE.197805221
[40] R. Naumann, D. Walz, S. M. Schiller, W. Knoll, J. Electroanal. Chem.
2003, 550–551, 241. doi:10.1016/S0022-0728(03)00013-5
[41] G. Valincius, D. J. McGillivray, W. Febo-Ayala, D. J. Vanderah,
J. J. Kasianowicz, M. Lo¨sche, J. Phys. Chem. B 2006, 110, 10213.
doi:10.1021/JP0616516
[9] E. Sackmann, Science 1996, 271, 43. doi:10.1126/SCIENCE.271.
5245.43
[10] J. Spinke, J. Yang, H. Wolf, M. Liley, H. Ringsdorf, W. Knoll,
Biophys. J. 1992, 63, 1667. doi:10.1016/S0006-3495(92)81742-3
[11] L. K. Tamm, H. M. McConnell, Biophys. J. 1985, 47, 105.
doi:10.1016/S0006-3495(85)83882-0
[12] M. Winterhalter, Curr. Opin. Colloid Interface Sci. 2000, 5, 250.
doi:10.1016/S1359-0294(00)00063-7
[42] P. P. Atanasova, V. Atanasov, I. Ko¨per, Langmuir 2007, 23, 7672.
doi:10.1021/LA7002854
[43] V. Atanasov, N. Knorr, R. S. Duran, S. Ingebrandt, A. Offenha¨user,
W. Knoll, I. Ko¨per, Biophys. J. 2005, 89, 1780. doi:10.1529/
BIOPHYSJ.105.061374
[44] I. K. Vockenroth, P. P. Atanasova, A. T. A. Jenkins, I. Ko¨per,
[13] N. Fertig, A. Tilke, R. H. Blick, J. P. Kotthaus, J. C. Behrends, G. ten
Bruggencate, Appl. Phys. Lett. 2000, 77, 1218. doi:10.1063/1.1289490
[14] T. J. Jeon, N. Malmstadt, J. J. Schmidt, J. Am. Chem. Soc. 2006, 128,
42. doi:10.1021/JA056901V
[15] I. K. Vockenroth, C. Ohm, J. W. F. Robertson, D. J. McGillivray,
M. Lo¨sche, I. Ko¨per, Biointerphases 2008, 3, FA68. doi:10.1116/
1.2912097
[16] I. Ko¨per, S. M. Schiller, F. Giess, R. Naumann, W. Knoll, Adv. Planar
Lipid Bilayers 2006, 3, 37. doi:10.1016/S1554-4516(05)03002-4
[17] S. M. Schiller, R. Naumann, K. Lovejoy, H. Kunz, W. Knoll, Angew.
Chem. 2003, 42, 208. doi:10.1002/ANIE.200390080
[18] S. Terrettaz, M. Mayer, H. Vogel, Langmuir 2003, 19, 5567.
doi:10.1021/LA034197V
Langmuir 2008, 24, 496. doi:10.1021/LA7030279
[45] L. H. He, J. W. F. Robertson, J. Li, I. Karcher, S. M. Schiller, W. Knoll,
R. Naumann, Langmuir 2005, 21, 11666. doi:10.1021/LA051771P
[46] D. J. McGillivray, G. Valincius, D. J. Vanderah, W. Febo-Ayala, J. T.
Woodward, F. Heinrich, J. J. Kasianowicz, M. Losche, Biointerphases
2007, 2, 21. doi:10.1116/1.2709308
[47] J. A. Dura, D. J. Pierce, C. F. Majkrzak, N. C. Maliszewskyj, D. J.
McGillivray, M. Lo¨sche, K. V. O’Donovan, M. Mihailescu, U. Perez-
Salas, D. L. Worcester, S. H. White, Rev. Sci. Instrum. 2006, 77,
074301. doi:10.1063/1.2219744
[48] S. Krueger, C. W. Meuse, C. F. Majkrzak, J. A. Dura, N. F. Berk,
M. Tarek, A. L. Plant,Langmuir 2001, 17, 511. doi:10.1021/LA001134T
[49] B. Schuster, D. Pum, O. Braha, H. Bayley, U. B. Sleytr, Biochim.
Biophys. Acta 1998, 1370, 208.
[19] I. K. Vockenroth, P. P. Atanasova, A. T. A. Jenkins, I. Ko¨per,
Langmuir 2008, 24, 496. doi:10.1021/LA7030279
[20] C. A. Keller, B. Kasemo, Biophys. J. 1998, 75, 1397. doi:10.1016/
S0006-3495(98)74057-3
[50] S. Cheley, M. S. Malghani, L. Song, M. Hobaugh, J. E. Gouaux,
J. Yang, H. Bayley, Protein Eng. 1997, 10, 1433. doi:10.1093/
PROTEIN/10.12.1433