Cellular Functions and Dynamics of Reggie Proteins

(1)

Cellular Functions and Dynamics of Reggie Proteins

Dissertation

zur Erlangung des akademischen Grades des Doktors der Naturwissenschaften (Dr. rer. nat.)

an der Universität Konstanz

Mathematisch-Naturwissenschaftliche Sektion Fachbereich Biologie

vorgelegt von Matthias F. Langhorst

Tag der mündlichen Prüfung: 17.11.2006 1. Referentin: Prof. Dr. C. Stürmer

2. Referentin: Prof. Dr. I. Adamska 3. Referent: Prof. Dr. Dr. W. Neupert

Konstanzer Online-Publikations-System (KOPS) URL: http://www.ub.uni-konstanz.de/kops/volltexte/2007/2136/

URN: http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-21367

(2)

Nature composes some of her loveliest poems for the microscope

Theodore Roszak

(3)

Table of contents

Table of contents I

Abbreviations IV

1 Introduction 1

1.1 Cellular multitasking – a problem of specificity in time and space 1 1.2 Microdomain organization of the plasma membrane– a strategy to maintain spatial

segregation and specificity 2

1.2.1 Compartmentalization by actin - lateral confinement zones 2

1.2.2 Protein-based microdomains 3

1.2.2.1 Clathrin-coated pits 3

1.2.2.2 Caveolae 4

1.2.3 Lipid-based microdomains: rafts floating on the sea of lipids 5 1.2.4 Mixing and mingling of membrane microdomains and the advent of new organizing factors 6 1.3 Scaffolding microdomains and beyond – the function of reggie/flotillin proteins 7

1.3.1 Discovery of the reggies/flotillins 7

1.3.2 Structure of the reggies 7

1.3.3 Subcellular localization and trafficking of the reggies 9 1.3.4 Tissue distribution and expression during development 12 1.3.5 Cellular function of reggie proteins 14 1.3.6 Reggie proteins in health and disease 16 1.3.7 Rivalling caveolae - microdomains scaffolded by reggie and other SPFH proteins 17

1.4 Aim of this work 21

2 Cellular dynamics of reggie proteins 22

2.1 Accumulation of FlAsH/Lumio Green in active mitochondria can be reversed by β-

mercaptoethanol for specific staining of tetracysteine-tagged proteins 22

2.1.1 Abstract 22

2.1.2 Introduction 22

2.1.3 Material and Methods 23

2.1.3.1 Reagents 23

2.1.3.2 Cloning of reggie-1-tetracysteine expression vectors 23

2.1.3.3 Cell culture and transfection 23

2.1.3.4 Labelling of HeLa and N2a cells 23

2.1.3.5 Fluorescence Microscopy 24

2.1.4 Results and Discussion 24

2.1.4.1 FlAsH/Lumio Green accumulates in active mitochondria 24 2.1.4.2 Specific staining of tetracysteine-tagged reggie in the presence of β-mercaptoethanol 26

2.1.5 Acknowledgements 27

2.2 Linking membrane microdomains to the cytoskeleton: Regulation of the lateral mobility of

reggie-1/flotillin-2 by interaction with actin 28

2.2.1 Synopsis 28

2.2.3 Experimental 29

2.2.3.1 Antibodies, plasmids and reagents, cells and transfection 29 2.2.3.2 Immunofluorescence and fluorescence microscopy 30

2.2.3.3 In vitro actin binding assay 30

2.2.3.4 Electron microscopy 30

2.2.3.5 Fluorescence recovery after photobleaching (FRAP) experiments 30

2.2.4 Results 31

2.2.4.1 Heterologous reggie-1/flotillin-2 associates with filamentous actin in cells 31 2.2.4.2 Interaction of reggie-1/flotillin-2 with actin is mediated by the SPFH domain 31

(4)

2.2.4.3 In vitro binding of the SPFH domain of reggie-1/flotillin-2 to F-actin 33 2.2.4.4 Binding to actin regulates lateral mobility of reggie-1/flotillin-2 at the plasma membrane 33

2.2.5 Discussion 38

2.3 Trafficking of reggie/flotillin proteins: regulated vesicular cycling revealed by total internal

reflection microscopy 41

2.3.1 Abstract 41

2.3.3.1 Antibodies, reagents and cells 42

2.3.3.2 Immunofluorescence and confocal microscopy 42

2.3.3.4 TIRF imaging and analysis of membrane-near trafficking 43

2.3.4 Results 43

3 Cellular functions of reggie proteins 54

3.1 PrP^c capping in T cells promotes its association with the lipid raft proteins reggie-1 and

reggie-2 and leads to signal transduction 54

3.1.1 Abstract 54

3.1.3 Methods 55

3.1.3.1 Cell culture 55

3.1.3.2 Antibodies 56

3.1.3.3 Determination of intracellular Ca²⁺ concentration 56

3.1.3.4 Isolation of lipid rafts 56

3.1.3.5 Immunoprecipitation experiments and cell lysates 56 3.1.3.6 Gel electrophoresis and immunoblotting 56

3.1.3.7 Immunocytochemistry 57

3.1.4 Results 57

3.1.4.1 Coexistence of PrP^c, reggie-1 and -2, fyn and lck in lipid rafts 57 3.1.4.2 PrP^c crosslinking induces capping and colocalization with reggie-1 and -2 59 3.1.4.3 Crosslinked PrP^c colocalizes with Thy-1, CD3 and F-actin in the cap 59 3.1.4.4 Coaggregation of PrP^c and reggie in peripheral blood T lymphocytes 61 3.1.4.5 Colocalization of PrP^c, reggie-1, reggie-2 and lck at the EM level 62 3.1.4.6 Coimmunoprecipitation of PrP^c with reggie-1, fyn and lck 62 3.1.4.7 PrP^c crosslinking induces transmembrane signalling 63 3.1.4.8 Internalisation of PrP^c and reggie-1 and -2 66

3.1.5.1 Rafts, PrP^c and the reggies 68

3.1.5.2 Signalling by PrP^c 69

3.1.5.3 Internalisation of PrP^c and the reggies 70

3.1.5.4 Conclusion and perspective 70

3.2 Preformed reggie/flotillin caps: stable priming platforms for macro-domain assembly in T

cells 72

3.2.1 Abstract 72

3.2.3 Methods 74

(5)

3.2.3.1 Antibodies, plasmids and reagents 74

3.2.3.4 Fluorescence recovery after photobleaching (FRAP) experiments 74 3.2.3.5 Superantigen-induced immunological synapse formation 75 3.2.3.6 Immunocytochemistry and confocal microscopy 75

3.2.3.7 ConA-induced cell spreading 75

3.2.3.8 Single cell Ca²⁺ imaging during spreading 75 3.2.3.9 Stimulated cell lysates and co-immunoprecipitation 75

3.2.4 Results 76

3.2.4.1 Membrane localization and lateral mobility of reggie-1-EGFP in PC12 cells 76 3.2.4.2 Stabilization of reggie-1 in preformed caps in Jurkat T cells 78 3.2.4.3 Recruitment of signalling molecules to stable reggie caps during T cell activation 78 3.2.4.4 The C-terminus of reggie-1 is essential for its incorporation into the preformed cap 80 3.2.4.5 A trans-negative reggie-1 mutant inhibits macrodomain assembly after T cell stimulation 80 3.2.4.6 Trans-negative reggie-1 impairs stimulation-induced T cell spreading 81 3.2.4.7 Reggie-1 regulates Vav localization 83

3.3 Reggies/flotillins regulate Rho GTPase signalling during neurite outgrowth via CAP/ponsin 89

3.3.1 Abstract 89

3.3.3.1 Antibodies and reagents 91

3.3.3.3 Cell lysates, GTPase assays and western blotting 91

3.3.3.4 Microscopy 92

3.3.4 Results 92

4 Conclusions and outlook 100

5 Summary / Zusammenfassung 102

5.1 Summary 102

5.2 Zusammenfassung 103

6 Literature 105

7 Curriculum vitae and list of publications 118

8 Note on contributions 121

9 Acknowledgements 123

(6)

Abbreviations

2-ME 2(β)-mercaptoethanol

AB antibody

ABC transporter ATP binding cassette transporter AD Alzheimer’s disease AKAP A kinase anchoring protein

AM acetoxymethylester

APP amyloid precursor protein ATP adenosine triphosphate BACE β-site of APP cleaving enzyme

BAPTA 1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid

BFA Brefeldin A

BHK baby hamster kidney BSA bovine serum albumine

BSE bovine spongiform encephalopathy CAP c-Cbl-associated protein

CD2AP CD2-associated protein cDNA complementary DNA CHO chinese hamster ovary CJD Creuzfeld-Jakob-Disease

ConA Concanavalin A

CTX Cholera toxin

CytD Cytochalasin D

DAB 3,3'-diaminobenzidine

DAPI 4',6-diamidino-2-phenylindole DNA deoxyribonucleic acid

DMEM Dulbecco’s modified Eagle’s medium DMSO dimethyl sulfoxide

DRM detergent-resistant membranes E.coli Escherichia coli

ECL enhanced chemiluminescence

EDT ethanedithiol

EGF epidermal growth factor

EGFP enhanced green fluorescent protein EGTA ethylene glycol tetra acetic acid EM electron microscopy

ERK1/2 extracellular regulated kinase 1/2 ESA epidermal surface antigen EST expressed sequence tag

FACS fluorescence-activated cell sorting FAK focal adhesion kinase

FALI fluorophore-assisted light inactivation FCS foetal calf serum

FITC fluorescein

FlAsH fluorescein arsenical helix binder

FRAP fluorescence recovery after photobleaching

(7)

GFP green fluorescent protein Glut glucose transporter

GPI glycosylphosphatidylinositol GSH glutathione (reduced state) GST glutathione-S-transferase

HA hemagglutinin

HBSS HEPES-buffered salt solution

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HIV human immunodeficiency virus

HRP horse radish peroxidase

Ig immune globulin

IgCAM cell adhesion molecule of the immunoglobulin superfamily IGF insulin-like growth factor

IgG immune globulin G IgM immune globulin M

IRM interference reflection microscopy JNK c-Jun N-terminal kinase

LAT linker of activated T cells LDL low density lipoprotein

LM light microscopy

LSM laser scanning microscope mAB monoclonal antibody

MAPK mitogen activated protein kinase MβCD methyl-β-cyclodextrin

MEM modified Eagle’s medium

mRNA messenger RNA

NA numerical aperture NRK normal rat kidney pAB polyclonal antibody PBL peripheral blood lymphocytes PBS phosphate buffered saline PCR polymerase chain reaction PKA protein kinase A

PKB protein kinase B PKC protein kinase C

PrP prion protein

Rb retinobalstoma protein RBD ras-binding domain

ReAsH resorufin arsenical helix binder RNA ribonucleic acid

ROI region of interest

RT-PCR reverse-transcription polymerase chain reaction SDM standard deviation of mean

SDS sodium dodecyl sulfate

SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SEE Staphylococcus enterotoxin E

SEM standard error of mean

(8)

siRNA short interfering RNAs

SNARE soluble N-ethylmaleimide sensitive factor attachmentreceptor SoHo sorbin homology

SPFH stomatin/prohibitin/flotillin/HflK&C SV40 simian virus 40

TBS tris-buffered saline TCR T cell receptor

TIRF total internal reflection fluorescence Tris trishydroxymethylaminomethane WASP Wiskott-Aldrich syndrome protein ZAP70 ζ-associated protein kinase of 70 kDa

(9)

1 Introduction

1.1 Cellular multitasking – a problem of specificity in time and space

Every single cell in a multicellular organism has to continuously integrate information, from both its surrounding and from within the cell. For the processing of these information, it can only rely on a limited set of signalling modules, which are interconnected and often used by more than one cell surface receptor. In spite of this seemingly insuperable task, all cells achieve specificity in all incoming and outgoing signals, thus ensuring the smooth interplay of cells necessary for the survival of a multicellular organism.

To achieve specificity in handling of incoming and outgoing signals in the face of a limited number of potential processing pathways, the cell has to separate these pathways both in time and space. By anchoring components of signalling pathways to specific subcellular locations, cells manage to perform opposite actions in different regions, like disassembly of the actin cytoskeleton at the trailing edge and assembly of new actin networks at the leading edge of migrating cells. Different kinetics during processing help to separate incoming signals reaching their receptors simultaneously. Local differences in the balance between activating and inactivating components of a signalling pathway lead to different kinetics of the same process, and is thus again based on spatial segregation of signalling molecules. This mechanism can be extremely fine tuned as shown in the case of Ca²⁺ signalling, where within one single cell, different signalling patterns of different duration can co-exist as specific combinations of Ca²⁺-pumps, Ca²⁺-channels and Ca²⁺-buffers in different regions of the cell result in different kinetics of changes in the local Ca²⁺-concentration (Berridge 2000). Such a spatial segregation of signalling proteins can be accomplished in different ways. In simple cases, proteins are only present on specific compartments and this distribution is established during protein synthesis, maturation and trafficking. This compartmentalization is then permanent and cannot be dynamically regulated. Indirect anchorage by specialized and specifically localized anchoring proteins solves this problem as these anchoring proteins only act upon posttranslational modifications of either their interacting proteins or themselves thus enabling dynamic regulation. A well known example of such an indirect anchorage is the regulation of protein kinase A (PKA) signalling, where a variety of PKA anchoring proteins (AKAPs), each specifically localized, is used to restrict the enzymatic activity of active PKA isoforms to specific cellular compartments, thus guiding broad-band enzymes towards specific targets and physically separate them from undesirable substrates (Bauman and Scott 2002). The same principle can also be applied to complexes of soluble proteins. To avoid unregulated diffusion of soluble components, they can be trapped by docking- or scaffolding- proteins, which have binding sites for various components of one signalling pathway. The multiprotein complexes assembled by these signalling scaffolds are big enough to avoid free diffusion through the crowded environment of the cytosol, thus preventing contact with other signalling pathways and enhancing the interaction of the bound components at the same time.

(10)

An example of this strategy are the scaffolds supporting the mitogen activated protein kinase (MAPK) cascades, like Ste5p in yeast. Assembly of all members of the MAPK cascade into one module by a scaffold not only enhances the local concentration of the components and thus signalling efficiency, but also infers complex kinetic regulatory properties on the function of the module (Morrison and Davis 2003).

In many cases spatial specificity in signalling is achieved by binding of signalling components to cellular membranes, which avoids free diffusion through the cell altogether. Even many scaffold proteins like LAT (linker for activated T cells) are bound to membranes.

Furthermore, active, membrane-bound signalling complexes can be guided through the cell by means of membrane trafficking. Although unwanted interactions with cytosolic proteins can be largely reduced in membrane-bound protein complexes and diffusion is restricted to two dimensions, the maintenance of spatial information and the need to avoid interactions with other membrane-bound cascades pose new challenges for the maintenance of spatial and temporal specificity.

1.2 Microdomain organization of the plasma membrane– a strategy to maintain spatial segregation and specificity

The classical view of the plasma membrane envisioned a sea of excess lipids in which the proteins can freely diffuse. This fluid mosaic model of the plasma membrane was originally proposed by Singer and Nicolson in 1972 (Singer and Nicolson 1972). However, free diffusion of transmembrane or membrane associated proteins without any restriction would severely hamper the maintenance of spatial specificity of membrane-bound processes. On a large scale, cell polarity requires diffusion barriers, e.g. between the axon and the cell body of a neuron. On a smaller scale even processes in the same region of the cell have to be separated, like in the active zone of a synapse. Without a localization mechanism, after receiving a signal by ligand binding, the spatial information of this signal would be quickly lost due to random diffusion of the receptor molecule.

During the last 20 years the classical view on the organization of the plasma membrane was largely extended by various reports describing specialized regions within the plasma membrane, where exchange of components with the surrounding is regulated and restricted.

Our today’s view of the plasma membrane describes it as a crowded patchwork of protein or lipid-based microdomains, actin-anchored lateral confinement zones and other specialized areas which are connected by small areas of free lateral diffusion – thus, instead of a homogenous sea of lipids with freely floating proteins in it, we think of the plasma membrane more as an archipelago of specialized islands with regulated exchange of inhabitants.

1.2.1 Compartmentalization by actin - lateral confinement zones

Single particle tracking experiments revealed the compartmentalization of the plasma membrane into small zones of random diffusion which are surrounded by an elastic diffusion barrier (Kusumi and Sako 1996). Membrane proteins with sufficient kinetic energy can

(11)

overcome these barriers, giving rise to a behaviour of so-called hop diffusion, with proteins undergoing random diffusion within a confinement zone and eventually hopping over the barrier to the next confinement zone (Kusumi and Sako 1996). The lateral confinement zones observed in single particle tracking experiments with transmembrane proteins were shown to be built up by the membrane cytoskeleton, the layer of actin cytoskeleton which is in close contact with the cytoplasmic leaflet of the plasma membrane. Disruption of the cytoskeleton increased, stabilization of the cytoskeleton decreased the areas of free diffusion. Truncation of the cytoplasmic tail of transmembrane proteins generally increased their mobility. These observations suggested that the movement of transmembrane proteins is confined by the trapping of their cytoplasmic tail by the membrane cytoskeleton (Ritchie and Kusumi 2004).

Surprisingly, also lipids of the outer leaflet showed hop-diffusion behaviour similar to transmembrane proteins. Transmembrane proteins which are actually anchored to the membrane cytoskeleton serve as anchored diffusion barriers, thus conveying the diffusion barriers also to the outer leaflet (Fujiwara et al. 2002). Thus, while the membrane cytoskeleton serves as a fence defining borders between areas of free diffusion, anchored transmembrane proteins serve as fence pickets.

This compartmentalization of the plasma membrane was shown to be universal, with only the size of the transient confinement zones varying between different cell types (Ritchie and Kusumi 2004). The compartmentalized mosaic formed by the membrane cytoskeleton gives the cell a template over which to design microdomains. A lipid-protein-complex which is larger than the inter picket spacing will be efficiently trapped in this confinement zone, thus largely abolishing lateral diffusion of the complex (Kusumi and Sako 1996; Ritchie and Kusumi 2004). Proteins with the ability to form large oligomeric complexes are therefore efficient organizers of stable microdomains, as their oligomers are readily trapped and immobilized in confinement zones. Furthermore, by controlling actin dynamics the cell can efficiently regulate the size of confinement zones and even move protein complexes trapped in the fences of the membrane cytoskeleton (Kusumi and Sako 1996).

1.2.2 Protein-based microdomains 1.2.2.1 Clathrin-coated pits

Long before the fluid mosaic model of the plasma membrane was proposed, electron microscopists had already described exceptions from a homogenous organization of the plasma membrane. One of the earliest membrane domains to be described in 1964 were endocytic invaginations of the plasma membrane, which had a distinct “ bristle coat” at the cytoplasmic leaflet (Roth TF and Porter 1964). Endocytic cargo was concentrated in these invaginations, which pinched of to give rise to coated endosomes. These coated pits were also shown to be the site where cell surface receptors like the LDL receptor accumulated after ligand binding and before endocytosis via coated vesicles (Goldstein et al. 1979). Already in 1969 electron microscopists were able to show, that the “bristle coat” had a regular structure composed of hexagons and pentagons (Kanaseki and Kadota 1969). This coat was shown to

(12)

consists of only one protein named clathrin (Pearse 1976). But the clathrin lattice is not in direct contact with the membrane. A variety of adaptor proteins links the coat, which helps to induce curvature in the membrane and thus invagination of the coated pits, to the membrane (Robinson 2004). These adaptor proteins additionally bind to cytoplasmic signals on cell surface receptors, which are thereby concentrated in clathrin coated pits. For endocytosis to occur, the invaginated clathrin coated pits finally bud of the plasma membrane with the help of the GTPase dynamin. Thus, clathrin-coated pits are the prototype of a membrane domain characterized by the accumulation of a proteinaceous scaffold which imposes a specific function, namely endocytosis, on this membrane domain (Roth MG 2006).

1.2.2.2 Caveolae

In the 1950s another type of plasma membrane microdomain had already been described by electron microscopists (Parton 2003). In several epithelial and endothelial cells numerous flask-shaped invaginations of the plasma membrane was observed. These “little caves” or caveolae were only 50 – 100 nm in size and lacked a distinguishable coat on the cytoplasmic side. Sometimes these invaginations were found in grapelike clusters or even forming tubular systems. From their morphology, a role in vesicular trafficking was immediately proposed and studies on ultrastructural quantification of gold-labelled probes spoke for a role of caveolae in pinocytosis and transcytosis. But it wasn’t until the discovery of the caveolin coat proteins in the late 1980s, that this and other functions of caveolae could be clarified (Parton 2003).

Caveolin-1, -2 and -3 are proteins of only 22 kDa. Caveolin-1 forms large homo- and hetero- oligomeric complexes (with caveolin-2) and its expression is necessary and sufficient for the formation of caveolae. Caveolin-3 is most similar to caveolin-1 and its expression is restricted to striated muscle. The other two caveolins are widely expressed with two notable exceptions:

lymphocytes and neurons are devoid of caveolin-1 expression and thus of caveolae. In addition to their proteinaceous coat, caveolae are enriched in certain lipids like cholesterol and sphingolipids, making them a specialized type of lipid raft (Cohen et al. 2004).

The cloning of the caveolins finally enabled researchers to unambiguously define the function of caveolae. Using EGFP-tagged caveolin-1, it could be shown that caveolae are mediating a novel, specialized pathway for endocytosis. The simian virus 40 (SV40) enters the host cell via caveolae and is subsequently delivered to “caveosomes” – pH neutral endocytic compartments unrelated to the classical endosomal pathways - and finally to the ER. This way, the virus also escapes the degradative lysosomal pathway (Pelkmans et al. 2001).

Accordingly, several components of vesicle formation (like dynamin) and docking (SNARE proteins) are localized to caveolae (Henley et al. 1998; Schnitzer et al. 1995a). Finally, caveolin-1 knock-out mice show defects in albumin uptake (Schubert et al. 2001). Thus, these reports confirmed earlier assumptions based on caveolar morphology, that these specialized invaginations of the plasma membrane do play a role in vesicular trafficking.

(13)

Additionally, caveolae are platforms for signal transduction. Via a so-called scaffolding domain, caveolin-1 binds to various signalling molecules like Gα, Ha-Ras or Src kinases. This binding of caveolins to signalling molecules also influences their activity. Caveolin preferably binds to the inactive forms of Gα, Ha-Ras or Src kinases and peptides derived from the scaffolding domain have inhibitory effects on these signalling proteins in vitro. Thus, the caveolins recruit signalling molecules into specialized membrane microdomains forming preassembled signalling complexes, but simultaneously the caveolins inhibit the activity of these complexes to avoid uncued activity (Okamoto et al. 1998).

1.2.3 Lipid-based microdomains: rafts floating on the sea of lipids

Cells can choose from hundreds of lipids when building their membranes. The functional consequences of the vast variety of lipids with different biophysical properties encountered in cellular membranes are largely not understood (Dowhan 1997). At a specific temperature a particular lipid species will undergo a phase transition from a solid-ordered (or gel) phase, where the acyl chains are tightly packed, to a liquid-disordered phase, where the acyl chains no longer pack in a rigid straight conformation. Upon addition of cholesterol a third biophysical state is possible, where the acyl chains are tightly packed like in solid-ordered states but the lateral mobility is still high like in liquid-disordered states. This additional phase was thus termed liquid-ordered. Due to the tight package of acyl chains, this phase is highly unfavourable for unsaturated phospholipids, as their kinked acyl chains do not allow tight packing. Importantly, in model membranes of appropriate composition consisting of e.g.

sphingomyelin, unsaturated phospholipids and cholesterol, liquid-ordered and liquid- disordered phases can coexist (de Almeida et al. 2003). This observation led to the suggestion that such phase separations might also exist in cellular membranes. Liquid-ordered “lipid rafts” are thought to float as microdomains on a sea of a liquid-disordered surrounding (Simons and Ikonen 1997). Proteins with membrane anchors favouring an ordered environment are supposed to preferentially associate with liquid-ordered rafts, these include GPI-anchored proteins and lipid-modified proteins like the double-acylated Src-kinases, but also some transmembrane proteins (Simons and Toomre 2000). However, the existence of stable, liquid-ordered domains in living cells is questionable (Munro 2003). A variety of recent studies using different biophysical approaches showed that large scale phase separation is not observable in cellular membranes. Rather, the liquid-ordered domains observed were small (10 nm) and unstable (with a lifetime < 0.1 ms) (Hancock 2006; Sharma et al. 2004;

Subczynski and Kusumi 2003). These small domains can only accommodate 3-5 proteins and are by far not stable enough to serve as signalling scaffolds. But proteins favouring an ordered environment can increase the stability of such spontaneously forming raft domains, and by protein-protein-interactions, larger and more stable liquid-ordered domains can be formed in cells (Hancock 2006). One well studied example is the signalling complex surrounding engaged T cell receptors. It is probably built up by protein-driven fusion of the small, transient raft domains into one large macrodomain, which is stabilized by extensive protein-

(14)

protein-interactions and binding to the membrane cytoskeleton (reviewed in (Rodgers et al.

2005)). The stable macrodomain finally assembled exhibits a high degree of order as recently directly visualized using the order-sensitive fluorescent dye laurdan (Gaus et al. 2005), thus representing a liquid-ordered phase.

Another complicating fact in the research on lipid microdomains is the fact, that different experimental approached often lead to confusing results. Certain membrane components cannot be solubilized under certain conditions even at high concentrations of detergents. As these detergent-resistant membranes (DRMs) are enriched in cholesterol and sphingolipids, one often assumes that lipid-rafts and DRMs are synonymous. Thus, an ever increasing number of proteins was shown to be associated with DRMs and thus assumed to reside in rafts. But as DRMs originate from detergent-treatment, they do not correspond to any structure present in the cell before. Therefore, the small, unstable lipid rafts existing in living cells have most probably only little resemblance to the insoluble complexes purified after detergent-extraction (reviewed in (Lichtenberg et al. 2005)).

1.2.4 Mixing and mingling of membrane microdomains and the advent of new organizing factors

The borders between these different types of membrane microdomains are diffuse. Although spontaneous phase separation of lipids exists, it needs proteins to stabilize lipid domains.

Furthermore, any protein might induce the accumulation of certain lipids in its vicinity due to the biophysical properties of its membrane anchor (Epand 2004). Oligomeric proteins like caveolins can bind preferentially to certain lipids like cholesterol, thus caveolae are a protein microdomain with the lipid composition of a lipid microdomain. In addition, the lateral diffusion of any kind of microdomain large enough will be controlled by the “picket fence” of the membrane cytoskeleton. Binding of microdomain organizing proteins to the cytoskeleton can further increase the domain’s stability.

Although caveolae were - thanks to their extraordinary morphology - one of the first membrane microdomains to be discovered, other proteins most probably fulfil functions very similar to those of caveolins. The crowded environment of the plasma membrane requires organizing factors, providing anchorage for protein complexes and thus spatial specificity.

Any membrane-associated, oligomeric protein can readily provide scaffolds, which recruit multi-protein complexes to specific membrane locations and control their activity. Excellent candidates for such organizing factors of membrane microdomains are the oligomeric proteins of the SPFH protein family, especially the founding members reggie/flotillin and stomatin, as described in the following review.

(15)

1.3 Scaffolding microdomains and beyond – the function of reggie/flotillin proteins

1.3.1 Discovery of the reggies/flotillins

“Third time is a charm” the saying goes and accordingly the reggie/flotillin proteins were independently “discovered” three times. Madeleine Duvic and co-workers were the very first to discover part of one of the proteins during a screen for the antigen of the monoclonal antibody ECS-1 in 1994. They identified a cDNA coding for a N-terminally truncated version of reggie-1/flotillin-2 of 42 kDa (see below), which they called epidermal surface antigen (ESA) (Schroeder et al. 1994). This name was given up later as it was shown that this protein is not the true antigen recognized by ECS-1 (Hazarika et al. 1999). In a screen for proteins upregulated in retinal ganglion cells during axon regeneration after optic nerve lesion in goldfish, we identified in 1997 two proteins of 47 kDa which we called reggie-1 and -2 (Lang et al. 1998; Schulte et al. 1997). In the same year the group of Michael Lisanti identified two proteins associated with the “floating” detergent-resistant membrane fraction from mouse lung tissue which they called flotillin-1 (= reggie-2) and flotillin-2 (= reggie-1) (Bickel et al.

1997). Although flotillins became the more commonly used name for the proteins, we will stick to reggie-1 and -2, since we believe that our names and numbers are more appropriate in light of the physiological relevance of the two proteins.

1.3.2 Structure of the reggies

The reggie proteins are highly conserved with about 64% homology between fly and man (Galbiati et al. 1998; Malaga-Trillo et al. 2002). Reggie-like proteins even exist in some bacteria, plants and fungi (Borner et al. 2005; Edgar and Polak 2001), although these most probably arose by convergent evolution (Rivera-Milla et al. 2006).

In humans, the gene encoding reggie-1 is located on chromosome 17 (17q11-12) (Cho et al.

1995). It is a single-copy gene consisting of 11 exons giving rise to a protein of 47 kDa. The human reggie-2 gene is a single-copy gene located on chromosome 6 (6p21.3). It encompasses 15 kb with 13 exons coding for a protein of 47 kDa (Edgar and Polak 2001).

The reggies are members of the SPFH (Stomatin/Prohibitin/Flotillin/HflK&C) protein superfamily, whose members share a domain of similar sequence but unknown function, the so-called SPFH domain at their N-terminus (Tavernarakis et al. 1999). In contrast to other SPFH proteins, the C-terminus of the reggies harbours a unique flotillin domain, which is characterized by several repeats of glutamic acid and alanine (EA repeats) and which is predicted to potentially form coiled coil structures (Figure 1.1) (Bickel et al. 1997; Schroeder et al. 1994).

In contrast to earlier reports, the reggies do not possess a transmembrane domain. Both C- and N-terminus are facing the cytosol (Gkantiragas et al. 2001; Morrow et al. 2002), the hydrophobic region might interact with but does apparently not span the membrane.

Anchoring to the cytoplasmic leaflet of membranes is accomplished by acylation. Reggie-1 is myristoylated at Gly2 and palmitoylated mainly at Cys4 and to a minor degree at Cys19 and

(16)

Cys20 (Neumann-Giesen et al. 2004). Reggie-2 lacks a myristoylation site but is palmitoylated at Cys34 (Morrow et al. 2002) and potentially also at Cys5 and Cys17.

Recently, the 3-dimensional structure of the SPFH domain of reggie-1 has been solved (Miyamoto et al. 2004) and is available in MMDB (Chen J et al. 2003). This structure encompasses the region from aa 43 to aa 173 and indicates that the SPFH domain of the reggies is a compact, ellipsoid-globular structure containing 4-5 alpha helices and 6 beta strands. The flotillin domain, which has not been solved yet, is predicted to harbour several alpha helices, some of which might be involved in coiled coil formation.

The C-terminal part of reggie-1 is essential for the formation of homo- and most probably also hetero-oligomers ((Neumann-Giesen et al. 2004) and G Solis and CAO Stuermer, unpublished observations) which may be mediated by the coiled coils in the region of the EA repeats.

Chemical crosslinking experiments demonstrated that the smallest building block of reggie oligomers are tetramers (G Solis and CAO Stuermer, unpublished observations).

The initial description of reggie-1 as a 42-kd cytosolic protein raises the question whether this and maybe other splice variants exist. Indeed, EST database searches yield multiple expressed sequence tags for this N-terminally truncated form lacking the membrane anchor. However, we and others have not found any major splice variants of reggie-1 in western blots or RT- PCRs of various mammalian cell lines or tissues (Edgar and Polak 2001; Solomon et al.

2002). In a recent report, a splice variant of reggie-2 based on EST sequences was proposed, which lacks exon 4 (Lopez-Casas and del Mazo 2003), but again a major expression of such a variant in mammalian cells was so far never observed. However, two variants of reggie-1 were recently reported in Drosophila, differing in 39 bp encoded by a short, alternatively spliced exon. These two splice variants are differentially expressed during development with the longer form being predominantly expressed during embryonic and larval development, while the shorter variant predominates in the adult fly (Hoehne et al. 2005).

Due to sequence similarities the reggies are considered members of the SPFH protein family including stomatin, podocin, prohibitin and the bacterial HflK/HflC proteins (Tavernarakis et al. 1999). A recent analysis of the evolutionary relationships between all SPFH proteins showed that the observed sequence similarities in the SPFH domain must have arisen by convergent evolution (Rivera-Milla et al. 2006). Several structural hallmarks are remarkably similar between all SPFH proteins (Figure 1.1). They share a hydrophobic domain in their N- terminus which is often preceded by a palmitoylation site. In case of the reggies, stomatin and podocin, this hydrophobic domain is not spanning the membrane but is suggested to form a horseshoe-like structure with both N- and C-termini facing the cytosol (Gkantiragas et al.

2001; Huber et al. 2003; Morrow et al. 2002; Roselli et al. 2002; Salzer et al. 1993), while prohibitin possesses a transmembrane domain. Similarly, all SPFH proteins share a stretch of EA-repeats in their C-termini, which is extended by the flotillin domain in case of the reggies.

The first EA repeat at the end of the SPFH domain (the only EA repeat in other SPFH proteins) was shown to be essential for oligomerization of stomatin (Snyers et al. 1998). The formation of oligomers is another hallmark shared by all SPFH proteins. Oligomers were

(17)

shown for stomatin (Snyers et al. 1998), podocin (Huber et al. 2003) prohibitin (Back et al.

2002; Tatsuta et al. 2005) and the reggies (Neumann-Giesen et al. 2004).

These similar structural features suggest a related function for all SPFH proteins. Considering their widespread distribution, their function is supposed to be basic and important.

Interestingly, the structural hallmarks described above are reminiscent of an unrelated protein – caveolin. Although quite different in domain structure, caveolin is predicted to adopt a hairpin-like structure with a central hydrophobic domain interacting with but not spanning the membrane and both N- and C-termini facing the cytosol (Okamoto et al. 1998; Spisni et al.

2005). It is palmitoylated on multiple cysteine residues (Dietzen et al. 1995) and it forms oligomers (Monier et al. 1995). These oligomers are supposed to form the structural scaffold of caveolae, a well-defined subtype of lipid microdomains important for signalling, specialized endocytic processes and transcytosis (Cohen et al. 2004).

1.3.3 Subcellular localization and trafficking of the reggies

In single cells the reggies are most prominently found at the plasma membrane, where they associate with the inner leaflet via acylations (Figure 1.2) (Neumann-Giesen et al. 2004).

Figure 1.1:

Alignment of reggie-1 and -2 with SPFH proteins and comparison with caveolin. The SPFH- and flotillin- domain are indicated, acylation sites, hydrophobic domains and EA repeats are shown as structural features.

(18)

Being insoluble in Triton-X-100 and other commonly used non-ionic detergents at 4°C, the reggies are considered as lipid raft proteins and are often used as lipid raft markers. But unlike many other proteins considered to be raft-associated, the reggies form stable clusters at the plasma membrane readily observable by electron microscopy and immuno-gold or DAB- based staining (Figure 1.2) (Kokubo et al. 2003; Lang et al. 1998; Stuermer et al. 2001). The reggie clusters are surprisingly uniform in size with an estimated diameter of 100 nm (Kokubo et al. 2003; Stuermer et al. 2001) and consist most probably of reggie-1 and -2 homo- and hetero-oligomers. They are quite widely spaced along most of the plasma membrane but become more closely spaced at cell-cell-contact sites and after crosslinking of associated cell surface molecules like GPI-anchored Thy-1 or PrP^c (Stuermer et al. 2001;

Stuermer et al. 2004).

Figure 1.2:

a) Confocal image of a PC12 cell stained with monoclonal antibody against reggie-1. Reggie-1 is detected in puncta along the plasma membrane and additionally at the centrosome (bar = 5 µm); b) Grazing section of the plasma membrane of a PC12 cell with double immunogold labeling of reggie-2 (10 nm gold grains) and GPI- anchored Thy-1 (5nm gold grains) after in vivo crosslinking of Thy-1. Note the labeling of ≈ 0.1 µm large microdomains, where both proteins are clustered (bar = 0.1 µm); c) Model of reggie scaffolds with associated proteins. Reggie oligomers serve as stable scaffolds for multiprotein complex assembly. The reggie clusters are anchored to the actin cytoskeleton but also recruit the machinery for regulation of the cytoskeleton.

(19)

Although it was repeatedly claimed that the reggies reside within caveolae (Bickel et al. 1997;

Volonte et al. 1999), we and others have convincingly shown that the reggie clusters are distinct from caveolae. First of all, the reggies are expressed and reggie clusters are detected in cell types which do not express caveolins - most importantly neurons and lymphocytes (Lang et al. 1998; Stuermer et al. 2001); caveolins and reggies show distinct expression patterns both in adult tissues (Evans et al. 2003) and in developing organisms (Pandur et al.

2004). Even in cells that do express both caveolin and reggies, reggie microdomains are clearly distinct and located outside of caveolae (Lang et al. 1998; Souto et al. 2003; Stuermer et al. 2001).

In addition to their preferred localization at the plasma membrane, the reggies can be found at various vesicular compartments inside the cell. In many cell types, a prominent localization of both reggie proteins at the pericentrosomal region is observed, sometimes overlapping with recycling endosomes (Figure 1.2) ((Gagescu et al. 2000; Solomon et al. 2002) and our own unpublished observations). The reggies also co-localize with markers for lysosomes (Stuermer et al. 2004) and are found in multivesicular bodies (de Gassart et al. 2003; Langui et al. 2004).

Like other raft-associated proteins they also localize to lipid-rich droplets (lipid bodies) (Liu P et al. 2004; Reuter et al. 2004a).

Although the localization pattern of reggie-1 and -2 is largely overlapping in many cell types, specific differences in some cell types and on particular organelles are observed. In many cells, reggie-2 is found more prominently on intracellular organelles than reggie-1. In 3T3 fibroblasts, reggie-2 is predominantly found on intracellular granules but relocates to the plasma membrane upon differentiation to adipocytes (Liu J et al. 2005). Similarly, in RAW 264.7 macrophages (which despite a contrary report (Slaughter et al. 2003) express both reggies), reggie-2 is almost exclusively found on intracellular vesicular organelles while reggie-1 is found predominantly at the plasma membrane (our own unpublished observations). In phagocytic cells reggie-2 was also observed on phagosomes (Garin et al.

2001) where it accumulates during maturation of the phagosome by fusion with endosomes (Dermine et al. 2001). The most important difference concerning the localization of the two reggies seems to be the translocation of reggie-2 to the nucleus. In a recent report, Santamaria et al. have described a cell-cycle dependent translocation of reggie-2 and PTOV-1 to the nucleus, the nuclear localization of both proteins being most pronounced at the beginning of S-phase (Santamaria et al. 2005). Reggie-1, however, was not observed to translocate to the nucleus in this study (Santamaria et al. 2005). This matches our own observation that reggie-2 can sporadically be found in the nucleus of e.g. PC12 cells (our own unpublished observations). The function of reggie-2 in the nucleus is currently unclear. It did not localize to any structure identifiable in the EM, but reggie overexpression had a mitogenic effect in PC3 cells (Santamaria et al. 2005). Taken together, these data suggest that reggie-1, being irreversibly myristoylated, is the more immobile of the two, residing predominantly at the plasma membrane. The reversibly palmitoylated reggie-2 on the other hand seems to be more versatile, shuttling between the plasma membrane and various intracellular organelles. Thus,

(20)

the common usage of reggie-2/flotillin-1 as a marker protein for plasma membrane lipid rafts is questionable.

Conflicting reports exist about the biosynthetic pathway of the reggies. Gkantiragas et al.

described sphingomyelin-enriched microdomains at the Golgi complex in NRK, CHO and HeLa cells, which contained reggie-2 (Gkantiragas et al. 2001). Reggie-2 localization at the plasma membrane was shown to be Brefeldin A-sensitive, indicating that reggie-2 travels through the Golgi in its biosynthetic pathway and becomes detergent-resistant on its way.

However, Morrow et al. showed that trafficking of reggie-2 is Brefeldin A- and Sar1- insensitive in BHK cells (Morrow et al. 2002), suggesting a Golgi-independent trafficking pathway similar to K-Ras, F3/Contactin or TC10 (Apolloni et al. 2000; Bonnon et al. 2003;

Watson et al. 2003). Due to their relatively small size and the lack of transmembrane domains or signal sequences, the reggies associate with membranes via co- or post-translational modifications like reversible palmitoylation and in case of reggie-1 myristoylation. So it is conceivable that the reggies might use both Golgi-dependent and -independent vesicular or even non-vesicular pathways to reach the plasma membrane or other intracellular membranes.

Little is known about the endocytic trafficking of the reggies. They are localized at recycling endosomes (Gagescu et al. 2000), lysosomes and multivesicular bodies (de Gassart et al.

2003; Langui et al. 2004; Stuermer et al. 2004), thus they may participate in endocytic trafficking. A recent report identified an unconventional pathway for pinocytosis and endocytosis of GPI-anchored proteins which is apparently dependent on reggie-2, but different from clathrin- and caveolin-mediated endocytosis (Glebov et al. 2006).

There are several ways by which reggie localization might be regulated. The reversibility of palmitoylation (Bijlmakers and Marsh 2003) of both reggies can regulate their affinity for membranes especially in case of reggie-2 which is not myristoylated. Moreover, reggie function might be regulated by proteolytic cleavage since a calpain-mediated cleavage of reggie-1 but not reggie-2 in platelets was reported (Mairhofer et al. 2002).

1.3.4 Tissue distribution and expression during development

All stable cell lines we have tested so far express at least reggie-1. Reggie-2 exhibits a more restricted expression, but is still widely expressed. The stability of reggie-2 is strongly linked to the presence of reggie-1 as downregulation of reggie-1 by specific siRNAs in mammalian cell lines reduces protein levels of reggie-2. On the other hand, siRNA-mediated knock-down of reggie-2 does not impair reggie-1 stability (M Hoegg and CAO Stuermer, unpublished observations). Furthermore, Drosophila reggie-1 knock-out mutants lose reggie-2 protein (Hoehne et al. 2005). These observations suggest that the stability of reggie-2 requires the presence of reggie-1, but reggie-1 can exist without reggie-2.

The expression of the reggies during differentiation was investigated in several cell culture models. In 3T3 fibroblasts reggie-2 expression is apparently upregulated during the formation of cell-cell-contacts (Lopez-Casas and del Mazo 2003). Differentiation of osteoclasts induces a strong upregulation of reggie-2 expression (Ha et al. 2003). Reggie-1 expression is

(21)

upregulated during in vitro differentiation of C2C12 skeletal myoblasts (Volonte et al. 1999).

During differentiation of 3T3 fibroblasts to adipocytes reggie-2 expression is enhanced while reggie-1 expression remains unchanged (Bickel et al. 1997) and reggie-2 translocates from intracellular compartments to the plasma membrane (Liu J et al. 2005). In contrast, differentiation of PC12 cells does not affect the expression of both reggies (Volonte et al.

1999). Thus, reggie expression seems to increase during differentiation of various cells in culture. Moreover, both reggies are strongly upregulated by retinal ganglion cells during axon regeneration after optic nerve lesion (Lang et al. 1998; Schulte et al. 1997).

There are only two comprehensive reports on reggie expression during development in vertebrates. Due to a genome duplication in fish there are two copies of each reggie gene in zebrafish (Malaga-Trillo et al. 2002), but one reggie-1 gene was rendered non-functional during evolution. In good correlation to the ubiquitous expression in stable cell lines, reggie-1a, reggie-2a and -2b are expressed ubiquitously during the early stages of zebrafish development (von Philipsborn et al. 2005). Upon segmentation, the expression pattern becomes more restricted. Reggie-2a is expressed in differentiating neurons in the brain, spinal cord and neurogenic placodes, while reggie-2b is expressed in head mesoderm, neural crest derivatives and along somite boundaries. Reggie-1a is highly expressed in domains overlapping with the expression pattern of both reggie-2 genes except at the somites where it complements the pattern of reggie-2b. Immunostainings using reggie antibodies stain all fibre tracts in the developing nervous system (von Philipsborn et al. 2005), indicating an involvement of reggie-1a and -2a in neuronal differentiation as expected due to their identification in regenerating axons.

Morpholino knock-down of reggie expression in the developing zebrafish leads to severe morphological defects starting early during gastrulation. This emphasizes the important role of the reggies during early development (E Riviera-Milla, E Malaga-Trillo and CAO Stuermer, unpublished observations).

An expression pattern similar to the one observed in zebrafish was reported for reggie-2 in the developing Xenopus (Pandur et al. 2004). Due to the partial duplication of the Xenopus laevis genome there are also two copies of the reggie-2 gene in frogs. Both genes are highly and ubiquitously expressed during early stages of development. During neural plate formation, expression is enhanced in the neural ectoderm and later on in the neural tube. From late tailbud stages on the two reggie-2 genes are differentially expressed. Flotillin-1a (reggie-2a) is expressed in several neural crest derivates including the olfactory pit and cranial ganglia, and in the dorsal regions of the neural tube including the primary neurons in the dorsal spinal cord. Flotillin-1b (reggie-2b) expression is restricted to small domains of the dorsal neural tube and a low level of expression is found in the branchial arches (Pandur et al. 2004).

Unfortunately, the expression pattern of reggie-1 was not investigated in this study.

In Drosophila, reggie-1 and -2 are highly expressed in the developing and adult nervous system (Galbiati et al. 1998) with particularly high expression in axons at the root of fibre tracts where strong fasciculation is required. Misexpression of reggie-1 and -2 in the eye

(22)

imaginal disc leads to severe mistargeting of specific cell-adhesion molecules of the immunoglobulin superfamily (IgCAMs) resulting in an irregular ommatidial pattern.

Similarly, misexpression in the wing imaginal disc leads to an extension of the wingless signal and disrupts normal wing development (Hoehne et al. 2005).

There are no comprehensive studies on the expression of the reggies in adult organisms, but reggie-1 is apparently widely expressed in many different tissues with particular high expression in brain, while reggie-2 expression is slightly more restricted (Bickel et al. 1997;

Volonte et al. 1999; von Philipsborn et al. 2005).

In summary, reggie expression seems to be essential for all cells in culture and during early stages of development. The available data on interference with reggie function during development from zebrafish and flies point to an important role of the reggies in regulating formation of correct cell-cell contacts during morphogenesis. During later developmental stages and in the adult organism, the reggies are still widely expressed but with a particularly high expression in the central nervous system.

1.3.5 Cellular function of reggie proteins

Despite their ubiquitous expression and their evolutionarily high conservation, the function of the reggies is still unclear. They were implicated in signal transduction, vesicle trafficking and cytoskeleton rearrangement and a variety of proteins were shown to interact with the reggies.

The reggies were isolated by co-immunoprecipitation with the monoclonal antibody M802 which was later shown to recognize the fish homologue of Thy-1 (Deininger et al. 2003; Lang et al. 1998; Schulte et al. 1997). Fish Thy-1 co-localizes with the reggies (Reuter et al.

2004b). The close association with Thy-1 is conserved in different mammalian cell types like PC12 and lymphocytes (Stuermer et al. 2001). Remarkably, the reggies also co-localize and can be co-immunoprecipitated with other GPI-anchored proteins like F3/contactin (Stuermer et al. 2001) and PrP^c (Stuermer et al. 2004). This suggests an involvement of the reggies in signal transduction by GPI-anchored proteins across the plasma membrane. Accordingly, the reggies seem to be quite closely associated with the Src-family kinases lck and fyn (Liu J et al. 2005; Slaughter et al. 2003; Stuermer et al. 2001) as shown by both co- immunoprecipitation and co-localization at the LM and EM level. Several large transmembrane proteins were also co-immunoprecipitated with the reggies including ABCA1 (Bared et al. 2004), an ABC transporter implicated in cholesterol transport to high density lipoprotein particles. Co-immunoprecipitation of the thrombin receptor PAR-1 with reggie-1 from melanoma cell lines (Hazarika et al. 2004) and identification of an interaction between Neuroglobin and reggie-2 in a yeast-two-hybrid screen (Wakasugi et al. 2004) suggest a function of the reggies in G-protein coupled receptor signalling.

Considering the identification of the reggies as proteins upregulated during axon regeneration, control of cytoskeletal dynamics seems a good guess for reggie function. Indeed, overexpression of full-length reggie-1 or of the cytosolic 42 kDa variant induces filopodia formation in several cell types (Hazarika et al. 1999; Neumann-Giesen et al. 2004). A direct

(23)

link between the reggies and the regulation of the actin cytoskeleton was reported by Kimura et al. who described the direct interaction between the sorbin homology (SoHo) domain present in adaptor proteins of the vinexin-family and reggie-2 (Kimura et al. 2001). The vinexin-family of adaptor proteins consists of vinexin-α and -β, CAP/ponsin (c-Cbl associated protein) and ArgBP2 (Kioka et al. 2002). These ubiquitously expressed proteins are characterized by one SoHo domain at their N-terminus and three SH3 domains in their C- terminal region. While the SoHo domain provides the means of membrane recruitment via its interaction with the reggies, the vinexin family proteins bind to a variety of proteins via their SH3 domains, e.g. to the ubiquitin-ligase and adaptor protein c-Cbl, the tyrosine kinase c-Abl and main regulators of cytoskeletal dynamics like vinculin, afadin and the regulators of small GTPases Grb4 and Sos (Kioka et al. 2002). Thus, via the interaction with vinexin family members, the reggies might recruit multiprotein signalling complexes to membrane microdomains to direct cytoskeletal dynamics. Reggie-dependent recruitment of a CAP/c-Cbl complex was shown to be essential for the insulin-receptor stimulated insertion of Glut4 glucose transporters into the plasma membrane. CAP-mediated recruitment of Cbl into lipid rafts stimulates the cdc42 family GTPase TC10 via a CrkII-C3G complex, and this pathway necessarily amends the well-known PI3-kinase dependent signalling pathways downstream of the insulin receptor (Baumann et al. 2000; Kimura et al. 2001). Similarly, the recruitment of an Cbl/Pyk2 complex to plasma membrane microdomains was shown to be essential for neuritogenesis in differentiating PC12 cells (Haglund et al. 2004). A role of the reggies in control of cytoskeletal remodelling might thus account for the observed upregulation of the reggies during differentiation (see above).

The reggies seem to play an important role in T cell signalling. Both reggies are expressed in B and T lymphocytes (Solomon et al. 2002; Stuermer et al. 2001). In T cells, reggie-1 and -2 are associated with the Src-kinase lck and fyn as shown by co-immunoprecipitation (Slaughter et al. 2003; Stuermer et al. 2001; Stuermer et al. 2004; Tu et al. 2004). Reggie-2 binds fyn regardless of its activity as it also binds to the kinase-dead mutant (Liu J et al.

2005). Furthermore the reggies are associated with the adaptor protein LAT (Slaughter et al.

2003). The association with LAT and lck apparently increases after stimulation of the cell by CD3/CD28 co-stimulation; furthermore vimentin and IKKβ associate with reggie complexes after activation of the cell (Slaughter et al. 2003; Tu et al. 2004). Both in stable cell lines and in peripheral lymphocytes, reggie-1 and -2 exhibit a strikingly polarized localization in resting cells, accumulating in one aspect of the cell forming a “preformed cap” (Rajendran et al.

2003; Stuermer et al. 2004). Stimulation of the cells by crosslinking of cell surface components leads to the accumulation of signalling molecules like lck, LAT and the TCR/CD3 complex in the region of the preformed reggie cap (Rajendran et al. 2003; Stuermer et al. 2004) reflecting the increased biochemical association. Thus, the preformed caps provided by the reggies might act as an organizing centre for T cell activation.

(24)

1.3.6 Reggie proteins in health and disease

Lipid rafts were implicated in a variety of pathological conditions ranging from arteriosclerosis to Ebola-infections (Simons and Ehehalt 2002). Entry of different viruses and other pathogens also occurs via lipid rafts, the most prominent example being the binding of HIV’s gp120 to CD4 and chemokine receptors which leads first to raft clustering and then to fusion of the virus envelope with the plasma membrane. Considering the proposed functions of the reggies, an involvement in pathogen entry seems very likely. Indeed, several reports implicate the reggies in host cell invasion by Plasmodium. Invasion of erythrocytes by Plasmodium falciparum can be inhibited by cholesterol depletion suggesting that lipid rafts are the site of invasion (Samuel et al. 2001). During formation of the parasitophorous vacuole (PV), several raft-associated proteins like CD59 and Gαs are recruited from the plasma membrane of the host cell to the PV membrane (Lauer et al. 2000). Similarly both reggies are recruited to the PV in erythrocytes (Murphy et al. 2004). Preliminary data suggest a similar translocation of reggie-2 to the PV during infection of hepatocytes by Plasmodium berghei (V Heussler, personal communication). Considering the vast remodelling of cell function and membrane trafficking in Plasmodium-infected cells, hijacking the reggie scaffolds and their function might be of particular importance for the parasite.

Some reports hint to a role of the reggies in the pathogenesis of neurodegenerative diseases like Parkinson and Alzheimer’s disease (AD). The formation of senile plaques is a major hallmark of both diseases, which consist in case of AD of the peptide Aβ. Aβ is generated from APP, a transmembrane precursor protein, by successive cleavage by two proteases. Both the γ-secretase complex as well as the β-secretase BACE are concentrated in lipid microdomains (Abad-Rodriguez et al. 2004; Vetrivel et al. 2004) while their substrate, APP, is largely excluded from lipid rafts (Abad-Rodriguez et al. 2004; Parkin et al. 2003). Thus the partitioning of the enzymes responsible for Aβ generation (but not their substrate APP) into lipid rafts normally prevents the production of large amounts of Aβ (Kaether and Haass 2004). Primitive senile plaques in non-demented persons show strong reggie-2 labelling and in AD patients both reggies show significantly increased staining of the cortex (Kokubo et al.

2000). An ultrastructural study showed an accumulation of reggie-2 in lysosomes of neurons having neurofibrillary tangles (a second hallmark of AD) (Girardot et al. 2003). Similarly, in neurons of transgenic mice overexpressing human APP and presenilin-1, Aβ and reggie-2 accumulated in multivesicular bodies (Langui et al. 2004). Accordingly, a direct binding of reggie-1 to the intracellular domain of APP was shown (Chen TY et al. 2006). Thus, the progression of Alzheimer’s disease is accompanied by an accumulation of the reggies at sites of Aβ production and secretion. Similarly, an upregulation of reggie-2 expression in the substantia nigra of Parkinson patients was recently reported (Jacobowitz and Kallarakal 2004).

Some strong hints are linking reggie microdomains to the progression of prion diseases. These diseases (e.g. BSE, CJD or scrapie) are caused by an aggregation of misfolded cellular prion