Marzieh Fanaei-Phd Thesis.Pdf

Investigating the role of tetraspanins in multinucleated giant cell formation using recombinant proteins

A thesis submitted in part fulfillment of the requirements for the degree of Doctor of Philosophy

Marzieh Fanaei

Department of Molecular Biology and Biotechnology

University of Sheffield

July 2013 Abstract

Tetraspanins are a superfamily of membrane proteins found in almost all multicellular organisms. Of the 33 human tetraspanins identified so far, most remain largely uncharacterised. Tetraspanin proteins are reportedly involved in numerous biological processes occurring at the cell surface including adhesion, membrane fusion and bacterial and viral entry into host cells. These proteins are thought to facilitate and/or regulate formation of functional complexes at certain functional sites on the cell membrane where these processes take place. The large extracellular domain of human tetraspanin proteins is a disulfide-containing protein domain that appears to be essential to tetraspanin interactions with other proteins on the cell surface. Recombinant tetraspanin EC2 domains have emerged as alternative tools to investigate the role of tetraspanins in biological processes as diverse as bacterial adhesion to epithelial cells, HIV-1 infection of macrophages, assembly of the photoreceptor membranes in the retina and multinucleated giant cell formation.

The aim of this study was to characterise the role of several new tetraspanins in multinucleated giant cell formation in monocytes by production of recombinant tetraspanin EC2 domains in bacteria. Expression of recombinant proteins was optimised to maximise the yield of soluble protein and recombinant proteins were characterised using biochemical and biophysical methods such as Western blotting and circular dichroism. The functional activity of recombinant tetraspanin proteins was determined in three different monocyte fusion systems including Concanavalin A- induced monocyte fusion into multinucleated giant cells (MGC), Burkholderia thailandensis-induced MGC formation and Receptor activator of nuclear factor kappa-B (RANKL)-induced multinucleated osteoclast formation. Monocytes are reportedly sensitive to the presence of bacterial contaminants in recombinant protein preparations obtained from a bacterial expression system, therefore various measures were adopted to minimise and rigorously exclude the contribution of bacterial contaminants to recombinant protein activity in monocyte fusion assays.

Acknowledgements

I would like express my deepest gratitude to my supervisors Dr Lynda Partridge and Dr Peter Monk for providing invaluable academic guidance and unfailing support and encouragement throughout this project. Without their diligence and commitment, this work would not have been possible. I must also thank my PhD advisors Dr Ewald Hettema and Prof Per Bullough for their knowledgeable advice and constructive criticism.

I would also like to thank lab members past and present, John Palmer, Atiga Elgawidi, Tom Champion, Daniel Cozens, Andrew O’Leary, Fawwaz Ali, Jenny Ventress, Rachel Bell, Noha Hassuna, Arunya Jiraviriyakul, Hanan Niaz, Kathryn Wareham, Sean Fang and Qi Ding for providing a supportive and enjoyable work environment and lasting friendships.

Practical advice and collaborations with other members of MBB and other department also proved to be invaluable to me. I must thank Dr Qaiser Sheikh, Paul Brown, Tom Minshull, Dr Paul Gokhale, Dr Stephen Brown and Dr Rosemary Staniforth in this regard. I would also like to acknowledge the contribution of volunteers who donated blood for my experiments including several members of our group and the MBB department.

Finally, I would like to thank my parents, Ali and Maryam for all the selfless sacrifices they have made for me. I owe everything to them.

Abbreviations

A Absorbance ADAM A disintegrin and a metalloproteinase agm African green monkey AIDS Acquired immunodeficiency syndrome Ampr Ampicillin resistance ANOVA Analysis of variance

B. pseudomallei Burkholderia pseudomallei B. thailandensis Burkholderia thailandensis BBN BSA/BSS/Sodium azide BMM Bone marrow-derived macrophages BSA Bovine serum albumin BSS Balanced salts solution

CCL2 Chemokine C-C ligand 2 CD Circular dichroism CD Cluster of Differentiation cDNA complementary DNA CDV Canine distemper virus C. elegans Caenorhabditis elegans CFSE Carboxyfluorescein diacetate CHAPS 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate Con A Concanavalin A

DC-STAMP Dendritic cell-specific transmembrane protein DMEM Dulbecco’s modified eagles medium DNA Deoxyribonucleic acid dNTP Deoxyribonucleotide triphosphate

EC1 Extracellular 1 EC2 Extracellular 2 ECL Enhanced chemiluminescence E. coli Escherichia coli EDTA Ethylenediaminetetraacetic acid EGFP Enhanced green fluorescent protein ELISA Enzyme-linked immunosorbent assay env (viral) envelope protein ER Endoplasmic reticulum

FACS Fluorescence activated cell sorting FBGC Foreign body giant cell FCS Foetal calf serum FITC Fluorescein isothiocyanate fM femtomolar

GFP Green fluorescent protein GMCSF Granulocyte-macrophage colony stimulating factor GST Glutathione-s-transferase

iii h human HBSS Hank’s Balanced salts solution HCV Hepatitis C virus HEC-1-B Human endometrial carcinoma-1-B HEK293 Human embryonic kidney 293 His polyhistidine HIV-1 Human immunodeficiency virus-1 HRP Horse radish peroxidase HTLV-1 Human T-cell leukaemia virus type-1

IC Intracellular IC50 50% inhibitory concentration IFN-γ Interferon-γ Ig Immunoglobulin IL Interleukin IPTG Isopropyl- β-D-1-thiogalactopyranoside

Kb Kilobasepairs KDa Kilodaltons

LAL Limulus Amoebocyte Lysate LB Lysogeny broth LDH Lactate dehydrogenase LGC Langhans giant cell LPS Lipopolysaccharide

µg Micrograms m Murine mA milliAmps mg milligrams MBP Maltose-binding protein ml Millilitre mM Millimolar MCSF Macrophage colony stimulating factor MDM MCSF-dependent macrophages MFR/SIRPα Macrophage fusion receptor/signal-regulatory protein-α MGC Multinucleated giant cells MHC-1 Major histocompatibility complex-1 MMP9 Matrix metalloproteinase 9 mRNA messenger ribonucleic acid MW Molecular weight

N. meningitides Neisseria meningitides NFκB Nuclear factor κ B NFATc1 Nuclear factor of activated T-cells, cytoplasmic 1 ng nanograms nm nanometre nM nanomolar

OC-STAMP Osteoclast stimulatory transmembrane protein OD Optical density OS Outer segments

PAMPs Pathogen-associated molecular patterns

PBMC Peripheral blood mononuclear cells PBS Phosphate-buffered saline PCR Polymerase chain reaction pg picograms pM picomolar PMX Polymyxin Poly (I:C) dsRNA Polyinosinic:polycytidylic acid double-stranded RNA PRR Pattern recognition receptor Pro-HB-EGF Pro-heparin-binding epidermal growth factor PSG17 Pregnancy-specific glycoprotein17

RANK Receptor activator of nuclear factor kappa-B RANKL Receptor activator of nuclear factor kappa-B ligand RCF Relative centrifugal force RDS Retinal degeneration slow RNA Ribonucleic acid ROM-1 Rod outer-segment membrane protein-1 rpm Revolutions per minute

SDS Sodium dodecyl sulphate SDS-PAGE Sodium dodecyl sulphate polyacrylamide gel electrophoresis SEM Standard error of the mean siRNA small interfering ribonucleic acid SOB Super optimal broth SOC Super optimal broth with catabolite repression SRB Sulforhodamine B

TAE Tris-acetate-EDTA TB Terrific broth TBS Tris-buffered saline TBST Tris-buffered saline/Tween20 TCA Trichloroacetic acid TEM/TERM Tetraspanin-enriched microdomain TEMED Tetramethylethylenediamine TEV Tobacco etch virus TGF-β Transforming growth factor-beta TH T helper cell TLR Toll-like receptor Tm Melting temperature TM Transmembrane TMB 3,3',5,5'-Tetramethylbenzidine TNF-α Tumour necrosis factor-alpha Tris Tris (hydroxymethyl) methylamine Tspan Tetraspanin

UP1a Uroplakin 1a UP1b Uroplakin 1b UPII Uroplakin 2 UPIIIa Uroplakin 3a UV Ultraviolet

V Volume w Weight

Amino acid abbreviations

Name Three letter code One letter code

Alanine Ala A Arginine Arg R Asparagine Asn N Aspartic acid Asp D Cysteine Cys C Glutamic acid Glu E Glutamine Gln Q Glycine Gly G Histidine His H Isoleucine Ile I Leucine Leu L Lysine Lys K Methionine Met M Phenylalanine Phe F Proline Pro P Serine Ser S Threonine Thr T Tryptophan Trp W Tyrosine Tyr Y Valine Val V

List of Tables

Page Chapter 1 Table 1.1 Generation of monocyte-derived multinucleated giant cells in vitro 33 Table 1.2 A summary of in vitro studies utilising recombinant tetraspanin 41 EC2 domain to investigate tetraspanin functions Table 1.3 A summary of LPS reduction techniques 46

Chapter 2 Table 2.1 Laboratory equipment 50 Table 2.2 Computer software 50 Table 2.3 Electrophoresis gels for DNA and protein analysis 51 Table 2.4 Buffers 51 Table 2.5 Antibodies 52 Table 2.6 Bacterial strains 53 Table 2.7 Bacterial growth media 53 Table 2.8 Antibiotics 54 Table 2.9 cDNA and protein sequences 54 Table 2.10 Oligonucleotide primers 55 Table 2.11 Restriction endonucleases 56 Table 2.12 PCR, digest and ligation mix composition 56

Chapter 3 Table 3.1 Cysteines and disulfide bonds in GST and tetraspanin EC2 domains 75

Chapter 4 Table 4.1 Inhibitory activity of GST and GST-CD9 in various studies 103 Table 4.2 Relationship between recombinant protein yield and LPS content 105 Table 4.3 Reduction of recombinant protein LPS content using a detergent 107 wash method

vii

List of Figures

Page Chapter 1 Figure 1.1 Structural features of tetraspanin proteins 2 Figure 1.2 Structure of tetraspanin EC2 domains 8 Figure 1.3 Biological membrane fusion 15 Figure 1.4 Monocyte differentiation/activation pathways 24 Figure 1.5 Mechanisms of monocyte fusion 30

Chapter 2 Figure 2.1 Separation of the mononuclear fraction by density gradient 67 centrifugation Figure 2.2 Reducing adherent cell density 68 Figure 2.3 Image acquisition 69

Chapter 3 Figure 3.1 Important features of the pGEX-KG vector used for recombinant 73 tetraspanin expression Figure 3.2 Amplification of eight tetraspanin EC2 DNA sequences by PCR 77 Figure 3.3 Analytical restriction endonuclease digest of pGEX-KG plasmids 78 carrying EC2 DNA fragments Figure 3.4 SDS-PAGE analysis of recombinant GST-tagged tetraspanin protein 80 expression Figure 3.5 Purification of recombinant tetraspanins using glutathione affinity 82 chromatography Figure 3.6 Detection of soluble and insoluble GST-tagged material in whole cell 83 lysates following recombinant protein induction Figure 3.7 Production of four-cysteine recombinant tetraspanins in E. coli 84 Rosetta-gami strain Figure 3.8 Production of recombinant proteins with six or eight cysteines in E. 85 coli Rosetta-gami: Analysis by SDS-PAGE and densitometry Figure 3.9 Optimisation of tetraspanin protein expression in E. coli SHuffle 87 strain Figure 3.10 Production of recombinant proteins with six or eight cysteines in E. 88 coli SHuffle strain: Analysis by SDS-PAGE and Western blotting Figure 3.11 Western blotting of recombinant tetraspanin proteins with anti-GST 90 antibody Figure 3.12 Western blotting of recombinant tetraspanin proteins with a range 93 of anti-tetraspanin antibodies Figure 3.13 Analysis of recombinant GST-CD82EC2 by ELISA method 94

viii

Chapter 4 Figure 4.1 Immunofluorescence images of monocytes stimulated with Con A 101 and recombinant GST and GST-CD9 Figure 4.2 Effects of recombinant GST and CD9 protein on Con A-induced 102 multinucleated giant cell (MGC) formation Figure 4.3 LPS content of recombinant proteins used in this study 105 Figure 4.4 ELISA method to compare reactivity of GST-CD9 constructs with 109 various antibodies Figure 4.5 Effect of LPS reduction on recombinant protein activity in the MGC 110 assay Figure 4.6 Dose-dependent inhibitory effects of LPS on Con A-induced MGC 112 formation Figure 4.7 Significant inhibitory effects of LPS on MGC formation 114 Figure 4.8 Immunofluorescence images of human monocytes treated with 115 various doses of LPS Figure 4.9 Contribution of bacterial contaminants to recombinant protein 117 activity in Con A-induced MGC formation Figure 4.10 Analysis of native and boiled GST-CD9 proteins by Western blotting 118 Figure 4.11 Effect of polymyxin B sulphate on the inhibition of MGC formation 120 by LPS

Chapter 5 Figure 5.1 Effect of GST and GST-CD9 proteins and LPS on bacterial adhesion 127 to HEC-1-B cells Figure 5.2 Testing of a panel of recombinant tetraspanins in the bacterial 128 adhesion assay Figure 5.3 Induction of osteoclastogenesis in adherent human PBMC using the 131 cytokines RANKL and MCSF Figure 5.4 Recombinant protein activity in the in vitro osteoclastogenesis assay 132 Figure 5.5 LPS activity in the in vitro osteoclastogenesis assay 134 Figure 5.6 Recombinant protein and LPS activity in Burkholderia thailandensis- 137 induced fusion of J774.2 cells

Table of Contents

Abstract i Acknowledgements ii Abbreviations iii Amino acid code vi List of Tables vii List of Figures viii Table of Contents x

Chapter 1: Introduction 1 1.1 The tetraspanin superfamily of proteins 1 1.1.1 Structural features of tetraspanins and associated functions 3 1.1.1.1 Transmembrane domains 3 1.1.1.2 Extracellular domains 5 1.1.1.3 Intracellular domains 7 1.1.2 Tetraspanin patterns of expression and modes of action 9 1.1.2.1 The tetraspanin web theory 10 1.2 The biological process of fusion and the role of tetraspanins 14 1.2.1 Virus-host cell fusion and virus-induced syncytium formation 16 1.2.2 Outer segment formation in photoreceptor cells: an intracellular 18 fusion event 1.2.3 Cell-cell fusion 19 1.2.3.1 Sperm-egg fusion 20 1.2.3.2 Developmental fusion events 21 1.2.3.3 Mononuclear phagocyte fusion and functions 23 1.2.3.3.1 Monocyte activation pathways 27 1.2.3.3.2 Mechanisms of monocyte fusion 29 1.2.3.3.3 Studying monocyte fusion in vitro 31 1.2.3.3.4 Regulation of monocyte fusion by tetraspanins 34 1.3 Studying tetraspanin functions using recombinant EC2 domains 38 1.3.1 Issues regarding the use of recombinant proteins in immunological 41 research 1.3.2 Documented effects of bacterial contaminants on monocytes and 42 monocyte-derived giant cells 1.3.3 Dealing with recombinant protein contamination 44 1.3.3.1 Measuring the level of contamination 45 1.3.3.2 LPS reduction techniques 45

1.3.3.3 LPS neutralisation 47 1.3.3.4 Recombinant protein inactivation 47 1.3.3.5 Alternative expression systems 48 1.3.4 Aims 48

Chapter 2: Materials and methods 50 2.1 Materials 50 2.1.1 Laboratory equipment and instrumentation 50 2.1.2 Computer software 50 2.1.3 Laboratory reagents and buffers 51 2.1.3.1 Electrophoresis gels 51 2.1.3.2 Buffers 51 2.1.3.3 Antibodies 52 2.1.4 Cloning materials 53 2.1.4.1 Bacterial strains 53 2.1.4.2 Growth media 53 2.1.4.3 Antibiotics 54 2.1.4.4 cDNA and protein sequences 54 2.1.4.5 Oligonucleotide primers 55 2.1.4.6 Restriction enzymes and buffers 56 2.1.4.7 PCR, digest and ligation mix composition 56 2.2 Methods 57 2.2.1 General molecular biology methods 57 2.2.1.1 DNA Sequencing 57 2.2.1.2 Plasmid DNA purification 57 2.2.1.3 Preparation of glycerol stocks 57 2.2.1.4 Agarose gel electrophoresis 58 2.2.1.5 SDS-PAGE 58 2.2.1.6 Coomassie staining and de-staining 58 2.2.1.7 Western blotting 59 2.2.1.8 Probing nitrocellulose membranes 59 2.2.1.9 Bradford assay 59 2.2.1.10 standard ELISA 59 2.2.2 Sub-cloning tetraspanin EC2 domains 60 2.2.2.1 Determination of EC2 domain boundaries 60 2.2.2.2 PCR primer design 60 2.2.2.3 PCR amplification of tetraspanin EC2 DNA 60 2.2.2.4 PCR product purification 61

2.2.2.5 Restriction digests 61 2.2.2.6 Ligation of amplified EC2 fragments into pGEX vector 62 2.2.2.7 Amplification of ligated vectors 62 2.2.3 Recombinant protein expression 63 2.2.3.1 Expression in E. coli Rosetta-Gami strain 63 2.2.3.2 Expression in E. coli SHuffle strain 63 2.2.4 Bacterial cell lysis by sonication 64 2.2.5 Affinity purification 64 2.2.6 Dialysis 65 2.2.7 LPS reduction using Triton X-114 66 2.3 Tissue culture methods 67 2.3.1 Multinucleated Giant Cell (MGC) formation in peripheral blood 67 monocytes 2.3.1.1 Isolation of peripheral blood mononuclear cells 67 2.3.1.2 Concanavalin A-induced MGC formation assay 68 2.3.1.3 Immunofluorescence cell staining 69 2.3.1.4 Immunofluorescence imaging 69 2.3.1.5 Quantifying giant cell fusion 70 2.3.1.6 Statistical methods 70 2.3.1.7 Osteoclastogenesis assay 71 2.3.2 Burkholderia thailandensis-induced MGC formation in J774.2 71 macrophage-like cells 2.3.2.1 Passage of adherent J774.2 cells 71 2.3.2.2 Burkholderia thailandensis-induced MGC formation assay 71 2.3.3 Bacterial adhesion assay 72

Chapter 3: Production of recombinant tetraspanin EC2 proteins 73 3.1 Introduction 3.1.1 Expression of disulfide bond-containing proteins in E. coli 74 3.2 Results 77 3.2.1 Sub-cloning of tetraspanin EC2 domain cDNA into pGEX-KG vector 77 3.2.2 Expression and purification of recombinant tetraspanin proteins in 79 E. coli Rosetta-gami strain 3.2.2.1 Recombinant tetraspanin expression 79 3.2.2.2 Recombinant protein purification using glutathione affinity 81 chromatography 3.2.2.3 Production of recombinant tetraspanins containing four 83 cysteines

xii

3.2.2.4 Production of recombinant tetraspanins containing six or eight 85 cysteines 3.2.3 Expression and purification of recombinant tetraspanin proteins in 86 E. coli SHuffle strain 3.2.3.1 Optimising recombinant protein expression 86 3.2.3.2 Production of recombinant tetraspanins containing six or eight 87 cysteines 3.2.4 Analysis of recombinant tetraspanin protein folding 89 3.2.4.1 Western blotting with anti-GST antibody 89 3.2.4.2 Western blotting with anti-tetraspanin antibodies 91 3.2.4.3 Analysis of recombinant GST-CD82EC2 protein by ELISA 92 3.3 Discussion 95

Chapter 4: The role of tetraspanins in Concanavalin A-induced 99 Multinucleated Giant Cell (MGC) formation 4.1 Introduction 99 4.2 Results 99 4.2.1 Establishing negative and positive controls 99 4.2.2 Measuring LPS contamination levels in recombinant protein 104 preparations 4.2.3 Reducing LPS contamination using a detergent wash method 106 4.2.4 ELISA analysis of recombinant GST-CD9 proteins 108 4.2.5 Effect of LPS reduction on recombinant protein activity in the MGC 109 assay 4.2.6 Effects of pure LPS on MGC formation 111 4.2.7 Contribution of bacterial contaminants to recombinant protein 116 inhibitory activity in the MGC assay 4.2.8 Effect of polymyxin B on the inhibitory activity of LPS and LPS- 118 contaminated recombinant proteins 4.3 Discussion 121

Chapter 5: Recombinant tetraspanin protein activity in other functional 125 assays 5.1 Introduction 125 5.2 Results 126 5.2.1 Bacterial adhesion to epithelial cell 126 5.2.1.1 Effect of control proteins and LPS on bacterial adhesion 126 5.2.1.2 Effect of a panel of recombinant tetraspanins on bacterial 128

xiii

adhesion 5.2.2 Recombinant protein activity in RANKL-induced osteoclastogenesis 129 5.2.3 Effect of recombinant proteins on the fusion of a murine 135 macrophage-like cell line induced by the bacterium Burkholderia thailandensis 5.3 Discussion 138

Chapter 6: General Discussion 140

Bibliography 146

Appendix 1: Multiple sequence alignment of 33 human tetraspanins 165

xiv

Chapter 1

Chapter 1: Introduction

1.1 The tetraspanin superfamily of proteins Tetraspanins are a superfamily of small (20-30KDa) transmembrane proteins expressed on the cell surface and on the intracellular membranes of eukaryotic cells (Hemler, 2008; Berditchevski and Odintsova, 2007). The existence of a new protein superfamily came to light in 1990 when the myeloma-associated antigen ME491 (CD63) was discovered to have strong sequence homology to the Sm23 protein of the human pathogen Schistosoma masoni (Wright et al., 1990) as well as to the target of an anti-proliferative antibody (CD81) and the B-lymphocyte antigen CD37 (Oren et al., 1990). Since then, the superfamily has grown to include members in a wide range of organisms studied including 33 members in humans, 37 in the fruit fly Drosophila melanogaster, 20 in the nematode worm Caenorabditis elegans and 23 in the sea anemone Nematostella vectensis (DeSalle et al., 2010; Huang et al., 2005). Tetraspanins have also been discovered in plants, multicellular fungi and a limited number of unicellular organisms such as the parasitic protozoan Entamoeba histolytica (Huang et al., 2005; Charrin et al., 2009).

Phylogenetic analyses suggest that all tetraspanins likely evolved from a single ancestral gene by gene duplication and subsequent sequence divergence (Garcia- Espana et al., 2008; DeSalle et al., 2010). The level of amino acid sequence conservation between species homologs varies widely, for example, Tspan5 is 100% in humans, mice and cows whilst human and murine Tspan22 (also called RDS) only share 36% identity. However, the general protein architecture of tetraspanins (figure 1.1) has been highly conserved through evolution and serves to distinguish tetraspanins from other four-span transmembrane proteins such as the L6 family (Hemler, 2008; Huang et al., 2010).

Chapter 1

A) B)

Figure 1.1 Structural features of tetraspanins proteins.

A) A generic schematic structure of tetraspanins. Tetraspanins are type III integral membrane proteins (Imai et al., 1992) with four transmembrane domains (labelled 1-4), three cytoplasmic regions (purple, green and brown lines) and two extracellular domains of unequal size, EC1 (light blue) and EC2 (light green, yellow and orange). According to the crystallographic structure of CD81 EC2 (Kitadokoro et al., 2001), this domain is composed of five helices (A-E), of which helices A, B and E are a constant feature of all tetraspanins (Seigneuret et al., 2001). The variable subdomain of the EC2 (orange oval) is stabilised by a number of disulfide bridges (dashed lines) between cysteine residues (grey circles). This region is predicted to be highly variable in secondary structure and tertiary fold in tetraspanins (Seigneuret et al., 2001), hence the C and D helices of CD81 EC2 have not been labelled in this drawing. The helices of transmembrane domains 3 and 4 extend beyond the membrane and are contiguous with helices A and E (light green and orange, respectively) in the EC2 domain (Kovalenko et al., 2005; Seigneuret, 2006; Min et al., 2006). The intracellular regions contain membrane-proximal cysteine residues (grey circles) which can be palmitoylated (Figure based on Hemler, 2005 and information in articles cited in the text).

B) A molecular model of tetraspanin UPIa alongside an electron density map (blue) of its single- span partner protein UPII. The intracellular regions are missing from this model. Notable features are the overall rod shape of the tetraspanin and the packing of EC1 against the conserved helices (A, B and E) of the EC2 domain.

Chapter 1

1.1.1 Structural features of tetraspanins and associated functions Human tetraspanin proteins range in size between 204 and 355 amino acid residues (Charrin et al., 2009). A schematic representation of tetraspanin proteins is shown in figure 1.1 A, illustrating membrane topology and conserved structural features. This drawing is based in part on a molecular model of the tetraspanin UPIa (figure 1.1 B) as well as the sources cited in the figure legend. The structure and function of the transmembrane, extracellular and intracellular regions is discussed in more detail in the following sections.

A multiple sequence alignment of all 33 human tetraspanins, illustrates the level of amino acid conservation within each of these domains (Appendix 1). As is evident from this alignment, transmembrane domains 1, 2 and 3 and the short intracellular loop are highly conserved whilst TM4, the N- and C-terminii and the EC1 are more divergent. The EC2 domain has highly variable regions within a conserved framework, with important functional implications (Seigneuret et al., 2001).

1.1.1.1 Transmembrane domains In 28 human tetraspanin sequences analysed, the highly conserved TM1-3 contain a heptad repeat motif (abcdefg)n where positions a and d are occupied by apolar residues (Kovalenko et al., 2005). Molecular modelling of tetraspanins CD9 and CD81 suggests that conserved bulky residues in TM1 pack tightly against small residues in TM2 and vice versa, forming classic ‘knobs-into-holes’ interactions (Kovalenko et al., 2005; Seigneuret, 2006). These features suggest that the tetraspanin transmembrane region is an α-helical coiled coil, a stable structural fold found in numerous other proteins (Mason and Arndt, 2004). A few partially conserved hydrophilic residues may be involved in intramolecular hydrogen bonding or ionic interactions, further stabilising the coiled coil (Stipp et al., 2003). However, molecular models of CD9 and CD81 do not specify such a role for these residues (Kovalenko et al., 2005; Seigneuret et al., 2006).

The study of the uroplakin tetraspanins illustrates the involvement of transmembrane domains in tetraspanin function. The uroplakins UPIa and UPIb are a pair of closely related tetraspanins expressed almost exclusively in the mammalian urothelium, a multi-layered epithelium lining most of the urinary tract (Sun et al., 1996). The

Chapter 1 urothelium functions as an extremely efficient permeability barrier, able to maintain steep water and urea gradients (Liang et al., 2001). The apical (outward-facing) surface of the urothelium is covered by plaques, natural 2-dimensional crystals composed of heterodimers of the tetraspanin uroplakins UPIa and UPIb with their non-tetraspanin single-span partners UPII and UPIIIa, respectively. An intermediate resolution structure of native murine uroplakin plaques (figure 1.1 B) proposes that each heterodimer is a tightly packed five-helix bundle with TM3 and 4 of the tetraspanin providing contacts with the non-tetraspanin helix (Min et al., 2006). The tall head domains of UPII and UPIIIa cap the shorter EC2 domains of UPIa and UPIb, respectively (figure 1.1 B.). As predicted for CD9 and CD81 (Kovalenko et al., 2005; Seigneuret, 2006), the helices of uroplakin TM3 and 4 extend beyond the plasma membrane and are contiguous with helices A and E in the stalk region of the EC2 domain (Min et al., 2006).

UPIa was also discovered to be a receptor for type-1 fimbriated uropathogenic E. coli, directly binding to the FimH protein present at the tip of bacterial fimbriae (Zhou et al., 2001). FimH binding to uroplakin plaques triggers phosphorylation of the UPIIIa cytoplasmic tail and downstream signalling events culminating in bacterial invasion into the urothelial substrata (Thumbikat et al., 2009). Comparison of the 3D structure of native and FimH-bound uroplakin plaques showed that FimH binding to UPIa induces global conformational changes in all four uroplakins including concerted twisting movements of the transmembrane domains (Wang et al., 2009). Thus a mechanism is suggested whereby a UPIa-ligand binding event at the cell surface is transmitted across the plasma membrane through transmembrane domain conformational changes, as is observed in G-protein coupled receptor signalling (Wang et al., 2009; Pierce et al., 2002).

The strength of intramolecular interactions between tetraspanin transmembrane domains was demonstrated by an experiment where co-expression of tetraspanin CD82 TM1 and mutant CD82 lacking TM1 rescued the latter protein’s cell surface expression (Cannon and Cresswell, 2001). Tetraspanin transmembrane domains may also be involved in homophilic tetraspanin associations, for example conserved glycines residues in the heptad motif of TM1 and 2 were predicted to be involved in

Chapter 1

CD9 dimerisation (Kovalenko et al., 2005). Oligomerisation is a common feature of proteins containing heptad repeats (Parry et al., 2008).

1.1.1.2 Extracellular domains A characteristic feature of tetraspanin proteins is the presence of two extracellular domains of unequal size (Charrin et al., 2009). In human tetraspanins, the small extracellular domain (EC1), flanked by TM1 and TM2, ranges in size between 13 and 31 amino acids whilst the large extracellular domain (EC2), flanked by TM3 and 4, is typically 69 to 132 amino acids long (Garcia-Espana et al., 2008).

The EC1 domain does not appear to have regular secondary structure although in some tetraspanins, short helical regions (figure 1.1b) or small β-strands may be present (Seigneuret, 2006). Nonetheless, molecular modelling supports the idea that the EC1 packs against a structurally conserved fold of the EC2 domain, as shown in figure 1.1 B (Min et al., 2006; Seigneuret, 2006). The functional importance of EC1 remains unclear, perhaps because it is somewhat masked structurally by the considerably larger EC2 domain (figure 1.1b). In any case, the absence of EC1 does not seem to impact on the structural integrity of the EC2 domain (Kitadokoro et al., 2001; Rajesh et al., 2012) or its functional activity in various in vitro assays (section 1.3).

Tetraspanin EC2 domains are highly variable in length and sequence (Garcia-Espana et al., 2008) but have a number of characteristic features, the most important of which is the presence of several cysteines and conserved cysteine-containing motifs (Stipp et al., 2003). The CCG motif is found in most tetraspanin superfamily members, including all 33 human proteins (Huang et al., 2005; De Salle et al., 2010). Likewise, almost all tetraspanins discovered so far have an even number of cysteines in the EC2 domain, with the exceptions being the retinal tetraspanins RDS (Retinal degeneration slow) and ROM-1 (Rod outer segment membrane protein-1) and plant tetraspanins (DeSalle et al., 2010). Interestingly, it appears that four-cysteine tetraspanins in animals evolved in a convergent pattern, i.e. by the loss of cysteine pairs from six- or eight- cysteine tetraspanins (DeSalle et al., 2010). This would suggest that cysteine pairs in the EC2 are involved in intramolecular disulphide bridge formation. 1. resolution crystal structure of human CD81 EC2 (Kitadokoro et al., 2001) and various lines of experimental evidence including mutational analysis suggest that disulfide bridges are

Chapter 1 essential for proper folding and function of the EC2 domain (for example Goldberg et al., 1998; Higginbottom et al., 2003).

According to the crystal structure, human CD81 EC2 is composed of five helices (A-E), where helices A and E form a stalk region (figure 1.2 A) stabilised by hydrophobic contacts and two salt bridges (Kitadokoro et al., 2001). In silico analysis of multiple tetraspanins from a wide range of organisms proposes that helices A, B and E are in fact conserved structural elements within the EC2 (Seigneuret et al., 2001; Boucheix and Rubinstein, 2001). A disulfide bridge between cysteine residues close to helices B and E naturally delimits these conserved helices (figure 1.2 B), dividing the EC2 into constant (blue) and variable subdomains (orange). Figure 1.2 B also highlights the fact that all the cysteines in the EC2 are located in the variable subdomain (with the exception of the odd cysteines of RDS/ROM-1 which are involved in intermolecular bonding) (Loewen and Molday, 2000). A recent NMR study of human CD81 EC2 suggests that the region corresponding to helix D in the crystal structure, adopts a highly dynamic loop conformation that is crucially involved in binding to the CD81 ligand, HCV glycoprotein E2 (Rajesh et al., 2012). Interestingly, EC1 and the variable subdomain of EC2 appear to be preferential sites for new intron insertion in the tetraspanin superfamily (Garcia-Espana et al., 2008).

The EC2 domain (and in particular the variable subdomain of EC2) appears to be of great functional importance as it determines specific tetraspanin interactions with partner proteins. In humans, at least 48 missense mutations or in-frame deletions have been described in the RDS gene, resulting in a broad spectrum of retinal degenerative diseases (Boon et al., 2008). 77% of these mutations occur in the EC2 domain and of those 78% are located in the variable subdomain. Tetraspanin-partner interaction sites for RDS/ROM-1, CD9/pro-HB-EGF (pro-heparin binding-epidermal growth factor), CD9/PSG17 (Pregnancy glycoprotein 17), CD81/HCV (Hepatitis C virus) E2 envelope glycoprotein and CD151/α3β1 integrin have all been mapped to the variable subdomain of EC2 (Hasuwa et al., 2001; Ellerman et al., 2003; Higginbottom et al., 2000; Kazarov et al., 2002; Ding et al., 2005). In addition, residues within the variable subdomain of CD9 are required for murine sperm-egg fusion though the CD9 partner has not yet been identified (Higginbottom et al., 2000; Zhu et al., 2002). This has led to the hypothesis that the constant subdomain and conserved

Chapter 1 cysteines in the variable subdomain provide a stable ‘structural scaffold’ onto which can be grafted variable structural motifs for specific interactions with a wide range of partner proteins (Huang et al., 2005; De Salle et al., 2010).

Most tetraspanins contain N-linked (asparagine-linked) glycosylation sites in the EC2 domain (and less frequently in the EC1 domain), (Martin et al., 2005). Such sites are not conserved between different family members but there is some conservation in individual members between species, for example the CD9 glycosylation site is 100% conserved between mouse, rat, primate and bovine CD9 (Maecker et al., 1997). Tetraspanin glycosylation patterns differ in different tissues and/or upon cell activation and may be functionally important (Martin et al., 2005). For example, recognition of tetraspanin UPIa by bacterial FimH protein (mentioned in section 1.1.1.1) is dependent on the glycosylation of a single amino acid residue in the EC2 domain (Zhou et al., 2001). However, RDS folding, stability and function was shown to be unaffected by mutation of the single glyscosylation site in its EC2 domain (Kedzierski et al., 1999).

1.1.1.3 Intracellular domains Tetraspanin proteins contain three intracellular regions; a typically short N-terminus, a small intracellular loop located between transmembrane domains 2 and 3 and a C- terminal tail of variable length (figure 1.1 A and Appendix 1). All three intracellular regions of tetraspanins typically contain several membrane-proximal cysteines which can be palmitoylated (Charrin et al., 2009). Palmitoylation has been shown to facilitate tetraspanin interactions with such membrane components as cholesterol (Charrin et al., 2003), gangliosides (Odintsova et al., 2006) and various proteins including other tetraspanins. A palmitoylation-deficient mutant of tetraspanin CD151 associated normally with its primary partner α3β1 integrin but showed markedly diminished binding to other membrane proteins (including tetraspanins CD9 and CD63) and a more diffuse cell surface distribution (Yang et al., 2002). Similarly, loss of CD81 palmitoylation only slightly reduced its association with its primary partner EWI-2 but almost completely prevented its cell-surface association with CD151 (Zhu et al., 2012). In addition, palmitoylation was found to stabilise CD9 interaction with CD81 and modify its dynamics on the cell surface (Charrin et al., 2002; Espenel et al., 2008). Thus palmitoylation may contribute to tetraspanin-tetraspanin interactions but is

Chapter 1 dispensible for binding to specific partners. Palmitoylation of multiple cysteines is predicted to lend an amphipathic character to the CD81 N-terminal helix (Seigneuret et al., 2001).

C Variable D subdomain

HOOC

NH 2

Figure 1.2 Structure of tetraspanin EC2 domains.

A) A generalised schematic structure of tetraspanin EC2 domains. Helices A, B and E (blue) are structurally conserved features of the EC2 domain. The variable region (orange oval) is delimited by a disulfide bridge (dashed lines) between a highly conserved pair of cysteines at the ends of helices B and E.

B) Crystallographic structure of a soluble human CD81 EC2 domain. The helices A, B and E form a constant subdomain (blue) whilst helices C and D (red) form a variable subdomain stabilised by disulfide bridges (yellow). (Adapted from Seigneuret, 2006).

Chapter 1

The C-terminal tail of human tetraspanins is highly divergent in length and sequence (see Appendix 1). The C-terminal tail of RDS (being the longest in human tetraspanins) appears to be an intrinsically disordered domain capable of inducing membrane fusion in vitro (Boesze-Battaglia et al., 2000; Ritter et al., 2005). Upon membrane binding, a region within the C-terminus with homology to the influenza fusion peptide adopts an amphipathic helical conformation (Edrington et al., 2007a). The fusogenic activity of RDS is regulated by several proteins which bind directly to the C-terminus, including calmodulin and melanoregulin (Boesze-Battaglia et al., 2007; Edrington et al., 2007b). Finally, signal sequences within the RDS C-terminus facilitate targeting to specialised membraneous structures (outer segment membranes) within photoreceptor cells in the retina where photo-transduction takes place (Boesze-Battaglia et al., 2007). The fusogenic function of RDS C-terminal region appears to be unique amongst human tetraspanins as the closest RDS homolog, ROM-1 does not promote fusion in vitro (Boesze-Battaglia and Gretzula, 2003).

Mutation of three C-terminal amino acid residues altered a number of CD9-dependent functions such as microvilli formation, inhibition of cell-cell adhesion and detergent solubility (Wang et al., 2011). This was attributed to aberrant oligomerisation of CD9 on the cell surface. The eight amino acid C-terminal tail of CD151 appears to be important in regulating the functions of its associated partner protein α β1 integrin (Liu et al., 2007). C-terminal deletion or exchange impaired α β1-dependent cellular network formation, migration and cell spreading without altering CD151-α β1 association (Zhang et al., 2002).

The C-terminal tail of several human tetraspanins contains tyrosine-based sorting motifs which may facilitate targeting to intracellular compartments such as lysosomes (Stipp et al., 2003). However, the functionality of these motifs has only been demonstrated for the tetraspanin CD63, which has a predominantly intracellular localisation (Berditchevski and Odintsova, 2007; Charrin et al., 2009).

1.1.2 Tetraspanin patterns of expression and modes of action As a superfamily, tetraspanin proteins are expressed almost ubiquitously on the surface of animal cells (Boucheix and Rubinstein, 2001). Some tetraspanins are present on almost all cell types whereas others are restricted to particular cells and

Chapter 1 tissues (Kitadokoro et al., 2001). Tetraspanin expression levels may vary depending on the differentiation stage of particular cell types, for example, CD37 expression increases dramatically during B cell maturation (Maecker et al., 1997). Concurrent with their expression patterns, some tetraspanins are ascribed multiple roles in membrane biology whilst others perform highly specialised functions. The uroplakin tetraspanins UPIa and UPIb (discussed in section 1.1.1.1) are prime examples of specialised tetraspanins, with a restricted distribution and well-defined functions. The tetraspanins RDS and ROM-1, exclusively expressed in the retina, provide another example (section 1.2.2). It is interesting to note that RDS, ROM-1 and the uroplakin tetraspanins are only distantly related to other human tetraspanins (Seigneuret, 2006).

It has proven difficult to assign specific functions to other superfamily members (such as CD9, CD81 and CD63) since they seem to be involved in a myriad of processes including sperm-egg fusion, giant cell formation, adhesion, motility, cancer metastasis and viral infections (Hemler, 2005). A further complicating factor is the apparent functional redundancy and compensation observed for several tetraspanins. In Drosophila melanogaster, deletion of late bloomer, a tetraspanin expressed on motor neurons, leads to delayed neuromuscular synapse formation. Simultaneous deletion of late bloomer and two other tetraspanins also expressed on motor neurons further enhances this phenotype suggesting that these tetraspanins functionally compensate for the deletion of late bloomer. Strikingly, simultaneous deletion of six other Drosophila tetraspanins yields no discernible phenotype (Fradkin et al., 2002). In mice, deletion of CD9 leads to female infertility, a phenotype that is partially rescued by injection of mouse CD81 mRNA into CD9-/- oocytes (Le Naour et al., 2000; Miyado et al., 2000; Kaji et al., 2000; kaji et al., 2002). Recombinant CD9 EC2 protein inhibited murine sperm-egg fusion but only when pre-incubated with eggs and not sperm, indicating that CD9 interacts laterally (in cis) with a protein on the egg surface rather than in trans with a sperm ligand or counter-receptor (Zhu et al., 2002).

1.1.2.1 The tetraspanin web theory In fact, very few in trans interactions have so far been reported for tetraspanins (Hemler, 2008). The tetraspanin CD81 is a co-receptor for hepatitis C virus (HCV), binding specifically and tightly to HCV glycoprotein E2 (Higginbottom et al., 2003).

Chapter 1

The uroplakin tetraspanin UPIa has already been mentioned as a receptor for the FimH protein of bacterial fimbria (Zhou et al., 2001). There is some evidence to suggest that murine CD9 is a receptor for pregnancy-specific glycoprotein 17 (PSG-17) (Waterhouse et al., 2002) but the significance of this interaction remains unclear.

On the other hand, tetraspanins appear to have a great propensity for lateral interactions with other proteins on the cell surface (Charrin et al., 2009). Numerous proteins are co-immunoprecipitated with tetraspanins including several members of the immunoglobulin and integrin superfamilies, various growth factor receptors, membrane-bound proteases, a limited number of cytoplasmic proteins and also other tetraspanins (Boucheix and Rubinstein, 2001; Hemler, 2003; Hemler, 2008; Boesze- Battaglia et al., 2007; Charrin et al., 2009).

Lateral molecular interactions of tetraspanins have been classed into primary, secondary and tertiary types (Boucheix and Rubinstein, 2001). Primary interactions are the most robust, being stable in highly hydrophobic detergents such as Triton X- 100 (Kovalenko et al., 2005). They constitute direct, specific interactions between tetraspanins and their partners since they can be stabilised by chemical cross-linking, are of high stoichiometry and usually occur early in biosynthesis (Hemler, 2003; Stipp et al., 2003; Martin et al., 2006). The 1:1 association of tetraspanins UPIa and UPIb with their respective non-tetraspanin partners UPII and UPIIIa (as detailed in section 1.1.1.1) fulfils all the above criteria (Tu et al., 2002; Min et al., 2006). The same tetraspanin can form primary interactions with different partners on the same cell or on different cell types, for example, CD151 interacts with several integrins in various cell lines and CD81 interacts with CD19 in B cells and with the T-cell specific molecules CD4/CD8 in T-cells (Hemler, 2008; Levy and Shoham, 2005). A few examples also exist of tetraspanin-tetraspanin primary associations resulting in the formation of homo- and hetero-meric complexes (Loewen and Molday, 2000; Kovalenko et al., 2004).

Lateral association with tetraspanins can regulate the biosynthesis and cell surface expression of primary partner proteins (Berditchevski and Odintsova, 2007). In the absence of CD81, CD19 expression at the surface of primary B cells is dramatically decreased and the CD19 molecules that do reach the surface are differentially glycosylated, leading to impaired B cell receptor signalling. Ectopic expression of

Chapter 1 human CD81 in CD81-/- primary B cells restores cell surface localisation of CD19, demonstrating the involvement of CD81 (Shoham et al., 2006). Similarly, silencing of the predominantly intracellular tetraspanin CD63 leads to re-distribution of its partner protein H+/K+-ATPase to the cell surface (Codina et al., 2005). In addition, the functional activities of several proteins are directly modulated by association with tetraspanins including ligand binding (CD151 and α3β1 integrin) and proteolytic activity (CD9 and ADAM17 metalloprotease) (Nishiuchi et al., 2005; Gutierrez-Lopez et al., 2012).

Heterophilic tetraspanin-tetraspanin associations (via palmitoylated membrane- proximal cysteines) facilitate formation of secondary complexes (Hemler, 2003). Such complexes are recovered using less stringent detergents such as Brij-97 and are thus likely to constitute molecular interactions of a weak or transitory nature (Odintsova et al., 2006; Kovalenko et al., 2005; Martin et al., 2005). Finally, large tertiary complexes containing several tetraspanins, primary partners and a multitude of other proteins may be isolated using mild detergents such as CHAPS (Hemler, 2003). The majority of reported tetraspanin interactions probably fall into this category, being an indirect association between a given tetraspanin and proteins directly associated with another tetraspanin or tetraspanin-partner (Hemler, 2003; Hemler, 2005).

The precise mechanism by which direct or indirect association with a given tetraspanin might regulate the function of a protein remains vague at present. Various biophysical techniques such as fluorescence and immunoelectron microscopy have shown that tetraspanins and associated proteins cluster together in discrete microdomains on the cell surface called tetraspanin-enriched microdomains (TERMs/TEMs) or the ‘tetraspanin web’ (Rubinstein et al., 1996; Berditchevski et al., 2002). TEMs are relatively small in size (~0.2µm2) and are similar to but distinct from another type of membrane microdomain, the lipid raft (Nydegger et al., 2006; Charrin et al., 2003; Hemler, 2005).

Uroplakin plaques on the surface of bladder epithelial cells may be said to be highly specialised TEMs with four major proteins and a defined function as a permeability barrier (Min et al., 2006). It is suggested that in other cell types, inclusion in (or indeed exclusion from) TEMs can regulate the function of a given protein by controlling its

Chapter 1 access to certain areas in the plasma membrane where receptors, signalling molecules and other proteins involved in a particular process are clustered. Examples of such sites are oocyte microvilli (sites of sperm-egg fusion), hemidesmosomes (sites of epithelial cell attachment to the basement membrane), immunological synapses (sites of antigen presentation) and viral entry or egress sites (discussed in section 1.2.1) (Kaji et al., 2000; Sterk et al., 2000; Wright et al., 2004; Singethan et al., 2008).

Tetraspanins were originally described as ‘molecular adaptors’, regulating the organisation of associated proteins through the formation of TEMs (Maecker et al., 1997). More recently, TEMs have been shown to be highly dynamic structures such that their composition and spatial organisation changes upon cellular stimulation, antibody cross-linking, ligand engagement or overexpression/deletion of TEM components (Barreiro et al., 2008; Singethen et al., 2008). This would highlight the contribution of non-tetraspanin proteins to TEM formation and suggest that the regulatory relationship between tetraspanin and non-tetraspanin components of TEMs is of a somewhat reciprocal and therefore more subtle nature.

The existence of microdomains containing multiple tetraspanins also provides an explanation for many experimental observations including apparent functional redundancy. For example, knockdown, overexpression or deletion of different superfamily members, antibody cross-linking of different tetraspanins and recombinant soluble EC2 domains belonging to several tetraspanins induce similar cellular responses such as inhibition of virus infection or bacterial entry into epithelial cells (Ho et al., 2006; Green et al., 2011; Weng et al., 2009; Gordon-Alonso et al., 2006). Thus targeting any of several tetraspanins within TEMs is thought to have global effects on TEM organisation which in turn would affect cellular functions of TEM- associated proteins under specific circumstances. It can also explain how tetraspanins can modify the activities of proteins with which they associate indirectly, through another tetraspanin and why identical co-immunoprecipitation profiles are often obtained using antibodies against different tetraspanins (Liu et al., 2007; Boucheix and Rubinstein, 2001).

Chapter 1

1.2 The biological process of fusion and the involvement of tetraspanins Fusion is a process by which two membranes merge together to form a continuous bilayer (Martens and McMahon, 2008). Membrane merger is involved in cell-cell fusion, intracellular fusion events (protein trafficking, organellar synthesis, exocytosis, enveloped virus release and budding) as well as extracellular fusion events (virus-host cell fusion, endocytosis) (Stein et al., 2004). These diverse processes are thought to proceed through a number of common steps shown in figure 1.3. The fusing membranes must first recognise and adhere to each other. The membranes are brought into close apposition by the action of fusogenic membrane proteins (fusogens). Fusogens are best characterised in vesicle trafficking and virus-induced fusion where they undergo conformational changes to bring the fusing bilayers into close proximity thus lowering the kinetic barrier to fusion (Harrison, 2008). This is followed by the formation of a fusion intermediate (hemi-fusion stalk) and finally a fusion pore which allows content mixing between the two membranous compartments (Sapir et al., 2008).

Tetraspanins have been implicated in various fusion events and as is characteristic of the superfamily, many have redundant and/or overlapping roles in fusion (Fanaei et al., 2011). The fact that tetraspanins have been found to enhance as well as inhibit fusion in different fusion systems suggests that they do not directly mediate the steps involved in fusion but regulate the function of fusion mediators such as adhesion molecules (e.g. integrins), signalling proteins, cytoskeletal proteins and fusogens (e.g. viral envelope proteins). An exception to this rule is the human tetraspanin RDS which has a fusion domain located in the C-terminus (section 1.2.2). Although monocyte- monocyte fusion is the main focus of this thesis, a review of other fusion events and the involvement of tetraspanins therein will help to elucidate their role.

Chapter 1

Figure 1.3 Biological membrane fusion. Diverse biological fusion events are thought to proceed through a number of common steps as outlined here. The actual membrane merger event is mediated by fusogen proteins which undergo conformational changes to bring opposing membranes into close proximity to allow lipid mixing. In some types of fusion such as viral fusion events, membrane recognition, adhesion and fusion activity reside within the same protein, the viral fusion protein. Based on figures in Chernomerdik, 2008 and Sapir et al., 2008.

Chapter 1

1.2.1 Virus-host cell fusion and virus-induced syncytium formation In enveloped viruses, the viral capsid (genetic material and associated proteins) is wrapped in one or more membranes derived from the host cell as newly assembled viral progeny bud through the plasma membrane of infected cells (Klasse et al., 1998). This envelope incorporates virus-encoded envelope (env) proteins which mediate attachment to and fusion with a new host cell to start another infection cycle (Pique et al., 2000). As well as virus-host cell fusion, env proteins can mediate cell-cell fusion leading to the formation of giant multinucleated cells called syncytia (Gordon-Alonso et al., 2006). Syncytium formation is a prominent mode of viral spread for some viruses (e.g. HTLV-1, see below) but for other viruses it may simply be a by-product of viral envelope protein fusogenic activity (Fais et al., 1997). Syncytium formation may shield viruses from adaptive immune responses and there is evidence that syncytia shed greater numbers of virions in vitro but it remains unclear whether syncytium formation is ultimately beneficial to virus spread in vivo (Gordon-Alonso et al., 2006; Weng et al., 2009). In any case, the existence of several mechanisms by which viral env protein fusogenicity is regulated suggests that controlled (as opposed to maximal) fusion is advantageous to enveloped virus spread (Rein et al., 1994; Pique et al., 2000).

Tetraspanins have been implicated in the life cycle of several enveloped viruses including human immunodeficiency virus type-1 (HIV-1), human T-cell leukaemia virus type-1 (HTLV-1), canine distemper virus (CDV) and hepatitis C virus (HCV) (Martin et al., 2005). The role of tetraspanins in viral-induced fusion events is perhaps best viewed in the context of TEMs (section 1.1.2.1) since multiple tetraspanins are involved and microscopy techniques have revealed the localisation of various viral proteins in tetraspanin-enriched areas as well as global rearrangements within TEMs upon viral protein expression (see below). Tetraspanins are also enriched in budding virions e.g. CD9 and CD81 in influenza and CD63 in HIV-1 virions (Shaw et al., 2008; Sato et al., 2008). As a general mechanism, tetraspanins probably regulate access of viral proteins or whole virions to functional areas of the plasma membrane where receptor/co-receptors and other machinery required for virus entry, assembly or release are clustered (Singethan et al., 2008).

The retrovirus HTVL-1 is largely reliant on cell-cell contact and syncytium formation to spread in vivo and in vitro (Pique et al., 2000). A number of antibodies against anti- CD82 antibodies inhibit HTLV-1-induced syncytium formation (Fukudome et al., 1992;

Chapter 1

Imai et al., 1992). Using mild immunoprecipitation conditions, CD82 was shown to associate with HTLV-1 env proteins. Co-expression of CD82 inhibited HTLV-1 (but not HIV-1) env-induced syncytium formation (Pique et al., 2000). Tetraspanin association with HTLV-1 env might depend on viral Gag protein (as it does in HIV-1) since Gag and CD82 colocalise in distinct membrane patches (Mazurov et al., 2006) but the precise role of CD82 in HTLV-1 infection remains unclear.

Unlike HTLV-1, syncytium formation is not required for HIV-1 spread in vitro and is typically a feature of late-stage AIDS (Acquired immunodeficiency syndrome) (Gordon-Alonso et al., 2006). However, HIV-1-induced syncytium formation is of clinical significance as it can deplete large populations of T-cells in vivo (Maas et al., 2000). Overexpression of CD9 and CD63 inhibits HIV-1-induced syncytium formation whilst silencing of CD9, CD63 and CD81 and an anti-CD81 antibody enhances fusion, suggesting that tetraspanins are negative regulators of this process (Weng et al., 2009; Gordon-Alonso et al., 2006). But the mechanism of this inhibition remains to be established. The tetraspanin CD81 is known to be constitutively associated with the primary HIV-1 receptor CD4 and is recruited to areas of cell-cell contact where HIV-1- mediated fusion takes place (Gordon-Alonso et al., 2006). CD9, CD63 and CD81 also co-localise with viral Gag protein which targets newly synthesised env to virus assembly sites (Nydegger et al., 2006; Weng et al., 2009). Multimerisation of Gag proteins at virus assembly sites resulted in local confinement of CD9 (krementsov et al., 2010). Hence, the active recruitment of tetraspanins to viral assembly sites by viral Gag may be a mechanism to prevent excessive env-mediated fusion.

Overexpression of CD9 in two cell lines enhanced CDV-induced syncytium formation whilst an anti-CD9 antibody (K41) inhibited this process (Loffler et al., 1997; Schmid et al., 2000). This inhibition was dependent upon the extracellular domain of CDV haemaglutinin (H) which facilitates attachment to host cells and regulates the fusogenic activity of the viral fusion protein (F) (Singethan et al., 2006). Interestingly, other experiments have suggested that modulating the affinity of the H protein for its receptor is a means of regulating CDV F protein-induced fusion (Plattet et al., 2005). K41 also inhibits CDV budding from infected cells and causes clustering of CD9 at cell- cell contact areas from which CDV proteins were specifically excluded (Singethan et al., 2008).

Chapter 1

The entry of HCV into host cells appears to be a complex process involving a number of cell surface molecules including tetraspanin CD81 which binds directly to viral envelope glycoprotein E2 (Pileri et al., 1998; Higginbottom et al., 2000). Unlike other viruses discussed here, the viral envelope of HCV does not fuse with the plasma membrane nor does it induce syncytium formation. Instead, following attachment, the intact virion is internalised into host cells by endocytosis. The viral envelope then fuses with intracellular endosomal membranes to release viral genetic material into the cytosol (Zeisel et al., 2011). The role of CD81 in this intracellular fusion event is uncertain, however mutation of the CD81 binding site in the E2 protein prevented E2- induced fusion with endosomal membranes (Kobayashi et al., 2006).

1.2.2 Outer segment formation in photoreceptor cells: an intracellular fusion event In the vertebrate retina, light transduction takes place in specialised membranes in the outer segments (OS) of photoreceptor (rod and cone) cells. OS are made up of stacks of flattened membranous sacs (disks or lamellae) containing the light-sensitive pigment rhodopsin (Boesze-Battaglia et al., 2003). To counter the damaging effects of light transduction, OS are continually renewed by two fusion events; at one end of the OS, nascent disks resulting from membrane invaginations fuse to create closed or partially closed sacs whilst at the other end, old disks are shed by fusion with the plasma membrane (Boesze-Battaglia et al., 1998; Tam et al., 2004).

The closely related tetraspanins RDS and ROM-1 are crucial to OS membrane integrity and renewal. These tetraspanins are exclusively expressed in photoreceptor cells where they localise to the disk rims (sites of disk closure and shedding) (Boesze- Battaglia et al., 1998; Loewen et al., 2000). RDS-/- mice lack outer segments whilst RDS+/- mice develop highly disorganised whorls of membrane suggesting that RDS is essential for organised disk formation (Loewen and Molday, 2000). In humans, more than 90 different mutations have been described in the RDS gene, resulting in a broad spectrum of retinal degenerative diseases (Boon et al., 2008). One specific RDS mutation (L185P) causes retinitis pigmentosa exclusively in patients who co-carry a ROM-1 gene mutation (Kajiwara et al., 1994). ROM-1-/- mice show only mild abnormalities in OS formation and enhanced photoreceptor cell death (Clarke et al., 2000).

Chapter 1

Currently, the mechanism of RDS and ROM-1 function is unknown. As mentioned in section 1.1.1.3, the C-terminal domain of RDS incorporates a short peptide capable of inducing membrane fusion in vitro as well binding sites for a number of proteins which regulate the fusogenicity of this peptide (Edrington et al., 2007a/b; Boesze- Battaglia et al., 2007). Thus RDS could potentially facilitate OS membrane fusion events. Co-expression of RDS and ROM-1 protein in a cell-free system enhances the fusogenicity of RDS (Boesze-Battaglia et al., 2007). RDS and ROM-1 associate laterally into homo- and hetero-tetrameric complexes. Uniquely amongst human tetraspanins, RDS and ROM-1 possess an unpaired cysteine residue in the EC2 domain which mediates intermolecular disulfide bridge formation between tetrameric complexes (Loewen and Molday, 2000). Such disulfide-linked complexes may contribute to disk flattening (by bridging the gap between opposing membranes in a disk) and/or stabilise the highly curved disk rims (Goldberg et al., 2001). Meanwhile, ROM-1 is likely to have an accessory role, potentially modulating RDS fusogenicity and/or multimerisation (Boesze-Battaglia et al., 2007; Loewen and Molday, 2000).

1.2.3 Cell-cell fusion Cell-cell fusion is fundamental to the development of most eukaryotic organisms from yeast, plants and fungi to humans (Sapir et al., 2008). In contrast to other fusion events, cell-cell fusion remains poorly understood (Chen et al., 2007). This is perhaps not surprising as cell-cell fusion is much more than just the merger of two biological membranes. Helming and Gordon (2009) have identified a number of steps likely to be involved in monocyte fusion but these can be generalised to other types of cell-cell fusion such as muscle cell fusion (Chen and Olson, 2005). Firstly, cells must reach a fusion-competent status, in response to developmental or environmental signals. Fusion-competent cells then have to move towards each other, a process that requires cytoskeletal rearrangements and modulation of cell adhesion to the substratum. Then, in common with other types of fusion, the fusing membranes must recognise each other and attach, bringing the respective bilayers into close contact for fusion to take place (Helming and Gordon, 2009). Interference with any of these steps (and not just membrane fusion itself) could potentially enhance or inhibit cell-cell fusion.

The characterisation of other fusion systems has also benefited from the identification of fusogens which directly mediate membrane fusion (e.g. viral fusogens). Only in

Chapter 1 trophoblast cell-cell fusion has a fusogen been identified, as discussed in the following sections.

1.2.3.1 Sperm-egg fusion In sexually-reproducing animals, fusion of haploid gametes (sperm and oocyte) results in the formation of a diploid zygote which is the first cell of a new organism (Sapir et al., 2008; Wilson and Snell, 1998). Tetraspanin CD9 is essential for murine sperm-egg fusion as CD9 knockout mice have severely reduced fertility in females but not males (Le Naour et al., 2000; Miyado et al., 2000; Kaji et al., 2000). A number of anti-CD9 antibodies inhibited sperm-egg fusion (Chen et al., 1999; Le Naour et al., 2000; Miyado et al., 2000; Miller et al., 2000) whilst injection of mouse or human CD9 (90% homology to mouse CD9) mRNA into CD9-/- mouse oocytes restored fusion competence (Zhu et al., 2002; Kaji et al., 2002). Soluble CD9 EC2 proteins also inhibited murine sperm-egg fusion but only when pre-incubated with eggs not sperm, suggesting that CD9 regulates the function of a laterally-associated molecule on the egg surface (Zhu et al., 2002).

During fertilisation, oocyte microvilli undergo rearrangements to capture the sperm head and treatment with latrunculin B (a small molecule inhibitor which alters microvilli morphology) inhibits sperm-egg fusion (Kaji et al., 2000; Runge et al., 2007). Interestingly, microvilli have been observed in other fusion processes such as viral, myoblast, mononuclear phagocyte and syncytiotrophoblast fusion (Wilson and Snell, 1998). Microvilli have a small radius of curvature which is known to enhance fusogenicity of lipid bilayers (Martens and McMahon, 2008). CD9 molecules become densely concentrated on oocyte microvilli during fertilisation and lack of CD9 results in short and thick microvilli with a larger radius of curvature suggesting that CD9 promotes or stabilises normal microvilli morphology (Kaji et al., 2000; Runge et al., 2007). Direct association with the EWI-2 and EWI-F proteins may serve to link CD9 and (also CD81) to the actin core of oocyte microvilli, as has been shown in other cell types (Runge et al., 2007; Sala-Valdez et al., 2006).

Other evidence suggests that CD9 regulates sperm-egg adhesion. In murine CD9-/- eggs, more sperm adhere to the oocyte but the strongest mode of adhesion is lost. This type of adhesion is thought to maintain close apposition of sperm-egg membranes to

Chapter 1 allow fusion events to take place (Jegou et al., 2011). In other cell types, adhesion molecules are concentrated on microvilli (Wilson and Snell, 1998). Thus altered microvilli morphology and/or altered distribution of certain CD9-associated adhesion molecules (e.g. integrins) in CD9-/- oocytes may contribute to this phenomenon. In conclusion, the exact function of CD9 in sperm-egg fusion is not clear and may well involve several different mechanisms (Rubinstein et al., 2006a).

There is also some evidence for involvement of murine CD81 in sperm-egg fusion. Endogenous oocyte levels of CD81 are lower than CD9 and correspondingly CD81-/- female mice only have a 40% reduction in fertility (Kaji et al., 2002; Rubinstein et al., 2006b). Monoclonal anti-CD81 antibodies have a modest effect on fusion as does a soluble mCD81 EC2 construct (Rubinstein et al., 2006b). Mouse CD81 has ~50% homology to mouse CD9. CD81 may partly compensate for the absence of CD9 since CD9-/-/CD81-/- double knockout female mice (but not single knockouts) are completely infertile and overexpression of CD81 partially rescues fusion competence of CD9-/- oocytes (Rubinstein et al., 2006b; Kaji et al., 2002). Antibodies against CD9, CD151 and the primary CD151 partner α β1 integrin inhibit human sperm-egg fusion and all three proteins co-localise in patches on the oocyte surface (Ziyyat et al., 2006). CD9 is thought to control formation of these patches but their importance in the fusion process remains to be determined.

1.2.3.2 Developmental fusion events In higher organisms, fusion of somatic cells is an integral part of embryonic development. In mammals, fusion of trophoblasts leads to the formation of the syncytiotrophoblast layer of the placenta which acts as an exchange barrier between maternal and foetal blood (Larsson et al., 2008). Intriguingly, trophoblast fusion in primates and rodents appears to be mediated by syncytins, endogenous retroviral envelope proteins related to class I viral fusogens such as HIV-1 env (Chen and Olson, 2005; Oren-Suissa and Podbilewicz, 2007).

Embryonic muscle development in vertebrates occurs in two distinct phases; initially, cells of the myoblast lineage fuse to form small multinucleated myotubes, the building blocks of skeletal muscle. Further myoblast-myotube or myotube-myotube fusion generates large mature myotubes that may contain hundreds or thousands of nuclei

Chapter 1

(Chen and Olson, 2005). In post-natal development and adult life, muscle stem cells (satellite cells) can differentiate and fuse with existing myotubes to promotes further muscle growth, repair and regeneration (Abmayr and Pavlath, 2012).

CD9 and CD81 are expressed at high levels on skeletal muscle. In a murine myoblast cell line, CD9 expression and complex formation with β1 integrin was up-regulated as cells became confluent and began to differentiate into multinucleated myotubes. Anti- CD9 and anti-CD81 antibodies delayed myotube formation and had an additive inhibitory effect on human and murine myoblast cell lines suggesting some functional redundancy between the two tetraspanins. CD9 overexpression in the human myoblast cell line promoted formation of giant cells more reminiscent of virus- induced syncytia (Tachibana and Hemler, 1999). The association of CD9 and β1- integrin on the surface of myoblasts may have a functional significance. β1-deficient murine myoblasts show defective fusion in vivo and in vitro and importantly, cell- surface expression of CD9 is also abolished in these cells (Schwander et al., 2003).

Chapter 1

1.2.3.3 Mononuclear phagocyte fusion and functions Mononuclear phagocytes (monocytes) are hematopoietic cells of the myeloid lineage, originating in the bone marrow. Circulating monocytes migrate into tissues where they undergo differentiation in response to environmental cues (Mosser and Edwards, 2008). Monocyte differentiation results in the formation of mature tissue macrophages, inflammatory dendritic cells and three different types of multinucleated giant cells called Langhans giant cells (LGC), foreign body giant cells (FBGC) and osteoclasts (figure 1.4). Resident tissue macrophages have pleiotropic functions including maintaining tissue homeostasis by clearance of old, damaged or infected cells and cellular debris, phagocytosis of pathogens, promotion of wound healing and initiation of host adaptive immune responses by antigen presentation (Kumar and Jack, 2006; Mosser and Edwards, 2008). Dendritic cells are specialised for antigen capture and presentation to T lymphocytes. In addition, they express and secrete various molecules and cytokines and are thus central to T cell development (Blanco et al., 2008). Monocyte-derived dendritic cells are only a small subset of the total dendritic cell population which is very heterogenous in origin, morphology, tissue localisation and function (Rivollier et al., 2004; Blanco et al., 2008). Monocyte-derived dendritic cells can be trans-differentiated into multinucleated osteoclasts by appropriate stimuli, as shown in figure 1.4 (Coury et al., 2008).

The three different types of giant cells mentioned above are morphologically and functionally distinct (Anderson et al., 2000). However, they usually differentiate in the presence of large or poorly digestible materials that cannot be destroyed by conventional mechanisms of uptake and intracellular digestion in resident tissue macrophages (Jay et al., 2007). In the case of LGCs, the inciting agent may be a pathogenic organism (e.g. M. tuberculosis, Leishmania and Schistosoma species) or unknown factors (such as in sarcoidosis and Crohn’s disease) (Anderson, 2000; Gasser and Most, 1999; Helming and Gordon, 2009, do Nascimento et al., 2008; Okamoto et al., 2003; Fais et al., 1997). FBGCs on the other hand form as part of the immune response to non-infectious foreign materials such as asbestos, silica, sutures, medical implants and transplanted organs (McNally and Anderson, 1995; Anderson, 2000; Cui et al., 2006). In vivo, LGC and FBGC formation takes place in the context of granuloma development where various immune cells aggregate around the inciting agent and are enclosed in a fibrous capsule in later stages (Lay et al., 2007; Anderson, 2000).

Chapter 1

Figure 1.4 Monocyte differentiation/activation pathways.

Broadly speaking, multipotent haematopoietic stem cells give rise to two distinct progenitor cells, the common lymphoid progenitor and the common myeloid progenitor from which mononuclear phagocytes (monocytes) are descended. Monocytes themselves have the potential to differentiate into several types of mono- or multinucleated cells when stimulated with particular combinations of cytokines and TLR agonists (see text for detailed explanation). Information available in the articles cited in the main text was collated to construct this scheme. References not cited in the main text: Martinez et al., 2006; Stirewalt and Radich, 2003.

Chapter 1

In osteoclastogenesis, circulating monocytes settle directly on bone and differentiate into mononuclear pre-osteoclasts which then fuse to become mature multinucleated osteoclasts containing on average 5 to 8 nuclei in humans (Vaananen and Laitalia- Leinonen, 2008; Ha et al., 2011). Mammalian bone is composed of organic (mainly collagen) and inorganic matter (crystalline hydroxyapatite, Ca5(PO4)3(OH), deposited by mononuclear cells called osteoblasts (Boyle et al., 2003). Osteoclast activity is essential to resorb (digest) old bone damaged by stress-induced micro-fractures and to regulate calcium and phosphate homeostasis in response to environmental changes (Vaananen and Laitalia-Leinonen, 2008). Defects in osteoclast formation lead to osteopetrosis where bones are thick and brittle whilst increased osteoclast formation/activity is observed in osteoporosis and inflammatory bone diseases such as rheumatoid arthritis and periodontal disease (Boyle et al., 2003; Helming and Gordon, 2009). In addition, in certain types of cancer such as multiple myeloma and metastatic prostate cancer, pathological osteoclast formation and activity can lead to tissue destruction (Boyle et al., 2003; Miyamoto et al., 2012). Thus elucidation of the mechanisms and molecules involved in osteoclast formation and activity are of considerable medical importance.

The cytokines macrophage colony stimulating factor (MCSF) and receptor activator of NF-κB ligand (R NKL) in particular are crucial to osteoclast formation. MCSF promotes osteoclast precursor survival and proliferation and upregulates expression of RANK, the cell surface receptor for RANKL on pre-osteoclasts (Asagiri and Takayanagi, 2007). RANKL is expressed (mainly as a membrane-bound form) by osteoblasts and also stromal cells and T lymphocytes present on or near bone surfaces (Maruoti et al., 2011). RANKL binding to RANK triggers multiple signal transduction pathways which converge on activation of the transcription factor NFATc1 (Nuclear factor of activated T cells 1) (Takayanagi, 2005). NFATc1 induces expression of a range of molecules involved in osteoclast differentiation and function such as fusion factors and several digestive enzymes (Boyle et al., 2003). Multinucleated osteoclasts undergo cytoskeletal rearrangements resulting in the formation of a tightly sealed compartment, the resorption pit, underneath the osteoclast where hydrogen ions and digestive enzymes can be safely released into the extracellular environment to attack bone (Teitelbaum, 2006). The products of bone digestion (collagen fragments and

Chapter 1 minerals) are taken up by osteoclasts and further processed or released back into blood (Boyle et al., 2003).

The transmembrane proteins DC-STAMP (Dendritic cell-specific transmembrane protein) and OC-STAMP (Osteoclast stimulatory transmembrane protein) are essential for osteoclast fusion (Yagi et al., 2005; Miyamoto et al., 2012). In both DC-STAMP-/- and OC-STAMP-/- mice, mononuclear osteoclasts are present and express the usual complement of osteoclast effector molecules but they are much less efficient at resorption. Thus it appears that multinucleation is necessary for efficient osteoclast function. Multinucleation increases cell size and therefore the surface area underneath which safe and efficient extracellular digestion of bone can take place (Ishii and Saeki, 2008). Very large osteoclasts also possess lower resorptive capacity per nucleus than smaller multinucleated osteoclasts highlighting the importance of osteoclast size regulation (Bar-Shavit, 2007). Mature osteoclasts also lose their phagocytic capacity (Takami et al., 2002).

Similar to osteoclasts, FBGCs appear to be efficient at extracellular degradation, creating cracks in medical implants and biopolymers which shortens the lifetime of the implant in vivo (McNally et al., 2002). Recent research has focused on modifying the surface chemistry of implants to make them more resistant to FBGC differentiation (Anderson, 2008). Interestingly, FBGC formation is also abrogated in both DC-STAMP-/- and OC-STAMP-/- mice suggesting that these molecules are universal regulators of monocyte fusion (Yagi et al., 2005; Miyamoto et al., 2012).

FBGCs are distinguishable from LGCs by virtue of their larger size and random nuclei distribution in contrast to the circular peripheral arrangement of small numbers of nuclei (~15) in LGCs, as shown in figure 1.4. LGCs were first described in the tubercule granuloma more than a hundred years ago (Anderson, 2000). In tuberculosis, mycobacterium-infected macrophages release a variety of cytokines and chemokines to recruit other immune cells (e.g. circulating monocytes, T lymphocytes and dendritic cells) to the infection site. The presence of mycobacteria, mycobacterial products (such as lipomannans) and the cocktail of cytokines produced by all these cells induce LGC and granuloma formation (Gasser and Most, 1999; Puissegur et al., 2007).

Chapter 1

Granulomas are thought to sequester pathogens and inflammatory cells and mediators thereby limiting the spread of infection and inflammation to healthy tissues (Byrd, 1998; Lay et al., 2007; Co et al., 2004). However, the function of LGCs within the granuloma is not well understood and data is currently lacking regarding the intracellular and extracellular digestive capacity of LGCs. In an in vitro model of tuberculosis infection, M. tuberculosis localised to the centre of LGCs and this correlated with restricted bacterial growth and cell-cell spread (Byrd et al., 1998). However, in another infectious granulomatous disease, melioidosis (caused by Burkholderia bacterial strains), monocyte-derived giant cell formation appears to facilitate cell-cell spread (Kespichayawattana et al., 2000).

1.2.3.3.1 Monocyte activation pathways The terms monocyte/macrophage differentiation, activation and polarisation are used somewhat interchangeably to refer to the process whereby the monocyte precursor is stimulated to develop into a mature functionally active cell (Gordon and Martinez, 2010). Monocyte activation as a whole follows one of two routes (figure 1.4); the classical (M1) pathway and the alternative (M2) pathway (Helming and Gordon, 2008). The M1 pathway is principally mediated by the cytokine interferon-γ (IFN-γ, major product of T helper cells type 1 or TH1) and microbial products which bind to monocyte toll-like receptors (TLRs) whilst the cytokines interleukin-4 and -13 (IL-4 and IL-13, products of T helper cells type 2 or TH2) mediate the M2 pathway. Classically activated macrophages efficiently destroy intracellular pathogens by production of reactive oxygen species and other mediators of inflammation whilst M2 activation is associated with attenuation of acute inflammatory responses and wound healing as well as less efficient pathogen destruction (Mosser, 2003; Gordon and Martinez, 2010).

With regard to giant cells, IFN-γ potently enhances LGC formation and inhibits FBGC formation whilst IL-4 has the opposite effect (McNally and Anderson, 1995; Takashima et al., 1993; Moreno et al., 2003; Byrd, 1998). IL-13 also induces FBGC formation, probably via the same signalling pathway as IL-4 since they both bind to the IL-4 receptor α-chain and do not have additive or synergistic effects when utilised together (DeFife et al., 1997; Herbert et al., 2004). Thus it may be argued that LGCs and FBGCs are the results of monocyte differentiation along two distinct pathways. On the other

Chapter 1 hand, in an in vivo model of FBGC formation in rats, the early period after implantation of foreign material was characterised by presence of small LGC-type giant cells whilst in later periods medium sized mixed-type giant cells were observed and finally large FBGC became the dominant cell type present (van der Rhee et al., 1978). This led the authors to conclude that LGCs may be the precursors of FBGCs. The transition between the two types of giant cells may reflect the changing profile of cytokines present at various stages of granuloma development. Both cell types are sometimes observed together in certain granulomatous lesions such as cutaneous sarcoidosis lesions (Okamoto et al., 2003).

IL-4 and IL-13 are also potent inhibitors of RANKL-induced osteoclast formation in vitro and in vivo, down-regulating expression of NFATc1, the master transcription factor of osteoclastogenesis (Yamada et al., 2007; Cheng et al., 2011). Both cytokines also stimulate production of osteoprotegerin in osteoblasts, a decoy receptor which inhibits RANK/RANKL interaction (Palmqvist et al., 2006; Yamada et al., 2007). In vitro, osteoclast precursors are only responsive to IL-4 during the first 48 hours of RANKL-induced differentiation suggesting that IL-4 regulates the lineage commitment of precursors (Moreno et al., 2003). On the other hand, the principal mediators of the M1 pathway, IFN-γ and a number of microbial products (TLR agonists), also inhibit the early stages of osteoclast differentiation (Gao et al., 2007, Hotokezaka et al., 2007). Thus the development of the pre-osteoclast appears to fall outside both pathways of macrophage activation. This is not unexpected since, under normal physiological conditions, the osteoclast has an essentially non-immune function i.e. bone resorption.

However, under inflammatory conditions, pre-osteoclasts can differentiate into multinucleated osteoclasts, as part of the M1 pathway (figure 1.4). TLR agonists and other potent inflammatory mediators such as TNF-α (Tumour necrosis factor-alpha) and IL-1 are known to exacerbate bone destruction in inflammatory diseases such as rheumatoid arthritis and periodontal disease, as will be discussed in section 1.3.2 (Hotokezaka et al., 2007; Ha et al., 2011).

The research discussed above highlights the plasticity of the monocyte precursor and its acute sensitivity to environmental signals (Mosser and Edwards, 2008). A further example of this plasticity is observed in the granulomatous disease Langerhans cell

Chapter 1 histiocytosis in which the T cell cytokine IL-17A mediates trans-differentiation of monocyte-derived inflammatory dendritic cells into osteoclast-like giant cells capable of resorbing bone (Coury et al., 2008).

1.2.3.3.2 Mechanisms of monocyte fusion ccording to the ‘cellocytosis’ model proposed by Vignery (2005), monocyte fusion is reminiscent of phagocytosis, the mechanism by which macrophages engulf and internalise apoptotic cells or foreign materials. Thus when two monocytes encounter one another, one of the pair takes the lead and internalises the other monocyte which is now enclosed in two plasma membranes (figure 1.5a). These membranes then fuse with the cell membrane of the engulfing monocyte and the two cells become integrated. The delay between internalisation and membrane disappearance allowed intact internalised monocytes to be visualised inside another monocyte by electron microscopy (Vignery et al., 1989).

Similarly, McNally and Anderson (2005) propose that monocyte fusion resembles macrophage phagocytosis of complement-opsonised particles with the involvement of the endoplasmic reticulum (ER). Markers of ER-mediated phagocytosis were noted to localise at fusion interfaces and pharmacological inhibition of ER components such as calcium-independent phospholipase A2 attenuated IL-4-induced FBGC formation as did inhibition of mannose receptor, a classic phagocytic receptor that is up-regulated during FBGC formation (McNally and Anderson, 2005; McNally et al., 1996).

However, fluorescence microscopy studies appear to dispute a phagocytic mechanism for monocyte fusion. Takito et al. (2012) transfected the murine monocyte cell line RAW264.7 with EGFP-actin (enhanced green fluorescent protein-actin) and tracked fusion of EGFP-actin-negative and EGFP-actin-positive cells into osteoclasts by time- lapse confocal microscopy, as represented in figure 1.5b (lower panel). The fluorescence signal disappears from a small area in the plasma membrane of the EGFP- actin-positive cell (assumed to be the initial site of fusion with the EGFP-actin-negative cell) and as fusion proceeds, this fluorescence-negative area expands suggestive of an expanding fusion pore (see figure 1.3). Finally, the fluorescence signal completely disappears from the contact region between the two cells. Phase contrast images (figure 1.5b, upper panel) confirm that the two cells have now merged. In contrast to

Chapter 1 the phagocytic model, these images suggest that membrane mixing (at the initial fusion site) is immediately followed by content mixing.

Time

Figure 1.5 Mechanisms of monocyte fusion.

a) The cellocytosis model proposed by Vignery (2005). A leader monocyte (blue) internalises other monocytes (red) that come into contact with it, reminiscent of macrophage phagocytosis of whole cells or particles.

b) Time-lapse study of osteoclast fusion observed by confocal microscopy (Takito et al., 2012). Top panel represents phase contrast time-lapse images. Two multinucleated osteoclasts come into close contact. The plasma membrane disappears from a small area of contact (yellow arrow) which gradually becomes larger. White dots represent nuclei. The bottom panel

represents fluorescence images obtained from the same field. The cell on the left is transfected with EGFP-actin (green) whilst the other cell is EGFP-actin-negative. A gap appears in the plasma membrane of the EGFP-actin-positive cell (yellow arrow) and gradually expands.

Chapter 1

One aspect of the cellocytosis model which is supported by other evidence is that monocyte fusion is heterophilic, i.e. the fusing partners are not identical and have distinct properties. The transmembrane protein DC-STAMP is essential for osteoclast fusion as monocytes from DC-STAMP-/- mice are unable to form multinucleated osteoclasts. However, fusion can still occur between wild-type and DC-STAMP-/- monocytes suggesting that monocyte fusion is of a heterophilic nature (Yagi et al., 2005). In another time-lapse study of osteoclast fusion, Ishii and Saeki (2008) observed that highly motile multinucleated ‘founder’ cells seek and initiate fusion with mononucleated ‘follower’ cells, reminiscent of drosophila myoblast fusion (Chen et al., 2007). Founder and follower monocytes may be distinguished by the expression levels of molecules involved in fusion such as DC-STAMP (Mensah et al., 2010; Miyamoto et al., 2012).

1.2.3.3.3 Studying monocyte fusion in vitro Numerous methods are employed in the literature to generate and study monocyte- derived giant cells in vitro (Anderson, 2000; Okamoto et al., 2003). A summary of the methods used by some of the studies mentioned in previous sections is outlined in Table 1.1. There are significant differences in how the monocyte precursor is obtained, the stimuli used to induce giant cell formation and the culture time employed. A disadvantage of most of these methods is that expensive cytokines such as RANKL and MCSF are often used and the assay is performed over several days.

An established method of generating multinucleated giant cells involves the incubation of freshly isolated human monocytes with the Jack bean lectin Concanavalin A (Con A) (Table 1.1). Lectins are plant-derived proteins which specifically recognise the carbohydrate moieties of glycosylated membrane proteins (Smith and Revel, 1972). Con A is reported to activate T lymphocytes and cells of the monocyte/macrophage lineage, resulting in pro-inflammatory cytokine production in vivo and in vitro (Gantner et al., 1995). It induces rapid aggregation of adherent monocytes and fusion is observed approximately 2 hours after Con A addition (Hulme, R., PhD thesis). After 72 hours incubation, 70-80% of monocytes have fused to form multinucleated giant cells (MGCs) with the characteristic features of Langhan’s giant cells (LGCs) (Parthasarathy et al., 2009). This system offers several advantages including simplicity, reproducibility and the absence of a monocyte proliferation step

Chapter 1

(Table 1.1). There is evidence that maintaining adherent monocytes in culture results in maturation towards a macrophage phenotype and a progressive loss of fusion capacity (Most et al., 1997). Thus using freshly isolated monocytes avoids possible cellular changes which may adversely affect fusion. The time course for Con A-induced fusion is also short, in line with LGC formation in in vivo experimental models of granuloma formation (Most et al., 1997).

The effect of Con A on MGC formation is dose-dependent but the mechanisms for its action are not clear (Takashima et al., 1993). Con A treatment results in up-regulation of several cytokines including TNF-α, IL-1b and IL-6 (Takeda et al., 2003). The supernatant of Con A-stimulated monocyte cultures also induces MGC formation supporting the involvement of soluble factors (Most et al., 1990). Takashima et al. (1993) noted a positive correlation between the levels of the monocyte-derived cytokine TNF-α in the culture medium and the monocyte fusion rate and an anti-TNF- α antibody inhibited Con -induced MGC formation. However, purified TNF-α by itself was unable to induce MGC formation and other stimulants such as bacterial lipopolysaccharide (LPS) which also activates monocytes to produce TNF-α, cannot induce MGC formation (Takashima et al., 1993; Abe et al., 1984; Enelow et al., 1992; Lazarus et al., 1990). Thus other factors must be involved. The T-cell derived cytokine IFN-γ caused an enhancement of Con A-induced MGC fusion whilst an anti-IFN-γ antibody had an inhibitory effect supporting the involvement of contaminating T lymphocytes on MGC formation (Takashima et al., 1993; Most et al., 1990; Okamoto et al., 2003). An anti-TNF-α antibody completely supressed the stimulatory effect of IFN- γ suggesting that TNF-α operates downstream of IFN-γ (Takashima et al., 1993). IFN-γ is also implicated in in vivo LGC formation (Belosevic et al., 1989; Chensue et al., 1992).

Chapter 1

In vitro assay system Comments References

LGC formation Adherent human PBMC + Closer to physiological conditions (?) Weinberg et al., stimulated with IFN-γ alone or + Uses fresh human cells 1984; Byrd et al., in combination with TNF-α and 1998; do α-calcitriol for 7-14 days − Time-consuming Nascimento et al., − Lack of reproducibility reported with 2008 some cytokines (Takashima et al., 1993; Most et al., 2005) MGC formation Adherent human PBMC + Quick Takashima et al., stimulated with Concanavalin A + Cheap 1993; (Con A) for 3 days. Giant cells + Highly reproducible Takeda et al., 2003; mostly resemble LGCs. + Uses fresh human monocytes Parthasarathy et al., 2009; Falzoni et al., − Uses a non-physiological stimulus 1995 Human (U937) or murine + Physiological stimulus (Burkholderia- Kespichayawattana (RAW264.7, J774.2) induced fusion observed in vivo) et al., 2000; Boddey macrophage-like cell lines et al., 2007; infected with Burkholderia − Immortalised cell line, may differ Suparak et al., 2011 species (B. pseudomallei, B. significantly from primary cells thailandensis) for 6-18 hours. − Effects of test reagents on bacterial growth, uptake and toxicity must be excluded FBGC formation Adherent PBMC cultured in + Uses primary human monocytes McNally et al., 1995; serum for 3 days. Cells then DeFife et al., 1997; stimulated with IL-4 and/or IL- − Expensive cytokines McNally and 13 for a further 4-7 days. − Time-consuming Anderson, 2002 Murine bone marrow or spleen + High yield of cells obtained Yagi et al., 2005; mononuclear cells incubated Yagi et al., 2007; with MCSF (and TGF-β) for 2-3 − Time-consuming Yamada et al., 2007; days to obtain MCSF-dependent − Requires animals and expensive macrophages (MDMs). Cells cytokines then incubated with MCSF or − Results may not be applicable to GMCSF in combination with IL- humans 4, IL-13 or both for 4-6 days. Osteoclastogenesis Adherent PBMC incubated with + Uses primary human cells Hiasa et al., 2009; recombinant RANKL/MCSF for Moreno et al., 2003 5-14 days, with frequent media − Time-consuming changes. Cultures may be − Difficult to separate effects on supplemented with Vitamin D proliferation and fusion and/or the synthetic steroid − Expensive cytokines dexamethasone. MDMs (see above) incubated − Results may not be applicable to Cheng et al., 2011; with RANKL and MSCF for 3-5 humans Ha et al., 2011; Lee days − Requires animals and expensive et al., 2008; Yamada cytokines et al., 2007; Yi et al., 2006

Chapter 1

Osteoblasts are isolated from + Closer to physiological conditions Ha et al., 2008; Liu murine calvarias. MDMs (see et al., 2009; Yamada above) are co-cultured with − Difficult to determine effects on et al., 2007 osteoblasts for 6-7 days. individual components Cultures may be supplemented − Requires animals with Vitamin D and/or the − Results may not be applicable to synthetic steroid humans dexamethasone. Murine RAW267.4 cells + Quick and convenient Hotozeka et al., incubated with RANKL for 3-5 + MCSF not needed as cells naturally 2007; Iwai et al., days to induce divide in culture 2007; Ishii et al., osteoclastogenesis 2006 − Immortalised cell line, may differ significantly from primary cells − Difficult to distinguish effects on proliferation and fusion − Results may not be applicable to humans

Table 1.1 Generation of monocyte-derived multinucleated giant cells in vitro.

Abbreviations; GM-CSF, Granulocyte-macrophage colony stimulating factor, IL, Interleukin, MCSF, Macrophage colony stimulating factor, MDM, MCSF-dependent macrophages, PBMC, Peripheral blood mononuclear cells, RANKL, Receptor activator of nuclear factor kappa-B ligand, TGF-β, Transforming growth factor-beta.

1.2.3.3.4 Regulation of monocyte fusion by tetraspanins In addition to the principal mediators discussed in section 1.2.3.3.1, numerous other molecules have been implicated in monocyte-derived giant cell formation including various cell-surface receptors/ligands (e.g. integrins, immunoglobulins), transcription factors, signal transducers (e.g. Diacylglycerol kinase), cytokines, enzymes (e.g. matrix metalloproteinases), microbial products (e.g. lipopysaccharide, peptidoglycan, flagellin), Vitamins D and E and hormones (e.g. parathyroid hormone) and finally several tetraspanins (Anderson et al., 2008; Anderson, 2000; Takami et al., 2002; Ha et al., 2008; Takayanagi, 2005; McNally and Anderson, 2005; Helming and Gordon, 2009; Parthasarathy et al., 2009). Given the involvement of giant cells in numerous disease states, molecules which regulate giant cell formation are important therapeutic targets (Helming and Gordon, 2009).

Tetraspanins are of particular interest in this regard since multiple members are expressed on monocytes and they participate in several other fusion events as fusion regulators. Expression of at least 18 tetraspanins has been reported on primary

Chapter 1 human monocytes (Wright et al., 2004) whilst the mouse monocyte cell line RAW264.7 expresses at least 12 tetraspanins (Iwai et al., 2007).

During LGC and osteoclast formation (induced by Con A and RANKL, respectively), CD9 and CD81 proteins levels were found to increase in human and murine monocytes and monocyte cell lines (Takeda et al., 2003; Ishii et al., 2006; Iwai et al., 2007; Yi et al., 2006). CD9 localised to areas of cell-cell contact before fusion but there was diminished cell surface expression of CD9 in multinucleated LGCs and osteoclasts (Takeda et al., 2003; Ishii et al., 2006). A number of anti-CD9 antibodies enhanced Con A-induced LGC formation and one anti-CD9 antibody also enhanced RANKL-induced osteoclast formation (Takeda et al., 2003; Parthasarathy et al., 2009). In addition, monocytes from CD9-/- mice show enhanced LGC and osteoclast formation in vitro and in vivo (Takeda et al., 2003). These results suggest that CD9 is an active down- regulator of monocyte fusion although the mechanisms of this inhibition are not clear.

Anti-CD81 antibodies also enhance monocyte fusion and produced an additive effect when combined with anti-CD9 treatment (Takeda et al., Parthasarathy et al., 2009). Similar to the CD9-/- knockout, monocytes from CD81-/- mice show an enhanced capacity to fuse in response to stimuli. In CD9/CD81 double knockout mutants, spontaneous LGC formation is observed in lung tissue and there is enhanced osteoclastogenesis resulting in reduced bone density. These phenotypes are not observed in single knockout mice suggesting that CD9 and CD81 have redundant functions in monocyte fusion, similar to sperm-egg and myoblast fusion (Takeda et al., 2003).

There is also evidence for CD63 involvement in LGC formation. CD63 protein levels increased upon Con A stimulation of primary human monocytes (Takeda et al., 2003) and a number of anti-CD63 antibodies inhibited LGC formation (Parthasarathy et al., 2009). In contrast to anti-CD9 and -CD81 antibodies, recombinant human CD9 and CD63 EC2 domains produced as GST (gluthatione-s-transferase)-tagged proteins inhibited LGC formation whilst human CD81 EC2, mouse CD9 EC2 or GST had no effect (Takeda et al., 2003; Parthasarathy et al., 2009). These results suggest that the mode of action of recombinant EC2 domains is distinct from anti-tetraspanin antibodies. These proteins did not have anti-proliferative or cytotoxic effects and did not alter monocyte adhesion or aggregation properties (Parthasarathy et al., 2009).

Chapter 1

Parthasarathy et al. (2009) propose that EC2 domains may intercalate into existing tetraspanin-enriched microdomains (TEMs) on the monocyte surface thereby disrupting the lateral molecular associations of tetraspanins. Recombinant EC2 domains have been utilised by several groups to study tetraspanin functions as will be discussed in more detail in section 1.3.

Takeda et al. (2003) have shown that the effect of anti-tetraspanin antibodies on LGC formation could not be attributed to Fc receptor crosslinking since F(ab’)2 fragments lacking the Fc portion had the same enhancing effect. In addition, no interference with cell adhesion, aggregation, proliferation or Con A-induced cytokine production was observed (Takeda et al., 2003). Considering their involvement in other fusion processes at a post-adhesion step, it is conceivable that tetraspanins regulate the function of a laterally-associated membrane protein involved in monocyte fusion.

A potential partner molecule is Matrix metalloproteinase 9 (MMP9). IL-4-induced FBGC formation is reduced in MMP9-null murine macrophages in vitro and in vivo (MacLauchlan et al., 2009). Ablation of CD9 and CD81 in murine macrophages resulted in increased MMP9 expression and activity (Takeda et al., 2008) suggesting that these tetraspanins may inhibit monocyte fusion by supressing MMP9 (Oursler, 2010). Interestingly, Con A stimulation of fibroblasts results in MMP9 and MMP2 secretion (Amin et al., 2007) so matrix metalloproteinases could potentially be involved in Con A-induced MGC formation as well.

Members of the ADAM (A disintegrin and a metalloprotease) protein family are also promising candidates for several reasons. Several ADAM proteins are expressed on the surface of monocytes and monocyte-derived giant cells and a number of ADAM proteins have already been implicated in monocyte (LGCs and osteoclasts) and myoblast fusion (Yagami-Hiromasa et al., 1995; Verrier et al., 2004; Namba et al., 2001; Ainola et al., 2009; Ishizuku et al., 2011). Some ADAM proteins contain putative fusion peptide domains with homology to viral fusogens (Namba et al., 2001). Finally, several tetraspanins have been shown to associate directly with ADAM proteins and to regulate their trafficking, cell surface expression or enzymatic functions in other cell types and organisms (Arduise et al., 2008; Dornier et al., 2012).

Chapter 1

Results contrary to those discussed above have been obtained using the murine monocyte cell line, RAW264.7 and murine MCSF-dependent macrophages (MDMs, see Table 1.1). In both cell types, treatment with anti-CD9 antibody inhibited rather than enhanced RANKL-induced osteoclast formation (Ishii et al., 2006; Yi et al., 2006). In RAW264.7 cells, CD9 siRNA (small interfering RNA) also inhibited osteoclastogenesis whilst CD9 overexpression resulted in the spontaneous formation of giant cells which did not express classic osteoclast markers (Ishii et al., 2006). Since antibody treatment also inhibited the formation of mononucleated osteoclasts, the authors concluded that CD9 does not specifically regulate the fusion event but has a general effect on monocyte differentiation into osteoclasts. The differences between these results and those obtained by the Takeda and Parthasarathy groups may be due to considerable variations in the culture systems employed (Table 1.1). Additionally in the RAW264.7 system, mRNA expression of Tspan5 was upregulated whilst Tspan13 expression was downregulated and siRNA knockdown of Tspan5 and Tspan13 had opposite effects, inhibiting and enhancing osteoclastogenesis, respectively (Iwai et al., 2007).

An interesting observation is that fusing murine monocytes develop interdigitations at sites of cell-cell contact (Vignery et al., 2005). These are reminiscent of the microvilli- like structures which are thought to be sites of sperm-egg fusion (section 1.2.3.1). During fertilisation, CD9 molecules became densely concentrated on oocyte microvilli and latrunculin B, a small-molecule inhibitor of actin polymerisation which alters microvilli morphology, inhibited sperm-oocyte fusion (Kaji et al., 2000; Runge et al., 2007). Latrunculin A (a related compound) similarly inhibited IL-4-induced FBGC formation (DeFife et al., 1999). It remains to be determined whether tetraspanins are similarly present on monocyte microvilli-like structures and if so whether they are associated with other molecules involved in monocyte fusion.

Chapter 1

1.3 Studying tetraspanin functions using recombinant EC2 domains Although numerous tetraspanins have been identified to date, most remain largely uncharacterised (Hemler, 2008). This can be attributed to the difficulty in generating monoclonal anti-tetraspanin antibodies in mice and relatively mild phenotypes in tetraspanin knockout organisms including multiple knockouts (Hemler, 2005; Fradkin et al., 2002). As an alternative, several studies have utilised recombinant tetraspanin EC2 domains to investigate tetraspanin functions in vitro (summarised in Table 1.2).

The large extracellular domain (EC2) of tetraspanins is the site for the majority of specific tetraspanin-partner interactions and is crucial to tetraspanin functions (section 1.1.1.2). Almost all anti-tetraspanin antibodies map to the EC2 region (Hemler, 2005). Recombinant tetraspanin EC2 domains are commonly produced as GST- or His (poly Histidine)-tagged fusion proteins in bacterial expression systems (Higginbottom et al., 2003; Parthasarathy et al., 2009). Western blotting and ELISA (Enzyme-linked immunosorbent assay) experiments using conformation-sensitive anti-tetraspanin antibodies are used when possible to demonstrate that these proteins are correctly folded and that the conformational antibody epitope is lost upon mutation of EC2 cysteines or chemical reduction of disulfide linkages (Higginbottom et al., 2003; Flint et al., 2006).

As can be seen from Table 1.2, a number of controls are included to ensure that effects of tetraspanin EC2 domains are specific and dependent on a correctly-folded epitope and to rule out the contribution of the fusion tag to these effects. These controls include the use of different fusion tags, free GST, free (cleaved) EC2 domains, mutant and heat-inactivated protein and EC2 domains of other tetraspanins from the same species or tetraspanin homologs from other species.

The precise mode of action of recombinant EC2s is not clear. Zhu et al. (2002) proposed that EC2 domains compete with endogenous tetraspanins for binding to a tetraspanin partner on the cell surface but since they lack the other tetraspanin domains they cannot functionally substitute for the endogenous protein leading to the effects observed in vitro. In several of the studies listed in Table 1.2, multiple tetraspanin EC2s had the same inhibitory action in the assay system employed (Higginbottom et al., 2003; Ho et al., 2006; Parthasarathy et al., 2009; Green et al.,

Chapter 1

2011) suggesting that they have a more general impact on the organisation of tetraspanin enriched microdomains (TEMs).

Microscopy techniques revealed that GST-human CD9EC2 (but not heat-inactivated protein) modified the dynamics of endogenous tetraspanin (CD9 and CD151) and non- tetraspanin (integrin receptors ICAM-1 and VCAM-1) TEM components, on the surface of endothelial cells (Barreiro et al., 2008). This lead the authors to suggest that recombinant CD9 EC2 intercalates into existing TEMs and disrupts binding of membrane-bound endogenous CD9 to laterally-associated tetraspanins and partners thus producing global effects on TEM organisation and function.

Higginbottom et al. used short synthetic peptide analogues of the variable subdomain of CD9, CD81 and CD63 EC2 to complement the data obtained using recombinant EC2 proteins (Higginbottom et al., 2003). In addition, short peptide analogues of the human CD9 EC2 region were found to potently inhibit bacterial adhesion to epithelial cell lines and are currently being explored as potential topical treatments for the prevention of bacterial infections in skin burns (Mr Daniel Cozens, Department of Infection and Immunity, University of Sheffield Medical School, internal communication). As with the recombinant EC2 domains, the mode of action of these peptides is currently unclear. The significant inhibitory activity of these peptides in the nanomolar concentration range and the lack of effect with control peptides (scrambled versions of the wild-type) suggest specificity.

Assay system Recombinant proteins Results References

A. Binding assays CD9/PSG17 GST None Ellerman et al., receptor-ligand GST-hCD9EC2 Bound to PSG17 2003 binding (by ELISA) GST-hCD9EC2 *mut None HCV glycoprotein GST-hCD81EC2 Binding Higginbottom E2/CD81 binding GST-aCD81EC2 None et al., 2000 (by ELISA) GST-hCD81EC2 muts Binding (some) RDS/RDS and GST-hRDSEC2 Successful pull-down Ding et al., RDS/ROM-1 GST-hRDSEC2 muts Successful pull-down (some) 2005 interactions (fusion MBP-hRDSEC2 Successful pull-down Chakraborty et tag pull-down) MBP-hRDSEC2 muts Successful pull-down al., 2012

Chapter 1

Assay system Recombinant proteins Results Reference

B. functional activity assays In vitro fertilisation GST None Zhu et al., 2002 assay to study GST-mCD9EC2 Inhibited fusion murine sperm-egg His-mCD9EC2 Inhibited fusion binding/fusion GST None Higginbottom GST-hCD63EC2 None et al., 2003 GST-mCD9EC2 Inhibited binding/fusion GST-hCD9EC2 Inhibited binding/fusion GST-hCD81EC2 Inhibited binding/fusion GST-hCD9EC2 muts Some inhibited fusion In vitro MGC GST None Takeda et al., formation assay GST-mCD9EC2 None 2003 induced by Con A GST-hCD9EC2 Inhibited MGC formation GST None Parthasarathy GST-mCD9EC2 None et al., 2009 GST-hCD81EC2 None His-hCD81EC2 None GST-hCD9EC2 Inhibited MGC formation GST-hCD63EC2 Inhibited MGC formation His-hCD9EC2 Inhibited MGC formation His-hCD63EC2 Inhibited MGC formation In vitro assay to GST None Green et al., study bacterial GST-hCD9EC2 Inhibited 2011 adhesion to GST-hCD63EC2 Inhibited epithelial cells GST-hCD81EC2 Inhibited GST-hCD151EC2 Inhibited In vitro HIV-1 GST None Ho et al., 2006 infection of GST-hCD9EC2 Inhibited infection monocyte-derived GST-hCD63EC2 Inhibited infection macrophages GST-hCD81EC2 Inhibited infection (CCR5-tropic) GST-hCD151EC2 Inhibited infection In vitro HCV None None Flint et al., infection of a GST-hCD81EC2 Inhibited infection 2006 hepatic cell lines GST-aCD81EC2 Partially inhibited infection GST-rCD81EC2 None GST-mCD81EC2 None GST (cleaved) None Rajesh et al., His-hCD81EC2 (GST Inhibited infection 2012 cleaved) His-hCD81EC2 Inhibited infection (mammalian) His-hTspan31EC2 None (mammalian) Microscopic study GST-hCD9EC2 Inhibited transmigration Barreiro et al., of integrin receptor GST-hCD9EC2 muts None 2005 dynamics on endothelial cells GST-hCD9EC2 Modified receptor dynamics Barreiro et al., GST-hCD9EC2 (heat- No effect 2008 inactivated) In vitro astrocyte GST-SCIP (unrelated None Kelic et al., growth regulation protein) 2001 GST-mCD81EC2 Blocked growth cycle arrest

Chapter 1

Table 1.2 A summary of in vitro studies utilising recombinant tetraspanin EC2 domains to investigate tetraspanin functions.

Abbreviations: a, African green monkey, Con A, Concanavalin A, EC2, Extracellular domain 2, ELISA, enzyme-linked immunosorbent assay, GST, Glutathione-s-transferase, HCV, Hepatitis C virus, His, polyhistidine-tag, HIV-1, Human immunodeficiency virus–type 1, h, human, MBP, maltose-binding protein, MGC, Multinucleated giant cell, m, murine, *mut, mutant, muts, mutants, r, rat, RDS, Retinal degeneration slow.

1.3.1 Issues regarding the use of recombinant proteins in immunological research Soluble recombinant proteins are valuable tools in molecular biology. Immunological research has long benefited from the use of recombinant cytokines (e.g. MCSF and TNF-α) and in recent years, protein-based pharmaceuticals such as engineered anti- tumour antibodies, recombinant allergens and therapeutic cytokines have become more prevalent (Zeiler et al., 1997; Rosenberg et al., 1985; Pavlou and Belsey, 2005).

The majority of these recombinant proteins are produced in a bacterial expression system such as E. coli (Escherichia coli) (Sarantos et al., 2013). A bacterial host offers several advantages including rapid cell growth, ease of manipulation of bacterial genetics and growth and expression conditions, high yields and low costs (Sorenson and Mortensen, 2005; Lobstein et al., 2012). However, a major drawback as far as immunological research is concerned is the presence of contaminating bacterial products in the purified recombinant protein preparation. These can include bacterial lipopolysaccharide (LPS), peptidoglycans, lipoproteins, nucleic acids and other proteins such as flagellin and porins (Wakelin et al., 2006; Iwasaki and Medzhitov, 2010). In an in vivo setting, these molecules would only be present at high concentrations in the event of bacterial colonisation/infection hence the mammalian immune system has evolved to react to these molecules as ‘danger signals’ (Hellman et al., 2000).

The responsiveness of immune cells to microbes and microbial products is mediated by pattern-recognition receptors (PRRs) which recognise pathogen-associated molecular patterns (PAMPs), molecular features exclusive to pathogen-derived products and not found in the host (Kumar, 2009). An important class of pattern- recognition receptors are Toll-like receptors (TLRs), integral membrane proteins with

Chapter 1 homology to the Toll protein of Drosophila (Tsan and Gao, 2004). So far, ten human TLRs have been characterised, mediating immunity to pathogens of bacterial, viral, parasitic and fungal origin as well as potential endogenous PAMPs such as products of necrotic cells (Akira et al., 2003; Mosser and Edwards, 2008). Various co-receptors and adaptor proteins can enhance TLR ligand binding and signalling activities (Akira 2003; Drexler and Foxwell, 2010). TLR signal transduction culminates in the transcription of pro-inflammatory cytokines (e.g. TNFα, IL-1 and IL-6) and other effectors of the immune system which mediate the response to infection (Zhang et al., 2011).

Since bacterial PAMPs are often present in recombinant protein preparations, there is a risk that the observed effects of recombinant proteins on immune cells is in fact due, partly or entirely, to the contaminating agent (Erridge, 2010). This has been found to be the case for a number of recombinant human proteins such as rhHsp70, rhHSP60 and gp96 (Bausinger et al., 2002; Gao and Tsan, 2003; Reed et al., 2003).

1.3.2 Documented effects of bacterial contaminants on monocytes and monocyte-derived giant cells Human and murine monocytes and monocyte-derived cells express multiple TLRs (Takami et al., 2002; Visintin et al., 2001). In particular, monocytes express relatively high levels of the bacterial lipopolysaccharide (LPS) receptor TLR4 and the co- receptors CD14 (itself a classic monocyte marker) and CD11/CD18, making these cells acutely sensitive to LPS (Sabroe et al., 2002; Fenton and Golenbock, 1998). Freshly isolated monocytes have a slightly higher binding affinity for LPS than differentiated macrophages (Gessani et al., 1993). LPS doses as low as 1pg/ml and exposure times as short as 5 to 10 minutes are sufficient to activate monocytes whilst maximal activation may be observed at 100-200pg/ml LPS (Fenton and Golenbock, 1998; Gallay et al., 1993; Gao and Tsan, 2003).

In vivo and in vitro, LPS-stimulated monocytes release powerful pro-inflammatory mediators to initiate an acute inflammatory response to the pathogen, including reactive oxygen species, pro-inflammatory cytokines (TNF-α, IL-1 and IL-6), prostaglandins and leukotrienes (Watson et al., 1994; Takami et al., 2002; Tsuzuki et al., 2001). These molecules mediate the immune response to control the infection and

Chapter 1 also increase monocyte viability by inhibiting programmed cell death (Mangan et al., 1991). A number of other TLR ligands such as peptidgoglycans (TLR2), poly(I:C) dsRNA (TLR3), flagellin (TLR5) and CpG unmethylated DNA (TLR9) can induce similar responses in monocytes (in Takami et al., 2002). Massive overproduction of pro- inflammatory mediators by activated monocytes can result in septic shock and multiple organ failure in man and other animals therefore the role of TLR ligands in monocyte activation and function has been the subject of intense research (Guiterrez- Ramos and Bluthmann, 1997; Wakelin et al., 2006). In contrast, limited data is available on the role of TLR ligands on monocyte-derived giant cell formation.

Doses of LPS (TLR4 ligand) as high as 10µg/ml did not induce LGC formation in vitro but inhibited Con A-induced LGC formation (reported in Takashima et al., 1993). On the other hand, previous data from our group showed an enhancement of Con A- induced giant cell formation at very high LPS doses but no inhibitory effects (reported in Parthasarathy et al., 2009). Similar to LGC formation, a high dose of LPS by itself is unable to induce differentiation of murine bone marrow-derived macrophages into osteoclasts (Liu et al., 2009). The effects of LPS and other TLR ligands in in vitro osteoclastogenesis assays appear to be highly dependent on the differentiation stage of the monocyte precursor. If added at the start of the culture period, LPS and other TLR ligands such as peptidoglycan, poly(I:C) dsRNA, flagellin and CpG unmethylated DNA strongly inhibit RANKL-induced osteoclast formation in a dose-independent manner (Takami et al., 2002; Zou and Bar Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009). Toxicity, anti-proliferative or pro-apoptotic effects have also been ruled out as contributing factors (Zou and Bar Shavit, 2002). However, TLR ligands down-regulate RANKL and MCSF receptor expression on osteoclast precursors in vitro (Ji et al., 2009; Hotozeka et al., 2007). A number of pro-inflammatory cytokines are also able to inhibit RANKL-induced osteoclastogenesis, suggesting that they operate downstream of TLR ligands (Takayanagi et al., 2002; Lee et al., 2008; Lee et al., 2010).

These inhibitory effects are unexpected given that TLR ligands and pro-inflammatory cytokines are potent inducers of pathological osteoclastogenesis and resorption in inflammatory bone diseases such as arthritis and periodontitis (Zhang et al., 2011). It is important to note that in vitro osteoclastogenesis assays normally lack the full

Chapter 1 complement of cells present on or near bone surfaces, such as osteoblasts, stromal cells and fibroblasts (Liu et al., 2009; Marouti et al., 2011). In addition, under inflammatory conditions, various immune cells such as neutrophils and B and T lymphocytes also infiltrate the bone microenvironment (Cvija et al., 2012; Mormann et al., 2008). Many of these cell types are responsive to TLR ligands or the products of TLR stimulation and can induce osteoclastogenesis directly or indirectly by producing RANKL, MCSF and other mediators (Liu et al., 2009; Cvija et al., 2012; Teitelbaum et al., 2006; Ji et al., 2009).

In contrast to their inhibition of monocyte to pre-osteoclast differentiation, TLR ligands promote pre-osteoclast fusion and survival in vitro, even in the further absence of RANKL. The mechanisms for this action are yet to be elucidated with studies variously suggesting the involvement of the transcription factor NF-κB, inflammatory cytokines and chemotactic proteins such as CXCL2 and CXCR4 (Zou and Bar Shavit, 2002; Takami et al., 2002; Suda et al., 2002; Itoh et al., 2003; Xing et al., 2011; Hotozeka et al., 2007; Ha et al., 2011). Whatever the mechanisms, the net effect of TLR stimulation in the in vivo bone environment appears to be an enhancement of osteoclastogenesis and activity. In light of the above findings, it is imperative to rule out the contribution of bacterial products to recombinant protein activity in in vitro giant cell formation assays.

1.3.3 Dealing with recombinant protein contamination Bacterial LPS is the major cell wall component of Gram negative bacteria, covering approximately 75% of the outer surface (Alexnader and Rietschel, 2001). LPS molecules are continuously liberated into the environment in which bacteria are growing. High concentrations of LPS are to be found in bacterial fermentation cultures and especially in bacterial lysates obtained for recombinant protein purification (Petsch and Anspach, 2000). In its natural state, LPS tightly associates with bacterial proteins, nuclei acids and phospholipids and is thus referred to as ‘endotoxin’ (Wakelin et al., 2006). These minor contaminants can bind other TLRs and activate monocytes independently of LPS, as discussed in section 1.3.2 (Takami et al., 2002). However, suitable methods for the removal of non-LPS contaminants from recombinant proteins and suppression of their effects in culture are not currently available. LPS contamination of therapeutic agents and in particular of bacterially- derived products is of major concern to the pharmaceutical industry (Hoffmann et al., 44

Chapter 1

2005). LPS has long been recognised as a pyrogen (fever-inducing compound) with the potential to cause septic shock and multiple organ failure. For these reasons, most research publications have concentrated on determining and reducing/eliminating the effects of contaminating LPS (Erridge, 2010).

1.3.3.1 Measuring the level of contamination The concentration of LPS in aqueous solutions can be measured with relative ease and accuracy using the Limulus Amoebocyte Lysate (LAL) chromogenic assay (Wakelin et al., 2006). Pure LPS is available commercially and it is common practice to use equivalent LPS concentrations to that found in the recombinant protein preparation to determine the contribution of LPS to recombinant protein activity (Wakelin et al., 2006). A disadvantage of the LAL assay is that it cannot detect the presence of other contaminants such as lipoproteins. Also, LPS-binding proteins such as LBP and net- positively charged proteins can effectively mask LPS leading to falsely low or negative results (Petsch et al., 1998; Erridge, 2010). As an alternative, cell-based assays are available which determine the ‘pyrogenicity’ of the test material by measuring pro- inflammatory cytokine release (Hoffmann et al., 2005). These assays have the potential to detect other minor contaminants but they are more laborious to carry out and do not quantify the levels of individual contaminant(s). Also, it cannot be ruled out that the recombinant protein itself stimulates cytokine release.

1.3.3.2 LPS reduction techniques A number of techniques are available to reduce LPS contamination in the protein sample, as detailed in Table 1.3. Following LPS reduction, an independent method must establish that the recombinant protein is still functionally active and that reagents introduced during the purification procedure (such as detergents) do not interfere with subsequent assays (Reichelt et al., 2006; Zimmerman et al., 2006). This is usually determined by examining the ability of the protein to bind a specific ligand, receptor, antibody or substrate or catalyse an enzymatic reaction (Read et al., 2003).

As can be seen from the table, a common drawback of most purification methods is substantial loss of recombinant protein. In addition, these techniques can reduce but not completely eliminate the LPS. Therefore complementary methods are often used to discriminate between the effects of LPS and those of the recombinant protein under

Chapter 1 investigation. These include neutralisation of LPS using LPS inhibitors and antagonists, blocking of LPS receptors and co-receptors and inactivation of the recombinant protein by boiling and protease treatment to determine residual activity in the solution.

Method Details Drawbacks References Phase separation Non-ionic detergent -Protein denaturation due to Aida and (also known as Exploits large temperature shifts Pabst, 1990; detergent hydrophobicity of LPS -Loss of hydrophobic proteins Reed et al., extraction) and temperature shifts by partitioning into the 2003 to drive endotoxin detergent phase particles into the -Necessity of residual detergent detergent phase whilst removal by gel filtration the protein remains in the aqueous phase Inclusion of Use of appropriate -Potential changes in protein Reichelt et detergent wash detergents to wash the conformation and/or activity al., 2006; step in affinity protein whilst still Zimmerman chromatography bound to the et al., 2006 chromatography column Ultrafilteration Use of dialysis -Unsuitable for high molecular Petsch and membranes to retain compounds (10KDa molecular Anspach, endotoxin aggregates weight cut-off) 2000; -Unsuitable if endotoxin is Anspach bound tightly to proteins and Hilbeck, 1995 Anion exchange Based on the binding of -Multiple rounds of binding Anspach negatively charged required (loss of recombinant and Hilbeck endotoxin to positively- protein) 1995 charged resins such as -Non-specific binding of protein DEAE to resin (diethylaminoethanol) -Unsuitable for net-negatively resin charged proteins Endotoxin- e.g. Polymyxin B affinity -Cellular effects/toxicity of Reed et al., selective affinity resin residual polymyxin B 2003; adsorption -Loss of protein (non-specific Anspach binding to resin) and Hilbeck, -Loss of protein (multiple 1995 rounds of binding) -Unsuitable for net-negatively charged proteins -Requires optimisation of pH and ionic strength Table 1.3 A summary of LPS reduction techniques. Abbreviations: LPS, Lipopolysaccharide, KDa, Kilodaltons

Chapter 1

1.3.3.3 LPS neutralisation At neutral pH, LPS molecules have a net negative charge and thus have an affinity for cationic compounds (Peterson et al., 1986). One such compound is polymyxin B, an antimicrobial polypeptide isolated from the Gram positive bacterium Bacillus polymyxa (Tsubery, 2000). Polymyxin B is an amphipathic molecule with multiple positively charged groups and a long hydrophobic tail and has been shown to be very effective at neutralising LPS in vitro and in vivo. Immobolised polymyxin B is used to purify LPS away from whole blood, tissue culture media or recombinant proteins (Aoki et al., 1994; Cardoso et al., 2007). Pre-incubation of soluble polymyxin B with pure LPS blocks LPS binding to and stimulation of human monocytes in in vitro experiments (Gallay et al., 1993; Gao et al., 2003). Some studies have cautioned against the use of polymyxin B, presenting evidence that the compound itself can stimulate cytokine production (Damais et al., 1987; Jaber et al., 1998), whilst others have found that its neutralisation of protein-bound LPS is variable (Loung et al., 2012).

Competitive inhibition of LPS can be achieved by inactive LPS variants from bacterial species such as R. sphaeroides and synthetic lipid A analogs (Fenton and Glockenbock, 1998; Alexander and Reitschel, 2001). Yet another approach is to functionally block the LPS receptor TLR4 or the co-receptor CD14 using antibodies (Lewthwaite et al., 2001; Wakelin et al., 2008; Giambartolomei et al., 2004) or to add soluble TLR4 to compete with the cell-bound receptor (Hyakushima et al., 2004). The latter only partially inhibited LPS-TLR4 interactions.

1.3.3.4 Recombinant protein inactivation The differential sensitivity of LPS and recombinant proteins to heat treatment can be exploited to determine whether the ‘active’ component of a recombinant protein solution is LPS or the protein itself. LPS is frequently described in the literature as being extremely heat-stable (Alexander and Reitschel, 2001) whilst recombinant proteins (including tetraspanins e.g. GST-hCD9EC2) can be efficiently inactivated by heating at 95˚C for 5 minutes (Barreiro et al., 2008). One report suggested that heating an LPS solution for one hour significantly reduced its cytokine-releasing activity (Gao et al., 2003). This may result from the excessively long incubation time or may be specific to the bacterial strain used since other similar reports were not found in the literature. Recombinant protein can also be inactivated by protease treatment, using

Chapter 1 immobolised proteinase K or other enzymes to which LPS will not be sensitive (Dudley et al., 2003; Bulut et al., 2005; Loung et al., 2010). A caveat is that commercial proteases are usually expressed in a bacterial system and so protease treatment may introduce significant bacterial contaminants, as was found in one study (Erridge, 2010).

1.3.3.5 Alternative expression systems Eukaryotic cells are generally more laborious and expensive to work with (Lobstein et al., 2012) therefore only a few instances exist in the literature where recombinant tetraspanins have been successfully produced in eukaryotic hosts. Full-length RDS was expressed and purified in Drosophila melanogaster S2 cells and in the yeast Pichia pastoris and characterised by circular dichroism (Vos et al., 2009). The EC2 domain of tetraspanin CD63 has previously been expressed as a recombinant polyhistidine- tagged protein in the human embryonic kidney cell line HEK293T by our group (A. Jiraviriyakul, PhD thesis). The EC2 domain of tetraspanin CD81 has also been expressed in the HEK293T (Human Embryonic Kidney) cell line and appeared fully active and folded (Rajesh et al., 2012). However, expression of recombinant CD9, CD151 and Tspan5 EC2 domains in mammalian (HEK293 and CHO cells) and S19 insect cells has proven problematic (John Palmer, Department of Molecular Biology and Biotechnology, University of Sheffield, personal communication).

1.4 Aims The present study was undertaken with the aim of expanding the repertoire of recombinant tetraspanin EC2 domains available for tetraspanin research. We were interested in particular in the role of poorly-investigated tetraspanins in multinucleated giant cell (MGC) formation. As mentioned in section 1.2.3.3.4, the expression of multiple tetraspanins has been demonstrated in monocytes, monocyte cell lines and monocyte-derived giant cells. However, the functions of only five tetraspanins have been explored in MGC formation and only three (CD9, CD81 and CD63) in any great detail. The redundancy of tetraspanin function, demonstrated in several systems including in monocyte fusion, would make it likely that other tetraspanins are also involved in MGC formation (Table 1.2).

Chapter 1

A panel of nine recombinant tetraspanin EC2 domains were produced as GST-tagged fusion proteins in a bacterial expression system. To increase recombinant protein yields, expression in a new host strain was optimised and recombinant proteins were purified and characterised by biochemical techniques such as Western blotting, ELISA and circular dichroism. Recombinant protein functionality was then determined in in vitro Concanavalin A (Con A)-induced MGC formation assay. Contrary to previous data by our group, recombinant protein activity in this assay system appeared to be adversely affected by bacterial contamination of the purified protein preparation. Therefore, a secondary aim was to reduce or exclude the contribution of bacterial contaminants using various measures.

We also aimed to extend the work of Green et al. (2011) in using recombinant tetraspanin EC2 domains to investigate the role of tetraspanins in bacterial adhesion to epithelial cells. In addition, recombinant protein activity in two other monocyte fusion systems, Burkholderia thailandensis-induced MGC formation and RANKL- induced osteoclast was determined and contrasted with the findings in the Con A- induced MGC formation assay.

Chapter 2

Chapter 2: Materials and methods

2.1 Materials

2.1.1 Laboratory equipment and instrumentation

Equipment Supplier

Centrifuges Eppendorf Minispin (12,100g) (Maximum RCF, ×g) Sigma 3K15 (5,400g) Beckman CP swinging-bucket rotor (14,000g) Beckman Avanti JA-25: JA-10.5 rotor (18,500g) JA-14 rotor (30,100g) JA-25.5 rotor (76,600g)

Microscopes Olympus CK40: Brightfield Inverted 10, 20 and 40× magnification objective

Nikon Eclipse E400: Brightfield and fluorescence Upright

InCell Analyzer 1000: Brightfield and fluorescence Inverted High content screening system

ImageXpress Micro: Fluorescence Inverted High content screening system Spectrophotometers Thermo Scientific NanoDrop 2000 (DNA quantitation)

Table 2.1 Laboratory equipment

2.1.2 Computer software

Software/database Used for Supplier BioEdit Viewing http://www.mbio.ncsu.edu/bioedit/bioedit.ht chromatogram files ml Clone Manager Suite Subcloning/primer Scientific & Educational Software 7 design Clustal Omega Multiple sequence http://www.ebi.ac.uk/Tools/msa/clustalo/ alignment (Protein) Ensembl genome Determination of http://www.ensembl.org/index.html browser tetraspanin EC2 boundaries Expasy Translate tool DNA to protein http://web.expasy.org/translate/

Chapter 2

translation GraphPad Prism6 Graph construction http://www.graphpad.com/scientific- and statistical analysis software/prism/ Image J Densitometry http://rsb.info.nih.gov/ij/ Irfanview Image processing http://www.irfanview.com/ SeqMan and seqAlign DNA sequence analysis DNAStar T-COFFEE Multiple sequence http://tcoffee.crg.cat/apps/tcoffee/do:regular alignment (protein) UniProtKB/Swiss- Determination of http://www.uniprot.org/ Prot bioinformatics tetraspanin EC2 portal boundaries

Table 2.2 Computer software

2.1.3 Laboratory reagents and buffers Ultrapure water was prepared using a PURITE NEPTUNE system (Scientific Laboratory Supplies Limited). Nuclease-free water was purchased from Gibco. Bacterial culture media, buffers, glassware, pipette tips and other equipment/reagents were sterilised by autoclaving at 121°C, 15psi for 20 min. Antibiotics, antibodies, buffers and other laboratory reagents (where stated) were filter-sterilised using 0.2µm disposable filters (Sartorius).

2.1.3.1 Electrophoresis gels

Gel Composition

Agarose gel (1%) 1g agarose dissolved in 100ml TAE buffer by microwaving SDS-PAGE separating 2.5ml water, 3ml 30% acrylamide, 1.9ml separating buffer (4×), 112µl gel (12.5%) 10% ammonium persulphate (made up in water), 5µl TEMED SDS-PAGE stacking 1ml water, 300µl 30% acrylamide, 444µl stacking buffer (4×), 28µl 10% gel (5%) ammonium persulphate (made up in water), 5µl TEMED

Table 2.3 Electrophoresis gels for DNA and protein analysis

2.1.3.2 Buffers

Buffer Recipe

Agarose gel running 242g Tris base, 57.1 ml glacial acetic acid, 100 ml 0.5 M EDTA made buffer/TAE (50×) up to 1L with water (pH8) BBN 0.1% sodium azide, 0.2% BSA dissolved in HBSS HBSS (modified) No Mg2+, No Ca2+, no phenol red (Lonza)

Chapter 2

Blocking buffer (Western 5% w/v semi-skimmed milk powder made in 1×TBS buffer, 0.15% blot) Tween 20 Blotting buffer (10×), no 30.3g Tris, 144g glycine made up to 1L in water. methanol Blotting buffer (1×) 100ml 10× blotting buffer, 200ml methanol, made up to 1L with water Coomassie gel de-stain 250ml methanol, 5ml acetic acid made up to 1L with water ELISA binding buffer 16ml 0.2M anhydrous sodium carbonate, 34ml 0.2M sodium (Carbonate-bicarbonate hydrogen carbonate made up to 200ml with water, pH9.6 buffer, 0.05M) ELISA blocking buffer PBS, 0.05% Tween 20, 1% w/v Milk powder ELISA wash buffer PBS, 0.05% Tween 20 Glutathione elution buffer 0.154g reduced glutathione dissolved in 50ml 50mM Tris-HCl, pH8 (25mM) Laemmli sample Novex® Tris-Glycine Native Sample Buffer (Invitrogen, LC2673) buffer/non-reducing (2×) Laemmli sample 4ml Tris-HCl (adjust pH to 6.8), 2.5ml 20% w/v SDS, 8ml glycerol, buffer/reducing (4×) 0.008g bromophenol blue, 4ml β-mercaptoethanol, made up to 20ml with water

PBS (10×) 80g NaCl, 2g KCl, 11.5g Na2HPO4, 2g KH2PO4 dissolved in water SDS-PAGE running buffer 30g Tris base, 144g glycine, 10g SDS, made up to 1L water (10×) SDS-PAGE separating gel 18.17g Tris base, 4ml 10% w/v SDS, pH adjusted to 8.8 and made up buffer (4×) to 100ml in water, stored at 4 0C SDS-PAGE stacking gel 6.06g Tris base, 4ml 10% w/v SDS, pH adjusted to 6.8 and made up buffer (4×) to 100ml in water, stored at 4 0C TBST 8.8g NaCl, 0.2g KCl, 3g Tris base, 500µl Tween 20, pH adjusted to 7.4 and made up to 1L in water

Table 2.4 Buffers

2.1.3.3 Antibodies

Working Antigen Clonality Label Source concentration Mouse anti-human CD9 Monoclonal - 1µg/ml (Andrews et al., 1981) (602.29) Rabbit anti-human Tspan2 Polyclonal - 1µg/ml Aviva Systems Biology (ARP46643) Rabbit anti-human Tspan3 Polyclonal - 1µg/ml Aviva Systems Biology (ARP46642) Rabbit anti-human Tspan4 Polyclonal - 1µg/ml Aviva Systems Biology (ARP44756) Rabbit anti-human Tspan5 Polyclonal - 1µg/ml Aviva Systems Biology (ARP47740) Rabbit anti-human Tspan8 Polyclonal - 1µg/ml Aviva Systems Biology

Chapter 2

(ARP46477) Rabbit anti-human CD82 Polyclonal - 1µg/ml Aviva Systems Biology (ARP63360) Kind gift, Prof. Eric Mouse anti-human CD82 Polyclonal - 1in 15,000 Rubinstein (TS82b) (INSERM U602) Rabbit anti-human Tspan31 Polyclonal - 1µg/ml Aviva Systems Biology (ARP46677) Rabbit anti-human Tspan32A Polyclonal - 2µg/ml Aviva Systems Biology (ARP46635) Rabbit anti-human Tspan32B Polyclonal 2µg/ml Aviva Systems Biology (ARP46636) Kind gift, Prof. Peter Mouse anti-Ok(a) antigen Polyclonal - 1 in 5 Andrews, University (TRA-1-85) of Sheffield Goat anti-mouse polyvalent immunoglobulins (G,A,M) Polyclonal FITC 1 in 250 Sigma (F1010) Rabbit anti-mouse IgG (whole Polyclonal HRP 1 in 5,000 Sigma molecule) (A9044) Mouse anti-rabbit Polyclonal HRP 1 in 7,000 Sigma immunoglobulins (A2074) (Muranova et al., Mouse isotype control (JC1) Monoclonal - 1µg/ml 2004) Rabbit anti-Schistosoma Polyclonal HRP 1 in 5,000 Sigma japonicum GST (A7340)

Table 2.5 Antibodies and concentrations used in Western blotting

2.1.4 Cloning materials 2.1.4.1 Bacterial strains

E. coli Strain Genotype

- - - Rosetta-gami B(DE3)pLysS F ompT hsdsB (rB mB ) gal dcm lacY1 ahpC (DE3) gor522::Tn10 (Novagen) trxB pLysSRARE (CamR, KanR, TetR) E. coli SHuffle T7 Express MiniF lysY (CamR) / fhuA2 lacZ::T7 gene1 [lon] ompT ahpC gal lysY (New England Biolabs, λatt::pNEB3-r1-cDsbC (SpecR, lacIq) ΔtrxB sulA11 R(mcr- cat. No. C3030H) 73::miniTn10--TetS)2 [dcm] R(zgb-210::Tn10 –TetS) endA1 Δgor ∆(mcrC-mrr)114::IS10

- - DH5α (Invitrogen, Cat. No. F φ80lacZ∆M15 ∆.(lacZYA-argF)U169 recA1 endA1 hsdR17(rk , + - 18263-012) mk ) phoA supE44 thi-1 gyrA96 relA1 λ

Table 2.6 Bacterial strains

2.1.4.2 Growth media

Medium Recipe LB (Lysogeny Broth) 10g tryptone, 5g yeast extract, 10g NaCl dissolved in 1L Milli-Q

Chapter 2

water (pH7), autoclaved LB solid agar 10g tryptone, 5g yeast extract, 10g NaCl, 15g bacteriological agar dissolved in 1L Milli-Q water (pH7), autoclaved. Agar was cooled to ~50°C before addition of appropriate antibiotics SOB (Super optimal broth) 4g tryptone, 1g yeast, 0.1g NaCl, 2ml 250mM KCL made up to 200ml with water, autoclaved

SOC (Super optimal broth 195ml SOB (autoclaved), 1ml 2M MgCl2 (autoclaved), 4ml 1M with catabolite repression) glucose (filter-sterilised)

TB (Terrific broth) 12g tryptone, 24g yeast extract, 4ml glycerol, 2.3g KH2PO4 and 12.5g K2HPO4 dissolved in 1L milli-Q water (pH7), autoclaved

TP (Tryptone phosphate 20g tryptone, 15g yeast, 8g NaCl, 2g Na2HPO4, 1g KH2PO4 dissolved broth) in 1L Milli-Q water (pH7), autoclaved

Table 2.7 Bacterial and tissue culture growth media

2.1.4.3 Antibiotics

Working concentration (µg/ml)

Antibiotic (50mg/ml stock) E. coli SHuffle E. coli Rosetta- DH5α gami Carbenicillin (in ultrapure water) 10 50 100 Chloramphenicol (in 100% ethanol) 34 34 - Kanamycin (in ultrapure water) - 15 - Tetracyclin (in 100% ethanol) - 34 -

Table 2.8 Antibiotic preparations

2.1.4.4 cDNA and protein sequences

RefSeq mRNA UniProtKB protein Final Expression Gene accession number accession number vector GST - P08515 pGEX-KG* CD9 NM_001769.3 P21926 pGEX-CD9EC2* Tspan2 NM_005725.4 O60636 pGEX-Tspan2EC2** Tspan3 NM_005724.5 O60637 pGEX-Tspan3EC2 Tspan4 NM_001025237.1 O14817 pGEX-Tspan4EC2 Tspan5 NM_005723.3 P62079 pGEX-Tspan5EC2* Tspan8 NM_004616.2 P19075 pGEX-Tspan8EC2* Tspan13 NM_014399.3 O95857 pGEX-Tspan13EC2 Tspan27 NM_002231.2 P27701 pGEX-Tspan27EC2

Tspan31 NM_005981.3 Q12999 pGEX-Tspan31EC2 Tspan32 NM_139022.2 Q96QS1 pGEX-Tspan32EC2

Chapter 2

Table 2.9 cDNA and protein sequences mRNA and protein accession numbers were obtained from the bioinformatics website http://www.uniprot.org/uniprot/. All sequences are human apart from GST which is the Sj26 protein from Schistosoma japonicum (Smith and Johnson, 1988). All cDNA clones were obtained from Geneservice Limited apart from Tspan27 which was purchased from OriGene (Cat. SC1077926). * cloning performed previously by our group, ** cloning performed by Mutagenex.

2.1.4.5 Oligonucleotide primers Primers were received in lyophilised form. Primers were reconstituted in sterile nuclease-free water to make 100µM stocks and stored at –20°C.

Name Sequence

*pGEX-F2 5’ GAGCTGTTGACAATTAATCATCGG 3’ *pGEX-R1 5’ CCGGGAGCTGCATGTGTCAGAGG 3’ *TS2GexEcoR1-F1 5’ TTAATTGAATTCTAATAGGCAAGGGGGTAGCTATCC 3’ *TS2GexHindIII-R1 5’ ACTGACAAGCTTCTACTTAACACTGATTATGGTC 3’ pGEX-TS3-XbaI-F 5’ CGTGAGATTCTAGACTACAGAGCAAAGGTG 3’ pGEX-TS3-HindIII-R 5’ GCGCCGGAAGCTTTCAGATTTCTTGTAGCTTCTTCAC 3’ pGEX-TS4-EcoR1-F 5’ GACTGGAATTCTATACACGGACAAGATTGAC 3’

pGEX-TS4-HindIII-R 5’ GTCACTAAGCTTTCAGTTCTCCTGAAGCCACAC 3’

*TS5GEXKGEcoRI-F1 5’ GGACGCGGAATTCTATTCAAAGACTGGATCAAAGACC 3’

*TS5GEXKGHindIII-R1 5’ CGCGCCAAGCTTTCAATTGTCCTGCAACCAC 3’ *TS8GEXKGEcoRI-F1 5’ GGACGCGGAATTCTATTCAAATCTAAGTCTGATCGC 3’

*TS8GEXKGNcoI-R1 5’ CGTCGCGGCCCATGGTCAATTTTTTGCCAAGAAGTC 3’ pGEX-TS13-EcoR1-F 5’ CGATGGAATTCTAAACCAGGAGCAACAGGGT 3’ pGEX-TS13-HindIII-R 5’ CGCGCCAAGCTTTCACTCTCCAGCATATTCTCC 3’

pGEX-TS25-EcoRI-F 5’ GACTGGAATTCTAGAACAGAAGCTGAATGAG 3’ pGEX-TS25-HindIII-R 5’ GTGACTAAGCTTTCAATTGGAATGAAACCACAG 3’ pGEX-TS26-EcoR1-F 5’ GTCAGGAATTCTATCCACTCAGCGGGCCCAG 3’

pGEX-TS26-NcoI-R 5’ GTCAGTCCATGGTCAGTTGTTGTGCAGCCACTT 3’

pGEX-TS27-EcoRI-F 5’ GACTGGAATTCTAAACATGGGCAAGCTGAAG 3’

pGEX-TS27-HindIII-R 5’ GTGACTAAGCTTTCAGTTCTCCTGCAGCCACGC 3’

pGEX-TS31-EcoRI-F 5’ GACTGGAATTCTAAACCGAAGCAAACAGACA 3’

pGEX-TS31-HindIII-R 5’ CTGTGAAAGCTTTCATTCGTCTGAATGCTTAAG 3’

*TS32GexEcoRI-F 5’ TTAATTGGAATTCTACACAGCCCCACCCCAGG 3’

*TS32GexHindIII-R 5’ CACGCTAAGCCTTCTACTGCTGGTGTGTCCTCAGG 3’

Chapter 2

Table 2.10 Forward and reverse primers used in sub-cloning and in sequencing The tetraspanin-specific sequence is shown in boldface. F designates forward (sense) primer and R designates reverse (anti-sense) primer. * primers designed previously by Mr. Ralph Hyde and produced by Genosys. All other primers were designed by the author and manufactured by Eurofins.

2.1.4.6 Restriction enzymes and buffers All enzymes and buffers were purchased from Promega.

Stock Enzyme Buffer concentration EcoR1 12units/µl E HindIII 10units/µl E NcoI 10units/µl H XbaI 8-12units/µl E EcoRI/HindIII - E EcoRI/NcoI - H HindIII/XbaI - E

Table 2.11 Endonuclease restriction enzymes

2.1.4.7 PCR, digest and ligation mix composition

PCR mix Digest mix Ligation mix

Component Volume Component Volume Components Volume (µl) (µl) (µl) Forward Template DNA Digested plasmid 2 1.5 4 primer (25µM) (450ng/µl) (see text) Digested PCR Reverse primer 2 Buffer (10×) 2 product (see 2 (25µM) text) NH buffer Ligation buffer 4 5 BSA (10mg/ml) 0.2 2 (10×) (10×) T4 DNA ligase MgCl (50mM) 2 Enzyme 1 1.5 1 2 (5units/µl) BIOTAQ DNA Nuclease-free polymerase 1 Enzyme 2 1.5 11 water (5units/µl) Nuclease-free dNTPs (10mM) 5 13.3 water Template DNA 1 (5ng/µl) Nuclease-free 32 water Total volume 50 Total volume 20 Total volume 20

Chapter 2

Table 2.12 PCR, digest and ligation mix composition BIOTAQ PCR kit and ligation kit were obtained from Bioline. Restriction enzymes, buffers and BSA were obtained from Promega. In single digests, 1.5µl of nuclease-free water was added instead of the second restriction enzyme. Restriction buffers are listed in Table 2.11. For ligation, a 3:1 molar ratio of insert (PCR product):vector was used (section 2.2.2.6) therefore the actual concentrations of vector and PCR product were different for each reaction but the volumes used were kept constant.

2.2 Methods

2.2.1 General molecular biology methods

2.2.1.1 DNA Sequencing pGEX-KG plasmid vectors encoding various tetraspanin constructs were sequenced using forward and reverse sequencing primers pGEX-F2 and pGEX-R1 (Table 2.10). Sequencing was performed by the Core Genomic Facility DNA sequencing service (University of Sheffield, UK) using Prism® Big Dye® terminator sequencing technology (ABI). Sequences were analysed using seqMan and seqAlign software (DNASTAR), the program BioEdit and the bioinformatics website, http://www.expasy.ch/tools/dna.html.

2.2.1.2 Plasmid DNA purification Plasmid DNA was prepared using QIAprep Spin Mini kits (Qiagen). Single colonies of transformed bacteria carrying the relevant construct were picked into 10ml LB (supplemented with appropriate antibiotics) and grown overnight at 37°C with shaking (300 rpm). Cells were harvested by centrifugation at 5,400g for 10 minutes at 4°C. Pellets were re-suspended in 500µl P1 buffer and transferred to a 2ml microfuge tube. 500µl buffer P2 was then added followed by 700µl buffer N3. Cell debris was pelleted by centrifugation at maximum speed (12,100g in an eppendorf minispin centrifuge) for 10 minutes. 750µl of supernatant was added to a minispin column (Qiagen) and centrifuged for 1 minute at full-speed. Supernatant was decanted and the rest of the supernatant was added to the column and spun, etc. The column was then washed with 500µl buffer PB and 750µl buffer PE. Plasmid DNA was eluted in 50µl sterile nuclease-free water and stored at -20°C.

2.2.1.3 Preparation of glycerol stocks Single colonies of transformed bacteria were picked into 5ml LB medium supplemented with the appropriate antibiotics and grown overnight at 37°C with shaking (200 rpm). 750µl of the overnight cultures were transferred to cryotube vials

Chapter 2

(Nunc) before addition of 250µl of 60% v/v glycerol (in ultrapure water). Tubes were briefly vortexed and flash-frozen in liquid nitrogen for 20-30 minutes before long- term storage at -80°C.

2.2.1.4 Agarose gel electrophoresis 1% agarose gels were made as described in Table 2.3. Once the gel had cooled sufficiently, 1µl/100ml v/v ethidium bromide (10mg/ml stock) was added and the gel was poured into a gel tray containing a comb. When the gel had set, it was placed in a gel tank filled with 1× TAE buffer (Table 2.4). Typically 10µl of sample DNA mixed with 10µl 2× DNA loading buffer (Bioline) was loaded into each well. Pre-stained DNA marker (Hyperladder I, Bioline) was also loaded. Gels were run for approximately 1 hour at 100V and visualised using a UV lightbox.

2.2.1.5 SDS-PAGE SDS-PAGE was carried out using the Mini-PROTEAN Tetra Cell system as instructed by the manufacturer (Biorad). 12.5% separating (resolving) gel and 5% stacking gel were prepared as described (Table 2.3). For non-reducing SDS-PAGE, sample protein was boiled with 4× non-reducing sample buffer (Table 2.4) for 3 min. For reducing SDS- PAGE, sample protein was boiled with 4× reducing loading buffer (Table 2.4) for 10 minutes. To prevent re-oxidation of reduced protein during electrophoresis, crystals of iodoacetamide were added to the reduced sample. Iodoacetaminde is commonly used in peptide mass spectrometry to alkylate free thiol groups (Lane et al., 1978; Andre et al., 2006). The reaction results in a colour change from blue to yellow which is indicative of a pH change thus the pH was re-adjusted using ammonia vapour until the colour changed back to blue (Dr. John Holt, University of Sheffield, personal communication). Excess iodoacetamide does not affect protein mobility (Lane et al., 1978). Typically, 5-7µl of marker (All Blue Precision Plus, Biorad) and 10-20µl of sample protein were loaded onto the gel. Gels were run at 200V constant for approximately 60-90 minutes.

2.2.1.6 Coomassie staining and de-staining SDS-PAGE gels were stained with coomassie R-250 brilliant blue stain (Biorad) for 1 hour on a rocking table. Gels were incubated with de-stain (Table 2.4) overnight at 4˚C, washed in de-ionised water and imaged using a scanner.

Chapter 2

2.2.1.7 Western blotting Once the gel had run as described in section 2.2.1.5, a gel sandwich was prepared containing, a sponge, blotting paper, the polyacrylamide gel, nitrocellulose membrane (Hybond ECL, Amersham), blotting paper and another sponge. All components were soaked in blotting buffer before the sandwich was assembled. Then it was placed in a blot module with the nitrocellulose membrane facing the anode. A block of frozen water was also placed in the module to keep the buffer cool. The gel was blotted at 250mA for 1 hour, then placed in blocking buffer overnight at 4˚C to block non-specific protein binding sites.

2.2.1.8 Probing nitrocellulose membranes Blocked membranes were probed with the appropriate primary antibody (diluted in blocking buffer) for 1-2 hours at room temperature, on a rocking table. Membranes were washed once with TBST and incubated with secondary HRP-linked antibody for 1 hour. Membranes were then washed with TBST for 20 minutes with several changes of buffer. Membranes were then incubated with SuperSignal West Pico chemiluminescent substrate (Thermoscientific), according to manufacturer’s instructions. Membranes were then placed in a cassette and exposed to CL-Xposure X- ray film (Pierce) for 10 seconds to 16 hours. The exposed film was immersed in developing solution for 30 seconds, briefly washed with tap water, immersed in fixative solution for 30 seconds and left to dry.

2.2.1.9 Bradford assay Protein standards were prepared by making serial dilutions of BSA (Promega) in PBS, in the range 0.0625-1.5 mg/ml. 10µl samples were applied to wells of a 96-well ELISA plate (Nunc MaxiSorp®). 200µl Bradford reagent (Protein Assay Dye Reagent, Bio-rad, diluted 1:5 with PBS) was added to each well. The plate was incubated on a plate mixer for 10 minutes. Absorbance values at A620nm were determined on a plate reader. BSA readings were used to plot a standard curve from which concentration of the test protein could be determined.

2.2.1.10 Standard ELISA Wells of a 96-well ELISA plate (Nunc MaxiSorp®) were coated with 100µl/well of the target antigen (1-10µg/ml, diluted in ELISA binding buffer) by overnight incubation at 4°C. The plate was sealed with adhesive plate sealer to prevent evaporation. All further incubations were performed at room temperature. Wells were washed twice with ELISA wash buffer and blocked for 2 hours with 150µl/well blocking buffer.

Chapter 2

Wells were then washed three times with wash buffer followed by addition of 50µl/well of primary antibody (diluted in blocking buffer). After 2 hours incubation, wells were washed three times with wash buffer and incubated with 50µl/well enzyme-linked secondary antibody (diluted in blocking buffer) for 1 hour. Wells were then washed three times with wash buffer and twice with distilled water before incubation with 50µl TMB substrate (Sigma) for 15 min on a plate mixer. The reaction was stopped by addition of 50µl/well 2M H2SO4 and the absorbance (A450nm) was read on a plate reader.

2.2.2 Sub-cloning tetraspanin EC2 domains 2.2.2.1 Determination of EC2 domain boundaries Full-length tetraspanin protein sequences of all 33 human tetraspanins were obtained in FASTA format from the bioinformatics databases UniProtKB/Swiss-Prot and Ensembl genome browser. To determine EC2 domain boundaries, protein sequences were first aligned using the multiple alignment programme Clustal Omega. Sequence homology, database annotations to protein sequences (above databases) and previous published works on tetraspanin protein topology (Todd et al., 1998; Robb et al., 2001) were used to manually delineate EC2 domains.

2.2.2.2 PCR Primer design Following determination of EC2 boundaries, the online tool NEBcutter V2.0 was used to check each EC2 sequence for the presence of restriction sites. Forward and reverse primer sets (Table 2.10) were designed with sequences complementary to the 5’ and 3’ ends of tetraspanin EC2 domains. Each primer incorporated an appropriate endonuclease restriction site (evident from the primer name) to enable enzymatic digestion of the amplified PCR fragment before ligation into a similarly digested pGEX- KG vector. Clone Manager software was used to check that the PCR-amplified sequence would be in-frame with the GST fusion tag upstream.

2.2.2.3 PCR amplification of tetraspanin EC2 DNA PCR mix was prepared as detailed in Table 2.12 in thin-walled PCR tubes. Tetraspanin EC2 DNA sequences were amplified from the original cDNA clone vectors (Table 2.9) in a thermal cycler, using the touchdown PCR method adapted from Don et al. (1991). In this technique, the annealing temperature of the first 5 cycles is set as the averaged melting temperature of the forward and reverse primer pair (Tm). The annealing

Chapter 2 temperature is then reduced by 2°C for the next set of 5 cycles and so on and so on, for a total of 30 cycles. Higher DNA yields were obtained compared to using standard PCR.

1. Denaturation 95°C, 2 minutes 2. Denaturation 95°C, 30 seconds Annealing Tm°C, 30 seconds 5 cycles Extension 72°C, 45 seconds 3. Denaturation 95°C, 30 seconds Annealing Tm -2°C, 30 seconds 5 cycles Extension 72°C, 45 seconds

…etc... 5 cycles × 3

7. Denaturation 95°C, 30 seconds Annealing Tm -10°C, 30 seconds 5 cycles Extension 72°C, 45 seconds 8. Final extension 72°C, 10 minutes 9. Hold 4°C

PCR reaction tubes were stored at -20°C before PCR purification.

2.2.2.4 PCR product purification PCR products were separated on 1% agarose gels and purified using Wizard SV Gel and PCR Clean-Up System (Promega), following manufacturer’s protocol. Purified DNA was typically eluted into 30-50µl of sterile nuclease-free water and kept at -20˚C for long-term storage.

2.2.2.5 Restriction digests Purified PCR products were digested with appropriate pairs of restriction enzymes (Table 2.11) to produce 5’ and 3’ sticky ends. pGEX-KG vector was similarly digested to produce sticky 5’ and 3’ ends for ligation. A restriction digest mix was prepared in thin-walled PCR tubes, as detailed in Table 2.12. Contents were mixed by pipetting and centrifuged briefly. Digestion was performed by overnight incubation in a thermal cycler, at 37°C. Restriction enzymes were then inactivated by incubation at 65°C for 15 minutes. Digestion reactions were gel purified using Wizard SV Gel and PCR Clean-Up System (Promega), following manufacturer’s protocol.

Chapter 2

2.2.2.6 Ligation of amplified EC2 fragments into pGEX-KG vector Following digestion and gel purification, PCR fragments were ligated into digested pGEX-KG vectors using the Quick-Stick DNA Ligation kit (Bioline). Ligation mixes were prepared as detailed in Table 2.12, containing a 3:1 molar ratio of insert:vector, according to the following equation:

( ) ( ) × molar ratio of = ng of insert ( )

Typically, 100ng of vector was used per reaction. Ligations were carried out at 25°C for three hours, according to manufacturer’s instructions.

2.2.2.7 Amplification of ligated vectors 20µl DH5α competent cells (Library Efficiency, Invitrogen) were thawed on ice. 1µl of the completed ligation reaction was added to the cells and gently mixed by tapping the tube. Cells were incubated on ice for 30 minutes and heat-shocked by incubation at 42˚C for 45 seconds. Tubes were re-incubated on ice for 2 minutes then 380µl SOC medium was added. Cells were incubated for 1 hour at 37˚C (225rpm shaking) and 50-100µl aliquots were plated on LB agar supplemented with 100µl carbenicillin. Following overnight incubation at 37˚C, six single colonies were picked into individual 10ml LB cultures containing carbenicillin (100µg/ml) and grown overnight at 37°C. 750µl of the overnight culture was added to 250µl 60% glycerol and flash frozen in liquid nitrogen for 20 minutes before long-term storage in a -80°C freezer. The rest of the culture was pelleted by centrifugation and plasmid DNA was purified as described earlier (section 2.2.1.2). Purified plasmids were digested with restriction enzymes to confirm presence of the amplified PCR fragment and sequenced (section 2.2.1.1) to ensure that the EC2 domains were in-frame with the GST tag upstream and no mutations were present.

Chapter 2

2.2.3 Recombinant protein expression Protein expression was carried out in two E. coli strains, E. coli Rosetta-gami B (DE3) pLysS (Novagen) and E. coli SHuffle T7 Express lysY (New England Biolabs).

2.2.3.1 Expression in E. coli Rosetta-gami strain Transformation was carried out as instructed by the manufacturer, with minor modifications. Briefly, 10µl aliquots of competent cells were thawed on ice for 5 minutes. 1µl tetraspanin expression plasmid (1ng/µl stock) was added to the cells, mixed by gentle tapping and incubated on ice for 5 minutes. Cells were heat shocked by incubation at 42˚C for 30 seconds in a water bath and transferred immediately back onto the ice. After 2 minutes, 190µl SOC medium was added and cells were grown out for an hour at 37˚C with shaking (250rpm). 25-100 µl of cells were plated on LB agar plates supplemented with the relevant selection antibiotics (Table 2.8) and grown overnight at 37˚C. Single colonies from overnight plates were inoculated into 5ml LB cultures (+antibiotics) and grown overnight at 37˚C with shaking (250rpm). 1ml aliquots of the overnight cultures were inoculated into 500ml LB (+antibiotics) cultures and grown overnight at 37˚C with shaking (100rpm). Following overnight culture, the absorbance (OD600) of bacterial cultures was measured and cultures were diluted with fresh LB, if necessary, to reach an OD600 of 0.6-0.8. Recombinant protein expression was induced by incubation with 0.1mM IPTG for 4 hours at 37˚C with shaking (250rpm). Bacteria were pelleted by centrifugation at 4,500g for 20 minutes at 4˚C. Supernatant was decanted and pellets were kept at -80˚C.

2.2.3.2 Expression in E. coli SHuffle strain Competent cells were transformed as instructed by the manufacturer, with minor modifications. Briefly, 10µl aliquots of cells were defrosted for 10 minutes on ice. 1µl tetraspanin expression plasmid (1ng/µl stock) was added to the cells, mixed by gentle tapping and incubated on ice for 30 minutes. Cells were heat shocked by incubation at 42˚C for 30 seconds in a water bath and transferred immediately back onto the ice. After 5 minutes, 190µl SOC medium was added and cells were grown out for 60 minutes at 30˚C with shaking (250rpm). 50-100µl of cells were plated on warmed selection plates containing the appropriate antibiotics (Table 2.8). Following overnight incubation at 30˚C, single colony transformants were picked into 5ml LB cultures (+antibiotics) and grown overnight at 30˚C (250rpm). 1ml of the overnight culture

Chapter 2 was inoculated into 500ml LB supplemented with the appropriate antibiotics and grown overnight at 30˚C with shaking (100rpm). Following overnight culture, the absorbance (OD600) of bacterial cultures was measured and cultures were diluted with fresh LB, if necessary, to reach an OD600 of 0.6-0.8. Recombinant protein expression was induced by incubation with 0.1mM IPTG for 4 hours at 30˚C with shaking (300rpm). Bacteria were pelleted by centrifugation at 4,500g for 20 minutes at 4˚C. Supernatant was decanted and pellets were kept at -80˚C.

In an attempt to increase the yield of soluble recombinant protein obtained from these cultures, various expression parameters such as IPTG concentration and temperature were altered as described in detail in Chapter 3.

2.2.4 Bacterial cell lysis by sonication Frozen pellets were defrosted on ice and gently re-suspended in 1× ice-cold PBS buffer containing 1in100 v/v dilution of Halt Protease Inhibitor Cocktail (Thermoscientific). Typically, 5-10ml 1×PBS buffer was used for every gram bacterial pellet. The cell suspension was split into two or more chilled universal tubes and each tube was sonicated in turn to prevent overheating of the sonication probe. Tubes were placed in a beaker containing 50% ice-water. Each tube was sonicated at an amplitude of 15microns for two 10 second bursts and this was repeated for 5-6 rounds. Lysed cells were centrifuged for 20 minutes at 24,000g and 4⁰C to pellet insoluble material and the supernatant was removed to a fresh chilled Falcon tube. A small sample was retained for future analysis. Recombinant protein was purified immediately or alternatively tubes were stored for a few days at -80⁰C until purification.

2.2.5 Affinity purification Recombinant protein was purified using glutathione-sepharose beads (G4B, GE Healthcare). Beads were gently re-suspended and the required volume of bead slurry (20% ethanol) was removed to an ice-cold 15ml Falcon tube. All centrifugation steps were carried out at 500g and 4˚C for 5 minutes. Beads were pelleted by centrifugation and the supernatant was discarded. Beads were then washed by re-suspending in 5 bead volumes of ice-cold PBS buffer and centrifuging. This step was repeated twice more. After the final step, a small volume of ice-cold PBS was added to the beads to allow re-suspension and dispensing. Approximately 100µl beads slurry (20%) was used for every gram of original bacterial pellet. Beads were added to defrosted

Chapter 2 bacterial lysates and incubated for 2 hours with gentle rotation on a blood mixer at room temperature. Supernantant (=flowthrough) was obtained by centrifugation and a small sample was retained for future analysis. Beads were then washed three times by re-suspending in 20 bead volumes ice-cold PBS and centrifuging (small samples of the supernatant were retained for analysis). Finally, beads were re-suspended in 25mM glutathione elution buffer (Table 2.4) and incubated at room temperature with gentle rotation on a blood mixer to elute the recombinant protein. Approximately 0.5- 1ml elution buffer was used for every gram of original bacterial pellet. Beads were pelleted by centrifugation and supernatant (=first elution) was removed to a fresh ice- cold tube. Beads were re-suspended once more in elution buffer (the volume of buffer used was half that used in the first elution) and incubated as before to elute any remaining recombinant protein. Beads were pelleted and supernatant (=second elution) was removed to a fresh ice-cold microfuge tube. The protein concentration in the first and second elution was determined visually by Bradford assay and if both samples contained the same protein concentration, the two samples were mixed. Samples were kept at 4˚C before dialysis.

2.2.6 Dialysis To avoid a freeze-thaw cycle, recombinant proteins were dialysed quickly after elution. Dialysis membranes (Medicell International Ltd Size 2, molecular weight cut- off 12-14KDa) were boiled for 10 minutes in a 2L solution of sodium bicarbonate/EDTA/water (20g/0.37g/1L). Membranes were rinsed with copious amounts of ultrapure water and boiled in a 2L solution of EDTA/water (0.37g/1L) and boiled for a further 10 minutes. The solution was cooled and membranes were transferred to a 70% ethanol solution for long-term storage at 4˚C. Immediately before dialysis, membranes were rinsed with copious amounts of ultrapure water and equilibrated in fresh ice-cold 1×PBS. To dialyse the protein, membranes were tied at one end and the protein sample was pipetted in. Membranes were then carefully tied at the other end and excess air was expelled. Tied membranes were put into a flask containing 5L fresh ice-cold 1×PBS. The flask was placed on a magnetic stirrer and incubated at 4˚C overnight. Membranes were then transferred to another flask with fresh 1×PBS and incubated for 4 hours at 4˚C with stirring. The 1×PBS solution was changed once more and membranes incubated for 4 hours at 4˚C with stirring. Finally,

Chapter 2 dialysed protein samples were aliquoted into 0.5ml chilled microfuge tubes and stored at -80°C. Small samples were removed for analysis by Bradford assay and SDS-PAGE.

2.2.7 LPS reduction using Triton X-144 Expression of GST R and GST-CD9 R was induced in E. coli Rosetta-gami strain, as detailed in section 2.2.3. Protein purification was commenced as before (section 2.2.4), with the following modifications. Crude cell lysate from each culture was divided into two equal aliquots and each aliquot was incubated with GST-sepharose beads for two hours. Beads were pelleted by centrifugation and the supernatant (=flowthrough) was discarded. At this point, one of the two aliquots was subjected to the LPS reduction technique outlined in Reichelt et al. (2006). Briefly, the tube with beads was equilibrated on ice for 5 minutes then incubated with 50 bead volumes of ice-cold 0.1% v/v Triton X-114/PBS at 4˚C for 2 minutes with end-to-end rotation. The other aliquot was treated the same way but incubated with PBS lacking detergent. Beads were pelleted by centrifugation at 500g for 5 minutes at 4˚C and washed with 20 bead volumes of PBS for three rounds, as before. Recombinant GST-tagged protein was then eluted and dialysed as previously.

Chapter 2

2.3 Tissue culture methods 2.3.1 Multinucleated Giant Cell (MGC) formation in peripheral blood monocytes 2.3.1.1 Isolation of peripheral blood mononuclear cells 30-400ml blood samples were obtained from anonymised ‘normal’ human volunteers, in the University of Sheffield Medical School OHPAT unit, in accordance with the South Sheffield Research Ethics Committee (protocol number SSREC/02/299). Peripheral blood was collected into sterile heparinised blood bags (Ask Ian Geary). Blood was diluted 1:1 in sterile HBSS (Gibco) and 30ml was carefully layered onto 15ml of room temperature Histopaque-1077 (Sigma, H8889), in a 50ml Falcon tube. Mononuclear cells were separated by gradient density centrifugation at 400g (1350rpm) in a swing- bucket rotor for 30 minutes with the brake disengaged. Using this technique, four distinct layers are obtained as below.

Platelet-rich plasma

Mononuclear fraction

Histopaque

Granulocytes and erythrocytes

Figure 2.1 Separation of peripheral blood mononuclear cells by density gradient centrifugation. The opaque mononuclear fraction contains monocytes as well as T and B lymphocytes.

The mononuclear fraction was carefully aspirated into a new Falcon tube using a plastic Pasteur pipette. Fractions from two columns were combined and topped up to 50ml with room temperature HBSS. The tubes were centrifuged at 200g for 15 minutes. Supernatant was carefully discarded and cell pellets were thoroughly re- suspended by flicking the tube. The tube was topped up with 50ml room temperature Histopaque and centrifuged at 200g for 10 minutes. Supernatant was discarded and cells were re-suspended by flicking before addition of 10-20ml pre-warmed RPMI- 1640 (10% FCS, 1% penicillin/streptomycin). A 50µl sample of the cell suspension

Chapter 2 was removed and mixed with 50µl 5% acetic acid to lyse contaminating erythrocytes. Cells were then counted on a haemocytometer.

2.3.1.2 Concanavalin A-induced MGC formation assay After counting, cells were diluted in pre-warmed RPMI-1640 (10% FCS/1% penicillin- streptomycin) and plated at a density of 0.75×106 cells/well into each well of a 96-well plate (Corning Cellbind). The plates were incubated at 37˚C/5% CO2 for 2 to 16 hours to allow monocytes to adhere to the plastic. Non-adherent cells were removed and 200µl fresh pre-warmed RPMI-1640 (1% penicillin-streptomycin/no FCS) was added to each well. Wells were washed by pipetting the medium up and down against the edge of the well to create a swirling motion. Using this technique, the adherent cell density in the centre of the well is uniformly reduced whilst at the edges, only a low cell density remains. The media was then removed and fresh media added to each well. The wells were inspected to see if the desired cell density had been reached. If cell density was still too high, a second wash step was performed as previously.

Washing to Medium cell density in the High cell density create a swirling middle of the well/low density across the well motion at the periphery

Figure 2.2 Reducing adherent cell density

To induce MGC formation, media was removed and wells were incubated with 150µl RPMI-1640 (10% FCS, 1% Penicillin-Streptomycin) containing 10µg/ml Concanavalin A (Con A, Sigma C5275). To test the effect of various reagents on Con A-induced MGC formation, the test reagent was diluted in RPMI-1640 (10% FCS, 1% Penicillin- Streptomycin) to give a 2× concentration of test reagent. This solution was then diluted 1:1 with RPMI-1640 (10% FCS, 1% Penicillin-Streptomycin) containing 20µg/ml Con A thus giving a solution of 10µg/ml Con A and 1× test reagent. Following

72 hours incubation at 37⁰C, 5% CO2, media was removed and pre-warmed HBSS was gently added to each well. Wells were inspected under a light microscope to see if

Chapter 2 there were a lot of floating cells and thus one more wash was needed. HBSS was then removed and cells were fixed by addition of 50µl/well paraformaldehyde (4% stock solution). After 5 minutes incubation at room temperature, paraformaldehyde was removed and wells were washed three times with HBSS. 200µl solution of sodium azide (0.1% in HBSS) was added to each well and plates were kept in foil at 4⁰C before staining.

2.3.1.3 Immunofluorescence cell staining Staining was carried out at room temperature. To stain the plasma membrane, wells were incubated with 50µl TRA-1-85 primary antibody (diluted 1in5 in BBN buffer) for 30 minutes. Wells were washed three times by removing and adding fresh PBS. Cells were then stained with 50µl/well FITC-labelled secondary antibody (F1010, diluted 1in250 in BBN) for 30 min and washed three times with PBS. Nuclei were counterstained with 50µl Hoescht 33342 (1µg/ml diluted in BBN) for 5-20 min followed by PBS washes as before. Finally, 200µl PBS containing 0.1% sodium azide was added to each well, plates were sealed with a plate sealer, wrapped in foil and stored at 4˚C.

2.3.1.4 Immunofluorescence imaging Plates were scanned using an ImageXpress Micro high content screening system. Using this automated system, the same area is imaged in each well eliminating the possibility of experimenter bias in choosing fields to image or count. 25 fields were imaged per well as shown in figure 2.3 using a 20× objective. For each field, two images were obtained using the blue (Hoescht) channel and the green (FITC) channel.

Chapter 2

Figure 2.3 Image acquisition Images are acquired from the grey area in each well which has a medium cell density optimal for MGC formation (figure 2.2).

2.3.1.5 Quantifying MGC formation The two images obtained for each field were merged and subsequently analysed using Irfanview softare. For each field, the total number of nuclei, the total number of nuclei in MGCs and the total number of MGCs were determined. An MGC was defined as a cell having 3 or more nuclei since two nuclei within a cell may result from nuclear division without cytokinesis. Data from all 25 fields in a well were combined before proceeding. For each well, Fusion index and Average MGC size was determined according to the following formulae (Parthasarathy et al., 2009):

number of nuclei in Cs Fusion index (%) = × 100 total number of nuclei

total number of nuclei in Cs Average MGC size = total number of MGCs

2.3.1.6 Statistical analysis The remaining analyses were carried out using built-in analysis tools in GraphPad Prism6 software. To reduce the effects of donor variability, for each donor, data were normalised to the Fusion index and Average MGC size obtained for the Con A control, which was assumed to be 100%. Before testing for statistical significance, normality tests were performed to determine whether the data followed a normal (Gaussian) distribution. If the data was normally distributed, a parametric Analysis of variance (ANOVA) test was performed followed by a post-test to compare individual data sets to each other or to the control. Otherwise, a non-parametric test was used with an appropriate post-test as recommended by GraphPad Prism6. This software was also used to display data graphically.

Chapter 2

2.3.1.7 Osteoclastogenesis assay PBMC were isolated as described, diluted in MEM-α media (containing Glutamax, 10% FCS, 1% penicillin-streptomycin) and plated at a cell density of 1×106 cells per well of a 96 well plate. Cells were incubated at 37˚C, 5% CO2 for 1-3 hours to allow monocyte to adhere. Following adhesion, cells were washed with pre-warmed un-supplemented MEM-α media, as described in section 2.3.1.2. Cells were then incubated with supplemented MEM-α media containing 25ng/ml RANKL (Peprotech) and 30ng/ml MCSF (Peprotech). The medium was replaced completely with fresh media containing RANKL/MCSF every 2-3 days. Following 14-21 days culture, cells were fixed with 4% w/v paraformaldehyde (in PBS) and stained using a TRAP staining kit (Sigma) as instructed by the manufacturer. Cell membranes and nuclei were additionally stained using TRA-1-85 antibody and Hoescht 33342 stain as described previously (section 2.3.1.3). Plates were imaged with the InCell Analyzer 1000 microscopy system using the bright-field (TRAP staining), DAPI (blue, nuclei) and GFP (green, cell membranes) channels and a 20× objective lens. Osteoclasts were defined as TRAP-stained cells containing 3 or more nuclei. Fusion indices were determined as for MGC formation (section 2.3.1.5).

2.3.2 Burkholderia thailandensis-induced MGC formation in J774.2 macrophage- like cells 2.3.2.1 Passage of adherent J774.2 cells J774.2 cells (murine macrophage cell line) were cultured in DMEM media (Gibco) containing 2mM glutamine with 10% FCS in a humidified atmosphere at 37˚C, 8% CO2. Cells were sub-cultured by discarding the medium, adding 10ml fresh medium and gently scraping the cells. The cell suspension was then diluted as necessary and transferred to new culture flasks.

2.3.2.3 Burkholderia thailandensis-induced MGC formation assay J774.2 cells were infected with Burkholderia thailandensis strain E264 as described previously (Suparak et al., 2005). 2×105 cells were seeded in wells of a 96-well plate and incubated overnight at 37˚C, 8% CO2. To test the effects of antibodies, recombinant proteins or other reagents on B. thailandensis-induced MGC formation, the test reagent was incubated with the cells for 1-16 hours at 37˚C, 8% CO2. Wells were then washed

Chapter 2 twice with PBS and a suspension of B. thailandensis cells was added to the cells at a multiplicity of infection (MOI) of 3:1. Following 2 hours incubation at 37˚C, 8% CO2, cells were washed with PBS and 500mg/ml kanamycin was added to cells and incubated for 2 hours to kill extracellular bacteria. After a further PBS wash, cells were fixed by incubation with 100% ethanol for 30 minutes and stained with Giemsa stain for an additional 30 minutes at room temperature. MGC formation was quantified as for Con A-induced MGC formation (section 2.1.3.5). Nuclei were counted manually in five random fields per well using the 40× objective on an Olympus CK40 bright-field microscope.

2.3.3 Bacterial adhesion assay An in vitro bacterial adhesion assay was carried out as described previously (Green et al., 2011). HEC-1-B cells (a human epithelial endometrium adenocarcinoma cell line)

5 were grown in EMEM media (Lonza) containing 10% FCS at 37˚C, 5% CO2. 1.5×10 cells were seeded onto coverslips in 24 well tissue culture plates and treated with 5% bovine serum albumin (BSA) for 30 minutes to block non-specific binding. Cells were washed twice with PBS and incubated with 500nM recombinant protein or other test reagents for 30 minutes at 37°C. Cells were then washed twice with PBS and incubated with bacteria at a multiplicity of infection (MOI) of 300 for 1 hour. A further two washes were performed to remove non-adherent bacteria and cells were fixed by incubation with 2% paraformaldehyde for 10 minutes. Cellular and bacterial DNA was stained with DAPI and the extracellular bacteria were additionally stained with an anti-Neisseria meningiditis antibody followed by an anti-mouse IgG FITC-labelled antibody. Bacterial adhesion is a pre-requisite for internalisation therefore the adhesion rate was determined as the number of internalised and adherent bacteria per 100 cells.

Chapter 3

Chapter 3: Production of recombinant tetraspanin EC2 proteins

3.1 Introduction Tetraspanins constitute a large superfamily of transmembrane proteins, widely expressed on almost all animal cells (Boucheix and Rubinstein, 2001). However, most family members remain essentially uncharacterised. The importance of the large extracellular domain (EC2) to tetraspanin function has already been highlighted (section 1.1.1.2). Currently monoclonal antibodies to native tetraspanins are unavailable for the majority of tetraspanins, due to poor antigenicity in mice (Hemler et al., 2005). Recombinant tetraspanin EC2 domains have been used as alternative tools to investigate tetraspanin functions in diverse processes (Table 1.2, section 1.3). The majority of these studies have utilised the GST (glutathione-s-transferase) gene fusion system (figure 3.1) to express and purify recombinant tetraspanin EC2 domains in E. coli (for example, Ho et al., 2006).

Figure 3.1 The pGEX-KG vector used for recombinant tetraspanin expression.

pGEX-KG vector features an IPTG-inducible promoter, from which recombinant protein overexpression can be induced. The plasmid encodes glutathione-s-transferase (GST) from the pathogenic eukaryote Schistosoma japonicum to facilitate glutathione-affinity purification. Downstream of GST is a glycine-rich linker (PGISGGGGG) followed by a thrombin cleavage sequence (LVPRGS) to enable protease cleavage of the fusion protein (Guan and Dixon, 1991).

The Ampicillin-resistance gene (Ampr) allows selection on Ampicillin media. Abbreviations: IPTG, Isopropyl- β-D-1-thiogalactopyranoside, bp, base pairs.

Chapter 3

The GST affinity tag is thought to enhance recombinant protein solubility and enables a relatively simple purification procedure using glutathione affinity resin (Guan and Dixon, 1991). In addition, GST-tagged EC2 domains can be used in pull-down studies to study tetraspanin/partner protein interactions (Ding et al., 2005). The aim of this chapter was to produce nine relatively understudied tetraspanins as recombinant N- terminal-tagged GST-EC2 fusion proteins in E. coli and to characterise their activity in different functional assays including multinucleated giant cell formation (Takeda et al., 2003; Parthasarathy et al., 2009) and bacterial adhesion to epithelial cells (Green et al., 2011).

3.1.1 Expression of disulfide bond-containing proteins in E. coli A potential issue with heterologous expression of recombinant eukaryotic proteins in bacteria is that, under normal physiological conditions, the reducing environment of the bacterial cytoplasm is not conducive to stable disulfide bond formation (Bessette et al., 1999; Baneyx and Mujicic, 2004). Table 3.1 shows the number of cysteines and putative disulfide bonds in GST and each of the tetraspanin EC2 domains which were selected for expression in this study. Several studies have demonstrated the importance of conserved cysteine residues to the proper folding and in vitro activity of tetraspanin EC2 domains (section 1.1.1.2). For example, mutagenesis of single or multiple cysteine residues in recombinant human tetraspanin CD81EC2 compromised its interaction with its protein partner, the E2 glycoprotein of Hepatitis C virus (HCV) (Higginbottom et al., 2000; Kitadokoro et al., 2001; Petracca et al., 2000).

Chapter 3

Number of putative Number of cysteines disulfide bonds GST 4 None *

Tspan2EC2 4 2

Tspan3EC2 4 2

Tspan4EC2 6 3

Tspan5EC2 8 4

Tspan8EC2 6 3

Tspan13EC2 4 2

CD82EC2 6 3

Tspan31EC2 6 3 Tspan32EC2 4 2

Table 3.1 Cysteines and disulfide bonds in GST and tetraspanin EC2 domains

*An X-ray crystallographic structure of the Schistosoma japonicum GST (also called Sj26) has shown that the protein has four solvent-exposed cysteines which do not take part in disulfide bond formation (Lim et al., 1994; Tudyka and Skerra, 1997).

According to the literature, protein mis-folding and aggregation is frequently encountered in the production of recombinant proteins containing 3 or more disulfide bonds in bacteria (Qiu et al., 1998). Five out of the nine recombinant tetraspanins that we wished to express potentially contain three or more disulfide bonds. Protein aggregates formed in E. coli (inclusion bodies) are insoluble and may be biologically inactive (Guan and Dixon, 1991; Sorensen and Mortensen, 2005). Therefore expression strain selection was particularly important to avoid recombinant protein misfolding and aggregation.

Disulfide bond formation in gram-negative bacteria is catalysed by the thioredoxin and glutaredoxin enzyme superfamilies (Lobstein et al., 2012; Baneyx and Mujicic, 2004). Under normal physiological conditions, the E. coli cytoplasmic thioredoxins TrxA and TrxC and glutaredoxins GrxA, GrxB and GrxC are maintained in a reduced state by the action of thioredoxin TrxB and glutaredoxin Gor. Thus these enzymes cannot catalyse disulfide bond formation and instead reduce disulfide bridges in other proteins, becoming oxidised themselves in the process (Bessette et al., 1999; Baneyx and Mujicic, 2004). The E. coli Origami strain and its derivative Rosetta-gami (Novagen) are trxB/gor null mutants with an enhanced capacity for cytoplasmic disulfide bond

Chapter 3 formation (Bar et al., 2006; Lobstein et al., 2012). Rosetta-gami strain has been used previously to produce recombinant tetraspanin EC2 proteins such as CD81EC2 from various species (Flint et al., 2006; Bankwitz et al., 2010), the sm-TSP-2 antigen of Schistosoma masoni (Pearson et al., 2012) and CD9, CD81 and CD63 (R. Hulme, PhD thesis). Therefore, as a starting point, this strain was to be used for tetraspanin overexpression.

In wild-type E. coli strains, disulfide bond formation takes place not in the cytoplasm but in the periplasmic space, catalysed by perisplasmic thioredoxin enzymes. DsbA catalyses disulfide bond formation between consecutive cysteines in polypeptides, as they are extruded into the periplasmic space (Bardwell et al., 1991). For disulfide bond formation between cysteine residues that are non-consecutive in the protein amino acid sequence (such as tetraspanin EC2 domain cysteines, see figure 1.2, Chapter 1) the activity of disulfide bond isomerase DsbC is further required to obtain the native conformation (Lobstein et al., 2012). Overexpression of ΔssDsbC (a DsbC mutant missing the periplasmic signal sequence) enhanced the yield of soluble, correctly folded and therefore functional cysteine-containing recombinant proteins (Bessette et al., 1999; Jurado et al., 2002; Lobstein et al., 2012). DsbC also acts as a chaperone to assist folding of non-cysteine-containing proteins (Tait and Strauss, 2011).

E. coli SHuffle strain (New England Biolabs) combines the desirable features of E. coli Rosetta-gami (i.e. the trx/B gor null mutations) with a chromosomal copy of ΔssDsbC (Lobstein et al., 2012). Recent work by our group showed that higher yields of soluble recombinant GST-CD9EC2 could be obtained using SHuffle strain compared to Rosetta-gami (Mr John Palmer, Final year project report, Department of Molecular Biology and Biotechnology, University of Sheffield). Therefore, this strain was also used in an attempt to obtain more soluble recombinant protein for functional studies.

Chapter 3

3.2 Results 3.2.1 Subcloning of tetraspanin EC2 domain cDNA into pGEX-KG vector Forward and reverse oligonucleotide primer sets encoding sequences complementary to the 5’ and 3’ ends of tetraspanin EC2 regions were designed as described in section 2.2.2.2. Each primer also incorporated an endonuclease restriction site to facilitate integration into the multiple cloning site of pGEX-KG vector. Tetraspanin EC2 domain DNA sequences (~200-400bp) were amplified from full-length human cDNA clones by polymerase chain reaction (PCR) (section 2.2.2.3). PCR products were analysed by agarose gel electrophoresis as shown in figure 3.2.

M 1 2 3 4 5 6 7 8 M

600bp 400bp

200bp

Figure 3.2 Amplification of eight tetraspanin EC2 DNA sequences by PCR.

Tetraspanin EC2 DNA was amplified from cDNA clone vectors by the touch-down PCR method (Don et al., 1991) in a thermal cycler (Chapter 2, section 2.2.2.3. 5µl of the completed PCR reaction was mixed with 1µl 6× DNA loading buffer and loaded onto a 1% agarose gel. 5µl DNA marker (Hyperladder I, Bioline) was also loaded. M: Marker, 1: Tspan3EC2 [277bp], 2: Tspan4EC2 [313bp], 3: Tspan13EC2 [221bp], 4: Tspan25EC2 [254bp], 5: Tspan26EC2 [426bp], 6: CD82EC2 [384bp], 7: Tspan31EC2 [260bp], 8: Tspan32EC2 [281bp]. The predicted molecular weights of the amplified fragments are shown in brackets, taking into account the restriction sites and other bases in the forward and reverse primers. PCR=polymerase chain reaction.

The amplified fragments (lanes 2-8) all appeared to be of correct size relative to each other as well as to the DNA marker. Tspan3EC2 (lane 1) appeared to slightly larger than predicted relative to the other tetraspanin EC2 domains. PCR products were then gel-purified and digested with restriction endonucleases to produce sticky ends for ligation into digested pGEX-KG vectors (~5000bp). Purified PCR products were sub- cloned into pGEX-KG vectors in-frame with an N-terminal GST affinity tag. To amplify the resulting recombinant vectors, competent DH5α cells were transformed with the products of the ligation reaction and grown on carbenicillin media to select

Chapter 3 transformants. Plasmid DNA was prepared from selected single colony E. coli cultures as described in section 2.2.1.2. Analytical restriction digests were performed to confirm presence of the cloned fragment. Similar results to those shown in figure 3.3 were obtained for pGEX-Tspan3EC2, pGEX-Tspan4EC2 and pGEX-Tspan32EC2. Cloning of Tspan26EC2 into pGEX-KG vector proved unsuccessful despite repeated attempts. pGEX-CD9EC2, pGEX-Tspan5EC2 and pGEX-Tspan8EC2 had been previously cloned in-house whilst pGEX-Tspan2EC2 was cloned by Mutagenex. DNA sequencing was carried out on all vectors to ensure no mutations were present.

M 1 2 3 4 5 6 7 8 9 10 11 12 M pGEX-KG 5000bp plasmid

600bp Excised EC2 200bp fragment

Figure 3.3 Analytical restriction endonuclease digest of pGEX-KG plasmids carrying EC2 DNA fragments.

Amplified plasmids were digested with appropriate pairs of restriction endonucleases to excise the EC2 fragment out of the vector (Chapter 2, section 2.2.2.5). Approximately 150ng plasmid was used in each digest and a 5µl sample of the reaction was loaded onto a 1% agarose gel. 7µl of DNA marker was also loaded. M: Marker (Hyperladder I, Bioline), 1-3: pGEX-Tspan13EC2 [221bp], 4-6: pGEX-Tspan25EC2 [254bp], 7-9: pGEX-CD82EC2 [384bp], 10-12: pGEX-Tspan31EC2 [260bp]. The predicted size of the excised EC2 domain is shown in brackets.

Chapter 3

3.2.2 Expression and purification of recombinant tetraspanin proteins in E. coli Rosetta-gami strain 3.2.2.1 Recombinant tetraspanin expression E. coli Rosetta-gami competent cells were transformed with plasmids encoding N- terminal GST-tagged EC2 domains of tetraspanins Tspan5, Tspan13, Tspan27 and Tspan31. Transformants were selected on media containing appropriate antibiotics and single colonies were used to seed starter cultures which were scaled up to 500ml (section 2.2.3). Recombinant protein expression was induced using IPTG (Isopropyl- β-D-1-thiogalactopyranoside). To follow the induction process, 1ml aliquots were removed from the culture at 0, 1, 2, 3 and 4 hours post-induction. Cells were harvested by centrifugation at 12,100g for 1 minute. Pellets were re-suspended in 100µl 1× reducing SDS-PAGE (Sodium dodecyl sulphate-polyacrylamide gel electrophoresis) gel loading buffer to lyse the cells and boiled at 100˚C for 3 minutes followed by centrifugation. Supernatants were analysed by SDS-PAGE and Western blotting.

As shown in figure 3.4, no GST-tagged material was detected before induction (time zero) whilst increasing levels were detected between 1 and 4 hours after induction suggesting that recombinant protein has been successfully induced from the IPTG- inducible promoter. Red arrows point to bands of the apparent correct size to be full- length GST-Tspan13EC2 (34.2KDa), GST-CD82EC2 (40.9KDa) and GST-Tspan31EC2 (35.5KDa). Lower molecular weight bands are also detected in the immunoblots and may be the result of recombinant protein degradation by bacterial proteases or premature termination of protein synthesis, as observed when other GST-tagged tetraspanin EC2 domains are purified (Higginbottom et al., 2003; Ding et al., 2005; Flint et al., 2006). The predicted molecular weight of GST-Tspan5EC2 is 40.6KDa but no band is apparent at this molecular weight in figure 3.4 A. Instead a diffuse staining pattern is observed which may be due to the presence of several mis-folded GST- Tspan5EC2 species with differential mobility on polyacrylamide gels, various degradation products of full-length GST-Tspan5EC2 or premature termination of recombinant protein synthesis. All samples were reduced and alkylated to prevent re- oxidation during SDS-PAGE (section 2.2.1.5) ruling out the first possibility.

Chapter 3

A) GST-Tspan5EC2 (40.6KDa) B) GST-Tspan13EC2 (34.2KDa)

Hours post-induction Hours post-induction

KDa M 0 1 2 3 4 KDa M 0 1 2 3 4 5 5 75 75 50 50

37 37

25 25 20 20

15 15

37 37

25 25

C) GST-CD82EC2 (40.9KDa) D) GST-Tspan31EC2 (35.5KDa)

Hours post-induction Hours post-induction KDa M 0 1 2 3 4 KDa M 0 1 2 3 4 5 5

75 75 50 50

37 37

25 25 20 20

15 15

37 37

25 25

Figure 3.4 SDS-PAGE analysis of recombinant GST-tagged tetraspanin protein expression.

10µl samples of lysed bacterial pellets obtained at various time points (see text) were separated

on a 12.5% SDS-PAGE gel. Total protein was detected by Coomassie staining (top image) whilst GST-tagged material was detected by Western blotting (bottom image) using an anti-GST antibody as described (sections 2.2.1.5 to 2.2.1.8). For Western blotting, samples were diluted 1in10. Blots were probed with anti-GST HRP-linked antibody. 7µl of marker (lane M, All blue Precision Plus, Biorad) was also loaded on each gel. Red arrows point to bands likely to be full-

length GST-tetraspaninEC2 proteins. See text for detailed explanation. Free GST has a molecular

3 weight of ~26.2KDa (Guan and Dixon, 1991). KDa=Kilodaltons, HRP=Horse radish peroxidase.

Chapter 3

3.2.2.2 Recombinant tetraspanin purification using glutathione affinity chromatography Recombinant protein expression was induced in 500ml bacterial cultures as described above. Following a 4 hour induction, cells were harvested by centrifugation and lysed by sonication. Lysates were centrifuged to pellet insoluble material and the supernatant was incubated with glutathione-sepharose beads to allow GST-tagged material to bind. Beads were pelleted by centrifugation and the supernatant (=flowthrough) was discarded. Beads were then washed three times with lysis buffer to remove contaminating proteins (=wash 1, 2 and 3). Retained proteins were eluted from the beads by incubation with 25mM reduced glutathione (=eluted material) and subsequently dialysed against 1×PBS buffer to remove free glutathione. A sample was kept at each purification stage for analysis by SDS-PAGE and Western blotting (figure 3.5).

Total protein concentration in the eluted material was determined by the Bradford assay (Chapter 2, section 2.2.1.9) and the proportion of full-length recombinant protein was determined from Coomassie-stained gels using densitometry analysis using ImageJ software. The concentration of purified GST-Tspan13EC2 (figure 3.5B, lane 6, red arrow) was determined to be 1mg/ml whilst for the other three recombinant proteins, total protein concentration was <100µg/ml. Despite evidence that recombinant protein expression was successfully induced in transformed bacteria (figure 3.4), little GST-Tspan5EC2, GST-CD82EC2 and GST-Tspan31EC2 was present in the soluble fraction of crude cell lysate (figure 3.5 A, C and D, lane 1). This is likely due to the way the samples were obtained. Figure 3.4 samples were obtained by re- suspending bacterial pellets in reducing SDS-PAGE sample buffer containing SDS to lyse the cells. SDS (sodium dodecyl sulphate) is an effective denaturant that can be used to extract recombinant proteins from insoluble protein aggregates (Singh and Panda, 2005). Therefore, the bands detected in figure 3.4 immunoblots represent ‘total’ (soluble and insoluble) recombinant protein whilst figure 3.5 samples represent only soluble recombinant protein.

Chapter 3

A) GST-Tspan5EC2 (40.9KDa) B) GST-Tspan13EC2 (34.2KDa)

KDa M 1 2 3 4 5 6 KDa M 1 2 3 4 5 6

75 75

50 50

37 37

25 25 20 20

37 37

25 25

C) GST-CD82EC2 (40.9KDa) D) GST-Tspan31EC2 (35.5KDa)

KDa M 1 2 3 4 5 6 KDa M 1 2 3 4 5 6

75 75 50 50

37 37

25 25 20 20

37 37

25 25

Figure 3.5 Purification of recombinant tetraspanins using glutathione affinity

chromatography.

Small aliquots were kept at each stage of purification and analysed by SDS-PAGE followed by Coomassie staining (top gels in A, B, C and D) or Western blotting (bottom gels in A, B, C and D).

10µl of each sample was boiled with 4× reducing SDS loading buffer at 95˚C for 5 minutes and loaded on the gel. 7µl of marker was also loaded. For Western blotting, samples were diluted 1in10 with 1× SDS loading buffer. M: Marker (All blue Precision Plus, Biorad), 1: soluble fraction of whole cell lysate, 2: flow-through, 3: first wash, 4: second wash, 5: third wash, 6: eluted material. Red arrows point to bands likely to be full-length material.

Figure 3.6 shows a direct comparison of the amount of recombinant protein present in the soluble and insoluble fractions of crude cell lysates. For all three tetraspanin constructs, most of the recombinant protein resides in the insoluble fraction. In addition, GST-Tspan5EC2 migrates as a lower molecular weight protein (~36KDa) than predicted (~40.6KDa) in both fractions. This indicates that the protein may be

Chapter 3 truncated but this must be confirmed by peptide mass spectrometry. It is notable that recombinant GST-CD9EC2 also migrates as a slightly lower molecular weight protein (figure 3.12, immunoblot B) but it is reactive with a conformation-sensitive antibody and active in functional assays (Chapter 4, figure 4.4; Chapter 5, figure 5.1).

KDa Ins Sol Ins Sol Ins Sol

50 37

GST-Tspan5 GST-CD82 GST-Tspan31 EC2 EC2 EC2

Figure 3.6 Detection of soluble and insoluble GST-tagged material in whole cell lysates following recombinant protein induction.

The pelleted material obtained after bacterial cell lysis (section 3.2.2.2) was re-suspended in 30ml 1× reducing SDS loading buffer (i.e. the same volume used in cell lysis) to solubilise insoluble material, as previously. The equivalent of 1µl of this solution was loaded onto an SDS- PAGE gel (labelled Ins=insoluble). 1µl of the soluble fraction of cell lysates was also loaded (labelled Sol=soluble) as well as 7µl of protein marker (All blue Precision Plus, Biorad). Western blotting was performed as previously and membranes were probed with HRP-labelled anti-GST polyclonal antibody.

It is interesting that GST-Tspan13EC2 which was successfully purified contains four cysteines in the EC2 domain (i.e. two putative disulfide bonds) whilst GST-CD82EC2 and GST-Tspan31EC2 contain six cystines (i.e. three putative disulfides) and GST- Tspan5EC2 contains eight cysteines (four putative disulfides) (Table 3.1). Thus the potential for misfolding and inclusion body formation seems to be greater with an increase in the number of cysteines and thus disulfides (Qiu et al., 1989). With this in mind, the expression of the other three tetraspanins containing four cysteines was attempted first before moving on to the six- and eight-cysteine tetraspanins.

3.2.2.3 Production of recombinant tetraspanins containing four cysteines Expression and purification of GST-Tspan2EC2, GST-Tspan3EC2 and GST-Tspan32EC2 was carried out as previously. GST and GST-CD9EC2 (Parthasarathy et al., 2009) were also expressed and purified, to be used as control proteins in other experiments

Chapter 3

(Chapter 4). Purified proteins were dialysed before analysis by SDS-PAGE, Bradford assay and densitometry (figure 3.7). The apparent size of all recombinant proteins (Coomassie- stained gel) is in agreement with the predicted molecular weight (table). Lower molecular weight bands are also present and likely result from protein degradation or early termination (Ding et al., 2005). Recombinant GST was found to be >99% pure by densitometry analysis (quantifying the proportion of protein that was present at the correct molecular weight in a Coomassie-stained SDS-PAGE gel). Recombinant tetraspanin purity was in the range 73-85%, in line with previous studies, where the purity of GST-CD9EC2 expressed in E. coli Rosetta-gami strain was 66% (R. Hulme, PhD thesis) compared to 73% obtained here. The average yield of recombinant protein was approximately 500µg protein per gram bacterial pellet.

KDa M 1 2 3 4 5 6 M

25 20

Recombinant protein Yield (µg recombinant Total recombinant Purity (predicted MW) protein/g bacterial pellet) protein (µg) (%)

GST (26.2KDa) 2784 7100 >99% GST-CD9EC2 (35.4KDa) 930 4650 73% GST-Tspan2EC2 (35.6KDa) 61 255 76% GST-Tspan3EC2 (38.9KDa) 284 794 76% GST-Tspan13EC2 (34.2KDa) 800 5000 85% GST-Tspan32EC2 (36.5KDa) 449 1798 84%

Figure 3.7 Production of four-cysteine recombinant tetraspanins in E. coli Rosetta-gami

strain: analysis by SDS-PAGE and densitometry

Approximately 1.5µg of dialysed purified protein solution was loaded in each well. Samples were reduced and alkylated before electrophoresis, as previously. 7µl of marker (Precision Plus Dual Colour, Bio-Rad) was also loaded. Following separation by SDS-PAGE, the gel was stained with Coomassie stain. Recombinant protein purity was determined using densitometry analysis. 1: GST, 2: GST-CD9EC2, 3: GST-Tspan2EC2, 4: GST-Tspan3EC2, 5: GST-Tspan13EC2, 6: GST- Tspan32EC2. MW=molecular weight, KDa=Kilodaltons.

Chapter 3

3.2.2.4 Production of recombinant tetraspanins containing six or eight cysteines Recombinant proteins were expressed and purified as previously. In order to obtain a higher concentration of protein, proteins were eluted in a 2-4 times lower volume of elution buffer. Purified proteins were analysed as above by SDS-PAGE, densitometry and Bradford assay as shown in figure 3.8. Compared to the four-cysteine tetraspanins (figure 3.7), protein yields were around 10 orders of magnitude lower (average ~40µg recombinant protein/g bacterial pellet). Protein purity was also considerably lower (~48% compared to 79% for four-cysteine proteins in figure 3.7), due mainly to the presence of lower molecular weight bands which are likely degradation products of the full-length protein.

KDa M 1 2 3 4 5 6 75 50

25 20

Recombinant protein Yield ( µg recombinant Total recombinant Purity (%) (predicted MW) protein/g bacterial pellet) protein (µg)

GST (26.2KDa) 2784 7100 92% GST-Tspan4EC2 (38.9KDa) - - 45% GST-Tspan5EC2 (40.6KDa) - - 55% GST-Tspan8EC2 (38.4KDa) 74 171 54% GST-CD82EC2 (40.9KDa) 73 729 52% GST-Tspan31EC2 (35.5KDa) 24 119 32%

Figure 3.8 Production of recombinant tetraspanins with six or eight cysteines in E. coli Rosetta-gami strain: analysis by SDS-PAGE and densitometry Recombinant GST-tagged tetraspanin EC2 domains were expressed in E. coli Rosetta-gami strain, and purified by glutathione-affinity chromatography, as previously. Free GST was

expressed and purified under the same conditions. 15µl samples of dialysed protein (3µl for GST) were incubated with 4× reducing SDS-loading buffer and separated by SDS-PAGE gel. 7µl of protein marker (All blue Precision Plus, Biorad) was also loaded. Gels were stained with Coomassie Brilliant Blue stain. M: Marker, 1: GST, 2: GST-Tspan4EC2, 3: GST-Tspan5EC2, 4:

GST-Tspan8EC2, 5: GST-CD82EC2, 6: GST-Tspan31EC2. MW=molecular weight, KDa=Kilodaltons. Red arrows point to bands which were assumed to be full-length recombinant GST-Tspan4EC2 and GST-Tspan5EC2 for the purpose of determining purity.

Chapter 3

Protein misfolding, partial folding or inherent instability can result in recombinant protein degradation by bacterial proteases (Baneyx and Mujicic, 2004). On the other hand, given the lower yields of these proteins, there may simply have been more bacterial proteases present per µg protein during purification. GST-Tspan4EC2 and GST-Tspan5EC2 migrated as lower molecular weight bands than expected. The concentration of GST-Tspan5EC2 was too low to determine by Bradford assay and apart from GST-CD82EC2, the total amount of recombinant proteins purified was not sufficient for functional studies.

3.2.3 Expression and purification of recombinant tetraspanins in E. coli SHuffle strain 3.2.3.1 Optimising recombinant protein expression Altering expression parameters such as temperature, IPTG concentration and induction length is reported to reduce protein aggregation and increase yields of soluble protein (Bar et al., 2006; Vasina and Baneyx, 1997; Sorensen and Mortensen, 2005). Therefore small-scale expression trials were conducted to find the optimum conditions for recombinant tetraspanin expression. E. coli SHuffle competent cells were transformed with plasmids encoding GST-Tspan5EC2 and GST-Tspan31EC2 (containing four and three putative disulfide bridges, respectively). 500ml cultures of transformed bacteria (OD600nm=0.6) was obtained (section 2.2.3). 10ml aliquots of the bacterial suspension were transferred to 100ml flasks and expression was induced under the conditions outlined in the table below. At the indicated time-points, cells were harvested, lysed by sonication and 10µl aliquots of the soluble fractions were analysed by Western blotting (figure 3.9).

Chapter 3

1 2 3 4 5 6 7 8 9 10 11 GST-Tspan5EC2 GST-Tspan31EC2

Relative protein yield

Temp Cell density Induction IPTG Lane o GST-TS5EC2 GST-TS31EC2 ( C) (OD600nm) (hours) (mM) 1 20 2.0 0 0.1 2.0 1.0 2 20 2.0 1 0.1 8.1 8.2 3 20 2.0 2 0.1 11.3 11.1 4 20 2.0 4 0.1 6.9 9.1 5 30 2.0 1 0.1 5.4 10.2 6 30 2.0 2 0.1 5.2 7.6 7 30 2.0 4 0.1 5.2 9.0 8 30 0.6 4 0.1 3.9 5.0 9 20 0.6 16 0.1 7.1 3.7 10 20 2.0 16 0.1 1.0 1.9 11 20 2.0 16 1 3.8 6.1

Figure 3.9 Optimisation of tetraspanin protein expression in E. coli SHuffle strain.

The following induction parameters were varied; temperature [Temp (°C)], cell density at

induction [OD600nm], length of induction [Induction (hours)] and concentration of inducing agent [IPTG (mM)]. Soluble cell lysates were separated by SDS-PAGE and analysed by Western blotting

with anti-GST HRP-linked antibody, as previously (top panels). Relative protein yield was estimated from immunoblots by densitometry. For both proteins, the highest yield was obtained under the same set of conditions (red box).Optimisation of growth media, using Tryptone phosphate broth, Terrific broth and culture additives such as sorbitol and betaine hydrochloride did not appreciably increase yield (data not shown).

3.2.3.2 Production of recombinant tetraspanins containing six or eight cysteines Following determination of the optimum expression conditions (Figure 3.9), recombinant tetraspanin expression was induced in 500ml bacterial cultures of E. coli SHuffle strain. Recombinant protein was purified using glutathione affinity resin, as previously. Protein was eluted in a 2-4 times lower volume of glutathione elution buffer in order to obtain sufficiently high concentrations for functional studies. On average, recombinant protein purity was higher (60%) than for recombinant tetraspanins expressed in E. coli Rosetta-gami strain (48%). Protein yield was also higher such that, apart from GST-Tspan5EC2, sufficient protein was obtained for

Chapter 3 testing in functional assays. However, one issue for concern was the presence of higher molecular weight material (~45KDa and 52KDa) in the GST-Tspan4EC2, GST- Tspan5EC2 and GST-Tspan8EC2 preparations. In immunoblots of reduced protein, these bands do not stain with anti-GST antibody (figure 3.11) or anti-tetraspanin antibodies (figure 3.12). Therefore, it is likely that they are contaminating bacterial proteins. For this reason, it was decided not to use these proteins in functional studies.

KDa M 1 2 3 4 5 M

Recombinant protein Yield (µg recombinant Total recombinant Purity (predicted MW) protein/g bacterial pellet) protein (µg) (%)

GST (26.2KDa) 928 3200 >99%

GST-Tspan4EC2 (38.9KDa) 14 243 57% GST-Tspan5EC2 (40.6KDa) 7 92 65% GST-Tspan8EC2 (38.4KDa) 27 889 73% GST-Tspan31EC2 (35.5KDa) 38 422 43%

Figure 3.10 Production of recombinant proteins with six or eight cysteines in E. coli

SHuffle strain: analysis by SDS-PAGE and densitometry

Recombinant GST-tagged tetraspanin EC2 domains were expressed in E. coli SHuffle strain under the optimised conditions arrived at previously (figure 3.9). Proteins were purified, dialysed, as previously. Free GST was expressed and purified under the same conditions. 15µl samples of dialysed protein (3µl for GST) were incubated with 4× reducing SDS-loading buffer and separated by SDS-PAGE gel. 7µl of protein marker (Precision Plus Dual Colour, Bio-Rad) was also loaded. Gels were stained with Coomassie Brilliant Blue stain. M: Marker, 1: GST, 2: GST- Tspan4EC2, 3: GST-Tspan5EC2, 4: GST-Tspan8EC2, 5: GST-Tspan31EC2. MW=molecular weight, KDa=Kilodaltons. Red arrows point to bands which were assumed to be full-length recombinant GST-Tspan4EC2, GST-Tspan5EC2 and GST-Tspan8EC2 for the purposes of determining protein purity.

Chapter 3

3.2.4 Analysis of recombinant protein folding 3.2.4.1 Western blotting with anti-GST antibody As a first step, recombinant tetraspanin proteins produced in this study were characterised using reducing and non-reducing SDS-PAGE followed by detection with an anti-GST polyclonal antibody which recognises linear as well as conformational (three-dimensional) epitopes (Flint et al., 2006). Disulfide bridges have been found to stabilise the native fold of numerous proteins particularly cell surface or secreted proteins (Fas et al., 2012). Mutation of any of the four cysteines within the EC2 domain of recombinant GST-tagged CD9EC2 and cysteine reduction using β-mercaptoethanol abolish recognition by a conformation-sensitive anti-CD9 antibody, highlighting the importance of cysteines to EC2 domain folding (Higginbottom et al., 2003; R. Hulme, PhD thesis). Reduction of oxidised cysteines in proteins tends to retard protein migration on SDS-PAGE gels due to loss of the compact three-dimensional structure and a consequent increase in hydrodynamic radius as the protein unfolds (Jurado et al., 2002; Bar et al., 2006). Thus analysis of non-reduced and reduced protein migration provides information on the contribution of disulfide bridges to protein folding. As shown in figure 3.11, non-reduced and reduced GST (immunoblot A) migrated at approximately the same rate (band at ~26KDa) and dimer formation previously reported for GST (Walker et al., 1993; McTigue et al., 1995) was not detected here. Therefore it is likely that differences between non-reduced and reduced GST-TspanEC2 migration patterns result from conformational changes in the tetraspanin EC2 domain itself and not the GST tag. Figure 3.12 also shows that for all recombinant tetraspanins, the migration patterns of the non-reduced (N) and reduced (R) protein are different. This suggests that some or all of the cysteines in the non- reduced protein are oxidised forming disulfide bridges. However it is unclear whether disulfide bridges have been formed between the correct pairs of cysteines in the protein and therefore whether the protein has native conformation.

For several non-reduced recombinant tetraspanins (GST-CD9EC2, GST-Tspan3EC2, GST-Tspan4EC2, GST-Tspan31EC2 and GST-Tspan32EC2), bands are observed at approximately 75KDa. The predicted molecular weight of homodimers for these proteins is 70.8-77.8KDa indicating that a proportion of these molecules may be present as dimers. Homo-meric oligomerisation has previously been reported for some tetraspanins. Full-length human RDS is present as a tetramer in cell lysates

Chapter 3 whilst human tetraspanins CD9, CD81 and CD151 are observed as homodimers on the cell surface (Loewen and Molday, 2000; Kovalenko et al., 2004). Recombinant tetraspanin EC2 domains have also been found to dimerise in solution. Analytical ultracentrifugation showed that purified polyhistidine-tagged human CD81EC2 produced in E. coli Rosetta-gami cells or in the human embryonic kidney cell line HEK293 forms stable dimers (Rajesh et al., 2012). The EC2 domain of the tetraspanin RDS expressed in the mammalian COS-1 cell line also forms homodimers (Ding et al., 2005).

KDa M N R N R N R N R N R N R N R M KDa 250 250 150 150 100 100 75 75 50 50

37 37

25 25 20 20

GST R GST-CD9 GST- GST- GST- GST- GST- EC2 R Tspan2 Tspan3 Tspan4 Tspan5 Tspan8 EC2 R EC2 R EC2 R EC2 SH EC2 SH

KDa M N R N R N R N R M KDa

250 250 150 150 100 100 75 75 50 50 37 37

25 25 20 20 GST- GST- GST- GST- Tspan13 CD82 Tspan31 Tspan32 EC2 R EC2 R EC2 SH EC2 R

Figure 3.11 Western blotting of recombinant tetraspanin proteins with anti-GST antibody.

Approximately 0.5µg purified recombinant protein was treated with non-reducing (N) or

reducing (R) SDS-PAGE loading buffer (containing β-mercapthoethanol), as previously. Reduced samples were additionally alkylated to prevent re-oxidation during electrophoresis (Chapter 2, section 2.2.1.5). Following separation by SDS-PAGE, bands were transferred to nitrocellulose membranes and recombinant protein was detected by immunoblotting with anti-GST HRP-linked

antibody. ‘R’ refers to proteins produced in E. coli Rosetta-gami strain whilst ‘SH’ refers to protein produced in E. coli SHuffle strain. Regarding the red arrows, see text for explanation.

Chapter 3

Interestingly, mutation of any of the six cysteines in the RDS EC2 region within the full-length protein results in RDS misfolding and aberrant oligomerisation (Goldberg et al., 1998). Very high molecular weight species are prominent in non-reduced samples of GST-Tspan2EC2, GST-Tspan4EC2, GST-TSpan5EC2, GST-CD82EC2 and GST- Tspan31EC2 (figure 3.11). These may represent higher order oligomers of correctly- folded protein or alternatively misfolded protein aggregates but their molecular association appears to be dependent on disulfide bridge formation as they are not observed when proteins are reduced. It is interesting that apart from GST-Tspan2EC2, all these recombinant tetraspanins contain 6 or more cysteines (i.e. 3 or more putative disulfide bridges). According to the literature, this would make them particularly prone to misfolding and aggregation (Qiu et al., 1998). In the reduced GST-Tspan4EC2 and GST-Tspan5EC2 preparations, the highest molecular weight bands (red arrows) are very faint compared to other bands. This may be due to incomplete protein reduction, early termination of protein translation or substantial protein degradation by bacterial proteases.

3.2.4.2 Western blotting with anti-tetraspanin antibodies Next, recombinant protein reactivity with a panel of antibodies against human tetraspanin EC2 domains was examined (figure 3.12). Apart from antibody J, all the antibodies used in figure 3.11 (obtained from Aviva system biology) have been raised against short synthetic peptide fragments corresponding to regions within tetraspanin EC2 domains. As far as the author is aware, this is the first time these antibodies have been tested against recombinant tetraspanins. As shown in figure 3.12, antibodies B, D, E, F, G, H and I only reacted against reduced recombinant tetraspanins whereas antibodies A and C predominantly reacted against reduced protein but also showed some reactivity with the non-reduced proteins. It is unclear whether the peptide used to elicit these antibodies had native conformation and thus if these antibodies recognise conformational (three-dimensional) or linear epitopes or both. So conclusions cannot be drawn about whether the recombinant proteins have native conformation.

TS82b (antibody J) is a conformation-sensitive antibody which recognises native human CD82 protein on the cell surface (Xu et al., 2009). This antibody recognised non-reduced but not reduced GST-CD82EC2 protein suggesting that the non-reduced

Chapter 3

CD82EC2 domain has native-like conformation and that disulfide bonds make an essential contribution to the native fold. There is also a band apparent at ~250KDa suggesting that at least some of the higher molecular weight material that reacted with anti-GST antibody (figure 3.11, GST-CD82EC2) is correctly folded. Antibodies A, B, E, F and in particular antibody C also react with lower molecular weight material likely to be truncated protein. Antibodies raised against the C-terminal boundary of tetraspanin EC2 domains would have been useful to determine if the expressed proteins are actually full-length.

3.2.4.3 Analysis of recombinant GST-CD82EC2 protein by ELISA

To investigate the contribution of disulfide bridges to GST-CD82EC2 folding, the reactivity of GST-CD82EC2 with three different antibodies was also analysed by ELISA method. Stock protein was boiled with reducing SDS-PAGE loading buffer and alkylated or boiled with SDS-PAGE non-reducing buffer. Native (non-denatured) protein was used as a positive control. As shown in figure 3.12, the conformation- sensitive anti-CD82 antibody TS82b reacted with native and non-reduced GST- CD82EC2 but not the reduced protein. This confirms the finding in Western blotting where only non-reduced protein reacted with this antibody (figure 3.12). For non- reduced GST-CD82EC2, the curve is bell-shaped suggesting that, at high concentrations, a component of the non-reducing loading buffer interferes with antibody binding or perhaps antibody-linked Horse radish perioxidase (HRP) enzyme catalysis but is diluted out at lower concentrations of protein. In fact, at concentrations below 0.5µg/ml, greater binding is detected for the non-reduced than native GST- CD82EC2. In contrast, anti-CD82 antibody ARP63360 recognised all three proteins but not GST showing that its recognition is specific to the CD82 EC2 domain. This antibody did not recognise non-reduced GST-CD82EC2 in Western blotting therefore ELISA method may be more sensitive at detecting protein-antibody interactions. Finally, all proteins reacted with anti-GST antibody as expected.

Chapter 3

N R N R N R N R

150 150 75 150 75 75 50 50 50 50 37 37 37 37 25 25 25 25 20

Protein: GST-Tspan2 GST-Tspan3 GST-Tspan4 GST-Tspan5 EC2 R EC2 R EC2 R EC2 SH Antibody: (A) (B) (C) (D)

N R N R N R 150 150 75 150 N R N R 50 75 75 37 50 50 25 37 37 20 25 25

Protein: GST-Tspan8 GST-CD82 GST-Tspan31 EC2 SH EC2 R EC2 SH Antibody: (E) (F) (G)

N R N R N R 150 150 250 75 75 150 50 50 75 37 37 50 37 25 25 25

Protein: GST-Tspan32 GST-Tspan32 GST-CD82 EC2 R EC2 R EC2 R Antibody: (H) (I) (J)

Figure 3.12 Western blotting of recombinant tetraspanin proteins with a range of anti- tetraspanin antibodies. 1-10µg of each recombinant protein was incubated with either non-reducing SDS-PAGE loading buffer (lanes labelled N) or with reducing SDS-PAGE loading buffer followed by alkylation to prevent re-oxidation of reduced cysteines (lanes labelled R). Samples were separated by SDS- PAGE and subjected to Western blotting as previously. Antibodies are as follows, A: ARP46643,

B: ARP46642, C: ARP44756, D: ARP46640, E: ARP46477, F: ARP63360, G: ARP46677, H: ARP46635, I: ARP46636, J: TS82b. Incubation with antibodies A-I were followed by anti-rabbit HRP-linked antibody whilst antibody J was followed by anti-mouse HPR-linked antibody. ‘R’ refers to proteins produced in E. coli Rosetta-gami strain whilst ‘SH’ refers to protein produced in E. coli SHuffle strain. N=non-reduced, R=reduced and alkylated.

Chapter 3

3.0

2.5

2.0 m

n 0

5 1.5

4 D O 1.0

0.5

0.0 2.0 1.0 0.5 0.2 0.1 0.05 P ro te in c o n ce n tra tio n ( g /m l)

G S T -C D 8 2 E C 2 (n a tiv e ) A n ti-C D 8 2 m o n o c lo n a l G S T -C D 8 2 E C 2 (n o n -re d u c e d ) a n tib o d y (T S 8 2 b ) G S T -C D 8 2 E C 2 (re d u c e d ) G S T -C D 8 2 E C 2 (n a tiv e ) G S T -C D 8 2 E C 2 (n o n -re d u c e d ) A n ti-C D 8 2 p o ly c lo n a l G S T -C D 8 2 E C 2 (re d u c e d ) a n tib o d y (A R P 6 3 3 6 0 ) G S T (re d u c e d ) G S T -C D 8 2 E C 2 (n a tiv e ) G S T -C D 8 2 E C 2 (n o n -re d u c e d ) A n ti-G S T p o ly c lo n a l G S T -C D 8 2 E C 2 (re d u c e d ) a n tib o d y (A 7 3 4 0 )

G S T (re d u c e d )

Figure 3.13 Analysis of recombinant GST-CD82EC2 by ELISA method.

ELISA assay was performed as described (Chapter 2, section 2.2.1.10) using recombinant GST and GST-CD82EC2 produced in E. coli Rosetta-gami strain. Proteins stock solutions were either untreated (native), boiled with SDS-PAGE non-reducing loading buffer for 5 minutes at 80˚ (non- reduced) or boiled with SDS-PAGE reducing loading buffer for 10 minutes at 80˚ and alkylated

(reduced). Stock solutions were then serially diluted in ELISA blocking buffer and adsorbed onto wells of an ELISA plate. Wells were incubated with the following antibody combinations as described in section 2.2.1.10; TS82b (1:1600) followed by anti-mouse HRP-linked secondary antibody (1:5000), ARP63360 (1µg/ml) followed by anti-rabbit HRP-linked secondary antibody

(1:7000) or anti-GST HRP-linked antibody (1:5000). Plates were developed with TMB substrate and the absorbance (A450nm) was determined using a plate reader. Data points are means of triplicate values from one experiment ± SEM. A bell-shaped dose-response curve was fitted to values for GST-CD82EC2 (non-reduced) treated with TS82b antibody whilst the remaining values were fitted to four-parameter (variable slope) dose-response curves.

Chapter 3

3.3 Discussion This chapter details the expression of the large extracellular domain of nine new tetraspanins as recombinant GST-tagged fusion proteins in E. coli. Yields of soluble recombinant protein expressed in E. coli Rosetta-gami strain were variable; tetraspanins containing four cysteines (two putative disulfide bonds) expressed reasonably well whilst for six or eight-cysteine (three or four putative disulfides) tetraspanins protein yield and purity was lower. A high proportion of the expressed material was present in the insoluble fraction of cell lysates indicating protein misfolding and aggregation into inclusion bodies. Expression in E. coli SHuffle strain resulted in greater protein purity and higher yields of protein which could be used for functional studies. In wildtype E. coli strains, the reducing environment of the cytoplasm does not support stable disulfide bond formation thus this post- translational modification is confined to the bacterial periplasmic space (Baneyx and Mujacic, 2004). The additional adaptations of the SHuffle strain such as constitutive cytoplasmic expression of the DsbC disulfide bond isomerase make it particularly suited to the expression of disulfide-containing proteins. In fact, cytoplasmic DsbC expression appeared to be more beneficial to production of soluble correctly folded recombinant protein than protein targeting to the periplasm where DsbC normally resides (Bassette et al., 1999).

In almost all recombinant tetraspanin preparations (figure 3.4 to 3.12), full-length recombinant protein purity was compromised by the presence of lower molecular weight bands ranging in size down to 26KDa (predicted size of free GST). This may be due to early termination of protein translation. On the other hand, proteolytic degradation of the full-length protein by bacterial proteases is commonly observed if the protein is misfolded or partially folded (Baneyx and Mujicic, 2004). Proteolytic cleavage may also occur during bacterial cell lysis when compartmentalised and/or regulated proteases are released into the lysis buffer (Goldberg and Dunn, 1988). Some post-lysis cleavage is likely attributable to instability of tetraspanin EC2 domains themselves since a number of the lower molecular weight bands are bigger than 26KDa. The presence of the glycine linker sequence in the modified pGEX-KG vector (figure 3.1) may also exacerbate the problem. The linker sequence is reported to improve the efficiency of proteolytic cleavage at the thrombin cleavage site immediately downstream, presumably by improving thrombin accessibility to the site

Chapter 3

(Guan and Dixon, 1990). By the same token, it may enhance accessibility to bacterial proteases or make the recombinant protein generally less stable. The glycine linker may also reduce steric hindrance by the GST tag therefore its absence may be detrimental to tetraspanin EC2 function. Another issue of concern is that some of the recombinant proteins (for example GST-Tspan4EC2 and GST-Tspan5EC2, figure 3.8) migrate as lower molecular weight proteins than predicted raising the possibility that they are truncated. The introduction of a second affinity tag such as a polyhistidine tag at the C-terminus of GST-tetraspaninEC2 proteins would enable a two-step purification procedure where GST-tagged proteins are isolated on a glutathione affinity column, the GST is enzymatically cleaved and the full-length EC2 domain is purified using immobilised metal-affinity chromatography (IMAC). Only full-length EC2 domains would be able to bind the IMAC column by virtue of the polyhistidine-tag and contaminating GST would also be removed resulting in greater protein purity. Thrombin cleavage of GST-CD9EC2 was found to be inefficient (R. Hulme, PhD thesis), therefore an alternative vector encoding a protease cleavage site specific for a different enzyme such as TEV (Tobacco etch virus) protease may be more appropriate.

Another significant issue is whether the recombinant proteins produced in this study have native conformation. Apart from TS82b mAb (specific for human CD81EC2 domain), conformation-sensitive antibodies are currently not available for the other tetraspanins expressed in this study, necessitating characterisation of these proteins by other methods such as circular dichroism (CD) spectroscopy. However, a high degree of protein purity (>95%) is required for circular dichroism to ensure that the signal is not coming from other contaminating proteins or degradation/truncation products. Large affinity tags such as GST (26KDa) also need to be removed from the target protein since they are expected to make to make a substantial contribution to the CD signal obtained (Kelly et al., 2005). Theoretically, it is possible to obtain separate CD spectra for free GST and GST-TspanEC2 proteins and subtract the former from the latter to determine the CD signal coming from the tetraspanin EC2 domains. When this was actually performed for GST-TspanEC2 proteins, the spectra proved difficult to interpret due to low protein purity (work carried out in collaboration with Dr Rosemary Staniforth, Department of Molecular Biology and Biotechnology, University of Sheffield).

Chapter 3

Analysis of recombinant tetraspanins by Western blotting with anti-GST antibody revealed that in some protein preparations including GST-CD82EC2, a considerable proportion of the GST-tagged material is present as high molecular material (≥150 KDa). These may represent misfolded protein aggregates, however blotting with the TS82b antibody showed that at least some of this material is correctly folded (figure 3.12, antibody J). The proportion of correctly-folded GST-CD82EC2 within the purified protein preparation could be determined using immunoprecipiatation with the TS82b antibody to precipitate correctly folded protein out of solution and detect any remaining protein (presumed not have native conformation) with an anti-GST antibody (Parthasarathy et al., 2009).

Protein regions involved in interactions with other proteins seem to be particularly susceptible to aggregation. Of interest was the finding that compared to other parts of the protein, these same regions are enriched in disulfide bonds (Fas et al., 2012). As discussed in section 1.1.1.2, tetraspanin-partner interaction sites for several tetraspanins have been mapped to the EC2 domain and in particular, the EC2 variable subdomain (Chapter 1, figure 1.2) which contains almost all the cysteine residues found in the EC2. Therefore, the presence of aggregates in recombinant tetraspanin EC2 preparations may reflect an intrinsic characteristic of these proteins rather than or in addition to being a consequence of misfolding in a bacterial expression system. Very large aggregates would also be insoluble thereby lowering yields of soluble recombinant protein (Sorensen and Mortensen, 2005). Considering these points, other approaches including recombinant protein refolding from inclusion bodies (Sadhev et al., 2008), periplasmic expression (Qiu et al., 1989) or expression in eukaryotic systems may not improve yields of soluble protein. In addition, aggregate removal by such methods as hydrophobic interaction chromatography will likely result in a considerable loss of biological material (Lu et al., 2009).

Alternative expression systems in which disulfide bond formation is better supported include mammalian, yeast or insect cells (Rajesh et al., 2012; Bar et al., 2006). Many tetraspanin EC2 domains contain potential N- and O-linked glycosylation sites therefore expression in yeast or insect cells may lead to heterogenous/non-native protein glycosylation (Qiu et al., 1998). Mammalian expression is therefore preferable and currently under investigation by our group but initial attempts to express GST-

Chapter 3

CD9EC2, GST-Tspan5EC2 and GST-CD151EC2 in mammalian human embryonic kidney 293T (HEK293T) and Chinese hamster ovary (CHO) have proven unsuccessful (Mr John Palmer, Department of Molecular Biology and Biotechnology, University of Sheffield, internal communication). It is notable that the heavily glycosylated EC2 domain of human CD63, when expressed as a recombinant GST-tagged protein in E. coli was recognised by conformation-sensitive anti-CD63 antibodies therefore lack of this post-translational modification in E. coli did not appear to adversely affect protein folding (Parthasarathy et al., 2009, Green et al., 2011).

Chapter 4

Chapter 4: The role of tetraspanins in Concanavalin A-induced Multinucleated Giant Cell formation

4.1 Introduction The tetraspanin superfamily of membrane proteins have been ascribed roles in several processes requiring membrane fusion (Fanaei et al., 2011). The involvement of tetraspanins in mononuclear phagocyte fusion leading to the formation of multinucleated giant cells (MGCs) has been the focus of our research group. Of interest was the finding that the large extracellular domains (EC2) of tetraspanins CD9 and CD63, produced as recombinant proteins in bacteria, could significantly and specifically inhibit MGC formation (Parthasarathy et al., 2009). Studies with anti- tetraspanin antibodies provided further evidence that these tetraspanins are indeed involved in MGC formation (discussed in section 1.2.3.3.4). Monoclonal anti- tetraspanin antibodies are currently not available for the majority of tetraspanins, therefore the aim of this chapter was to determine the involvement of several new tetraspanins in MGC formation using a panel of recombinant human tetraspanin proteins produced in bacteria (Chapter 3).

4.2 Results 4.2.1 Establishing negative and positive controls A number of controls are commonly utilised by studies investigating the functional activity of recombinant tetraspanin EC2 domains (Chapter 1, Table 1.2). Free GST (produced in the same expression host) is used as a negative control to show that the functional activities of recombinant GST-tagged EC2 domains is not due to GST or a bacterial contaminant or other component present in the purified protein preparation (for example Green et al., 2011). The GST tag can be cleaved from the recombinant protein using the thrombin cleavage site (Chapter 3, figure 3.1). However, relatively low yields of tetraspanin EC2 were obtained following thrombin digestion (R. Hulme, PhD thesis) therefore full-length tagged proteins were used in subsequent experiments.

Two independent studies have confirmed the inhibitory effects of GST-CD9EC2 in the MGC assay (Takeda et al., 2003; Parthasarathy et al., 2009) therefore this construct was adopted as a positive control. Establishing proper negative and positive controls

Chapter 4 was particularly important in this study, because some of the recombinant proteins to be tested were produced in a new expression system, E. coli SHuffle strain (NEB, C3030H). The adaptations of this strain allow for better folding of disulfide bridge- containing proteins which may result in higher levels of enzymatic or other activity (Lobstein et al., 2012). Therefore, it was of particular interest here to determine whether recombinant GST-CD9EC2 produced in the SHuffle strain had higher activity in the MGC assay than that made in the E. coli Rosetta-gami host.

To this end, GST and GST-CD9EC2 (herein referred to as GST-CD9) were expressed in both SHuffle and Rosetta-gami strains and purified using glutathione affinity chromatography (Chapter2, sections 2.2.3 to 2.2.6). ‘SH’ denotes proteins expressed in SHuffle strain whilst ‘R’ refers to proteins produced in Rosetta-gami. The effects of these control proteins on MGC formation was determined as described (Chapter 2, section 2.3.2). Briefly, peripheral blood mononuclear cells (PBMC) were purified from human blood and monocytes were isolated by adhesion to wells of a 96-well tissue culture plate. Monocyte fusion was induced by incubation with 10µg/ml Concanavalin A (Con A). To test recombinant activity in the MGC assay, 500nM protein was also added to the culture media containing Con A. Following 72 hours incubation at

37˚C/5% CO2, cells were fixed, stained and imaged (Chapter 2, sections 2.3.3 to 2.2.4) as shown in figure 4.1. Con A efficiently induced fusion of adherent monocytes (figure 4.1 A) into multinucleated giant cells (figure 4.1 B).

The extent of monocyte fusion in each well was assessed using two parameters; the fusion index (number of nuclei in MGC as a proportion of total numbers of nuclei) and average MGC size (average number of nuclei per giant cell) (Chapter 2, section 2.3.5). The results are shown in figure 4.2. In Con A-only wells, 88% of the nuclei were present in giant cells with an average of 13 nuclei per MGC which is comparable to previous findings (Parthasarathy et al., 2009). All four recombinant proteins significantly inhibited Con A-induced MGC formation (P<0.0001), causing a reduction in both parameters of fusion index and average MGC size, compared to the control. In addition, the inhibitory activity of recombinant proteins made in E. coli SHuffle strain (GST SH and GST-CD9 SH) was significantly higher than those made in E. coli Rosetta- gami strain (GST R and GST-CD9 R), as judged by the fusion index.

100

Chapter 4

(A) Unstimulated (B) Con A

(E) Con A + GST-CD9 SH (F) Con A + GST-CD9 R

Figure 4.1 Immunofluorescence images of monocytes stimulated with Con A and recombinant GST and GST-CD9.

Monocytes were stimulated to form MGCs as described in the text. Following 72 hours incubation, cells were fixed and incubated with Hoescht to stain nuclei (red) and TRA-1-85 primary antibody followed by anti-mouse FITC-labelled secondary antibody to stain the plasma membrane (green). Imaging was carried out using the 20× objective lens on an ImageXpress imaging system as described (Chapter 2, section 2.3.4). Representative false-colour images of each condition are shown. Typically, nuclei in 25 such fields were counted per well. Scale bar=100µm.

101

Chapter 4

120

100

) l

x ** o

e *** r

d 80

t ****

n n

****

o n c 60

****

i o

s ****

u 40

F ( 20

A H R H R n S T S 9 o T S 9 D S C G D -C G -C T T S S G G

120

100

)

o s r

n G o ****

c 60

M f

****

. ****

g 40 **** v

( A 20 0

A H R H R n S T S 9 o T S 9 D S C G D -C G -C T T S S G G

Figure 4.2 Effects of recombinant GST and GST-CD9 proteins on Con A-induced multinucleated giant cell (MGC) formation.

Adherent monocytes were incubated with 10µg/ml Concanavalin A (Con A) to induce MGC formation. To test the effects of recombinant proteins, 500nM protein was added at the same time and the assay was carried out as described in the text. MGCs were defined as cells containing three or more nuclei (Parthasarathy et al., 2009). The extent of fusion was quantified using the parameters of Fusion index and Average MGC size (Chapter 2, section 2.3.1.5). Results are means ± standard error of the mean (SEM). For each monocyte donor, the experiment was performed in duplicate or triplicate and the results were normalised against the Con A control and are thus presented as a percentage of the control which was taken to be 100%. Statistical significance was tested using ordinary one-way analysis of variance (ANOVA) with a Holm-Sidak post-test. N (number of independent experiments/donors)=3+, n (total number of replicates per condition)=6+, **P<0.01, ***P<0.001, ****P<0.0001.

102

Chapter 4

As shown in Table 4.1, the inhibitory activity of GST-CD9 R (assessed by the reduction in fusion index and average MGC size) was similar to that observed by other researchers using GST-CD9 produced in other E. coli strains (Takeda et al., 2003; Parthasarathy et al., 2009 and R. Hulme, PhD thesis). However, free GST which had previously been found to have no effect on MGC formation, significantly inhibited MGC formation in this study.

Approx. reduction Approx. reduction in

in fusion index (%) average MGC size (%)

GST

This study (GST R) 30 49

Takeda et al., 2003 0 Not determined

Parthasarathy et al., 2009 0 Not determined

R. Hulme, PhD thesis 5 5

GST-CD9 This study (GST-CD9 R) 40 59 Takeda et al., 2003 36 Not determined Parthasarathy et al., 2009 47 52 R. Hulme, PhD thesis 43 62

Table 4.1 Inhibitory activity of GST and GST -CD9 in various studies.

Recombinant GST and GST-CD9 were produced in E. coli BL21 strain (Parthasarathy et al., 2009), E. coli Rosetta-gami strain (R. Hulme, PhD thesis; present study) and an unknown strain (Takeda et al., 2003). Parthasarathy et al. and Takeda et al. had used 20µg/ml recombinant protein instead of 500nM. 500nM GST=13µg/ml, 500nM GST-CD9= 18.3µg/ml. Approximate values for the fusion index and average MGC size were obtained from bar chart graphs supplied by the authors. Values were converted to percentages of the Con A control.

Methodological differences between this study and the others mentioned could have contributed to these results, however this possibility was excluded as will be addressed in the discussion (section 4.3). GST-CD9 has two disulfide linkages therefore, in theory, the higher activity levels of GST-CD9 SH could be due to better folding in the E. coli SHuffle expression strain, resulting in higher levels of active protein (Chapter 3, section 3.1.2.5). However, GST-CD9 R and SH were found to have similar reactivity with a conformation-sensitive anti-CD9 antibody (section 4.2.4) and similar levels of activity in an independent functional assay (Chapter 5, figure 5.1).

103

Chapter 4

Next the role of contaminants was considered. As discussed in Chapter 1 (section 1.3.1 and 1.3.2), bacterial contaminants are inevitably present in recombinant protein preparations and are reported to have dramatic effects on monocyte differentiation and particularly osteoclast formation (Takashima et al., 1993; Zou and Bar Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009). The major bacterial contaminant in recombinant protein preparations is bacterial lipopolysaccharide (LPS) (Petsch and Anspach, 2000). In a number of studies, the functional activity attributed to recombinant proteins produced in bacteria has been found to be due to LPS contamination, with the apparent level of activity directly corresponding to the LPS content (Bausinger et al., 2002; Gao et al., 2003a; Gao et al., 2003b; Dudley et al., 2003; Pepys et al., 2005; Hreggvidsdottir et al., 2009). The level of contamination may depend on several factors including the bacterial expression host, expression conditions, cell lysis and purification conditions as well as the biochemical properties of the recombinant protein itself such as its hydrophobicity and isoelectric point (Anspach and Hilbeck, 1995). Thus it is possible for different proteins produced in the same expression system or the same protein produced in different expression systems to have different levels of bacterial contamination.

4.2.2 Measuring LPS contamination levels in recombinant protein preparations The Limulus Amoebocyte Lysate (LAL) chromogenic assay is widely used to quantify the concentration of LPS in protein solutions (Wakelin et al., 2006). The assay is based on the ability of LPS to bind and activate the Factor C pro-enzyme isolated from the horseshoe crab Limulus polyphemus (Muta et al., 1991). The LPS-activated enzyme then catalyses the cleavage of a synthetic peptide liberating a yellow-coloured compound, the concentration of which can be easily determined using a standard plate reader.

The assay was carried out as described in figure 4.3. The LPS content of recombinant proteins was calculated as ng/ml LPS in a 500nM protein solution, i.e. the protein concentration used in the MGC assay. It is apparent from figure 4.3 A that GST-CD9 proteins (SH and R) have 1.9 and 3.8 times the level of LPS, respectively, than GST (SH and R). LPS is an amphipathic molecule with a great propensity for binding to proteins through electrostatic and hydrophobic interactions (Petsch and Anspach, 2000). Therefore, much of the LPS in a recombinant protein solution is likely to be tightly

104

Chapter 4 bound to the protein itself (Reichelt et al., 2006). The parameters that determine the capacity of a particular protein to bind LPS are poorly defined. The molecular weight of the protein may be one such parameter (i.e. the bigger the protein the more LPS may bind), therefore LPS content was re-calculated as ng/ml LPS per µg recombinant protein to account for the molecular weight difference between GST and GST-CD9 (26.2 and 35.4KDa, respectively). This partly reduced the magnitude of the difference in LPS content between GST and GST-CD9 produced in the same strain (figure 4.3 B).

A) B)

4 0 2 .0 1 .8 9 3 4 .5 0

)

e t

t 1 .3 8

o n

n 3 0 1 .5

u S

s 1 8 .0 0

i 2 0 1 .0

m 0 .6 8 S

S 1 2 .4 0

n µ

1 0 ( 0 .5 r

( 0 .2 5 p 3 .2 5

0 0 .0

H R H R H R H R S T S S T S T S 9 9 T S 9 9 S G D D S G D D G C G C -C - -C - T T T T S S S S G G G G

Figure 4.3 LPS content of recombinant proteins used in this study.

The concentration of LPS in recombinant protein preparations was determined using a LAL

Chromogenic Endotoxin Quantitation kit (Pierce), following manufacturer’s protocol. Briefly, lyophilised LPS from E. coli 0114:B4 strain was reconstituted in ultrapure endotoxin-free water and diluted to give standard LPS stock solutions containing 0.1, 0.25, 0.5 and 1.0 EU/ml (Endotoxin Unit/ml). Recombinant proteins were also serially diluted in ultrapure water. 50µl

samples of the LPS standards, diluted recombinant proteins and ultrapure water (blank) were loaded in duplicate into wells of a 96well ELISA plate and the assay was performed as instructed. Absorbance at 405nm was measured on a plate reader and a standard curve was plotted from which the LPS concentration of samples (in EU/ml) could be extrapolated. LPS

concentrations were converted to ng/ml, assuming that 120pg of LPS from E. coli O111:B4 has an activity of 1 Endotoxin Units (EU) (Kruger, 1989).

105

Chapter 4

Protein yield LPS content (ng/ml Recombinant protein (µg/ml protein per gram LPS per µg protein) bacterial pellet)

GST SH vs GST-CD9 SH 1.4× higher 1.4× lower

GST R vs GST-CD9 R 2.4× higher 2.7× lower

Table 4.2 Relationship between recombinant protein yield and LPS content

Another factor may be recombinant protein yield. Assuming that every gram of bacterial pellet (from the same strain) contains a defined amount of LPS, then it seems plausible that the lower the expression levels of the recombinant protein (µg/g bacterial pellet), the higher will be the ratio of LPS to the recombinant protein. This notion is supported by data shown in Table 4.2.

Figure 4.3 also shows that GST and GST-CD9 produced in the SHuffle strain have 5.5 and 2.8 times greater levels of LPS, respectively, than those made in the Rosetta-gami strain. Differences in protein yield could not account for these results therefore other factors such as strain-specific variations in the quantity of LPS synthesised by the bacterium may be responsible. It is apparent that the inhibitory activity of the four proteins tested here broadly correlates with their level of LPS contamination. It was of interest to determine whether directly reducing the LPS contamination of the same protein expressed under the same conditions could result in a reduction in inhibitory activity in the MGC assay.

4.2.3 Reducing LPS contamination using a detergent wash method A summary of LPS reduction methods employed in the literature is provided in Table 1.3 (Chapter 1). Reichelt et al. (2006) have developed a simple and effective procedure for LPS reduction of affinity-tagged recombinant proteins. Crude cell lysate is obtained in the usual way and incubated with chromatography resin. The column is washed with a non-ionic detergent to dissociate LPS from resin-bound proteins. This is followed by extensive washing with binding buffer to remove residual detergent and contaminating bacterial proteins and the recombinant protein is eluted as usual. Using this technique, LPS reduction rates of at least 1000 fold were obtained for several GST-

106

Chapter 4 and His-tagged recombinant proteins (Reichelt et al., 2006; Zimmerman et al., 2006). A significant issue with other LPS reduction methods is the loss of recombinant protein, occurring through non-specific interactions with anion-exchangers, LPS-affinity resins or detergent (Table 1.3). However, using the detergent wash method, protein recovery rates close to 100% were achieved (Reichelt et al., 2006; Zimmerman et al., 2006). Presumably, the specific high-affinity interactions of recombinant protein with chromatography beads reduce the potential for protein loss through non-specific interactions. In addition, no apparent loss in protein integrity or functionality was observed in these two studies.

To produce recombinant GST and GST-CD9 with a reduced LPS content, these two constructs were expressed in E. coli Rosetta-gami strain, as previously and the original method of Reichelt et al. (2006) was followed to reduce the LPS (Chapter 2, section 2.2.7). Briefly, crude cell lysate was divided into two equal aliquots and one aliquot was washed with a 0.1% v/v solution of Triton X-114/PBS whilst the other one was treated with PBS only. Recombinant protein was eluted, dialysed and analysed by SDS- PAGE and Western blotting where no obvious differences were detected between detergent-treated (GST RX and GST-CD9 RX) and untreated protein (GST R and GST- CD9 R).

The LPS content of these proteins was determined using the LAL chromogenic assay, as outlined in the section above. As seen in Table 4.3, detergent washing reduced the LPS content of GST R and GST-CD9 R 20- and 9- fold, respectively. These clearance rates compare poorly with those reported previously (Reichelt et al., 2006; Zimmerman et al., 2006). Reichelt et al. (2006) speculate that at a temperature below its cloud point (22˚C), Triton X-114 dissociates LPS from the recombinant protein and drags it along the column. Recombinant tetraspanins tend to be hydrophobic in nature, binding to plastic surfaces including the sinter used in column purification (unpublished observations by our group). Thus to avoid protein loss, a batch purification method was followed throughout this study. Using this technique, chromatography beads are briefly incubated with detergent to dissociate LPS, then the beads are pelleted by centrifugation and the supernatant is supposed to contain the detergent and LPS. This process may not efficiently remove detergent and therefore

107

Chapter 4

LPS from the solution. However, the level of reduction achieved here was sufficient for the purposes of this study as will become apparent.

LPS (ng/ml in 500nM protein solution) Recombinant Fold reduction in protein Washed with Triton LPS levels Not washed X-114 GST R 3.25 0.16 ~20 GST-CD9 R 12.40 1.40 ~9

Table 4.3 Reduction of recombinant protein LPS content using a detergent wash method.

4.2.4 ELISA analysis of recombinant GST-CD9 proteins The reactivity of GST-CD9 SH, GST-CD9 R and GST-CD9 RX (produced in E. coli Rosetta- gami strain and washed with Triton X-114) with three different antibodies was tested using Enzyme-linked immunosorbent assay (ELISA). As shown in figure 4.4, the three proteins showed almost identical patterns of reactivity with the conformation-

sensitive monoclonal anti-CD9 antibody, 602.29 (similar EC50 values were obtained,

EC50=dosage of protein which elicits a half-maximal antibody response). This antibody only recognises correctly folded CD9 protein (Higginbottom et al., 2003; R. Hulme, PhD thesis) suggesting that all three protein preparations contain the same amount/proportion of correctly folded protein with a native conformation and that detergent treatment of GST-CD9 RX does not compromise protein structure/folding.

GST-CD9 SH showed a slightly lower reactivity (higher EC50 value) with anti-GST polyclonal antibody than the other two proteins. This may be due the presence of lower levels of truncated or partially degraded GST-CD9 in this preparation. Minimal binding was observed to an isotype-matched control antibody showing that binding to anti-GST and anti-CD9 antibodies is specific.

108

Chapter 4

GST-CD9 SH GST-CD9 R GST-CD9 RX EC50 – Anti-CD9 0.65µg 0.69µg 0.67µg

EC50 – Anti-GST 1.85µg 1.71µg 1.75µg

Figure 4.4 ELISA method to compare reactivity of GST-CD9 constructs with various antibodies.

The assay was carried out as described (Chapter 2, section 2.2.1.10). Briefly, purified GST-CD9 SH, R and RX proteins were serially diluted in ELISA binding buffer and used to coat triplicate wells of an ELISA microtitre plate. Following overnight incubation, wells were treated with 1) anti-CD9 monoclonal conformation-sensitive antibody (602.29, 2µg/ml) followed by an anti-mouse IgG HRP-linked antibody (A9044, 1in5000), 2) anti-GST HRP-linked antibody (A7340) or 3) an isotype control antibody (JC1) followed by the anti-mouse IgG HRP-linked antibody (1in5000). Plates were developed with the HRP substrate TMB and the absorbance (OD450nm) was determined using an ELISA plate reader. Data are mean ± SEM from one experiment carried out in duplicate. A four-parameter (variable slope) dose-response curve was fitted to each data set with bottom constrained to a value greater than zero. EC = protein amount which provokes antibody 50 binding half-way between the minimum and maximum values. OD=optical density.

4.2.5 Effect of LPS reduction on recombinant protein activity in the MGC assay Recombinant GST RX and GST-CD9 RX were tested in the MGC assay to find out if a direct reduction in LPS content would have an effect on the inhibitory activity of recombinant protein preparations. As shown in figure 4.5, recombinant GST RX and GST-CD9 RX inhibited MGC formation as measured by the fusion index (P<0.05), but

109

Chapter 4 the level of inhibition of these two constructs was significantly reduced compared to GST R and GST-CD9 R (P<0.05 and P<0.001, respectively). Thus, LPS reduction has a significant effect on the apparent activity of recombinant proteins in the MGC assay. This is in line with findings by several other studies in which ‘low-LPS’ protein preparations apparently lost their activity in cell-based assays (Gao and Tsan, 2003a; Gao and Tsan, 2003b, Pepys et al., 2005; Bausinger et al., 2002; Dudley et al., 2003; Hreggvidsdottir et al., 2009). This was not due to an impairment in recombinant protein structure (figure 4.4) nor activity since GST-CD9 RX had the same inhibitory effect as GST-CD9 SH and GST-CD9 R in an independent functional assay whilst GST SH, R and RX had no effect (Chapter 5, section 5.2.1).

120 * *** 100

) *

l *

e r

d 80

t ****

n n

i ****

o n

c 60

f i

u 40 %

F ( 20 0 A R X R X n T R 9 R o S T D 9 C S G -C D G T -C S T G S G

120 n s

100 *

)

s r

n G o ***

c 60

M f

****

g 40

( A 20

A R X R X n T R 9 R o S T D 9 C S G -C D G T -C S T G S G

Figure 4.5 Effect of LPS reduction on recombinant protein activity in the MGC assay.

The MGC assay was carried out as described previously. Recombinant proteins were tested at 500nM. ‘RX’ refers to recombinant protein produced in Rosetta-gami strain and washed with Triton X-114 to reduce LPS contamination. Data are means ± SEM, normalised to the Con A control (=100%). Statistical significance was determined using an ordinary ANOVA with a Holm-Sidak post-test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, ns, not significant. N=3+, n=7+ 110

Chapter 4

4.2.6 Effects of pure LPS on MGC formation Monocytes, the precursors of multinucleated giant cells (MGC) are known to be acutely sensitive to bacterial contaminants (Chapter 1, section 1.3.2). However, data on the effects of bacterial contamination on Concanavalin A-induced MGC formation is limited and conflicting. Takashima et al. (1993) reported that LPS inhibited Con A- induced MGC formation in a dose-dependent manner but did not provide details of the LPS concentrations used nor the levels of inhibition achieved. Several studies have reported that LPS by itself is incapable of inducing monocyte fusion (Abe et al., 1984; Enelow et al., 1992; Lazarus et al., 1990; Takashima et al., 1993). In contrast to this, previous data from our group showed that LPS doses from 1ng/ml to 2µg/ml had no effect on Con A-induced MGC formation (reported findings in Parthasarathy et al., 2009; R. Hulme, PhD thesis). At higher doses, LPS induced MGC formation by itself (V. Parthasarathy, PhD thesis) and enhanced the effects of Con A-induced MGC formation (reported in Parthasarathy et al., 2009). Similarly, Yanagishita et al. (2007) had observed that co-treatment of the murine macrophage-like cell line RAW246.7, resulted in enhanced MGC formation compared to Con A treatment alone.

Data presented in figure 4.5, suggests that LPS inhibits Con A-induced MGC formation since a reduction in the LPS content of recombinant proteins results in a reduction in inhibitory activity. Given these contrasting results, it was important to establish definitively whether pure LPS itself can inhibit Con A-induced MGC formation at concentrations found in the recombinant protein preparations used in this study. To this end, highly purified tissue culture-grade LPS from E. coli strain O111:B4 (Sigma, L4391) was tested in the MGC assay. This is the same strain used as a standard in the LAL assay and also by other researchers whose results have been mentioned here (Takashima et al., 1993; Takeda et al., 2003; Parthasarathy et al., 2009; R. Hulme, PhD thesis). LPS was serially diluted in RPMI-1640 media (10% FCS, 1% Penicillin- Streptomycin) and added to Con A-containing media to give solutions containing 10µg/ml Con A plus LPS concentrations from 10µM to 100pM. The purpose of this wide concentration range was to allow construction of dose response curves from which inhibition values could be interpolated (required for later analyses). The MGC assay was carried out as previously using monocytes from several donors. The results are shown in figures 4.6 to 4.8.

111

Chapter 4

1A) 1B)

1 2 0 1 2 0

1 0 0 1 0 0

) )

s r

8 0 r 8 0

n n

G o

6 0 o 6 0

4 0 g 4 0

( ( A 2 0 2 0 0 0 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5

[L P S ], lo g 1 0 M [L P S ], lo g 1 0 M

2A) 2B)

1 2 0 1 2 0 1 0 0 1 0 0

)

i o

r 8 0

s r

d 8 0

o G

6 0 o n

c 6 0

4 0 g

u 4 0

% F

(

( A 2 0 2 0 0 0 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5

[L P S ], lo g 1 0 M [L P S ], lo g 1 0 M

3A) 3B)

1 2 0 1 2 0

1 0 0 1 0 0

) e

)

s r

r 8 0 8 0

G o o 6 0

6 0 c

. o

4 0 g 4 0

( (

A 2 0 2 0 0 0 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5

[L P S ], lo g 1 0 M [L P S ], lo g 1 0 M

4A) 4B)

1 2 0 1 2 0

1 0 0 1 0 0

)

l x

r s

8 0 r 8 0

o G

6 0 o n

c 6 0

4 0 g 4 0

( ( A 2 0 2 0 0 0 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 [L P S ], lo g 1 0 M [L P S ], lo g 1 0 M

112

Chapter 4

Figure 4.6 Dose-dependent inhibitory effects of LPS on Con A-induced MGC formation.

The MGC assay was carried out as described. Briefly, adherent monocytes were incubated for 72 hours in medium containing 10µg/ml Con A alone (control) or Con A plus LPS in the concentration range 10µg/ml to 100pg/ml (1µM to 10pM). Results are means ± SEM, normalised to the Con A control (=100%). A four-parameter (variable slope) dose-response curve was fitted to graphs 1A, 2B, 3A, 3B, 4A and 4B using the least squares fitting method, with the top constrained to 100. Similarly, a three-parameter dose-response curve was fitted to graphs 1B and 2A. Graphs 1A/B, 2A/B and 3A/B are results obtained from 3 independent

donors (n=2+) whilst graphs 4A/B are the combined results of two independent donors (n=2+) due to missing values. To work out molar concentrations, the molecular weight of an LPS monomer was assumed to be 10KDa (Petsch and Anspach, 2000).

In all donors, LPS inhibited MGC formation in a dose-dependent manner as judged by the fusion index and average MGC size (figure 4.6). LPS appeared to be active in this assay over a wide range of concentrations (100pg/ml to 10µg/ml) with half-maximal inhibitory concentrations (IC50) of ~10ng/ml and ~0.6ng/ml for the fusion index and average MGC size, respectively. Thus LPS has a more dramatic effect on MGC size as indeed do other inhibitory reagents such as recombinant tetraspanins (This study and unpublished observations by our group). These IC50 values are comparable to RANKL-induced osteoclast formation from murine bone marrow-derived macrophages (BMMs), where an IC50 of approximately 3ng/ml was obtained (re- analysis of original data by Zou and Bar-Shavit (2002) supplied in the form of a bar chart graph).

Data shown in figure 4.6 were combined and presented as bar charts for statistical analysis (figure 4.7). At the lowest concentration tested, 100pg/ml, LPS significantly inhibited Con A-induced MGC formation as measured by the average MGC size (P<0.01). At 1ng/ml (100pM), LPS had a significant inhibitory effect on both parameters. This is somewhat expected given the acute sensitivity of monocytes to LPS and its documented effects on monocyte differentiation in other assays such as osteoclast formation (Chapter1, section 1.3.2). However, this result is contrary to previous data from our group where LPS doses as high as 2µg/ml had no effect in this assay (Parthasarathy et al., 2009; R. Hulme, PhD thesis).

113

Chapter 4

At the highest concentration tested (10µg/ml), LPS caused a 56% reduction in the fusion index and a 59% reduction in the average MGC size. This is in contrast to RANKL-induced osteoclast formation where 10ng/ml LPS totally inhibited fusion of human and murine monocytes (Liu et al., 2009; Takami et al., 2002).

120

100

) l

x ***

o e

r ****

d 80 t

n ****

c 60

o f ****

u 40

% F ( 20

0 l l l l l l A m m m m m m n / / / / / / o g g g g g g C p n n n   0 1 0 0 1 0 0 1 0 1 1 1

[L P S ]

120

100

)

z i

o **

s r

t **** C

n ****

G o

c 60 ****

f ****

o ****

g 40

( A 20

0 l l l l l l A m m m m m m n / / / / / / o g g g g g g C p n n n   0 1 0 0 1 0 0 1 0 1 1 1

[L P S ]

Figure 4.7 Significant inhibitory effects of LPS on MGC formation.

Data obtained in figure 4.6 were combined and presented as bar charts. Data are means ± SEM, normalised to the Con A control (=100%). Statistical significance was determined using ordinary one-way analysis of variance (ANOVA) with the Holm-Sidak post-test. **P<0.01, ***P<0.001, ****P<0.0001. N=5, n=8+.

114

Chapter 4

(A) Unstimulated (B) Con A

(E) Con A + 10ng/ml LPS (F) Con A + 100ng/ml LPS

(G) Con A + 1µg/ml LPS (H) Con A + 10µg/ml LPS

115

Chapter 4

Figure 4.8 Immunofluorescence images of human monocytes treated with various doses of LPS.

Monocytes were stimulated with Con A alone (B) or Con A plus 100pg/ml to 10µg/ml LPS (C-H). Cells were fixed, stained and imaged as described previously. The nuclei are shown in red and the membrane is shown in green. Scale bar=100µm.

4.2.7 Contribution of bacterial contaminants to recombinant protein inhibitory activity in the MGC assay Data presented in the previous sections showed that pure LPS has inhibitory effects on MGC formation and that reducing the LPS content of recombinant proteins also reduced their inhibitory activity in the MGC assay. To determine more precisely the contribution of LPS to recombinant protein activity, the LPS dose-response curves obtained in figure 4.6 were used to predict the inhibition levels of each protein solution if its activity was solely due to its LPS content.

In addition, the activity of native and boiled recombinant proteins was compared in the MGC assay (figure 4.9). LPS is an extremely heat-stable molecule whilst many proteins have been shown to be inactivated by incubation at high temperatures (Gamble et al., 1985; Smiley et al., 2001; Vogl et al., 2007; Drexler et al., 2009). This differential sensitivity to heat denaturation can also be exploited to discriminate between protein and LPS activity in a protein solution. In several studies, it was found that whilst protein activity was abolished by boiling, LPS activity remained intact therefore the residual activity of boiled protein solutions may be attributed to LPS or other heat-stable contaminants (Bulut et al., 2005; Vogl et al., 2007; Okamura et al., 2001).

As shown in figure 4.9, there was no significant difference between the inhibitory activity of native and boiled protein, as assessed by the fusion index or the average MGC size. This suggests that the inhibitory activity is due to a heat-stable component or components in the protein solution and not the recombinant protein itself. There was a significant difference between the actual and predicted inhibitory values obtained for GST SH (P<0.05) and GST-CD9 SH (P<0.001) but not for the other four proteins tested here. Heat treatment of GST SH and GST-CD9 SH reduced the inhibitory activity of these two constructs albeit not significantly suggesting that

116

Chapter 4 proteins produced in E. coli SHuffle strain may contain other heat-sensitive contaminants at high enough concentrations to have a noticeable effect on monocyte fusion.

Native GST Native GST-CD9

Boiled GST Boiled GST-CD9

Predicted response (based on LPS content)

120

100

)

x o

e * *** r

d 80

o n

c 60

u 40

F ( 20

H R X H R X S T R S 9 R T S T 9 D 9 S S G D -C D G G -C T -C T S T S G S G G

120

100

)

s r

G o

c 60

g 40

( A 20

H R X H R X S T R S 9 R T S T 9 D 9 S S G D -C D G G -C T -C T S T S G S G G

Figure 4.9 Contribution of bacterial contaminants to recombinant protein activity in Con A-induced MGC formation.

Data shown for native proteins are those presented in previous graphs in this chapter. For boiled GST and GST-CD9, proteins were boiled for 10 minutes at 95˚C before testing in the MGC assay at 500nM, as previously. For the predicted response based on LPS content, the dose- response curves in figure 4.6 were used to interpolate values for the inhibitory activity of each protein assuming its activity was solely due to its LPS content. Results were analysed by ordinary two-way ANOVA followed by a Holm-Sidak post-test. *P<0.05, **P<0.01, ***P<0.001. Native protein, N=2+, n=6+, Boiled protein, N=2+, n=5+, Predicted response, N=4, n=4.

117

Chapter 4

Complete inactivation of the recombinant protein by boiling must be confirmed by independent methods (Lewthwaite et al., 2001). In a previous study, boiling of GST- CD9EC2 for 5 minutes at 90˚C abolished its effect on cell surface dynamics (Barreiro et al., 2008). Similarly in this study, heat-treated but not native GST-CD9EC2 constructs were found to be inactive in the bacterial adhesion assay (Chapter 5, figure 5.1). However, protein inactivation/denaturation could not be further confirmed by Western blotting. As shown in figure 4.10 A, a conformation-sensitive anti-CD9 antibody (602.29) reacted with native but not boiled GST-CD9 SH, R and RX. But an anti-GST antibody which recognises linear as well as conformational epitopes also failed to recognise boiled proteins (figure 4.10 B). Since boiling results in the formation of visible protein aggregates it is likely that this material is not effectively solubilised by SDS loading buffer and/or fails to enter the polyacrylamide gel. Similarly, boiled material was not recognised by anti-CD9 or anti-GST antibody in an ELISA experiment, perhaps because it fails to adsorb effectively to the ELISA plate.

A) Anti-CD9 antibody B) Anti-GST antibody

KDa M SH R RX SH R RX KDa M SH R RX SH R RX

250 250 150 150 100 100 75 75 Dimer 50 50 37 37 Monomer 25 25 20 20

Native Boiled Native Boiled

Figure 4.10 Analysis of native and boiled GST-CD9 proteins by Western blotting.

Stock solutions of purified GST-CD9 SH, R and RX were boiled for 10 minutes at 95˚C. Native and boiled proteins were incubated with 2× non-reducing SDS-PAGE loading buffer for 8 minutes at 80˚C. Approximately 0.5µg of protein was separated by SDS-PAGE and subjected to Western blotting. Blots were probed with 1µg/ml anti-CD9 antibody (602.29) followed by an anti-mouse HRP-linked antibody or with anti-GST HRP-linked antibody.

118

Chapter 4

An unexpected finding was the differential migration of GST-CD9 SH compared to GST- CD9 R and GST-CD9 RX (figure 4.10 A and B). Staining with anti-CD9 antibody suggests that non-reduced GST-CD9 SH is mostly in oligomeric form. However, this does not affect protein reactivity with anti-CD9 antibody in ELISA (figure 4.4) nor functional activity in the bacterial adhesion assay (Chapter 5, figure 5.1) therefore it was assumed that GST-CD9 SH, R and RX preparations all contain the same amount of correctly folded and therefore functional material.

4.2.8 Effect of polymyxin B on the inhibitory activity of LPS and LPS- contaminated recombinant proteins LPS inhibitors such as polymyxin B have been used previously to provide direct evidence that LPS is the contaminant responsible for observed experimental effects (Lewthwaite et al., 2001; Wakelin et al., 2008). As mentioned in section 1.3.3.3 (Chapter 1), polymyxin B is a cationic cyclic peptide which binds to LPS through both hydrophobic and electrostatic interactions, although hydrophobic interactions with the lipid A region of LPS may predominate (Tsubery et al., 2002). Pre-incubation of LPS with polymyxin B blocks LPS binding to human monocytes thus preventing its pro-inflammatory effects such as TNF-α production (Gallay et al., 1993). Polymyxin B is potentially toxic to cells at high concentrations and may have unintended effects of its own on monocyte differentiation (Bi et al., 2001). The optimal working concentration of polymyxin B has been determined to be approximately 10µg/ml, in a number of studies (Gao et al., 2003b, Okamura et al., 2001; Tsuzuki et al., 2001).

At 10µg/ml, polymyxin B by itself did not have a measurable effect on MGC formation (figure 4.11). Pre-incubation of a high concentration of LPS (500ng/ml) with 10µg/ml polymyxin B completely abrogated its inhibitory effects on MGC formation. In some studies, polymyxin B was pre-incubated with cells before addition of LPS (Gao and Tsan, 2003a; Bausinger et al., 2002) and this method also efficiently neutralised the inhibitory effects of LPS (figure 4.11). Next, the effect of polymyxin B on recombinant protein activity in the MGC assay was tested using both polymyxin B incubation methods mentioned. Neither method was effective at neutralising the inhibitory activity of recombinant protein preparations. Instead, it appeared to enhance the inhibitory activity of all recombinant proteins such that too few giant cells were present in the culture to determine the rate of fusion. This result is surprising given

119

Chapter 4 that polymyxin B alone had no effect on MGC formation and fully neutralised the effects of a high dose of LPS (figure 4.11). The implications of this finding are discussed further in section 4.3. In light of all the data presented in this chapter, further experiments with other recombinant tetraspanin proteins (Chapter 3) were not pursued since it was deemed likely that the effects of the control proteins in the Con A-induced MGC formation assay are due to contaminants.

**** 120 ****

100

)

e r

d 80

t n

n ****

o n

c 60

o f

u 40

F ( 20

A X S S S P P P n M L L L o P + n C e X h M t P X M P

120 *

100

)

o s r **

G o

c 60

g 40

( A 20

A X S S S P P P n M L L L o P + n C e X h M t P X M P

Figure 4.11 Effect of polymyxin B sulphate on the inhibition of MGC formation by LPS.

PBMC were isolated and adherent monocytes were induced to form MGC using 10µg/ml Con A as previously. Polymyxin B sulphate powder (P4932, Sigma) was reconstituted in sterile HBSS. To test the effects of polymyxin B and LPS, monocytes were incubated with 10µg/ml Con A plus 10µg/ml polymyxin B (PMX) or 500ng/ml LPS (LPS). To neutralise LPS activity, 10µg/ml polymyxin B was incubated with 500ng/ml LPS for 30 minutes at 37˚C with frequent agitation before being added to cells (PMX + LPS). Alternatively, 10µg/ml polymyxin B sulphate was pre- incubated with adherent monocytes for 30 minutes at 37˚C before the addition of Con A plus LPS. Fusion indices and average MGC size were determined as previously. Data were normalised to the Con A control and plotted as means ± SEM. Statistical significance was determined using an ANOVA test with a Holm-Sidak post-test. *P<0.05, **P<0.01, ****P<0.0001, N=3, n=8+.

120

Chapter 4

4.3 Discussion The aim of this chapter was to test the effects of several new recombinant GST-tagged tetraspanin EC2 domains (Chapter 3) on Con A-induced MGC formation. Recombinant GST and GST-CD9 EC2 have been previously used as negative and positive controls in this assay, respectively (Takeda et al., 2003; Parthasarathy et al., 2009; R. Hulme., PhD thesis). Some of the recombinant tetraspanins to be tested had been produced in a new bacterial strain, E. coli SHuffle, therefore it was important to compare the activities of the controls proteins produced in SHuffle (denoted with SH) to those produced in Rosetta-gami (denoted with R) which had been tested previously (R. Hulme, PhD thesis). Contrary to previous results obtained by our group (Parthasarathy et al., 2009; R. Hulme. PhD thesis) and Takeda et al. (2003), GST (SH and R) significantly inhibited MGC formation and the activity of GST SH and GST-CD9 EC2 SH was higher than those produced in Rosetta-gami strain. It is notable that the GST encoded by the pGEX-KG vector has been shown to be an active enzyme, catalysing the conjugation of the synthetic compound 1-Chloro-2,4-dinitrobenzene (CDNB) to the free thiol group of glutathione (Kaplan et al., 1997). But it is unkown whether GST enzymatic activity would affect monocyte function. In any case, polyhistidine-tagged recombinant tetraspanins also inhibited MGC formation (Parthasarathy et al., 2009).

Also contrary to previous findings (Parthasarathy et al., 2009; R. Hulme. PhD thesis), pure LPS was found to inhibit Con A-induced MGC formation in a dose-dependent manner, as reported by Takashima et al. (1993). This result is not unexpected given the well-documented sensitivity of human and murine monocytes to TLR ligands such as LPS and other bacterial contaminants (Chapter 1, section 1.3.2). Indeed LPS is utilised in many studies to ‘prime’ or activate macrophages to become mature pro- inflammatory macrophages (also known as M1 or classically activated) (Martinez et al., 2006). TLR ligands have also been found to inhibit the differentiation of monocytes into pre-osteoclasts but enhance pre-osteoclast fusion into mature multinucleated osteoclasts (Takami et al., 2002; Zou and Bar Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009; Suda et al., 2002; Itoh et al., 2003; Xing et al., 2011).

To elucidate the contribution of LPS to recombinant protein activity, ‘low LPS’ versions of GST R and GST-CD9 EC2 R were obtained using a detergent wash method (Reichelt

121

Chapter 4 et al., 2006; Zimmerman et al., 2006). The LPS content of all constructs was determined using the LAL chromogenic assay. LPS is known to be highly resistant to heat denaturation whilst many proteins are heat-sensitive, therefore the activity of boiled recombinant protein solutions was also determined in the MGC assay. It was found that the inhibitory activity of different GST and GST-CD9 EC2 constructs in the MGC assay correlated positively with their LPS content, despite the demonstration of almost identical activity in other independent assays (Figure 4.4 and Chapter 5, figure 5.1). Further, the activity of native and boiled proteins did not differ significantly in the MGC assay even though boiled GST-CD9 EC2 was found to be inactive in the bacterial adhesion assay (Chapter 5, figure 5.1). Thus neither GST nor GST-CD9 EC2 appeared to have intrinsic activity in the MGC assay in this study.

The effects of LPS, but not recombinant GST and GST-CD9 EC2 constructs, were completely neutralised by the LPS inhibitor polymyxin B. It should be noted that LPS, in its natural state, is tightly associated with bacterial proteins and most of the LPS in a recombinant protein preparation is likely to be bound to the protein itself (Reichelt et al., 2006). It has been suggested previously that polymyxin B cannot effectively bind and neutralise protein-bound LPS (Bausinger et al., 2002; Erridge, 2010). The biophysical properties of the recombinant protein itself may determine the effectiveness of polymyxin B. An extensive literature search to find instances where the activity of contaminating LPS in a recombinant protein preparation was efficiently inhibited by polymyxin B yielded only two results. LPS activity was fully neutralised in recombinant autolysin and sonic hedgehog preparations (Lewthwaite et al., 2001; Wakelin et al., 2008). Interestingly, both these proteins are positively charged at neutral pH, with theoretical isoelectric points (pI) of 9.83 and 8.92 whilst the pI values for GST and GST-CD9EC2 are 6.1 and 5.96, respectively. Negatively charged proteins may compete with LPS for binding to the positively charged amine groups of polymyxin B (Landmann et al., 1995) through electrostatic interactions. However, it is unclear why polymyxin B addition to recombinant proteins should result in extremely low fusion rates when the molecule by itself has no effect on MGC formation (figure 4.8). It is possible that polymyxin B binding to proteins releases the protein-bound LPS into solution resulting in higher LPS activity. Instead of targeting the LPS molecule itself (PMX binds LPS with high stoichiometry), LPS receptors on the cell surface may be targeted using inactive LPS variants, synthetic LPS analougs or anti-LPS receptor

122

Chapter 4

(TLR4/CD14) antibodies (Fenton and Glockenbock, 1998; Alexander and Reitschel, 2001; Lewthwaite et al., 2001; Wakelin et al., 2008; Giambartolomei et al., 2004).

Even though direct evidence could not be obtained, the fact that LPS contamination levels correlate closely with the observed inhibitory activity of recombinant proteins preparations, that directly reducing LPS contamination reduces recombinant protein activity in the MGC assay and that this activity is heat-resistant supports the notion that LPS is responsible for some if not all these effects.

The discrepancy between results obtained in this study and those previously reported by our group (Parthasathy et al., 2009; R. Hulme, PhD thesis) was also addressed in the course of these experiments. A number of changes had been made to the assay protocol but these were not thought to be responsible for the differences observed. For example, PBMC were cultured in plastic 96-well tissue culture plates instead of glass-bottomed rectangular-shaped LabTek chambered slides with (Parthasarathy et al., 2009; R. Hulme, PhD thesis). As shown in figure 2.2 (Chapter 2), the circular shape of the well allows adherent monocyte cell density to be evenly reduced across the middle of the well by washing. The same area can then be imaged using an automated set up (Chapter 2, section 2.3.4). Not only is this system much more convenient but it eliminates the potential for experimenter bias in choosing fields to count nuclei. Monocyte adherence to tissue culture surfaces is thought to be a pre-requisite for monocyte fusion (Helming and Gordon, 2009) and variations in tissue culture surface chemistry were observed to influence monocyte adhesion and FBGC formation (McNally and Anderson, 2002). However, the same fusion rates were achieved when monocytes from the same donor were cultured in 96-well plates compared to LabTek slides (data not shown). Takeda et al. (2003) had also performed the MGC assay in 96 well plastic plates.

Variations in the PBMC isolation procedure may also have contributed to these differences, although great care was taken to emulate the same conditions. Given the documented sensitivity of freshly isolated monocytes to LPS (Gessani et al., 1993; Gallay et al., 1993), one would imagine that very specific conditions are required to obtain LPS-insensitive monocytes (Parthasarathy et al., 2009; R. Hulme, PhD thesis),

123

Chapter 4 conditions which could not be replicated here despite attempting several different variations of the PBMC isolation procedure.

A minor modification to the analysis method employed by previous workers was the normalisation of all data to the Con A control which was taken to be 100%. This approach was taken to minimise the influence of donor variability in fusion rates obtained with the Con A control and to simplify quoting of statics such as the percentage reduction in average MGC size by GST-CD9 R. Again, non-normalised data did not differ significantly to the normalised (data not shown).

Of interest is the fact that the activity of GST-CD9 EC2 in this study was very similar to that observed previously (Table 4.1) whereas GST which was found to inhibit MGC formation but was found to be inactive previously. Considering that protein yield in this study was inversely correlated with LPS content (Table 4.2), a plausible explanation may be that the GST obtained in previous studies was simply of higher yield and therefore lower LPS content (<100pg/ml) thus it apparently had no activity. The purified proteins used in previous studies in our lab are no longer available for testing in the LAL chromogenic assay but LPS contribution to the activity of GST-CD9 EC2 could have been ruled out simply by boiling which completely inactivates the protein (Chapter 5, figure 5.1). A detailed examination of the mechanisms by which LPS inhibits Con A-induced MGC formation is beyond the scope of this study however some general observations are made in the general discussion (Chapter 6).

124

Chapter 5

Chapter 5: Recombinant tetraspanin protein activity in other functional assays

5.1 Introduction

Recombinant tetraspanin EC2 proteins have been of value in deciphering tetraspanin functions in different areas of cell biology, as outlined in Table 1.2 (Chapter 1). In Chapter 4, it was shown that the inhibitory activity of recombinant tetraspanins on Con A-induced multinucleated giant cell formation in human peripheral blood monocytes is likely to be due to the presence of contaminants. Protein contamination with bacterial lipopolysaccharide (LPS) and other contaminants such as lipoproteins is an unavoidable consequence of using bacteria as protein expression hosts (Wakelin et al., 2006). Monocyte sensitivity to bacterial contaminants is due to high levels of expression of Toll-like receptors (TLRs) and their associated proteins which mediate binding and recognition of microbial contaminants (Takami et al., 2002; Visintin et al., 2001; Kumar, 2009). However, considered in the context of previous published and unpublished data by our group, the results obtained in Chapter 4 were unexpected (discussed in Chapter 4, section 4.3). Thus we wanted to confirm those findings in other monocyte fusion systems so that more concrete conclusions could be drawn. It was also of interest to determine if the recombinant tetraspanins had any intrinsic activity in the bacterial adhesion assay system which is expected to be unresponsive to bacterial contaminants.

125

Chapter 5

5.2 Results 5.2.1 Recombinant protein inhibition of bacterial adhesion to epithelial cells The emergence of antibiotic-resistant strains has spawned research into alternative mechanisms of preventing bacterial infections. To cause septicaemia, mucosal bacteria such as Neisseria meningitides must first cross the epithelial barrier that lines mucosal surfaces. Bacterial proteins interact with epithelial cell surface proteins to mediate initial attachment (adherence) to the cell followed by internalisation. Internalised bacteria can then multiply and spread (Capecchi et al., 2005). Thus preventing bacterial adhesion is a potentially useful way of reducing infection rates. In in vitro studies, it had been demonstrated that pre-incubation of epithelial cells with different anti-tetraspanin antibodies or GST-tagged recombinant tetraspanin EC2 domains, specifically inhibited bacterial adhesion to the human epithelial endometrium adenocarcinoma cell line, HEC-1-B cells (Green et al., 2011). In addition, short synthetic peptide analogues of the human CD9 EC2 domain also inhibited bacterial adhesion (unpublished observations, D. Cozens, Department of Infection and Immunity, University of Sheffield Medical School). Being of non-immune origin, HEC-1- B cells are unlikely to be responsive to bacterial contaminants. This presented an opportunity to the test the functionality of the recombinant tetraspanin proteins produced in this study (Chapter 3).

5.2.1.1 Effect of controls proteins and LPS on bacterial adhesion The bacterial adhesion assay was carried out by Mr Daniel Cozens (University of Sheffield, Medical School) as described previously (Green et al., 2011) whilst data analysis was performed by the author. As shown in figure 5.1 A, 500nM recombinant GST-CD9 SH, R and RX all significantly inhibited bacterial adhesion to HEC-1-B cells. Since these proteins have different levels of LPS contamination, this suggests that LPS does not contribute to adhesion inhibition. GST SH, R and RX had no inhibitory activity suggesting that the observed activity is attributable to the CD9 EC2 domain. It is also interesting that despite enhanced oligomer formation (figure 4.10, Chapter 4), GST- CD9 SH has the same level of activity in this assay as GST-CD9 R and GST-CD9 RX. Doses of LPS as high as 100ng/ml had no effect on bacterial adhesion (figure 5.1 B) whilst boiling of GST-CD9 SH and GST-CD9 R completely abolished inhibitory activity (figure 5.1 A), suggesting that this activity is due to a heat-sensitive component.

126

Chapter 5

120

100

)

a r

r 80 t

n *** *** ***

o i

c 60

h 40

% (

A 20

0 ) ) y H R X R X l R H R (B (B n S T S 9 o T S T 9 D 9 H R a S G S C D S i G G D - C 9 d -C T - 9 D e S T D C T S C - M S G - T G G T S S G G

120 n s n s n s

100

)

e l

a r

r 80

o i

c 60

h 40

( A 20 0 l l y l l m m m n / / / o g g g a n n n i 1 0 0 d S 1 0 e P S 1 L P S M L P L

Figure 5.1 Effect of GST and GST-CD9 proteins and LPS on bacterial adhesion to HEC-1-B cells.

The bacterial assay was carried out by Mr Daniel Cozens (University of Sheffield, Medical School) as described in section 2.3.3 (Chapter 2). 500nM recombinant protein or 1-100ng/ml LPS were incubated with cells for 30 minutes before washing and incubation of cells with bacteria (Neisseria meningitides ). Data analysis was performed by the author. Data are means ± SEM of two independent experiments performed in quadruplicate and normalised to the media only control. Adhesion rate is defined as the number of adherent and internalised bacteria per cell. Statistical significance was determined using ordinary ANOVA followed by a Holm-Sidak post-test. ***P<0.001, ns, not significant. Abbreviations: ‘SH’, produced in E. coli SHuffle strain, ‘R’, produced in E. coli Rosetta-gami strain, ‘RX’ produced in Rosetta-gami strain and washed with Triton X-114 detergent to reduce LPS contamination, LPS, bacterial lipopolysaccharide, (B), 500nM protein inactivated by boiling at 95˚C for 10 minutes.

127

Chapter 5

5.2.1.2 Effect of a panel of recombinant tetraspanin on bacterial adhesion

In the previous section, it was demonstrated that control reagents did not adversely impact upon the bacterial adhesion assay therefore the effects of six new recombinant tetraspanin EC2 domains on bacterial adhesion could be determined. The bacterial adhesion assay was carried by Mr Daniel Cozens (University of Sheffield Medical School) as described in figure 5.1. 500nM recombinant protein was pre-incubated with HEC-1-B cells for 30 minutes before infection with N. meningitides. Data analysis was performed by the author. As shown in figure 5.2, four out of the six recombinant tetraspanins tested (not including GST-CD9 R) significantly inhibited bacterial adhesion to HEC-1-B cells. It would be interesting to find out if recombinant protein inhibition correlates with native tetraspanin expression levels on the surface of HEC- 1-B cells.

120

100

)

a r

r 80 t

n n ****

**** **** o

o **** **** i

c 60 ****

f e

h 40

( A 20 0

ly H R H R R R R R H R n S S 9 2 3 3 2 S 2 T 3 o T S 9 D n n 1 8 1 a S G C a a n D 3 n i G D - p p a C n a d C T s s p - a - S T T s T p p e T - - T S s s M S G T T - T -T G S S T G - T G G S T S G S G G

Figure 5.2 Testing of a panel of recombinant tetraspanins in the bacterial adhesion assay.

GST-Tspan2 R was tested once in triplicate whilst the other proteins were tested in two or more independent experiments performed in triplicate or quadruplicate. Data are means ± SEM. Statistical significance was determined using ordinary ANOVA followed by a Holm-Sidak post-test. ***P<0.001, ****P<0.0001.

128

Chapter 5

Green et al. (2011) found that recombinant tetraspanins could inhibit the adherence of an additional four species of bacteria including Staphalococcus aureus. This suggested that recombinant proteins do not interact specifically with adhesion receptors on the cell surface nor with a specific bacterial protein involved in adhesion but rather disrupt the formation of complexes on the cell surface required for bacterial adhesion. The observation of inhibitory activity with four new tetraspanins in addition to ones already investigated by Green et al. (2011) supports this theory.

5.2.2 Recombinant protein activity in RANKL-induced osteoclastogenesis

The in vitro osteoclastogenesis assay offers an alternative method to investigate monocyte fusion (Chapter 1, Table 1.1). Osteoclasts are multinucleated giant cells derived from the fusion of monocytes at the bone surface. These cells are one half of the bone homeostasis system where osteoblasts continually deposit new bone and osteoclasts resorb old and damaged bone. The identification of the primary cytokines responsible for monocyte differentiation into osteoclasts has enabled research into osteoclast formation and function in vitro (Boyle et al., 2003). The assay involves incubation of adherent peripheral blood mononuclear cells (PBMCs) with the recombinant human cytokines, receptor activator of NF-κB ligand (RANKL) and macrophage colony stimulating factor (MCSF) for a period of 14-21 days, during which time monocyte proliferation, differentiation and multinucleation takes place. Aberrant osteoclastogenesis is implicated in several chronic human diseases such as rheumatoid arthritis therefore finding agents which can regulate osteoclast formation is of particular value.

Disruption of tetraspanin function using tetraspanin gene inactivation/deletion, RNA interference or anti-tetraspanin antibodies had indicated the involvement of tetraspanins CD9, CD81, Tspan13 and Tspan5 in osteoclast formation, as described in Chapter 1, section 1.2.3.4.4. Unpublished data by our group had shown that recombinant GST-CD9EC2 and GST-CD81EC2 enhanced osteoclastogenesis when tested at doses of 0.05nM and 0.05nM (R. Hulme, PhD thesis). At higher doses, recombinant protein treatment resulted in cell detachment. This finding seems intriguing particularly as GST-CD9EC2 inhibits Con A-induced MGC formation (Takeda et al., 2003; Parthasarathy et al., 2009). Therefore, we aimed to re-examine the role of recombinant proteins in the osteoclastogenesis assay.

129

Chapter 5

The assay was carried out as described (Chapter 2, section 2.3.2). It should be noted that mononucleated functional osteoclasts are also formed in the culture (Yagi et al., 2005) but we were primarily interested in osteoclast fusion therefore osteoclasts in this text refers only to multinucleated osteoclasts. As shown in figure 5.3 A, the fusion index in untreated monocytes and those treated with MCSF was close to zero whilst RANKL and RANKL/MCSF treatment resulted in fusion rates of ~11% and 75%, respectively. The average size of RANKL/MCSF-derived osteoclasts (figure 5.3 B) was relatively small (~7 nuclei per osteoclast), as expected (Vaananen and Laitalia- Leinonen, 2008; Ha et al., 2011). Since minimum osteoclast size is defined to be 3 nuclei per osteoclast, this parameter may not accurately reflect differences in fusion rate, as compared with the fusion index (figure 5.3 A) or number of fusion events (figure 5.3 C). Fusion events (total number of nuclei in giant cells per well – the number giant cells) may be a more convenient parameter than the fusion index (number of nuclei in giant cells as a percentage of total nuclei) especially under conditions where the total number of nuclei in wells (figure 5.3 D) is very large (e.g. MCSF only condition). MCSF treatment by itself resulted in a great increase in cell numbers, reflecting increased cell proliferation and survival (Asigiri and Takayanagi, 2007).

Next the effect of recombinant proteins on RANKL-induced osteoclast formation was determined. Since both CD9 and Tspan13 are reported to be involved in osteoclast formation (Takeda et al., 2003; Ishii et al., 2006; Yi et al., 2006; Iwai et al., 2007), the effects of recombinant GST-CD9 R and GST-Tspan13 R was examined and GST R was included as a negative control (‘R’ denotes that the protein was produced in E. coli Rosetta-gami strain). In preliminary experiments, 500nM protein was found to almost totally inhibit osteoclast formation, therefore dose-response experiments were carried out as shown in figure 5.4. All three recombinant proteins had surprisingly potent effects on osteoclast formation, even at concentrations as low as 500fM (femtomolar).

130

Chapter 5

A) B)

100 10

)

t e

% 80 8

(

)

u e

60 s 6

i l

o n 40 u 4 **

( **

u o

F 20 2

**** v A 0 **** **** 0 **** y L F F y L F F l K S S l K S S n C n C o N C o N C A M M A M M ia R / ia R / d L d L e K e K N N M A M A R R

C) D)

800 i 4000

l l

u ***

n t

600 3000

b 2000

400 d

u s

n *

F 1000

200 o

a c

t n s o

** T ** 0 *** 0

y L F F y L F F l K S S l K S S n C C n N C C o N o A M M a A M M ia R / i R / L d L d K e K e N N M M A A R R

Figure 5.3 Induction of osteoclastogenesis in adherent human PBMC using the cytokines RANKL and MCSF.

The osteoclastogenesis assay was carried out as described (Chapter 2, section 2.3.2). Briefly, adherent PBMC were stimulated with MEM-α media containing either 25ng/ml RANKL or 30ng/ml MCSF or both. Medium plus cytokines was renewed every 2-3 days. Following 21 days culture, cells were fixed and stained for simultaneous bright-field and immunoflouresence microscopy as described (Chapter 2, section 2.3.2). Osteoclasts were defined as TRAP (Tartrate- resistant acid phosphatase)-stained cells with three or more nuclei. Fusion indices and average MGC size were determined as described previously for Con A-induced MGC formation. Fusion events are defined as the number of giant cells subtracted from total number of nuclei in fused cells. The experiment was carried out in one donor in duplicate (RANKL/MCSF in quadruplicate). Non-normalised data are plotted as means ± SEM. Statistical significance was determined using ordinary ANOVA with a Holm-Sidak post-test. The line drawn across three bars in each graph indicates that those bars were compared to the fourth bar which was taken to be the control. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, ns, not significant. N=1, n=2.

131

Chapter 5

G ST R 120 G ST -C D 9 R 100

) G S T -T sp a n 1 3 R l

e r

d 80

o **** n

c 60

o f

u 40

F ( 20 0 F S M M M M M M M M M M M M M M M M M M M M M C n n n p p p f n n n p p p f n n n p p p f 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 M 5 5 0 5 5 0 5 5 0 / 0 5 0 5 5 0 5 0 5 5 0 5 0 5 5 L 5 5 5 5 5 5 K N A R

800

s 600 t

n *

v e 400 n

o **

s u F 200 *** *** *** *** *** *** *** *** **** **** **** **** **** **** **** **** **** **** **** 0 F S M M M M M M M M M M M M M M M M M M M M M C n n n p p p f n n n p p p f n n n p p p f 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 M 5 5 0 5 5 0 5 5 0 / 0 5 0 5 5 0 5 0 5 5 0 5 0 5 5 L 5 5 5 5 5 5 K N A R

Figure 5.4 Recombinant protein activity in the in vitro osteoclastogenesis assay. The assay was carried out as described previously. Recombinant proteins were serially diluted in growth media (containing RANKL/MCSF) to give concentrations in the range 500nM to 500fM (femtomolar). Following monocyte adherence to 96 well plates, medium was removed and diluted recombinant proteins were added to the plates. Medium plus RANKL/MCSF and recombinant proteins was renewed every 2-3 days. Following 21 days culture, cells were fixed, stained, imaged and analysed as previously. Data are normalised to the RANKL/MCSF control and plotted as means ± SEM. Statistical significance was determined using an ANOVA test with a Holm-Sidak post-test. *P<0.05, ** P<0.01, ***P<0.001, ****P<0.0001, N=1, n=2.

We next considered whether LPS contamination had an influence on this activity. As mentioned in the previous chapter, 10ng/ml LPS totally inhibits osteoclast formation of human and murine monocytes whilst at 10pg/ml it has a modest inhibitory effect (Takami et al., 2002; Liu et al., 2009). Of interest is the observation by several groups that the effects of LPS and other bacterial contaminants on osteoclast formation are highly time-dependent, i.e. they inhibit the early stages of monocyte to pre-osteoclast differentiation but enhance fusion of RANKL-stimulated pre-osteoclasts into

132

Chapter 5 multinucleated mature osteoclasts (Takami et al., 2002; Zou and Bar-Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009; Suda et al., 2002; Itoh et al., 2003; Xing et al., 2011; Ha et al., 2011). Zou and Bar-Shavit (2002) further observed that 20ng/ml LPS added to cells at the start of the culture period inhibited osteoclast formation in murine bone-marrow-derived macrophages whilst if it was added after 48 hours it had no effect and after 72 hours it actually enhanced osteoclast formation. Therefore, this may provide a window of opportunity where differentiating osteoclasts are insensitive to bacterial contaminants. A dose-response experiment was carried out where doses of pure LPS were added to monocytes at the start of the culture period or following 24, 48 or 72 hours of monocyte pre-incubation with RANKL/MCSF (figure 5.5).

For convenience, only fusion events were quantitated. Figure 5.5 A shows that no fusion was observed when 100pg/ml-10µg/ml LPS was added at the start of the culture period. Even at 10pg/ml LPS, the fusion rate was extremely low. A 500fM concentration of GST-CD9 R would theoretically contain 12.4fg (femtograms)/ml of LPS. Given the results in figure 5.5 A, it does not seem improbable that LPS contamination is, partially or entirely, responsible for the extreme potency of recombinant proteins in osteoclastogenesis inhibition. When cells were pre-incubated for 24 or 48 hours with RANKL/MCSF before LPS addition, fusion rates were higher but still significantly inhibited at all concentrations tested (figure 5.5 B and C). With a 72 hour pre-incubation period, 100ng/ml LPS significantly enhanced osteoclast formation, similar to the finding of Zou and Bar-Shavit (2002) where 20ng/ml LPS was tested. LPS did not have a significant effect at lower concentrations tested but this may reach significance with more repeats. The pattern of inhibition by LPS at various concentrations appears to be a complex one (figure 5.5 B, C and D). This suggests that LPS may have multiple targets on the cell surface and/or interfere with multiple pathways or factors such as cell adhesion, differentiation and proliferation. Since the osteoclastogenesis assay is carried out over a 14-21 day period and involves monocyte proliferation, initial small effects on factors such as cell viability or proliferation are likely to be amplified resulting in apparently extreme potency being observed. Going back to the culture method employed by R. Hulme (PhD thesis), it was noted that in some experiments monocyte populations were first expanded by pre-incubating cells with MCSF for 5-7 days. MCSF upregulates expression of the RANKL receptor RANK on the cell surface and promotes pre-osteoclast survival (Asagiri and Takayanagi, 2007).

133

Chapter 5

MCSF pre-incubation was not tested here but may have been responsible for the enhancing effects observed previously with recombinant tetraspanin proteins (R. Hulme, PhD thesis).

A) 0 hour B) 24 hour

200 150

)

l t

150 t

n r

r * * e

e 100

v n

* c

100 c

s 50

(

( F 50 F

0 0 F l l l l l l l F l l l l l l l S m m m m m m m S m m m m m m m C / / / / / / / C / / / / / / / g g g g g g g g M g g g M g g g / p p n n n   / p p n n n   L 0 0 1 0 0 1 0 L 0 0 1 0 0 1 0 1 0 1 0 1 1 0 1 0 1 K 1 1 K 1 1 N N A A R R

C) 48 hour D) 72 hour

150 250

* *

)

) s

s 200

n r

r * e

e 100

v n

n 150

i o

o 100

s 50

(

( F F 50 * 0 0 F l l l l l l l F l l l l l l l S m m m m m m m S m m m m m m m C / / / / / / / C / / / / / / / g g g g g g g g M g g g M g g g / p p n n n   / p p n n n   L 0 0 1 0 0 1 0 L 0 0 1 0 0 1 0 1 0 1 0 1 1 0 1 0 1 K 1 1 K 1 1 N N A A R R

Figure 5.5 LPS activity in the in vitro osteoclastogenesis assay.

The assay was carried out as described previously. Pure LPS from E. coli strain 0111:B4 (Sigma, L4391) was serially diluted in growth media (containing RANKL/MCSF) to give concentrations in the range 10µg/ml to 10pg/ml. For the ‘0 hour’ condition, adherent monocytes were directly incubated with LPS dilutions from the beginning of the culture period whilst for the 24, 48 and 72 hour conditions, adherent monocytes were pre-incubated with RANKL/MCSF for 24, 48 or 72 hours before removal of media and addition of new media containing RANKL/MCSF plus LPS dilutions. The media plus reagents on all cells was changed every 2-3 days. After 21 days culture, cells were fixed, stained, imaged and analysed as previously. Data are normalised to the RANKL/MCSF control for each individual time point (0, 24, 48 or 72 hour) since the number of fusion events for the RANKL/MCSF control differed slightly between the four time points. Data are means ± SEM. Statistical significance was determined using ordinary ANOVA followed by the Holm-Sidak post-test *P<0.05, **P<0.01, **P<0.001, ****P<0.0001. N=1, n=2.

134

Chapter 5

5.2.3 Effects of recombinant proteins on the fusion of a murine macrophage-like cell line induced by the bacterium Burkolderia thailandensis

As discussed in Chapter 1 (section 1.2.3.3), giant cell formation in the context of a granuloma is a feature of some bacterial infections such as tuberculosis and melioidosis. Melioidosis is a spectrum of infectious diseases endemic in Southeast Asia and parts of Australia, with a high fatality rate in severe cases (Utaisincharoen et al., 2006; Boddey et al., 2007). The causative agent of melioidosis, the gram-negative facultative bacterium Burkholderia pseudomallei, can also induce fusion of human and murine monocyte and macrophage-like cell lines in in vitro culture as well as in vivo (Kespichayawattana et al., 2000). The resulting multinucleated giant cells harbour numerous intracellular bacteria, reminiscent of Mycobacterium tuberculosis infections (Santanirand et al., 1999). It is thought that B. psudomallei cannot be effectively destroyed by macrophage phagocytosis and instead uses these cells to survive, multiply and spread (Suparak et al., 2005). Interestingly, chronic bone infections such as osteomyelitis and melioidotic arthritis are frequently encountered in melioidosis. In addition, B. pseudomallei–infected RAW264.7 cells (murine macrophage-like cell line) form multinucleated giant cells expressing certain osteoclast markers such as calcitonin receptor, cathepsin C and tartrate-resistant acid phosphatase (TRAP) as well as factors required for osteoclastogenesis such as the transcription factor NFATc1. However, lack of dentine resorption suggests that these cells are not functioning osteoclasts (Boddy et al., 2007).

Several membrane proteins known to be involved in the formation of other types of monocyte-derived giant cells are also implicated in B. pseudomallei-induced MGC formation in vitro (Suparak et al., 2011). The involvement of tetraspanins in this process is currently unknown. J774.2 murine cells were found to express high levels of murine CD9 and CD81 (Mrs Atiga Elgawidi, personal communication). Burkholderia- induced MGC formation offers an interesting and novel system with significant differences to both Con A-induced MGC formation and RANKL-induced osteoclastogenesis (Chapter 1, Table 1.1) to investigate the role of tetraspanins in monocyte fusion, with potentially relevant implications for medical treatment of melioidosis. Efforts to induce MGC formation in human macrophage cell lines such as THP-1 cells proved unsuccessful (Mrs Atiga Elgawidi, personal communication) therefore J774.2 murine macrophages were used. Human CD9EC2 domain shares

135

Chapter 5

~78% identity and 90% similarity to the murine protein and human CD9 can functionally substitute for the murine protein in mouse CD9-/- mouse oocytes which are defective in sperm-egg fusion (Zhu et al., 2002; Kaji et al., 2002). In addition, GST- murineCD9EC2 and GST-humanCD9EC2 both inhibit murine sperm-egg fusion in vitro (Higginbottom et al., 2003). However, a conformation-sensitive anti-human CD9 antibody (602.29) does not recognise the murine CD9 on the cell surface (personal communication, Dr Lynda Partridge, Department of Molecular Biology and Biotechnology, University of Sheffield). Therefore it was unknown whether GST- humanCD9EC2 would have functional activity in this assay. Nevertheless, this assay provided an opportunity to check the activity of the negative controls, i.e. GST and LPS in an alternative monocyte fusion system.

The assay was carried out as described (Chapter 2, section 2.3.2.3) using the non- pathogenic Burkholderia thailandensis strain (Boddey et al., 1997). The results are shown in figure 5.6 A. To allow comparison, data for Con A-induced MGC formation is presented in figure 5.6 B. Data presented in figure 5.6 A and C were obtained by Mrs Atiga Elgawidi (Department of Molecular Biology and Biotechnology, University of Sheffield) and analysed by the author. Figure 5.6 A shows that GST-CD9 R and RX, and to a lesser extent GST R, inhibited B. thailandensis-induced MGC formation as judged by the fusion index whilst all proteins had a significant effect on average MGC size. Similar to Con A-induced MGC formation (figure 5.6 B), GST R had a higher inhibitory activity in this assay than GST RX which may have been due to its higher LPS content. However, GST-CD9 R and RX had similar levels of activity despite different LPS contamination levels and also the LPS content of GST-CD9 RX is actually lower than GST R. The magnitude of the difference between inhibition by GST R and GST-CD9 proteins is greater in B. thailandensis-induced MGC formation. Thus GST-CD9 may have some intrinsic activity in this assay, independent of bacterial contamination and/or GST. Treatment with boiled protein (and loss of effect thereof) would confirm this. Pure LPS at concentrations of 10, 50 and 100ng/ml did not significantly inhibit B. thailandensis-induced MGC formation (figure 5.6 C). But at 50ng/ml which is higher than the LPS content of GST-CD9 R (12.4ng/ml), it seems to reduce MGC size. The experiment was only carried out once so repeats need to be done to confirm lack of effect. It would also be interesting to test the effects of mouse CD9EC2 in this assay.

136

Chapter 5

A) B. thailandensis-induced MGC formation

120 120 * n s

100 ** 100

)

l l

x **

s r

* r

d 80 80 **

n c

60 c 60 ****

****

. ****

o o

s ****

g ****

u 40 40

( ( A 20 20 0 0

X R X X R X e R R R e R R R n T 9 n T 9 o T D 9 o T D 9 S S S S n G -C D n G -C D G T -C G T -C S T S T G S G S G G

B) Con A-induced MGC formation

120 120 * *** n s

100 100 *

) e

) * l

l *

r r

80 d 80 t

t ****

n n

i ****

G o

o ***

c n

c 60 60

o f

****

g 40

u 40

v %

(

( A 20 20 0 0 A R X R X A R X R X n T R 9 R n T R 9 R o S T D 9 o S T D 9 C G S C D C G S C D - G - G T -C T -C S T S T G S G S G G

C) B. thailandensis-induced MGC formation

120 120

100 100

)

r r

80 t

d 80

c 60

n c

s g

40 v

u 40

(

A ( 20 20 0 0 l l l l l l e m m m e m m m n / / / n / / / o g g g o g g g n n n N n n n N 0 0 0 0 0 0 1 5 0 1 5 0 1 S S 1 S S P P S P P S L L P L L P L L

Figure 5.6 Recombinant protein and LPS activity in Burkholderia thailandensis-induced fusion of J774.2 cells.

Burkholderia thailandensis-induced MGC formation was carried out as described (Chapter 2, section 2.3.2.3). Briefly, J774.2 cells were pre-incubated with 500nM recombinant protein or 10- 100ng/ml LPS for 1-16 hours before washing and overnight incubation with bacteria to induce MGC formation. Extracellular bacteria were killed and cells were fixed, stained and counted as described. Data are means ± SEM, normalised to the media only (none) control. Data from 1 and 16 hour protein incubations were not significantly different and therefore were combined. Data in figure B are those shown previously (Chapter 4, figure 4.5). Statistical significance was determined using ordinary ANOVA followed by Holm-Sidak post-test *P<0.05, **P<0.01, **P<0.001, ****P<0.0001, ns=not significant. A) N=5, n=10, B) N=3+, n=7+, C) N=1+, n=2+.

137

Chapter 5

5.3 Discussion The aim of this chapter was to examine the activity of recombinant tetraspanin proteins in other cell-based functional assays including the bacterial adhesion assay, RANKL-induced osteoclast formation and Burkholderia-induced MGC formation. Several recombinant tetraspanins EC2 domain proteins including GST-CD9EC2 and GST-CD63EC2 were found to reduce bacterial adhesion to epithelial cells by approximately 50% (Green et al., 2011). In this study, it was found that an additional four recombinant tetraspanin EC2 domains similarly inhibited bacterial adhesion to epithelial cells (figure 5.2). LPS, GST and boiled GST-CD9EC2 had no effect on this assay therefore it is highly likely that this inhibition is specific to the tetraspanin EC2 domains (figure 5.1). Green et al. (2011) propose that recombinant tetraspanin EC2 domains probably do not have specific targets on the cell surface but have a general disruptive effect on the organisation of tetraspanin-enriched microdomains (TEMs) where adhesion receptors and other molecules involved in bacterial adhesion are clustered. Short synthetic peptide analogues of the human CD9EC2 region also potently inhibit bacterial adhesion to epithelial cell lines and are currently under investigation as agents for the topical treatment of skin burns to prevent bacterial infections (Mr Daniel Cozens and Miss Jennifer Ventress, Department of Infection and Immunity, University of Sheffield Medical School, internal communication). Based on the findings in this study, studies with EC2 domain peptide analogues may be extended to other tetraspanin EC2 domains to find more potent and/or stable peptides.

Next, the effects of tetraspanin proteins in two monocyte-derived giant cell formation assays was determined (figures 5.5 and 5.6). GST, GST-CD9EC2, GST-Tspan13EC2 and LPS all had extremely potent effects on RANKL-induced osteoclast formation such that fusion rates were dramatically reduced even at extremely low protein or LPS concentrations. Osteoclast formation involves monocyte proliferation, differentiation as well as fusion therefore interference with any of these processes may reduce fusion rates. Further, the effects of LPS on osteoclast formation were observed to be time- dependent, as reported previously (Takami et al., 2002; Zou and Bar-Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009; Suda et al., 2002; Itoh et al., 2003; Xing et al., 2011; Ha et al., 2011). It was also reported that monocyte pre-incubation for 48 hours with RANKL/MCSF rendered cells insensitive to subsequent stimulation

138

Chapter 5 with 20ng/ml LPS, at least as far as fusion rates were concerned (Zou and Bar-Shavit, 2002). In this study, stimulation with 10ng/ml LPS had no effect on the fusion of monocytes pre-incubated for 72 hours with RANKL/MCSF. This would have provided a window of opportunity to test the activity of recombinant tetraspanins in this assay. However, this effect was found to be dose- as well as time-dependent i.e. different fusion rates were observed with different concentrations of LPS making it unsuitable for testing of recombinant proteins containing different levels of LPS contamination. Comparison of recombinant protein effects on Burkholderia- and Con A-induced MGC formation (figure 5.6 A and B) showed similar results. However, the magnitude of the difference between inhibition levels by GST and GST-CD9EC2 proteins suggests that GST-CD9EC2 may have some intrinsic activity in this assay. LPS did not significantly inhibit Burkholderia-induced MGC formation but this must be confirmed with further repeats.

In conclusion, it appears that recombinant protein activity in monocyte-based assays in general is adversely affected by the presence of bacterial contaminants in the purified protein preparation, thus care must be taken to include adequate controls and thoroughly exclude the contribution of contaminants before attributing functional activity to recombinant proteins.

139

Chapter 6

Chapter 6: General Discussion

As a large protein superfamily, tetraspanins remain poorly investigated, despite evidence of involvement in many different biological processes (Charrin et al., 2009). This is attributed partly to a lack of monoclonal anti-tetraspanins antibodies which would allow investigation of tetraspanin functions on intact cells in vitro (Hemler, 2005). Another factor is tetraspanin functional redundancy and compensation such that even multiple tetraspanin knockout organisms often display mild phenotypic abnormalities (Fradkin et al., 2002). As alternative tools to anti-tetraspanin antibodies, several studies have utilised recombinant tetraspanin EC2 domains to investigate tetraspanin functions in vitro (summarised in Table 1.2, Chapter 1). The tetraspanin large extracellular domain (EC2) is relatively small compared to the extracellular domains of other membrane proteins such as immunoglobulins and integrins, ranging in size between 69 to 132 amino acids (Garcia-Espana et al., 2008). However, it is of great functional importance to tetraspanin functions, particularly specific tetraspanin-partner protein interactions (Hasuwa et al., 2001; Ellerman et al., 2003; Higginbottom et al., 2000; Kazarov et al., 2002; Ding et al., 2005).

In this study, large extracellular domains (EC2 domains) from nine human tetraspanins were expressed for the first time as recombinant GST-tagged N-terminal fusion proteins in E. coli Rosetta-gami and SHuffle strains. These strains have specific adaptations which are reported to enhance expression of disulfide bond-containing proteins in bacteria (Bar et al., 2006; Bessette et al., 1999; Lobstein et al., 2012). crystal structure of human CD81EC2 showed that the four cysteines in this protein take part in intramolecular disulfide bridge formation (Kitadokoro et al., 2001). In addition, mutational studies suggest that cysteine residues in the EC2 domain are crucial to proper folding and functionality of tetraspanin EC2 domains (Goldberg et al., 1998; Higginbottom et al., 2003). The tetraspanin EC2 domains expressed here contain between four and eight cysteine residues, i.e. between two to four putative disulfide bridges.

Expression of five proteins, GST-Tspan2EC2, GST-Tspan3EC2, GST-Tspan13EC2, GST- Tspan32EC2 and GST-CD82EC2, in E. coli Rosetta-gami strain was successful, in that sufficient amounts were obtained for functional studies. However, the remaining four proteins, GST-Tspan4EC2, GST-Tspan5EC2, GST-Tspan8EC2, GST-Tspan31EC2, proved

140

Chapter 6 more difficult to express and purified protein purity was relatively low. SDS-PAGE analysis followed by Western blotting showed that most of the expressed protein was present in the insoluble fraction of bacterial cell lysates. Protein mis-folding and aggregation is a common occurrence in the production of recombinant proteins containing 3 or more disulfide bonds in bacteria (Qiu et al., 1998). This is due to the reducing environment of the bacterial cytoplasm which does not support stable disulfide bridge formation (Bessette et al., 1999; Baneyx and Mujicic, 2004). Interestingly, GST-Tspan4EC2, GST-Tspan5EC2, GST-Tspan8EC2, GST-Tspan31EC2 all contain six to eight cysteines whilst the other proteins (apart from GST-CD82EC2) which expressed well contain four cysteines. However, some aggregate formation was also observed in the GST-CD9EC2 and GST-Tspan2EC2 preparations. Analysis of recombinant tetraspanins by Western blotting with anti-GST antibody revealed the presence of h gh m c a ma a (≥ 50 KDa) in a number of purified proteins. Blotting of GST-CD82EC2 with a conformation-sensitive antibody showed that at least some of this material is correctly folded (figure 3.12, antibody J). It is also interesting to note that for extracellular proteins, protein regions involved in protein-protein interactions are often enriched in disulfide bonds and seem to be particularly prone to aggregation (Fas et al., 2012). Therefore, aggregate formation may be an intrinsic property of tetraspanin EC2 domains. It remains to be determined whether such aggregates are biologically active. Dimer formation was also observed in a number of purified proteins and has been previously noted for wild-type tetraspanins on the cell surface as well as recombinant tetraspanin EC2 domains (Kovalenko et al., 2004; Rajesh et al., 2012; Ding et al., 2005).

Analysis of recombinant GST-Tspan4EC2 and GST-Tspan5EC2 by SDS-PAGE also revealed that these proteins may be truncated. This could be confirmed by peptide mass spectrometry (Zhou et al., 2001). Further, prominent higher molecular weight bands were present in the purified GST-Tspan4EC2, GST-Tspan5EC2 and GST- Tspan8EC2, which did not stain with anti-GST or anti-tetraspanin antibodies. Thus these bands were thought to be bacterial contaminants and these proteins were not used in functional studies. Conformation-sensitive anti-tetraspanin antibodies are currently not available for most tetraspanins. Therefore, to assess protein folding, other methods must be used. Recombinant tetraspanins were not amenable to characterisation by circular dichroism, due to interference from substantial amounts

141

Chapter 6 of lower molecular weight material which were assumed to be degradation products of the full-length GST-tagged protein. Use of a second affinity tag such as polyhistidine (as outlined in Chapter 3, section 3.3) would improve protein purity (although perhaps not yields) and allow the cleaved GST-free his-tagged protein to be characterised by biophysical methods such as circular dichroism and Nuclear magnetic resonance (NMR) spectroscopy (Rajesh et al., 2012). However, the protein concentrations required for NMR are significantly higher than those typically obtained for recombinant tetraspanin EC2 domains (this study and R. Hulme, PhD thesis).

The role of tetraspanin proteins in multinucleated giant cell formation has been a focus of our research group. As discussed in detail in Chapter 1, tetraspanin proteins are involved in a number of different fusion processes (summarised in Fanaei et al., 2011) such as sperm-egg fusion (Le Naour et al., 2000; Miyado et al., 2000; Kaji et al., 2000), monocyte-monocyte fusion (Takeda et al., 2003; Parthasarathy et al., 2009; Ishii et al., 2006; Yi et al., 2006), muscle cell fusion (Tachibana and Hemler, 1999), virus-host cell fusion (Fukudome et al., 1992; Imai et al., 1992; Gordon-Alonso et al., 2006; Weng et al., 2009; Nydegger et al., 2006; Singethan et al., 2008) and outer- segment membrane formation in the retina (Boesze-Battaglia et al., 2007). Apart from the tetraspanin RDS which functions in outer-segment membrane fusion, other human a pa d app a f c a ‘f g ’, m c wh ch ac v y induce membrane fusion by enabling close membrane apposition (Harrison, 2008; figure 1.3, Chapter 1). Instead, they appear to regulate fusion processes by controlling access to certain functional areas of the membrane where various molecules involved in fusion are assembled.

The involvement of tetraspanins in multinucleated giant cell (MGC) formation in monocytes is discussed in detail in Chapter 1 (section 1.2.3.3.4). Previous studies by our group had shown that recombinant GST-tagged EC2 domains of human tetraspanins CD9 and CD63 inhibited Concanavalin A-induced MGC formation (Parthasarathy et al., 2009). Here, we wished to determine if other less well- characterised tetraspanins were also involved in monocyte fusion by testing the activity of recombinant GST-tagged tetraspanin EC2 domains (produced in Chapter 3) in Con A-induced MGC formation in adherent human peripheral blood monocytes (PBMC) as well as in RANKL-induced osteoclast formation in adherent PBMC and B.

142

Chapter 6 thailandensis-induced MGC formation in the murine macrophage cell line J774.2. Unfortunately, it was found that protein contamination with bacterial products such as lipopolysaccharide (LPS) was partly or entirely responsible for the inhibitory effects of recombinant proteins in these assays (Chapter 4 and Chapter 5). Various techniques were used to minimise or exclude the contribution of bacterial contaminants such as LPS reduction by washing the protein with detergent during purification or addition of the LPS inhibitor polymyxin B to the culture medium. However, we could not confirm that recombinant tetraspanins had intrinsic activity in monocyte fusion assays.

These results were contrary to previous data by our group where pure LPS, at the concentrations found in recombinant proteins, had been inactive in Con A-induced MGC formation (Parthasarathy et al., 2009; R. Hulme, PhD thesis). However, the results in this study are perhaps not surprising given the wealth of available data on the sensitivity of monocytes to bacterial products at low concentrations (Sabroe et al., 2002; Fenton and Golenbock, 1998; Gallay et al., 1993; Gao and Tsan, 2003) and profound effects on monocyte differentiation into pro-inflammatory mature macrophages and RANKL-induced multinucleated osteoclasts (Watson et al., 1994; Takami et al., 2002; Tsuzuki et al., 2001; Zou and Bar Shavit, 2002; Ha et al., 2008; Hotokezata et al., 2007; Liu et al., 2009; Suda et al., 2002; Itoh et al., 2003; Xing et al., 2011; Hotozeka et al., 2007; Ha et al., 2011).

Less is known about the effects of LPS and other contaminants on Con A-induced MGC formation. Takashima et al. (1993) reported that LPS dose-dependently inhibited Con A-induced MGC formation, a result that was replicated in this study. In contrast, previous data by our group suggested that high concentrations of LPS actually enhance Con A-induced MGC formation (Parthasarathy et al., 2009). It is interesting to note that striking similarities are observed between Con A- and LPS-induced cell activation. B h ag ‘p m ’ ac va m cy a p -inflammatory mediators such as TNF-α, IL- β and IL-6 (Takeda et al., 2003; Gordon and Martinez, 2010). Con A also upregulates the expression of a number of Toll-like receptors (TLRs) including the LPS receptor TLR4 as well as downstream signalling mediators in murine macrophages (Sodhi et al., 2007). Therefore, Con A stimulated monocytes are expected to be even more sensitive to the effects of LPS than unstimulated cells. TNF-α ha b found to be necessary but not sufficient for monocyte fusion (Takashima et al., 1993).

143

Chapter 6

Further, LPS, TNF-α a d h p -inflammatory stimulants by themselves cannot induce MGC formation suggesting the involvement of other factors (Takashima et al., 1993; Abe et al., 1984; Enelow et al., 1992; Lazarus et al., 1990). Con A and LPS also stimulate T lymphocyte proliferation (Sodhi et al., 2007; Vogel et al., 1983). The T-cell derived cytokine IFN-y is implicated in Con A-induced MGC formation as well as LGC (La gha ’ g a c ) f ma in vivo (Sakai et al., 2011; Belosevic et al., 1989; Chensue et al., 1992; Most et al., 1990; Okamoto et al., 2003). Con A, IFN-γ a d LPS a induce homotypic adhesion in monocytes (monocyte-monocyte adhesion) which is thought to be necessary for fusion (Mentzer et al., 1986; Hayes et al., 1991; Lauener et al., 1990). Therefore one would expect LPS to enhance the effects of Con A in MGC formation but the opposite is in fact observed (this study and Takashima et al., 1993).

It is possible that excessive activation (via stimulation of TNF-α and other pro- inflammatory mediator production) has a cytotoxic effect on monocytes. It was observed that after 6 hours treatment with 100ng/ml LPS, 35% of murine macrophages in culture were apoptotic, an activity that correlated with TNF-α p d c (Xaus et al., 2000). It would be interesting to determine TNF-α c c a p a a f monocytes stimulated with Con A, LPS or both. LPS/recombinant protein cytotoxicity could also be determined using colorimetric assays. An example is the MTT assay which measures the conversion of the yellow MTT dye (3-(4,5-Dimethyl-2-thiazolyl)- 2,5-diphenyl-2H tetrazolium bromide) to a blue formazan product, a reaction carried out exclusively by living cells (Denizot and Lang, 1986). Alternatively, the LDH assay measures the presence of the cytosolic lactate dehydrogenase (LDH) enzyme in the growth media, an indicator of dead or dying cells (Decker and Lohmann-Matthes, 1988). LPS also stimulates T cell proliferation and high concentrations of pro- inflammatory mediators may have deleterious effects on T cells which thereby indirectly affecting monocyte fusion (see above). T cell proliferation could be quantified by FACS (fluorescence activated cell sorting)-based assays by following the dilution of a fluorescent dye such as CFSE (carboxyfluorescein diacetate) as it is distributed from the parent cell to daughter cells through multiple rounds of cell division (Lyons et al., 2000).

A general conclusion from the above discussion is that caution must be exercised when bacterially-derived products such as recombinant proteins are used in cell-

144

Chapter 6 based assays involving cells with known sensitivity to microbial contaminants. To circumvent the interference of bacterial contaminants in monocyte fusion assays, monocytes from Toll-like receptor (TLR)-knockout organisms could be used such as the TLR4-knockout mouse, C3H/HeJ (Takami et al., 2002). In this study, the activity of six new recombinant tetraspanins was determined in the bacterial adhesion assay, a cell-based system using the epithelial cell line HEC-1-B. These cells are of non-immune origin and were found to be insensitive to LPS. In addition, GST and boiled GST- CD9EC2 (used as negative controls) were found to be devoid of activity whilst native GST-CD9EC2 and four other recombinant tetraspanins, GST-Tspan3EC2, GST- Tspan13EC2, GST-CD82EC2 and GST-Tspan32EC2 significantly inhibited bacterial adhesion to HEC-1-B cells. One puzzling point is that anti-CD9 antibody and GST- CD9EC2 protein were found to have opposing effects on MGC formation (Takeda et al., 2003; Parthasarathy et al., 2009). This has been explained previously by suggesting that their modes of action are different, the antibody specifically binds to CD9 on the cell surface and prevents its interactions with other proteins whereas the recombinant protein has a generally disorganising effect on tetraspanin-enriched microdomains (Green et al., 2011; Parthasarathy et al., 2009). However, in the bacterial adhesion assay, CD9 targeting by RNA interference, anti-tetraspanin antibody or recombinant tetraspanin EC2 domains all produced the same effects, i.e. inhibition of adhesion (Green et al., 2011). Therefore one must conclude that the mode of action of GST- CD9EC2 in the two assays is yet again different, if GST-CD9EC2 has any activity at all in the MGC assay.

Regarding investigations into the role of tetraspanins in multinucleated giant cell formation, a more targeted approach than production of recombinant tetraspanins known to be expressed on monocytes may prove more fruitful. Transcriptomic and proteomic methods could be used to elucidate the profile of tetraspanins expressed during various stages of monocyte development into multinucleated giant cells. A comparison of tetraspanins profiles involved in Con A-induced MGC formation, LGC formation, osteoclast formation and Burkholderia-induced MGC formation may also yield interesting results. Subtractive screens (Yagi et al., 2005) could highlight key tetraspanins (with differential expression in monocytes and multinucleated giant cells) for further investigations by alternative methods such as RNA interference (Green et al., 2011).

145

Bibliography

Abe, E., Shiina, Y., Miyaura, C., Tanaka, H., Hayashi, A., Kanegasaki, S., Saito, M., Nisho, Y., Deluca, H. F. and Suda, T. (1984) Activation and fusion induced by lα,25- dihydroxyvitamin D3 and their relation in alveolar macrophages. Proceedings of the National Academy of Science USA 81, 7112-7116. Abmayr, S. M. and Pavlath, G. K. (2012) Myoblast fusion: lessons from flies and mice. Development 139, 641-656. Aida, Y. and Pabst, M. J. (1990) Removal of endotoxin from protein solutions by phase separation using Triton X-114. Journal of Immunological Methods 132, 191-195. Ainola, M., Li, T., Mandelin, J., Hukkanen, M., Choi, S. J., Salo, J., & Konttinen, Y. T. (2009) Involvement of a disintegrin and a metalloproteinase 8 (ADAM8) in osteoclastogenesis and pathological bone destruction. Annals of the Rheumatic Diseases 68, 427–434. Akira, S. (2003) Toll-like receptor signalling. The Journal of Biological Chemistry 278, 38105-38108. Alexander, C. and Rietschel, E. T. (2001) Bacterial lipopolysaccharides and innate immunity. Journal of Endotoxin Research 7, 167-202. Anderson, J. M. (2000) Multinucleated giant cells. Current Opinion in Hematology 7, 40-47. Anderson, J. M., Rodriguez, A. and Chang, D. T. (2008) Foreign body reaction to biomaterials. Seminars in Immunology 20, 86–100. Anspach, F. B. and Hilbeck, O. (1995) Removal of endotoxins by affinity sorbents. Journal of Chromatography 711, 81-92. Aoki, H., Kodama, M., Tani, T. and Hanasawa, K. (1994) Treatment of sepsis by extracorporeal elimination of endotoxin using polymyxin. The American Journal of Surgery 167, 412–417. Arduise, C., Abache, T., Li, L., Billard, M., Chabanon, A., Ludwig, A., Mauduit, P., Boucheix, C., Rubinstein, E. and Le Naour, F. (2008) Tetraspanins regulate ADAM10-mediated cleavage of TNF-alpha and epidermal growth factor. Journal of Immunology 181, 7002–7013. Asagiri, M. and Takayanagi, M. (2007) The molecular understanding of osteoclast differentiation. Bone 40, 251-264. Bankwitz, D., Steinmann, E., Bitzegeio, J., Ciesek, S., Friesland, M., Herrmann, E., Zeisel, M. B., Baumert, T. F., Keck, Z., Foung, S. K. L., Pecheur, E. and Pietschmann, T. (2010) Hepatitis C virus hypervariable region 1 modulates receptor interactions, conceals the CD81 binding site, and protects conserved neutralizing epitopes. Journal of Virology 84, 5751–5763. Bar, M., Bar-Ziv, R., Scherf, T. and Fass, D. (2006) Efficient production of a folded and functional, highly disulfide-bonded beta-helix antifreeze protein in bacteria. Protein Expression and Purification 48, 243–252. Bardwell, J. C. A., McGovern, K. and Beckwith, J. (1991) Identification of a protein required for disulfide bond formation in vivo. Cell 67, 581-589. Barreiro, O., Yáñez-Mó, M., Sala-Valdés, M., Dolores Gutiérrez-López, M., Ovalle, S., Higginbottom, A., Monk, P. N., Cabañas, C. and Sánchez-Madrid, F. (2005) Endothelial tetraspanin microdomains regulate leukocyte firm adhesion during extravasation. Blood 105, 2852-2861. Barreiro, O., Zamai, M., Yáñez-Mó, M., Tejera, E., López-Romero, P. and Monk, P. N. (2008) Endothelial adhesion receptors are recruited to adherent leukocytes by inclusion in preformed tetraspanin nanoplatforms. Journal of Biological Chemistry 183, 527- 542. Bar-Shavit, Z. (2007) The osteoclast: a multinucleated, hematopoietic-origin, bone- resorbing osteoimmune cell. Journal of Cellular Biochemistry 102, 1130-1139. 146

Bausinger, H., Lipsker, D., Ziylan, U., Manie, S., Briand, J., Cazenave, J., Muller, S., Haeuw, J., Ravanat, K., de la Salle, H. and Hanau, D. (2002) Endotoxin-free heat-shock protein 70 fails to induce APC activation. European Journal of Immunology 32, 3708-3713. Belosevic, M., Finebloom, D. S., van der Meide, P. H., Slayter, M. V. and Nacy, C. A. (1989) Administration of monoclonal anti-IFN-γ antibodies in vivo abrogates natural resistance of C3H/HeN mice to infection with Leishmania major. The Journal of Immunology 143, 266-274. Berditchevski, F., Gilbert, E., Griffiths, M. R., Fitter, S., Ashman, L. and Jenner, S. J. (2001) Analysis of the CD151·α3β1 integrin and CD151·tetraspanin interactions by mutagenesis. The Journal of Biological Chemistry 276, 41165–41174. Berditchevski, F., Odintsova, E., Sawada, S. and Gilbert, E. (2002) Expression of the palmitoylation-deficient CD151 weakens the association of α3β1 integrin with the tetraspanin-enriched microdomains and affects integrin-dependent signalling. Journal of Biological Chemistry 277, 36991–37000. Berditchevski, F. and Odintsova, E. (2007) Tetraspanins as regulators of protein trafficking. Traffic 8, 89–96 Bessette, P. H., Aslund, F., Beckwith, J. and Georgiou, G. (1999) Efficient folding of proteins with multiple disulfide bonds in the Escherichia coli cytoplasm. Proceedings of the National Academy of Sciences USA 96, 13703–13708. Blanco, P., Palucka, A. K., Pascual, V., Banchereau, J. (2008) Dendritic cells and cytokines in human inflammatory and autoimmune diseases. Cytokine and Growth Factor Reviews 19, 41-52. Boddey, J. A., Day, C. J., Flegg, C. P., Ulrich, R. L., Stephens, S. R., Beacham, I. R., Morrison, N. A. and Peak, I. R. A. (2007) The bacterial gene lfpA influences the potent induction of calcitonin receptor and osteoclast-related genes in Burkholderia pseudomallei- induced TRAP-positive multinucleated giant cells. Cellular Microbiology 9, 514- 531. Boesze-Battaglia, K., Lamba, O. P., Napoli, A. A., Sinha, S. and Guo, Y. (1998) Fusion between retinal rod outer segment membranes and model membranes: a role for photoreceptor peripherin/rds. Biochemistry 37, 9477-9487. Boesze-Battaglia, K., Stefano, F. P., Fenner, M., Napoli, A. A. (2000) A peptide analogue to a fusion domain within photoreceptor peripherin/rds promotes membrane adhesion and depolarization. Biochimica et Biophysica Acta 1463, 343-354 Boesze-Battaglia, K. and Gretzula, C. (2003) A role for Rom-1 in peripherin/rds dependent membrane fusion processes. Investigative Ophthalmology and Visual Sciences 44, 4259. Boesze-Battaglia, K., Song, H., Sokolov, M., Lillo, C., Pankoski-Walker, L., Gretzula, C., Gallagher, B., Rachel, R. A., Jenkins, N. A., Copeland, N. G., Morris, F., Jacob, J., Yeagle, P., Williams, D. S. and Damek-Poprawa, M. (2007) The tetraspanin protein peripherin-2 forms a complex with melanoregulin, a putative membrane fusion regulator. Biochemistry 46, 1256-1272. Boon, C. J. F., den Hollander, A. I., Hoyng, C. B., Cremers, F. P. M., Klevering, B. J. and Keunen, J. E. E. (2008) The spectrum of retinal dystrophies caused by mutations. Progress in Retinal and Eye Research 27, 213–235. Boucheix, C. and Rubinstein, E. (2001) Tetraspanins. Cellular and Molecular Life Sciences 58, 1189–1205. Boyle, W. J., Simonet, W. S. and Lacey, D. L. (2003) Osteoclast differentiation and activation. Nature 432, 337-342. Bucciantini, M., Giannoni, E., Chiti, F., Baroni, F., Formigli, L., Zurdo, J., Taddei, N., Ramponi, G., Dobson, C. M. and Stefani, M. (2002) Inherent cytotoxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature 416, 507-511. Bulut, Y., Michelsen, K. S., Hayrapetian, L., Naiki, Y., Spallek, R., Singh, M. and Arditi, M. (2005) Mycobacterium tuberculosis heat shock proteins use diverse Toll-like

147

receptor pathways to activate pro-inflammatory signals. Journal of Biological Chemistry 280, 20961–20967. Byrd, T. F. (1998) Multinucleated giant cell formation induced by IFN-γ/IL-3 is associated with restriction of virulent Mycobacterium tuberculosis cell to cell invasion in human monocyte monolayers. Cellular Immunology 188, 89-96. Cannon, K. S. and Cresswell, P. (2001) Quality control of transmembrane domain assembly in the tetraspanin CD82. EMBO Journal 20, 2443–2453. Capecchi, B., Adu-Bobie, J., Di Marcello, F., Ciucchi, L., Masignani, V., Taddei, A., Rappuoli, R., Pizza, M. and Aricò, B. (2005) Neisseria meningitides NadA is a new invasin which promotes bacterial adhesion to and penetration into human epithelial cells. Molecular Microbiology 55, 687-698. Cardoso, L. S., Araujo, M. I., Góes, A. M., Pacífico, L. G., Ricardo, R., Oliveira, R. and Oliveira, S. C. (2007) Polymyxin B as inhibitor of LPS contamination of Schistosoma mansoni recombinant proteins in human cytokine analysis. Microbial Cell Factories 6, 1. Chakraborty, D., Rodgers, K. K., Conley, S. M. and Naash, M. I. (2012) Structural characterization of the second intra-discal loop of the photoreceptor tetraspanin RDS. FEBS Journal 280, 127-138. Charrin, S., Manie, S., Oualid, M., Billard, M., Boucheix, C. and Rubinstein, E. (2002) Differential stability of tetraspanin/tetraspanin interactions: role of palmitoylation. FEBS Letters 516, 139-144. Charrin, S., Manie, S., Thiele, C., Billard, M., Gerlier, G., Boucheix, C. and Rubinstein, E. (2003) A physical and functional link between cholesterol and tetraspanins. European Journal of Immunology 33, 2479–489. Charrin, S., Le Naour, F., Silvie, O., Milhiet, P., Boucheix, C. and Rubinstein, E. (2009) Lateral organization of membrane proteins: tetraspanins spin their web. Biochemical Journal 420, 133–154. Chen, M.S., Tung, K. S., Coonrod, S. A., Takahashi, Y., Bigler, D., Chang, A., Yamashita, Y., Kincade, P. W., Herr, J. C., and White, J. M. (1999) Role of the integrin-associated protein CD9 in binding between sperm ADAM 2 and the egg integrin α6β1: implications for murine fertilization. Proceedings of the National Academy of Science USA 96, 11830–11835. Chen, E. H. and Olson, E. N. (2005) Unveiling the mechanisms of cell-cell fusion. Science 308, 369-391. Chen, E. H., Groteb, E., Mohler, W. and Vignery, A. (2007) Cell-cell fusion. FEBS letters 581, 2181–2193. Cheng, J., Liu, J., Shi, Z., Xu, D., Luo, S., Siegal, G. P., Feng, X. and Wei, S. (2011) Interleukin-4 inhibits RANKL-induced NFATc1 expression via STAT6: a novel mechanism mediating its blockade of osteoclastogenesis. Journal of Cellular Biochemistry 112, 3385-3392. Chensue, S. W., Terebuh, P. D., Warmington, K., S., Hershey, S. D., Evanoff, H. L., Kunkel, S. L. and Higashi, G. I. (1992) Role of IL-4 and IFN-γ in Schistosoma mansoni egg- induced hypersensitivity granuloma formation: Orchestration, relative contribution and relationship to macrophage function. The Journal of Immunology 148, 900-906. Clarke, G., Goldberg, A. F. X., Vidgen, D., Collins, L., Ploder, L., Schwarz, L., Molday, L. L., Rossant, J., Szél, A., Molday, R. S., Birch, D. G. and McInnes, R. R. (2000) Rom-1 is required for rod photoreceptor viability and the regulation of disk morphogenesis. Nature Genetics 25, 67-73. Co, D. O., Hogan, L. H., Il-Kim, S. and Sandor, M. (2004) T cell contributions to the different phases of granuloma formation. Immunology Letters 92, 135-142. Codina, J., Li, J. and DuBose T. D. Jr. (2005) CD63 interacts with the carboxy terminus of the colonic H+-K+-ATPase to decrease plasma membrane localization and 86Rb+ uptake. American Journal of Physiology-Cell Physiology 288, 1279–1286.

148

Coury, F., Annels, N., Rivollier, A., Olsson, S., Santoro, A., Speziani, C., Azocar, O., Flacher, M., Djebali, S., Tebib, J., Brytting, M., Egeler, R. M., Rabourdin-Combe, C., Henter, J., Arico, M. and Delprat, C. (2008) Langerhans cell histiocytosis reveals a new IL- 17A–dependent pathway of dendritic cell fusion. Nature Medicine 14, 81-87. Cui, W., Ke, J. Z., Zhang, Q., Ke, H., Chalouni, C. and Vignery, A. (2006) The intracellular domain of CD44 promotes the fusion of macrophages. Blood 107, 796-805. Damais, C., Jupin, C., Parant, M. and Chedid, L. (1987) Induction of human interleukin-1 production by polymyxin B. Journal of Immunological Methods 101, 51-56. Damek-Poprawa, M., Krouse, J., Gretzula, C. and Boesze-Battaglia, K. (2005) A novel tetraspanin fusion protein, Peripherin-2, requires a region upstream of the fusion domain for activity. Biological Chemistry 280, 9217-9224. Decker, T. and Lohmann-Matthes, M. (1988) A quick and simple method for the quantitation of lactate dehydrogenase release in measurements of cellular cytotoxicity and tumor necrosis factor (TNF) activity. Journal of Immunological Methods 15, 61-69. DeFife, K. M., Jenney, C. R., McNally, A. K., Colton, E. and Anderson, J. M. (1997) Interleukin- 13 induces human monocyte/macrophage fusion and macrophage mannose receptor expression. The Journal of Immunology 158, 3385-3390. DeFife, K. M., Jenney, C. R., Colton, E. and Anderson, J. M. (1999) Disruption of filamentous actin inhibits human macrophage fusion. FASEB 13, 823-832. Denizot, F. and Lang, R. (1986) Rapid colorimetric assay for cell growth and survival. Modifications to the tetrazolium dye procedure giving improved sensitivity and reliability. Journal of Immunological Methods 89, 271-277. DeSalle, S., Mares, R. and Garcia-España, A. (2010) Evolution of cysteine patterns in the large extracellular loop of tetraspanins from animals, fungi, plants and single- celled eukaryotes. Molecular Phylogenetics and Evolution 56, 486–491. Ding, X. Q., Stricker, H. M. and Naash, M. I. (2005) Role of the second intradiscal loop of peripherin/rds in homo and hetero associations. Biochemistry 44, 4897-4904. do Nascimento, M. P. P., de Campos Soares, A. M. V., Dias-Melicio, L. A., Parise-Fortes, M. R., Martins, R. A. R., Nakaira, E. T., Peracoli, M. T. S. (2008) Fungicidal activity of human monocyte-derived multinucleated giant cells induced in vitro by Paracoccidioides brasiliensis antigen. Mycopathologia 166, 25-33. Don, R. H., Cox, P. T., Wianright, B. J., Baker, K. and Mattick, J. S. (1991) ‘Touchdown’ PCR to circumvent spurious priming during gene amplification. Nuclei Acids Research 19, 4008. Dornier, E., Coumailleau, F., Ottavi, J. F., Moretti, J., Boucheix, C., Mauduit, P., Schweisguth, F. and Rubinstein, E. (2012) TspanC8 tetraspanins regulate ADAM10/Kuzbanian trafficking and promote Notch activation in flies and mammals. The Journal of cell biology 199, 481–496 Drexler, S. K. and Foxerll, B. M. (2010) The roll of Toll-like receptors in chronic inflammation. International Journal of Biochemistry and Cell Biology 42, 505-518. Dudley, A., McKinstry, W., Thomas, D., Best, J. and Jenkins, A. (2003) Removal of endotoxin by reverse phase HPLC abolishes anti-endothelial cell activity of bacterially expressed plasminogen kringle 5. BioTechniques 35, 724-732. Edrington, T. C., Lapointe, V. R., Yeagle, P. L., Gretzula, C. L. and Boesze-Battaglia, K. (2007a) Peripherin-2: an intracellular analogy to viral fusion proteins. Biochemistry 46, 3605-3613. Edrington, T. C., Yeagle, P. L., Gretzula, C. L. and Boesze-Battaglia, K. (2007b) Calcium- dependent association of calmodulin with the C-terminal domain of the tetraspanin protein peripherin/rds. Biochemistry 46, 3862-3871. Ellerman, D. A., Ha, C., Primakoff, P., Myles, D. G. and Dveksler, G. S. (2003) Direct binding of the ligand PSG17 to CD9 requires a CD9 site essential for sperm-egg fusion. Molecular Biology of the Cell 14, 5098–5103.

149

Enelow, R. I., Sullivan, G. W., Carper, H. T. and Mandell, G. L. (1992) Induction of multinucleated giant cell formation from in vitro culture of human monocytes with interleukin-3 and interferon-γ: comparison with other stimulating factors. American Journal of Respiratory Cell and Molecular Biology 6, 57-62. Erridge, C. (2010) Endogenous ligands of TLR2 and TLR4: agonists or assistants? Journal of Leukocyte Biology 87, 989-999. Espenel, C., Margeat, M., Dosset, P., Arduise, C., Le Grimellec, C., Royer, C. A., Boucheix, C., Rubinstein, E. and Milhiet, P. E. (2008) Single-molecule analysis of CD9 dynamics and partitioning reveals multiple modes of interaction in the tetraspanin web. Journal of Cell Biology 182, 765-776. Fais, S., Burgio, V. L., Capobianchi, M. R., Gessani, S., Pallone, F. and Belardelli, F. (1997) The biological relevance of polykaryons in the immune response. Immunology Today 18, 522-527. Falzoni, S., Munerati, M., Ferrari, D., Spisani, S., Moretti, S. and Di Virgilio, F. (1995) The purinergic P2Z receptor of human macrophage cells: characterisation and possible physiological role. Journal of Clinical Investigation 95, 1207-1216. Fan, J., Zhu, G. and Niles, R. M. (2010) Expression and function of CD9 in melanoma cells. Molecular Carcinogenesis 49, 85–93. Fanaei, M., Monk, P. N. and Partridge, L. J. (2011) The role of tetraspanins in fusion. Biochemical Society Transactions 39, 524-528. Fas, D. (2012) Disulfide bonding in protein biophysics. Annual Reviews in Biophysics 41, 63–79. Fenton, M. J. and Golenbock, D. T. (1998) LPS-binding proteins and receptors. Journal of Leukocyte Biology 64, 25-32. Flint, M., von Hahn, T., Zhang, J., Farquhar, M., Jones, C. T., Balfe, P., Rice, C. M. and McKeating, J. A. (2006) Diverse CD81 proteins support Hepatitis C virus infection. Journal of Virology 80, 11331-11342. Fradkin, L. G., Kamphorst, J. T., DiAntonio, A., Goodman, C. S. and Noordermeer, J. N. (2002) Genomewide analysis of the Drosophila tetraspanins reveals a subset with similar function in the formation of the embryonic synapse. Proceedings of the National Academy of Science 99, 13663–13668. Fukudome, K., Furuse, M., Imai, T., Nishimura, M., Takagi, S., Hinuma, Y. and Yoshie, O. (1992) Identification of membrane antigen C33 recognized by monoclonal antibodies inhibitory to human T-cell leukemia virus type 1 (HTLV-1)-induced syncytium formation: altered glycosylation of C33 antigen in HTLV-1-positive cells. Journal of Virology 66, 1394-1401. Gallay, P., Jongeneel, V., Barras, C., Burrtier, M., Baumgartner, J., Clauser, M. P. and Heumann, D. (1993) Short time exposure to lipopolysaccharide is sufficient to activate human monocytes. Journal of Immunology 150, 5086-5093. Gantner, A., Lerist, M., Lohse, A. W., Germann, P. G. and Tieg, G. (1995) Concanavalin A- induced T-cell-mediated hepatic injury in mice: the role of Tumor Necrosis Factor. Hepatology 21, 190-198. Gao, B. and Tsan, M. (2003a) Endotoxin contamination in recombinant human heat shock protein 70 (Hsp70) preparation is responsible for the induction of Tumor Necrosis Factor α release by murine macrophages. Journal of Biological Chemistry 278, 174- 179. Gao, B. and Tsan, M. (2003b) Recombinant human heat shock protein 60 does not induce the release of Tumor Necrosis Factor α from murine macrophages. Journal of Biological Chemistry 278, 22523–22529. Garcia-Espana, A., Chung, P., Sarkar, I. N., Stiner, E., Sun, T. and DeSalle, R. (2008) Appearance of new tetraspanin genes during vertebrate evolution. Genomics 91, 326–334.

150

Gasser, A. and Most, J. (1999) Generation of multinucleated giant cells in vitro by culture of human monocytes with Mycobacterium bovis BCG in combination with cytokine- containing supernatants. Infection and Immunity 67, 395–402 Giambartolomei, G. H., Zwerdling, A., Cassataro, J., Bruno, L., Fossati, C. A. and Philipp, M. T. (2004) Lipoproteins, not lipopolysaccharide, are the key mediators of the proinflammatory response elicited by heat-killed Brucella abortus. Journal of Immunology 173, 4635-4642. Goldberg, A. F. X., Loewen, C. J. R. and Molday, R. S. (1998) Cysteine residues of photoreceptor peripherin/rds: role in subunit assembly and autosomal dominant retinitis pigmentosa. Biochemistry 37, 680-685. Gordon, S. and Martinez, F. O. (2010) Alternative activation of macrophages: mechanism and functions. Immunity 32, 593-604. Gordon-Alonso, M., Yanez-Mo, M., Barreiro, O., Alvarez, S., Munoz-Fernandez, M. A., Valenzuela-Fernandez, A. and Sanchez-Madrid, F. (2006) Tetraspanins CD9 and CD81 modulate HIV-1-induced membrane fusion. The Journal of Immunology 177, 5129-5137. Green, L. R., Monk, P. N., Partridge, L. J., Morris, P., Gorringe, A. R. and Read, R. C. (2011) Cooperative role for tetraspanins in adhesion-mediated attachment of bacterial species to human epithelial cells. Infection and Immunity 79, 2241-2249 Guan, K. L. and Dixon, J. E. (1991) Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase. Analytical Biochemistry 192, 262–267 Gutierrez-Ramos, J. K. and Bluethmann, H. (1997) Molecules and mechanisms operating in septic shock: lessons from knockout mice. Immunology Today 18, 329-333. Ha, H., Lee, J., Kim, H., Kwak, H. B., Kim, H., Lee, S. E., Rhee, J. H., Kim, H. and Lee, Z. H. (2008) Stimulation by TLR5 modulates osteoclast differentiation through STAT/IFN-β1. The Journal of Immunology 180, 1382-1389. Ha, J., Lee, Y., Kim, H. (2011) CXCL2 mediates lipopolysaccharide-induced osteoclastogenesis in RANKL-primed precursors. Cytokine 55, 48-55. Hasuwa, H., Shishido, Y., Yamazaki, A., Kobayashi, T., Yu, X. and Mekada, E. (2001) CD9 amino acids critical for upregulation of Diphtheria toxin binding. Biochemical and Biophysical Research Communications 289, 782–790. Hayes, M. P., Enterline, J. C., Gerrard, T. L. and Zoon, K. C. (1991) Regulation of interferon production by human monocytes: Requirements for Priming for lipopolysaccharide-induced production. Journal of Leukocyte Biology 50, 176-181. He, B., Liu, L., Cook, G. A., Grgurevich, S., Jennings, L. K. and Zhang, X. A. (2005) Tetraspanin CD82 attenuates cellular morphogenesis through down-regulating integrin α6- mediated cell adhesion. Journal of Biological Chemistry 280, 3346–3354. Hellman, J., Loiselle, P. M., Tehan, M. M., Allaire, J. E., Boyle, L. A., Kurnick, J. T., Andrews, D. M., Kim, K. S. and Warren, H. S. (2000) Outer membrane protein A, peptidoglycan- associated lipoprotein, and murein lipoprotein are released by Escherichia coli bacteria into serum. Infection and Immunity 68, 2566–2572. Helming, L. and Gordon, S. (2008) The molecular basis of macrophage fusion. Immunobiology 212, 785-793. Helming, L. and Gordon, S. (2009) Molecular mediators of macrophage fusion. Trends in Cell Biology 19, 514-522. Hemler, M. E. (2001) Specific tetraspanin functions. The Journal of Cell Biology 155, 1103– 1107. Hemler, M. E. (2003) Tetraspanin proteins mediate cellular penetration, invasion and fusion events and define a novel type of membrane microdomain. Annual Reviews in Cellular and Developmental Biology 19, 397–422. Hemler, M. E. (2005) Tetraspanin functions and associated microdomains. Nature Reviews: Molecular Cell Biology 6, 801-811.

151

Hemler, M. E. (2008) Targeting of tetraspanin proteins – potential benefits and strategies. Nature reviews: Drug Discovery 7, 747-758. Herbert, D. R., Holscher, C., Mohrs, M., Arendse, B., Schwegmann, A., Radwanska, M., Leeto, M., Kirsch, R., Hall, P., Mossmann, H., Claussen, B., Forster, I. and Brombacher, F. (2004) Alternative macrophage activation is essential for survival during Schistomiasis and downmodulates T Helper 1 responses and immunopathology. Immunity 20, 623-635. Hiasa, M., Abe, M., Nakano, A., Oda, A., Amou, H., Kido, S., Takeuchi, K., Kagawa, K., Yata, K., Hashimoto, T., Ozaki, S., Asaoka, K., Tanaka, E., Moriyama, K. and Matsumoto, T. (2009) GM-CSF and IL-4 induce dendritic cell differentiation and disrupt osteoclastogenesis through M-CSF receptor shedding by up-regulation of TNF-α converting enzyme (TACE). Blood 114, 4517-4526. Higginbottom, A., Quinn, E. R., Kuo, C., Flint, M., Wilson, L. H., Bianchi, E., Nicosia, A., Monk, P. N., McKeating, J. A. and Levy, S. (2000) Identification of amino acid residues in CD81 critical for interaction with hepatitis C virus envelope glycoprotein E2. Journal of Virology 74, 3642–3649. Higginbottom, A., Takahashi, Y., Bolling, L., Scott A. Coonrod, S. A., White, J. M., Partridge, L. J. and Monk, P. N. (2003) Structural requirements for the inhibitory action of the CD9 large extracellular domain in sperm/oocyte binding and fusion. Biochemical and Biophysical Research Communications 311, 208–214. Ho, S., Martin, F., Higginbottom, A., Partridge, L. J., Parthasarathy, V., Moseley, G. W, Lopez, P., Cheng-Mayer, C. and Monk, P. N. (2006) Recombinant extracellular domains of tetraspanin proteins are potent inhibitors of the infection of macrophages by Human Immunodeficiency Virus type 1. Journal of Virology 80, 6487–6496. Hotokezaka, H., Sakai, E., Ohara, N., Hotokezaka, Y., Gonzales, C., Matsuo, K., Fujimura, Y., Yoshida, N. and Nakayama, K. (2007) Molecular analysis of RANKL-independent cell fusion of osteoclast-like cells induced by TNF-α, lipopolysaccharide, or peptidoglycan. Journal of Cellular Biochemistry 101, 122-134. Huang, S., Yuan, S., Dong, M., Su J., Yu, C., Shen, Y., Xie, X., Yu, Y., Yu, X., Chen, S., Zhang, S., Pontarotti, P. and Xu, A. (2005) The phylogenetic analysis of tetraspanins projects the evolution of cell–cell interactions from unicellular to multicellular organisms. Genomics 86, 674–684. Hyakushima, N., Mitsuzawa, H., Nishitani, C., Sano, H., Kuronuma, K., Konishi, M., Himi, T., Miyake, K. and Kuroki, Y. (2004) Interaction of soluble form of recombinant extracellular TLR4 domain with MD-2 enables lipopolysaccharide binding and attenuates TLR4-mediated signalling. The Journal of Immunology 173, 6949–6954. Imai, T., Fukudome, K., Takagi, S., Nagira, M., Furuse, M., Fukuhara, N., Nishimura, M., Hinuma, Y. and Yoshie, O. (1992) C33 antigen recognised by monoclonal antibodies inhibitory to human T cell leukemia virus type 1-induced syncytium formation is a member of a new family of transmembrane proteins including CD9, CD37, CD53, and CD63. The Journal of Immunology 149, 2879-2886. Ishii, M., Iwai, K., Koike, M., Ohshima, S., Kudo-Tanaka, E., Ishii, T., Mima, T., Katada, Y., Miyatake, K., Uchiyama, Y. and Saeki, Y. (2006) RANKL-induced expression of tetraspanin CD9 in lipid raft membrane microdomain is essential for cell fusion during osteoclastogenesis. Journal of Bone and Mineral Research 21, 965-976. Ishii, M. and Saeki, Y. (2008) Osteoclast cell fusion: mechanisms and molecules. Modern Rheumatology 18, 220-227. Ishizuka, H., García-Palacios, V., Lu, G., Subler, M. A., Zhang, H., Boykin, C. S., Choi, S. J., Zhao, L., Patrene, K., Galson, D. B., Blair, H. C., Hadi, T. M., Windle, J. J., Kurihara, N. and Roodman, G. D. (2011) ADAM8 enhances osteoclast precursor fusion and osteoclast formation in vitro and in vivo. Journal of Bone and Mineral Research 26, 169–181. Iwai, K., Ishii, M., Ohshima, S., Miyatake, K. and Saeki, Y. (2007) Expression and function of transmembrane-4 superfamily (tetraspanin) proteins in osteoclasts: Reciprocal

152

roles of Tspan-5 and NET-6 during osteoclastogenesis. Allergology International 56, 457-463. Iwai, K., Ishii, M., Ohshima, S., Miyatake, K. and Saeki, Y. (2008) Abundant expression of tetraspanin CD9 in activated osteoclasts in ovariectomy-induced osteoporosis and in bone erosions of collagen-induced arthritis. Rheumatology International 28, 225–231. Iwasaki, A. and Medzhitov, R. (2010) Regulation of adaptive immunity by the innate immune system. Science 327, 291-295. Jaber, B. L., Sundaram, S., Neto, M. C., King, A. J. and Pereira, B. J. (1998) Polymyxin-B stimulates tumor necrosis factor-alpha production by human peripheral blood mononuclear cells. International Journal of Artificial Organs 21, 269–273. Jay, S. M., Skokos, E., Laiwalla, F., Krady, M. and Kyriakides, T. R. (2007) Foreign body giant cell formation is preceded by lamellipodia formation and can be attenuated by inhibition of Rac1 activation. The American Journal of Pathology 171, 632-640. Jégou, A., Ziyyat, A., Barraud-Lange, V., Perez, E., Wolf, J. P., Pinceta, F. and Gourier, C. (2011) CD9 tetraspanin generates fusion competent sites on the egg membrane for mammalian fertilization. Proceedings of the National Academy of Science 108, 10946–10951. Ji, J., Park-Min, K., Shen, Z., Fajardo, R. J., Goldring, S. R., McHugh, K. P. and Ivashkiv, L. B. (2009) Inhibition of RANK expression and osteoclastogenesis by TLRs and IFN-γ in human osteoclast precursors. The Journal of Immunology 183, 7223–7233. Jurado, P., Ritz, D., Beckwith, J., De Lorenzo, V. and Fernández, L. A. (2002) Production of functional single-chain Fv antibodies in the cytoplasm of Escherichia coli. Journal of Molecular Biology 320, 1–10. Kaji, K., Oda, S., Shikano, T., Ohnuki, T., Uematsu, Y., Sakagami, J., Tada, N., Miyazaki, S. and Kudo, A. (2000) The gamete fusion process is defective in eggs of Cd9-deficient mice. Nature Genetics 24, 279–282. Kaji, K., Oda, S., Miyazaki, S. and Kudo, A. (2002) Infertility of CD9-deficient mouse eggs is reversed by mouse CD9, human CD9, or mouse CD81; Polyadenylated mRNA injection developed for molecular analysis of sperm-egg fusion. Developmental Biology 247, 327–334. Kajiwara, K., Berson, E. L. and Dryja, T. P. (1994) Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM-1 loci. Science 264, 1604- 1608. Kaplan, W., Husler, P., Klump, M., Erdhart, J., Sluis-Cremer, N. and Dirr, H. (1997) Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S- transferase: A detoxification enzyme and fusion-protein affinity tag. Protein Science 6, 399-406. Karamatic Crew, V., Burton, N., Kagan, A., Green, C. A., Levene, C., Flinter, F., Brady, R. L., Daniels, G. and Anstee, D. J. (2004) CD151, the first member of the tetraspanin (TM4) superfamily detected on erythrocytes, is essential for the correct assembly of human basement membranes in kidney and skin. Blood 104, 2217–2223. Kazarov, A. R., Yang, X., Stipp, C. S., Sehgal, B. and Hemler, M. E. (2002) An extracellular site on tetraspanin CD151 determines integrin–dependent cellular morphology. The Journal of Cell Biology 158, 1299–1309. Kedzierski, W., Bok, D. and Travis, G. H. (1999) Transgenic analysis of rds/peripherin N- glycosylation: effect on dimerization, interaction with Rom1, and rescue of the rds null phenotype. Journal of Neurochemistry 72, 430–438. Kelic, S., Levy, S., Suarez, C. and Weinstein, D. E. (2001) CD81 regulates neuron-induced astrocyte cell-cycle exit. Molecular Cell Neuroscience 17, 551–560. Kelly, S. M., Jess, T. M. and Price, N. C. (2005) How to study proteins by circular dichroism. Biochemica et Biophysica Acta 1751, 119-139. Kespichayawattana, W., Rattanachetkul, S., Wanun, T., Utaisincharoen, P. and Sirisinha, S. (2000) Burkholderia pseudomallei induces fusion and actin-associated membrane

153

protrusion: A possible mechanism for cell-to-cell spreading. Infection and Immunity 68, 5377-5384. Kitadokoro, K., Bordo, D., Galli, G., Petracca, R., Falugi, F., Abrignani, S., Grandi, G. and Bolognesi, M. (2001) CD81 extracellular domain 3D structure: insight into the tetraspanin superfamily structural motifs. EMBO Journal 25, 12–18. Klasse, P. J., Bron, R. and Marsh, M. (1998) Mechanisms of enveloped virus entry into animal cells. Advanced Drug Delivery Reviews 34, 65–91. Kobayashi, M., Bennett, M. C., Bercot, T. and Singh, I. R. (2006) Functional analysis of hepatitis C virus envelope proteins, using a cell-cell fusion assay. Journal of Virology 80, 1817–1825. Kovalenko, O. V., Yang, X., Kolesnikova, T. V. and Hemler, M. E. (2004) Evidence for specific tetraspanin homodimers: inhibition of palmitoylation makes cysteine residues available for cross-linking. Biochemical Journal 377, 407–417. Kovalenko, O. V., Metcalf, D. G., DeGrado, W. F. and Hemler, M. E. (2005) Structural organization and interactions of transmembrane domains in tetraspanin proteins. BioMed Central: Structural Biology 5, 11. Krementsov, D. N., Weng, J., Lambelé, M., Roy, H. N. and Thali, M. (2009) Tetraspanins regulate cell-to-cell transmission of HIV-1. Retrovirology 6, 64–73. Krementsov, D. N., Rassam, P., Margeat, E., Roy, N. H., Schneider-Schaulies, J., Milhiet, P. and Thali, M. (2010) HIV-1 assembly differentially alters dynamics and partitioning of tetraspanins and raft components. Traffic 11, 1401–1414. Kruger, D. (1989) Assessing the quality of medicinal products containing ingredients obtained by gene technology. Drugs Made Germany 32, 64–67. Kumar, S. and Jack, R. (2006) Origin of monocytes and their differentiation to macrophages and dendritic cells. Journal of Endotoxin Research 12, 278-284. Kumar, H., Kawai, T. and Akira, S. (2009) Toll-like receptors and innate immunity. Biochemical and Biophysical Research Communications 388, 621–625. Lagaudriere-Gesbert, C., Le Naour, F., Lebel-Binay, S., Billard, M., Lemichez, E., Boquet, P., Boucheix, C., Conjeaud, H. and Rubinstein, E. (1997) Functional analysis of four tetraspans, CD9, CD53, CD81 and CD82, suggests a common role in costimulation, cell adhesion and migration: only CD9 upregulates HB-EGF activity. Cellular Immunology 182, 105–112. Landmann, R., Scherer, F., Schumann, R., Link, S., Sansano, S. and Zimmerli, W. (1995) LPS directly induces oxygen radical production in human monocytes via LPS binding protein and CD14. Journal of Leukocyte Biology 57, 440-449. Lane, L. C. (1978) A Simple method for stabilizing protein-sulfhydryl groups during SDS- gel electrophoresis. Analytical Biochemistry 86, 655-664. Larsson, L., Bjerregaard, B. and Talts, J. F. (2008) Cell fusion in mammals. Histochemistry and Cell Biology 129, 551-561. Lauener, R. P., Geha, R. S. and Vercelli, D. (1990) Engagement of the monocyte surface antigen CD14 induces lymphocyte-associated antigen-1/intracellular adhesion molecule-1-dependent homotypic adhesion. The Journal of Immunology 145, 1390- 1394. Lay, G., Poquet, Y., Salek-Peyron, P., Puissegur, M. P., Botanch, C., Bon, H., Levillain, F., Duteyrat, J. L., Emile, J. F and Altare, F. (2007) Langhans giant cells from M. tuberculosis-induced human granulomas cannot mediate mycobacterial uptake. Journal of Pathology 211, 76–85. Lazarus, D., Yamin, M., McCarthy, K., Schneeberger, E. E. and Kradin, R. (1990) Anti-RMA, a murine monoclonal antibody, activates rat macrophages. II. Induction of DNA synthesis and formation of multinucleated giant cells. American Journal of Respiratory Cell and Molecular Biology 3, 103-111. Le Naour, F., Rubinstein, E., Jasmin, C., Prenant, M. and Boucheix, C. (2000) Severely reduced female fertility in CD9-deficient mice. Science 287, 319–321.

154

Lee, Y., Huang, H., Kim, H. J., Park, C. and Kim, H. (2008) The phosphatidylinositol 3-kinase- mediated production of interferon-β is critical for the lipopolysaccharide inhibition of osteoclastogenesis. Life Sciences 83, 369-376. Lee, B., Kim, T., Jun, J., Yoo, D., Woo, J., Choi, S. D., Lee, Y. H., Song, G. G., Sohn, J., Park-Min, K., Ivashkiv, L. B. and Ji, D. J. (2010) Direct inhibition of human RANK+ osteoclast precursors identifies a homeostatic function of IL-1β. The Journal of Immunology 185, 5926-5934. Levy, S. and Shoham, T. (2005) Protein-protein interactions in the tetraspanin web. Physiology 20, 218-224. Lewthwaite, J. O., Coates, A. R. M., Tormay, P., Singh, M., Mascagni, P., Poole, S., Roberts, M., Sharp, L. and Henderson, B. (2001) Mycobacterium tuberculosis chaperonin 60.1 is a more potent cytokine stimulator than chaperonin 60.2 (Hsp 65) and contains a CD14-binding domain. Infection and Immunity 69, 7349–7355. Liang, F.X., Riedel, I., Deng, F. M., Zhou, G., Xu, C., Wu, X. R., Kong, X. P., Moll, R. and Sun, T. T. (2001) Organization of uroplakin subunits: transmembrane topology, pair formation and plaque composition. Biochemical Journal 355, 13–18. Liu, L., He, B., Liu, W. M., Zhou, D., Cox, J. V. and Zhang, Z. A. (2007) Tetraspanin CD151 promotes cell migration by regulating integrin trafficking. The Journal of Biological Chemistry 282, 31631-31642. Liu, J., Wang, S., Zhang, P., Said-Al-Naief, N., Michalek, S. M. and Feng, X. (2009) Molecular mechanism of the bifunctional role of lipopolysaccharide in osteoclastogenesis. The Journal of Biological Chemistry 284, 12512-12523. Lobstein, J., Emrich, C. A., Jeans, C., Faulkner, M., Riggs, P. and Berkmen, M. (2012) SHuffle, a novel Escherichia coli protein expression strain capable of correctly folding disulfide bonded proteins in its cytoplasm. Microbial Cell Factories 11, 56. Loewen, C. J. R. and Molday, R. S. (2000) Disulfide-mediated oligomerization of peripherin/RDS and Rom-1 in photoreceptor disk membranes: Implications for photoreceptor outer segment morphogenesis and degeneration. Journal of Biological Chemistry 275, 370–5378. Loffer, S., Lottspeich, F., Lanza, F., Azorsa, D. O., ter Meulen, V. and Schneider-Schaulies, J. (1997) CD9, a tetraspan transmembrane protein, renders cells susceptible to canine distemper virus. Journal of Virology 71, 42–49. Luong, M., Zhang, Y., Chamberlain, T., Zhou, T., Wright, J. F., Dower, K. and Hall, J. P. (2012) Stimulation of TLR4 by recombinant HSP70 requires structural integrity of the HSP70 protein itself. Journal of Inflammation 9, 11. Lyons, A.B. (2000) Analysing cell division in vivo and in vitro using flow cytometric measurement of CFSE dye dilution. Journal of Immunological Methods 243 147- 154. McNally, A. K. and Anderson, J. M. (1995) Interleukin-4 induces foreign body giant cells from human monocytes/macrophages: differential lymphokine regulation of macrophage fusion leads to morphological variants of multinucleated giant cells. American Journal of Pathology 147, 1487-1499. McNally, A, K., DeFife, K. M. and Anderson, J. M. (1996) Interleukin-4-induced macrophage fusion is prevented by inhibitors of mannose receptor activity. The American Journal of Pathology 149, 975–985. McNally, A. K. and Anderson, J. M. (2002) β1 and β2 integrins mediate adhesion during macrophage fusion and multinucleated foreign body giant cell formation. American Journal of Pathology 160, 621-630. McNally, A. K. and Anderson, J. M. (2005) Multinucleated giant cell formation exhibits features of phagocytosis with participation of the endoplasmic reticulum. Experimental and Molecular Pathology 79, 126-135. Maas, J. J. J., Gange, S. J., Schuitemaker, H., Coutinho, R. A., van Leeuwen, R. and Margolick, J. B. (2000) Strong association between failure of T cell homeostasis and the

155

syncytium-inducing phenotype among HIV-1-infected men in the Amsterdam Cohort Study. AIDS 14, 1155-1161. MacLauchlan, S., Skokos, E. A., Meznarich, N., Zhu, D., H., Raoof, S., Shipley, M., Senior, R. M., Bornstein, P. and Kyriakides, T. M. (2009) Macrophage fusion, giant cell formation, and the foreign body response require matrix metalloproteinase 9. Journal of Leukocyte Biology 85, 617-626. Maecker, H. T., Todd, S. C. and Levy, S. (1997) The tetraspanin superfamily: molecular facilitators. FASEB 11, 428-442. Mangan, D. F., Welch, G. R. and Wahl, S. M. (1991) Lipopolysaccharide, Tumor Necrosis Factor-α, and IL-1β prevent programmed cell death (apoptosis) in human peripheral blood monocytes. The Journal of Immunology 146, 1541-1546. Marks, M. S., Ohno, H., Kirchhausen, T. and Bonifacino, S. J. (1997) Protein sorting by tyrosine-based signals: adapting to the Ys and wherefores. Trends in Cell Biology 7, 124–128. Martens, S. and McMahon, H. T. (2008) Mechanisms of membrane fusion: disparate players and common principles. Nature Reviews: Molecular Biology 9, 543-556. Martin, F., Roth, D. M., Jans, D. A., Pouton, W. C., Partridge, L. J., Monk, P. N. and Mosely, G. W. (2005) Tetraspanins in viral infections: a fundamental role in viral biology? Journal of Virology 279, 10839–10851. Martinez, F. O., Gordon, S., Locati, M. and Mantovani, A. (2006) Transcriptional profiling of the human monocyte-to-macrophage differentiation and polarization: new molecules and patterns of gene expression. The Journal of Immunology 177, 7303- 7311. Maruotti, N., Grano, M., Colucci, S., D’Onofrio, F. and Cantatore, F. P. (2011) Osteoclastogenesis and arthritis. Clinical and Experimental Medicine 11, 137–145. Mason, J. M. and Arndt, K. M. (2004) Coiled coil domains: stability, specificity, and biological implications. ChemBioChem 5, 170-176. Mazurov, D., Heidecker, G. and Derse, D. (2006) HTLV-1 Gag protein associates with CD82 tetraspanin microdomains at the plasma membrane. Virology 346, 194-204. Meerloo, T., Sheikh, M. A., Bloem, A. C., de Ronde, A., Schutten, M., van Els, C. A., Roholl, P. J., Joling, P., Goudsmit, J. and Schuurman, H. J. (1993) Host cell membrane proteins on human immunodeficiency virus type 1 after in vitro infection of H9 cells and blood mononuclear cells: An immuno-electron microscopic study. Journal of Genetic Virology 74, 129–135. Mensah, K. A., Ritchlin, C. T. and Schwarz, E. M. (2010) RANKL induces heterogeneous DC- STAMP(lo) and DC-STAMP(hi) osteoclast precursors of which the DC-STAMP(lo) precursors are the master fusogens. Journal of Cell Physiology 223, 76–83. Mentzer, S. J., Faller, D. V. and Burakoff, S. J. (1986) Interferon-γ induction of LFA-1- mediated homotypic adhesion of human monocytes. The Journal of Immunology 137, 108-113. Miller, B. J., Georges-Labouesse, E., Primakoff, P. and Myles, D. G. (2000) Normal fertilization occurs with eggs lacking the integrin alpha6beta1 and is CD9- dependent. Journal of Cell Biology 149, 1289–1296. Min, G., Wang, H., Sun, T. and Kong, X. (2006) Structural basis for tetraspanin functions as revealed by the cryo-EM structure of uroplakin complexes at 6-Å resolution. Journal of cell Biology 173, 975–983. Miyado, K., Yamada, G., Yamada, S., Hasuwa, H., Nakamura, Y., Ryu, F., Suzuki, K., Kosai, K., Inoue, K. Ogura, A., Okabe, M. and Mekada, E. (2000) Requirement of CD9 on the egg plasma membrane for fertilization. Science 287, 321–324. Miyamoto, H., Suzuki, T., Miyauchi, Y., Iwasaki, R., Kobayashi, T., Sato, Y., Miyamoto, K., Hoshi, H., Hashimoto, K., Yoshida, S., Hao, W., Mori, T., Kanagawa, H., Katsuyama, E., Fujie, A., Morioka, H., Matsumoto, M., Chiba, K., Takeya, M., Toyama, T. and Miyamoto, T. (2012) Osteoclast Stimulatory Transmembrane Protein and Dendritic Cell–Specific Transmembrane Protein cooperatively modulate cell-cell

156

fusion to form osteoclasts and foreign and body giant cells. Journal of Bone and Mineral Research 27, 1289-1297. Moreno, J. L., Kaczmarek, M., Keegan, A. D. and Tondravi, M. (2003) IL-4 suppresses osteoclast development and mature osteoclast function by a STAT6-dependent mechanism: irreversible inhibition of the differentiation program activated by RANKL. Blood 102, 1078-1086. Mosser, D. M. (2003) The many faces of macrophage activation. Journal of Leukocyte Biology 73, 209-212. Mosser, D. M. and Edwards, J. P. (2008) Exploring the full spectrum of macrophage activation. Nature Reviews Immunology 8, 958-969. Most, J., Neumayer, H. P. and Dierich, M. E. (1990) Cytokine-induced generation of multinucleated giant cells in vitro requires interferon-γ and expression of LFA- 1. European Journal of Immunology 20, 1661-1667. Most, J., Spotl, L., Mayr, G., Gasser, A., Sarti, A. and Dierich, M. P. (1997) Formation of multinucleated giant cells in vitro is dependent on the stage of monocyte to macrophage maturation. Blood 89, 662-671. Mustafa, T., Bjune, G., Jonsson, R., Hernandez Pando, R. and Nilsen, R. (2001) Increased expression of Fas ligand in human tuberculosis and leprosy lesions: a potential novel mechanism of immune evasion in Mycobacterial infection. Scandinavian Journal of Immunology 54, 630-639. Muta, T., Miyata, T., Misumi, Y., Tokunaga, F., Nakamura,T., Tohe, Y., Ikehara, Y. and Iwanaga, S. (1991) Limulus Factor C: An endotoxin-sensitive serine protease zymogen with a mosaic structure of complement-like epidermal factor-like, and lectin-like domains. The Journal of Biological Chemistry 266, 6554-6561. Namba, K., Nishio, M., Mori, K., Miyamoto, N., Tsurudome, M., Ito, M., Kawano, M., Uchida, A. and Ito, Y. (2001) Involvement of ADAM9 in multinucleated giant cell formation in blood monocytes. Cellular Immunology 213, 104–113. Nishiuchi, R., Sanzen, N., Nada, S., Sumida, Y., Wada, Y., Okada, M., Takagi, J., Hasegawa, H. and Sekiguchi, K. (2005) Potentiation of the ligand-binding activity of integrin α3β1 via association with tetraspanin CD151. Proceedings of the National Academy of Science 102, 1939–1944. Nydegger, S., Khurana, S., Krementsov, D. N., Foti, M. and Thali, M. (2006) Mapping of tetraspanin-enriched microdomains that can function as gateways for HIV-1. The Journal of Cell Biology 173, 795-807. Odintsova, E., Butters, T. D., Monti, E., Sprong, H., van Meer, G. and Berditchevski, F. (2006) Gangliosides play an important role in the organization of CD82- enrichedmicrodomains. Biochemical Journal 400, 315–325. Okamoto, H., Mizuno, K. and Horio, T. (2003) Monocyte-derived multinucleated giant cells and Sarcoidosis. Journal of Dermatological Science 31, 119-128. Oren, R., Takahashi, S., Doss, C., Levy, R. and Levy, S. (1990) TAPA 1, the target of an antiproliferative antibody, defines a new family of transmembrane proteins. Molecular Cell Biology 10, 4007–4015. Oren-Suissa, M. and Podbilewicz, B. (2007) Cell fusion during development. Trends in cell biology 17, 537-546. Oursler, M. J. (2010) Recent advances in understanding the mechanisms of osteoclast precursor fusion. Journal of Cellular Biochemistry 110, 1058–1062. Palmqvist, P., Lundberg, P., Persson, E., Johansson, A., Lundgren, I., Lie, A., Conaway, H. H. and Lerner, U. H. (2006) Inhibition of hormone and cytokine-stimulated osteoclastogenesis and bone resorption by interleukin-4 and interleukin-13 is associated with increased osteoprotegerin and decreased RANKL and RANK in a STAT6-dependent pathway. The Journal of Biological Chemistry 281, 2414-2429. Parry, D. A. D., Fraser, R. B. D. and Squire, J. M. (2008) Fifty years of coiled-coils and α- helical bundles: A close relationship between sequence and structure. Journal of Structural Biology 163, 258-269.

157

Parthasarathy, V., Martin, F., Higginbottom, A., Murray, H., Moseley, G. W., Read, R. C., Mal, G., Hulme, R., Monk, P. N. and Partridge, L. J. (2009) Distinct roles for tetraspanins CD9, CD63 and CD81 in the formation of multinucleated giant cells. Immunology 127, 237-248. Pavlou, A. K. and Belsey, M. J. (2005) The therapeutic antibodies market to 2008. European Journal of Pharmaceutics and Biopharmaceutics 59, 389–396. Pearson, M. S., Pickering, D. A, McSorley, H. J., Bethony, J. M., Tribolet, L., Dougall, A. M., Hotez, P. J. and Loukas, A. (2012) Enhanced protective efficacy of a chimeric form of the Schistosomiasis vaccine antigen Sm-TSP-2. PLoS Neglected Tropical Diseases 6, 1564. Peterson, A. A., Haug, A. and McGroarty, E. E. (1986) Physical properties of short- and long-O-antigen-containing fractions of lipopolysaccharide from Escherichia coli O111:B4. Journal of Bacteriology 165, 116-122. Petracca, R., Falugi, F., Galli, G., Rosa, D., Campagnoli, S., Burgio, V., Stasio, D. E., Giardina, B., Houghton, M., Abrignani, S. and Grandi, G. (2000). Structure-function analysis of Hepatitis C virus envelope-CD81 binding. Journal of Virology 74, 4824–4830. Petsch, D. and Anspach, F. B. (2000) Endotoxin removal from protein solutions. Journal of Biotechnology 76, 97-119. Pierce, K. L., Premont, R. T. and Lefkowitz, R. J. (2002). Seven-transmembrane receptors. Nature Reviews: Molecular Cell Biology 3, 639–650. Pileri, P., Uematsu, Y., Campagnoli, S., Galli, G., Falugi, F., Petracca, R., Weiner, A. J., Houghton, M., Rosa, D., Grandi, G. and Abrignani, S. (1998) Binding of hepatitis C virus to CD81. Science 282, 938–941. Pique, C. Lagaudriere-Gesbert, C., Delamarre, L., Rosenberg, A. R., Conjeaud, H. and Dokhelar, M. C. (2000) Interaction of CD82 tetraspanin proteins with HTLV-1 envelope glycoproteins inhibits cell-to-cell fusion and virus transmission. Virology 276, 455-465. Plattet, P., Rivalsa, J. P., Zubera, B., Brunnera, J. M., Zurbriggenb, A. and Wittek, R. (2005) The fusion protein of wild-type canine distemper virus is a major determinant of persistent infection. Virology 337, 312-326. Puissegur, M., Lay, G., Gilleron, M., Botella, L., Nigou, J., Marrakchi, H., Mari, B., Duteyrat, J., Guerardel, Y., Kremer, L., Barbry, P., Puzo, G. and Altare, F. (2007) Mycobacterial lipomannan induces granuloma macrophage fusion via a TLR2-dependent, ADAM9- and β1 integrin-mediated pathway. The Journal of Immunology 178, 3161- 3169. Qiu, J., Swartz, J. R. and Georgiou, G. (1998) Expression of active human tissue-type plasminogen activator in Escherichia coli. Applied Environmental Microbiology 64, 4891–4896. Rajesh, S., Sridhar, P., Tews, B. A., Fénéant, F., Cocquerel, L., Ward, D. G., Berditchevski, F. and Overduin, M. (2012) Structural basis of ligand interactions of the large extracellular domain of tetraspanin CD81. Journal of Virology 86, 9606-9616. Reed, R. C., Berwin, B., Baker, J. P. and Nicchitta, C. V. (2003) GRP94/gp96 elicits ERK activation in murine macrophages: A role for endotoxin contamination in NF-κB activation and nitric oxide production. Journal of Biological Chemistry 278, 31853- 31860. Reichelt, P., Schwarz, C. and Donzeau, M. (2006) Single step protocol to purify recombinant proteins with low endotoxin contents. Protein Expression and Purification 46, 483-488. Rein, A., Mirro, J., Gordon Haynes, J., Ernst, S. M. and Nagashima, K. (1994) Function of the cytoplasmic domain of a retroviral transmembrane protein: pl5E-p2E cleavage activates the membrane fusion capability of the murine leukemia virus Env protein. Journal of Virology 68, 1773-1781. Ritter, L. M., Arakawa, T. and Goldberg, A. F. (2005) Predicted and measured disorder in peripherin/rds, a retinal tetraspanin. Protein Peptide Letters 12, 677-686.

158

Rivollier, A., Mazzorana, M., Tebib, J., Piperno, M., Aitsiselmi, T., Rabourdin-Combe, C., Jurdic, P. and Servet-Delprat, C. (2004) Immature dendritic cell transdifferentiation into osteoclasts: a novel pathway sustained by the rheumatoid arthritis microenvironment. Blood 104, 4029-4037. Robb, L., Tarrant, J., Groom, J., Ibrahim, M., Li, R., Borobakas, B. and Wright, M. D. (2001) Molecular characterisation of mouse and human TSSC6: evidence that TSSC6 is a genuine member of the tetraspanin superfamily and is expressed specifically in haematopoietic organs. Biochimica et Biophysica Acta 1552, 31–41. Rosenberg, S. A. (2001) Progress in human tumour immunology and immunotherapy. Nature 411, 380-384. Rous, B. A., Reaves, B. J., Ihrke, G., Briggs, J. A., Gray, S. R., Stephens, D. J., Banting, G. and Luzio, J. P. (2002) Role of adaptor complex AP-3 in targeting wildtype and mutated CD63 to lysosomes. Molecular Biology of the Cell 13, 1071–1082. Rubinstein, E., Ziyyat, A., Prenant, M., Wrobel, E., Wolf, J. P., Levy, S., Le Naour, F. and Boucheix, C. (2006a) Reduced fertility of female mice lacking CD81. Developmental Biology 290, 351–358. Rubinstein, E., Ziyyat, A., Wolf, J. P., Le Naour, F. and Boucheix, C. (2006b) The molecular players of sperm–egg fusion in mammals. Seminars in Cell & Developmental Biology 17, 254–263. Runge, K. E., Evans, J. E., He, Z., Gupta, S., McDonald, K. L., Stahlberg, H., Primakoff, P. and Myles, D. G. (2007) Oocyte CD9 is enriched on the microvillar membrane and required for normal microvillar shape and distribution. Developmental Biology 304, 317-325. Sabroe, I., Jones, E. C., Usher, L. R., Whyte, M. K. B. and Dower, S. K. (2002) Toll-like receptor (TLR)2 and TLR4 in human peripheral blood granulocytes: A critical role for monocytes in leukocyte lipopolysaccharide responses. The Journal of Immunology 168, 4701–4710. Sakai, H., Okafuji, I., Nishikomori, R., Abe., J., Izawa, K., Kambe, N., Yasumi, T., Nakahata, T. and Heike, T. (2011) The CD40–CD40L axis and IFN-γ play critical roles in Langhans giant cell formation. International Immunology 24, 5-15. Sala-Valdes, M., Ursa, A., Charrin, S., Rubinstein, E., Hemler, M. E., Sanchez-Madrid, F. and Yanez-Mo, M. (2006) EWI-2 and EWI-F link the tetraspanin web to the actin cytoskeleton through their direct association with Ezrin-Radixin-Moesin proteins. The Journal of Biological Chemistry 281, 19665–19675. Santanirand, P., Harley, V. S., Dance, D. A. B., Drasar, B. S. and Bancroft, G. J. (1999) Obligatory role of Gamma Interferon for host survival in a murine model of infection with Burkholderia pseudomallei. Infection and Immunity 67, 3593-3600. Sapir, A., Avinoam, O., Podbilewicz, B. and Chernomordik, L. V. (2008) Viral and developmental cell fusion mechanisms: conservation and divergence. Developmental Cell, 14, 11-21. Sarantos, K. and Cleo, K. (2013) Analysis of the landscape of biologically-derived pharmaceuticals in Europe: Dominant production systems, molecule types on the rise and approval trends. European Journal of Pharmaceutical Sciences 48, 428– 441. Sato, T., Aoki, J., Misawa, N., Daikoku, E., Sano, K., Tanaka, Y. and Koyanagi, Y. (2008) Modulation of Human Immunodeficiency Virus Type 1 infectivity through incorporation of tetraspanin proteins. Journal of Virology 82, 1021-1033. Schmid, E., Zurbriggen, A., Gassen, U., Rima, B., ter Meulen, V. and Schneider-Schaulies, J. (2000) Antibodies to CD9, a tetraspan transmembrane protein, inhibit canine distemper virus-induced cell-cell fusion but not virus-cell fusion. Journal of Virology 74, 7554–7561. Schwander, M., Leu, M., Stumm, M., Dorchies, O. M., Ruegg, U. T., Schittny, J. and Muller, U. (2003) β1 integrins regulate myoblast fusion and sarcomere assembly. Developmental Cell 4, 673-685.

159

Seigneuret, M., Delaguillaumie, A., Lagaudrie`re-Gesbert, C. and Conjeaud, H. (2001) Structure of the tetraspanin main extracellular domain: A partially conserved fold with a structurally variable domain insertion. The Journal of Biological Chemistry 276, 40055–40064. Seigneuret, M. (2006) Complete predicted three-dimensional structure of the facilitator transmembrane protein and Hepatitis C virus receptor CD81: Conserved and variable structural domains in the tetraspanin superfamily. Biophysical Journal 90, 212–227. Sharma, S. K. (1986) Endotoxin detection and elimination in biotechnology. Biotechnology and Applied Biochemistry 8, 5–22. Shaw, M. L., Stone, K. L., Colangel, C. M., Gulcicek, E. E. and Palese, P. (2008) Cellular proteins in influenza virus particles. PLoS Pathogens 4, e1000085. Shoham, T., Rajapaksa, R., Kuo, C., Haimovich, J. and Levy, S. (2006) Building of the tetraspanin web: distinct structural domains of CD81 function in different cellular compartments. Molecular and Cellular Biology 26, 1373–1385. Singh, S. M. and Panda, K. P. (2005) Solubilization and refolding of bacterial inclusion body proteins. Journal of Bioscience and Bioengineering 99, 303-310. Singethan, K., Topfstedt, E., Schubert, S., Duprex, W. P., Rima, B. K. and Schneider- Schaulies, J. (2006) CD9-dependent regulation of Canine distemper virus-induced cell–cell fusion segregates with the extracellular domain of the haemagglutinin. Journal of General Virology 87, 1635–1642. Singethan, K., Muller, N., Schubert, S., Luttge, D., Krementsov, D. N., Khurana, R. S., Krohne, G., Schneider-Schaulies, S., Thali, M. and Schneider-Schaulies, J. (2008) CD9 clustering and formation of microvilli zippers between contacting cells regulates virus-induced cell fusion. Traffic 9, 924-935. Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D., Warren, J. T., Bokesch, H., Kenney, S. and Boyd, M. R. (1990) New colorimetric cytotoxicity assay for anticancer-drug screening. Journal of the National Cancer Institute 82, 1107- 1112. Smith, D. B. and Johnson, K. S. (1988) Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67, 31-40. Smith, S. B. and Revels, J. (1972) Mapping of Concanavalin A binding sites on the surface of several cell types. Developmental Biology 27, 434-441. Sodhi, A., Tarang, S. and Cesherwani, V. (2007) Concanavalin A induced expression of Toll- like receptors in murine peritoneal macrophages in vitro. International Immunopharmacology 7, 454, 463. Sørensen, H. P. and Mortensen, K. K. (2005) Soluble expression of recombinant proteins in the cytoplasm of Escherichia coli. Microbial Cell Factories 4, 1. Stefani, M. and Dobson, C. M. (2003) Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution. Journal of Molecular Medicine 81, 678-699. Stefano, F. P., Krouse, J., Marta, P. and Boesze-Battaglia, K. (2002) Heterologous expression of WT and mutant peripherin/rds in Madine Darby canine kidney cells: an assessment of fusogenic function. Experimental Eye Reseach 74, 267–283. Stein, K. K., Primakoff, P. and Myles, D. (2004) Sperm-egg fusion: events at the plasma membrane. Journal of Cell Science 117, 6269-6274. Sterk, L. M. T., Geuijen, C. A. W., Oomen, L. C. J. M., Calafat, J., Janssen, H. and Sonnenberg, A. (2000) The tetraspan molecule CD151, a novel constituent of hemidesmosomes, associates with the integrin α6β4 and may regulate the spatial organisation of hemidesmosomes. The Journal of Cell Biology 149, 969–982. Stipp, C. S., Kolesnikova, T. V. and Hemler, M. E. (2003) Functional domains in tetraspanin proteins. Trends in Biochemical Sciences 28, 106-112. Stirewalt, D. L. and Radich, J. P. (2003) The role of FLT3 in haematopoietic malignancies. Nature Reviews Cancer 3, 650-665.

160

Sun, T. T., Zhao, H., Provet, J., Aebi, U. and Wu, X. R. (1996) Formation of asymmetric unit membrane during urothelial differentiation. Molecular Biology Reports 23, 3–11. Suparak, S., Kespichayawattana, W., Haque, A., Easton, A., Damnin, S., Lertmemongkolchai, G., Bancroft, G. J. and Korbsrisate, S. (2005) Multinucleated giant cell formation and apoptosis in infected host cells is mediated by Burkholderia pseudomallei type III secretion protein BipB. Journal of Bacteriology 187, 6556-6560. Suparak, S., Muangsombut, V., Riyapa, D., Stevens, J. M., Stevens, M. P., Lertmemongkolchai, G. and Korbsrisate, S. (2011) Burkholderia pseudomallei- induced cell fusion in U937 macrophages can be inhibited by monoclonal antibodies against host cell surface molecules. Microbes and Infection 13, 1006- 1011. Tachibana, I. and Hemler, M. E. (1999) Role of transmembrane 4 superfamily (TM4SF) proteins CD9 and CD81 in muscle cell fusion and myotube maintenance. Journal of Cell Biology 146, 893–904. Tait, A. R. and Straus, S. K. (2011) Overexpression and purification of U24 from human herpes virus type-6 in E. coli: unconventional use of oxidizing environments with a maltose binding protein-hexahistine dual tag to enhance membrane protein yield. Microbial Cell Factories 10, 51. Takami, M., Kim, N., Rho, J. and Choi, Y. (2002) Stimulation by Toll-Like receptors inhibits osteoclast differentiation. Journal of Immunology 169, 1516-1523. Takashima, T., Ohnishi, K., Tsuyuguchi, I. and Kishimoto, S. (1993) Differential regulation of formation of multinucleated giant cells from Concanavalin A stimulated human blood monocytes by IFN-y and IL-4. The Journal of Immunology 150, 3002-3010. Takayanagi, H., Kim, S., Matsuo, K., Suzuki, H., Suzuki, T., Sato, K., Yokochi, T., Odak, H., Nakamurak, K., Ida, N., Wagner, E. F and Taniguchi, T. (2002) RANKL maintains bone homeostasis through c-Fos-dependent induction of interferon-β. Nature 416, 744-749. Takayanagi, H. (2005) Mechanistic insight into osteoclast differentiation in osteoimmunology. Journal of Molecular Medicine 83, 170-179. Takeda, Y., Tachibana, I., Miyado, K., Kobayashi, M., Miyazaki, T., Funakoshi, T., Kimura, H., Yamane, H., Saito, Y., Goto, H., Yoneda, T., Yoshida, M., Kumagai, T., Osaki, T., Hayashi, S., Kawa, I. and Mekada, E. (2003) Tetraspanins CD9 and CD81 function to prevent the fusion of mononuclear phagocytes. The Journal of Cell Biology 161, 945–956. Takeda, Y., He, P., Tachibana, I., Zhou, B., Miyado, K., Kaneko, H., Suzuki, M., Minami, S., Iwasaki, T., Goya, S., Kijma, T., Kumagai, T., Yoshida, M., Osaki, T., Toshihisa Komori, T., Mekada, E. and Kawase, I. (2008) Double deficiency of tetraspanins CD9 and CD81 alters cells motility and protease production of macrophages and causes chronic obstructive pulmonary disease-like phenotype in mice. Journal of Biological Chemistry 183, 26089-26097. Takito, J., Nakamura, M., Yoda, M., Tohmonda, T., Uchikawa, S., Horiuchi, K., Toyama, Y. and Chiba, K. (2012) The transient appearance of zipper-like actin superstructures during the fusion of osteoclasts. Journal of Cell Science 125, 662-672. Tam, B. M., Moritz, O. L. and Papermaster, D. S. (2004) The C terminus of peripherin/rds participates in rod outer segment targeting and alignment of disk incisures. Molecular Biology of the Cell 15, 2027–2037. Tarrant, J. M., Robb, L., Van Spriel, A. B., Wright, M. D. (2003) Tetraspanins: molecular organisers of the leukocyte surface. Trends in Immunology 24, 610-617. Teitelbaum, S. L. (2006) Osteoclasts; culprits in inflammatory osteolysis. Arthritis Research and Therapy 8, 201-208. Thumbikat, P., Berry, R. E., Zhou, G., Billips, B. K., Yaggie, R. E., Zaichuk, T., Sun, T. T., Schaeffer, A. J. and Klumpp, D. J. (2009) Bacteria-induced uroplakin signalling mediates bladder response to infection. PLOS Pathogens 5, 1000415.

161

Todd, S. C., Doctor, V. S., and Levy, S. (1998) Sequences and expression of six new members of the tetraspanin/TM4SF family. Biochimica et Biophysica Acta 1399, 101–104. Troelstra, A., Antal-Szalmas, P., de Graaf-Miltenburg, L. A., Weersink, A. J., Verhoef, J., Van Kessel, K. P. and Van Strijp, J. A. (1997) Saturable CD14-dependent binding of fluorescein-labelled lipopolysaccharide to human monocytes. Infection and Immunity 65, 2272-2277. Tsan, M. and Gao, B. (2004) Cytokine function of heat shock proteins. American Journal of Physiology-Cellular Physiology 286, 739-744. Tsubery, H., Ofek, I., Cohen, S. and Fridkin, M. (2000) The functional association of polymyxin B with bacterial lipopolysaccharide is stereospecific: studies on polymyxin B nonapeptide. Biochemistry 39, 11837-1184. Tsuzuki, H., Tani, T., Ueyama, H. and Kodama, M. (2001) Lipopolysaccharide: Neutralization by polymyxin B shuts down the signalling pathway of Nuclear Factor κB in peripheral blood mononuclear cells, even during activation. Journal of Surgical Research 100, 127-134. Tu, L., Sun, T. T. and Kreibich, G. (2002) Specific heterodimer formation is a prerequisite for uroplakins to exit from the endoplasmic reticulum. Molecular Biology of the Cell 13, 4221–4230. Utaisincharoen, P., Arjcharoen, S., Limposuwan, K., Tungpradabkul, S. and Sirisinha, S. (2006) Burkholderia pseudomallei RpoS regulates multinucleated giant cell formation and inducible nitric oxide synthase expression in mouse macrophage cell line (RAW 264.7). Microbial Pathogenesis 40, 184–189. Väänänen, H. K. and Laitala-Leinonen, T. (2008) Osteoclast lineage and function. Archives of Biochemistry and Biophysics 473, 132-138. Van Soest, S., Westerveld, A., De Jong, P. T. V. M., Bleeker-Wagenmakers, E. M. and Bergen, A. A. B. (1999) Retinitis Pigmentosa: Defined from a molecular point of view. Survey of Opthalmology 43, 321-334. Van Spriel, A. B. and Figdor, C. G. (2010) The role of tetraspanins in the pathogenesis of infectious diseases. Microbes and Infection 12, 106-112. Vasina, J. A and Baneyx, F. (1997) Expression of aggregation-prone recombinant proteins at low temperatures: a comparative study of the Escherichia coli cspA and tac promoter systems. Protein Expression and Purification 9, 211–218. Vignery, A. (1989) Macrophage multinucleation is accompanied by the expression of new soluble and membrane antigens in mice. The American Journal of Pathology 135, 565–570. Vignery, A. (2005) Macrophage fusion: are somatic and cancer cells possible partners? Trends in Cell Biology 15, 188-193. Visintin, A., Mazzoni, A., Spitzer, J. H., Wyllie, D. H., Dower, S. K. and Segal, D. M. (2001) Regulation of Toll-like receptors in human monocytes and dendritic cells. The Journal of Immunology 166, 249–255. Vogel, S. N., Hilfiker, M. L. and Caulfield, M. J. (1983) Endotoxin-induced lymphocyte proliferation. The Journal of Immunology 130, 1774-1779. Wakelin, S. J., Sabroe, I., Gregory, C. D., Poxton, I. R., Forsythe, J. L. R, Garden, O. J. and Howie, S. E. M. (2006) Dirty little secrets-Endotoxin contamination of recombinant proteins Immunology Letters 106, 1–7. Wakelin, S. J., Forsythe, J. L. R., Garden, O. J. and Howie, S. E. M (2008) Commercially available recombinant sonic hedgehog up-regulates Ptc and modulates the cytokine and chemokine expression of human macrophages: An effect mediated by endotoxin contamination? Immunobiology 213, 25–38. Wang, H., Kolesnikova, T. V., Denison, C., Gygi, S. P. and Hemler, M. E. (2011) The C- terminal tail of tetraspanin protein CD9 contributes to its function and molecular organization. Journal of Cell Science 124, 2702-2710.

162

Waterhouse, R., Ha, C. and Dveksler, G. S. (2002) Murine CD9 is the receptor for pregnancy-specific glycoprotein 17. Journal of Experimental Medicine 195, 277– 282. Watson, R. W. G., Redmond, H. P. and Bouchier, Hayes, D. (1994) Role of endotoxin in mononuclear phagocyte-mediated inflammatory responses. Journal of Leukocyte Biology 56, 95-103. Weinberg, J., B., Hobbs, M. M. and Misukonis, M. A. (1984) Recombinant human γ- interferon induces human monocyte polykaryon formation. Proceedings of the National Academy of Science USA 81, 4554-4557. Weng, J., Krementsov, D. N., Khurana, S., Roy, N. H. and Thali, M. (2009) Formation of syncytia is repressed by tetraspanins in Human Immunodeficiency Virus type 1- producing cells. Journal of Virology 83, 7467–7474. Weyermann, J., Lochmanna, D. and Zimmer, A. (2005) A practical note on the use of cytotoxicity assays. International Journal of Pharmaceutics 288, 369-376. Wilson, N. F. and Snell, W. J. (1998) Microvilli and cell-cell fusion during fertilization. Trends in Cell Biology 8, 93-96. Wright, M. D., Henkle, K. J. and Mitchel, G. F. (1990) An immunogenic Mr 23,000 integral membrane protein of Schistosoma mansoni worms that closely resembles a human tumour-associated antigen. Journal of Immunology 144, 3195-3200. Wright, M. D., Moseley, G. W. and van Spriel, A. B. (2004) Tetraspanin microdomains in immune cell signalling and malignant disease. Tissue Antigens 64, 533–542. Wrigley, J. D., Ahmed, T., Nevett, C. L. and Findlay, J. B. (2000) Peripherin/rds influences membrane vesicle morphology: Implications for retinopathies. Journal of Biological Chemistry 275, 13191–13194. Xaus, J., Comalada, M., Valledor, A. F., Lloberas, J., Lopez-Soriano, F., Argiles, J. M., Bogdan, C. and Celada, A. (2000) LPS induces apoptosis in macrophages mostly through the autocrine production of TNF-α. Blood 95, 3823-3831. Xu, C., Zhang, Y. H., Thangavel, M., Richardson, M. M., Liu, L., Zhou, B., Zheng, Y., Ostrom, R. S. and Zhang, X. A. (2009) CD82 endocytosis and cholesterol-dependent reorganization of tetraspanin webs and lipid rafts. FASEB Journal 23, 3273-3288. Yanagishita, T., Watanabe, D., Akita, A., Nakano, A., Ohshima, Y., Tamada, Y. and Matsumoto, Y. (2007) Construction of novel in vitro epithelioid cell granuloma model from mouse macrophage cell line. Archives of Dermatological Research 299, 399-403. Yagi, M., Miyamoto, T., Sawatani, Y., Iwamoto, K., Hosogane, N., Fujita, N., Morita, K., Ninomiya, K., Suzuki, T., Miyamoto, K., Oike, Y., Takeya, M., Toyama, Y. and Suda, T. (2005) DC-STAMP is essential for cell–cell fusion in osteoclasts and foreign body giant cells. The Journal of Experimental Medicine 202, 345-351. Yagi, M., Ninomiya, K., Fujita, N., Suzuki, T., Iwasaki, R., Morita, K., Hosogane, N., Matsuo, K., Toyama, Y., Suda, T. and Miyamoto, T. (2007) Induction of DC-STAMP by alternative activation and downstream signalling mechanisms. Journal of Bone and Mineral Research 22, 992-1001. Yamada, A., Takami, M., Kawawa, T., Yasuhara, R., Zhao, B., Mochizuki, A., Miyamoto, Y., Eto, T., Yasuda, H., Nakamichi, Y., Kim, N., Katagiri, T., Suda, T. and Kamijo, R. (2007) Interleukin-4 inhibition of osteoclast differentiation is stronger than that of interleukin-13 and they are equivalent for induction of osteoprotegerin production from osteoblasts. Immunology 120, 573-579. Yanez-Mo, M., Barreiro, O., Gordon-Alonso, M., Sala-Valdes, M. and Sanchez-Madrid, F. (2009) Tetraspanin-enriched microdomains: a functional unit in cell plasma membranes. Trends in Cell Biology 19, 434-446. Yang, X., Claas, C., Kraeft, S., Chen, L. B., Wang, Z., Kreidberg, J. A. and Hemler, M. E. (2002) Palmitoylation of tetraspanin proteins: modulation of CD151 lateral interactions, subcellular distribution, and integrin-dependent cell morphology. Molecular Biology of the Cell 13, 767–781.

163 van der Rhee, H. J., Hillebrands, W. and Daems, W. T. (1978) Are Langhans giant cells precursors of foreign-body giant cells? Archives of Dermatology Research 263, 13- 21. Verrier, S., Hogan, A., McKie, N. and Horton, M. (2004) ADAM gene expression and regulation during human osteoclast formation. Bone 35, 34-46. Yagami-Hiromasa, T., Sato, T., Kurisaki, T., Kamijo, K., Nabeshima, Y. and Fujisawa-Sehara, A. (1995) A metalloprotease-disintegrin participating in myoblast fusion. Nature 377, 652–656. Yi, T., Kim, H., Cho, J., Woo, K., M., Ryoo, H., Kim, G. and Baek, J. (2006) CD9 regulates osteoclastogenesis via regulation of p44/42 MAPK activity. Biochemical and Biophysical Research Communications 347,178–184. Zeiler, T., Taivainen, A., Rytkonen, M., Rautiainen, J., Karjalainen, H., Mantyjarvi, R., Tuomisto, L. and Virtanen, T. (1997) Recombinant allergen fragments as candidate preparations for allergen immunotherapy. Journal of Allergy and Clinical Immunology 100, 721-727. Zeisel, M. B., Fofana, I., Fafi-Kremer, S. and Baumert, T. F. (2011) Hepatitis C virus entry into hepatocytes: molecular mechanisms and targets for antiviral therapies. Journal of Hepatology 54, 566-576. Zemni, R., Bienvenu, T., Vinet, M. C., Sefiani, A., Carrié, A., Billuart, P., McDonell, N., Couvert, P., Francis, F., Chafey, P., Fauchereau, F., Friocourt, G., des Portes, V., Cardona, A., Frints, S., Meindl, A., Brandau, O., Ronce, N., Moraine, C., van Bokhoven, H., Ropers, H. H., Sudbrak, R., Fryns, J. P., Beldjord, C. and Chelly, J. (2000) A new gene involved in X-linked mental retardation identified by analysis of an X2 balanced translocation. Nature Genetics 24, 167–170. Zhang, X. A., Kazarov, A. R., Yang, X., Bontrager, A. L., Stipp, C. S. and Hemler, M. E. (2002) Function of the tetraspanin CD151-α6β1 integrin complex during cellular morphogenesis. Molecular Biology of the Cell 13, 1–11. Zhang, P., Liu, J., Xu, Q., Harber, G., Feng, X., Michalek, S. M. and Katz, J. (2011) TLR2- dependent modulation of osteoclastogenesis by Porphyromonas gingivalis through differential induction of NFATc1 and NF-κB. The Journal of Biological Chemistry 286, 24159–24169. Zhou, G., Mo, W. J., Sebbel, P., Min, G., Neubert, T. A., Glockshuber, R., Wu, R., Sun, T. T. and Kong, X. P. (2001) Uroplakin Ia is the urothelial receptor for uropathogenic Escherichia coli: evidence from in vitro FimH binding. Journal of Cell Science 114, 4095–4103. Zhu, G., Miller, B. J., Boucheix, C., Rubinstein, E., Liu, C. C., Hynes, R. O., Myles, D. G. and Primakoff, P. (2002) Residues SFQ (173-175) in the large extracellular loop of CD9 are required for gamete fusion. Development 129, 1995-2002. Zhu, Y., Luo, Y., Cao, M., Liu, Y., Liu, X., Wang, W., Wu, D., Guan, M., Xu, Q., Ren, H., Zhao, P. and Qi, Z. (2012) Significance of palmitoylation of CD81 on its association with tetraspanin-enriched microdomains and mediating hepatitis C virus cell entry. Virology 429, 112–123. Zimmerman, T., Frère, C. P., Satzger, M., Raba, M., Weisbach, M., Döhn, K., Popp, A. and Donzeau, M. (2006) Simultaneous metal chelate affinity purification and endotoxin clearance of recombinant antibody fragments. Journal of Immunological Methods 314, 67-73. Ziyyat, A., Rubinstein, E., Monier-Gavelle, F., Barraud, V., Kulski, O., Prenant, M., Boucheix, C., Bomsel, M. and Wolf, J. (2006) CD9 controls the formation of clusters that contain tetraspanins and the integrin α6β1, which are involved in human and mouse gamete fusion. Journal of Cell Science 119, 416-424. Zöller, M. (2009) Tetraspanins: push and pull in supressing and promoting metastatis. Nature Reviews Cancer 9, 40-55. Zou, W. and Bar-Shavit, Z. (2002) Dual modulation of osteoclast differentiation by lipopolysaccharide. Journal of Bone and Mineral Research 17, 1211-1218.

164

Appendix 1: Multiple sequence alignment of 33 human tetraspanins

Full-length protein sequences were obtained in FASTA format from the bioinformatics portal Uniprot available at www.uniprot.org. In cases where multiple isoforms were reported, the dominant isoform as annotated in the database was chosen for the alignment. Sequences were aligned using the multiple alignment programme T-COFFEE available at www.tcoffee.org. Minor differences were observed using the Clustal Omega multiple alignment software available at http://www.ebi.ac.uk/Tools/msa/clustalo/. The Clustal Omega alignment proved to be more accurate for prediction of EC2 domain boundaries (Chapter 2, section 2.2.2.1), however the T-COFFEE alignment is displayed here for the visual effectiveness of the output. For simplicity, tetraspanin sequences are labelled with the Tspan number. Other common tetraspanin names are outlined in the table below.

Tetraspanin Alternative names Tetraspanin Alternative names Tetraspanin Alternative names Tspan1 NET-1, TM4-C Tspan12 NET-2 Tspan23 ROM-1 Tspan2 NET-3 Tspan13 NET-6 Tspan24 CD151 Tspan3 TM4-A Tspan14 DC-TM4F2 Tspan25 CD53 Tspan4 NAG-2, TM4SF7 Tspan15 NET-7 Tspan26 CD37 Tspan5 NET-4, TM4SF9 Tspan16 TM4-B Tspan27 CD82 Tspan6 TM4D Tspan17 FBOXO23 Tspan28 CD81 Tspan7 CD231, TALLA-1, A15 Tspan18 Tspan29 CD9 Tspan8 CO-029, TM4SF3 Tspan19 Tspan30 CD63 Tspan9 NET-5 Tspan20 UPIb, UPKb Tspan31 SAS Tspan10 Oculospanin Tspan21 UPIa, UPKa Tspan32 Tscc6, Protein PHEMX Tspan11 Tspan22 RDS, peripherin-2 Tspan33 Penumbera

165

Figure key

BAD AVG GOOD = Strength of residue alignment with a derived primary consensus sequence (*)100% conserved, (:)Highly conserved, (.)Conserved, (TM1-4)Transmembrane 1 to 4, (EC1, EC2)Extracellular 1 & 2, (IC)Intracellular Tspan1 MQC------FSFIKTMMILFNLLIFLCGAALLAVGIWVSIDGASF Tspan2 MGRF------RGGLRCIKYLLLGFNLLFWLAGSAVIAFGLWFRFGG-AI Tspan3 MGQC------GITSSKTVLVFLNLIFWGAAGILCYVGAYVFITYDDY Tspan4 MAR------ACLQAVKYLMFAFNLLFWLGGCGVLGVGIWLAATQGSF Tspan5 MSGKHY------K--GPEVSCCIKYFIFGFNVIFWFLGITFLGIGLWAWNEKGVL Tspan6 MASPSRRL------Q--TKPVITCFKSVLLIYTFIFWITGVILLAVGIWGKVSLENY Tspan7 MASRRM------E--TKPVITCLKTLLIIYSFVFWITGVILLAVGVWGKLTLGTY Tspan8 MAG------VSACIKYSMFTFNFLFWLCGILILALAIWVRVSNDSQ Tspan9 MAR------GCLCCLKYMMFLFNLIFWLCGCGLLGVGIWLSVSQGNF Tspan10 MEEGERSPLLSQETAGQKPLSVHRPPTSGCLGPVPREDQAEAWGCSCCPPETKHQALSGTPKKGPAPS--LSPGSSCVKYLIFLSNFPFSLLGLLALAIGLWGLAVKGSL Tspan11 MAHYKTE------Q--DDWLIIYLKYLLFVFNFFFWVGGAAVLAVGIWTLVEKSGY Tspan12 MARE------DSVKCLRCLLYALNLLFWLMSISVLAVSAWMRDYLNNV Tspan13 MVCG------GFACSKNCLCALNLLYTLVSLLLIGIAAWGIGFG--- Tspan14 MHYYRY------S--NAKVSCWYKYLLFSYNIIFWLAGVVFLGVGLWAWSEKGVL Tspan15 MPRGDSEQV------RYCARFSYLWLKFSLIIYSTVFWLIGALVLSVGIYAEVERQKY Tspan16 MAEIH------TPYSSLKKLLSLLNGFVAVSGIILVGLGIGGKCGGASL Tspan17 MPGKHQHF------Q--EPEVGCCGKYFLFGFNIVFWVLGALFLAIGLWAWGEKGVL Tspan18 MEGD------CLSCMKYLMFVFNFFIFLGGACLLAIGIWVMVDPTGF Tspan19 MLRN------NKTIIIKYFLNLINGAFLVLGLLFMGFGAWLLLDRNNF Tspan20 MAKDNS------TVRCFQGLLIFGNVIIGCCGIALTAECIFFVSDQHSL Tspan21 MASAAAA------E--AEKGSPVVVGLLVVGNIIILLSGLSLFAETIWVTADQYRV Tspan22 MALLKVKF------DQKKRVK-LAQGLWLMNWFSVLAGIIIFSLGLFLKIELRKR Tspan23 MAPVLPLVL------PLQPRIR-LAQGLWLLSWLLALAGGVILLCSGHLLVQLRHL Tspan24 MGEFNEK------K--TTCGTVCLKYLLFTYNCCFWLAGLAVMAVGIWTLALKSDY Tspan25 MGM------SSLKLLKYVLFFFNLLFWICGCCILGFGIYLLIHN-NF Tspan26 MSAQE------SCLSLIKYFLFVFNLFFFVLGSLIFCFGIWILIDKTSF Tspan27 MGS------ACIKVTKYFLFLFNLIFFILGAVILGFGVWILADKSSF Tspan28 MGV------EGCTKCIKYLLFVFNFVFWLAGGVILGVALWLRHDPQTT Tspan29 MPV------KGGTKCIKYLLFGFNFIFWLAGIAVLAIGLWLRFDSQTK Tspan30 MAVE------GGMKCVKFLLYVLLLAFCACAVGLIAVGVGAQLVLSQ- Tspan31 MVCG------GFACSKNALCALNVVYMLVSLLLIGVAAWGKGLG--- Tspan32 MGPWSRV------R------VAKCQMLVTCFFILLLGLSVATMVTL-TYFGAHF Tspan33 MARRPRAPAAS------GEE--FSFVSPLVKYLLFFFNMLFWVISMVMVAVGVYARLMKHAE * : .

N-terminus (IC) TM1 EC1

166

Tspan1 LKIFG-PL-SSS----AMQFVN---VGYFLIAAGVVVFALGFL-G--CYGAKTES-----KCALV-TFFFILLLIFIAEV------AAA-VVALVYTTMAEHF----LT Tspan2 KELSSE---DKS----PEYFYV---GLYVLVGAGALMMAVGFF-G--CCGAMRES-----QCVLG-SFFTCLLVIFAAEV------TTG-VFAFIGKGVAIRH----VQ Tspan3 DHFFED-----V------YTLI---PAVVIIAVGALLFIIGLI-G--CCATIRES-----RCGLA-TFVIILLLVFVTEV------VVV-VLGYVYRAKVENE----VD Tspan4 ATLSS------S----FPSL-S---AANLLIITGAFVMAIGFV-G--CLGAIKEN-----KCLLL-TFFLLLLLVFLLEA------TIA-ILFFAYTDKIDRY----AQ Tspan5 SNISSI-----T----DLGGFD---PVWLFLVVGGVMFILGFA-G--CIGALREN-----TFLLK-FFSVFLGIIFFLEL------TAG-VLAFVFKDWIKDQ----LY Tspan6 FSLLNE------K-ATN---VPFVLIATGTVIILLGTF-G--CFATCRAS-----AWMLK-LYAMFLTLVFLVEL------VAA-IVGFVFRHEIKNS----FK Tspan7 ISLIAE------N-STN---APYVLIGTGTTIVVFGLF-G--CFATCRGS-----PWMLK-LYAMFLSLVFLAEL------VAG-ISGFVFRHEIKDT----FL Tspan8 AIFGSE---DVG-----SSSYV---AVDILIAVGAIIMILGFL-G--CCGAIKES-----RCMLL-LFFIGLLLILLLQV------ATG-ILGAVFKSKSDRI----VN Tspan9 ATFSP------S----FPSL-S---AANLVIAIGTIVMVTGFL-G--CLGAIKEN-----KCLLL-SFFIVLLVILLAEL------ILL-ILFFVYMDKVNEN----AK Tspan10 GSD---LG---G-----PLPAD---PMLGLALGGLVVSAVSLA-G--YLGALCEN-----TCLLR-GFSGGILAFLVLEA------VAG-ALVVALWGPLQDS----LE Tspan11 LSV---LA---S-----STFAA---SAYILIFAGVLVMVTGFL-G--FGAILWER-----KGCLS-TYFCLLLVIFLVEL------VAG-VLAHVYYQRLSDE----LK Tspan12 LTLTAE-T-RVEEAVILTYFPV---VHPVMIAVCCFLIIVGML-G--YCGTVKRN-----LLLLA-WYFGSLLVIFCVEL------ACG-VWTYEQELMVPVQ----WS Tspan13 ------LI-S------SL---R---VVGVVIAVGIFLFLIALV-G--LIGAVKHH-----QVLLF-FYMIILLLVFIVQF------SVS-CACLALNQEQQGQ----LL Tspan14 SDLTKV-----T----RMHGID---PVVLVLMVGVVMFTLGFA-G--CVGALREN-----ICLLN-FFCGTIVLIFFLEL------AVA-VLAFLFQDWVRDR----FR Tspan15 KTLE------S-AFLA---PAIILILLGVVMFMVSFI-G--VLASLRDN-----LYLLQ-AFMYILGICLIMEL------IGG-VVALTFRNQTIDF----LN Tspan16 TNVLGL-S-S------AYLLH---VGNLCLVMGCITVLLGCA-G--WYGATKES-----RGTLL-FCILSMVIVLIMEV------TAA-TVVLLFFPIVGDVALEHTF Tspan17 SNISAL-----T----DLGGLD---PVWLFVVVGGVMSVLGFA-G--CIGALREN-----TFLLK-FFSVFLGLIFFLEL------ATG-ILAFVFKDWIRDQ----LN Tspan18 REIVAA-----N-----PLLLT---GAYILLAMGGLLFLLGFL-G--CCGAVREN-----KCLLL-FFFLFILIIFLAEL------SAA-ILAFIFRENLTRE----FF Tspan19 LTAFDE-N-N------HFIVP---ISQILIGMGSSTVLFCLL-G--YIGIHNEI-----RWLLI-VYAVLITWTFAVQV------VLS-AFIITKKEEVQQL----WH Tspan20 YPLLE-AT-DND------DIYG---AAWIGIFVGICLFCLSVL-G--IVGIMKSS-----RKILL-AYFILMFIVYAFEV------ASC-ITAATQQDFFTPN------Tspan21 YPLMG-VS-GKD------DVFA---GAWIAIFCGFSFFMVASF-G--VGAALCRR-----RSMVL-TYLVLMLIVYIFEC------ASC-ITSYTHRDYMVSN----PS Tspan22 SDVMNN-----S----E-S-HF---VPNSLIGMGVLSCVFNSLAGKICYDALDPAKYARWKPWLK-PYLAICVLFNIILF------LVA-LCCFLLRGSLENT----LG Tspan23 GTFLAP-----S----C-QFPV---LPQAALAAGAVALGTGLV-G--VGA-SRASL----NAALYPPWRGVLGPLLVAGTAGGGGLLVVGLGLALALPGSLDEA----LE Tspan24 ISL---LA---S-----GTYLA---TAYILVVAGTVVMVTGVL-G--CCATFKER-----RNLLR-LYFILLLIIFLLEI------IAG-ILAYAYYQQLNTE----LK Tspan25 GVLFH------N----LPSL-T---LGNVFVIVGSIIMVVAFL-G--CMGSIKEN-----KCLLM-SFFILLLIILLAEV------TLA-ILLFVYEQKLNEY----VA Tspan26 VSFVGL-----A----FVPLQI---WSKVLAISGIFTMGIALL-G--CVGALKEL-----RCLLG-LYFGMLLLLFATQI------TLG-ILISTQRAQLERS----LR Tspan27 ISVLQT-----S----SSSLRM---GAYVFIGVGAVTMLMGFL-G--CIGAVNEV-----RCLLG-LYFAFLLLILIAQV------TAG-ALFYFNMGKLKQE----MG Tspan28 NLLYLELGDKPA----PNTFYV---GIYILIAVGAVMMFVGFL-G--CYGAIQES-----QCLLG-TFFTCLVILFACEV------AAG-IWGFVNKDQIAKD----VK Tspan29 SIFEQE-T-NNN----NSSFYT---GVYILIGAGALMMLVGFL-G--CCGAVQES-----QCMLG-LFFGFLLVIFAIEI------AAA-IWGYSHKDEVIKE----VQ Tspan30 TIIQGA-----T-----PGSLL----PVVIIAVGVFLFLVAFV-G--CCGACKEN-----YCLMI-TFAIFLSLIMLVEV------AAA-IAGYVFRDKVMSE----FN Tspan31 ------LV-S------SI---H---IIGGVIAVGVFLLLIAVA-G--LVGAVNHH-----QVLLF-FYMIILGLVFIFQF------VIS-CSCLAINRSKQTD----VI Tspan32 AVIRRASL-EKN----PYQAVHQWAFSAGLSLVG-LLTLGAVL-S--AAATVREA-----QGLMA-GGFLCFSLAFCAQV------QVV-FWRLHSPTQVEDA----ML Tspan33 AALAC------LAVD---PAILLIVVGVLMFLLTFC-G--CIGSLREN-----ICLLQ-TFSLCLTAVFLLQL------AAG-ILGFVFSDKARGK----VS : :

EC1 TM2 IC TM3 EC2

167

Tspan1 LLV--VPAIKKDYGSQ------EDFTQVWNTTMKGLKCCGFT-NYTDFEDSPYF-KEN------SAFPPFC-CNDNVT-N-T--A--N Tspan2 TMY--EEAYN-DYLKDR------GKG--NGTLITFHSTFQCCGKE-SSEQVQP------Tspan3 RSI--QKVYK-TYNGTN------PDAASRAIDYVQRQLHCCGIH-NYSDWENTDWF-KE------TKNQSVPLSC-CRETAS-NCNG----- Tspan4 QDL--KKGLH-LYGTQG------NVGLTNAWSIIQTDFRCCGVS-NYTDWFEVYN------ATRVPDSC-CLEFSE-S------Tspan5 FFI--NNNIR-AYRDD------IDLQNLIDFTQEYWQCCGAF-GADDWNLNIYF-NCTDSNA------SRERCGVPFSC-CTKDPA-ED---VI-N Tspan6 NNY--EKALK-QYNSTG------D-YRSHAVDKIQNTLHCCGVT-DYRDWTDTNYY-S------EKGFPKSC-CKLED---CTP--Q-- Tspan7 RTY--TDAMQ-TYNGND------E--RSRAVDHVQRSLSCCGVQ-NYTNWSTSPYF-L------EHGIPPSC-CMNET--DCNP--Q-- Tspan8 ETL--YENTK-LLSATG------ESEKQFQEAIIVFQEEFKCCGLVNGAADWGNNF------QHYPELCACLDKQR-PCQ------Tspan9 KDL--KEGLL-LYHTEN------NVGLKNAWNIIQAEMRCCGVT-DYTDWYPVLG------ENTVPDRC-CMENSQ-G------Tspan10 HTL--RVAIA-HYQD------D-PDLRFLLDQVQLGLRCCGAA-SYQDWQQNLYF-NCSSP------GVQACSLPASC-CIDPRE-DG--ASV-N Tspan11 QHL--NRTLAENYGQPG------A-TQITASVDRLQQDFKCCGSN-SSADWQHSTYI-LLRE------AEGRQVPDSC-CKTVVV------Tspan12 DMVTLKARMT-NYGL------PRYRWLTHAWNFFQREFKCCGVV-YFTDWLEMTEM------DWPPDSC-CVREFP-GCSK----- Tspan13 EVG------WNNTASARNDIQRNLNCCGFR-SVNPN-----DT-C------LAS-CVKSD------Tspan14 EFF--ESNIK-SYRDD------IDLQNLIDSLQKANQCCGAY-GPEDWDLNVYF-NCSGASY------SREKCGVPFSC-CVPDPA-QK---VV-N Tspan15 DNI--RRGIE-NYYD------DLDFKNIMDFVQKKFKCCGGE-DYRDWSKNQYH-DCSAPG------PLACGVPYTC-CIRNTT-E----VV-N Tspan16 VTL--RKNYR-GYNEPD------D--YSTQWNLVMEKLKCCGVN-NYTDFSGSSFE-MT------TGHTYPRSC-CKSIGSVSCDG--R-- Tspan17 LFI--NNNVK-AYRDD------IDLQNLIDFAQEYWSCCGAR-GPNDWNLNIYF-NCTDLNP------SRERCGVPFSC-CVRDPA-ED---VL-N Tspan18 --T--KELTK-HYQGNN------DTDVFSATWNSVMITFGCCGVN-GPEDFKFASVF-RLLT------LDSEEVPEAC-CRREPQ-SRDGVLLSR Tspan19 DKI--DFVIS-EYGSKDK------PEDITKWTILNALQKTLQCCGQH-NYTDWIKNKNK-EN------SGQVPCSC-TKSTLR-K------Tspan20 LFL--KQMLE-RYQNNSPPNNDDQWKNNGVTKTWDRLMLQDNCCGVN-GPSDWQKYTSAFRTEN------NDADYPWPRQC-CVMNNL-KEP--LN-L Tspan21 LIT--KQMLT-FYSADTD------QGQELTRLWDRVMIEQECCGTS-GPMDWVNFTSAFRAAT------PEVVFPWPPLC-CRRTGN-FIP--LN-E Tspan22 QGL--KNGMK-YYRDTDT------PGRCFMKKTIDMLQIEFKCCGNN-GFRDWFEIQWI-SNRYLDFSSKEVKDRIKSNVDGRYLVDGVPFSC-CNPSSP-RP---CI-Q Tspan23 EGL--VTALA-HYKDTEV------PGHCQAKRLVDELQLRYHCCGRH-GYKDWFGVQWV-SSRYLDPGDRDVADRIQSNVEGLYLTDGVPFSC-CNPHSP-RP---CL-Q Tspan24 ENL--KDTMTKRYHQPG------H-EAVTSAVDQLQQEFHCCGSN-NSQDWRDSEWI-RSQE------AGGRVVPDSC-CKTVVA------Tspan25 KGL--TDSIH-RYHSD------NSTKAAWDSIQSFLQCCGIN-GTSDWTSG------PPASC-P------Tspan26 DVV--EKTIQ-KYGTNP------EETAAEESWDYVQFQLRCCGWH-YPQDWFQVLIL-RGNG------SEAHRVPCSC-YNLSAT-ND---ST-I Tspan27 GIV--TELIR-DYNSSR------EDSLQDAWDYVQAQVKCCGWV-SFYNWTDNAEL-MNR------PEVTYPCSC-EVKGEE-DN---SL-S Tspan28 QFY--DQALQ-QAVVDD------DANNAKAVVKTFHETLDCCGSS-TLTALTTSVLK-N------N------Tspan29 EFY--KDTYN-KLKTKD------E--PQRETLKAIHYALNCCGLA-GGVEQFIS------D------Tspan30 NNF--RQQME-NY------PKNNHTASILDRMQADFKCCGAA-NYTDWEKIPSM------SKNRVPDSC-CINVTV-GCGI----- Tspan31 NAS-----W------WVMSNKTRDELERSFDCCGLF-NLTTLYQQDYDF-C------TAI-CKSQS------Tspan32 DTY--DLVYEQAMKGT-S------HVRRQELAAIQDVFLCCGKK-SPFSRLGSTEAD------L-CQGE------Tspan33 EII--NNAIV-HYRDD------LDLQNLIDFGQKKFSCCGGI-SYKDWSQNMYF-NCSEDNP------SRERCSVPYSC-CLPTPD-QA---VI-N ***

EC2

168

Tspan1 ET------C------TKQKAHDQK--VEGCFNQLLYDIRT--NAVTVGGV-AAGIGGLELAAMIVSMYLYCN------Tspan2 -T------CP------KELLG--HKNCIDEIETIISV--KLQLIGIV-GIGIAGLTIFGMIFSMVLCCA-IR-----NS------Tspan3 ------SLAHPSDLY--AEGCEALVVKKLQE--IMMHVIWA-ALAFAAIQLLGMLCACIVLCRRSR-----DPAYEL------Tspan4 ------CGLHAPGTWW--KAPCYETVKVWLQE--NLLAVGIF-GLCTALVQILGLTFAMTMYCQ-VV-----K--A------Tspan5 TQ------CGYD--ARQ---KPEVDQQIVIY--TKGCVPQFEKWLQD--NLTIVAGI-FIGIALLQIFGICLAQNLVSD-IE-----AV------Tspan6 ------RDADKVN--NEGCFIKVMTIIES--EMGVVAGI-SFGVACFQLIGIFLAYCLSRA-IT-----N------Tspan7 DL------H------NLTVAATKVN--QKGCYDLVTSFMET--NMGIIAGV-AFGIAFSQLIGMLLACCLSRF-IT-----A------Tspan8 ------SYNGKQVY--KETCISFIKDFLAK--NLIIVIGI-SFGLAVIEILGLVFSMVLYCQ-IG------Tspan9 ------CGRNATTPLW--RTGCYEKVKMWFDD--NKHVLGTV-GMCILIMQILGMAFSMTLFQH-IH-----R--T------Tspan10 DQ------CGFG--VLR---LDADAAQRVVY--LEGCGPPLRRWLRA--NLAASGGY-AIAVVLLQGAELLLAARLLGA-LA-----ARRGAA------Tspan11 -R------CGQ------RAHPSNIYKVEGGCLTKLEQFLAD--HLLLMGAV-GIGVACLQICGMVLTCCLHQR-LQ------Tspan12 QA------H------QEDLSDLY--QEGCGKKMYSFLRGTKQLQVLRFL-GISIGVTQILAMILTITLLWA-LY-----YDRREP------Tspan13 ------HS--CSPCAPIIGEYAGE--VLRFVGGI-GLFFSFTEILGVWLTYRYRNQ-K------DPR------Tspan14 TQ------CGYD--VRI---QLKSKWDESIF--TKGCIQALESWLPR--NIYIVAGV-FIAISLLQIFGIFLARTLISD-IE-----AVKA------Tspan15 TM------CGYK--TI---DKERFSVQDVIY--VRGCTNAVIIWFMD--NYTIMAGI-LLGILLPQFLGVLLTLLYITR-VE-----DIIMEH------Tspan16 ------DVSPNVIH--QKGCFHKLLKITKT--QSFTLSGS-SLGAAVIQRWGSRYVAQ------A------Tspan17 TQ------CGYD--VRLKLVRGELEQQGFIH--TKGCVGQFEKWLQD--NLIVVAGV-FMGIALLQIFGICLAQNLVSD-IK-----AVKANW------Tspan18 EE------C------LLGRSLFLN--KQGCYTVILNTFET--YVYLAGAL-AIGVLAIELFAMIFAMCLFRG------Tspan19 WF------C------DEPLNATY--LEGCENKISAWYNV--NVLTLIGI-NFGLLTSEVFQVSLTVCFFKN-IK-----N------Tspan20 EA------C------KLGVPGFYH--NQGCYELISGPMNR--HAWGVAWF-GFAILCWTFWVLLGTMFYWSR------Tspan21 EG------C------RLGHMDYLF--TKGCFEHIGHAIDS--YTWGISWF-GFAILMWTLPVMLIAMYFY------Tspan22 YQ------ITNN--SAH-YSYDHQTEELNLW--VRGCRAALLSYYSS--LMNSMGVV-TLLIWLFEVTITIGLRYLQTS-LD-----GVSNPE------Tspan23 NR------LSDS--YAH-PLFDPRQPNQNLW--AQGCHEVLLEHLQD--LAGTLGSM-LAVTFLLQALVLLGLRYLQTA-LE-----GLGGVI------Tspan24 -L------CGQ------RDHASNIYKVEGGCITKLETFIQE--HLRVIGAV-GIGIACVQVFGMIFTCCLYRS-LK------Tspan25 ------SDRK--VEGCYAKARLWFHS--NFLYIGII-TICVCVIEVLGMSFALTLNCQ-ID-----K--T------Tspan26 LDKVILPQLSRLGHLARSRHSADICAVPAESHIY--REGCAQGLQKWLHN--NLISIVGI-CLGVGLLELGFMTLSIFLCRN-LD-----HVYN------Tspan27 VR------KGFCEAPGNRTQSGNHPEDWPVY--QEGCMEKVQAWLQE--NLGIILGV-GVGVAIIELLGMVLSICLCRH-VH-----SEDY------Tspan28 -L------CPS------GSNIISNLF--KEDCHQKIDDLFSG--KLYLIGIA-AIVVAVIMIFEMILSMVLCCG-IR-----N------Tspan29 -I------CPK------K-DVLETFT--VKSCPDAIKEVFDN--KFHIIGAV-GIGIAVVMIFGMIFSMILCCA-IR-----RN------Tspan30 ------NFNEKAIH--KEGCVEKIGGWLRK--NVLVVAAA-ALGIAFVEVLGIVFACCLVKS-IR-----S------Tspan31 ------PT--CQMCGEKFLKHSDE--ALKILGGV-GLFFSFTEILGVWLAMRFRNQ-K------DPR------Tspan32 ------EAA--REDCLQGIRSFLRT--HQQVASSLTSIGLA-LTVSALLFSSFLWFA-IRCGCSLDRKGKYTLTPRACGRQPQEP Tspan33 TM------CGQG--MQA---FDYLEASKVIY--TNGCIDKLVNWIHS--NLFLLGGV-ALGLAIPQLVGILLSQILVNQ-IK-----DQIKLQ------*

EC2 TM4 C-terminus (IC)

169

Tspan1 ------Tspan2 ------Tspan3 ------Tspan4 ------Tspan5 ------Tspan6 ------Tspan7 ------Tspan8 ------Tspan9 ------Tspan10 ------YGPGARG------Tspan11 ------Tspan12 ------GTDQMMSLKNDNSQHLSC------Tspan13 ------Tspan14 ------Tspan15 ------S------Tspan16 ------Tspan17 ------S-K-WNDDFENHWLTPTISEVLSTAGPQQNSLTGAPGP Tspan18 ------Tspan19 ------Tspan20 ------Tspan21 ------Tspan22 ------E-S--ES------Tspan23 ------D-A--GG------Tspan24 ------Tspan25 ------Tspan26 ------Tspan27 ------Tspan28 ------Tspan29 ------Tspan30 ------Tspan31 ------Tspan32 SLLRCSQGGPTHCLHSEAVAIGPRGCSGSLRWLQESDAAPLPLSCHLAAHRALQGR------Tspan33 ------L------

C-terminus (IC)

170

Tspan1 ------LQ Tspan2 ------RDVI Tspan3 ------LITGGTYA Tspan4 ------DTYCA Tspan5 ------RASW Tspan6 ------NQYEIV Tspan7 ------NQYEMV Tspan8 ------NK Tspan9 ------GKKYDA Tspan10 ------EDRA------GPQSPSP-----G------AP---PAAKPARG Tspan11 ------RHFY Tspan12 ------PSVE-----LLKPSLSRIFEHTSMANSFN---THFEMEEL Tspan13 ------ANPSAFL Tspan14 ------GHHF Tspan15 ------V--TDGLL------GPGAKPSVE-----A------AGT--GCCLCYPN Tspan16 ------GLELLA Tspan17 APPSRHVFFGLGGLY------PEPTFKNW Tspan18 ------IQ Tspan19 ------IIHAEM Tspan20 ------IEY Tspan21 ------TML Tspan22 ------E--SQGWLLERSVPETWK-AFLESVKKLGKGNQVEAE-----G------AD---AGQAPEAG Tspan23 ------E--TQGYLFPSGLKDMLKTAWLQ----GGVACRPAPEEAPPGE------AP---PKEDLSEA Tspan24 ------LEHY Tspan25 ------SQTIGL Tspan26 ------RLARYR Tspan27 ------SKVPKY Tspan28 ------SSVY Tspan29 ------REMV Tspan30 ------GYEVM Tspan31 ------ANPSAFL Tspan32 ------SR-----G------GLSGCPERGLSD Tspan33 ------Y------N------Q---QHRADPWY

171