Structural Investigations of Lipid Model Biomembranes by ... · von Phosphatgruppen zwischen...

Structural Investigations of Lipid Model

Biomembranes by FTIR ATR Spectroscopy:

I) Interaction with the Uncoupling Agent 2,4,5-Trichlorophenol

II) Investigation of Immobilized Mitochondrial Creatine Kinase

Struktur-Funktions-Untersuchungen an Lipid-Modellbiomembranen

mittels FTIR ATR Spektroskopie:

I) Wechselwirkung mit dem Entkoppler 2,4,5-Trichlorophenol

II) Studie an immobilisierter mitochondrialer Kreatinkinase

Dissertation

zur Erlangung des Akademischen Grades

Doktorin der Naturwissenschaften

an der Fakultät für Naturwissenschaften und Mathematik

der Universität Wien

vorgelegt von

Mag. Monira Siam Wien, Juni 2001

3

Contents

Zusammenfassung 11 Summary 15

1 INTRODUCTION: FTIR-ATR AS A TOOL FOR THE INVESTIGATION OF BIOLOGICAL MEMBRANES 19

2 PRINCIPLES OF FTIR-ATR SPECTROSCOPY 21

2.1 Fourier Transform Infrared (FTIR) - Spectroscopy 21

2.1.1 Vibration of Molecules 21

2.1.2 FTIR-Spectrometers 23

2.1.3 Characteristic Bands of Trichlorophenols, Phospholipids and Proteins 25

2.2 Attenuated Total Reflection 28

2.2.1 Total Reflection and Evanescent Wave 28

2.2.2 Relative Electric Field Components 29

2.2.3 Quantitative Determination 30

2.2.3.1 Effective Thickness 30

2.2.3.2 Dichroic Ratio 31

2.2.3.3 Determination of Sample Orientation 32

2.2.3.4 Surface concentrations of Thin Films 35

2.2.4 SBSR Set Up 37

3 MITOCHONDRIAL MEMBRANES 39

3.1 Structure and Function of Mitochondrial Membranes 39

3.1.1 Assembly and Composition 39

3.1.2 The pH Gradient and ATP Production 40

3.1.3 Uncoupling 40

3.2 Penetration of Ions and Polar Solutes into/through Membranes 41

3.3 Energy Management: Phosphocreatine-Circuit 46

4 PREPARATION OF MODEL MEMBRANES

4

ON INTERNAL REFLECTION ELEMENTS 49

4.1 Preparation of the Substrates 49

4.2 Preparation of Monolayers 49

4.3 Preparation of Bilayers 50

5 INTERACTION OF 2,4,5-TRICHLOROPHENOL

WITH PLANAR LIPID MODEL MEMBRANES 53

5.1 Introduction 53

5.1.1 Chlorophenol-Membrane-Interaction: Narcosis and Uncoupling 53

5.1.2 Kinetic Schemes of Uncoupling 54

5.2 Materials 55

5.3 Methods 56

5.3.1 Long Term Measurements 56

5.3.2 Time resolved Measurements 56

5.3.3 Determination of Molar Absorption Coefficients 59

5.3.3.1 IR: Transmission and ATR-Measurements 59

5.3.3.2 UV VIS 60

5.4 Results 60

5.4.1 UV VIS Measurements 60

5.4.2 IR Transmission Measurements 61

5.4.2.1 Phenol and Phenoxide Spectra 61

5.4.2.2 Molar Absorption Coefficients and Integrated Molar Absorption Coefficients 66

5.4.3 Characterization of the Prepared Model Membranes 69

5.4.4 Long Term Measurements of Lipid-TCP Interaction 70

5.4.4.1 Interaction with DPPA Monolayers 70

5.4.4.2 Interaction with DPPA-POPC-Bilayers 72

5.4.5 Time resolved Measurements Of Lipid-TCP Interaction 75

5.4.5.1 Interaction with DPPA Monolayers 75

5.4.5.2 Interaction with DPPA-POPC-Bilayers 81

5.4.6 Survey and Comparison 87

5.5 Discussion 91

5.5.1 Stacks and multilayer formation 91

5

5.5.2 Evidence for a Heterodimer 92

5.5.3 Effects Exerted on the Membranes 93

5.5.4 Statements about Orientation of TCP in/on the Model Membranes 94

5.5.5 Thermodynamic Considerations 98

5.5.6 Conclusion 99

6 INVESTIGATION OF IMMOBILIZED MITOCHONDRIAL CREATINE KINASE 101

6.1 Introduction 101

6.2 Experimental Section 104

6.2.1 Materials 104

6.2.2 AFM and EM Measurements 105

6.2.3 FTIR ATR Spectra Acquisition, Spectra Manipulation and Special Attachments 106

6.2.4 Survey of the Experiments 107

6.3 Results 112

6.3.1 Bilayers on Ge: Quantification and Stability 112

6.3.2 Adsorption of Mitochondrial Creatine Kinase 121

6.3.3 1H-2H-exchange of Immobilized Mitochondrial Creatine Kinase 123

6.3.4 Activity Control Measurements 129

6.3.5 Interaction with Magnesium ADP 134

6.3.6 Binding of CL(Beef Heart) to Adsorbed Mitochondrial Creatine Kinase 137

6.4 Discussion 140

7 APPENDIX 147

7.1 Ad Section 2: 147


7.2.1 Results of the UV VIS Measurements 150

7.2.2 Results of the Membrane Preparation 153

7.2.3 Further Results of Time Resolved Measurements 154


7.3.1 AFM Image of a DPPA Bilayer 161

7.3.2 EM: Image of a Replica of the Immobilized Mib-CK 162

7.3.3 Tables 163

7.3.3.1 Properties of the prepared phospholipid layers 163

7.3.3.2 Results of the immobilization of Mib-CK 169

6

8 REFERENCES 173

Lebenslauf 187

Publikationsliste 189

7

„Also in müßiger Tätigkeit, wenn Sie schon nicht anders

wollen“, sagte er lächelnd. „Ich treibe zu meinem Vergnügen

etwas Naturwissenschaften, das ist das Ganze.“

(aus „Innocens“ von Ferdinand von Saar)

9

Danksagung

Die vorliegende Arbeit wurde am Institut für Physikalische Chemie (Biophysikalische Chemie) an

der Universität Wien unter Anleitung von Prof. Urs Peter Fringeli im Zeitraum von April 1996 bis

Juni 2001 durchgeführt.

Mein besonderer Dank gilt Herrn Prof. Urs Peter Fringeli für die Aufnahme in seine

Arbeitsgruppe, die wissenschaftliche Anregung und Kritik und die freie und selbständige

Arbeitsatmosphäre. Meinen Kollegen, Mag. Gerald Reiter, Dr. Dieter Baurecht und Mag.

Michael Schwarzott, möchte ich für das angenehme Arbeitsklima, die ständige Hilfsbereitschaft

und die vielen wertvollen Diskussionen rund um eine Tasse dampfenden Kaffee danken. Herr Mag.

Gerald Reiter hat mich in die praktischen Einzelheiten der Membranpräparation und der FTIR-ATR

Messtechnik mit viel Liebe zum Detail eingeführt und war bei der Behebung von Schwierigkeiten

immer behilflich. Herr Dr. Dieter Baurecht war immer zur Stelle, um Probleme mit der EDV zu

lösen. Von Frau Dr. Beate Escher und Herrn Dr. René Hunziker (EAWAG, Dübendorf, CH)

stammt die Anregung die Wechselwirkung zwischen Trichlorophenol und Phospholipidmembranen

zu untersuchen. René Hunziker danke ich für die freundliche Unterstützung bei den Untersuchungen.

Die theoretischen Aufbereitung der Infrarot-Spektren stützt sich auf Beiträge von Herrn Prof.

Alfred Karpfen und Frau Dr. Alexandra Simperler (Institut für Theoretische Chemie bzw. Institut

für Organische Chemie Universität Wien, A). Herrn Prof. Alfred Karpfen und Herrn Prof. Werner

Mikenda (Institut für Organische Chemie Universität Wien, A) gilt mein Dank für die Diskussion

der Ergebnisse aus den quantenchemischen Rechnungen.

Herr Prof. Theo Wallimann und Herr Dr. Uwe Schlattner (Institut für Zellbiologie, ETH Zürich,

CH) haben die mitochondriale Kreatinkinase kostenlos zur Verfügung gestellt und mir damit ein

lohnendes Untersuchungsobjekt in die Hand gegeben. Darüberhinaus haben sie ermöglicht, daß Herr

Dr. Peter Tittmann, EM Aufnahmen des immobilisierten Enzymes durchgeführt hat. Die AFM-

Aufnahmen des DPPA-Bilayers wurden von Frau Dr. Erika Györvary (Zentrum für

Ultrastrukturforschung, Universität für Bodenkultur, Wien, A) durchgeführt. Ihnen allen möchte ich

an dieser Stelle ebenfalls meinen herzlichen Dank aussprechen.

Dr. Graham Muir hat die Durchsicht des englischen Textes besorgt. Ihm, allen meinen Freunden

und Freundinnen, und insbesondere meiner Familie sei gedankt für das Verständnis und die

moralischen Unterstützung, die mir erst die Kraft für die Durchführung dieser Arbeit gegeben haben.

11

Zusammenfassung

Als Modellsysteme für Zellmembranen können Bilayer aus Phospholipiden verwendet

werden. In der vorliegenden Dissertation wurden an solchen Modellmembranen mittels

abgeschwächter Totalreflexion (ATR) im Infrarotbereich Struktur- und

Funktionsuntersuchungen durchgeführt. Dazu wurden die Phospholipide auf Germanium-

Platten immobilisiert. Diese dienen dabei gleichzeitig als Träger des planaren Lipidlayers

und als Lichtleiter für die interne Reflexion. In der vorliegenden Arbeit wurde die Methode

auf zwei Fragestellungen angewendet.

Die erste Fragestellung beschäftigt sich mit der Aufklärung der Wechselwirkung des

ökotoxischen 2,4,5-Trichlorophenols mit Phospholipidmembranen. Für die zweite

Fragestellung wurde die Methode auf das Membranprotein mitochondriale Kreatinkinase

angewendet. Das Protein wurde an Modellmembranen aus negativ geladenen Phospholipiden

immobilisiert und unter möglichst nativen Bedingungen folgende Fragen abgeklärt: Art der

Bindung an die Modellmembranen, Struktur des Proteins an der Membran in wäßrigem

Millieu, Aktivität des immobilisierten Enzymes und Wechselwirkung mit dem Substrat

MgADP.

12

I) Wechselwirkung mit dem Entkoppler 2,4,5-Trichlorophenol

Trichlorophenole können als schwache Säuren mit großer Hydrophobizität Protonen durch

eine Lipidmembran schleusen und damit den Protonengradienten abbauen. Sie entkoppeln so

den Protonentransport von der ATP-Produktion und stören daher den Energiestoffwechsel

der Zelle. Je nach der experimentell gefundenen Kinetik wird zwischen Klasse-1- und

Klasse-2- Entkopplern unterschieden. Klasse-1-Entkoppler zeigen eine Kinetik 1.Ordnung,

Klasse-2-Entkoppler eine 2. Ordnung. Nach Finkelstein [1] kommt es bei Klasse-2-

Entkoppler zur Bildung von Heterodimeren aus protonierter und deprotonierter Form. 2,4,5-

Trichlorophenol (TCP) kann als Klasse-2-Entkoppler eingestuft werden [2]. Mittels Fourier

Transform Infrarot (FTIR) ATR Spektroskopie sollte nun die Heterodimerbildung des TCP

auf Phospholipidmembranen nachgewiesen werden. Dafür wurden zwei Membrane mit

unterschiedlichen Oberflächeneigenschaften hergestellt: a) Monolayer aus Dipalmitoyl-

phosphatidsäure (DPPA), die ihre hydrophoben Ketten als Adsorptionsfläche zur Verfügung

stellen und b) planare Bilayer aus DPPA und Palmitoyloleoylphosphatidylcholin (POPC), bei

denen die polaren Kopfgruppen exponiert werden. DPPA Monolayer wurden auf die

Germanium-Platte mittels Langmuir-Blodgett (LB) Methode transferiert und DPPA/POPC

Bilayers mittels LB/Vesikel Methode [3] aufgebaut. TCP adsorbierte aus 1 bis 3 mmol/L

Lösungen bei pH 6.0 an beide Oberflächen. Zeitaufgelöste Messungen ergaben zwei

unterschiedliche Adsorptionsprozesse, von denen einer dem TCP und der andere der

deprotonierten Form, dem Phenoxid, zugeordnet wurden. Die Adsorption von TCP

(repräsentiert durch einen Peak bei 1080 cm-1) war deutlich schneller als die des Phenoxids

(repräsentiert durch einen Peak bei 1352 cm-1). Die gemessenen Oberflächenkonzentrationen

für das Phenoxid zeigen, daß es an den Modellmembranen angereichert wurde. Phenoxid und

ein Teil des Phenols bleiben adsobiert, nachdem man die TCP-Lösungen durch Puffer

13

ersetzt. Für die retardierten Spezies wurde ein molares Verhältnis Phenol/Phenoxid von ca.

1/1 bei pH 6.0 bestimmt (bei einem pKa 7.0 für TCP). Die vorliegenden Ergebnisse

unterstützen damit die These der Heterodimerbildung.

II) Studie an immobilisierter mitochondrialer Kreatinkinase

Kreatinkinasen (CK) sind eine Gruppe von Isoenzymen (EC 2.7.3.2), die die Übertragung

von Phosphatgruppen zwischen ATP/Kreatin einerseits und ADP/Phosphokreatin (PCr)

andererseits katalysieren. Sie kommen überall dort vor, wo in Wirbeltierzellen innerhalb

kurzer Zeit große Energiemengen benötigt werden (Muskeln, Herzmuskel, Gehirn)[4]. Die in

den Mitochondrien auftretenden basischen CKs (Mib-CK) bilden Oktamere und Dimere,

wobei die Oktamere, die an den Membranen gebundenen aktiven Enzyme darstellen. Das

Oktamere ist mit 345 kDa ein großes Protein, das ca. würfelförmig gebaut ist (P 422). Die

genaue Struktur einer Mib-CK wurde 1996 von Fritz-Wolf et al. [5] aufgeklärt. Mib-CK

zeigen unspezifisch gute Bindung an diversen Membranen, die bei höherem Anteil an

negativ geladenen Lipiden, wie dem in Mitochondrienmembranen häufig auftretenden

Cardiolipin (CL), ansteigt [6, 7]. Dies wurde ausgenützt, um mitochondriale Kreatinkinase

an negativ geladene Phospholipidmembranen zu adsorbieren. Dazu wurden symmetrische

Bilayer aus Dipalmitoylphosphatidsäure (DPPA) und asymmetrische Bilayer mit CL als

äußerer Membran eingesetzt. Symmetrische DPPA Bilayer wurden mittels Langmuir-

Blodgett (LB) Methode auf die Germanium-Platte übertragen, während asymmetrische

DPPA/CL Bilayers mittels LB/Vesikel Methode [3] aufgebaut wurden.

An diesen Bilayern wurden zwei recombinante sarkomere Mi-CK, die vom Huhn stammende

ch Mib-CK und die vom Menschen stammende hu Mib-CK, untersucht.

Beide Enzyme ließen sich mit eine Belegungsdichte von 40% bis 60% in aktiver Form

14

immobilisieren. Die beobachtete schnelle Adsorption deutet auf elektrostatische

Wechselwirkungen zwischen den an Lysin und Arginin reichen Oberflächen des Proteins

und den negativ geladenen Kopfgruppen von CL und DPPA hin. Das bestätigt auch die

verstärkte Stabilität der negativ geladenen Bilayer, sobald Mib-CK adsorbiert ist. Mib-CK

kompensiert durch seine positiv geladene Oberfläche die elektrostatische Abstoßung der

Kopfgruppen. Für zusätzliche hydrophobe Wechselwirkungen mit den Acylketten der

Phospholipide konnte kein Hinweis gefunden werden. Die Untersuchung des 1H-2H-

Austausches des immobilisierten Enzymes ergab einen Gehalt von 27 ± 6% an Amid-

Protonen, die für das Lösungsmittel schwer zugänglich sind. Die zugehörige Zeitkonstante

beträgt etwa eine Woche. Dieser Wert entspricht dem durch Röntgenstrukturanalyse

ermittelten Anteil an α-Helix von 33%. Für den Anteil an β-Faltblättern im immobilisierten

Enzyme ergibt eine Analyse der IR-Spektren jedoch einen Wert unter den in der Literatur

beschriebenen 13% [5]. Die an CL-Monolayern gefundene 2D-Kristallbildung [8] hat im

vorliegenden System nicht stattgefunden. Darüberhinaus wurden FTIR ATR Differenz-

Spektren vom MgADP*Enzym-Komplex der ch Mib-CK und hu Mib-CK gemessen. Sie

bestätigen, daß Asp, Glu, Tyr und Arg an der Bindung von MgADP beteiligt sind.

Außerdem, zeigt die Antwort im Amid I/I’ -Bereich eine Zunahme von α-Helices auf Kosten

von ungeordneten Teilen und β-Faltblättern.

15

Summary

Membrane of cells can be modeled by phospholipid bilayers. In this study polarized Fourier

Transform IR Attenuated Total Reflection (FTIR ATR) spectroscopy was used for structural

investigations of such lipid model membranes. The phospholipid membranes were

immobilized on germanium crystals functioning as substrate for the lipid layers and as

waveguides for the internal reflection. Two problems were examined. First, the interaction

between phospholipid layers and 2,4,5-trichlorophenol, a contaminant of considerable

environmental concern, was investigated. Second, the method was applied to study the

membrane protein mitochondrial creatine kinase. The protein was immobilized on bilayers of

negatively charged phospholipids. The study addressed the following questions: the lipid

protein interaction, the structure and activity of the immobilized protein and the interaction

with one of its substrates, namely MgADP.

I) Interaction with the Uncoupling Agent 2,4,5-Trichlorophenol

Trichlorophenols are weak acids of high hydrophobicity and able to transport protons across

the mitochondrial membrane. Thus the proton motive force is dissipated and the ATP

production decreased. Depending on the observed kinetics we differentiate between class-1

and class-2 uncouplers. Class-1 uncouplers follow a first order kinetic, whereas for class-2

uncouplers a second order kinetic is found. Finkelstein [1] proposed that the class-2

uncoupler forms a heterodimer, consisting of an ion and a neutral species. For 2,4,5-

trichlorophenol a second order kinetic was determined [2]. Thus the formation of the

heterodimer of 2,4,5-trichlorophenol (TCP) should be confirmed by trapping it on supported

lipid layers. Two model surfaces were examined: a monolayer of dipalmitoylphosphatidic

acid (DPPA) and a planar bilayer of DPPA and of palmitoyl oleoyl phosphatidylcholin

16

(POPC). The DPPA monolayer was transferred by the Langmuir-Blodgett (LB) method to

the germanium plate. Whereas, DPPA/CL bilayers were built up using the LB/vesicle

method [3]. TCP was adsorbed from 1 to 3 mmol/L solutions at pH 6.0 to the lipid layers

leading to superficial layers at the water/lipid layer interface. Difference spectra show an

effect to DPPA acyl chains even when it is covered with POPC. Time resolved

measurements revealed two distinct adsorption processes, which were assigned to TCP and

its deprotonated anion (phenoxide), respectively. The adsorption of TCP (represented by a

peak at 1080 cm-1) was faster than the one of its phenoxide (represented by a peak at 1352

cm-1). The observed surface concentration of phenoxide reveals that it was enriched at the

model membranes. Phenoxide and phenol were retained after replacing the TCP solution

with pure buffer. For the retained species we estimated a phenol-phenoxide molar ratio of 1

at pH 6.0 (pKa 7 for TCP). Thus, a strong evidence for heterodimer formation is given.

II) Investigation of Immobilzed Mitochondrial Creatine Kinase

Creatine kinases are a group of the isoenzymes (EC 2.7.3.2) that transphosphorylate ATP and

creatine to ADP and phosphocreatine (PCr) or vice versa. They are found in vertebrate cells

with a high demand of energy [4]. Basic mitochondrial creatine kinases (Mib-CK) form

dimers and octamers. The latter are the membrane binding species. Octamers have a mass of

345 kDa and a cubic like shape (P422). The structure of the protein was solved in 1996 by

Fritz-Wolf et al. [5]. Mib-CKs adsorb to membranes the better the more negatively charged

lipids they contain [6, 7]. Here, FTIR ATR spectroscopy was applied to investigate the

immobilization of mitochondrial creatine kinase on negatively charged bilayers, made from

dipalmitoyl phosphatidic acid (DPPA) and cardiolipins (CL). Recombinant proteins of two

different species, chicken (ch Mib-CK) and human

17

(hu Mib-CK) were examined. The DPPA bilayer was transferred by the Langmuir-Blodgett

(LB) method to the germanium plate which was used as internal reflection element. On the

other hand, asymmetric DPPA/CL bilayers were built up by the LB/vesicle method [3].

Density of coverage estimated for the immobilized enzyme on the bilayer was found between

40% and 60%. The enzyme was immobilized in active form. The observed fast adsorption

process is predominantly due to electrostatic interactions between the lysine and arginine rich

surfaces of the protein and the negatively charged headgroups of CL and DPPA. Mib-CK

clearly enhanced the stability of the bilayers in buffer due to the compensation of the

electrostatic repulsion. The 1H-2H-exchange of immobilized protein was observed and a

content of 27 ± 6% of amid protons that are weakly accessible for the solvent were found.

The time constant of their exchange rate is about one week. This matches with the 33% α-

helices, observed by x-ray diffraction [5]. However, the reported 13% β-sheets could not be

detected for the immobilized enzyme by FTIR. Besides, the formation of 2D crystals, which

was observed with monolayers of CL [8], was not found. Furthermore, its interaction with

the substrate MgADP is reported. FTIR ATR difference spectra of the MgADP* Mib-CK-

complex confirm that Asp, Glu, Tyr and Arg are involved in the MgADP binding. The

response in the amide I/I’ area points to an increase of α-helices on the cost of unordered

parts and β-sheets.

1 Introduction

19

1 Introduction:

FTIR-ATR as a Tool for the Investigation of Biological Membranes

The major components of biological membranes, lipids, proteins and sugars, show infrared

signals, which respond to changes in conformation and interaction. Therefore, Fourier

Transform Infrared (FTIR) spectroscopy has been employed for structure-function

investigations on biological membranes and model systems, such as lipid bilayers [e.g. 9,

10]. In this work two investigations of model systems for mitochondrial membranes by

means of Attenuated Total Reflection (ATR) FTIR spectroscopy are reported. Section 2

gives a short introduction into the theory of FTIR ATR spectroscopy. FTIR-ATR

spectroscopy has several advantages compared to FTIR transmission measurements for the

investigation of biological membranes [11].

1) A lot of preparation methods for biological membranes and simplified model systems,

such as membrane proteins, membrane fragments from tissues and supported lipid layers,

e.g. on an internal reflection element, are described in literature [9, 10, 12, 13, 14, 15]. They

are often self-assembling processes and allow the sample to be immobilized in a known

orientation. Polarized measurements then gain detailed information about the orientation or

changes in orientation. 2) The membranes can be studied in a physiological equivalent

environment. In particular, water can be used as a bulkphase even if the bands of interest are

close to the water bands. The reason for this is, that the sensitivity falls off exponentially

with the distance from the surface of the waveguide. 3) For mono- and submonolayer

spectroscopy internal reflection elements with multiple reflections (MIRE) can be used to

enhance the absorption. The number of reflections of a MIRE depends on its thickness.

MIRE with up to 50 internal reflections are commercially available. 4) It is possible to

expose the membrane to a changing environment (pH, ionic strength, substrates, etc.) using

1 Introduction

20

flow through systems. 5) FTIR-ATR measurements reveal integrated mean information.

Section 3 summarizes some characteristics of mitochondrial membranes. Two lipid model

biomembranes were developed from this basis and prepared on multiple internal reflection

elements (MIRE) as described in section 4.

One of the questions investigated was the interaction of the uncoupling agent 2,4,5-

trichlorophenol with planar lipid layers supported on Ge plates functioning as MIRE. Section

5 describes the results of these experiments. In brief, we found strong indication for the

formation of phenoxide-phenol heterodimers on dipalmitoyl phosphatidic acid (DPPA)

monolayers as well as on DPPA/palmitoyloleoyl phosphatidylcholine (POPC) bilayers.

The second question addressed in this study concerned immobilized mitochondrial creatine

kinase (Mi-CK) on negatively charged bilayers. These model membranes were investigated

to gain information about the lipid-protein interaction, as well as the solvent accessibility and

protein stiffness during hydrogen-deuterium exchange. Additionally, the substrate MgADP

was studied to gain information about its interaction with the active center of the enzyme.

Section 6 deals with these results.

2 Principles of FTIR-ATR Spectroscopy

21


2.1 FOURIER TRANSFORM INFRARED (FTIR) - SPECTROSCOPY

2.1.1 VIBRATION OF MOLECULES

An electromagnetic wave in the region of infrared (IR) (with wavelengths λ between 780 nm

and 2 000 µm, wavenumbers ~ν between 12 800 cm-1 and 5 cm-1) stimulates molecular

vibration (and rotation). Energy will only be absorbed if there is a net change of dipole

moments due to the vibrational motion of the molecule, i.e. a transition dipole moment. If the

atoms of the molecule are modeled as balls and the bonds between them as perfect springs,

any vibration can be described classically by Hooke´s law for a spring [16, 17]. The

frequency f for the resulting vibration is given in equation (1).

The potential and kinetic energies of this harmonic oscillator can be introduced into the

Hamilton operator to give the simplest quantum mechanical model to describe vibrational

energies. The Eigen values (energy levels) are depicted in equation (2). Together with the

selection rules for the IR absorption process (∆n = ±€1), this leads to equidistant energy

levels.

(1) fk

=1

2π µ

with k ... force (spring) constant,

µ ... reduced masses,

µ =+

m mm m

1 2

1 2

h ... Planck´s constant

f ... frequency of the vibration

n ... vibration quantum number

n = 0, 1, 2, 3,...

E h f nv = ⋅ ⋅ +

12 (2)


22

At room temperature almost all molecules are in the vibrating ground state (n = 0). The

observed wavenumbers ~ν for the first harmonics can be calculated from the absorbed energy

∆E = E1-E0 = h ⋅c ⋅ ~ν = h ⋅€f combined with equation (1) to give equation (3).

Therefore, wavenumbers (and frequencies) depend on the strength of the bonds between the

atoms and their masses. They alter if bonds are weakened (e.g. through formation of

hydrogenbonds) or for isotop exchange (e.g. 1H-2H-exchange).

For complex molecules (with more than 3 atoms) vibrations can be divided into different

classes depending on whether only a few atoms (e.g. a functional group) of the molecule or

the whole molecule is involved. The first class give rise to characteristic absorption bands for

functional groups, whereas, the latter make up the so called „fingerprint“ of the molecule.

Furthermore, characteristic bands can be subdivided into vibrations changing the bond

lengths, like symmetric stretching (νs) and antisymmetric stretching (νas), or the bond angles,

like bending (δ) vibrations. Vibrations for repeated units can be coupled, like the wagging (γ)

vibration of CH2-groups.

Generally, a non-linear molecule with N atoms has 3 N-6 (resp. 3 N-5 for linear molecule),

normal modes (types) of vibration. These normal modes and their absorption frequencies can

be estimated by means of quantum chemical calculations. But even in very small molecules

this means one has to solve a many-body problem, which can only be done approximately.

The following approximations are commonly used: the Born-Oppenheimer approximation

~ν ... wavenumber of the infrared light

c ... vacuum velocity of light

k ... force constant

µ ... reduced masses

~νπ µ

=⋅

12 c

k (3)


23

(separation of the motion of electrons and nuclei), simplification of the function describing

electron orbitals with density function theory (DFT), and estimations for exchange

correlation energies of the electrons. Program packages like the GAUSSIAN 98 calculate

iteratively energetically optimized positions of the atoms within the molecule (geometry

opitimization), force constants (strength of bonds) between them, dipole moments, normal

vibration modes, and frequencies. The calculated frequencies can be used to check and

interpret spectra of samples that resemble the gaseous state, as expected for diluted CCl4

solutions. Spectra of higher concentrations and/or in solvents (e.g. H2O) that exert

hydrophobic or hydrophilic forces can differ from calculated results, because intermolecular

interactions between solute and solvent are not included.

2.1.2 FTIR-SPECTROMETERS

As an attendant phenomenon of the computational evolution, Fourier Transform infrared

(FTIR) spectrometers caught on against dispersive spectrometers since the 1980´s. FTIR

spectrometers are a) the faster working, b) the more exact devices with c) the higher signal-

to-noise ratio. The advantage a) is often named as Felgett´s advantage and evolves from the

fact that all wavelengths are measured at the same time. Secondly, a FTIR spectrometer

works with a laser light measuring the position of the moveable mirror, which allows

simultaneous calibration of wavelengths (Connett´s advantage). c) is based on the

Jacquinot´s advantage. There is no need for entrance slits or monochromators and therefore,

measurements at higher intensities are possible [16, 18, 19, 20].

These advantages can be utilized, since the Fourier Transformation of the resulting

interferograms to spectra can be calculated nowadays within a few seconds. The

polychromatic IR beam of the source is split and sent through an interferometer (e.g. a


24

Michelson interferometer) before it probes the sample [17, 20]. A moveable mirror reflects

the beam in a way that it superimposes and interferes with the original one. In the resulting

beam wavelengths are increased or decreased depending on the position of the moveable

mirror, thus producing an interferogram. The absorption of the sample alters the

interferogram, which is recorded by a detector. From the interferograms usually single

channel (SCh) intensity spectra are calculated with the Fast Fourier algorithm (FFT). Using

equation (4) and (5), SCh intensity spectra can be converted into transmission (T) or

absorbance (A) spectra, with I0 representing the intensity of the background or reference

spectra and I the intensity of the sample spectrum. We preferred the absorbance spectra

because sample concentrations (c) can be estimated by Lambert-Beer´s equation (5) for a

known molar absorption coefficient ε and a given sample thickness d.

During the last decades FTIR spectroscopy has been increasingly used for investigations of

biomolecules, like proteins and DNA [10, 21]. High signal-to-noise ratio and fast

accumulation of scans made it possible to get high resolved difference spectra. Applications

of FTIR spectroscopy include investigating kinetic pathways or interactions between proteins

and their environment, like embedded membrane proteins and surrounding lipids.

Particularly, for investigations of biological membranes, the functional group specificity of

FTIR spectroscopy profits a lot if coupled to a surface sensitive technique. Attenuated total

reflection (ATR) represents such a surface sensitive technique and therefore, FTIR-ATR

measurements gained importance as a tool to investigate biological membranes [11].

(4) II

T c d

010= = − ⋅ ⋅ε

(5) A T= c d= − ⋅ ⋅log ε


25

2.1.3 CHARACTERISTIC BANDS OF TRICHLOROPHENOLS, PHOSPHOLIPIDS AND PROTEINS

To illustrate the theoretical section above, characteristic bands of the substances under

investigation are briefly described.

For trichlorophenols, the stretching vibrations ν(OH) are very intense, as was measured in

CCl4 solution (see Fig. 14 in section 5) but cannot be found in spectra of aqueous solutions

due to the high absorption of water in this region. As there are only two aromatic CH-groups

C-H-stretching signals are too small to be detected. 2,4,5-Trichlorophenol shows bands at

1488 cm-1 (22% C-H bending vibration, 11% of C-O-stretching vibration), at 1400 cm-1

(16% of C-O-H bending vibration) and at 1080 cm-1 (49% ring breathing vibration, and 16%

C-Cl stretching vibration). Clearly, all these bands are within the fingerprint region, thus

none of them is solely made up of a vibration of one specific group. The extent of the various

vibrations (normal modes) leading to the signals as calculated in this work are listed in Table

2 (see section 5).

Phospholipids contain a glycerol backbone, acyl chains of fatty acids and a phosphatidic

group. Fig. 6 (see section 4) displays the polarized (parallel: || and perpendicular: ⊥ ) FTIR

ATR absorbance spectra of cardiolipin (CL), dipalmitoyl phosphatidic acid (DPPA) and

palmitoyloleoyl phosphatidylcholine (POPC). The following groups can easily be identified

by the bands found in the spectra: CH3- and CH2-stretching bands at 2959 cm-1 (νas(CH3),

small peak), 2917 cm-1 (νas(CH2), large and sharp peak), 2878 cm-1 (νs(CH3), very small

peak only in ||-polarized spectra), 2849.4 cm-1 (νs(CH2), large and sharp peak), C=O-

stretching band of the esters at 1741 cm-1, the bending vibration δ(CH2) at 1468 cm-1. The

broad bands at 1180 cm-1 and at 1096 cm-1 are dominated by P-O-stretching bands of the

phosphate group (νas(PO2-) and νs(PO2

-)). DPPA/air spectra show a specific sawtooth pattern


26

with 7 little peaks on the broad band of νas(PO2-) due to a concerted wagging vibration of the

14 CH2-groups of the acyl chains in all-trans conformation [22]. For POPC the quaternary

amine group of choline leads to an absorption band at 971 cm-1. Spectra from CL and POPC

were taken from outer leaflets of a bilayer in aqueous environment (reference:

monolayer/buffer). Thus negative water bands indicate that water is displaced by the outer

leaflet.

Infrared spectra of Proteins [e.g. 23] are characterized by the amide bands: amide A ν(NH)

at 3300 cm-1, amide B at 3100 cm-1 (within the water region), amide I at 1650 cm-1, amide II

at 1550 cm-1, amide III at 1300 cm-1, amide IV at 725 cm-1, amide V at 625 cm-1, amid VI at

600 cm-1 and amide VII at 200 cm-1. The amide I band and the amide II band are caused by

combined vibrations. Contributing normal coordinates to the potential energy distribution of

amide I are: C=O-stretching vibration (70-85%), C-N-stretching vibration (10-20%) and C-

CN-stretching vibration (10%). There are small contributions from the N-H-bending

vibration, as well. The band is sensitive for secondary structure elements, α-helix, β-sheet

and turns, and thus often used for the analysis of proportions of α-helix and β-sheet in a

protein or peptide [24, 25]. Random coils can be recognized by a peak with the maximum at

1645cm-1. An α-helix produces a bell shaped band with maxima between 1656 and 1648 cm-

1. Whereas, for β-sheets a strong band with a peak maximum at 1624-1633 cm-1 and for

antiparallel sheets an additional small band at 1675-1685 cm-1 is found. To the amide II band

a contribution of 40-60% N-H-bending vibration, 18-40% C-N-stretching vibration and 10%

C-C-stretching vibration to the potential energy distribution was calculated. The high

contribution of the N-H-bending vibration makes it more sensitive for 1H-2H-exchange.

Additionally, side chains have characteristic frequencies, from which information of specific


27

interactions can be gained, e.g. at the active center of an enzyme (a compilation of

frequencies and molar absorbances is given in [23]). Furthermore, some side chain functional

groups absorb in the amide I and/or amide II region. For a quantitative analysis of secondary

structure elements the complex signals are often deconvoluted and the components extracted

by curve fitting [26], a method which is, however, prone of artifacts. Nevertheless, the

distinctions in the shape of the amide I band can be used for qualitative analysis and to study

the kinetics of aggregation [27] (see Fig.1) or unfolding processes of proteins and

polypeptides. The predictions of such investigations can be increased using reversible folding

and unfolding that allow periodic stimulation and applying modulation technique [28, 29].

Fig. 1. Curve Fitting results of the amide I’-band of human calcitonin in monomeric (A) and aggregated form (B). The figure is taken from [27]. Fits are based on four Gaussian components for due to the three secondary structures (assignment/center wavenumber/HWHH: random coil/1650 cm-1/20 cm-1; α-helix /1635 cm-1/ 8 cm-1; β-sheet/1618 cm-1/ 9 cm-1; νas(COO-)/1590 cm-1/ 29 cm-1 and the νas (COO-) of Asp15. (A) Simulation of a FTIR transmission spectrum recorded after 0.25 h, fitted predominantly with a broad Gaussian band at 1650 cm-1 due to random coil structure and a small α-helical component at 1635 cm-1. (B) Simulation of the last measured spectrum showing in addition to the random coil component significant components due to α-helix (1635 cm-1) and β-sheet (1618 cm-1). For further details see [27].


28

2.2 ATTENUATED TOTAL REFLECTION

2.2.1 TOTAL REFLECTION AND EVANESCENT WAVE

When light approaches an interface with an incidential angle θ from the denser medium 1

(refractive index n1) to the rarer medium 2 (refractive index n2; n1>n2), internal total

reflection (and no transmittance) will occur above a critical angle θc. For this case the angle

of the refracted light φ is 90° and the critical angle θc can be calculated from Snell´s law by

means of equation (6).

The denser medium works as waveguide. It is known from theory (both Fresnel´s and

Maxwell´s equations) as well as from experiments [30, 31, 32] that the electromagnetic wave

is propagating into the adjacent medium. There, the so called evanescent wave falls off

exponentially with the distance from the interface. The electric field amplitude E of any

direction (x,y,z) at the interface E0 (z = 0) decreases by a factor 1/e at the penetration depth

dp.

with

In equation (8) λ1=λ/n1 denotes the wavelength in the denser medium 1 and n21=n2/n1 is the

ratio of the refractive indices [33, 34].

If the adjacent medium is capable of absorbing energy from this field, we speak of attenuated

total reflection (ATR).

sin θ c

nn

n= =2

121 (6)

(7) E E ex y z x y z

zd p

, , , ,= ⋅−

0

(8) dn

p =−

λ

π θ1

221

22 sin


29

2.2.2 RELATIVE ELECTRIC FIELD COMPONENTS

Quantitative calculations and statements about the orientation have to start with the

computation of relative electric field components Er (x,y and z) for the adjacent medium 2 at

z = 0. Schematic representation are given in Fig. 2.

Harrick [35] developed terms (9)-(14) for nonabsorbing media from Fresnel´s equations for

a) a sample resembling a bulk phase with a thickness d much greater than the penetration

depth dp. In case b) the evanescent wave penetrates the bulk medium 3 through a thin film

(e.g. a lipid monolayer) with d << dp. θ denotes the angle of incidence, nik the ratio of the

refractive indices ni/nk of medium i and medium k, where i, k = 1,2,3 stand for the ATR plate

(1), the thin film (2) or the bulk medium (case a) 2 or case b) 3). The electric field can be

assumed to be constant and is controlled more by media 1 and 3 than 1 and 2. Thus,

equations (12) to (14) hold, where Er0,x,2 and Er

0,y,2 are dependent to n31 = n3/n1 and the thin

film influences only E r0,z,2 via n32 = n3/n2. Errors due to the approximations used herein are

usually within the experimental error for a) if the d > 0.5µm and for b) if d < 20 nm [36].

Fig. 2. ATR set up. Directions of electric field components of the incident light and the evanescent wave: θ, angle of incidence; E|||||||| and E⊥⊥⊥⊥ denote the parallel and perpendicular polarized electric field components of the incidence light; Ex, Ey and Ez, electric field components of the evanescent wave in the direction of the coordinate system fixed onto the ATR plate (x,y,z). E|||||||| results in Ex, and Ez; E⊥⊥⊥⊥ gives rise to Ey.


30

a)

b)

Though, the equations are developed for non-absorbing media they are good approximations

for weak absorbers, too. E. g. for Ge in contact with water and θ = 45° the upper limit was

found at ε€⋅ c = αmax = 1 000 cm-1 [34].

2.2.3 QUANTITATIVE DETERMINATION

2.2.3.1 Effective Thickness

To enable the application of Lambert-Beer´s law Harrick [33: pp 41-63] introduced the

concept of effective thickness de. The quantity deiso (as dex

iso, deyiso and dez

iso ) represents the

thickness of an isotropic sample needed for an equal amount of absorption measured as

transmission experiment. In equation (15) n21, dp and θ have their usual meanings. Er0,2 are

(9)

( )E

EE

n

n n nx

r x0 2

0 2

1

221

2

212

212

212

2

1 1, ,

, ,

,

cos sin

sin= =

⋅ −

− + −||

θ θ

θ

(10)

( )E

EE n n n

zr z0 2

0 2

1 212

212

212

2

1 1, ,

, ,

,

sin cos

sin= =

− + −||

θ θ

θ

(11) E

EE n

yr y0 2

0 2

1 212

2

1, ,

, ,

,

cos= =

−⊥

θ

(12)

( )E

EE

n

n n nx

r x0 2

0 2

1

231

2

312

312

312

2

1 1, ,

, ,

,

cos sin

sin= =

⋅ −

− + −||

θ θ

θ

(13)

( )E

EE

n

n n nz

r z0 2

0 2

1

322

312

312

312

2

1 1, ,

, ,

,

sin cos

sin= =

⋅

− + −||

θ θ

θ

(14) E

EE n

yr y0 2

0 2

1 312

2

1, ,

, ,

,

cos= =

−⊥

θ


31

the relative electric field components for x,y and z direction. deiso depends on the sample

position and thickness, given by the initial distance from the interface zi and the final

distance zf.

For a bulk medium with zf = ∞ starting at zi = 0 equation (15) simplifies to equation (16).

Whereas, for an immobilized layer of the thickness d (zi = 0) and d << dp equation (15) can

be simplified to equation (17), using a taylor expansion.

2.2.3.2 Dichroic Ratio

For quantification ATR samples have to be measured with parallel (||) and perpendicular

(⊥) polarized light. These measurements allow orientation analysis, too. As the relevant

quantity, the dichroic ratio R can be calculated either from integrated absorption ( A d⊥ ||∫ ,~

~

~νν

ν

1

2

)

or from peak heights (A||,⊥).

Experimental values of R can be compared to theoretically expected values and interpreted in

(15) d

n dE

zd

zd

re

iso p i

p

f

p=

⋅−

− −

210 22

2 2cos

exp exp,θ

dn d

E re

iso p=⋅

210 22 cos ,θ

(16)

(17) d

n dE

dd

n dEr r

eiso p

p=

⋅− −

=21

0 221

0 221

2cos

expcos, ,θ θ

(18) R A d A d= || ⊥∫ ∫~ ~

~

~

~

~

ν νν

ν

ν

ν

1

2

1

2

= A||/A⊥


32

terms of orientation, provided one knows the ultrastructure of the sample.

Considering an isotropic sample, Riso equals 1 only in case of a transmission experiment. For

ATR measurements, however, Riso differs from unity according to equation (19) and depends

on the relative field components Er. Er for a given system can be calculated as shown in

section 2.2.2 (Relative Electric Field Components) with the equations (9)- (11) and equations

(12)-(14), respectively.

Riso, therefore, is influenced by the relative refractive indices nik and the angle of incidence θ.

For a bulk and weak absorbing rarer medium one obtains equation (20). In this case Riso = 2.0

for θ = 45° (with cos245° = ½), regardless of the value for n21.

Riso for thin films is not an absolutely constant over the whole spectrum. Equation (21) shows

the pronounced dependence of Riso for the relative refractive indices (via n324!) as deduced

from the thin film approximation. Changes of the refractive index of the bulk medium due to

anomalous dispersion thus have a high impact on R.

2.2.3.3 Determination of Sample Orientation

The intensity of light absorption ∆I depends on the mutual orientation of the transition dipole

moment v

M associated with a vibrational mode and the exciting electric field vE . In equation

RE E

Ex z

y

isor r

r=

+02

02

02

, ,

,

(19)

(20)

Rn

n n

n

nbulkiso =

−+ ⋅ −

=⋅ − −

− − ⋅2 2 1

21

221

2

221

2 221

2

2 212

221

2 2

sinsin sin

( cos )

( cos cos )θ

θ θ

θ

θ θ

(21) R

n nn nthin film

iso =+ ⋅ −+ ⋅ −

sin sinsin sin

232

4 231

2

231

2 231

2

θ θθ θ


33

(22) the spatial components of the vectors are denoted as Mx, My, Mz and Ex, Ey, Ez,

respectively.

FTIR ATR polarization measurements utilizes the dichroic ratio R, a dimensionless relative

quantity, that can be determined conveniently. With ∆I ⊥,|| ∝ A⊥,||, equation (18) reads as

where mx y z, ,2 denote the ensemble mean values of the unit vector in direction of the

transition dipole moment. For a given ultrastructure and a given spatial orientation of the

transition dipole moment in the molecule R measures the orientation of the sample.

For a simple uniaxial orientation, equation (24) can be derived.

The segmental order parameter Sseg is frequently used to characterize molecular ordering and

also is related with cos2 α , as depicted in equation (26).

The calculation reveals Sseg = 1 for perfect alignment of the segmental axis with the z-axis

(α = 0°) and Sseg = -1/2 for perfect orientation in the xy-plane (α = 90°). For isotropic

(22) ( )∆I E M

E M E M E M E M E Mx x y y z z

∝ ⋅

∝ ⋅ ⋅ = + +

v v

v v v v

2

2 2 2 2... cos ( , ) ( )

(23) R

E m E m E E m m

E mx x z z x z x y

y y

=+ + ⋅2 2 2 2

2 2

2

(24) R

EE

EE

xr

yr

zr

yr

= +−

02

02

02

02

2

221

,

,

,

,

coscos

αα

(25) cos , ,

, , ,

2 0 0

0 0 02α =

⋅ −+ ⋅ −

R E EE R E E

yr

xr

zr

yr

xr

(26) S seg = −32

12

2cos α


34

arrangement one obtains cos2 α = 1/3 and α = 55°, thus Sseg = 0.

Values for the dichroic ratio R depend on the orientation and on the electric field. Isotropic

samples therefore can display different Rs depending on the refracting indices and the angle

of incidence. The angle of incidence is set to 45° and for Ge deviations are negligible,

because of its high refractive index. But the refractive index of the bulk phase water (n3) can

vary with the wavelength between 1.4 and 1.2 due to anomalous dispersion. Another crucial

point is the refractive index of the thin film (n2), which usually is taken as 1.45 for lipids.

Whereas, for compounds with a high amount of chlorine and for phenole usually higher

refractive indices are found (1.50 - 1.55 [37]). Therefore, we set the refractive index for

model membranes with adsorbed chlorophenol at 1.5.

Plots of the mean angle α and Sseg vs. the dichroic ratio R for systems under study and

wavenumbers of marker bands for lipids (νs(CH2): 2850 cm-1, bulk = air and bulk = water),

for the amid I band (1645 cm-1) and the phenoxide (1352 cm-1) and phenol (1080 cm-1)

marker bands are shown in the Appendix 7.1. For the calculation of the electric field

components Ex, Ey and Ez, the thin film approximation was used.

Once the orientation of the transition dipole moment with respect to the ATR plate has been

determined, mean molecular orientations can be stated if the position of the transition

moment in the molecule and the orientation of the molecules on the ATR plate are known.

The following discussion is focused on the liquid crystalline ultrastructure (LCU) because

lipid layers are characterized by this type of ultrastructure.


35

In the LCU model the molecules are randomly arranged around a space-fixed axis, e.g. the

acyl chains of lipid molecules are oriented around the z-axis (uniaxial orientation perpen-

dicular to the ATR plate). They are free rotating around the molecular axis and there can be a

fluctuation of the angle between the molecular and the space fixed axis (see Fig. 3).

If Sseg is interpreted as superposition of three uniaxial orientations, as depicted in equation

(27), it suits for the complex

(and fluctuating)

orientation described

above and pictured in Fig. 3.

Sseg is the product of an order

parameter for the side chains

(Ssc), one for the molecular

axis (Smol) and one for the tilt

angle (Stilt). The FTIR ATR

spectrum of DPPA/air in Fig.

6 shall be discussed as an

example. Comparing the ⊥-

spectrum and the ||-spectrum the CH2-stretching bands in the ⊥-spectrum are slightly larger

than in the ||-spectrum. This shows the high order of the DPPA monolayer, with the acyl

chains of the lipid perpendicular to the surface of the Ge plate. Because the transition dipole

moments of the CH2-stretching vibrations are in plane with the Ge-plate, no absorption in z-

direction occurs. An isotropic arrangement would lead to 1.17 larger ||-bands (n1= 4.0 (Ge),

n2= 1.45 (DPPA), n3= 1.0 (air), angle of incidence θ€= 45°).

2.2.3.4 Surface concentrations of Thin Films

z

x

y

ti ai

mi

θ α

Fig. 3. Superposition of three uniaxial distributions. mi : instantaneous orientation of the unit vector of the transition moment. ai : instantaneous orientation of the segmental molecular axis. ti : tilted average orientation of the molecular axis. Angles: θt: tilt angle between z axis and ti; θai : angle between ti and ai; θ: angle between ai and the transition moment mi . By means of polarized ATR spectra the mean angle α between the mi and the z-axes is measured.

S S S Sseg sc mol tilt= ⋅ ⋅ (27)


36

From either the adsorption spectra of parallel (||) or perpendicular (⊥) polarization the surface

concentration Γ can be calculated with the help of the effective thickness de,⊥,|| by equation

(28). It is possible to use either absorbances (peak heights) A ~, ,maxν ⊥ || and molar absorption

coefficients ε ν~max , or integral absorbances A d⊥ ||∫ ,

~

~

~νν

ν

1

2

, integrated between the wavenumbers ~ν1

and ~ν2 and integral molar absorption coefficients ε νν

ν

⊥ ||∫ ,~

~

~d1

2

, respectively. The number of

active internal reflections N and the number of equal absorbing groups ν are introduced into

equation (28), too.

For isotropic samples de simply resembles deiso and can be calculated from equation (15).

Whereas, for the quantification of oriented samples the ultrastructure must be known to

determine de. For the LCU model Wenzl et al. (for details see [3] and [34]) derived

expression (29) from equation (28), using the dichroic ratio R to relate de and deiso.

For the determination of A d|| ⊥∫ ,~

~

~νν

ν

1

2

spectra were usually taken as measured, because the

contribution of bulk signals was found to be within the experimental error. The thickness d

(28)

Γ = ⋅ =⋅ ⋅ ⋅

=⋅ ⋅ ⋅

⊥ ||

⊥ ||

⊥ ||

⊥ ||

∫

∫c d

AN d

A d

N d dee

~, ,max

~max

, ,

,~

~

, ,~

~

~

~

ν

ν

ν

ν

ν

νν ε

ν

ν ε ν

1

2

1

2

and d de e y, ,⊥ = with d d de e x e z, , ,|| = +

(29)

Γ =

⋅

⋅ ⋅ ⋅=

⋅

⋅ ⋅ ⋅ ⋅− + ⋅

⊥ ⊥∫

∫

∫

∫⊥ ⊥

A d d

d N d

A d d

d N d

E

ER

E

Ee e

iso

xr

zr

yr

zr

~

~

~

~

~

~

~

~~

~

~

~,

,

,

,, ,

ν

ν ε ν

ν

ν ε ν

ν

ν

ν

νν

ν

ν

ν1

2

1

2

1

2

1

2

32 0

2

02

02

02


37

was taken from published x-ray data, namely 2.5 nm/lipid layer and 9 nm for the height of

the mitochondrial creatine kinase or iteratively estimated for trichlorophenol studies. The

number of active internal reflections N was calculated from the dimensions of the Ge plate

and the SBSR cell. For the calculation of the relative electric field components of the three

layer system MIRE/membrane/bulk aqueous environment the thin film approximation and

equations (12)-(14) were used. Anomalous dispersion of the refractive index of the bulk

(aqueous buffer) was taken into account.

2.2.4 SBSR SET UP

One of the tricky problems of high resolved (difference) IR spectroscopy are the reliability of

reference spectra (background stability) and water vapor incompensations (long purging

times). To solve this problems the ATR setup was converted into a pseudo-double beam

instrument [38]. A special flow through cell (SBSR cell) subdivides the MIRE into two

compartments, which are independently accessible. A Single Beam is used to probe either

the Sample compartment or the Reference compartment (SBSR). This can be achieved by a

chopper, that covers sequentially one of the compartments. Another facility is represented by

a lift, that alternately changes the position of the plate, so that the beam passes either the

reference or the sample compartment. The FTIR-ATR-SBSR-setup used for the reported

investigations, is pictured in section 5.3.1 Long Term Measurements (Fig. 9) and 5.3.2 Time

resolved Measurements (Fig. 10).


38

3 Mitochondrial Membranes

39


3.1 STRUCTURE AND FUNCTION OF MITOCHONDRIAL MEMBRANES

3.1.1 ASSEMBLY AND COMPOSITION

Mitochondria are the only organelles of a cell with two membranes enclosing the so called

intermembrane space. Inner and outer membrane have different functions and thus differ in

their compositions. The inner membrane is highly folded into a series of internal ridges

called cristae bounding the mitochondrial matrix space. There, the citric acid cycle and fatty

acid oxidation occur. The products NADH and FADH2 are then used to generate ATP

(adenosine triphosphate) through the oxidative phosphorylation which takes place in the

inner membrane [39, 40]. Thus the assembly and composition of the inner membrane has to

guarantee the formation and maintenance of the pH gradient which enables ATP production.

The outer membrane separates the mitochondrion from the cytosol. Its function is the

controlled exchange of matter with the cytosol, the uptake of pyruvate and fatty acids and the

emission of ATP for a variety of purposes of the cell. Therefore, all necessary pore- and

transport- proteins as well as stabilizing ones are found. Outer mitochondrial membranes

have lipid compositions comparable to other organelles with about 50% of the total lipid

phosphorus coming from phosphatidylcholine (PC) and only some percents from cardiolipin

(CL). At pH 7 PC, or lecithin, is a zwitter ionic (neutral) lipid and CL, or

diphosphatidylglycerol, is one of the anionic phospholipids. Inner mitochondrial membranes

contain 20-30% of CL and low amounts of cholesterol [41, 42, 43]. As integral proteins the

enzymes of the respiratory chain, ATP-ADP translocase and other carriers are found.

Whereas, on the side facing the intermembrane space among others mitochondrial creatine

kinase (Mi-CK) is bound [4]. With the help of Mi-CK it is possible to convert excess of ATP


40

into phosphocreatine (PCr) which then is distributed and stored (details see 3.3 Energy

Management: Phosphocreatine-Circuit). To mimic the inner mitochondrial membrane we

used planar bilayers consisting of a monolayer of dipalmitoylphosphatidic acid (DPPA)

immobilized on the internal reflection element (a Germanium plate) and a CL monolayer as a

second leaflet. The CL was then exposed to a buffer solution of Mi-CK. It is known from the

literature that Mi-CK adsorbs readily to monolayers with high amounts of CL [6, 7]. To

study the interactions of lipids with the uncoupler 2,4,5-trichlorophenol we used a layer of

palmitoyloleoyl phosphatidylcholine (POPC) as outer leaflet. This resembles an outer

mitochondrial membrane quite well. Lecithins were also used in partitioning studies of

phenols between water and liposomes e.g. by Beate Escher et al.[44].

3.1.2 THE PH GRADIENT AND ATP PRODUCTION

In the following, the chemiosmotic theory is outlined. It was established for the first time by

Peter Mitchell in 1961 [45]. The enzymes of the respiratory chain transfer electrons step-by-

step to O2. Simultaneously, NADH-Q reductase, cytochrome reductase and cytochrome

oxidase release protons into the intermembrane space, thus generating a proton gradient. This

proton gradient exerts together with the membrane potential an electrochemical potential

difference of 0.224 V, called the proton motive force. The proton motive force is utilized by

the ATP synthase to synthesize ATP from ADP (adenosine diphosphate) and phosphate. Any

leakage of the membrane or any other mechanism which dissipates the pH gradient will lead

to an uncoupling of oxidation and phosphorylation of ADP to ATP.

3.1.3 UNCOUPLING

The proton gradient can be short-circuited by weak organic (hydrophobic) acids and bases

(e.g. 2,4-dinitrophenol, chlorophenols, and others) which carry protons across the inner

mitochondrial membrane (section 5). On the one hand they are useful tools in metabolic


41

studies. On the other hand, they destroy the energy supply of cells and organism and thus

they are toxic. The heat produced by the uncoupling of the oxidative phosphorylation can be

biologically useful. The process is utilized in the brown adipose tissues of hibernating

animals and newborn mammals and humans.

3.2 PENETRATION OF IONS AND POLAR SOLUTES INTO/THROUGH MEMBRANES

The efficiency of hydrophobic weak acids as proton shuttles depends on the ability of the

anion and the acid itself to permeate the membrane. Phospholipid membranes act as

permeation barriers for polar solutes and ions, thus restricting their flux. In general, the

permeability for a special entity is given by its permeability coefficient P, the product of the

partition coefficient Km and diffusion coefficient D within the membrane, divided by the

thickness of the membrane d, as depicted in equation (30) [46],[47].

Km depends on the free energy that is necessary to move the entity through the membrane,

the transfer energy ∆Gt. For a dipole molecule and an ion ∆Gt is ruled by the electric

potential energy profile of the membrane. Making use of the Boltzmann distribution, we

derive equation (31). cm and cw denote the concentration of the particle in the membrane and

in the bulk phase, respectively.

Six main contributions to ∆Gt are listed below.

1) The electrostatic work required to place an ion inside a bilayer membrane, i.e. the Born

(30) PK D

d=

⋅m

(31) Kcc

Gk Tm

m

w

t= = −⋅

exp

∆

∆Gq

r rBornm m w w

= −

2

081 1

πε ε ε (32)


42

energy ∆GBorn and the image energy. The first one is caused by the difference of the dipolar

polarization of the bulk phase (water) and the membrane. Whereas, the latter term describes

the perturbation of the electrostatic field exerted by the ions. The energy ∆GBorn, based on the

Born model, can be expressed with equation (32) by the charge of the ion q, the radius of the

ion in the membrane rm (approximately the bare ion), and in the water rw (including the

hydration shell), the vacuum permittivity ε0 and the relative permittivities for the

membrane εm and εw. The following term for the image energy given as in [48] and describes

approximately the perturbation of the ion exerted on the electric field.:

∆GI,ion is the image energy depending on the position x in the membrane of an ion with the

charge q and the radius r. The membrane has the thickness d and x has a value between

r and d-r. Further abbreviations: ϑε εε ε

=−+

w m

w m, with εm the relative permittivity of the

membrane (e.g. for hydrocarbons 2) and εw the relative permittivity of water (78.5) and d the

thickness of the membrane.

For a dipole molecule the analogous equation (34) was developed for the electrostatic energy

∆Gwm by Volkov [49]. Second, equation (35) resembles the image energy ∆GI,dip as derived

by Arekalian [and cited in 49].

∆GqkT x d n

rd

nxd

d nrd

nxd

n n

nn

n n

nnI,ion

m=

⋅⋅ ⋅

− ++

−+

+−

−−

=

∞

=

∞

=

∞

=

∞

∑∑ ∑∑2 2 2

112

2 2

1141 1 1ϑ

εϑ ϑ

ϑϑ ϑ

(33)

(34)

( )( )∆Gp

lwm w m

w m

=−

+ +

2

0312 2 1 2 1πε

ε εε ε


43

Here p is used for the dipolar moment and l for the effective dipole size. All other variables

have the same meaning as before.

2) Phospholipid membranes and biomembranes posses an internal dipole potential of +100

to +300 mV [50, 51] resulting most probably from oriented carbonyl groups of the

phospholipids. It decreases transfer energies for anions and increases them for cations.

Membrane potentials calculated by Flewelling [50] are depicted in Fig. 4.

3) Most biomembranes contain negatively charged phospholipids. They are electrically

neutralized by counterions which built up a surface potential at the membrane-water-

interface, that can be described with the Gouy-Chapman theory. Protons also function as

counterions at such negatively charged membranes and thus the local pH at the membrane-

water-interface is smaller than the bulk pH. This alters the apparent pKa values of any group

or molecule located at the surface. The difference can be predicted by the Gouy-Chapman

theory. Hydrophobic ions and amphiphilic molecules can be used to probe the surface

potential, if the intensity of signals for bound and unbound species are different and if the

surface potential modulates their binding.

( )∆Gp

l i

i

iI,dip

m= −

−

−

=

∞

∑212

22 1

2

03

2 1

31πε ε

ϑ

(35)


44

4) The transmembrane potential is defined as the difference in the electric potentials of the

two bulk phases separated by the membrane. It is caused by different permeabilities (due to

either semipermeable membranes, flux differences or active transport) for ions and their

counterion, respectively, leading to excess concentrations on one side and thus to a potential.

This potential can be described by a term based on the Nernst equation. Transmembrane

potentials are utilized to convert chemical energy into electric energy and vice versa. These

Fig. 4. Membrane potential components and total potential profiles for monovalent anions and cations. The profiles are based on equation (32) and equation (33). (A) combined energy profile (solid line) together with dipole potential profiles (broken line). (B) Resulting total potential energy profiles giving a good fit to the experimental data for tetraphenylphosphonium, a cation, and tetraphenylboron, an anion. Calculation and experiments were done by Flewelling [50]. Parameters used: half-thickness of the lipid bilayer tail region = 18 Å, headgroup region thickness = 8 Å, dipole layer location = 24 Å, distance between the point dipoles = 8.13 Å, dipole strength = 0.85 D, radius of the ion = 4.2 Å, εr

water = 78, εrlipid = 2, neutral energy (hydrophobic effect) = -7.5 kcal/mol.


45

„energized membranes“ allow to store energy, e.g. as proton motive force (see chapter 3.1.2

The pH Gradient and ATP Production).

5) There is a solvophobic contribution to the Gibbs free energy. It is gained through

entropic profit in the water phase if a hydrophobic ion or dipolar molecule moves into the

membrane. As an estimation Volkov et al. [49] calculated it from equation (36) in terms of

intersurface tensions using Uhlig´s equation.

A different approach was used by Flewelling [50], who used a function depending on the

position x of the ion with the radius r in the membrane. The membrane is characterized by

the half-thickness of the lipid bilayer tail region t and the headgroup region thickness h.

where WN0 is defined as the free energy of transfer for a uncharged entity of the same size and

the exponent xN = -2(t + h/4 - x)/r. This functional dependence for xN gives a neutral energy

transition region of the width 2r (ion radius) centered the effective electrical thickness

te = t + h/4.

6) And last but not least, often hydrogen bonds to the solvent have to be broken.

Obviously, by FTIR ATR measurements only adsorption to and insertion into the model

membrane can be observed. But these are the first steps of the whole permeation process.

Adsorption and insertion are insensitive for the transmembrane potential but are influenced

by all other phenomena that contribute to the Gibbs free energy.

(36) ( )γ γ γ γ γo,w o,m w,m w m− = ⋅ −sign

(37) ( )W xW

xNN

N=+

0

1 10


46

3.3 ENERGY MANAGEMENT: PHOSPHOCREATINE-CIRCUIT

For most of the energy requiring processes of organism ATP4- is the universal energy

currency in biological systems. ATP4-, normally bound to Mg2+, is hydrolyzed into MgADP-,

inorganic phosphate Pi and protons [39, 40]. The reaction, as depicted in Scheme 1,

generates a free energy of -28 kJ/mol at pH 7 and 25°C [52].

Because a lot of reactions are regulated by ATP and ADP, their concentrations are kept

between narrow limits (ATP 3-5mmol/L, ADP < 0.4mmol/L). Thus a phosphagen buffer

system is found at sites of high and fluctuating energy demands. In vertebrates PCr is used as

a potent phosphagen with a free energy of hydrolysis of -43.1 kJ/mol. It is formed at excess

of ATP, as shown in Scheme 2.

Adenosin-O

O-

O

O

P PP

OO

O

O

O- O-

O-

Mg2+

+ H2O

Adenosin-O

O-

O

O

P P

OO

O-Mg2+

O- + H+ +HO P

O

O-

O-

which can be written shortly as:

H+ +PiMgADP- +MgATP2-

Scheme 1. Hydrolysis of MgATP2- to MgADP- at pH 7.0.


47

Levels of PCr in the cytosol can be quite high. PCr is distributed and stored at energy

consuming sites and therefore it is used as temporal (for energy bursts) and spatial energy

buffer [53, 4]. The interconversion of ATP and PCr is carried out by creatine kinases.

Creatine kinases are a group of isoenzymes compartmentalized specifically at the site of

either energy production or consumption. In vertebrates soluble creatine kinases (CKc) can

be found in the cytosol as dimers MM and BB of the M (muscle) or B (brain) monomers or

as mixed dimer MB (Mr 80-86 kDa). On energy demand of muscles or nerves, the cytosolic

creatine kinases transfer phosphate from PCr to ADP. In contrast, the less soluble

mitochondrial creatine kinases (Mi-CK), which form octamers (Mr ca. 340 kDa), are located

on the inner mitochondrial membrane facing the intermembrane space. Their function is to

transfer phosphate groups from excess ATP to creatine, thus producing PCr. This keeps the

ATP concentration low at the inner mitochondrial membrane and facilitates the synthesis of

ATP by ATP synthase. ATP concentrations can be maintained at about the same level even

when huge amounts of ATP are consumed. Furthermore, the system prevents acidification

because the phosphorylation from PCr to ADP consumes protons, produced by hydrolysis of

ATP. The circuit, as depicted in Fig. 5 (taken from [4]), closes with the creatine (Cr) brought

H3C

COO-

H2C

NC

N

NH2

H

H3C

COO-

H2C

NC

NH2

NH2

O-

O-

O

P+ +

+ H+

MgATP2-

MgADP-

Mi-CK

CKc

Scheme 2. Phosphate transfer between MgATP2- and phosphocreatine (PCr). Creatine (Cr) is phosphorylated by mitochondrial creatine kinases (Mi-CK) using MgATP2- (left). Cytosolic creatine kinases (CKc) catalyze the regeneration of MgATP2- from MgADP- and PCr (right).


48

back to the mitochondrion, where it is phosphorylated again.

Fig. 5. The PCr circuit model for tissues with high and fluctuating energy requirement. The scheme is taken from the review of Wallimann et al. [4]. A complex regulated network is presented where the CK/PCR system is proposed to fulfil important functions (i) for ‘temporal energy buffering’, (ii) for ‘spatial energy buffering’, directing intracellular ‘energy flux’, (iii) for kjeeping the concentration of ADP low and preventing a net loss of adenine nucleotides, (iv) for removing protons and thus preventing acidification within a cell, and for undirectly stimulating glycogenolysis and glycolysis by Pi released as a consequence of net PCr hydrolysis, (v) for cotrolling local ATP/ADP ratios and thus also for increasing the thermodynamic efficiency of ATP hydrolysis. ATP is derived from to major synthetic pathways, oxidative phosphorylation (shown at the lower half) and glycogenolysis or glycolysis (shown at the left upper middle). Four major compartments for CK isoenzymes are indicated: (i) ‘cytosolic’ CK (CKc) functionally coupled to glycolysis on the producing side of the PCr circuit (at the left upper middle); (ii) mitochondrial CK (Mi-CK), functionally coupled to oxidative phosphorylation also on the producing side of the PCr circuit (at the lower half); (iii) ‘cytosolic’ CK, specifically associated (CKa) with subcellular structures at sites of high and fluctuating ATP utilization on the receiving side of the PCr circuit where it is functionally coupled to the corresponding ATPases (at top left) and (iv) strictly soluble cytosolic CK (CKc) freely equilibrating PCr/Cr and ATP/ADP pools of the cytosol (at the right upper middle). As represented schematically, the relative pool sizes of PCr and Cr are much larger than those of adenine nucleotides. On the mitochondrial side the functional and structural compartmentation of Mi-CK octamers at the inner/outer mitochondrial membrane ‘energy transfer contact sites’ are depicted. The Mi-CK octamer is shown here to interact with adenine nucleotide translocator (ANT) of the inner mitochondrial membrane (IM), possibly via an involvement of cardiolipin ( ), and with voltage -gated ion-selective pores (P) of the outer membrane (OM), to form an efficient multienzyme trans-membrane ‘energy-channeling’ complex. For further details see [4].

4 Preparation of Model Membranes on Internal Reflection Elements

49


4.1 PREPARATION OF THE SUBSTRATES

As substrates for the model membranes trapezoid Ge crystals (5 × 2 × 0.1 to 0.2 cm) or Si

plates (1 × 1 × 0.05 cm) were used. The Ge plates are the multiple internal reflection

elements (MIRE) for ATR measurements. Whereas, on Si plates small samples for atomic

force microscopy and electron microscopy were immobilized.

At the beginning of an experiment, each side of the Ge plate was polished by means of a 0.25

µm diamond paste for 10 min. Subsequently, the plate was cleaned with ultrapure water and

ethanol until there were no visible impurities left. In order to remove small traces of organic

compounds the Ge plate was cleaned for 3 min in a high-voltage glow discharge unit

(Harrick Sci. Corp.). The Ge plate was considered to be clean if the ν(CH2) bands at ~2920

and ~2850 cm-1 disappeared completely in the FTIR ATR spectrum (single beam mode).

Si plates were polished with paper towels saturated with ethanol or ultrapure water and put

into the plasma cleaner for 5 min. From Ge MIRE it is known that this procedure removes all

organic compounds and prepares the surface for DPPA immobilization.

4.2 PREPARATION OF MONOLAYERS

Monolayers were prepared by means of the Langmuir-Blodgett (LB)-technique described

first by Blodgett (1937) and Gaines (1966).

About 25µl of a 1 mg/ml dipalmitoylphosphatidic acid (DPPA)/CHCl3 solution was spread

on a film balance (NIMA technology, Coventry GB), filled with an aqueous subphase

containing 0.1 mmol/L CaCl2. After the solvent evaporated, the film was compressed to

30 mN/m and checked for 5 min for stability. The final area was about 100 cm2, the substrate

area about 20 cm2. Then the DPPA was transferred at 23 ± 2°C and at a surface pressure of


50

30 ± 0.2 mN/m, pulling out the Ge plate with a dipper speed of 2 mm/min (0.8 cm2/min).

Polarized FTIR ATR spectra of the monolayer against air were recorded and quantitatively

analyzed for surface concentration and orientation (reference: clean Ge/air). The checked

DPPA coated Ge was mounted in a flow through (SBSR) cell and filled with buffer for the

subsequent experiment.

4.3 PREPARATION OF BILAYERS

Symmetric DPPA bilayers were obtained from a second transfer of a DPPA film to the

monolayer. Accordingly, after the first LB transfer the Ge plate was completely pulled out

from the subphase. While the layer on the Ge-plate dried, the film was compressed to

30 mN/m and checked for stability again. Then the plate was dipped into the subphase

reaching a little beaker, set in the pit in the middle of the trough. For a successful transfer,

the meniscus of the water/lipid interface points below the level of the surface. As a rule,

dipping protocols and transfer ratios clearly show a good transfer of the second layer. But it

is necessary to prevent the bilayer from any contact with air or else it would collapse.

Therefore, after the trough was emptied the beaker with the plate was taken out of the film

balance and the plate was mounted into a flow through (SBSR) cell, all fully immersed in

buffer. FTIR ATR spectra of the bilayers were measured and the quantitative analysis of the

integrated absorbance (reference: clean Ge/buffer) at 2850cm-1 confirmed the bilayer

formation.

For the preparation of asymmetric bilayers with palmitoyloleoyl phosphatidyl choline

(POPC) as outer leaflet the Langmuir-Blodgett/vesicle method as described in [3] was used.

Alternatively, cardiolipin (CL) was used. A vesicle solution was prepared by sonication of

the lipid in an appropriate buffer (pH 7.0 or pH 6.0). 20 µl of a 10 mg/ml lipid/CHCl3


51

solution was dried in a small sample tube and about 3 ml buffer were added to give a final

lipid concentration of 0.7-1.0 mg/ml. The solution was sonicated under N2 purge for ½ h to 1

h at temperatures above the melting temperature of the lipid (POPC: 25°C<T<32°C, CL

(E.Coli): 35°C<T<42°C, CL(bh): 25°C<T<32°C). Solutions were almost completely

transparent and used within 20 min. They were slowly pumped (0.2 ml/min) over a DPPA

monolayer immobilized on the Ge plate. Then a bilayer was formed by spontaneous

adsorption of lipid from small unilamellar vesicles at adsorption temperatures of 18°C

(POPC and CL (E.Coli)) or 25°C CL(bh). Typically, the adsorption stopped after about 45

min. After 1 h the vesicles solution was exchanged by buffer solution. The formation of

multilayers, reported by Wenzl was found neither for POPC, nor for CL. Whereas, the latter

is due to electrostatic forces between the CL molecules, the first is most probably caused by

K+ ions and low ionic strength of the 25 mmol/L K-phosphate buffer (75 mmol/L KCl). Thus

the washing procedure, described in [3] was not found necessary. Every step of the bilayer

formation was checked in situ by ATR FTIR spectra, like the spectra of lipid layers pictured

in Fig. 6.


52

wavenumber / cm-1

1000120014001600180020003000

abso

rban

ce /

AU

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

CL/buffer

DPPA/air

POPC/buffer

Fig. 6. Parallel (||||||||) and perpendicular (⊥⊥⊥⊥) polarized FTIR ATR absorbance spectra of lipid layers under investigation. DPPA monolayers (middle) were transferred by the Langmuir-Blodgett technique at 30 mN/m from an aqueous subphase (0.1 mmol/L CaCl2) to a trapezoidal Ge plate. Γ = 3.9 ± 0.1×10-10 mol cm-2, corresponding to an area of 43 Å2 per molecule, which is close to the value of 43.8 Å2 reported by Demel et al. [66], dichroic ratio R = 0.93, order parameter Smol = 0.99 (reference: clean Ge plate; active internal reflections Nact = 43, integration between: 2867-2832cm-1). The signal shown was downscaled by f = 0.5 representing now the DPPA monolayer of one side of the Ge plate. Then bilayers were produced with either an outer leaflet of CL (top) or POPC (bottom) monolayer adsorbed from vesicles solutions (clipid = 0.6 mg/ml) to a DPPA monolayer; CL: Γ = 1.9 ± 0.1×10-10 mol cm-2; mean molecular area Am = 87 Å2, dichroic ratio R = 1.16, Smol = 0.61; active internal reflections Nact= 16.17; POPC: Γ = 2.3 ± 0.1×10-10 mol cm-2; mean molecular area Am = 74 Å2, in good agreement with 68 Å2, as reported by Evans [67]; dichroic ratio R = 1.33, Smol = 0.43 ; active internal reflections Nact= 18.36; (reference in both cases: DPPA monolayer in buffer) All surface concentrations were calculated with the thin film approximation (see 2.2.3.4 Surface concentrations of Thin Films) for the νs(CH2) at 2850cm-

1 with ε νd~∫ = 5.22×103 m mol-1 , with an angle of incidence θ = 45 and refractive indices n1= 4, n2= 1.43, n3 = 1 (DPPA/air) or 1.41 (CL/buffer and POPC/buffer);

5 Interaction of 2,4,5-Trichlorophenol with Planar Lipid Model Membranes

53


5.1 INTRODUCTION

5.1.1 CHLOROPHENOL-MEMBRANE-INTERACTION: NARCOSIS AND UNCOUPLING

Chlorophenols are an active component of a great number of pesticides. The quantities

applied coupled with their long degradation times makes them a major class of environmental

contaminants. In principal, their toxicity is caused by a narcosis based mechanism or by their

respiratory uncoupling potential. There exist a number of investigations comparing the

toxicity of different substituted benzenes and phenols on various biological systems to

elucidate quantitative structure-activity relationships (QSAR), e.g. [54], [55], [56], [57], [58].

The aim of these studies was to find simple, reliable and general molecular descriptors, based

on physical parameters like the octanol-water partition coefficient or the constant of Hammet,

that allows one to give an estimation for the toxicity of any compound without testing it on a

living system. The more that is known concerning the mechanistic details of the interaction

that affect the organism, the better is the chance to find the correct molecular descriptor.

Both narcosis and respiratory uncoupling are governed by the interaction between the

compound and the cell membrane. Whereas, narcotic agents alter the physical properties of

cell membranes through incorporation, uncoupling agents permeate membranes and destroy

the proton gradient essential for ATP production. As weak acids of high hydrophobicity, they

transport protons across membranes via dissociation at the proton poor side and proton uptake

at the proton rich side of the membrane ([59], [60]). From this point of view, the interaction of

uncouplers and model membranes, like phospholipid vesicles, black lipid membranes and

planar lipid layers, can be used to gain insight into molecular mechanism of the narcotic effect

([61], [62]) as well as of the uncoupling process.


54

5.1.2 KINETIC SCHEMES OF UNCOUPLING

In general, kinetic schemes of the uncoupling process consist of the following steps:

(i) partition of the uncoupling agent into the

membrane; (ii) translocation of the

protonated species through the membrane;

(iii) delivering of the proton at the interface

of the side with the high pH; (iv)

backdiffusion in the membrane and (v)

uptake of protons at the interface from the

side with the low pH to close the

protonophoric shuttle. Fig. 7 displays such a

kinetic scheme.

It is documented (e.g. in kinetic studies for

substituted phenols [2, 63], that

protonophoric uncouplers may exhibit a first order or second order kinetic and, therefore, they

are divided into class-1 and class-2 uncouplers. An explanation was given by Finkelstein [1],

who proposed that the second order kinetic is due to the formation of a heterodimer (AHA-),

e.g. consisting of a phenolic and a phenoxide entity for the chlorophenols (see Fig. 7 (ivb)). In

kinetic studies Escher et al. [2] found that the heterodimer predominates the uncoupling

process for 2,4,5-trichlorophenol (see Fig. 8). Therefore, we have chosen this compound to

prove spectroscopically the formation and existence of the phenol-phenoxide dimer.

membrane

A-

H+

HA

HA

AHA-

HA

H+

HA

AHA-AqueousphasepH1 < pH2

AqueousphasepH2 > pH1

HA

HA

(i)

(i)

(v)

(iii)

(ii)

(iva)

(ii)

(ivb)

A-

Fig. 7. Kinetic scheme of uncoupling. A hydrophobic weak acid HA permeates into the membrane -(i), (ii)- and dissociates at the interface to the bulk phase with a high pH value, releasing a proton (iii). Backdiffusion (iv) can be carried out either by an A-- or an AHA- -entity (with iva: class-1 uncoupler and ivb: class-2 uncoupler). The cycle closes with the uptake of a proton at the proton rich interface (v).


55

5.2 MATERIALS

KOH, KCl, CCl4 and n-hexane were obtained from Merck with p.a. grade. K2HPO4, KH2PO4,

dipalmitoylphosphatidic acid (DPPA) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

(POPC) were purchased from Fluka, whereas TCP came from Riedel de Haen with 99 %

purity. HPLC analysis confirmed that the given specification is correct. The remaining 1% are

mainly 2,4,6-trichlorophenol, tetrachlorophenols and dichlorophenols. Chemicals were used

without further purification. Water used was purified with an Elga filtering system to give a

specific resistance of 18.2 MΩ cm. The buffer employed was a 25 mmol/L potassium

phosphate buffer with a total concentration of 100 mmol/L K+ at pH 6.0.

Cl

Cl

Cl H

H

OH

1492 cm-1

1073 cm-1

δ9

8

3

2

7

1

5

4

6

11

10

13

12

γ

Fig. 8. Formula of 2,4,5-trichlorphenol (TCP). The numbers at the atoms in the scheme refer to the interpretation of the potential energy distribution given in Table 2. The position of dipole transition moments for 1492 cm-1 (calculated)/1488 cm-1(experimental) and 1073 cm-1 (calculated)/1080 cm-1 (experimental) as calculated from a potential energy distribution with B3LYP/6-311++ G(d, p). Angles to the axes O-C1-C4-Cl (dotted line) are denoted as γ (16°; 1492 cm-1) and δ (81°; 1073 cm-1).


56

5.3 METHODS

5.3.1 LONG TERM MEASUREMENTS

For long term measurements a SBSR-ATR-setup with a chopper as pictured in Fig. 9 was

mounted in a BRUKER IFS 25 FTIR spectrometer. A self-constructed SBSR cell (flow-

through cuvette) made of Delrin was used consisting of 2 compartments for each side of the

MIRE. The compartments are sealed from one another with Viton O-rings and shaped as

shown in Fig. 9 A (43 mm long and 7.5 mm broad). To prevent the adsorption of TCP to

hydrophobic surfaces on the one hand and contaminations from softeners a steal lid and glass

capillaries were used. TCP solutions of about 3€mmol/L in buffer were prepared by

sonication for 10 min and injected with a glass syringe. The adsorption process was monitored

with parallel polarized ATR spectra taken

every hour for 10-14 h. Spectra from

adsorbed TCP in the presence of TCP in

the bulkphase and of washed adsorbents

were taken. For filling and washing with

buffer solutions a peristaltic pump (flow

rate: 0.2 ml/min) was used.

5.3.2 TIME RESOLVED MEASUREMENTS

TCP solutions of concentrations between

1 and 3€ mmol/L in buffer were prepared

by sonication and pumped by means of a

peristaltic pump (flow rate: 0.5 ml/min) over either a DPPA monolayer or a DPPA/POPC-

bilayer (preparation see section 4; characterization see 5.4.3) Early experiments showed that

P

LS POL

Chopper

DET

SR

A

B

Fig. 9. SBSR-ATR setup for long term measurements. A) Sealing rings separate the sample (S) and the reference (R) compartment. B) Light path through a Ge plate mounted in a SBSR-flow through cell (LS: light source, POL: polarizer, DET: detector). Only one side of the lipid coated plate (opposite to the steal lid) was used for investigations.


57

Ismapren (material for the tubes) and Delrin (material for the cuvette) absorb high amounts

of TCP. A test of tube materials revealed that TCP does not adsorb to Viton. Thus it was

used in the later experiments together with a steal lid opposite to the Ge plate. Time resolved

experiments with Viton tubes were carried out twice, the first set with a conducting

connection between the Ge plate and the steal lid, the second time with an insulation of the Ge

plate. Potentials were measured at the end of experiments and found to be 0 V for the first

case and -0.43 V and -0.34 V for the insulated plate with DPPA monolayer and DPPA/POPC

bilayer, respectively.

Time resolved measurements were measured for 0.5 h with a BRUKER IFS66 making use of

a software controlled polarizer, which allowed recording of parallel and perpendicular

polarized spectra (15 each in turn) at intervals of one minute. The adsorption was monitored

until equilibrium and documented with a SBSR-ATR spectrum. This was achieved with a

SBSR-ATR-lift attachment, as sketched in Fig. 10. Then the TCP solution was exchanged for

buffer,

DET

ATR

M1

M2 M3

M4M5

M6M7

L

POLF

A

L

S

R

B

Fig. 10. Single-Beam-Sample-Reference (SBSR) ATR attachment. (A) The focus in the sample compartment is displaced to the position F by the planar mirrors M1 and M2. The off-axis parabolic mirror M3 produces a parallel beam with a diameter of one centimeter, i.e. half of the height of the MIRE. The cylindrical mirror M4 focuses the light to the entrance face of the MIRE. M5, which has the same shape as M4, reconverts it to parallel light by directing the light via the planar mirror M6 through the polarizer POL. The light is then focused to the detector DET by the off-axis parabolic mirror M7. (B) Alternating change from sample to reference is performed by computer controlled lifting and lowering of the ATR cell body.


58

the adsorbate was washed for about 15 min and a SBSR-ATR spectrum was taken.

During the whole experiment the pH and the concentration of TCP were monitored. As shown

in Fig. 11, the solutions were pumped with a peristaltic pump into the SBSR-ATR-cell.

Afterwards they passed the concentration control unit (ccu) and a micro-flow-through-

electrode (Hamilton/Orion). The ccu consisted of a flow through cell (standard IR-CaF2 cell

with a 200 µm spacer) mounted in a home built attachment with two mirrors and holders for

the waveguides. The waveguides were coupled to a diode array UV VIS spectrometer (Zeiss

specord S10).

For the analysis of the time resolved spectra with respect to TCP adsorption peak heights at

1352 cm-1 and 1080 cm-1 in the parallel and perpendicular spectra were determined. Data from

one run were fitted with a function f(t) of one of the following types: f(t)=g⋅(1-exp(-kt)),

f(t)=d+g ⋅(1-exp(-kt)), f(t)=d+g⋅(1-exp(-kt))+l⋅ t (g...scaling factor, d...offset, k...time constant,

l...slope). Fits were calculated with SigmaPlot 3.0. The type of f(t) was chosen by trial and

error to give the best fit (smallest Rsqr value). Usually, data points for up to 1000 min for

parallel polarization were measured. The matching perpendicular polarized data were

available by interpolation. The result was used to calculate the surface concentration Γ and the

dichroic ratios R for 1352 cm-1 (indicating strong bound phenoxide) and for 1080 cm-1

(indicating the loosely bound phenol) as a function of time. The impact of TCP to the lipid

layer was evaluated by integrating νs(CH2) bands at 2850 cm-1. The surface concentration Γ

and the dichroic ratios R were calculated. To determine the time dependence data were fitted

with functions of the type f(t)=d+g ⋅exp(-kt).


59

5.3.3 DETERMINATION OF MOLAR ABSORPTION COEFFICIENTS

5.3.3.1 IR: Transmission and ATR-Measurements

Molar absorption coefficients for the typical bands of the two species of TCP were first

determined from transmission measurements of solutions at pH 3 and pH 11. Therefore, TCP

was solved either in diluted HCl or diluted KOH to give concentrations between 0.5-

3 mmol/L and measured in a CaF2 cell with a 10 µm mylar spacer. The coefficients were

refined by measuring TCP solutions in potassium phosphate buffer at pH 6.0. Curve fitting

was used to separate the intense 1450 cm-1 band. Fitting parameters were checked to give the

correct phenol/phenoxide ratios, calculated from the dissociation constant pK 6.94 [2]. Also

ATR measurements of TCP solution at pH 6.0 and a clean Ge surface were used determine

molar absorption coefficients and to detect unspecific adsorption.

UV-Vis-spectrometer

TCP-sol.

Waste

pH-meter

concentrationcontrol unit

Det &Com-puter

POL

DetLS

LS

P

B

C

A

Fig. 11 Extended setup to monitor pH and concentration. After the sample passed the ATR cell (A) the concentration was monitored with VIS waveguides connected to a UV VIS spectrometer (B). A micro-flow-through-electrode was used to control the pH (C). Abbreviations: LS = light source, Det = Detector, P = peristaltic pump, POL = polarizer;


60

5.3.3.2 UV VIS

A 2.5 ml quartz cuvette with 1 cm pathlength was filled with 1800 µl -1980 µl buffer and

measured as reference. As standards for calibration 20-100 µl of a TCP stock solution were

injected with a pipette to give concentrations of between 9-100 µmol/L. The molar absorption

coefficients were evaluated for the peak heights at 293 nm and 204.5 nm (with D2 filter:

229 nm). To determine sample concentrations 5-20 µl sample was injected with a Hamilton

syringe into buffer, leading to concentrations of about 30 µmol/L. TCP solutions of

concentrations in the mmol/L range were performed using a IR standard cuvette with CaF2

windows and either a 0.1 or 0.2 mm spacer to keep the intense signals at 203 nm below 1

absorbance unit. UV VIS measurements were performed at room temperature by means of a

diode array spectrometer (Zeiss Specord S10) between 200 and 700 nm with 200

accumulations. An integration time of 50 ms and a resolution of 2.4 nm was used.

5.4 RESULTS

5.4.1 UV VIS MEASUREMENTS

UV VIS measurements were used to verify TCP concentrations and to check for association at

pH 6 and 11.3 and oxidation in diluted KOH with pH 11.0-11.9. The spectra, like the one for

pH 5.9 depicted in Fig. 53 (Appendix), show a small peak at 293 nm and a intense peak at 203

nm at pH 6.0. Deprotonation leads to peak shifts. At pH 11.1 the small peak was found at

312 nm and the intense one at 209 nm and a new peak was emerging at 244.5 nm. A perfect

Lambert-Beer behavior was found for concentrations between 0.01 to 3.0 mmol/L (in buffer at

pH 6.0) and 0.03 and 3.0 mmol/L (in diluted KOH at pH 11.3). Lambert-Beer plots using

peakheights at 293.2 nm, 229.3 nm (no maximum) and 203.3 nm Fig. 54 as well as 311.5 nm,

244.5 nm and 209 nm Fig. 55 are shown in the Appendix. For concentrations higher than


61

1.3 mmol/L at pH 6.0 the absorbance exceeds 1 AU, thus no linearity can be found

(d = 0.2 mm). Molar absorbance coefficients are listed in Table 1. Thus it seems unlikely, that

association of TCP in solutions with concentrations up to 3 mmol/L at pH 6.0 occurs.

Table 1. Molar absorption coefficients εεεε from UV VIS measurements. 2,4,5-trichlorophenol was dissolved in 25 mmol/L K-phosphate buffer pH 6.0 or diluted KOH pH 11.3, respectively.

pH 6.0 6.0 6.0 11.3 11.3 11.3 wavelength λ in nm

203.3 229.3 293.2 209 244.5 311.5

ε in cm2/mol

3.7±0.1 ×107

6.8±0.5 ×106

2.5± 0.1×106

2.9±0.2 ×107

9.0±0.7 ×106

4.3±0.2 ×106

Secondly, no changes in UV VIS spectra (see Fig. 56, Appendix) could be detected upon

oxidation with dry air for 2.75 h at ambient temperature of TCP dissolved in diluted KOH.

This proves that TCP solutions were stable within the experimental duration.

5.4.2 IR TRANSMISSION MEASUREMENTS

5.4.2.1 Phenol and Phenoxide Spectra

Transmission spectra of TCP solutions of acidic and alkaline pH values were used to identify

TCP as acid (HA) and as phenoxide (A-). The results are shown in Fig. 12. At pH 11.3 (Fig.

12 a) three intense absorption bands appear at 1455 cm-1, 1367 cm-1 and 1292 cm-1, which are

typical for the deprotonated state. These bands are absent at pH 3.0, i.e. in the protonated state

of TCP (Fig. 12 c). Furthermore, it was possible to show that the change in the FTIR spectra

are reversible acidifying an alkaline TCP solution or vice versa. Based on these results, the

ATR spectra of TCP adsorbed to a bilayer as shown in Fig. 12 b are interpreted as a

superposition of the acid and phenoxide spectra1. However, a frequency shift of about 10 cm-1

1 ) Materials for cuvette and tubings were chosen to guarantee that no contamination from the system could


62

of typical absorption bands of A- is observed, comparing A- in solution and adsorbed to the

membrane, respectively. Pronounced phenoxide formation at pH 6.0 was found to be specific

for lipid layers and could be detected neither on clean Ge nor on clean ZnSe surfaces, as is

depicted in Fig. 13 (a: TCP solution in CCl4; b: DPPA/POPC coated Ge; c: DPPA on Ge; d:

ZnSe; e: Ge;). Fig. 13 e reveals that the signals of TCP in the bulk were small compared to

adsorbed TCP and thus spectra were quantified without substraction of these signals.

The comparison of TCP dissolved in CCl4 with adsorbed and bulk phase TCP reveals shifts

and broadenings for a number of peaks, e.g. at 1400 cm-1and 1186 cm-1. Especially, the peak

at 1186 cm-1, prominent for TCP in CCl4, should be mentioned, because it is strongly reduced

or even erased in aqueous environment. This indicates a contribution of bending vibrations of

the phenolic OH group to these bands, which are sensitive to intermolecular and

intramolecular H bonding and solvent interaction.

For assignment of bands and determination of their transition dipole moments we performed

quantum chemical calculations with the GAUSSIAN 98 [64] suites of programs to gain

normal modes of TCP. Therefore, geometry optimization was executed at the B3LYP level of

density function theory and with a 6-311++ G(d,p) basis set.

occur. Enrichment of impurities (no more than 1%) of the TCP are very unlikely. By comparing the results with

spectra from the Aldrich IR spectra catalog the main components 2,4,6 TCP and tetrachlorophenols can be

excluded.


63

Table 2. Experimental and calculated wavenumbers for 2,4,5 trichlorophenol. Results of a quantum chemical calculation (BECKE3LYP/6-311++ G(d, p)). The experimental and calculated wavenumbers, calculated integral molar absorption coefficient ε νν

th d~ ~∫ (

calculated band intensity), potential energy distribution and the angle α for the dipole transition moment are given.

Amax (experimental)

[cm-1]

also found at

[cm-1]

Amax (calculated)

[cm-1]

ε ννth d~ ~∫

[km mol-1]

potential energy distribution2

angle α3 in °

1079 10804 10745

1073 74 δ(ring) 49%, -ν(C2Cl) 16%, ν(C6C1) 8%, ν(C2C1)7%

81

1130 1128b 1135 35 ν(C4C5) 19%, δ(CH) 18%+14%(-), -ν(C4Cl) 13%, ν(C5C6) 10%,-ν(C5Cl) 9%

-

1186 1208 136 δ(C1OH) 37%, ν(C1C6,2)17%+ 9%(-), -δ(CH) 13%+6%

25

1248 1260 16 δ(CH) 25%+23%, -ν(C4C3) 23%, -ν(C1C2) 8%, δ(C1OH) 7%

-

1283 30 ν(C1O)31%, -ν(C2C3) 29%,δ(C3H) 11%, -ν(C4C5) 8%

-

1327 85 ν(C5C6; C2C3) 23%+10%(-), -ν(C4C5,3) 18%+10%(-), δ(C1OH) 14%, ν(C1C2) 10%, -ν(C1O) 5%

-

1400 1419 15 ν(C5C6) 19%, δ(C1OH)16%, -ν(C4C3) 14%, ν(C1O) 7%, -ν(C2C3)7%

-

1462 to 1466 (BFA)

? -

1488 1492 260 δ(C3H) 22%, -δ(C6H) 15%,-ν(C2C3) 12%, ν(C1O) 11%, ν(C4C5,3) 10%+8% (-), ν(C1C6,2) 8%+6% (-);

-

1565b 1599 16 1600b 1627 3212/3214 ν(CH13,10):

89%+10% (-)/10%+89%

3767 ν(OH) 100% -... out of phase

2 ) numbers of atoms refer to Fig. 8, ν for stretching, δ for bending vibrations

3 ) α denotes the angle between the transition dipole moment and the molecular axes OC1C4Cl

4 ) 25 mmol/L potassium phosphate buffer

5 ) n-hexan and CCl4


64

The output was read by the program gar2ped (gaussian results to potential energy

distribution) [65], which calculated the normal modes and the potential energy distribution.

Results for the tricholorophenol are displayed in Fig. 14 A and B. The measured frequencies

for TCP in CCl4 match well with the calculated one, thus the result encourages the usage of

the calculated dipole transition moments for orientation analysis. The potential energy

distributions are given in Table 2.

Fig. 12. Comparison of IR TR spectra of 3 mmol/L TCP solutions at pH 11.3 and 3.0 with ATR IR spectra of TCP adsorbed to a bilayer. (a) 3.0 mmol/L TCP in diluted KOH pH 11.3; (b) TCP adsorbing from a 2.9 mmol/L solution in 25 mmol/L potassium phosphate buffer pH 6 (ctotal (K+) 100 mmol/L) to a DPPA/POPC bilayer; (c) 3.0 mmol/L TCP in diluted HCl pH 3.0; (b) can be interpreted as superposition of (a) representing A- and (c) representing HA. Peak shifts of about 10 wavenumbers between (a) and (b) may indicate changes in hydrogen bonds; TR measurement conditions: CaF2 cuvette with d = 10 µm, ambient temperature; ATR measurements: Ge as MIRE, active reflections Nact= 18.36, angle of incidence θ = 45°,T = 25°C; this parallel polarized ATR spectra was taken from an adsorption study with a DPPA/POPC bilayer. It was down scaled by f = 0.2.

wavenumber/cm-1

1000120014001600

abso

rban

ce/A

U

-0.002

0.000

0.002

0.004

0.006

0.008

a)

b)

c)

pH 11.3

pH 3.0

adsorbed to a bilayer


65

wavenumber/cm-1

1000120014001600

abso

rban

ce/A

U

0.00

0.01

0.02

0.03

a

b

cde

Fig. 13. Comparison of IR ATR absorbance spectra of 2,4,5-trichlorophenol (TCP) solutions in contact with different surfaces and IR transmission (TR) spectra of TCP in CCl4 (a). The parallel polarized spectra of TCP adsorbed to a DPPA/POPC bilayer (b), a DPPA monolayer (c; down scaled by f = 0.7), a clean ZnSe (d) and a clean Ge (e) show the preference of adsorption to lipophilic surfaces and the increased adsorption of phenoxide by lipid layers. In all cases the bulk is about 3 mmol/L TCP in 25 mmol/L potassium phosphate buffer pH 6 (ctotal (K+) 100 mmol/L). Compared to the TCP IR TR spectra adsorbed TCP reveals spectra with peak broadening (especially at 1400 cm-1) and peak shifts. Spectra with lipid layers are superpositions of the HA and A- spectra on the one hand and effects on the lipid layer spectra. TR measurement conditions: CaF2 cuvette, pathlength 10 µm, ambient temperature; ATR measurement conditions: either Ge (b, c, e) or ZnSe (d) as MIRE, active reflections Nact 18.36 (b, e), 16.15 (c) and 10.45 (d), angle of incidence θ = 45°,T = 25°C;


66

5.4.2.2 Molar Absorption Coefficients and Integrated Molar Absorption Coefficients

The peak at 1080 cm-1 was taken as a measure for loosely adsorbed TCP, i.e. phenolic TCP.

As a result of a normal coordinate analysis (see Table 2) the peak can be assigned to a

combined vibration, with 49 % of C-C bend vibrations („ring breathing“) and with 16% C-Cl

wavenumber/cm-1

10001200140016001800200030004000

0.01

0.02

0.03

0.04

-0.00

abso

rban

ce/A

U A

wavenumber/cm-1

10001200140016001800200030004000

calc

.ban

d in

tens

ity/1

03 ⋅ m⋅ m

ol-1

0

100

200

300B

Fig. 14. Experimental and calculated IR spectra of 2,4,5-trichlorophenol (TCP). (A) absorbance TR IR spectra of 10 mmol/L TCP in CCl4 (measurement conditions: CaF2 cuvette, pathlength 50 µm, ambient temperature); (B) IR band intensities ( ε νν~

~d∫ ) as calculated with BECKE3LYP/6-311++ G(d, p)


67

stretching vibration as major components. For the determination of ε νν~~d∫ ATR absorbance

spectra of standard solutions between 0.8 and 3.1 mmol/L at pH 6.0 were integrated between

1085 cm-1 and 1074 cm-1. Peak heights were found to be less influenced by the baseline and

the signal-noise-ratio. Therefore, ε ν~ were determined for peak heights measured from a

straight line between 1092 cm-1 and 1067 cm-1, and used whenever the spectra displayed

nonlinear baselines and/or small signal-noise-ratio.

The prominent peak at 1446 cm-1, indicating phenoxide formation and strong bound

adsorbates is located in a broad band. Therefore, the determination of ε νν~~d∫ started with a

curve fitting. The parameters for the curve fitting were tested to give a phenoxide to phenol

ratio for TR absorbance spectra of 0.11 as can be calculated from the pKa of 6.94 for the

solution of 2.2 mmol/L TCP at pH 6.0. Then the parameter set was used to separate and

integrate the peaks in TR absorbance spectra of TCP solutions at pH 11.0 and pH 6.0. The

following parameters (wavenumber/half width broadness/shape of the peak) were used:

1367 cm-1/9.5 cm-1/15%Lorentz+85%Gauss; 1405 cm-1 /45 cm-1/Gauss; 1455 cm-1/20 cm-

1/Lorentz; 1475.3 cm-1/15 cm-1/Gauss. From the slope of the linear regression of the resulting

integrals (CF Int) for 1455cm-1 plotted vs. the concentration (see Fig. 15) ε νν~~d∫ was

determined. However, using the ε νν~~d∫ for 1446 cm-1 the peak position was used as a flexible

parameter, due to shifts of peak maxima for adsorbed TCP.

The small and well separated peak at 1352 cm-1 was found always together with the one at

1446 cm-1.


68

Therefore, the ε νν~~d∫ and ε ν~ were

determined from the quantified

adsorbates, using signals

integrated between 1361 cm-1 and

1340 cm-1 or peakheights with a

straight baseline set between

1365 cm-1 and 1337 cm-1.

Quantification of TCP adsorbates

from time resolved measurements

were determined from peak

heights at 1352 cm-1. Molar

absorption coefficients and integrated molar absorption coefficients together with their

parameters are listed in Table 3. The integrated molar absorption coefficient of the 1446 cm-1

peak is 40 times higher than the one for the 1080 cm-1 peak.

Table 3. Molar absorption coefficients and integrated molar absorption coefficients.

Amax at [cm-1]

shifts to used to quantify (species)

ε ν~ in cm2mol-

1

straight baseline set between:

ε νν~~d∫

in cm mol-1

integration method:

1080 1074 (in n-hexan)

HA 1.85±0.08 ×105

1092 -1067 cm-1

7.2± 0.6 ×105

1085 - 1074 cm-1(6)

1352 1367 (in diluted KOH)

A- 1.5±0.2 ×105

1365-1337 cm-1

1.4± 0.2 ×106

1361-1340 cm-1(6)

1446 1455 (in diluted KOH)

A- --------------

------------- 2.88±0.06 ×107

after curve fitting7; at 1455.5 cm-1

6) Peak area above a straight baseline with limits as noted.

7 ) peak shape: Lorentz; HWB:19.9 cm-1

y = 0.0294xR2 = 0.9986

00.020.040.060.080.1

0.120.140.160.180.2

0 2 4 6 8

c/mM

CF

Int/A

Ucm

-1

Fig. 15. Data points and linear regression for the determination of ε νν~

~d∫ . 4 TCP solutions in dilute KOH were measured in a CaF2 cell (d = 10.3 µm). The peak at 1455 cm-1 was separated from the band between 1510 cm-1 and 1320 cm-1 with curve fitting (CF) and the resulting integrals (CF Int) were plotted against the concentration.


69

5.4.3 CHARACTERIZATION OF THE PREPARED MODEL MEMBRANES

Lipid layers were quantified by integrating νs(CH2) bands between 2867-2832 cm-1 (DPPA) or

between 2869.9-2832 cm-1 (POPC) and applying the thin film approximation as described in

section 2.2.3 Quantitative Determination using an integrated molar absorption coefficient

ε νd~∫ of 5.22 ×105 cm/mol [3]. Order parameters were calculated as given in section 2.2.3.3

Determination of Sample Orientation. The LB-technique leads to DPPA monolayers with high

reproducible surface concentrations Γ of (4.2 ± 0.3) ×€10-10 mol/cm2, corresponding to an

area of (40 ± 3) Å2 per molecule. For the molecular order parameters Smol values of

1.00 (± 0.03) were determined, representing perfect xy-orientation for the transition dipole

moment of the νs(CH2) and z-orientation for the molecular axis. These results are in good

agreement with the value given by Demel et al. [66], who determined 43.8 Å2 per molecule at

the air-water interface at pH 7.0. Integrated absorption of parallel (int App) and perpendicular

(int Avp) polarized spectra and calculated results, as well as mean values (MW) and standard

deviations (σ) are listed in Table 7 (Appendix). Values of integrated absorption depend on the

number of active reflections Nact, and thus on the thickness of the plate and the length of lipid

coverage.

The results for the preparation of the second leaflet of POPC scatter within 14% for Γ and

11% for the dichroic ratio R. For the calculated areas per POPC molecule values in the range

of 86 ± 14 Å2 were found. Values between 68 and 75 Å2 are expected from the literature [67],

thus the results indicate that some of the second leaflets were loosely packed. The order

parameter S reflects less order in the POPC layers than for DPPA monolayer. There is a kink

in the chain, because one of the acyl chains of POPC is unsaturated. Furthermore, there is

more fluctuation in chain conformations because the preparation was achieved at a

temperature above the melting temperature (25°C > tm = -5°C ). However, the fact that the


70

high sensible order parameter Smol ranges from 0.2 to 0.4, must be additionally due to the

course of preparation. Thus the results rather characterize the individual preparations than the

preparation technique, which clearly needs further optimization to suite the electric potential

(Ge facing a steal lid), K+ ions and low ionic strength of the buffer system. Data in Table 7

(Appendix) were listed for each of the compartment separately to show the reproducibility of

one run, which is an essential prerequisite for background compensation.

5.4.4 LONG TERM MEASUREMENTS OF LIPID-TCP INTERACTION

5.4.4.1 Interaction with DPPA Monolayers

Fig. 16 displays SBSR absorption spectra of 2,4,5-trichlorophenol (TCP) adsorbed to a DPPA

monolayer after 1 h (a) and 12 h (b) of exposure. The prominent peak at 1446 cm-1 is

accompanied by shoulders at 1487 cm-1 and 1462 cm-1. Spectra (a) and (b) display a broad

shoulder at 1400 cm-1 and small well separated peaks at 1352 cm-1, 1080 cm-1 and 1060 cm-1.

Two bands are localized at 1248 cm-1 and 1127 cm-1. All but the peak at 1400 cm-1 are

observed from the beginning and remain, if the adsorbates are washed with buffer (a, b, c).

Especially, the peaks at 1446 cm-1 and 1352 cm-1 are unchanged. However, the difference

spectra (d) displays major decreases for all other bands. Fig. 17 depicts the C-H stretching

region of the spectra a, b, and c of Fig. 16. SBSR absorption spectra resemble difference

spectra of the disturbance of the monolayer through TCP interaction. There is no change of

the CH2 stretching signals after washing as shown by Fig. 17 c. The parallel polarized (||)

spectra show sigmoid band shapes, whereas, the perpendicular (⊥) polarized spectra show

negative bands at 2915 cm-1 and 2848 cm-1. Obviously, a shift of the ν(CH2) bands to higher

wavenumbers occurs upon TCP interaction. Both figures show that there is no surface

concentration enhancement due to longer exposure.


71

wavenumber/cm-1

10001200140016001800

abso

rban

ce/A

U

-0.02-0.010.000.010.020.030.040.050.06

c

b

a

d

Fig. 16. Polarized IR ATR absorbance spectra between 1850 and 900 cm-1 of 2,4,5-trichlorophenol (TCP) adsorbed on a DPPA monolayer. Parallel (||||||||) and perpendicular (⊥⊥⊥⊥) polarized spectra after 1 h (a) and 12 h (b) exposure of the monolayer to a 2.9 mmol/L 2,4,5-TCP solution. The TCP solution was replaced with 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100mmol/L) and parallel (||||||||) and perpendicular (⊥⊥⊥⊥) polarized spectra were measured (c). A prominent peak at 1445 cm-1 (dashed line) was found. Only a small decrease of phenoxide and phenol after removing the TCP solution can be detected. Whereas, the band at 1400 cm-1 diminishes. For (c) the thin film approximation yields a surface concentration of TCP (HA) ΓHA at 1080 cm-1 4.8×10-10 mol cm-2 and of TCP- (A-) ΓA- at 1446 cm-1 5.9×10-10 mol cm-2 (HA/A- ratio 0.8). (d) displays the parallel (||||||||) and perpendicular polarized (⊥⊥⊥⊥) difference spectra (c)-(b). Measurement conditions: trapezoidal Ge as MIRE, angle of incidence θ = 45°, active reflections Nact = 16.15; refractive indices: n1= 4.0 (Ge), n2= 1.50 (membrane), n3= 1.31 (H2O, 1446 cm-1) and n3= 1.26 (H2O, 1080 cm-

1).


72

5.4.4.2 Interaction with DPPA-POPC-Bilayers

The FTIR ATR SBSR absorbance spectra of TCP adsorbing to a DPPA/POPC bilayer for

57 min, 10 h and 16 h time are shown in Fig. 18. Prominent bands of TCP are localized at

1487 cm-1, 1462 cm-1, 1400 cm-1, 1248 cm-1, 1200 cm-1, 1127 cm-1 and 1080 cm-1. With

increasing time a remarkable band appears at 1446 cm-1. Replacing the TCP buffer solution by

pure buffer leads to spectra b of Fig. 19. Most primarily observed bands of TCP have vanished

to a great extent, except the 1446 cm-1 band. A pure DPPA/POPC lipid bilayer in contact with

wavenumber/cm-1

28002850290029503000

abso

rban

ce/A

U

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

c

b

a

Fig. 17. Effect on the CH2 stretching vibrations of the DPPA monolayer by TCP adsorption. SBSR spectra of a DPPA monolayer exposed to 2.9 mmol/L 2,4,5-TCP solution as sample (S) and a DPPA monolayer against buffer as reference (R) directly provide the difference signal generated by the influence of TCP on the acyl chains. Parallel (||||||||) polarized spectra indicate a peak shift, whereas and perpendicular (⊥⊥⊥⊥) polarized spectra show a decrease. There is no changing in time (a: after 1 h, b: after 12 h) or with replacing TCP with buffer (c). The dashed line marks the Amax of νas(CH2) of DPPA. Measurement conditions: Ge trapezoid as IRE, angle of incidence θ = 45°, active reflections Nact = 16.15, (a) and (b): bulk concentration of TCP = 2.9 mmol/L; (c): 25 mmol/L potassium phosphate buffer pH 6 (ctotal (K+) 100 mmol/L) as bulk;


73

pure buffer served as reference (SBSR mode). The C-H stretching regions exhibit significant

differences with respect to the shapes of parallel polarized (||) and perpendicular (⊥) polarized

spectra. Sigmoid band shapes appear with ||-polarized light, while ⊥-polarized light leads to

negative CH2-stretching bands. The corresponding minima of the νas(CH2) and νs(CH2) are

found at 2919 cm-1, and 2850 cm-1, respectively, i.e. closer to the wavenumbers observed with

DPPA (νas(CH2): 2917cm-1; νs(CH2): 2850cm-1), than with POPC (νas(CH2): 2923 cm-

1; νs(CH2): 2853 cm-1). Furthermore, negative bands at 1741 cm-1 (ν(C=O), ester group), and

at about 1100 cm-1 (superposition of νs (PO2-) and ν (P-O-C), phosphatidic ester/acid group)

can be seen [9].

wavenumber/cm-1

1000120014001600180020003000

abso

rban

ce/A

U

-0.01

0.00

0.01

0.02

a

b

A

Fig. 18. Polarized IR ATR absorbance spectra of 2,4,5-trichlorophenol (TCP) adsorbing to a lipid bilayer. (a) Parallel (||) polarized and perpendicular (⊥⊥⊥⊥) polarized spectra after 1 h and (b) parallel (||) polarized spectra after 10.3 h. Calculated surface concentration of the phenol (TCP (HA); 1080 cm-1) ΓHA = 1.1×10-9 mol cm-2; dichroic ratio R (1080 cm-1) = 1.7; Measurement conditions: Trapezoidal Ge MIRE, angle of incidence θ = 45°; active internal reflections Nact = 18.36; refractive indices: n1= 4.0 (Ge), n2= 1.50 (membrane), n3= 1.31 (H2O, 1446 cm-1) and n3= 1.26 (H2O, 1080 cm-1).


74

wavenumber/cm-1

1000120014001600180020003000

abso

rban

ce/A

U

-0.01

0.00

0.01

0.02a

b

Fig. 19. Comparison of polarized IR ATR absorbance spectra of a lipid bilayer in contact (a) and after the contact (b) with a 2.9 mmol/L 2,4,5-trichlorophenol (TCP) solution. (a) Parallel (||||||||) and perpendicular (⊥⊥⊥⊥) polarized spectra of the DPPA/POPC bilayer; Calculated surface concentrations for phenol (TCP (HA); 1080 cm-1) and phenoxide (TCP- (A-); 1446 cm-1): ΓHA = 1.14 ± 0.22 ×10-9 mol cm-2 and ΓA- = 1.6 ± 0.1 × 10-10 mol cm-2. (b) TCP solution replaced with 25 mmol/L potassium phosphate buffer pH 6 (ctotal (K+) 100 mmol/L). The peak at 1446 cm-1 indicates phenoxide and is marked with a dashed line. Calculated surface concentrations for phenol (TCP (HA); 1080 cm-1) and phenoxide (TCP- (A-); 1446 cm-1): ΓHA = 1.6 ± 0.3 × 10-10 mol cm-2 and ΓA- = 1.9 ± 0.1 × 10-10 mol cm-2. Thus AHA- -heterodimers may be retained in or on the bilayer. Measurement conditions: Trapezoidal Ge MIRE, angle of incidence θ = 45°; active internal reflections Nact = 18.36; refractive indices: n1= 4.0 (Ge), n2= 1.50 (membrane), n3= 1.31 (H2O, 1446 cm-1) and n3= 1.26 (H2O, 1080 cm-1).


75

5.4.5 TIME RESOLVED MEASUREMENTS OF LIPID-TCP INTERACTION

5.4.5.1 Interaction with DPPA Monolayers

The adsorption of TCP to DPPA monolayers was monitored in situ with parallel and

perpendicular polarized spectra for concentrations between 1 and 3 mmol/L. Fig. 20 displays

the resulting spectra for 2.15 mmol/L (-0.42V). The peak at 1446 cm-1 (with its shoulders)

increases together with the peaks at 1352 cm-1, 1080 cm-1 and 1060 cm-1. The latter three have

molar absorption coefficients of similar magnitude. At the equilibrium (after 3 h) the two

peaks at 1080 cm-1 and 1060 cm-1 are of about the same size, although in the beginning the

higher peaks are found at 1080 cm-1. The peak at 1650 cm-1 is a small incompensation of the

water δ(OH) band maybe caused by a small air bubble. As depicted in Fig. 21, the

determination of the surface concentration Γ(HA) and Γ(A-) reveals that the first one is

reaching its saturation value slightly faster than the latter one. These saturation values are of

about the same value for both species, resulting in a HA/A- ratio of 1.3 (150 min).

Spectra similar to those of Fig. 20 together with a series of SBSR spectra as Fig. 57

(Appendix) were obtained for 0.8 (-0.42V), 0.95 (0V,i), 1.5 (-0.42V), 2.15 (-0.42V) and

2.9 mmol/L (0V,i). Peak heights at 1080 cm-1 and 1352 cm-1 from these spectra were used to

determine saturation values of the surface concentrations of adsorbed HA (ΓHA) and A- (ΓA-)

at equilibrium with dissolved TCP as well as of strongly bound adsorbates. Values for ΓHA

and ΓA- after the contact of 2.15 mmol/L TCP for 1 h were found to be 6.3€×€10-10 and

5.5 ×€10-10 in the presence of TCP, and 4.9 ×€10-10 and 4.4 ×€10-10 in the absence of TCP.

Therefore, the ratio ΓHA / ΓA- is 1.1 in both cases and 80% of the adsorbed TCP is retained

after the washing procedure. Saturation values with and without TCP in the bulk solution for

all experiments are listed in

Table 4 (Appendix). The surface concentrations were also calculated as a function of time and


76

the results are shown and in Fig. 21 (for 2.15 mmol/L; results for 1.5 (-0.42V) and 2.9 mmol/L

(0V,i) can be found in the Appendix: Fig. 58, Fig. 59). A similar time behavior for both

species is obeserved. To obtain the curves, peak heights at 1352 cm-1 and 1080 cm-1 were

fitted with functions of the typ f(t) = d + g⋅(1-exp (-k⋅t)) + l⋅t. The resulting fit parameters,

which are listed in

Table 8 (Appendix), can be assigned to the following meanings: d is the onset-signal from

adsorbed TCP from the previous run, g is a scaling factor, k is the time constant of the signal

increase coupled to the adsorption process and l is a linear increase or decrease of the signal,

caused by a slow overlaying process. The linear factor l was only essentially for long term

exposure.

wavenumber / cm-1

1000120014001600180020003000

abso

rban

ce /

AU

0.00

0.02

0.04

0.06A

2 min

10 min

18 min

29 min

143 min

after washing


77

wavenumber / cm-1

1000120014001600180020003000

abso

rban

ce /

AU

0.00

0.02

0.04B

3 min12 min20 min30 min

153 min

after washing

48 min

Fig. 20. IR ATR absorbance spectra of the adsorption of TCP to a DPPA monolayer. A TCP solution (cTCP= 2.15 mmol/L) in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L) was pumped into a flow through cell8 with 0.5 ml/min at 25°C. For 0.5 h parallel (||) polarized and perpendicular (⊥) polarized ATR spectra were measured in turn every 64 s. After about 1 h the TCP solution was exchanged for buffer and the adsorbate was washed for 15 min. Parallel (||) and perpendicular (⊥) polarized SBSR-ATR spectra were recorded. (A) parallel (||) polarized absorbance spectra of the washed adsorbate (top) and of the adsorption process arranged by ascending times from the bottom to the top (2 (bottom), 10, 18, 29, and 143 min). (B) perpendicular (⊥) polarized spectra of the washed adsorbate (top) and of the adsorption process arranged by ascending times from the bottom to the top (3 (bottom), 12, 20, 30 and 153 min). A dashed line marks the peak at 1446 cm-1. Reference: DPPA monolayer against 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L); Measurement conditions: Ge trapezoid as MIRE, angle of incidence θ = 45°, active internal reflections Nac t= 13.46;

8Cell with steal lid. A potential of -0.42 V was measured (Ge as kathode).


78

Regarding the results gained for 2.9 mmol/L (0 V, i) and 2.1 mmol/L (-0.42 V), the time

constants k for both polarizations are of the same size in each case within the error

methodically implicated. They are only slightly different for both wavenumbers 1352 cm-1 and

1080 cm-1. Thus, there is an isotropic adsorption of both species to the DPPA monolayers

with a k for 1352 cm-1 (A-) between 0.02 and 0.03 min-1 and a k for 1080 cm-1 (HA) between

0.04 and 0.06 min-1 from TCP solutions with csol(TCP) > 2 mmol/L. For the 1.5 mmol/L TCP

solution peaks were quite small and thus the low signal-to-noise ratio prevents reasonable

interpretation of the fit results.

t / min0 30 60 90 120 150 180

surfa

ce c

once

ntra

tion

/ 10-1

0 mol

·cm

-2

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Fig. 21. Surface concentrations of phenol (HA) and phenoxide (A-) on a DPPA monolayer vs. time. The peak heights at 1080 cm-1 and 1352 cm-1 were determined and fitted. The fit results were used to calculate the surface concentration Γ with the thin film approximation. Curves were determined with the help of f(t), whereas squares (for HA) and circles (for A-) denote distinct values, calculated with measured and interpolated peak heights. Reference: DPPA monolayers against buffer. Molar absorption coefficients: ε HA (1080 cm-1) 1.85 ± 0.08 ×€105; ε A- (1352cm-1) 1.5 ± 0.2 ×€105; angle of incidence θ = 45°, active internal reflections Nact= 13.46; T = 25°C, U = -0.42V; concentration of TCP csol: 2.1 mmol/L. The dotted lines represent the standard deviations of the curves (includes standard deviation of fit parameters and of the molar absorption coefficient).


79

The CH2-stretching bands of the absorbance spectra (SCh of the sample compartment)

displayed in Fig. 20 indicate a loss of DPPA molecules of 20% of the DPPA monolayer.

These spectra were calculated using the initial DPPA monolayer for a series of three

concentrations (0.8, 1.5, 2.15 mmol/L) as reference. Because SBSR spectra represent

difference spectra for the lipid bands, we had to analyze spectra referring to the uncovered

plate/buffer. That was achieved by adding the spectrum of the DPPA layer (scaled to take the

air/buffer change and the lower Nact into account). Then the surface concentration was

calculated from the νs(CH2) as a function of time. Again matching of time was obtained by

interpolation of parallel polarized (||) data. The results are shown in Fig. 22, displaying a loss

of 3% of the lipid molecules accompanied with a decrease of the dichroic ratio from 0.90 to

0.88 during the exposure to 2.15 mmol/L TCP. These results are equivalent to the loss,

observed when pure buffer is pumped over a DPPA monolayer. Thus if data of the sample

compartment are compared with data of the reference compartment, no additional effect was

found. Similar analysis was done for every experiment and results were found to be in the

same range, except when the time of exposure was several hours. Table 10 (Appendix)

summarizes surface concentration Γlipid, dichroic ratios R and order parameters Smol for

different times for the two compartments, separately.


80

t/min0 20 40 60 80 100 120 140 160 180

Γ lipid /

10-1

0 mol

·cm

-2

1.0

2.0

3.0

4.0

t/min0 20 40 60 80 100 120 140 160 180

dich

roic

ratio

R

0.86

0.87

0.88

0.89

0.90

0.91

0.92

Fig. 22. Surface concentration and dichroic ratio of DPPA during the adsorption of TCP vs. time. The DPPA monolayer was exposed to 2.15 mmol/L TCP in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal(K+) 100 mmol/L). The integrals of νs(CH2) were determined between 2867 cm-1 and 2832 cm-1 and fitted with f(t)=d+g ⋅ ⋅ ⋅ ⋅(exp(-k ⋅ ⋅ ⋅ ⋅t)). The fitted results were used to calculate the surface concentration Γ with the thin film approximation. Curves were determined with the help of f(t), whereas, filled circles denote distinct values, calculated with measured and interpolated integrals. (integral molar absorption coefficients ε νd~∫ =

5.22€×105 cm mol-1, angle of incidence θ = 45°, number of active internal reflections Nact= 13.46, T = 25°C, U = -0.42V) The dotted lines represent the standard deviations of the curves (includes standard deviation of fit parameters and of the molar absorption coefficient).


81

5.4.5.2 Interaction with DPPA-POPC-Bilayers

In Fig. 23 A and B ATR spectra measured during the exposure of a 2 mmol/L TCP to a

DPPA/POPC-bilayer (0 V, i) are shown, representative of results observed with different TCP

concentrations. Typically a broad band between 1500 and 1400 cm-1 arises together with a

sharp peak at 1080 cm-1. Then a peak at 1446 cm-1 is emerging. Spectra were analyzed with

respect to the peaks at 1080 cm-1 (for the phenol) and at 1352 cm-1 (for the phenoxide), using

f(t)=d+g(1-exp(-k⋅t) to fit the data. The time constant k shows only small differences fitting

either the parallel (||) or perpendicular polarized (⊥) spectra, but a significant one between the

two species. For 1080 cm-1 we found 0.249 min-1 (||), 0.263 min-1(⊥) and for 1352 cm-1

0.048 min-1(||), 0.059 min-1(⊥). In Table 9 (Appendix) these results are listed together with

parameters gained from three similar experiments. The functions f(t) were then taken to

calculate the Γ curves, depicted in Fig. 24. Distinct values for Γ were calculated from ⊥- data

(taken from the ⊥-polarized spectra, as displayed in Fig. 23 B) and ||-values from an

interpolation using the fit result from data points taken from ||-polarized spectra (as displayed

in Fig. 23 A). The same procedure was used to analyze the two further experiments. Data are

listed in

Table 4, Table 9 (Appendix) and Table 10 (Appendix).

The νs(CH2)-band in the ATR spectra were integrated and the effect of TCP on the bilayer

was analyzed. For the determination of the dichroic ratio R the spectra of Fig. 23 A and B

(reference: bilayer before absorption/buffer) could not be used directly, because for the CH2-

stretching region they represent difference spectra. Furthermore, analysis of the signals in this

region were found to agree more likely to DPPA frequencies than to POPC frequencies.

Therefore, we had to analyze spectra referring to the uncovered plate/buffer. That was


82

achieved by adding the spectra of the POPC layer as well as the spectra of the DPPA layer

(scaled to take the air/buffer change and the lower Nact into account). The νs(CH2) of

||||||||-polarized and ⊥⊥⊥⊥-polarized spectra were integrated between 2867 cm-1 and 2832 cm-1. The

integrals were fitted with the function f(t)=d+g(-exp(-kt)) and the results were used to

calculate dichroic ratios R and the surface concentrations as functions of time. R(t) and Γlipid(t)

are shown in Fig 25. For distinct values the interpolated ||||||||-values and integrals of ⊥ ⊥ ⊥ ⊥-polarized

spectra were used. The surface concentration Γlipid as calculated with the thin film

approximation for the bilayer was 6.5 × 10-10 mol cm-2.

wavenumber/cm-1

1000120014001600180020003000

abso

rban

ce/A

U

0.00

0.02

0.04

0.06A

2 min6 min

11 min17 min

23 min

30 min

after washing


83

wavenumber/cm-1

1000120014001600180020003000

abso

rban

ce/A

U

0.00

0.01

0.02

0.03B

after washing

31 min24 min

18 min12 min

7 min3 min

previous page and above

Fig. 23. IR ATR absorbance spectra of the adsorption of TCP to a DPPA/POPC bilayer. A TCP solution (cTCP= 2.0 mmol/L) in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L) was pumped into a flow through cell9 with 0.5 ml/min at 25°C. For 0.5 h parallel (||||||||) polarized and perpendicular (⊥⊥⊥⊥) polarized ATR spectra were measured in turn every 64 s. After about 1 h the TCP solution was exchanged for buffer and the adsorbate was washed for 15 min. Parallel (||||||||) and perpendicular (⊥⊥⊥⊥) polarized SBSR-ATR spectra were recorded. (A) parallel (||||||||) polarized absorbance spectra of the washed adsorbate (top) and of the adsorption process arranged by ascending times from the bottom to the top (2 (bottom), 6, 11, 17, 23 and 30 min). (B) perpendicular (⊥⊥⊥⊥) polarized spectra of the washed adsorbate (top) and of the adsorption process arranged by ascending times from the bottom to the top (3 (bottom), 7, 12, 18, 24, 31 min). A dashed line marks a peak emerging at 1446 cm-1. Reference: DPPA/POPC bilayer against 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L); Measurement conditions: Ge trapezoid as MIRE, angle of incidence θ = 45°, active internal reflections Nac t= 13.46;

9Cell with steal lid. A conducting connection lead to a small current and a potential of 0 V.


84

Comparison with values determined for the layers separately after their preparation, show that

there is only a very small loss (see Table 7; ΓDPPA = 4.5 × 10-10 mol cm-2, R = 0.94; ΓPOPC=

2.3 × 10-10 mol cm-2, R = 1.6) due to exposure to TCP. Furthermore, this loss of lipids in the

t/min0 10 20 30 40 50 60 70su

rface

con

cent

ratio

n (H

A) /1

0-10 m

ol·c

m-2

0.0

5.0

10.0

15.0

20.0

surfa

ce c

once

ntra

tion

(A- )/

10-1

0 mol

·cm

-2

0.0

2.0

4.0

6.0

8.0

Fig. 24. Calculated surface concentrations ΓΓΓΓ for phenol and phenoxide on a DPPA/POPC bilayer vs. time. Open squares denote values for Γ(HA) and refer to the left axis. Whereas, filled circles indicate values for Γ(A-), referring to the right axis. The surface concentrations Γ were determined with the thin film approximation, using peak heights Amax at 1080 cm-1 for Γ(HA) and a molar absorption coefficient of ε€= 1.85 ±€0.08 × 105 cm2 mol-

1, as well as peak heights Amax at 1352 cm-1 for Γ(A-) with ε€= 1.5 ±€0.2 × 105 cm2 mol-1. Data from the peak heights were fitted with f(t)=d+g(1-exp(-kt)). These functions were used to calculate the Γ curves. For the calculation of discrete values for Γ see text above. Values after 57 min: ΓHA = 1.65 × 10-9 mol cm-2 and ΓA- = 0.43 × 10-9 mol cm-2. Dotted lines represent the standard deviation of Γ(t), derived from standard deviations of the molar absorption coefficient and the error of the fit/data relation. Reference: DPPA/POPC bilayer against 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L); Measurement conditions: Ge trapezoid as MIRE, angle of incidence θ = 45°, active internal reflections Nact = 13.46; refractive indices: n1= 4.0 (Ge), n2= 1.50 (membrane), n3= 1.31 (H2O, 1352 cm-1) and n3= 1.26 (H2O, 1080 cm-1).


85

sample compartment was comparable to the one found for reference compartment. The change

of the dichroic ratio R from 1.09 to 1.16 (a mean value for the bilayer) is also only slightly

higher than the effect we found for the reference compartment. Generally, the displacement of

lipid molecules was about twice what could be detected for DPPA monolayers for the whole

concentration series. These results are listed in Table 10 (Appendix).

t/min0 20 40 60su

rface

con

cent

ratio

n/10

-10 m

ol·c

m-2

2.0

4.0

6.0

8.0


86

t/min0 10 20 30 40 50 60 70

dich

roic

ratio

R

1.06

1.08

1.10

1.12

1.14

1.16

above and previous page

Fig. 25. Surface concentration and dichroic ratio of the ννννs(CH2) for a DPPA/POPC bilayer exposed to TCP. The bilayer was exposed to 2.0 mmol/L TCP in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal(K+) 100 mmol/L). The integrals of νs(CH2) were determined between 2869.9cm-1 and 2832 cm-1 after DPPA and POPC spectra had been added. Their values were fitted with f(t)=d+g ⋅ ⋅ ⋅ ⋅(exp(-k ⋅ ⋅ ⋅ ⋅t). The surface concentrations Γlipid (previous page) were calculated using the fit results and the thin film approximation. Curves were determined with the help of f(t), filled circles denote distinct values, calculated with measured and interpolated integrals. For the dichroic ratio R (above) the interpolated ||-values (fit result of f(t)=d+g(-exp(-kt)) were devided by the integrals of ⊥-polarized spectra (filled circles). Whereas, the curve represents the ratio of the two fit results: f(t,||)/f(t,⊥). (int. molar absorption coefficient ε νd~∫ : 5.22×105 cm mol-1, angle of incidence θ = 45°, active internal reflections Nact = 13.46; refractive indices: n1= 4.0 (Ge), n2= 1.45 (membrane), n3= 1.41 (H2O, 2850 cm-1)).


87

5.4.6 SURVEY AND COMPARISON

In the following a survey of the results of the various experiments described above are given.

One objective is to show correlations between the long term and the time resolved

measurements. The second one addresses the different adsorption of TCP to DPPA

monolayers and DPPA/POPC-bilayers.

Fig. 26 displays the calculated surface concentrations of adsorbed phenol HA in equilibrium

with the concentration of TCP in the solution (csol) as determined with UV VIS. The surface

concentrations of HA is a linear function of the concentration and independent of an

additional potential. The bilayer absorbs 2.6 times the amount of HA of a monolayer. There is

a reasonable agreement between long term experiments and the time resolved runs for the

phenol surface concentrations on mono- and bilayers.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

0 0.5 1 1.5 2 2.5 3 3.5csol / mmol/L

ΓΓΓΓ/10-9 mol cm-2

DPPA (0V,i)

DPPA (-0.42V)

Bilayer (0V,i)

Bilayer (-0.34V)

DPPA (long term)

Bilayer (long term)

Bilayer (1h, 0V, no i))

Fig. 26. Surface concentration Γ Γ Γ Γ of phenol as determined at 1080 cm-1 for mono- andbilayers. TCP concentrations of the solution (csol) as checked with an UV VIS spectrometer.Γ was calculated with the thin film approximation, using ε = 1.85 ±€0.08 ×€105 cm2 mol-1,θ = 45°, refractive indices: n1= 4.0 (Ge), n2= 1.50 (membrane) and n3= 1.26 (H2O, 1080 cm-

1). Linear regressions for the Γphenol of the bilayer (0V,i) and one for the DPPA monolayer(U=-0.42V) were drawn.


88

The comparison of phenoxide surface concentrations is more sophisticated. Data are shown in

Fig. 27. First, the bilayer clearly shows saturation for phenoxide surface concentrations of 5 -

6 ×€10-10 mol cm-2 at TCP concentrations (csol) higher than 1.5 to 2.0 mmol/L. Whereas, data

for the DPPA monolayer do not suggest a saturation up to 3 mmol/L. However, there are not

enough data to give a definite conclusion concerning the monolayers. The surface

concentration of phenoxide of the long term exposure to the bilayer is small compared to the

data measured in time resolved measurements.

Furthermore, one can see the influence of the potential and current applied on the adsorption

behavior of the phenoxide. The highest amount of phenoxide was determined in the absence

of a potential and of a current (Ge opposite a lid of Delrin). Using a steal lid, we determined

a negative potential for the insulated Ge plate between -0.3 and -0.4 V. The lowest amount of

phenoxide was found in case of the noninsulated plate at a potential of 0 V and a low current.

01.002.00

3.004.00

5.00

6.007.008.00

0 0.5 1 1.5 2 2.5 3 3.5csol / mmol/L

ΓΓΓΓ/10-10 mol cm-2

DPPA (0V,i)

DPPA (-0.42V)

Bilayer (0V,i)

Bilayer (-0.34V)

DPPA (long term)

Bilayer (long term)

Bilayer (1h, 0V, no i)

Fig. 27. Surface concentration Γ Γ Γ Γ of phenol as determined from the peak height at 1352 cm-

1 and integrals at 1446 cm-1 for mono- and bilayers. TCP concentrations of the solution (csol)as checked with an UV VIS spectrometer. Γ was calculated with the thin film approximation,using ε = 1.5 ±€0.2 ×€105 cm2 mol-1, θ = 45°, refractive indices: n1= 4.0 (Ge), n2= 1.50(membrane) and n3= 1.31 (H2O, 1080 cm-1).


89

This is only contradictory at first sight, because in the latter case there was an electrochemical

reaction of Ge and H2O to GeO and GeO2 supplying the system with protons. Thus the

phenoxide recombined to phenol to a certain amount instead of adsorbing to the lipid layer.

Surface concentrations of phenol (HA) and phenoxide (A-) were also determined after the

exchange of TCP solution for buffer to quantify strong bound TCP. While, surface

concentrations of HA adsorbed to bilayers are diminished, there is no effect of washing for

DPPA monolayers. The strong bound adsorbents always contain about the same amount of

phenol (HA) and phenoxide (A-). Additional amounts of phenol (HA) are loosely bound to the

bilayer and can easily be washed away. Data are listed in

Table 4.


90

Table 4. Survey of surface concentrations of TCP as phenol and phenoxide adsorbed to DPPA monolayers and POPC-DPPA bilayers.

Lipid layer (experimental condition)

time of exposure

in h

cTCP in mmol/L

ΓHA in 10-10 mol/cm2

ΓA- in 10-10 mol/cm2

HA/A- -ratio

DPPA (long term) 1 2.90 7.3 7.1 1.0 12 2.90 8.2 7.1 1.2 DPPA-TCP (long term) 12 - 4.8 5.9 0.8 DPPA (0 V, i) 0.5 0.95 0.8 0.5 1.6 1 2.90 4.4 3.8 1.2 12 2.90 7.1 6.3 1.1 16 - 7.0 4.9 1.4 DPPA (-0.42 V) DPPA (-0.42 V, 0.8 mmol/L)

0.5 - 1.5 0.9 1.7

DPPA (-0.42 V, 1.5 mmol/L)

0.5 1.50 2.3 1.4 1.6

1.3 - 2.6 2.0 1.3 DPPA (-0.42 V, 2.15 mmol/L)

0.5 2.15 5.4 4.1 1.3

1 2.15 6.0 5.3 1.1 2.5 2.15 6.3 5.5 1.1 2.5 - 4.9 4.4 1.1 DPPA-POPC (long term)

1 1.60 10.6 0.02

16 1.60 11.4 1.6 7.1 DPPA-POPC-TCP (long term)

16 - 1.6 1.9 0.8

DPPA-POPC (0 V, i) 0.5 1.05 9.9 0.8 12 0.5 2.00 16.4 3.7 4.4 1 2.00 18.4 4.7 3.9 1 - 5.3 3.9 1.4 0.5 2.70 26.3 4.2 6.3 1 2.70 25.6 4.7 5.4 13 2.7010 19.4 5.3 3.7 DPPA-POPC (-0.34 V) 0.5 0.80 6.8 3.0 2.3 0.5 1.45 12.0 5.5 2.2 1 1.45 12.4 6.0 2.1 1 - 5.3 5.2 1.0 0.5 2.00 14.5 6.0 2.4 1 2.00 14.5 6.1 2.4 1 - 6.3 5.2 1.2

10) TCP destabilized the bilayer, which lost 60% of its lipids during the exposure to 2.70 mmol/L TCP.


91

5.5 DISCUSSION

5.5.1 STACKS AND MULTILAYER FORMATION

To visualize the amount of TCP adsorbed to the layers, surface concentrations of TCP

monolayers can be calculated from the molecular size of TCP. The size was estimated from

interatomic distances O7-Cl11 (9.31 Å) and Cl9-Cl12 (9.87 Å), resulting from geometry

optimization, on the one hand, and the van der Waals radius of Cl (1.8 Å), on the other hand.

We assume the same size for phenol and phenoxide and an overall shape of an elliptic

cylinder for the molecules. Then a maximum and a minimum area per molecule were

estimated as Amax 62.5 Å2 per molecule (benzen ring parallel to the surface of the membrane

surface) and Amax 10.2 Å2 per molecule (benzen ring perpendicular to the surface of the

membrane). Therefore, the surface concentration of a monolayer of TCP (both entities

together) can be between 2.7 × 10-10 mol cm-2 or 1.6 × 10-9 mol cm-2, depending on the

orientation of the molecules. As listed in

Table 4, only for TCP bulk concentrations < 1 mmol/L and time of exposure of 0.5 h on a

DPPA monolayer, we found the sum of the surface concentrations (ΓHA + ΓA-) to be lower

than 2.7 × 10-10 mol cm-2. Most of the experiments yielded surface concentrations between the

two limits. Thus dense packing of tilted but upright oriented TCP molecules may occur. For

bulk concentrations of TCP of more than 2 mmol/L exposed to a DPPA/POPC bilayer the

surface concentrations exceeded 1.6 ×€10-9 mol cm-2. In these cases the formation of

multilayers took place. The loss of lipids is about 10% for DPPA monolayers and up to 38%

for DPPA/POPC bilayers. As shown in Table 10 (Appendix), it is not significantly increased

for the sample compartment compared to the reference compartment for the DPPA

monolayers. The figures show that TCP layers on the top of DPPA monolayers were formed.

Whereas, for DPPA/POPC bilayers and for higher TCP concentrations, as well as longer times


92

of exposure there is an increased loss of lipids for the sample compartment indicating an

embedding into the bilayers. This is supported by the fact, that POPC can displace TCP

adsorbed on the DPPA monolayer readily leading to about half the surface concentration of a

second leaflet formed on a pure DPPA monolayer (data not shown).

5.5.2 EVIDENCE FOR A HETERODIMER

We found an unexpected high amount of A-, retained at the lipid-water interfaces, resulting in

a HA/A- ratio between 0.8 and 1.7 for monolayers and strongly bound adsorbents on bilayers.

The absorption coefficients used to quantify the two membrane bound species, were

determined with errors of about 10% (see Table 3). For the determination of TCP phenoxide,

TCP in diluted KOH was measured as TR spectra. These spectra were used for calibration

after a curve fitting. It should be noted that systematic deviations may occur upon adsorption

to the membrane. For A-, there are distinct shifts and broadenings of peaks indicating matrix

effects and changes of hydrogen bonding. The surface concentrations of A- are thus

approximations. Nevertheless, the HA/A- ratios of about 1 support heterodimers as strongly

bound adsorbents.

Only for the bilayer in contact with TCP solution the HA/A- ratios are significantly higher

than 1, ranging from 2.1 to 12.4 (see

Table 4). Thus, we observed a loosely bound population of HA on DPPA/POPC bilayers,

which required the presence of phenolic TCP in the bulk. On monolayers the adsorption of

phenol and phenoxide was always equally fast (kHA,monolayer between 0.04-0.06 min-1, kA-

,monolayer between 0.02-0.03 min-1). Whereas, on bilayers the time constants for the phenol

adsorption and the phenoxide formation and adsorption were quite different. The adsorption

of phenol was 5-9 times faster at the DPPA/POPC bilayer than at the DPPA monolayer. The

adsorption of phenoxide was accelerated only by a factor 2 (see


93

Table 8 and Table 9). Thus compared to the DPPA monolayers, with its hydrocarbon chains

facing the aqueous environment, the bilayer bound the phenol more efficiently than the

phenoxide. Hence, the headgroups of the POPC are involved in the first step of TCP

adsorption. This is confirmed by the reversible decrease observed for n(C=O) ester at

1741 cm-1 and the band of the phosphate group at 1100 cm-1 (see Fig. 19).

5.5.3 EFFECTS EXERTED ON THE MEMBRANES

In the CH2-stretching region between 3000 cm-1 and 2800 cm-1 and for the C=O stretching

band an immediate and partly reversible answer of the bilayer to the adsorption of TCP is

visible in Fig. 18. The formation of TCP phenoxide containing layers on the phospholipid

bilayers induces most probably field effects, which lead to a disturbance of the inner DPPA

layer in respect to its order and absorption characteristic. The sigmoid band shape in parallel

polarized (||) spectra evolves from band shifts and the negative band in perpendicular (⊥)

polarized spectra indicates a loss only of CH2-groups oriented perpendicular to the membrane

surface. As documented for quite a number of lipids [68, 69], blue shifts like the one observed

in parallel polarized (||) spectra indicate a loss of order in the acyl chain. The acyl chains get

tilted and thus CH2-groups move out of the plane parallel to the surface of the membrane (that

is the xy-plane). Hence, the absorption of E⊥⊥⊥⊥ is decreased, which causes an increase of the

dichroic ratio R and a negative signal in the ⊥−⊥−⊥−⊥−spectrum. A similar effect remained for the

DPPA monolayer after washing with buffer (Fig. 17). A negative band at 1741 cm-1 with a

slightly sigmoid shape for both polarizations was observed at the C=O stretching frequency.

This effect is of comparable size but nearly reversible for the bilayer. The ester stretching

vibrations for the undisturbed POPC outer layer and the DPPA layers in buffer was found at

1738 -1737 cm-1 . Thus, a band shift to lower wavenumbers occurred, which again can be


94

attributed to loss of order, as it was e.g. detected for bilayer-nonbilayer transitions [69].

Alternatively, the shift might indicate formation of hydrogen bonding [70]. Comparing Fig. 16

and Fig. 18, only with the bilayer negative bands at 1100 cm-1 and 1000 cm-1, where the

phosphate groups absorbs, can be seen. After washing with buffer the disturbance vanishes.

Thus, during the first step of adsorption the phosphatidic part of POPC headgroups interacts

with phenolic TCP. The headgroup can readily form hydrogen bonds, which compete with

water molecules and anions and lead to a red shift of the P-O stretching vibrations of the

phosphatidic group. Washing away the phenolic excess, only heterodimer like structures are

retained and therefore, effects on the phosphate vibrations vanish. The formation of hydrogen

bonds between phenol and phenoxide in the heterodimer may cause the shifts of bands with

high contribution of ν(C-O-) as observed for 1455 cm-1 (to 1446 cm-1) and 1367 cm-1 (to

1352 cm-1). Hydrogen bonding of the phenolic species HA to POPC headgroups might be the

reason, why POPC acts as a barrier against deprotonation. If any, then only a negligible loss of

POPC occurred, as no negative bands for the νas(N-C, choline group) at 970 cm-1 could be

detected. However, because of the looser packing of the lipid with one unsaturated acyl chain

embedding of TCP can occur.

5.5.4 STATEMENTS ABOUT ORIENTATION OF TCP IN/ON THE MODEL MEMBRANES

Orientation analysis needs inputs concerning the knowledge of the ultrastructure and the

dichroic ratio, as well as information of the orientation of the transition dipole moment in the

molecule. In this study we assumed stacks or monolayers of TCP in liquid crystalline

ultrastructure. TCP signals were quite small (about 5 mAU high) and set on a complex

background, representing effects on the lipid signals. Therefore, values of dichroic ratios

differ, if estimated from integrals or peakheights. For the peakheights any variation in peak


95

widths, influenced from neighbor peaks, are excluded. Resembling the more reliable

quantities, peakheights were used for the following analysis. The orientation of the transition

dipole moments were taken from the DFT calculation. As mentioned above and displayed in

Fig. 14, the results for the phenol (HA) are in good agreement with the experimental spectra.

This enables us to take the orientation of the transition moments for a given frequency from

the computation.

To indicate the orientation we estimated the angle between the transition moment and the

O-C1-C4-Cl axes. In Fig. 8 the results for 1492 cm-1(calculated)/1488 -1(experimental) and

1073 cm-1 (calculated)/1080 cm-1(experimental) are pictured. However, assignment for the

phenoxid (A-) is not as straight forward, because DFT is known to poorly describe even

simple phenoxide [71]. That might be reasoning from solvent effects or from interaction with

counterions [72]. Nonella proposed ion and counterion complexes. The frequencies he

calculated were resembling experimental spectra to a high extent. The complexation and

dimerization lead to downshifts of about 90 cm-1 for the frequency ν6 (from 1357 cm-1 to

1270 cm-1) with a high percentage (64%) of ν(CO). For the 2,4,5-trichlorophenoxide

measured in diluted KOH we also found that the calculated frequencies did not match with the

experimental ones. Based on the work of Nonella, we decided to shift down the frequencies

1605 cm-1 to 1455 cm-1 and 1549 cm-1 to 1352 cm-1, both with a high percentage ν(CO). The

result is depicted in Fig. 28 and shows a good agreement of the intensity pattern. For a first

approximation, we then assigned the experimental frequencies to normal vibrations and used

the information about the orientation of their transition dipole moments in the molecule for a

spatial analysis.


96

Fig. 28. Experimental and calculated IR spectra of 2,4,5-trichlorophenoxide. (A) absorbance TR IR spectra of 3 mmol/L TCP in 2 mmol/L KOH (measurement conditions: CaF2 cuvette, pathlength 50 µm, ambient temperature); (B) IR band intensities ( ε νν~

~d∫ ) as calculated with BECKE3LYP/6-311++ G(d, p). Frequencies marked with an asterix (*) were shifted down (1605 cm-1/-150 cm-1, 1549 cm-1/180 cm-1) to fit the experimental data, taking a possible dimerization of K+-phenoxide complexes into account [72].

wavenumber/cm-1

100012001400160018000.002

0.004

0.006

0.008

abso

rban

ce/A

U A

wavenumber/cm-1

10001200140016001800

calc

.ban

d in

tens

ity/1

03 ⋅ m⋅ m

ol-1

0100200300400500600

B *

*


97

On DPPA monolayers TCP adsorbs and deprotonates rapidly and leads to isotropic 1/1

phenol/phenoxide layers. For the orientation analysis of the phenol (HA) the peak at 1080 cm-

1 and at 1488 cm-1, and of the phenoxide (A-) the peak at 1352 cm-1 and 1045 cm-1 were used.

Dichroic ratios for 1080 cm-1 started by 1.4 (isotrop) and went up to 1.6-1.7 within an

exposure time of about 1 h and concentrations of 2-3mmol/L. For 1352 cm-1 dichroic ratios

leveled off at 1.7-1.8 (Riso 1.5). Based on dichroic ratios for 1080 cm-1 of 1.7 and of 1.8 for

1488 cm-1, we determined a mean angle (α) between the z-axes and the transition dipole

moments of 50°. Taking the angle between the O-C1-C4-Cl axes of the HA molecules and the

transition dipole moments of 81° (1080 cm-1) into account, the O-C1-C4-Cl axes encloses an

angle of about 30° with the z-axes. Whereas, this angle was found to be about 45° for A-,

using an angle of 1° for the transition dipole moment of the 1352 cm-1-band. There was a

slight decrease of the angle after the washing procedure although surface concentrations

decreased, too. During the adsorption process the TCP molecules slightly raise up to allow a

tighter packing.

The slow deprotonation process on the DPPA/POPC bilayer succeeds an adsorption of

ordered phenolic TCP molecules. For 1080 cm-1 instantaneously dichroic ratios between 1.8-

2.0 were found. The R-values for 1352 cm-1 of 1.8-1.9 show a slight loss of order due to

phenoxide formation. Calculation of mean angles between z-axes and the O-C1-C4-Cl axes of

TCP entities lead to about 35° for HA and 48° for A-, which is a quite similar results for the

orientation on bilayers than found for DPPA monolayers. However, it should be mentioned

that the initial values of 46° to 50° for the mean angles of the transition dipole moments for

equilibrated TCP layers are quite close to the value of 55° found for isotropic behavior.


98

5.5.5 THERMODYNAMIC CONSIDERATIONS

The high adsorption of the phenoxide is only surprising at first sight. The Gibbs free energy

for the transfer energy of the phenoxide (and the phenol) into the membrane calculated from

the following four contributions reveals a partition highly favoring the membrane.

For the Born energy as given in section 3.2 we obtained 69 kJ/mol with a bare radius of

0.49 nm and a hydrated radius of 0.6 nm. The image energy is a function of the distance x (see

sec. 3.2 equation 33). Thus the electrostatic potential for an ion varies with the distance x as

given in Fig. 29. For phenol (HA) the electrostatic contribution of the transfer energy is about

-1.6 kJ/mol.

x (in 0.1 nm)

0 10 20 30 40 50

∆GBo

rn+∆

Gim

age (

x) in

kJ/

mol

0

10

20

30

40

50

60

70

80

εm=2εw=78.5d=5 nmr=0.49 nm

Born energy (for an monovalent ion with r = 0.049 nm)

Fig. 29. Potential energy ∆∆∆∆G of the phenoxide as a function of its position x in a membrane of the thickness d (= 5 nm). The plot shows the result of equations (32) and (33) (section 3.2). The external phase is assumed to be an aqueous solution with the relative permittivity εw of 78.5. For the membrane a relative permittivity εm of 2 was used. The radius for the phenoxide was taken from DFT calculations for the TCP (half of the longest distance in the molecule).Center of the membrane at 2.5 nm;

The solvophobic contribution estimated with Uhlig´s equation is -95 kJ/mol and equal for HA


99

and A-. Schellenberg et al. determined an octanol-water partition coefficient of 15 500 for

TCP [as cited in 44]. From this value a transfer energy of -23.9 kJ/mol (for 298 K) can be

calculated. Thus Uhlig´s equation overestimates the solvophobic effect.

Hydrogen bonds between HA or A- and water molecules are broken but in return hydrogen

bonds between HA and A- in the heterodimer are formed. For the first one we used a value of

23 kJ/mol calculated by Schütz [73] for phenol-water dimer. For the latter we can take into

account the energy of intermolecular hydrogen bonds of 17 kJ/mol [70] and the proton

transfer energy of 19 kJ/mol [74].

Furthermore, the dipolar potential of a phospholipid membrane of about +0.2 V caused by its

ester groups [50], decreases the transfer energy for the phenoxide or any other anion (see

section 3.2). For the DPPA monolayer a higher dipole potential can be expected, because it is

not compensated and only shielded by the buffer. Together with hydrophobic interactions

between TCP and DPPA this might explain the stability of adsorbents on the monolayers. For

the first step the electrostatic field seems to be less relevant than the effect of the POPC

headgroups, which accelerate TCP (HA) binding.

From this considerations heterodimer formation is straight forward. On the one hand, because

the larger the ion the smaller is the Born energy. On the other hand, because the

intermolecular hydrogen bonds compensate for the water-TCP hydrogen bonds.

5.5.6 CONCLUSION

Using FTIR ATR spectroscopy we investigated supported lipid layers exposed to the

uncoupler TCP. The results obtained enable us to give a detailed picture of the adsorption to

the membrane, the first step of uncoupling. The adsorption starts with a hydrogen bond

formation from phenolic TCP to the phosphatidic group of the phospholipids of the bilayer.


100

This forces the TCP phenols to adsorb in a ordered manner. In the TCP phenol layers they

partly deprotonate to or enrich of phenoxide. Heterodimers are formed, which can be deduced

from the following facts. It is not possible to adsorb an infinite number of phenoxide to the

bilayer surface and it is not possible to retain an infinite number of phenol at the bilayer

surface. Strong bound TCP species have a 1/1 molar ratio. Hydrogen bonds from HA to

phosphatidic groups are transferred to A-. The heterodimers (or heterodimer equivalent

structures) were found to be less ordered than the initial adsorbed HA.

The TCP disturbs the bilayer mainly because the adsorbed phenoxide leads to a change of the

membrane potential and partly because it is inserted in the bilayer. There is only a small loss

of lipids and changes observed in the spectra are partly reversible for the bilayer.

Comparing the sample and the reference compartment, no additional loss of lipids caused by

TCP was found (see Table 8, Appendix) on DPPA monolayers. Furthermore, phenol and

phenoxide adsorb together. Thus heterodimers are the species formed in a hydrophobic

environment or a hydrophobic/water interface. The heterodimers disturb the DPPA-monolayer

predominantly because the adsorbed phenoxide leads to a change of the potentials exerted on

the lipid molecules of the monolayer. This alters their molar absorbances and was concluded

from the dichroic ratios determined by polarization measurements (see Fig. 17 and Fig. 20).

6 Investigation of Immobilized Mitochondrial Creatine Kinase

101


6.1 INTRODUCTION

Creatine kinases (CK) are a group of the isoenzymes that transphosphorylate ATP and

creatine to ADP and N-phosphoryl creatine (phosphocreatine; PCr) and vice versa. In vivo

the octameric Mib-CK is located in the intermembrane space of the mitochondrion, bound to

the inner membrane and participating in the phosphocreatine (PCr)/ATP circuit or shuttle

[75, 76, 4 and section 3.3 Energy Management: Phosphocreatine-Circuit]. It converts excess

ATP in the intermembrane space to PCr. The latter is an energy rich compound that is pooled

to create a spatial and temporal energy buffer (independent from ATP concentrations). PCr is

distributed to sites of high energy demand, where the dimeric cytosolic CK catalyses the

transphosphorylation leading to ATP. Thus CKs are found in cells and tissues with a high

demand of energy, like skeletal or cardiac muscle, brain and photoreceptor cells, as well as

spermatozoa and electrocytes.

In 1996, Fritz-Wolf et al. [5] resolved the structure of ch Mib-CK and ch Mib-CK-ATP-

complex by x-ray diffraction and calculated an electron density map of a resolution of 0.5

nm. Their results were the first gained for a phosphokinase and elucidated the structure at an

atomic level. They confirmed what was already known from electron microscopy (EM)

measurements [8]. The protein has a cubic like shape (point group 422) with a side length of

9 nm and a channel in the center with a diameter of 2 nm. A detailed overview of structure-

function relationships is given in [77]. Bottom and top faces are enriched of positive charged

amino acid side chains. These are the putative binding sites for the Mib-CK to mitochondrial

membranes, which have a high content of the negatively charged phospholipid CL [e.g. 43].

It was known from binding studies of Mib-CKs and model membranes like vesicles [6, 78,


102

79] and monolayers [7] containing negatively charged phospholipids, that the protein adsorbs

to such lipid layers that mimic mitochondrial membranes. Furthermore, Mib-CKs are found

at places where inner and outer mitochondrial membranes are in close proximity [4]. It even

bridges vesicles to monolayers [80] if they contain negatively charged phospholipids like CL.

Additionally, it was observed that Mib-CK forms 2D crystals on CL layers which can be

determined by EM using negative staining with uranyl acetate [81, 82].

Thus we have chosen negatively charged planar bilayers immobilized on an ATR MIRE

which provide either CL or DPPA as an outer leaflet for the adsorption of Mib-CK.

Adsorption of ch Mib-CK and hu Mib-CK to bilayers of different rigidity were realized with

tightly packed DPPA-bilayers (at a surface pressure of 30 mN/m), on the one hand, and

asymmetric bilayers with two CLs of different content of unsaturated acyl chains as outer

leaflets, on the other hand. A high degree of unsaturated acyl chains leads to less ordered and

loosely packed layers. We used CL isolated from E. Coli membranes which has only 20%

unsaturated acyl chains. Whereas, the third phospholipid chosen was CL from beef heart

which has only unsaturated acyl chains. It was inferred from DSC measurements [79] and

monolayer studies [7] that hydrophobic interactions contribute to the binding of Mib-CK to

lipid layers. Electron spin resonance (ESR) measurements [83], on the other hand, pointed to

a pure electrostatic interaction between lipid molecules and protein.

FTIR ATR measurements in this study were carried out with a special attachment that allows

quasi simultaneous measurements of a sample and a reference channel (single-beam-sample-

reference (SBSR)-technique) [11, 84]. This technique enabled not only in situ measurements

of the activity of immobilized enzyme. But it was also utilized to record difference FTIR

spectra for the investigation of the Mib-CK*MgADP-complex. It was shown before by small

angle x-ray scattering [85], by ESR [86] and FTIR [87] that MgADP (and MgATP) induces


103

greater effects upon binding than creatine. For arginine kinase (AK), a protein which has a

high homology in sequence with Mib-CK but consists of monomers, the x-ray structure of its

transition state analog complex with MgADP and nitrate (TSA-complex) was determined

[88]. Comparing the results from the Mib-CK*ATP-complex [5] with the AK*TSA-

complex, the most pronounced difference was found for the loop 315-325 (Mib-CK) and

309-319 (AK), respectively. The loop connects the two domains of the monomer. Thus the

closure of a cleft is presumed [89] during the formation of the active state and confirmed by

the smaller radius of gyration for the Mib-CK*MgADP-complex than for the Mib-CK itself

or the creatine complex [85]. Investigations of the reaction mechanism and the active site of

CKs using variety of methods have been documented in literature [e.g. 53 (a review of the

basics), 90, 91, 92, 93, 94, 95]. For dimeric CK, FTIR measurements with caged compounds

(ADP, ATP, phosphate and nitrate) have been reported [87, 96]. Recently, this method was

also applied to mitochondrial creatine kinase of rabbit hearts [97]. Resulting from

deactivation due to mutation Trp223 [94], Arg91 [93], and the acidic cluster Glu226,Glu227,

Asp228 [92] were found to participate in the transphosphorylation (numbers refer to the

sequence of ch MibCK11). Whereas, Cys 278 is steering the substrate to the active center

[91], but does not play an active role by the catalysis. 1 of 10 Tyr was found by UV-Vis

measurements to be involved in the reaction [98].

As we investigated the functional and mechanical stability of the immobilized Mib-CK, the

data collected shed some light on the limits for the design of a biosensor using a comparable

immobilization technique for Mib-CK.

11) equivalent to numbers of dimeric CK - 4


104

6.2 EXPERIMENTAL SECTION

6.2.1 MATERIALS

Water was ultrapure (Elga) with a specific resistance of 18.2 MΩ cm-1. DPPA, CL from

Escherichia Coli membranes (CL(E.Coli); Lot:328381/1 294, ShpB713, 525 97: 970716 and

970812), PCr, ADP and NADP were purchased from Fluka AG, CL from beef heart was

obtained from Sigma (CL(bh); Lot: 17H8392). Whereas, glucose-6-phosphate-

dehydrogenase and hexokinase were products from Roche-Boehringer Mannheim.

Chemicals to prepare buffers as NaH2PO4, Na2HPO4, NaCl, CaCl2, MgCl2 (all from Merck),

tris-(hydroxymethyl)-aminomethan (Merck), triethanolamin (Riedel de Haen) and 2-

mercaptoethanol (Fluka) were of p.a. grade. 2H2O was purchased from Merck and Aldrich.

The ch Mib-CK, as well as the hu Mib-CK were isolated from Escherichia Coli strain

BL21(DE3)pLysS transformed with the expression vector pRF23 (for ch Mib-CK) and

pUS01 (for hu Mib-CK) as described before [82, 99]. Solutions were produced with protein

concentrations of about 5 mg/ml. They were pressed through a filter (Millipore, Durapore

2µm: SLGV 004NL), divided into aliquots and frozen with liquid nitrogen. These aliquots

were kept at -18°C. Gel permeation chromatography revealed a stable octamer content. Prior

to immobilization experiments, the stocksolutions were thawed and then centrifuged at 15°C

for 15 min with 6000 × g to separate higher aggregated protein from the octamer. Solution

for the adsorption were prepared by diluting the supernatant with buffer to give the

appropriate concentrations of about 0.5 mg/ml or 0.05 mg/ml, respectively. As buffers 10

mmol/L phosphate, 50 mmol/L NaCl or 20 mmol/L tris-(hydroxymethyl)-aminomethan

(tris); both with 1 mmol/L 2-mercaptoethanol (2-ME) and pH 7.0 were used. Protein

concentrations were determined by a modified method of Bradford [100]. Enzyme activity

and ATP concentrations were measured with the coupled enzyme assay [101, 102].


105

Chemicals and biochemicals were used without further purification.

6.2.2 AFM AND EM MEASUREMENTS

Additionally to FTIR ATR measurements, atomic force microscopy (AFM) of the DPPA

bilayer and electron microscopy (EM) of the immobilized enzyme were performed.

Therefore, small DPPA bilayers on Si plates (10 mm × 10 mm× 0.5 mm) were immobilized

as described in section 4. To some preparations enzyme was adsorbed from a 0.5 mg/ml to

0.6 mg/ml protein solution. Si plates with immobilized DPPA were mounted in a home built

small flow through cell and the ch Mib-CK solutions were slowly pumped (0.2 ml/min) into

the cell and left there for 45 min. Then the membrane was washed with buffer and afterwards

the buffer was exchanged for a substrate-buffer-solution.

A Nanoscope III (Digital Instruments, Inc., Santa Barbara, CA) AFM in the contact mode

was used for the sample surface imaging in a liquid environment. The scanner head D (12

µm scan range) was applied with a liquid cell. An oxid-sharpened silicon nitride probe with a

nominal spring constant of 0.06 N/m was used to the scanning of samples. To further prevent

the probe from modifying the sample surface, the applied force was first minimized on a

small scanning area. Care was taken to keep the sample under liquid during analysis. In this

way a DPPA bilayer was imaged in a 20 mmol/L phosphate buffer pH 7.0. However, the tip

moved the protein and thus it was not possible to get images of adsorbed ch Mib-CK.

To analyze the ultrastructure of immobilized Mib-CK EM was applied. Immobilization of

DPPA bilayers and protein was achieved as described above for AFM measurements. Before

they were used to prepare replicas for EM measurements, the samples were washed with

distilled water to remove salts. Water was gently removed with a paper towel. Then the

samples were frozen in liquid nitrogen. They were freeze dried for 3.5 h at 3 × 10-7 mbar and

-80°C. Shadowing was performed with Pt (45°, d = 1.5 nm) and C (90°, 15 nm). The replicas


106

were removed from the sample by hydrofluoric acid, put on a copper grid and scanned in the

EM (transmission mode).

6.2.3 FTIR ATR SPECTRA ACQUISITION, SPECTRA MANIPULATION AND SPECIAL

ATTACHMENTS

Infrared spectra were recorded within 4000-800 cm-1, a resolution of 2 or 4 cm-1 with a FTIR

spectrometer (Bruker IFS 25) equipped with a gold grid polarizer on a KRS-5 substrate and a

mercury-cadmium-telluride (MCT) detector. For steady state measurements an integration

time of at least 15 min was chosen, for time resolved measurements of 1 min. Spectra were

recorded by the single-beam-sample-reference (SBSR) method [11, 84]. Therefore, a SBSR-

ATR mirror attachment consisting of a cell with two cuvettes, a chopper and a special mirror

set described in detail in section 5 (see Fig. 9) was used. The mirrors were adjusted to get

maximal energy and to achieve an angle of incidence of 45°. The spectra obtained, i.e. S/R

and -lg(S/R), are called SBSR transmittance and absorbance spectra, respectively. SBSR

absorbance spectra were corrected with SBSR absorbance spectra representing the previous

difference of S- and R-compartment. Two experiments with DPPA/CL(E.Coli)-bilayers were

done with a one compartment Delrin® cell and hence a simple ATR-attachment.

Conventional absorbance spectra were computed from single channel spectra with reference

single channel spectra representing the previous state of the sample. In some cases spectra

were smoothened and compensated for water vapor and CO2. All calculations and

manipulations were done with OPUS (Bruker).

The sample- (S-) and reference- (R-) compartments were each separately connected to a

peristaltic pump (Ismatec). The cuvette was coupled to water-filled Cu-plates connected to a

thermostat (Julabo MH F25).


107

6.2.4 SURVEY OF THE EXPERIMENTS

(a) Adsorption of Mib-CK on Ge (MIRE). A 0.6 mg/ml enzyme solution in 20 mmol/L

phosphate buffer pH 7.0 with 100 mmol/L NaCl and 1 mmol/L 2-mercaptoethanol (2-ME)

was pumped through the S-compartment. Whereas, pure buffer solution was pumped through

the R-compartment. The protein solution was changed to buffer after 1 h and polarized

SBSR absorbance spectra were recorded. The SBSR technique was used for in situ activity

measurements with 20 mmol/L ADP/PCr solution and 10 mmol/L MgCl2. The amid II band

was used to calculate the surface concentration of the protein with equation (29) (section 2).

Details see below.

(b) Building bilayers on Ge (MIRE). The Ge plate was activated by high-voltage glow

discharge for 3 min and equilibrated in H2O for at least 10 min. Then it was used as substrate

for DPPA mono- and bilayers by means of the Langmuir-Blodgett method as described in

detail in section 4.

Asymmetric bilayers were prepared by the Langmuir Blodgett (LB)/Vesicle method [3]. The

DPPA monolayer was transferred to the ATR plate as described above, then the coated dry

plate is mounted in a liquid sample cell for flow-through experiments. Completion of the

bilayer was performed by spontaneous adsorption of phospholipid molecules from a

vesicular solution as described in section 4.

Two CLs from different species and thus with different contents of unsaturated acyl chains

were used to obtain bilayers of different rigidity. CL from E. Coli membranes contains about

20% unsaturated acyl chains (44% 16:0, 19% 18:0, 4% 18:1, 14% 18:2 as specified by

Fluka). Whereas, all acyl chains of CL from beef heart are unsaturated (92% 18:2, 6% 18:1,

0.7% 18:3, 0.6% 16:1, 0.4% 16:0; taken from [103]). It was possible to prepare CL (E.Coli)

and CL(bh) bilayers using 2H2O instead of 1H2O with comparable results.


108

(c) Adsorption of Mib-CK on bilayers. The bilayers were exposed to enzyme solutions and

the adsorption process was monitored with the spectrometer. After 1 h the solution was

replaced by buffer and polarized FTIR ATR spectra were measured. These spectra were used

to calculate surface concentrations by means of equation (29) from the amid II band (in

1H2O) or after 1H-2H-exchange from the amid I’ band (in 2H2O). As integral absorption

coefficients ε νd~∫ of 2.74 × 107 (amid I’; integrated with a straight baseline between: 1698-

1595 cm-1) and ε νd~∫ of 8.25 × 106 (amid II; integrated with a straight baseline between:

1585-1500 cm-1) were taken from [14] and determined with the antibiotic peptide

alamethicin [104].

(d) 1H-2H-exchange of the system. 2H2O buffer solutions (either with phosphate or tris-

(hydroxymethyl)-aminomethan (tris)) were prepared, pumped into the cell, and the 1H-2H-

exchange was observed by FTIR ATR measurements for about 20 hours. The buffer

solutions were adjusted to pH* = 6.6 (pD = 7.0) using the relation given by Glasoe et al.

[105].

(e) FTIR in situ activity control measurements. The activity of the enzyme was checked by

the reverse reaction using ADP and PCr as substrates. Before the substrate solution was

pumped into all compartments of the cell, the tubes were changed to ensure that no protein is

adsorbed there. The concentration decrease of PCr was monitored and the activity was

verified by the decay of bands at 1117 cm-1 and 980 cm-1. These bands vanish if the N-P

bond in PCr is broken. The activity was measured right after enzyme adsorption and up to a

few days later. The SBSR technique compensates the signals of the substrate mixture and the

hydrolysis of ATP with the signal of the reference channel. Thus SBSR absorbance spectra


109

directly display the difference signal stemming from the substrate consumption of the

enzyme. From these signals we estimated specific activities under the following

assumptions:

- The immobilized enzyme on the plate can be treated like solubilized enzyme as a first

approximation. This neglects that there are different conditions relating convection and

diffusion and the limited accessibility of the immobilized enzyme for the substrates.

- With the described method the activity of the whole system is observed, but only the

enzyme on the plate can be quantified by FTIR ATR spectroscopy. The contribution to the

activity of adsorbed enzyme anywhere but on the plate was estimated to be 70-75% to the

whole measured activity. This value is slightly lower than the surface ratio of the whole

system to the plate (lid of Delrin and the holes in the lid for the flow (6 cm2/(6+25) cm2 =

20%). Thus if the enzyme is adsorbed in a comparable density, it has the same specific

activity, too12. The activities found were thus scaled down by fact_plate/act_system of 0.35-0.25.

- The volume of one of the compartments is estimated to be 85 µL (± 20%, depending on the

strength the cell is tightened)

- Quantification is based on an absorption coefficient ε determined from an ATR absorbance

spectra of 20 mmol/L PCr water solution (1 point calibration). ε at 1117 cm-1 is 9.3 ± 0.5 ×

105 cm2 mol-1 13 and ε at 978 cm-1 is 7.6 ± 0.4 × 105 cm2 mol-1 14.

- The protein-lipid-layer is thought as a homogenous layer of the given thickness (sum of the

known thickness for lipid layers and protein; can vary ± 10% with influence on the

12) Note: These parts cannot be changed after the CL adsorption, thus they will have CL adsorbed on the surface. But the DPPA-bilayer preparation does not include a step, where lipids are flushed through the cuvette, thus there the enzyme adsorbs just to the Delrin surface. No differences were found and hence the Mib-CK adsorbs to Delrin in an active form. 13) 9.8 × 105 cm2 mol-1 straight baseline fixed at 1825 cm-1

14 ) 8.0 × 105 cm2 mol-1straight baseline fixed at 1825 cm-1


110

concentration < 0.5%). Then the effective thickness for parallel polarization deff,|| was

calculated starting either at zi = 0 nm (for the calibration; deff,|| at 1117 cm-1 = 6.03 × 10-5 cm

and deff,|| at 978 cm-1 = 6.66 × 10-5 cm) or at zi = 14.3 nm (protein + lipid bilayer: for activity

measurements; deff,|| at 1117 cm-1 = 5.74 × 10-5 cm and deff,|| at 978 cm-1 = 6.37 × 10-5 cm).

For the uncoated MIRE deff,|| is 1.055 times higher at 1117 cm-1 and 1.045 times higher at 978

cm-1. Changes due to the change of refractive indices of the Ge/protein-lipid-layer/bulk-

solution and the Ge/bulk-solution system are neglected.

- The initial velocity of the reaction v0 can be calculated with equations (38) and (39),

respectively from the results of the fitting functions for data sets. Beside |dc(t)/dt| for t = 0,

for v0 |dA(t)/dt| for t = 0 can be used (c is concentration of a product or substrate with a

stochiometric factor of 1 and A is the observed absorbance). Data were either fitted by an

exponential function or by a second order polynomial function 15 .

where b1 is the gradient of the polynom. The initial velocity v0 = dc t

dt( )= 0

or dA t

dt( )= 0

is

given as b1 or s.k. Then the specific activity asp can be calculated from equation (40) with the

help of the following parameters:

15 ) some programs (like Excel) can only fit with polynomials. The methods are equivalent as long as the function is not highly curved.

(38) c t s k tdc t

dts k k t

dc tdt

s k

( ) exp( )( )

exp( )

( )

= ⋅ − ⋅

= − ⋅ ⋅ − ⋅

== − ⋅

0

Here s is the saturation value and k the velocity constant.

(39) c t b b t b tdc t

dtb b t

dc tdt

b

( )( )

( )

= + ⋅ + ⋅

= + ⋅ ⋅

==

0 1 22

1 2

1

2

0


111

Vsystem .... volume of the compartment(s), where the reaction takes place (85µL/compartment)

menzyme .... mass of the enzyme; can be calculated with the surface concentration Γ, the

surface of the spectroscopic determined enzyme

Γ .... surface concentration of the enzyme

Oplate ... enzyme coated surface of the plate 3 cm2/compartment

Mr.... molecular mass of the protein (345 kDa for ch Mib-CK and 433 kDa for hu Mi b-CK)

deff ... effective thickness

ε ... absorption coefficient

N... number of active reflections

ν ... number of functional groups absorbing at the given wavenumber

(f) SBSR-FTIR spectra of the Mib-CK-MgADP complex. A 10 mmol/L Na2ADP/MgCl2

solution in 2H2O buffer (either phosphate or tris) at pH* 6.6 was brought in contact with the

immobilized enzyme. After equilibration for 0.5 h polarized FTIR ATR SBSR spectra were

recorded. Compensation for the ADP signals was achieved using the signal from the R-

compartment. SBSR difference spectra from the SBSR absorbance spectra

(MgADP*enzyme-enzyme) were calculated.

(40) [ ]

[ ]

[ ]

a

dc tdt

V

m

dc tdt

V f

M O

dc tdt

=dA t

dtd N

spµ

ε ν

mol

mg min

mol

cm mincm

mol

cm

g

molcm

mol

cm min

AU

mincm

AU cm

mol

33

22

3 2

⋅

=

=⋅

=

=

=

⋅ ⋅

⋅

⋅

=

=

⋅⋅

⋅

⋅ ⋅

( )

( )

( ) ( )

0

1000

0

0 0 1

system

enzyme

system act_plate/act_system

r plate

eff

with

Γ


112

6.3 RESULTS

6.3.1 BILAYERS ON GE: QUANTIFICATION AND STABILITY

The three kinds of model membranes which were prepared differ in stability and rigidity.

Fig. 9 (section 4, DPPA/air, CL), Fig. 30 (DPPA bilayer) and Fig. 31 (CL bh) show polarized

FTIR ATR spectra for the model membranes obtained. Fig. 30 also displays single channel

spectra of parallel (||) and perpendicular (⊥) polarization to put the impression of water bands

across. Lipid layers of DPPA and CL are characterized by intense bands for the CH3- and

CH2-stretching vibrations between 2900 and 2700 cm-1 with an dichroic ratio of about 1.

Furthermore, the ν(C=O)-band at 1740 cm-1, the δ(CH2) at 1450 cm-1 and between 1300 and

1200 cm-1 and at 1100 cm-1 the broad band for the ν(P-O) vibrations can be seen. DPPA as

bilayer/buffer as well as monolayer/air show additionally the γ(CH2) set on the ν(P-O) band.

For the quantification of DPPA and CL from E. Coli the νs(CH2)-bands at 2850 cm-1

(integrated between 2832-2867 cm-1) were evaluated using the thin film approximation and a

ε νd~∫ of 5.22 ± × 105 cm mol-1. Whereas, CL from beef heart is characterized by rather

small ν(CH2) bands due to the high content of unsaturated acyl chains (see Fig. 31). Thus we

evaluated the surface concentration using the ν(C=O)-band at 1740 cm-1 from the ester

groups. For this case the ε νd~∫ was determined from DPPA monolayers and a value of 1.10

± 0.07 × 107 cm mol-1 (integrated between 1769.5-1678 cm-1) was found.

DPPA mono- and bilayers of surface concentrations of Γ = 4.1 ± 0.4 × 10-10 mol cm-2,

corresponding to a molecular area of 41 ± 4 Å2 were obtained, as determined with the thin

film approximation from the ATR absorbance spectra. These values are in good agreement

with 43.8 Å2 found by x-ray measurements for a DPPA monolayer at the air/water interface

at pH 7.2 and 30 mN/m [66]. We determined a molecular order parameter Smol of 0.93 ±


113

0.06.

||

⊥

||

⊥

AbsorbanceUnits

0.3

0.2

0.1

0.0

-0.1

Fig. 30. Polarized FTIR ATR spectra of a DPPA bilayer. Top: single channel spectra of the DPPA bilayer and clean Ge plate of the S-compartment, respectively, in 20 mmol/L phosphate, 100 mmol/L NaCl, buffer pH 7.0; Bottom: correspondent absorbance spectra; Reference: clean Ge plate/buffer mounted in the SBSR cell; Γ = 4.4 × 10-10 mol/cm2 (Amol = 38 Å2/molecule, R = 0.9, Smol = 0.99; angle of incidence θ€=45°, Nact = 24.5; n1= 4.0, n2= 1.45, n3= 1.41);

All DPPA layers were achieved by means of the Langmuir-Blodgett technique. As a control

measurement a DPPA bilayer transferred to a Si plate was examined by AFM. A smooth

surface at 5 nm height with some tiny holes were found in a patch of 5 × 5 µm (Appendix:

Fig. 61). Taken together, this reveals that planar bilayers of high order were obtained. DPPA-

bilayers are the most stable but also the most rigid model membranes used in this study.

They exert the highest surface charge density (-1qe/41 Å2 at pH 7.0).


114

1

23

4

Fig. 31 Polarized FTIR ATR spectra of CL(bh). 1(||) and 2(⊥): a CHCl3 solution of CL(bh) was spread homogeneously on a Ge crystal slowly. Thereby the CHCl3 evaporated. Reference: clean plate/air; spectra were scaled down (f = 0.5, Nact = 22.7) for comparison with following spectra for the outer layer; 3(||) and 4(⊥): outer layer of CL(bh) constituted on a DPPA monolayer from a 0.7 mg CL(bh) /mL 20 mmol/L tris, 150 mmol/L NaCl (2H2O) vesicles solution pH* 6.6; Γ = 1.12×10-10 mol/cm2 (Amol = 150 Å2/molecule, R = 1.6, Sseg = 0); Reference: DPPA/buffer (angle of incidence θ€=45°, Nact = 32.3; n1= 4.0, n2= 1.45, n3= 1.33);

Bilayers with an outer layer made from CL of E.Coli or CL of beef heart were prepared with

the LB-Vesicle-method. We obtained surface concentrations Γ of 1.6 ± 0.3 × 10-10 mol cm-2,

mean area per molecule 104 ± 15 Å2, a mean dichroic ratio of 1.25 ± 0.09 and a mean

molecular order parameter of 0.45 ± 0.05 for CL (E.Coli), which contains about 20%

unsaturated acyl chains. For the evaluation of the surface concentration Γ the mean number

of CH2 groups was calculated to be 62 from the given composition of acyl chains (including


115

groups from the glycerin backbone). The result depended on the particular lot of CL used for

the preparation and can be influenced by the ionic strength of the buffer. The high surface

charge density (-2qe/100 Å2 at pH 7.0) and the possibility of CL to adopt hexagonal H II

phases [106, 107, 108] clearly leads to instability. Because of their unsaturated acyl chains

CL layers are less stiff and less ordered. This effect is most pronounced for the CL (beef

heart), which is composed of nearly 100% unsaturated acyl chains. Fig. 31 shows FTIR ATR

spectra of CL(bh) multilayers formed from a CHCl3 solution and the CL(bh) outer leaflet of

an asymmetric bilayer. Typically, small CH-stretching bands compared to the ν(C=O)-band

were found. The evaluation of the ν(C=O)-band resulted in a surface concentration Γ of 1.10

± 0.08 × 10-10 mol cm-2, mean area per molecule 150 ± 12 Å2. It exhibits the lowest surface

charge density (-2qe/150 Å2 at pH 7.0). A mean dichroic ratio of 1.53 ± 0.06 and a mean

segmental order parameter of -0.05 ± 0.02 was found. Thus there is either no mean

orientation for the four ester groups or the angle is about 55° with respect to the bilayer

normal (z-axes).

For a CL with four saturated acyl chains a molecular area of 80 Å2 is expected as minimum

value. The CLs used in this experiments are from natural sources and thus are not single

compounds. With the given mean degree of unsaturation of 20% (CL E.Coli) or 100% (CL

bh) the results found are quite reasonable16. Two unsaturated chains out of four will increase

the molecular area from 80 to about 150 Å2 per molecule (based on results for DPPA and

POPC). 20% unsaturation will lead to about 110 Å2 per molecule. Thus the measured values

show that the areas occupied can be deduced to the acyl chains.

Because of the electrostatic repulsion between their head groups, DPPA and CL model

membranes were less stable then the DPPA/POPC-bilayers reported before [3]. However,

16) no data in literature were found


116

DPPA-bilayers were more stable than asymmetric bilayers. A survey of the results of all

preparations used are given in Table 11 (Appendix).

DPPA/CL(E.Coli)-bilayers lose about 15% within 15 h without pumping and up to 30%

when buffer is pumped with 10 mm/min (0.2 ml/min) for about 5 h. Fig. 32 to Fig. 35

display the CH-stretching region for different duration times and compare the signals from

R- and S-compartment. The kinks found in the signals can be attributed to frequencies of

DPPA. This indicates the loss of bilayer fragments containing DPPA and seems typical for

DPPA/CL(E.Coli)- and DPPA/CL(bh)-bilayers. Fig. 32 to Fig. 35 clearly show, that the

asymmetric bilayers of the sample (S)-compartments were stabilized by the protein. Mib-CK

exhibits protection against the electrostatic repulsion.

However, for DPPA bilayers the protection of the protein against loss of lipids is not as

pronounced as for CL. In the course of the experiment the DPPA bilayer of the R-

compartment (without enzyme) lost 40% of its lipids as deduced from the negative signals of

CH stretching region within 5 days. Whereas, the lipid loss in the S-compartment was about

30% as can be seen in Table 13 (Appendix). Table 12 (Appendix) shows, that there is no

significant change of dichroic ratios after adsorption of protein to the DPPA-bilayer. Thus

the protein does not disturb the order of the acyl chains.

If a phospholipid bilayer is not as rigid as a DPPA-bilayer, a protein that is naturally found

embedded in the membrane should be able to sink into the model membrane. This will in

turn disturb the lipid layer and can be recognized by shifts of the CH-stretching bands and

decrease of the Smol (and the dichroic ratio R). Such effects were detected for neither of the

proteins and for none of the asymmetric DPPA/CL bilayers. Thus FTIR ATR spectra show

Mib-CK proteins adsorbed at the surface, without penetration into the hydrophobic region of

the lipid bilayer.


117

Fig. 32. CH-stretching bands of the CL (E.Coli) layer in the reference (R) - compartment during the experiment. Both polarizations are shown (solid line: pp=||, dashed line: vp=⊥). Top: after the preparation (dichroic ratio R = 1.3, surface concentration Γ = 1.53 ± 0.03 × 10-10 mol cm-2, corresponding to a molecular area of 108 ± 3 Å2, calculated with the thin film approximation; θ 45°, Nact= 36.7, n1= 4.0, n2= 1.45, n3= 1.41); Buffer: 20 mmol/L phosphate, 0.1 mol/L NaCl, pH 7.0; Middle: after 15 h (ca. 1 h with pumping at 0.2 ml/min for 1H2H -exchange), right before start of protein adsorption in the sample (S) compartment (loss of lipids: 34%). Bottom: CL-layer of the R-compartment after ch Mib-CK adsorption in the S-compartment (17 h after the bilayer formation); Buffer for and after enzyme adsorption: 10 mmol/L phosphate, 0.05 mol/L NaCl, 1 mmol/L 2-ME, (either in 1H2O (pH 7.0) or 2H2O (pH* 6.6)) Reference: DPPA monolayer/buffer; The ⊥ spectrum shows two kinks which come from the decrease of DPPA-CH-stretching-bands.


118

Fig. 33. CH-stretching bands of the CL (E.Coli) layer in the R- and S-compartment during the experiment. Pairs of conventional absorbance spectra from the S-compartment and R-compartment are shown (solid lines: ||, dashed lines: ⊥). First pair (top): signal from S-compartment after the preparation (dichroic ratio R = 1.3, Γ = 1.53 ± 0.03 × 10-10 mol cm-

2, corresponding to a molecular area of 108 ± 3 Å2, calculated with the thin film approximation; θ 45°, Nact 36.7, n1=4, n2=1.45, n3=1.41); Buffer: 20 mmol/L phosphate, 0.1 mol/L NaCl, pH 7.0; Reference: DPPA monolayer/buffer; Second pair: after 17 h (ca. 2.5 h with pumping at 0.2 ml/min for 1H2H-exchange and enzyme adsorption). The spectra include CH-stretching bands of ch Mib-CK. Reference: DPPA monolayer/buffer; Third pair: CH-stretching bands of ch Mib-CK and decrease and differences of the lipid bilayer of the S-compartment due to enzyme adsorption; Reference: DPPA/CL-bilayer/buffer before enzyme adsorption (S-compartment); Bottom: Decrease and differences of the lipid bilayer of the R-compartment; Buffer for and after enzyme adsorption: 10 mmol/L phosphate, 0.05 mol/L NaCl, 1 mmol/L 2-ME, (either in 1H2O (pH 7.0) or 2H2O (pH* 6.6)) The arrows mark the positions of the kinks mentioned in Fig. 32. Reference: DPPA/CL-bilayer/buffer before enzyme adsorption (R-compartment);


119

Fig. 34. CH-stretching bands of the CL (bh)-layer in the R-compartment during the experiment. Pairs of polarized spectra are shown (solid lines: ||, dashed lines: ⊥). First pair (top): signal from R-compartment after the bilayer preparation (buffer: 20 mmol/L tris, 0.15 mol/L NaCl in 2H2O pH* 6.6, surface concentration Γ = 1.10 ± 0.01 × 10-10 mol cm-2, corresponding to a molecular area of 150 ± 5 Å2, dichroic ratio R = 1.5, calculated with the thin film approximation from the νs(C=O); θ 45°, Nact= 32.3, n1= 4.0, n2= 1.45, n3= 1.33); Second pair: after 10 h in contact with CL(bh) vesicle solution at 10°C (buffer as above); Third pair: CH-stretching signals of the R-compartment 43 h later and after the hu Mib-CK adsorption in the S-compartment, as well as one 2H 1H- and one 1H 2H-exchange. There is a loss of all CL (beef heart) and of about 60% of the DPPA. This is most probably due to a decrease in ionic strength of the buffer (20 mmol/L tris, 1 mmol/L 2-ME in 1H2O pH 7.0) used for adsorption. Reference: DPPA/buffer; Bottom: CH-stretching bands of the DPPA-monolayer against air (to compare the position and size of signals). Reference: Ge/air;


120

Fig. 35. CH-stretching bands of the CL (bh)-layer in the S-compartment during the experiment. Pairs of polarized spectra are shown (solid: || (except bottom), dashed: ⊥). First pair (top): signal from S-compartment after the bilayer preparation (buffer: 20 mmol/L tris, 0.15 mol/L NaCl in D2O pH* 6.6, surface concentration Γ = 1.10 ± 0.01 × 10-10 mol cm-2, corresponding to a molecular area of 150 ± 5 Å2, dichroic ratio R = 1.5, calculated with the thin film approximation from the νs(C=O); θ€=€ 45°, Nact = 32.3, n1= 4.0, n2= 1.45, n3= 1.33). Second pair: after 10 h in contact with CL(bh) vesicle solution at 10°C (buffer as above); Third pair: CH-stretching signals of the S-compartment after hu Mib-CK adsorption in the S-compartment and one 2H1H- and one 1H2H-exchange. The protein and the CL(bh) are contributing to the CH-stretching bands shown. There is a small loss of DPPA (kink in the ⊥ spectra). However, compared to the observed loss displayed in Fig. 34 hu Mi -CK stabilizes the bilayer. Bottom: CH-stretching bands of DPPA against air (to compare the position and size of signals; ⊥-spectra downscaled by f=0.5).


121

6.3.2 ADSORPTION OF MITOCHONDRIAL CREATINE KINASE

It is well known that Mib-CK bind to lipid layers like vesicles [6, 78, 79] or monolayers [7,

80] containing negatively charged phospholipids, e.g. CL. In this study the adsorption to

negatively charged bilayers which provide either CL or DPPA as outer leaflet was observed

by FTIR ATR spectroscopy. As a control, adsorption to Ge was investigated too.

Comparison of Fig. 38 and Fig. 39 shows that FTIR ATR spectra of adsorbed ch Mib-CK and

hu Mib-CK are quite indistinguishable. Thus the features of their spectra can be described

together. At 3300 cm-1 the ν(NH) is found and between 3000 and 2800 cm-1 the various CH

stretching vibrations absorb. In Fig. 38 and Fig. 39 the SBSR absorbance spectra display not

only the protein signal in this region, but also reflect differences of the lipid bilayers of the

R- and S-compartment. The observed increase at 2900 cm-1 and 2850 cm-1 during the course

of 1H-2H-exchange is caused by the loss of lipids in the R-compartment.

The spectra of Mib-CK in 2H2O are clearly dominated by the amide I/I’ band (1700-1600 cm-

1 Amax: 1647 cm-1) and amide II (1600-1500 cm-1; Amax: 1549 cm-1) band, as can be seen in

Fig. 38 and Fig. 39. In the course of 1H-2H-exchange the νas(COO-) (Asp and Glu) at

1586 cm-1 and νring(C-C) (Tyr) at 1515 cm-1 are emerging and the amid II band shifts to 1450

cm-1. This new band is called amide II’ and it overlaps now with the δ(CH2) and partly with

the νs(COO-). 1H2HO will also absorb in this region. The amide I band is only changing

slightly by 1H-2H-exchange.

Fig. 36 depicts the increase of the amide bands in the course of adsorption to various surfaces

as measured with Mib-CK. The signals are related to the increase of the phosphate band

measured during the exchange of buffers with different phosphate concentrations.


122

t / min

0 20 40 60 80

Γ ×

10-1

3 / m

ol·c

m-2

0

2

4

6

8

10

Abso

rban

ce a

t 107

7 cm

-1/ m

AU

0

2

4

6

8

10

12

14

16

18

hu/DPPA-bilayerphosphate (buffer exchange)ch/Ge

ch/DPPA-bilayerch/DPPA-CL(E.Choli)

Fig. 36. Time course of adsorption of Mib-CK. Solutions of 0.5 mg Mib-CK /ml 20 mmol/L phosphate, 100 mmol/L NaCl pH 7.0 were pumped into the cuvette. Solid lines resulted from curve fit with the function y(t) = y(t=∞)·(1-exp(-k·t). Either chicken (ch) or human (hu) Mib-CK was used and exposed to the following surfaces: ch to clean Ge ( , k = 0.8 min-1), ch to a DPPA/CL(E. Coli) bilayer ( , k = 0.2 min-1), ch to a DPPA bilayer ( , k = 0.6 min-1) and hu to DPPA-bilayer ( , k = 0.6 min-1). The amide I or amide II signals were scaled to the known surface concentrations at the end of the adsorption. To characterize the time constant of the system, the absorbance values for the phosphate band ( , k ca. 0.3 min-

1) are shown.

Within the limits of the flow rate there is no difference in the adsorption behavior for hu and

ch Mib-CK and the ch Mib-CK differentiates only weakly between Ge, DPPA and

CL(E.Coli). Fig. 36 reveals that the adsorption kinetic is of the same magnitude as the flow

rate chosen (100 mm/min, k ca. 0.3 min-1). This is in the range given by Stachowiak et al

[78] who determined k2 = 7.6 × 10-3 s-1 for the slow association to biotinylated CL vesicles


123

immobilized on avidin. Enzyme solutions were in contact with the lipid layers for 0.75 - 1.0

h. After that period the protein solution was exchanged for buffer and polarized FTIR ATR

spectra were measured. The amount of adsorbed protein was independent of the protein

concentration in solution. Equivalent results were obtained for 55 µg protein/ml buffer

solution and 0.6 mg protein/ml buffer solution as can be seen from Fig. 37, Fig. 38 and Table

13 (Appendix). Surface concentrations were calculated either from the amide I’ (for protein

after 1H-2H-exchange) or from amide II band (in water). The amid II band was corrected for

the contribution of νasCOO- using the νsCOO- for the determination of COO- groups present

in the protein. The results of 12 immobilization experiments for the initial surface

concentration were found between 7.9 and 11.5 × 10-13 mol cm-2. From these values surface

coverage densities of 40-60% for the protein were computed, using a area/molecule of 932 =

8 649 Å2 (for ch Mib-CK, determined by x-ray diffraction [5]). However, calculating surface

coverage densities for hu Mib-CK with 1452 = 21 025 Å2 as determined by EM [82], would

lead to values >100%. Two interpretations for this result are possible: either partially

multilayers were formed or the protein is packed more tightly in our preparations than in

[82]. Details are listed in Table 14 (Appendix). From this table one can see that the protein

surface concentration decreases 10-20% in 2-3 days. For 1 day protein loss is neglectible.

6.3.3 1H-2H-EXCHANGE OF IMMOBILIZED MITOCHONDRIAL CREATINE KINASE

Exchange of 1H to 2H shifts the N1H-stretching band and the amide bands to lower

wavenumbers. N2H-stretching will be found at ca. 2500 cm-1, where it usually cannot be

detected due to the solvents absorption. Whereas, the shifts for the amide I band is only in

the range of 10 cm-1, the amide II band is moved to 1450 cm-1. It is indicated as amide II’.

The evaluation of the rate of the 1H-2H exchange reveals information about the flexibility


124

and solvent accessibility of the protein. In most cases the adsorption of Mib-CK to the

phospholipid bilayers was followed by the 1H-2H exchange. In some cases the enzyme was

adsorbed from a 2H2O buffer solution, e.g. Fig. 37 shows polarized FTIR ATR spectra of ch

Mib-CK 20 h after it had been solved in 2H2O buffer solution and adsorbed to a

DPPA/CL(bh) bilayer. In spite of this long period, there is a prominent N1H-stretching band

at 3300 cm-1, stemming from N1H-groups that were not exchanged to N2H-groups.

AbsorbanceUnits

0.2

0.0

0.1

-0.1

||

⊥

||

⊥

Fig. 37. Polarized FTIR ATR spectra of ch Mib-CK on a DPPA/CL(bh) bilayer. Top: single channel spectra of ch Mib-CK immobilized on the bilayer; the stocksolution was centrifuged and an aliquot was diluted in 2H2O buffer to give 55 µg ch Mib-CK/ml 19 mmol tris, 144 mmol/L NaCl pH*=6.6; Bottom: (conventional) absorbance spectra of the immobilized protein; Γ = 1.0 × 10-12 mol/cm2 (Amol = 130 Å2/molecule, coverage 52%, R=1.7, Sseg = 0); Note the high amount of unexchanged N1H-groups though protein was solved in 2H2O for 20 h. A surface concentration of 1.2 × 10-10 mol/cm2 (R = 1.62) was calculated. This corresponds to 37% related to the amide groups. Reference: DPPA/CL(bh) against buffer; angle of incidence θ€= 45°, Nact = 32.31; n1= 4.0, n2= 1.4, n3= 1.32 (at 1645 cm-1) and n3= 1.27 (at 3300 cm-1);


125

1

2,||

3

||

⊥||

⊥

Fig. 38. Polarized FTIR ATR SBSR absorbance spectra of ch Mib-CK on a DPPA bilayer. Top: ch Mib-CK adsorbed to a DPPA-bilayer from a 0.6 mg protein/ml. Stocksolution was centrifuged and an aliquot was diluted in buffer (20 mmol/L phosphate, 100 mmol/L NaCl, 1 mmol/L 2-ME, pH 7.0); contact time 45 min at T 25°C; Reference: DPPA bilayer/buffer; Middle: after 2 h contact with 2H2O buffer solution. Bottom: as above, but with 17 h contact with 2H2O buffer solution; for the enzyme surface concentration Γ = 9 × 10-13 mol/cm2 (Amol = 136 Å2 /molecule, coverage 47%, R=1.6, Sseg = 0) was found; the amount of unexchanged N1H-groups is equivalent to Γ = 1.09 × 10-9 mol/cm2 (R=1.45), i.e. 39% of the amide groups; Reference: DPPA-bilayer against buffer (angle of incidence θ€= 45°, Nact = 24.47; n1= 4.0, n2= 1.4, n3= 1.32 (1645 cm-1/2H2O) n3= 1.33 (1550 cm-

1/1H2O), n3 = 1.27 (3300 cm-1/2H2O);

Comparison of Fig. 38 and Fig. 39 (bottom) reveals that both, ch Mib-CK as well as hu Mib-

CK, contain N1H-groups although the enzyme has been in contact with 2H2O buffer solution

for 17 h (ch Mib-CK) and 18.5 h (hu Mib-CK), respectively. Fig. 38 and Fig. 39 (top) display

initial polarized FTIR ATR spectra in 1H2O and both figures show spectra of the enzyme

after a short period of 1H-2H exchange (middle).


126

||

⊥

||

||

⊥

⊥

Fig. 39. Polarized FTIR ATR SBSR absorbance spectra of hu Mib-CK on a DPPA-bilayer. Top: hu Mib-CK adsorbed to a DPPA-bilayer from a 0.3 mg protein/ml; the stocksolution was centrifuged and an aliquot was diluted in buffer (20 mmol/L phosphate, 100 mmol/L NaCl, 1mmol/L 2-ME, pH 7.0); contact time: 45 min (necessary ca. 7-10 min); Middle: after 4 h contact with 2H2O buffer solution (20 mmol/L phosphate, 100 mmol/L NaCl, 1mmol/L 2-ME, pH* 6.6); Bottom: as above, but with 18.5 h contact with 2H2O buffer solution; the surface concentration of the protein was determined to Γ = 9 × 10-13 mol/cm2 (Amol = 137 Å2 /molecule, coverage 46%, R=1.6, Sseg = 0). For the unexchanged N1H groups a value of Γ = 7.2 × 10-10 mol/cm2, corresponding to 33% of all amide groups was determined. Reference: DPPA-bilayer against buffer (angle of incidence θ=45°, Nact = 19.23; n1= 4.0, n2=1.4, n3=1.32 (Amid I/2H2O) n3=1.33 (Amid II/1H2O);

The time course of the 1H-2H exchange for one preparation of ch Mib-CK is depicted in Fig.

40. It displays peak areas of the amide II-, N1H-stretching and amide II’-band of immobilized

ch Mib-CK in contact with a 2H2O buffer solution. As expected, for the amide II-, N1H-

stretching-band curves of similar shape and for the amide II’-band of inverted shape, can be

seen in Fig. 40. For further analyses, the logarithm of the peak areas was plotted vs. time


127

(Fig. 41). The logarithmic function does not generate a straight line. Thus the kinetic

mechanism cannot be described with one exponential function but relies on different

processes, each with its own time dependent function. This results reflects, that N1H-groups

of a complex molecule as the Mib-CK will have different accessibility and exchange rates for

2H2O dependent on local flexibility. Small exchange rates are pointing to structures that are

not flexible, like β-sheets and α-helices. The amount and time constant of the slowest

exchanging N1H-groups were determined from the linear regression of the logarithm of peak

areas at the end of the exchange between 318 to 552 min (see Fig. 41 B). The time constant

was calculated from the slope of the linear regression and is about one week (5 ± 2 d). The

amount of „resistant“ N1H-groups at t = 0 min was determined from the value (lnA(t = 0))

for the intercept of the linear regression.

Fig. 40. 1H2H exchange of the ch Mib-CK adsorbed to a DPPA-bilayer. Time course of the peak area of the NH-stretching -, the amide II and amide II’ -band.

Integration limits: 3387-3217 cm-1 (NH), 1595-1525.5 cm-1 (amide II) and 1491-1414 cm-1 (amide II’).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 200 400 600t/min

int.

Abs

orpt

ion/

cm-1

Amid 2NH stretchAmid 2'


128

Fig. 41. Logarithm of the peak areas of ννννsN1H ( ), amide II ( ) and amide II’( ) taken from Fig. 40. (A) time course between 38 and 600 min of a 1H2H exchange. (B) linear regression for 318 to 552 min for the estimation of the time constant and amount of slow exchanging (resistant) N1H-groups; νsN1H: y(t) = -1.24 × 10-4 x - 0.530 (Rsqr =0.78); amide II: y(t) = -1.06× 10-4 x - 1.110 (Rsqr =0.90); amide 2’: y(t) = -1.54 × 10-4 x - 0.605 (Rsqr = 0.93);

In the equilibrated state polarized FTIR ATR spectra were measured and an isotropic

behavior for N1H-groups was determined. Thus we used Riso = 1.64 (n1= 4.0, n2= 1.4,

n3= 1.27; 3300 cm-1) to calculate the surface concentration of the N1H-groups, applying the

thin film approximation. The result was related to the whole amount of amide groups as

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0 200 400 600

t/min

ln (i

nt. A

bsor

banc

e)

ln (A m id 2'-a-b)

ln (NH)

ln A m id 2

(A)

(B)

-1.4-1.2

-1-0.8-0.6-0.4

200 300 400 500 600t/min

ln (i

nt. A

bsor

banc

e)


129

determined from the amide I/I’-band and values of 27 ± 6 % were found. That corresponds

well to the amount of 33% for α-helices as determined by the x-ray analysis and indicates

that N1H-groups of the α-helices are responsible for the slow 1H-2H exchange.

6.3.4 ACTIVITY CONTROL MEASUREMENTS

Beside the mechanical stability, we also determined the activity by means of in situ FTIR

ATR measurements. Fig. 42 shows clearly that the production or consumption of PCr will

lead to easy detectable positive (production) or negative (consumption) peaks at 1117 cm-1

and 980 cm-1.

1

2

1117cm-1

980cm-1

1611cm-1

1395cm-1

Fig. 42. Parallel polarized ATR absorbance spectra of 20 mmol/L phosphocreatine (PCr; 1) and 20 mmol/L creatine (Cr; 2) in H2O. The solution was pumped into one compartment of a SBSR cuvette. Reference: water in the same compartment; angle of incidence θ = 45° and Nact = 18.36; the 2 peaks of PCr at 1117 cm-1 and 980 cm-1 are isolated and were used for in situ activity measurements. For the prominent peaks of the PCr spectrum the position of the maximum is given.


130

Whereas, changes between ADP and ATP will give differences of minor amplitude, because

their IR spectra are quite similar (see Fig. 43). Thus, the negative bands at 1117 cm-1 and 980

cm-1 were used for evaluation for the in situ activity control measurements.

We have chosen the reverse reaction at pH 7.0 and PCr and ADP as substrates. This direction

avoids major problems due to ATP hydrolysis. However, the signals of the substrate-mixture

as well as changes due to hydrolysis were compensated with the signals from the R

compartment using SBSR technique. Fig. 44 displays the resulting SBSR absorbance spectra

of one activity control measurement. From such SBSR absorbance spectra the absorptions at

1117 cm-1 or 978 cm-1 vs. time was determined.

1

2

1241cm-1

1229cm-1

1070cm-1

Fig. 43. Parallel polarized ATR absorbance spectra of 4 mmol/L ATP (1) and 4 mmol/L ADP (2) in H2O. The solution was pumped into one compartment of a SBSR cuvette. Reference: water in the same compartment. Angle of incidence θ = 45° and Nact= 18.36. Bands of ATP and ADP overlap to a great extend. Thus bands of PCr were used for quantification.

Initial reaction velocities were gained from curve fits of the decrease of the Amax at 1117 cm-1


131

or 978 cm-1, like the one depicted inFig. 45. Specific activities were calculated by means of

Equ. (40) and listed in Table 5. For the reasons described before the specific activities are

rough approximations and shall only be considered within these type of experiments.

At the beginning of an experiment specific activities of about 20 µmol/mg·min at pH 7.0 and

40 µmol/mg·min at pH 5.0 were found. For two experiments the activity control was

performed every day and a decrease of 50% for the first 24 h due to aging was found even in

the presence of 2-ME. Datas are listed in Table 5. Furthermore, the influence of 2H2O is

documented. 2H2O exhibits a isotope effect and slows down the reaction to about 1 µmol/mg

min. In 1H2O the enzyme recovers its activity again.

The specific activity of the enzyme in solution was determined with the reverse reaction at

pH 7.0 and ambient temperature (22-26°C) using the coupled enzyme assay. For the

ch Mib-CK initial activities of 150 µmol/mg·min and 112 µmol/mg·min of the high

concentrated stocksolutions (2-5 mg/ml) were found. These values are comparable to the

Vmax values of 144 µmol/mg·min [92] and 95 µmol/mg·min [94] for ch Mib-CK and 130

µmol/mg·min [109] for hu Mib-CK measured before. An aliquot of the diluted enzyme

solution used for adsorption was analyzed at the end of every experiment series. Aging was

observed and after the 2-5 days of experimental time only 40-60 µmol/mg·min were

measured, though 1 mmol/L 2-ME was added to all buffers. Thus both, the immobilized

enzyme and the solubilized enzyme show decrease in specific activity.

6 In

vest

igat

ion

of Im

mob

ilized

Mito

chon

dria

l Cre

atin

e Ki

nase

13

2

Tab

le 5

. Spe

cific

act

iviti

es o

f im

mob

ilize

d M

i b-C

K. I

nitia

l vel

ociti

es d

c/dt

(t=0

) and

spec

ific

activ

ity w

ere

calc

ulat

ed w

ith

equa

tion

(38)

and

(40)

from

Am

ax at

111

7 cm

-1 o

r 978

cm

-1 a

nd li

sted

toge

ther

with

the

surf

ace

conc

entra

tion

Γ an

d th

e m

easu

rem

ent c

ondi

tions

expe

rimen

t dc

/dt (

t=0)

in

10-6

m

ol/c

m³m

in

spec

ific

activ

ity17

in

µm

ol/m

g m

in re

mar

ks

PCr/A

DP/

Mg

in m

mol

/L

Γ in

10-1

3 m

ol/c

m²

f ac

t_pl

ate

/act

_sys

ch

Mi b-

CK

/DPP

A-

CL(

bh)

0.8

161st

day

, 1 H2O

;100

mm

ol/L

tris

pH

7.0

20

/4/4

10

.0

0.25

hu M

i b-C

K /D

PPA

-C

L(bh

) 0.

044

1.28

1st d

ay,

2 H2O

,20

mm

ol/L

tris

0.

15 m

ol/L

NaC

l pH

* 6.

6

8.9

0.37

0.

17

62nd

day

, 1 H2O

10

0 m

mol

/L tr

is

10/1

0/10

8.

1

hu/D

PPA

-bila

yer

0.57

18

1st d

ay, 1 H

2O, 1

00 m

mol

/L

phos

phat

e

pH 7

.0

20/4

/4

0.9

0.35

ch M

i b-C

K /D

PPA

-bi

laye

r al

l mea

sure

men

ts in

20

mm

ol/L

pho

spha

te

0.1

mol

/L N

aCl a

t pH

5.3

1.

59

431s

t day

;1 H2O

20

/20/

20

9.0

0.3

0.

65

182nd

day

; 1H

2O

0.04

5 1

3rd d

ay ;

2 H2O

0.

265

74th

day

;1 H2O

0.

544

165th

day

; 1H

2O

0.88

4 26

6th d

ay; 1

H2O

17) V

alue

s hav

e un

certa

intie

s of a

bout

20%

due

to si

gnal

inte

nsity

and

fit p

roce

dure

.


133

0 min

8 min

16 min

24 min

31 min

39 min

Fig. 44. SBSR FTIR ATR absorbance spectra of a 10 mmol/L ADP/PCr/Mg2+ solution in contact with immobilized hu Mib-CK. First Line: spectrum of a 10 mmol/L substrate/product-mixture calculated from the individual spectra of the substances in 1H2O. From the bottom: A substrate mixture in 100 mmol/L tris pH 7.0 was pumped into R- and S-compartment. Difference SBSR absorbance spectra at the indicated reaction times are shown and typical difference bands are marked with lines (ATP/ADP: 1250 cm-1; neg. bands from decrease of PCr: 1117 cm-1, 978 cm-1).

0.002.00

4.006.00

8.0010.00

0 10 20 30 40t/min

cPC

r / m

mol

L-1

Fig. 45. Decrease of the concentration of PCr and fit used for the estimation of specific activity. hu Mib-CK immobilized on DPPA/CL(bh)-bilayer was in contact with a 10 mmol/L PCr/ADP/Mg2+ 1H2O solution 100 mmol/L tris at pH 7.0. For the evaluation of the concentration of PCr SBSR absorption spectra from Fig. 44 were used. Data points are displayed as filled circles. Data were fitted with the polynomial function y = b0 + b1· t + b2· t2. The solid line shows the result of the fit. With v0 = dc/dt (t=0) = 0.17 × 10-6 mol/(cm³min) a specific activity of 6 µmol/mg·min is derived by means of equation (40) (Γ = 8.1 × 10-13 mol/cm2, Vsys = 0.17 cm3, Oplate = 6 cm2, Mr 433 × 103 g/mol).


134

6.3.5 INTERACTION WITH MAGNESIUM ADP

FTIR spectra are sensitive to alterations upon changes of secondary structure caused by

substrate binding. They will also show the response of specific amino acid side chains

involved in substrate binding. Fig. 46 displays SBSR difference spectra of ch Mib-CK and hu

Mib-CK measured in the presence of 10 mmol/L MgADP. The MgADP solution was

pumped into the S- and the R-compartment. Thus its signal is compensated. The SBSR

absorbance spectrum measured previous to the exposure to MgADP was subtracted. Spectra

were recorded with two polarizations after the system was equilibrated. The first three

spectra compare parallel polarized SBSR difference spectra of two preparations of ch Mib-

CK and one of hu Mib-CK.

Whereas, the lower three show the corresponding spectra of perpendicular polarization. In

the middle, the spectra of water vapor (down scaled) is depicted, which clearly reveals that

measured frequencies are protein signals. The response signals of the ch Mib-CK

preparations are similar to one another. The different size can be attributed to the differences

in the number of active reflection Nact and the surface concentration Γ. The wavenumbers

and signs (to describe the direction) of the difference signals are listed in Table 6. To decide

if a band is positive or negative, it was related to the overall shape in proximity. Table 6 also

lists wavenumbers and assignment of the absorption of amino acid side chain and amide

bands taken from literature [23]. Shifts observed may be due to the overlapping of

unresolved bands. Besides, appearing as a shoulder of a higher peak shifts bands as well.

Although, there are distinctions between the three difference spectra, they share a lot of

features. A small positive band at 1400 cm-1 is found and can be assigned to contribution of

Glu and Asp. Also a very broad negative signal at 1450 cm-1, the amide II’ band is present

and can be deduced to changes of the amide backbone of the protein. Amino acid side chains


135

of Tyr cause the negative signals at 1514-1518 cm-1. Whereas, the two small but good

resolved negative bands at 1586-1590 cm-1 and 1607 cm-1 show the contribution of Arg.

Furthermore, two steep minima at 1627 cm-1 and at 1652 cm-1, as well as a positive

maximum at 1637 cm-1 are found. These are amid I’ frequencies and thus indicate again

changes of secondary structure elements. Thus FTIR spectra show the participation of Glu,

Asp, Arg and Tyr in the binding of the MgADP. The difference signals in the amid I’ region

infer that the substrate binding leads to a higher content of α-helices on the cost of β-sheets

and unordered parts of the enzyme.

1

2

3

4

5

6

7

Fig. 46. Polarized FTIR ATR spectra of the MgADP complex of ch Mib-CK and hu Mib-CK (2000-1200 cm-1). Comparison of ch Mib-CK and hu Mib-CK and two preparations of immobilized ch Mib-CK. Spectra were taken in the presence of 10 mmol/L MgADP. Compensation is achieved via SBSR technique. 1(||) and 5 (⊥) hu Mib-CK adsorbed to a DPPA/CL(bh)-bilayer (Nact 32.3); buffer 20 mmol/L tris, 150 mmol/L NaCl in 2H2O pH 6.6; 2(pp) and 6(vp) from ch Mib-CK adsorbed to a DPPA/CL(bh)-bilayer (Nact 32.3); buffer 20 mmol/L tris, 150 mmol/L NaCl in 2H2O pH 6.6; 3(pp) and 7(vp) from ch Mib-CK adsorbed to a DPPA-bilayer (Nact 24.5); buffer: 20 mmol/L phosphat, 100 mmol/L NaCl in 2H2O pH 6.6; 4 = watervapor spectra;

6 In

vest

igat

ion

of Im

mob

ilized

Mito

chon

dria

l Cre

atin

e Ki

nase

13

6

Tabl

e 6.

Wav

enum

bers

and

ass

ignm

ent o

f pea

ks o

f the

SB

SR d

iffer

ence

spec

tra

of th

e M

gAD

P-co

mpl

ex o

f ch

Mib

-CK

and

hu

Mib

-CK

. W

aven

umbe

rs a

nd a

ssig

nmen

t wer

e ta

ken

from

[23]

. Mea

sure

d si

gnal

s wer

e re

late

d by

list

ing

the

obse

rved

shift

s. Th

e si

gn (+

and

-) g

ives

the

diff

eren

ce w

ith

resp

ect t

o th

e sp

ectru

m in

the

vici

nity

. ++

or –

was

use

d fo

r int

ense

sign

als.

--- w

as u

sed

for m

issi

ng, a

nd ?

for a

mbi

guou

s sig

nals

.

hu M

ib-C

K o

n

DPP

A/C

L(bh

)-bi

laye

r ch

Mib

-CK

on

DPP

A/C

L(bh

)-bi

laye

r ch

Mib

-CK

on

DPP

A-b

ilaye

r

wav

enum

ber

in c

m-1

as

sign

men

t sh

ift in

cm

-1

sign

sh

ift in

cm

-1

sign

sh

ift in

cm

-1

sign

1405

14

07 ν s

CO

O-

(Glu

,Asp

) -5 -7

+ -1

.4-3

.4+

-1.4

-3

.4

+

1450

sec

. stru

ctur

e 0

- (v

ery

broa

d)

+5-

(ver

y br

oad)

+0

.8

-

1487

sec

.stru

ctur

e 0

+ 0

+ --

- --

- 15

15 ν

ringC

C a

nd

dCH

(Tyr

) -1

- +3

.5-

+3.5

-

1541

νrin

gCC

(Tyr

)

15

45 s

ec. s

truct

ure,

Trp

0

- +2

- --

- --

- 15

67 ν

asC

OO

- (Glu

) +3

+ --

---

- --

- --

- 15

84 ν

asC

OO

- (Asp

) +6

?- (

shou

lder

) +2

- +1

-

158

6 ν s

CN

3H5+ (A

rg)

+4?

- (sh

ould

er)

0

-1

16

08 ν

asC

N3H

5+ (Arg

) -1

.5-

-1.5

- -1

.5

- 16

27 a

mid

I’ (t

urns

, β-

shee

t) -1

--

--

--

1637

am

id I’

(α-h

elix

) -1

++

0++

0

++

1652

am

id I’

(ran

dom

) 0

--

0--

0

--

1672

am

id I’

(tur

ns)

---

---

0-

? ?

1692

am

id I’

(β-s

heet

) 0

- 0

- ?

? 17

41 ν

CO

OH

(te

rmin

al),

este

r (lip

ids)

0

- 0

- --

- --

-


137

6.3.6 BINDING OF CL(BEEF HEART) TO ADSORBED MITOCHONDRIAL CREATINE KINASE

The sarcomeric Mib-CK from chicken as well as from human was found to bind CL(bh)

presented as vesicle solution. The FTIR ATR spectra thus confirm the ability of Mib-CK to

bridge lipid membranes as shown by Rojo et al. [80]. Fig. 47 shows CL(bh) binding to ch

Mib-CK that occurred from lipid that was set free from the tubings. At the same time the

CL(bh) adsorbs to immobilized ch Mib-CK, the spectra of the R-compartment shown on the

top of Fig. 47, reveal a net loss of lipids. The surface concentration of adsorbed CL(bh) was

calculated with the thin film approximation from the Amax of ν(C=O) at 1740cm-1. A Γ=

6.4 × 10-11 mol cm-2, comprising 60% of the CL(bh) monolayer shown in the middle, and a

dichroic ratio R of 1.79 was found. We assume that CL(bh) adsorbs as bilayer patches to the

ch Mib-CK. Thus the spectra show that about 30% of the immobilized protein are covered

with CL(bh). Fig. 48 shows the result of immobilized hu Mib-CK exposed to a diluted

(40µg/ml) CL(bh) vesicle solution. Again only for the S-compartment binding of CL(bh) is

observed, while the R-compartment loses lipids. The surface concentration of the hu Mib-CK

was Γ= 8.2 × 10-13 mol cm-2, corresponding to a density of coverage of about 40%18. Spectra

are conventional absorbance spectra and show a slight incompensation for 1H2HO and 2H2O.

For the CL(bh) adsorbed on the protein a surface concentration of Γ= 4.3 × 10-11 mol cm-2

was calculated with the thin film approximation from the Amax of ν(C=O) at 1740 cm-1. This

value corresponds to about 22% of a CL(bh) bilayer, inferred from results of the CL(bh)

outer layers formed on DPPA. Thus half of the surface of the protein may now be covered

with a CL(bh) bilayer.

18 ) calculated with a side length of 9 nm (see also 6.4 Discussion)


138

It is discussed in literature [82], whether substrates and products enter and leave the active

center via the central channel or a so called „backdoor“ mechanism. If CL(bh) is adsorbed on

the top of the enzyme it will block the central channel at least partly. Thus a comparison of

the activities prior and after the exposure to the CL(bh) vesicle solution should give smaller

1

2

3

4

5

6

Fig. 47. Immobilized ch Mib-CK is able to bind CL(bh). Polarized FTIR ATR absorbance spectra from the R-(1:||, 2: ⊥) and S-compartment (5:||, 6: ⊥) after additional contact to lipid in the S-compartment. The lipid was washed out from the tubings used for the bilayer preparation. In the middle (3:||,4: ⊥) the CL(bh) signal gained quite after the preparation of the CL(bh) outer leaflet in the S-compartment is shown. Reference: R: bilayer/buffer (2H2O); S: bilayer+ch Mib-CK/buffer (2H2O); Surface concentration of the ch Mib-CK: Γ= 1.00 × 10-12 mol cm-2, corresponding to a density of coverage of 52%. Spectra are conventional absorbance spectra. The surface concentration of adsorbed CL(bh) was calculated with the thin film approximation from the Amax of ν(C=O) at 1740cm-1. A Γ= 6.9 × 10-11 mol cm-2 (60% of the CL(bh) layer shown in the middle) and a dichroic ratio R of 1.79 was found. Thus the spectra show a loosely packed second layer on top of the immobilized protein (θ = 45°, Nact= 32.31 , n1= 4.0, n2= 1.45, n3= 1.33 (2H2O);


139

values for the latter. But in situ FTIR activity measurements revealed similar activities of the

enzyme with and without adsorbed CL(bh). Whereas, a blocking should lead to a decrease of

about 50%. However, the CL(bh) may be adsorbed as single molecules on top of the protein

at the corners rather than at the central channel. Furthermore, the adsorbed CL(bh) was not

bound tightly, it loses an amount of 20% within 5 h.

1

2

3

4

Fig. 48. Immobilized hu Mib-CK is able to bind CL(bh). Polarized FTIR ATR absorbance spectra from the R-(1: ||, 2: ⊥) and S-compartment (3: ||,, 4: ⊥,) after additional contact to CL vesicles. A vesicle solution (40 µg CL/ml 20 mmol/L tris 150 mmol/L NaCl) was pumped through the R and S compartment; references: R: bilayer/buffer (1H2O); S: bilayer+ch Mib-CK/buffer (1H2O); Surface concentration of the hu Mib-CK: Γ= 8.2 × 10-13 mol cm-2, corresponding to a density of coverage of 42%. Spectra are conventional absorbance spectra with a slight incompensation for 1HO2H. The surface concentration of adsorbed CL(bh) was calculated with the thin film approximation from the A max of ν(C=O) at 1740cm-1 and found to be Γ= 4.3 × 10-11 mol cm-2 (22% of a CL(bh) bilayer). Furthermore, a dichroic ratio R of 1.88 was found. (θ = 45°, Nact= 32.31 , n1= 4.0, n2= 1.45, n3= 1.26 (1H2O);


140

6.4 DISCUSSION

We have prepared negatively charged planar bilayers immobilized on an ATR MIRE which

provide either CL or DPPA as outer leaflet for the adsorption of Mib-CK. The adsorption of

the protein was observed and found a fast process leading to surface coverage densities of

40-60% for ch Mib-CK. The variation of the surface concentrations Γ 1.2 to 0.9 × 10-12

mol/cm2 for different preparations was independent from the surface charge density, which

varied between 2 and 4 × 10-21 C/Å2 for the outer layers (see Fig. 49). Furthermore, there was

no response to the order of the outer leaflet, but Fig. 50 shows grouping due to the different

Smol values of the phospholipids used and the variation of the preparations. For hu Mib-CK

surface concentrations Γ 9 ×10-13 mol/cm2 were found. Applying a side length of 145 Å as

given in [82] would result in surface coverage densities of more than 100%. Thus either the

hu Mib-CK was packed more tightly on the bilayers than on CL monolayers or partially

multilayers were formed. Adsorption occurred as fast as the exchange of solutions at the flow

through rate chosen and hence, it was not possible to resolve the adsorption kinetic.

Additionally, we found unspecific adsorption of Mib-CK to surfaces like Ge, glass and

plastic (Delrin, Ismapren). The first was observed by FTIR ATR, whereas, the latter two

can be deduced from activity measurements. This shows that Mib-CK binds also to

hydrophobic surfaces and adsorption to surfaces can occur not only via electrostatic

interactions.

The phospholipid bilayers made up solely of negatively charged phospholipids were found

less stable than the reported DPPA/POPC bilayers [3]. This is due to electrostatic repulsion

of their headgroups.


141

20

30

40

50

60

70

80

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035

IzI/Amol in 1.6x10-19C/Ang2

cove

rage

(ch

Mi-C

K) /

%

Fig. 49. Correlation test for the influence of surface charge density exhibited by the outer layer on the adsorption of ch Mib-CK. The surface concentrations Γ of the prepared phospholipid layers were used to calculate a molecular area Amol. Furthermore, the different charges of CL (-2) and DPPA (-1) are included in |z|/Amol. The percentage of coverage of protein is (NA/Γ)/Amol*100, with Amol = 932 Å2 (from x-ray data [5]). The surface concentration of the protein was determined by the amide I’ or amide II band, using the thin film approximation. The signs correspond to the different outer layers; circle: CL(bh); square: CL (E.Coli); triangle: DPPA;

20

30

40

50

60

70

80

0.00 0.20 0.40 0.60 0.80 1.00 1.20

Smol(Lipid)

cove

rage

(MiC

K) /

%

Fig. 50. Correlation test for the influence of the order of the phospholipid bilayer on the adsorption of ch Mib-CK. Smol of the prepared phospholipid layers was calculated by equation. 26 and 27 (section 2). The three types of layers spilt in three groups. For CL(bh) Smol is 0.06, for CL(E.Coli) values between 0.48 and 0.64 were found and for DPPA values are about 1. Every square resembles the initial surface concentration of one immobilization experiment.

The Mib-CK clearly enhanced the stability of the whole model membrane, as can be expected


142

from their lysine and arginine enriched putative binding motives at the C terminal for the

phospholipids [77]. Fig. 47 and Fig. 48 show that the R-compartment is losing lipids (mainly

DPPA), whereas the immobilized protein in the S-compartment prevents the CL outer leaflet

and additionally binds CL. It is described in literature that Mib-CK can bridge membranes [4]

and lipid monolayers to vesicles [80]. Thus the observed binding of CL (bh) may be taken as

a proof for the fact that the protein is immobilized in a native like structure (as octamer). The

proposed hydrophobic interaction with the lipid layer or an embedding of the enzyme into

the layer [7, 79] was observed neither with the asymmetric DPPA/CL-bilayers nor with

DPPA-bilayers (transferred at 30 mN/m). Whereas, the latter are tightly packed and surely

too rigid to allow an embedding, this is not the case for the CL outer leaflets. Therefore, one

would expect changes in the frequencies and dichroic ratio of typical lipid bands, like CH-

stretching and C=O-stretching bands caused by disturbance from embedded protein. The

FTIR ATR spectra did not reveal such effects. These findings point to an adsorption to the

phospholipid membranes predominantly due to electrostatic interactions. Furthermore, the

adsorption should take place in an ordered manner, because the binding sites are located on

the top and on the bottom of the cubic like shaped enzyme and should be preferable in

contact with the negatively charged bilayer surface.

This generally will have consequences for polarized measurements resulting in bands with

dichroic ratios R ≠ Riso. The polarized FTIR ATR spectra of the enzyme did not show

polarized bands. For the amide I/I’ and the amide II bands of ch Mib-CK on phospholipid

bilayers we found dichroic ratios R of 1.72 ± 0.08. The values for ch Mib-CK on Ge and for

hu Mib-CK on phospholipid bilayers were slightly smaller. The interpretation in terms of the

order of the adsorbed protein is complicated by two facts: (i) The refractive index of the

protein-lipid layer is not precisely known. It was estimated between 1.40 and 1.45. For


143

isotropic samples the thin film approximation gives for the amide I’-band dichroic ratios of

1.76 and 1.65, respectively. For the amide II values of 1.79 and 1.67, respectively, were

calculated. (ii) Because of the huge number of amide groups (3240 for ch Mib-CK and 3590

for hu Mib-CK) contributing to the signal and the shape (cubic symmetry; P422) of the

enzyme. In spite of the structural integrity of the protein, amide groups will be randomly

distributed in space and polarization effects can easily be canceled. Thus in this case it is not

possible to detect an ultrastructure, like the formation of 2D crystals, by amid bands.

The analysis shall be continued with the Arg and Lys of the binding site. Only 8% of all

amino acids are Arg and 6% are Lys. Their frequencies in 1H2O are 1673 cm-1 and 1633 cm-

1, and 1629 cm-1 and 1526 cm-1, respectively. Arg absorbs at 1608 cm-1 and 1586 cm-1 in

2H2O. Clearly, these bands cannot be resolved from the amide I/I’ and amide II bands

without curve fitting. A curve fitting of a signal as complex as the amide I/I’ will not be

precise enough for the determination of dichroic ratios R. Arg and Lys can only be separately

observed in the FTIR ATR spectra if they are chemically modified. However, the

modification may disturb the native structure of the enzyme to a great extent.

For this reasons, the determination of the ultrastructure of ch Mib-CK on DPPA bilayers was

performed by electron microscopy. It revealed no 2D crystals but an amorphous tightly

packed protein surface (Fig. 62, Appendix) or probably a liquid crystalline phase. This is in

contrast to the results of EM measurements, which show that Mib-CK (ch as well as hu)

builds 2D crystals [81, 82] when it is exposed to CL monolayers.

We suggest that the contact to air and the interaction with heavy metal ions used for negative

staining might be necessary for 2D crystal formation. These are the only parameters which

differ between the immobilization on ATR plates and adsorption of Mib-CK to CL

monolayers used in [81, 82].


144

FTIR can be used to detect secondary structure elements and it is especially sensitive for β-

sheets. Thus we analysed the FTIR ATR spectra to gain informations about the secondary

structure of immobilized Mib-CK. For crystallized ch Mib-CK 13% antiparallel β-sheets,

33% α-helices and 53% unordered regions as well as a 310-helix were determined by x-ray

diffraction [5]. These values were used to calibrate the amide I/I’ bands that are typical for a

certain secondary structure. Then they were summed to give a synthetic amide I’ band. Fig.

51 compares the measured amide I’ signals with bands constructed from signals of the model

compound poly-L-lysin. Depending on the pH, poly-L-lysin adopts either a structure with

α-helices (pH < 6), β-sheets (pH > 10), or meta stabile helices only. The latter was used to

model the contribution of unordered regions. In general, the frequencies of typical secondary

structure elements can be different in a protein from polypeptides. Thus we also used data for

literature [23], comprising mean values of a set of proteins, to construct a synthetic spectrum

of a amid I’ band with 13% antiparallel β-sheet. Comparing the results with the measured

spectra we find that there is most probably a lower content of β-sheets than 13%. Fig. 51 also

displays the complexity of amide bands arising from secondary structure and from the amino

acid side chain absorptions. The 8% Arg and 13-14% acidic amino acids the Mib-CKs

contain were included in the calculation. The overall shape of the amide I/I’ band observed in

this work is similar to the one reported for solubilized active dimeric CK from rabbit muscle

[110], for which the content of antiparallel β-sheets is 14% and of α-helices is 31% as

determined by x-ray diffraction [111]. Therefore, the frequencies for antiparallel β-sheets

might be shifted from 1619 cm-1 to higher values in CKs or the contents of β-sheets may

vary depending on whether the enzyme is solubilized or crystalline.


145

νasCOO- (Asp,Glu)amid I': randomamid I': β-sheet

amid I': turns

νasCN3H5+(Arg)

νsCN3H5+(Arg)amid I': β-sheet

amid I': α-helix

Fig. 51. Synthetic and measured amid I’ bands to estimate the secondary structure elements of immobilized ch Mib-CK. Top: FTIR absorbance spectra [112] of poly-L-lysin in 2H2O at pH* 2 (α-helix), pH* 8 (random, transition state) and pH* 12 (β-sheet, antiparallel) were calibrateted with their proportion in ch Mib-CK as determined by x-ray [5]. The components (α-helix: 33%, random, transition state: 54%, β-sheet, antiparallel: 13%) and their sum (first spectrum) are shown. Middle: ||- and ⊥- FTIR ATR absorbance spectra of ch Mib-CK adsorbed to a DPPA/CL(bh) bilayer. Bottom: mean values of amid I’ datas from proteins as cited in [23] as well as Arg and Asp were used to construct a spectrum with the secondary structure contributions as known from x-ray diffraction and Arg and Asp contributions.

The question arises whether the enzyme was bound in a native form. For the function of the

enzyme in vivo 2D crystal formation does not seem to be essential. The Mib-CK will cluster

rather with ANT and porin than with one another [4]. Thus the fact, that 2D crystal formation

cannot be detected in systems immersed in buffer, does not necessarily mean that the protein

was denaturated. Facts that point to an immobilization in the native form are its ability to

bind CL and activity of the enzyme. It was possible to detect activity in situ with the FTIR

for up to 6 days. The activity was found to decay and given as specific activity the initial

values are about 20 µmol/mg·min. A straight forward comparison with data measured for


146

solubilzed enzyme is not possible. The difference to Vmax values of 95-140 µmol/mg·min

usually found [92, 94, 113] for the solubilized enzyme may be deduced to convection and

diffusion, which differ greatly for both systems. Furthermore, the accessibility for the

substrates to immobilized enzyme will be reduced if the enzyme adsorbs in tightly packed

patches. On the other hand, the slow 1H-2H exchange and the high content of N1H groups

even after a few days show that the structure was rigid, an unusual behavior for denaturated

proteins. Furthermore, the „resistant“ N1H-groups can be assigned to the α-helices of the

protein.

The immobilized enzyme was exposed to MgADP and FTIR ATR difference spectra of the

MgADP*Mib-CK-complex were recorded, applying the SBSR technique. The results

confirm that Glu and Asp, Tyr and Arg are involved in the binding of MgADP. Additionally,

we found an increase at 1637 cm-1 and a decrease at 1627 cm-1 and 1652 cm-1. Thus we

consider that an α-helix, as well as unordered parts and β-sheets are affected through the

complex formation. The result shows that a form with unordered parts transformed

to structures that are similar to α-helices.

During the experiment series lasting for a few days the immobilized enzymes were exposed

to mechanical stress of 6-8 h of pumping with velocities of about 10 mm/min. This led to

losses of ca. 30% of the initial adsorbed protein. For immobilizations used 1-2 days the loss

was found to be neglectible. Using in situ FTIR ATR SBSR measurements it was possible to

determine the functional stability of the immobilized enzyme within a few days. Generally,

we found a decay of activity of 50% from the initial value to values measured 24 hours later.

Hence, the data collected shed some light on the limits with which designers of biosensors

are confronted when working with immobilized enzymes.

7 Appendix

147

7 Appendix

7.1 AD SECTION 2:

Dependence of R, Sseg and the mean angle α from refractive indices used for the calculation

of the components of the electric field (thin film approximation)

0

10

20

30

40

50

60

70

80

90

0 0.5 1 1.5 2 2.5

R

alfa

/[°]

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

S seg

alfa2850 [°] airSseg2850 air

n1(Ge)= 4.0n2(thin film)= 1.45n3(air)=1.0

0

10

20

30

40

50

60

70

80

90

0 0.5 1 1.5 2 2.5

R

alfa

/[°]

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

S seg

alfa2850 [°] water

Sseg2850 water

n1(Ge)= 4.0n2(thin film)= 1.45n3(water)= 1.41

A

B B

7 Appendix

148

0

10

20

30

40

50

60

70

80

90

0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25

R

alfa

/[°]

-0.60

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

S seg

alfa1645 [°] Sseg1645

n1(Ge)= 4.0n2(thin film)= 1.45n3(water)= 1.32

0

10

20

30

40

50

60

70

80

90

0 0.5 1 1.5 2 2.5

R

alfa

/[°]

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

S seg alfa1352 [°]

Sseg1352

n1 (Ge) = 4.0n2 (thin film) = 1.5n3 (water) = 1.30

C

D

7 Appendix

149

0

10

20

30

40

50

60

70

80

90

0 0.5 1 1.5 2 2.5

R

alfa

/[°]

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

S seg

alfa1080 [°]Sseg1080

n1 (Ge) = 4.0n2 (thin film) = 1.5n3 (water) = 1.26

Fig. 52. The interdependences of the dichroic ratio R, segmental order parameter Sseg and the mean angle of the transition dipole moment (alfa) in °. Sseg and alfa were computed using equation (25) and (26). The electric field components were calculated with the thin film approximation (angle of incidence θ = 45°; refractive indices as indicated in the diagrams). A and B show the results for 2850 cm-1 (CH-stretching region), C for the amide I/I’ band and C; D for two marker bands for the trichlorophenol. Filled triangles are values for the Sseg and refer to the right axes. Whereas, open triangles are values of the mean angle (alfa) and refer to the left axes. The results illustrates the effect of changes of the refractive index of bulkwater due to anomalous dispersion.

E

7 Appendix

150

7.2 AD SECTION 5:

7.2.1 RESULTS OF THE UV VIS MEASUREMENTS

0

0.1

0.2

0.3

0.4

0.5

0.6

200 250 300 350 400 450 500

wavelength in nm

Abs

orba

nce

pH1.8

pH5.9

pH7.3

pH11.1

Fig. 53. UV VIS absorbance spectra of 2,4,5-trichlorophenol at different pH values. TCP solutions (c=2.9 mmol/L) in 25 mmol/L K-phosphate buffer or KOH (2 mmol/L) were mixed and/or adjusted to the pH, indicated in the diagram with HCl (1.5 mol/L). Then the sample was injected into a CaF2 cuvette with a 50µm spacer.

7 Appendix

151

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.5 1 1.5 2 2.5 3 3.5 4

c/mM

Abs

orpt

ion

(blc

) /A

U

293.24 nm (blc)229.26 nm (blc)203.3 nm (blc)Linear (229.26 nm (blc))Linear (293.24 nm (blc))

Fig. 54. Lambert-Beer plot for TCP solutions at pH 6. Peakheights at 293.2 nm (peak maximum), 229.3 nm (beside a maximum) and at 203.3 nm (peak maximum) were determined using a baseline correction (450 nm set to 0). Cuvette: CaF2, d = 0.209mm;

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.5 1 1.5 2 2.5 3

c/mM

Abs

orpt

ion/

AU

311.5 nm244.5 nm209 nmLinear (209 nm)Linear (244.5 nm)Linear (311.5 nm)

Fig. 55. Lambert-Beer plot for TCP dissolved in 2 mmol/L KOH. Peakheights at 311.5 nm (peak maximum), 244.5 nm (peak emerging) and at 209 nm (peak maximum) were determined using a baseline correction (450 nm set to 0). Cuvette: CaF2, d = 0.20mm;

7 Appendix

152

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

200 250 300 350 400 450 500

wavelength in nm

Abs

orba

nce

20 min (no dry air)

with dry air purge for: 15 min, 35 min, 165 minwithout dry air purge: after 2 has well as a control (0.01 mmol/L)

Fig. 56. Effects of aging and oxygen (dry air) tested with UV VIS spectra. 2,4,5-trichlorophenoxide (c = 1 mmol/L) in diluted KOH (pH 11) were treated with dry air for 2.75 h. Aliquots of 20 µL were taken after 15, 35 and 165 min, injected into 1.98 mL KOH and spectra were measured. For comparison spectra of aliquots of untreated solutions after 20 min and 2 h and a spectrum measured for standardization of phenoxide is shown. The latter one was rescaled and confirms that there is no severe loss and no alteration.

7 Appendix

153

7.2.2 RESULTS OF THE MEMBRANE PREPARATION

Table 7. Properties of prepared model membranes at 25°C. Number of active reflections Nact and integrated absorbances for parallel (int App) and perpendicular (int Avp) polarization are listed. The dichroic ratios R and surface concentration G, as well as the area per molecule (Amol) and the molecular order parameter (Smol) were calculated by means of equation (27)and (29). DPPA date Nact int App int Avp R Γ/Γ/Γ/Γ/10-10

in mol/cm² Amol in Å2

Smol

27/09/2000 30.0 0.328436 0.350112 0.938 4.6 36 0.97

11/10/2000 30.9 0.313705 0.338217 0.928 4.3 39 1.04

02/10/2000 29.8 0.31348 0.33468 0.937 4.5 37 0.98

16/10/2000 31.5 0.311361 0.332985 0.935 4.2 40 0.99

11/07/1998 27.6 0.265755 0.28655 0.927 4.0 41 1.04

25/11/1998 26.9 0.261088 0.281189 0.929 4.1 40 1.03

25/02/1999 43 0.385087 0.412189 0.934 3.8 44 0.99

09/11/1999 36.3 0.352545 0.376747 0.936 4.1 40 0.98

MW 0.933 4.2 40 1.00

σ 0.004 0.3 3 0.03

POPC

date compartment19

Nact int App int Avp R Γ/Γ/Γ/Γ/10-10 in mol/cm²

Amol in Å2

Smol

03/10/2000 Rr 13.5 0.095025 0.051426 1.9 2.0 83 0.0220

Sr 13.5 0.102655 0.063815 1.6 2.3 73 0.1920

17/10/2000 Rr 13.5 0.089047 0.062898 1.4 2.1 79 0.36

Sr 13.5 0.092505 0.065127 1.4 2.2 76 0.35

26/02/1999 Rr 18.4 0.093260 0.060560 1.5 1.6 107 0.25

Sr 18.4 0.097508 0.063170 1.5 1.6 105 0.25

03/03/1999 Rf 18.4 0.127301 0.095466 1.3 2.3 74 0.43

Sf 18.4 0.10239 0.076111 1.4 1.8 92 0.42

MW 1.5 2.0 86 0.28

σ 0.2 0.3 14

19 ) abbreviations: Nact... number of active reflections, Int App and Int App ...integrated absorbances parallel (pp) and perpendicular polarized (vp), R ... reference compartment, S... sample compartment, r ... rear side, f ... front side of the plate 20) The vesicle solution was left in the SBSR-flow-through cell for 14 h.

7 Appendix

154

7.2.3 FURTHER RESULTS OF TIME RESOLVED MEASUREMENTS

wavenumber / cm-1

1000120014001600180020003000

abso

rban

ce /

AU

-0.02

0.00

0.02

0.04

0.06A

38 min (S)

37 min (R)

38 min (SBSR)142 min (R)

143 min (S)

143 min (SBSR)

wavenumber / cm-1

1000120014001600180020003000

abso

rban

ce /

AU

0.00

0.02

0.04B

48 min (S)

47 min (R)

48 min (SBSR)

152 min (R)

153 min (S)

153 min (SBSR)

Fig. 57. Polarized ATR absorbance spectra of sample (S) and reference (R) compartment and SBSR ATR absorbance spectra. TCP solution (cTCP= 2.1 mmol/L) in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L) was in contact with a DPPA monolayer at 25°C (-0.42V). (A) parallel (||) polarized absorbance spectra measured after exposure of 38 min and 143 min. (B) perpendicular (⊥) polarized spectra after exposure of 48 min and 153 min. A dashed line marks a peak emerging at 1446 cm-1. Reference: DPPA-monolayer against 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L); Measurement conditions: Ge trapezoid as IRE, angle of incidence θ = 45°, active internal reflections Nact= 13.46;

7 Appendix

155

t / min0 200 400 600 800 1000

surfa

ce c

once

ntra

tion

/10-1

0 mol

·cm

-2

0.00

2.00

4.00

6.00

8.00

10.00

t / min0 20 40 60 80 100

surfa

ce c

once

ntra

tion

/10-1

0 mol

·cm

-2

0.00

2.00

4.00

6.00

8.00

10.00

Fig. 58. Long term and short term behavior of the surface concentrations of phenol (HA) and phenoxide (A-) on a DPPA monolayer. TCP (csol= 2.9 mmol/L) was pumped over a DPPA monolayer (0.05 ml/min). From the spectra (not shown) peak heights at 1080 cm-1 and 1352 cm-1 were determined and fitted with f(t)=d+g⋅⋅⋅⋅(1-exp(-k⋅⋅⋅⋅t)+l⋅⋅⋅⋅t. The fit results (see

Table 8) were used to calculate the surface concentration Γ with the thin film approximation. Curves were determined with the help of f(t), whereas squares (for HA) and circles (for A-) denote distinct values, calculated with measured and interpolated peak heights. Molar absorption coefficients: HA (1080 cm-1) 1.85±0.08 ×105; A-(1352cm-1)1.5±0.2 ×105; angle of incidence θ = 45°, active internal reflections Nact= 13.46; T = 25°C, U = 0.0V, no insulation of the Ge; dotted lines represent the standard deviations of the curves (includes standard deviation of fit parameters and of the molar absorption coefficient).

7 Appendix

156

t / min0 20 40 60 80 100

surfa

ce c

once

ntra

tion

/ 10-1

0 mol

·cm

-2

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Fig. 59. Surface concentrations of phenol (HA) and phenoxide (A-) on a DPPA monolayer vs. time. The peak heights at 1080 cm-1 and 1352 cm-1 were determined and fitted. The fit results were used to calculate the surface concentration Γ with the thin film approximation. Curves were determined with the help of f(t), whereas squares (for HA) and circles (for A-) denote distinct values, calculated with measured and interpolated peak heights. Molar absorption coefficients HA (1080 cm-1) 1.85±0.08 ×105; A-(1352cm-1)1.5±0.2 ×105; angle of incidence θ = 45°, active internal reflections Nact= 13.46; T = 25°C, U = -0.42V; concentration of TCP csol: 1.5 mmol/L. The dotted lines represent the standard deviations of the curves (includes standard deviation of fit parameters and of the molar absorption coefficient).

7 Appendix

157

Table 8. Parameters of the functions f(t)=d+g(1-exp(-k⋅⋅⋅⋅t)+l⋅⋅⋅⋅t used to fit the peak heights at 1352 and 1080 cm-1.

run wavenumber in cm-1 andpolarization

g in 10-3AU

k in min-1

d in 10-4AU

l in min-1

Rsqr σ ×10-5

2.9 mmol/L (0 V,i)

1352 pp/|| (>30min)

1.461.50

0.0220.026

1.901.50

---1.16×10-6

0.957 4.9688.03

1352 vp/⊥ 8.89 0.022 1.32 --- 3.453 1080 pp/||

(>30min) 1.581.47

0.0430.048

3.353.14

---1.86×10-6

9.3288.801

1080 vp/⊥ (>30min)

1.051.00

0.0360.038

2.782.77

---6.71×10-7

0.965

7.7797.771

2.1 mmol/L (-0.42 V)

1352 pp/|| 1.65 0.034 4.70 --- 0.971 9.706

1352 vp/⊥ 0.81 0.032 3.04 --- 0.93 7.39 1080 pp/|| 1.80 0.051 8.55 --- 0.967 11.04 1080 vp/⊥ 0.96 0.059 6.79 --- 0.927 8.3371.5 mmol/L (-0.42V)

1352 pp/|| 0.65 0.053 1.46 -4.51×10-6 0.874 4.948

1352 vp/⊥ 0.29 0.149 0.04 -1.43×10-6 0.615 5.012 1080 pp/|| 0.98 0.063 2.94 -6.31×10-6 0.879 8.03 1080 vp/⊥ 0.69 0.084 2.23 -4.54×10-6 0.796 7.384

7 Appendix

158

Table 9. Parameters of the functions f(t)=d+g(1-exp(-k⋅⋅⋅⋅t)+l⋅⋅⋅⋅t used to fit the peak heights at 1352 and 1080cm-1.

run wavenumber in cm-1 andpolarization

g in 10-3AU

k in min-1

d in 10-4 AU

l in min-1

Rsqr σ

2 mmol/L(0 V,i)

1352 pp/|| 1.64 0.048 0.75 0 0.98 6.458×10-

5

1352 vp/⊥ 0.84 0.059 0.90 0 0.947 5.709×10-

5

1080 pp/|| 6.19 0.249 9.37 0 0.983 2.495×10-

4

1080 vp/⊥ 3.45 0.263 4.82 0 0.915 2.49×10-4

2.7 mmol/L(0 V,i)

1352 pp/|| 0.60 0.019 13.1 187e-9 0.982 40.63×10-

6

1352 vp/⊥ 0.68 0.012 1.76×10-3/7.64×10-4

0.605 4.49×10-5

1080 pp/|| 8.96 0.525 0.35 2.76×10-3 0.977 2.86

1080 vp/⊥ 3.82 0.412 4.41×10-3/2.47×10-3

0.797 4.71×10-6

1.45 mmol/L(-0.34)

1352 pp/|| 1.05 0.066 0.116 0.95 7.341×10-

5

1352 vp/⊥ 0.52 0.067 6.75 0.89 5.345×10-

5

1080 pp/|| 3.175 0.787 14.56 1.948×10-

5 0.98 1.493×10-

4

1080 vp/⊥ 1.142 0.436 15.62 6.189×10-

6 0.924 8.478×10-

5

2.1 mmol/L(-0.34)

1352 pp/|| (quadratic)

b0: 2.009 b1: 13.23×10-

6

b2: -150.5×10-

5

0.845 4.359×10-

5

1352 vp/⊥ (quadratic)

b0: 1.153 b1: -95.83×10-

9

b2: -58.69×10-

5

0.316 7.588×10-

5

1080 pp/|| 3.57 1.072 2.77e-3 0.9881.059×10-

4 1080 vp/⊥ 2.04 1.57 1.37e-3 0.89 6.043×10-

5

7 Appendix

159

Table 10. Change of characteristic data of the lipid layers during exposure to TCP. The course of the surface concentration Γ, the dichroic ratio R and the order parameter Smol is given for all four experiments for the reference compartment (R) and the sample compartment (S), separately. Experiment series are indicated by the type of layer and electrostatic conditions (potential, current). Four to five distinct dates from every series are documented.

Experiment conditions

Γ in 10-10 mol cm-2

R Smol

sample /compartment: R S R S R S DPPA (0 V,i) before exposure 4.6 4.6 0.86 0.86 0.99 0.99 1.0 mmol/L TCP

for 30 min 4.4 4.4 0.89 0.91 0.96 0.93

2.9 mmol/L TCP for 48 min (0.5ml/min)

4.2 4.4 0.90 0.92 0.94 0.91

2.9 mmol/L TCP for 13 h (0.05ml/min)

2.7 3.2 1.11 1.10 0.67 0.68

DPPA (-0.42 V) before exposure 4.3 4.3 0.85 0.85 1.01 1.01 after 0.8 mmol/L TCP

washed with buffer 4.1 3.9 0.87 0.88 0.98 0.97

1.5 mmol/L TCP for 80 min

3.9 3.7 0.89 0.87 0.95 0.98

after 1.5 mmol/L TCP washed with buffer

3.8 3.7 0.91 0.89 0.93 0.96

2.1 mmol/L TCP for 150 min

3.6 3.5 0.92 0.89 0.91 0.96

DPPA/POPC bilayer (0 V, i)

before exposure 6.9 6.8 1.08 1.05 0.74 0.71


6.4 6.5 1.12 1.15 0.66 0.63

after 2.7 mmol/L TCP for 60 min (0.5and 0.2 ml/min)

6.4 5.5 1.11 1.16 0.67 0.62

after 2.7 mmol/L TCP for 14 h (0.05ml/min)

5.9 2.7 1.13 1.22 0.65 0.55

DPPA/POPC bilayer (-0.34 V)

before exposure 6.3 6.4 1.00 0.95 0.80 0.87


6.0 5.8 1.00 1.00 0.81 0.81

after 1.5 mmol/L TCP for 30 min

6.0 4.8 1.00 1.09 0.81 0.70


6.0 4.8 1.00 1.09 0.81 0.70

after 2.0 mmol/L TCP for 58 min

5.9 4.0 1.00 1.14 0.81 0.64

7 Appendix

160

Fig. 60. Polarized ATR absorbance spectra of sample (S) and reference (R) compartment and SBSR ATR absorbance spectra. TCP solution (cTCP= 2.0 mmol/L) in 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L) was in contact with a DPPA monolayer at 25°C (0V, i) for 50 min. Then the TCP solution was exchanged for buffer. Top: strong adsorbed TCP; Bottom: before washing; includes loosely bound TCP; (A) parallel (||) polarized absorbance spectra; (B) perpendicular (⊥) polarized spectra after exposure of A dashed line marks a peak emerging at 1446 cm-1. Reference: DPPA/POPC bilayer against 25 mmol/L potassium phosphate buffer pH 6.0 (ctotal (K+) 100 mmol/L); Measurement conditions: Ge trapezoid as IRE, angle of incidence θ = 45°, active internal reflections Nact= 13.46;

S

SBSR

R

S

SBSR

R

A

S

SBSR

R

S

SBSR

R

B

7 Appendix

161

7.3 AD SECTION 6:

7.3.1 AFM IMAGE OF A DPPA BILAYER

Fig. 61. AFM image of a DPPA bilayer immobilized on Si in buffer environment. Top: „Roughness“ and height of a 1µm x 1µm patch of the bilayer. Bottom: AFM top view image of a 5µm x 5µm patch and cross section at positions in line with a hole that allows to determine the depth of the bilayer. The images were produced with a Nanoscope III (Digital instruments, Inc. Santa Barbara, CA) in the contact mode. Scanner head: D (12 µm scan range); Probe: oxide-sharpened silicon nitride; spring constant: 0.06 N/m; buffer: 20 mmol/L phosphate, 0.1 mol/L NaCl, pH 7.0;

7 Appendix

162

7.3.2 EM: IMAGE OF A REPLICA OF THE IMMOBILIZED MIB-CK

Fig. 62. EM image of a replica of ch Mib-CK adsorbed to a DPPA-bilayer. The model membrane was immobilized on a Si-plate (10 x 10 x 0.5 mm³), frozen in N2 (l) and freeze- dried. Afterwards it was shadowed with Pt (45°, d=1.5 nm ) und C (90°,d=15 nm); The replica was removed with HF and put on a copper grid. The picture shows a part of the replica with a branched valley of a smoother surface. This surface belongs to the DPPA bilayer. Whereas, the protein causes the rougher surface. From the shadows at the edges a distance of about 12 nm between the two surfaces can be determined.

16

3

7.3.

3 TA

BLE

S

7.3.

3.1

Prop

ertie

s of

the

prep

ared

pho

spho

lipid

laye

rs

Tab

le 1

1. P

rope

rtie

s of t

he p

repa

red

phos

phol

ipid

laye

rs. Γ

was

cal

cula

ted

with

the

thin

film

app

roxi

mat

ion

eith

er fr

om th

e ν s

CH

2 (ε

νd~∫

=

5219

40 c

m/m

ol b

etw

een

2832

-286

cm

-1) o

r νC

=O(E

ster

) (ε

νd~∫

=1.1

0 x

107 c

m/m

ol, 1

769.

5-16

78cm

-1 c

alib

rate

d w

ith D

PPA

) int

egra

ted

with

m

etho

d B

(OPU

S).

type

of l

ayer

da

te o

f

prep

arat

ion

A pp

in c

m-1

A vp

in c

m-1

R

Γ ΓΓΓ

in m

ol/c

m²

A mol

in

Ang²

S mol

TR

Nac

t

DPP

A-M

L/ai

r

a

ngle

α 9

0°

8/

26/9

6 0.

3485

34

0.39

4402

0.

883.

42E-

10

49

1.34

99

%40

.73

9/

25/9

6 0.

3309

36

0.37

689

0.88

3.19

E-10

52

1.

39

94%

41.1

8

12

/2/9

6 0.

3812

84

0.39

7407

0.

963.

99E-

10

42

0.85

96

%41

.36

4/

7/97

0.

3928

06

0.40

9553

0.

964.

19E-

10

40

0.85

98

%40

.55

2/

2/98

0.

3823

44

0.40

6775

0.

944.

04E-

10

41

0.96

98

%40

.18

2/

15/0

0 0.

3626

37

0.38

4842

0.

944.

11E-

10

40

0.95

94

%37

.6

2/

26/0

0 0.

3257

73

0.33

9669

0.

964.

09E-

10

41

0.85

98

%34

.48

mea

n va

lues

0.93

4.08

E-10

41

0.

89

97%

stan

d. d

ev.

0.

047.

53E-

12

1 0.

06

2%

7 Ap

pend

ix

16

4

type

of l

ayer

da

te o

f

prep

arat

ion

A pp

in c

m-1

A vp

in c

m-1

R

Γ ΓΓΓ in

mol

/cm

²

A mol

in

Ang²

S mol

TR

N

of th

e la

yer

DPP

A-

Bila

yer/b

uffe

r

angl

e α

90°

R9/

2/98

0.

5240

46

0.57

7906

0.

91

3.92

E-10

420.

94

105%

1

S

0.57

0025

0.

6495

09

0.88

4.

40E-

1038

0.99

98

%

2

R11

/16/

98

0.46

826

0.50

6155

0.

93

3.35

E-10

500.

91

104%

1

S

0.47

4759

0.

5146

14

0.92

3.

41E-

1049

0.92

98

%

2

R5/

5/99

0.

4986

33

0.55

9843

0.

89

3.72

E-10

450.

97

95%

1

S

0.49

2086

0.

5515

13

0.89

3.

66E-

1045

0.97

93

%

2

R8/

12/9

9 0.

4529

4 0.

511

0.89

4.

37E-

1038

0.97

10

3%

1

S

0.46

1982

0.

5294

69

0.87

4.

52E-

1037

1.00

91

%

2

mea

n va

lues

3.

92E-

10

0.90

4.

10E-

1041

0.96

98

%

stan

d. d

ev.

11.7

8

0.

02

3.78

E-11

40.

03

5% c

16

5

type

of l

ayer

da

te o

f

prep

arat

ion

A pp

in c

m-1

A vp

in c

m-1

R

Γ ΓΓΓ

in m

ol/c

m²

A mol

in

Ang²

S mol

lo

ss

CL(

E.C

holi)

/buf

fer

ca. 2

0% u

nsat

. fat

ty a

cids

an

gle

α 90

°

flow

thro

ugh

8/27

/96

0.17

7247

0.15

5881

1.14

1.

99E-

1083

0.

64

flow

thro

ugh

9/26

/96

0.14

4592

0.11

4016

1.27

1.

54E-

1010

8 0.

50

9/

27/9

6 0.

1296

330.

0960

654

1.35

1.

34E-

1012

4 0.

4213

% in

14h

flow

thro

ugh

12/4

/96

0.18

7192

0.16

7206

1.12

2.

09E-

1080

0.

48

12

/5/9

6 0.

1602

90.

1366

971.

17

1.78

E-10

93

0.37

15%

in 1

4h

R

4/9/

97

0.33

7146

0.25

8985

1.30

1.

56E-

1010

6 0.

46

S

0.32

1525

0.24

9932

1.29

1.

50E-

1011

1 0.

48

R

2/5/

98

0.11

3113

0.08

8124

81.

28

1.05

E-10

157

0.48

S

0.10

9377

0.08

5124

11.

28

1.02

E-10

163

0.48

mea

n va

lues

1.

49E-

10

1.26

1.

69E-

1010

1 0.

46

stan

d. d

ev.

23.9

2 0.

07

2.76

E-11

16

0.04

7 Ap

pend

ix

16

6

type

of l

ayer

da

te o

f

prep

arat

ion

A pp

in c

m-1

A vp

in c

m-1

R

Γ ΓΓΓ

in m

ol/c

m²

A mol

in

Ang²

S seg

CL(

bh)/b

uffe

r c

a.10

0% u

nsat

. fat

ty a

cids

R

2/17

/00

0.29

2281

0.

1810

39

1.61

1.10

E-10

15

1 -0

.02

S

0.29

8366

0.

1860

25

1.60

1.12

E-10

14

8 -0

.03

R

2/28

/00

0.28

3203

0.

1916

62

1.48

1.09

E-10

15

2 -0

.09

S

0.28

4797

0.

1930

4 1.

481.

10E-

10

151

-0.0

9

R n

ach

10h

0.30

5903

0.

2003

1.

531.

17E-

10

142

-0.0

6

S

0.24

342

0.16

337

1.49

9.37

E-11

17

7 -0

.08

mea

n va

lues

1.

09E-

10

1.53

1.09

E-10

15

4 -0

.06

stan

d. d

ev.

7.20

0.

067.

82E-

12

12

0.03

7 Appendix

167

Table 12. Dichroic ratios for different bands of DPPA of the R-and S-compartment during an experiment series. Time t starts with the spectra measured after the filling of the SBSR cell.

R-compartment:

Label / t in h 0 21 24 43 92 137

vs(CH2) 0.92 0.99 0.98 0.98 1.01 1.02

Ester 1.13 1.00 1.17 1.10 1.15 1.11

CH wagg 0.92 0.97 1.00 1.02 1.02 1.03

S-compartment:

Label / t in h 0 21 24 43 92 137

vs(CH2) 0.92 0.94 0.91 0.92 0.92 0.92

Ester 1.17 1.22 1.13 1.11 1.17 1.11

CH wagg 0.99 0.98 1.03 1.01 1.00 1.03

7 Appendix

168

Table 13. Loss of DPPA of a bilayer in the R- and S-compartment during an

experiment series. The νs(CH2) of parallel (pp) and perpendicular polarization (vp) were

integrated.

R-compartment: t[h] pp vp loss pp[%] loss. vp[%]

0 0.470079 0.509267

21 0.405572 0.409962 -13.7 -19.5

24 0.344361 0.350922 -26.7 -31.1

43 0.329432 0.33486 -29.9 -34.2

92 0.30821 0.305278 -34.4 -40.1

137 0.300236 0.292946 -36.1 -42.5

S-compartment: t[h] pp vp loss pp[%] loss vp[%]

0 0.477302 0.517568

21 0.409983 0.435604 -12.8 -14.5

24 0.356429 0.392317 -24.2 -23.0

43 0.339172 0.370626 -27.8 -27.2

92 0.327142 0.357142 -30.4 -29.9

137 0.324761 0.352428 -30.9 -30.8

16

9

7.3.

3.2

Res

ults

of t

he im

mob

ilizat

ion

of M

i b-C

K

Tab

le 1

4. R

esul

ts o

f the

imm

obili

zatio

n of

Mi b-

CK

on

diff

eren

t pho

spho

lipid

mem

bran

es. T

he ta

ble

show

s the

inte

gral

s of a

mid

e I/I

’(ε

νd~∫

= 2.

74 x

107 c

m/m

ol; b

etw

een

1698

cm

-1 a

nd 1

595

cm-1

) and

am

ide

II ba

nds (

ενd~

∫=

8.25

x 1

06 cm

/mol

bet

wee

n 15

85 c

m-1

-150

0 cm

-1)

as d

eter

min

ed in

the

vario

us e

xper

imen

ts. T

hey

wer

e us

ed to

cal

cula

te th

e di

chro

ic ra

tios R

and

surf

ace

conc

entra

tions

Γ (b

y th

e th

in fi

lm

appr

oxim

atio

n, a

pply

ing

the

LCU

mod

el).

For c

onve

nien

ce th

e Γ

valu

es w

ere

conv

erte

d in

to a

mol

ecul

ar a

rea

and

the

hypo

thet

ical

side

leng

th o

f a

squa

re is

giv

en a

s am

ol. T

his v

alue

shou

ld b

e co

mpa

red

to th

e si

de le

ngth

of 9

3 A

ng. f

or M

i b-C

K a

s det

erm

ined

by

x-ra

y di

ffra

ctio

n [5

]. Th

is v

alue

w

as u

sed

to c

alcu

late

a d

ensi

ty o

f cov

erag

e in

% (d

eter

min

ed a

rea

per m

olec

ule/

theo

retic

al a

rea

per m

olec

ule)

. Am

ide

II ba

nds w

ere

corr

ecte

d fo

r the

co

ntrib

utio

n of

νas

CO

O- fr

om A

sp. P

rote

in c

once

ntra

tions

as d

eter

min

ed w

ith th

e m

etho

d de

scrib

ed in

[100

] and

indi

cate

d by

M-B

rad.

type

of l

ayer

and

type

of M

i b-C

K

date

of

prep

arat

ion

mg/

ml P

rote

in

Nac

t A p

p

in c

m-1

A vp

in c

m-1

R

Γ Γ Γ Γ

in m

ol/c

m²

a mol

in A

ng

% o

f

cove

rage

DPP

A/C

L(E.

chol

i) 8/

28/9

6 0.

50

16.1

7 2.

7547

11.

6683

81.

651.

05E-

12

126

55

ch M

i b-C

K 8/

29/9

6 no

min

al

3.

0489

91.

9106

41.

601.

18E-

12

119

61

DPP

A/C

L(E.

chol

i) 9/

27/9

6 0.

50

16.1

7 2.

8016

51.

6451

71.

701.

06E-

12

125

55

ch M

i bCK

9/28

/96

Lsg.

von

089

6

2.77

026

1.62

474

1.71

1.05

E-12

12

6 55

DPP

A/C

L(E.

chol

i) 12

/5/9

6 0.

50

16.1

7 2.

4279

61.

3093

81.

858.

89E-

13

137

46

ch M

i bCK

12/6

/96

Lsg.

von

089

6

2.36

081.

2939

81.

828.

75E-

13

138

46

DPP

A/C

L(E.

chol

i) 4/

10/9

7 0.

55

36.7

1 2.

3277

091.

3859

891.

681.

15E-

12

120

60

17

0

ch M

i bCK

4/11

/97

nom

inal

6.93

783

3.84

082

1.81

1.13

E-12

12

1 59

4/

13/9

7

6.

5576

53.

8773

81.

691.

10E-

12

123

57

4/

14/9

7

6.

0908

93.

4290

91.

781.

00E-

12

129

52

DPP

A/C

L(E.

chol

i) 2/

5/98

0.

22

18.3

6 0.

7951

570.

4746

851.

687.

87E-

13

145

41

ch M

i bCK

2/6/

98 M

-Bra

d

0.68

1289

0.40

0129

1.70

6.70

E-13

15

7 35

2/

7/98

0.

4052

960.

2858

261.

424.

27E-

13

197

22

1.72

0.08

DPP

A-Bi

laye

r 9/

4/98

0.

32

24.4

7 1.

0446

320.

6467

941.

627.

86E-

13

145

41

ch M

i bCK

9/6/

98

0.83

0918

0.53

1811

1.56

6.33

E-13

16

2 33

DPP

A-Bi

laye

r 11

/17/

98

0.63

24

.47

1.12

5778

0.83

2273

1.57

9.00

E-13

13

6 47

ch M

i bCK

11/2

0/98

nom

inal

1.13

9733

0.75

3137

1.59

8.79

E-13

13

7 46

11

/22/

98 =

M-B

rad

1.

0898

750.

6952

361.

658.

29E-

13

142

43

DPP

A-Bi

laye

r 5/

6/99

0.

55

24.4

7 1.

1293

830.

6420

661.

768.

24E-

13

142

43

ch M

i bCK

5/7/

99 M

-Bra

d

2.67

409

1.56

716

1.71

6.68

E-13

15

8 35

5/

8/99

nom

inal

0.6

3

0.74

1671

0.41

6759

1.78

5.39

E-13

17

6 28

17

1

1.64

0.07

DPP

A-Bi

laye

r 8/

13/9

9

19.2

3 0.

9094

50.

5919

791.

548.

87E-

13

137

46

hu M

i bCK

8/13

/99

0.55

2.63

462

1.62

969

1.62

7.71

E-13

14

7 40

8/

14/9

9 no

min

al

2.

5936

31.

6116

11.

617.

60E-

13

148

40

8/

16/9

9 0.

30

0.

7731

140.

4780

151.

627.

40E-

13

150

39

M

-Bra

d

DPP

A/C

L(bh

) 2/

19/0

0 0.

05

32.3

1 5.

0932

92.

9606

41.

729.

61E-

13

131

50

ch M

i bCK

2/20

/00

M-B

rad

5.

2754

93.

1245

11.

691.

00E-

12

129

52

DPP

A/C

L(bh

) 2/

28/0

0 0.

02

32.3

1 1.

5550

60.

9641

991.

618.

86E-

13

137

46

hu M

i bCK

2/29

/00

M-B

rad

4.

7283

62.

8375

21.

678.

13E-

13

143

42

Ge

23/1

2/19

98 (S

) 0.

46 (

Was

te, M

-

Brad

)

24.4

7 0.

9192

320.

5588

791.

646.

87E-

13

155

36

ch M

i bCK

23/1

2/19

98 (R

) 0.

63 (n

omin

al)

0.

0589

10.

0346

681.

704.

35E-

14

618

2

8 References

172

8 References

173

8 References

[1] Finkelstein A. Weak-acid uncouplers of oxidative phosphorylation. Mechanism of action

on thin lipid membranes, Biochim. Biophys. Acta 205, 1970: 1-6

[2] Escher BI, Snozzi M, Schwarzenbach RP. Uptake, Speciation, and Uncoupling of

Substituted Phenols in Energy Transducing Membranes, Environ. Sci. Techn. 30 (10),

1996: 3071-3079

[3] Wenzl P, Fringeli M, Goette J, Fringeli UP. Supported Phospholipid Bilayers Prepared

by the „LB/Vesicle Method“: A Fourier Transform Infrared Attanuated Total Reflection

Spectroscopic Study on Structure and Stability, Langmuir 10, 1994: 4253-4264

[4] Wallimannn T, Wyss M, Brdizka D, Nicolay K, Eppenberger H, Intracellular

compartimentation, structure and function of creatine kinase isoenzymes in tissues with

high and fluctuating energy demands: the ‘phosphocreatine circuit’ for cellular energy

homeostasis, Biochem. J 281, 1992: 21-40

[5] Fritz-Wolf K, Schnyder T, Wallimann T, Kabsch W, Structure of mitochondrial creatine

kinase, Nature 381, 1996: 341-345

[6] Müller M, Moser R, Cheneval D, Carafoli E, Cardiolipin is the Membrane Receptor for

Mitochondrial Creatine Phosphokinase, J. Biol. Chem. 260 (6), 1985: 3839-3843

[7] Rojo M, et al. Interaction of mitochondrial creatine kinase with model membranes, FEBS

Letters 281, 1991: 123-129

[8] Schnyder T, Cyrklaff M, Fuchs K, Wallimann T, Crystallization of Mitochondrial

Creatine Kinase on negatively charged Lipid Layers, J. Struc. Biol. 112, 1994: 136-147

8 References

174

[9] Fringeli UP, Günthard HH. Infrared Membrane Spectroscopy. In: Grell E., ed. Membrane

Spectroscopy, Springer, Berlin, 1981: 270-332

[10] Tamm LK, Tatulian SA. Infrared Spectroscopy of Proteins and Peptides in Lipid

Bilayers, Quarterly Reviews of Biophysics 30 (Nov 97), 1997: 365-429

[11] Fringeli UP. In Situ Infrared Attenuated total Reflection Membrane Spectroscopy. In:

Mirabella FM, ed. Internal Reflection Spectroscopy - Theory and Application, Marcel

Dekker, New York, 1992: 255-324

[12] Kuhn H., Functionalized monolayer assembly manipulation, Thin solid Films 99, 1983:

1-16

[13] Ohlsson PA, et al. Liposome and proteoliposome fusion onto solid substartes, studied

using atomic force microscopy, quartz crystal microbalance and surface plasmon

resonance: Biological activities of incorporated components, Bioelectrochem.

Bioenergetics 38, 1995: 137-148

[14] Fringeli UP, Apell HJ, Fringeli M, Läuger P. Poalrized infrared absorption of Na+/K+-

ATPase studied by attenuated total reflection spectroscopy, Biochim. Biophys. Acta 984,

1989: 301-312

[15] Puu G, Gustafson I. Planar lipid bilayers on solid supports from liposomes - factors of

importance for kinetics and stability, Biochim. Biophys. Acta 1327, 1997: 149-161

[16] Günzler H, Heise MH. IR-Spektroskopie (3.Auflage), VCH, Weinheim, 1996

[17] Colthup NB, Daly LH, Wiberley SE. Introduction to Infrared and Raman Spectroscopy,

3.Ed. Academic Press, London, 1990

[18] Griffiths PR. Chemical Infrared Fourier Transform Spectroscopy, Chemical Analysis

43, J. Wiley&Sons, New York, 1975

8 References

175

[19] Glasser L. Fourier Transform for Chemists, Topics in Chemical Instrumentation, J.

Chem. Edu. 64 (10,11,12), 1987: part I A228-A233, part II A260-266, part III A306-313

[20] Perkins WD. Fourier Transform-Infrared Spectroscopy, Topics in Chemical

Instrumentation, J. Chem. Edu. 63 (1), 1986 and 64 (11,12), 1987: part I A5-A10, part II

A269-A271, part II A 269-A 305

[21] Gremlich HU, Yan B. Infrared and Raman Spectroscopy of Biological Materials. In:

Pratical Spectroscopy Series 24, Marcel Dekker, New York, 2000

[22] Zbinden R, Infrared Spectroscopy of High Polymers, Academic Press, New York, 1964

[23] Goormatigh E, Cabiaux V, Ruysschaert J-M. Determination of Soluble and Membrane

Protein Structure by Fourier Transform Infrared Spectroscopy. In: Hilderson HJ, Ralson

GB, eds. Subcellular Biochemistry 23 „Physicochemical Methods in Study of

Biomembranes“, Plenum Press, New York, 1994, 329-362

[24] Byler MD, Susi H. Examination of Secondary Structure of Proteins by Deconvoluted

FTIR Spectra, Biopolymers 25, 1986: 469-487




Biomembranes“, Plenum Press, New York,1994: 405-450




Biomembranes“ ,Plenum Press, New York, 1994, 363-403

8 References

176

[27] Bauer, HH, Müller M, Goette J, Merkle HP, Fringeli UP, Interfacial Adsorption and

Aggregation Associated Changes in Secondary Structure of Human Calcitonin Monitored

by ATR-FTIR spectroscopy, Biochemistry 33, 1994: 12276-12282

[28] Porth I, Molekulare Strukturänderungen von Ribonuklease A bei der Entfaltung und

Rückfaltung untersucht mit Modulations-FTIR-ATR-Spektroskopie, Diplomarbeit,

Universität Wien, 1999

[29] Müller M, Buchet R, Fringeli UP, 2D-FTIR ATR Spectroscopy of Thermo-Induced

Periodic Secondary Structural Changes of Poly-(L)-lysine: A Cross-Correlation Analysis

of Phase-Resolved Temperature Modulation Spectra, J. Phys. Chem. 100, 1996: 10810-

10825.

[30] Sélényi P. Sur léxistence et observation des ondes luminescenses sphériquees

inhomogènes, Compt. Rend. 157, 1913: 1408

[31] Wood RW. Physical Optics, 3rd ed. Macmillan, New York, 1934

[32] Goos F, Hänchen H. Ein neuer und fundamentaler Versuch zur Totalreflexion, Ann.

Phys. 1, 1947: 333

[33] Harrick JM. Internal Reflection Spectroscopy, Harrick Sc. Co.,Ossining, 1975

[34] Fringeli UP. ATR and Reflectance IR Spectroscopy: Applications. In: Lindon JC,

Tranter GE, Holmes JL, eds. Encyclopedia of Spectroscopy and Spectrometry, Academic

Press, San Diego, 2000: 58-75

[35] Harrick JN. Electric Field Strengths at Totally Reflecting Interfaces, J. Opt. Soc. Am.

55, 1965: 851

[36] Picard F, et al. Quantitative Orientation Measurements in Thin Films by Attanuated

Total Reflection Infrared Spectroscopy, Biophysical Journal 76, 1999: 539-551

[37] Handbook of Chemistry and Physics, 55th Ed. CRC Press, 1974

8 References

177

[38] Fringeli UP, Goette J, Reiter G, Siam M, Baurecht D, Structural Investigation of

Oriented Membrane Assemblies by FTIR-ATR Spectroscopy. In: Haseth JA (ed.),

Fourier Transform Spectroscopy: 11th International Conference, 1998: 729-747

[39] Alberts B, et al. Molekularbiologie der Zelle (2. Auflage, übersetzt von Lothar Jaenicke

et al.), VCH, Weinheim, 1990: 403-431

[40] Stryer L. Biochemistry (3. Edition), Freeman & Co., New York, 1988: 397-424

[41] Hostetler KY. Polyglycerolphospholipids: phosphatidylglycerol, diphosphatidylglycerol

and bis(monoacyl-glycero)phosphate. In: Hawthorne JN, Ansell GB. New Comprehensive

Biochemistry 4, Phospholipids, Elsevier, Amsterdam, 1982: 240-241

[42] Voelker DR. Lipid assembly into cell membranes. In: Vance DE, Vance JE, eds. New

Comprehensive Biochemistry 20, Biochemistry of Lipids, Lipoproteins and Membranes,

Elsevier, Amsterdam, 1991: 493

[43] Cullis PR, Hope MJ, Physical properties and functional roles of lipids in membranes, In:

Vance DE, Vance J (eds.), Biochemistry of Lipids, Lipoproteins and Membranes,

Elsevier, Amsterdam, 1991, 1-41

[44] Escher BI, Schwarzenbach RP. Partitioning of Substituted Phenols in Liposome-Water,

Biomembrane-Water and Octanol-Water Systems, Environ. Sci. Techn. 30 (1), 1996:

260-270

[45] Mitchell P. Keilin´s respiratory chain concept and its chemiosmotic consequences,

Science 206, 1979: 1148-1159

[46] Gennis RB. Biomembranes-Molecular Structure and Function, Springer, New York,

1989: 235-269

[47] Ackermann T, Physikalische Biochemie, Springer, Berlin, 1992: 150-194

8 References

178

[48] Neumcke B, Läuger P. Nonlinear Electrical Fields in Lipid Bilayer Membranes,

Biophys. J. 9, 1969: 1160-1170

[49] Volkov AG, Paula S, Deamer DW. Two mechanism of permeation of small neutral

molecules and hydrated ions across phospholipid bilayers, Bioelectrochem. Bioenerg. 42,

1997: 153-160

[50] Flewelling RF, Hubbell WL. The Membrane Dipole Potential in a Total Membrane

Potential Model, Application to Hydrophobic Ion Interactions with Membranes, Biophys.

J. 49, 1986: 541-552

[51] Gawrisch K, et al., Membrane dipole potentials, hydration forces, and the ordering of

water at membrane surfaces, Biophys. J. 61, 1992: 1213-1223

[52] Karlson P, Detlef Doenecke D, Koolman J. Kurzes Lehrbuch der Biochemie (14.

Auflage), Thieme Verlag, Stuttgart, 1994

[53] Kenyon GL, Reed GH, Creatine Kinase: Structure-Activity Relationship, Advances in

enzymology and related areas of molecular biology 54, 1983: 367-426

[54] Cronin MTD, Zhao YH, Yu RL. pH-dependence and QSAR analysis of the toxicity of

phenols and anilines to Daphnia magna, Environ. Toxicol. 15 (2), 2000: 140-148

[55] Schultz TW. Structure-Toxicity Relationship for Benzenes Evaluated with Tetrahymena

pyriformis, Chem. Res. Toxicol. 12,1999: 1262-1267

[56] Warne MA, Osborn D, Lindon JC, Nicholson JK. Quantitative structure-toxicity

relationships for halogenated substituted-benzenes to Vibrio fischeri, using atom-based

semi-empirical molecular-orbital descriptors, Chemosphere 38 (14), 1999: 3357-3382

[57] Argese E, Bettiol C, Giurin G, Miana P. Quantitative structure-activity relationships for

the toxicity of chlorophenols to mammalian submitochondrial particles, Chemosphere 38

(10), 1999: 2281-2292

8 References

179

[58] Bearden AP, Schultz TW, Comparison of Tetrahymena and Pimephales toxicity based

on mechanism of action, SAR QSAR Environ. Res. 9 (3-4), 1998: 127-153

[59] McLaughlin S, J.P. Dilger, Transport of protons across membranes by weak acids, Ann.

Rev.of Biophys. Biophys.Chem. 19, 1989: 113-136

[60] Terada H, Uncouplers of Oxidative-Phosphorylation, Environ. Health Perspect. 87,

1990: 213-218

[61] Schöpflin M, Fringeli UP, Perlia X. A Study on the Interaction of Local Anesthetics

with Phospholipid Model Membranes by Infrared ATR Spectroscopy, J. Am. Chem. Soc.

109, 1987: 2375-2380

[62] Shibata A, Ikawa K, Terada H. Site of Action of the Local Anesthetic Tetracaine in a

Phosphatidylcholine Bilayer with Incorporated Cardiolipin, Biophys. J 69, 1995: 470-477

[63] Escher BI, Hunziker R, Schwarzenbach RP. Kinetic Model To Describe the Intrinsic

Uncoupling Activity of Substituted Phenols in Energy Transducing Membranes, Environ.

Sci. Techn. 33 (4), 1999: 560-570

[64] Frisch MJ, Trucks GW, Schlegel HB, Scuseria CE, Robb MA, Cheeseman JR,

Zakrzewski JA, Montgomery JA Jr., Stratmann RE, Burant JC, Dapprich S, Millam JM,

Daniels AD, Kudin KN, Strain MC, Farkas O, Tomasi J, Barone V, Cossi M, Cammi R,

Mennucci B, Pomelli C, Adamo C, Clifford S, Ochterski J, Petersson CA, Ayala PY, Cui

Q, Morokuma K, Malick DK, Rabuck AD, Raghavachari K, Foresman JB, Cioslowski J,

Ortiz JV, Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Gomperts R, Martin

RL, Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Gonzalez C,

Challacombe M, Gill PM, Johnson B, Chen W, Wong MW, Andres JL, Gonzalez C,

Head-Gordon M, Replogle ES, Pople JA, Gaussian 98, Revision A.6; Gaussian Inc. :

Pittsburgh, PA, 1998

8 References

180

[65] Martin JML, van Alsenoy C, University of Antwerpen, 1995

[66] Demel RA, Yin CC, Lin BZ, Hauser H., Monolayer characteristics and thermal behavior

of phosphatidic acid, Chem. Phys. Lipids 60, 1992: 209

[67] Evans RW, Willams MA, Tinoco J. Surface areas of 1-palmitoyl phosphatidylcholines

and their interaction with cholesterol, Biochem. J. 245, 1987: 455-462

[68] Cameron DG, Casal HL, Mantsch HH. Characterization of the Pretransition in 1,2-

Dipalmitoyl-sn-glycero-3-phosphocholine by Fourier Transform Infrared Spectroscopy,

Biochemistry 19, 1980: 3665-3672

[69] Mantsch HH, Martin A, Cameron DG. Characterization by Infrared Spectroscopy of the

Bilayer to Nonbilayer Phase Transition, Biochemistry 20, 1981: 3138-3145

[70] Pimentel GC, McClellan AL. The Hydrogen bond, Freeman & Co., San Francisco,

1960, 138;

[71] Nwobi O, Higgins J, Zhou X, Liu R. Density functional calculation of phenoxyl radical

and phenolat anion: an examination of DFT methods, Chem. Physics Letter 272, 1997:

155-161

[72] Nonella M, Suter HU, Formation of Anion-Counterion Complexes Can Explain

Vibrational Properties of the Phenolate Anion in Solution, J. Phys. Chem. A 103, 1999:

7867-7871

[73] Schütz M, Bürgi T, Fischer T, Leutwyler S. Intermolecular bonding and vibrations of

phenol H2O (D2O), J. Chem. Phys. 98 (5), 1993: 3763-3776

[74] Schreiber V, et al. Temperature effect on proton-transfer equilibrium and IR spectra of

chlorophenol-tributylamine systems, Faraday Trans. 92 (14), 1996: 2555-2561

[75] Bessman S, Geiger PJ, Transport of Energy in muscle: The Phosphorycreatine Shuttle,

Science 211, 1981. 448-452

8 References

181

[76] Kenyon GL, Creatine kinase shapes up, Nature 381, 1996: 281-282

[77] Schlattner U, Forstner M, Eder M, Stachowiak O, Fritz-Wolf K, Wallimann T,

Functional aspects of the X-ray structure of mitochondrial creatine kinase: A molecular

physiology approach, Mol. Cell. Biochem. 184, 1998: 125-140

[78] Stachowiak O, Dolder M, Wallimann T, Membrane-Binding and Lipid-Vesicle Cross-

Linking Kinetics of the Mitochondrial Creatine Kinase Octamer, Biochemistry 35 (48),

1996. 15522-15528

[79] Vacheron M-J, Clottes E, Chautard C, Vial C, Mitochondrial Creatine Kinase

Interaction with Phospholipid Vesicles, Arch. Biochem. Biophys. 344 (2), 1997: 316-324

[80] Rojo M, Hovius R, Demel R, Nicolay K, Wallimann T, Mitochondrial Creatine Kinase

Mediates Contact Formation between Mitochondrial Membranes, J. Biol. Chem., 1991:

20290-20295

[81] Schnyder T, Cyrklaff M, Fuchs K, Wallimann T, Crystallization of Mitochondrial

Creatine Kinase on negatively charged Lipid Layers, J. Struc. Biol. 112, 1994: 136-147

[82] Schlattner U, Eder M, Dolder M, Khuchua ZA, Strauss AW, Wallimann T, Divergent

Enzyme Kinetics and Structural Properties of the Two Human Mitochondrial Creatine

Kinase Isoenzymes, Biol. Chem. 381, 2000: 1063-1070

[83] Cheneval D, Carafoli E, Powell GL, Marsh D, A spin label electron-spin resonance

study of the binding of mitochondrial creatine kinase to cardiolipin, Eur. J. Biochem.

186, 1989: 415-419

[84] Fringeli UP, et al., ATR Spectroscopy of Thin Films, In: Nalwa HS (ed.), Thin Films

Handbook, Academic Press, San Diego, 2001

8 References

182

[85] Forstner M, Kriechbaum M, Laggner P, Wallimann T, Structural Changes of Creatine

Kinase upon Substrate Binding, Biophys. J. 75,1998:1016-1023

[86] Liu Z-J, Zhou J-M, Spin-labelling probe on conformational change at the active sites of

creatine kinase during denaturation by guanidine hydrochloride, Biochim. Biophys. Acta

1253, 1995: 63-68

[87] Raimbault C, Buchet R, Vial C, Changes of creatine kinase secondary structure induced

by the release of nucleotides from caged compounds. An infrared difference-spectroscopy

study, Eur. J. Biochem. 240,1996: 134-142

[88] Zhoufa G, Somasundaram T, Blanc E, Parthsarathy G, Ellington WR, Chapman MS,

Transition state structure of arginine kinase: implications for catalysis of bimolecular

reactions, Proc. Natl. Acad. Sci. USA 95, 1998: 8449-8454

[89] Zhou G, Ellington R, Chapman MS, Induced Fit in Arginin Kinase, Biophys. J. 78,

2000: 1541-1550

[90] Travers F, Barman TE, Cryoenzymic Studies on the Transition-State Analog Complex

Creatine Kinase ADPMg Nitrate Creatine, Eur. J. Biochem. 110, 1980: 405-412

[91] Buechter DD, Medzihradszky KF, Burlingame AL, Kenyon GL, The Active Site of

Creatine Kinase. Affinity Labaling of Cysteine 282 with N-(2,3-epoxypropyl)-N-

amidinoglycine, J. Biol. Chem. 267 (4), 1992: 2173-2178

[92] Eder M, Stolz M, Wallimann T, Schlattner U, A Conserved Negatively Charged Cluster

in the Active Site of Creatine Kinase Is Critical for Enzymatic Activity, J. Biol. Chem.

275 (35), 2000: 27094-27099

[93] Edmiston PL, Schavolt KL, Kersteen EA, Moore NR, Charles L. Borders Jr., Creatine

kinase: a role for arginine-29 in creatine binding and active site organization, Biochim

Biophys. Acta 1546, 2001: 291-298

8 References

183

[94] Gross M, Furter-Graves EM, Wallimann T, Eppenberger HM, Furter R, The tryptophan

residues of mitochondrial creatine kinase: Roles of Trp-223, Trp-206, and Trp-264 in

active-site and quaternary structure formation, Protein Sci. 3, 1994: 1058-1068

[95] Clottes E, Vial C, Discrimination between the Four Tryptophan Residues of MM-

Creatine Kinase on the Basis of the Effect of N-Bromosuccinimde on Activity and

Spectral Properties, Arch. Biochem. Biophys. 329 (1), 1996: 97-103

[96] Hagemann H, Marcillat O, Buchet R, Vial C, Mg-ADP binding site in wild-type

creatine kinase and in mutants. The role of aromatic residues probed by Raman and

Infrared Spectroscopy, Biochemistry 36, 2000: 9251-9256

[97] Granjon T, Vacheron M-J, Vial C, Buchet R, Structural Changes of Mitochondrial

Creatine Kinase upon Binding of ADP, ATP, or Pi, Observed by Reaction-Induced

Infrared Difference Spectra, Biochemistry 40, 2001: 2988-2994

[98] Leydier C, Clottes E, Couthon F, Marcillat O, Vial C, Involvement of a tyrosine reidue

in the ADP binding site of creatine kinase. A second derivative UV-Vis spectroscopy

study, Biochem. Molec. Biol. Int. 41, 1997: 777-784

[99] Furter R, Kaldis P, Furter-Graves EM, Schnyder T, Eppenberger HM, Wallimann T,

Expression of active octameric chicken cardiac mitochondrial creatine kinase in

Escherichia Coli, Biochem. J. 288, 1992: 771-775

[100] Read SM, Northcote DH, Minimization of variation in the Response to Different

Proteins of the Coomassie Blue G Dye-Binding Assay for Protein, Analyt. Biochem. 116

1981: 53-64

[101] Eisenthal R, Danson MJ, Enzyme Assays. A practical Approach, IRL Press, Oxford

(1995)

8 References

184

[102] Gerhardt W, Creatine Kinase, In: Methods of Enzymological Analysis, Vol III,

Bergmeyer HU (ed.), VCH, Weinheim, 1989 (3.ed): 508-539

[103] Hoch FL, Cardiolipins and biomembrane function, Biochim. Biophys. Acta 1113,

1992: 71-133 (Table III)

[104] Fringeli UP, Distribution and Diffusion of Alamethicin in a Lecithin/Water Model

Membrane System, J. Membrane Biol. 54, 1980: 203-212

[105] Glasoe PK, Long FA, Use of glass electrodes to measure acidities in deuterium oxide,

J. Phys. Chem. 64, 1960: 188-190

[106] Ioannou PV, Cardiolipins: Their Chemistry and Biochemistry, Prog. Lipid Res. 17,

1979: 279-318

[107] Seddon JM, R. D. Kaye RD, Marsh D, Induction of the Lammellar-Inverted Hexagonal

Phase Transition in Cardiolipin by Protons and Monovalent Cations, Biochim. Biophys.

Acta. 734, 1983: 347-352

[108] de Kruijff B, et al., Further Aspects of Ca2+-Dependent Polymorphism of Bovine Heart

Cardiolipin, Biochim. Biophys. Acta 693, 1982: 1-12

[109] Wyss M, James P, Schlegel J, Wallimann T, Limited Proteolyses of Creatine Kinase.

Implication for Three-Dimensional Structure and for Conformational Substates,

Biochemistry 32,1993:10727-10735

[110] Raimbault C, Couthon F, Vial C, Buchet R, Effects of pH and KCl on the formation of

creatine kinase from rabbit muscle. Infrared, circular dichroic and fluorescence studies,

Eur. J. Biochem 234, 1995: 570-578

[111] Rao MJK, Bujacz G, Wlodawer A, Crystal structure of rabbit muscle creatine kinase,

FEBS Lett. 439, 1998:133-137

8 References

185

[112] Schwarzott M, Untersuchung von Strukturänderungen bei Poly-L-Lysin und L-Lysin

an Lösungen unterschiedlicher pD-Werte mittels FTIR-Transmissions-Spektroskopie,

Diplomarbeit, Universität Wien, 1998

[113] Wyss M, James P, Schlege J, Wallimann T, Biochemistry 32, 1993: 10727-10735

Structural Investigations of Lipid Model Biomembranes by ... · von Phosphatgruppen zwischen...

Documents

Transcript of Structural Investigations of Lipid Model Biomembranes by ... · von Phosphatgruppen zwischen...