Rights / License: Research Collection In Copyright - Non ...263/eth-2… · membrane protein transporters, operating at lower rates than ion channels and requiring higher numbers

Research Collection

Doctoral Thesis

Investigations on planar lipid bilayers in nano-pores to study thefunction of membrane proteins

Author(s): Studer, André Olivier

Publication Date: 2009

Permanent Link: https://doi.org/10.3929/ethz-a-005901902

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-005901902

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

Doctoral Thesis ETH No. 18473

Investigations on planar lipid bilayers in nano-pores to study

the function of membrane proteins

A dissertation submitted to

ETH ZURICH

For the degree of

DOCTOR OF SCIENCES

Presented by

ANDRÉ OLIVIER STUDER

Dipl. Natw. ETH Zürich

born 02.03.1978

citizen of

Winterthur ZH, Switzerland

accepted on the recommendation of

Prof. Dr. Fritz K. Winkler, examiner

Prof. Dr. Kaspar Locher, co-examiner

Prof. Dr. Janos Vörös, co-examiner

Dr. Louis X. Tiefenauer, co-examiner

2009

Table of Contents

iii

Table of Contents

Table of Contents....................................................................................... iii

Abbreviations ............................................................................................ vii

Summary .....................................................................................................ix

Zusammenfassung ......................................................................................xi

Acknowledgments.................................................................................... xiii

1. Introduction .............................................................................................1

1.1 The structure of the cell membrane ........................................................................ 1

1.2 The lipid bilayer...................................................................................................... 2

1.3 Proteoliposomes as model systems......................................................................... 4

1.4 Heterologous expression systems for electrophysiology........................................ 6

1.5 Synthetic planar lipid bilayers ................................................................................ 7

1.6 Aim of the thesis ................................................................................................... 10

2. Characterization of free-standing lipid bilayers in nanopores .........12

2.1 Measurement setup ............................................................................................... 12

2.2 Chip with nanopores ............................................................................................. 13

2.2.1 Chip production ............................................................................................. 13

2.2.2 Chip types ...................................................................................................... 14

2.3 Lipid bilayer formation......................................................................................... 15

2.3.1 Painting .......................................................................................................... 15

2.3.2 Müller – Montal ............................................................................................. 17

2.3.3 Liposome on chip adsorption......................................................................... 17

2.4 Lipid bilayer characterization ............................................................................... 18

2.4.1 Electrochemical impedance spectroscopy ..................................................... 18

2.4.2 Lipid bilayer capacitance ............................................................................... 21

2.4.3 Lipid bilayer resistance .................................................................................. 22

2.5 Lipid bilayer stability............................................................................................ 23

2.6 Bilayers in nanopores: concluding remarks and perspectives .............................. 27

3. Examination of the usefulness of bilayers in arrays of nanopores for

functional studies using model proteins ..................................................29

3.1 Melittin.................................................................................................................. 29

3.1.1 Interaction of melittin with lipid bilayers ...................................................... 30

Table of Contents

iv

3.1.2 Voltage dependence of melittin induced conductivity .................................. 31

3.2 Hemolysin from Staphylococcus aureus .............................................................. 32

3.2.1 α-HLY characterization................................................................................. 33

3.2.2 Single channel current ................................................................................... 34

3.2.3 Onset of α-HLY pore formation depends on monomer concentration.......... 35

3.2.4 The chip pore diameter has an impact on the measured lag time .................. 36

3.2.5 Two-sided symmetric toxin addition ............................................................. 38

3.3 Diffusion through aqueous pores .......................................................................... 38

3.3.1 Quantitative sodium diffusion measurements................................................ 39

3.3.2 Diffusion experiments with calcium.............................................................. 42

3.4 Using model proteins: concluding remarks and perspectives............................... 43

4. Measurements with the voltage gated sodium channel )aChBac....45

4.1 NaChBac of Bacillus halodurans ......................................................................... 45

4.1.1 The ion selective pore .................................................................................... 47

4.1.2 The gate.......................................................................................................... 47

4.1.3 The voltage sensor ......................................................................................... 48

4.2 Expression, purification and reconstitution of NaChBac ..................................... 49

4.2.1 Expression vectors ......................................................................................... 49

4.2.2 Small-scale test expression ............................................................................ 49

4.2.3 Large-scale over-production .......................................................................... 50

4.2.4 Reconstitution ................................................................................................ 51

4.3 Formation of bilayers with integrated NaChBac .................................................. 52

4.3.1 Müller – Montal ............................................................................................. 52

4.3.2 Nystatin / ergosterol vesicle fusion................................................................ 53

4.4 Measurements with reconstituted NaChBac......................................................... 59

4.5 Measuring NaChBac: concluding remarks and perspectives................................ 63

5 Conclusions and outlook ........................................................................64

6. Materials and methods..........................................................................67

6.1 Molecular biology methods .................................................................................. 67

6.1.1 NaChBac constructs....................................................................................... 67

6.1.2 Restriction enzyme digest .............................................................................. 67

6.1.3 Agarose gel electrophoresis ........................................................................... 67

6.1.4 Transformation of competent cells ................................................................ 67

6.1.5 Plasmid preparation ....................................................................................... 68

Table of Contents

v

6.2 Microbiology methods.......................................................................................... 68

6.2.1 Expression in E. Coli ..................................................................................... 68

6.2.2 Harvesting and disruption.............................................................................. 68

6.2.3 Purification..................................................................................................... 69

6.3 Protein analysis and characterization.................................................................... 69

6.3.1 Protein concentration determination .............................................................. 69

6.3.2 SDS-PAGE .................................................................................................... 69

6.3.3 Static light-scattering ..................................................................................... 70

6.4 (Proteo-) liposome formation................................................................................ 70

6.4.1 Reconstitution ................................................................................................ 70

6.4.2 Extrusion........................................................................................................ 70

6.4.3 Protein / lipid ratio determination .................................................................. 71

6.4.4 Dynamic light scattering ................................................................................ 72

6.5 Chip characterization ............................................................................................ 72

6.5.1 Light microscope ........................................................................................... 72

6.5.2 SEM ............................................................................................................... 72

6.5.3 Contact angle measurements ......................................................................... 72

6.6 Lipid bilayer formation......................................................................................... 72

6.6.1 Painting .......................................................................................................... 72

6.6.2 Müller – Montal ............................................................................................. 72

6.6.3 Nystatin-Ergosterol vesicle fusion................................................................. 72

6.7 Electrochemical methods...................................................................................... 73

6.7.1 EIS ................................................................................................................. 73

6.7.2 Normal pulse voltammetry ............................................................................ 73

6.7.3 Linear sweep voltammetry............................................................................. 73

6.7.4 Chronoamperometry ...................................................................................... 73

6.7.5 Four-electrode setup ...................................................................................... 73

6.8 Diffusion ............................................................................................................... 73

7. References...............................................................................................75

Appendix.....................................................................................................85

Diffusion: formulas and calculations .......................................................................... 85

List of publications ....................................................................................87

Abbreviations

vii

Abbreviations

ABC ATP-binding cassette

AFM atomic force microscopy

α-HLY alpha-hemolysin of Staphylococcus aureus

ATP adenosine triphosphate

BLM bilayer lipid membrane or black lipid bilayer

β-OG detergent n-octyl beta-D-glucopyranoside

cDNA complementary DNA

CHO Chinese hamster ovary

DDM n-dodecyl beta-D-maltoside

DLS dynamic light-scattering

DOPC 1,2-dioleoyl-sn-glycero-3-phosphocholine

DOPG 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)]

DPhPC 1,2-diphytanoyl-sn-glycero-3-phosphocholine

EIS electrochemical impedance spectroscopy

HEK293 human embryonic kidney cell line

HRP horseradish peroxidase

IPTG isopropyl-beta-D-thiogalactopyranoside

IMAC immobilized metal ion affinity chromatography

ISE ion selective electrode

M-M Müller – Montal method

MOPS 3-[N-morpholino] propanesulfonic acid

NaChBac voltage gated sodium channel of Bacillus halodurans

pBLM1 polymer-cushioned bilayer lipid membranes

PEG-DMA poly-(ethylene glycol) dimethacrylate

PFT pore forming toxins

PMMA polymethyl methacrylate

POPC 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

POPE 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine

PTFE polytetrafluoroethylene

QCM quartz crystal microbalance

Rpm revolutions per minute

Abbreviations

viii

SEM scanning electron microscope

sBLM supported bilayer lipid membrane

SLS static light-scattering

SPC soy PC

SPR surface plasmon resonance

STIF surface tension induced fusion

tBLM tethered bilayer lipid membrane

Tris tris(hydroxymethyl)-aminomethane

U enzymatic unit (1 U = 1 µmol min-1)

USL unstirred layer

v/v volume per volume

w/v weight per volume

w/w weight per weight

Summary

ix

Summary

Planar lipid bilayers were recognized as model systems for cellular membranes almost

50 years ago. They offer the basic properties such as membrane thickness and lateral

lipid mobility; additionally they provide a natural environment for isolated membrane

proteins. The possibility to tightly control the composition of the artificial membrane

and the free access to both sides of the bilayer makes them ideally suited to investigate

the function of isolated membrane proteins. A significant disadvantage is their poor

stability that varies with the used lipid and the formation technique. Based on the

assumption that an increased ratio of the circumference to bilayer area has a stabilizing

effect, small pores were arranged in arrays. A major part of my thesis was to investigate

whether lipid bilayers formed in arrays of small pores are more stable than in larger

pores and thus improve the conditions to investigate the function of membrane proteins.

Our arrays are based on a thin silicon nitride membrane. I report about the stabilizing

effect on membranes by reducing the pore diameter from 800 nm to 200 nm that was

observed for all lipids tested. In the case of the natural lipid soy PC, bilayers were stable

for at least two days in the smaller pores, which corresponds to an improvement by a

factor of about 30 compared to the larger pores. The precise determination of the

number of bilayer spanned pores in the array was affected by the electrical noise

introduced by the silicon-based device. But the successful ion channel recording from

the pore forming toxin α-hemolysin assembled from seven monomers, proved the

presence of a lipid bilayer in chip pores. The very large total pore area allowed the

incorporation of a large number of toxin molecules which enabled us to measure

directly the amount of sodium ions diffusing through the protein pores using ion

selective electrodes. This demonstrated that the setup is also suitable to investigate

membrane protein transporters, operating at lower rates than ion channels and requiring

higher numbers of protein molecules to make solute transport measurable.

Different methods to form protein containing lipid bilayers were explored and adapted.

The fusion of proteoliposomes with a preformed bilayer using the nystatin-ergosterol

method resulted in the successful incorporation of a voltage gated sodium channel. We

found that this method was limited to larger chip pores and fusion events were not

observed with bilayers smaller than 30 µm2. In using this method, a compromise has to

be accepted between the chance of liposome fusion in larger pores and sufficient bilayer

Summary

x

stability in smaller pores. Furthermore the incorporation of protein molecules to high

densities cannot be achieved. This would be possible if lipid bilayers on a nanopore chip

could be directly formed from proteoliposomes, but this method was not successfully

implemented so far. However, investigations on protein channels and pores are perfectly

possible and the exceptional stability of the lipid bilayers spanning small pores

improves the measurement possibilities. The duration of an experiment is no longer

highly limited by the lifetime of the lipid bilayer and the possible exchange of the buffer

solutions allows using a prepared nanopore array chip for multiple measurements.

Zusammenfassung

xi

Zusammenfassung

Planare Lipiddoppelschichten wurden bereits vor fast 50 jahren als Modellsystem für

die Zellmembran beschrieben. Sie besitzen die selben Grundeigenschaften, zu denen

beispielsweise die Membrandicke und die laterale Mobilität der Lipidmoleküle gehören;

zusätzlich bieten sie eine natürliche Umgebung für aufgereinigte Membranproteine. Die

Möglichkeit, die Zusammensetzung dieser künstlichen Membran exakt zu definieren,

sowie der uneingeschränkte Zugang zu beiden Seiten der Doppelschicht, macht sie zu

einem idealen Kandidaten um die Funktionen von aufgereinigten Membranproteinen zu

untersuchen. Der grosse Nachteil ist ihre geringe Stabilität, die sowohl vom benutzten

Lipid sowie der angewandten Herstellungstechnik abhängt. Basierend auf der Annahme,

dass sich ein grösseres Verhältnis zwischen Umfang zu Fläche stabilisierend auf die

Doppelschicht auswirkt, wurden kleine Poren in Reihen angeordnet. Ein wesentlicher

Teil meiner Doktorarbeit bestand in der Ermittlung ob planare Lipiddoppelschichten,

welche in regulär angeordneten, kleinen Poren gebildet werden, stabiler sind und somit

die Untersuchung von isolierten Membranproteinen vereinfachen.

Unsere regulär angeordneten Reihen von Nanoporen wurden in einer dünnen Membran

aus Siliziumnitrid angefertigt. Durch eine Reduktion des Porendurchmessers von

800 nm auf 200 nm konnten unterschiedliche Lipidmembranen stabilisiert werden. Im

Falle des natürlich vorkommenden Lipidgemisches Soja PC waren die Doppelschichten

in den kleinen 200 nm Poren mindestens 2 Tage stabil, dies entspricht einer 30 fachen

Verlängerung verglichen mit den 800 nm Poren. Die genaue Bestimmung der Anzahl an

Lipiddoppelschichten wurde aufgrund der spezifischen Chipeigenschaften

beeinträchtigt. Aber durch die Porenbildung des Toxins α-hemolysin in zahlreichen

Chipporen konnte das Vorhandensein ebensovieler Doppelschichten bewiesen werden.

Die grosse Gesamtfläche ermöglichte den Einbau von vielen Toxin Molekülen,

wodurch die erleichterte Diffusion von Natriumionen mit ionenselektiven Elektroden

gemessen werden konnte. Der Chip bietet somit die Möglichkeit, Membranprotein

Transporter zu untersuchen, welche langsamer als Ionenkanäle arbeiten und in einer

grösseren Anzahl in der Lipiddoppelschicht vorhanden sein müssen damit der

Stofftransport detektierbar wird.

Verschiedene Methoden um Lipidmembranen mit Proteinen herzustellen wurden

geprüft und modifiziert. Die Verschmelzung von Protein-enthaltenden Liposomen mit

vorgeformten Doppelschichten wurde mit der Nystatin-Ergosterol Methode realisiert.

Zusammenfassung

xii

Diese Methode war aber auf Lipidmembranen beschränkt welche grösser als 30 µm2

waren. Somit muss ein Kompromiss eingegangen werden zwischen erfolgreicher Fusion

in grossen Poren und Stabilität in kleineren Poren. Ausserdem eignet sich diese

Methode nicht zum Einfügen von einer grossen Anzahl von Proteinmolekülen. Dies

wäre möglich falls die Doppelschicht direkt von Protein-enthaltenden Liposomen

gebildet werden könnte, bis jetzt konnte diese Methode aber nicht realisiert werden.

Andererseits ist die Untersuchung von einzelnen Ionen-Kanälen sehr gut möglich und

die aussergewöhnlich hohe Stabilität der Lipidmembranen vereinfacht die Untersuchung

von isolierten Membranproteinen. Die Dauer einer Messung wird nicht mehr so stark

von der Lebensdauer der Lipiddoppelschicht eingeschränkt und ein Austauschen der

wässrigen Lösungen ermöglicht das mehrmalige Verwenden der gleichen künstlichen

Zellmembran.

Acknowledgments

xiii

Acknowledgments

First of all, I am very grateful to my doctoral advisor Prof. Dr. Fritz K. Winkler, who

gave me the opportunity to carry out my thesis at the Paul Scherrer Institut. He was

willing to discuss and offered help whenever I needed it.

Special thanks go to my supervisor Dr. Louis X. Tiefenauer who introduced me in the

material science aspects of the project and provided an insight into product development

processes. I appreciated our regularly occurring discussions and the encouraging words.

I am likewise thankful to Prof. Dr. Kaspar Locher and Prof. Dr. Janos Vörös who

readily agreed to examine my work; for their continuous interest in the study, the

discussions and contributions.

At PSI, I very much enjoyed the working atmosphere and the well equipped

laboratories. Many thanks go to all current and former members of the structural

biology group, Andrea Prota, Anke Weisbrich, Christian Kambach, Daniel Frey,

Dominik Frei, Guido Capitani, Hanna Sundström, Jean-Jacques Hefti, John Missimer,

Martin Schärer, Michel Steinmetz, Monica Balsera, Monika Brunner, Natacha Olieric,

Peter Hasler, Rolf Jaussi, Rubén Buey, Saša Bjelic, Sophie Demarche and Xiao-dan Li.

Furthermore I would like to thank Antonietta Gasperina and Domenico Lupo for the

support in the protein expression and purification and, particularly Arnaud Javelle,

Christian de Groot, David Langenegger, Felix Grünewald and Remo Studer; it was

good to have them around.

Deepest thanks go to my parents for their support and their encouragement during my

whole studies and to Anita for her love and patience.

1. Introduction

1

1. Introduction

1.1 The structure of the cell membrane

Cell membranes operate as barriers to compartmentalize living cells. The plasma

membrane encloses the cell and separates the cytoplasm from the outside world; inside

the cell, membrane-enclosed organelles enable complex concerted biochemical

reactions by maintaining specific conditions with respect to composition and

concentration. This allows for example the controlled degradation of food particles or

waste molecules in lysosomes at a low pH, the generation of the chemical high-energy

ATP compound in mitochondria or the processing of proteins in the Golgi apparatus.

Despite their different function, cell membranes have a common structure: amphiphilic

lipids form a thin layer that includes protein molecules and represents a barrier to ion

and solute transport. From measurements on the surface area that lipids occupy at the

air-water interface, the double layer nature of cell membranes was discovered already in

1925 [1]. But it took several additional years until a model was developed that described

correctly the assembly of the proteins. In the “fluid mosaic model” developed by Singer

and Nicholson in 1972 the lipid bilayer forms a matrix and proteins are either adsorbed

to the surface or may span the membrane [2]. The lipid molecules in the bilayer are

mobile producing a dynamic, fluid membrane that is impermeable to most water-soluble

compounds. The associated membrane proteins are responsible for many essential and

specific functions such as solute transport, signalling and ATP synthesis. The numerous

tasks require many different membrane proteins; it is estimated that one third of the

proteins encoded in a cell genome are membrane proteins [3]. Thus the initial model

that proposed membrane proteins to be dispersed and low concentrated needed to be

updated. Also the idea that the hydrophobic core of a membrane spanning protein

matches the bilayer thickness was not correct [4]. The “fluid mosaic model” has been

further modified and has become more mosaic than fluid; specific proteins and lipids

can form domains and clusters and the elastic properties of the bilayer allow a variable

bilayer thickness adjacent to the membrane proteins that possess shorter or longer

hydrophobic regions compared to the undisturbed membrane (Fig. 1.1).

1. Introduction

2

1.2 The lipid bilayer

A membrane patch of 1 µm2 area is assembled by approximately 3 million lipid

molecules. Lipids are of amphiphilic nature; that is they have a hydrophilic water loving

headgroup and a hydrophobic water repellent tail. For this reason they spontaneously

self assemble in an aqueous environment to form the double layered structure, with the

headgroups toward the aqueous phase and the tails buried in the interior. The most

abundant lipids in mammalian membranes are the phospholipids: they contain a

phosphate group that connects the polar head group either to a sphingosine

(sphingomyelin) bonded to a fatty acid, or via a glycerol backbone

(glycerophospholipids) to two fatty acid chains (Fig. 1.2).

Fig. 1.1 | Model of the cell membrane structure

The drawing shows a three-dimensional view of a cell membrane.

1. Introduction

3

The nonpolar tails differ in length and are usually in the range between 14 and 24

carbon atoms. Most common are C16 (palmitic), C18 (stearic) and C20 (arachidic). The

degree of unsaturation varies highly, but the most common species are 18:1

(9-cis = oleic), 18:2 (9-cis, 12-cis = linoleic), 18:3 (9, 12, 15-(all)cis = α-linolenic) and

20:4 (5, 8, 11, 14-(all)cis = arachidonic). Depending on the saturation grade of the fatty

acid chains and the nature of the polar headgroup, the packing of the molecules in a

membrane varies. The optimal cross sectional area that is required by these two regions

allows classifying the lipids into cones, cylinders and inverted cones. Cylindrical

phospholipids favour the formation of planar bilayers whereas the wedge-shaped

lysophospholipids with only one tail preferentially form micellar structures. Thus these

different structural properties help to stabilize regions of low or high curvature in a

membrane or to pack adjacent to membrane proteins. Individual phospholipid molecules

diffuse in the lateral plane of a bilayer leaflet with a diffusion coefficient of about

10-8 cm2 s-1 [5] and they rotate about their long axis. This two-dimensional fluidity that

depends on the temperature is precisely regulated in a cell by its lipid composition.

Shorter hydrocarbon chains and cis-double bonds result in weaker interactions and

lower packing densities such that the membrane remains fluid at lower temperatures.

Spontaneous movements of lipids from one bilayer leaflet to the other (“flip-flop”)

occur very rarely. However, rapid transbilayer movement of phospholipids can be

mediated by specific exchange proteins and translocators that actively maintain an

asymmetric lipid composition in a cell [6] which is important for many functions. One

example is the recognition of dead cells by macrophages in animals. The negatively

charged phosphatidylserine is normally located in the cytosolic monolayer of the plasma

membrane. Upon apoptosis this lipid is rapidly translocated and exposed on the cell

Fig. 1.2 | Phospholipid chemical structure and lipid composition of different membranes

The parts of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and the approximate percentage

of total lipid per weight of different membranes are displayed. Modified from Alberts et al. 2002 [5].

1. Introduction

4

surface where it is recognized directly by the macrophage phosphatidylserine receptor

[7]. The recognition and phagocytosis of apoptotic cells protects the organism from the

exposure to intracellular compounds leading to inflammation.

1.3 Proteoliposomes as model systems

In order to study the function of channel and transport proteins, the formation of

compartments is essential. Phospholipid vesicles (liposomes) with incorporated,

purified membrane proteins (proteoliposomes) can serve as excellent model systems

(Fig. 1.3 A). Liposomes spontaneously form from diverse phospholipids and lipid

mixtures that are dispersed in an aqueous buffer, but these vesicles are multilamellar

and of different size. Several procedures yield the formation of proteoliposomes that are

similar in size and contain a single type of membrane protein. Their ability to efficiently

separate the encapsulated solutes from the bulk solution allows monitoring protein

mediated substrate flow. In such assays the substrate changes in the vesicle lumen need

to be assessed which is achieved by applying different techniques. The conventional

strategy is to initiate uptake by adding the (radioisotopically-labeled) substrate and

measure the transported amount after separating the proteoliposomes from the medium.

During the separation and washing steps, adherence to the matrix, breakage or leakage

might occur and produce an enhanced variability. This was circumvented in the case of

a reconstituted calcium pump using phosphate inside the liposomes to precipitate the

transported Ca2+ [8]. Alternatively, the fluorescent change of indicators, confined to the

lumen and sensitive to substrate concentration can be directly measured (Fig. 1.3 B).

For example a Zn2+ metal transporter activity was measured using proteoliposomes

preloaded with the dye fluozin-1 [9]. In another study proton conductance was

Fig. 1.3 | Illustration of a (proteo-) liposome

A) Cross section of a liposome B) Substrate transport results in a fluorescence change of

the indicator.

1. Introduction

5

measured using a pH sensitive dye and fluorescence quenching has been used to assay

the water permeability of liposomes [10]. Stopped-flow analysis of transmembrane

fluxes allows quantifying the transport kinetics. If the solute transport is coupled to ATP

hydrolysis such as in ABC transporters, a reduced ATPase activity is related to a lower

transport rate [11]. However ATP turnover is not in an obligatory manner coupled to ion

transport; the so called P-type pumps can operate at variable stoichiometry due to

uncoupling and slippage [12]. Thus to quantitatively understand transport mechanisms

the translocated species need to be determined.

Using proteoliposomes to investigate the function of membrane proteins imposes

several difficulties and limitations. (1) The amount of functional protein per liposome

and their relative orientation are difficult to determine. (2) Homogeneous size

populations are hardly obtained and variations in the diameter result in different

concentration dependent responses of the indicator. (3) The high surface area to volume

ratio may require a rapid detection system such as a stopped-flow apparatus. (4) Finally,

the internal volume is not directly accessible.

If the proteoliposome is large enough (e.g. giant liposome), the translocation of charged

solutes can be assayed by the patch-clamp technique [13]. After the immobilization of

giant liposomes, a patch of the liposome surface is excised by gentle suction using a

pipette and a micromanipulator (Fig. 1.4). A tight seal between the micropipette and the

membrane patch is necessary to assure that the flow of charged particles (current)

occurs only through open channels. Transmembrane current can thus be measured using

this “inside-out” configuration under different holding potentials [14].

Fig. 1.4 | Patch-clamp recording technique

1. Introduction

6

In this setting, it is possible to alter the solution on either side of the membrane to test

the channel behaviour on various solutes. This method permits to analyze the gating

properties and ion selectivity of channels at the level of single molecules. Expression,

solubilisation and purification of a membrane protein selected for such studies represent

critical steps toward proteoliposome formation.

1.4 Heterologous expression systems for electrophysiology

It appears much simpler to perform patch-clamp measurements directly on the native

membrane or on the membrane of cells that express the protein of interest. In order to

successfully characterize a specific membrane protein, several conditions need to be

satisfied. The application of patch-clamp methods requires (1) the direct accessibility of

the membrane; (2) the ability to form a tight seal between the membrane and the pipette;

a host cell that is (3) able to express the recombinant protein in the cell membrane and

(4) which does not express endogenous channels that can influence the measurement.

As a result, several heterologous expression systems are used for the functional

characterization of ion channels. The ideal expression system of membrane proteins to

perform patch clamp experiments is not available. Depending on the type of channel to

be analyzed, possible interference by endogenous ion channels and the feasibility of its

functional production including post translational modifications set the limits in

choosing the cell line. Advantages and disadvantages of some commonly used systems

will be discussed in the following paragraph, with the objective to show which of the

aforementioned critical requirements are not completely fulfilled and why they

complicate patch clamp measurements.

The facile genetic manipulation and the relatively large size of Saccharomyces

cerevisiae cells have made it a promising candidate for patch clamp measurements.

Since the cell wall impedes direct electrophysiology measurements, it is inevitable to

first produce protoplasts by digesting the cell wall using mycolytic enzymes [15].

Difficulties in the formation of tight seals and numerous endogenous ion channels and

pumps reduce the achievements made using this model system [16]. Human embryonic

kidney cells (HEK293) also serve as an efficient expression system in the study of

channels from many sources. Stable transfection allows the generation of a large

number of cells that express the protein at a level that is high enough to perform

electrophysiology measurements. However, the constitutive expression of endogenous

receptors and channels may influence the measurement. The presence of various ion

1. Introduction

7

channels in HEK293 cells were reported but it has been assumed that the endogenous

channels produce a negligible background current. For example contradictory results

concerning endogenous calcium channels were reported; Jiang et al. identified several

channel types [17], while Avila et al. found no contribution of calcium channels to the

signal [18]. Other native ionic currents were reported for this system such as sodium,

chloride and potassium currents that emerge at different times during cell growth [19].

Thus whole cell channel currents of an expressed exogenous protein might be the sum

of different channel types, and endogenous conductances need to be suppressed using

customized experimental conditions. Consequently the understanding of the endogenous

channels of a system is a prerequisite for functional characterization of heterologously

expressed channels. Another popular cell line for patch clamp measurements are the

epithelial-like Chinese hamster ovary (CHO) cells. Non-transfected CHO cells produce

low currents suggesting the expression of low quantities of endogenous ion channels;

these cells are readily transfected with cDNA encoding the protein of interest and the

large cells are very easy to patch [20]. Different reports about the channel occupancy

suggest variability in their expression that depends on the cell cycle or the cell line.

Disagreements in the existence of chloride channels, voltage-sensitive sodium channels

and calcium channels [20,21] make large control experiments necessary.

1.5 Synthetic planar lipid bilayers

Working with artificial lipid bilayers and purified membrane proteins provide the

benefit to define the membrane composition and allows free access to both sides. Thus

they are attractive to investigate the physical properties, the structure and function of

biological membranes. The first experiment with a bilayer was performed in 1665 by

the physicist Robert Hooke who studied soap films, as outlined by Ti Tien [22]. Hooke

observed “black holes” that spontaneously formed in soap films or bubbles, which were

sharply separated from the rainbow-coloured part. The thickness of those “black holes”

was estimated by Isaac Newton a few years later to be 9.5 nm. These black holes

resemble lipid membranes with respect to composition and thickness, with the

important difference, that the two soap monolayers coating a water layer have their

hydrophobic tails oriented towards the air. In 1962, the formation of bilayer membranes

spanning the aperture between two compartments filled with a saline solution was

reported. They were generated from lipids which were extracted from white matter and

formed by methods analogous to those employed for the generation of soap films.

1. Introduction

8

Regions of the film that scarcely reflected light exhibited mechanical and electrical

properties similar to those attributed to the cell membrane [23]. Such black spots which

were located next to rainbow-coloured interference patterns correspond to regions

featuring a thickness below the wavelength of visible light and destructive interference

makes them appear black: The first artificial so-called “black lipid membrane” or

“bilayer lipid membrane” (BLM) was formed (Fig. 1.5 A). In the following years such

artificial biomembranes formed in an aperture from lipids dissolved in a hydrocarbon

solvent (for detailed information see chapter 2.3.1) were used as model systems to

explore cell membranes. BLMs were used to characterize membrane-active peptides

such as gramicidin [24] and alamethicin [25], to study the ion selectivity of protein

pores [26] and gated channels [27], and to investigate receptor-mediated signal

transduction [28,29] and the effect of changes in lipid bilayer properties on the function

of membrane proteins [30]. But BLMs are extremely fragile and generally have a life

span of only several hours [31]. Their mechanical stability was improved by the use of

several additives; for example cholesterol within the bilayer decreases the mobility of

the first few CH2 groups of the hydrocarbon chains of the lipid molecules and makes the

lipid bilayer less deformable upon thermal or mechanical stress [32]. Artificial cell

walls formed from polysaccharide derivatives that bear hydrophobic anchor groups and

cover the membrane [33] helped to increase its life span and hydrophobic polymers

were investigated for their stabilization effect [34]. Other strategies involved the use of

polymerizable lipids [35] or highly branched hydrophobic chains [36], but the influence

of these modifications on the lipid mobility and bilayer thickness limits their usefulness.

Jeon et al. encapsulated the bilayer in a hydrogel by inducing PEG-DMA

polymerization in the buffer after bilayer formation [37]. The resulting branched

polymer increased the stability of the bilayer by a factor of four, but the mesh size of

7 nm limits the diffusion of particles to the membrane to small molecules.

Increased stability resulted from the preparation of membranes on solid supports

(Fig 1.5 B). The spreading of liposomes on hydrophilic materials is one method that

results in such supported bilayer lipid membranes (sBLM) that were successfully

formed on glass, mica, silicon and various metals [38]. Their robustness and the

proximity to the supporting material allows the use of surface-sensitive techniques for

investigations such as atomic force microscopy (AFM), quartz crystal microbalance

(QCM) or surface plasmon resonance (SPR). Thus they are well suited to analyze lipid

domain formation [39], protein adsorption and self-assembly [40,41] or intermembrane

1. Introduction

9

interactions such as membrane fusion events [42]. Since one monolayer is in direct

contact with the support, strong interactions reduce the mobility and only the upper

leaflet is free to move. On the other hand, a weak interaction results in a fluid membrane

that is susceptible to damage. Another disadvantage is its limited use for studying

membrane proteins. The minimal space between the lipid bilayer and the solid support

impedes the functional incorporation of transmembrane proteins with hydrophilic

domains on both sides. Also electrochemical measurements are restricted; the small

reservoir limits the number of ions that can be transported and therefore the

determination of transport or diffusion rates [38].

The risk of protein denaturation caused by the contact with the support can be avoided

by using polymeric materials residing on the solid substrate to separate the membrane

(Fig. 1.5 C). In these polymer-cushioned bilayer lipid membranes (pBLM), the

polymeric support, being typically less than 100 nm thick, needs to be carefully adjusted

to guarantee complete wetting of the surface. Additionally, the interactive forces

between the lipid membrane and the support need to be weak to avoid a de-wetting and

the formation of pinning centres [43]. Polymer cushions were formed by covalently

coupling dextran to glass [44], PEG layers on alumina [45] or glycoacrylate with a

covalently linked lipid analogue facing the aqueous space to assist in bilayer formation

[46].

A related strategy to lift the lipid bilayer off the solid support is to use hydrophilic

spacers containing a hydrophobic end group that integrates into the membrane

(Fig. 1.5 D). The separating distance of these tethered bilayer lipid membranes (tBLM)

can be controlled via the bifunctional spacer unit. In addition to phospholipids that are

coupled to the surface via their head groups [47,48], also other lipophilic molecules

were used to anchor the bilayer such as cholesterol [49], a single phytanoic acid [50]

and even whole transmembrane proteins [51]. The key problem with tBLM is to achieve

a high surface coverage without any defects that is needed to obtain a high electrical

resistance. The quality of the membrane depends on the concentration of tethers; a low

density guarantees mobility whereas a high density is needed for the formation of stable

lipid membranes [52]. Thus the right balance needs to be found between tether

concentration, spacer length and hydrophilic end group.

1. Introduction

10

1.6 Aim of the thesis

Progress in the production and purification of complex membrane proteins as well as in

protein crystallography resulted in the availability of increasing numbers of high

resolution structures [54]. To understand the mechanistic details of how such proteins

work, quantitative functional data are equally important. In addition, the screening for

compounds interacting with potential drug targets located in cell membranes makes

assay systems for measuring the function of reconstituted membrane proteins highly

demandable. No simple method is available but in general the protein of interest is first

reconstituted either in proteoliposomes or in an artificial lipid bilayer on a support and

is then analyzed using diverse techniques. The major advantage in using free standing

planar lipid bilayers is that both, the “inner” and “outer” compartments are accessible

and controllable. The major drawback is their low mechanical stability. Using solid

materials as supports for lipid membranes, their stability is enhanced. However, the

resulting close proximity of the bilayers to the supporting surface has severe limitations.

Incorporation of large transmembrane spanning proteins is nearly impossible and

quantitative measurements in the very small cleft between the bilayer and surface are

difficult.

Fig. 1.5 | Planar lipid bilayer model systems

A) Bilayer lipid membrane (BLM) formed in an aperture of the support. The two monolayers on either

side of the support forming a hybrid bilayer with the covalently bound silanes meet in the pore to form a

fluid lipid bilayer. B) Solid supported bilayer lipid membrane (sBLM). C) Polymer-cushioned bilayer

lipid membrane (pBLM). D) Tethered bilayer lipid membranes (tBLM).

Adapted from Richter et al. 2006 [53]

1. Introduction

11

In our group a thin silicon nitride membrane with arrays of nanopores has been

developed that serves as support for BLMs. These arrays of nanopores provide a large

total bilayer area, with an increased circumference to bilayer area ratio compared to a

single pore of similar bilayer area, which it thought to improve the stability. A large

total pore area was designed to allow the incorporation of a large number of protein

molecules needed to measure signals from the many membrane proteins which have

lower transport or turnover rates than ion channels. The fabrication procedure for the

chips is well defined and highly reproducible allowing a commercial production;

furthermore the device is sufficiently robust to become integrated in automated

measurement system.

With the objective to develop a generally usable septum for functional assays of

membrane proteins using such nanostructured supports, the goal of this thesis was to

investigate its basic characteristics. First the spontaneous formation of lipid bilayers

within the nanopores and the expected improved stability needed to be assessed. Next,

the possibility of the incorporation of suitable protein quantities into the lipid bilayer

needed to be explored and assay methods for measuring solute transit needed to be

investigated. Different methods to form protein containing lipid bilayers were foreseen

to be evaluated and the feasibility of measuring the activity of a complex membrane

protein using this device should be demonstrated.

2. Characterization of free-standing lipid bilayers in nanopores

12

2. Characterization of free-standing lipid bilayers in

nanopores

The development of activity assays to investigate membrane proteins requires that

several conditions are met. (1) A robust and well defined support should allow the easy

formation of lipid bilayers. (2) These bilayers need to be free-standing with access on

both sides, to allow easy adjustment of the buffer composition and to perform diffusion

measurements. (3) The formation of stable bilayers using (4) different lipids or mixtures

permits long term measurements of proteins that are reconstituted in bilayers of a

composition comparable to their native environment. (5) A large bilayer area offers the

advantage to incorporate a high number of protein molecules if the signal needs to be

enhanced. The study of ion channels is already possible at the level of a single molecule

using electrochemical methods. But to investigate ion pumps and transporters that move

ions at much lower rates across the membrane, thousands of active protein molecules

are required to obtain a measurable current in the pA-range [38]. A large bilayer area is

also a prerequisite for measuring the translocation of uncharged substances which are

electrochemically not detectable. We have designed a silicon-based membrane support

for the formation of such free-standing lipid bilayers, because the fabrication procedure

[55] is highly reproducible and the robust surface can easily be modified using

chemically activated hydrophobic silanes [56]. We attempt to enhance the stability of

the bilayer by increasing the contact area to the support. Therefore we split a large area

into arrays of regularly arranged pores. In the first phase of the project, the formation of

lipid bilayers using different preparation techniques and lipids was evaluated and

criteria for successful bilayer formation were established. The effect of the pore size,

preparation technique and lipid composition on the bilayer stability were investigated.

Furthermore the influence of a breakdown of a bilayer in a single nanopore on the

measured resistivity was assessed.

2.1 Measurement setup

Several measurement cells fulfilling different requirements were designed. For example

to observe bilayer formation with a microscope, the chip was assembled horizontally in

a cell with a glass bottom. In all other setups the chip was sandwiched vertically

between a set of two identical half cells allowing easy access and stirring of both


13

compartments. They are manufactured form the transparent plastic material polymethyl

methacrylate (PMMA) to provide a direct view on the chip surface, a necessity to check

for air bubbles and for a convenient bilayer formation. The operation volume is

confined by the cylindrical reservoir that needs to be filled to above the perpendicular

hole making the connection to the counterpart. The different cells cover the applicable

volume range from 50 µL to 10 mL. Silicone fittings (indicated by an arrow) between

the two half cells tightly seal the chip (Fig. 2.1).

2.2 Chip with nanopores

2.2.1 Chip production

Chips with different layouts with respect to the nanopore arrays and pore size were

fabricated [55] by our project partner (Axetris Microsystem, Sarnen, Switzerland).

These are 6 x 6 mm2 wide and have a squared silicon nitride membrane of

0.5 x 0.5 mm2 in the centre which is 300 nm thick (Fig. 2.2). Starting with a 380 µm

thick silicon wafer (1,0,0) with 300 nm silicon nitride layers on both sides, a thin resist

was spun on the front side and structured either by photolithography for pores with

diameter 800 nm and bigger or by electron beam lithography for the smaller pores. By

standard reactive ion etching the pores were transferred through the silicon nitride layer

to the silicon surface. Subsequently the backside of the waver was structured similarly

using photolithography followed by wet etching in 40 % KOH at 70°C up to the silicon

nitride layer on the front side. The predefined dimensions of the opening and the waver

thickness result in the release of a 0.25 mm2 silicon nitride membrane containing the

pores. The surface of the chips was hydrophobically silanized [57] using a 1 to 1

mixture by volume of (tridecafluoro-1,1,2,2-tetrahydro-octyl) trichlorosilane and

(tridecafluoro-1,1,2,2-tetrahydro-octyl) dimethyl-chlorosilane in the gas phase.

Advancing contact angles of distilled water were in the range of 110° to 115°.

Fig. 2.1 | Disassembled measurement cell


14

2.2.2 Chip types

During the advancement of the project different chip layouts with respect to the number

of pores, pore diameter, pitch and total pore area were required. To investigate whether

the lipid bilayer stability depends on the chip pore size, to assess the resistance of a

single open chip pore or to evaluate the effect of the pore spacing on the diffusion

properties, various series of chips were manufactured. The technical specifications of

the different chips are summarized in Table 2.1; due to grouping into the categories

some chips appear more than once.

Fig. 2.2 | Chip with nanopores.

A) Chip layout, drawn not to scale B) Light microscope image of a chip with 400

nm pores. The squared silicon nitride membrane containing the regularly

arranged dark pores appears bright. SEM images of a chip with 200 nm pores

from the surface (C) and of a cross section (D)


15

Chip identifier Pore diameter (µm)

Number of pores per chip

Pitch1 (µm)

Total pore area (µm2)

Array of pores: PA800 0.8 32’000 1.6 16085 PA400 0.4 240’000 0.8 30159 PA200 0.2 960’000 0.4 30159 Single pore SP0.8 0.8 1 -- 0.503 SP1.6 1.6 1 -- 2.011 SP3.2 3.2 1 -- 8.043 SP6.4 6.4 1 -- 32.17 SP12.8 12.8 1 -- 128.7 SP25.6 25.6 1 -- 514.7 Constant total pore area CA0.8 0.8 1024 1.6 514.7 CA1.6 1.6 256 3.2 514.7 CA3.2 3.2 64 6.4 514.7 CA6.4 6.4 16 12.8 514.7 CA12.8 12.8 4 40 514.7 SP25.6 25.6 1 -- 514.7 Four pores varying pitch and diameter 4P0.8/40 0.8 4 40 2.011 4P0.8/4.8 0.8 4 4.8 2.011 4P0.8/2.8 0.8 4 2.8 2.011 4P1.6/40 1.6 4 40 8.043 4P1.6/5.6 1.6 4 5.6 8.043 4P1.6/3.6 1.6 4 3.6 8.043 4P3.2/40 3.2 4 40 32.17 4P6.4/40 6.4 4 40 128.7 CA12.8 12.8 4 40 514.7

2.3 Lipid bilayer formation

2.3.1 Painting

In the majority of the reported studies on BLMs, they have been formed by the painting

technique [23], a rather easy and successful method. The lipid is dissolved in a non-

volatile hydrocarbon, usually n-decane but bilayer formation from n-hexane [58] or

n-tetradecane [59] solutions has also been reported. The chip is immersed in an aqueous

solution and a small quantity of the lipid mixture is spread on one side over the aperture

using a (single-bristle) brush or a modified plastic pipette. Excess lipid solution is 1 The pitch is the distance from the center of a pore to the center of the adjacent pore

Table 2.1 | Chip type description


16

thought to spread away on the wettable support [60] and subsequent thinning of the

hydrocarbon-phospholipid plug within the aperture results in the spontaneous formation

of a lipid bilayer (Fig. 2.3). It is assumed that hydrocarbon tails of adjacent lipid

molecules are pulled together by van der Waals forces and that the organic solvent is

being pushed towards the pore border; the bilayer grows in a zipper-like manner [61]

until it spans the majority of the aperture. This growing process can be observed with a

microscope if viewed under reflected light. Single or several black spots emerge within

the lipid film that continuously grow and merge to cover large parts of the aperture [62].

The residual hydrocarbon solution forms a relatively thick annulus (Fig. 2.3 B), also

called the Plateau-Gibbs border [63], which connects the lipid bilayer and the support.

This process of thinning is crucial and can only take place if the hydrocarbon solution

drains. Otherwise the amount of painting solution needs to be exactly adjusted to the

aperture size [64]. To avoid this difficulty the bilayer can also be forced to thin by

applying a hydrostatic pressure [65], or by reshaping it with a brush or an air bubble

[66]. These painted bilayers still may contain residual solvent and even mobile solvent

lenses were observed, regions where the bilayer is substantially thickened [67]. The size

of the annulus depends in part upon the applied volume of the lipid solution and the

diameter of the aperture, the contact angles α and β are related to the interfacial tensions

[68] and thus are influenced by the support material and the organic solvent.

Fig. 2.3 | Illustration of a painted lipid bilayer.

The cross-sectional view of a single pore: A) Silanized chip membrane

B) Residual organic solvent forming the annulus C) Lipid bilayer

Adapted from White, S.H. et al. 1976 [69]


17

2.3.2 Müller – Montal

The Müller – Montal (M-M) method [70] for the formation of a lipid bilayer is

technically more challenging but offers the advantage that asymmetric bilayers can be

formed and that the residual hydrocarbon solvent can be minimized. Starting with the

buffer level below the aperture on both sides of the support (Fig. 2.4), the lipid

dissolved in a volatile solvent such as chloroform or pentane is added to the buffer in

both compartments. This results in the formation of lipid monolayers which spread at

the air-water interfaces. The bilayer is formed by sequentially raising the buffer levels

above the aperture. Prior to monolayer folding the chip needs to be pretreated with a

hydrocarbon; the presence of a nonpolar solvent which can form the annulus is a

necessity [69]. Using selected solvents such as squalene [71] or other long-chain

hydrocarbons allows the formation of nearly solvent-free bilayers since these solvents

tend to be excluded from the bilayer structure [72].

2.3.3 Liposome on chip adsorption

Inducing (giant) unilamellar vesicles to spread on a hydrophilic surface seems to be an

attractive and simple route to form a lipid bilayer. The use of an organic solvent can be

avoided and this one-step method would be suited for automation. The formation of

bilayers from vesicles involves several critical steps: (I) vesicle adsorption to the solid

Fig. 2.4 | Sequence of illustrations showing M-M bilayer formation.

The cross-sectional view of a single pore: A) Monolayers formed at the air-

water interface contain a nonpolar solvent that wets the chip surface B) After

raising the right-hand water level C) Asymmetric lipid bilayer (see differently

coloured lipid headgroups)

Adapted from White, S.H. et al. 1976 [69]


18

support, (II) rupture and (III) spreading into planar bilayers that stretch over a pore [53].

Consequently it is very demanding to achieve a high surface coverage, a prerequisite to

obtain a high membrane resistance needed for electrical measurements.

We pursued several strategies to induce the rupture of adsorbed liposomes such as

varying the lipid composition [73], applying an osmotic shock or adding specific ions

which facilitate the process [74,75]. Bilayer formation was monitored using

electrochemical impedance spectroscopy (EIS). This method allows us to measure the

amount of pore-spanning bilayers to a certain extent, but it is not possible to determine

the number of adsorbed vesicles and respectively the portion which does not break

during an experiment. An additional difficulty was the lack of chips with only few pores

until recently. Having only PA800, PA400 and PA200 with several thousand pores at

our disposal initially we were not able to achieve a high level of sealing. As total

resistance is only sensitive to single pore openings at nearly complete sealing, this

method is not useful to monitor improvements at moderate levels of sealing. To fully

develop this method to become applicable on chips with arrays of pores, a major effort

will be necessary. To my knowledge the highest sealing that was achieved using the

liposome adsorption method on a porous support was around 100 mega-Ohms; after

extensively functionalizing the support [76,77].

2.4 Lipid bilayer characterization

2.4.1 Electrochemical impedance spectroscopy (EIS)

Formation of a free standing lipid bilayer can be verified by EIS. It involves the

application of a sine wave perturbation of different frequencies and measuring the

amplitude and phase shift of the response wave. The measurement setup is modelled as

an electrical circuit consisting of resistances (R) and capacitances (C) which describe

the individual elements of the setup. Thus the buffer, the electrodes and the chip with

and without a bilayer need to be described by electrical units (Fig. 2.5).


19

The buffer solution is represented by a single resistance RS which reflects its ionic

conductivity. The electrodes are combined to a capacitance CE and resistance RE in

parallel [78,79] that characterize the capacitive behaviour of the electrode-solution

interface and its resistance to charge transfer. The equivalent circuit for the chip is a

capacitance, Ctotal with and CSNM without a lipid bilayer, in parallel to a resistance RM.

Subsequent mathematical analysis of the frequency dependent impedance using the

described model allows determining the values of the individual circuit components.

Since the capacitive behaviour of the describing elements is not ideal, fitting of the

phase angle is not perfect. The values of interest are the changes of the chip capacitance

and resistance that occur after lipid bilayer formation. RM is low if many pores are open

and ions can pass the chip unhindered but it reaches giga-ohms if all pores are blocked.

A high resistance may thus originate from a lipid bilayer but also from multilayers or a

lipid plug. However, the capacitance value allows drawing conclusions about the

thickness of the barrier and is thus a measure for lipid bilayer formation (eq. 2.1).

d

AC εε0= (eq. 2.1)

The membrane capacitance depends directly on the bilayer area A and the hydrocarbon

thickness d, the two constants are the permittivity of free space (ε0 = 8.85·10-12 F m-1)

and the dielectric constant of the lipid membrane (ε). Prior to bilayer formation the

determined capacitance (CSNM) originates from the thin silicon nitride layers of the chip.

After the formation of a lipid bilayer, Ctotal is measured which is the sum of CSNM and

the effective membrane capacitance (CBLM): Ctotal = CSNM + CBLM since the support and

the bilayer are arranged in parallel. To determine CBLM it is necessary to measure CSNM

Fig. 2.5 | Electrical equivalent circuit model describing the measurement setup


20

in advance, due to variations between individual chips this needs to be done for each

chip. The specific capacitance i.e. the bilayer capacitance per area is the standardized

value directly related to the thickness.

The impedance spectra of the electrodes in 0.1 M KCl (I) a bare PA800 chip (II) and the

chip with a painted POPC bilayer (III) are shown in figure 2.6 (A). The impact of the

electrodes on the impedance is only visible in the low frequency range of the bare chip.

After painting, RE and CE can be neglected and an easier equivalent circuit can be used

to fit the data. This is shown in figure 2.6 (B) where a four electrode setup was used to

compensate the potential drop on the electrodes [80].

Consequently the bare chip PA800 (IV) can be fitted using the same equivalent circuit

as for the chip with a painted bilayer (V) and the same results for the bilayer resistance

and capacitance are obtained independently from the measurement setup (Table 2.2).

Fig. 2.6 | Electrochemical impedance spectroscopy

The measured data points and the fitted curves using the indicated

equivalent circuits are shown for (I) the electrodes, (II),(IV) the bare

PA800 chip and (III),(V) a painted POPC bilayer using a two-electrode

(A) and a four-electrode (B) setup. Elements of the equivalent circuit

arranged in parallel are put in brackets.


21

Bare chip With BLM Measurement setup RM (kΩ) CSNM (pF) RM (GΩ) Ctotal (pF)

2 electrode 580 ± 70 708 ± 212 10 ± 2.4 1056 ± 47 4 electrode 385 ± 8 644 ± 84 7.5 ± 2.5 1059 ± 63

2.4.2 Lipid bilayer capacitance

It was shown that the specific capacitance of lipid bilayers decreases linearly with

increasing number of carbon atoms in the mono-unsaturated fatty acid of the lipid [72].

Assuming that the dielectric constant of the membrane equals that of a long chain

hydrocarbon (ε = 2.1, [81]) and that the membrane thickness is in the range of a typical

biological membrane (d = 3.5 – 4.2 nm, [82]), leads to specific membrane capacitance

values ranging from 0.53 µF cm-2 to 0.44 µF cm-2. However, most reported values using

artificial lipid bilayers are in the range of 0.5 – 0.9 µF cm-2. This could lead to the

assumption that their thickness is lower compared to biological membranes, probably

due to a different lipid packing in the absence of membrane proteins. A second

explanation could be a higher dielectric constant of artificial bilayers due to enclosed

water molecules. Hydrocarbon-containing bilayers formed by painting show a specific

capacitance of 0.45 µF cm-2 and bilayers formed by the M-M method 0.9 µF cm-2 [70].

The observed difference is not only believed to originate from differences in bilayer

thickness but also from water molecules that may penetrate deeper into the hydrocarbon

region of bilayers without intercalated hydrocarbon solvents. Geometrical factors such

as curvature which leads to a lower lipid packing density of one monolayer compared to

the other could also explain the higher specific capacitance values observed for artificial

bilayers.

The determination of the membrane capacitance is challenging since the bare chip

possesses already a capacitance ranging from 0.6 - 1 nF. Using the upper limit of

0.9 µF cm-2 for the specific capacitance to calculate the expected bilayer capacitance of

chip PA200 offering the largest pore area, a maximum value of 270 pF is predicted. The

high value of CSNM makes a precise capacitance determination very difficult and only

for chips with a large pore area (high porosity) where CBLM contributes around one

fourth to the measured total capacitance, the determined specific capacitance value

approaches the expected 0.9 µF cm-2 (Table 2.3).

Table 2.2 | R and C values determined by EIS measurements.


22

Chip identifier Porosity [%]

Lipid Preparation method

Spec. capac. [µF cm-2]

PA800 6.4 DPhPC MM 2.2 ± 0.8 POPC painting 2.3 ± 0.6 PA200 12 POPC MM 1.6 ± 1.2 POPC painting 1.5 ± 0.1

No difference in the capacitance of bilayers formed from DPhPC or POPC could be

detected and the specific membrane capacitance of the nearly solvent free bilayers

formed by the Müller – Montal method was comparable to that of bilayers formed by

painting. Most likely the high shunt capacitance of the SNM prevents the detection of

small differences in the bilayer capacitance. The estimation of the overall bilayer

formation success rate on chips with arrays of pores is also affected by this factor. The

contribution of a bilayer in a single 800 nm pore to the total capacitance is

approximately 4.5 fF and it is impossible to detect a single failed thinning of a

hydrocarbon-phospholipid plug. Since no big variations in the total capacitance were

observed by repeated bilayer formation using the same chip, it can be stated that

reproducibly the same amount of bilayers was retained. The high capacitance value

suggests a large number of thinned bilayers but accurate numbers cannot be determined

due to the measurement errors induced by the large total capacitance.

2.4.3 Lipid bilayer resistance

Fitting analysis of the impedance measurements after bilayer formation on our chips

yielded membrane resistance values above 1 GΩ for all lipids tested. Such a high

sealing is needed for the analysis of membrane proteins with electrophysiological

methods and is generally achieved applying the painting or Müller – Montal method.

The breakdown of a lipid bilayer has an enormous impact on the total bilayer resistance

and a single open pore can unquestionably be recognised. At applied potentials between

± 100 mV the current – voltage dependence is linear and the ionic resistance of the

electrolyte filled nanopore is related to the membrane thickness (lp) and pore diameter

(dp) via [83]

2

)8.0(4

p

pp

d

dlR

πκ

+= (eq. 2.2)

Table 2.3 | Capacitance values determined for lipid bilayers on nanopore chips.

Lipid bilayers were formed from DPhPC or POPC by painting or the Müller – Montal

method (MM). The porosity is calculated from the total pore area / SNM area, n = 3.


23

where κ is the conductivity of the electrolyte. The equation is derived from the

resistance of a cylindrical pore with lp >> dp, taking the effects at the pore entrance and

exit into account since the pore length is comparable to the diameter. Using a

conductivity standard solution with κ = 12.15 mS cm-1 results in a calculated resistance

of a single open 800 nm pore of about 1.5 MΩ. The measured value for chip SP0.8

(RM = 2.2 MΩ) is slightly higher (Fig. 2.7).

2.5 Lipid bilayer stability

Initially it appeared plausible to us that one way to enhance the stability of free standing

lipid bilayers was to increase the ratio of bilayer-support contact area compared to the

total bilayer area. To test this assumption, chips with arrays of small pores were

fabricated. Later there was evidence from two reported studies to suggest that bilayers

are more stable in smaller pores. First, a higher resistance to disturbance by electrical

forces with decreasing pore size was demonstrated by measuring the minimum voltage

required to break a lipid bilayer formed on Teflon [84]. In this study not only the

circumference to bilayer area was changed but also the total bilayer area was reduced.

This complicates the interpretation and the observed stabilizing effect cannot be directly

attributed to the increased ratio of circumference to bilayer area (discussed in more

Fig. 2.7 | Resistance of single chip pores

RM values were determined in a conductivity standard solution

(κ = 12.15 mS cm-1) using the 2 electrode setup and the specified

fitting model for the bare chip.


24

detail in the following paragraph). Second, the resistance of bilayers to a mechanical

force applied by an atomic force microscope cantilever was monitored [85]. Bilayers

suspended on 300 nm pores were mechanically stable and retained their elastic

behaviour whereas bilayers on 1 µm pores did not.

The breakdown voltage of a lipid bilayer can be determined using different

experimental protocols. Rectangular voltage pulses of short duration can be applied

which are incremented until the bilayer disrupts [86]. This has the disadvantage that the

details of the procedure (e.g. number of steps and pulse length), influence the

determined breakdown voltage [87]. This is less critical when applying a linearly rising

potential. But it was shown that the lifetime also depends on the slope of the voltage

increase and breakdown occurs at higher potentials using faster rising signals [87].

Consequently the reported breakdown voltages from different studies can hardly be

used to estimate the effect of the pore size on the bilayer stability. For example for

POPC bilayers formed on Teflon the reported values vary from 490 mV in a 105 µm

pore [87] to 270 mV in a 50 µm pore [84] and an increase in stability by reducing the

pore diameter is not evident. In the earlier mentioned study where this trend was

reported, the experimental conditions did in fact not change. But the addressed problem

that the total pore area was varied prevents drawing clear conclusions. According to the

transient aqueous pore model presented by Weaver et al. [88] which describes the

electrical breakdown of lipid bilayers, the total bilayer area plays an important role. It

assumes that occasionally small pores are formed by thermal fluctuations but that the

critical energy barrier is rarely exceeded and the pore tends to reseal. By applying a

transmembrane potential this barrier is reduced and pores in the bilayer exceeding a

critical size leading to unlimited pore expansion and breakdown are formed more

frequently. The rate of formation of such critical size pores depends amongst other

things on the total membrane area. Thus, the reported increase in stability of almost

100 % by reducing the pore diameter from 600 µm to 2 µm [84] might just be due to the

reduction of the total bilayer area by a factor of 90’000.

We measured the breakdown voltage of a bilayer formed from the lipid mixture

POPC/POPE (7/3, w/w) on different chip types in 150 mM NaCl, 20 mM NaAc pH 5.5

by linearly rising the potential with a rate of 1 mV per second (Fig. 2.8) using the four -

electrode setup. In the control experiments the bare chips were scanned and from the

obtained linear voltage - current response (Fig. 2.8 A) the resistances of the chips were


25

calculated using Ohm’s Law. Compared to the values found by EIS (Fig. 2.8 B) the

determined membrane resistances are slightly lower but the overall picture is perfectly

maintained and the PA800 chip with the lowest porosity shows the highest resistance.

Using this measurement setup allows performing DC measurements since the non-

polarisable electrodes are able to sustain an extensive charge flow over time. The

current through a bilayer (Fig. 2.8 C) remains low up to high potentials and increments

slowly until a sudden increase of the transmembrane current indicates bilayer rupture. In

a series of repetitive experiments (n = 5) this breakdown voltage was determined for the

individual chips (Fig. 2.8 D). No dependence on the pore diameter was observed.

Probably the small variation of the pore diameter from 800 nm to 200 nm does not

significantly improve the bilayer stability in an electric field. Studies using the chip

series with different pore diameters ranging from 0.8 µm to 25.6 µm that maintain a

constant total pore area (Table 2.1) could answer the question whether an increased

circumference to pore area has a stabilizing effect on the lipid bilayer under an applied

potential.

Fig. 2.8 | Breakdown voltage measurement

The voltage - current relationship was measured for three different chip types

by linearly increasing the applied potential (A). From the determined slope

RSNM was calculated and compared to the values obtained from EIS (B).

Bilayer breakdown (VBreak) was observed as a sudden increase in the current

(C) and did not depend on the pore diameter (D).


26

The lifetime of lipid bilayers in the absence of applied perturbations was determined by

monitoring the membrane resistance using EIS. A bilayer was considered stable as long

as RM remained above the threshold of 1 GΩ, a resistance that would be sufficient for

single ion-channel recording. A single open chip pore already results in a greatly

reduced membrane resistance of 10 MΩ (Fig. 2.6). The stability of bilayers formed on

800 nm pores was found to depend strongly on the nature of the lipid [89,56]. Painted

bilayers formed from POPE were stable for more than 6 days, POPC and DPhPC

bilayers for 2 days and the naturally occurring soy PC mixture only for 2 hours. A high

stability of bilayers formed from lipid mixtures is important to allow an adaptation of

the lipid composition to that of the native environment of the membrane protein being

explored. Using chips with smaller pores the stability of lipid bilayers was highly

improved, especially those formed from naturally occurring lipid mixtures and

independently of the preparation method (Fig. 2.9).

Lipid bilayers formed by the M-M method were less stable compared to painted

bilayers, resistance values above 1 GΩ were only observed for hours. The influence of

the lipid chain type on the stability was determined using chip PA200 and different PC

lipids in 50 mM phosphate buffer at pH 7.4 (Fig. 2.10).

Fig. 2.9 | Stability of bilayers formed from lipid mixtures

From four separate preparations the mean ±1S.D. of the 3

best is given in hours.


27

Membrane resistance values of POPC bilayers were above the threshold (green line) for

4 hours, the less stable DOPC and DPhPC membranes ruptured within 2 hours.

2.6 Bilayers in nanopores: concluding remarks and perspectives

The usefulness of nanopore arrays in a silicon nitride membrane for the formation of

stable lipid bilayers was analyzed. The fabrication of the structured support is highly

reproducible and allows the formation of regularly arranged and well-separated pores

with a low aspect ratio (pore diameter/pore length) and in a predefined number. The

disadvantage of using a silicon-based support membrane is the relatively high

capacitance of the device. This prevents a precise capacitance determination of the lipid

bilayer by introducing electrical noise resulting in higher signal error [84]. The

breakdown of a bilayer in a single pore can be easily detected, but the determination of

the total bilayer area via the measured capacitance is affected. The accuracy of the

measurements does not allow us to identify small changes in the capacitance. For

instance using PA800 as support and assuming an amount of 10 % lipid plugs in the

pores that failed to thin out would result in a capacitive change of 15 pF. This value lies

still within the measurement error (Table 2.2). A reduction of the chip capacitance could

be achieved by coating the surface with a low dielectric material such as parylene [90],

but this changes directly the surface chemistry and could influence the bilayer stability.

Another strategy would be to passivate only the backside of the chip. Since the two

silicon nitride layers are connected in series, the total chip capacitance would be

reduced just below the value exhibited by the backside.

Fig. 2.10 | Stability of PC bilayers formed by M-M on PA200


28

The nanopore system allows the formation of BLMs using different preparation

methods and lipid types. Depending on the application, a solvent containing bilayer can

easily be painted or a virtually solvent-free bilayer can be formed by the more laborious

and delicate Müller – Montal method. Reducing the pore size from 800 nm to 200 nm

improved the stability of the free standing lipid bilayers for all lipids tested, in the case

of the natural lipid soy PC even by a factor of about 30. This allows performing long

term measurements such as passive diffusion over a bilayer or to investigate membrane

proteins which may require specific lipid mixtures to be functional. Additionally the

BLMs were stable under an applied potential up to 250 mV; consequently

electrophysiological measurements are possible within a relatively large potential range.

As already mentioned a much more desirable way to form lipid bilayers on a nanopore

chip would be directly from (proteo-) liposomes. To realize this as a robust and reliable

method, still major research and development needs to be made. Large surface

modifications may be necessary to guarantee a high coverage and sealing.

3. Examination of the usefulness of bilayers in nanopores for functional studies using model proteins

29

3. Examination of the usefulness of bilayers in arrays of

nanopores for functional studies using model proteins

To corroborate the effective formation of lipid bilayers in most of the nanopores and to

prove that the achieved sealing is large enough to measure single channel currents we

decided to use well-characterized proteins as model systems. Toxins, for example, offer

the advantage that they are soluble in water and incorporate spontaneously into

preformed bilayers. We chose to study by electrochemical techniques the spontaneous

insertion of representatives from two types of pore forming toxins (PFT’s) found in

nature: the bee venom peptide melittin which inserts into lipid bilayers in a α-helical

conformation (α-PFT) and α-hemolysin (α-HLY) from Staphylococcus aureus which

inserts a long loop that becomes part of an oligomeric β-barrel (β-PFT).

3.1 Melittin

Melittin is a cationic, amphiphilic polypeptide composed of 26 amino acids (Fig. 3.1 A).

The carboxy-terminal region is hydrophilic whereas the amino-terminal region is

predominantly hydrophobic which makes this peptide highly water-soluble, but also

capable to associate with natural and artificial membranes [91]. The binding of melittin

to membranes and its lytic activity depend on the lipid composition of the bilayer. It has

been shown that negatively charged lipids in membranes inhibit lysis and that melittin

binding to zwitterionic lipids is much lower than to negatively charged lipids [92,93].

After many years of research aimed at understanding the insertion mechanism of

melittin and its structure within biological membranes, it has been proposed that

melittin molecules form most likely tetrameric pores within bilayers [94]. The melittin

monomers appear to bend the external monolayer-sheet via strong ionic interactions

with the lipid headgroups such that a toroidal hole, delineated by both the peptides and

the lipid headgroups (Fig. 3.1 B) is formed [95].


30

3.1.1 Interaction of melittin with lipid bilayers

Melittin is able to form tetramers in solution under certain conditions such as high ionic

strength or in the presence of phosphate at physiological concentrations [98]. Typically

conformationally disordered monomers bind to lipid headgroups and are considered to

undergo a structural change, forming a α-helix that is oriented parallel to the membrane

surface and partially inserted such that the structure of the bilayer is perturbed [99]. To

initiate lysis, a collective action of several monomers is necessary. In POPC vesicles the

required amount was estimated to be one melittin per 200 lipid molecules [93]. At low

melittin concentrations individual pores were detected under an applied electric

potential load, demonstrating that also channels can be formed [91]. In some cases the

observed discrete conductance fluctuations of single pores had variable amplitudes,

indicating heterogeneity in the properties of the pore structure [100].

We performed current measurements after the addition of 300 nM melittin to a POPC

bilayer in 0.5 M NaCl at a constant potential of 150 mV. Under these conditions the

frequency of the pore opening and the lifespan of pores are high. The relatively high

voltage leads to large fluctuations in melittin-induced conductance across the bilayer,

resulting in highly variable current steps as reported earlier [101]. Discrete stepwise

conductance fluctuations were observed over a long time period (Fig. 3.2 A). In the

corresponding current distribution histogram a peak separation of 10 pA is found

(indicated by the vertical dotted lines) which corresponds to a single pore conductance

of 67 pS (Fig. 3.2 B). This value is compatible with a previously reported single pore

conductance of 110 pS measured at the high salt concentration of 1.8 M NaCl [101].

Fig. 3.1 | Schematic representation of melittin

A) Topology plot and helical wheel projection generated using TEXtopo [96]. Transmembrane helix

was predicted using the HMMTOP server [97]. B) Schematic drawing of the toroidal pore model [94].


31

3.1.2 Voltage dependence of melittin induced conductivity

It is known that the probability for open melittin pores increases with higher positive

transmembrane potentials [101]. If melittin is added only to one side (cis) of the lipid

bilayer, an increase in the conductance is only observed if the voltage is positive in cis

relative to the trans compartment. The induced increase of the permeability at a higher

potential has been assigned to a voltage-dependent structural rearrangement of melittin

within the membrane. It is assumed that melittin penetrates the bilayer in the adsorbed

state but does not extend across the membrane until a trans-negative voltage is applied

(Fig 3.1 B). The positive charge in the amino-terminal region has been shown to be

important for this voltage dependent orientation [102].

We took advantage of the preference of melittin for negatively charged lipids and added

100 nM of the peptide to the cis-side of a DOPG bilayer. We observed an asymmetric

current-voltage relation with an exponential increase of the conductance at positive

potentials above 130 mV (Fig. 3.3) whereas the conductance profile of a BLM before

melittin addition exhibits a pure ohmic behaviour.

Fig. 3.2 | Current fluctuations induced by melittin channels

A) Current measurement across POPC bilayers formed on chip PA200, 45 min after

the addition of 300 nM melittin and B) the corresponding current amplitude

distribution histogram. The measurement was performed in 0.5 M NaCl, 5 mM Tris

pH 7.5 at a holding potential of 150 mV. The arrow indicates a peak separation of

10 pA corresponding to a conductivity of 67 pS.


32

3.2 Hemolysin from Staphylococcus aureus

In contrast to melittin pores, the spatial structure of heptameric α-HLY pores in lipid

membranes is known at atomic resolution [103] (Fig. 3.4 A) although the precise

mechanism of heptamer assembly and insertion is still poorly understood. The water

soluble, 33.2 kD α-HLY monomer binds to cell membranes in a species-specific

manner. For instance, rabbit erythrocytes are lysed at very low concentrations of about

1 nM suggesting a specific binding site, whereas for human erythrocytes non specific

adsorption of α-HLY is indicated since a toxin concentration > 1 µM is needed for lysis

[104]. So far specific binding sites could not be identified, but a preferred binding of

α-HLY monomers to phosphocholine head groups has recently been reported [105].

The adsorbed monomers are assumed to oligomerize on the membrane surface by lateral

diffusion and form a heptameric, non-lytic prepore (Fig. 3.4 C). In a final step the

central regions of the subunits translocate across the membrane and irreversibly form a

functional, water-filled pore. The diameter of the pore in cross-section ranges from 4.6

nm down to 1.4 nm at its narrowest site (Fig. 3.4 B), resulting in a large single channel

conductance that linearly depends on the conductivity of the electrolyte [106].

Fig. 3.3 | Current – voltage curve

I-V curve of a painted DOPG-bilayer formed on chip PA800 in 0.25 M

NaCl, 5 mM Tris pH 7.5 before () and after () the addition of 100 nM

melittin to the cis compartment.


33

3.2.1 αααα-HLY characterization

Functional α-HLY is a heptamer that assembles on the lipid bilayer membrane. To

assess its oligomeric state in solution α-HLY was analysed using static light scattering

(SLS) (Fig. 3.5).

α-HLY eluted at about 18 mL on the gel filtration column which corresponds to a mass

of 32.3 kD with a polydispersity of 1.000 as calculated by the software. This excellent

agreement with the mass of a monomer clearly indicates that no oligomerization occurs

in solution and that monomers are likely to bind individually to the bilayer.

Fig. 3.4 | Structure and supposed assembly mechanism of the αααα-hemolysin pore

A) Lateral view of the heptameric α-HLY pore with one subunit coloured in red. B) Surface

representation of a bisected channel [PDB access code 7AHL, Song 1996]. C) Proposed pore

formation mechanism. Adapted from Kawate, T. et al. 2003 [107].

Fig. 3.5 | Gel filtration chromatogram from the SLS measurement

The thick horizontal line within the peak represents the calculated mass of the

sample. The dashed line is the measured UV signal; the blue line is the scatter

signal.


34

3.2.2 Single channel current

To further confirm that our setup can be used to monitor membrane channels we

measured the current across POPC bilayers formed on chip PA200 in the absence and

presence of α-HLY (Figure 4).

At an applied potential of 50 mV, bilayers in a sodium-acetate buffer at pH 5.5 are

stable for hours and show a constant transmembrane current of 20 pA which

corresponds to a membrane resistance of 2.5 GΩ (Fig. 3.6A, curve I). After α-HLY

addition a stepwise increase of the current was observed (Fig. 3.6A, curve II). The

conductance increments in steps of 9 pA as indicated by the dotted lines (Fig. 3.6B) and

each step is attributed to the formation of an individual stable heptameric pore. The

measured single pore conductance of 180 pS corresponds well with the reported value

of 7.2 pA at a holding potential of 30 mV and 250 mM KCl [108]. In order to

discriminate between bilayer rupture and α-HLY pore formation we measured a POPC

bilayer in 20 mM MOPS buffer at pH 7. Under this condition the bilayer is less stable

and ruptures occasionally occur within hours. The initial membrane resistance was in

the G-Ohm range and continuously decreased to the still high value of 0.7 GΩ within 8

hours (Fig. 3.6A, curve III), presumably due to weakening of the interaction to the

support. The sudden rupture of a lipid bilayer in a pore was identified as a current jump

of 200 pA, a value which is much higher than the observed steps of 9 pA after α-HLY

Fig. 3.6 | Monitoring pore formation of αααα-HLY

A) Current measured across a POPC bilayer under different conditions. A buffer containing

20 mM NaAc, 150 mM KCl at pH 5.5 was used to record the current across the bilayer with (II)

and without (I) the addition of 200 nM α-HLY (indicated by the arrow). In 20 mM MOPS,

150 mM KCl at pH 7.0 the lipid bilayer was instable (III). B) Blow up of the indicated region of

curve A II.


35

addition. The expected current through an open 200 nm chip pore (eq. 2.2) under an

applied potential of 50 mV would be 4000 pA. Presumably residual lipid material

located at the pore wall reduces the actual current. This strongly indicates that we are

able to observe the formation of single α-HLY pores in bilayers suspended in the

different chip pores and that these bilayers arranged in arrays are independent from one

another as substantiated by the observed breakdown of an individual bilayer.

3.2.3 Onset of αααα-HLY pore formation depends on monomer concentration

To further identify the factors influencing α-HLY pore formation in POPC bilayers, we

varied the concentration of the protein as well as the pore diameters and monitored the

current across the BLM over time (Fig. 3.7). The toxin was added after 1000 s to the

cis-side followed by 1 minute of rigorous stirring. The transmembrane current was

recorded at a holding potential of 50 mV.

At a first glance the onset of pore formation depends on the α-HLY concentration and

the chip pore size. As expected we observed that the height of current steps is

independent of the chip pore diameter, thus confirming that α-HLY pores and not single

bilayer ruptures are observed. We further found that protein pore formation occurs

earlier and more frequently at higher monomer concentration. For instance using chip

PA200 and 500 nM α-HLY the lag time is 2400 s whereas at 1 µM α-HLY the first

protein pore appears already 400 s after α-HLY addition (Fig. 3.7 B). This lag time is

comparable with the reported value of 40 s for free-standing DPhPC bilayers with

Fig. 3.7 | Concentration dependence of αααα-HLY pore formation

Current measurement across POPC bilayers formed on A) chip PA800 and B) chip PA200 at the

indicated α-HLY concentrations. The measurements were performed in 150 mM KCl with 20 mM

sodium-acetate buffer pH 5.5 at a holding potential of 50 mV. α-HLY was added 1000 s after start

followed by 1 minute of stirring, visible as disturbances in the curves.


36

190 nm pore diameter to which 1.5 µM α-HLY was added under a low transmembrane

pressure [109]. Unexpectedly the lag time also depends on the chip pore size. Whereas a

concentration of 200 nM is sufficient to induce pore formation on chip PA800 within the

experimental time (Fig. 3.7 A), 500 nM are needed to see this effect with bilayers on

chip PA200 (Fig. 3.7 B). Compared to the previously described α-HLY experiment

(Fig. 3.6 B) the solutions were stirred and the toxin was not injected near the membrane

surface which might explain the longer lag times for channel formation.

3.2.4 The chip pore diameter has an impact on the measured lag time

In a more systematic study we determined the time until a continuous pore formation

process was observed after the addition of 300 nM α-HLY to POPC bilayers in arrays of

pores with various diameters (Fig. 3.8).

Continuous pore formation was observed 21 minutes after monomer addition to bilayers

on chip PA800, after 220 minutes for chip PA400, and after 402 min for chip PA200.

Thus, the lag time for the formation of conducting α-HLY heptamers is strongly

dependent on the area of the individual chip pores and to first approximation this

dependence appears linear with inverse pore area (1.8 times shorter for a 4-fold

increased pore area, 19 times shorter for a 16-fold larger pore area). Such dependence

appears plausible under conditions where the rate limiting step is the adsorption of a

sufficient number of monomers which are competent to assemble into a conducting

Fig. 3.8 | Onset of αααα-HLY pore formation depends on chip pore size.

Current measurement across POPC bilayers formed on chips with various

pore diameters as indicated. In each experiment 300 nM α-HLY was

added after 1000 s, followed by 1 minute of stirring. The measurements

were performed in 150 mM KCl, 20 mM sodium-acetate buffer pH 5.5.


37

pore. This is illustrated in Fig. 3.9 where an assumed distribution of surface-adsorbed

monomers at some early time-point is shown with different pore sizes superimposed.

We assume that only monomers adsorbed to the lipid bilayer in the pores are mobile and

that proteins adsorbed to the support area are in an inactive state. Oligomerization can

thus only occur through the assembly of monomers bound to the same chip pore.

Lateral diffusion and/or the structural rearrangement leading to α-HLY pores cannot be

rate-limiting under our conditions as this would not be consistent with the observed pore

size dependence. However, it is presently not fully clear why (productive) adsorption

appears so slow under our conditions. It certainly is not simply diffusion-limited.

Fig. 3.9 | Effect of the chip pore on the formation of αααα-HLY heptamers

In this scheme the monomer binding pattern (red spots) and the total pore area are identical, but

the pore diameter varies. We assume that the monomers are mobile only in bilayer areas. In

800 nm pores nine monomers are present allowing the formation of a heptameric pore as shown

in the cross-sectional view beside. In each of the four 400 nm pores two monomers are present,

whereas only half of the 200 nm pores contain 1 monomer in average (X = 0.5). These

considerations illustrate that the probability of pore formation at a chosen time point is

proportional to the pore area.


38

3.2.5 Two-sided symmetric toxin addition

We expected that the symmetric addition of identical α-HLY portions does not

influence the observed lag time compared to the addition of only one portion to one

side, since all monomers assembling into a pore must be on the same side in our model.

Thus the time required to form a heptamer is not changed but the subsequent pore

formation rate should be doubled. In repeated experiments we found that the lag time is

highly reduced if the same toxin amount was added on both sides. Pore formation

started once 75 min and in a second experiment 73 min after the addition of 250 nM

α-HLY to cis on chip PA800, as expected (Fig. 3.7A). The symmetric addition of

250 nM α-HLY per side resulted in lag times of 150 min and 196 min. Contrary to our

expectation, this experiment revealed that monomers bound on both sides of the bilayer

might interact and retard pore formation. During oligomer assembly the transmembrane

domains are protected from proteolysis in a pre-insertion state [110]. These occluded

loops might penetrate the lipid bilayer far enough to interact with monomers bound to

the opposite side. An alternative possibility is that monomers bound to the membrane

induce bilayer curvature. White et al. controlled the insertion of α-HLY by applying a

positive transmembrane pressure and were able to remove inserted heptamers under

negative pressures [109].

3.3 Diffusion through aqueous pores

Diffusion of ions across aqueous pores can be described as follows. An initial

concentration gradient (∆C0) equilibrates with time in an exponential manner depending

on the total pore area (A), pore length (L) and the compartment volumes on the cis (Vc)

and trans (Vt) side. In smaller volumes concentration changes emerge faster; a large

gradient result in continuous changes and the total pore area reflects the solute amount

that is able to cross the barrier (Fig. 3.10 A). The apparent diffusion coefficient is given

by [111]:

+=

∆

∆

⋅=

tcapp

VVL

Awith

tC

C

tD

11

)(ln

1 0 ββ (eq. 3.1)

Since the pore area of the septum is precisely known we can calculate the constant β for

our setup.


39

Sodium ion diffusion across bare chips was measured using a sodium selective

electrode. The apparent diffusion coefficients, derived using equation 3.1, were much

lower compared to the reported values for sodium ions (DNa+ = 1.33·10-5 cm2 s-1 [112] ).

The discrepancy was 9 fold using the measurement cell with a 100 µL volume and 83

fold using the measurement cell with 1.5 mL (Fig. 3.10 A). Unstirred layers (USL),

known to be present parallel to a membrane, can be assumed to be responsible for these

observations. Their thickness can vary between 20 µm and 500 µm depending on

stirring speed, cell geometry and the solute diffusion coefficient [113,114]. The

dependence on the stirring speed was demonstrated by measuring the diffusion of

calcium ions across bare chips using a calcium selective electrode (Fig. 3.10 B). The

observed discrepancies in the diffusion coefficient (73 and 88 fold reduction) relate to

USL thicknesses of 169 µm and 202 µm respectively (see appendix for equations used).

3.3.1 Quantitative sodium diffusion measurements

In order to demonstrate that net ion transport across protein pores can be quantified, we

added 100 mM NaCl to the cis compartment and monitored the sodium ion

concentration in the trans-compartment using a sodium selective electrode

(Fig. 3.11 A). All experiments were performed using the same measurement cell with a

Fig. 3.10 | Solute diffusion across bare chips measured using ion selective electrodes (ISE).

A) The effect of the total pore area and the compartmental volumes on the measured sodium ion time

course. 100 mM sodium was added on cis and the concentration was measured on trans, the volumes

were stirred at 150 rpm. Despite the smaller total area of chip PA800, half of the sodium equilibrium

concentration is obtained within 75 min using a 100 µL volume. B) Calcium diffusion over chip

PA800, Vcis = Vtrans = 3 mL, 20 mM gradient at the indicated stirring speed. Unstirred layers were

calculated to be 169 µm at 500 rpm and 202 µm at 250 rpm.


40

volume of 1.5 mL on the cis and trans side and a stirring speed of 150 rpm using a

5 mm long stirring bar with a diameter of 2 mm.

To minimize the effect of interfering ions, the measurements were performed in 5 mM

diammonium hydrogenphosphate and 50 mM ammonium chloride at pH 7.5. This buffer

composition contributed a relatively low signal background and painted POPC bilayers

were stable for at least 6 hours on chip PA200. We found experimentally a detection

limit for Na+ of about 10 µM under these conditions (Fig. 3.11 B). The lag time of

α-HLY pore formation with 500 nM and PA200 using this buffer was 35 min, similar to

the previously determined value of 38 min (Fig. 3.7B). The measured diffusion

coefficient for the bare chip PA200 (Fig. 3.12) was Dapp = 2.1·10-7 cm2 s-1,

approximately 60 times lower compared to the reported value. For this setup an USL

thickness of 80 µm is calculated to explain the observed slowed diffusion [116].

Fig. 3.11 | Diffusion setup condition

A) Illustration of the test assembly. The chip is sandwiched between two half cells,

sodium chloride is added cis and Na+ concentration is measured trans. B) Standard

curve of the MI-425 ion selective electrode in 5 mM (NH4)2HPO4 and 50 mM NH4Cl

at pH 7.5. The slope factor of the electrode is close to the ideal 59 mV per decade

and the cell constant was -20 mV [115].


41

The time course of the sodium concentration increase appears linear due to the marginal

changes in the sodium gradient during the measurement. After bilayer formation the

sodium concentration was monitored over time and the observed diffusion coefficient

was highly reduced. From this coefficient the apparent membrane permeability was

calculated (eq. 3.2) as P´Na = 7.14·10-9 cm s-1 assuming a sodium partition coefficient of

K = 6.8·10-7 and a lipid bilayer thickness of δ = 4 nm [113].

δKD

P meas=′ (eq. 3.2)

This value is 70 times higher than a previously reported sodium permeability of

PNa = 1.0·10-10 cm s-1 measured across liposome membranes [117]. The authors

hypothesized an inhibitory effect of bent bilayers in liposomes on ionic permeability

since they found a large variation up to 104 depending on the vesicle size. The higher

observed sodium permeability for planar bilayers can be explained similarly by the

reduced curvature. In parallel experiments we added different concentrations of α-HLY

to POPC bilayers present in the pores of a P200 chip and after waiting 2.5 hours in

order to complete protein pore formation, the sodium concentration measurement in the

trans compartment was started. The determined structure factor β (eq. 3.1) was constant

Fig. 3.12 | Diffusion across POPC bilayers with incorporated αααα-HLY

Measurements were performed on chip PA200 in 5 mM (NH4)2HPO4 and

50 mM NH4Cl at pH 7.5. 100 mM NaCl was added to the cis compartment

followed by the addition of different α-HLY concentrations to the same

side. After 2.5 h the concentration of Na+ was monitored in the trans

compartment with an ISE. The measured diffusion coefficients (Dmeas) are

given in 10-9 cm2 s-1 for: bare chip 210; 800 nM α-HLY 24; 500 nM

α-HLY 10.8 and 0 nM α-HLY (control) 4.2.


42

during the diffusion measurement, indicating that pore formation was indeed nearly

completed. At higher α-HLY concentrations higher permeabilities were observed,

reflecting a higher pore density in the lipid bilayer (Fig. 3.12). The number of α-HLY

pores in the bilayer has been estimated from the measured diffusion coefficients

according to eq. 3.1, assuming that all diffusive transport is through the protein pores

featuring a cylindrical shape with a length of 10 nm and a diameter of 4 nm. Thus,

addition of 500 nM α-HLY leads after 2.5 hours to the formation of about 60’000

protein pores, whereas about 150’000 protein pores are formed at a concentration of

800 nM. Since 960’000 pores are present on this chip, on average one heptameric

α-HLY pore is formed in 7 % of all nano-bilayers at 500 nM α-HLY and in 15 % of

them at 800 nM. In this estimate the presence of USL was not considered and the

diameter of the simplified cylindrical protein pore was chosen rather large. Thus the

obtained numbers of protein pores represent the lower limit and the effective amount

might be much higher.

3.3.2 Diffusion experiments with calcium

A challenge in using ion selective electrodes to measure ion diffusion was to find a

buffer with no ions interfering with the ion selective electrode yet permitting the

formation of stable lipid bilayers. The previously described sodium selective electrode

is differently affected by Ag+, Li+, K+ and NH4+ and a background signal has to be

accepted. The plan was to use an ion that can be determined with high selectively.

Measurement of the divalent calcium ion is nearly unaffected by interfering ions. All

indicated selectivities (Ca2+/ion) are larger or equal to 10’000 promising a low

background signal. The experimentally found detection limit was 1 µM as quoted by the

supplier. Interactions with the phospholipids in the bilayer can cause breakdown and

asymmetric addition results in increased membrane conductance [118]. The addition of

100 mM CaCl2 on cis was compensated by the addition of the same amount of MgCl2 on

trans, and bilayers remained stable at the expense of some background signal. Upon the

addition of 250 µM α-HLY no increase of the calcium concentration on the trans side

was detected within 12 hours. The bilayer resistance after this time was 20 MΩ in this

specific experiment. Assuming a α-HLY conductance of 180 pS suggests that only 300

heptamers were formed in the POPC bilayer on chip PA800. This can be explained by

the findings of Bashford et al. that divalent cations inhibit the permeabilizing action of


43

pore forming agents such as α-HLY [119]. For this reason no further experiments with

calcium ions were performed.

3.4 Using model proteins: concluding remarks and perspectives

The insertion of melittin and α-HLY into bilayers and the formation of pores have been

monitored. Distinct mechanisms contribute to the induced ion permeability increase of

the lipid bilayer. The alpha helical peptide melittin bends the bilayer and forms a

toroidal type pore that is delineated by lipid headgroups. This pore is transient and

opening depends on the transmembrane potential. On the other hand α-HLY heptamers

form a stable transmembrane beta barrel structure with a hydrophilic interior. This water

filled barrel-stave type pore allows the diffusion of various ions and also larger solutes

down their concentration gradient. Consequently this setup should also be suitable to

investigate other peptides and proteins of interest that exhibit various structural motifs.

We additionally showed that a large number of membrane proteins could be integrated.

In the case of α-HLY (> 5 estimated protein pores per µm2), which assembles from

seven monomers that individually bind to the bilayer, this also proved that in most chip

pores a single lipid bilayer had formed. Since the bilayer carrying pores are isolated

from one another, heptamer formation occurs primarily in different chip pores prior to

the formation of a second protein pore in the same bilayer. Thus, each observed current

step relates to a separate lipid bilayer in the majority of cases. Assuming that the lytic

effect of α-HLY evokes from a combined action of several protein pores by local

bilayer perturbation, the separation of the proteins into different compartments prevents

an early bilayer collapse. With the high number of protein molecules in the bilayer we

were able to follow directly the net diffusion of sodium ions through the protein pores

using ion selective electrodes. Thus this setup should also be suitable to investigate the

transport of electro-neutral species which is much more demanding since no

potentiometric or amperometric methods can be applied. Since the translocation rate of

transporters is much lower compared to the ion flow across channels, signal

amplification can be achieved by measuring the combined action of many proteins.

Assuming a hypothetical transporter which has a transport rate of 104 ions s-1, a

measurement volume of 50 µl and that a detectable signal (∆Cion = 3 µM) is obtained

within 3 hours, this requires that 840’000 transporters are integrated in the lipid bilayer

or approximately one in every pore of chip PA200. This is achievable if the bilayer

could be formed in a single step directly from proteoliposomes and if the lipid to protein


44

ratio is maintained. The controlled delivery of this amount of proteins to a preformed

bilayer within a reasonable time seems rather questionable. If the surface density of

membrane proteins is known, unitary permeabilities can be determined. Such

quantitative determinations are important to investigate interactions of ligands and

proteins with integral membrane proteins aimed at investigating dynamic biological

processes of interest.

4. Measurements with the voltage gated sodium channel /aChBac

45

4. Measurements with the voltage gated sodium channel

)aChBac

The diverse activities of integral membrane proteins in signalling and transport across

cellular membranes make them very important in sustaining life functions. In many

diseases, a pathway controlled by a membrane protein is in a disarranged state.

Therefore it is not surprising that many membrane proteins are major drug targets and

are in the focus of researchers and thus pharmaceutical industry. They often are

composed of several transmembrane helices and functional domains on either side of

the membrane thus their recombinant larger scale expression is often difficult. To

maintain them in their functional conformation, the lipid environment can be essential

but often can be substituted by detergent micelles. Compared to the toxins used to probe

the usefulness of lipid bilayers formed in nanopores as described in Chapter 3,

membrane proteins are not soluble in the absence of detergents and do not insert

spontaneously into lipid bilayers. Special methods are required to insert such purified

membrane proteins into artificial bilayers. Ideally, one would like to precisely control

the amount of protein inserted in the bilayer. This would allow introducing only one

molecule to measure single channel currents or to work with a large number of proteins

to be able to measure transport quantitatively and to calculate unitary transport rates. To

establish such methods on the nanoporous support, we have chosen the voltage gated

sodium channel from Bacillus halodurans (NaChBac) for the following reasons: (1)

compared to transporters, the assumed large ion transport rate of the channel allows

easier signal detection from individual molecules using electrophysiological recording

methods, (2) NaChBac activity was so far only recorded from cellular membranes of

heterologous expression systems, (3) structural research on such sodium channels is

currently carried out in-house with the main objective to understand how Na+ selectivity

is achieved and (4) the production and purification of this bacterial channel is well

established in our laboratory.

4.1 )aChBac of Bacillus halodurans

With the tetrameric voltage gated sodium channel from Bacillus halodurans, we have

chosen a relatively simple ion channel protein. In contrast, voltage gated sodium

channels of eukaryotes consist of one very large polypeptide chain with four

homologous domains which form a pseudo-tetrameric transmembrane channel


46

(Fig. 4.1 A). NaChBac consists of 274 amino acids with sequence similarities to voltage

gated calcium channels and forms a homo-tetramer in the membrane (Fig. 4.1 B). This

assembly is shared by the structurally well characterized tetrameric potassium channels

which are so far the best studied and understood voltage gated ion channels and serve as

model to describe the basic mechanisms of this protein class [120,121]. The four

domain architecture of voltage gated channels is shown in figure 4.1C. Each domain

contains six transmembrane helices (S1-S6) arranged symmetrically around a central

pore which is formed by helices S5 and S6, whereas helices S1-S4 of each domain form

the voltage sensor [122].

Prior to the discovery of NaChBac in Bacillus halodurans [27] it was believed that

voltage gated sodium channels exist only in higher organisms where they contribute to

the electrical excitability of cells. Shortly thereafter, two other voltage-gated sodium

selective channels of bacterial origin were identified after analyzing 11 NaChBac

homologs [123]. Their biological function in prokaryotes is still unknown but they are

believed to play a role in bacterial mobility, chemotaxis and pH homeostasis as sodium

re-entry limb of the sodium cycle [124].

Fig. 4.1 | Voltage gated channels

A) Alpha subunit of the human voltage gated sodium channel (Nav1.1). B) Voltage gated sodium

channel of Bacillus halodurans (NaChBac). C) Structure of a Shaker family voltage dependant

potassium channel viewed from the external side (Paddle-chimaera channel, PDB access code 2R9R,

Long, S.B. et al. 2007 [121]). The topology plots were generated using TEXtopo [96]


47

Voltage-gated ion channels of this superfamily can be divided into three main

functional parts: (I) the ion selective pore (II) the channel gate and (III) the voltage

sensor.

4.1.1 The ion selective pore

The central pore provides a pathway for charged ions to cross the lipid bilayer. The

selectivity filter is a narrow site in this water filled pore where specific ion binding sites

interact with the dehydrated permeant ions. Ions that do not bind firmly do not permeate

whereas very strongly binding ions act as pore blocker. Thus the function of the

selectivity filter is to assure a high selectivity combined with a high conduction rate. If

the channel is perfectly selective for sodium, an applied sodium gradient across the

membrane can be opposed with a defined electrical potential E to completely prevent

ion flow; this reverse potential can be calculated using the Nernst equation (eq. 4.1)

where F is the Faraday constant and z the charge of the ion.

[ ][ ]

=

in

out

Ion

Ion

Fz

RTE ln (eq. 4.1)

The ion flow and current respectively is proportional to the net driving force, which is

the difference between the applied voltage and the reversed potential.

Like in potassium channels, it is assumed that the transmembrane pore of NaChBac is

formed by helices S5 and S6 of the four subunits and that the loop connecting the two

helices forms the selectivity filter (Fig. 4.1). NaChBac has been determined to be highly

sodium selective with a permeability of sodium over potassium (PNa/PK) of ≈ 170 [125].

A triple mutation in the pore loop changes the selectivity from sodium to the divalent

ion calcium PCa/PNa ≈ 133 (Fig. 4.1 B, green residues), whereas each single of these

mutations is sufficient to change the channel into a sodium and calcium permeant pore

[125]. It is believed that no fast N-type inactivation mechanism for NaChBac exists but

that rearrangements of the pore causes a slow C-type inactivation of the channel [126].

4.1.2 The gate

Several residues were identified which are critical in the voltage dependent opening of

the NaChBac channel gate. For example a glycine in helix 6 was identified as gating

hinge (Fig. 4.1 B, brown residue). Its substitution by a proline which induces a bend in

the α-helix resulted in a channel which favoured the open state [127]. It was concluded

that bending at this residue is important for channel opening. However, the details of the


48

structural rearrangements of gating and their coupling to voltage sensor movements

remain unclear.

4.1.3 The voltage sensor

The conductance of NaChBac depends on the transmembrane potential; the channel is

closed at relatively high negative potentials and opens upon depolarization of the

membrane. The voltage of half-maximal activation is -24 mV and the probability that

the channel is in an open state rises with increasing depolarization [27]. In order to

respond to membrane potential changes, a voltage sensor is needed that detects changes

in the electric field and converts it to a structural response of the channel. Electric

charges and dipoles experience a force in an electric field and are key candidates to

build up such a voltage sensor. Since in all voltage gated ion channels helix S4 contains

a stretch of positively charged residues, this segment was believed to play a major role

and this has been confirmed by many mutagenesis studies in different such channels.

How S4 motion is coupled to channel gating is still not resolved in detail. A plethora of

work was carried out on potassium channels and several models emerged which try to

explain the interaction of the voltage sensor movement with the conformational change

of the pore [128]. But in the absence of detailed structural information of the closed

state the details remain controversial [122,129].

Helix S4 of NaChBac contains four positively charged residues (Fig. 4.1 B, blue

residues) which are separated by two hydrophobic residues. By systematically

neutralizing the charges in mutants it could be shown that these residues are important

in voltage sensing [130]. Upon depolarization, the inward directed force on the

positively charged helix is reduced, resulting in its release and outward movement. A

coupled conformational change results in the opening of the channel. This movement of

charge across the membrane was measured in the absence of sodium and calcium [131].

Consequently no ionic current was recorded and only the translocation of charge

originating from structural rearrangements (the so-called gating charge) was detected in

the potential range where the sensor responds. A total movement of 16 elementary

charges per channel was estimated [131]; thus each voltage sensor is involved in the

movement of four charges relative to the full membrane electric field which is very

similar to the gating charges measured for K+ channels. Other conserved acidic residues

in helix 2 and 3 were also shown to be important for voltage-sensing; whether they

operate via inter-domain interactions to make the structure more rigid or by generating

local electric fields is not clear [132].


49

4.2 Expression, purification and reconstitution of )aChBac

4.2.1 Expression vectors

Three different expression vectors were used (by courtesy of Xiao-Dan Li, PSI); their

characteristics are summarized in Table 4.1.

Identifier Vector Promoter Tag Selection marker

NaChBac

I9.1C pET-22b(+) T7 C-term. His6 Ampicillin Wild type I9.1N pET-15b T7 N-term. His6 Ampicillin Wild type I9.1NC pET-15b T7 N-term. His6 Ampicillin Truncated2

The coding sequence of NaChBac is under the control of the T7 promoter with the lac

operator sequence just downstream. The T7 RNA polymerase itself is located in the

chromosome of the host cell under the control of the lacUV5 promoter. Prior to

induction, only small amounts of the polymerase are present in the cell and production

of the target protein is further prevented by binding of the lac repressor to the operator.

Induction by IPTG leads to the production of the T7 polymerase and unblocking of the

T7 promoter, resulting in high levels of target protein expression [133].

Plasmid was isolated from 2 mL bacterial overnight cultures and the size of the insert

was controlled after a digest with the restriction enzymes NdeI and XhoI on a 1%

Agarose gel.

4.2.2 Small-scale test expression

In a first screen the expression of I9.1C was tested in the three E. coli strains

BL21(DE3), C41(DE3) and C43(DE3). Since the overexpression of some membrane

proteins has been reported to produce toxic effects in the BL21(DE3) host, the mutant

C41(DE3) and the double mutant host C43(DE3) with improved expression levels of

toxic proteins were also included in the screen [134]. Highest cell densities were

obtained with strain C43(DE3) and NaChBac was successfully expressed by all hosts as

verified by analyzing lysed cells with SDS-PAGE and Western blotting. Accordingly all

NaChBac constructs were expressed in C43(DE3) cells. After rupturing the cells,

membranes were isolated and solubilized in 1% DDM. The supernatant was loaded on a

PD-10 gravity-flow column packed with the Co2+-loaded chelating sepharose fast flow

2 25 residues of the carboxyl terminus are missing

Table 4.1 | Expression vectors.


50

affinity media. Unbound samples were removed by a wash step and NaChBac was

eluted using buffers described in Materials and Methods. The SDS-PAGE and Western

blot disclose the successful purification of all constructs (Fig. 4.2). The calculated

molecular weight of the NaChBac monomer is 32 kDa but the migration position on the

gel is below 30 kDa [135].

4.2.3 Large-scale over-production

The expression of the proteins and their purification is described in detail in Materials

and Methods. In brief, 12 times a one liter auto-induction medium [133] was inoculated

with 15 mL of an over night culture of C43(DE3) cells. In contrast to other nutrient

solutions no IPTG is needed to induce gene expression in ZYP-5052 [133] and the cells

were grown for 22h – 24h at 30°C (Fig. 4.3 A). After preparation of the membrane

fraction and its solubilisation in 1% DDM an affinity chromatography (IMAC) followed

by gel filtration was performed on the ÄKTAxpressTM (Fig. 4.3 B). The collected

fractions were analyzed using SDS-PAGE and Western blot.

Fig. 4.2 | Expression and purification of )aChBac constructs

A) SDS-PAGE and Western blot of the three constructs. 1: Solubilized membrane fraction

2: Unbound fraction (flow through) 3: Eluted fraction. B) SDS-PAGE of the purified

proteins and indicated standards.


51

Peak fractions A2-A5 were pooled and concentrated prior to the determination of the

protein concentration. Typically yield of purified protein per liter culture medium was

0.4 mg.

4.2.4 Reconstitution

NaChBac in elution buffer was mixed with the desired buffer containing 0.03 % DDM

and added to dried lipid resulting in 10 mg mL-1 lipid solution with a protein to lipid

ratio of approximately 1:5000 (mol/mol) and 1:100 (w/w) respectively. After dialysis

for 48 hours against a buffer without detergent the proteoliposomes were extruded ten

times to the desired size. The effective lipid to protein ratio was determined by

measuring the protein amount with a modified Lowry assay and the phosphate groups

were determined as described in Materials and Methods. In the majority of cases the

original lipid to protein ratio was maintained but the total amount was lower, most

likely due to material loss during extrusion. The size distribution of the proteoliposomes

was measured by dynamic light scattering (Fig. 4.4).

Fig. 4.3 | Purification of I9.1C from a 12 liter preparation

A) Bacterial growth was followed by regular OD600 measurements B) Chromatogram of the IMAC

followed by gel filtration.


52

Also other reconstitution conditions and methods were explored. These involved

detergent removal with BioBeads [136] or using the detergent β-OG. As found later this

did not yield functional channels.

4.3 Formation of bilayers with integrated )aChBac

The size of proteoliposomes is in general too small to apply electrophysiological

techniques. Therefore, the channel protein needs to be transferred to a planar lipid

bilayer or a bilayer has to be formed from proteoliposomes. This could be done either

by the M-M method, by the fusion of proteoliposomes to painted bilayers or by direct

adsorption of liposomes on a support. The last-mentioned technique was not further

explored because we were not able to achieve a high level of sealing in initial trials.


The simultaneous assembly of lipids and proteins into a bilayer with the M-M method

requires the initial self assembly of a lipid-protein monolayer at the air-water interface

[137]. One way to achieve this is starting with lipid-protein complexes in organic

solvents [138]. The purified protein is extracted into an organic phase using charged

lipids or surfactants forming ion pairs with the protein. Since this method involves a

volatile organic solvent, conditions need to be found under which the protein is not

irreversibly damaged. Another method is the spontaneous formation of lipid-protein

monolayers at the air-water interface from a proteoliposome suspension [139]. Upon

disintegration of the liposomes at the interface, a lipid monolayer is formed. Further

lipid incorporation stops when the monolayer has reached a certain packing density. The

Fig. 4.4 | Proteoliposome analysis

N-tagged NaChBac (I9.1N) was reconstituted in SPC proteoliposomes with a protein to

lipid ratio of 1:100 (w/w). The proteoliposomes were extruded to 100 nm by repeatedly

passing the solution through a filter of defined pore size. Starting with 800 nm diameter

the pore size was decreased in steps.


53

surface pressure increase depends on the vesicle concentration in the buffer [139].

Protein densities in the monolayer were found to match the lipid-protein ratio of the

used proteoliposomes within a factor of two [140]. Both methods, solvent spreading and

formation from a liposome suspension, yield lipid films of identical properties [141].

The M-M method is probably not suitable for delicate membrane proteins which do not

tolerate the conformational distortions in a lipid monolayer.

We were able to form stable lipid bilayers from proteoliposome suspensions on chips

containing arrays of pores. C-tagged NaChBac (I9.1C) was reconstituted with E.coli

polar lipids using a high protein to lipid ratio of 1:10 (w/w). In the case that the tetramer

is not stable at the air-water interface, this increases the probability that enough

monomers are in the same chip pore. The lack of channel activity in functional tests as

described below was attributed to the harsh conditions during bilayer formation. Partial

denaturation at the air-water interface might yield inactive protein or cause improper

incorporation. We were not able to discriminate these two possibilities. This failure with

the M-M method directed our effort toward the development of other methods. In

retrospect another possible reason why NaChBac turned out to be inactive in these

experiments, may have been its reconstitution with β-OG and the use of BioBeads for

detergent removal. The reconstitution condition which yielded functional channels (as

found in later experiments, see below) was never explored with the M-M procedure.

4.3.2 )ystatin / ergosterol vesicle fusion

Fusion of vesicles with lipid bilayers is an essential process in living cells. In the

biosynthetic-secretory pathway for example, proteins are transported from the

endoplasmic reticulum through the Golgi apparatus to lysosomes or to the plasma

membrane of the cell. This highly organized traffic is mediated by particular membrane

proteins, but fusion itself is not a very specific process. To induce fusion of liposomes

with an artificial planar bilayer, an osmotic gradient across the lipid bilayer combined

with open channels in the vesicle have been shown to be important and sufficient [142].

Divalent cations (Ca2+) increase the fusion rate if negatively charged lipids are used, but

they are not mandatory. Furthermore the absence of a hydrocarbon solvent in the bilayer

is not essential and fusion occurs with both types of bilayers regardless if they are

formed by the M-M method or by painting [143]. Unlike casual integration of protein-

detergent micelles into preformed planar bilayers, real incorporation of vesicle

membranes by fusion is suggested by the observation that channels insert as groups


54

rather than individually [144]. Woodbury introduced the surface tension induced fusion

model (STIF) which is able to explain the experimental findings [145]. He postulates

that a higher surface tension in the vesicle membrane compared to the lipid bilayer is

necessary to make fusion energetically favorable. This is achieved by generating a

hydrostatic pressure inside a vesicle which is induced by fast influx of ions through

open channels into the vesicle adhering to the bilayer with an osmotic gradient across it.

Because not all protein channels induce fusion, it is essential to make the vesicles

equally fusigenic independently from the protein of interest. An excellent candidate is

the pore forming antibiotic nystatin [146,147]. It binds to ergosterol and forms a barrel

consisting of 8 to 10 monomers with a conductance of about 5 pS [148]. The advantage

is that the sterol is required to hold the channel together. Thus the vesicle containing

several nystatin-ergosterol complexes together with the protein of interest is fusigenic.

But upon vesicle fusion to a lipid bilayer without ergosterol, the ergosterol diffuses into

the bilayer and the nystatin channels fall apart. Thus the nystatin-ergosterol method

allows the incorporation of virtually any membrane protein avoiding a background

signal. Additionally, fusion events can be recognized through a current signal caused by

the nystatin channels which makes it easier to interpret negative experiments.

Initial fusion experiments with this system on chip PA800 were not successful. After

verifying the nystatin activity by measuring its characteristic UV absorption at 416 nm

[149] and screening several lipid compositions the system was tested on a “classical”

black lipid bilayer in a micrometer aperture. We think that the capacity to assimilate

lipid material from the vesicle is limited in small pores and that this may prevent fusion.

This hypothesis was corroborated later when chips with larger pores became available

(see Table 4.2).

Nystatin-ergosterol vesicle fusion with a “classical” lipid bilayer

A hole was formed in a 0.03 mm thick polypropylene foil using a syringe needle.

Bilayer formation was verified using EIS. In contrast to the chip, the plastic foil does

not posses a measurable capacitance and fitting using an R(RC)(RC) equivalent circuit

was not possible. After bilayer formation (RM >10GΩ) a membrane capacitance of

30 ± 5 pF was determined. The effective bilayer area was calculated to be 3.34 E-5 cm2

by assuming a specific capacitance of 0.9 µF cm-2. This corresponds to a bilayer in a

pore with a diameter of 65 µm (Fig. 4.5).


55

In a first control experiment (Fig. 4.6) a POPC bilayer was formed which contained

ergosterol at a ratio 8:2 (mol POPC/mol ergosterol). Nystatin dissolved in dry methanol

was added to one side of the bilayer as indicated by the arrow. The final concentration

of 20 µg mL-1 was sufficient to monitor a continuous increase of the membrane

conductance. After a short lag time nystatin channels were spontaneously formed in the

bilayer. The small steps expected from single channels are not observable as the

sensitivity of the measurement is too low.

The second control experiment included the fusion of nystatin channel containing

vesicles (activated) with a bilayer containing ergosterol. Upon a fusion event several

nystatin channels are inserted into the bilayer and remain stable [147] (Fig. 4.7). The

bilayer was painted from a POPC/ergosterol (8:2 mol/mol) solution in decane.

Liposomes formed from the mixture POPC/POPE/ergosterol (4:4:2 molar ratio) were

Fig. 4.6 | )ystatin channel formation in a POPC bilayer containing ergosterol

Soluble nystatin binds to ergosterol containing bilayer and forms stable pores resulting in a

continuous increase of the transmembrane current at a holding potential of 50 mV.

Fig. 4.5 | Light microscope image of a perforated plastic foil

From the measured capacitance (30 pF) the lipid bilayer area was

calculated to correspond to that of the red circle.


56

extruded to 200 nm and 400 nm. A 1:1 mixture of the differently sized vesicles was

incubated with nystatin to allow time for channel formation. Upon addition to one side

of the bilayer fusion was induced by applying a urea gradient and with intense stirring.

The step height varied from large 20 pA jumps to only small abrupt increases in the

range of 3 pA. This is assumed to reflect the different number of nystatin channels

which are incorporated in the bilayer upon a single fusion.

The fusion of activated vesicles with a POPC bilayer which does not contain ergosterol

results in a sudden increase of the bilayer conductance depending on the number of

nystatin channels. Thereafter ergosterol diffuses into the lipid bilayer resulting in a

decay of nystatin channels and a subsequent reduction of the transmembrane current

(Fig. 4.8). The shape of the fusion peak depends on the nystatin and ergosterol

concentration [150].

Fig. 4.7 | Liposome fusion with a POPC bilayer containing ergosterol

Fusion of nystatin-activated liposomes with an ergosterol containing bilayer results in the

incorporation of several channels per fusion event. This results in a stepwise increase of the

transmembrane current at a holding potential of 50 mV. The step height reflects the amount of

channels which are integrated in one fusion event.


57

The molecular mechanism of nystatin channel decay is not yet clear. Helrich et al.

propose a model [151] in which the hundreds of tiny channels that sum up to the large

peak current decay in a correlated manner. If nystatin channels would close

independently an exponential decay curve is expected. Thus the linear decrease is

explained by the assumption that nystatin channels form at the boundary of a circular

sterol superlattice. Radial diffusion of ergosterol from this domain is constant which

produces a linear decrease of the recorded current. Another explanation was suggested

by de Planque et al [150]. They assume that nystatin-ergosterol channel conductance is

much higher compared to the very low reported nystatin-cholesterol conductance. Thus,

the peak current is generated by only a small number of channels (one to two) which

gradually inactivate upon diffusion of the associated ergosterol. Consequently the

inactivation rate depends on the location of the channel in the patch after fusion. The

finding that the peak current does not correlate with the vesicle surface area [152]

strongly supports the superlattice model.

We analyzed the fusion characteristics of differently sized liposomes with respect to the

ease of fusion. No striking differences were observed between liposomes extruded to

100 nm, 200 nm or 400 nm. No fusion was observed using 50 nm liposomes, most

probably because the potentially existing peaks are too small and too fast to be detected.

Comparing peak current and vesicle size disclosed that smaller vesicles tend to induce

smaller spikes but peak height seems not to correlate with the surface area (Fig. 4.9).

The approximately linear relationship to the vesicle diameter could be confirmed [151].

Fig. 4.8 | Liposome fusion with a POPC bilayer

Liposomes formed from the mixture POPC / POPE / ergosterol (4:4:2 molar ratio) were extruded to

200 nm and activated with nystatin. At a holding potential of 100 mV fusion events with a POPC

bilayer result in current spikes. Characteristic is an abrupt current increase that reflects the fast

fusion event followed by a slow decrease to the baseline as a result of nystatin channel decay.


58

Nystatin-ergosterol vesicle fusion with chips

Initially, experiments on chips with arrays of nanopores failed using liposome sizes

from 600 nm to 50 nm. We assumed that a bilayer formed in an 800 nm pore may not

be able to assimilate a 100 nm liposome which corresponds to an increase of more than

6 % of the bilayer area. Reduction of the vesicle size does not help since fusion peaks of

the smaller 50 nm vesicles cannot be detected. Having chips available with a larger pore

area the feasibility of vesicle fusion was tested. We used chips with a constant total pore

area to be able to compare the probability of fusion (Table 4.2).

Fig. 4.9 | Average fusion peak current compared to vesicle size

A) Liposomes formed from the mixture POPC / POPE / ergosterol (4:4:2 molar ratio)

were extruded ten times to 400 nm, 200 nm and 100 nm. Polydispersity was measured

using dynamic light scattering. The peak current is not proportional to the square of

the vesicle radius. B) Average fusion peak generated with the indicated vesicle size.

Measurement was performed at a holding potential of 100 mV (n ≥ 5).


59

Chip type Liposome diameter (nm)

Pore diameter (nm)

Liposome area to pore area (%)

Success

Foil 400 > 100’000 < 0.006 +++ SP25.6 100 25’600 0.006 + CA12.8 100 12’800 0.024 + CA6.4 100 6’400 0.097 + CA6.4 200 6’400 0.39 - CA3.2 100 3’200 0.39 - CA1.6 100 1600 1.56 - PA800 100 800 6.25 ---

According to the STIF model fusion takes place if the increased surface tension of the

vesicle caused by the hydrostatic pressure can be reduced upon integration into the lipid

bilayer. The required tension in a vesicle was estimated to 30 µN cm-1 [145]. On the

other hand a steep increase in the lateral pressure upon a reduction of the area per lipid

molecule was reported [153]. But it can be assumed that not the whole material is

absorbed within the planar lipid bilayer. The edge of the aperture and the annulus

provide also space for the accumulation of the added lipid material. And if the

individual monolayers are mobile on the support, assimilation should not be a problem.

This could explain the findings that fusion is only possible above a defined bilayer to

vesicle area threshold on the chip. The thin silicon nitride membrane and its

hydrophobically silanized surface do not facilitate vesicle fusion, whereas on the thick

plastic foil small amounts of additional material incorporation seems to be negligible.

4.4 Measurements with reconstituted )aChBac

A lipid bilayer was formed across a small hole in a plastic foil by painting. The fusion

of activated proteoliposomes with this “classical” lipid bilayer results in a large fusion

peak, evoked by the combined conductance of the nystatin channels and the open

voltage gated sodium channel (Fig. 4.10 A).

Table 4.2 | Qualitative analysis of fusion probability.

(+++) High success rate, often multiple fusion peaks (+) seldom, mostly single fusion events

(-) no fusion observed within three days (---) no fusion observed within several weeks


60

The decrease of the current after a fusion event (indicated with red circles) is attributed

to the decay of nystatin channels. The residual transmembrane conductance is generated

by the introduced NaChBac channels. Inactivation results in a discrete stepwise

decrease (indicated by a series of blue dashed lines) which allows us to estimate the

single channel current to 120 pS (Fig. 4.10 B). This conductance corresponds to a flow

of 75 million sodium ions per second through one channel molecule, a value that is one

order of magnitude higher compared to the reported 12 pS for NaChBac which has been

expressed and measured in CHO cells [27]. The observed inactivation time for the

bacterial channel is very slow compared to other sodium channels which close within

milliseconds and NaChBac corresponds rather to the mammalian non-inactivating

sodium channels [154]. Such a long open state was also reported by Pavlov et al., who

found that NaChBac inactivation depend strongly on the applied voltage with time

constants in the range of 100 s to 100 ms [126].

The number of functional channel proteins per liposome can be estimated from the

finding that two to three channels are incorporated simultaneously by a fusion event

(Fig. 4.11 A). With a protein to lipid ratio of 1:5000 (mol/mol) and an assumed area per

lipid molecule of 63 Å2 [152], the theoretical number of NaChBac monomers in a

200 nm liposome is 80. Assuming a random orientation, 10 channels could be measured

from the total 20 functional tetramers. From this assessment a reconstitution success of

25 % has resulted. A random orientation was confirmed by measuring the

transmembrane current at positive and negative potentials (Fig. 4.11).

Fig. 4.10 | Fusion of )aChBac containing liposomes with a POPC bilayer

A) The fusion of proteoliposomes extruded to 200 nm with a painted bilayer generates a large

transmembrane conductance at a holding potential of 100 mV. Decay of the nystatin channels

(red circles) after fusion results in a current decrease. B) Amplitude histogram of the data from

A indicates a mean single channel current of 12 ± 3 pA. Fusion conditions as described in 6.6.3


61

NaChBac was reported to be sensitive to calcium channel blockers and completely

insensitive to the sodium channel blocker TTX [27]. A high sensitivity to

1,4-dihydropyridines (DHPs) resembles the pharmacological sensitivity of the long-

lasting (L-type) calcium channels. DHPs have a high affinity for the inactivated state of

the channel [155].

After vesicle fusion occurred six channels were detected in the lipid bilayer in this

specific measurement. Thereafter the transmembrane current was monitored at 80 mV

before and 2 minutes after the addition of the potent blocker nimodipine (Fig. 4.12).

Fig. 4.11 | Channel currents at positive and negative potentials

A holding potential of 80 mV indicates two channels that open and

close. With a potential of -80 mV three channels are detected.


62

The successful blocking of the channel strongly indicates that the monitored

transmembrane current fluctuations originate from opening and closing events of

NaChBac. Experiments with two individual proteoliposome preparations resulted in the

transfer of similar amounts of NaChBac per fusion event and the observed current

jumps were always around 12 pA at a holding potential of 100 mV. This value is

unusually high compared to other conductance values reported for voltage gated sodium

channels, but not beyond the transport efficiency of selective channels. For example the

unit conductance of a calcium-activated potassium channel measured in the oocyte was

in the range of 123 pS [156]. Interestingly the same protein reconstituted into a lipid

bilayer was found to exhibit a single channel conductance of 262 pS [157]. The authors

attributed the higher conductance to an increased ion concentration and the use of

charged phospholipids. The lipid composition and pH can influence the surface charge

which in turn affects the concentration of the permeant ions near the channel and

consequently single channel conductance [158]. Another possible explanation could be

that the selectivity filter built up from the four monomers is not correctly assembled in

the liposomes during the reconstitution from detergent micelles. If the selectivity would

be reduced, different types of ions could pass the protein unhindered which would result

in a larger current. This hypothesis could be tested by measuring the reversed potential

under asymmetric sodium conditions or by comparing the single channel current

generated in the buffer with potassium instead of sodium.

Fig. 4.13 | Sensitivity of )aChBac to nimodipine

At a holding potential of 80 mV channels close slowly. After the addition of

nimodipine (see chemical structure inserted) to a final concentration of

15 µM most of the channels get quickly blocked.


63

4.5 Measuring )aChBac: concluding remarks and perspectives

The nystatin ergosterol method to incorporate membrane proteins into a planar lipid

bilayer was successfully applied. With this method the purified bacterial voltage gated

sodium channel NaChBac was measured for the first time in an artificial lipid bilayer.

Fusion events can be observed while avoiding a permanent background signal which is

an important requirement to evaluate the functional state of the protein of interest. The

number of protein incorporated into the bilayer can be controlled to a certain degree via

the lipid to protein ratio of the proteoliposomes, the amount of vesicles that is added and

the stirring conditions. But a high channel density is difficult to reach since the

ergosterol content of the bilayer increases upon continuous fusion and will eventually

prevent nystatin channel decay. Thus, single channel measurements have been clearly

demonstrated but there is little prospect to investigate the combined function of many

transporters incorporated into bilayers using this method. For this purpose the fusion

rate needs to be increased and at the same time fusion to the same chip pore would need

to be prevented to avoid ergosterol accumulation. Another disadvantage is that the lipid

mixture is preset to induce vesicle fusion but these lipids might not be the appropriate

environment for many membrane proteins. The initial objective to establish a method

which allows the formation of bilayers with integrated proteins on nanoporous chips is

thus only partially achieved. Fusion was only observed if the bilayer area exceeded a

certain threshold that is located somewhere between 8 µm2 and 32 µm2. A compromise

has to be accepted between the chance of liposome fusion in larger pores and sufficient

bilayer stability in smaller pores. For this method the advantage of the silicon nitride

support is lost and the drawbacks such as the high chip capacitance and manufacturing

costs dominate. Probably a surface chemistry that allows proper lipid movement can

improve the fusion efficiency on the chip and allows the transfer of channels to bilayers

formed in smaller pores. If the chip capacitance can be reduced as well, research on ion

channels is possible with the nystatin ergosterol method using a well defined support

that allows easy and stable bilayer formation. The optimal solution to form protein

containing lipid bilayers on a nanopore chip is seen in a direct fusion of

proteoliposomes to the support.

5. Conclusions and outlook

64

5 Conclusions and outlook

The use of nanopore arrays in a silicon nitride support offers several advantages

compared to other systems aiming to investigate functional properties of membrane

proteins. The aspired basic characteristics as described in chapter 2 have been achieved:

(1) The well defined support allows the easy formation of (2) free-standing and (3)

stable lipid bilayers using (4) different lipids or mixtures and offers (5) a large total

bilayer area. Using the toxins melittin and α-HLY we showed that the achieved sealing

is high enough to measure single channel currents and that lipid bilayers are formed in

the majority of the chip pores. The separated bilayers can collapse individually and the

resulting increase in the conductance can clearly be discriminated from protein induced

current steps. We observed that heptamer formation of α-HLY on chips with arrays of

pores was very slow which we explain with the relatively small area of the individual

chip bilayers compared to other studies. The achieved protein density in the BLM was

high enough (> 5 estimated protein pores per µm2) to measure quantitatively the

diffusion of sodium ions within 2 hours. And the feasibility of the incorporation of more

complex membrane proteins was demonstrated using NaChBac and the nystatin

ergosterol method.

With the objective to integrate this support into an automated device for functional

studies of diverse membrane proteins, additional improvements need to be achieved. In

the present state the chip is not suitable to measure transmembrane currents below a

resolution of 2 pA and with a high temporal resolution. A strong reduction of the chip

capacitance is required which might, for example, be achieved by passivating the back

side of the chip. The initial choice to use a silicon-based material to investigate the

effect of the chip pore size on the lipid bilayer stability was based on the established

processing methods. They guarantee the production of a well defined and reproducible

device. If by further modifications [159] the capacitance cannot be reduced sufficiently,

the use of a new material with a low dielectric constant could also be considered.

100 µm pores were successfully laser-drilled into PMMA [64] and PTFE [160], or

formed through gold/SU8 layers [161]. Smaller apertures in the sub-micrometer range

were fabricated in thinned glass substrate [109,162] and polymer films [163]. For

examining the lipid bilayer permeation of drug candidates or measuring the protein-

mediated transport of uncharged solutes, the nanoporous chip is potentially an excellent

support that permits long term measurements thanks to exceptional bilayer stabilities. In


65

the latter case, however, better methods need to be found and developed that allow the

formation of bilayers containing the proteins of interest at high density together with a

high membrane resistance.

Other crucial points with regard to automation and high throughput screening are

automatic bilayer formation and individual addressable recording areas. This involves

the embedding into a microfluidic system with integrated electrodes at best. Hansen et

al. reported the reproducible formation of vertical lipid bilayers in 300 µm aperture

arrays by an automated Müller – Montal method with a success rate above 95 % [164].

Suzuki et al. demonstrated the recording of gramicidin channels in 44 out of 96 wells

after the formation of bilayers by passing organic solvent containing lipids across

horizontal apertures [165]. Their setup involved a relatively simple microfluidic system

at the expense of proper bilayer formation. Previously they reported a bilayer formation

success rate of 90 % and a low specific capacitance of 0.17 µF cm-2 [166]. Most

probably locally different conditions in different recording areas result in improper

thinning and remaining lipid plugs that cause a low capacitance and lack ion channel

recording. A different strategy is followed by Zagnoni et al. who form horizontal BLMs

by the air exposure technique [64]. An initially thick lipid plug is forced to thin while

exposed to air, addition of buffer results then in a stable lipid bilayer. They developed a

platform with 12 separated apertures of 100 µm in diameter that can be addressed

individually via integrated electrodes [160]. The achieved success rate was 50 % and the

authors suggested possible improvements through using a robotic injection approach. It

seems that bilayer thinning in vertical apertures is a critical issue.

Our main goal remains to create an efficient device for measuring the function of

diverse membrane proteins which includes the development of a robust and reliable

method to form protein containing lipid bilayers. Ideally, this will comprises (1)

automated bilayer formation, (2) long-term stability of the bilayer membrane, (3) easy

controlling of the protein amount in the bilayer and (4) a large total bilayer area. As

shown in my thesis, a large total bilayer area together with an excellent stability is

achieved by forming bilayers in arrays of small pores. Using the nystatin ergosterol

method to introduce membrane proteins has permitted to control the amount of

introduced protein to a certain degree via the stirring speed and the proteoliposome

composition. However, a compromise between lipid bilayer stability and vesicle fusion

efficiency had to be accepted. In addition, the protein density needed to measure slow

transport can hardly be obtained with this method. An alternative to delivering


66

membrane proteins to preformed lipid bilayers is the direct formation of protein

containing bilayers on the support. This is in principle possible with the M-M method,

but its applicability is limited to rather robust membrane proteins. An alternative is the

direct adsorption of proteoliposomes to the support. This method would comply with

the stated requirements but so far no high sealing has been reported with this approach,

even after extensive surface modification that should guarantee high bilayer coverage.

Therefore, major efforts in research and development are still needed to reach the goal.

6. Materials and methods

67


All lipids were obtained from Avanti polar lipids (USA). Melittin (prod. nr. 63650),

nystatin (prod. nr. 74721) ergosterol (prod. nr. 45480) n-decane (≥ 99.8%) were

obtained from Fluka (Switzerland). α-HLY (prod. nr. H9395) was from Sigma.

6.1 Molecular biology methods

6.1.1 )aChBac constructs

Construct Number of amino acids Molecular weight (Da)

I9.1N 294 33602 I9.1C 280 32262 I9.1NC 266 30509

6.1.2 Restriction enzyme digest

Digest was performed at 37°C in a 20 µL scale for 1 hour. 20 U of the restriction

enzymes NdeI and XhoI obtained from /ew England Biolabs were used in buffer 4

provided by the supplier.

6.1.3 Agarose gel electrophoresis

0.4 g Agarose was solubilized in 40 mL boiling TBE buffer (89 mM Tris, 89 mM boric

acid, 2 mM EDTA) to produce a 1 % Agarose gel. After the addition of 0.01 % ethidium

bromide (4 µL) the gel was run in TBE buffer at a constant voltage of 100 V.

6.1.4 Transformation of competent cells

Cells were transformed by the heat shock method. 100 µL cells were mixed with 10 ng

plasmid DNA and incubated on ice for 20 minutes. After 2 minutes at 42°C the cells

were placed back on ice for 3 minutes. After the addition of 400 µL 2xTY medium,

incubation at 37°C for 45 minutes preceded the plating on LB-agar plates containing


68

ampicillin. After incubation over night at 37°C the grown colonies were used to

inoculate 2 mL 2xTY medium or 50 mL LB medium for recombinant protein

production or plasmid preparation respectively.

6.1.5 Plasmid preparation

Plasmid preparation from 50 mL overnight cultures was performed with the GenEluteTM

Plasmid Midiprep Kit from Sigma.

6.2 Microbiology methods

6.2.1 Expression in E. Coli

All chemicals used for culture media including antibiotics were obtained from Gerbu.

General media LB medium: 10 g l-1 tryptone 10 g l-1 NaCl 5 g l-1 yeast extract 2xTY medium: 16 g l-1 tryptone 5 g l-1 NaCl 10 g l-1 yeast extract

The auto-induction medium ZYP-5052 was prepared freshly. Prior to mixing ZY, NPS

and 5052, 1 mM MgSO4 was added to ZY to prevent precipitation.

ZYP-5052 ZY: 10 g l-1 tryptone 5 g l-1 yeast extract NPS: 50 mM Na2HPO4 50 mM KH2PO4 25 mM (NH4)2SO4 5052: 0.5 % (v/v) glycerol 0.2 % (w/v) α-lactose 0.05 % (w/v) glucose

12 times 1 liter of ZYP-5052 containing 100 µg mL-1 ampicillin was inoculated with

15 mL from the over night culture. Cells were grown for 1 day at 30°C.

6.2.2 Harvesting and disruption

Cells were harvested by centrifugation at 5000 rpm in a RC-3B Plus (Sorvall) centrifuge

at 4°C for 20 minutes. The cell pellet was resuspended in lysis buffer containing

5 µg mL-1 DnaseI (Fermentas) and 1 tablet of the complete EDTA free protease

inhibitor cocktail (Roche). Cells were disrupted by passing them 3 times through an


69

EmulsiFlex-C3 (Avestin) homogeniser. Unlysed cells were removed by centrifugation

at 6000 rpm using an A8.24 rotor in a Hi-Cen® 21 C centrifuge (Herolab). Total

membrane was isolated by centrifugation of the supernatant at 40’000 rpm using a TI 45

rotor in an OptimaTM L-90 K centrifuge (Beckmann-Coulter). The pellet was stored at -

80°C if processing was not continued.

Lysis buffer pH 8.0: 50 mM NaH2PO4 10 mM imidazole 300 mM NaCl 10 % (v/v) glycerol

6.2.3 Purification

The pellet was resuspended in lysis buffer containing 1.5 % DDM and stirred on ice for

1.5 hours. After centrifugation at 40’000 rpm for 1 hour, the supernatant was collected

and loaded on a Ni2+-loaded 1 mL HiTrapTM Chelating HP column on the

ÄKTAxpressTM system (Amersham Biosciences). Unbound sample was removed by

washing with lysis buffer containing 0.05 % DDM. Elution was performed with lysis

buffer containing 300 mM imidazole and 0.03 % DDM. The elution peak was collected

in a loop and injected on a 120 mL HiLoadTM Superdex 75 (Amersham Biosciences)

column. With a flow rate of 0.8 mL min-1 the protein was eluted using lysis buffer

without imidazole but containing 0.03 % DDM.

6.3 Protein analysis and characterization

6.3.1 Protein concentration determination

The protein concentration was determined via UV absorption at 280 nm using a

/anoDrop® /D-1000 Spectrophotometer (NanoDrop Technologies). The extinction

coefficient and molecular weight of the protein were calculated from their amino acid

sequence using the ProtParam tool3.

6.3.2 SDS-PAGE

Protein samples were analyzed using 12 % separating gel with a 4 % stacking gel. They

were poured in a Mighty Small Multiple Gel Caster (Hoefer) and were run in a SE 300

MiniVE unit (Hoefer) at a constant voltage of 200 V.

Coomassie staining: SED gels were stained with Coomassie blue R-250 staining

solution and destained in ddH2O containing acetic acid and ethanol.

3 http://www.expasy.ch/tools/protparam.html


70

Western blot: Proteins were transferred to an Immobilon-P PVDF membrane

(Millipore) using a SemiPhor T70 Semi-Dry Transfer Unit (Hoefer) at a constant current

of 0.8 mA cm-2 membrane. After washing and blocking, the membrane was incubated

for 30 minutes with 1:5000 of the penta-His HRP conjugate (Quiagen) antibody. An

image of the blot was captured using the ECLTM Western blotting detection reagents and

an Amersham Hyperfilm ECL (GE Healthcare) on a Curix 60 (AGFA).

6.3.3 Static light-scattering

The analysis of the oligomeric state of α-HLY in solution was performed by static light-

scattering. 0.12 mg protein in 120 µL buffer (20 mM NaAc, 150 mM KCl, pH 5.5) was

loaded on a SuperdexTM 200 10/300 gel filtration column (Amersham Biosciences) on

an Agilent 1100 Series HPLC system (Agilent Technologies). The mass of the protein

and monodispersity were measured using the miniDAWNTM light scattering and an

Optilab® refractometer and analyzed using the Astra software version 4.90.08 (Wyatt).

6.4 (Proteo-) liposome formation

6.4.1 Reconstitution

The lipids dissolved in chloroform were dried under a stream of nitrogen. To 10 mg

lipids 1 mL buffer was added and lipids were allowed to hydrate for 1 hour at room

temperature prior to extrusion.

In the case of proteoliposome formation the protein was first added to buffer containing

detergent and then to the dried lipid.

Method % detergent in buffer detergent removal Dialysis4 0.03 2x against buffer 1:100 (v/v) 10 h each BioBeads 1.5 - 3 2x 0.75g BioBeads for 5 h each

6.4.2 Extrusion

The lipid suspension with multilamellar vesicles was passed through a polycarbonate

membrane with a defined pore size (Whatman) using a LipexTM extruder (Northern

Lipids). Starting with a membrane containing 800 nm pores, the filter was exchanged

after 5 passages and replaced by a membrane with smaller pores after the 10th passage.

To form 50 nm liposomes the used filters were 800, 600, 400, 200, 100 and 50 nm.

4 Spectra/por molecularporous membrane MWCO 12-14 kDa


71

6.4.3 Protein / lipid ratio determination

Phosphate test

The lipid concentration was measured via its phosphorus. In the presence of magnesium

nitrate the phospholipid phosphorus was converted to inorganic phosphate at high

temperatures. The addition of hydrochloric acid, ascorbic acid and ammonium

molybdate generated a phosphomolybdate complex. Absorption at 750 nm was

correlated to a standard curve generated from KH2PO4 to determine the concentration:

Add 0.03 mL (I) to the samples in borosilicate glass vials and incubate for 10 minutes at

250°C. After vials cooled down add 0.3 mL (2), vortex and leave for 10 minutes at

95°C. Add 0.7 mL of a mix (one part (III) and 6 parts (IV)) and incubate for 20 minutes

at 37°C prior to analysis on the spectrophotometer.

(I) 10 % Mg(NO3)2 in 95% EtOH (II) 0.5 N HCl (III) 10 % Ascorbic acid (IV) 0.42 % (NH4)6Mo7O24 in 1 N H2SO4

Lowry protein determination

Protein incorporated in liposomes was measured using a modified Lowry method. It

involves precipitate formation with sodium deoxycholate and trichloroacetic acid. After

resuspension in SDS under alkaline conditions the addition of copper(II)sulphate and

folin reagent results in a reaction with the protein nitrogens and the subsequent

reduction of the folin reagent. Absorption at 750 nm was correlated to a standard curve

generated from BSA to determine the concentration:

To 1 mL ddH2O in eppendorf tubes add liposomes containing approximately 10-100 µg

protein. Add 0.1 mL 0.15 % sodium deoxycholate and wait 10 minutes before the

addition of 0.1 mL 72 % CCl3COOH. After incubation for 1 hour at room temperature

centrifuge samples for 20 minutes and remove supernatant. After drying, the samples

are resuspended in 0.5 mL ddH2O and incubated for 10 minutes with 0.5 mL (I). After

the addition of 0.25 mL (II) and waiting for at least 30 minutes the samples can be

measured on the spectrophotometer.

(I) mix equal volumes of 0.1 % CuSO4, 10 % Na2CO3, 0.2% Na2C4H4O6 10 % SDS 0.8 N NaOH ddH2O (II) mix 1 volume 2 N folin reagent 5 volumes ddH2O


72

6.4.4 Dynamic light scattering

Liposome size distributions were measured either on the DynaProTM temperature

controlled micro sampler or on the DynaProTM plate reader (Wyatt) and analyzed using

the DY/AMICS software. Some measurements were also performed on a /icomb

particle size analyzer.

6.5 Chip characterization

6.5.1 Light microscope

Bright field images of the chip were recorded using an Axiophot microscope (Zeiss) and

a 10x objective.

6.5.2 SEM

Inspection was performed using a Supra 55 VP field emission scanning microscope

(Zeiss).

6.5.3 Contact angle measurements

Advancing contact angles of distilled water were measured using a goniometer system

G2 (Krüss GmbH)

6.6 Lipid bilayer formation

6.6.1 Painting

Painted lipid bilayers were formed from 10 mg mL-1 lipid in decane solutions.


The chips were pretreated with a pentane/hexadecane solution (9:1 v/v) before being

sandwiched between the two PMMA compartments. A small volume (5 µL) of lipid

dissolved in pentane (10 mg mL-1) was added to the buffer in both compartments.

6.6.3 )ystatin-Ergosterol vesicle fusion

All experiments were performed in 20 mM NaCH3CO2 with 150 mM NaCl at pH 5.5. A

lipid bilayer was painted from a POPC/POPE (7:3 w/w) mixture in decane

(10 mg mL-1). (Proteo-)liposomes of the composition POPC/POPE/ergosterol

(4/4/2 molar ratio) in 20 mM NaCH3CO2, 150 mM NaCl and 0.5 M urea at pH 5.5 were

extrude to a defined size. 100 µL of a 10 mg mL-1 (proteo-)liposome solution were


73

activated with 10 µL nystatin dissolved in dry methanol (3 mg mL-1) and incubated for

1 hour. 20 µL activated vesicles were added to one side per mL buffer and urea was

added to the same side to a final concentration of 0.5 M to induce fusion.

6.7 Electrochemical methods

6.7.1 EIS

Electrochemical impedance spectroscopy was performed using an Autolab PGSTAT12

instrument (EcoChemie) equipped with a FRA-module. The spectra were recorded from

1 MHz to 0.01 Hz at 0 V potential applying the 10 mV amplitude between two platinum

wires (diameter 1 mm). The spectra were analysed using FRA software.

6.7.2 )ormal pulse voltammetry

The current-voltage characteristics of a melittin treated bilayer were recorded by

scanning the potential range from 0.15 V to -0.15 V using the Autolab. Starting from a

0 V base potential voltage pulses were applied for 70 ms and collected samples from the

last 20 ms of the pulse were averaged. The voltage is defined as positive to the trans

side, the current measures positive charges translocated from the cis- to the trans side.

6.7.3 Linear sweep voltammetry

The breakdown of lipid bilayers was measured using the Autolab starting at 0 V and

increasing the potential in steps of 0.45 mV to 1 V with a scan rate of 1 mV s-1.

6.7.4 Chronoamperometry

Current flow through the bilayer was recorded over time using the Autolab. Sampling

time was limited by the amounts of recorded points and varied from 0.1s to 3s

depending on the measurement.

6.7.5 Four-electrode setup

The potential difference across the bilayer was maintained using two Hg/HgO XR400

reference electrodes (IG, Switzerland). They were positioned close to the membrane

with two ambient platinum wires as working and counter electrode.

6.8 Diffusion

Diffusion was monitored using ion selective electrodes connected to the pX-module of

the Autolab instrument.


74

Electrode Type Manufacturer Type NH3 Microelectrodes (Bedford, USA) MI-740 combination Na+ Sartorius (Germany) PY-IO3 combination Omega/Newport (Germany) ISE-8765 combination Mettler Toledo (Switzerland) DX-233 ISE InLab301 reference Microelectrodes (Bedford, USA) MI-425 combination Ca2+ Microelectrodes (Bedford, USA) MI-600 ISE MI-409F Reference

7. References

75

7. References

[1] Gorter, E. and Grendel, F. 1925: On bimolecular layers of lipoids on the chromocytes of the blood. J Exp Med, 41:439-443

[2] Singer, S.J. and Nicolson, G.L. 1972: The fluid mosaic model of the structure of cell membranes. Science, 175:720-731

[3] Engelman, D.M. 2005: Memranes are more mosaic than fluid. /ature, 438:578-578

[4] Mouritsen, O.G. and Bloom, M. 1984: Mattress model of lipid-protein interactions in membranes. Biophys J, 46:141-153

[5] Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., and Walter, P. in: "Molecular Biochemistry of The Cell", ch. 10: Membrane Structure, 4th ed. (Ed.: Gibbs, S. and Jenner, K.), Garland Science, 2002

[6] Sahu, S.K., Gummadi, S.N., Manoj, N., and Aradhyam, G.K. 2007: Phospholipid scramblases: An overview. Arch Biochem Biophys, 462:103-114

[7] Brouckaert, G., Kalai, M., Krysko, D.V., Saelens, X., Vercammen, D., Ndlovu, M., Haegeman, G., D'Herde, K., and Vandenabeele, P. 2004: Phagocytosis of necrotic cells by macrophages is phosphatidylserine dependent and does not induce inflammatory cytokine production. Mol Biol Cell, 15:1089-1100

[8] MacLennan, D.H., Ostwald, T.J., and Stewart, P.S. 1974: Structural components of the sarcoplasmic reticulum membrane. Ann / Y Acad Sci, 227:527-536

[9] Chao, Y. and Fu, D. 2004: Kinetic study of the antiport mechanism of an Escherichia coli Zinc transporter, ZitB. J Biol Chem, 279:12043-12050

[10] Javelle, A., Lupo, D., Ripoche, P., Fulford, T., Merrick, M., and Winkler, F.K. 2008: Substrate binding, deprotonation, and selectivity at the periplasmic entrance of the Escherichia coli ammonia channel AmtB. Proc /atl Acad Sci USA, 105:5040-5045

[11] Gerber, S., Comellas-Bigler, M., Goetz, B.A., and Locher, K.P. 2008: Structural basis of trans-inhibition in a Molybdate / Tungstate ABC transporter. Science, 321:246-250

[12] Berman, M.C. 2001: Slippage and uncoupling in P-type cation pumps; implications for energy transduction mechanisms and regulation of metabolism. Biochim Biophys Acta, 1513:95-121

[13] Neher, E. and Sakmann, B. 1976: Single-channel currents recorded from membrane of denervated frog muscle fibres. /ature, 260:799-802

[14] Riquelme, G., Lopez, E., Garcia-Segura, L.M., Ferragut, J.A., and Gonzalez-Ros, J.M. 1990: Giant liposomes: a model system in which to obtain patch-clamp recordings of ionic channels. Biochemistry, 29:11215-11222

[15] Terpitz, U., Raimunda, D., Westhoff, M., Sukhorukov, V.L., Beaugθ, L., Bamberg, E., and Zimmermann, D. 2008: Electrofused giant protoplasts of Saccharomyces cerevisiae as a novel system for electrophysiological studies on membrane proteins. Biochim Biophys Acta, 1778:1493-1500

[16] Bertl, A., Bihler, H., Kettner, C., and Slayman, C.L. 1998: Electrophysiology in the eukaryotic model cell Saccharomyces cerevisiae. Pflugers Arch - Eur J Physiol, 436:999-1013

[17] Jiang, B., Sun, X., Cao, K., and Wang, R. 2002: Endogenous KV channels in human embryonic kidney (HEK-293) cells. Mol Cell Biol, 238:69-79

7. References

76

[18] Avila, G., Sandoval, A., and Felix, R. 2004: Intramembrane charge movement associated with endogenous K+ channel activity in HEK-293 cells. Cell Mol /eurobiol, 24:317-330

[19] Kurejova, M., Uhrik, B., Sulova, Z., Sedlakova , B., Krizanova, O., and Lacinova, L. 2007: Changes in ultrastructure and endogenous ionic channels activity during culture of HEK 293 cell line. Eur J Pharmacol, 567:10-18

[20] Gamper, N., Stockand, J.D., and Shapiro, M.S. 2005: The use of Chinese hamster ovary (CHO) cells in the study of ion channels. J Pharmacol Toxicol Methods, 51:177-185

[21] Skryma, R., Prevarskaya, N., Vacher, P., and Dufy, B. 1994: Voltage-dependent ionic conductances in Chinese hamster ovary cells. Am J Physiol Cell Physiol, 267:C544-C553

[22] Tien, H.T. and Ottova, A.L. 2001: The lipid bilayer concept and its experimental realization: from soap bubbles, kitchen sink, to bilayer lipid membranes. J Memb Sci, 189:83-117

[23] Müller, P., Rudin, D.O., Ti Tien, H., and Wescott, W. 1962: Reconstitution of cell membrane structure in vitro and its transformation into an excitable system. /ature, 194:979-980

[24] Hladky, S.B. and Haydon, D.A. 1972: Ion transfer across lipid membranes in the presence of gramicidin A: I. Studies of the unit conductance channel. Biochim Biophys Acta, 274:294-312

[25] Mueller, P. and Rudin, D.O. 1968: Action potentials induced in biomolecular lipid membranes. /ature, 217:713-719

[26] Tombola, F., Carlesso, C., Szabo, I., de Bernard, M., Reyrat, J.M., Telford, J.L., Rappuoli, R., Montecucco, C., Papini, E., and Zoratti, M. 1999: Helicobacter pylori vacuolating toxin forms anion-selective channels in planar lipid bilayers: Possible implications for the mechanism of cellular vacuolation. Biophys J, 76:1401-1409

[27] Ren, D., Navarro, B., Xu, H., Yue, L., Shi, Q., and Clapham, D.E. 2001: A prokaryotic voltage-gated sodium channel. Science, 294:2372-2375

[28] Murayama, T., Oba, T., Katayama, E., Oyamada, H., Oguchi, K., Kobayashi, M., Otsuka, K., and Ogawa, Y. 1999: Further characterization of the type 3 ryanodine receptor (RyR3) purified from rabbit diaphragm. J Biol Chem, 274:17297-17308

[29] Bubien, J.K., Cornwell, T., Bradford, A.L., Fuller, C.M., DuVall, M.D., and Benos, D.J. 1998: alpha-Adrenergic receptors regulate human lymphocyte amiloride-sensitive sodium channels. Am J Physiol Cell Physiol, 275:C702-C710

[30] Lundbaek, J.A. 2006: Regulation of membrane protein function by lipid bilayer elasticity - a single molecule technology to measure the bilayer properties experienced by an embedded protein. J Phys: Condens Matter, 18:S1305-S1344

[31] Reimhult, E. and Kumar, K. 2008: Membrane biosensor platforms using nano- and microporous supports. Trends Biotechnol, 26:82-89

[32] Papahadjopoulos, D., Cowden, M., and Kimelberg, H. 1973: Role of cholesterol in membranes effects on phospholipid-protein interactions, membrane permeability and enzymatic activity. Biochim Biophys Acta, 330:8-26

7. References

77

[33] Moellerfeld, J., Prass, W., Ringsdorf, H., Hamazaki, H., and Sunamoto, J. 1986: Improved stability of black lipid membranes by coating with polysaccharide derivatives bearing hydrophobic anchor groups. Biochim Biophys Acta, 857:265-270

[34] Ash, P.S., Bunce, A.S., Dawson, C.R., and Hider, R.C. 1978: The effect of synthetic polymers on the electrical and permeability properties of lipid membranes. Biochim Biophys Acta, 510:216-229

[35] Benz, R., Prass, W., and Ringsdorf, H. 1982: Black lipid membranes from polymerizable lipids. Angew Chem Int Ed, 21:368-369

[36] Redwood, W.R., Pfeiffer, F.R., Weisbach, J.A., and Thompson, T.E. 1971: Physical properties of bilayer membranes formed from a synthetic saturated phospholipid in n-decane. Biochim Biophys Acta, 233:1-6

[37] Jeon, T.J., Malmstadt, N., and Schmidt, J.J. 2006: Hydrogel-encapsulated lipid membranes. J Am Chem Soc, 128:42-43

[38] Janshoff, A. and Steinem, C. 2006: Transport across artificial membranes - an analytical perspective. Anal Bioanal Chem, 385:433-451

[39] Reviakine, I., Simon, A., and Brisson, A. 2000: Effect of Ca2+ on the morphology of mixed DPPC / DOPS supported phospholipid bilayers. Langmuir, 16:1473-1477

[40] Richter, R.P., Lai Kee Him, J., Tessier, B., Tessier, C., and Brisson, A. 2005: On the kinetics of adsorption and two-dimensional self-assembly of annexin A5 on supported lipid bilayers. Biophys J, 89:3372-3385

[41] Glasmästar, K., Larsson, C., Höök, F., and Kasemo, B. 2002: Protein adsorption on supported phospholipid bilayers. J Colloid Interface Sci, 246:40-47

[42] Fix, M., Melia, T.J., Jaiswal, J.K., Rappoport, J.Z., You, D., Söllner, T.H., Rothman, J.E., and Simon, S.M. 2004: Imaging single membrane fusion events mediated by SNARE proteins. Proc /atl Acad Sci USA, 101:7311-7316

[43] Sackmann, E. and Tanaka, M. 2000: Supported membranes on soft polymer cushions: fabrication, characterization and applications. Trends Biotechnol, 18:58-64

[44] Elender, G., Kühner, M., and Sackmann, E. 1996: Functionalisation of Si/SiO2 and glass surfaces with ultrathin dextran films and deposition of lipid bilayers. Biosens Bioelectron, 11:565-577

[45] Deme, B. and Marchal, D. 2005: Polymer-cushioned lipid bilayers in porous alumina. Europ Biophys J, 34:170-179

[46] Hwang, L.Y., Götz, H., Knoll, W., Hawker, C.J., and Frank, C.W. 2008: Preparation and characterization of glycoacrylate-based polymer-tethered lipid bilayers on benzophenone-modified substrates. Langmuir, 24:14088-14098

[47] Naumann, R., Baumgart, T., Gräber, P., Jonczyk, A., Offenhässer, A., and Knoll, W. 2002: Proton transport through a peptide-tethered bilayer lipid membrane by the H+-ATP synthase from chloroplasts measured by impedance spectroscopy. Biosens Bioelectron, 17:25-34

[48] Becucci, L., Guidelli, R., Peggion, C., Toniolo, C., and Moncelli, M.R. 2005: Incorporation of channel-forming peptides in a Hg-supported lipid bilayer. J Electroanal Chem, 576:121-128

[49] Jeuken, L.J.C. 2008: AFM study on the electric-field effects on supported bilayer lipid membranes. Biophys J, 94:4711-4717

7. References

78

[50] Terrettaz, S., Mayer, M., and Vogel, H. 2003: Highly electrically insulating tethered lipid bilayers for probing the function of ion channel proteins. Langmuir, 19:5567-5569

[51] Giess, F., Knoll, W., Naumann, R., Heberle, J., and Friedrich, M.G. 2004: The protein-tethered lipid bilayer: a novel mimic of the biological membrane. Biophys J, 87:3213-3220

[52] Koper, I. 2007: Insulating tethered bilayer lipid membranes to study membrane proteins. Mol BioSyst, 3:651-657

[53] Richter, R.P., Berat, R., and Brisson, A.R. 2006: Formation of solid-supported lipid bilayers: An integrated view. Langmuir, 22:3497-3505

[54] White, S.H. 2004: The progress of membrane protein structure determination. Protein Sci, 13:1948-1949

[55] Heyderman, L.J., Ketterer, B., Bächle, D., Glaus, F., Haas, B., Schift, H., Vogelsang, K., Gobrecht, J., Tiefenauer, L., Dubochet, O., Surbled, P., and Hessler, T. 2003: High volume fabrication of customized nanopore membrane chips. Microelectron Eng, 67-68:208-213

[56] Han, X., Studer, A., Sehr, H., Geissbühler, I., Di Berardino, M., Winkler, F.K., and Tiefenauer, L.X. 2007: Nanopore arrays for stable and functional free-standing lipid bilayers. Adv Mater, 19:4466-4470

[57] Schift, H., Saxer, S., Park, S., Padeste, C., Pieles, U., and Gobrecht, J. 2005: Controlled co-evaporation of silanes for nanoimprint stamps. /anotechnology, 16:171-175

[58] Wilk, S.J., Goryll, M., Laws, G.M., Goodnick, S.M., Thornton, T.J., Saraniti, M., Tang, J., and Eisenberg, R.S. 2004: Teflon-coated silicon apertures for supported lipid bilayer membranes. Appl Phys Lett, 85:3307-3309

[59] Huang, C., Wheeldon, L., and Thompson, T.E. 1964: The properties of lipid bilayer membranes seperating two aqueous phases: Formation of a membrane of simple composition. J Mol Biol, 8:148-160

[60] Suzuki, H., Kato-Yamada, Y., Noji, H., and Takeuchi, S. 2004: Planar lipid membrane array for membrane protein chip. Micro Electro Mechanical Systems, 2004 17th IEEE International Conference on (MEMS)272-275

[61] Tien, H.T. and Dawidowicz, E.A. 1966: Black lipid films in aqueous media: A new type of interfacial phenomenon. J Colloid Interface Sci, 22:438-453

[62] Tien, H.T. and Dawidowicz, E.A. 1966: Black lipid films in aqueous media: A new type of interfacial phenomenon : Experimental techniques and thickness measurements. J Colloid Interface Sci, 22:438-453

[63] Tien, H.T. 1968: Black lipid membranes at bifaces: Formation characteristics, optical and some thermodynamic properties. J Gen Physiol, 52:125-144

[64] Sandison, M.E., Zagnoni, M., and Morgan, H. 2007: Air-exposure technique for the formation of artificial lipid bilayers in microsystems. Langmuir, 23:8277-8284

[65] Suzuki, H., Tabata, K.V., Noji, H., and Takeuchi, S. 2007: Electrophysiological recordings of single ion channels in planar lipid bilayers using a polymethyl methacrylate microfluidic chip. Biosens Bioelectron, 22:1111-1115

[66] Wilburn, J.P., Wright, D.W., and Cliffel, D.E. 2006: Imaging of voltage-gated alamethicin pores in a reconstituted bilayer lipid membrane via scanning electrochemical microscopy. The Analyst, 131:311-316

7. References

79

[67] Malmstadt, N., Jeon, T.J., and Schmidt, J.J. 2008: Long-lived planar lipid bilayer membranes anchored to an in situ polymerized hydrogel. Adv Mater, 20:84-89

[68] White, H.S. 1972: Analysis of the torus surrounding planar lipid bilayer membranes. Biophys J, 12:432-445

[69] White, S.H., Petersen, D.C., Sidney, S., and Yafuso, M. 1976: Formation of planar bilayer membranes from lipid monolayers, a critique. Biophys J, 16:481-489

[70] Montal, M. and Mueller, P. 1972: Formation of bimolecular membranes from lipid monolayers and a study of their electrical properties. Proc /atl Acad Sci USA, 69:3561-3566

[71] Simon, S.A., Lis, L.J., MacDonald, R.C., and Kauffman, J.W. 1977: The noneffect of a large linear hydrocarbon, squalene, on the phosphatidylcholine packing structure. Biophys J, 19:83-90

[72] Benz, R., Fröhlich, O., Läuger, P., and Montal, M. 1975: Electrical capacity of black lipid films and of lipid bilayers made from monolayers. Biochim Biophys Acta, 394:323-334

[73] Hamai, C., Yang, T., Kataoka, S., Cremer, P.S., and Musser, S.M. 2006: Effect of average phospholipid curvature on supported bilayer formation on glass by vesicle fusion. Biophys J, 90:1241-1248

[74] Kim, Y.H., Rahman, M., Zhang, Z.L., Misawa, N., Tero, R., and Urisu, T. 2006: Supported lipid bilayer formation by the giant vesicle fusion induced by vesicle-surface electrostatic attractive interaction. Chem Phys Lett, 420:569-573

[75] Puu, G. and Gustafson, I. 1997: Planar lipid bilayers on solid supports from liposomes - factors of importance for kinetics and stability. Biochim Biophys Acta, 1327:149-161

[76] Schmidt, E.K., Nurnabi, M., Bushby, R.J., and Steinem, C. 2008: Electrically insulating pore-suspending membranes on highly ordered porous alumina obtained from vesicle spreading. Soft matter, 4:250-253

[77] Hennesthal, C. 2002: Membrane-suspended nanocompartments based on ordered pores in alumina. ChemPhysChem, 3:885-889

[78] Moulton, S.E., Barisci, J.N., Bath, A., Stella, R., and Wallace, G.G. 2004: Studies of double layer capacitance and electron transfer at a gold electrode exposed to protein solutions. Electrochim Acta, 49:4223-4230

[79] Franks, W., Schenker, I., Schmutz, P., and Hierlemann, A. 2005: Impedance characterization and modeling of electrodes for biomedical applications. IEEE Trans Biomed Eng, 52:1295-1302

[80] Kalinowski, S. and Figaszewski, Z. 1995: A four-electrode potentiostat-galvanostat for studies of bilayer lipid membranes. Meas Sci Technol, 6:1050-1055

[81] Fettiplace, R., Andrews, D.M., and Haydon, D.A. 1971: The thickness, composition and structure of some lipid bilayers and natural membranes. J Membr Biol, 5:277-296

[82] Mitra, K., Ubarretxena-Belandia, I., Taguchi, T., Warren, G., and Engelman, D.M. 2004: Modulation of the bilayer thickness of exocytic pathway membranes by membrane proteins rather than cholesterol. Proc /atl Acad Sci USA, 101:4083-4088

7. References

80

[83] Han, A., Schürmann, G., Mondin, G., Bitterli, R.A., Hegelbach, N.G., de Rooij, N.F., and Staufer, U. 2006: Sensing protein molecules using nanofabricated pores. Appl Phys Lett, 88:093901-1-093901-3

[84] Mayer, M., Kriebel, J.K., Tosteson, M.T., and Whitesides, G.M. 2003: Microfabricated teflon membranes for low-noise recordings of ion channels in planar lipid bilayers. Biophys J, 85:2684-2695

[85] Simon, A., Girard-Egrot, A., Sauter, F., Pudda, C., Picollet D'Hahan, N., Blum, L., Chatelain, F., and Fuchs, A. 2007: Formation and stability of a suspended biomimetic lipid bilayer on silicon submicrometer-sized pores. J Colloid Interface Sci, 308:337-343

[86] Meier, W., Graff, A., Diederich, A., and Winterhalter, M. 2000: Stabilization of planar lipid membranes: A stratified layer approach. Phys Chem Chem Phys, 2:4559-4562

[87] Kramar, P., Miklavcic, D., and Lebar, A.M. 2007: Determination of the lipid bilayer breakdown voltage by means of linear rising signal. Bioelectrochemistry, 70:23-27

[88] Weaver, J.C. and Chizmadzhev, Y. 1996: Theory of electroporation: A review. Bioelectrochem Bioenerg, 41:135-160

[89] Tiefenauer, L.X. and Studer, A. 2008: Nano for bio: Nanopore arrays for stable and functional lipid bilayer membranes (Mini Review). Biointerphases, 3:FA74-FA79

[90] James, C.D., Reuel, N., Lee, E.S., Davalos, R.V., Mani, S.S., Carroll-Portillo, A., Rebeil, R., Martino, A., and Apblett, C.A. 2008: Impedimetric and optical interrogation of single cells in a microfluidic device for real-time viability and chemical response assessment. Biosens Bioelectron, 23:845-851

[91] Tosteson, M.T. and Tosteson, D.C. 1981: The sting. Melittin forms channels in lipid bilayers. Biophys J, 36:109-116

[92] Lee, T., Mozsolits, H., and Aquilar, M. 2001: Measurement of the affinity of melittin for zwitterionic and anionic membranes using immobilized lipid biosensors. J Pept Res, 58:464-476

[93] Benachir, T. and Lafleur, M. 1995: Study of vesicle leakage induced by melittin. Biochim Biophys Acta, 1235:452-460

[94] Allende, D., Simon, S.A., and McIntosh, T.J. 2005: Melittin-induced bilayer leakage depends on lipid material properties: evidence for toroidal pores. Biophys J, 88:1828-1837

[95] Yang, L., Harroun, T.A., Weiss, T.M., Ding, L., and Huang, H.W. 2001: Barrel-stave model or toroidal model? A case study on melittin pores. Biophys J, 81:1475-1485

[96] Beitz, E. 2000: TEXtopo: shaded membrane protein topology plots in LATEX. Bioinformatics, 16:1050-1051

[97] Tusnady, G.E. and Simon, I. 2001: The HMMTOP transmembrane topology prediction server. Bioinformatics, 17:849-850

[98] Wilcox, W. and Eisenberg, D. 1992: Thermodynamics of melittin tetramerization determined by circular dichroism and implications for protein folding. Protein Sci, 1:641-653

[99] Dawson, C.R., Drake, A.F., Helliwell, J., and Hider, R.C. 1978: The interaction of bee melittin with lipid bilayer membranes. Biochim Biophys Acta, 510:75-86

7. References

81

[100] Hanke, W., Methfessel, C., Wilmsen, H.U., Katz, E., Jung, G., and Boheim, G. 1983: Melittin and a chemically modified trichotoxin form alamethicin-type multi-state pores. Biochim Biophys Acta, 727:108-114

[101] Pawlak, M., Stankowski, S., and Schwarz, G. 1991: Melittin induced voltage-dependent conductance in DOPC lipid bilayers. Biochim Biophys Acta, 1062:94-102

[102] Kempf, C., Klausner, R.D., Weinstein, J.N., Van Renswoude, J., Pincus, M., and Blumenthal, R. 1982: Voltage-dependent trans-bilayer orientation of melittin. J Biol Chem, 257:2469-2476

[103] Song, L., Hobaugh, M.R., Shustak, C., Cheley, S., Bayley, H., and Gouaux, J.E. 1996: Structure of Staphylococcal alpha - hemolysin, a heptameric transmembrane pore. Science, 274:1859-1865

[104] Hildebrand, A., Pohl, M., and Bhakdi, S. 1991: Staphylococcus aureus alpha-toxin. Dual mechanism of binding to target cells. J Biol Chem, 266:17195-17200

[105] Valeva, A., Hellmann, N., Walev, I., Strand, D., Plate, M., Boukhallouk, F., Brack, A., Hanada, K., Decker, H., and Bhakdi, S. 2006: Evidence that clustered phosphocholine head groups serve as sites for binding and assembly of an oligomeric protein pore. J Biol Chem, 281:26014-26021

[106] Menestrina, G. 1986: Ionic channels formed by Staphylococcus aureus alpha-toxin: Voltage-dependent inhibition by divalent and trivalent cations. J Membr Biol, 90:177-190

[107] Kawate, T. and Gouaux, E. 2003: Arresting and releasing Staphylococcal alpha-hemolysin at intermediate stages of pore formation by engineered disulfide bonds. Protein Sci, 12:997-1006

[108] Hemmler, R., Bose, G., Wagner, R., and Peters, R. 2005: Nanopore unitary permeability measured by electrochemical and optical single transporter recording. Biophys J, 88:4000-4007

[109] White, R.J., Ervin, E.N., Yang, T., Chen, X., Daniel, S., Cremer, P.S., and White, H.S. 2007: Single ion-channel recordings using glass nanopore membranes. J Am Chem Soc, 129:11766-11775

[110] Walker, B., Krishnasastry, M., Zorn, L., and Bayley, H. 1992: Assembly of the oligomeric membrane pore formed by Staphylococcal alpha - hemolysin examined by truncation mutagenesis. J Biol Chem, 267:21782-21786

[111] Cussler E.L. in: "Diffusion: Mass transfer in fluid systems", ch. 2: Diffusion in dilute solutions, 2nd (Ed.: Varma, A), Cambridge University Press, 1997

[112] Vanysek, Petr in: "CRC Handbook of chemistry and physics", ch. 5: Ionic conductivity and diffusion at infinite dilution, 88th ed. (Ed.: Linde, David R.), CRC press, 2007

[113] Ohki, Shinpei and Spangler, Robert A in: "The structure of biological membranes", ch. 10: Passive and facilitated transport, 2nd ed. (Ed.: Yeagle, Philipp L.), CRC press, 2005

[114] Pohl, P., Saparov, S.M., and Antonenko, Y.N. 1998: The size of the unstirred layer as a function of the solute diffusion coefficient. Biophys J, 75:1403-1409

[115] Kane, P. and Diamond, D. 1997: Determination of ion-selective electrode characteristics by non-linear curve fitting. Talanta, 44:1847-1858

[116] Bohrer, M.P. 1983: Diffusional boundary layer resistance for membrane transport. Ind Eng Chem Fund, 22:72-78

7. References

82

[117] Wylie Nichols, J., Hill, M.W., Bangham, A.D., and Deamer, D.W. 1980: Measurement of net proton-hydroxyl permeability of large unilamellar liposomes with the fluorescent pH probe, 9-aminoacridine. Biochim Biophys Acta, 596:393-403

[118] Laclette, J.P. and Montal, M. 1977: Interaction of calcium with negative lipids in planar bilayer membranes. Influence of the solvent. Biophys J, 19:199-202

[119] Bashford, C.L., Alder, G.M., Menestrina, G., Micklem, K.J., Murphy, J.J., and Pasternak, C.A. 1986: Membrane damage by hemolytic viruses, toxins, complement, and other cytotoxic agents. A common mechanism blocked by divalent cations. J Biol Chem, 261:9300-9308

[120] Lee, S.Y., Banerjee, A., and MacKinnon, R. 2009: Two Separate Interfaces between the Voltage Sensor and Pore Are Required for the Function of Voltage-Dependent K+ Channels. PLoS Biol, 7:e1000047-

[121] Long, S.B., Tao, X., Campbell, E.B., and MacKinnon, R. 2007: Atomic structure of a voltage-dependent K+ channel in a lipid membrane-like environment. /ature, 450:376-382

[122] Bezanilla, F. 2008: How membrane proteins sense voltage. /at Rev Mol Cell Biol, 9:323-332

[123] Koishi, R., Xu, H., Ren, D., Navarro, B., Spiller, B.W., Shi, Q., and Clapham, D.E. 2004: A superfamily of voltage-gated sodium channels in bacteria. J Biol Chem, 279:9532-9538

[124] Ito, M., Xu, H., Guffanti, A.A., Wei, Y., Zvi, L., Clapham, D.E., and Krulwich, T.A. 2004: The voltage-gated Na+ channel NaVBP has a role in motility, chemotaxis, and pH homeostasis of an alkaliphilic Bacillus. Proc /atl Acad Sci USA, 101:10566-10571

[125] Yue, L., Navarro, B., Ren, D., Ramos, A., and Clapham, D.E. 2002: The cation selectivity filter of the bacterial sodium channel, NaChBac. J Gen Physiol, 120:853-

[126] Pavlov, E., Bladen, C., Winkfein, R., Diao, C., Dhaliwal, P., and French, R.J. 2005: The pore, not cytoplasmic domains, underlies inactivation in a prokaryotic sodium channel. Biophys J, 89:232-242

[127] Zhao, Y., Yarov-Yarovoy, V., Scheuer, T., and Catterall, W.A. 2004: A gating hinge in Na+ channels: A molecular switch for electrical signaling. /euron, 41:859-865

[128] Bezanilla, F. 2005: Voltage-gated ion channels. IEEE Trans /anobioscience, 4:34-48

[129] Swartz, K.J. 2008: Sensing voltage across lipid membranes. /ature, 456:891-897

[130] Chahine, M., Pilote, S., Pouliot, V., Takami, H., and Sato, C. 2004: Role of arginine residues on the S4 segment of the Bacillus halodurans Na+ channel in voltage-sensing. J Membr Biol, 201:9-24

[131] Kuzmenkin, A., Bezanilla, F., and Correa, A.M. 2004: Gating of the bacterial sodium channel, NaChBac: Voltage-dependent charge movement and gating currents. J Gen Physiol, 124:349-356

[132] Blanchet, J., Pilote, S., and Chahine, M. 2007: Acidic residues on the voltage-sensor domain determine the activation of the NaChBac sodium channel. Biophys J, 92:3513-3523

7. References

83

[133] Studier, F.W. 2005: Protein production by auto-induction in high-density shaking cultures. Protein Expr Purif, 41:207-234

[134] Miroux, B. and Walker, J.E. 1996: Over-production of proteins in Escherichia coli: Mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Membr Biol, 260:289-298

[135] Nurani, G., Radford, M., Charalambous, K., O'Reilly, A.O., Cronin, N.B., Haque, S., and Wallace, B.A. 2008: Tetrameric bacterial sodium channels: Characterization of structure, stability, and drug binding. Biochemistry, 47:8114-8121

[136] Rigaud, J.L., Mosser, G., Lacapere, J.J., Olofsson, A., Levy, D., and Ranck, J.L. 1997: Bio-Beads: An efficient strategy for two-dimensional crystallization of membrane proteins. J Struct Biol, 118:226-235

[137] Montal, M., Darszon, A., and Schindler, H. 1981: Functional reassembly of membrane proteins in planar lipid bilayers. Q Rev Biophy, 14:1-79

[138] Darszon, A., Vandenberg, C.A., Schoenfeld, M., Ellisman, M.H., Spitzer, N.C., and Montal, M. 1980: Reassembly of protein-lipid complexes into large bilayer vesicles: Perspectives for membrane reconstitution. Proc /atl Acad Sci USA, 77:239-243

[139] Schindler, H. 1979: Exchange and interactions between lipid layers at the surface of a liposome solution. Biochim Biophys Acta, 555:316-336

[140] Schürholz, T. and Schindler, H. 1991: Lipid-protein surface films generated from membrane vesicles: selfassembly, composition, and film structure. Eur Biophys J, V20:71-78

[141] Salesse, C., Ducharme, D., and Leblanc, R.M. 1987: Direct evidence for the formation of a monolayer from a bilayer. An ellipsometric study at the nitrogen-water interface. Biophys J, 52:351-352

[142] Cohen, F.S., Akabas, M.H., Zimmerberg, J., and Finkelstein, A. 1984: Parameters affecting the fusion of unilamear phospholipid vesicles with planar bilayer membranes. J Cell Biol, 98:1054-1062

[143] Anzai, K., Ogawa, K., Ozawa, T., and Yamamoto, H. 2001: Quantitative comparison of two types of planar lipid bilayers - folded and painted - with respect to fusion with vesicles. J Biochem Biophys Methods, 48:283-291

[144] Miller, C. and Racker, E. 1976: Ca++-induced fusion of fragmented sarcoplasmic reticulum with artificial planar bilayers. J Membr Biol, 30:283-300

[145] Woodbury, D.J. and Hall, J.E. 1988: Role of channels in the fusion of vesicles with a planar bilayer. Biophys J, 54:1053-1063

[146] Ermishkin, L., Kasumov, K., and Potzeluyev, M. 1976: Single ionic channels induced in lipid bilayers by polyene antibiotics amphotericin B and nystatine. /ature, 262:698-699

[147] Woodbury, D.J. and Miller, C. 1990: Nystatin-induced liposome fusion. A versatile approach to ion channel reconstitution into planar bilayers. Biophys J, 58:833-839

[148] Kleinberg, M.E. and Finkelstein, A. 1984: Single-length and double-length channels formed by nystatin in lipid bilayer membranes. J Membr Biol, 80:257-269

[149] Domb, A.J., Linden, G., Polacheck, I., and Benita, S. 1996: Nystatin-dextran conjugates: Synthesis and characterization. J Polymer Sci, 34:1229-1236

7. References

84

[150] De Planque, M.R.R., Mendes, G.P., Zagnoni, M., Sandison, M.E., Fisher, K.H., Berry, R.M., Watts, A., and Morgan, H. 2006: Controlled delivery of membrane proteins to artificial lipid bilayers by nystatin-ergosterol modulated vesicle fusion. IEE Proc /anobiotechnol, 153:21-30

[151] Helrich, C.S., Schmucker, J.A., and Woodbury, D.J. 2006: Evidence that nystatin channels form at the boundaries, not the interiors of lipid domains. Biophys J, 91:1116-1127

[152] Woodbury, D.J. 1999: Nystatin - ergosterol method for reconstituting ion channels into planar lipid bilayers. Methods Enzymol, 294:319-339

[153] White, H.S. and King, G.I. 1985: Molecular packing and area compressibility of lipid bilayers. Proc /atl Acad Sci USA, 82:6532-6536

[154] Crill, W.E. 1996: Persistent sodium current in mammalian central neurons. Ann Rev Physiol, 58:349-362

[155] Yamakage, M. and Namiki, A. 2002: Calcium channels - basic aspects of their structure, function and gene encoding; anesthetic action on the channels - a review. Can J Anesth, 49:151-164

[156] Adelman, J.P., Shen, K.Z., Kavanaugh, M.P., Warren, R.A., Wu, Y.N., Lagrutta, A., Bond, C.T., and an North, R. 1992: Calcium-activated potassium channels expressed from cloned complementary DNAs. /euron, 9:209-216

[157] Pérez, G., Lagrutta, A., Adelman, J.P., and Toro, L. 1994: Reconstitution of expressed KCa channels from Xenopus oocytes to lipid bilayers. Biophys J, 66:1022-1027

[158] Rostovtseva, T.K., Aguilella, V.M., Vodyanoy, I., Bezrukov, S.M., and Parsegian, V.A. 1998: Membrane surface-charge titration probed by Gramicidin A channel conductance. Biophys J, 75:1783-1792

[159] Schmidt, C., Mayer, M., and Vogel, H. 2000: A chip-based biosensor for the functional analysis of single ion channels. Angew Chem Int Ed, 39:3137-3140

[160] Zagnoni, M., Sandison, M.E., and Morgan, H. 2009: Microfluidic array platform for simultaneous lipid bilayer membrane formation. Biosens Bioelectron, 24:1235-1240

[161] Cheng, Y., Bushby, R.J., Evans, S.D., Knowles, P.R., Miles, R.E., and Ogier, S.D. 2001: Single ion channel sensitivity in suspended bilayers on micromachined supports. Langmuir, 17:1240-1242

[162] Fertig, N., Meyer, C., Blick, R.H., Trautmann, C., and Behrends, J.C. 2001: Microstructured glass chip for ion-channel electrophysiology. Phys Rev E, 64:040901-1-4

[163] Schift, H., Bellini, S., and Gobrecht, J. 2006: Perforated polymer membranes fabricated by nanoimprint. Microelectron Eng, 83:873-875

[164] Hansen, J.S., Perry, M., Vogel, J., Vissing, T., Hansen, C.R., Geschke, O., Emneus, J., and Nielsen, C.H. 2009: Development of an automation technique for the establishment of functional lipid bilayer arrays. J Micromech Microeng, 19:025014-1-11

[165] Suzuki, H., Pioufle, B., and Takeuchi, S. 2009: Ninety-six-well planar lipid bilayer chip for ion channel recording fabricated by hybrid stereolithography. Biomed Microdevices, 11:17-22

[166] Le Pioufle, B., Suzuki, H., Tabata, K.V., Noji, H., and Takeuchi, S. 2008: Lipid bilayer microarray for parallel recording of transmembrane ion currents. Anal Chem, 80:328-332

Appendix

85

Appendix

Diffusion: formulas and calculations

The following derivation of the diffusion coefficient is adapted from Cussler [111] and

Bohrer [116].

Assuming that no membrane potential exists, the main driving force is the concentration

gradient and Fick’s law in a steady diffusion is:

Fick’s first law: x

CDJ i ∂

∂−= (A1.1)

Fick’s second law: x

CD

t

C

∂∂

=∂

∂ 2= 0 (A1.2)

Consequently diffusion across a thin film occurs from the higher

concentration Ccis (x<0) to the lower concentrated solution Ctrans

(x>l). With the solute concentration (boundary conditions) at

x0 = Ccis and xl = Ctrans the concentration profile of the solute is

found by integrating the second law (A1.2) twice: C(x) = a + bx.

X = 0 C(x) = a = Ccis

X = l C(x) = a + bl = Ctrans ⇔ b = (Ctrans –Ccis)/l

Concentration profile: ( )l

xCCCxC cistranscis −+=)(

The flux across the thin film is given by Fick’s first law (A1.1) and can be calculated

knowing the concentration profile: l

CCD

l

CCD

x

CDJ transciscistrans −

=−

−=∂∂

−=

Unlike a film or porous chip, a lipid membrane is chemically different from the

solution. Using a partition coefficient H, the concentration profile and flux across the

membrane can be calculated using the boundary conditions: x0 = H·Ccis and xl =

H·Ctrans.

Concentration profile: ( )l

xCCHCHxC cistranscis −+⋅=)(

Flux across the membrane: l

CCDHJ transcis −

= where (D·H)/l is the permeability.

Assuming that the flux across the chip reaches quickly its steady state value the overall

mass balance is: JAt

CV

t

CV cis

cistrans

trans ⋅=∂

∂−=

∂

∂ with A = total pore area.

Change in the gradient with time:

+−=

+−⋅=

⋅−

⋅−=

∂∂

−∂

∂=−

∂∂

transciscistrans

transciss

trans

s

cis

scistranscis

VVCC

l

AD

VVjA

V

jA

V

jA

t

Ctrans

t

CCC

t

11)(

11)(( )

Can also be written as )(( ) cistranstranscis CCDCCt

−=−∂∂

β with

+=

transcis VVl

A 11β

X

Ccis l

Ctrans

Appendix

86

This differential equation has the solution: DtAetC β−=∆ )( and can be solved using the

initial condition: 000)0( transcis

Dt CCAAeC −===∆ −=

β to:

+=−=−∆=∆ −−

transcis

Dttranscistranscis

Dt

VVl

AwitheCCtCtCoreCtC

11)()()()( 00

0 βββ

Since we can directly measure the various concentrations at certain time points and the

geometric factor β is known, the diffusion coefficient can be found.

Compared to the reported sodium diffusion coefficient DNa+= 1.33·10-5 cm2s-1 [112] the

observed diffusion is 64 times slower. This is due to unstirred layers.

Unstirred layers

The total resistance to transport (RT) is the sum of the membrane resistance (RM) and

the boundary layer resistance (RB) on each side of the membrane. RT = RM + 2RB. From

the observed diffusion (Fig. A1), the thickness of the unstirred layers can be calculated.

1681.25.1

0025.02224.118613

2 −−⋅

⋅ === − scmRsEcm

cmDVA

totobs

tot

β

RM can be calculated for a given solute if the pore radius (r), pore length (L), diffusion

coefficient (DNa+) and pore density (η) are known:

122122

61.49256.1884.35348.1

53 −−−

+

=−⋅⋅−

−== scm

cmEcmEscmE

cmE

rD

LR

/am

ηπ

2 RB = Rtot-RM = 590.8 s cm-1

The thickness of the unstirred layer is δ = DNa+·RB = 79.6 µm.

Fig. A1 | Sodium diffusion over bare chip PA200

From a linear regression (red line) the observed diffusion coefficient is

calculated: Slope = 2.8066E-6 (R = 0.998), β = 13.4 Dapp = 2.1E-7 cm2s-1

List of publications

87

List of publications

Reviewed papers

Philipp S. Angerer, André Studer, Bernard Witholt and Zhi Li. Oxidative Polymerization of a Substituted Phenol with Ion-Paired Horseradish Peroxidase in an Organic Solvent. Macromolecules 2005; 38:6248-6250

Xiaojun Han, André Studer, Harald Sehr, Isabelle Geissbühler, Marco Di Berardino, Fritz K. Winkler and Louis X. Tiefenauer. /anopore Arrays for Stable and Functional Free-Standing Lipid Bilayers. Advanced Materials 2007; 19:4466-4470

Louis X. Tiefenauer and André Studer. /ano for bio: /anopore arrays for stable and functional lipid bilayer membranes (Mini Review). Biointerphases 2008; 3(2):FA74-FA79

André Studer, Xiaojun Han, Fritz K. Winkler and Louis X. Tiefenauer. Formation of individual protein channels in lipid bilayers suspended in nanopores. Colloids and Surfaces B: Biointerfaces 2009 accepted

Conference talks

Functional assays for membrane proteins using nanopore array chips. Workshop on „Dynamics of artificial and biological membranes“, Gomadingen (Germany) 2006

Measuring membrane protein-mediated transport across lipid bilayers. The tenth world congress on Biosensors, Shanghai (China) 2008

Poster presentations

Stable planar lipid bilayers in nanopores. Biosurf VII – Functinal Interfaces for Directing Biological Response, Zurich (Switzerland) 2007

Monitoring ion diffusion and transport across free-standing lipid bilayers. TETHMEM workshop on “Novel Model Systems for Bimolecular Lipid Membranes” Tegernsee (Germany) 2007

Stable free-standing lipid bilayers in nanopores for bioanalytical applications. The tenth world congress on Biosensors, Shanghai (China) 2008

Monitoring integration of proteins into free-standing lipid bilayers. USGEB annual meeting on “Membranes in Motion”, Interlaken (Switzerland) 2009

Documents

Rights / License: Research Collection In Copyright - Non ...263/eth-2… · membrane protein transporters, operating at lower rates than ion channels and requiring higher numbers