Research Collection

Doctoral Thesis

Structural and functional analysis of the neuronal mitochondrial fission factor GDAP1 and its homolog GDAP1L1

Author(s): Wagner, Konstanze Marion

Publication Date: 2009

Permanent Link: https://doi.org/10.3929/ethz-a-005978239

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

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

ETH Library DISS. ETH Nr. 18423

Structural and functional analysis of the neuronal mitochondrial fission factor GDAP1 and its homolog GDAP1L1

A dissertation submitted to ETH ZÜRICH

for the degree of Doctor of Science

presented by Konstanze Marion Wagner

Dipl. Biochemie, Friedrich-Schiller-Universität Jena, Germany

Born 14.10.1980 German

accepted on the recommendation of

Prof. Dr. Ueli Suter, examiner Dr. Imre Berger, co-examiner Dr. Uwe Konietzko, co-examiner Dr. Axel Niemann, co-examiner

June 2009

TABLE OF CONTENTS

Table of Contents

List of Figures ...... iv

List of Tables...... vi

Abbreviations ...... vii

Summary ...... 1

Zusammenfassung ...... 3

1.
Introduction ...... 5

1.1
The vertebrate nervous system ...... 5
1.2
Charcot-Marie-Tooth disease (CMT)...... 7
1.3
Mitochondrial dynamics: fusion and fission...... 8
1.4
GDAP1...... 11
1.4.1
GDAP1 contains structural features of glutathione S-transferases...... 14
1.4.1.1
The family of glutathione S-transferase enzymes...... 14
1.4.1.2
Cytosolic glutathione S-transferase enzymes ...... 17
1.4.1.3
Theta class glutathione S-transferase enzymes...... 17
1.4.1.4
Glutathione S-transferases and mitochondria...... 18
1.4.2
GDAP1 contains features of tail-anchored proteins of the mitochondrial outer membrane...... 19 1.5
GDAP1L1 – the GDAP1 paralog ...... 21
1.6
Aim of the study...... 24

2.
Materials and Methods ...... 25

2.1
Frequently used solutions and Equipment...... 25
2.1.1
Solutions ...... 25
2.1.2
Equipment...... 26 2.2
Molecular biology methods ...... 27
2.2.1
cDNA amplification...... 27
2.2.2
Restriction digest ...... 27
2.2.3
DNA gel electrophoresis...... 28
2.2.4
Ligation...... 28
2.2.5
Transformation of E.coli...... 28
2.2.6
DNA purification ...... 28
2.2.7
Cloning of constructs ...... 29
2.2.7.1
Cloning of GDAP1 disease- and point mutations ...... 29
2.2.7.2
Cloning of GDAP1 deletion mutants ...... 30
2.2.7.3
Cloning of hGDAP1 ...... 31
2.2.7.4
Cloning of FLAG-tagged GDAP1 and hGDAP1...... 31
2.2.7.5
Cloning of constructs used to analyze the C-terminal region of GDAP1 ...... 31
2.2.7.6
Cloning of chimera GDAP1-Vamp1B and GDAP1-Omb5 ...... 31

i TABLE OF CONTENTS

2.2.7.7
Cloning of GDAP1L1 constructs ...... 32
2.2.7.8
Cloning of GDAP1 and GDAP1L1 constructs for insect cell expression...... 32
2.2.8
Sequencing analysis...... 33
2.3
Cell culture...... 34
2.3.1
Cell culture solutions and materials...... 34
2.3.2
Cell lines ...... 34
2.3.3
Primary cell cultures ...... 35
2.3.4
Transient transfection...... 36
2.4
Methods to analyze topology and localization of GDAP1/GDAP1L1 constructs...... 36
2.4.1
Fluorescence protease protection assay ...... 36
2.4.2
Subcellular localization and protease digest...... 36
2.4.3
In vitro transcription and translation...... 36
2.4.4
MOM integration assay and immuno-precipitation...... 37
2.4.5
Analysis of GDAP1L1 localization: screen of stress-inducing agents ...... 37
2.5
Protein isolation from mouse/ rat tissue...... 38
2.6
Biochemical methods...... 39
2.6.1
Glutathione S-transferase activity assays...... 39
2.6.2
Glutathione peroxidase activity assays ...... 40
2.6.3
Detection of ROS...... 40
2.6.4
Measurement of protein carbonyls...... 40
2.6.5
Measurement of the mitochondrial membrane potential ...... 40
2.7
Immunochemical methods...... 41
2.7.1
Immunoblot analysis...... 41
2.7.2
Immunocytochemistry ...... 42
2.7.3
Immunohistochemistry ...... 43
2.8
Bioinformatic analysis...... 44

3.
Results...... 45

3.1
Analysis of recessive and dominant GDAP1 disease mutations...... 45
3.1.1
In contrast to the majority of recessive disease-causing mutations dominant mutations induce mitochondrial fragmentation...... 45
3.1.2
GDAP1 and disease mutant-induced mitochondrial fission is blocked by co- expression of dominant-negative Drp1...... 45
3.1.3
The expression of dmGDAP1’s impairs mitochondrial fusion...... 48
3.1.4. Expression of dmGDAP1’s leads to damaged mitochondria...... 50
3.2
GDAP1 structural analysis...... 52
3.2.1
Analysis of the overall GDAP1 domain structure ...... 52
3.2.2
Detailed analysis of the GDAP1 tail-anchor...... 52
3.2.2.1
GDAP1 is a tail-anchored protein with a single transmembrane domain ...... 54
3.2.2.2
The classical tail-anchored proteins GDAP1 and Fis1 but not Mfn2 integrate into isolated mitochondria-enriched membrane preparations...... 56
3.2.2.3
Positively charged amino acids surrounding the TMD determine GDAP1 mitochondrial targeting and function...... 58
3.2.2.4
The correctly ordered amino acid sequence of HD1, but not of the TMD, is critical for mitochondrial fission activity...... 60
3.2.2.5
The specific nature of the GDAP1 TMD and its C-terminus is not required for fission activity...... 61
3.2.2.6
TMD hydrophobicity does not influence GDAP1-induced fission of mitochondria ...... 62
3.2.3
Detailed analysis of the GST features of GDAP1 ...... 63
3.2.3.1
Analysis of the putative catalytic centre of GDAP1...... 63
3.2.3.2
GDAP1 is an active glutathione S- transferase...... 65
3.2.4
Three-dimensional structure of GDAP1 ...... 67

ii TABLE OF CONTENTS

3.3
GDAP1L1...... 69
3.3.1
GDAP1L1 is exclusively expressed in the central nervous system...... 69
3.3.2
GDAP1L1 is a cytosolic protein...... 70
3.3.3
The putative TMD of GDAP1L1 is able to integrate into the MOM ...... 72
3.3.4
The TA of GDAP1 targets GDAP1L1 to mitochondria ...... 75
3.3.5
The HD1 of GDAP1L1 cannot substitute for the HD1 of GDAP1 ...... 76
3.3.6
GDAP1L1 is translocated to mitochondria upon treatment with menadione...... 77
3.3.7
Detailed analysis of the effect of MQ on GDAP1L1 subcellular distribution...... 79
3.3.8
Analysis of the naphthoquinone pathway and EA on GDAP1L1 subcellular distribution ...... 80
3.3.9
MQ induced redistribution of GDAP1L1 is blocked by NAC ...... 81

4.
Discussion ...... 84

4.1
Dominant and recessive GDAP1 disease mutations differ in their mode of action.....84
4.2
GDAP1 is a tail-anchored protein...... 87
4.3
GDAP1 is an active glutathione S-transferase...... 89
4.4
GDAP1L1 subcellular distribution depends on oxidative stress ...... 94
4.4.1
GDAP1L1 is predominantly expressed in the CNS...... 94
4.4.2
GDAP1L1 contains features of a tail-anchored protein...... 96
4.4.3
Quinone-induced ROS stress directs GDAP1L1 to mitochondria...... 97

5.
References...... 101

6.
Appendix...... 111

Acknowledgement...... 114

Curriculum Vitae...... 116

iii LIST OF FIGURES

List of Figures

Figure 1.1 Schematic representation of PNS and CNS myelination 5 Figure 1.2 Schematic overview highlighting proteins that are mutated in CMT 6 Figure 1.3 Mitochondria are dynamic organelles 7 Figure 1.4 Unique challenges for mitochondria in neurons 8 Figure 1.5 GDAP1 disease mutations 10 Figure 1.6 GDAP1 promotes mitochondrial fission 12 Figure 1.7 Homology modeled structures of the amino terminal thioredoxin-like domains of GDAP1 and GDAP1L1 17 Figure 1.8 TA protein membrane insertion occurs post release from the ribosome 18 Figure 1.9 Features that determine the targeting of TA proteins to the MOM or ER 19 Figure 1.10 Structural comparison of GDAP1 and GDAP1L1 22

Figure 3.1 Recessive and dominant inherited GDAP1 disease mutations differ in their ability to induce mitochondrial fission 44 Figure 3.2 Fission induced by GDAP1 and mutations is blocked by Drp1 mut 45 Figure 3.3 Fission induced by dominantly inherited mutated forms of GDAP1 cannot be counterbalanced by mitofusion 1 or 2 overexpression 48 Figure 3.4 Dominantly inherited mutated forms of GDAP1 impair the mitochondrial transmembrane potential 49 Figure 3.5 GDAP1 deletion mutants display impaired fission activity 52 Figure 3.6 GDAP1-EGFP is peripherally attached to the MOM 53 Figure 3.7 GDAP1 membrane topology 54 Figure 3.8 Integration assay of in vitro synthesized TA-proteins 56 Figure 3.9 TA-associated critical basic aa for GDAP1 localization and function 57 Figure 3.10 Role of TMD and HD1 aa sequences in GDAP1-mediated fission 59 Figure 3.11 Effect of TMD length on mitochondrial targeting and function 60 Figure 3.12 Function of TA in GDAP1 fission activity 61 Figure 3.13 Influence of TMD hydrophobicity on GDAP1 fission activity 62 Figure 3.14 GST-N related GDAP1 point mutations display a loss of fission 63 Figure 3.15 GDAP1 expression constructs for recombinant expression 65 Figure 3.16 Recombinant expression constructs of GDAP1 disease mutations 66 Figure 3.17 Recombinant GDAP1 expression constructs for X-ray analysis 67 Figure 3.18 GDAP1L1 is expressed in the central nervous system 68 Figure 3.19 GDAP1L1 is a cytosolic protein 70 Figure 3.20 Integration assay of in vitro-synthesized GDAP1L1 constructs 72 Figure 3.21 The TA of GDAP1 localizes GDAP1L1 to mitochondria 73

iv LIST OF FIGURES

Figure 3.22 The HD1 of GDAP1l1 cannot substitute for the HD1 of GDAP1 74 Figure 3.23 Redistribution of GDAP1L1 to mitochondria upon MQ treatment 76 Figure 3.24 Effect of MQ on GDAP1L1 subcellular distribution 77 Figure 3.25 Analysis of the effect of quinones and EA on GDAP1L1 distribution 78 Figure 3.26 NAC rescue of MQ induced ROS 80

Figure 4.1 Model of possible GDAP1 membrane topologies 86 Figure 4.2 Hypothetic model of GDAP1L1 translocation due to MQ treatment 96

v LIST OF TABLES

List of Tables

Table 1.1 Substrate specificities of human glutathione S-transferases 13

Table 2.1 Frequently used buffers and solutions 23 Table 2.2 Primers used to generate disease and point mutations 28 Table 2.3 Primers used to generate GDAP1 deletion mutants 28 Table 2.4 Primer used to generate GDAP1L1 constructs 30 Table 2.5 Primer used for recombinant expression of GDAP1/GDAP1L1 30 Table 2.6 Primer used for sequencing reactions 31 Table 2.7 Sequencing PCR 32 Table 2.8 Solutions and materials used in cell culture 32 Table 2.9 Cell lines used for this study 33 Table 2.10 Working dilutions of used stress inducing agents 36 Table 2.11 Assay conditions for spectrophotometric GST activity assays 37 Table 2.12 Primary antibodies for western blot analysis 39 Table 2.13 Secondary antibodies for western blot analysis 40 Table 2.14 List of primary antibodies used for immunofluorescence 41 Table 2.15 List of secondary antibodies used for immunofluorescence 41 Table 2.16 URL/programs used in this study 42

Table 3.1 Characteristics of TMDs of selected MOM TA proteins and mutants 63 Table 3.2 35S-labeled GSH binding assay 65 Table 3.3 GST activities of GDAP1 288X and GDAP1 318X 66

Table 6.1 Primer used to generate constructs used in this study 98

vi ABBREVIATIONS

Abbreviations

m Mitochondrial transmembrane potential AA Amino acid ADOA Autosomal dominant optic atrophy Ala Alanine AP Alkaline phosphatase CDNB 1-Chloro-2,4-dinitrobenzene CMT Charcot-Marie-Tooth disease CNS Central nervous system dmGDAP1 Dominant GDAP1 disease mutations DNA Deoxyribonucleic Acid DNMQ 2,3-dimethyoxy-1,4-naphthoquinone DTT Dithiothreitol Drp1 Dynamin-related protein 1 Drp1 mut Dominant negative Drp1 EA Ethacrynic acid EGFP Enhanced green fluorescent protein EPNP 1,2-Epoxy-3-(4-nitrophenoxy) propane ER Endoplasmic reticulum FCS Fetal calf serum Fig. Figure GDAP1 Ganglioside-induced differentiation-associated protein 1 GDAP1L1 Ganglioside-induced differentiation-associated protein 1 Like 1 GSH Glutathione GST Glutathione S-transferase HD1 Hydrophobic domain 1 HD2 Hydrophobic domain 2 HQ Hydroquinone HRP Horseradish peroxidase IMS Intermembrane space kDa Kilodalton MFN1 Mitofusin 1 MFN2 Mitofusion 2 Min Minutes ml Mililiter mM Milimolar μl Microliter

vii ABBREVIATIONS

μM Micromolar MOM Mitochondrial outer membrane MQ Menadione NAC N-acetyl-L-cysteine nm nanometer PBS Phosphate buffered saline PDI Proteindisulfide isomerase pNBC p-Nitrobenzylchloride pNPA p-Nitrophenylacetate PNS Peripheral nervous system PCR Polymerase chain reaction PFA Paraformaldehyd PMSF Phenylmethansulfonylfluoride rmGDAP1 Recessive GDAP1 disease mutations ROS Reactive oxygen species Rf Relative fluorescence Rpm rounds per minute Sec Seconds Ser Serine SDS Sodium dodecyl sulphate SQ Semiquninone TA Tail-anchor TBHQ tert-buthylhydroquinone TMD Transmembrane domain Tris Tris-(hydroxymethyl)-aminomethan Tyr Tyrosine v/v Volume per volume w/v Weight per volume

viii SUMMARY

Summary

With a prevalence of 1 in 2,500 people affected, Charcot-Marie-Tooth disease (CMT) is the most frequent inherited peripheral neuropathy in humans. It comprises a large group of genetically heterogeneous hereditary motor and sensory neuropathies affecting the peripheral nervous system (PNS) without severe central nervous system (CNS) phenotype. Mutations in the ganglioside-induced differentiation-associated protein 1 (GDAP1) are associated with various types of CMT. Today, 33 CMT-associated GDAP1 alterations, including point mutations and changes leading to truncated proteins have been described, causing axonal (no demyelination, axonal loss), demyelinating (demyelination, onion bulb formation) and intermediate (characteristics of axonal and deymelinating phenotypes) forms of CMT. Both, PNS neurons and Schwann cells express GDAP1 suggesting that both cell types may contribute to the motor and sensory neuropathy phenotype. Recent studies could show that GDAP1 is involved in mitochondrial dynamics and induces mitochondrial fission. Although originally identified as a recessively inherited disease three mutations are inherited in a dominant mode. In this study I show that the similar clinical manifestations of recessive as well as dominant GDAP1 disease mutations are caused by different hindrance of normal mitochondrial dynamics. Based on its domain structure GDAP1 was proposed to belong to a novel GDAP1 class of glutathione S-transferase (GST) enzymes, although recent studies could not detect any glutathione (GSH)-binding or GST activity of bacterial-expressed recombinant protein. In this study I demonstrate that insect cell-expressed recombinant GDAP1 288X is a catalytically active GST enzyme. It displays high GSH-conjugating activity against various substrates typically found for theta class GST enzymes, a subgroup of the cytosolic GST family. Furthermore, GDAP1 contains two hydrophobic domains at the C-terminal region that display features of a tail-anchor (TA) protein of the mitochondrial outer membrane (MOM). Here I define GDAP1 as a TA protein of the MOM. I could demonstrate that GDAP1 carries a single transmembrane domain (TMD) that is, together with the adjacent basic residues, critical for MOM targeting. The flanking N-terminal region containing the other hydrophobic domain is located on the cytosolic leaflet of the MOM. TMD sequence, length, and high hydrophobicity do not influence GDAP1 fission function if MOM targeting is maintained. The basic amino acids bordering the TMD in the cytoplasm, however, are required for both targeting of GDAP1 as part of the TA and GDAP1-

1 SUMMARY mediated fission. Thus, this GDAP1 region contains critical overlapping motifs defining intracellular targeting by the TA concomitant with functional aspects. Expression of GDAP1 is found in tissues of the PNS and CNS. However, dysfunctional GDAP1 leads to peripheral neuropathies without severe CNS phenotype. This implies that either GDAP1 function is compensated by another protein expressed in the CNS or that GDAP1 exerts different functions in the CNS compared to the PNS. GDAP1-like 1 (GDAP1L1) is a closely related paralog of the mitochondrial fission factor GDAP1 and belongs to the same novel GDAP1 class of GST enzymes. This study further aims at determining the potential of GDAP1L1 to compensate for dysfunctional GDAP1 proteins in the CNS. I show that unlike GDAP1, GDAP1L1 is expressed in the CNS but not in the PNS. Although I confirmed the potential of the putative TA domain of GDAP1L1 to integrate into the MOM, under normal cellular conditions GDAP1L1 is not localized to mitochondria. However, I could show that GDAP1L1 is translocated to mitochondria upon treatment with the stressor menadione. The exact functional mechanisms, the cellular relevance and its implications on neuronal cells and GDAP1 function needs further investigation.

Taken together, I present in this study new insights into the cellular functions of GDAP1 and the structural features crucial for GDAP1 activity in neuronal cells. I further provide evidence that the GDAP1 paralog GDAP1L1 may be able to compensate for dysfunctional GDAP1 in the CNS.

2 ZUSAMMENFASSUNG

Zusammenfassung

Charcot-Marie-Tooth (CMT) Krankheiten gehören mit einer Prävalenz von 1:2500 im Menschen zu der am häufigsten auftretenden Gruppe von heterogen vererbbaren motorischen und sensorischen Neuropathien, die spezifisch das periphere Nervensystem (PNS) schädigen. Mutationen in GDAP1 (ganglioside-induced differentiation-associated protein 1) führen entweder zu einer demyelinisierenden (Demyelinisierung, “Zwiebelschalenbildung”), einer axonalen (keine Demyelinisierung, Verlust von Axonen) oder einer intermediären (Eigenschaften vom axonalen und demyelinisierenden Phänotyp) Form von CMT. Bisher wurden 33 Mutationen in CMT Patienten beschrieben, die zu Punktmutationen oder verkürzten Proteinvarianten führen. Sowohl PNS Neuronen wie auch Schwann Zellen exprimieren GDAP1. Dies lässt vermuten, dass beide Zelltypen zu dem Krankheitsbild im PNS beitragen. Aktuelle Studien zeigen, dass GDAP1 in mitochondriale Dynamik-Prozesse involviert ist und mitochondriale Fragmentierung induziert. Ursprünglich wurde angenommen, dass Mutationen in GDAP1 zu rezessiv vererbter CMT führen, jedoch wurden später auch drei dominant vererbte Krankheitsmutationen in GDAP1 identifiziert. In der vorliegenden Studie konnte ich zeigen, dass die vergleichbaren klinischen Symptome der rezessiv- und der dominant vererbten GDAP1 Mutationen durch unterschiedliche Veränderung normaler mitochondrialer Dynamik-Prozesse hervorgerufen werden. Aufgrund der theoretisch vorausgesagten Domänenstruktur von GDAP1 wurde dieses Protein einer neuen Familie von Glutathion S-Transferase (GST) Enzymen zugeordnet. Jedoch konnten verschiedene Studien keine Glutathion (GSH)-Bindung oder GST- Aktivität von bakteriell hergestellten rekombinanten Protein nachweisen. In dieser Arbeit zeige ich, dass in Insektenzellen recombinant hergestelltes GDAP1 288X ein aktives GST- Enzym ist. Die Substratspezifität von GDAP1 ist typisch für die Theta-Klasse der GST’s. Im Gegensatz zu den meisten bekannten GST’s besitzt GDAP1 zwei hydrophobe Domänen in der C-terminalen Region. Diese Region hat strukturelle Ähnlichkeiten zu membranverankerten Proteinen der äusseren mitochondrialen Membran (MOM). In dieser Studie definiere ich GDAP1 als ein membranverankertes Protein der MOM, dass eine einzige Transmembrandomäne (TMD) besitzt. Diese TMD ist zusammen mit angrenzenden basischen Aminosäuren essentiell für den Transport zur MOM. Die flankierende N-terminale Region inklusive der anderen hydrophoben Domäne und der GST-Domänen befindet sich im Zytoplasma. Sowohl die Aminosäuresequenz der TMD, als auch die Länge und Hydrophobizität der TMD beeinflussen die mitochondriale Fragmentierungsaktivität von GDAP1 nicht, solange das Protein weiterhin in die MOM integriert ist. Die basischen Aminosäuren auf der zytoplasmatischen Seite der TMD

3 ZUSAMMENFASSUNG werden hingegen für den korrekten Transport an die MOM und für die GDAP1-induzierte mitochondriale Fragmentierung benötigt. Somit beinhaltet die membranverankerte Region von GDAP1 Aminosäuremotive, welche essentiell für den Transport und die Funktion des Proteins sind. Obwohl GDAP1 in Zellen des PNS und des ZNS (zentrales Nervensystem) exprimiert ist, führen Krankheitsmutationen von GDAP1 fast ausschliesslich zu einer Schädigung des PNS. Folglich wird entweder die Funktion von GDAP1 durch ein anderes Protein im ZNS kompensiert oder GDAP1 hat eine andere Funktion im ZNS als im PNS. GDAP1-like 1 (GDAP1L1) weisst eine hohe Identität zu GDAP1 auf und wird der gleichen, neuen GDAP1-Klasse von GST-Enzymen zugeordnet. Das Ziel dieser Studie ist es weiterhin zu zeigen, ob GDAP1L1 für dysfunktionales GDAP1 im ZNS kompensieren kann. Meine Resultate belegen, dass im Gegensatz zu GDAP1 GDAP1L1 exklusiv im ZNS, und nicht im PNS, exprimiert ist. Obwohl die potenzielle membranverankerte Domäne von GDAP1L1 in die MOM integrieren kann, ist GDAP1L1 unter normalen zellulären Bedingungen zytoplasmatisch. Wird die Zelle jedoch mit dem Stressreagenz Menadione behandelt, ändert sich die zelluläre Verteilung von GDAP1L1. Das Protein kolokalisiert mit den Mitochondrien. Die exakten Mechanismen von GDAP1L1, die zelluläre Relevanz und die Auswirkungen auf neuronale Zellen erfordern weitere Untersuchungen.

In dieser Studie präsentiere ich neue Einblicke in die zellulären Funktionen von GDAP1 und analysiere strukturelle Besonderheiten, welche essentiell für die Aktivität von GDAP1 sind. Desweiteren präsentiere ich erste Experimente, welche die Hypothese unterstützen, dass GDAP1L1 die Funktion von dysfunktionalem GDAP1 im ZNS kompensieren kann.

4 INTRODUCTION

1. Introduction

1.1 The vertebrate nervous system

The nervous system is a highly complex communication network that deals with information about the vertebrate’s surrounding and internal state. Throughout evolution, vertebrates developed the most complex nervous system crucial for sensory perception, motor activity, and higher cognitive functions such as learning, memory and social behavior. The nervous system is subdivided into the central nervous system (CNS), consisting of brain and spinal cord that process information, and the peripheral nervous system (PNS), formed by peripheral nerves that distribute information. The CNS is connected with an organism’s surroundings and its internal state via sensory fibers originating from the PNS. The major cellular components of the nervous tissue are neurons and glial cells. Neurons are the hallmark of communicating cells. They conduct electrical signals and are further involved into the processing and transmission of electrically and chemically encoded information. Glial cells are much more abundant than neurons. There are three types of supporting cells in the CNS: oligodendrocytes, astrocytes and microglia. In the PNS, Schwann cells constitute the major neuroglial component. Glial cells provide support and nutrition to neurons, maintain their homeostasis, and participate in signal transmission. Furthermore, in adult organisms the predominant function of glial cells is the formation of myelin sheaths that enwrap axons. Myelination of the CNS is carried out by oligodendrocytes, while the PNS is myelinated by Schwann cells (Fig. 1.1). Both cell types form myelin sheaths by producing myelin, a lipid-rich membrane that repeatedly wraps around a stretch of axon (Barres and Raff, 1994). Oligodendrocytes extend multiple processes, each of which contacts one axon. In contrast, the myelinating cells of the PNS, Schwann cells, establish a one-to-one relationship with a single axon (Jessen and Mirsky, 2005). Myelination allows the fast saltatory conduction of electrical signals towards their target cell. This mode of transmission is possible due to the unique myelin structure found around axons. Myelin internodes, discontinuos insulation units along individual axons, are separated by myelin- free segments –the nodes of Ranvier- where the axolemma is exposed to the extracellular milieu. Electrical currents exclusively flow at the nodes of Ranvier. Thus, electrical impulses, called action potentials, jump from node to node, achieving a conduction velocity of up to 100 m/sec. Conduction velocities of unmyelinated axons of the same diameter are up to 10 times slower (Squire, 2002).

In the PNS, axons and Schwann cells are anatomically and functionally closely connected and tightly regulate each other. While early in development Schwann cell precursers are

5 INTRODUCTION dependent on axonal signals for survival, later, myelinating Schwann cells affect axonal properties (Jessen, 2004). Furthermore, the axon cytoskeleton, organelle content, and rates of axonal transport are tightly regulated by myelinating Schwann cells (Arroyo and Scherer, 2000; Edgar and Garbern, 2004). Consequently, the functional integrity of Schwann cells and the accompanying axon is tightly regulated, and a breakdown of this elaborate system usually leads to peripheral neuropathies. Hereditary neuropathies are genetically heterogeneous and affect neurons and/or Schwann cells. The motor and sensory neuropathy Charcot-Marie-Tooth disease (CMT) is the most common form of inherited neuropathy (Parman et al., 2004).

Figure 1.1. Schematic representation of PNS and CNS myelination. Myelination of the CNS is carried out by oligodendrocytes while the PNS is myelinated by Schwann cells. In the PNS, Schwann cells myelinate one segment of the axons (green), whereas in the CNS (blue), the oligodendrocyte forms multiple myelin internodes (T. Turnherr, ETH Diss, 2007).

6 INTRODUCTION

1.2 Charcot-Marie-Tooth disease (CMT)

In the CMT group, mutations in several different genes cause similar disease phenotypes. Conversely, different mutations of the same gene can result in different disease phenotypes (Berger et al., 2006; Niemann et al., 2006). The clinical features of this disorder were first described by Jean-Martin Charcot, Pierre Marie and Howard Henry Tooth in 1886. Patients usually present in the first or second decade of life with distal muscle athrophy in the legs, areflexia, foot deformity, and steppage gait. With progression of the disease, hands and arms are also involved in most cases. Symptoms and progression of the disease can vary. Based on electrophysiological studies and clinical/histopathological data of nerve biopsies, CMT has been subcategorized into two distinct neuropathies: First, a demyelinating form (CMT1, CMT3, CMT4) characterized by a reduced nerve conduction velocity (NCV) and segmental de- and remyelination events usually characterized by onion bulb formations. Secondary effects also involve axonal loss and muscle athrophy due to the tight interaction between Schwann cells and axons (Suter and Scherer, 2003). Second, an axonal form (CMT2) that is associated with normal or almost normal NCVs and a loss of myelinated axons leading to a reduction of the compound muscle action potential amplitude (CMAP; Zuchner and Vance, 2006). However, the categorization into dymelinating and axonal forms is not absolute; intermediate forms displaying features of both types are also known (Suter and Scherer, 2003). One group of proteins for which mutant forms lead to CMT comprises factors involved in the regulation of mitochondrial dynamics processes.

Figure 1.2. Schematic overview highlighting proteins that are mutated in CMT. The localization of the normal proteins is depicted. Proteins have been assigened to Schwann cells and/or neurons, respectively, when expression and the observed form of CMT overlap (Niemann et al., 2006).

7 INTRODUCTION

1.3 Mitochondrial dynamics: fusion and fission

Various neuropathies caused by mutations in mitochondrial dynamics factors reinforce the notion that neurons are particularly prone to defects in mitochondrial dynamics, a process described in the following section. Mitochondria are dynamic organelles (Fig. 1.3; Chan, 2006b; Herzig and Martinou, 2008; Hoppins et al., 2007; Kiefel et al., 2006; McBride et al., 2006; Yaffe, 1999). A continuous process of fusion and fission events is crucial to maintain the mitochondrial morphology at a given moment, the mitochondrial functionality, and to allow transport within a cell. Mitochondrial fusion is a coordinated process that allows the fusion of the outer and the inner mitochondrial membrane (Chan, 2006b; de Brito and Scorrano, 2008; Hoppins et al., 2007; Knott et al., 2008). This leads to the exchange of lipids, proteins and mitochondrial DNA (mtDNA). Thereby mitochondrial fusion lowers the risk that mutations in the mtDNA accumulate within a single organelle. Such an accumulation of mutated mtDNA and dysfunctional gene products could in turn interfere with mitochondrial functionality and lead to more damage (Chan, 2006a). To avoid the accumulation of damaged mitochondrial material, fission continuously generates smaller mitochondrial units. If the separated mitochondrial unit is unable to fuse back, it can be degraded by the process of autophagy (Twig et al., 2008b). Thus, continuous fusion and fission events maintain a homogenous mitochondrial population within a cell and allow quality control (Dimmer and Scorrano, 2006; Kim et al., 2007; Twig et al., 2008b).

Mitochondrial fragmentation becomes most prominent during apoptosis (Dimmer and Scorrano, 2006; Herzig and Martinou, 2008; Suen et al., 2008). The mitochondrial fragmentation occurs during or prior to the release of pro-apoptotic factors such as cytochrome C. Although mitochondrial fission and the release of pro-apoptotic factors from mitochondria can be separated (Cassidy-Stone et al., 2008; Karbowski et al., 2006), a

Figure 1.3. Mitochondria are dynamic organelles constantly undergoing fusion and fission. Depicted are selected factors that are known to influence the mitochondrial morphology. Grey, major mammalin mitochondrial dynamic regulators (conserved from yeast to mammalian); *, yeast proteins; (), yeast homologs of the corresponding mammalian factors.

8 INTRODUCTION tight association between the mitochondrial morphology and the apoptotic progression has been demonstrated in various studies (Herzig and Martinou, 2008; Suen et al., 2008). Increased mitochondrial fusion blocks or delays the intrinsic apoptotic pathway. Loss of fission resulting in elongated mitochondria, as the fusion process is still ongoing, is also protective against the induction of apoptosis. Cells with fragmented mitochondria, caused by increased fission or decreased fusion, are more susceptible to apoptotic stimuli (Herzig and Martinou, 2008; Suen et al., 2008).

However, mitochondrial fission is not always destructive or deleterious. Fission is needed to allow mitochondrial inheritance into the daughter cells (Yaffe, 1999). In addition, fission influences mitochondrial transport and distribution in neuronal cells (Frank, 2006; Niemann et al., 2006). The overexpression of the fission factor Drp1 (Dynamin-related protein1, or Dlp1; Dnm1 in yeast) conducts more mitochondria into dendrites, and increases synapse formation in cultured hippocampal neurons (Li et al., 2004). Similar effects of Bcl-xL are Drp1-dependent (Li et al., 2008). This suggests that mitochondrial fission supports neural differentiation and maintenance (Baloh, 2008; DiMauro and Schon, 2008; Frank, 2006; Niemann et al., 2006).

Detmer and Chan speculated that due to the extreme dimensions of neurons, especially long peripheral nerves, they are especially vulnerable to alterations in the tight regulation of mitochondrial fusion and fission (Fig. 1.4; Detmer and Chan, 2007). This dependence of neurons propably stems from their high energy demands and the special importance of proper mitochondrial distribution: mitochondria are concentrated in several neuronal regions, including pre- and postsynaptic sites (Detmer and Chan, 2007). Several proteins involved in the regulation of mitochondrial dynamics have been found to cause peripheral neuropathies when the cognate genes are mutated. Mutations in the mitochondrial fusion factors MFN2 (mitofusin 2) and OPA1 (optical atrophy 1) lead to the neurodegenerative disorders Charcot- Figure 1.4. Unique challenges for mitochondria in neurons. Shown is a short neuron with mitochondria Marie-Tooth disease (CMT) 2A and to autosomal (brown ovals). Mitochondria travel long distances from the cell body to dendritic and axonal processes dominant optical atrophy (ADOA), respectively and are involved in neuronal activity and calcium buffering, respecitvely (Chan et al., 2006). (Baloh, 2008; Chan, 2006a; Frank, 2006; McBride

9 INTRODUCTION et al., 2006; Niemann et al., 2006). Over 40 mutations in MFN2 have been associated with CMT2A, an autosomal dominant disorder that manifests in axonal loss that is in some cases associated with optical degenerations (Zuchner et al., 2004). Analysis of CMT2A alleles in mice provided functional insights into this dominantly inherited axonopathy. Many CMT2A alleles of MFN2 are non-functional in terms of fusion when expressed alone (Detmer and Chan, 2007). However, this loss of fusion can be efficiently compensated by wildtype Mfn1 (but not Mfn2) due to the ability of Mfn1 and Mfn2 to form hetero-oligomeric complexes. This implies that the phenotypic severity of the disease depends predominantly on the levels of Mfn1 in the patients. Interestingly, even in peripheral nerves mitochondrial fusion defects are only partial – only the longest nerves show structural aberrations in their mitochondrial outer and inner membranes, along with swelling that is suggestive of mitochondrial dysfunction (Detmer and Chan, 2007). Insights into Mfn2-regulated processes that eventually lead to neurodegeneration were provided by a study, in which mice lacking Mfn2 showed a highly specific degeneration of Purkinje cell neurons in the cerebellum resulting in cerebellar ataxia (Detmer and Chan, 2007). Mfn2- deficient Purkinje cells failed to support dentritic outgrowth, particularly of dentritic spines, which represent the sites of synaptic connections. Interestingly, Li et al. (2004) could show that overexpression of the mitochondrial fission factor Drp1 in rat hippocampal neurons promotes synaptogenesis. In contrast, the fusion-promoting dominant negative Drp1 or OPA1 reduce the amount of mitochondria found in the close synaptic vicinity, which leads to a decreased densitiy of spines and synapses. Hence, the proper intracellular distribution of mitochondria is critical for an intact physiology of neuronal cells. Mfn2 has additionally been associated with vascular proliferative diseases, diabetes mellitus, and obesity, supporting the finding that Mfn2 is implicated in signaling pathways largely independent of its role in mitochondrial fusion (de Brito and Scorrano, 2008). In addition to the loss of peripheral nerve function, a subset of patients with CMT2A have optic atrophy, suggesting that MFN2 and OPA1 mutations can lead to overlapping clinical outcomes. Heterozygous mutations in OPA1 lead to the most common heritable form of optic atrophy (Delettre et al., 2000). This disorder is characterized by the degeneration of retinal ganglion cells, the axon of which form the optic nerve. OPA1 mutations have been associated with reduced ATP production and reduced mtDNA content. Fibroblasts lacking OPA1 display fragmented mitochondria, defects in respiration, aberant christae structure, and increased susceptibility to apoptosis (Abdelmohsen et al., 2004; Olichon et al., 2002). Furthermore, heterozygous mouse models show a progressive decline in retinal ganglion cell number and aberrations of axons in the optic nerve (Davies et al., 2007). Despite this, the exact pathogenic mechanism of OPA1 mutations still remains unclear.

10 INTRODUCTION

Disease-causing mutations in mitochondrial fission factors have been reported for Drp1 and GDAP1. A single case study reported a mutation in human Drp1 (Waterham et al., 2007). The patient, who died a month after birth, exhibited elongated mitochondria that were aggregated around the nucleus. The symptoms resembled those seen in ADOA and CMT2A, but the Drp1-related disorder is characterized by a much earlier disease onset and a more severe disease phenotype. Previous reports linked Drp1 and hFis1 to mitochondrial as well as peroxisomal fission (Schrader, 2006). It was speculated that this is the reason for the multi-syndrome phenotype, but this awaits further clarification (Waterham et al., 2007). Mutations in GDAP1 are known to lead to axonal, intermediate and demyelinating forms of CMT. As the following study focused on GDAP1 and the disease-causing mechanisms of GDAP1 mutations, this protein will be discussed in detail in the next section (1.4).

1.4 GDAP1

Mutations in the GDAP1 gene (Ganglioside-induced differentiation associated protein 1) lead to early onset CMT with demyelinating, intermediate or axonal forms (Baxter et al., 2002; Cuesta et al., 2002; Nelis et al., 2002). Autosomal recessive GDAP1 mutations are associated with a severe phenotype that is characterized by a rapidly progressing weakness that usually begins before the age of 3 years and eventually leads to an inability to walk by the second or third decade of life (Bernard et al., 2006). The demyelinating GDAP1-caused CMT phenotype was associated with reduced motor and sensory NCV, and patients exhibited severe symmetric weakness and atrophy of feet and hands. The axonal CMT phenotype caused by different mutations in GDAP1 was characterized by hoarse voice, vocal cord paresis and reduced CMAP but showed no impairment of the NCV and no

Figure 1.5. GDAP1 disease mutations. CMT-causing GDAP1 disease mutations lead to different disease phenotypes in patients: Intermediate phenotype, red; demyelinating phenotype, blue; axonal phenotype, black; dashed area, enlarged interdomain with two additional alpha helices not found in classical GST enzymes.

11 INTRODUCTION demyelination events were found. Recent studies found that the axonal phenotype is often additionally accompanied by diaphragm dysfunction (Sevilla et al., 2008; Stojkovic et al., 2004). First mutations found in the chromosomal region 8q13-21.1 encoding GDAP1 were found predominantly in the mediterranean region (Azzedine et al., 2003; Baxter et al., 2002; Birouk et al., 2003; Cuesta et al., 2002; Senderek et al., 2003). Later on, case studies reported also from patients with GDAP1 mutations in other european countries (Ammar et al., 2003; Baxter et al., 2002; Stojkovic et al., 2004) and from all over the world (Chung et al., 2008; Claramunt et al., 2005). Up to now, 33 mutations in GDAP1 were found; 17 point mutations (Fig. 1.5), five nonsense mutations resulting in a stop codon, nine frameshift mutations that result in a premature stop codon, and two mutations in the splice sites of intron 3 and 4 (Ammar et al., 2003; Auer-Grumbach et al., 2008; Azzedine et al., 2003; Barankova et al., 2007; Baxter et al., 2002; Biancheri et al., 2006; Birouk et al., 2003; Boerkoel et al., 2003; Bouhouche et al., 2007; Chung et al., 2008; Claramunt et al., 2005; Cuesta et al., 2002; De Sandre-Giovannoli et al., 2003; Di Maria et al., 2004; Georgiou et al., 2006; Kabzinska et al., 2006; Kabzinska et al., 2005; Kabzinska et al., 2007; Parman et al., 2004; Rougeot et al., 2008; Senderek et al., 2003; Sevilla et al., 2003; Sevilla et al., 2008; Stojkovic et al., 2004; Xin et al., 2008). Due its demyelinating, axonal and intermediate phenotypes, the GDAP1 gene does not fit the traditional CMT classification (either demyelinating or axonal; 1.2) that is based on clinical features of patients. Although CMT caused by mutations in GDAP1 was initially considered to be an autosomal recessive disorder, three mutations were identified that are expressed in patients in a heterozygous mode and segregate dominantly (R120W, T157P, Q218E; Bernard et al., 2006; Claramunt et al., 2005; Cuesta et al., 2002). The so far identified dominant GDAP1 mutations show a later onset of disease and have a much milder disease phenotype compared to the recessively inherited alleles. Dominantly inherited mutated forms have been reported to induce mitochondrial aggregation when overexpressed as GFP-fusion protein (Pedrola et al., 2008). However, Brambillasca et al. (2007) showed that large tags such as GFP-tags interfere with membrane integration of C-terminal tail-anchored proteins, a class of proteins to which GDAP1 shows high structural similarities (1.4.2).

The GDAP1 gene consists of six exons and spans 13.9 kb of DNA. GDAP1 was identified as a transcript that was up-regulated after ganglioside-induced cholinergic differentiation of the mouse neuroblastoma cell line Neuro2a (Liu et al., 1999). Consistently, an increase in GDAP1 expression was also found in P19 cells, an embryonal carcinoma cell line which can differentiate to a neural cell lineage in an appopriate environment. Highest expression levels of GDAP1 in vivo were found in mouse brain and spinal cord (Liu et al., 1999). RT-

12 INTRODUCTION

PCR analysis by Cuesta et al. (2002) and Western blot analysis of the protein levels by Niemann et al. (2005) confirmed the nervous system-enriched expression levels. GDAP1 is expressed most prominently in different regions of the CNS, including the olfactory bulb, thalamus, cortex, cerebellum and spinal cord, but no obvious CNS phenotype was found in patients (1.5). Consistent with the consequences of GDAP1 mutations leading to motor and sensory neuropathies, we also found GDAP1 expression in sciatic nerve and dorsal root ganglia (Niemann et al., 2005). Neurons and Schwann cells express GDAP1 (Niemann et al., 2005) suggesting that both cell types contribute to the disease. However, no obvious correlation between the position and nature of the mutations and the electrophysiologic or neuropathologic phenotype or the cellular origin could be detected. Co-localization experiments with organelle marker proteins and cell fractionation approaches defined GDAP1 as a mitochondrial protein. Proteinase K digestion assays with crude mitochondrial extract of transfected COS-7 cells showed that the N-terminal part of the protein containing the GST-N and GST-C domains is oriented towards the cytosol (Niemann et al., 2005). Subsequent carbonate and high salt treatment demonstrated that GDAP1 is an integral membrane protein of the MOM (Niemann et al., 2005). Studies performed by Niemann et al. (2005) showed that GDAP1 is a regulator of the mitochondrial network. It promotes mitochondrial fragmentation and is involved in mitochondrial fission processes (Fig. 1.6) without inducing apoptosis, affecting overall mitochondrial activity, or interfering with mitochondrial fusion (Niemann et al., 2005). The mitochondrial fusion proteins mitofusion 1 and 2 and a dominant negative version of Drp1 can counterbalance GDAP1-dependent fission. GDAP1-specific knockdown by RNA interference results in a tubular mitochondrial morphology. Truncated forms of GDAP1 that are found in patients who have CMT have lost their mitochondrial targeting signal. Furthermore, disease-associated GDAP1 point mutations

Figure 1.6. GDAP1 promotes mitochondrial fission. Untransfected and control transfected COS-7 cells display a mixed and tubular mitochondrial morphology (a,b), while GDAP1 shifts the mitochondrial architecture to fragmented mitochondria (c,d). Quantification of the mitochondrial morphology by classifying five distinct mitochondrial morphologies (e) (Niemann et al., 2005).

13 INTRODUCTION have been shown to display a reduced mitochondrial fragmentation activity (Niemann et al., 2005). These data indicate that both features of GDAP1 – proper localization and the regulation of mitochondrial dynamics – are likely to be crucial for the functional integrity of myelinated peripheral nerves. On a structural level GDAP1 (1) contains GST-N and GST-C domains and thus fullfills the prerequisites for a putative activity as a GST (glutathione S-transferase)-enzyme and (2) contains a C-terminal membrane anchor that could act as a classical tail-anchor for integration into the mitochondrial outer membrane. As the following study was performed with specific focus on these structural features I will next introduce the field of GST-enzymes and tail-anchored proteins.

1.4.1 GDAP1 contains structural features of glutathione S-transferases GDAP1 is a 358 amino acid long protein with a molecular weight of 41 kDa. Computer analysis identified two hydrophobic domains at the C-terminus. With a N-terminal thioredoxin fold domain (GST-N) and a C-terminal substrate-binding domain (GST-C) the protein encoded by GDAP1 shows high homologies to GST’s (Marco et al., 2004; Shield et al., 2006). Despite the sequence similarity to GST proteins, evolutionary studies revealed certain features that are absent in canoical GST’s and proposed that GDAP1, GDAP1-like1 (GDAP1L1, a GDAP1 paralog) and a homologous protein from Drosophila melanogaster belong to a novel, still undescribed class of GST’s. In addition to a distinct primary sequence GDAP1 is characterized by an enlarged interdomain between the GST domains absent in all other GST’s (two predicted, additional
- helices between
4 and
5), and it is anchored into the mitochondrial outer membrane (MOM) by a C-terminal transmembrane domain, whereas most GST’s are cytosolic enzymes and lack these domains (Marco et al., 2004; Shield et al., 2006). In silico analysis of GDAP1 performed by Shield et al. (2006) suggest that the proposed GDAP1 GST family constitutes a subgroup of the cytosolic GST class with highest similarity to the Zeta, Omega and Theta GST classes (Marco et al., 2004). Furthermore, Shield et al. claim from cross-linking experiments of recombinant bacterial expressed GDAP1
334-358 that GDAP1 has a dimeric structure comparable to other cytosolic GST’s, the ability to dimerize is a prerequisite for enzymatic activity of GST proteins. The capability to form dimers was confirmed by co-immunoprecipitation studies of M. Rüegg (Rüegg, 2009). However, so far, no evidence of GST activity or glutathione-binding ability was found (Pedrola et al., 2005; Shield et al., 2006).

1.4.1.1 The family of glutathione S-transferase enzymes As mentioned in 1.4.1, GDAP1 contains structural features of GST enzymes. GST’s constitute an ancient protein superfamily with multiple roles in eukaryotic organisms that

14 INTRODUCTION are defined by all members of a class having high sequence similarity and common structural and functional features. GSTs are ubiquitous enzymes that occur in high concentrations in most human cells. In plants and mammals, GSTs are the principal phase II enzymes in metabolic detoxification processes (Sheehan et al., 2001). Phases I and II involve the conversion of a lipophilic, non-polar xenobiotic into a more water-soluble and therefore less toxic metabolite, which can then be eliminated from the cell (phase III). Phase I is catalysed mainly by the cytochrome P450 system. Phase II enzymes catalyze the conjugation of activated xenobiotics to an endogenous water-soluble substrate, such as glutathione (GSH), UDP-glucuronic acid, or glycine. Quantitatively, GST-mediated conjugation of GSH is the major phase II reaction (Sheehan et al., 2001). Originally, GST proteins were discovered as proteins that display enzymatic activity with the substrate CDNB (1-chloro-2,4-dinitrobenzene). CDNB is commonly used as prototypic substrate for GST’s. However, several classes of GST enzymes exhibit low (or no) activity with CDNB (Board et al., 2000; Tan et al., 1996). In fact, many GST’s display overlapping substrate selectivities and substrate specificity/ activity does not always correlate with sequence similarities or GST class (Table 1.1).

The main chemistry of GSTs is to catalyze a nucleophilic attack by the reduced tripeptide GSH (-Glu-Cys-Gly) on non-polar compounds that contain an electrophilic centre (a carbon, nitrogen or sulphur atom). This results in the formation of a more soluble, non- toxic peptide derivative that is ready to be excreted or compartimentalized by phase III enzymes, such as the transmembrane ATPase GS-X pump (Frova, 2006). In addition to these classical functions GST’s can also exhibit peroxidase, isomerase, and thiol transferase activities (Board et al., 2000), or function in a non-catalytic mode via binding of non-substrate ligands and modulation of signaling processes (Hayes et al., 2005). Furthermore, it is well known that their role extends to non-stress metabolism, as leukotriene and prostaglandine biosynthesis (Fernandez-Canon and Penalva, 1998; Kanaoka et al., 1997) and the catabolism of aromatic amino acids (Fernandez-Canon and Penalva, 1998; Thom et al., 2001). Hence, the original view of GST’s as solely detoxification enzymes that detoxify xenobiotics such as chemical carcinogens, environemtal pollutants, and antitumor agents, or inactivate endogenous
,-unsaturated aldehydes, quinones, epoxides, and hydroperoxides formed as secondary metabolites during oxidative stress has changed (Hayes et al., 2005).

Mammalian GST enzymes (Table 1.1) are grouped into three main subfamilies according to their sequence similarity, which is over 60 % within one family, but can be below 30 % between members of different families. The sequence identity also reflects their homologies in biochemical features, substrate specificity and subcellular localization (Hayes et al., 2005). Two groups, the cytosolic, also called canonical, and the

15 INTRODUCTION mitochondrial GSTs comprise soluble enzymes that share some similarities in their three- dimensional structure but are only distantly related. The third familiy comprises microsomal GST’s, now termed MAPEG (membrane-associated proteins in eiconsanoid and glutathione metabolism (Hayes et al., 2005).

Table 1.1. Substrate preferences of human glutathione S-transferases (adopted from Hayes et al., 2005)

Family Class, enzyme Substrates or reaction**

5-ADD, BCDE, BPDE, Busulfan, Chlorambucil, DBADE, Cytosolic Alpha, A1-1 DBPDE, BPhDE, N-a-PhIP Alpha, A2-2 CuOOH, DBPDE, 7-chloro-4-nitrobenz-2-oxa-1,3-diazole Alpha, A3-3
5-ADD,
5-pregnene-3,20-dione, DBPDE Alpha, A4-4 COMC-6, EA, 4-hydroxynonenal, 4-hydroxydecenal Alpha, A5-5 not done trans-4-phenyl-3-buten-2-one, BPDE, CDE, DBADE, trans- Cytosolic Mu, M1-1 stilbene oxide, styrene-7,8-oxide COMC-6, 1,2-dichloro-4-nitrobenzene, aminochrome, dopa O- Mu, M2-2 quinone, PGH2  PGE2

Mu, 3-3 BCNU, PGH2  PGE2 Mu, 4-4 CDNB Mu, 5-5 low for CDNB

Cytosolic Sigma, S1-1 PGH2  PGE2 BCNU, butadiene epoxide, CH Cl , EPNP, ethylene oxide, EA, p- Cytosolic Theta, T1-1 2 2 nitrobenzyl chloride, p-nitrophenethyl bromide Theta, T-2-2 CuOOH, menaphthyl sulfate dichloroacetate, fluoroacetate, 2-chloropropionate, Cytosolic Zeta, Z1-1 malelyacetoacetate Cytosolic Omega, O1-1 monomethylarsonic acid, dehydroascorbic acid Omega, O2-2 dehydroascorbic acid CDNB, CuOOH, (S)-15-hydroperoxy-8,11,14-cis-6-trans- Mitochondrial Kappa, K1-1 eicosatetraenoic acid CDNB, LTA  LTC , (S)-15-hydroperoxy-8,11,14-cis-6-trans- MAPEG gp I, MGST2 4 4 eicosatetraenoic acid gp I, FLAP Nonenzymatic binding of arachidonic acid

gp I, LTC4S LTA4  LTC4 CDNB, LTA  LTC (S)-15-hydroperoxy-8,11,14-cis-6-trans- MAPEG gp II, MGST3 4 4, eicosatetraenoic acid MAPEG gp IV, MGST1 CDNB*, CuOOH, hexachlorobuta-1,3-diene

gp IV, PGES1 PGH2  PGE2 *Activity increased by treating enzyme with N-ethylmaleimide. **Abbreviations:
5-ADD,
5-androstene-3,17-dione; BCDE, benzo[g]chrysene diol epoxide; BCNU, 1,3-bis(2- chloroethyl)-1-nitrosourea; BPDE, benzo[a]pyrene diol epoxide; BPhDE, benzo[c]phenanthrene diol epoxide; CDE, chryene-1,2-diol 3,4-epoxide; COMC-6, crtotonyloxymethyl-2-cyclohexenone; DBADE, dibenz[a,h]anthracene diol epoxide; DBPDE, dibenzo[a,l]pyrene diol epoxide; EA, ethacrynic acid; EPNP, 1,2-epoxy-3-(p-nitrophenoxy)propane; N-a-PhIP, N-acetoxy-2-amino-1-methyl-6-phenylimidazol[4,5-b]pyridine.

16 INTRODUCTION

1.4.1.2 Cytosolic glutathione S-transferase enzymes The subfamily of cytosolic GST’s is by far the most abundant. For instance, in humans 15- 20 different cytosolic GST genes have been identified (Hayes et al., 2005). The family of cytosolic GST’s is again subgrouped into numerous classes, based on distinct criteria such as amino acid/ nucleotide sequence identity, physical gene structure and immunoreactive properties. Up to now, seven classes of mammalian cytosolic GST’s, namely Alpha, Mu, Pi, Sigma, Theta, Zeta and Omega are recognized (Mannervik, 1995). Despite pronounced onverall sequence divergence, all GST proteins for which structures are known show striking levels of structural conservation, displaying a common three- dimensional organization (Frova, 2006). Furthermore, all soluble GST’s are biologically active as dimers. All GST enzymes are composed of the same basic protein fold, which consists of two distinct ligand-binding domains: an N-terminal domain (GST-N domain), consisting of  strands and
helices as secondary elements that adopt a thioredoxin fold, and an all-helical C-terminal domain (Armstrong, 1997). Both domains together constitute the catalytically active site. The GST-N domain provides the among all GST’s highly conserved binding-site for GSH and contains specific residues critical for GSH binding and GST activity. In particular, a highly conserved Tyr (for Alpha/Mu/Pi) or Ser (for Theta/Zeta) is crucial for catalytic activation of GSH. Additionally, for certain GST’s the ‘lock and key’ motif constitutes an important hydrophobic interaction in the GST-N domain. Here the side chain of an aromatic amino acid (usually Tyr or Phe) protrudes from the GST-N domain of one subunit into a hydrophobic pocket formed by the GST-C domain of the other GST protein in the dimer (Armstrong, 1997; Frova, 2006). The binding-site for the hydrophobic substrate is located in the GST-C domain. This domain is less conserved compared to the GST-N domain and is primarily formed by residues with non-polar side chains (Frova, 2006).

GDAP1, despite being an integral protein of the mitochondrial outer membrane does not belong to the group of mitochondrial GST enzymes. Evolutionary studies and sequence analysis suggest that the novel GDAP1 GST family constitutes a subgroup of the cytosolic GST class and displays high similarities to the cytosolic GST enzyme subclass of theta class enzymes (Fig. 1.7; Marco et al., 2004; Shield et al., 2006).

1.4.1.3 Theta class glutathione S-transferase enzymes Theta class GST enzymes were originally discovered as a novel class of GST’s that lack activity towards CDNB, the ‘universal’ GST substrate (Meyer et al., 1991). High enzymatic activities were found towards EPNP (1,2-epoxy-3-(p-nitrophenoxy)propane), pNBC (p-nitrobenzyl chloride), p-nitrophenethyl bromide, and epoxyeicosatrienoic acids

17 INTRODUCTION

(Spearman et al., 1985). This class is further distinguishable from other classes by their inability to bind GSH or S-hexyl-GSH affinity matrices (Sheehan et al., 2001). Overall sequence homology of Theta class GST’s to Alpha/Mu/Pi classes is low. Furthermore, in contrast to the latter classes theta class Figure 1.7. Homology modelled structures of the amino terminal thioredoxin- like domains of GDAP1 and GDAP1L1. The conserved active site serine GST’s have a catalytically positioned at the amino terminal end of alpha helix 1 is shown in stick format (Shield et al. 2006). active serine rather than a tyrosine in the GST-N domain (Frova, 2006; Meyer et al., 1991; Sheehan et al., 2001). Additionally, unlike Alpha/Mu/Pi enzymes, the predominantly hydrophobic intersubunit interactions lack a ‘lock and key’ motif and the hydrophobic pocket (1.4.1.2) is less pronounced in this class (Armstrong, 1997; Sheehan et al., 2001).

1.4.1.4 Glutathione S-transferases and mitochondria As described under 1.4 the putative GST enzyme GDAP1 is localized to mitochondria. Mitochondria play a key role in energy metabolism utilizing oxygen, and in the regulation of apoptosis. Reactive oxygen species (ROS) are constantly produced under physiological conditions, and multiple enzymes such as glutathione peroxidases, glutathione reductases, glutaredoxins, thioredoxins, and GST’s exist to protect the organelle and its macromolecules against chemicals and oxidative stress. The mammalian mitochondrial class of GST’s, also named Kappa GST enzymes, are dimeric proteins that exhibit GSH-dependent conjugating and peroxidase activity with model substrates, thus sharing common catalytical features with the canonical GST’s. Mouse, rat, and humans posses only a single Kappa GST. Although no mitochondrial GST’s are found in the cytoplasm, the protein was also detected in peroxisomes. Since mitochondria and peroxisomes are involved in lipid metabolism and high amounts of reactive oxygen species are produced in the close vicinity of these organelles, it is highly suggestive that Kappa class GST’s play a crucial role in β-oxidation of fatty acids and in the detoxification of lipid peroxides (Frova, 2006; Hayes et al., 2005). Although one could speculate that GDAP1, due its mitochondrial localization, may belong to this group of

18 INTRODUCTION

GST’s, bioinformatic analysis of Marco et al. (2004) proposed a close relation to cytosolic GST enzymes and not to membrane-associated GST classes such as kappa GST’s.

Interestingly, recently cytosolic GST enzymes have been identified (i.e. GST-Pi) that contain mitochondrial targeting signals and thus are present in mitochondria as well as in the cytosol and the nucleus in mammalian cell lines. Overexpression of GST Pi diminished the cytotoxicity of rotenone or antimycin A, suggesting that it directly protects mitochondria against oxidative stress (Goto et al., 2009).

Finally, the GSH metabolism has been shown to be involved in the modulation of the mitochondrial morphology, generally in the context of studies using ethacrynic acid (EA). EA is a substrate and potent inhibitor of GST enzymes that is able to spontaneously react with GSH leading to its well known GSH depletory effects in both cytosol and mitochondria (Seyfried et al., 1999). Short-term exposure of cells to EA results in extensive mitochondrial fusion and a subsequent formation of a complete mitochondrial reticulum (Bowes and Gupta, 2005; Soltys and Gupta, 1994). It was suggested that this might be due to alterations in the mitochondrial GSH levels, changed mitochondrial transmembrane potential, apoptotic events, or to inhibition of proteins that are involved in the regulation of mitochondrial dynamics (Awasthi et al., 1993; Bowes and Gupta, 2005; Soltys and Gupta, 1994).

1.4.2 GDAP1 contains features of tail-anchored proteins of the mitochondrial outer membrane GDAP1 contains a C-terminal transmembrane domain that integrates the protein into the mitochondrial outer membrane (MOM; Niemann et al., 2005). Many proteins involved in the regulation of mitochondrial dynamics are similarly located at the MOM and contain a C-terminal membrane tail anchor (TA). For example, mitofusins span the MOM twice with the N- and the C-terminus facing the cytosol (Rojo et al., 2002). A specific class of TA proteins, however, termed here “classical” TA proteins, have a cytosolic N-terminal part that is membrane- embedded via a single hydrophobic Figure 1.8. Membrane insertion of tail-anchored proteins occures after release from the ribosome. The C-terminal transmembrane segment close (less than transmembrane region is so close to the carboxy terminus that it emerges from the ribosome only upon termination of translation thirty amino acids) to the C-terminus (Fig. and is therefore unlikely to interact with signal recognition particles (Borgese et al., 2003). 1.8). This TA-domain is sufficient for

19 INTRODUCTION efficient posttranslational targeting and integrates into the membrane (Borgese et al., 2003; Borgese et al., 2001; Wattenberg and Lithgow, 2001). Notable examples of classical TA-proteins of the MOM include small import receptors of the preprotein translocase of the outer mitochondrial membrane (TOM complex; Neupert and Herrmann, 2007). Furthermore, various proteins involved in mitochondrial dynamics belong to the group of TA proteins (Table 3.1) including the mitochondrial fission factors Fis1 (Borgese et al., 2007) and Mff (Gandre-Babbe and van der Bliek, 2008), and the proapoptotic protein Bak, which also regulates mitochondrial morphology in non-apoptotic conditions (Karbowski et al., 2006).

Although TA-proteins lack a N-terminal targeting signal, they specifically integrate into a limited number of organelles such as the endoplasmic reticulum (ER), mitochondria, peroxisomes, and chloroplasts (Borgese et al., 2007). The best-studied ER- and MOM- targeted TA proteins share structural and mechanistic features but strongly differ in other aspects. Specificity of targeting of classical TA proteins to the MOM is mediated by basic residues flanking a short TMD, while in ER-targeted proteins the TMD is in general longer and flanked by neutral or acidic amino acids (Fig. 1.9; Borgese et al., 2003; Borgese et al., 2001; Egan et al., 1999; Kuroda et al., 1998; Wattenberg and Lithgow, 2001). ER-targeted TA proteins with more hydrophobic TMDs require accessory factors for translocation across a lipid bilayer, while TMDs with limited hydrophobicity can translocate without assistance (Brambillasca et al., 2006). Chaperones and a targeting factor for assisted TA protein ER integration have been identified (Abell et al., 2007; Stefanovic and Hegde, 2007). The mechanism of TA-MOM protein integration remains a matter of debate. Studies suggest that Tom40 is an essential component during the integration process of classical TA proteins (Borgese et al., 2007) but Setoguchi et al. (2006) found that MOM- targeted TA proteins even with a more hydrophobic TMD are still capable of unassisted translocation. Despite progress in understanding the underlying cellular machinery of TA protein targeting and insertion, the functional role of the tail region remains elusive (Borgese et al., 2007; Borgese et al., 2003). The exchange of the TA of Fis1 with the TA of Tom5 or Tom6 demonstrated Figure 1.9. Features that determine the targeting of TA proteins to the MOM or the ER. MOM targeted proteins are characterized by a rather that the TA is required for targeting short TMD regions (<20 residues), flanked by basic aa (Borgese et al., 2003). of Fis1 but not for its activity in

20 INTRODUCTION yeast. Yet, these overexpressed fusion proteins were partially integrated into the TOM complex, suggesting that the TA of Tom5 and Tom6 might additionally act as an assembly signal (Habib et al., 2003). Further clarification of the important issue whether the TA of classical TA proteins is, in addition to its function in membrane anchoring, of other functional significance, requires appropriate assays that assess protein activity in response to alterations of the TA.

1.5 GDAP1L1 – the GDAP1 paralog

GDAP1 disease mutations cause axonal, intermediate or demyelinating forms of the peripheral neuropathy CMT. Expression of GDAP1 is found in tissues of the PNS and CNS (Niemann et al., 2005; Pedrola et al., 2005). This implies that either GDAP1 function is somewhat compensated by another protein in the CNS, or that GDAP1 exerts different functions in the CNS compared to the PNS. Evolutionary studies found genes very similar to GDAP1 in several vertebrate species including mammals, birds and fishes. With a protein sequence homology of 60 % one obvious paralog of GDAP1 was identified, Ganglioside-induced differentiation associated protein-like 1 (GDAP1L1; Fig. 1.10; Marco et al., 2004). GDAP1L1 consists of six exons that span 33.1 kb of DNA and is located at the chromosomel region 20q12-q13. It is a 368 amino acid protein with a molecular weight of 42 kDa. Comparable to GDAP1, GDAP1L1 contains two hydrophobic domains at the outermost C-terminal region (Marco et al., 2004). However, membrane integration could was not previously confirmed. Additionally, some programs predict a third transmembrane domain just upstream of the other two hydrophobic domains (Marco et al., 2004). Due to its high sequence similarity to GDAP1 with the highly conserved GST-N and GST-C domains, the similarly enlarged interdomain, and the C-terminal hydrophobic domains, GDAP1L1 belongs to the same novel GDAP1- class of GST enzymes as GDAP1. Yet, bacterially expressed recombinant protein does not display GST-activity (Shield et al., 2006). B. Angst found three potential splice variants for mouse GDAP1L1 in the EST database, but could confirm only isoform A (367 aa long), the primary and functionally relevant form and isoform C (81 amino acid long N-terminal fragment), a truncated form that constitutes only half of the GST-N domain (Angst, 2005; Marco et al., 2004). Isoform B (370 amino acid long) differs from isoform A only by three additional residues (Arg-Leu-Gln inserted after Lys60) but could not be confirmed by quantitative RT-PCR or Western blot analysis and thus does not represent a true isoform (Angst, 2005). Quantitative RT-PCR analysis identified an expression of GDAP1L1 in the CNS in cortex, thalamus, cerebellum, olfactory bulb and spinal cord (Angst, 2005). Analysis of mRNA expression levels in the PNS, namely in dorsal root ganglia and sciatic nerves remained

21 INTRODUCTION

Figure 1.10. Structural comparison of GDAP1 and GDAP1L1. (A) Domain structure of GDAP1 and GDAP1L1 and (B) amino aicd sequence alignment of GDAP1 and GDAP1L1 proteins. GDAP1 is the upper sequence with alpha x indicating the prototypic GST alpha-helices; the two additional interdomain helices specific for the GDAP1 family are displayed in red. Hydrophobic domains and potential transmembrane domains are colored corresponding to the domain structure under (A). inconclusive and will be evaluated on protein level in this study. In contrast to GDAP1, overexpressed GDAP1L1 does not localize to mitochondria. GDAP1L1 is ubiquitously distributed throughout the cytosol (Angst, 2005). These findings are surprising, as the carboxy terminal region of GDAP1L1 is highly similar to the corresponding region of GDAP1 including all features essential for tail-anchored protein integration. Further subcellular localization studies on the endogenous population of GDAP1L1 protein in cell lines failed, as none of the tested cell lines (COS-7, HeLa, MSC80, HEK293, IEC, HaCaT, PaTu, T896, MC 3T3, 3T6, NIH 3T3, BHK, Caki, P19, CV1, RN22, ISC, D6P2T, SH- Sy5y, N1E-115, Neuro2a) expressed detectable levels of endogenous protein (Angst, 2005). Interstingly, subcellular localization studies revealed that in contrast to isoform A and C, isoform B displays a distinct mitochondrial pattern (Angst, 2005). GDAP1L1 subcellular distribution might also be regulated, for instance by alteration of the chaperone repertoire in cells, through posttranslational modifications, or via external stimuli that trigger certain intracellular responses. Modulations due to external apoptotic stimuli have been found for the anti-apoptotic protein Bcl-2 and the pro-apoptotic protein Bax, both are, like GDAP1, TA proteins of the MOM. Bcl-2 is targeted to the endoplasmic reticulum and mitochondria. However upon interaction of its N-terminal domain with a

22 INTRODUCTION

MOM-localized binding partner (FKB38) it preferentially inserts into the MOM (Borgese et al., 2003). Although TA proteins are usually localized to membraneous structures, under normal cellular conditions Bax resides in the cytosol. Integration into the MOM takes exclusively place in response to an apoptotic signal. The C-terminal region of Bax occupies a hydrophobic pocket, which provides a binding site for the BH3 domain of Bcl-2 family proteins. Upon interaction the C-terminal tail is displaced and free to insert into the MOM. Subsequently, the pro-apoptic activity of Bax is switched on (Borgese et al., 2003).

23 INTRODUCTION

1.6 Aim of the study

Recent studies showed that GDAP1 is a regulator of the mitochondrial network; it promotes mitochondrial fragmentation and is involved in mitochondrial fission processes. Mutations in GDAP1 lead to demyelinating, axonal and intermediate forms of Charcot- Marie-Tooth disease, a peripheral motor and sensory neuropathy without severe CNS phenotype. It is well established that the nervous system is especially sensitive to mitochondrial dysfunction and several neuropathies are linked to proteins involved in the fusion and fission machinery of mitochondria. Despite ongoing research, the exact functional mechanisms of GDAP1 still remain largely unclear and further characterization of its structure and function will provide the basics for uncovering its role in mitochondrial dynamics and its effect on peripheral nerves. The aim of this study was to elucidate the structure-function relationships of GDAP1 in the peripheral nervous system.

I demonstrate that recessive and dominant inherited GDAP1 disease mutations, although leading to similar clinical symptoms, differ in their mode of action – impairment of mitochondrial fission vs. blocked mitochondrial fusion. The dominant inherited mutated forms of GDAP1 interfere with mitochondrial fusion. This results in the accumulation of mitochondrial damage, as illustrated by increased ROS load and an uneven mitochondrial transmembrane potential (
m).

GDAP1 was defined as the founder of a new GST family, yet bacterially expressed recombinant protein does not display GST-activity. In this study I could show that insect cell expressed recombinant GDAP1 displays GST activity against several substrates tested, and thus is an active GST enzyme.

With its two C-terminal hydrophobic domains that integrate GDAP1 into the MOM GDAP1 contains structural features untypical for classical GST enzymes. This study was further aimed at determining the precise topology of GDAP1, the motifs involved in MOM targeting and integration, and the relationship between tail-anchor structure and the fission function of the protein.

Expression of GDAP1 is found in tissues of the PNS and CNS. This implies that either GDAP1 function is somewhat compensated by another protein in the CNS, or that GDAP1 exerts different functions in the CNS compared to the PNS. Thus, in this study I elucidated whether GDAP1L1, the paralog of GDAP1 expressed exclusively in the CNS is also able to fragment mitochondria and therefore could compensate for the function of the mutated GDAP1 in the CNS.

24 MATERIALS & METHODS

2. Materials and Methods

2.1 Frequently used solutions and Equipment

2.1.1 Solutions

If it is not mentioned differently, the water used to prepare the solutions (ddH2O) is deionized and sterile filtered water, which was purified by a water purifying apparatus (Purelab classic, ELGA LabWater, Switzerland). All chemicals were purchased from Sigma-Aldrich (USA) if not mentioned differently.

Table 2.1. Frequently used buffers and solutions Solution Description

1 x Loading buffer for SDS- 50 mM Tris/HCl pH 6.8, 100 mM DTT, 2% (w/v) SDS, PAGE 0.1% (w/v) Bromphenolblue, 10% (v/v) Glycerin

Ampicillin Stock solution: 50 mg/ml in ddH2O, sterile filtered with a 0.2 μm syringe filter (Sartorius, Goettingen, Germany), stored at -20°C for 1 month. Working concentration: 100 μg/ml for liquid LB- medium, 50 μg/ml for LB-culture plates.

Kanamycin Stock solution: 10 mg/ml in ddH2O, sterile filtered with a 0.2 μm syringe filter (Sartorius, Goettingen, Germany), stored at -20°C for 1 month. Working concentration: 20 μg/ml for liquid LB- medium, 25 μg/ml for LB-culture plates.

IPTG 100 mg/ml in ddH2O X-gal 50 mg/ml in dimethylsulfoxide LB-Medium/ LB-culture plates 1% (w/v) Bacto
Tryptone, 0.5% (w/v) Bacto
Yeast Extract (Becton Dickinson, Sparks, USA), 1% (w/v) NaCl, (for plates: 15% (w/v) agar) Loading buffer for agarose gel 20% (w/v) Ficoll PM400, 50 mM EDTA pH 8.0 electrophoresis Cell fractionation buffer 210 mM mannitol, 70 mM sucrose, 1 mM EDTA, 10 mM Hepes- NaOH pH 7.5

PBS 140 mM NaCl, 2.5 mM KCl, 6.5 mM Na2HPO4, 1.5 mM KH2PO4, pH 7.25 Running buffer for SDS-PAGE 25 mM Tris-base, 190 mM Glycin, 0.1% SDS (w/v) (Lämmli buffer) TAE buffer 40 mM Tris-acetate, 2 mM EDTA TBS buffer 20 mM Tris/HCl pH 7.5, 150 mM NaCl TE buffer 10 mM Tris/HCl pH 8.0, 1 mM EDTA pH 8.0

25 MATERIALS & METHODS

2.1.2 Equipment

Centrifuges Tabletop centrifuge type Pico17 (Heraeus, Zurich, Switzerland) Cooled tabletop centrifuge type Fresco17 (Heraeus, Zurich, Switzerland) Cooled tabletop centrifuge type Megafuge 1.0R (Heraeus, Zurich, Switzerland) Cooled High Speed centrifuge type Z36HK (Hermle, Wehingen, Germany) Cooled Superspeed centrifuge type RC-5C (Sorvall, Zurich, Switzerland)

Incubators

CO2 Incubator type HERAcell 150 (Thermo Scientific, Pittsburgh, USA) Incubator for bacteria culture type FED53 (Binder, Tuttlingen, Germany) Incubator shaker for bacteria culture type Lab-Therm LT-X (A. Kühner AG, Birsfelden, Switzerland)

Cell culture hood Laminar Flow type BDK-SK 1500 (BDK, Sonnenbühl, Germany)

Microscopes For visualizing immunofluorescence two Zeiss fluorescence microscopes were used: Axioplan imaging (Zeiss, Oberkochen, Germany) with following filter sets (Zeiss): Fs2 for DAPI, Fs17 for Cy2 and Alexa488, Fs43 for Cy3. Observer. D1 (Zeiss, Oberkochen, Germany) with the following filter sets (Zeiss): Fs49 for DAPI, Fs38 for Cy2 and Alexa488, Fs43 for Cy3, Fs50 for Cy5. Pictures were taken using the digital camera AxioCam MRm (Zeiss, Oberkochen, Germany) and visualized by Adobe’s Photoshop CS3 for Macintosh. For co-localization analysis, a confocal microscope was used: LSM 510-NLO (Zeiss, Oberkochen, Germany). Images were visualized using Imaris and Selima image processing software (Bitplane AG, Technopark Zurich, Switzerland) and Adobe’s Photoshop CS3 for Macintosh.

Photometer Eppendorf
Biophotometer (Vaudaux, Schönenbuch, Switzerland)

UV-transilluminator UVI-tec Gel Documentation (312 nm wavelength) (Witec AG, Littau, Switzerland) UV-transilluminator (365 nm wavelength) (OMNI LAB, Mettmenstetten, Switzerland)

26 MATERIALS & METHODS

Microtome Rotary Microtome HM 330 (Microm, Heidelberg, Germany)

PCR apparatus RoboCycler
96 Gradient (Stratagene, La Jolla, USA) T Professional Basic Gradient (Biometra, Göttingen, Germany)

Homogenizer U200S control Sonic Homogenizer (50-60 Hz) (IKA Labortechnik, Staufen, Germany) Dounce Tissue Grinder (7 ml) (Wheaton, Millville, USA) Polytron PT1200 (Kinematica AG, Switzerland)

SDS-PAGE and Western Blot Vertical electrophoresis system type SE300 (Hoefer Inc., San Francisco, USA) Vertical Gel System (C.B.S., Del Mar, USA) Mini Trans-Blot cell (Biorad, Hercules, USA)

Film Processor Medical Film Processor RG II (Fujifilm, Sint-Niklaas, Belgium)

2.2 Molecular biology methods

2.2.1 cDNA amplification Total RNA from SH-SY5Y cells was isolated using the RNeasy Mini Kit (Qiagen, Düsseldorf, Germany). First strand cDNA synthesis was performed from 400 ng of isolated total RNA. 2 μl of random hexamer primers (25 μM), 1 μl dNTP’s (10 mM each) and 12 μl RNase-free water were added, the reaction was incubated at 65°C and after 5 min immediately placed on ice and quickly centrifuged. 4 μl of 5x First Strand Buffer (Invitrogen), 2 μl 0.1 M DTT and 1 μl RNase Out (Invitrogen) were added and incubated for 2 min at 42°C. After addition of 1 μl RT Superscript II (Invitrogen) the reaction was incubated for 10 min at 25°C and 50 min at 42°C. Finally the reaction was inactivated at 70°C for 15 min.

2.2.2 Restriction digest All enzymes used for the restriction digest were purchased from New England Biolabs (Beverly MA, USA). Digests were performed using 5 units enzyme/μg DNA, 1.5 μl of the

27 MATERIALS & METHODS appropriate reaction buffer and 1.5 μl BSA (1 mg/ml) in a final reaction volume of 15 μl. The reaction was incubated according to the manufacturers instructions for 2 h.

2.2.3 DNA gel electrophoresis For DNA separation 1–1.8% (w/v) agarose gels were used. The agarose (Invitrogen, Carlsbad, USA) was melted in TAE buffer (2.1.1) and 100 ng/ml ethidiumbromide was added. The DNA probes were mixed with 6x loading buffer (2.1.1) and loaded. As molecular weight standard the 2-log ladder (NEB, Beverly, USA) was used. The electrophoresis was performed at 120 V for 30 - 60 min and the gel was analyzed using UV light in a transilluminator. In order to purify DNA fragments the gel was laid on a transilluminator with UV light of higher wavelength (2.1.2) to minimize DNA damage. The DNA fragment of the right size was cut out of the gel and purified with a Gel Extraction Kit (Macherey-Nagel, Düren, Germany).

2.2.4 Ligation All ligations performed in this study were performed with DNA containing complementary overhangs. For the ligation reaction a 3-6 fold molar surplus of the insert was added to the vector (10-100 ng). 1 μl of T4 DNA ligase (Roche Diagnostics, Indianapolis, USA) and 2 μl of the appropriate reaction buffer were added (total volume 20 μl). For pGEM-T vector ligations, the pGEM-T vector systems Kit (Promega, Madison, USA) was used. The reaction volume was filled up with ddH2O and incubated for 2 h at room temperature.

2.2.5 Transformation of E.coli

200 μl of CaCl2 competent Dh5
E. coli cells were thawn on ice. 2 μl of the ligation reaction or 100 ng plasmid DNA were added and incubated on ice. After 30 min a heat shock was performed for 45 seconds at 42°C. 1 ml of LB-medium (2.1.1) was added and the samples were incubated for 1 h at 37°C. After centrifugation (5000g, 1 min) 1 ml of the supernatant was discarded and the pellet was resuspended in the remaining 200 μl. This suspension was plated on LB-culture plates (2.1.1) containing the appropriate antibiotic. The plates were incubated over night at 37°C. As negative control an opened but empty vector was transformed. As control for the activity of the antibiotics untransformed bacteria were plated.

2.2.6 DNA purification Clones of transformed bacteria were picked from the agar plates and grown over night at 37°C on a shaker (250 rpm) in 3 ml LB-medium containing the respective antibiotic (2.1.1). DNA was extracted from the 3 ml cultures according to the protocol of the Nucleospin Miniprep Kit (Macherey-Nagel, Düren, Germany). The purified DNA was subsequently

28 MATERIALS & METHODS used for restriction digests, gel electrophoretic analysis and ligations (2.2.3-4). To obtain high amounts of purified DNA of those samples the 3 ml cultures were diluted 1:500 in 100 ml LB-medium containing the respective antibiotic and incubated over night on a shaker (250 rpm) at 37°C. The following DNA extraction and purification was performed according to the protocol of the Nucleobond PC500 Kit (Macherey-Nagel, Düren, Germany).

2.2.7 Cloning of constructs All primers used in this study were purchased from Microsynth (Balgach, Switzerland) and are listed in Table 6.1. All following PCR reactions were performed with the PicoMaxx High Fidelity PCR System (Stratagene, La Jolla, USA) according to the manufactures instructions if not indicated differently. All cDNAs were cloned using the pGEM-T vector (Promega, Madison, USA), subcloned into pcDNA3.1 (Invitrogen, Carlsbad, USA) or EGFP-C1/N1 (Clonetech Laboratories, Mountain View, USA) and verified by sequencing (2.2.8). Mfn1-10xMyc, Mfn1(K88T)-10xMyc, Mfn2-16xMyc, Mfn2(K109A)-16xMyc, Drp1-6xMyc and Drp1(K38A)-6xMyc were provided by D. Chan (California Institute of Technology, Pasadena, USA). MtDsRED2 and mtEGFP were obtained from Clontech Laboratories (Mountain View, USA), the HA-C1 vector was provided by A. Hirschy (ETH Zurich, Zurich, Switzerland). The constructs GDAP1 (in pGEM-T), GDAP1 (in pcDNA3.1), EGFP-GDAP1, GDAP1-EGFP, EGFP-HD1, EGFP-TMD, EGFP-HD1-TMD, M116H, R310Q and GDAP1L1 were generated by A. Niemann, hFis1, R161H, R282C and T288fs290X were cloned by M. Rüegg, R120Q by K. Schwerdtfeger and EGFP-HD2- HD1(GDAP1L1) by B. Angst. All GDAP1 constructs of the following sections are cloned on the basis of either GDAP1 in pGEM-T or GDAP1 in pcDNA 3.1 as template.

2.2.7.1 Cloning of GDAP1 disease- and point mutations GDAP1 point mutations were generated according to the protocol of the QuikChange Site- Directed Mutagenesis Kit (Stratagene, La Jolla, USA). Table 2.2 summarizes the primer pairs used to generate the specific constructs based on the template GDAP1 in pGEM-T vector.

29 MATERIALS & METHODS

Table 2.2. Primers used to generate disease and point mutations Disease/Catalytic centre Tail-anchor TMD length

Construct Primer Construct Primer Construct Primer

R120W 7+8 K291A 77+78 TMD+1 25+26 T157P 9+10 K313A 79+80 TMD+3 27+28 Q218E 11+12 K314A 81+82 TMD+5 29+30 G271R 13+14 R315A 83+84 TMD-1 31+32 L239F 15+16 K318A 85+86 S130C 17+18 R341A 87+88 M116T 19+20 R342A 89+90 D149Y 21+22 R343A 91+92 Y29A 67+68 K314AR315A 93+94 S34A 69+70 K313K314R315A 95+96 S36A 71+72 R341AR342A 97+98 S37A 73+74 R341R342R343A 99+100 F68A 75+76 GDAP1hy 37-42

2.2.7.2 Cloning of GDAP1 deletion mutants Table 2.3 summarizes the primer pairs used to construct the deletion mutants of GDAP1. The constructs
N-terminus and
C-terminus were generated in one PCR step using appropriate forward and reverse primer. The constructs
GST-N,

4-helix,
interdomain,
GST-C and
HD1 were generated in two steps. First, the N- and C- terminal fragment excluding the deletion were amplified by PCR separately and cloned in to pGEM-T. In the PCR step, restriction sites for XhoI were introduced. In a second step the fragments were digested with XhoI and NotI and corresponding fragments were ligated. The construct
short_interdomain was created by overlapping PCR.

Table 2.3. Primers used to generate GDAP1 deletion mutants Construct N-terminal C-terminal Final PCR step fragment fragment
N-terminus 47+2
GST-N 48+49 50+2

4-helix 1+51 52+2
interdomain 1+53 54+2
short_interdomain 1+55 56+2 1+2
GST-C 1+57 58+2
HD1 1+59 60+2
C-terminus 1+62

30 MATERIALS & METHODS

In a first PCR step a N-and C-terminal fragment excluding the interdomain region were generated with primers containing complementary ends. In a second PCR step, the PCR products of the first PCR were used as template and the construct was amplified with the forward primer of the N-terminal fragment and the reverse primer of the C-terminal fragment.

2.2.7.3 Cloning of hGDAP1 hGDAP1 was amplified by RT-PCR (2.2.1) from SY5Y cells using primer pair 3/4 and cloned into the pGEM-T vector. For subcloning into pcDNA3.1 vector BamHI and SalI restriction sites were inserted by PCR using the primer pair 5/2.

2.2.7.4 Cloning of FLAG-tagged GDAP1 and hGDAP1 GDAP1-FLAG and hGDAP1-FLAG were generated by PCR using the primer pairs 1/63 and 5/6 respectively. GDAP1-FLAG+8 was amplified in two successive PCR steps, the first step containing the primer pair 1/23 and the second step with the primer pair 1/24.

2.2.7.5 Cloning of constructs used to analyze the C-terminal region of GDAP1 The constructs TMDscr and HD1scr were generated by PCR in three steps. The first step (using the primer pairs 1/43 and 44/2 for TMDscr and 1/35 and 33/2 for HD1scr) produces a C-terminal and a N-terminal fragment excluding the TMD or HD1 region respectively. The second PCR step elongates the products of the first PCR using the primer pairs 1/45 and 46/2 for the construct TMDscr and 1/34 and 36/2 for HD1scr. The third PCR step results in an amplification of the pooled products of step two with the primer pair 1/2 for both constructs. TMD-EGFP and HD1-TMD-EGFP were generated by PCR using the primer pairs 64/65 and 64/66 respectively.

2.2.7.6 Cloning of chimera GDAP1-Vamp1B and GDAP1-Omb5 Vamp1B cDNA was amplified by RT-PCR form mouse brain cDNA (provided by P. Berger, PSI, Villigen, Switzerland) using the primer pair 101/102. GDAP1-Vamp1B was generated in two PCR steps. In the first PCR step, the N-terminal GDAP1 fragment was amplified using the primer 1/104 and the C-terminal Vamp1B fragment usind the primer 103/102. The second PCR amplifies the in step one with overlapping ends generated GDAP1 and Vamp1B PCR products using the primer pair 1/102. hOMb5 cDNA was amplified from HeLa cell cDNA (P. Berger, PSI, Villigen, Switzerland) using the primer pair 105/106. FLAG-hOMb5 was generated by PCR usind the primer pair 156/106. GDAP1-hOMb5 was generated in two PCR steps comparable to GDAP1-Vamp1B with the primer pairs 1/108 and 107/106 in a first PCR step and 1/106 in a second PCR step. GDAP1-rOMb5 was generated via PCR amplification on the template GDAP1- hOMb5/pGEM-T with the consecutively used primer pairs 1/109, 1/110 and 1/111.

31 MATERIALS & METHODS

2.2.7.7 Cloning of GDAP1L1 constructs EGFP-HD1(GDAP1L1) and EGFP-HD3-HD2-HD1(GDAP1L1) were generated by PCR using appropriate primer pair (Table 2.4) and cloned in to pEGFP-N1 vector. The constructs G-SF(TMD) and SF-G(TMD) were amplified in two PCR steps, G-SF(HD2) and G-SF(HD1) in three PCR steps as described in section 2.2.7.5 (Table 2.4).

Table 2.4. Primer used to generate GDAP1L1 constructs Construct N-terminal C-terminal Final PCR step fragment fragment EGFP-HD1 114+113 EGFP-HD3-HD2-HD1 115+113 G-SF(TMD) 1+119 118+113 1+113 SF-G(TMD) 112+117 116+2 112+2 PCR1 1+126 PCR1 124+113 G-SF(HD2) 1+113 PCR2 1+127 PCR2 125+113 PCR1 112+122 PCR1 120+2 SF-G(HD1) 112+2 PCR2 112+123 PCR2 121+2

2.2.7.8 Cloning of GDAP1 and GDAP1L1 constructs for insect cell expression GDAP1 288X, GDAP1 318X, GDAP1 16_309X, GDAP1 16_288X, GDAP1 point mutants, GDAP1L1 fl, GDAP1L1 340X and GDAP1L1 310X were generated using appropriate primer pairs that contain a N-terminal BstEII or BamHI site and a C-terminal RsrII site for subcloning into the insect cell expression vector pBDO (Table 2.5). GDAP1- MBP (maltose binding protein) constructs were created using the method of SLIC-cloning (Li and Elledge, 2007) with primer pairs listed in Table 2.5. GDAP1-MBP constructs were cloned into the pBDO vector or the pMALc-vector with or without a linker sequence (_li) between the MBP protein and GDAP1.

Table 2.5. Primer used for recombinant expression of GDAP1/GDAP1L1 GDAP1/GDAP1L1 expression constructs Disease mutant expression constructs Construct Primer Template Construct Primer

GDAP1 288X 128+129 S34A 288X 128+135/1337/138

GDAP1 318X 128+130/131 R120Q 288X 128+135/1337/138 GDAP1 16_309X 132+133 R120W 288X 128+135/1337/138 GDAP1 16_288X 132+134 T157P 288X 128+135/1337/138 GDAP1L1 fl 151+157/158 L239F 288X 128+135/1337/138 GDAP1L1 340X 151+1554/155 G271R 288X 128+135/1337/138 GDAP1L1 310X 151+152/153 R282 288X 128+136/1337/138 MBP-GDAP1 Vector 140+141 GDAP1 288X/pBDO S34A 318X 128+130/131 288X Insert 139+142 pMALc

32 MATERIALS & METHODS

MBP-GDAP1_li Vector 140+147 GDAP1 288X/pBDO R120Q 318X 128+130/131 288X Insert 139+146 pMALc MBP-GDAP1 Vector 140+141 GDAP1 318X/pBDO R120W 318X 128+130/131 318X Insert 139+142 pMALc MBP-GDAP1_li Vector 140+147 GDAP1 318X/pBDO T157P 318X 128+130/131 318X Insert 139+146 pMALc MBP-GDAP1 Vector 142+143 pMALc L239F 318X 128+130/131 288X Insert 141+144 GDAP1 288X/pBDO MBP-GDAP1_li Vector 145+146 pMALc R161H 318X 128+130/131 288X Insert 147+148 GDAP1 288X/pBDO MBP-GDAP1 Vector 142+143 pMALc G271R 318X 128+130/131 318X Insert 141+144 GDAP1 318X/pBDO MBP-GDAP1_li Vector 142+143 pMALc R282C 318X 128+130/131 318X Insert 141+144 GDAP1 318X/pBDO

2.2.8 Sequencing analysis To verify the cloned constructs, sequencing was performed. The primer used for the sequencing reactions are listed inTable 2.6.

Table 2.6. Primer used for sequencing reactions Primer Sequence SP6 5’-ATTTAGGTGACACTATAG-3’ T7 5’-TAATACGACTCACTATAGGG-3’ EGFP_C1-for 5’-TGCTGGAGTTCGTGA-3’ GDAP1_f224 5’-AAAGGTGCGCTTGGTAATTG-3 GDAP1_f435 5’-TTAATGCCCGATGAAGGAAG-3’ GDAP1_f666 5’-ATTGCCAAACAGAAGCGACT-3’ GDAP1_f950 5’-TGAGCGTGTCTTGAAGAGAAA-3’ GDAP1_r319 5’-CGCATAAACCAAGGCTCATT-3’ GDAP1_r507 5’-TTGGCAAGGAGTCAAGAAGC-3’ GDAP1_r771 5’-TTTCCACCTGATCCAAGACC-3’ GDAP1_r1027 5’-GGCAGCACCGCAGAGATTA-3’ hGDAP1_f325 5’-TTAATGCCTGATAAAGAAAG-3’ hGDAP1_f556 5’-ATTGCAAAACAGAAACGACT-3’ hGDAP1_f840 5’-CGAGCGTGTCTTGAAGAGAAA-3’ hGDAP1_r397 5’-TTGGCAAGGAGTCAAGCAGC-3’ hGDAP1_r661 5’-TTTCAACCTGATCCAAGACT-3’ hGDAP1_r917 5’-GGCAGCACTGCAGAGATTA-3’ GDAP1L1_f155 5’-GGACCCAGTCCTTCAGCTC-3’ GDAP1L1_f355 5’-GACTACGTGGAACGCACCTT-3’ GDAP1L1_f532 5’-AAGTATGCCACGGCTGAGAT-3’ GDAP1L1_f741 5’-GCTGGAGAAGAGGAAACTGG-3’ GDAP1L1_f908 5’-GAGCACGATGACGTGAGCTA-3’ GDAP1L1_r180 5’-CTTCTGCGAGCTGAAGGACT-3’ GDAP1L1_r375 5’-GAAGGTGCGTTCCACGTAGT-3’ GDAP1L1_r575 5’-GTGGCGTTGGCTAAATGTCT-3’ GDAP1L1_r771 5’-CCCCTCATTCTCCAGTTTCC-3’ GDAP1L1_r989 5’-ACAGCAGAGTGGTGTGGATG-3’

The sequencing PCR was performed as shown in Table 2.7. 20 μl PCR product were precipitated with 2 μl sodium acetate pH 5.3 and 55 μl ethanol absolute. After

33 MATERIALS & METHODS centrifugation (30 min, 17000xg) the supernatant was removed and the pellet was washed with 60 μl 70 % (v/v) ethanol (10 min, 17000xg). The pellet was dried for 15 min at room temperature in the dark and resuspended in 30 μl HiDi Formamide (Applied Biosystems, Warringtion, UK). The samples were denatured for 3 min at 95°C and immediatley transfered on ice. Sequencing was performed by the ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Warrington, UK).

Table 2.7. Sequencing PCR Material Amount Temperature Time Cycles DNA 100ng/μl 1 μl 95°C 1 min 1x Primer 2.5 μM 1 μl 95°C 30 s Reaction Mix 2 μl 50°C 30 s 25x 5x buffer 3 μl 60°C 4 min

ddH2O 13 μl 4°C 1 min 1x Total 20 μl

2.3 Cell culture

2.3.1 Cell culture solutions and materials

Table 2.8. Solutions and materials used in cell culture Solution Chemicals (Company)

PBS see 2.1.1 Trypsin solution 0.25 % Trypsin/EDTA (for cell lines; Invitrogen, Carlsbad, USA) 10x Trypsin (for primary neurons; Gibco, Carlsbad, USA) Serum (FCS) Fetal Calf Serum (Brunschwig, Amsterdam, Netherlands) Antibiotics Penicillin/Streptomycin (P/S; Invitrogen, Carlsbad, USA) Growth Medium Cell lines: Dulbecco’s modified Eagle’s medium containing GLUTAMAX (#31966, Invitrogen, Carlsbad, USA) + 10 % FCS (+ 1 % antibiotics if indicated) Primary Neurons (neuron culture medium): Dulbecco’s modified Eagle’s medium containing 4500 mg glucose/L, 110 mg sodium pyruvate/L, L- glutamine (#D6429, Sigma) + B27 supplement (50x, Gibco, Carlsbad, USA) + Albumax I (25x (125 mg/ml), Gibco, Carlsbad, USA) +1 % Antibiotics Culture plates 150 cm2 (
140 mm), 55 cm2 (
100 mm), 21 cm2 (
60 mm) 8 cm2 (
35 mm) and 12-well (4 cm2 each well) plastic culture plates (Nunc, Rochester, USA)

2.3.2 Cell lines All cell lines used in this study were cultured in DMEM supplemented with 10% FCS in a humidified atmosphere at 37°C, 5% CO2 (Table 2.9). If not otherwise indicated no antibiotics were used to maintain the cell lines.

34 MATERIALS & METHODS

Table 2.9. Cell lines used for this study Cell Line Description African green monkey (Cercopithecus aethiops) kidney fibroblast-like cell line COS-7 suitable for transfection by vectors requiring expression of SV40 T antigen. Human (homo sapiens) epithelial cell line derived from a cervix HeLa adenocarcinoma. Cell line derived from a spontaneous mouse (mus musculus) brain N1E-115 neuroblastoma tumor. Human (homo sapiens) brain neuroblastoma cell line derived from a 4 year old SH-SY5Y female patient.

2.3.3 Primary cell cultures Mouse hippocampal and cortical neurons were prepared as follows: In a first step glial cells from P0 to P2 C57BL/6 mice (Janvier, Strassburg, France) were prepared for a feeder cell layer to culture hippocampal neurons. The mice were decapitated, meninges were removed from the cortical halfs and the cortical halfs were stored in HBSS+ buffer (Invitrogen, Carlsbad, USA). The cortices were washed once in HBSS- (Invitrogen, Carlsbad, USA), the wash buffer was aspirated and 5 ml HBSS-/ 0.5 ml 10x Trypsin/EDTA were added. The cortices were incubated for 5 minutes in a 37°C waterbath before the supernatant was removed. 5 ml of DMEM +10% FCS were added and the cortices were carefully triturated with a 1 ml pipette until they were completely homogenized. The homogenate (1 cortex per plate) was distributed to 55 cm2 culture dishes and supplemented with DMEM (+10% FCS +1% P/S) to a final volume of 10 ml.

The cells were grown in a humidified atmosphere at 37°C, 5% CO2. Confluent glia cells can be splitted (maximum 1:4) for amplification. For use as feeder layer a confluent dish of glia cells is used for two 12-well plates. The cells were split into the 12-well plates two days before adding hippocampal neurons. The hippocampus of P0 C57BL/6 mice (Janvier, Strassburg, France) was isolated, stored in a tube containing ice-cold HBSS+ and treated comparable to the glia cells above. After trituration with a 1 ml pipette additional medium was added (500 μl per hippocampus). The cells were counted in a Neubauer chamber (LO Laboroptik AG, Friedrichsdorf, Germany) and diluted to obtain a final concentration of 50’000 to 100’000 cells/ml. 100 μl of this cell suspension was plated on poly-L-lysine coated 18mm glass coverslips (Thermo

Scientific, Braunschweig, Germany) and incubated for 1 hour at 37°C, 5% CO2. The coverslips were then transferred upside down into 12-well plates containing the glial cells in 1 ml neuron culture medium. To avoid the direct contact of the coverslip on the glial cells paraffin droplets were placed in the wells as spacers before the glial cells were seeded. 3 days after plating cytosine arabinoside (Ara C, final concentration 15 μM) was added to prevent the outgrowth of glia cells and possible contaminations of fibroblasts.

35 MATERIALS & METHODS

2.3.4 Transient transfection COS-7 cells were transfected using Fugene 6 (Roche, Indianapolis, USA), HeLa, N1E-115 and SH-SY5Y cells using Lipofectamine2000 (Invitrogen, Carlsbad, USA) according to the manufacturers' protocols. For Lipofectamine2000 transfections the medium was changed 4 h after transfection to avoid cytotoxic effects.

2.4 Methods to analyze topology and localization of GDAP1/GDAP1L1

2.4.1 Fluorescence protease protection assay The assay was performed according to the protocol of Lorenz et al. (2006). COS-7 cells were grown on 12-well plates and analyzed 24 h post tranfection with indicated constructs. Cells were carefully washed twice in 1 ml KHM buffer (110 mM potassium acetate, 20 mM HEPES, 2 mM MgCl2) at room temperature, subsequently incubated with 0.5 ml of 50 μM digitonin and 250 μM trypsin in KHM buffer and imaged immediately. Images were taken at indicated time points using a Zeiss Observer microscope with a Zeiss MRM camera and AxioVision 4 software, and processed by Adobe Photoshop.

2.4.2 Subcellular localization and protease digest Two 55 cm2 plates of HeLa cells were transfected with either GDAP1 or GDAP1-FLAG expression constructs. After 24 hours, the cells were washed once with ice-cold cell fractionation buffer (210 mM mannitol, 70 mM sucrose, 1 mM EDTA, 10 mM Hepes- NaOH pH 7.5) and subsequently scraped from the culture plates in 0.5 ml cell fractionation buffer per plate using a Cell Scraper (23 cm, NUNC, USA). For each construct cells of the two transfected plates were combined with the cells of four untransfected HeLa cell plates (to increase cell volume) and dounced 20 times with a 7 ml Dounce Tissue Grinder (Wheaton, USA). The dounced cell suspension was centrifuged for 5 min at 1’000 g to remove cell debris and the supernatant (post-nuclear supernatant, S1) was again centrifuged 20 min at 10’000g to pellet mitochondria (mitochondria enriched fraction P1). Freshly isolated mitochondria were resuspended in 140 μl cell fractionation buffer and samples of 20 μl were treated with the indicated concentrations of proteinase K (10 mg/ml stock solution, 50 μg/ml final concentration), digitonin (16 mg/ml stock solution) and SDS. After 30 min on ice the digest was stopped with 5 mM PMSF (Olichon et al., 2002). SDS-PAGE and Western blotting procedures were performed as described in 2.7.1.

2.4.3 In vitro transcription and translation To synthesize radioactive labeled proteins we performed an in vitro trancription/translation from the respective cDNAs in pcDNA3.1 vector (including a T7 promotor) using the TNT T7 Quick Coupled reticulocyte system (Promega, Madison, USA) according to the

36 MATERIALS & METHODS manufactures protocol. For radioactive detection we added radioactively labeled 35S- methionine. To 20 μl of reticulocyte mastermix 0.5 μg plasmid DNA, 1 μl of 35S- methionine (2 μCi, HARTMAN ANALYTIC, Braunschweig, Germany) and 3.5 μl diethyldicarbonate (DEPC)-treated ddH2O (to inactivate RNases) was added. The in vitro transcription/translation reaction was incubated for 90 min at 30°C. The in vitro translates were directly used for the MOM integration assay and immuno-precipitation experiments (2.4.4) and analyzed by autoradiography using ENHANCE (Perkin Elmers, Boston, USA) according to the manufacturer’s protocol.

2.4.4 MOM integration assay and immuno-precipitation The protocols for membrane integration, differential centrifugation and the carbonate wash were adapted from Niemann, et al (2005). The integration assay was performed as follows: 1.5 ml post-nuclear supernatant (see 2.4.2; around 800 μg total protein obtained from six confluent 55 cm2 HeLa cell plates) was incubated with in vitro-synthesized proteins (10 μl of the transcription/translation reaction) for 60 minutes at 4°C. Two subsequent centrifugation steps were performed for 20 minutes at 10’000 g yielding the mitochondria enriched fractions P1 and P2, respectively. Equal volumes of supernatants and pellets were analyzed by autoradiography. To test for membrane integration of selected in vitro- translated constructs, the first mitochondrial pellet (P1) from the differential centrifugation procedure was equally split and resuspended in four different buffers: the cell fractionation buffer as control, 1 M sodium chloride, 0.1 M sodium carbonate pH 11, or cell fractionation buffer containing 0.1 % TX-100. The samples were centrifuged again for 20 min at 10’000 g. Equal volumes of the supernatants and pellets were analyzed by Western blot. To confirm the integration of in vitro-translated GDAP1-FLAG this radiolabeled protein was digested with 50 μg/ml proteinase K for 30 min at 4°C without (pure in vitro- translat) or after membrane integration (60 min incubation with post nuclear supernatant on ice). The digest was stopped with 5 mM PMSF for 10 min. The immuno-precipitation was done with the anti-FLAG M2 antibody in IP-buffer (50 mM Tris pH 7.5, 150 mM NaCl, 0.1 % SDS, 1 % Triton X-100, 1 mM PMSF, 1:100 Proteinase Inhibitor Cocktail). Therefore samples were incubated with 10 μl of the anti-FLAG M2 antibody at 4°C for 2 h. 50 μl of protein A-sepharose beads (GE Healthcare) were added for an additional hour. Samples were washed three times in IP-buffer, boiled at 90°C for 10 min in SDS-loading buffer (2.1.1) and analyzed by autoradiography.

2.4.5 Analysis of GDAP1L1 localization: screen of stress-inducing agents 24 h post GDAP1L1, GDAP1 or pcDNA3.1 transfection COS-7 and N1E-115 cells were treated with various stress inducing agents as listed in Table 2.10 for 2 h at 37°C. Cells were further processed as described in 2.7.2. For indicated cells the antioxidant N-acetyl-L-

37 MATERIALS & METHODS cysteine (NAC, 10 mM) was applied 3 min before and during menadione (MQ) application.

Table 2.10. Working dilutions of used stress inducing agents Reactive agent Stock (mM) Stock dilution Final Storage concentration (μM) Actinomycin D 10 DMSO 40 -20°C Staurosporin 1 DMSO 0.5 4°C m-Chlorophenylhydrazone 10 DMSO 10 -20°C (CCCP) Ethacrynic acid (EA) 10 mg/ml EtOH 50 μg/ml -20°C Rotenone (ROT) 10 DMSO 10 -20°C Menadione (MQ) 40 DMSO 20 -20°C tert-Butylhydroquinone 0.1 DMSO 100 -20°C (TBHQ) 2,3-Dimethoxy-1,4- 40 DMSO 20 -20°C naphtoquinone (DMNQ) tert-buthylhydroperoxide 7780 H O 1 4°C (TBHP) 2

Dithiotreitol (DTT) 500 H2O 250 -20°C

H2O2 9880 (30% w/w) - 100 4°C

N-Ethylmaleimide (NEM) 500 H2O 250 -20°C DMSO, dimethylsulfoxide, EtOH, ethanol absolute.

The MOM integration of GDAP1L1 upon MQ treatment was analyzed using the Qproteome Cell Compartment Kit (Qiagen, USA). 24 h post transfection HeLa cells of confluent 9.5 cm2 6-well plates were treated with 20 μM MQ and, if indicated, with 10 mM NAC for 1 or 2 h. Cells were harvested in culture medium after a 5 min treated with 200 μl trypsin (0.25 %) and processed according to the manufacturer’s protocol.

2.5 Protein isolation from mouse/ rat tissue

Tissues of C57BL/6 mice (Janvier, Strassbourg, France) and Sprague Dawley rats (Janvier, Strassbourg, France) were isolated, flash frozen in liquid nitrogen and stored at -80°C. The tissues samples were homogenized in tissue lysis buffer (100 mM Tris pH 8, 100 mM NaCl, 0.1 % TX-100, 2 mM EDTA, 1:1000 PMSF (200 mM), 1:100 protease inhibitor cocktail) using a Polytron PT homogenizer. Protein concentrations were determined using the Bio-Rad Protein Assay (Bio-Rad Laboratories, München, Germany).

38 MATERIALS & METHODS

2.6 Biochemical methods

2.6.1 Glutathione S-transferase activity assays GST activities against CDNB were determined using the Glutathione S-transferase Assay Kit (Sigma, Saint Louis, USA) according to the manufacturers protocol. Activities against ethacrynic acid, EPNP, pNBC, pNPA and 7-Chloro-4-nitrobenzeno-2-oxa-1,3-diazol were measured as described by Habig et al. (1974) and Widersten et al. (1995) and are summarized in Table 2.11. To analyze thiol transferase activity the assay using 2- Hydroxyethyldisulphide was performed as described by Axelsson et al. (1978). All assays were performed at room temperature in a reaction volume of 200 μl using flat bottom, 96- well PS Microplates (Greiner bio-one, USA) for measurements over 350 nm and UV-star microplates (Greiner bio-one, USA) for measurements below 350 nm. All assays were tested for the recombinant proteins GDAP1X288 and GDAP1X318 (2.2.7.8). Both constructs were recombinant expressed in insect SF21 cells and purified via affinity chromatography (Ni-NTA colums) and size exclusion by C. Bieniossek (Institute of Molecular Biology and Biophysics, ETH Zurich) and stored at 4°C in sample buffer (100 mM NaCl, 20 mM HEPES pH 7.5, 1 mM DTT, 1 mM EDTA) at a working concentration of 2.5 mg/ml. As positive control for all GST assays rat liver was prepared as described in section 2.5 and used in a stock solution of 30 mg/ml in sample buffer.

Table 2.11. Assay conditions for spectrophotometric GST activity assays Substrate Company Substrate GSH Buffer Wavelength Extinction mM* mM nm coefficient (stock) mM-1cm-1 1-Chloro-2,4- 0.1 M KPo Sigma 1 2 340 9.6 dinitrobenzene (CDNB) (pH 6.5) p-Nitrophenylacetate 1 KPo Sigma 5 400 8.3 (pNPA) (10) (pH 6.5) p-Nitrobenylchloride 0.25 KPo Fluka 5 310 1.9 (pNBC) (100) (pH 6.5) 0.2 KPo Ethacrynic acid (EA) Sigma 0.25 270 5 (10) (pH 6.5) 7-Chloro-4- 0.2 NaAc nitrobenzeno-2-oxa-1,3- Sigma 0.5 419 14.5 (10) (pH 5.0) diazol 1,2-Epoxy-3-(4- Fisher 2 KPo nitrophenoxy)propane 5 360 0.5 Scientific (100) (pH 6.5) (EPNP) 50mM Tris (pH 7.6), Cumene hydroperoxide Cayman n.d. n.d. 340 6.22 5mM EDTA 50mM Tris (pH 7.6), t-buthyl hydroperoxide Fluka n.d. n.d. 340 6.22 5mM EDTA *Substrate stocks are prepared in ethanol. KPo, potassium phosphate; NaAc, sodium acetate; n.d., not declared in the assay kit.A

39 MATERIALS & METHODS

2.6.2 Glutathione peroxidase activity assays Glutathione peroxidase activity against cumene hydroperoxide and t-buthyl hydroperoxide were determined using the Glutathione Peroxidase Assay Kit (Cayman Chemical Company, USA) according to the manufacturers protocol (Table 2.11). The constructs used are described in 2.6.1. The decrease of absorbance at 340 nm was measured at room temperature using a photometer (2.1.2)

2.6.3 Detection of ROS Detection of ROS was performed using the Image-iT Live Green Reactive Oxygen Species Detection Kit (Invitrogen, Carlsbad, USA). N1E-115 cells treated with 20 μM menadione and/or 1 mM N-acetyl-L-cysteine (NAC) for 2h were washed once in 37°C warm HBSS+, + subsequently incubated with 15 μM carboxy-H2DCFDA (diluted in HBSS ) and immediately analyzed by epifluorescence microscopy (2.7.2).

2.6.4 Measurement of protein carbonyls Cells grown on 9.5 cm2 6-well plates (NUNC, USA) were treated with 20 μM menadione for 2 h, washed with PBS and lysed in 200 μl ice-cold lysis buffer (20 mM Tris pH 7.4, 140 mM NaCl, 2 mM EDTA, 1 % Triton X-100, 50 mM DTT, 1:100 proteinase inhibitor cocktail). Protein carbonyls were measured using the Oxyblot Protein Oxidation Detection Kit (Chemicon International, Temecula, USA). 10 μl of the cell lysate were incubated at room temperature with 10 μl 12 % SDS and 20 μl DNP-solution. After 15 min 15 μl neutralization-solution was added and 20 μl of the reaction were analyzed by SDS-page and Western blot. Blots were quantified using ImageJ and normalized against GAPDH.

2.6.5 Measurement of the mitochondrial membrane potential

To label mitochondria MitoTrackerH2XRos (Molecular probes, Oregon, USA) was used following the manufactures recommendations. The MitoTracker was added to the medium in a final concentration of 0.5 μM for 30 min. Cells were incubated in culture medium without dye for 10 min before fixation (2.7.2). To determine the relative uptake of

MitoTrackerH2XRos we analyzed single plane confocal pictures using ImageJ. The relative fluorescence (RF) intensity of MitoTrackerH2XRos per area was measured. For the GDAP1 disease mutants for each picture the RF/area of a transfected cell was divided by the RF/area of an untransfected cell. The average and standard deviation of at least 15 pictures were determined using a two-tailed unpaired t-test. The RF/area for the GDAP1 deletion mutants was normalized against cells transfected with the deletion mutant and treated with 10 μM CCCP to avoid intensity artifacts due to the aggregation of mitochondrial induced by those mutants.

40 MATERIALS & METHODS

2.7 Immunochemical methods

2.7.1 Immunoblot analysis Protein samples were first separated by SDS-PAGE and transferred on a nitrocellulose membrane. To block unspecific binding sites the membrane was incubated in 5 ml 10 % low fat milk powder in TBS (blocking buffer) for 1 h at room temperature. The specific primary antibody was diluted according to Table 2.12 in 2 ml blocking buffer and the membrane was incubated with the antibody dilution for 2 h (room temperature) or over night (4°C). After washing the membrane 3x in TBS with 0.05 % Tween20 and 2x in TBS the membrane was incubated with 2 ml of the secondary antibody dilution (Table 2.13) in blocking buffer for 1 h at room temperature. The membrane was washed as described before. For detection with alkaline phosphatase the membrane was additionally washed with AP buffer (110 mM Tris pH 9.5, 100 mM NaCl). Alkaline phosphatase conjugates were detected with CDP-Star (Roche, Indianapolis, USA), horseradish peroxidase conjugates were detected with ECL Western Blotting Detection Reagents (Amersham Biosciences, NJ, USA) using Fuji Medical X-ray films (Fujifilm, Sint-Niklaas, Belgium).

Table 2.12. Primary antibodies for western blot analysis 1. Antibody Species Dilution Company

Calnexin rabbit polyclonal 1:1000 A. Helenius, ETH Zurich Drp1 (Dlp1) mouse monoclonal 1:1000 BD Biosciences FLAG (M2) mouse monoclonal 1:1000 Sigma GAPDH mouse monoclonal 1:20000 HyTest GDAP1(1.3) rabbit polyclonal 1:5000 Pineda GDAP1(3.2) rabbit polyclonal 1:5000 Pineda GDAP1L1 2c-2 rabbit polyclonal 1:5000 Pineda HA rat monoclonal 1:1000 Roche Diagn. GmbH hFis1 rabbit polyclonal 1:500 Alexis Biochemicals PDI mouse monoclonal 1:500 StressGen Porin mouse monoclonal 1:1000 Calbiochem
-actin mouse monoclonal 1:1000 Sigma Cytochrome C mouse monoclonal 1:1000 BD Biosciences OPA 1 mouse monoclonal 1:1000 Becton Dickinson GFAP mouse monoclonal 1:1000 Chemicon Neurofilament160 mouse monoclonal 1:1000 Sigma Matrin rabbit polyclonal 1:5000 J. Senderek, ETH Zurich
-III-tubulin mouse monoclonal 1:400 -tubulin mouse monoclonal 1:1000 Sigma Myc rabbit monoclonal 1:1000 Roche

41 MATERIALS & METHODS

Table 2.13. Secondary antibodies for western blot analysis 2. Antibody Conjugate Species Dilution Company

anti-mouse Alkaline Phosphatase goat 1:10000 Jackson Laboratories anti-mouse Horseradish Peroxidase goat 1:20000 Jackson Laboratories anti-rabbit Alkaline Phosphatase goat 1:10000 Jackson Laboratories anti-rabbit Horseradish Peroxidase goat 1:20000 Jackson Laboratories anti-rat Alkaline Phosphatase goat 1:10000 Jackson Laboratories

2.7.2 Immunocytochemistry Primary antibodies for immunofluorescence are listed in Table 2.14, secondary antibodies in Table 2.15. Cells were fixed at the indicated time points. Immunofluorescence procedures were performed as described previously (Niemann et al., 2005). To label mitochondria, cells were incubated with MitoTracker Red (Molecular Probes) prior to fixation according to the manufacturer’s recommendations. Cells were observed either with a Zeiss Axioplan microscope equipped for epifluorescence and a Zeiss MRM camera or with a confocal inverted microscope (Zeiss LSM 520-NLO) using argon and helium- neon lasers. All images were imported into Photoshop CS (Adobe) for pseudo-coloring, merging, cropping and linear contrast adjustment. To determine the percentage of colocalization of the GDAP1 signal with the mitochondrial marker MitoTracker, we analyzed single plane confocal images using the IMARIS colocalization tool (Bitplane AG). To reduce background signals for the quantitative analysis, the thresholds for the colocalization studies were set at “30” for the green channel and “40” for the red channel as recommended by the software. 85% of the GDAP1 signal colocalized with the MitoTracker signal. For better comparison, this value was set to 100%. The average and the standard deviation of three experiments with 15 pictures per condition were determined, and statistical significance assessed with a two-tailed unpaired t-test. For quantification of the mitochondrial morphology, 450 to 600 transfected cells were counted and categorized into five distinct mitochondrial morphologies: Aggregated, tubular, mixed, vesicular and fragmented. For clarity reason, only the percentage of fragmented mitochondria is shown. Results are shown as average and standard deviation of the percentage of cells with fragmented mitochondrial morphology from three independent transfection experiments. Statistical significance: Two-tailed unpaired t-test. Error bars: Standard deviation; *P<0.05; **P<0.01.

42 MATERIALS & METHODS

Table 2.14. List of primary antibodies used for immunofluorescence 1. Antibody Species Dilution Dilution Company Cells Tissue Cytochrome C mouse monoclonal 1:1000 BD Biosciences GDAP1(1.3) rabbit 3 1:1000 Pineda GDAP1(3.2) rabbit 2 1:1000 1:500 Pineda GDAP1L1 2c-1 rabbit 1:1000 Pineda GDAP1L1 2c-3 rabbit 1:500 Pineda Porin mouse monoclonal 1:500 Calbiochem Giantin mouse monoclonal 1:8000 HP. Hauri, Uni Basel HA rat monoclonal 1:200 Roche Diagn. GmbH Myc rabbit monoclonal 1:1000 Roche PDI mouse monoclonal 1:200 A. Helenius, ETH Zurich
-tubulin mouse monoclonal 1:1000 Sigma FLAG (M2) mouse monoclonal 1:1000 Sigma GFAP mouse monoclonal 1:100 Chemicon Neurofilament160 mouse monoclonal 1:200 Sigma

Table 2.15. Secondary antibodies used in immunofluorescence 2. Antibody Species Dilution Dilution Company Cells Tissue Alexa488 anti rabbit goat 1:500 Molecular Probes Alexa488 anti mouse goat 1:500 1:200 Molecular Probes Cy3 anti rabbit goat 1:500 1:200 Jackson Laboratories Cy2 anti rat goat 1:500 Jackson Laboratories Cy3 anti rat goat 1:500 Jackson Laboratories Cy3 anti mouse goat 1:500 Jackson Laboratories Cy5 anti rabbit goat 1:500 Jackson Laboratories Biotinylated anti rabbit donkey 1:100 Jackson Laboratories Cy3 Streptavidin-biotin anti mouse 1:100 Jackson Laboratories

2.7.3 Immunohistochemistry Cryo sections of mouse brain were provided by H. Pohl (Institute of Cell Biology, ETH Zürich, Switzerland). Paraffin embedded rat brain was provided by M. Rüegg (Institute of Cell Biology, ETH Zürich, Switzerland). 5 μm paraffin sections were prepared using a Rotary Microtome HM 330. Sections were rehydrated as follows: three times 5 min in xylene, three times 5 min in ethanol absolute, 5 min in 70 % ethanol and two times 5 min in ddH2O. Rehydrated sections were incubated with 0.2 % Triton X-100/PBS for 10 min. After washing with PBS the sections were incubated with 10 % goat serum/PBS (blocking

43 MATERIALS & METHODS buffer) for 1 h at room temperature followed by incubation with primary antibodies (Table 2.14) in blocking buffer overnight at 4°C. After washing in PBS samples were exposed to appropriate secondary antibodies (Table 2.15) for 1 h at room temperature, incubated with DAPI (0.2 μg/ml in 70 % ethanol; Roche, USA) and mounted in Immu-Mount (ThermoScientific, Pittsburgh, USA). Cryo sections were dried for 1 h at room temperature, fixed for 10 min in 4 % PFA/PBS and further processed as described for paraffin sections. Microscopy was performed as described in section 2.7.2.

2.8 Bioinformatic analysis

Table 2.16 lists all URLs and programs used in this study.

Table 2.16. URL/programs used in this study URL name/Program URL/ Company

PubMed http://www.ncbi.nlm.nih.gov/entrez/query.fcgi Blast Search http://www.ncbi.nlm.nih.gov/BLAST http://hits.isb-sib.ch/cgi-bin/PFSCAN Domain Prediction http://smart.embl-heidelberg.de Subcellular Localization Prediction http://psort.nibb.ac.jb/form2m.htl Adobe Photoshop CS3 for Macintosh (Adobe Systems Image Processing Incorporated, San Jose, USA) Sequence Analysis CodonCode Aligner (Codon Code Coorperation, Dedham, USA) Immunostaining Quantification IMARIS colocalization tool (Bitplane AG) QuantityOne
(Biorad, Hercules, USA) Western Blot Quantification ImageJ64 (NIH, USA)

44 RESULTS

3. Results

3.1 Analysis of recessive and dominant GDAP1 disease mutations

3.1.1 In contrast to the majority of recessive disease-causing mutations, dominant mutations induce mitochondrial fragmentation The study of GDAP1 mutations leading to peripheral neuropathies might yield important information about specific structural features that are required for proper function of the protein. Niemann et al. (2005) tested a subset of five recessive disease mutations (rmGDAP1’s) for fission activity. Each rmGDAP1 showed some impairment in the ability to induce fragmentation, albeit to different degrees. To complete this picture I further analyzed the fission activity of five additional rmGDAP1s (M116T, S130C, D149Y, L239F and G271R; Fig. 3.1A). In agreement with our previous data D149Y, L239F and G271R displayed a variable reduction of fission activity (Fig. 3.1B g-i; C). Interestingly, M116T and S130C retained full fission activity. The “missing” fission phenotype of S130C can be explained by a second gene defect in the same patient. A mutation in the 3’splice site of the GDAP1 gene (Kabzinska et al., 2005) could be the cause of the CMT disease. In addition to the known rmGDAP1’s, recent studies showed that certain CMT-causing GDAP1 mutations are inherited in an autosomal dominant mode (R120W, T157P, Q218E; Fig. 3.1A). The autosomal dominantly inherited mutations (dmGDAP1’s) showed a milder CMT phenotype compared to rmGDAP1’s (Chung et al., 2008; Claramunt et al., 2005). Analysis of the fission activity revealed no impairment in the ability to induce mitochondrial fragmentation in COS-7 cells (a cell line without endogenous GDAP1 expression) or SH-Sy5y cells (a cell line that endogenously expresses GDAP1; Fig. 3.1B j- l; C). I conclude that (1) the analysis of additional rmGDAP1s confirms the findings of Niemann et al. (2005) in that rmGDAP1’s are defective in the ability to induce mitochondrial fragementation. (2) In contrast, dmGDAP1’s can phenotypically (visually observed fragmentation) not be distinguished from wildtype GDAP1.

3.1.2 GDAP1 and disease mutant-induced mitochondrial fission is blocked by co- expression of dominant-negative Drp1 Previous experiments have shown that the overexpression of dmGDAP1’s induces the same amount of fragmented mitochondria as wildtype GDAP1. However, it is not clear whether this fragmentation results from increased mitochondrial fission or reduced mitochondrial fusion. Thus in a first approach I tested whether the induction of

45 RESULTS

Figure 3.1. Recessive and dominant inherited GDAP1 disease mutations differ in their ability to induce mitochondrial fission. (A) Schematic representation of the GDAP1 protein with recently characterized recessive (red) and dominant (blue) disease mutations. (B) Transfected COS-7 cells were counterstained with MitoTracker prior to fixation and stained against GDAP1. Representative confocal pictures of controls (a-c), GDAP1 (d-f), D149Y (g-i) and T157P (j-l) are shown. Bars, 10 µm. (C) The mitochondrial morphology was quantified for all depicted disease mutations. The effect of dominant mutations in COS-7 cells was analysed via double transfections with wildtype GDAP1. We counted at least 500 cells per condition. The cells were grouped into five distinct mitochondrial morphologies. (D) The expression levels of the tested GDAP1 constructs were comparable to GDAP1 wildtype in COS-7 and SH-Sy5y cells (quantified as ratio GDAP1/Actin signals in cell lysates of sister plates).

fragmentation by GDAP1 and its disease mutants can be counterbalanced by proteins that block fission. Dynamin-related protein 1 (Drp1) plays a central role in the fission process of mitochondria. However, Drp1 overexpression alone does not alter the overall

46 RESULTS

Figure 3.2. Fission induced by GDAP1 and GDAP1 disease mutations is blocked by Drp1 mut overexpression. (A) Cos-7 cells were transiently co-transfected with expression constructs coding for GDAP1 or mutated forms of GDAP1 (recessive mode of inheritance R120Q, R310Q, dominant mode R120W, T157P) and myc-tagged Drp1 (f-j) or a mutated form of Drp1 (k-o) that blocks mitochondrial fission, or with an empty vector (pcDNA, a-e). Cells were fixed 24 h post transfection and stained against GDAP1 (green fluorescence) and myc (red fluorescence). Figure f-o shows the overlay of both signals to illustrate the mitochondrial morphology in co-transfected cells. (B) The mitochondrial morphology was quantified in 300 transfected cells (n=3; 100 cells per experiment). The graph represents the average percentage of cells with fragmented mitochondrial morphology and the standard error. Drp1 does not alter the fission activity of GDAP1 or tested mutants, while the dominant-negative mutant of Drp1 blocks mitochodrial fission induced by GDAP1 and all tested mutants. Stars indicated the significant difference induced by Drp1 mut coexpression compared to pcDNA control co-transfected cells (Student’s t-Test p-value *<0.05). Bars, 10 µm.

mitochondrial morphology (Yoon, 2004). A mutated form, dominant-negative Drp1 (Drp1mut) is known to block mitochondrial fission (Smirnova et al., 1998) and thus shifts the mitochondrial morphology to more elongated mitochondria. Our group has previously shown that the co-expression of GDAP1 and Drp1 does not influence GDAP1-induced fission, while GDAP1 co-expression with Drp1mut blocked the mitochondrial fission triggered by GDAP1 (Niemann et al., 2005). Here I co-expressed dmGDAP1’s and rmGDAP1’s with Drp1 or Drp1mut and analyzed again the mitochondrial morphology. Western blot analysis confirmed comparable expression levels of all transfected constructs

47 RESULTS

(not shown). Comparable to wildtype GDAP1, the co-expression of rmGDAP1’s and dmGDAP1’s with Drp1 did not significantly influence GDAP1-induced mitochondrial fragmentation (Fig. 3.2A f-j; B). In contrast, the co-expression of Drp1mut significantly inhibited the fragmentation of mitochondria (Fig. 3.2A k-o; B). This is again independent from the mode of inheritance. The amount of fragmented mitochondria for R310Q is not significantly reduced by co-expression with Drp1mut as the disease mutation R310Q per se has already lost the GDAP1 typical fragmentation activity (Niemann et al., 2005); Fig. 3.2A c,h,m). In summary, the mitochondrial morphology is not altered upon co-expression of wildtype GDAP1, rmGDAP1’s or dmGDAP1’s with Drp1 while Drp1mut blocks GDAP1 and mutant-induced mitochondrial fragmentation.

3.1.3 The expression of dmGDAP1’s impairs mitochondrial fusion In the cell, continuous fission and fusion events determine the mitochondrial morphology, and increasing mitochondrial fission results in more fragmented mitochondria as demonstrated for GDAP1 overexpression (Fig. 3.3A a). I could show that the induction of fragmentation by GDAP1 and its mutants is inhibited by proteins that block fission (3.1.2). Here I tested whether the fission-induction of GDAP1 and its mutated forms can be counterbalanced by co-expression of the fusion factors mitofusin (Mfn) 1 or 2. As controls I used mutated, inactive forms of Mfn1 and 2, and an empty vector control. I confirmed that GDAP1 expression levels and the expression of the mitofusins are identical in the different combinations tested (Niemann et al., 2009, submitted; not shown). As we have shown previously, GDAP1-induced fission can be counterbalanced by co-expression of Mfn1 or Mfn2, but not with the inactive mutant mitofusin proteins (Niemann et al., 2005; Fig. 3.3A a, f, k; B). Moreover, the fragmentation induced by the rmGDAP1 R120Q can be reverted by coexpression of Mfn1 or Mfn2, but not by the inactive mutated mitofusin variants (Fig. 3.3A b, g, l; B; not shown). The rmGDAP1 R310Q does not significantly induce mitochondrial fission compared to pcDNA3.1 control transfected cells (Fig. 3.3B). Thus no change in the amount of fragmented mitochondria was observed upon coexpression of Mfn1 and 2 (Fig. 3.3A c, h, m; B). However, under these conditions the percentage of cells with highly fused mitochondria increased, which confirms the increased fusion caused by overexpression of Mfn1 or Mfn2 (not shown). The overexpression of the dmGDAP1 in COS-7 cells leads to mitochondrial fragmentation similar or even more extensive than seen with GDAP1 overexpression itself (Fig. 3.3A d, e; B). With these mutants, co-expression of Mfn1 or Mfn2 does not change the mitochondrial morphology. These results suggest that the overexpression of dmGDAP1 impairs the mitochondrial fusion process. To validate this, A. Niemann performed polyethylene glycol (PEG)-based cell fusions (Niemann et al., 2009, submitted) to assess fusion of differentially labeled mitochondria.

48 RESULTS

The fusion was not altered if GDAP1 or rmGDAP1’s were co-expressed with mtEGFP and mtDsRed. However, co-expression of the dmGDAP1’s R120W or T157P (not shown) reduced the fusion of mitochondria from mtDsRed and mtEGFP labeled cells. We conclude that the expression of GDAP1 or rmGDAP1 does not interfere with mitochondrial fusion, whereas dmGDAP1’s impair normal mitochondrial fusion.

Figure 3.3. Fission induced by dominantly inherited mutated forms of GDAP1 cannot be counterbalanced by mitofusin 1 or 2 overexpression. (A) COS-7 cells were transiently co-transfected with expression constructs coding for GDAP1 or mutated forms of GDAP1 (recessive mode of inheritance R120Q, R310Q, dominant mode R120W, T157P) and myc-tagged mitofusin 1 (f-j) or 2, or mutated forms of mitofusin 1 (k-o) and 2 inactive for fusion, or with an empty vector (pcDNA, a-e). Cells were fixed 24 h post transfection and stained against GDAP1 (green fluorescence) and myc (red fluorescence). Figure a-o shows the overlay of both signals to illustrate the mitochondrial morphology in co-transfected cells. (B) The mitochondrial morphology was quantified in 300 transfected cells (n=3; 100 cells per experiment). The graph represents the average percentage of cells with a fragmented mitochondrial morphology and the standard error. GDAP1 and R120Q-induced fragmentation can be counterbalanced by Mfn1 or Mfn2 co-expression. R310Q does not induce fission and is therefore not altered upon mitofusion co-expression. The expression of dominantly inherited mutated forms of GDAP1 cannot be counterbalanced by Mfn1 or Mfn2 co-expression. Stars indicated the significant difference induced by mitofusin coexpression compared to pcDNA control co-transfected cells (Student’s t-Test p-value *<0.05). Bars, 10 µm.

49 RESULTS

Expression of dmGDAP1’s leads to damaged mitochondria Mitochondrial fusion is impaired if the mitochondrial transmembrane potential (ΔΨm) is lost or strongly reduced (Kim et al., 2007; Mattenberger et al., 2003; Meeusen et al., 2004; Twig et al., 2008a). I therefore tested if the expression of GDAP1, rmGDAP1’s or dmGDAP1’s disturbs ΔΨm. The uptake of the dye MitoTrackerH2XRos into mitochondria is dependent on mitochondrial activity. Overall dye uptake was unchanged in cells expressing GDAP1 or any mutated form of GDAP1 tested, compared with nontransfected HeLa cells (Fig. 3.4). However, higher magnification images revealed that the dye labeling was sometimes uneven in cells that

Figure 3.4. Dominantly inherited mutated forms of GDAP1 impair the mitochondrial transmembrane potential. (A) HeLa cells were transiently transfected with GDAP1 (c-f), autosomal recessive inherited (R120Q, R310Q (g-h)) and dominant inherited (R120W, T157P (i-l)) forms of GDAP1. 24 h post transfection cells were incubated with MitoTracker H2XROS prior fixation and stained against GDAP1. Treatment with the protonophore cccp demonstrates that the MitoTracker uptake is dependent on an active mitochondrial transmembrane potential (b). The relative fluorescence intensity of MitoTracker in transfected cells (dashed green area) was identical to the fluorescence intensity in untransfected cells of the same image (dashed blue area). (B) The observation shown on the representative confocal images (a–d and g-j) was quantified using ImageJ (NIH) on 15 independent confocal pictures (total 20 to 28 transfected cells) from three independent experiments (average and standard error). Dashed white boxes in c and i indicate the area of the blow-ups shown in e, f and k, l, respectively. Arrows in k and l point to GDAP1-positive structures which are not labeled by Mitotracker. Bars, 10 µm.

50 RESULTS overexpressed dmGDAP1. We detected GDAP1-positive structures with a stronger labeling (not shown) and those that were only weakly labeled by the dye (Fig. 3.4A k, l). This uneven labeling was detected with all dmGDAP1’s, but not in all cells. Moreover it was never observed in GDAP1 or rmGDAP1 expressing cells (Fig. 3.4A e, f). Impaired mitochondrial fusion can account for the heterogeneous
m observed in dmGDAP1 expressing cells. Thus, we asked whether the fragmentation that is induced by overexpression of dmGDAP1’s also induced other damage. A. Niemann could show that compared to transfection with wildtype GDAP1 and rmGDAP1’s, cells expressing dmGDAP1’s have higher ROS levels. This further confirms that mitochondria are damaged in cells that express dmGDAP1’s. In summary, the expression of dmGDAP1’s leads to impaired mitochondrial fusion accompanied by mild damage to mitochondria.

51 RESULTS

3.2 GDAP1 structural analysis

3.2.1 Analysis of the overall GDAP1 domain structure Truncations of GDAP1 interfere with its mitochondrial targeting. Furthermore, missense mutations of the protein, although still mitochondrial, have lost their fission inducing activity. Despite this, further information regarding specific domains such as the GST domains or structural features crucial for GDAP1 fission activity, targeting and function is scarce. To link observed functions to specific protein regions, I constructed various GDAP1 deletion mutants (Fig. 3.5A). Transfection of COS-7 cells revealed that (1) apart from the deletion of the transmembrane domain (TMD) all deletion mutants are still mitochondrial (Fig. 3.5B) and (2) only deletions of the N- and C-terminal part were able to induce mitochondrial fragmentation comparable to that observed with wildtype GDAP1 (Fig. 3.5B d-f, C). All other deletion mutants appear to lack fission-inducing activity and lead to a perinuclear aggregation of mitochondria (Fig. 3.5C). The expression levels of the tested GDAP1 deletion mutants were comparable to the GDAP1 wildtype protein in transfected COS-7 cells (not shown). To assess if the extensive mitochondrial aggregation causes (or is secondary due to) a loss of

m I next analyzed

m. As before I used the dye MitoTrackerH2XRos, whose uptake into mitochondria is dependent on mitochondrial activity (Chen and Cushion, 1994). Niemann et al. (2005) showed previously that transient transfection with an GDAP1 expression construct does not change the dye uptake and thus does not influence the

m. Dye uptake was also unchanged in the tested GDAP1 deletion mutants (Fig. 3.5D). The protonophore cccp is known to abolish

m and was used as control for reduced

m. As expected, HeLa cells treated with cccp displayed a reduced uptake of MitoTrackerH2XRos

(Fig. 3.5 D a,b). To exclude that the observed MitoTrackerH2XRos-labeling in cells overexpressing a deletion mutant is an artifact due to the aggregated mitochondrial phenotype, I repeated the assay but treated the deletion mutant-expressing cells with cccp (Fig. 3.5D d,f,h,j,l,n,p). In all conditions tested the dye uptake was strongly reduced. This indicates that the observed dye uptake in not-cccp treated, deletion mutant-overexpressing cells is due to an active

m.

3.2.2 Detailed analysis of the GDAP1 tail-anchor As the generation and analysis of various domain deletions of GDAP1 proved to be inconclusive (3.2.1) I continued with a detailed analysis of specific structural features of GDAP1 starting at the C-terminus of the protein. Similar to mitofusins, GDAP1 contains two hydrophobic stretches at the C-terminus and is an integral MOM protein with the more C-terminal hydrophobic domain sufficient for mitochondrial targeting (Niemann et al.,

52 RESULTS

2005). The following experiments aimed at determining the precise topology of GDAP1, the motifs involved in MOM targeting and integration, and the relationship between TA structure and the fission function of the protein.

Figure 3.5. GDAP1 deletion mutants display impaired fission activity. (A) Schematic representation of the GDAP1 deletion mutants. Numbers indicate the first and last deleted amino acid. (B) COS-7 cells transiently transfected with wildtype GDAP1 (a-c), delN- terminus (d-f; C) and delC-terminus (C) display a predominantly fragmented mitochondrial morphology while delGST-N (g-i), delInterdomain (j-l) and delHD1 (m-o) display a highly aggregated mitochondrial architecture. Bars, 10 µm. (C) The influence on the mitochondrial architecture was quantified for all deletion mutants. Between 500 to 600 cells per condition from three independend experiments were counted. The cells were grouped into five mitochondrial classifications. (D) The mitochondrial uptake of the dye MitoTrackerH2XRos into HeLa cells is dependent on an active mitochondrial membrane potential (a) and can be blocked with the protonophore cccp (b). HeLa cells were transfected with the various GDAP1 deletion mutants for 24 h and labeled with

MitoTrackerH2XRos (c-p’). To avoid intensity artefacts due to highly aggregated mitochondria in most of the deletion mutants a sisterplate was transfected with the corresponding construct and treated with cccp to monitor a loss of the membrane potential in cells with accumulated mitochondria (d,f,h,j,l,n,p). No difference between untransfected cells, cells expressing GDAP1 and cells expressing deletion mutants was found. Bars, 10 µm.

53 RESULTS

3.2.2.1 GDAP1 is a tail-anchored protein with a single transmembrane domain To determine the topology of GDAP1, I required an appropriate epitope-tagged version of the protein. Large C-terminal tags severely impair the posttranslational translocation of ER-targeted TA proteins (Brambillasca et al., 2006). Similarly, I found that C-terminal tags that significantly extend GDAP1, for example EGFP, cause extensive mitochondrial aggregation and a significant loss of mitochondrial localization compared to untagged GDAP1 (Fig. 3.6A g-i). Using the fluorescence protease protection assay established by Lorenz et al. (2006) I could further show that such EGFP-tagged proteins attach only peripherally to mitochondria (Fig. 3.6B). In this assay, the cell membrane of COS-7 cells was initially permeabilized with low digitonin concentrations that do not affect organelle membrane integrity (Lorenz et al., 2006). Cytosolic proteins including the cytosolic fraction of GDAP1-EGFP are released (Fig. 3.6B c); only the GDAP1-EGFP associated with mitochondria remains detectable but is lost over time of the protease digest (Fig. 3.6B d). The mitochondrial marker mtDsRED is not washed out or degraded by the protease (Fig. 3.6B e-g). These results indicate that GDAP1-EGFP is partially associated with mitochondria but the long-extended C-terminus fails to translocate across the MOM. To avoid these artifacts, I exchanged eight C-terminal amino acids (aa) of GDAP1 for the short FLAG-tag (DYKDDDDK) to maintain the length of the C-terminus (Fig. 3.7C). As previously shown (Fig. 3.5) deletion of the C-terminal aa does not influence the

Figure 3.6. GDAP1-EGFP is peripherally attached to the MOM. (A) COS-7 cells were transiently transfected with either mtEGFP (a-c), wt GDAP1 (d-f), or the C-terminal EGFP-tagged construct GDAP1-EGFP (g-i). Fifteen hours after start of transfection, cells were co-stained with MitoTracker to analyze mitochondrial localization. GDAP1-EGFP shows only partial mitochondrial localization. In addition, the expression of GDAP1-EGFP causes mitochondrial aggregation. (B) Upper panel: Cartoon of a cell before and after permeabilization of the cell membrane with low concentrations of digitonin showing the release of cytosolic proteins (a, a’). Subsequent treatment with trypsin digests cytosol-exposed membrane-bound protein parts (a’’; yellow dots, cytosolic proteins; blue dots, cytosolic parts of MOM-attached proteins; green dots, proteins of the intermembrane space; red dots, proteins of the matrix). The used digitonin and trypsin concentrations do not affect the mitochondrial membrane integrity (Lorenz et al., Nat. Methods 3, 205-210, 2006). Lower panel (b-g): GDAP1-EGFP and mtDsRED expressing COS-7 cells were permeabilized with 50 µM digitonin and treated in parallel with 250 µM trypsin for the indicated time points. Images were taken after permeabilization and trypsin digest. The cytosolic GDAP1-EGFP signal is washed out early due to permeabilization of the cell (c). Only the GDAP1-EGFP associated with mitochondria remains detectable but is lost over time of the protease digest (b-d). The mitochondrial targeted marker mtDsRED is not washed out or degraded by the protease (e-g). These results indicate that GDAP1-EGFP partly associates with mitochondria but the long-extended C-terminus fails to translocate across the MOM. Bars, 10 µm.

54 RESULTS fragmentation activity of GDAP1. HeLa cells were transfected with the GDAP1-FLAG construct and mitochondria isolated after 24 hours, digested with proteinase K in the presence or absence of the detergent digitonin, followed by Western blot analysis (Fig. 3.7A; Olichon et al., 2002). Without detergent, the anti-Flag antibody detected a proteinase K-resistant fragment of 5 kDa, consistent with the expected molecular weight of a peptide containing the C-terminus and the more C-terminally located hydrophobic segment of GDAP1 (Fig. 3.7B, arrowhead). This fragment and a control protein located in the intermembrane space (IMS), OPA1, were digested upon membrane solubilization by increasing digitonin concentrations. Comparable results were obtained with a C-terminally FLAG-tagged full-length GDAP1 (not shown). Together with our previous results demonstrating that the more C-terminally located hydrophobic domain and its immediate flanking regions are sufficient for GDAP1 targeting to mitochondria (Niemann et al., 2005), these findings establish that GDAP1 integrates into the MOM as a classical TA protein with a single transmembrane domain and the C-terminus located in the IMS. Consequently, I have renamed the most C-terminal hydrophobic segment of GDAP1 to transmembrane domain (TMD) and the other hydrophobic portion to hydrophobic domain 1 (HD1).

Figure 3.7. GDAP1 membrane topology. (A) Graphic of experimental strategy. (B) Western blot of mitochondria-enriched fractions derived from GDAP1-FLAG-transfected HeLa cells treated with proteinase K for 30 minutes at 4°C with increasing digitonin concentrations. Without digitonin, the cytosolic N-terminal part of GDAP1-FLAG is degraded (arrow) and only the C-terminal 5 kDa fragment is detected (arrowhead). After membrane permeabilization, the C-terminal fragment is gradually degraded confirming its location in the IMS. The IMS protein Opa1 served as control. (C) Confocal immunofluorescence microscopy of transfected COS-7 cells expressing GDAP1-FLAG (a-d). Co-stainings: MitoTracker (mitochondria). Bars, 10 µm.

55 RESULTS

3.2.2.2 The classical tail-anchored proteins GDAP1 and Fis1 but not Mfn2 integrate into isolated mitochondria-enriched membrane preparations To test whether GDAP1 can integrate into membranes like a classical TA-protein, I established an integration assay similar to assays published by Henderson et al. (Henderson et al., 2007) and Setoguchi et al. (Setoguchi et al., 2006). I added radiolabeled in vitro translated GDAP1, Fis1, Mfn2, or luciferase to the post-nuclear supernatant of Hela cells (Fig. 3.8A). After an hour of incubation at 4°C, the soluble fraction (S1) and a mitochondrial-enriched fraction (P1) were separated by centrifugation. Mfn2 and luciferase remained in S1. In contrast, GDAP1 and Fis1 co-sedimented with the mitochondrial marker porin in two serial centrifugation steps (P1, P2). Radiolabeled GDAP1 and Fis1 integrated with an efficiency of 25 ± 2% and 28 ± 2% (s.d.), respectively, under the given conditions, and remained membrane-bound after the second centrifugation step (P2). To evaluate the role of the cytosol-exposed region of GDAP1 on the integration, I used EGFP fused to HD1 and TMD together (EGFP-HD1-TMD; aa 291-358), or the TMD alone (EGFP-TMD; aa 311-358). Both constructs contain the flanking basic aa and the TMD C-terminal aa (Fig. 3.8). Comparable to full-length GDAP1, the chimeric proteins also integrated efficiently (25% ± 3%) and co-sedimented with porin. Hence, the TMD (including its flanking basic aa and the C-terminus) is sufficient for membrane integration. To rule out that GDAP1 is only peripherally associated with membranes, we repeated the original integration assay but treated P1 with 1 M sodium chloride or 0.1 M carbonate buffer (pH 11; Niemann et al., 2005). Consistent with membrane integration, GDAP1 could be sedimented again in the second centrifugation step. GDAP1 was only released into the supernatant upon treatment with detergent, comparable to what was observed for the integral mitochondrial membrane protein porin. EGFP-HD1- TMD and EGFP-TMD behaved identically to the GDAP1 full-length protein in this assay (Fig. 3.8B). In contrast, although the GDAP1-EGFP fusion protein can be co-sedimented with the mitochondrial-enriched fractions P1 and P2, it is released into the supernatant upon treatment with high salt or carbonate (Fig. 3.8C). This supports the previous observation (Fig. 3.7) that the C-terminal EGFP-fusion interferes with the integration into the MOM. Finally, to confirm the GDAP1 membrane integration in our in vitro assay, I performed a proteinase K digest of membrane-integrated radiolabeled GDAP1-FLAG. Upon immuno-precipitation with the anti-FLAG antibody the full-length GDAP1-FLAG was detected in the control reaction without protease. In the presence of proteinase K I precipitated a proteinase K-resistant fragment of 5 kDa (Fig. 3.8D). This fragment represents the membrane-protected C-terminal region of GDAP1-FLAG confirming the proper integration of in vitro translated GDAP1-FLAG into membranes, comparable to the integration shown in Figure 3.7. Without membrane integration GDAP1-FLAG is

56 RESULTS degraded by proteinase K and no proteinase resistant fragment can be precipitated (Fig. 3.8D). Taken together, these results support the suggested topology of GDAP1 in the MOM as a classical TA-protein with a single TMD. This is backed up by the following additional evidence: First, GDAP1 integrated in membranes like a classical TA-protein, comparable to Fis1 in our in vitro assay. Second, in contrast, the two TMD-containing MOM protein Mfn2 did not integrate. Third, the TMD of GDAP1 was sufficient to mediate integration.

Figure 3.8. Integration assay of in vitro-synthesized TA-proteins. (A) The post-nuclear supernatant of HeLa cells was incubated with the in vitro-translates of GDAP1, Fis1, Mfn2, Luciferase, EGFP-HD1-TMD or EGFP-TMD. Mitochondria were enriched in a differential centrifugation approach (S1, P1, S2, P2). GDAP1, Fis1, EGFP-HD1-TMD and EGFP-TMD showed co-sedimentation with the mitochondrial marker porin. Mfn2 and Luciferase remained in the supernatant. (B) The post-nuclear supernatant of HeLa cells was incubated with the in vitro-translated GDAP1, Fis1, EGFP-HD1-TMD and EGFP-TMD and the mitochondrial pellet was resuspended in buffer (control), in 1 M NaCl, 0.1 M carbonate (pH 11), or in buffer with 0.1 % TritonX-100, and centrifuged to separate the soluble protein supernatants (S) from membranous pellets (P). All in vitro-translated proteins and the integral membrane protein porin could be sedimented and became soluble only upon treatment with detergent. Unlike the integral membrane proteins cytochrome C can only be sedimented under control conditions. (C) In vitro translated GDAP1-GFP was processed as described for (B). Upon treatment with high salt or carbonate GDAP1-GFP was extracted as was the intermembrane space protein cytochrome C. (D) Immuno-precipitation of in vitro translated GDAP1-FLAG with an anti-FLAG antibody without and after membrane integration and with or without proteinase K (50 mg/ml) digest. The upper arrow points to the undigested immuno-precipitated full length GDAP1-FLAG. The lower arrows indicates the 5 kD fragment still detected after membrane integration and protease digest.

57 RESULTS

3.2.2.3 Positively charged amino acids surrounding the TMD determine GDAP1 mitochondrial targeting and function Basic aa flanking the TMD are essential for MOM targeting of classical TA proteins, whereas neutral or acidic residues at the C-terminus of TA proteins lead to a predominantly ER localization (Borgese et al., 2003; Borgese et al., 2001; Kuroda et al., 1998). Since both hydrophobic domains of GDAP1 are flanked by basic aa, I examined their effect on mitochondrial localization by systematic exchanges to neutral aa, either singly or in combination (Fig. 3.9A). Subcellular localization was analyzed in transiently transfected COS-7 cells using MitoTracker and anti-PDI (for ER) as organelle markers on single plane confocal images. MtEGFP control-transfected and wildtype (wt) GDAP1 transfected COS-7 cells showed a typical mitochondrial pattern (Fig. 3.9B, a-h). ER localization distinguishable from mitochondrial staining, especially focusing on the periphery of cells, was not detected. Single exchanges of basic aa either N- or C-terminal of the TMD did not alter the strict mitochondrial localization (Fig. 3.9B, i-l; q-t, 10C), nor did the exchange of two basic residues N-terminal of the TMD (GDAP1[K314A R315A], Fig.

3.9C) or flanking HD1 (GDAP1[K291A R310Q], Fig. 3.9C). Only alteration of three basic aa N- terminal of the TMD led to mistargeting to the ER (GDAP1[K313A K314A R315A], Fig. 3.9B m-p; 3.9C). In contrast, the exchange of two residues C-terminal or flanking the TMD

(GDAP1[R341A R342A]; GDAP1[K314A R342A]) already led to significant loss of mitochondrial localization (Fig. 3.9C). This effect was even more pronounced in the triple mutant

GDAP1[R341A R342A R343A] C-terminal of the TMD (Fig. 3.9B u-x; 3.9C). I also observed that the mitochondrial morphology was altered compared to wt GDAP1 when some mutants were expressed (Fig. 3.9B, D). This was not due to different expression levels as determined by Western blots (Wagner et al., 2009). Quantification revealed that all mutant proteins with aa exchanges between HD1 and the TMD showed a significant reduction in fission activity compared to wt GDAP1, shifting the morphology to more tubular and aggregated mitochondria (Fig. 3.9B i-p). Hence, the mutated aa are critical for both correct targeting and GDAP1-mediated no alteration in mitochondrial morphology and targeting (Fig. 3.9D). Taken together, (1) Mitochondrial targeting of GDAP1 is not critically dependent on individual basic aa within the TA; (2) C-terminally TMD-flanking clusters of basic residues play a key role in targeting; (3) A cluster of basic aa N-terminally bordering the TMD is also required for mitochondrial localization; strikingly, even individual residues of this cluster are concurrently crucial for GDAP1 function; (4) The residues immediately adjoining HD1 N- or C-terminally (K291, R310) are not required for targeting, but R310 is required for fission. Interestingly the mutation R310Q was found in patients with CMT (Niemann et al., 2005).

58 RESULTS

Figure 3.9. TA-associated critical basic residues for GDAP1 localization and function. (A) Schema of GDAP1 and aa sequence of C-terminal part of wt GDAP1 (GDAP1) and mutants (bold letters: altered aa sequence). (B) Confocal immunofluorescence microscopy of transfected COS-7 cells expressing mtEGFP as control (a-d) wt GDAP1 (e-h) and mutant proteins (i-x). Co-stainings: MitoTracker (mitochondria), PDI (ER). Bars, 10 µm. (C) Quantification of GDAP1 and mutants colocalized with MitoTracker. (D) Quantification of mitochondrial morphology.

59 RESULTS

3.2.2.4 The correctly ordered amino acid sequence of HD1, but not of the TMD, is critical for mitochondrial fission activity Next, I examined the role of the aa sequences of the TMD and HD1 on mitochondrial targeting and fission activity. To this end I constructed GDAP1 mutants with scrambled TMD and HD1 (TMDscr, HD1scr) and a mutant lacking HD1 (Fig. 3.10A). TMD deletion causes loss of mitochondrial targeting and activity (Niemann et al., 2005). TMDscr strictly colocalized with mitochondria (Fig. 3.10B a-c; Wagner et al., 2009) and displayed full fission activity (Fig. 3.10C) indicating that the native aa sequence is not critical. Scrambling or deletion of HD1 (Fig. 3.10A) did not interfere with mitochondrial targeting (Wagner et al., 2009), but extensive mitochondrial aggregation and tubulation was observed (Fig. 3.10B d-i). Expression of both of these constructs leads to a significant reduction of cells with fragmented mitochondrial morphology compared to wt GDAP1 (Fig. 3.10C). Thus, both the presence of HD1 and its correct aa sequence are essential for GDAP1 fission activity. MOM targeting of TA proteins has been suggested to depend on short TMDs (Borgese et al., 2001; Isenmann et al., 1998). GDAP1-variants with different TMD lengths (minus 1, plus 1, 3 or 5 aa) but unchanged mean hydrophobicity are still targeted to the mitochondria (Fig. 3.11). The variant with five additional aa is slightly but significantly mislocalized to the ER. None of these proteins displayed detectable alterations in fission activity. Thus, the TMD length influences the specificity of MOM targeting but as sufficient protein appears to be still targeted to the mitochondria, GDAP1 fission activity is not impaired.

Figure 3.10. Role of TMD and HD1 amino acid sequences in GDAP1-mediated fission. (A) Schematic representation of the aa sequence of the TMD and HD1 of wt GDAP1 (GDAP1) and mutants TMDscr, HD1scr and HD1del. (B) Distinct mitochondrial localization of mutant proteins in transfected COS-7 cells (a-i). (C) Quantification of mitochondrial morphology. Fission activity was lost by HD1 scrambling or deletion (n = 500). Bars, 10 µm.

60 RESULTS

Figure 3.11. Effect of TMD length on mitochondrial targeting and fission activity. (A) Schema of GDAP1 TMD aa sequence and constructs with altered TMD length. TMD hydrophobicities are given in brackets. (B) Confocal immunofluorescence analysis of transfected COS-7 cells reveals mitochondrial targeting for all recombinant proteins (a-l), albeit for TMD+5 with reduced efficiency (g- i). (C) Quantification of mitochondrial localization. Significant mislocalization was detected for TMD+5. (D) Analysis of fragmentation-inducing activity of mutants revealed no significant difference compared to wt GDAP1 (GDAP1; n=500). Bars, 10 µm.

3.2.2.5 The specific nature of the GDAP1 TMD and its C-terminus is not required for fission activity I next asked whether the particular TMD and the flanking GDAP1 C-terminus are critical for fission-inducing activity. Thus, I constructed a chimeric protein in which the TMD and the C-terminus of GDAP1 are replaced by that of VAMP1B, a tail-anchored MOM protein with no known function in mitochondrial fission (Isenmann et al., 1998). The chimeric protein was correctly targeted to mitochondria (Fig. 3.12B a-c; Wagner et al., 2009) and retained full fission activity (Fig. 3.12C). I conclude that there is no functional relevance of the primary sequence of the GDAP1 TMD and its neighboring C-terminus in GDAP1- mediated mitochondrial fission.

Figure 3.12. Function of TA in GDAP1 fission activity. (A) Construction of chimera GDAP1-VAMP1B. (B) COS-7 cells transfected with GDAP1-VAMP1B display mitochondrial localization and fragmented mitochondria comparable to full-length GDAP1 (a-c). (C) Quantification reveals no difference in fission activity between GDAP1 and GDAP1-Vamp1B (n=500). Bars, 10 µm.

61 RESULTS

3.2.2.6 TMD hydrophobicity does not influence GDAP1-induced fission of mitochondria Wattenberg et al. (Wattenberg et al., 2007) showed previously that the hydrophobic/hydrophilic balance is crucial for the targeting of TA proteins. However, so far nothing is known about whether the TA hydrophobicity also influences the proper function of TA proteins. Hence, I tested whether TMD hydrophobicity has an impact on GDAP1 fission function. Computer analysis revealed that TA mitochondrial dynamics factors can be divided into one group with high TMD hydrophobicity and one with limited hydrophobicity (Table 3.1; Brambillasca et al., 2006). Both GDAP1 and VAMP1B contain a strongly hydrophobic TMD. Thus, the previously observed full fission-inducing activity of GDAP1-VAMP1B and TMDscr might be attributed to similar TMD hydrophobicities. To test whether a less hydrophobic TMD would have an effect, I constructed the chimera GDAP1-OMb5 containing the less hydrophobic TA of mitochondrial cytochrome b5 (Table 3.1). This chimera was only partially targeted to mitochondria, inducing reduced levels of mitochondrial fragmentation (Fig. 3.13B) and with significant proportion of GDAP1-OMb5 directed to the ER (Fig. 3.13A a-c). Control experiments revealed a comparable ER mislocalization for wt hOMb5 and ratOMb5 in our system rendering this approach suggestive but inconclusive (not shown). Hence, I generated GDAP1hy, a GDAP1 variant with four TMD aa exchanged for less hydrophobic aa (Table 3.1). GDAP1hy colocalized exclusively with MitoTracker (Fig. 3.13A d-f), confirming that a less hydrophobic TMD still retains MOM specificity (Brambillasca et al., 2006) and full mitochondrial fission activity (Fig. 3.13B). I conclude that GDAP1 fission activity is not influenced by its TMD hydrophobicity.

Figure 3.13. Influence of TMD hydrophobicity on GDAP1 fission activity. (A) COS-7 cells expressing either GDAP1-OMb5 (a- c) or GDAP1hy (d-f) were co-stained with MitoTracker. Both proteins show mitochondrial localization, with some mislocalization of GDAP1-OMb5 to the ER. (B) Quantification of mitochondrial morphology (n=500). Bars, 10 µm.

62 RESULTS

Table 3.1. Characteristics of TMD’s of selected MOM TA proteins and mutants thereof.

Total hydrophobicity: Sum of values of TMD between basic aa (in bold), according to GES hydrophobicity scale (Engelman et al., 1986). Italics indicate aa between flanking basic aa and the longest hydrophobic stretch (not included in the calculation). TMDs used here, in blue; mutants generated, in grey. *protein is not a classical TA protein.

3.2.3 Detailed analysis of the GST features of GDAP1

3.2.3.1 Analysis of the putative catalytic centre of GDAP1 The protein encoded by GDAP1 shows high sequence similarities to GST’s. Detailed bioinformatic approaches revealed that GDAP1 is related to cytosolic Zeta, Omega and Theta GST classes (Alves et al., 2006; Frova, 2006; Marco et al., 2004; Shield et al., 2006; Tan et al., 1996; Winayanuwattikun and Ketterman, 2005). However, so far no evidence of GST-activity or glutathione-binding ability for bacterial expressed recombinant protein was found. To obtain information about whether the putative GST-activity of GDAP1 is related to its fission inducing activity, I introduced point mutations into the highly conserved amino acids (aa) of the GST-N domain, which is the potential glutathione binding domain (Y29A, S34A, S36A, S37A, F68M; Fig. 3.14A). It is well known that in classical GST’s a Tyr or Ser at the N-terminal part of the predicted first α-helix of the GST-N domain constitute a key part of the catalytic centre. Consistent with this, Shield et al. performed in silico modeling experiments and proposed Ser34 as the conserved active site residue of GDAP1 (Shield et al., 2006). Furthermore, a highly conserved Phe N- terminal of the first beta-strand of the GST-N domain (corresponding to Phe68 of GDAP1) is also known to be of crucial functional relevance. This residue constitutes a central part of the lock-and-key motif necessary for GST dimerization and catalytic activity. Transfection of COS-7 cells revealed that all GST-N domain related point mutations still display mitochondrial localization (Fig. 3.14B) but differ in their ability to induce mitochondrial fragmentation. All tested mutations of the GST-N domain lost the ability to induce mitochondrial fission compared to wildtype GDAP1. In contrast, the mitochondrial

63 RESULTS

Figure 3.14. GST-N domain related GDAP1 point mutants display a loss of fission activity. (A) Schematic representation of the catalytic centre mutants of GDAP1. (B) COS-7 cells transiently transfected with wildtype GDAP1 (a-c) display a predominantly fragmented mitochondrial morphology while the mitochondrial morphology of the depicted mutations of the catalytic centre shifts to a more elongated and aggregated mitochondrial architecture (d-l). Bars, 10 µm. (C) The influence on the mitochondrial architecture was quantified for all point mutants. 500 cells per condition from three independend experiments were counted. The cells were grouped into five mitochondrial classifications. All tested point mutants lost their ability to induce mitochondrial fragmentation (D) The

mitochondrial uptake of the dye MitoTrackerH2XRos into HeLa cells is dependend on an active mitochondrial membrane potential (a) and can be blocked with the protonophore cccp (b). HeLa cells were transfected with the catalytic centre mutations of GDAP1 for 24

h and labeled with MitoTrackerH2XRos (c-q). To avoid intensity artefacts due to highly aggregated mitochondria in most of the point mutants sisterplates were transfected with the corresponding construct and treated with cccp to monitor a loss of the membrane potential in cells with accumulated mitochondria (e,h,k,n,q). No difference between untransfected cells, cells expressing GDAP1 and cells expressing the mutations in the GST-N domain was found. Bars, 10 µm. morphology shifted to a more elongated and aggregated architecture (Fig. 3.14B d-l, C). This shift was especially evident for the mutation S34A (Fig. 3.14C) which was proposed to be the active side residue in the GST-N domain of GDAP1 (Shield et al., 2006). The

64 RESULTS expression levels of the tested GDAP1 point mutants were comparable to that of the GDAP1 wildtype protein in transfected COS-7 cells (not shown). Mitochondrial fusion is impaired if the mitochondrial transmembrane potential (∆ψm) is lost or strongly reduced (Kim et al., 2007; Mattenberger et al., 2003; Meeusen et al., 2004; Twig et al., 2008a). I therefore asked whether the expression of these fission-inactive point mutations causes a loss of ∆ ψm. The uptake of the dye MitoTrackerH2XRos into mitochondria is dependent on an active ∆ψm and can be blocked by the protonophore cccp (Fig. 3.14D a,b). Overall dye uptake was essentially unchanged in cells expressing GDAP1 (previously shown by Niemann et al. (2005)) or any mutated form of GDAP1 tested, compared to untransfected HeLa cells (Fig. 3.14D). To avoid intensity artifacts due to extremely aggregated mitochondria I additionally examined dye uptake in mutant transfected cells treated with cccp (Fig. 3.14D d,f,h,j,l). All point mutant expressing-cells lost their ∆ψm upon cccp treatment, indicating that the observed fluorescence of

MitoTrackerH2XRos in mutant overexpressing cells is due to an active ∆ψm.

3.2.3.2 GDAP1 is an active glutathione S- transferase Exchange of key aa of the proposed catalytic centre of GDAP1 completely abolishes the fission activity of GDAP1. One could speculate that the fission inducing activity is related to an unknown GST activity of GDAP1. To be catalytically active, GST-enzymes must form dimers (Hayes et al., 2005). The ability of GDAP1 to form homodimers was confirmed by Shield et al. (2006) via cross linking experiments, and by M. Rüegg using various immunoprecipitation approaches (Rüegg, 2009). However, so far no evidence for glutathione (GSH) binding or GST activity of bacterial expressed recombinant protein has been found (Shield et al., 2006). This could be due to its untypical GST structure – its two hydrophobic domains at the C-terminus and its localization at the MOM – resulting in too little GDAP1 being obtained from bacterial systems, wrong folding of the protein or missing of correct posttranslational modifications. To address this question I changed to a eukaryotic baculovirus based expression system using SF21 cells. Two different variants of GDAP1, GDAP1 288X lacking both C-terminal hydrophobic domains, and GDAP1 318X lacking only the TMD, were recombinantly produced by C. Bieniossek (ETH Zürich) and I. Berger (EMBL Grenoble; Fig. 3.15). First we analyzed the ability of GDAP1 to bind GSH, a prerequisite for GST activity. Therefore, we tested both constructs for their ability to bind GSH-sepharose. Neither of the Figure 3.15. GDAP1 expression constructs for recombinant expression. Schematic representation of the GDAP1 protein constructs truncated proteins was able to bind GDAP1 318X and GDAP1 288X used for recombinant protein expression in insect cells for use in the GST assays. either in a common “pull down” or in

65 RESULTS pull-down experiments combined with cross-linking (not shown). This outcome was not unexpected as in silico modeling approaches revealed a close relationship to Theta-class GST’s (Shield et al., 2006), which do not bind to GSH affinity matrices (Sheehan et al., 2001). Furthermore, I performed direct binding studies using radioactive 35S-labeled GSH in combination with a pulldown of recombinant GDAP1 318X and GDAP1 288X. The radioactivity of the input (recombinant GDAP1 incubated with 35S-labeled GSH), all intermediate steps, and the elution fraction were measured using a scintillation counter. While I eluted only low amounts of radioactivity with GDAP1 318X compared to control conditions (pulldown without GDAP1) we eluted around 2.5 times more radioactivity with GDAP1 X288 compared to the control (one representative experiment is shown in Table 3.2). However, compared to the radioactivity detected in the input fraction the eluted radioactivity was extremely low. To allow the direct comparison of all performed experiments, an appropriate protein control (such as GDAP1
GST-N) for normalization of the data needs to be included into the assay. First trials to express recombinant GDAP1
GST-N were not successful and would need to be further optimized.

Table 3.2. 35S-labeled GSH binding assay control GDAP1 288X GDAP1 318X Input (cpm) 47.776 46.645 40.668 final wash (cpm) 12 14 15 Elution (cpm) 111 253 130 Representative radioactivity countings detected in the GSH-binding assay. Cpm, counts per minute.

Finally, I performed GST activity assays to detect GSH-dependent enzyme activities using substrates that have previously been shown to be substrates for GST enzymes (Table 3.3). In contrast to previous assays performed with the bacterial expressed protein, I found significant specific activities of GDAP1 X288 with the substrates 1,2-Epoxy-3-(4- nitrophenonxy) propane (EPNP), p-Nitrobenzylchloride (pNBC) and ethacrynic acid (EA). For the model substrate 1-Chloro-2,4-dinitrobenzene (CDNB) I found only very low specific activities. However, Theta class GST enzymes are known to lack activity against CDNB (Sheehan et al., 2001). No activity was found for GDAP1 318X, the construct containing the HD1. This is in consistence with my GSH-binding data (Table 3.2). On a highly speculative note, this might be due to an autoinhibitory mechanism triggerd by HD1 via posttranslational modifications, such as HD1 phosphorylation-dependent activation/ inhibition of the GST activity. CMT-causing mutations of GDAP1 are likely to yield important information about the newly identified GST activity of GDAP1 and its relevance to the disease phenotype. Thus, I cloned various GDAP1 mutations as short (GDAP1 288X; Fig. 3.16A) or long (GDAP1 318X; Fig. 3.16B) version in the insect cell expression vector. As negative control I

66 RESULTS additionally cloned the proposed catalytic centre mutations S34A (Fig. 3.16 green). Recombinant expression of the disease mutations was performend by C. Bieniossek (ETH Zürich). However so far limited protein was ultimately purified, suggesting a folding deficit of these mutant proteins. Further optimization of the purification process of the disease mutant forms will be required to obtain sufficient material for GST activity assays.

Table 3.3. GST activities of GDAP1 288X and GDAP1 318X Substrate Specific activity (nmol/min per mg of protein) GDAP1 288X (# assays) GDAP1 318X (# assays) 1-Chloro-2,4-dinitrobenzene (CDNB) 11 ± 4 (12) n.d. (10) Cumene hydroperoxide n.d. (18) n.d. (12) t-Buthyl hydroperoxide n.d. (8) n.d. (8) 2-Hydroxyethylsulphide n.d. (8) n.d. (8) Ethacrynic acid (EA) 26 ± 2 (8) n.d. (8) p-Nitrobenzylchloride (pNBC) 96 ± 3 (8) n.d. (8) p-Nitrophenylacetate (pNPA) n.d. (8) n.d. (8) 7-Chloro-4-nitrobenzeno-2-oxa-1,3-diazol n.d. (8) n.d. (8) 1,2-Epoxy-3-(4-nitrophenoxy) propane (EPNP) 820 ± 116 (10) n.d. (8) Data are shown as means ± standard deviation; n.d. indicates where specific activity was <10.

3.2.4 Three-dimensional structure of GDAP1 For in-depth understanding of the structure-function relationship of GDAP1 and its functional role as GST protein, it is necessary to reveal the three-dimensional structure of GDAP1. Recombinant expression of the GDAP1 constructs, crystallization, and X-ray analysis are in progress, in collaboration with C. Bieniossek (ETH Zürich, Switzerland) and I. Berger (EMBL Grenoble, France). In a first approach we used the two GDAP1 constructs GDAP1 288X and GDAP1 318X (Fig. 3.16). Recombinant soluble protein was obtained via the baculovirus based system MultiBac using insect cells and purified. Gelfiltration assays showed that the variant containing HD1 (GDAP1 318X) exists as dimer and monomer, and that both forms dynamically interconvert. This further confirms that GDAP1 is able to form dimers, generally a prerequisite for GST activity (3.2.3). Circular dichroism spectra of both constructs Figure 3.16. Recombinant expression constructs of GDAP1 disease mutations. Schematic representation of the GDAP1 disease mutations indicate a lack of unstructured regions. introduced into the recombinant truncated GDAP1 constructs GDAP1 318X and GDAP1 288X. Green, proposed key amino acid of the Thus both variants are potential catalytic centre as negative control; Red, tested recessive disease mutations; Blue, tested dominant disease mutations. candidates for crystallization.

67 RESULTS

However, first crystallization experiments with both constructs yielded no crystals. Consequently, we continued with two new strategies: (1) Cloning of more stable, Edman degradation and mass- spectrometry verified versions of the GDAP1 constructs lacking the N-terminal region (GDAP1 17-309 and GDAP1 17-288; Fig. 3.17A) and (2) Cloning of GDAP1 expression constructs N-terminally fused to a maltose binding protein (MBP) to facilitate crystallization (Fig. 3.17B). So far, first crystals were obtained with MBP- GDAP1 318L. However, further analysis of these crystals showed that they are do not diffract, and thus are not usable for determination of the three dimensional structure via X- ray analysis.

Figure 3.17. Recombinant GDAP1 expression constructs for X-ray analysis. Schematic representation of the recombinant GDAP1 constructs. (A) Mass spectrometry and Edman degradation verified constructs. (B) GDAP1 expression constructs N-terminally fused to a maltose binding protein (MBP).

68 RESULTS

3.3 GDAP1L1

Although GDAP1 is expressed in the peripheral nervous system (PNS) and in the central nervous system (CNS), GDAP1 disease mutations lead to peripheral neuropathies without severe CNS phenotype (1.4). This implies, that either GDAP1 function is somewhat compensated by another protein in the CNS or that GDAP1 exerts different functions in the CNS compared to the PNS. Here I investigate, whether the GDAP1 paralog GDAP1L1 is able to compensate for dysfunctional GDAP1 in the CNS.

3.3.1 GDAP1L1 is exclusively expressed in the central nervous system It is known that GDAP1 is most prominently expressed in different regions of the CNS and the PNS (Niemann et al., 2005). Western blot analysis of mouse neuronal tissues revealed that GDAP1L1 is expressed in various regions of the CNS including cortex, cerebellum,

Figure 3.18. GDAP1L1 is expressed in the central nervous system. (A) 12 µg of protein lysate from the indicated neuronal tissue was analyzed by Western blot. Except in DRG and sciatic nerve GDAP1L1 was detected in all neuronal tissues at the predicted size of 42 kDa (arrowhead). (B) Endogenous GDAP1L1 was detected in neuronal regions (visualized by the neuron specific marker NeuN) of the CNS on adult rat paraffin sections of cerebellum (a-c), cortex (d-f) and hippocampus (g-i). (C) On paraffin section of adult rat cerebellum specific signals for GDAP1L1 (a-c) and GDAP1 (j-l) were detected in purkinje cells (a-f). Only the signal of GDAP1 colocalizes with the mitochondrial marker porin (m-o), GDAP1L1 displays a cytosolic staining (d-f). The perinuclear ring detected in purkinje cells stained for GDAP1L1 is unspecific as the pre-incubation with the peptide used for immunization (250 µg/ml) only blocks the cytoplasmic but not the perinuclear signal (g-i). The mitochondrial signal of GDAP1 is specific, as peptide blocking abolishes all signals (p- r). Bar, 10 µm.

69 RESULTS thalamus, olfactory bulb and spinal cord. However, no expression of GDAP1L1 was detected in the PNS (Fig. 3.18A). These data are in agreement with the mRNA expression profiling (Angst, 2005). The specificity of the antibody was confirmed by successful blocking of the GDAP1L1 band by pre-incubation of the serum with the antigenic peptide (not shown). Next I analyzed the expression and localization of GDAP1L1 in various tissues of the CNS in detail. I could confirm the Western blot analysis of CNS expression (Fig. 3.18A) using saggital paraffin sections of adult rat cerebellum (Fig. 3.18 B a-c), cortex (d-f) and hippocampus (g-i). Here endogenous GDAP1L1 expression is predominantly found in neuronal regions as visualized by the colocalization with the neuron-specific marker NeuN (Fig. 3.18B). As described in detail in previous sections (3.2), GDAP1 localizes to the mitochondrial outer membrane (MOM; Niemann et al., 2005). Mutations in GDAP1 lead to sensory and motor neuropathies without an obvious disease phenotype in the CNS. To identify whether GDAP1L1 is similarly distributed in neuronal cells I examined the localization of endogenous GDAP1L1 and GDAP1 in cerebellara Purkinjeb cells of rat cbrain saggital sections. Interestingly, GDAP1L1 is not localized to specific cellular structures such as mitochondria, as visualized with the mitochondrial marker protein porin but is rather ubiquitously expressed throughout the cellsd (Fig. 3.18Ce a-f). In contrast,f GDAP1 colocalizes completely with porin, indicating a mitochondrial localization of GDAP1 in Purkinje cells of the CNS (Fig. 3.18C j-o). This is consistent with previous findings that identified GDAP1 as a MOM protein in the PNS by colocalization studies of rat dorsal root ganglia (Niemann et al., 2005). The staining specificity was confirmed by successful blocking of the staining by pre-incubation of the serum with the antigenic peptide (Fig. 3.18C g-i; p-r). I conclude that GDAP1L1 is expressed in the CNS, but not in the PNS, in contrast to GDAP1, which is expressed in both parts of the nervous system. One could speculate that GDAP1L1 expression compensates for mutated GDAP1 versions in the CNS and thus prevents disease phenotypes there. However, first localization studies do not reveal a mitochondrial localization of GDAP1L1.

3.3.2 GDAP1L1 is a cytosolic protein GDAP1 localizes to mitochondria. As described in section 3.2 the protein integrates into the MOM with its C-terminally located TMD constituting the tail-anchor of GDAP1. The MOM localization was confirmed in cells endogenously expressing GDAP1 as SH-Sy5y or N1E-115 cells, cells overexpressing GDAP1 as HeLa or COS-7 cells, and primary neurons (Niemann et al., 2005). The C-terminal domain structure of GDAP1L1 has a high similarity to the corresponding region of GDAP1 with a potential transmembrane domain

70 RESULTS

Figure 3.19. GDAP1L1 is a cytosolic protein. (A) Schematic representation of the GDAP1L1 domain structure. (B) COS-7 cells, 24 h post transfection with GDAP1L1 expression constructs. GDAP1L1 does not colocalize with the Golgi complex marker gigantin (a-c), the ER marker PDI (d-f) or the mitochondrial marker cytochrome C (g-i) on single plane confocal pictures. Bars, 10 µm. (C) Western blot analysis of isolated and cultured primary neurons and glia cells. In contrast to GDAP1 GDAP1L1 is not expressed in cultured cells. Control represents transient transfected COS-7 cells overexpressing either GDAP1 or GDAP1L1 (a). Western blot analysis of 2 day old mouse hippocampus (HiC) and Cortex. As control COS-7 cells transiently overexpressing either GDAP1 or GDAP1L1 was used. In contrast to GDAP1, GDAP1L1 is not expressed in two day old mice. (b).

(TMD) at the outermost C-terminal region and a hydrophobic domain N- terminally flanking the putative TMD (Fig. 3.19A). It seemed likely that, like GDAP1, GDAP1L1 would be a mitochondrial protein. To test this hypothesis I determined the precise subcellular localization of GDAP1L1 in an overexpression system. In COS-7 cells that were transfected with an untagged full-length expression construct I found a predominantly cytosolic distribution (Fig. 3.19B). To exclude a localization to membranous compartments including the Golgi complex (marker gigantin; Fig. 3.19B a-c), the ER (marker protein disulfide isomerase (PDI); Fig. 3.19B e- f) or mitochondria (marker cytochrome C; Fig. 3.19B g-i) I performed colocalization studies. No appreciable overlap was found with any of the tested membraneous structures.

To exclude potential localization artifacts due to overexpression, I next examined the endogenous GDAP1L1 expression in various cell lines. In all cell lines tested, no endogenous GDAP1L1 expression was detected by Western blot analysis (Angst, 2005). Thus, to further support the GDAP1L1 intracellular localization studies I tested cultures of primary mouse hippocampal and cortical neurons and glial cells for endogenous expression of GDAP1L1. No specific GDAP1L1 staining could be detected in any of the primary celltypes tested. In contrast, GDAP1 endogenous expression was nicely detected in primary neurons of mouse hippocampus (Fig. 3.19C a). As Western blot control I used lysates of COS-7 cells overexpressing either GDAP1L1 or GDAP1 full-length protein. Subsequent Western blot analysis of mouse hippocampus and cortex at the developmental stage P2 revealed, that in contrast to GDAP1, which is expressed at P2 (Fig. 3.19C b),

71 RESULTS

GDAP1L1 is only expressed in adult mice (Fig. 3.18A). Hence, primary neurons prepared from P0 mice brain cannot be used as endogenous system for GDAP1L1 subcellular localization studies. To summarize, despite the structural similarity of the TMD region of GDAP1 and GDAP1L1, the latter protein does not localize to membranous compartments but is rather ubiquitously distributed throughout the cell.

3.3.3 The putative TMD of GDAP1L1 is able to integrate into the MOM As described in section 3.2.2 GDAP1 is a classical tail-anchored protein. Basic amino acids (aa) flanking the TMD constitute a key part in targeting to the MOM (3.2.2 and Wagner et al., 2009). Since a computer-aided comparison of the tail-anchor region of GDAP1L1 and GDAP1 (Fig. 3.20 A) revealed identical targeting motifs in the primary sequence of the two proteins, I further examined the putative MOM integration properties of GDAP1L1. In a first approach I tested whether GDAP1L1 is able to integrate into membranous preparations in vitro using our integration assay (3.2.2.2 and Wagner et al., 2009; Fig. 3.20B). In contrast to GDAP1, which co-sedimented with the mitochondrial marker protein porin and remained membrane-bound after the second centrifugation step (P2), GDAP1L1 remained in supernatant S1 (Fig. 3.20C). Bax, another well known tail-anchored protein of the MOM is distributed throughout the cytosol under control conditions. Only upon a certain apoptotic stimulus does Bax undergoe a conformational change and integrates into the MOM (Borgese et al., 2003). To evaluate the integration properties of the GDAP1L1 putative transmembrane and hydrophobic domains independent from the conformational state of the full-length protein, I used EGFP fused to HD2, HD1 and TMD (EGFP-HD2-TMD), HD1 and TMD (EGFP- HD1-TMD) or the TMD alone (EGFP-TMD). All constructs contain the flanking basic aa and the TMD C-terminal residues whose analogs in GDAP1 are crucial for targeting (3.2.2 and Wagner et al., 2009). Interestingly, unlike full-length GDAP1L1, all three chimeric proteins integrated efficiently and co-sedimented with porin (Fig. 3.20D a). Hence, the TMD alone (including its flanking basic residues and the C-terminus) alone is sufficient for membrane integration. To rule out that the EGFP-tagged constructs are only peripherally associated with the membranes, I repeated the original integration assay (3.2.2.2) but treated mitochondrial pellet P1 with 1 M sodium chloride or 0.1 M carbonate buffer (pH 11; Niemann et al., 2005). Consistent with membrane integration, the chimeric constructs could be sedimented again in the second centrifugation step in yields comparable to those obtained for the MOM integral protein porin (Fig. 3.20D b). The constructs were only released into the supernatant upon treatment with detergent. Finally, as our assay is performed with mitochondrial-enriched fractions that cannot be perfectly separated from other membranous structures such as ER, I performed

72 RESULTS colocalization studies of the EGFP-tagged GDAP1L1 constructs with the mitochondrial marker MitoTracker. As previously shown mtEGFP (Fig. 3.20E a) and GDAP1 (Fig.3.20E b) display a typical mitochondrial pattern (Wagner et al., 2009), while full-length GDAP1L1 (Fig. 3.20E c) is ubiquitously found throughout the cytosol (Fig. 3.19). As expected, all EGFP-tagged hydrophobic domain variants strictly colocalize with mitochondria (Fig. 3.20E d-l). Taken together, these results suggest that the C-terminal hydrophobic domains have the potential to anchor GDAP1L1 to mitochondria, although the full-length protein does not display mitochondrial localization.

73 RESULTS

Figure 3.20. Integration assay of in vitro-synthesized GDAP1L1 constructs. (A) Schematic representation of the tail-anchored region of GDAP1L1 compared to GDAP1. (B) Scheme of the membrane integration assay using in vitro-synthesized proteins. The post-nuclear supernatant of HeLa cells was incubated with the in vitro-translates of relevant proteins. Mitochondria were enriched in a differential centrifugation approach (S1, P1, S2 and P2). (C) Integration assay of GDAP1L1 and GDAP1 as control. In contrast to GDAP1, GDAP1L1 showed no co-sedimentation with the mitochondrial marker porin but remained in the supernatant. (D a) Integration assay of EGFP-TMD, EGFP-HD1-TMD and EGFP-HD2-TMD. All tested constructs were co-sedimented comparable to porin. (b) The post-nuclear supernatant of HeLa cells was incubated with the in vitro-translated EGFP-tagged GDAP1L1 constructs. The mitochondrial pellet was resuspendend in buffer (control), 1 M NaCl, 0.1 M carbonate (pH 11) or in buffer containing 0.1 % TritonX-100 and centrifuged to separate the soluble protein supernatants (S) from the membranous pellets (P). All in vitro-translated proteins and the integral membrane protein porin could be sedimented and became soluble upon treatment with detergent. (E) Displayed are COS-7 cells, 24 h post transfection with various constructs. mtEGFP (a) and GDAP1 (b) display a typical mitochondrial pattern while GDAP1L1 (c) is distributed throughout the cytosol. EGFP-tagged GDAP1L1 hydrophobic domain constructs (d-l) colocalize with the mitochondrial marker MitoTracker on single plane confocal pictures. Bars, 10 µm.

74 RESULTS

3.3.4 The TA of GDAP1 targets GDAP1L1 to mitochondria I next asked whether the particular TMD and the flanking GDAP1L1 C-terminus are sufficient to integrate full-length proteins into the MOM and thus act as a classical TA. The protein GDAP1 provides a mean to test this. I could previously show that the TA of GDAP1 can be exchanged with TA’s of different TA-MOM proteins without impairing its fission-inducing function (Wagner et al., 2009). Thus I constructed a chimeric protein in which the TMD including the C-terminus of GDAP1L1 was fused to GDAP1 and thereby replacing the TA of GDAP1 (G-L1(TMD); Fig. 3.21A). The chimeric protein G-L1(TMD) was correctly targeted to mitochondria (Fig. 3.21B d-f) and was still able to induce mitochondrial fragmentation albeit slightly reduced compared to wildtype GDAP1 (Fig. 3.21C). GDAP1L1 even though predominantly cytosolic causes a slight increase in the amount of fragmented mitochondria compared to control transfected cells. However,

Figure 3.21. The TA of GDAP1 localizes GDAP1L1 to mitochondria. (A) Schematic representation of the exchange of the TA-domain of GDAP1 and GDAP1L1. (B) COS-7 cells transfected with mtEGFP (a), wildtype GDAP1 (b), wildtype GDAP1L1 (c) and the chimera G-L1(TMD) (d-f) and L1-G(TMD) (g-i). Except GDAP1L1 all constructs colocalize with the mitochondrial marker MitoTracker. Bars, 10 µm. (C) Quantification of the mitochondrial morphology. The amount of fragmented mitochondria induced by G-L1(TMD) is slightly reduced to wildtype GDAP1. GDAP1L1 induced fission is significantly lower compared to GDAP1 and the amount of fragmented mitochondria of L1-G(TMD) is not altered compared to control levels. (D) The post-nuclear supernatant of HeLa cells was incubated with the in vitro-translated chimeric constructs. The mitochondrial pellet was resuspendend in buffer (control), 1 M NaCl, 0.1 M carbonate (pH 11) or in buffer containing 0.1 % TritonX-100 and centrifuged to separate the soluble protein supernatants (S) form the membranous pellets (P). L1-G(TMD) and the integral membrane protein porin could be sedimented and became soluble upon treatment with detergent indicating proper membrane integration. In contrast, GDAP1 containing the TMD of GDAP1L1 becomes already soluble upon treatment with carbonate.

75 RESULTS fragmentation induced by wildtype GDAP1 was still significantly higher (∼40 % GDAP1 vs. ∼15 % GDAP1L1). Furthermore I used this approach to analyze the behavior of GDAP1L1 if artificially anchored to the MOM by replacing its putative TMD with the well-characterized TA of GDAP1 (Fig. 3.21A). As expected, the so created construct L1- G(TMD) is also efficiently targeted to mitochondria (Fig. 3.21B g-i). However, the artificial mitochondrial targeting does not increase the number of cells with a predominantly fragmented mitochondrial morphology (Fig. 3.21C). To analyze the mode of localization of the chimeric constructs to mitochondria, peripheral vs. integral, I performed an alkaline and high salt wash as described in Figure 3.20. While GDAP1L1 containing the TA of GDAP1 co-sediments with the membranous pellets and becomes only soluble upon treatment with detergent, the still fission active G-L1(TMD) is already released into the supernatant upon treatment with carbonate. This finding was unexpected as I could previously show that a loss of targeting results in a loss of GDAP1 fission activity (Wagner et al., 2009). In summary, GDAP1L1, if artificially localized to mitochondria, does not influence the mitochondrial architecture. Furthermore, the putative TA of GDAP1L1 is able to target GDAP1 to mitochondria without impairing the fission activity of GDAP1.

3.3.5 The HD1 of GDAP1L1 cannot substitute for the HD1 of GDAP1 Previous studies aimed at elucidating the functional role of HD1 for GDAP1 revealed that both the presence of HD1 and the correct primary sequence are essential for GDAP1

Figure 3.22. The HD1 of GDAP1L1 cannot substitute for the function of the GDAP1 HD1. (A) Detailed scheme of HD1 (a) and schematic representation of the constructs with exchanged HD1 (b; letters in bold represent conserved aa). (B) Transfected COS-7 cells 24 h post transfection. G-L1(HD1) displays a distinct mitochondrial localization as visulalized by the colocalization with the mitochondrial marker MitoTracker (a-c). L1-G(HD1) is ubiquitously expressed throughout the cytosol and shows no colocalization with mitochondria (d-f). Bars, 10 µm. (C) Quantification of the mitochondrial morphology. Upon exchange of the HD1 the fission activity was lost in both constructs. 76 RESULTS activity (3.2.2 and Wagner et al., 2009). The HD1 of GDAP1L1 (HD1-L1) shows high sequence similarity to the HD1 of GDAP1 (HD1-G) (Fig. 3.22A a; b) and thus could be involved in similar functional mechanisms. Therefore, I next examined whether HD1-L1 is able to substitute for HD1-G function. I created two constructs representing GDAP1 and GDAP1L1 with exchanged HD1’s. Analysis of the subcellular distribution revealed no effect of the exchanged HD1 on the localization. G-L1(HD1) is comparable to wildtype GDAP1 strictly localized to mitochondria (Fig. 3.22B a-c) while L1-G(HD1) similar to wildtype GDAP1L1 is distributed throughout the cytosol (Fig. 3.22B d-f). However on a functional level both constructs completely lost their fission inducing activity (Fig. 3.22C). To conclude, HD1-L1 is not able to substitute for HD1-G function.

3.3.6 GDAP1L1 is translocated to mitochondria upon treatment with menadione The previous results suggest that GDAP1L1, albeit a cytosolic protein, might be targeted to specific membranes upon an external stimulus. Similar mechanisms are well characterized e.g. for the MOM protein Bax; an apoptotic stimulus leads to a conformational shift that exposes its TA and thus causes an integration of Bax into the MOM (Borgese et al., 2003). To test this hypothesis I exposed N1E-115 cells to various reagents that are known to induce cellular stress. N1E-115 cells, although not expressing GDAP1L1, endogenously express its close relative GDAP1 and thus represent the best system available. To verify that I was monitoring viable cells despite the exposure to oxidative stressors, I also followed subcellular localization of cytochrome C. Cytochrome C is a mitochondrial- specific protein that is released into the cytosol in cells that undergo apoptosis. Actinomycin D, staurosporine, ethacrynic acid, N-ethylmaleimide (not shown), TBHQ (Fig. 3.23A d-e’), TBHP (Fig. 3.23A f-g’), rotenone (Fig. 3.23A h-I’), dithiotreitol (DTT,

Fig. 3.23A j-k’), H2O2 (Fig. 3.23A l-m’) and the protonophore m-chlorophenylhydrazone (cccp, Fig. 3.23A n-o’) did not alter the subcellular localization of GDAP1L1 or GDAP1. In contrast, a 2 h treatment with 20 μM menadione shifted the ubiquitously distributed GDAP1L1 to a distinct localization that bears certain similarities to the corresponding mitochondrial cytochrome C staining pattern. The cytotoxic compount menadione (2- methyl-1,4-naphthoquinone; MQ) has been extensively used as oxidative stressor in various studies (Abdelmohsen et al., 2004; Gant et al., 1988; Watanabe et al., 2004). GDAP1 localization is not altered upon treatment with menadione (Fig. 3.23A p), as it is still predominantly found on mitochondria. I next determined the exact subcellular distribution of GDAP1L1. In N1E-115 cells that were transfected with an untagged full-length GDAP1L1 expression construct and subsequently treated with menadione, I found GDAP1L1 in distinct structures with tubular and vesicular appearance throughout the cytosol (Fig. 3.23B g). Colocalization studies with the mitochondrial marker cytochrome C revealed a significant overlap between

77 RESULTS

GDAP1L1 and cytochrome C positive structures (Fig. 3.23B j-l). No appreciable overlap was found with other membranous compartments, including the Golgi complex visualized by gigantin (Fig. 3.23B a-c) or the ER marker protein PDI (Fig. 3.23B d-f). Taken together these results suggest that the cytosolic protein GDAP1L1 is redistributed to mitochondria upon a specific ROS stress in N1E-115 cells. To further support the redistribution of GDAP1L1 to mitochondria, I used a subcellular fractionation approach. Cytosolic proteins, membranes, and nuclei were separated without

Figure 3.23 .

78 RESULTS mechanical membrane rupture and analyzed by Western blot. However, no redistribution of GDAP1L1 to the membranous fraction was observed with this assay (Fig. 3.23C). Hence, the exact nature of the GDAP1L1 association with the MOM remains unclear requires further clarification

3.3.7 Detailed analysis of the effect of MQ on GDAP1L1 subcellular distribution To assess the redistribution of GDAP1L1 to mitochondria in dependence of the time of menadione treatment and the concentration of MQ used, I performed a time- and concentration course. With 20 µM MQ, GDAP1L1 begins to relocalize to mitochondria at

Figure 3.24. Effect of menadione on GDAP1L1 subcellular distribution. (A) Transfected N1E-115 cells, 24 h after transfection with GDAP1L1 and following mendadione treatment at the indicated time and (B) at the indicated menadione concentration. Cells were costained with a the mitochondrial marker cytochrome C that additionally allows for detection of apoptotic events. The majority of GDAP1L1 is redistributed to mitochondria 2 h after treatment with 20 µm menadione (A p-r; B j-l). Bars, 10 µm.

79 RESULTS around 1 h of treatment (Fig. 3.24A m-o). After 2 h, the majority of GDAP1L1 protein is found to colocalize with mitochondria (Fig. 3.24A p-r). However, after 3 h of MQ treatment the cells begin to undergo apoptosis indicated by a release of cytochrome C into the cytoplasm (Fig. 3.24A t) and thus can no longer be analyzed anymore. To assess the critical MQ concentration crucial for triggering the redistribution of GDAP1L1, I treated N1E-115 cells with concentrations of 0 – 100 µM MQ for 2 h. A small amount of GDAP1L1 colocalizes with cytochrome C at a concentration of 10 µM (Fig. 3.24B g-i), while the major fraction of GDAP1L1 protein is redistributed to mitochondria at a concentration of 20 µM (Fig. 3.24B j-l). Higher MQ concentrations lead to increased cell death, hence the relocalization can no longer be monitored anymore (Fig. 3.24B m-r). To summarize, GDAP1L1 is predominantly localized at mitochondria if cells are treated with 20 µM MQ for 2 h.

3.3.8 Analysis of the naphthoquinone pathway and EA on GDAP1L1 subcellular distribution The putative glutathione S-transferase (GST) GDAP1L1 is redistributed to mitochondria upon MQ-induced ROS stress. The metabolism of quinones such as MQ, glutathione (γ-L- glutamyl-L-cysteine-glycine; GSH) and GST’s is remarkably intertwined (Bellomo et al., 1990; Gant et al., 1988; Watanabe et al., 2004). One major route of quinone elimination is the GST- mediated conjugation of GSH to quinones. The second important elimination pathway is the production

of H2O2 via redox cycling of the quinones. MQ can redox cycle, and it can be conjugated to GSH and thus be marked for export from the cell. In contrast, another well known naphthoquinone, 2,3-dimethyoxy-

Figure 3.25. Analysis of the effect of quinones and EA on the distribution of GDAP1L1. Transfected N1E-115 cells, 24 h post transfection with GDAP1L1 expression constructs and subsequent 2 h treatment with the indicated reagents were analyzed on single plane confocal pictures. 2,3-dimethoxy-1,4- naphtoquinoe (DNMQ, 20 µM) lead to a redistribution of GDAP1L1 to mitochondria as visualized by the colocalization with the mitochondrial marker cytochrome C (a-f, see arrowheads). In contrast, tert-buthylhydroquinone (TBHQ, 100 µM; g-i) and ethacrynic acid (EA, 50 µg/ml; j-l) had no impact on the subcellular distribution of GDAP1L1.Bars, 10 µm.

80 RESULTS

1,4-naphthoquinone (DNMQ) can only redox cycle, while the quinone tert- buthylhydroquinone (TBHQ) acts primarily through conjugation of GSH (Watanabe et al., 2004). Thus, DNMQ and TBHQ provide a means to test whether the subcellular redistribution of GDAP1L1 to mitochondria is predominantly dependent on the intracellular ratio of GSSG/GSH (GSSG – oxidized GSH) or due to the activity of reductase-mediated redox cycling. Interestingly, DNMQ alone was able to trigger the relocalization of GDAP1L1 (Fig. 3.25 a-f) while the treatment with TBHQ had no effect on the cytosolic distribution of GDAP1L1 (Fig. 3.25 g-i). These data suggest, that the GDAP1L1 distribution is predominantly dependent on the redox cycling part of the quinone induced ROS stress.

Ethacrynic acid (EA), an oxidative stress-causing agent and potent inhibitor of the GST family of enzymes, has been shown to affect mitochondria in that it induces mitochondrial fusion in cultured cells (Bowes and Gupta, 2005). Hence I was interested in the effect of EA on the subcellular distribution of GDAP1L1. Treatment of GDAP1L1 transfected HeLa cells with EA did not alter the cytosolic staining pattern of the protein (Fig. 3.25 j-l).

3.3.9 MQ induced redistribution of GDAP1L1 is blocked by NAC

The antioxidant N-acetyl-L-cysteine (NAC), a free radical scavenger, is a precurser of the ROS scavenger GSH (Criddle et al., 2006; Watanabe et al., 2004). High intracellular GSH levels are crucial to maintain a reduced environment. To assess a potential protective effect of NAC on MQ induced redistribution of GDAP1L1, I first examined the ROS levels under our experimental conditions. The membrane permeable probe H2DCFDA was used to measure ROS produced by MQ treatment in live N1E-115 cells. ROS in the cells converted the nonfluorescent dye into fluorescein, which emits green fluorescence. Without MQ treatment, no ROS was detected in control cells and cells expressing GDAP1L1 (Fig. 3.26 A a-d). After 2 h of MQ treatment, ROS production was significantly increased in GDAP1L1 expressing cells (Fig. 3.26A e, f). NAC significantly reduced MQ- induced ROS levels (Fig. 3.26A g-j) and increased the viability of the cells (not shown). To further evaluate the effect of the ROS-scavenging effect of NAC on GDAP1L1 subcellular distribution, I analyzed GDAP1L1 expressing cells untreated, with MQ, and with MQ plus NAC treatment. As previously observed, the cytosolic distribution of GDAP1L1 changed to a distinct mitochondrial pattern upon treatment with MQ (Fig. 3.26B a-f). Interestingly, NAC inhibited the MQ-induced redistribution of GDAP1L1 to mitochondria and instead the cytosolic localization was maintained (Fig. 3.26B g-l). Moreover, NAC treated cells displayed an overall healthy morphology and had no increased cell death rate as observed for only MQ treated cells (not shown).

81 RESULTS

Figure 3.26. NAC rescue of MQ induced ROS. (A) N1E-115 cells were treated with 20 µM menadione (MQ) and indicated concentrations of N- acetyl–L- cysteine (NAC) for 2 h and

stained with H2CFDA. NAC prevents ROS induced by MQ treatment (d, e). Bars, 10 µm. (B) N1E-115 cells 24 h post transfection with GDAP1L1 expression constructs and subsequent 2 h treatment with MQ and indicated concentrations of NAC. In untreated cells GDAP1L1 is distributed throughout the cytosol (a-c). MQ treatment redistributes GDAP1L1 to mitochondria visualized by colocalization with the mitochondrial marker cytochrome C (d-f). Parallel treatment with NAC prevents MQ triggerd redistribution of GDAP1L1 (g-l). Bars, 10 µm. (C) N1E- 115 cells were treated with 20 µM MQ for 2 h. Protein carbonyls in the total cell lysates were derived with 2,4- dinitrophenylhydrazine (DNP) and analyzed by western blot. GAPDH was used as loading control (a). Levels of protein carbonyls were quantified against the amount of GAPDH in the same lysates (b).

To examine, whether MQ increases protein oxidation and to analyze a potential protective effect of GDAP1L1 and GDAP1 I measured protein carbonyls in total cell lysates of GDAP1L1 or GDAP1 transfected N1E-115 cells. As seen in Fig. 3.26C a, more protein carbonyls were formed by MQ treatment. However, no protetive effect of GDAP1L1 or GDAP1 expression on the amount of protein carbonyls could be detected. To quantify the generation of protein carbonyls I normalized the samples against GAPDH. Although MQ treated cells have a slightly increased amount of protein carbonyls, no significant difference was detected (Fig. 3.26C b). Taken together,

82 RESULTS the results suggest, that (1) the ROS-scavenging action of NAC prevents MQ-induced redistribution of GDAP1L1 to mitochondria and (2) expression of GDAP1L1 or GDAP1 or MQ treatment do not significantly alter the levels of oxidized proteins in N1E-115 cells.

83 DISCUSSION

4. Discussion

GDAP1 is a CNS (central nervous system) and PNS (peripheral nervous system) enriched mitochondrial fission factor dependent on the fission factors Fis1 and Drp1 (Niemann et al., 2005; Niemann et al., 2009, submitted). Mutations in GDAP1 lead to demyelinating, axonal and intermediate forms of CMT (Charcot-Marie-Tooth disease), a peripheral motor and sensory neuropathy without severe CNS phenotype (Baxter et al., 2002; Cuesta et al., 2002). Although originally characterized as a recessively inherited disease, recent studies reported CMT-causing GDAP1 mutations that are inherited in an autosomal dominant mode (Chung et al., 2008; Claramunt et al., 2005). Based on its domain structure GDAP1 was proposed to belong to a novel GDAP1 class of GST (glutathione S-transferase) enzymes (Marco et al., 2004; Shield et al., 2006), although recent studies could not detect any GSH (glutathione)-binding or GST activity of bacterial expressed recombinant protein (Pedrola et al., 2005; Shield et al., 2006). Furthermore, GDAP1 contains two hydrophobic domains at the C-terminal region that display features of a TA (tail-anchor) protein of the MOM (mitochondrial outer membrane). In this study I analyzed the different disease- causing mechanisms of the recessive and dominant mutations and the putative GST- activity of GDAP1. On a structural level I characterize GDAP1 as a TA protein of the MOM and provide first experiments for the determination of the three-dimensional structure.

Expression of GDAP1 is found in tissues of the PNS and CNS (Cuesta et al., 2002; Liu et al., 1999; Niemann et al., 2005). This implies that either GDAP1 function is compensated by another protein only expressed in the CNS or that GDAP1 exerts different functions in the CNS compared to the PNS. GDAP1L1 is a closely related paralog of the mitochondrial fission factor GDAP1 and belongs to the same novel GDAP1 class of GST enzymes (Marco et al., 2004). Unlike GDAP1, GDAP1L1 is expressed in the CNS but not in the PNS. This study further aims at determining the potential of GDAP1L1 to compensate for dysfunctional GDAP1 proteins in the CNS.

4.1 Dominant and recessive GDAP1 disease mutations differ in their

mode of action

It is well-established that the nervous system is especially sensitive to mitochondrial dysfunction (Bossy-Wetzel et al., 2003). The extreme dimensions of long peripheral nerves makes them most vulnerable to changes in mitochondrial dynamics and alterations in this highly complex system can lead to neuropathies such as the neurodegenerative disorder

84 DISCUSSION

CMT. Mutations in GDAP1 lead to CMT and are a frequent cause of demyelinating, intermediate, and/or axonal forms of CMT (Baxter et al., 2002; Cuesta et al., 2002; Nelis et al., 2002). GDAP1 is involved in the regulation of mitochondrial dynamics and induces mitochondrial fission (Niemann et al., 2005). To understand how GDAP1 influences the mitochondrial fission activity we analyzed the CMT-causing recessively (rmGDAP1) and dominantly (dmGDAP1) inherited GDAP1 mutations. Niemann et al. (2005) showed that a subset of five tested rmGDAP1’s have lost their ability to induce mitochondrial fragmentation to various degrees. To complete this picture I analyzed the fission activity of five additional rmGDAP1’s (M116T, S130C, D149Y, L239F and G271R). In agreement with previous data from A. Niemann, D149Y, L239F and G271R display a variable reduction in fission activity. The disease mutations M116T and S130C show a slight impairment in their ability to induce mitochondrial fragmentation, although this reduction is not significant (3.1.1). Recently, in addition to the known rmGDAP1’s, some studies identified CMT causing GDAP1 mutations that are inherited in an autosomal dominant mode (R120W, T157P, Q218E; Chung et al., 2008; Claramunt et al., 2005). The dmGDAP1’s lead to a later disease onset and display a much milder CMT phenotype in patients compared to rmGDAP1’s. In contrast to rmGDAP1’s, overexpression of the three dmGDAP1’s in COS-7 cells (no endogenous GDAP1 expression) and SH-Sy5y-cells (endogenous GDAP1 expression) leads to a highly fragmented mitochondrial morphology that is comparable to the mitochondrial morphology observed with overexpressed wildtype GDAP1 (3.3.1). Thus, we next aimed to understand the difference between the rmGDAP1’s and dmGDAP1’s.

The shape of mitochondria depends on the balance of fusion and fission events and A. Niemann could previously show that GDAP1 regulates mitochondrial dynamics processes by inducing mitochondrial fission and not by blocking mitochondrial fusion (Niemann et al., 2005). However, so far it is not clear whether the observed fragmentation of the GDAP1 disease mutants is due to similar mechanisms. To test this, I first analyzed, whether the induction of mitochondrial fission can be counterbalanced by proteins that block fission. For all following analyses I choose two rmGDAP1’s, R310Q with the lowest residual fission activity and R120Q with the highest residual fission activity (of all rmGDAP1’s so far tested for fission activity). With regard to dmGDAP1’s I chose the two at that timepoint identified mutations R120W and T157P, but the experiments were repeated with Q218E, which gave identical results as the other dmGDAP1’s (not shown). We therefore assume that the mechanisms presented here are identical for mutations with the same mode of inheritance. Dynamin-related protein 1 (Drp1) plays a central role in the fission process of mitochondria but its overexpression does not alter the mitochondrial morphology. In contrast, a mutated form, dominant-negative Drp1 (Drp1 mut), is known to block

85 DISCUSSION mitochondrial fission (Smirnova et al., 1998) and leads to extensively elongated mitochondria, as the mitochondrial fusion process is still ongoing. Coexpression of GDAP1, rmGDAP1’s or dmGDAP1’s and the mitochondrial fission factor Drp1 has no effect on the balance between mitochondrial fission and fusion. This was expected, as Drp1 per se does not change the mitochondrial architecture – Drp1 is an abundant protein in the cell (Yoon, 2004). In contrast, Drp1 mut is able to block GDAP1 and mutant induced fission, independent of the mode of inheritance (3.1.2). Thus, if the fission process is blocked by Drp1 mut, GDAP1 and all mutants thereof are incapable to induce mitochondrial fission. Next, I analyzed whether the fission-inducing activity of GDAP1 and its mutants can be counterbalanced by the co-expression of fusion factors. The simultaneous overexpression of a fusion factor can increase mitochondrial fusion and thus reverts the mitochondrial fragmentation phenotype. Coexpression studies of GDAP1, rmGDAP1 or dmGDAP1 with the fusion promoting factors Mfn1 (mitofusin 1) or Mfn2 (mitofusin 2) revealed that mitochondrial fission induced by GDAP1 and rmGDAP1’s can be counterbalanced with co-expression of Mfn1 or Mfn2, but not with mutated, inactive mitofusin proteins (3.1.3). Interestingly, coexpression of dmGDAP1’s and Mfn’s does not alter the amount of fragmented mitochondria. These results suggest that the overexpression of dmGDAP1 impairs the mitochondrial fusion process. Hence, similar clinical manifestations caused by rmGDAP1’s and dmGDAP1’s are due to different modes of action. It is well established that mitochondrial fusion is impaired if the mitochondrial transmembrane potential (

m) is lost or strongly reduced (Kim et al., 2007; Mattenberger et al., 2003; Meeusen et al., 2004; Twig et al., 2008a). Hence, we tested GDAP1, rmGDAP1 and dmGDAP1 transfected cells for their

m. All mitochondria showed an overall normal

m (3.1.4). Despite an overall normal

m, an uneven labeling of mitochondrial structures was observed for all dmGDAP1’s. This heteogeneous

m could be due to the impaired mitochondrial fusion of dmGDAP1’s (3.1.3). Furthermore, A. Niemann found that compared to wildtype GDAP1 and rmGDAP1 transfected cells dmGDAP1 transfected cells have further deficits such as elevated ROS levels and reduced fusion (Niemann et al., 2009, submitted). Interestingly, the expression of mutant Mfn2 leads to impaired mitochondrial transport (Baloh et al., 2007). MFN2 mutations result in axonal forms of CMT, sometimes with optic atrophy (Zuchner et al., 2006). This supports the idea that dominant inherited mutations in GDAP1 and in MFN2 result in similar pathogenesis during the development of axonal forms of CMT. Nevertheless the dominant inherited Q218E mutation shows features of axonal loss and primary demyelination (Chung et al., 2008), which were not detected in patients with mutations in MFN2. Further case reports will have to show whether the dmGDAP1 R120W and T157P are exclusively axonal (Claramunt et al., 2005). Patients with

86 DISCUSSION recessively inherited mutations in GDAP1 show similarly demyelinating, axonal or both types of pathogenesis, but an early onset of the disease (Niemann et al., 2006). In the peripheral nervous system Schwann cells and neurons express GDAP1 (Niemann et al., 2005). The clinical manifestations suggest that the mutations in GDAP1 impair either Schwann cells, neurons, or both with a clear correlation to the type of mutation and mode of inheritance (Niemann et al., 2006). We conclude that similar clinical symptoms are caused either by reduced mitochondrial fission or impaired mitochondrial fusion due to different mutations in the GDAP1 locus and are dependent on the mode of inheritance.

4.2 GDAP1 is a tail-anchored protein

Disease-relevant C-terminal truncations of GDAP1 interfere with its mitochondrial targeting (Niemann et al., 2005). Mutations in the GST-N domain, the interdomain, and the GST-C domain found in patients with CMT underline the importance of these cytosolic domains for GDAP1 function (Niemann et al., 2006). However, my initial approach – the linking of observed functions to specific protein regions via a series of deletion mutants – rendered inconclusive. Any structural changes performed (apart from small N-and C- terminal deletions) resulted in a fission-incapable protein that caused extensive mitochondrial aggregation in the perinuclear region (3.2.1). The mitochondrial clusters have an intact
m (3.2.1) and are composed of an accumulation of hyperfused mitochondria, as shown by electron microscopy analyses (in collaboration with Prof. Elsässer, Universität Marburg) of selected deletion mutants (
GST-N,
GST-C,
HD1; not shown). These data suggest that my initial approach, gaining insights into the structure- function relationship of GDAP1, cannot be realized by the analysis of crude domain deletions. Hence, I changed the strategy and continued with a detailed analysis of specific structural features starting at the C-terminal region of GDAP1.

Similar to Mfn’s, GDAP1 contains two hydrophobic stretches at the outermost C-terminus and is an integral MOM protein. The more C-terminal domain is sufficient for mitochondrial targeting (Niemann et al., 2005) and resamples similarities to a certain class of proteins termed tail-anchored (TA) proteins. Several mitochondrial dynamic factors are TA-proteins (Table 3.1), and thus I reasoned that the putative GDAP1-TA domain might play a functional role in mitochondrial fission activity. Little is known about the functional relevance of TA domains. In yeast, Habib et al. (Habib et al., 2003) replaced the TA- domain of Fis1 by the TA-domain of TOM5 and TOM6. Both chimeric proteins could restore the mitochondrial morphology phenotype in
fis1 yeast strains, indicating that the TA of the mitochondrial fission factor Fis1 does not have a functional role. Yet, the Fis1- TOM6 fusion protein was targeted to and stabilized the TOM complex in
tom6 yeast

87 DISCUSSION strains, leaving the possibility that the TA domain of Tom6 has functional relevance (Habib et al., 2003). In the following study I analyzed the topology of GDAP1 and its mitochondrial targeting features in conjunction with mitochondrial fission. I first aimed at clarifying the mode of MOM-integration. The protease-protection assays revealed that GDAP1 spans the MOM once with its C-terminal TMD. The N-terminal part with the GST domains is located in the cytosol. The short C-terminal tail is located in the IMS (3.2.2.1). These findings are in agreement with the definition of classical TA-proteins that span the membrane with a single transmembrane domain close to the C-terminus (Borgese et al., 2003). Furthermore, I show that GDAP1 integrates post-translationally into membranes in vitro comparable to the integration of the classical TA protein Fis1 (3.2.2.2). I found that GDAP1 integrates into the membrane exclusively dependent on the TA-domain and not on other cytosolic domains like an N-terminal targeting sequence, the GST-domains, or HD1. In contrast, the MOM protein Mfn2, which contains two TMD’s, did not integrate in this assay system (3.2.2.2). A similar integration of in vitro translated TA-proteins into the MOM has previously been demonstrated in digitonin-permeabilized HeLa cells (Setoguchi et al., 2006). My results further demonstrate that the HD1 does not span the MOM. Whether the HD1 is located in the cytosol or is embedded within the bilayer plane cannot be discriminated (modeled in Fig. 4.1). The classification of GDAP1 as a TA protein of mitochondria is further supported by the following results: First, MOM TA protein targeting is regulated by basic amino acids flanking the TMD (Borgese et al., 2003). In concurrence with this fact, I found that mitochondrial targeting of GDAP1 is critically dependent on clusters of positively charged residues surrounding the TMD (3.2.2.3). Second, targeting to the MOM is dependent on a short TMD (Borgese et al., 2003). Indeed, a five amino acid long extension of the TMD of GDAP1 leads to aberrant targeting of the extended protein to the ER (3.2.2.4). Third, the TA-domain of Figure 4.1. Model of possible GDAP1 membrane topologies. Experimentally confirmed is a single transmembrane span with the C- GDAP1 and its adjacent C- terminus in the intermembrane space and the N-terminus in the cytosol. Within this topology two different arrangements of the HD1 are represented. terminal sequences can be Crossed-out: Experimentally disproved topology. MOM, mitochondrial outer membrane; blue, GST-N; red, GST-C; yellow, HD1; green, TMD. replaced by a heterologous TA-

88 DISCUSSION domain and the chimeric protein still maintains mitochondrial targeting (3.2.2.6). Fourth, like for other TA-proteins, GDAP1-targeting and integration into the membrane is impaired by large C-terminal tags like EGFP (3.2.2.1; (Brambillasca et al., 2006). In contrast to the clusters of basic residues surrounding the TMD of GDAP1, the basic residues K291 and R310, adjoining the more N-terminally located HD1, are not required for MOM targeting. They are, however, essential for GDAP1-mediated fission activity (3.2.2.3). Similarly, HD1 deletion or scrambling of its primary sequence severely impair fission activity of the mutant GDAP1 proteins but without affecting mitochondrial localization (3.2.2.4). These experiments indicate a key role for HD1 in GDAP1 function without major involvement in mitochondrial targeting and GDAP1 insertion into the MOM. Indeed the R310Q mutant protein, which has lost the ability to induce mitochondrial fission, was found in CMT patients (Niemann et al., 2005).

If I apply a very restricted definition, which limits the TA to the TMD, my results reveal no effect of the domain on the fission activity. However, the common definition of the TA- domain includes the TMD and the flanking amino acids concurrently needed for correct targeting (Borgese et al., 2003). Intriguingly, my results show that the cluster of basic residues N-terminally bordering the TMD is not only required for mitochondrial localization of GDAP1, but that these amino acids are also of functional importance (3.2.2.3). Even individual residues of this cluster are concomitantly crucial for GDAP1- induced mitochondrial fission. These findings reveal an overlap of residues belonging to the targeting information of the TA-domain and the fission activity of GDAP1. I conclude that TA-domains in higher eukaryotes can, as shown here for GDAP1, not only serve as a targeting sequence but can also have crucial functional relevance for a mitochondrial dynamics factor.

4.3 GDAP1 is an active glutathione S-transferase

Mutations in GDAP1 have been linked to CMT. However, little is known about the pathogenic mechanism and how mutations in this protein lead to the observed defects in Schwann cells, neurons or both. Functional analyses by A. Niemann revealed that GDAP1 is involved in mitochondrial dynamics processes and induces mitochondrial fission (Niemann et al., 2005). Shifting the balance of fusion and fission is a major regulatory mechanism, and long, myelinated peripheral nerves are particularly dependent on the proper function of this complex system. As a consequence, an alteration of this dynamic network by mutations in GDAP1 leads to motor and sensory neuropathies. While by now it is well established that GDAP1 is a regulator of mitochondrial dynamics, it remains elusive how the GST-features of GDAP1 might by involved in the regulation of

89 DISCUSSION the mitochondrial architecture. Bioinformatic approaches revealed that GDAP1, despite containing structural features of canonical GST enzymes, does not belong to any of the known GST-classes and rather belongs to a new class termed GDAP1-class GSTs (Marco et al., 2004). This GST-class is closely related to the cytosolic Zeta, Omega and Theta class GST proteins (Marco et al., 2004; Shield et al., 2006). Previous studies using bacterially produced recombinant truncated GDAP1 (GDAP1
320-358, Pedrola et al., 2005; GDAP1
290-358, GDAP1
334-358, Shield et al., 2006) could not find evidence for GSH-binding or prototypic GSH-dependent enzyme activity with the substrate CDNB (1- chloro-2,4-dinitrobenzene). CDNB is the most commonly used GST substrate. The sensitive reaction involves an aromatic substitution in which GSH displaces chloride to produce a thioether S-2,4-dinitrophenylglutathione (Cohen et al., 1964). However, several classes of GST, including Zeta, Omega and Theta classes exhibit low (or no) activity with CDNB (Tan et al., 1996). Alternative substrates utilized by different classes of GSTs, including those most closely related to GDAP1 (Zeta, Omega, Theta) were also tested without success using bacterially expressed recombinant truncated GDAP1 (Shield et al., 2006).

Recent estimates indicate that perhaps one-third to one-half of all prokaryotic proteins cannot be overexpressed in bacteria in soluble form (Edwards et al., 2000; Stevens, 2000). This number is even higher for the expression of eukaryotic proteins, as stated by several independend high-throughput studies (Braun et al., 2002; Hammarstrom et al., 2002; Shih et al., 2002). On the basis of this knowledge we changed our strategy and used the eukaryotic baculovirus-based insect cell expression system Multibac (Bieniossek et al., 2008) to express recombinant truncated GDAP1 288X (GDAP1
289-358). We could not detect a binding of GDAP1 to GSH. This is in consistence with previous reports, which could not detect a binding of the Theta class GST’s (a subgroup of GST enzymes with high similarity to the GDAP1 class) to immobilized GSH (Sheehan et al., 2001; Spearman et al., 1985). Interestingly, we detected significant GST enzymatic activity with the substrates 1,2-epoxy-3-(4-nitrophenoxy)propane (EPNP), p-nitrobenzylchloride (pNBC) and ethacrynic acid (EA; 3.2.3.2). GST’s display overlapping substrate specificities, a feature that makes it difficult to identify isoenzymes solely on the basis of their catalytic properties. However, certain GST classes have preferences for specific substrates. Shield et al. (2006) claim that GDAP1 is closely related to the Zeta, Omega or Theta class and structural features from in silico modeling approaches strongly suggest a close relationship to the Theta class of GST enzymes (Shield et al., 2006). This is also consistent with our detected substrate specificities. Omega class GST’s are known to show particularly high GSH conjugation activities towards CDNB, an activity not detected for GDAP1 (Sheehan et al., 2001). Zeta class GSTs are known to have low activities with CDNB but also with EA and additionally display GSH peroxidase activity with both cumene and t-buthyl

90 DISCUSSION hydroperoxides (Sheehan et al., 2001). For GDAP1 we could not detect any GSH peroxidase activity. In contrast, Theta class GST’s have been shown to display no activity with CDNB but have high activities towards EPNP (Meyer et al., 1991) and pNBC (Spearman et al., 1985). This activity pattern is closest to the activities detected with GDAP1. Furthermore, the detected activities of GDAP1 are within the range of activities that were also detected with different Theta-class GST enzymes such as GST5-5 (i.e. specific activity for pNBC: 86 (GST5-5, Meyer et al., 1991) and 96 (GDAP1); 3.2.3.2). pNBC is widely used as xenobiotic model substrate (Widersten, 1995) and is typically found as substrate for Theta class GST enzymes (Spearman et al., 1985). Usually, enzyme- catalyzed reactions neutralize the electrophilic substrate site and render the products more water-soluble. Subsequent cleavage of the glutamate and glycine residues is followed by N-acetylation of the resulting cysteine conjugate. Once formed, these conjugates are eliminated from the cell by the trans-membrane MRP (multidrug resistance protein; Habig et al., 1974; Hayes et al., 2005). Epoxides such as EPNP are converted into less reactive products ready for excretion by GST-mediated catalysis of epoxide ring opening reactions (Meyer et al., 1991). EPNP belongs to the group of reactions that involve conjugation to GSH with activated alkenes (Michael additions) and epoxides. Corresponding endogenous substrates are compounds that may arise from normal cellular constituents by oxidative processes, such as radical reactions, lipid peroxidation and oxygenations catalyzed by the cytochrome P450 system (Mannervik, 1995). Exogenous compounds corresponding to EPNP comprise especially a large number of epoxides such as environmental carcinogens. GDAP1 may be involved in detoxification processes of endogenous compounds such as highly reactive byproducts arising from oxidative chain reactions in mitochondria, and thus helps to retain the cell in a reduced state. GDAP1 was also found to display GSH-conjugating activity with EA. EA can act as both a substate and inhibitor of GST enzymes (Awasthi et al., 1993). Because of this inhibitory activity it is currently used as a chemotherapeutic enhancing agent (Tew et al., 1998). The GST activity detected with EA provides a first link between the GST activity of GDAP1 and its regulation of the mitochondrial network. A few reports have commented on the role of EA in mitochondrial dynamics, where EA was shown to cause mitochondrial fusion leading to the formation of a complete mitochondrial reticulum (Awasthi et al., 1993; Bowes and Gupta, 2005). EA is a potent inducer of oxidative stress due to its
,- unsaturated ketone that spontaneously reacts with GSH, leading to GSH depletion in the cytosol and mitochondria (Seyfried et al., 1999). Bowes et al. (2005) claim that the extensive mitochondrial fusion observed in response to EA treatment is not dependent on its GSH-depletion properties but is due to alkylation of one or more specific cysteine residues in proteins that are involved in mitochondrial fusion and fission processes. More specifically, they suggest that on the fission site EA inactivates proteins crucial for fission

91 DISCUSSION such as Drp1, a protein responsible for the regulation of outer membrane fission by mechanisms of GTP-hydrolysis (Legesse-Miller et al., 2003). Mutants of Drp1 show a similar fusion morphology as observed with EA. Recent studies further support the hypothesis of a mitochondrial fission block caused by EA treatment (Bowes and Gupta, 2008). GDAP1 is involved in mitochondrial dynamics and induces mitochondrial fission dependent on Drp1 and Fis1 (Niemann et al., 2005). One could speculate that the GDAP1 fission activity is regulated by its GST-activity. First insights into the close relation between GDAP1-induced mitochondrial fission and its GSH conjugation activity were obtained by studies of GDAP1 mutants that contain point mutations in highly conserved amino acids of the GST-N domain. This N-terminal thioredoxin fold is the among all GST’s highly conserved binding-site for GSH. Specific residues critical for GSH binding or GST activity, in particular a serine that is crucial for catalytic activation of GSH, were exchanged. All introduced point mutations in the GST-N domain (Y29A, S34A, S36A, S37A and F68S) completely abolished the GDAP1-typical fission activity in COS-7 cells. This supports our hypothesis that the fission activity of GDAP1 may be related to its GST activity (3.2.3.1). However, these mutations also lead to an increased amount of mitochondrial aggregates that accumulate in perinuclear regions. This effect was especially evident for the point mutation S34A; around 50 % of all cells contain mitochondrial aggregates not observed in the wildtype protein. Interestingly, Shield et al. (2006) proposed that this serine residue positioned at the amino terminal end of helix 1 in the GST-N domain constitutes the conserved active site residue for the GDAP1 class of GST proteins. Initial electron microscopic studies revealed that the mitochondrial aggregations are composed of normal mitochondria, indicating a loss of fission activity and display a dearranged inner mitochondrial membrane (not shown, in collaboration with Prof. Elsässer, Universität Marburg). This suggests that an accumulation of dysfunctional GDAP1 leads to the extensive perinulear mitochondrial aggregation. However, the exact mechanisms remains elusive and further research is required to clarify the connection between GDAP1 fission- and GST activity.

The above mentioned GST activities were detected with the truncated GDAP1 288X, a construct lacking HD1 and TMD. In contrast, a longer construct only lacking the TMD, GDAP1 318X (GDAP1
319-358), does not display GSH-conjugating activities. On a highly speculative note this may be due to an autoinhibitory mechanism triggerd by HD1. HD1 could modulate GST activity due to primary sequence specificities, hydrophobicity dependent interactions, or posttranslational modifications such as phoshorylation. Autoinhibitory mechanisms have been described for proteins such as the transcription factor interferon regulatory factor (IRF-3). Phosphorylation of specific Ser or Thr residues results in an unfolding of an N- and C-terminal autoinhibitory structure. The introduction of a negative charge, through either phosphorylation or mutagenesis, exposes a

92 DISCUSSION hydrophobic surface for interaction with CBP/p300 (Lin et al., 1999; Panne et al., 2007). Initial database searches revealed no conserved motifs for posttranslational modifications in the HD1 (not shown). Studies on the effect of HD1 mutations on GDAP1 fission activity implicate a crucial role for HD1. Deletion as well as scrambling of HD1 (without changing the overall HD1 hydrophobicity) resulted in a complete loss of GDAP1 fission activity (3.2.2; Wagner et al., 2009). Thus, the primary sequence structure of HD1 alone is essential for full GDAP1 fission activity. To further study this issue I will create a HD1 scrambled construct for recombinant expression in insect cells. If HD1 inhibits GDAP1 GST activity, this construct should now display GST enzyme activity. Another approach to test this hypothesis would be to measure GDAP1 enzyme activity of the active short version in the presence of a peptide that corresponds to the HD1 and thus would inhibit enzymatic activity.

CMT causing mutations of GDAP1 are likely to yield important information about the recently identified GST activity of GDAP1 and its relevance for the disease phenotype. Furthermore, analysis of the GST activity of the disease mutants will reveal whether the GST activity is differentially affected by the mode of inheritance (recessive (rmGDAP’s) vs dominant (dmGDAP1’s)) and whether the residual fission activity of rmGDAP1’s is in congruence with a diminished GST activity. Most disease-related GDAP1 mutations are found in the GST-C domain and in the region between the GSH- and the substrate-binding domain, with an accumulation in the interdomain loop that is not found in canonical GST’s. Only a few mutations are found in the GST-N domain (P78L) and close to the HD1 (R310Q). To assess the influence of the disease-causing mutations on the GST activity, we recombinantly expressed a selection of the disease mutants in the short (GDAP1 288X) and the long (GDAP1 318X) version. However, up to now too little amounts of protein could ultimately be purified, suggesting a folding deficit of the recombinant proteins. We will proceed with a large scale expression of the disease mutants in order to obtain sufficient amounts of soluble protein for the GST activity assays. Alternatively, we will express the GDAP1 disease mutants as MBP (maltose binding protein) fusion proteins. This will faciliate the purification process and could also enhance the correct folding of the disease mutations (Smyth et al., 2003).

For in-depth understanding of the structure-function relationship of GDAP1 and its functional role as a GST protein it is necessary to reveal the three-dimensional structure of GDAP1. In silico modeling approaches with a three-dimensional structure would also allow a screen for additional potential GST substrates. First crystallization attempts using recombinant truncated GDAP1 288X and GDAP1 318X expressed in insect cells yielded no crystals (3.2.4). As GDAP1 seems not to be conducive to more conventional crystallization strategies, we changed to a different system using N-terminal MBP fusions.

93 DISCUSSION

MBP is a part of the maltose/maltodextrin system of E.coli and known to be advantageous in terms of increased expression, enhanced solubility, protection from proteolysis, improved protein folding, and protein purification by affinity chromatography. Furthermore, MBP aids structure determination of proteins that present particular experimental challenges (Smyth et al., 2003). Here, the length of the flexible linker region is of crucial importance. This region is considered to be the largest detriment to successful crystallization of fusion proteins with large affinity tags due to the conformational heterogeneity introduced by the linker region (Smyth et al., 2003). Therefore, we constructed MBP-tagged versions of GDAP1 288X and GDAP1 318X without and with a linker region (7 amino acid flexible linker; 3.2.4). However, first crystals obtained with this approach were not well ordered and diffracting, and could thus not be used to determine the three-dimensional structure via X-ray crystallography. We will proceed with further optimization of the conditions for protein purification and crystallization. Finally, we also identified more stable, Edman degradation and mass spectrometry verified versions of GDAP1 288X and GDAP1 318X that have a shortened N-terminus (3.2.4). These constructs will also be tested for crystallization and subsequent X-ray crystallograpy.

4.4 GDAP1L1 subcellular distribution depends on oxidative stress

Although expression of GDAP1 is found in tissues of the PNS and CNS (Cuesta et al., 2002; Liu et al., 1999; Niemann et al., 2005), GDAP1 disease mutations cause axonal, intermediate or demyelinating forms of the peripheral neuropathy CMT without severe CNS phenotype (Niemann et al., 2006). This implies, that either GDAP1 function is somewhat compensated by another protein only expressed in the CNS or that GDAP1 exerts different functions in the CNS compared to the PNS. GDAP1L1 is a closely related paralog of the mitochondrial fission factor GDAP1 and belongs to the same new GDAP1 class of GST enzymes (Marco et al., 2004). On a highly speculative note this renders GDAP1L1 a potential candidate for the compensation of GDAP1 function in the CNS and could thus be the reason for the lacking CNS phenotype of CMT patients.

4.4.1 GDAP1L1 is predominantly expressed in the CNS In order to compensate for a GDAP1 loss of function in the CNS, GDAP1L1 needs to meet the criteria of mitochondrial localization in CNS tissue. Consistent with mRNA expression data from B. Angst (Angst, 2005) on protein levels, GDAP1L1 is most prominently expressed in tissues of the CNS, such as cortex, cerebellum, thalamus, olfactory bulb and spinal cord, whereas no expression was detected in tissues of the PNS (3.3.1). Further analysis of specific brain regions revealed that GDAP1L1 is predominantly found in

94 DISCUSSION neuronal regions of the cerebellar molecular, purkinje cell and granular cell layers, and in granule cells of the dentate gyrus of the hippocampus and pyramidal cells of the cortex – regions where GDAP1 expression was also detected (3.3.1). Cell culture studies by B. Angst with COS-7 cells showed that GDAP1L1 does not co- localize with mitochondria on a subcellular level but is rather distributed throughout the cytsol. These data are in agreement with my subcellular localization studies; I could not detect any appreciable overlap of GDAP1L1 in COS-7 cells with membraneous structures including mitochondria, the Golgi complex and the ER (3.3.2). However, it was speculated that the predominant cytosolic distribution could be due to overexpression artifacts in cell lines that do not endogenously express the protein (Angst, 2005). So far, no suitable cell line system was found, as several frequently used cell lines, such as Hek and Hela as well as neuroblastoma and schwannoma cell lines, were tested negative for endogenous GDAP1L1 expression (Angst, 2005). Furthermore, primary mouse hippocampal and cortical neuron cultures were also tested negative for endogenous GDAP1L1 expression (3.3.2). Protein expression analysis revealed that in contrast to adult mice, which have strong GDAP1L1 expression in CNS tissue, cortical and hippocampal regions of two day old mice are negative (3.3.2). This may explain the lack of endogenous GDAP1L1 expression in primary neurons, as the neurons used for my studies are prepared from one day old mice. In contrast, GDAP1 expression in the CNS is already found in two day old mice, and primary hippocampal and cortical neurons are positive for GDAP1 expression (3.3.2). Thus, if GDAP1L1 is able to compensate for GDAP1 loss of function in the CNS, this molecular mechanism is either only essential for adult organisms or GDAP1L1 is able to revert accumulated deficits due to dysfunctional GDAP1 proteins. Finally, to obtain first information about the localization of endogenous GDAP1L1, I changed to an adult in vivo system using rat brain saggital paraffin sections. Purkinje cells of the cerebellum contain large neuronal bodies that allow for subcellular localization studies of proteins. Interestingly, consistent with our previous results from cell culture overexpression systems, GDAP1L1 is not localized to specific membraneous structures but appears to be cytosolic (3.3.1). In contrast, using the same system, GDAP1 clearly co- localizes with the mitochondrial marker protein porin. Thus, under normal cellular conditions GDAP1L1, although expressed in cells in the CNS where GDAP1 is also detected, is not localized to mitochondria. This alone would suggest that GDAP1L1 is unlikely to perform functions similar to those of the MOM protein GDAP1. However, although not localized to mitochondria, GDAP1L1 induces minor mitochondrial fission (3.3.4). GDAP1 induced mitochondrial fragmentation is still much higher (40 % of GDAP1 transfected cells display fragmented mitochondria vs 15 % in GDAP1L1 transfected cells) but compared to the amount of fragmented mitochondria found in control transfected cells (4 %) this increase is significant.

95 DISCUSSION

4.4.2 GDAP1L1 contains features of a tail-anchored protein I defined GDAP1 as a TA protein that spans the MOM once with its C-terminal TMD (4.2). The TMD and its bordering basic residues in the IMS are crucially involved in mitochondrial targeting and membrane insertion. Positively charged amino acids flanking the TMD on the cytosolic side control both, mitochondrial targeting and the fission function of the protein (Wagner et al., 2009). Although GDAP1L1 appears to be a cytosolic protein, a comparison of the primary sequence structure of GDAP1 and GDAP1L1 revealed the presence of very similar membrane targeting motifs. TMD and HD1 are comparable in length and contain basic amino acids flanking the domains. Thus, the essential criteria for putative mitochondrial targeting of GDAP1L1 are fulfilled. However, in contrast to GDAP1 and other classical MOM targeted TA proteins such as Fis1 (Wagner et al., 2009), full-length GDAP1L1 does not post-translationally integrate into membranes in vitro. Another MOM targeted TA protein, Bax, also resides in the cytosol under normal cellular conditions. The C-terminal tail occupies a hydrophobic pocket, which is thought to provide a binding site for the so-called BH3 domain of the Bcl- 2 family of proteins. Upon binding of the Bcl-2 family members the C-terminal TA is displaced and ready for MOM integration, switching on the pro-apoptotic activity of the protein Bax (Borgese et al., 2003; Nechushtan et al., 1999; Suzuki et al., 2000). These data led to the speculation that the TA of GDAP1L1 could also be hidden and only exposed and ready for MOM integration upon certain external stimuli. To support this hypothesis I tested the ability of the putative TA alone to integrate into the MOM. I found, that the putative TMD of GDAP1L1 alone is able to anchor N-terminal residues (here an EGFP- tag) into the MOM in our in vitro assay as well as in transiently transfected COS-7 cells (3.3.3). Based on the knowledge that TA’s can be readily exchanged between different TA proteins without affecting overall functionality of the protein (Borgese et al., 2007; Brambillasca et al., 2006; Wagner et al., 2009) and the fact that the TMD of GDAP1L1 is able to act as TA for protein tags, I further tested whether the TA of GDAP1L1 is sufficient to integrate full- length proteins into the MOM and thus act as a classical TA. The paralog GDAP1 provides a mean to test this. I could previously show that the TA of GDAP1 can be exchanged with TA’s of different TA-MOM proteins without impairing its fission-inducing function (Wagner et al., 2009). I found that the TA of GDAP1 and its adjacent C-terminal sequences can be replaced by the corresponding TA-region of GDAP1L1. The chimeric protein still maintains mitochondrial targeting and GDAP1 fission activity, albeit slighty reduced (3.3.4). Thus, if exposed, the TA of GDAP1L1 is able to integrate proteins into the MOM. Furthermore, this approach allows to monitor the effect of GDAP1L1 on mitochondria if artificially targted to these organelles. Therefore I exchanged the TA of GDAP1L1 with

96 DISCUSSION the corresponding TA-region of GDAP1. The chimeric protein is correctly integrated into the MOM but does not alter the mitochondrial morphology (3.3.4). Moreover, in contrast to wildtype GDAP1L1 that is, although cytosolic, able to induce mitochondrial fragmentation, this chimeric protein has lost this activity. A possible explanation could be that (1) the TA-region of GDAP1L1 is crucial for the observed fission activity of the protein or (2) due to the artificial TA the overall structural integrity of GDAP1L1 is disrupted, leading to a loss of function.

The TA sequence requirements for Bax translocation have been extensively investigated (Nechushtan et al., 1999). Essential for the regulated binding of Bax to the MOM is a Ser residue in the TMD (Ser 184). If deleted or replaced with Ala or Val, Bax constitutively localizes to the MOM. Structural studies of Suzuki et al. (2000) revealed that an H-bond between Ser184 and Asp98 stabilizes the binding of the TA in the hydrophobic pocket. The TMD of GDAP1L1 contains two Ser residues at the positions Ser345 and Ser350. It will be interesting to test whether GDAP1L1 variants with mutations at these positions act via a similar mechanism and also constitutively localize to mitochondria.

Similar to GDAP1, GDAP1L1 contains a second hydrophobic domain (HD1) close to the C-terminal TMD. This domain is highly conserved between both proteins and thus could be involved in similar functional mechanisms. GDAP1 HD1 has been shown to be of crucial importance for the proper function of GDAP1, since deletion of HD1 or altering the correct primary sequence severely impairs GDAP1 fission activity (Wagner et al., 2009). Furthermore, I speculated (4.3) that HD1 modulates the enzymatic GST activity of GDAP1 via an autoinhibitory mechanism, as a protein version containing the HD1 does not display GST activity. Despite the similarity of the HD1’s, the fission activity of both proteins was completely abolished upon an exchange of the HD1 between both proteins (3.3.5). This strengthens the importance of HD1 for proper function of the proteins. My data suggest that the C- terminal regions of GDAP1L1 and GDAP1 differ in their structural organization. The TA of GDAP1 is exposed and can readily insert into the MOM, whereas the TA of GDAP1L1 seems to be hidden under normal cellular conditions. This suggests that despite the HD1 similarities, both might regulate certain functions in a different structural environment, with stringent requirements for features that cannot be compensated by the HD1 of the other protein.

4.4.3 Quinone-induced ROS stress directs GDAP1L1 to mitochondria As mentioned in section 4.4.2, certain TA proteins such as Bax can reside constitutively in the cytosol and become integrated into a membrane in response to a signal. GDAP1L1

97 DISCUSSION displays high similarities to Bax in terms of its putative TA, its cytosolic localization and its assumed need for dimerization (it belongs to the GDAP1 class of GST’s and GST enzymes are usually active as dimers). To test whether GDAP1L1 is regulated by conformational changes due to external stimuli comparable to Bax, I exposed N1E-115 cells transiently overexpressing GDAP1L1 to various cellular stressors (TBHQ, DTT, CCCP, TBHP, actinomycin D, staurosporine, N- etylmaleimide, menadione, H2O2), and analyzed its subcellular localization (3.3.6). As in the CNS GDAP1 and GDAP1L1 are usually co-expressed (4.4.2), N1E-115 cells, although not endogenously expressing GDAP1L1, express its close relative GDAP1 and thus represent the best system available. Strikingly, exposure to menadione (MQ) led to a translocation of GDAP1L1 to distinct membraneous structures (3.3.6; Fig. 4.1). Subcellular localization studies using markers for membraneous compartments such as the Golgi complex, ER and mitochondria identified the GDAP1L1-positive structures as mitochondria although proper MOM integration could not yet be confirmed. This translocation was monitored 2 h after MQ treatment; prolonged exposure lead to cytochrome C release and subsequent cell death (3.3.7). The effect of MQ on the GDAP1L1 localization was not observed in other cell lines such as COS-7 or HeLa cells that do not endogenously express GDAP1 (not shown). This suggests that the MQ-induced GDAP1L1 translocation to mitochondria may depend on GDAP1. To test this, one could perform transient co-transfections of GDAP1 and GDAP1L1 in cells with no endogenous expression of both proteins and assess the effect of MQ on GDAP1L1 subcellular localization. Alternatively, the subcellular localization of GDAP1L1 upon MQ treatment can be assessed in N1E-115 cells under GDAP1 knock-down conditions. To exclude translocation artifacts and strengthen the argument that the putative TA of GDAP1L1 is crucial for the translocation, it will be necessary to assess the effect of MQ on a GDAP1L1 variant lacking the TA.

The exogenous stressor MQ is a napthoquinone. Quinones are either eliminated via their GST-mediated conjugation to GSH or by their metabolic degradation to a semiquinone (a process also called redox cycling) by various flavoenzymes such as NADPH cytochrome

P450 reductase, NADH cytochrome b5 reductase and NADH-ubiquinone reductase. The semiquinone product is oxidized back to the parent quinone in the presence of molecular oxygen, thereby generating ROS such as superoxide anion radical, hydrogen peroxide and others (Abdelmohsen et al., 2004; Castro et al., 2007; Gerasimenko et al., 2002). In addition, this drug can be converted to its semiquinone by non-enzymatic reactions with various thiols and, in the presence of oxygen, ROS can also be formed (Chung et al., 1999). Furthermore, MQ prevents mitochondrial Ca2+ uptake leading to elevations in the cytosolic Ca2+ concentration, which also contributs to cell death (Nicotera et al., 1992). MQ has been shown to be a potent inducer of the mitochondrial permeability pore (PTP).

98 DISCUSSION

Activation of the PTP triggers an open pore state, possibly causing swelling of the mitochondrial matrix and rupture of the MOM with release of apoptogenic factors such as cytochrome C from the intermembrane space (Martinou et al., 2000). GDAP1L1 was proposed to be a GST enzyme. Thus, one could speculate that GDAP1L1 is involved in the conjugation of GSH to MQ, marking MQ for cell export. While MQ can both, redox cycle and it can be conjugated to GSH, two related quinone compounds, DNMQ and TBHQ can exclusively redox cycle or be degraded via conjugation to GSH, respectively. I could demonstrate that MQ triggers GDAP1L1 translocation via its redox cycling properties, as only DNMQ treatment lead to comparable changes in the subcellular localization of GDAP1L1 (3.3.8).

Figure 4.2. Hypothetic model of GDAP1L1 translocation due to menadione treatment. (A) GDAP1L1 is distributed throughout the cytosol under normal reduced cellular conditions (green background). The TA of GDAP1L1 is hidden and is not able to integrate into the MOM. (B) Upon menadione (MQ) treatment the TA is exposed and is able to integrate into the MOM (oxidised cellular environment indicated by orange background).

The GST-substrate ethacrynic acid (EA) is able to spontaneously react with GSH leading to its known GSH-depletory effects in cytosol and mitochondria (Seyfried et al., 1999). GSH is involved in the maintenance of low cellular ROS levels, and depletion of GSH causes an increase in cellular ROS (Bowes and Gupta, 2005). However, ROS generated by EA treatment did not alter the subcellular localization of GDAP1L1 in N1E-115 cells (3.3.8). One could speculate that ROS levels generated by EA or other agents tested are, despite being deleterious for cell survival, low compared to ROS generate due to MQ treatment. However, for reagents such as TBHP no difference in the amount of ROS compared to MQ-treated cells was observed (not shown). This suggests that the translocation of GDAP1L1 to mitochondria is not initiated by ROS in general, but is rather due to specific features of the naphtoquinone pathway.

The antioxidant NAC, a free radical scavenger, is a precursor of the ROS scavenger GSH (Criddle et al., 2006; Watanabe et al., 2004). The rate-limiting step in GSH synthesis is catalyzed by the enzyme glutamate cysteine ligase, which is readily induced by quinones

99 DISCUSSION such as MQ. High intracellular GSH levels are crucial to maintain a reduced environment that allows for normal cellular activity. To assess a potential protective effect of NAC on MQ-induced GDAP1L1 translocation, I analyzed the effect of MQ on GDAP1L1 overexpressing N1E-115 cells in parallel with NAC treatment. NAC was able to significantly reduce the amount of newly generated ROS in our experimental settings, and furthermore prevented the MQ-induced redistribution of GDAP1L1. This suggests that the MQ-induced alterations in the reduced cellular environment (shift to a more oxidized environment) are essential for the translocation of GDAP1L1 to the MOM. Oxidative stress is defined as a disturbance in the oxidant-antioxidant balance in favor of the former and thus causing harmful damages to membranes, proteins and DNA. It is believed that oxidative stress is related to the natural process of aging and to disorders such as atherosclerosis, cancer, Alzheimer’s disease and other neurodegenerative diseases (Costa and Moradas-Ferreira, 2001). The translocation of GDAP1L1 to mitochondria upon MQ treatment suggests that GDAP1L1 may be involved in cell survival mechanisms due to oxidative stress. This is further supported by the knowledge that mitochondria play a key role in energy metabolism utilizing oxygen, and as a consequence high amounts of ROS are constantly produced under physiological conditions in the close vicinity of these organelles (Goto et al., 2009). First experiments, however, revealed that on the protein level MQ treatment causes slightly increased protein oxidation, but neither GDAP1L1 nor GDAP1 expression had a protective effect on these oxidation levels.

We hypothesize, that GDAP1L1 may compensate for a loss of GDAP1 function in the CNS. Recent in vitro studies performed in our lab further support this speculation. A. Niemann could show that in N1E-115 cells transiently overexpressed GDAP1L1 is able to rescue a loss of fission activity due to GDAP1 knock-down. Although the original amount of fragmented mitochondria was not completely restored, this experiment showed that GDAP1L1 has the potential to compensate for loss of GDAP1 fission activity in the CNS. Future experiments will aim at determining the subcellular localization of GDAP1L1 under GDAP1 knock-down conditions. To verify our hypothesis it will be necessary to validate and complement our in vitro data with in vivo studies in mouse models. Presently, a “straight” and a conditional GDAP1 knock-out are being bred in our mouse facility. It will be interesting to determine whether GDAP1L1 is upregulated on protein level and/or translocated to mitochondria in GDAP1-deficient CNS neurons and supporting cells, and might therefore compensate for the missing GDAP1 function. Furthermore, one could generate a GDAP1L1/GDAP1 double knock-out mouse using a commercially available GDAP1L1 gene trap line. If GDAP1L1 is able to compensate for GDAP1 function in the CNS a double knock-out would additionally to the PNS phenotype also display a CNS phenotype.

100 REFERENCES

5. References

Abdelmohsen, K., Patak, P., Von Montfort, C., Melchheier, I., Sies, H., and Klotz, L.O. (2004). Signaling effects of menadione: from tyrosine phosphatase inactivation to connexin phosphorylation. Methods Enzymol 378, 258-272. Abell, B.M., Rabu, C., Leznicki, P., Young, J.C., and High, S. (2007). Post-translational integration of tail-anchored proteins is facilitated by defined molecular chaperones. J Cell Sci 120, 1743-1751. Alves, C.S., Kuhnert, D.C., Sayed, Y., and Dirr, H.W. (2006). The intersubunit lock-and- key motif in human glutathione transferase A1-1: role of the key residues Met51 and Phe52 in function and dimer stability. Biochem J 393, 523-528. Ammar, N., Nelis, E., Merlini, L., Barisic, N., Amouri, R., Ceuterick, C., Martin, J.J., Timmerman, V., Hentati, F., and De Jonghe, P. (2003). Identification of novel GDAP1 mutations causing autosomal recessive Charcot-Marie-Tooth disease. Neuromuscul Disord 13, 720-728. Angst, B. (2005). Expression and Subcellular Localization Analysis of Ganglioside- induced Differentiation-Associated Protein 1-Like 1 (GDAP1L1). In Institute of Cell Biology (Zurich, ETH Zurich). Armstrong, R.N. (1997). Structure, catalytic mechanism, and evolution of the glutathione transferases. Chem Res Toxicol 10, 2-18. Arroyo, E.J., and Scherer, S.S. (2000). On the molecular architecture of myelinated fibers. Histochem Cell Biol 113, 1-18. Auer-Grumbach, M., Fischer, C., Papic, L., John, E., Plecko, B., Bittner, R.E., Bernert, G., Pieber, T.R., Miltenberger, G., Schwarz, R., et al. (2008). Two novel mutations in the GDAP1 and PRX genes in early onset Charcot-Marie-Tooth syndrome. Neuropediatrics 39, 33-38. Awasthi, S., Srivastava, S.K., Ahmad, F., Ahmad, H., and Ansari, G.A. (1993). Interactions of glutathione S-transferase-pi with ethacrynic acid and its glutathione conjugate. Biochim Biophys Acta 1164, 173-178. Azzedine, H., Ruberg, M., Ente, D., Gilardeau, C., Perie, S., Wechsler, B., Brice, A., LeGuern, E., and Dubourg, O. (2003). Variability of disease progression in a family with autosomal recessive CMT associated with a S194X and new R310Q mutation in the GDAP1 gene. Neuromuscul Disord 13, 341-346. Baloh, R.H. (2008). Mitochondrial dynamics and peripheral neuropathy. Neuroscientist 14, 12-18. Baloh, R.H., Schmidt, R.E., Pestronk, A., and Milbrandt, J. (2007). Altered axonal mitochondrial transport in the pathogenesis of Charcot-Marie-Tooth disease from mitofusin 2 mutations. J Neurosci 27, 422-430. Barankova, L., Vyhnalkova, E., Zuchner, S., Mazanec, R., Sakmaryova, I., Vondracek, P., Merlini, L., Bojar, M., Nelis, E., De Jonghe, P., et al. (2007). GDAP1 mutations in Czech families with early-onset CMT. Neuromuscul Disord 17, 482-489. Barres, B.A., and Raff, M.C. (1994). Control of oligodendrocyte number in the developing rat optic nerve. Neuron 12, 935-942.

101 REFERENCES

Baxter, R.V., Ben Othmane, K., Rochelle, J.M., Stajich, J.E., Hulette, C., Dew-Knight, S., Hentati, F., Ben Hamida, M., Bel, S., Stenger, J.E., et al. (2002). Ganglioside-induced differentiation-associated protein-1 is mutant in Charcot-Marie-Tooth disease type 4A/8q21. Nat Genet 30, 21-22. Bellomo, G., Thor, H., and Orrenius, S. (1990). Modulation of cellular glutathione and protein thiol status during quinone metabolism. Methods Enzymol 186, 627-635. Berger, P., Niemann, A., and Suter, U. (2006). Schwann cells and the pathogenesis of inherited motor and sensory neuropathies (Charcot-Marie-Tooth disease). Glia 54, 243-257. Bernard, R., De Sandre-Giovannoli, A., Delague, V., and Levy, N. (2006). Molecular genetics of autosomal-recessive axonal Charcot-Marie-Tooth neuropathies. Neuromolecular Med 8, 87-106. Biancheri, R., Zara, F., Striano, P., Pedemonte, M., Cassandrini, D., Stringara, S., Manganelli, F., Santoro, L., Schenone, A., Bellone, E., et al. (2006). GDAP1 mutation in autosomal recessive Charcot-Marie-Tooth with pyramidal features. J Neurol 253, 1234-1235. Bieniossek, C., Richmond, T.J., and Berger, I. (2008). MultiBac: multigene baculovirus- based eukaryotic protein complex production. Curr Protoc Protein Sci Chapter 5, Unit 5 20. Birouk, N., Azzedine, H., Dubourg, O., Muriel, M.P., Benomar, A., Hamadouche, T., Maisonobe, T., Ouazzani, R., Brice, A., Yahyaoui, M., et al. (2003). Phenotypical features of a Moroccan family with autosomal recessive Charcot-Marie-Tooth disease associated with the S194X mutation in the GDAP1 gene. Arch Neurol 60, 598-604. Board, P.G., Coggan, M., Chelvanayagam, G., Easteal, S., Jermiin, L.S., Schulte, G.K., Danley, D.E., Hoth, L.R., Griffor, M.C., Kamath, A.V., et al. (2000). Identification, characterization, and crystal structure of the Omega class glutathione transferases. J Biol Chem 275, 24798-24806. Boerkoel, C.F., Takashima, H., Nakagawa, M., Izumo, S., Armstrong, D., Butler, I., Mancias, P., Papasozomenos, S.C., Stern, L.Z., and Lupski, J.R. (2003). CMT4A: identification of a Hispanic GDAP1 founder mutation. Ann Neurol 53, 400-405. Borgese, N., Brambillasca, S., and Colombo, S. (2007). How tails guide tail-anchored proteins to their destinations. Curr Opin Cell Biol. Borgese, N., Colombo, S., and Pedrazzini, E. (2003). The tale of tail-anchored proteins: coming from the cytosol and looking for a membrane. J Cell Biol 161, 1013-1019. Borgese, N., Gazzoni, I., Barberi, M., Colombo, S., and Pedrazzini, E. (2001). Targeting of a tail-anchored protein to endoplasmic reticulum and mitochondrial outer membrane by independent but competing pathways. Mol Biol Cell 12, 2482-2496. Bossy-Wetzel, E., Barsoum, M.J., Godzik, A., Schwarzenbacher, R., and Lipton, S.A. (2003). Mitochondrial fission in apoptosis, neurodegeneration and aging. Curr Opin Cell Biol 15, 706-716. Bouhouche, A., Birouk, N., Benomar, A., Ouazzani, R., Chkili, T., and Yahyaoui, M. (2007). A novel GDAP1 mutation P78L responsible for CMT4A disease in three Moroccan families. Can J Neurol Sci 34, 421-426.

102 REFERENCES

Bowes, T., and Gupta, R.S. (2008). Novel mitochondrial extensions provide evidence for a link between microtubule-directed movement and mitochondrial fission. Biochem Biophys Res Commun 376, 40-45. Bowes, T.J., and Gupta, R.S. (2005). Induction of mitochondrial fusion by cysteine- alkylators ethacrynic acid and N-ethylmaleimide. J Cell Physiol 202, 796-804. Brambillasca, S., Yabal, M., Makarow, M., and Borgese, N. (2006). Unassisted translocation of large polypeptide domains across phospholipid bilayers. J Cell Biol 175, 767-777. Braun, P., Hu, Y., Shen, B., Halleck, A., Koundinya, M., Harlow, E., and LaBaer, J. (2002). Proteome-scale purification of human proteins from bacteria. Proc Natl Acad Sci U S A 99, 2654-2659. Cassidy-Stone, A., Chipuk, J.E., Ingerman, E., Song, C., Yoo, C., Kuwana, T., Kurth, M.J., Shaw, J.T., Hinshaw, J.E., Green, D.R., et al. (2008). Chemical inhibition of the mitochondrial division dynamin reveals its role in Bax/Bak-dependent mitochondrial outer membrane permeabilization. Dev Cell 14, 193-204. Castro, F.A., Herdeiro, R.S., Panek, A.D., Eleutherio, E.C., and Pereira, M.D. (2007). Menadione stress in Saccharomyces cerevisiae strains deficient in the glutathione transferases. Biochim Biophys Acta 1770, 213-220. Chan, D.C. (2006a). Mitochondria: dynamic organelles in disease, aging, and development. Cell 125, 1241-1252. Chan, D.C. (2006b). Mitochondrial fusion and fission in mammals. Annu Rev Cell Dev Biol 22, 79-99. Chen, F., and Cushion, M.T. (1994). Use of fluorescent probes to investigate the metabolic state of Pneumocystis carinii mitochondria. J Eukaryot Microbiol 41, 79S. Chung, K.W., Kim, S.M., Sunwoo, I.N., Cho, S.Y., Hwang, S.J., Kim, J., Kang, S.H., Park, K.D., Choi, K.G., Choi, I.S., et al. (2008). A novel GDAP1 Q218E mutation in autosomal dominant Charcot-Marie-Tooth disease. J Hum Genet 53, 360-364. Chung, S.H., Chung, S.M., Lee, J.Y., Kim, S.R., Park, K.S., and Chung, J.H. (1999). The biological significance of non-enzymatic reaction of menadione with plasma thiols: enhancement of menadione-induced cytotoxicity to platelets by the presence of blood plasma. FEBS Lett 449, 235-240. Claramunt, R., Pedrola, L., Sevilla, T., Lopez de Munain, A., Berciano, J., Cuesta, A., Sanchez-Navarro, B., Millan, J.M., Saifi, G.M., Lupski, J.R., et al. (2005). Genetics of Charcot-Marie-Tooth disease type 4A: mutations, inheritance, phenotypic variability, and founder effect. J Med Genet 42, 358-365. Cohen, A.J., Smith, J.N., and Turbert, H. (1964). Comparative detoxication. 10. The enzymic conjugation of chloro compounds with glutathione in locusts and other insects. Biochem J 90, 457-464. Cory, S., and Adams, J.M. (2002). The Bcl2 family: regulators of the cellular life-or-death switch. Nat Rev Cancer 2, 647-656. Costa, V., and Moradas-Ferreira, P. (2001). Oxidative stress and signal transduction in Saccharomyces cerevisiae: insights into ageing, apoptosis and diseases. Mol Aspects Med 22, 217-246.

103 REFERENCES

Criddle, D.N., Gillies, S., Baumgartner-Wilson, H.K., Jaffar, M., Chinje, E.C., Passmore, S., Chvanov, M., Barrow, S., Gerasimenko, O.V., Tepikin, A.V., et al. (2006). Menadione-induced reactive oxygen species generation via redox cycling promotes apoptosis of murine pancreatic acinar cells. J Biol Chem 281, 40485-40492. Cuesta, A., Pedrola, L., Sevilla, T., Garcia-Planells, J., Chumillas, M.J., Mayordomo, F., LeGuern, E., Marin, I., Vilchez, J.J., and Palau, F. (2002). The gene encoding ganglioside-induced differentiation-associated protein 1 is mutated in axonal Charcot- Marie-Tooth type 4A disease. Nat Genet 30, 22-25. Davies, V.J., Hollins, A.J., Piechota, M.J., Yip, W., Davies, J.R., White, K.E., Nicols, P.P., Boulton, M.E., and Votruba, M. (2007). Opa1 deficiency in a mouse model of autosomal dominant optic atrophy impairs mitochondrial morphology, optic nerve structure and visual function. Hum Mol Genet 16, 1307-1318. de Brito, O.M., and Scorrano, L. (2008). Mitofusin 2: a mitochondria-shaping protein with signaling roles beyond fusion. Antioxid Redox Signal 10, 621-633. De Sandre-Giovannoli, A., Chaouch, M., Boccaccio, I., Bernard, R., Delague, V., Grid, D., Vallat, J.M., Levy, N., and Megarbane, A. (2003). Phenotypic and genetic exploration of severe demyelinating and secondary axonal neuropathies resulting from GDAP1 nonsense and splicing mutations. J Med Genet 40, e87. Delettre, C., Lenaers, G., Griffoin, J.M., Gigarel, N., Lorenzo, C., Belenguer, P., Pelloquin, L., Grosgeorge, J., Turc-Carel, C., Perret, E., et al. (2000). Nuclear gene OPA1, encoding a mitochondrial dynamin-related protein, is mutated in dominant optic atrophy. Nat Genet 26, 207-210. Detmer, S.A., and Chan, D.C. (2007). Functions and dysfunctions of mitochondrial dynamics. Nat Rev Mol Cell Biol 8, 870-879. Di Maria, E., Gulli, R., Balestra, P., Cassandrini, D., Pigullo, S., Doria-Lamba, L., Bado, M., Schenone, A., Ajmar, F., Mandich, P., et al. (2004). A novel mutation of GDAP1 associated with Charcot-Marie-Tooth disease in three Italian families: evidence for a founder effect. J Neurol Neurosurg Psychiatry 75, 1495-1498. DiMauro, S., and Schon, E.A. (2008). Mitochondrial disorders in the nervous system. Annu Rev Neurosci 31, 91-123. Dimmer, K.S., and Scorrano, L. (2006). (De)constructing mitochondria: what for? Physiology (Bethesda) 21, 233-241. Edgar, J.M., and Garbern, J. (2004). The myelinated axon is dependent on the myelinating cell for support and maintenance: molecules involved. J Neurosci Res 76, 593-598. Edwards, A.M., Arrowsmith, C.H., Christendat, D., Dharamsi, A., Friesen, J.D., Greenblatt, J.F., and Vedadi, M. (2000). Protein production: feeding the crystallographers and NMR spectroscopists. Nat Struct Biol 7 Suppl, 970-972. Egan, B., Beilharz, T., George, R., Isenmann, S., Gratzer, S., Wattenberg, B., and Lithgow, T. (1999). Targeting of tail-anchored proteins to yeast mitochondria in vivo. FEBS Lett 451, 243-248. Engelman, D.M., Steitz, T.A., and Goldman, A. (1986). Identifying nonpolar transbilayer helices in amino acid sequences of membrane proteins. Annu Rev Biophys Biophys Chem 15, 321-353.

104 REFERENCES

Fernandez-Canon, J.M., and Penalva, M.A. (1998). Characterization of a fungal maleylacetoacetate isomerase gene and identification of its human homologue. J Biol Chem 273, 329-337. Frank, S. (2006). Dysregulation of mitochondrial fusion and fission: an emerging concept in neurodegeneration. Acta Neuropathol 111, 93-100. Frova, C. (2006). Glutathione transferases in the genomics era: New insights and perspectives. Biomol Eng 23, 149-169. Gandre-Babbe, S., and van der Bliek, A.M. (2008). The Novel Tail-anchored Membrane Protein Mff Controls Mitochondrial and Peroxisomal Fission in Mammalian Cells. Mol Biol Cell 19, 2402-2412. Gant, T.W., Rao, D.N., Mason, R.P., and Cohen, G.M. (1988). Redox cycling and sulphydryl arylation; their relative importance in the mechanism of quinone cytotoxicity to isolated hepatocytes. Chem Biol Interact 65, 157-173. Georgiou, D.M., Nicolaou, P., Chitayat, D., Koutsou, P., Babul-Hirji, R., Vajsar, J., Murphy, J., and Christodoulou, K. (2006). A novel GDAP1 mutation 439delA is associated with autosomal recessive CMT disease. Can J Neurol Sci 33, 311-316. Gerasimenko, J.V., Gerasimenko, O.V., Palejwala, A., Tepikin, A.V., Petersen, O.H., and Watson, A.J. (2002). Menadione-induced apoptosis: roles of cytosolic Ca(2+) elevations and the mitochondrial permeability transition pore. J Cell Sci 115, 485-497. Goto, S., Kawakatsu, M., Izumi, S.I., Urata, Y., Kageyama, K., Ihara, Y., Koji, T., and Kondo, T. (2009). Glutathione S-transferase Pi localizes in mitochondria and protects against oxidative stress. Free Radic Biol Med. Habib, S.J., Vasiljev, A., Neupert, W., and Rapaport, D. (2003). Multiple functions of tail- anchor domains of mitochondrial outer membrane proteins. FEBS Lett 555, 511-515. Habig, W.H., Pabst, M.J., and Jakoby, W.B. (1974). Glutathione S-transferases. The first enzymatic step in mercapturic acid formation. J Biol Chem 249, 7130-7139. Hammarstrom, M., Hellgren, N., van Den Berg, S., Berglund, H., and Hard, T. (2002). Rapid screening for improved solubility of small human proteins produced as fusion proteins in Escherichia coli. Protein Sci 11, 313-321. Hayes, J.D., Flanagan, J.U., and Jowsey, I.R. (2005). Glutathione transferases. Annu Rev Pharmacol Toxicol 45, 51-88. Henderson, M.P., Billen, L.P., Kim, P.K., and Andrews, D.W. (2007). Cell-free analysis of tail-anchor protein targeting to membranes. Methods 41, 427-438. Herzig, S., and Martinou, J.C. (2008). Mitochondrial dynamics: to be in good shape to survive. Curr Mol Med 8, 131-137. Hoppins, S., Lackner, L., and Nunnari, J. (2007). The machines that divide and fuse mitochondria. Annu Rev Biochem 76, 751-780. Isenmann, S., Khew-Goodall, Y., Gamble, J., Vadas, M., and Wattenberg, B.W. (1998). A splice-isoform of vesicle-associated membrane protein-1 (VAMP-1) contains a mitochondrial targeting signal. Mol Biol Cell 9, 1649-1660. Jessen, K.R. (2004). Glial cells. Int J Biochem Cell Biol 36, 1861-1867. Jessen, K.R., and Mirsky, R. (2005). The origin and development of glial cells in peripheral nerves. Nat Rev Neurosci 6, 671-682.

105 REFERENCES

Kabzinska, D., Kochanski, A., Drac, H., Rowinska-Marcinska, K., Ryniewicz, B., Pedrola, L., Palau, F., and Hausmanowa-Petrusewicz, I. (2006). A novel Met116Thr mutation in the GDAP1 gene in a Polish family with the axonal recessive Charcot-Marie-Tooth type 4 disease. J Neurol Sci 241, 7-11. Kabzinska, D., Kochanski, A., Drac, H., Ryniewicz, B., Rowinska-Marcinska, K., and Hausmanowa-Petrusewicz, I. (2005). Autosomal recessive axonal form of Charcot- Marie-Tooth Disease caused by compound heterozygous 3'-splice site and Ser130Cys mutation in the GDAP1 gene. Neuropediatrics 36, 206-209. Kabzinska, D., Saifi, G.M., Drac, H., Rowinska-Marcinska, K., Hausmanowa-Petrusewicz, I., Kochanski, A., and Lupski, J.R. (2007). Charcot-Marie-Tooth disease type 4C4 caused by a novel Pro153Leu substitution in the GDAP1 gene. Acta Myol 26, 108- 111. Kanaoka, Y., Ago, H., Inagaki, E., Nanayama, T., Miyano, M., Kikuno, R., Fujii, Y., Eguchi, N., Toh, H., Urade, Y., et al. (1997). Cloning and crystal structure of hematopoietic prostaglandin D synthase. Cell 90, 1085-1095. Karbowski, M., Norris, K.L., Cleland, M.M., Jeong, S.Y., and Youle, R.J. (2006). Role of Bax and Bak in mitochondrial morphogenesis. Nature 443, 658-662. Kiefel, B.R., Gilson, P.R., and Beech, P.L. (2006). Cell biology of mitochondrial dynamics. Int Rev Cytol 254, 151-213. Kim, I., Rodriguez-Enriquez, S., and Lemasters, J.J. (2007). Selective degradation of mitochondria by mitophagy. Arch Biochem Biophys 462, 245-253. Knott, A.B., Perkins, G., Schwarzenbacher, R., and Bossy-Wetzel, E. (2008). Mitochondrial fragmentation in neurodegeneration. Nat Rev Neurosci 9, 505-518. Kuroda, R., Ikenoue, T., Honsho, M., Tsujimoto, S., Mitoma, J.Y., and Ito, A. (1998). Charged amino acids at the carboxyl-terminal portions determine the intracellular locations of two isoforms of cytochrome b5. J Biol Chem 273, 31097-31102. Legesse-Miller, A., Massol, R.H., and Kirchhausen, T. (2003). Constriction and Dnm1p recruitment are distinct processes in mitochondrial fission. Mol Biol Cell 14, 1953- 1963. Li, H., Chen, Y., Jones, A.F., Sanger, R.H., Collis, L.P., Flannery, R., McNay, E.C., Yu, T., Schwarzenbacher, R., Bossy, B., et al. (2008). Bcl-xL induces Drp1-dependent synapse formation in cultured hippocampal neurons. Proc Natl Acad Sci U S A 105, 2169-2174. Li, M.Z., and Elledge, S.J. (2007). Harnessing homologous recombination in vitro to generate recombinant DNA via SLIC. Nat Methods 4, 251-256. Li, Z., Okamoto, K., Hayashi, Y., and Sheng, M. (2004). The importance of dendritic mitochondria in the morphogenesis and plasticity of spines and synapses. Cell 119, 873-887. Lin, R., Mamane, Y., and Hiscott, J. (1999). Structural and functional analysis of interferon regulatory factor 3: localization of the transactivation and autoinhibitory domains. Mol Cell Biol 19, 2465-2474. Liu, H., Nakagawa, T., Kanematsu, T., Uchida, T., and Tsuji, S. (1999). Isolation of 10 differentially expressed cDNAs in differentiated Neuro2a cells induced through controlled expression of the GD3 synthase gene. J Neurochem 72, 1781-1790.

106 REFERENCES

Lorenz, H., Hailey, D.W., and Lippincott-Schwartz, J. (2006). Fluorescence protease protection of GFP chimeras to reveal protein topology and subcellular localization. Nat Methods 3, 205-210. Mannervik, B. (1995). Human Glutathione Transferases: Classification, Tissue Distribution, Structure, and Functional Properties (Sweden, European Commission). Marco, A., Cuesta, A., Pedrola, L., Palau, F., and Marin, I. (2004). Evolutionary and structural analyses of GDAP1, involved in Charcot-Marie-Tooth disease, characterize a novel class of glutathione transferase-related genes. Mol Biol Evol 21, 176-187. Martinou, J.C., Desagher, S., and Antonsson, B. (2000). Cytochrome c release from mitochondria: all or nothing. Nat Cell Biol 2, E41-43. Mattenberger, Y., James, D.I., and Martinou, J.C. (2003). Fusion of mitochondria in mammalian cells is dependent on the mitochondrial inner membrane potential and independent of microtubules or actin. FEBS Lett 538, 53-59. McBride, H.M., Neuspiel, M., and Wasiak, S. (2006). Mitochondria: more than just a powerhouse. Curr Biol 16, R551-560. Meeusen, S., McCaffery, J.M., and Nunnari, J. (2004). Mitochondrial fusion intermediates revealed in vitro. Science 305, 1747-1752. Meyer, D.J., Coles, B., Pemble, S.E., Gilmore, K.S., Fraser, G.M., and Ketterer, B. (1991). Theta, a new class of glutathione transferases purified from rat and man. Biochem J 274 ( Pt 2), 409-414. Nechushtan, A., Smith, C.L., Hsu, Y.T., and Youle, R.J. (1999). Conformation of the Bax C-terminus regulates subcellular location and cell death. EMBO J 18, 2330-2341. Nelis, E., Erdem, S., Van Den Bergh, P.Y., Belpaire-Dethiou, M.C., Ceuterick, C., Van Gerwen, V., Cuesta, A., Pedrola, L., Palau, F., Gabreels-Festen, A.A., et al. (2002). Mutations in GDAP1: autosomal recessive CMT with demyelination and axonopathy. Neurology 59, 1865-1872. Neupert, W., and Herrmann, J.M. (2007). Translocation of proteins into mitochondria. Annu Rev Biochem 76, 723-749. Nicotera, P., Bellomo, G., and Orrenius, S. (1992). Calcium-mediated mechanisms in chemically induced cell death. Annu Rev Pharmacol Toxicol 32, 449-470. Niemann, A., Berger, P., and Suter, U. (2006). Pathomechanisms of mutant proteins in Charcot-Marie-Tooth disease. Neuromolecular Med 8, 217-242. Niemann, A., Ruegg, M., La Padula, V., Schenone, A., and Suter, U. (2005). Ganglioside- induced differentiation associated protein 1 is a regulator of the mitochondrial network: new implications for Charcot-Marie-Tooth disease. J Cell Biol 170, 1067- 1078. Olichon, A., Emorine, L.J., Descoins, E., Pelloquin, L., Brichese, L., Gas, N., Guillou, E., Delettre, C., Valette, A., Hamel, C.P., et al. (2002). The human dynamin-related protein OPA1 is anchored to the mitochondrial inner membrane facing the inter- membrane space. FEBS Lett 523, 171-176. Panne, D., McWhirter, S.M., Maniatis, T., and Harrison, S.C. (2007). Interferon regulatory factor 3 is regulated by a dual phosphorylation-dependent switch. J Biol Chem 282, 22816-22822.

107 REFERENCES

Parman, Y., Battaloglu, E., Baris, I., Bilir, B., Poyraz, M., Bissar-Tadmouri, N., Williams, A., Ammar, N., Nelis, E., Timmerman, V., et al. (2004). Clinicopathological and genetic study of early-onset demyelinating neuropathy. Brain 127, 2540-2550. Pedrola, L., Espert, A., Valdes-Sanchez, T., Sanchez-Piris, M., Sirkowski, E.E., Scherer, S.S., Farinas, I., and Palau, F. (2008). Cell expression of GDAP1 in the nervous system and pathogenesis of Charcot-Marie-Tooth type 4A disease. Journal of cellular and molecular medicine 12, 679-689. Pedrola, L., Espert, A., Wu, X., Claramunt, R., Shy, M.E., and Palau, F. (2005). GDAP1, the protein causing Charcot-Marie-Tooth disease type 4A, is expressed in neurons and is associated with mitochondria. Hum Mol Genet 14, 1087-1094. Rojo, M., Legros, F., Chateau, D., and Lombes, A. (2002). Membrane topology and mitochondrial targeting of mitofusins, ubiquitous mammalian homologs of the transmembrane GTPase Fzo. J Cell Sci 115, 1663-1674. Rougeot, C., Chabrier, S., Camdessanche, J.P., Prieur, F., d'Anjou, M.C., and Latour, P. (2008). Clinical, electrophysiological and genetic studies of two families with mutations in the GDAP1 gene. Neuropediatrics 39, 184-187. Rüegg, M. (2009). Functional analysis of Ganglioside-induced differentiation-associated protein 1 (GDAP1), a mitochondrial dynamics regulator, in neural cells. In Institute of Cell Biology (Zurich, ETH Zurich). Schrader, M. (2006). Shared components of mitochondrial and peroxisomal division. Biochim Biophys Acta 1763, 531-541. Senderek, J., Bergmann, C., Ramaekers, V.T., Nelis, E., Bernert, G., Makowski, A., Zuchner, S., De Jonghe, P., Rudnik-Schoneborn, S., Zerres, K., et al. (2003). Mutations in the ganglioside-induced differentiation-associated protein-1 (GDAP1) gene in intermediate type autosomal recessive Charcot-Marie-Tooth neuropathy. Brain 126, 642-649. Setoguchi, K., Otera, H., and Mihara, K. (2006). Cytosolic factor- and TOM-independent import of C-tail-anchored mitochondrial outer membrane proteins. Embo J 25, 5635- 5647. Sevilla, T., Cuesta, A., Chumillas, M.J., Mayordomo, F., Pedrola, L., Palau, F., and Vilchez, J.J. (2003). Clinical, electrophysiological and morphological findings of Charcot-Marie-Tooth neuropathy with vocal cord palsy and mutations in the GDAP1 gene. Brain 126, 2023-2033. Sevilla, T., Jaijo, T., Nauffal, D., Collado, D., Chumillas, M.J., Vilchez, J.J., Muelas, N., Bataller, L., Domenech, R., Espinos, C., et al. (2008). Vocal cord paresis and diaphragmatic dysfunction are severe and frequent symptoms of GDAP1-associated neuropathy. Brain 131, 3051-3061. Seyfried, J., Soldner, F., Schulz, J.B., Klockgether, T., Kovar, K.A., and Wullner, U. (1999). Differential effects of L-buthionine sulfoximine and ethacrynic acid on glutathione levels and mitochondrial function in PC12 cells. Neurosci Lett 264, 1-4. Sheehan, D., Meade, G., Foley, V.M., and Dowd, C.A. (2001). Structure, function and evolution of glutathione transferases: implications for classification of non- mammalian members of an ancient enzyme superfamily. Biochem J 360, 1-16.

108 REFERENCES

Shield, A.J., Murray, T.P., and Board, P.G. (2006). Functional characterisation of ganglioside-induced differentiation-associated protein 1 as a glutathione transferase. Biochem Biophys Res Commun 347, 859-866. Shih, Y.P., Kung, W.M., Chen, J.C., Yeh, C.H., Wang, A.H., and Wang, T.F. (2002). High-throughput screening of soluble recombinant proteins. Protein Sci 11, 1714- 1719. Smirnova, E., Shurland, D.L., Ryazantsev, S.N., and van der Bliek, A.M. (1998). A human dynamin-related protein controls the distribution of mitochondria. J Cell Biol 143, 351-358. Smyth, D.R., Mrozkiewicz, M.K., McGrath, W.J., Listwan, P., and Kobe, B. (2003). Crystal structures of fusion proteins with large-affinity tags. Protein Sci 12, 1313- 1322. Soltys, B.J., and Gupta, R.S. (1994). Changes in mitochondrial shape and distribution induced by ethacrynic acid and the transient formation of a mitochondrial reticulum. J Cell Physiol 159, 281-294. Spearman, M.E., Prough, R.A., Estabrook, R.W., Falck, J.R., Manna, S., Leibman, K.C., Murphy, R.C., and Capdevila, J. (1985). Novel glutathione conjugates formed from epoxyeicosatrienoic acids (EETs). Arch Biochem Biophys 242, 225-230. Squire, L.R., J.R. Roberts, N.C. Spitzer, M.J. Zigmond, S.K. McConnell, and F.E. Bloom (2002). Fundamental Neuroscience. Academic Press. Stefanovic, S., and Hegde, R.S. (2007). Identification of a targeting factor for posttranslational membrane protein insertion into the ER. Cell 128, 1147-1159. Stevens, R.C. (2000). Design of high-throughput methods of protein production for structural biology. Structure 8, R177-185. Stojkovic, T., Latour, P., Viet, G., de Seze, J., Hurtevent, J.F., Vandenberghe, A., and Vermersch, P. (2004). Vocal cord and diaphragm paralysis, as clinical features of a French family with autosomal recessive Charcot-Marie-Tooth disease, associated with a new mutation in the GDAP1 gene. Neuromuscul Disord 14, 261-264. Suen, D.F., Norris, K.L., and Youle, R.J. (2008). Mitochondrial dynamics and apoptosis. Genes Dev 22, 1577-1590. Suter, U., and Scherer, S.S. (2003). Disease mechanisms in inherited neuropathies. Nat Rev Neurosci 4, 714-726. Suzuki, M., Youle, R.J., and Tjandra, N. (2000). Structure of Bax: coregulation of dimer formation and intracellular localization. Cell 103, 645-654. Tan, K.L., Chelvanayagam, G., Parker, M.W., and Board, P.G. (1996). Mutagenesis of the active site of the human Theta-class glutathione transferase GSTT2-2: catalysis with different substrates involves different residues. Biochem J 319 ( Pt 1), 315-321. Tew, K.D., O'Brien, M., Laing, N.M., and Shen, H. (1998). Coordinate changes in expression of protective genes in drug-resistant cells. Chem Biol Interact 111-112, 199-211. Thom, R., Dixon, D.P., Edwards, R., Cole, D.J., and Lapthorn, A.J. (2001). The structure of a zeta class glutathione S-transferase from Arabidopsis thaliana: characterisation of a GST with novel active-site architecture and a putative role in tyrosine catabolism. J Mol Biol 308, 949-962.

109 REFERENCES

Twig, G., Elorza, A., Molina, A.J., Mohamed, H., Wikstrom, J.D., Walzer, G., Stiles, L., Haigh, S.E., Katz, S., Las, G., et al. (2008a). Fission and selective fusion govern mitochondrial segregation and elimination by autophagy. EMBO J 27, 433-446. Twig, G., Hyde, B., and Shirihai, O.S. (2008b). Mitochondrial fusion, fission and autophagy as a quality control axis: the bioenergetic view. Biochim Biophys Acta 1777, 1092-1097. Wagner, K.M., Ruegg, M., Niemann, A., and Suter, U. (2009). Targeting and function of the mitochondrial fission factor GDAP1 are dependent on its tail-anchor. PLoS ONE 4, e5160. Watanabe, N., Dickinson, D.A., Liu, R.M., and Forman, H.J. (2004). Quinones and glutathione metabolism. Methods Enzymol 378, 319-340. Waterham, H.R., Koster, J., van Roermund, C.W., Mooyer, P.A., Wanders, R.J., and Leonard, J.V. (2007). A lethal defect of mitochondrial and peroxisomal fission. N Engl J Med 356, 1736-1741. Wattenberg, B., and Lithgow, T. (2001). Targeting of C-terminal (tail)-anchored proteins: understanding how cytoplasmic activities are anchored to intracellular membranes. Traffic 2, 66-71. Wattenberg, B.W., Clark, D., and Brock, S. (2007). An artificial mitochondrial tail signal/anchor sequence confirms a requirement for moderate hydrophobicity for targeting. Biosci Rep 27, 385-401. Widersten, B.M.a.M. (1995). Human Glutathione Transferases: Classification, tissue distribution, structure, and functional properties. In advances of drug metabolism in man, G.N.F. g.m. Pacifici, ed., pp. 409-457. Winayanuwattikun, P., and Ketterman, A.J. (2005). An electron-sharing network involved in the catalytic mechanism is functionally conserved in different glutathione transferase classes. J Biol Chem 280, 31776-31782. Xin, B., Puffenberger, E., Nye, L., Wiznitzer, M., and Wang, H. (2008). A novel mutation in the GDAP1 gene is associated with autosomal recessive Charcot-Marie-Tooth disease in an Amish family. Clin Genet 74, 274-278. Yaffe, M.P. (1999). The machinery of mitochondrial inheritance and behavior. Science 283, 1493-1497. Yoon, Y. (2004). Sharpening the scissors: mitochondrial fission with aid. Cell Biochem Biophys 41, 193-206. Zuchner, S., De Jonghe, P., Jordanova, A., Claeys, K.G., Guergueltcheva, V., Cherninkova, S., Hamilton, S.R., Van Stavern, G., Krajewski, K.M., Stajich, J., et al. (2006). Axonal neuropathy with optic atrophy is caused by mutations in mitofusin 2. Ann Neurol 59, 276-281. Zuchner, S., Mersiyanova, I.V., Muglia, M., Bissar-Tadmouri, N., Rochelle, J., Dadali, E.L., Zappia, M., Nelis, E., Patitucci, A., Senderek, J., et al. (2004). Mutations in the mitochondrial GTPase mitofusin 2 cause Charcot-Marie-Tooth neuropathy type 2A. Nat Genet 36, 449-451. Zuchner, S., and Vance, J.M. (2006). Molecular genetics of autosomal-dominant axonal Charcot-Marie-Tooth disease. Neuromolecular Med 8, 63-74.

110 APPENDIX

6. Appendix

Table 6.1. Primer used to generate constructs used in this study # Construct Sequence 1 Bam_GDAP1-for 5’-CGGATCCATGGCTCGGAGGCAGGAC-3’ 2 Sal_GDAP1-rev 5’-CGTCGACTAGAAATAATTTGGTCTGG-3’ 3 hGDAP1cDNA-for 5’-CAAGATGGCTGAGAGGC-3’ 4 hGDAP1cDNA-rev 5’-GATCCCAACAAACCTAG-3’ 5 Bam_hGDAP1-for 5’-CGGATCCATGGCTGAGAGGCAGGAAG-3’ 6 FLAG_hGDAP1-rev 5’-CTACTTGTCGTCGTCCTTGTAGTCTCTGGGTCTAAATGCTAA-3’ 7 R120W-for 5’-GGAAGCATGTATTACCCATGGGTGCAGCATTACCGAG-3’ 8 R120W-rev 5’-CTCGGTAATGCTGCACCCATGGGTAATACATGCTTCC-3’ 9 T157P-for 5’-GATCCCCGCTTATGCACCCACAAGGATTCGCAGCC-3’ 10 T157P-rev 5’-GGCTGCGAATCCTTGTGGGTGCATAAGCGGGGATC-3’ 11 Q218E-for 5’-GAGTTGGAGAAGGTCTTGGATGAGGTGGAAACTGAATTACAAAG-3’ 12 Q218E-rev 5’-CTTTGTAATTCAGTTTCCACCTCATCCAAGACCTTCTCCAACTC-3’ 13 G271R-for 5’-GCAAGACGAAACTGGGGACATAGAAAGAGACCGAACTTGG-3’ 14 G271R-rev 5’-CCAAGTTCGGTCTCTTTCTATGTCCCCAGTTTCGTCTTGC-3’ 15 L239F-for 5’-GGAGGGCAACCAGCCTTGGTTCTGTGGAGAATCCTTTACAC-3’ 16 L239F-rev 5’-GTGTAAAGGATTCTCCACAGAACCAAGGCTGGTTGCCCTCC-3’ 17 S130C-for 5’-CCGAGAGCTTCTTGACTGCTTGCCAATGGATGCTTACAC-3’ 18 S130C-rev 5’-GTGTAAGCATCCATTGGCAAGCAGTCAAGAAGCTCTCGG-3’ 19 M116T-for 5’-GCCCGATGAAGGAAGCACGTATTACCCACGGGTGCAGC-3’ 20 M116T-rev 5’-GCTGCACCCGTGGGTAATACGTGCTTCCTTCATCGGGC-3’ 21 D149Y-for 5’-GCATCCTGAGCTCACTGTGTATTCCATGATCCCCGCTTATGC-3’ 22 D149Y-rev 5’-GCATAAGCGGGGATCATGGAATACACAGTGAGCTCAGGATGC-3’ 23 FLAG+8_1-for 5’-GATCTGACTACAAGGACGACGACGACAAGTAGGGCC-3’ 24 FLAG+8_2-rev 5’-CTACTTGTCGTCGTCGTCCTTGTAGTC-3’ 25 TMD+1-rev 5’-CATTCCTACTACAAGCAAACCAACCAC-3’ 26 TMD+1-for 5’-GGTTTGCTTGTAGTAGGAATGGGATATTTTGC-3’ 27 TMD+3-rev 5’-CATTCCTACTACTGCTACAAGCAAACCAACCAC-3’ 28 TMD+3-for 5’-GGTTTGCTTGTAGCAGTAGTAGGAATGGGATATTTTGC-3’ 29 TMD+5-rev 5’-CATTCCTACTACCGCTGCTACTACAAGCAAACCAACCAC-3’ 30 TMD+5-for 5’-GGTTTGCTGTAGTAGCAGCGGTAGTAGGAATGGGATATTTTGC-3’ 31 TMD-1-rev 5’-CATTCCAAGCAAACCAACCAC-3’ 32 TMD-1-for 5’-GGTTTGCTTGGAATGGGATATTTTGC-3’ 33 HD1scr_1-rev 5’-GTGTAGAATTGAAAACGATCTTGTTAAATGTTTTTC-3’ 34 HD1scr_1-for 5’-GATCGGAGCGGTGTTATTAAACCGGGTGGCCAAAAAAAGGGC-3’ 35 HD1scr_2-rev 5’-CTCCGATCGCTGGCACTAAATGTGTAGAATTGAAAACGATC-3’ 36 HD1scr_2-for 5’-CAATTCTACACATTTAGTGCCAGCGATCGGAGCGGTGTTATTAAACC-3’ 37 GDAP1hy_1-for 5’-CCAAAAGTTCTTGGCAGCACCGCTGTGGTTGGTTTGCTTG-3’ 38 GDAP1hy_1-rev 5’-CAAGCAAACCAACCACAGCGGTGCTGCCAAGAACTTTTGG-3’ 39 GDAP1hy_2-for 5’-GGTTGGTTTGCTTGTAGGAAGCGGATATTTTGCATTTATGC-3’ 40 GDAP1hy_2-rev 5’-GCATAAATGCAAAATATCCGCTTCCTACAAGCAAACCAACC-3’ 41 GDAP1hy_3-for 5’-GGTTGGTTTGCTTGTAGGAAGCGGATATGCTGCATATATGC-3’ 42 GDAP1hy_3-rev 5’-GCATATATGCAGCATATCCGCTTCCTACAAGCAAACCAACC-3’ 43 TMDscr_1-rev 5’-CACAAAAAGTGCCATGCCAACTTTTGGGGCCCTTTTTTTGGC-3’ 44 TMDscr_1-for 5’-CTTGTGGGACTTTTGGTGAGAAGGAGACTTGGCAGCATG-3’ 45 TMDscr_2-rev 5’-CCAAAATAGCTCATACCGGTCACAAAAAGTGCCATGCCAAC-3’ 46 TMDscr_2-for 5’-CCAAAATAGCTCATACCGGTCACAAAAAGTGCCATGCCAAC-3’ 47 delN-terminus-for 5’-CGGATCCATGCCGCCGGACAAGGAGGTCCAC-3’ 48 delGST-N_1-for 5’-GCCGGCGAACGTGGCGAGAA-3’ 49 delGST-N_1-rev 5’-CTCGAGCTTGTCCGGCGGGCCTTCAA-3’ 50 delGST-N_2-for 5’-CTCGAGACTTTCCTGGATGAAAGAACA-3’ 51 del
4-helix_1-rev 5’-CTCGAGCTGCTCAAGATAATCAATGATC-3’ 52 del
4-helix_2-for 5’-CTCGAGTTGCATCCTGAGCTCACTGTG-3’ 53 delInterdomain_1-rev 5’-CTCGAGAATGCAGCCATGTGTGTAAGC-3’ 54 delInterdomain_2-for 5’-CTCGAGGACAATGTCAAATACTTGAAG-3’ 55 delsInterdomain_1-rev 5’-AGGATTCGCAGCCAGATTGGTAAGTCAAAGCTGCTTGAT-3’ 56 delsInterdomain_2-for 5’-GTCATGATCAAGCAGCTTTGACTTACCAATCTGGCTGCG-3’ 57 delGST-C_1-rev 5’-CTCGAGAATTTTCTTCAAGTATTTGACA-3’ 58 delGST-C_2-for 5’-CTCGAGACATTTAACAAGGTTTTAGGA-3’ 59 delHD1_1-rev 5’-CTCGAGTTTTCTCTTCAAGACACGCTC-3’ 60 delHD1_2-for 5’-CTCGAGCGGGTGGCCAAAAAAAGGGCCCC-3’ 61 delHD1_2-rev 5’-GGCACCCCAGGCTTTACACT-3’ 62 delC-terminus-rev 5’-CTCGAGCTAGCCAAGTCTCCTTCTGAAAAG-3’ 63 FLAG_GDAP1-rev 5’-CTACTTGTCGTCGTCGTCCTTGTAGTCTCTGGGTCTAAGTGCTAA-3’ 64 TMD-GFP-for 5’-GGATCCATGGTGGCCAAAAAAAGGGCCCC-3’ 65 TMD-GFP-rev 5’-GCGGCCGCTTTACTTGTACAGCTCGTCCA-3’ 66 HD1-TMD-GFP-rev 5’-GGATCCATGGCGAAGGTTTTAGGACATGTC-3’

111 APPENDIX

67 Y29A-for 5’-GGTCCACCTCATTCTGGCCCACTGGACGCACTCCTTC-3’ 68 Y29A-rev 5’-GAAGGAGTGCGTCCAGTGGGCCAGAATGAGGTGGACC-3’ 69 S34A-for 5’-GTACCACTGGACGCACGCCTTCAGCTCTCAAAAGG-3’ 70 S34A-rev 5’-CCTTTTGAGAGCTGAAGGCGTGCGTCC-3’ 71 S36A-for 5’-CTGGACGCACTCCTTCGCCTCTCAAAAGGTGCGC-3’ 72 S36A-rev 5’-GCGCACCTTTTGAGAGGCGAAGGAGTGCGTCCAG-3’ 73 S37A-for 5’-GACGCACTCCTTCAGCGCTCAAAAGGTGCGCTTGG-3’ 74 S37A-rev 5’-CCAAGCGCACCTTTTGAGCGCTGAAGGAGTGCGTC-3’ 75 F68S-for 5’-GCACAATGAGCCTTGGAGTATGCGCTTGAACTCAGC-3’ 76 F68S-rev 5’-GCTGAGTTCAAGCGCATACTCCAAGGCTCATTGTGC-3’ 77 K291A-for 5’-CGTGTCTTGAAGAGAAAAACATTTAACGCGGTTTTAGGACATGTC-3’ 78 K291A-rev 5’-GACATGTCCTAAAACCGCGTTAAATGTTTTTCTCTTCAAGACACG-3’ 79 K313A-for 5’-GCATTCCGGGTGGCCGCAAAAAGGGCCCCAAAAGTTC-3’ 80 K313A-rev 5’-GAACTTTTGGGGCCCTTTTTGCGGCCACCCGGAATGC-3’ 81 K314A-for 5’-GCATTCCGGGTGGCCAAAGCAAGGGCCCCAAAAGTTC-3’ 82 K314A-rev 5’-GAACTTTTGGGGCCCTTGCTTTGGCCACCCGGAATGC-3’ 83 R315A-for 5’-GCATTCCGGGTGGCCAAAAAAGCGGCCCCAAAAGTTC-3’ 84 R315A-rev 5’-GAACTTTTGGGGCCGCTTTTTTGGCCACCCGGAATGC-3’ 85 K318A-for 5’-GCCAAAAAAAGGGCCCCAGCAGTTCTTGGCAGCACCC-3’ 86 K318A-rev 5’-GGGTGCTGCCAAGAACTGCTGGGGCCCTTTTTTTGGC-3’ 87 R341A-for 5’-GCATTTATGCTTTTCGCAAGGAGACTTGGCAGCATGATATTAG-3’ 88 R341A-rev 5’-CTAATATCATGCTGCCAAGTCTCCTTGCGAAAAGCATAAATGC-3’ 89 R342A-for 5’-GCATTTATGCTTTTCAGAGCGAGACTTGGCAGCATGATATTAG-3’ 90 R342A-rev 5’-CTAATATCATGCTGCCAAGTCTCGCTCTGAAAAGCATAAATGC-3’ 91 R343A-for 5’-GCATTTATGCTTTTCAGAAGGAGGGCACTTGGCAGCATGATATTAG-3’ 92 R343A-rev 5’-CTAATATCATGCTGCCAAGTGCCCTTCTGAAAAGCATAAATGC-3’ 93 K314A_R315A-for 5’-GCATTCCGGGTGGCCAAAGCAGCGGCCCCAAAAGTTC-3’ 94 K314_R315A-rev 5’-GAACTTTTGGGGCCGCTGCTTTGGCCACCCGGAATGC-3’ 95 K313_K314_R315A-for 5’-GCATTCCGGTGTGCCGCAGCAGCGGCCCCAAAAGTTC-3’ 96 K313_K314_R315A-rev 5’-GAACTTTTGGGGCCGCTGCTGCGGCCACCCGGAATGC-3’ 97 R341_R342A-for 5’-GCATTTATGCTTTTCGCAGCGAGACTTGGCAGCATGATATTAG-3’ 98 R341_R342A-rev 5’-CTAATATCATGCTGCCAAGTCTCGCTGCGAAAAGCATAAATGC-3’ 99 R341_R342_R343A-for 5’-GCATTTATGCTTTTCGCAGCGGCACTTGGCAGCATGATATTAG-3’ 100 R341_R342_R343A-rev 5’-CTAATATCATGCTGCCAAGTGCCGCTGCGAAAAGCATAAATGC-3’ 101 Vamp1B_90-for 5’-CCTCCGGAGAGGAACAGAC-3’ 102 Vamp1B_568-rev 5’-CTGCATTAGGCAAGGAGGAG-3’ 103 Vamp_GDAP1-for 5’-CAAAAAAAGGGCCCCAAAAATGATGATCATGCTGGGAG-3’ 104 GDAP1_Vamp-rev 5’-CTCCCAGCATGATCATCATTTTTGGGGCCCTTTTTTTG-3’ 105 hOmb5_102-for 5’-CGGTGGAGAGGCAGTATGTC-3’ 106 hOmb5_601-rev 5’-GGATGCACTTTCTAACTTCAGCA-3’ 107 hOmb5_GDAP1-for 5’-CAAAAAAAGGGCCCCAAAAAGTTGCTGGGCATATTGG-3’ 108 GDAP1_hOmb5-rev 5’-CCAATATGCCCAGCAACTTTTTGGGGCCCTTTTTTTG-3’ 109 GDAP1_rOmb5_1-rev 5’-CACGATGGGGACAATCCAATATGCCCATTTTGGGGCCCTTTTTTTGGCCAC-3’ 110 GDAP1_rOmb5_2-rev 5’-GACGATACAGGAAACCTATAAGAATAGCACCCACGATGGGGACAATCCAAT-3’ 111 GDAP1_rOmb5_3-rev 5’-GTCGACTCAGGATTTGCTGTCAGCCCAGAAGTGACGATACAGGAAACCTATAAG-3’ 112 Bam_GDAP1L1-for 5’-GGATCCATGGCGACCCCC-3’ 113 Sal_GDAP1L1-rev 5’-CAGCTGCTAGATGTATTTTTTC-3’ 114 GDAP1L1_TMD-for 5’-GGATCCCTGGTCAAACGGAAAC-3’ 115 GDAP1L1_HD1-3-for 5’-GGATCCGGGCAGACGTGCGAG-3’ 116 GDAP1L1_exTMD-for 5’-CGGAAACCGCCAAAAGTTCTTGGCAGCACC-3’ 117 GDAP1L1_exTMD-rev 5’-GCCAAGAACTTTTGGCGGTTTCCGTTTGAC-3’ 118 GDAP1_exTMD-for 5’-AAAAGGGCCCCATCCTTCTTTGGGGCATCC-3’ 119 GDAP1_exTMD-rev 5’-CCCAAAGAAGGATGGGGCCCTTTTTTTGGC-3’ 120 GDAP1L1_exHD2_1-for 5’-GTGCTGCCAACAGCGTTCCGACTGGTCAAAC-3’ 121 GDAP1L1_exHD2_2-for 5’-ATATTAATCTCTGCGGTGCTGCCAACAGCGTTCCGAC-3’ 122 GDAP1L1_exHD2_1-rev 5’-ATTGTTGACATGGCCCAGGACTTTCCGGAAGG-3’ 123 GDAP1L1_exHD2_2-rev 5’-CGCAGAGATTAATATATTGTTGACATGGCCCAGGAC-3’ 124 GDAP1_exHD1_1-for 5’-GTTATCCCCAACGCATTCCGGGTGGCCAAA-3’ 125 GDAP1_exHD1_2-for 5’-CTGCTGTCAGCTTTATCCCCAACGCATTC-3’ 126 GDAP1_exHD1_1-rev 5’-GGTGTGGATGTCTCCTAAAACCTTGTTAAATG-3’ 127 GDAP1_exHD1_2-rev 5’-TGACAGCAGAGTGGTGTGGATGTCGCCAAG-3’ 128 GDAP1_X288-for 5’-GGAATTCGGTAACCGGAGGCAGGACGAGGCGCGGGCC-3’ 129 GDAP1_X288-rev 5’-GGATCCGCGGACCGTTAGTGATGGTGATGGTGATGAGATC-3’ 130 GDAP1_X318_1-rev 5’-TGATGGTGATGTTTTGGGGCCCTTT-3’ 131 GDAP1_X318_2-rev 5’-CGCGGACCGTTAGTGATGGTGATGGTGATGTTTTGGGGCCCT-3’ 132 GDAP1_X16_309-for 5’-CGGATCCATGGTTGAAGGCCCGCCGGAC-3’ 133 GDAP1_X16_309-rev 5’-GGTGATGTGCTGTTGGCAGCACCGC-3’ 134 GDAP1_X16_288-rev 5’-CCGCGGACCGTTAGTGATGGTGATGGTGATGTGCTGTTGG-3’ 135 GDAP1_X288mut_1-rev 5’-TTTTCTCTTCAAGACACG-3’ 136 GDAP1_X288_R282C_1-rev 5’-TTTTCTCTTCAAGAC-3’ 137 GDAP1_X288mut_2-rev 5’-GATGGTGATGTTTTCTCTTCAAG-3’ 138 GDAP1_X288mut_3-rev 5’-GGATCCGCGGACCGTTAGTGATGGTGATGGTGATG-3’ 139 pBDO_MBP-for 5’-CGGGCGCGGATCCTCGAGATGAAAACTGAAGAAGGTAAACTG-3’ 140 MBP_pBDO-rev 5’-CAGTTTACCTTCTTCAGTTTTCATCTCGAGGATCCGCGCCCG-3’ 141 MBP_GDAP1-for 5’-AATTCGAGCTCGGTACCCGGCGGTAACCGGAGGCAGGACGAG-3’

112 APPENDIX

142 GDAP1_MBP-rev 5’-CTCGTCCTGCCTCCGGTTACCGCCGGGTACCGAGCTCGAATT-3’ 143 GDAP1_pMALc-for 5’-CATCACCATCACCATCACTAACGGGGATCCATCGAGGGTAGG-‚ 144 pMALc_GDAP1-rev 5’-CCTACCCTCGATGGATCCCCGTTAGTGATGGTGATGGTGATG-3’ 145 pMAL+_GDAP1-for 5’-CATCACCATCACCATCACTAACCTGAATTCAGTAACCTAACC-3’ 146 GDAP1_MBP+-rev 5’-CTCGTCCTGCCTCCGGTTACCCCTACCCTCGATGGATCCCCG-3’ 147 MBP+_GDAP1-for 5’-CGGGGATCCATCGAGGGTAGGGGTAACCGGAGGCAGGACGAG-3’ 148 pMALc+_GDAP1-rev 5’-GGTTAGGTTACTGAATTCAGGTTAGTGATGGTGATGGTGATG-3’ 149 MBP-pBDO_2-rev 5’-CAGTTTACCTTCTTCAGTTTTCATCTCGAGGATCCGCGCCCGATGGTG-3’ 150 MBP_GDAP1_2-for 5’-AATTCGAGCTCGGTACCCGGCGGTAACCGGAGGCAGGACGAGGCGCGG-3’ 151 GDAP1L1_rec-for 5’-CGGATCCATGGCGACCCCCAACAAC-3’ 152 GDAP1L1_X310_1-rev 5’-GATGGTGATGGGCAAAGCGTCTCTGGACCCTC-3’ 153 GDAP1L1_X310_2-rev 5’-TAGTGATGGTGATGGTGATGGGCAAAGCGTCTC-3’ 154 GDAP1L1_X340_1-rev 5’-GATGGTGATGGGATGGCGGTTTCCGTTTGACCAG-3’ 155 GDAP1L1_X340_2-rev 5’-TAGTGATGGTGATGGTGATGGGATGGCGGTTTCCG-3’ 156 FLAG-hOMb5_for 5’-TAGGATCCATGGACTACAAGGATGACGATGACAAGGCGACTGCGGAAGCTAGCGGC-3 157 GDAP1L1_fl_1-rev 5’-GATGGTGATGGATGTATTTTTTCTTGAGGTAC-3’ 158 GDAP1L1_fl_2-rev 5’-TAGTGATGGTGATGGTGATGGATGTATTTTT-3’

113 ACKNOWLEDGEMENT

Acknowledgement

I would like to acknowledge all the people who contributed to or supported this work in one way or another.

First of all I want to thank Dr. Axel Niemann. Not only for your excellent supervision and your endless patience in teaching me all relevant laboratory skills, answering (sometimes annoying) questions and never declining any help but even more for your unbeatable optimism and confidence in our project during all the ups and downs of the last 3
years.

I would like to thank Prof. Ueli Suter. I’m very grateful for the opportunity to work in his well-equipped unit and for the supervision of this work.

A special thank goes to my “Doctor-Brother” Marcel. Sharing the lab with you was a great pleasure and our discussions about science and politics, especially the german vs swiss part, will remain unforgettable.

I would also like to thank Dr. Imre Berger who was not only willing to participate in my thesis committee but was also an enthusiastic collaborator who motivated me enourmously with his always positive attitude for our “GDAP1”.

A big thank you further goes to Dr. Christoph Bieniossek for the excellent collaboration, the huge amount of work he put into our project and his valuable suggestions for a lot of experiments.

I would also like to thank Dr. Uwe Konietzko, who participated in my thesis committee and was always interested in my work.

Furthermore I would like to thank my current lab members Paris and Nina for the great working atmosphere, for numerous scientific and private discussions during working hours and beyond and for all the fun we had fighting our battle through the PhD.

Many thanks go to Dr. Ned Mantei, the true scientist, for reviewing my thesis and your irreplaceable help for any scientific or computer problem.

Thanks go to former lab members Christian Somandin and Francois Castagner. You were great friends and always ready to help during my first month in the lab.

114 ACKNOWLEDGEMENT

Now there remains the part to thank all people from the Suter unit, the ICB and also the ICB supporting staff. I met so many great people there that I’m afraid I can’t credit each and every one for their contribution to a really nice atmosphere and for so many scientific and private discussions.

Last but not least I’m extremely greatful to Matthias for his constant support and his belief in me during all the ups and downs of the last 3
years. And of course I want to especially thank my parents and my brothers for their steady love and support.

115 CURRICULUM VITAE

Curriculum Vitae

Name Konstanze Marion Wagner Date/Place of birth 14.10.1980 / Arnstadt Citizenship German Marital status Single Education 01/2006 – today ETH Zurich, Switzerland PhD Thesis; Institute of Cell Biology Titel: “Structural and functional characterization of GDAP1, a neuronal mitochondrial fission factor involved in Charcot-Marie Tooth disease” Thesis supervisors: Prof. Dr. U. Suter, Dr. A. Niemann

10/1999 – 01/2005 Friedrich – Schiller - Universität Jena Diplom: Biochemistry/Molecular Biology

06/2004 – 12/2004 Böhringer Ingelheim, Biberach Diploma Thesis Title: “Genechip array analysis of asthma und COPD”

07/2002 – 06/2003 University of Queensland, Brisbane/Australia Graduate Diploma: Molecular Biology

09/1991 – 07/1999 Neideck - Gymnasium Arnstadt Abitur Work experience 05/2005 – 07/2005 Roland Berger Strategy Consulting, Munich Intern: Business analysis and due diligence

02/2005 – 04/2005 The Boston Consulting Group, Vienna/Austria Intern: Pricing strategy

08/2003 – 10/2003 Procter & Gamble, Frankfurt European Research & Development Internship Program Project leader: Line-extension in the field of hygienic materials

05/2001 – 07/2002 Hans – Knöll Institut Jena Research assistant: Gene-chip array Development

08/2001 – 09/2001 EnTec Jena Research assistant: Pharmaceutical research on hormone therapy

Skills/Competences

Languages German -mother tongue English -fluent Spanish -basics Russian -basics Laboratory methods Biochemical-, molecular-, cell- and microbiological standard techniques, confocal microscopy, animal handling and dissection

116 CURRICULUM VITAE

Publications/Poster

Wagner, K., M. Ruegg, A. Niemann, and U. Suter. 2009. Targeting and Function of the Mitochondrial Fission Factor GDAP1 are Dependent on ist Tail-Anchor. PLoS ONE. 4(4): e5160

Niemann, A., K. Wagner, M. Ruegg, and U. Suter. 2009. GDAP1 induces fission without increasing the risk of apoptosis: new insights for Charcot-Marie-Tooth disease. Submitted to J Cell Biol (under review)

Wagner, K., A. Niemann, and U. Suter. 2006. Charcot-Marie-Tooth disease, a peripheral neuropathy caused by mutations in the GDAP1 gene. ZNZ PhD retreat Valens

Wagner, K., C. Bieniossek, A. Niemann, and U. Suter. 2007. Biochemical analysis of the tail-anchor of GDAP1, a neuronal mitochondrial fission factor. ZNZ Symposium Zurich

Wagner, K., C. Bieniossek, I. Berger, A. Niemann, and U. Suter. 2008. Biochemical analysis of the tail-anchor of GDAP1, a neuronal mitochondrial fission factor. ZNZ Symposium Zurich

117