Proquest Dissertations

THE ROLE OF CASP AND SORTING NEXIN 27 IN IMMUNITY

Adam J. MacNeil

Submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy

Dalhousie University Halifax, Nova Scotia December 2008

395 Wellington Street 395, rue Wellington Ottawa ON K1A0N4 Ottawa ON K1A0N4 Canada Canada

Your file Votre reference ISBN: 978-0-494-50156-6 Our file Notre reference ISBN: 978-0-494-50156-6

NOTICE: AVIS: The author has granted a non L'auteur a accorde une licence non exclusive exclusive license allowing Library permettant a la Bibliotheque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par Plntemet, prefer, telecommunication or on the Internet, distribuer et vendre des theses partout dans loan, distribute and sell theses le monde, a des fins commerciales ou autres, worldwide, for commercial or non sur support microforme, papier, electronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in et des droits moraux qui protege cette these. this thesis. Neither the thesis Ni la these ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent etre imprimes ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.

In compliance with the Canadian Conformement a la loi canadienne Privacy Act some supporting sur la protection de la vie privee, forms may have been removed quelques formulaires secondaires from this thesis. ont ete enleves de cette these.

While these forms may be included Bien que ces formulaires in the document page count, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. Canada DALHOUSIE UNIVERSITY

To comply with the Canadian Privacy Act the National Library of Canada has requested that the following pages be removed from this copy of the thesis:

Preliminary Pages Examiners Signature Page (pii) Dalhousie Library Copyright Agreement (piii)

Appendices Copyright Releases (if applicable) DEDICATION

This thesis is dedicated to my loving mother and father, Charlene and Harold MacNeil.

iv TABLE OF CONTENTS

LIST OF TABLES x

LIST OF FIGURES xi

ABSTRACT xiii

LIST OF ABBREVIATIONS xiv

ACKNOWLEDGMENTS xvii

CHAPTER 1 INTRODUCTION 1

CHAPTER 2 SORTING NEXIN 27 INTERACTS WITH THE CYTOHESIN-ASSOCIATED

SCAFFOLDING PROTEIN (CASP) IN LYMPHOCYTES 6

2.1 ABSTRACT 7

2.2 INTRODUCTION 7

2.3 MATERIALS AND METHODS 9

2.3.1 CELLS AND LYSATES 9

2.3.2 DNA AND RECOMBINANT PROTEINS 10 2.3.3 GST PULLDOWNS, WESTERN ANALYSIS AND ESI-TANDEM MASS SPECTROMETRY 11

2.3.4 CO-IMMUNOPRECIPITATION 11

2.3.5 IMMUNOCYTOCHEMISTRY AND TRANSFECTIONS 12

2.4 RESULTS 13

2.4.1 SNX27 INTERACTS WITH THE CARBOXY-TERMINAL PDZ-BINDING MOTIF OF CASP IN LYMPHOCYTES 13

2.4.2 PHYSICAL INTERACTION WITH CASP is MEDIATED BY THE PDZ DOMAIN OF SNX27 14

2.4.3 SNX27 CO-IMMUNOPRECIPITATES WITH CASP AND CYTOHESIN-1 IN LYMPHOCYTES 14

v 2.4.4 ENDOGENOUS SNX27 CO-LOCALIZES WITH CASP AT EARLY ENDOSOMES IN YT CELLS AND RECRUITS TRANSFECTED CASP TO ENDOSOMES IN HEK-293 CELLS 15

2.5 DISCUSSION 22

2.6 ACKNOWLEDGEMENT 24

2.7 REFERENCES 25

2.8 LINKING PARAGRAPH 28

CHAPTER 3 POLARIZATION OF ENDOSOMAL SNX27 IN MIGRATING AND

TUMOR-ENGAGED NATURAL KILLER CELLS 29

3.1 ABSTRACT 30

3.2 INTRODUCTION 30

3.3 MATERIALS AND METHODS 32

3.3.1 CELLS AND ANTIBODIES 32

3.3.2 IMMUNOCYTOCHEMISTRY 33

3.3.3 CONJUGATION ASSAYS 33

3.4 RESULTS 33 3.4.1 ENDOGENOUS SNX27 LOCALIZES TO THE EARLY ENDOSOMES IN NK-92 CELLS 33

3.4.2 SNX27 POLARIZES TO THE IMMUNOLOGICAL SYNAPSE IN VITRO AND IS DISTINCT FROM THE CYTOTOXIC PERFORIN- CONTAINING GRANULES 34

3.4.3 SNX27 POLARIZES TO THE LEADING EDGE OF MIGRATING NK-92 CELLS 35

3.5 DISCUSSION 41

3.6 ACKNOWLEDGEMENT.. 44

3.7 REFERENCES 45

3.8 LINKING PARAGRAPH 48

vi CHAPTER 4 GENE DUPLICATION IN EARLY VERTEBRATES RESULTS IN TISSUE- SPECIFIC SUBFUNCTIONALIZED ADAPTOR PROTEINS: CASP AND GRASP 49

4.1 ABSTRACT 50

4.2 INTRODUCTION 51

4.3 MATERIALS AND METHODS 53

4.3.1 IDENTIFICATION AND RETRIEVAL OF CASP AND GRASP HOMOLOGUES 54

4.3.2 ANALYSIS OF CASP AND GRASP GENE STRUCTURES AND PROTEIN ALIGNMENTS 54

4.3.3 ALIGNMENTS AND PHYLOGENETIC ANALYSIS OF GRASP AND CASP HOMOLOGUES 55

4.3.4 ANALYSIS OF AMPHIOXUS, ZEBRAFISH, AND HUMAN GENOMIC LOCI 56

4.4 RESULTS 57

4.4.1 CASP AND GRASP SHARE HIGHLY SIMILAR FUNCTIONAL DOMAINS, MOTIFS, AND GENOMIC STRUCTURE ..57

4.4.2 MAMMAL, BIRD, AND REPTILE CASP SHARE A UNIQUE AND RECENTLY ACQUIRED AMINO ACID IN THE COILED-COIL MOTIF 58

4.4.3 THE PDZBM OF VERTEBRATE CASPS AND GRASPS SHOW CONSIDERABLE CONSERVATION 59

4.4.4 PHYLOGENETIC ANALYSIS IDENTIFIES A CASP/GRASP GENE DUPLICATION PREVIOUS TO THE VERTEBRATE RADIATION 61

4.4.5 EXAMINATION OF GENETIC LOCI IN VARIOUS CHORDATES SHOWS THAT GRASP is THE MOST LIKELY ANCESTRAL GENE AND THAT THE ANCESTRAL NR4A GENE WAS CO-DUPLICATED 62

4.5 DISCUSSION 74

vii 4.6 ACKNOWLEDGEMENTS 79

4.7 REFERENCES 80

4.8 LINKING PARAGRAPH 84

CHAPTER 5 GETTING A GRASP ON CASP: PROPERTIES AND ROLE OF

THE CYTOHESIN-ASSOCIATED SCAFFOLDING PROTEIN IN IMMUNITY....85

5.1 ABSTRACT 86

5.2 INTRODUCTION 87

5.2.1 DISCOVERY AND EARLY YEARS 87

5.3 TRANSCRIPTIONAL ACTIVATION AND TRANSCRIPT PROPERTIES 88

5.4 PROTEIN PROPERTIES AND INTERACTIONS 91

5.4.1 THE CYTOHESIN/ARNO FAMILY 92

5.4.2 SNX27 AND THE SORTING NEXIN FAMILY 94

5.5 INTRACELLULAR LOCALIZATIONS AND FUNCTIONAL ACTIVITIES 99 5.6 FUNCTIONAL ASSESSMENT IN THE CONTEXT OF IMMUNITY: KNOCKOUTS AND RNAI 107

5.6.1 ROLE IN IMMUNE CELL POLARIZATION 108

5.6.2 ROLE IN CELL MIGRATION 110

5.7 ADAPTOR PROTEIN FAMILY: CASP AND GRASP 115

5.8 PROGRESS TO DATE AND OUTLOOK 116

5.9 ACKNOWLEDGEMENT.. 118

5.10 REFERENCES 119

5.11 LINKING PARAGRAPH 129

CHAPTER 6 SUPPORTING AND PRELIMINARY RESULTS AND DISCUSSION 130

6.1 SNX27 RECRUITMENT OF CASP TO ENDOSOMES 130

viii 6.1.1 CASPAPDZBM HAS ALTERED INTRACELLULAR DISTRIBUTION 130

6.1.2 MATERIALS AND METHODS 130

6.2 ADDITIONAL INTERACTIONS: CASP IS CLEAVED BY GRANZYME B IN VITRO 132

6.2.1 CASP AS A TARGET FOR GRB AND FURTHER

INTERACTIONS 132

6.2.1 MATERIALS AND METHODS 133

6.3 GRASP INTERACTS WITH SNX27 IN VITRO 137

6.4 LINKING PARAGRAPH 139 CHAPTER 7 CONCLUSION 140

REFERENCES 145

APPENDIX A DEVELOPMENT OF SNX27 ANTISERUM 159

APPENDIX B COPYRIGHT PERMISSIONS 162

ix LIST OF TABLES

Table 2.1 ESI-tandem mass spectrometry identification of SNX27 interacting with the carboxy terminus of CASP. 17

Table 5.1 Summary of CASP binding partners, related localizations and functions. 98

x LIST OF FIGURES

Figure 2.1 Identification of proteins interacting with the carboxy terminus of CASP 18

Figure 2.2 SNX27 interaction with the carboxy terminus of CASP is mediated via its PDZ domain in vitro 19

Figure 2.3 Co-immunoprecipitation of endogenous SNX27 with CASP and cytohesin-1 and detection of SNX27 in lymphocytes 20

Figure 2.4 Imrnunocytochemistry shows partial co-localization of endogenous CASP with SNX27 in YT cells 21

Figure 3.1 Endogenous SNX27 co-localizes with early endosomes and not with lysosomal or Golgi markers in NK-92 cells 37

Figure 3.2 Imrnunocytochemistry of killer :target conjugates shows SNX27 polarization to the IS in vitro 38

Figure 3.3 SNX27 polarization to the IS in conjugates is distinct from the similarly polarized perforin-containing cytotoxic granules in NK-92 cells 39

Figure 3.4 SNX27 polarizes to the leading edge, opposite the CD 18 enriched uropod of migrating NK-92 cells 40

Figure 4.1 A comparison of the domain distribution of human CASP and

GRASP proteins 65

Figure 4.2 Comparison of CASP and GRASP gene structures 66

Figure 4.3 Alignment of vertebrate CASP and GRASP protein sequences in the PDZ-binding motif region 6 8 Figure 4.4 Phylogenetic analysis of vertebrate CASP and GRASP proteins including invertebrate homologues 69

Figure 4.5 Comparison of gene loci from amphioxus, zebrafish, and human genomes 71

Figure 4.6 A simplified model depicting the sequence of events and relative evolutionary time line (left) for duplication of the primordial GRASP and NR4A locus 73

xi Figure 5.1 Domain and motif architecture in the CASP interactome. Protein domain distributions of CASP, the cytohesin/ARNO family, and SNX27 97

Figure 5.2 A model implicating CASP in putative novel polarized hematopoietic endosomal sorting events 106

Figure 5.3 A model depicting the role of CASP expression in antigen presentation by DCs 114

Figure 6.1 SNX27 recruits transfected CASP to the endosomal compartment,

dependent on the PDZ-PDZbm interaction 131

Figure 6.2 Recombinant CASP is cleaved by granzyme B in vitro 135

Figure 6.3 Predicted CASP proteolytic cleavage site by granzyme B 136

Figure A.l New antibodies for SNX27 immunoprecipitate endogenous protein 160

Figure A.2 SNX27a antiserum detects endogenous, endosomal SNX27 in YT cells 161

xn ABSTRACT

Cytohesin-associated scaffolding protein (CASP) is a novel human adaptor protein, expressed exclusively in immune cells. CASP participates in the assembly and recruitment of protein complexes associated with intracellular trafficking and signaling. I have discovered sorting nexin 27 (SNX27) as a new interaction partner for CASP using affinity chromatography and mass spectrometry, and further confirmed this interaction using co-immunoprecpitation (IP). The respective domains and motifs required for this interaction were determined to be the carboxy-terminal PDZ-binding motif (PDZbm) of CASP and the PDZ domain of SNX27. Cytohesin-1 was also detected in SNX27 IPs, indicating that a complex of all three proteins could occur in vivo. This links a protein family involved in vesicle initiation to a protein involved in sorting and trafficking vesicular cargo. Using immunocytochemistry and transfection studies, I also characterized the subcellular localization of this interaction at the early endosomal compartment of lymphocytes. Targeted deletion of the CASP PDZbm showed that this motif is required for SNX27 recruitment of transfected CASP to endosomes, demonstrating a functional significance for the interaction. In NK cells, endosomal SNX27 was found to change its cytoplasmic distribution from scattered to polarized at the immunological synapse during conjugation with target tumor cells. This apical polarization of SNX27 was also detected in migrating lymphocytes and is consistent with mouse models of CASP knockout where a deficiency in lymphocyte migration to inflammation sites, and in killing infected and/or tumor cells is observed. The relationship between CASP and GRASP, a related neuronal protein, was characterized by comparing both their gene structures and functional motifs across vertebrate organisms. I determined the evolutionary relationship of these proteins using phylogenetics and comparative analysis of the conservation of genes near each locus in various chordates. CASP and GRASP were determined to be the products of a relatively recent gene duplication event in early vertebrates. I also demonstrate that CASP is a putative proteolytic target for granzyme B and provide an integrative critical review of all CASP literature to date.

xiii LIST OF ABBREVIATIONS

5-HT4a 5-hydroxytryptamine 4a ADP adenosine diphosphate APC antigen presenting cell APvF ADP-ribosylation factor ARNO ARF nucleotide-binding site opener B41 four point one component FERM domain BFA brefeldin A BSA bovine serum albumin bp basepair CASP cytohesin-associated scaffolding protein CC coiled-coil motif CD cluster of differentiation cDNA complementary DNA CTL cytotoxic T lymphocyte Cybr cytohesin binder and regulator CYTIP cytohesin interacting protein DC dendritic cell DDC duplication-degeneration-complementation model DGKC diacylglycerol kinase C, EE early endosome EEA1 early endosomal antigen 1 EGF epidermal growth factor ER endoplasmic reticulum ESI electrospray ionization FBS fetal bovine serum FERM Four point one ezrin radixin moesin domain GALNT5 N-acetylgalactosaminyltransferase 5 GDP guanosine diphosphate GEF guanine nucleotide-exchange factor GIRK G protein-gated inwardly rectifying potassium channel GPD2 glycerol-3-phosphate dehydrogenase 2 GRASP GRP1-associated scaffolding protein GrB granzyme B GRP1 general receptor for phosphoinositides GRSP1 GRP1-binding protein GST glutathione ^-transferase GTP guanosine triphosphate GTPase GTP phosphatase HA hemagglutinin HRP horseradish peroxidase HSC hematopoietic stem cell hsp90 heat shock protein 90 IP immunoprecipitation

XIV kb kilobases kDa kilodalton Kir3 potassium (K) inwardly rectifying channel KO knockout ICAM-1 Inter-Cellular Adhesion Molecule 1 IL interleukin IPCEF1 Interaction protein for cytohesin exchange factors 1 IPTG isopropyl-P-D-thiogalactopyranoside IS immunological synapse LAMP-2 lysosomal-asscociated membrane protein 2 LAMP-3 lysosomal-asscociated membrane protein 3 LETMD1 LETM1 domain containing 1 LFA-1 lymphocyte function-associated antigen 1 LPS lipopolysaccharide LSM laser scanning microscope mGluR metabotropic glutamate receptor MTOC microtublue organizing center mRNA messenger RNA Mrtl methamphetamine-responsive transcript MHC major histocompatibility complex MS mass spectrometry NCBI national center for biotechnology information NFAT nuclear factor of activated T cells NK natural killer cell NR4A nuclear receptor 4A PAGE polyacrylamide gel electrophoresis PBL peripheral blood lymphocytes PBS phosphate-buffered saline PBMC peripheral blood mononuclear cells PDZ PSD-95/Dlg/ZO-l PDZbm PDZ-binding motif PH pleckstrin homology domain PHA phytohemagglutinin PI phospatidylinositol PI3K PI 3-kinase PI3P PI (3) phosphate PIP2 PI (4,5) phosphate PIP3 PI (3,4,5) phosphate PKC protein kinase C PMA phorbol 12-myristate 13-acetate PSCDBP pleckstrin homology Sec7 and coiled-coil domain-binding protein Ptdlns phospatidylinositol PX phox homology domain RA Ras associated domain RBC red blood cell RNAi RNA interference

XV SAPAP3 SAP90/PSD-95-associated protein 3 SDS sodium dodecyl sulfate SE sorting endosome Sec7 domain with homology to yeast Sec7 protein SMAC supramolecular activation cluster SNX sorting nexin TcR T cell receptor TGN trans Golgi network Th T helper cell TNF tumour necrosis factor TrkC receptor tyrosine kinase TX triton-x VCAM-1 vascular cell adhesion molecule 1

xvi ACKNOWLEDGEMENTS

I would like to thank my supervisor, Dr. Bill Pohajdak, for providing me with this opportunity, for his enthusiasm for research, and helpful expertise and advice. I would like to thank my committee members, Drs. Vett Lloyd and Vanya Ewart for their time and direction. I would also like to thank Dr. David Hoskin, Dr. Deborah Burshtyn, Dr. Marc Mansour for their contributions and help, as well as lab members Cameron Starratt, Asra'a Abidali, Dr. Olga Hyrtsenko, Dr. Andy Haigh, and Will MacDonald. I thank my mother and father for their tremendous support, my brother and sister, Kevin and Jenny, and finally my best friend, sounding board, and soulmate, Lori McEachern.

xvii CHAPTER 1: INTRODUCTION

The specialized roles of unique cells and tissues are driven by selective gene expression and manifest in the functional role of their resultant proteins. This paradigm is well-established and forms the foundation on which molecular cell biology stands in its aim to decipher the mechanisms that produce the variation and diversity of tissues present in living organisms. The most direct way to understand what makes a particular cell type unique is to study those proteins that are only present therein. Through these observations, we enhance our understanding of how cells operate at the most basic level and why they are capable of unique functions. It is through this pursuit that we can more intelligently and creatively treat illness and understand the consequences (positive or negative) of our interactions with the environment.

Today, the genomes of many organisms have been sequenced, including that of humans. Genome sequencing provides a catalogue of content, which in turn provides source material for a directory of gene expression. The product of this work is an index of which proteins are expressed and where. Often however, the more pressing issue is what the function of each protein is. Science has been examining this concept for most of the past century, but genome sequencing has placed our foot on the accelerator and provided a comprehensive index of variables to work with. Prior to this knowledge, many genes were sequenced independently and those proteins specifically expressed in particular tissues were discovered the hard way. The following thesis is the product of studies on one such protein which was discovered in our laboratory and termed cytohesin- associated scaffolding protein (CASP) (Dixon, Sahely et al. 1993; Mansour, Lee et al.

2002). CASP is a novel protein exclusively expressed in cells of the immune system

1 which are derived from hematopoietic stem cells (HSCs). Characterized as an adaptor or scaffolding protein, CASP is a relatively small (40 kDa) molecule that participates in the assembly of larger molecular complexes. By binding its various targets and bringing them into proper proximity or conformation, adaptor proteins facilitate or enhance the further functions of their respective binding partners. CASP is composed of a series of smaller modules called domains and/or motifs that provide unique binding properties in a particular orientation relative to one another. As an adaptor protein, CASP contains no catalytic domains with intrinsic activity, but rather a series of protein-protein interaction domains and motifs.

The interaction of CASP (also known as Cybr, CYTIP, and PSCDBP) with cytohesins has resulted in the bulk of what was known about this protein's function as well as its various names in the literature (Mansour, Lee et al. 2002; Tang, Cheng et al.

2002; Boehm, Hofer et al. 2003). The four-protein mammalian cytohesin/ARNO family interacts with CASP through the respective coiled-coiled motifs of each protein

(Mansour, Lee et al. 2002). The cytohesin/ARNOs are a group of guanine nucleotide- exchange factors (GEFs), which function to activate molecular switches termed ADP- ribosylation factors (ARFs). ARFs cycle between active (GTP-bound) and inactive

(GDP-bound) conformations determined by the phosphorylation state of their associated

guanosines (D'Souza-Schorey and Chavrier 2006). Cytohesins facilitate the exchange of

bound GDP for GTP, thereby activating the ARF. Active ARFs participate in signalling

and vesicle initiation on various membranes by recruiting coat proteins to their insertion

site (D'Souza-Schorey and Chavrier 2006). Together, this series of proteins regulate

2 intracellular traffic and signalling processes associated with a variety of biological functions including cell adhesion in lymphocytes (Kolanus, Nagel et al. 1996).

Cells of the human immune system, including lymphocytes, provide screening and protection against foreign infection and cancer and must traffic the various tissues and noisy microenvironments present in individual organisms. The organization and distribution of proteins, lipids, and organelles within an individual cell is a highly structured, but fluid process. A living cell is in a constant state of flux, with new proteins being continually synthesized, and damaged or unneeded molecules recycled or discarded. Maintenance of this process is incredibly complex on its own* but lymphocytes add a further layer of complexity to this dynamic intracellular organization, as they travel throughout an organism and encounter other cells which they then interact and communicate. This migration and communication requires a directional orientation that lymphocytes achieve via a process called polarization (Sanchez-Madrid and del Pozo

1999). Polarization results in the asymmetrical distribution of proteins and organelles within the cell and achieves directionality for the cell so that it may carry out a function at a particular intracellular location or move within its microenvironment in a defined direction. The process of distributing and re-distributing the cell's components is an active one that employs an abundance of proteins and lipids that function in either moving or clustering components at an appropriate location, or in signalling processes in response to specific stimuli (Sanchez-Madrid and del Pozo 1999). For example, during engagement of an infected or transformed cell, a specialized lymphocyte called the T cell, forms a spatially defined structure at the site of contact with the tumour cell, called the immunological synapse (IS). Membrane components involved in recognition and cell

3 activation, such as the T cell receptor (TcR), are clustered in a central region, while cell adhesion components, such as integrins, are clustered in a zone peripheral thereto

(Orange, Harris et al. 2003).

All blood cells are derived through hematopoiesis and originate in the bone marrow from hematopoietic stem cells (HSCs). HSCs subsequently differentiate into common progenitor cells for the myeloid and lymphoid lineages, producing a spectrum of cells that cumulatively compose the immune system's active machinery. The myeloid lineage includes macrophages, neutrophils, dendritic cells (DCs), and other leukocytes, as well as those cells giving rise to platelets and erythrocytes or red blood cells (RBCs). The lymphoid lineage includes T cells, B cells, natural killer (NK) cells, as well as lymphoid- derived DCs. Each cell type fills a particular role in a remarkably cooperative and resourceful system that collectively serves as the body's defence against infection and cancer. The immune system is classically divided into two branches, innate and adaptive.

Innate immunity is an organism's first line of defence. It requires no past exposure to a specific target or antigen, but is activated through a broad and general recognition of the presence of foreign materials. The adaptive immune system is a potent mechanism which essentially learns as the organism develops and proceeds through exposure to pathogens.

This branch of immunity employs a specific molecular recognition system that responds and amplifies based on the antigen or specific foreign protein encountered. These two branches of immunity are not entirely mutually exclusive, but often work together to form an efficient and comprehensive defence. This cooperative mechanism relies on effective communication though soluble factors secreted by various cells as well as physical cell- cell interactions. The versatility and efficiency of this system depends on competent and

4 rapid communication among various cells types in each scenario in order for uncompromised immunity and ultimately a healthy organism.

The objective of the following research at its inception was to discover and characterize new interaction partners for CASP in an effort to further understand its molecular role in cells of the immune system. This objective was met through the discovery and characterization of sorting nexin 27 (SNX27) as a new binding partner for

CASP and the characterization of CASP as a potential substrate for granzyme B (GrB) proteolysis. It then progressed further into a functional characterization of changes in the subcellular distribution of SNX27 in lymphocytes engaged in specialized immune activities dependent on polarization, including interactions of natural killer (NK) cells with tumour cells and lymphocyte migration. This research also developed into an analysis and description of the origins of a small two-member protein family comprised of CASP and a paralogous protein, GRP1-associated scaffolding protein (GRASP), which is primarily expressed in the nervous system. This research culminated in my writing and publishing a comprehensive review article on the properties and role of CASP in

immunity. The following chapters will present my research in a publication style format with article manuscripts separated by brief transition sections.

5 CHAPTER 2

SORTING NEXIN 27 INTERACTS WITH THE CYTOHESIN-ASSOCIATED SCAFFOLDING

PROTEIN (CASP) IN LYMPHOCYTES

Adam J. MacNeil, Marc Mansour, and Bill Pohajdak

Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada

Manuscript submitted May 8, 2007 Available online May 30, 2007 Published in Biochemical and Biophysical Research Communications 359(4): 848-853.

6 2.1 ABSTRACT

CASP is a small cytokine inducible protein, primarily expressed in hematopoetic cells, which associates with members of the cytohesin/ARNO family of guanine nucleotide-exchange factors. Cytohesins activate ARFs, a group of GTPases involved in vesicular initiation. Functionally, CASP is an adaptor protein containing a PDZ domain, a coiled-coil, and a potential carboxy terminal PDZ-binding motif that we sought to characterize here. Using GST pulldowns and mass spectrometry, we identified the novel interaction of CASP and sorting nexin 27 (SNX27). In lymphocytes, CASP's PDZ- binding motif interacts with the PDZ domain of SNX27. This protein is a unique member of the sorting nexin family of proteins, a group generally involved in the endocytic and intracellular sorting machinery. Endogenous SNX27 and CASP co-localize at the early endosomal compartment in lymphocytes and also in transfection studies. These results

suggest that endosomal SNX27 may recruit CASP to orchestrate intracellular trafficking

and/or signaling complexes.

2.2 INTRODUCTION

Trafficking of proteins and organelles within cells is a well-controlled and

regulated process involving many molecular interactions. Transport vesicles govern much of this movement by trafficking their cargo to the appropriate organelle or

membrane site. There are a multitude of proteins involved in the dynamic steps of vesicle

formation, sorting, and fusion (Bonifacino and Rojas 2006). The initiation of vesicle

formation is performed by a series of related proteins, called ARFs, belonging to the Ras

superfamily of small GTPases (D'Souza-Schorey and Chavrier 2006). Cytoplasmic GDP-

7 bound ARFs are inserted into the appropriate membrane through exchange of bound GDP for GTP via the Sec7 domain of a family of guanine nucleotide-exchange factors (GEFs)

(Jackson and Casanova 2000; Hawadle, Folarin et al. 2002). These GEF proteins, grouped on size and sensitivity to brefeldin A, have other protein-interacting domains including a coiled-coil motif (Pacheco-Rodriguez, Moss et al. 2005). We have previously characterized an adaptor protein named CASP (Cybr, CYTIP, PSCDBP, B3-1) that can interact with several members of the cytohesin/ARNO family of GEFs through its coiled- coil motif (Mansour, Lee et al. 2002). CASP is primarily expressed in hematopoetic cells with the highest expression found in cytokine-stimulated lymphocytes (Tang, Cheng et al.

2002; Coppola, Barrick et al. 2006; Watford, Li et al. 2006). This interaction, expression, and charcterization of CASP has also been substantiated by others (Tang, Cheng et al.

2002; Boehm, Hofer et al. 2003).

A structurally related adaptor protein called GRASP (GRP1-associated scaffolding protein) has been described in neuronal cells (Nevrivy, Peterson et al. 2000;

Kitano, Kimura et al. 2002). Like CASP, GRASP (also known as Tamalin), binds ARNO and ARN03/GRP1 through its coiled-coil motif, has a PDZ domain, and a functional carboxy terminal PDZ-binding motif (PDZbm) (Kitano, Kimura et al. 2002). PDZ domains have been well described in many proteins and usually interact with the carboxy terminus (PDZbm) of their binding partner. GRASP's PDZbm can bind to its own PDZ domain to form homodimers and to that of S-SCAM (Kitano, Yamazaki et al. 2003; Sugi,

Oyama et al. 2007). Most GRASP-interacting proteins are expressed in neuronal cells

indicating a cell-type restricted function for these complexes.

8 A group of multi-domain containing proteins referred to as sorting nexins (SNX) are primarily involved in endocytic trafficking (Worby and Dixon 2002; Carlton, Bujny et al. 2005). All of the SNX proteins have a phosphoinositide binding domain (Phox, PX)

(Seet and Hong 2006) and most have various other protein-interacting domains (Carlton,

Bujny et al. 2005). Only one member of the SNX family, SNX27, contains a PDZ domain. In neurons, SNX27's PDZ domain interacts with the 5-hydroxytryptamine 4a (5-

HT4a) receptor (Joubert, Hanson et al. 2004). SNX27, also known as Mrtl

(methamphetamine-responsive transcript 1), appears to be expressed in many tissues including the brain (Kajii, Muraoka et al. 2003) and in lymphocytes, as has been recently shown (Rincon, Santos et al. 2007) and we confirm here. We report a novel interaction in which CASP's carboxy PDZbm can directly bind with SNX27's PDZ domain. CASP co- localizes with SNX27 at the early endosomal compartment of lymphocytes and in transfection studies using HEK 293 cells suggesting a role for this interaction in endocytic trafficking and/or signaling.

2.3 MATERIALS AND METHODS

2.3.1 CELLS AND LYSATES

Human NK/T cell line YT was cultured in RPMI1640 (Gibco) with 10% FBS and antibiotics while Raji cells were cultured similarly but in 5% FBS. NK-92 cells (a gift from Dr. D. Burshtyn, University of Alberta) was grown in RPMI 1640 supplemented with 12.5% FBS, 12.5% horse serum (Gibco), 50uM p-mercaptoethanol (Sigma), 100 units/mL of IL-2 (PeproTech), and antibiotics. HEK 293 cells were cultured in DMEM

(Gibco) with 10% FBS, 1% non-essential amino acids, HEPES, and antibiotics. Human

9 peripheral blood lymphocytes (PBL) were separated from RBCs by centrifugation in a

Histopaque gradient (Sigma). All cells were lysed in 0.5% Triton-X 100, 0.5% NP-40,

5% Glycerol, 50mM Tris (pH 7.4), 150mM NaCl, and Complete protease inhibitors

(Roche).

2.3.2 DNA AND RECOMBINANT PROTEINS

GST-CASP and ASRF constructs were made by amplifying CASP cDNA using a common forward primer 5'-GTTACAGGAACATCGTCTGC-3' and reverse primer 5'-

AGGAACCTGCATCTTTGTTA-3', corresponding to the complete carboxy tail of

CASP, and 5'-TTATTATTCTCATCTTTCCACAGCAC-3' for the ASRF tail.

Amplicons were subcloned into pGEX-1 (AMRAD). Recombinant protein was expressed in Rosetta E.coli by IPTG induction (ImM, 2-3 hours) and purified on glutathione beads (Sigma).

SNX27 cDNA was cloned from YT cells using a reverse transcriptase PCR kit

(Invitrogen). The forward primer 5'-ATGGCGGACGAGGACGGGGAAGGGATT-3' was used with reverse primer 5'-CTAGGTGGCCACATCTCTCTGCTGTGACCT-3' to clone the full SNX27a cDNA into the pCRII plasmid. The 6His-SNX27 plasmid was made by subcloning an £coi?/-digested fragment into the His6 pET32-C vector

(Novagen). Recombinant protein was purified on Ni-NTA agarose (Qiagen). All vectors were confirmed by DNA sequencing.

10 2.3.3 GST PULLDOWNS, WESTERN ANALYSIS AND ESI-TANDEM MASS SPECTROMETRY

Cleared lysates (2-3x107 cells/ml) were incubated with equal amounts of purified recombinant proteins (GST-CASP and ASRF) on Glutathione beads at 4°C for 2-16 hours. Beads were washed extensively with PBS-Triton X, heated at 37°C for 45 minutes in IX SDS reducing buffer, and supernatants were electrophoresed on 10% SDS-PAGE gels. Gels were stained with Coomassie Blue R-250 (Sigma).

All bands of interest were excised and analyzed by ESI-tandem mass spectrometry

(Applied Biosystems Q TRAP) at the DalGEN Proteomics Facility (Atlantic Research

Centre, Dalhousie University). Resulting peptides were searched using MASCOT and/or

SWISSPROT databases and corresponding proteins were identified.

For GST pulldowns with His6-SNX27, lOOng of pre-cleared protein was incubated with equal amounts of GST recombinants. Beads were washed, resuspended in loading buffer and supernatants were electrophoresed on a 10% SDS PAGE gel. Proteins were transferred onto PVDF membrane (Amersham, GE) and SNX27 recombinant protein was detected with S Protein HRP Conjugate (Novagen) using standard chemiluminescent (Santa Cruz) procedures. GST-CASP recombinants were detected with anti-GST (Santa Cruz) and anti-rabbit HRP (Santa Cruz).

2.3.4 CO-IMMUNOPRECIPITATION

Lysates (2-3x107 cells/ml) were cleared and 5jag of CASP-specific anti-PSCDBP

(Abeam) or PBS (control) was added and incubated at 4°C for 2-4 hours. Protein G-Plus

Agarose beads (Santa Cruz) were added and incubated for 2 hours at 4°C. Isolated proteins were transferred onto PVDF membrane and SNX27 was detected with SNX27

11 antiserum and anti-rabbit HRP secondary. SNX27 antiserum (a generous gift from Dr. T.

Nishikawa, Tokyo Medical and Dental University, Japan) was raised in rabbit against rat

SNX27 (Mrtl), and to demonstrate its cross-reactivity with human SNX27, 50 \xg of lysate from YT, human PBLs, and rat spleen lymphocytes was probed with SNX27 antiserum (1:500) and anti-rabbit HRP. For immunoprecipitations from YT, Raji, and rat spleen lymphocytes, lOul of SNX27 antiserum was used. Cytohesin-1 was detected with monoclonal anti-Cytohesin clone 2E11 (Dr. Sylvain Bourgoin, Universite Laval).

2.3.5 IMMUNOCYTOCHEMISTRY AND TRANSFECTIONS

For immunocytochemistry, 2-5x105 cells self-adhered to poly-L-lysine coated

slides (Lab Scientific) were fixed with 4% paraformaldehyde and permeabilized with

0.2% Triton-X 100 in PBS (PBS-TX) for 5 minutes. Slides were blocked with 1% BSA in PBS-TX followed by primary antibodies at room temperature for 30 minutes. SNX27 antiserum and CASP-specific antibodies 2F9 (Ascenion) were used at 1:200 and anti-

EEA1 (BD Biosciences) was used at 1:50. Slides were then washed before applying Cy3

and Alexa 488 conjugated secondary antibodies (Jackson ImmunoResearch and

Molecular Probes) and washed extensively before application of VectaShield mounting medium (Vector Laboratories).

For transfections, HEK 293 cells were seeded in 6-well plates with glass

coverslips and transfected with 4^ig of full-length CASP construct (pJ3H vector) with an

N-terminal HA tag using Lipofectamine 2000 (Invitrogen). Cells were then grown for

24-48 hours in the presence of 10% FBS supplemented DMEM without antibiotics.

Immunocytochemistry followed using monoclonal anti-HA (Santa Cruz) and SNX27

12 antiserum. All cells were viewed and imaged using an LSM 510 laser scanning confocal microscope with a 63X oil objective lens (Zeiss).

2.4 RESULTS

2.4.1 SNX27 INTERACTS WITH THE CARBOXY-TERMINAL PDZ-BINDING MOTIF OF CASP IN LYMPHOCYTES

Since CASP is primarily expressed in lymphocytes, we set out to identify proteins interacting with the putative carboxy terminal PDZbm of CASP in these cells. The carboxy terminal amino acids of CASP (S-R-F) correspond to a class I PDZbm (S/T-X-

O) (Hung and Sheng 2002), yet no partner had been identified. A frequently used approach for identifying proteins interacting with a carboxy PDZbm involves the use of truncated mutants (Joubert, Hanson et al. 2004; Malmberg, Andersson et al. 2004). By deleting these few but critical amino acids, binding to a respective PDZ domain is often disrupted. To this end, GST fusion proteins corresponding to the full carboxy terminal tail and a truncated (ASRF) mutant were produced and purified (Figure 2.1 A). Fusion proteins were incubated with YT cell lysates (Figure 2.IB) or human peripheral blood lymphocyte (PBL) lysates (Figure 2.1C) and three putative proteins interacting with GST-

CASP but not with ASRF were identified in GST pulldown assays (Figure 2.1B-C). ESI- tandem mass spectrometry analysis of the isolated protein bands identified them as variants of SNX27 (Table 2.1). This experiment was repeated five times with similar results. Until recently, SNX27 (Mrtl) was reported to be predominantly expressed in several brain tissues (Kajii, Muraoka et al. 2003), whereas we, and others (Rincon, Santos et al. 2007), have isolated and detected the protein in various lymphocytes as well.

13 2.4.2 PHYSICAL INTERACTION WITH CASP IS MEDIATED BY THE PDZ DOMAIN OF SNX27

The SNX27 cDNA was cloned from YT cells and a 6His fusion protein harbouring its full PDZ domain and a portion of the PX domain was designed and produced (Figure 2.2A). In vitro GST pulldown assays were conducted using the GST-

CASP PDZbm constructs (Figure 2.1 A) and the PDZ domain of SNX27. The 6His-

SNX27 PDZ domain fusion protein co-precipitated with GST-CASP but not GST-

CASPASRF, confirming a direct physical interaction between the PDZbm of CASP and the PDZ domain of SNX27 (Figure 2.2B). SNX27a and b isoforms, representing splice variants with slightly different carboxy termini, both contain a PDZ domain capable of interacting with the PDZbm of CASP.

2.4.3 SNX27 CO-LMMUNOPRECIPITATES WITH CASP AND CYTOHESIN-1 IN LYMPHOCYTES

SNX27 was also found to co-immunoprecipitate with endogenous CASP in YT lysates using an anti-CASP antibody, indicating that both proteins interact in their native conformations (Figure 2.3A). SNX27 antiserum was used to detect SNX27 in

immunoprecipitations (IPs). Cell lysates from YT, human PBLs, and rat spleen

lymphocytes were examined by Western analysis and an equivalent band was detected in

each (Figure 2.3B). SNX27 antiserum also recognized SNX27 in IPs using YT, Raji, and

rat spleen lymphocytes (Figure 2.3C). CASP could not be detected in SNX27 IPs, most

likely due to weak CASP antibodies on Western blots or possibly due to interference from the IP antibody IgG heavy chain, which also blocked further attempts to detect CASP in

IPs using mass spectrometry. However, the CASP binding partner, cytohesin-1, was

14 detected in SNX27 IPs (Figure 2.3C) indicating CASP's likely presence in this IP

complex and suggesting that all three proteins may simultaneously interact in vivo. It is

very unlikely that SNX27 could interact with cytohesin-1 without the presence of CASP

since the GST-CASP construct used to bind SNX27 did not contain the cytohesin/ARNO

interacting coiled-coil motif, precluding cytohesin's presence unless bound to a protein

other than CASP (Figure 2.1 A). As expected, mass spectrometry analysis of protein at the molecular weight of cytohesin-1 from GST pulldown gels (Figure 2.1B-C), did not

detect cytohesin-1 despite the presence of SNX27.

Altogether, these data indicate that SNX27, via its PDZ domain, binds the carboxy terminal PDZbm of CASP in vitro and in lymphocytes, and that this interaction is occurs

through the classical PDZ mode of binding.

2.4.4 ENDOGENOUS SNX27 CO-LOCALIZES WITH CASP AT EARLY ENDOSOMES IN YT CELLS AND RECRUITS TRANSFECTED CASP TO ENDOSOMES IN HEK-293 CELLS

SNX27 and CASP partially co-localized in vivo using the anti-CASP antibody

2F9 and SNX27 antiserum in YT cells (Figure 2.4A-C). SNX27 showed a distinct

tubulo-vesicular localization. CASP was partially co-localized with SNX27 at tubulo-

vesicular structures but also exhibited a diffuse cytoplasmic distribution. The largest

amount of co-localization was found at the larger endosome structures. In outlying

regions, away from these larger structures, there were fewer apparent co-localizations

observed.

Antibody markers were then used to identify the subcellular compartment in

which CASP and SNX27 might interact in lymphocytes. The vesicular structures where

15 CASP and SNX27 interact were identified as early endosomes as SNX27 was found to co-localize with the early endosomal marker EEA1 in YT cells (Figure 2.4D-F). SNX27 did not co-localize with either lysosomal markers, Lamp-2 or Lamp-3, nor did it co- localize with the Golgi complex marker, Giantin (data not shown). The presented data is representative of many observations using several additional cell types (NK-92 and Raji, data not shown).

To further confirm whether these two proteins were co-localizing, transfection experiments were conducted in HEK 293 cells using a vector encoding full length CASP with an N-terminal HA tag. Whereas endosomal SNX27 protein is natively found in HEK

293 cells, CASP is not. The transfected CASP protein again co-localized with endogenous SNX27 in these cells (Figure 2.4G-I). Analysis of the HEK 293 cells indicated that the proteins again specifically co-localized at large structures of the early endosomal compartment (Figure 2.4J-L). This result may suggest that SNX27 participates in recruiting CASP protein to endosomal membranes where they interact and may take part in trafficking and/or signaling mechanisms. Cumulatively, these results

suggest an interaction between SNX27 and CASP at the early endosomes of lymphocytes.

16 Table 2.1 ESI-tandem mass spectrometry identification of SNX27 interacting with

the carboxy terminus of CASP.

Banda Source Protein (MASCOT)b Accession no.c Peptides % Coverage MOWSE Scored

1 YT SNX27 gi|31742501 7 16 238

2 YT SNX27 gi|31742501 7 14 324

3 human PBL SNX27 gi|20140140 10 18 374

aNumbers correspond to those indicated in Fig. 1B-C.

bMASCOT database search was performed.

cNCBInr database accession numbers are listed.

dMOWSE Score >48 (1-2) or >38 (3) indicates significance (p<0.05).

17 A CASP EEEESRF

359aa GST-CASP ^EEEESRF GST-CASPASRF ]-EEEE B // ° // _JL <^_ KDa

83 83

62 62 3H 47 5 • - 47.5

32.5 32.5

Figure. 2.1 Identification of proteins interacting with the carboxy terminus of CASP. (A) Diagram of full-length CASP and GST recombinant proteins. (B) GST pulldown using recombinant CASP proteins as bait in YT cell lysate. Arrows indicate bands retained by GST-CASP but not the truncated ASRF mutant and not present in the original GST fusion protein purifications. (C) GST-pulldown using recombinant CASP proteins as bait in human PBL lysate. Arrow indicates a band retained by GST-CASP but not the truncated ASRF mutant.

18 (B)528aa SNX27 |i£fl PX RA/B41 1 (A) 541aa 6HIS.SNX27 h^^l Eg

B GST-CASP + GST-CASPASRF - + 6His-SNX27 + +

Figure. 2.2 SNX27 interaction with the carboxy terminus of CASP is mediated via its PDZ domain in vitro. (A) Diagram of full-length SNX27 (A and B isoforms) and 6His- SNX27 fusion protein containing the full PDZ domain. (B) GST pulldown with recombinant CASP constructs (Fig. 1A) using 6His-SNX27 purified protein. Recombinant SNX27 is indicated by an arrow. GST-CASP proteins were detected with anti-GST antibodies, lower panel. A smaller truncation band from the original purification of the GST constructs from E.coli is present in both lanes and had no effect on the experiment.

19 anti-CASP Protein G Plus KDa & 72 J ,<<* Cell Lysate -6 <^ Q0<> KDa 55

40 WB: anti-SNX27

0) B iv WB: anti-SNX27 ^-_ CO (^ z Cell Lysate A KDa 55 X 170 •^1 40

24 WB: anti-Cytohesin-1

WB: anti-SNX27

Figure 2.3 Co-immunoprecipitation of endogenous SNX27 with CASP and cytohesin-1 and detection of SNX27 in lymphocytes. (A) IP was performed with and without anti- CASP antibodies. SNX27 was detected using SNX27 antiserum. (B) SNX27 antiserum detects SNX27 in YT, human PBL, and rat spleen lymphocyte lysates. (C) SNX27 antiserum IPs SNX27 in YT, Raji, and rat spleen lymphocytes. Cytohesin-1 is also detected in SNX27 IPs. A control lane representative of IPs conducted on lysates without the addition of SNX27 antiserum is indicated.

20 t

CO o> CM

LU X

Figure 2.4 Immunocytochemistry shows partial co-localization of endogenous CASP with SNX27 in YT cells. (A-C) CASP and SNX27 co-localize using anti-CASP antibody 2F9 and SNX27 antiserum in YT cells. (D-F) Endogenous SNX27 co-localizes with EEA1 marker in YT cells. (G-I) HEK 293 cells transiently transfected with full-length HA-CASP show co-localization with endogenous SNX27. (J-L) Endogenous SNX27 co- localizes with the early endosomal marker EEA1 in HEK 293 cells.

21 2.5 DISCUSSION

One approach toward further characterizing the unique functions of a specific cell- type is to better understand the proteins and interactions exclusive to those cells. Here we describe a novel protein interaction in lymphocytes in which SNX27, via its N-terminal

PDZ domain, binds the carboxy terminus of CASP. This interaction requires the classical

PDZ mode of binding. Such PDZ interactions are associated with large protein complex assembly and targeting for signaling and subcellular transport (Hung and Sheng 2002).

SNX27 is a unique member of the sorting nexin family in that it is the only SNX that contains a PDZ domain. SNX27 was originally detected as a stimulant-inducible transcript in neurons (Kajii, Muraoka et al. 2003) and its PDZ domain was subsequently shown to interact with the PDZbm of the 5-HT4a receptor (Joubert, Hanson et al. 2004).

Recently SNX27 has also been found in lymphocytes where it interacts with

Diacylglycerol kinase L, (DGKQ via its PDZ domain (Rincon, Santos et al. 2007). Here we also show SNX27 expression in lymphoid cells and its interaction with CASP, a protein widely accepted as being restricted to hematopoetic cells and primarily expressed in lymphocytes. Interestingly, both CASP and 5-HT4a receptors have similar carboxy termini (5-HT4a: -ESLESCF and CASP: -EEEESRF), but 5-HT4 receptors are not found in lymphocytes. The tail of DGK^ (-EDQETAV) differs further from the previous two

PDZbm yet stills conforms to the class I PDZbm and is expressed in many tissues. This suggests a cell-type specific role for SNX27, dependent upon the interacting proteins in each respective cellular milieu and raises the question of whether CASP and DGK^

22 compete for binding to the PDZ domain of SNX27 in certain lymphocytes and what this competition may affect.

All members of the sorting nexin family contain a PX domain that usually binds

PtdIns(3)P (PI3P) (Seet and Hong 2006). These products of PI 3-kinase, are often enriched at endosomes, but are also found at other membranes (Seet and Hong 2006).

The early endosomal localization of many SNX family members has been well described.

Consistent with previous findings for transfected SNX27 (Joubert, Hanson et al. 2004;

Rincon, Santos et al. 2007), we also show this protein's endogenous localization to the early endosomes and detect an overlap with CASP at this compartment in NK/T cells

(Fig. 4). Since SNX27 is found at early endosomes regardless of the presence of CASP protein (Fig. 4J-L), it is likely that SNX27 participates in recruitment of CASP to the early endosomal membranes where they may function in coordinating intracellular traffic.

CASP's coiled-coil binding partners, cytohesin/ARNO family members, have a pleckstrin homology (PH) domain that usually bind PtdIns(3,4,5)P3 (PIP3) and

PtdIns(4,5)P2 (PIP2), phospholipids, generally found at the plasma membrane (Czech

2003). However, ARNO (cytohesin-2) has also been shown to localize to endosomes and to play a role in intra-endosomal acidification via V-ATPase (Hurtado-Lorenzo, Skinner et al. 2006). ARNO's coiled-coil domain was also shown to be involved in apical endocytosis of polarized epithelial cells (Shmuel, Santy et al. 2006). Interestingly, we have detected low levels of the CASP binding partner, cytohesin-1, in a complex with immunoprecipitated SNX27 (Fig. 3C). The CASP-SNX27 interaction we report could support a mechanism where members of the cytohesin/ARNO family are dynamically

23 recruited to early endosomal membranes for vesicular trafficking in lymphocytes but this remains to be fully explored.

In summary, we have characterized a cell-type specific and novel interaction between CASP and SNX27 in lymphocytes and determined that these proteins interact at the early endosomal compartment. From transfection studies in cells not expressing endogenous CASP we have concluded that SNX27 likely recruits CASP to endosomal membranes where they may participate in coordinating vesicular traffic and/or signaling complexes.

2.6 ACKNOWLEDGEMENT

Funding for this research was from the Natural Sciences and Engineering Research

Council of Canada (NSERC).

24 2.7 REFERENCES

Boehm, T., S. Hofer, P. Winklehner, B. Kellersch, C. Geiger, A. Trockenbacher, S. Neyer, H. Fiegl, S. Ebner, L. Ivarsson, R. Schneider, E. Kremmer, C. Heufler and W. Kolanus (2003). Attenuation of cell adhesion in lymphocytes is regulated by CYTIP, a protein which mediates signal complex sequestration. Embo J 22(5): 1014-24.

Bonifacino, J. S. and R. Rojas (2006). Retrograde transport from endosomes to the trans- Golgi network. Nat Rev Mol Cell Biol 7(8): 568-579.

Carlton, J., M. Bujny, A. Rutherford and P. Cullen (2005). Sorting nexins—unifying trends and new perspectives. Traffic 6(2): 75-82.

Coppola, V., C. A. Barrick, S. Bobisse, M. C. Rodriguez-Galan, M. Pivetta, D. Reynolds, O. M. Howard, M. E. Palko, P. F. Esteban, H. A. Young, A. Rosato and L. Tessarollo (2006). The scaffold protein Cybr is required for cytokine-modulated trafficking of leukocytes in vivo. Mol Cell Biol 26(14): 5249-58.

Czech, M. P. (2003). Dynamics of phosphoinositides in membrane retrieval and insertion. Arum Rev Physiol 65: 791-815.

D'Souza-Schorey, C. and P. Chavrier (2006). ARF proteins: roles in membrane traffic and beyond. Nat Rev Mol Cell Biol 7(5): 347-58.

Hawadle, M. A., N. Folarin, R. Martin and T. R. Jackson (2002). Cytohesins and centaurins control subcellular trafficking of macromolecular signaling complexes: regulation by phosphoinositides and ADP-ribosylation factors. Biol Res 35(2): 247-65.

Hung, A. Y. and M. Sheng (2002). PDZ domains: structural modules for protein complex assembly. J Biol Chem 277(8): 5699-702.

Hurtado-Lorenzo, A., M. Skinner, J. El Annan, M. Futai, G. H. Sun-Wada, S. Bourgoin, J. Casanova, A. Wildeman, S. Bechoua, D. A. Ausiello, D. Brown and V. Marshansky (2006). V-ATPase interacts with ARNO and Arf6 in early endosomes and regulates the protein degradative pathway. Nat Cell Biol 8(2): 124-36.

25 Jackson, C. L. and J. E. Casanova (2000). Turning on ARF: the Sec7 family of guanine- nucleotide-exchange factors. Trends Cell Biol 10(2): 60-7.

Joubert, L., B. Hanson, G. Barthet, M. Sebben, S. Claeysen, W. Hong, P. Marin, A. Dumuis and J. Bockaert (2004). New sorting nexin (SNX27) and NHERF specifically interact with the 5-HT4a receptor splice variant: roles in receptor targeting. J Cell Sci 117(Pt 22): 5367-79.

Kajii, Y., S. Muraoka, S. Hiraoka, K. Fujiyama, A. Umino and T. Nishikawa (2003). A developmentally regulated and psychostimulant-inducible novel rat gene mrtl encoding PDZ-PX proteins isolated in the neocortex. Mol Psychiatry 8(4): 434-44.

Kitano, J., K. Kimura, Y. Yamazaki, T. Soda, R. Shigemoto, Y. Nakajima and S. Nakanishi (2002). Tamalin, a PDZ domain-containing protein, links a protein complex formation of group 1 metabotropic glutamate receptors and the guanine nucleotide exchange factor cytohesins. JNeurosci 22(4): 1280-9.

Kitano, J., Y. Yamazaki, K. Kimura, T. Masukado, Y. Nakajima and S. Nakanishi (2003). Tamalin is a scaffold protein that interacts with multiple neuronal proteins in distinct modes of protein-protein association. J Biol Chem 278(17): 14762-8.

Malmberg, E. K., C. X. Andersson, M. Gentzsch, J. H. Chen, A. Mengos, L. Cui, G. C. Hansson and J. R. Riordan (2004). Bcr (breakpoint cluster region) protein binds to PDZ-domains of scaffold protein PDZK1 and vesicle coat protein Mint3. J Cell Sci 117(Pt 23): 5535-41.

Mansour, M., S. Y. Lee and B. Pohajdak (2002). The N-terminal coiled coil domain of the cytohesin/ARNO family of guanine nucleotide exchange factors interacts with the scaffolding protein CASP. J Biol Chem 277(35): 32302-9.

Nevrivy, D. J., V. J. Peterson, D. Avram, J. E. Ishmael, S. G. Hansen, P. Dowell, D. E. Hruby, M. I. Dawson and M. Leid (2000). Interaction of GRASP, a protein encoded by a novel retinoic acid-induced gene, with members of the cytohesin family of guanine nucleotide exchange factors. J Biol Chem 275(22): 16827-36.

Pacheco-Rodriguez, G., J. Moss and M. Vaughan (2005). Cytohesin-1: structure, function, and ARF activation. Methods Enzymol 404: 184-95.

Rincon, E., T. Santos, A. Avila-Flores, J. P. Albar, V. Lalioti, C. Lei, W. Hong and I. Merida (2007). Proteomic identification of sorting nexin 27 as a diacylglycerol

26 kinase zeta -associated protein: new diacylglycerol kinase roles in endocytic recycling. Mol Cell Proteomics 6(6): 1073-1087.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta 1761(8): 878-96.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta.

Shmuel, M., L. C. Santy, S. Frank, D. Avrahami, J. E. Casanova and Y. Altschuler (2006). ARNO through its coiled-coil domain regulates endocytosis at the apical surface of polarized epithelial cells. J Biol Chem 281(19): 13300-8.

Sugi, T., T. Oyama, T. Muto, S. Nakanishi, K. Morikawa and H. Jingami (2007). Crystal structures of autoinhibitory PDZ domain of Tamalin: implications for metabotropic glutamate receptor trafficking regulation. Embo J26(8): 2192-2205.

Tang, P., T. P. Cheng, D. Agnello, C. Y. Wu, B. D. Hissong, W. T. Watford, H. J. Ahn, J. Galon, J. Moss, M. Vaughan, J. J. O'Shea and M. Gadina (2002). Cybr, a cytokine-inducible protein that binds cytohesin-1 and regulates its activity. Proc Natl Acad Sci USA 99(5): 2625-9.

Watford, W. T., D. Li, D. Agnello, L. Durant, K. Yamaoka, Z. J. Yao, H. J. Ahn, T. P. Cheng, S. R. Hofmann, T. Cogliati, A. Chen, B. D. Hissong, M. R. Husa, P. Schwartzberg, J. O'Shea J and M. Gadina (2006). Cytohesin Binder and Regulator (Cybr) Is Not Essential for T- and Dendritic-Cell Activation and Differentiation. Mol Cell Biol 26(17): 6623-6632.

Worby, C. A. and J. E. Dixon (2002). Sorting out the cellular functions of sorting nexins. Nat Rev Mol Cell Biol 3(12): 919-31.

27 2.8 LINKING PARAGRAPH

The following chapter resulted from an investigation into changes in the intracellular localization of SNX27 and CASP during polarized events in lymphocytes.

Others had shown changes in CASP localization in dendritic cells during T cell priming.

Here, we specifically examine cytotoxic cells during engagements/contacts with tumour cells and during cell migration. This chapter was the first report on specific changes in sorting nexin distribution and was published soon after chapter 2, both in Biochemical and Biophysical Research Communications.

28 CHAPTER 3

POLARIZATION OF ENDOSOMAL SNX27 IN MIGRATING AND TUMOR-ENGAGED NATURAL KILLER CELLS

Adam J. MacNeil and Bill Pohajdak

Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada

Manuscript submitted June 25, 2007 Available online July 16, 2007 Published in Biochemical and Biophysical Research Communications 361(1): 146-150. 3.1 ABSTRACT

Polarization is a critical mechanism for the proper functioning of many cell types.

For lymphocytes, it is essential in a variety of processes, including migration from the blood to other tissue sites and vice versa. In NK cells and CTLs, the cytotoxic granule delivery mechanism requires polarization for granule movement to the immunological synapse (IS), in killing tumor and virus-infected cells. Recently, it has become apparent that endosomes are also involved in the cytotoxic mechanism. Using an in vitro conjugation approach, we show that in NK-92 cells, endosomal sorting nexin 27 (SNX27) polarizes to the IS during tumor cell engagement in a distinct compartment adjacent to the cytotoxic granules. We also show that SNX27 polarizes to the apical membrane, opposite the uropod, during NK cell migration. These previously unreported results indicate that

SNX27 is a participant in NK cell polarization, as a mediator or target of the mechanism.

3.2 INTRODUCTION

Lymphocytes are mobile, migrating cells that require dynamic intracellular machinery to quickly change direction and migrate in and out of blood vessels and to specific tissues in response to extracellular cues such as chemokines and other effector ligands. A cell defines its direction by polarizing many of its molecular components and organelles to one end of the cell or the other (Sanchez-Madrid and del Pozo 1999;

Ludford-Menting, Oliaro et al. 2005; Krummel and Macara 2006). Cellular polarization is a dramatic event and a large number of intracellular molecules and organelles participate in the redistribution of cytoplasmic contents (Montoya, Sancho et al. 2002).

30 In establishing polarization, a cell can be orientated to perform necessary functions at the appropriate membrane site or move in a particular direction depending on the extracellular signals detected (Sanchez-Madrid and del Pozo 1999). For a cell to migrate from one tissue site to another, polarization is critical in defining the direction of movement. Cytotoxic lymphocytes have a specialized mechanism, degranulation, which also requires intracellular polarization (Montoya, Sancho et al. 2002). In the event a competent NK cell or CTL encounters a tumor or virus-infected cell, the killer cell must polarize and move cytotoxic granules to the appropriate membrane site for degranulation to occur at the immunological synapse (Trambas and Griffiths 2003).

Sorting Nexin 27 (also known as Mrtl) is a member of the SNX family of proteins, which are typically involved in the endocytic sorting machinery (Carlton, Bujny et al. 2005). SNX27 is unique amongst this family of proteins as it is the only member containing a PDZ domain. Many PDZ domain containing proteins have been implicated in cell polarity events and if their expression is reduced, can inhibit polarization (Ludford-

Menting, Oliaro et al. 2005). We have recently described a PDZ domain-mediated interaction between SNX27 and CASP (Cytohesin Assocaited Scaffolding Protein) and localized this interaction to the endosomal compartment (MacNeil, Mansour et al. 2007).

CASP knockout studies have also indicated that this protein is involved in lymphocyte migration and possibly in tumor cell cytolysis (Coppola, Barrick et al. 2006). Recent studies have also described the essential role played by proteins of the endosomal compartment in delivering the contents of cytotoxic granules to the IS for cell-mediated cytotoxicity (Menager, Menasche et al. 2007).

31 Here we show previously unreported polarization of endosomal SNX27 to the IS in NK cells conjugated to target tumor cells. Endosomal SNX27 is detected in a compartment distinct from, but in the same vicinity as the cytotoxic granules. We also show that SNX27 polarizes to the leading, apical edge during lymphocyte migration.

These results support the recent description of CASP knockout models in which polarized events such as cell migration and tumor cytolysis are inhibited.

3.3 MATERIALS AND METHODS

3.3.1 CELLS AND ANTIBODIES

Human cells lines NK-92 and YTS cells were a gift from Dr. D. Burshtyn,

(University of Alberta). NK-92 was grown in RPMI 1640 supplemented with 12.5%

FBS, 12.5% horse serum (Gibco), 50 uM p-mercaptoethanol (Sigma), 100 units/mL of

IL-2 (PeproTech), and antibiotics. YTS and K562 cells were cultured in RPMI 1640

(Gibco) with 10% FBS and antibiotics.

SNX27 antiserum (a gift from Dr. T. Nishikawa, Tokyo Medical and Dental

University, Japan) was used at 1:200 while subcellular marker antibodies used were: anti-EEAl (BD Biosciences); anti-perforin (BioLegend); anti-KIM185 (Celltech

Therapeutics); anti-Lamp-2 and 3 (Developmental Studies Hybridoma Bank) were a gift from Dr. Vett Lloyd (Dalhousie University); and anti-Giantin was a gift from Dr. H.-P.

Hauri (University of Basel, Switzerland).

32 3.3.2 IMMUNOCYTOCHEMISTRY

For immunocytochemistry, 2-5x105 cells were allowed to adhere to poly-L-lysine coated slides (Lab Scientific) for 15 minutes at 37°C. Cells were then fixed with 4% paraformaldehyde. Cells were permeabilized with 0.2% Triton-X 100 in PBS (PBS-TX) for 5 minutes. Slides were blocked with 1% BSA in PBS-TX and primary antibodies/antisera were incubated at room temperature for 30 minutes. Slides were then washed before applying Cy3 and Alexa 488 conjugated secondary antibodies (Jackson

ImmunoResearch and Molecular Probes). Finally, slides were washed extensively before application of VectaShield mounting medium (Vector Laboratories).

3.3.3 CONJUGATION ASSAYS

For conjugation assays, 5xl05 killer cells (NK-92 or YTS) were combined with the target cell line (K562) at a 1:1 ratio and centrifuged for five minutes at very low speed for conjugate formation. Conjugates were incubated for either 15 or 30 minutes at 37°C prior to incubation on poly-L-lysine coated slides and immunocytochemistry, as described above. All cells were viewed and imaged using an LSM 510 laser scanning confocal microscope with a 63X oil objective lens (Zeiss).

3.4 RESULTS

3.4.1 ENDOGENOUS SNX27 LOCALIZES TO THE EARLY ENDOSOMES IN NK-92 CELLS

The subcellular distribution of endogenous SNX27 was consistent with the early endosomal marker EEA1 in NK-92 cells with co-localizations occurring at the larger structures of this compartment (Figure 3.1 A-C). This is consistent with previous work in

33 our lab showing SNX27 localization to the early endosomes in YT and HEK 293 cells and the work of others in various distinct cell lines (MacNeil, Mansour et al. 2007).

SNX27 did not co-localize with either lysosomal markers, Lamp-2 or Lamp-3 (Figure

3.1D-E), nor did it co-localize with the Golgi complex marker, Giantin (Figure 3. IF).

3.4.2 SNX27 POLARIZES TO THE IMMUNOLOGICAL SYNAPSE IN VITRO AND IS DISTINCT FROM THE CYTOTOXIC PERFORIN-CONTAINING GRANULES

NK and T cell-specific delivery of cytotoxic granules to the immunological synapse (IS) for effective destruction of tumor and virus-infected cells is a well- established and essential phenomenon for a properly functioning immune system

(Trambas and Griffiths 2003). Recent work has also shown that other endosome- originating vesicular structures containing essential effectors of the exocytic mechanism must also polarize during this event for effective delivery of the "lethal hit" to occur

(Menager, Menasche et al. 2007). To investigate changes in the distribution of SNX27 during activation of cytotoxic cells, we used NK-92 with K562 target cells in an in vitro killentarget conjugate approach in which cell types were easily distinguished by size or using a cytotoxic-specific anti-perforin marker. Interestingly, endosomal SNX27 was found to polarize to the IS in conjugates (Figure 3.2A). This endosomal polarization was only observed in NK-92 cells and not in the target K562, where endosomes were uniformly distributed (Figure 3.2B).

We have previously characterized an interaction between the adapter protein

CASP and SNX27 in lymphocytes (MacNeil, Mansour et al. 2007). Others have shown that the protein CASP may play a role at the immunological synapse in dendritic cells

(Hofer, Pfeil et al. 2006). We investigated CASP in cytotoxic lymphocytes using the

34 same approach but its localization did not appear to change during conjugation to the target cell (data not shown). However, this result may not be conclusive as effective antibody detection of endogenous CASP in cells is often difficult and problematic as this adaptor protein can be buried in protein complexes.

To investigate whether polarized SNX27 co-localizes with granules containing the pore-forming cytolytic protein perforin, conjugates using NK-92 (Figure 3.3) or YTS (not shown) as killers were analyzed. In both instances, SNX27 endosomes polarized to the IS in the same vicinity of the perforin-containing granules, but were clearly not co-localized

(Figure 3.3A-C). As anticipated, K562 target cells showed no perforin staining and no

SNX27 polarization to the IS in conjugates. In unconjugated resting killer cells, SNX27 and perforin were not polarized and again did not co-localize in NK-92 (Figure 3.3D-F) or YTS cells (data not shown). Altogether, these results demonstrate a cell-type restricted polarization of endosomal SNX27, distinct from the cytotoxic granules, to the IS in cell- mediated cytotoxicity.

3.4.3 SNX27 POLARIZES TO THE LEADING EDGE OF MIGRATING NK-92 CELLS

Recently, a function in lymphocyte migration for the SNX27-interacting protein

CASP has been described using several in vivo models based on CASP knockout mice

(Coppola, Barrick et al. 2006). To investigate whether the observed polarization of endosomal SNX27 was also a feature of migrating cells, the marker CD 18 was employed to identify the uropod structure of cells undergoing migration. Endosomal SNX27 was again polarized, in this instance to the leading edge of migrating NK-92 cells, opposite the uropod (Figure 3.4A-C). In cells not displaying a uropod structure, SNX27 was found

35 to be unpolarized (Figure 3.4D-F). The same results were observed using an antibody for

CDlla at the uropod (data not shown). We have previously shown that CASP and

SNX27 co-localize in resting lymphocytes (MacNeil, Mansour et al. 2007); however, endogenous CASP was again difficult to detect in migrating cells. It is clear that these proteins are interacting preceding polarization, while interaction during polarization is unclear using this approach.

Taken together, these data support a role for the active polarization of endosomal

SNX27 during lymphocyte migration and during formation of the immunological synapse. This, with the previously reported loss of lymphocyte trafficking in CASP knockout mice (Coppola, Barrick et al. 2006), suggests the SNX27-CASP interaction may play a role in lymphocyte polarization.

36 Figure 3.1 Endogenous SNX27 co-localizes with early endosomes and not with lysosomal or Golgi markers in NK-92 cells. SNX27 co-localized with EEA1 (A-C). No co-localization was found with lysosomal markers Lamp-2 (D) and Lamp-3 (E) or with the Golgi marker Giantin (F).

37 Target

Z plane: +4|jm

Figure 3.2 Immunocytochemistry of killer:target conjugates shows SNX27 polarization to the IS in vitro. (A-B) Polarization of SNX27 endosomes to the IS in NK-92 cells during conjugation with the target cell K562. Polarization of two NK-92 cells is shown in A. Cell types are labeled in B (NK: NK-92; Target: K562) where the same conjugate in shown at an optical section 4um higher in the Z-plane and demonstrating that SNX27 is not polarized in the target cell K562.

38 o

"E* o O Target SNX27 Perforin

Figure 3.3 SNX27 polarization to the IS in conjugates is distinct from the similarly polarized perforin-containing cytotoxic granules in NK-92 cells. (A) SNX27 and perforin polarize to the IS while the K562 target cell (labeled 'Target') does not show SNX27 polarization and does not show perforin staining. (B) SNX27 and perforin show distinct localizations and no polarization in unconjugated NK-92 cells.

39 o (0

SNX27 CD18

Figure 3.4 SNX27 polarizes to the leading edge, opposite the CD 18 enriched uropod of migrating NK-92 cells. (A-C) A migrating NK-92 cell is shown in which SNX27 is polarized opposite the CD 18 marker (uropod). (D-F) A non-migrating (stationary) NK- 92 cell is shown, lacking the bright uropod structure, and SNX27 is found to be non polarized.

40 3.5 DISCUSSION

Cell mediated cytotoxicity involves polarization of lytic granules, unique in NK cells and cytotoxic T cells, to the immunological synapse and subsequent delivery of lytic mediators to the target cell contact site (Trambas and Griffiths 2003). SNX27 endosomes polarize similarly to lytic granules but are completely distinct components (Figure 3.3).

This is consistent with recent studies showing endosomal polarization of effectors of the cytolytic mechanism, hMuncl3-4 and Rab27a, alongside, but distinct from, the lytic granules (Menager, Menasche et al. 2007). Work has shown that these effectors are critical in the delivery of granule contents (perforin and granzymes) to the target cell but apparently only play a role at the very last step in the delivery mechanism (Hong 2005).

This leads to the question of what role, if any, polarized SNX27 might play at the IS? It is thought that while exocytosis of granule contents is taking place, endocytosis in adjacent and central regions of the contact site also plays a significant regulatory role in the development, maintenance, and eventual dissipation of the IS (Krummel and Macara

2006). Receptors, such as the TcR (Lee, Dinner et al. 2003), and other membrane components are recycled or destroyed during target cell engagement creating high turnover in the area in and around the IS. SNX27 could serve as a part of this 'vesicular cycle' as molecular components of the IS are endocytosed and sorted accordingly.

Lymphocyte migration is also a polarized event. Phosphatidylinositol 3-kinase

(PI3K) is localized to the front of a chemotaxing cell where it can create a polarized PIP3 population (Funamoto, Meili et al. 2002; Sasaki and Firtel 2005), recruit PH and PX domain containing proteins, and participate in the onset of motility and lamellipodia formation (Santy, Ravichandran et al. 2005). All members of the SNX family contain a

41 PX domain (Seet and Hong 2006) and PH domains are found in members of the cytohesin/ARNO family of Arf guanine nucleotide- exchange factors (GEFs). We have previously shown that Cytohesin-1 and ARNO interact with CASP (Mansour, Lee et al.

2002). SNX27 and CASP also have PDZ domains and many PDZ domain proteins, including scribble, crumbs3, and par3, are important in establishing asymmetry in migrating CTLs and epithelial cells (Ludford-Menting, Oliaro et al. 2005). In fact, reduced expression of scribble inhibits polarization and morphological changes important for formation of the uropod (Ludford-Menting, Oliaro et al. 2005). We have shown that the PDZ domain protein SNX27 changes its distribution and polarizes to the leading edge during NK cell migration (Figure 3.4). SNX27 is likely either a target of the polarizing machinery or possibly even a mediator of the mechanism directly.

Recently, two mouse CASP knockouts have been produced and described

(Coppola, Barrick et al. 2006; Watford, Li et al. 2006). Although both knockouts show minimal effects in the development of an apparently normal immune system, differences were found in stress conditions. One of the knockouts showed limited immune cell migration to inflammation sites in an aseptic peritonitis model (Coppola, Barrick et al.

2006). Also, reduced lymph node enlargement and larger tumors were found when knockouts were infected with the Moloney murine sarcoma/leukemia virus (Coppola,

Barrick et al. 2006). These results point to a role for CASP in lymphocyte migration and/or tumor cytolysis and are consistent with our data presented here, indicating active polarization of a CASP-interacting protein, SNX27, during migration and tumor cell engagement. CASP was also investigated in migrating and conjugated NK/T cells, but was difficult to detect in activated cells and showed no apparent polarization using this

42 approach. This may indicate that the CASP-SNX27 interaction precedes polarization of endosomes, perhaps as part of a signal transduction event. Alternatively, it is also a possibility that CASP protein participating in polarization is buried in a large complex that blocks detection of the protein with antibodies.

Of late, a small molecule inhibitor of Arf activation by GEF Sec7 domains (e.g.

ARNO) has been developed and described (Viaud, Zeghouf et al. 2007). This inhibitor acts on Arf-1 and ARNO by producing a non-functional complex of these proteins and blocking the downstream effects of Arf-1 activation. Interestingly, this inhibition disrupts

ARNO-dependent cell migration in MDCK cells. As mentioned above, SNX27's interaction partner, CASP, is also a binding partner of the ARNO family of GEF proteins

(Mansour, Lee et al. 2002; Tang, Cheng et al. 2002). ARNO is also found at early endosomes (Hurtado-Lorenzo, Skinner et al. 2006) and plays a role in endocytosis at the apical membrane of polarized epithelial cells suggesting that ARNO is targeted to this membrane domain during polarized cellular events such as migration (Shmuel, Santy et al. 2006). Given the polarized distribution of endosomal SNX27 described here, we hypothesize that SNX27 endosome movement during cell migration could participate in recruitment of ARNO to the apical membrane.

In summary, we have shown that endosomal SNX27 polarizes to the apical membrane in NK cell migration and during formation of the immunological synapse.

These results are consistent with knockout models of the SNX27-interacting protein,

CASP, and also inhibition of ARNO-mediated Arf activation studies on cell migration, and point to new avenues of experimentation in the mechanisms of lymphocyte polarization and migration.

43 3.6 ACKNOWLEDGEMENT

Funding for this research was from the Natural Sciences and Engineering Research

Council of Canada (NSERC).

44 3.7 REFERENCES

Carlton, J., M. Bujny, A. Rutherford and P. Cullen (2005). Sorting nexins-unifying trends and new perspectives. Traffic 6(2): 75-82.

Funamoto, S., R. Meili, S. Lee, L. Parry and R. A. Firtel (2002). Spatial and temporal regulation of 3-phosphoinositides by PI 3-kinase and PTEN mediates chemotaxis. Cell 109(5): 611-23.

Hofer, S., K. Pfeil, H. Niederegger, S. Ebner, V. A. Nguyen, E. Kremmer, M. Auffinger, S. Neyer, C. Furhapter and C. Heufler (2006). Dendritic cells regulate T-cell deattachment through the integrin-interacting protein CYTIP. Blood 107(3): 1003- 9.

Hong, W. (2005). Cytotoxic T lymphocyte exocytosis: bring on the SNAREs! Trends Cell Biol 15(12): 644-50.

Krummel, M. F. and I. Macara (2006). Maintenance and modulation of T cell polarity. Nat Immunol 7(11): 1143-9.

Lee, K. H., A. R. Dinner, C. Tu, G. Campi, S. Raychaudhuri, R. Varma, T. N. Sims, W. R. Burack, H. Wu, J. Wang, O. Kanagawa, M. Markiewicz, P. M. Allen, M. L. Dustin, A. K. Chakraborty and A. S. Shaw (2003). The immunological synapse balances T cell receptor signaling and degradation. Science 302(5648): 1218-22.

Ludford-Menting, M. J., J. Oliaro, F. Sacirbegovic, E. T. Cheah, N. Pedersen, S. J. Thomas, A. Pasam, R. Iazzolino, L. E. Dow, N. J. Waterhouse, A. Murphy, S. Ellis, M. J. Smyth, M. H. Kershaw, P. K. Darcy, P. O. Humbert and S. M. Russell (2005). A network of PDZ-containing proteins regulates T cell polarity and morphology during migration and immunological synapse formation. Immunity 22(6): 737-48.

45 MacNeil, A. J., M. Mansour and B. Pohajdak (2007). Sorting nexin 27 interacts with the Cytohesin associated scaffolding protein (CASP) in lymphocytes. Biochem Biophys Res Commun 359(4): 848-853.

Menager, M. M, G. Menasche, M. Romao, P. Knapnougel, C. H. Ho, M. Garfa, G. Raposo, J. Feldmann, A. Fischer and G. de Saint Basile (2007). Secretory cytotoxic granule maturation and exocytosis require the effector protein hMuncl3-4. Nat Immunol 8(3): 257-67.

Montoya, M. C, D. Sancho, M. Vicente-Manzanares and F. Sanchez-Madrid (2002). Cell adhesion and polarity during immune interactions. Immunol Rev 186: 68-82.

Sanchez-Madrid, F. and M. A. del Pozo (1999). Leukocyte polarization in cell migration and immune interactions. Embo J18(3): 501-11.

Santy, L. C, K. S. Ravichandran and J. E. Casanova (2005). The DOCK180/Elmo complex couples ARNO-mediated Arf6 activation to the downstream activation of Racl. CurrBiol 15(19): 1749-54.

Sasaki, A. T. and R. A. Firtel (2005). Finding the way: directional sensing and cell polarization through Ras signalling. Novartis Found Symp 269: 73-87; discussion 87-91, 223-30.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta 1761(8): 878-96.

46 Trambas, C. M. and G. M. Griffiths (2003). Delivering the kiss of death. Nat Immunol 4(5): 399-403.

Viaud, J., M. Zeghouf, H. Barelli, J. C. Zeeh, A. Padilla, B. Guibert, P. Chardin, C. A. Royer, J. Cherfils and A. Chavanieu (2007). Structure-based discovery of an inhibitor of Arf activation by Sec7 domains through targeting of protein-protein complexes. Proc Natl Acad Sci USA 104(25):10370-10375.

47 3.8 LINKING PARAGRAPH

The following chapter was the product of a series of comparative studies on CASP and a structurally similar protein called GRASP. CASP and GRASP, while similar in domains and motifs, were considered to be isolated to immune and neural tissues respectively. By thoroughly examining all available CASP and GRASP homologues, we determined an interesting putative evolutionary time point where gene duplication had produced this pair of adaptor proteins and carefully examined conserved and novel functionally relevant features across vertebrates. This chapter is published in the Journal of Molecular Evolution.

48 CHAPTER 4

GENE DUPLICATION IN EARLY VERTEBRATES RESULTS IN TISSUE-SPECIFIC

SUBFUNCTIONALIZED ADAPTOR PROTEINS: CASP AND GRASP

Adam J. MacNeil, Lori McEachern, and Bill Pohajdak

Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada

Manuscript submitted April 18, 2008 Accepted June 9, 2008 and available online July 4, 2008 Published in the Journal of Molecular Evolution 67(2): 168-178.

49 4.1 ABSTRACT

CASP and GRASP are small cytoplasmic adaptor proteins that share highly similar protein structures as well as an association with the cytohesin/ARNO family of guanine nucleotide exchange factors (GEFs) within the immune and nervous systems respectively. Each contains an N-terminal PDZ domain, a central coiled-coil motif, and a carboxy terminal PDZ-binding motif (PDZbm). We set out to further characterize the relationship between CASP and GRASP by comparing both their gene structures and functional motifs across several vertebrate organisms. CASP and GRASP not only share significant protein structure but also share remarkably similar gene structure, with six of eight exons of equal length and relative position. We report on the addition of a unique amino acid within the coiled-coil motif of CASP proteins in several species. We also examine the Class I PDZbm, which is highly conserved across all classes of vertebrates, but shows a functionally relevant mutation in the CASP proteins of several species of fish. Further, we determine the evolutionary relationship of these proteins using both phylogenetics and by comparative analysis of the conservation of genes near each locus in various chordates including amphioxus. We conclude that CASP and GRASP are the products of a relatively recent gene duplication event in early vertebrate organisms and that the evolution of the adaptive immune system and complex brain most likely contributed to the apparent subfunctionalization of these proteins into tissue-specific roles.

50 4.2 INTRODUCTION

Adaptor or scaffolding proteins are soluble cytoplasmic molecules that mediate the assembly of larger protein complexes. They generally have no catalytic domains but rather a series of domains and/or motifs that facilitate protein-protein interactions, essentially acting like glue holding a larger collection of proteins together so that they may carry out their required functions. The cytohesin associated scaffolding protein

(CASP, aka Cybr, CYTIP, PSCDBP, B3-1) and GRP-1 associated scaffolding protein

(GRASP, aka tamalin) are two such adaptor proteins. Each is similarly composed of an

N-terminal PDZ domain, a central coiled-coil motif and a carboxy terminal PDZ-binding motif (PDZbm) (Figure 4.1). Our group and others have demonstrated that the coiled-coil motif of each binds members of the Cytohesin/ARNO family and regulates their localization and vesicle trafficking properties (Nevrivy, Peterson et al. 2000; Kitano,

Kimura et al. 2002; Mansour, Lee et al. 2002; Tang, Cheng et al. 2002; Boehm, Hofer et al. 2003). Each also has a very similar PDZbm, and while we have shown that CASP is recruited to endosomes by SNX27, others have reported that GRASP binds S-SCAM and also forms homodimers (via its PDZ domain) through this motif (Kitano, Yamazaki et al.

2003; MacNeil, Mansour et al. 2007; Sugi, Oyama et al. 2007). Finally, the N-terminal

PDZ domain of GRASP has been shown to interact with several distinct neuronal proteins including mGluRs, SAPAP3, and TrkC while the PDZ domain of CASP remains uncharacterized (Kitano, Kimura et al. 2002; Kitano, Yamazaki et al. 2003; Kimura,

Kitano et al. 2004; Esteban, Yoon et al. 2006). While there are similarities in structure, function, and interaction partners for these two adaptor proteins, their expression pattern is quite different. GRASP is considered a neuronal specific protein, supported by

51 northern analysis and its PDZ domain binding partners, while CASP is exclusive to cells of the hematopoeitic lineage and is highly expressed in cells of the immune system

(Nevrivy, Peterson et al. 2000; Kitano, Kimura et al. 2002; Tang, Cheng et al. 2002;

Watford, Li et al. 2006). In view of these reports, it appears as though CASP and

GRASP are adaptor proteins with several similar properties that may function in assembly of protein complexes tailored to either the immune or nervous systems respectively.

Protein or gene families consist of evolutionarily related members that have diverged in some capacity to perform particular functions that require unique binding and/or localization of the protein (eg. integrins, reticulons, argonautes). The means by which these families accumulate new members and grow is through gene duplication events followed by divergence. Gene duplication leads to opportunities for the development of novelties in patterns of expression and molecular pathways. Upon duplication, unless an immediate advantage is provided, divergence will commence and mutations slowly accumulate in redundant regions of one, or possibly both, duplicates.

Novelty arises when a new function is realized or both genes become essential to perform the original function in processes called neo-functionalization and sub-functionalization respectively (Lynch and Conery 2000). However, the most common outcome of this process is non-functionalization, wherein one duplicate eventually degenerates into a pseudogene (Lynch and Conery 2000).

Interestingly, the classical model describing the predicted outcome for gene duplicates has been thought to underestimate the number of duplicated genes that are actually being retained (Nowak, Boerlijst et al. 1997; Lynch and Force 2000). To explain this, an alternative model referred to as the Duplication-Degeneration-Complementation

52 (DDC) model was described whereby both copies are retained through complementary loss of independent subfunctions resulting in both copies being required to perform the ancestral function (Force, Lynch et al. 1999). This model of subfunctionalization, based on the modular nature of protein function and gene structure (including enhancers and silencers), supports the surprisingly high number of duplicated genes that have been retained. To date, every eukaryotic genome that has been sequenced contains duplicated genes. Some gene families are very old and very large containing numerous members with conserved homologues present in distantly related species. Here we describe and provide evidence in support of a very small and relatively new gene and/or protein family consisting of two vertebrate adaptor proteins, CASP and GRASP. We outline the conserved structural similarities at the protein and gene level as well as compare functional protein domains and motifs across vertebrate species. We also characterize the gene duplication event that produced this small family in early vertebrates by providing a phylogenetic analysis and examining the genetic loci of each in various chordate organisms. Finally, we describe the possible evolutionary forces of purifying selection that may have contributed to subfunctionalization of these two specialized proteins in aspects of the immune and nervous systems.

4.3 MATERIALS AND METHODS

4.3.1 IDENTIFICATION AND RETRIEVAL OF CASP AND GRASP HOMOLOGUES

Protein or corresponding gene sequences were retrieved from the NCBI Entrez

Protein database by BlastP (protein-protein) searching with various CASP or GRASP sequences to create a comprehensive list of related proteins. Sequences obtained were:

53 human (Homo sapiens) CASP NPJ304279 and GRASP NP_859062, chimpanzee (Pan troglodytes) CASP XP_515844 and GRASP XP_001141173, monkey (Macaca mulatto)

CASP XPOO1088144 and GRASP XP_001085174, dog (Canis familiaris) CASP

XP_545482 and GRASP XP_850335, cow (Bos taurus) CASP NP_001095713 and

GRASP XP_583750, mouse (Mus musculus) CASP NP631939 and GRASP

NP_062391, rat (Rattus norvegicus) CASP NP_001012086 and GRASP NP_620249, bird

(Gallus gallus) XP_426609, horse (Equus caballus) XP001491328, zebrafish (Danio rerio) NP_001035013 and XP_001338854, frog (Xenopus laevis) GRASP AAH42351, and tetraodon fish (Tetraodon nigroviridis) CAG09325.

Reptile (Anolis carolinensis) predicted proteins were retrieved by searching with highly conserved regions using the program BLAT (Kent 2002) from the UCSC Genome

Bioinfomatics site (http://genome.ucsc.edu/): CASP scaffold_294:518717-544354 and

GRASP scaffold_42:647354-661853. BLAT was also used to obtain sequences from frog

(Xenopus tropicalis) CASP scaffold_51:2726251-273 8128 and GRASP scaffold_130:726861-748128, and partial sequences from fugu pufferfish (Fugu rubripes)

-1 chrUn: 128190000-128192603 and -2 chrUn:106622086-106623619. Shark

(Callorhinchus milii) partial sequences were obtained from IMCB (http://esharkgenome. imcb.a-star.edu.sg): CASP scaffold AAVX01314667.1 and GRASP scaffold

AAVX01284196.1.

4.3.2 ANALYSIS OF CASP AND GRASP GENE STRUCTURES AND PROTEIN ALIGNMENTS

Gene information was retrieved from NCBI Entrez Nucleotide, where exon sizes are reported and intron sizes were deduced by comparison to the genomic sequence.

54 Accession numbers for human transcripts are: CASP NM004288 and GRASP

NM181711. Gene IDs are CASP 9595 and GRASP 160622. Determination of the protein domain/motif coded for by each exon was resolved by comparison to the corresponding protein sequence above. All protein alignments were prepared using the

ClustalW program. Analysis of the unique CASP codon was completed by examining the corresponding nucleotide sequence using the program Gene Runner.

4.3.3 ALIGNMENTS AND PHYLOGENETIC ANALYSIS OF GRASP AND CASP HOMOLOGUES

Vertebrate CASP and GRASP homologues as well as two invertebrate (B. floridae and C. elegans) homologues were aligned using the ClustalW program. The resultant alignment (Phylip format) was adjusted manually using BioEdit (Ibis

Biosciences) with minimal gap introduction in regions of poor alignment and finally used to infer the phylogentic relationship of these proteins using the maximum likelihood program RAxML (Stamatakis, Ludwig et al. 2005). Trees were subjected to 100 rounds of bootstrapping with C. elegans as an outgroup. C. elegans was chosen as an outgroup because its putative CASP/GRASP homologue was found to show the greatest similarity to chordate CASP/GRASPs, in both sequence and domain/motif structure, in comparison to other putative protostome homologues. Trees were subsequently viewed using the

TreeView program (Page 1996). All protein sequences used in the analysis are listed above except the amphioxus homologue which was from the JGI Branchiostoma floridae vl.O database: Protein ID 86734, and the C. elegans homologue which is from NCBI

Entrez Protein: NP492164.1. Partial or incomplete sequences were not used in the analysis.

55 4.3.4 ANALYSIS OF AMPHIOXUS, ZEBRAFISH, AND HUMAN GENOMIC LOCI

Amphioxus gene data were retrieved from the JGI Branchiostoma floridae vl.O database. (http://genome.jgi-psf.org/Amphioxus). The GRASP/CASP homologue is located at position 1,290,516 of scaffold 154 (protein ID = 86734). The CD63 and

LETMD1 homologues are located on scaffold 154 at positions 987,129 (protein ID =

124708) and 1,116,476 (protein ID = 124710) respectively. The amphioxus genome currently contains two highly similar scaffolds with NR4A and GPD2 homologues.

Several additional genes are repeated between both scaffolds, suggesting either a sequence assembly error, or a recent duplication of a portion of the genome. On both scaffolds the NR4A and GPD2 homologues are immediately adjacent to one another. On scaffold 97, the GPD2 homologue is at position 550,434 (protein ID = 281135) and the

NR4A homologue is at position 573,506 (protein ID = 223708). On scaffold 128, the

GPD2 homologue is at position 192,057 (protein ID = 123751) and the NR4A2 homologue is at position 180,397 (protein ID = 84128).

Zebrafish gene data were obtained on the NCBI Entrez Gene database

(http://www.ncbi.nlm.nih.gov/sites/entrez?db=gene). On zebrafish chromosome 23, the

Gene IDs are as follows: GRASP: 558950, NR4A1: 431720, CD63: 321461, and

LETMD1: 567014. CASP is located adjacent to the zebrafish GALNT5 homologue on chromosome 5, with the following Gene IDs: CASP: 798407 and GALNT5: 798489. On chromosome 6, a gene with similarity to CASP is located adjacent to GPD2 and NR4A2.

The Gene IDs are: CASP-like: 792667, GPD2: 751628, andNR4A2: 436679.

56 Human gene data were also retrieved from the NCBI Entrez Gene database. In humans, CASP and surrounding genes are located on chromosome 2. The Gene IDs are as follows: CASP: 9595, GALNT5: 11227, GPD2: 2820, and NR4A2: 4929. GRASP and surrounding genes are located on chromosome 12, with the following Gene IDs: GRASP:

160622, NR4A1: 3164, CD63: 967, andLETMDl: 25875.

4.4 RESULTS

4.4.1 CASP AND GRASP SHARE HIGHLY SIMILAR FUNCTIONAL DOMAINS, MOTIFS, AND GENOMIC STRUCTURE

The degree of structural similarity between human CASP and GRASP proteins is most apparent when considering the variety and relative distribution of functional domains and motifs in these proteins. While both are similar in size, differing by 36 amino acids (or approximately 10%) in the longer GRASP, they share an N-terminal PDZ domain, a central coiled-coil motif, and a carboxy-terminal PDZbm (Figure 4.1). These domains and motifs typically facilitate the assembly of larger protein complexes, a characteristic of adaptor or scaffolding proteins.

Analysis of the genomic regions coding for human CASP and GRASP reveals a further layer of similarity. At first glance, the genes appear somewhat similar in exon/intron distribution with six smaller exons (49-102 bp, Table 1) arranged between a somewhat larger first exon and a final exon of more than 1 kb, but CASP is spread over some 30kb while GRASP is less than a third its size at nearly 9kb (Figure 4.2A).

However, examination of exon sizes shows that of those internal six, five are identical in size and relative position with one other only differing by the addition of three bases or one codon (Table 1 and Figure 4.2A). While the size of their respective introns has

57 diverged over time, many critical exons have remained completely conserved. Not surprisingly, these highly conserved exons specifically code for the PDZ domain and coiled-coil motif of both CASP and GRASP, regions that have been shown to be critical in mediating protein-protein interactions (Table 1). Altogether, the similarities observed between CASP and GRASP both at the protein level and also at the genomic level point to a potential common ancestry for these two proteins.

4.4.2 MAMMAL, BIRD, AND REPTILE CASP SHARE A UNIQUE AND RECENTLY ACQUIRED AMINO ACID IN THE COILED-COIL MOTIF

Given the observed structural and sequence similarities, we further examined

CASP and GRASP proteins by assembling a record of all available CASPs, GRASPs, and potentially related proteins from various genomic databases. To assess the diversity and conservation within the group, alignments were produced using the ClustalW program.

As expected, the most highly conserved regions were found in the characterized functional domains and motifs, with less similarity in the flanking regions.

We also aimed to determine if the additional codon or amino acid encoded by exon six of human CASP (mentioned above) was a signature feature of CASP proteins.

In fact, when examining an alignment of the region containing this amino acid, it became clear that there was a trend in its occurrence. Of the vertebrate classes, only mammalian

(R), bird (K) and reptile (Q) CASP proteins contained this extra amino acid (Figure 4.2B).

These three classes of vertebrates (amniotes) form a monophyletic group and share a unique common evolutionary ancestor. This ancestral animote is the most likely origin of this additional amino acid. Given that all other vertebrate GRASP/CASPs (including

58 fishes and amphibians) lacked this feature, it was clear that this was an addition to CASP proteins at the divergence of amniotes.

Although only hypothetical, if you assume reptiles have retained the ancestral codon, it is interesting that a single base change can accommodate the changes observed in amino acids for birds and mammals (Figure 4.2C). An additional redundant base change reflects the codons currently found in birds and mammals. The actual codon of the ancestral organism could have been any of those in figure 4.2C, or possibly even another codon altogether.

Interestingly, the CASP coiled-coil is considered to be longer than that of

GRASP. The addition of an amino acid within the highly conserved and functionally relevant coiled-coil motif of CASP may play a unique role in the structure and function of the CASP coiled-coil as well as in the evolutionary development of a longer coiled-coil motif.

4.4.3 THE PDZBM OF VERTEBRATE CASPS AND GRASPS SHOW CONSIDERABLE CONSERVATION

Alignment of the carboxy-terminal region of vertebrate CASP and GRASP proteins reveals a remarkable degree of conservation in another functionally relevant protein-interacting motif (Figure 4.3). This region of human CASP and GRASP is characterized as a PDZ-binding motif (PDZbm) and can interact with competent PDZ domains by insertion of the C-terminal motif into the well-characterized carboxylate- binding loop or pocket of a particular PDZ domain (Nourry, Grant et al. 2003; Jemth and

Gianni 2007). Our group has showed this previously for the PDZbm of CASP with the

PDZ domain of SNX27 (MacNeil, Mansour et al. 2007) and others have documented this

59 for GRASP and TrkC among others (Kitano, Yamazaki et al. 2003; Esteban, Yoon et al.

2006; Sugi, Oyama et al. 2007). Mammalian CASPs and GRASPs have a particularly notable degree of conservation in the PDZbm reaching as far back as 15 amino acids of almost completely conserved residues (Figure 4.3). Even if the entire vertebrate panel is considered, the degree of conservation is considerable with several residues 100% identically conserved amongst many conserved substitutions (Figure 4.3). Although the final 3-4 amino acids of PDZbms are most often discussed and used to variously group these motifs into four classes (I, II, III, IV), it is well known that additional residues near the carboxy-terminus are also important for interactions with PDZ domains (Pegan, Tan et al. 2007). The PDZbm of vertebrate CASPs and GRASPs conform very well to the

Class I consensus sequence of -S/T-X-0 (Hung and Sheng 2002; Jelen, Oleksy et al.

2003; Nourry, Grant et al. 2003), with the terminal hydrophobic residue only varying between three amino acids [L/F/V] over twenty-three proteins. This indicates that this motif has not only retained a similar sequence and structure, but that functionally its activity and binding partners are likely also very well conserved throughout vertebrate organisms. Interestingly, some fish proteins have lost the final few amino acids of the

PDZbm (Figure 4.3). This change would have functional consequences for specific protein interactions, as experimental evaluations have shown that this type of mutation impairs the binding ability of PDZbms (Joubert, Hanson et al. 2004; Malmberg,

Andersson et al. 2004; MacNeil, Mansour et al. 2007), and may represent a molecular difference between fishes and other vertebrate organisms in subcellular functioning.

60 4.4.4 PHYLOGENETIC ANALYSIS IDENTIFIES A CASP/GRASP GENE DUPLICATION PREVIOUS TO THE VERTEBRATE RADIATION

To further investigate the similarities and apparent relatedness of CASP and

GRASP proteins, a phylogenetic analysis was conducted on vertebrate proteins and invertebrate homologues for which full sequences were available. This analysis was conducted on ClustalW alignments using the maximum likelihood program RAxML in order to visualize the inferred evolutionary relationship between these proteins

(Stamatakis, Ludwig et al. 2005). The resultant tree was well supported statistically and robust with relatively high bootstrap values (Figure 4.4). Interestingly, the tree revealed a paralogous relationship between CASP and GRASP proteins with each set of proteins separating into distinct clades with the two invertebrate proteins grouping outside (Figure

4.4). The node indicated by an asterisk represents the inferred point of gene duplication for an ancestral gene, giving rise to a small family of paralogous adaptor proteins. This node is extremely well supported with a bootstrap value of 99 (Figure 4.4). Amphioxus

{Branchiostoma) is a chordate organism and is considered to be one of the most closely related extant invertebrate organisms to modern vertebrates. Other closely related chordates examined, including Ciona intestinalis (sea squirt) and Strongylocentrotus purpuratus (sea urchin), lacked a CASP/GRASP homologue possibly due to gene loss or incomplete sequencing. Based on the inferred tree, the duplication event most likely took place in an organism of early vertebrate ancestry (Figure 4.4).

It is interesting that the two zebrafish proteins included in the analysis grouped in well supported distinct clades because, as it is well known, teleost fish have undergone genome duplication events that can confound the relatedness of protein families. It would seem that the CASP/GRASP gene duplication most likely occurred prior to genome

61 duplication. Notably, there is another partially homologous protein to fish CASP that is likely representative of a divergent gene following genome duplication (discussed below).

Also of interest is that the sole bird species {Gallus) examined currently has only one member of this protein family (unique amongst vertebrates), and based on the inferred tree and also on the analysis of the extra codon in exon six, this protein appears to be a

CASP. However, the Gallus genome still contains gaps and unassigned genomic scaffolds, including a portion the genomic region where GRASP might be expected to be found, and it remains possible that the region containing the GRASP gene has not yet been sequenced.

4.4.5 EXAMINATION OF GENETIC LOCI IN VARIOUS CHORDATES SHOWS THAT GRASP is THE MOST LIKELY ANCESTRAL GENE AND THAT THE ANCESTRAL NR4A GENE WAS CO-DUPLICATED

An examination of the genes adjacent to CASP and GRASP in zebrafish (£). rerio), frog (X. tropicalis), bird (G. gallus), and humans (H. sapiens) quickly revealed several similarities conserved across these species. In all species examined, GRASP was located near the nuclear receptor NR4A1 (aka NGFI-B) gene, while CASP was located in close proximity to the closely related nuclear receptor NR4A2 (aka Nurrl). NR4A1 and

NR4A2 are members of the NR4A family of orphan nuclear receptors, which also contains a third member, NR4A3. The close association of CASP and GRASP with

NR4A2 and NR4A1 respectively, suggests that the ancestral CASP/GRASP gene was adjacent to an NR4A homologue, and that the two genes were duplicated together.

Notably, Gallus gallus, which does not appear to have GRASP according to the current gene assembly, is also lacking an NR4A1 gene.

62 In addition to NR4A2, CASP is also consistently located in close proximity to

GPD2 and GalNT5. Interestingly, in zebrafish, which has undergone a genome duplication, the CASP homologue is located adjacent to GalNT5 on chromosome 5, while a similar CASP-like gene is retained on chromosome 6 adjacent to GPD2 and NR4A2.

GRASP is located in close proximity to CD63 and LETMD1, as well as NR4A1, on zebrafish chromosome 23, and these genes are also located together within a larger region on human chromosome 12 (Figure 4.5).

In amphioxus, the ancestral GRASP/CASP gene is located on scaffold 154, within a 300 kb region that also contains the amphioxus homologues of CD63 and LETMD1.

CD63 and LETMD1 are located in very close proximity to GRASP in zebrafish, and on the same chromosome as GRASP in humans, indicating that following gene duplication, the GRASP locus has retained similarity to the ancestral gene arrangement. The amphioxus assembly has two highly similar scaffolds containing an NR4A homologue.

Within these two scaffolds a high number of genes are repeated, suggesting either a sequence assembly error or a recent duplication in amphioxus. It seems unlikely that this duplication was present in the common ancestor of amphioxus and vertebrates, as this would mean that amphioxus have retained two very highly similar genome regions, containing a large number of the same genes, throughout their evolution. On both scaffolds, the NR4A homologue is located immediately adjacent to the GPD2 gene, which is closely associated with CASP in the vertebrate lineage. As NR4A was likely duplicated with the GRASP/CASP homologue, we hypothesize that these two gene regions were located close together prior to duplication (hypothetical intermediate, Figure

4.5). Gene duplication then produced two copies of each of the GRASP/CASP and

63 NR4A homologues, which subsequently evolved into GRASP, CASP, NR4A1, and

NR4A2. Interestingly, homologues of genes that are adjacent to NR4A3 in zebrafish and humans (e.g. syntaxin 17 and thioredoxin domain containing 4) are also located near

NR4A and GPD2 in amphioxus (data not shown). Thus it appears that following the duplication with GRASP/CASP, one of the nuclear receptors duplicated a second time to produce the third family member. This is consistent with a phylogenetic tree of the nuclear receptor family, which demonstrates a recent split between the NR4A2 and

NR4A3 subgroups, with a more ancestral connection to the NR4A1 subgroup (Kapsimali,

Bourratera/. 2001).

64 CASP CC . laiHB 359 GRASP CC ^H 395

Figure 4.1 A comparison of the domain distribution of human CASP and GRASP proteins. As depicted, both proteins show significant similarity in domain/motif composition and relative position. Both have class I (S/T-X-O) PDZbms, where X represents any amino acid and O represents a hydrophobic residue. Domains/motifs are not drawn to scale.

65 Figure 4.2 Comparison of CASP and GRASP gene structures. A. Human CASP and GRASP exon/intron gene structures and relative sizes are shown with solid lines connecting exons of identical length and relative position in both genes. Exon six (boxed) differs by the addition of only three bases (one codon) in CASP. Exons and introns are drawn to scale. B. Alignment of vertebrate CASP and GRASP protein sequences in a region coding for the coiled-coil motif (indicated amino acid position specific to human CASP). Highlighted is the signature additional amino acid present in all mammalian, avian, and reptilian CASP sequences corresponding to the boxed exon six outlined in A. Due to incomplete sequencing or partial sequence availability, some species (e.g. frog-X laevis; horse) are only represented by one protein in this alignment. C. A hypothesized evolutionary course of base changes in the additional CASP codon and corresponding amino acid of exon six (outlined above), assuming the reptile codon (or amino acid) has remained unchanged since the common ancestor. Of interest is that one base change in each case can accommodate the observed change in amino acids, while one additional redundant change in the codon reflects the current sequence.

66 CASP

1 kb

B 148 182 Human-CASP VDLIRSSGNLLTIETLNGTMILKGTELEAKLQVLK Chimpanzee-CASP VDLIRSSGNLLTIETLNGTMILKGTELEAKLQVLK Monkey-CASP VDLIRSSGNLLTIETLKGTMILKBTELEAKLQVLK Horse-CASP VDLIRSSGNLLTIETLNGTMILKBTELEAKLQVLK Cow-CASP VDLIRSSGNLLTIETLNGTMILKSTELEAKLQVLK Dog-CASP VDLIRSSGNLLTIETVNGTMILKSAELEAKLQVLK Mouse-CASP VDLIRSSGNLLTIETLNGTMIHRBAELEAKLQTLK Rat-CASP VDLIRSSGNLLTIETLNGTMIHRBAELEAKLQTLK Bird-CASP VDLIKSSGNXLKLETVNGALFLRWJELETKLQLLK Reptile-CASP VDMFKLSRNLL|LDWNEALI IKG^IELETKLQLLK Monkey-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Dog-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Rat-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Mouse-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK COW-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Chimpanzee-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Human-GRASP VDIIKASGNVLRLETLYGTSIRK-AELEARLQYLK Reptile-GRASP VEIIKASGNVLRLETLYGTSIRR-AELEARLQYLK Prog-laevis VEMIKASGNTIRLETVYGSAIRR-AELEARIQYLK Frog-GRASP VELIKASGNTIRLETVYGSAIRR-AELEARIQYLK Fugu-1 VQLIRACGNTVRLETVYSDSIRK-AELEARLSYLK Tetradon-2 VQLIRACGNTVRLETVYSDSIRK-AELEARLSYLK Zebrafish-1 VQLIKSSGNNIRLETVYSDSIRK-AELEARLQYLK Frog-CASP VDLIRASGNYLRIEAVNGTKIRK- SELEAKLLFLK Tetradon-1 LDLVRKSDNSIIMMETVCGSKVKQ- IELEKRMSVLK Fugu-2 LDLIKESTNSLMMETVSGSKVKQ-IELEKRMSVLK Zebrafish-2 IELIRESSNTLKLETVSGSVMKR-IELEKKMHYLK

Class amino acid codon Bird K A A t6 K tAA A Reptile Q C A A 1 R C G A

Mammal R A 6 A

Figure 4.2 Comparison of CASP and GRASP gene structures.

67 PDZbm concensus Class I

Human-GRASP FIPGLNRSLEEEE Chimpanzee-GRASP FIPGVNRSLEEEE Mouse-GRASP FIPGLNRSLEEEEE Rat-GRASP FIPGLNRSLEEEE: Monkey-GRASP FIPGLNRSLEEEE: Cow-GRASP FIPGLNRSLEEEE* Frog-GRASP FIPGLNRNLEEE Frog.laevis FIPGLNRNLEEE Dog-CASP FIPGLHRAVEEEEE Horse-CASP FIPGLHRAVEEEE: Rat-CASP FIPGLHRAVEEEE: Mouse-CASP FIPGLHRAVEEEE M Monkey-CASP FIPGLHRAVEEEE jj Chimpanzee-CASP FIPGLHRAVEEEE:S Human-CASP FIPGLHRAVEEEE p Cow-CASP FIPGLHHAVEEEEj Shark-CASP FIPGLNRAVEEDEg Bird-CASP FIPGLHRALEEEEg Reptile-CASP FIPGLYRAVEEEE* Frog-CASP XIPGLHRSVEEEI Shark-GRASP FIPGLNRPVEEEEg Zebrafish-1 YIPGLNRPLEEEE Fugu-1 KFPGLNQPMEEEE Zebrafish-2 FIPGLNHSVEEEECS- Tetradon-1 LLPGLQRSVEEEE Fugu-2 FLPGLQRSVEEE

Figure 4.3 Alignment of vertebrate CASP and GRASP protein sequences in the PDZ- binding motif region. This region is highly conserved across vertebrate species and those sequences containing the characteristic carboxy-terminal three amino acids are highlighted. These PDZbms conform well to the Class I sequence of -S/T-X-O indicating that this motif has retained a common functional binding affinity throughout the vertebrate species. Specifically, the terminal hydrophobic amino acid of the PDZbm only varies between L, F, and V amongst all vertebrates. Interestingly, some fish proteins have lost portions of the PBZbm indicating a functional difference in these proteins. Due to incomplete sequencing or partial sequence availability, some species (e.g. frog-X laevis; horse) are only represented by one protein in this alignment. * = identical; : = conserved substitution

68 Figure 4.4 Phylogenetic analysis of vertebrate CASP and GRASP proteins including invertebrate homologues. Sequences were aligned using the ClustalW program. A phylogenetic tree was constructed from this alignment using the RAxML maximum likelihood program and visualized using TreeView. Branch labels indicate the robustness of the tree over 100 bootstrap replicates. Protein sequences used in the analysis corresponded to completely sequenced regions of high similarity over 366 amino acid positions. Species with partial or incomplete sequences were not included in the analysis. Branch lengths indicate the number of amino acid substitutions per site. The * corresponds to the predicted point of a gene duplication event in early vertebrates leading to distinct paralogous CASP (dark box) and GRASP (light box) clades. Notably, the bird species {Gallus gallus) does not currently have a corresponding GRASP protein, and are only represented in the CASP clade.

69 C.elegans

Amphioxus

Zebrafish-2

Frog CASP

Bird-CASP i—4 .69

Mouse CASP 100

Rat-CASP

100 - MonKey-CASP

t30 Human-CASp

I Chimpanzee CASP

• Zebrafish-1

Frog-GRASP 91

Rat-GRASP 72 96 Mouse-GRASP

100 Monkey-GRASP

Human-GRASP

100 Chimpanzee-GRASP

Figure 4.4 Phylogenetic analysis of vertebrate CASP and GRASP proteins including invertebrate homologues.

70 Figure 4.5 Comparison of gene loci from amphioxus, zebrafish, and human genomes. A representative model of the predicted sequence of evolutionary events linking these loci and the most likely common origin of CASP and GRASP is shown. Analysis of the genomic region around the GRASP/CASP loci shows similarity in several of the adjacent genes across all species. A hypothetical intermediate between amphioxus and zebrafish illustrates that the GRASP/CASP gene region most likely came together with the NR4A gene region prior to duplication. A gene duplication event led to distinct CASP and GRASP genes in early vertebrates with some of the adjacent genes being retained near respective CASP and GRASP loci. Interestingly, NR4A (nuclear receptor 4A) was likely duplicated along with GRASP/CASP resulting in NR4A1 and NR4A2 adjacent to GRASP and CASP respectively. A genome duplication in teleost fish led to a divergent CASP-like gene still near NR4A2 while the GRASP duplicate was lost. GRASP is still located near CD63 and LETMD1 in vertebrates, as in amphioxus, suggesting that the GRASP gene region most closely reflects the ancestral gene locus. The amphioxus homologue of CASP and GRASP is located on scaffold 154 of the present genome assembly (JGI).

71 Amphioxus (Branchiostoma floridae) scaffold 154 scaffold 97/128

CD63 LETMD1 pra-GRASP/CASP NR4A GPD2 M 313 kb-

hypothetical intermediate

CD63 LETMD1 pre-GRASP/CASP NR4A GPD2

Gene Duplication Zebrafish (Danio rerio) chrom 23 chrom 5

CD63 LETMD1 NR4A1 GRASP CASP GALNT5 Genome Duplication (Gene Loss) (Gene Divergence) chrom 6

•-=> —- CASP-like GPD2 NR4A2

Human (Homo sapiens) chrom 12 chrom 2

LETMD1 GRASP NR4A1 CD63 NR4A2 GPD2 GALNT5 CASP

Figure 4.5 Comparison of gene loci from amphioxus, zebrafish, and human genomes.

72 primordial condition pre-vertebrate • D NR4A GRASP

Gene Duplication vertebrate radiation • D complex brain /v adaptive immunity subfunctionalization NR4A1 GRASP NR4A2 CASP

Figure 4.6 A simplified model depicting the sequence of events and relative evolutionary time line (left) for duplication of the primordial GRASP and NR4A locus. Interestingly, the gene duplication occurred near the same time point as the evolution of both a complex brain and adaptive immunity in vertebrates, perhaps supplying the stimulus for these two adaptor proteins to specialize in brain function (GRASP) and adaptive immunity (CASP) as their spatial subfunctionalization implies today.

73 4.5 DISCUSSION

Gene duplication is a significant contributor in generating molecular novelty by providing new materials on which selective forces can act and drive molecular evolution

(Holland, Garcia-Fernandez et al. 1994; Sidow 1996). Protein or gene families are spawned from an original ancestral gene through duplication events and subsequent divergence (Holland, Garcia-Fernandez et al. 1994). Whether through tandem, chromosomal, or entire genome duplications, identical clones of genes are created that subsequently diverge from the original under one of several evolutionary scenarios

(Lynch and Conery 2000). This divergence can happen at somewhat varied rates depending on the selective forces applied on an organism at a given time, but generally a fate for the duplicated gene is thought to be reached within several million years (Lynch and Conery 2000).

Many gene families code for proteins of highly similar structure with each individual protein serving in a comparable, but slightly specialized physiological role

(e.g. Hox, Wnt). Three outcomes are generally recognized as possibilities following gene duplication: i. non-functionalization: pseudogenization followed by the death of one copy; ii. neo-functionalization: one copy acquires a new function; and iii. subfunctionalization: both copies are required to perform the full ancestral function

(Lynch and Conery 2000). While the first is by far the most common outcome, the latter two are strong driving forces in evolution, producing genes with new and/or specialized functions (Lynch and Conery 2000).

Inherently, evolution leaves clues as to the common origin of highly similar genes.

This appears to be the case for CASP and GRASP, as both show a significant degree of

74 conservation in the most functionally critical exons, those coding for protein-interacting motifs. The extent of CASP and GRASP gene similarity is well demonstrated by the fact that three quarters of their exons are identical in size and position in each gene (Figure

4.2A). Even the ancestral gene in the most distant chordate relative to the mammalian genes has five of eight of these exons, and they have also remained of identical length and relative position (data not shown). This measure of similarity does not occur randomly and strongly supports a gene duplication event producing these paralogues. This, together with the presence of similar domains and motifs, including a highly conserved Class I

PDZbm, suggests common functionality along with common ancestry. Indeed, it is well known that both CASP and GRASP interact with members of the cytohesin/ARNO family of Arf GEFs, each via the coiled-coil motif. This associates both proteins in common cellular pathways related to intracellular trafficking and signalling functions

(Kitano, Kimura et al. 2002; Mansour, Lee et al. 2002; Tang, Cheng et al. 2002). Given the similarity in PDZbm structure, more congruent binding partners are a strong possibility.

In the case of CASP and GRASP, gene duplication and divergence appears to have resulted in subfunctionalization of these two adaptor proteins. While both appear to function in aspects of intracellular trafficking/signalling pathways, CASP is most highly expressed in the immune system while GRASP in found most highly expressed in the nervous system (Nevrivy, Peterson et al. 2000; Kitano, Kimura et al. 2002; Mansour, Lee et al. 2002; Tang, Cheng et al. 2002; Coppola, Barrick et al. 2006; Watford, Li et al.

2006; MacNeil, Mansour et al. 2007; Ogawa, Miyakawa et al. 2007). This is a classic example of subfunctionalization in which an ancestral gene gives rise to two copies that

75 eventually acquire specialized roles in a specific set of tissues. It is likely that the ancestral gene (perhaps similar to the amphioxus gene today) performed all of the required functions of CASP and GRASP combined. Often, the segregation of duplicate proteins over time occurs through complementary loss of spatial domains of expression, such as non-coding regulatory elements, resulting in genes that are optimized specialists in a particular tissue (Force, Lynch et al. 1999). This typically involves the loss of enhancers and/or silencers that drive expression following induction of relevant signalling pathways (Force, Lynch et al. 1999). In line with its expression profile, GRASP has several neuronal specific binding partners for the PDZ domain while interacting proteins for the PDZ domain of CASP have yet to be elucidated (Kitano, Kimura et al. 2002;

Kitano, Yamazaki et al. 2003). An educated guess might suggest them to be immune specific.

The PDZbm of vertebrate CASPs and GRASPs shows a considerable degree of conservation as well. In fact, most vertebrate species examined showed a class I PDZbm

(Figure 4.3) indicating a common degree of binding affinity and functional ability has been retained in this motif. It is interesting to note, however, that all three fish species examined have lost several of the terminal amino acids of the PDZbm of one protein.

This form of mutation would render the PDZbm non-functional, hindering protein interactions. Based on our phylogentetic analysis of these fish proteins, versions lacking the PDZbm are most likely all CASP-like proteins. We have previously characterized the

PDZbm of human CASP and its interaction with, and recruitment by, SNX27 in lymphocytes (MacNeil, Mansour et al. 2007). This subtle mutation in fish CASP proteins indicates that a functionally relevant protein-protein interaction has likely been altered in

76 fish, potentially demonstrating a very specific molecular difference in cellular functioning between the immune systems of fish and that of other vertebrate organisms. Notably, despite this loss, a considerable degree of conservation has been maintained in the region immediately upstream of the final few residues, possibly indicating a further unknown role for this acidic tail structure.

Interestingly, CASP and GRASP do not appear to be the only products of this specific duplication event. Two of the three members of the NR4A nuclear receptor family are closely linked to both GRASP and CASP in mammals, with a similar but slightly different scenario present in zebrafish due to the genome duplication in teleost fishes (Christoffels, Koh et al. 2004; Jaillon, Aury et al. 2004) (Figure 4.5). The occurrence of one duplicated gene in close genomic proximity to another, while their respective duplicates are also closely linked to one another at a distinct genomic position, suggests a common duplication event is likely to have resulted in both sets of paralogues.

Interestingly, the current amphioxus genome assembly, places the GRASP/CASP and

NR4A homologues on distinct scaffolds. However, other genes (GPD2, CD63, and

LETMD1) that are closely linked to each in both amphioxus and several vertebrates, lead us to propose a hypothetical assembly wherein these two gene regions are in very close proximity at some evolutionary point prior to the duplication event resulting in CASP,

GRASP, NR4A1 and NR4A2 (Figure 4.5). Conspicuously, birds are the only vertebrate class that lack both GRASP and NR4A1, consistent with the loss of an entire chromosomal region containing these two duplicate genes in a vertebrate class noted for its comparatively small genome (Hughes and Piontkivska 2005). Finally, we hypothesize that the duplication event producing these genes most likely occurred at the same time,

77 involving at least these two syntenic genes and possibly others whose paralogues have since become non-functionalized and eventually lost.

Phylogenetic analysis shows that the most likely point for a gene duplication event resulting in CASP/GRASP paralogues is just prior to the radiation of the vertebrate lineage (Figure 4.4). Given the apparent subfunctionalized expression of these two genes in the immune and nervous systems respectively, the timing of this duplication event might be suggestive of the selective pressures leading to specialized expression.

Interestingly, both adaptive immunity and complex brain structure are thought to have begun evolving near the time of jawed vertebrates. While all multicellular organisms are considered to have some form of an innate immune system, invertebrates lack an adaptive immune system. The jawless vertebrates (hagfish and lamprey) meanwhile, have convergently evolved a similar but very different adaptive mechanism evidenced by the absence of B and T cell receptor proteins (Cooper and Alder 2006; Pancer and Cooper

2006; Osorio and Retaux 2008). Expansion of the brain from a neural network of relatively simple reflex pathways to a structure of more complex capacity also followed the evolution of jaws (Osorio and Retaux 2008). In terms of evolutionary timing, the beginnings of adaptive immunity and expansion in brain complexity occurred roughly in the same period as the evolution of these specialized adaptor proteins for acquired immunity and brain function (Figure 4.6). These highly selectable and advantageous traits likely provided the stimulus that drove these duplicated genes to become spatially subfunctionalized with one specializing in molecular events benefiting neural cells and the other playing a parallel role but tailored to cell types of the immune system.

78 Together, they have evolved to become a novel and very small family of adaptor proteins functioning in tissues unique to vertebrate organisms.

4.6 ACKNOWLEDGEMENTS

Funding for this research was from the Natural Sciences and Engineering Research

Council of Canada (NSERC). We would like to thank the laboratory of Dr. Alastair

Simpson for helpful suggestions and discussion.

79 4.7 REFERENCES

Christoffels, A., E. G. Koh, J. M. Chia, S. Brenner, S. Aparicio and B. Venkatesh (2004). Fugu genome analysis provides evidence for a whole-genome duplication early during the evolution of ray-finned fishes. Mol Biol Evol 21(6): 1146-51.

Cooper, M. D. and M. N. Alder (2006). The evolution of adaptive immune systems. Cell 124(4): 815-22.

Esteban, P. F., H. Y. Yoon, J. Becker, S. G. Dorsey, P. Caprari, M. E. Palko, V. Coppola, H. U. Saragovi, P. A. Randazzo and L. Tessarollo (2006). A kinase-deficient TrkC receptor isoform activates Arf6-Racl signaling through the scaffold protein tamalin. J Cell Biol 173(2): 291-299.

Force, A., M. Lynch, F. B. Pickett, A. Amores, Y. L. Yan and J. Postlethwait (1999). Preservation of duplicate genes by complementary, degenerative mutations. Genetics 151(4): 1531-45.

Holland, P. W., J. Garcia-Fernandez, N. A. Williams and A. Sidow (1994). Gene duplications and the origins of vertebrate development. Dev Suppl: 125-33.

Hughes, A. L. and H. Piontkivska (2005). DNA repeat arrays in chicken and human genomes and the adaptive evolution of avian genome size. BMC Evol Biol 5(1): 12.

Hung, A. Y. and M. Sheng (2002). PDZ domains: structural modules for protein complex assembly. J Biol Chem 277(8): 5699-702.

80 Jaillon, 0., J. M. Aury, F. Brunet, J. L. Petit, N. Stange-Thomann, E. Mauceli, L. Bouneau, C. Fischer, C. Ozouf-Costaz, A. Bernot, S. Nicaud, D. Jaffe, S. Fisher, G. Lutfalla, C. Dossat, B. Segurens, C. Dasilva, M. Salanoubat, M. Levy, N. Boudet, S. Castellano, V. Anthouard, C. Jubin, V. Castelli, M. Katinka, B. Vacherie, C. Biemont, Z. Skalli, L. Cattolico, J. Poulain, V. De Berardinis, C. Cruaud, S. Duprat, P. Brottier, J. P. Coutanceau, J. Gouzy, G. Parra, G. Lardier, C. Chappie, K. J. McKernan, P. McEwan, S. Bosak, M. Kellis, J. N. Volff, R. Guigo, M. C. Zody, J. Mesirov, K. Lindblad-Toh, B. Birren, C. Nusbaum, D. Kahn, M. Robinson-Rechavi, V. Laudet, V. Schachter, F. Quetier, W. Saurin, C. Scarpelli, P. Wincker, E. S. Lander, J. Weissenbach and H. Roest Crollius (2004). Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature 431(7011): 946-57.

Jelen, F., A. Oleksy, K. Smietana and J. Otlewski (2003). PDZ domains - common players in the cell signaling. Acta Biochim Pol 50(4): 985-1017.

Jemth, P. and S. Gianni (2007). PDZ domains: folding and binding. Biochemistry 46(30): 8701-8.

Kapsimali, M., F. Bourrat and P. Vernier (2001). Distribution of the orphan nuclear receptor Nurrl in medaka (Oryzias latipes): cues to the definition of homologous cell groups in the vertebrate brain. JComp Neurol 431(3): 276-92.

Kent, W. J. (2002). BLAT-the BLAST-like alignment tool. Genome Res 12(4): 656-64.

Kimura, K., J. Kitano, Y. Nakajima and S. Nakanishi (2004). Hyperpolarization- activated, cyclic nucleotide-gated HCN2 cation channel forms a protein assembly with multiple neuronal scaffold proteins in distinct modes of protein-protein interaction. Genes Cells 9(7): 631-40.

81 Kitano, J., Y. Yamazaki, K. Kimura, T. Masukado, Y. Nakajima and S. Nakanishi (2003). Tamalin is a scaffold protein that interacts with multiple neuronal proteins in distinct modes of protein-protein association. J Biol Chem 278(17): 14762-8.

Lynch, M. and J. S. Conery (2000). The evolutionary fate and consequences of duplicate genes. Science 290(5494): 1151-5.

Lynch, M. and A. Force (2000). The probability of duplicate gene preservation by subfunctionalization. Genetics 154(1): 459-73.

MacNeil, A. J., M. Mansour and B. Pohajdak (2007). Sorting nexin 27 interacts with the Cytohesin associated scaffolding protein (CASP) in lymphocytes. Biochem Biophys Res Commun 359(4): 848-853.

Nourry, C, S. G. Grant and J. P. Borg (2003). PDZ domain proteins: plug and play! Sci STKE 2003(179): RE7.

Nowak, M. A., M. C. Boerlijst, J. Cooke and J. M. Smith (1997). Evolution of genetic redundancy. Nature 388(6638): 167-71.

Ogawa, M., T. Miyakawa, K. Nakamura, J. Kitano, K. Furushima, H. Kiyonari, R. Nakayama, K. Nakao, K. Moriyoshi and S. Nakanishi (2007). Altered sensitivities to morphine and cocaine in scaffold protein tamalin knockout mice. Proc Natl Acad Sci USA 104(37): 14789-94.

82 Osorio, J. and S. Retaux (2008). The lamprey in evolutionary studies. Dev Genes Evol 218(5): 221-35.

Page, R. D. (1996). TreeView: an application to display phylogenetic trees on personal computers. Comput Appl Biosci 12(4): 357-8.

Pancer, Z. and M. D. Cooper (2006). The evolution of adaptive immunity. Annu Rev Immunol 24: 497-518.

Pegan, S., J. Tan, A. Huang, P. A. Slesinger, R. Riek and S. Choe (2007). NMR studies of interactions between C-terminal tail of Kir2.1 channel and PDZ1,2 domains of PSD95. Biochemistry 46(18): 5315-22.

Sidow, A. (1996). Gen(om)e duplications in the evolution of early vertebrates. Curr Opin Genet Dev 6(6): 715-22.

Stamatakis, A., T. Ludwig and H. Meier (2005). RAxML-III: a fast program for maximum likelihood-based inference of large phylogenetic trees. Bioinformatics 21(4): 456-63.

83 4.8 LINKING PARAGRAPH

The following chapter is a comprehensive review article that integrates all available published literature on CASP, from gene transcription through biological significance. CASP has been examined in several immune cell types and two mouse knockout models have been produced. Despite this, CASP's function had remained elusive. Here I propose new hypotheses on the function of CASP based on a critical analysis of the available literature. This chapter developed from my preliminary exam paper and is accepted and in press as a review article to be published in Immunology and

Cell Biology.

84 CHAPTER 5

GETTING A GRASP ON CASP: PROPERTIES AND ROLE OF THE CYTOHESIN-

ASSOCIATED SCAFFOLDING PROTEIN IN IMMUNITY

Adam J. MacNeil and Bill Pohajdak Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H4J1, Canada

Manuscript submitted July 8, 2008 Accepted August 28, 2008 and available online September 30, 2008 Published in Immunology and Cell Biology: in press.

85 5.1 ABSTRACT

CASP is a novel human adaptor protein that participates in the assembly and recruitment of protein complexes associated with intracellular trafficking and signaling.

Due to its exclusive expression in cells of hematopoietic origin, CASP has attracted attention from many groups of researchers as a potential key contributor to molecular mechanisms governing cells of the immune system. Functional characterization of CASP has involved a wide range of experimental approaches and provided broad and interesting insights that, collectively, distinguish CASP as an important contributor for a fully functioning and rapidly responsive immune system. Protein interaction studies have demonstrated that CASP interacts with members of the ARF-activating cytohesin family and with a unique PDZ domain-containing member of the sorting nexin family of endocytic trafficking proteins. Physiological knockout studies however, have revealed that CASP may not be an essential protein in immunity under normal conditions, but rather a streamlining protein that greatly ameliorates the immune system's efficiency under circumstances of significant stress. Interestingly, an evolutionarily related neuronal protein, called GRASP, may further participate in CASP-related functions in immune cells, conferring a level of redundancy in associated molecular pathways. In this review, we summarize and critically review the current literature, bringing together common themes from a variety of studies that, when considered together, provide new insights into the nature and significance of CASP function in the broad context of immunity.

86 5.2 INTRODUCTION

It is well established that the unique features of specific tissues are primarily a result of selective gene expression and the corresponding functions of their protein products. The distinctive expression of the cytohesin-associated scaffolding protein

(CASP) in cells originating through hematopoiesis has placed CASP directly in the scope of researchers interested in the molecular mechanisms governing cells of the immune system. In this review, we chronicle and critically review the existing literature for

CASP, from discovery and early description, to protein function, localization, and biological significance. By compiling and analyzing this wide array of CASP research, we aim to provide new insights and a more complete perspective on the function of CASP in immunity. This review should greatly contribute to the advancement and focus of future research with this protein, as well as allowing new researchers in the field to more easily appreciate the full range of literature available.

5.2.1 DISCOVERY AND EARLY YEARS

Prior to the sequencing of the human genome, novel and interesting transcripts were not as close to our fingertips as many of them are today. The CASP cDNA was originally cloned and sequenced in our laboratory in the early 1990s from a subtractive hybridization of human natural killer (NK) cell-enriched transcripts minus those of a T helper cell line. Of the several hundred clones obtained with this technique, 13 were new, previously undiscovered gene transcripts. One of these NK-specific clones was named

B3-1, which today corresponds to a slightly truncated CASP cDNA (Dixon, Sahely et al.

1993). This transcript was missing a portion of the 5' end, leading to a predicted

87 translation product cut short by 35 amino acids at the N-terminus. The full transcript was finally elucidated based on genomic clones and the human genome sequencing project, and was ultimately later confirmed in our lab by 5' rapid amplification of cDNA ends

(RACE). The 29 kb CASP gene (Entrez Gene: PSCDBP, also known as Cybr and

CYTIP) is localized on human chromosome 2 band qll.2 (Kim 1999), comprises eight exons giving a transcript with a 1077 bp open reading frame (ORF) corresponding to a polypeptide of 359 amino acids.

5.3 TRANSCRIPTIONAL ACTIVATION AND TRANSCRIPT PROPERTIES

In the research community, CASP is widely accepted as a hematopoietic cell specific transcript, meaning it is exclusively found in those cell types that originate from hematopoietic stem cells. A few exceptions have been reported in various experiments

(lung, kidney, heart, liver and testis), but given that vasculature and blood cells are found in nearly all tissues in one form or another, the potential for contaminated tissue is high.

No exceptions have been confirmed in specific cell lines. The expression profile for

CASP on Unigene (NCBI) is consistent with what has been shown experimentally, with expression highest in the lymph, lymph nodes, blood, bone marrow, spleen, and thymus.

This hematopoietic-specific feature of CASP expression was first described after northern analysis of a panel of cell lines, which showed lymphoid (NK/T cell) expression that could be stimulated through PHA/PMA (phytohemagglutinin/phorbol 12-myristate 13- acetate) treatment and/or IL-2 treatment (Dixon, Sahely et al. 1993). This expression profile has since been further supported in the literature through additional organ-derived and tissue-specific northern analyses and also microarray studies showing increased

88 expression in response to IL-2 and/or IL-12 in peripheral blood mononuclear cells

(PBMCs) and NK3.3 cells (Tang, Cheng et al. 2002). The highest expression was found through synergistic activity of both IL-2 and IL-12 (Tang, Cheng et al. 2002). CASP expression has also been shown to be induced either independently or synergistically

(with IL-2 or IL-12) by a variety of other cytokines including IL-7, IL-15 and IL-18

(Coppola, Barrick et al. 2006; Watford, Li et al. 2006).

In addition to NK cells, T cells, and PBMCs, CASP expression has also been found in the lymphoblastoid cell line 721, in the T helper cell line Jurkat (near undetectable when resting, but increased upon stimulation with PMA/PHA), and in

CD 14+ monocyte derived dendritic cells (DCs) (Boehm, Hofer et al. 2003). Expression

(mRNA and protein) in DCs is dramatically higher in mature versus immature DCs when isolated cells are stimulated to mature by a cocktail of cytokines (IL-1(3, IL-6 and TNF-a) along with prostaglandin E2, or using lipopolysaccharide (LPS) as the stimulus (Boehm,

Hofer et al. 2003). CASP expression was also later confirmed by real-time PCR in a variety of DC subsets including both CD 14+ (monocyte-derived) and CD34+ (stem cell- derived) DCs, along with crawl-out cells from cultured skin explants and epidermal

Langerhans cells (Hofer, Pfeil et al. 2006). Again, in this study CASP mRNA levels were dramatically higher in mature DCs when compared to immature cells.

CASP expression is also highly regulated in T cells. Several microarray studies have shown that expression is differentially upregulated during positive selection in the thymus (Huang, Li et al. 2004; Mick, Starr et al. 2004; Watford, Li et al. 2006).

Interestingly, northern analysis has also shown that induction of CASP expression is affected by the stage of development for T cells; less differentiated CD3- thymocytes

89 express CASP more highly than the more differentiated CD3+ cells, which show relatively little change in expression when stimulated by IL-12 and IL-18 (Thl polarizing conditions) (Coppola, Barrick et al. 2006). Further, cell populations in the spleen, including more mature T cells, B cells, and macrophages, show a decrease in expression of CASP following IL-12 and IL-18 stimulation, supporting the hypothesis that the differentiation stage plays a significant role in the responsiveness of immune cells to IL-

12 and IL-18 in the context of CASP expression (Coppola, Barrick et al. 2006).

Microarray, northern analysis, and real-time PCR studies have demonstrated that CASP expression in T cells is at least double in Thl polarized conditions when compared to Th2 polarized conditions (Rogge, Bianchi et al. 2000; Tang, Cheng et al. 2002; Watford, Li et al. 2006). This is consistent with the earliest reports of CASP's Thl polarizing IL-2 and

IL-12 inducibility. Further, in T cells (Jurkat), CASP mRNA (and protein) is increased following T cell Receptor (TcR) engagement and also has effects on TcR-dependent downstream events, such as enhancement of Nuclear Factor of Activated T cell (NFAT) activity (Chen, Coffey et al. 2006). This increase in NFAT activity is dependent on Vav,

JNK, and p38 signaling which are also enhanced following CASP transcription (Chen,

Coffey et al. 2006).

The upstream region of the CASP gene contains potential binding sites for several lymphoid-specific transcription factors, including AP-1 and NFAT (Mansour 2002). This sets up a potential positive feedback loop between CASP and NFAT transcription during

T cell activation and proliferation (Woodrow, Clipstone et al. 1993), given that CASP expression has been shown to increase NFAT activity (Chen, Coffey et al. 2006) and

NFAT may act on the CASP promotor, though this has not been examined

90 experimentally. CASP expression also enhances AP-1 activation. NFAT and AP-1 can act cooperatively (NFAT-AP-1) on many promoters, such as those for the IL-2 and IL-4 genes to promote transcription (Macian, Garcia-Rodriguez et al. 2000). These transcription factors may also play a key role in the specific regulation of CASP gene transcription.

In view of these reports, transcription of the CASP gene appears to be differentially regulated and specifically controlled within the hematopoietic cell lineage.

Expression is basally low in many of these cell types, but is highly and rapidly responsive to, and regulated by, many cytokines and other soluble effectors of the immune system.

From T cells to dendritic cells, CASP expression is readily inducible by specific stimulation of each respective cell type. This attribute is surely a clue into to the nature of in vivo CASP function at the protein level.

5.4 PROTEIN PROPERTIES AND INTERACTIONS

CASP is characterized as an adaptor protein, meaning that it does not contain any known catalytic domains, but rather multiple protein-protein interaction domains and motifs that participate in the assembly of larger protein complexes. The relatively small

40 kDa protein is characterized by an N-terminal PDZ domain (PSD-95 / Dig / ZO-1), an unusually long central coiled-coil motif (or leucine zipper) and a carboxy-terminal PDZ- binding motif (PDZbm) (Figure 5.1).

91 5.4.1 THE CYTOHESIN/ARNO FAMILY

Most of what is currently known about the intracellular function of CASP can be attributed to the discovery of its first binding partner, cytohesin-1. This interaction was originally reported in our laboratory using a yeast two-hybrid system and was deduced to be mediated by the interaction of each protein's coiled-coil motif based on the sequence of the corresponding clones isolated using this technique (Mansour, Lee et al. 2002).

This specific and coiled-coil motif-dependent interaction was also confirmed by in vitro protein binding assays and in vivo using transfection and immunoprecipitation (IP) experiments (Mansour, Lee et al. 2002). Interestingly, this study also showed that CASP could interact with other members of the cytohesin family of proteins (ARNO/cytohesin-

2 and ARN03/GRPl/cytohesin-3) demonstrating that CASP may act as an adaptor for all members of this family (Table 5.1) (Mansour, Lee et al. 2002). These interactions have also been reported by other groups using co-IP studies (Tang, Cheng et al. 2002; Boehm,

Hofer et al. 2003). Interestingly, if coincidentally, cytohesin-1, like CASP, was also discovered and sequenced in our laboratory (named B2-J) as part of the original subtractive hybridization mentioned above (Liu and Pohajdak 1992).

The cytohesin/ARNO family of guanine nucleotide exchange factors (GEFs) comprises a four-member group of mammalian proteins encoded by the PSCD(l-4) genes. The family members are cytohesin-1 (B2-1), ARNO (cytohesin-2), ARN03

(GRP1, cytohesin-3), and cytohesin-4. All are characterized by an N-terminal coiled-coil motif, a central Sec7 domain, and a C-terminal Pleckstrin Homology (PH) domain

(Figure 5.1). As mentioned above, the cytohesin coiled-coil motif mediates the family's interaction with CASP in hematopoietic cells, but it has also been shown to interact with

92 other proteins including GRASP, GRSP1, IPCEF1, and Gaq via this motif (Nevrivy,

Peterson et al. 2000; Klarlund, Holik et al. 2001; Venkateswarlu 2003; Laroche, Giguere et al. 2007). The Sec7 domain is responsible for this family's GEF activity and facilitates the exchange of bound GDP for GTP on ADP ribosylation factors (ARFs), a family of six proteins belonging to the Ras superfamily of small GTPases (D'Souza-Schorey and

Chavrier 2006). Once activated, ARFs can promote vesicle formation upon insertion into the appropriate membrane site and recruitment of necessary vesicle coat proteins (Jackson and Casanova 2000; Hawadle, Folarin et al. 2002). The PH domain is generally considered to be a lipid-binding domain, as it has been shown to interact preferentially with membrane sites rich in PtdIns(4,5)P2 (PIP2) and/or PtdIns(3,4,5)P3 (PIP3) phosphoinositols, which are typically found at the plasma membrane (Czech 2003).

However, this domain has also been shown to have protein binding properties, confounding its true role within some proteins (Cohen, Honda et al. 2007; Hofmann,

Thompson et al. 2007; Li, Chiang et al. 2007).

The cytohesins are a multifunctional and ubiquitous group that have been variously implicated in several biological roles including intracellular trafficking and signaling through ARF activation (Chardin, Paris et al. 1996), cytoskeletal rearrangement

(Li, Chiang et al. 2007), endocytic trafficking (Hurtado-Lorenzo, Skinner et al. 2006;

Shmuel, Santy et al. 2006), cell adhesion through p2 integrin signaling (Kolanus, Nagel et al. 1996; Geiger, Nagel et al. 2000; Perez, Mitchell et al. 2003), and recently insulin signaling in Drosophila (Fuss, Becker et al. 2006) and insulin resistance in human and mouse hepatic cells (Hafner, Schmitz et al. 2006). The functional relationship between

CASP and the cytohesin/ARNO family will be discussed below.

93 5.4.2 SNX27 AND THE SORTING NEXIN FAMILY

Recently, our group has discovered a new CASP-interacting protein, sorting nexin

27 (SNX27) (Figure 5.1). This protein was identified using ESI-tandem mass spectrometry (MS) after being isolated from human lymphocyte cell lysates in GST pulldown experiments. The PDZ-binding motif (PDZbm) of CASP and the PDZ domain of SNX27 were shown to be required for binding (MacNeil, Mansour et al. 2007).

Endogenous proteins were shown to interact via co-immunoprecipitation studies in lymphocytes. Notably, cytohesin-1 was also detected in SNX27 immunopreciptates, demonstrating that CASP may link members of the cytohesin/ARNO family to SNX27 in vivo. This is a functionally intriguing result as it links a group of proteins that activate

ARFs by stimulating vesicle formation, with a member of the endocytic trafficking family of sorting nexins. Interestingly, there are two SNX27 isoforms generated by alternative splicing (SNX27a and SNX27b) differing only slightly at the carboxy-terminus. Both are capable of interacting with CASP, yet it is not known what role each variant plays in lymphocytes.

SNX27 was originally detected as a stimulant-inducible transcript in rat neurons, where it was upregulated after acute exposure to methamphetamines (MAP) and cocaine, and designated (MAP)-responsive transcript 1 (Mrtl) (Kajii, Muraoka et al. 2003). It was later shown to interact with the PDZbm of both the 5-HT4a (hydroxytryptamine or serotonin) receptor (Joubert, Hanson et al. 2004) and more recently the G protein-gated inwardly rectifying potassium channels (GIRK or Kir3) (Lunn, Nassirpour et al. 2007) in neurons. SNX27 has also recently been isolated in lymphocytes by another group which found that it interacts with diacylglycerol kinase C, (DGKQ (Rincon, Santos et al. 2007).

94 It has become apparent, through several independent reports, that SNX27 is not a neural- specific transcript, as was initially reported (Kajii, Muraoka et al. 2003), but is a fairly ubiquitous protein that interacts with a variety of other proteins dependent upon the particular cell-type in question. This would suggest a cell-type specific role for its interaction with CASP (as well as neural receptors), given the limited and regulated expression of CASP.

The sorting nexin family comprises some 33 proteins in mammals that have an assortment of roles in endocytic trafficking and signaling pathways, from internalization to pro-degradative and membrane recycling mechanisms (Worby and Dixon 2002;

Carlton, Bujny et al. 2005; Cullen 2008). The characterizing feature of all SNXs is the

PX (phox homology) domain, which confers a lipid binding property to members of the family and contributes to their localization. Not all SNX-PX domains have been examined for phospholipid preference, but those that have generally show a strong affinity for the PtdIns(3)P (PI3P), including SNX27 (Limn, Nassirpour et al. 2007). Not surprisingly, this phosphoinositol (PI) is often enriched at endosomes given the overwhelming association of the SNX family with this compartment (Table 5.1) (Cozier,

Carlton et al. 2002; Hanson and Hong 2003; Choi, Hong et al. 2004; van Kerkhof, Lee et al. 2005). In the case of SNX27, together with the PX domain, the PDZ domain has also been shown to enhance its localization to endosomes (Rincon, Santos et al. 2007).

Aside from the PX domain, the SNXs are generally a multidomain family with most having other domains that surely aid in conferring a particular functionality for each respective protein in the endocytic pathway (Carlton, Bujny et al. 2005). Interestingly however, SNX27 is the only member of the SNX family that contains a PDZ domain,

95 making this protein unique amongst its group. The C-terminal half of SNX27 is the least characterized. This region contains both a putative Ras-Associated (RA) and a putative

B41 domain that overlap considerably. RA domains are involved in binding and regulating activated Ras or Ras-related small GTPases, while B41 is a large component of the FERM domain. FERM domains have been implicated in binding various membrane and cytoskeletal elements (Stanasila, Abuin et al. 2006; Terawaki, Maesaki et al. 2006) including actin (Lee, Bellin et al. 2004). Two other members of the SNX family, SNX17 and SNX31, contain a similar region (B41) (Seet and Hong 2006). This region has been shown to be involved in SNX17 cargo binding ability (Burden, Sun et al. 2004; Knauth,

Schluter et al. 2005; van Kerkhof, Lee et al. 2005; Czubayko, Knauth et al. 2006) while putative family member, SNX31 (NP689841), has not yet been studied. At this time, it is not known what role either of these putative domains play in SNX27 function.

The N-terminal PDZ domain of CASP appears to be the final frontier in terms of fully characterizing binding partners for CASP's known domains (Figure 5.1). However, the PDZ domain of CASP has been difficult to work with in the laboratory. CASP variants containing the PDZ domain can be produced in great quantities in E.coli expression systems, but we, and others (Stiffler, Chen et al. 2007), have found these recombinant proteins to be insoluble. It is also remains possible that other proteins, aside from the cytohesin/ARNO family and SNX27, interact with the respective coiled-coil motif and the PDZbm, as these motifs can be promiscuous and interact with several distinct proteins (Figure 5.1).

96 cytohesins

! OP ASP' • j CASP KfKfl cc IB i m

Figure 5.1 Domain and motif architecture in the CASP interactome. Protein domain distributions of CASP, the cytohesin/ARNO family, and SNX27. All four members (cytohesin-1, ARNO, ARN03, and cytohesin-4) of the mammalian family are identical in domain distribution. SNX27 exists as two isoforms with modified carboxy termini. GRASP, with a domain composition the same as CASP, may also play a role in potentially redundant pathways. Domains: PDZ: PSD-95 / Dig / ZO-1; CC: coiled-coil, PH: pleckstrin homology, PX: phox homology, Sec7: domain with homology to S. cerevisiae protein SEC7, RA: Ras Associated, B41: component of FERM.

97 Table 5.1 Summary of CASP binding partners, related localizations and functions.

CASP Binding Domains/motifs Localizations PI -binding Functions Partner mediating domain'1 cytohesin-1 CC : CC dynamic: plasma membrane, PH GEF , regulation of cell endosomal, vesicular, cytoplasmic adhesion

cytohesin-2 CC : CC plasma membrane, endosomal PH GEF, apical endocytosis, (ARNO) endosomal acidification, cell migration

cytohesin-3 CC : CC plasma membrane, cytoplasmic PH GEF, plasma membrane (ARN03/GRP1) signalling

Sorting Nexin 27 PDZ : PDZbm endosomal, vesicular PX receptor sorting and (SNX27) recycling, polarization

aphosphoinositol (PI); pleckstrin homology (PH); phox homology (PX).

Guanine-nucleotide exchange factor (GEF).

98 5.5 INTRACELLULAR LOCALIZATIONS AND FUNCTIONAL ACTIVITIES

The intracellular localization of CASP was originally described in our laboratory following transfection of COS-1 cells (Mansour, Lee et al. 2002). Full-length CASP constructs expressed in these cells showed a perinuclear localization characterized by tubulovesicular structures. Because the cytohesin/ARNO family had recently been found to be at times associated with the Golgi complex (Lee, Mansour et al. 2000; Lee and

Pohajdak 2000), CASP localization in these cells was compared to several known Golgi- related proteins. CASP did not co-localize with these markers, nor an endoplasmic reticulum (ER) marker (Mansour, Lee et al. 2002). It did, however, show a partial overlap with an ER-Golgi intermediate marker (ERGIC-53), which was somewhat enhanced upon treatment of the cells with brefeldin A (BFA) (Mansour, Lee et al. 2002).

BFA causes dissociation of the Golgi complex into diffuse vesicular structures

(Fullekrug, Sonnichsen et al. 1997), and upon treatment, CASP and ERGIC-53 were redistributed to similar structures and in the same vicinity, but still appeared to be distinct

(Mansour, Lee et al. 2002).

CASP was also investigated in COS-1 cells for co-localization with cytohesin-1.

When transfected independently, these proteins did not appear to localize to the same compartment, with cytohesin-1 showing a diffuse cytoplasmic distribution while CASP was again distributed in perinuclear vesicular structures (Mansour, Lee et al. 2002).

When expressed together, however, cytohesin-1 recruited CASP away from vesicular structures and into a cytoplasmic co-localization, dependent on the coiled-coil interaction

(Mansour, Lee et al. 2002). It should be noted however, that a C-terminal myc tag was used in the CASP expression construct for these studies. This would disrupt any potential

99 interactions mediated by the, then uncharacterized, PDZbm (e.g. with SNX27), and could have contributed to a mislocalization of the expressed protein in these cells.

Cytohesin-1 and ARN03 were previously shown to translocate to the plasma membrane of CHO-T and PC-12 cells (Venkateswarlu, Gunn-Moore et al. 1999), and

CHO-T and COS-1 cells (Venkateswarlu, Gunn-Moore et al. 1998; Langille, Patki et al.

1999) respectively, in response to epidermal growth factor (EGF). Since then, this EGF- responsive translocation has also been observed for ARNO in NIH-3T3 cells, dependent on its PH domain (Varnai, Bondeva et al. 2005). CASP and cytohesin-1 were examined in EGF-stimulated COS-1 cells and while cytohesin-1 translocated and recruited co- expressed CASP to plasma membrane ruffles, CASP did not translocate from perinuclear structures when expressed alone (Mansour, Lee et al. 2002). This indicates that cytohesin-1 (and possibly all cytohesin/ARNOs) can recruit CASP to separate subcellular locations, including the cytoplasm and membrane, dependent upon the coiled-coil interaction and, in this case, cytohesin-1-related signals (via EGF).

CASP localization has also been investigated in Jurkat cells using a transfected N- terminal Ig-tagged construct (Boehm, Hofer et al. 2003). In these cells, CASP showed a cytoplasmic and vesicular localization and also associated with the cell cortex, a cytoplasmic domain directly below the plasma membrane, usually rich in actin filaments.

The latter was dependent on the substrate onto which these cells adhered (Boehm, Hofer et al. 2003). These studies were carried out using Jurkat cells, adherent to several substrates, including poly-L-lysine, fibronectin, and ICAM-1 (Boehm, Hofer et al. 2003).

A relatively small amount of CASP was recruited to the cell cortex/membrane when cells were bound to fibronectin and ICAM-1 (integrin ligands) (Montoya, Sancho et al. 2002),

100 but not poly-L-lysine, while in all conditions CASP maintained a distinct vesicular localization overall with an apparently diminished cytoplasmic pool (Boehm, Hofer et al.

2003). This indicated that translocated CASP was likely derived from the cytoplasmic

CASP pool and not the vesicular-associated pool. Further analysis of the identity of the relatively large vesicular compartment with which CASP associated was not carried out in this study. Also, localization of the relatively uncommon Ig tag (without CASP cDNA) was not demonstrated in these cells, making it difficult to evaluate any tag- associated effects.

Through point mutations in conserved residues of the N-terminal PDZ domain,

CASP was shown to require a functional PDZ domain for localization to the cell cortex in fibronectin-adherent Jurkat cells (Boehm, Hofer et al. 2003). Though no binding partner has yet been elucidated for the PDZ domain of CASP, this was the first evidence of functional capability for this domain, and implicated the domain in localizing CASP to a membrane-associated subcellular compartment.

Jurkat cells adherent on an integrin ligand substrate and treated with PMA induced sequestration of transfected CASP away from the actin-related cortex to the cytoplasm, and showed a notable general decrease in the adherence of cells (Boehm, Hofer et al.

2003). This effect was also observed in cells co-expressing cytohesin-1 and CASP, and demonstrated that CASP could act on, and sequester, cytohesin-1 from the membrane/cortex to the cytoplasm. Cytohesin-1 has previously been associated with promotion and/or regulation of integrin-dependent cell adhesion (Kolanus, Nagel et al.

1996; Nagel, Zeitlmann et al. 1998; Geiger, Nagel et al. 2000), though the precise mechanism is not fully understood. While certain stimuli appear to cause cytohesin-1

101 recruitment of CASP to the membrane (EGF and cell adhesion-associated events),

(Mansour, Lee et al. 2002; Boehm, Hofer et al. 2003) others (PMA) show the opposite, in which CASP recruits cytohesin-1 away from the membrane/cortex (Boehm, Hofer et al.

2003). It should also be noted that phosphorylation of ARNO, and possibly other cytohesins, by protein kinase C (PKC) can influence the subcellular localization of these proteins and down regulate their exchange activity, adding a further layer of signaling- dependent recruitment and function (Santy, Frank et al. 1999). This begins to illustrate the dynamic relationship between the cytohesin/ARNO family and CASP and points to differentially regulated roles for these interactions, likely dependent on both intracellular signaling and the particular cytohesin family member, as well as further protein components (such as CASP-PDZ binding partners, ARF(s), and others) associated with the particular complex implicated.

The catalytic functional activity of the CASP-cytohesin/ARNO complex has scarcely been examined. However, there is one catalytic domain to consider in the complex. As mentioned previously, the Sec7 domain of the cytohesin/ARNO family catalyzes the exchange of bound GDP for GTP on ARFs, thereby activating the small

GTPases and promoting vesicle formation (Hawadle, Folarin et al. 2002; Casanova 2007) and ARF-related signaling (D'Souza-Schorey and Chavrier 2006; Kolanus 2007). In vitro activity assays have shown that CASP enhances the GEF activity of both cytohesin-1 and

ARNO on ARF-1, but only over a very narrow concentration range (Tang, Cheng et al.

2002). Select ARFs are localized on various membranes (e.g. plasma, Golgi, endocytic) and show preference for certain ARF GEFs (Casanova 2007), but CASP-enhanced activity would only take place if all three proteins were localized to the same subcellular

102 compartment, underscoring the importance of properly assessing the localization of these players during cellular events.

CASP localization has also been investigated in dendritic cells (DCs) where it is localized to the cytoplasm in a very diffuse distribution (Hofer, Pfeil et al. 2006).

However, as in Jurkat T cells, when DCs are bound to fibronectin, a very small pool of

CASP moves to the periphery of the cytoplasm and overlaps with the plasma membrane.

This effect is also observed when DCs are bound by antibodies for ICAM-1, VCAM-1, and CD 18 but not CD 11a, CD lie, CD49d (VLA-4) or poly-L-lysine, indicating that this subtle recruitment may be induced by a specific group of adhesion-related molecules

(Hofer, Pfeil et al. 2006). To examine the role CASP might have in DC adherence, CASP expression was silenced using siRNA (Hofer, Pfeil et al. 2006). Normal immature DCs

(low endogenous CASP expression) adhere more effectively than mature DCs (high endogenous CASP expression), but silencing of CASP enhanced the adherence of mature

DCs to fibronectin, indicating a role for CASP in facilitating a reduction in DC adherence

(Hofer, Pfeil et al. 2006). In an in vivo model, CMSP-silenced DCs showed enhanced binding to T cells but limited effects on binding to endothelial cells and demonstrated that

CASP accumulated at the contact site of DCs and T cells (Hofer, Pfeil et al. 2006).

Finally, this study demonstrated that silencing of CASP in DCs inhibited their ability to efficiently detach from T cells during an in vitro T cell priming assay (Hofer, Pfeil et al.

2006). DCs are professional antigen presenting cells (APCs) that take up antigen from their local environment, migrate to the lymph nodes and present that antigen in the context of class II MHC to T cells in the adaptive immune response. Thus, it appears that in DCs, CASP translocation to the membrane contact site (immunological synapse, IS) is

103 critical for efficient cell-cell interaction and detachment, allowing DCs to rapidly contact and subsequently leave T and B cells in their search for a T or B cell receptor corresponding to the presented antigen. The efficiency of this process (attach-detach, repeat) is critical in the immune response as it involves a vast number of cell-cell interactions during priming and migration to tissue sites.

Recently, our lab has reported the localization of CASP in NK-92 and YT cells, where CASP co-localizes with SNX27 at early endosomes (MacNeil, Mansour et al.

2007). This was demonstrated both endogenously and in a transfection system using

HEK 293 cells. Again, CASP is distributed in cytoplasmic tubulovesicular structures in

NK/T cells and shows co-localization with SNX27 at larger structures of the endosomal compartment (MacNeil, Mansour et al. 2007). This was the first report that localized

CASP to an organelle system in an endogenous context. Previous reports commented on the cytoplasmic distribution of CASP and its association with the membrane/cortex during certain events, yet neglected to determine the identity of the membrane compartment, as seen in Jurkat cells (Boehm, Hofer et al. 2003). CASP's association with this compartment is likely dependent on its interaction with SNX27. Studies in a non-hematopoietic cell type, HEK 293, demonstrated that SNX27 localized to the early endosomes regardless of CASP expression, and upon transfection, N-terminally tagged

CASP was recruited to these same structures with considerable overlap (MacNeil,

Mansour et al. 2007). Further work has demonstrated that knockout of the CASP

PDZbm, responsible for the interaction with SNX27, redistributed CASP away from the endosomal network and into a diffuse cytoplasmic distribution. These data indicate that

SNX27 recruits CASP to lymphocyte endosomes. As mentioned above, we have also

104 detected cytohesin-1 in SNX27 immunoprecipitations (MacNeil, Mansour et al. 2007), suggesting that a link between SNX27 and the cytohesin/ARNO family may be mediated through the CASP adaptor protein (Figure 5.2). Of note, ARNO has been detected in early endosomes (Hurtado-Lorenzo, Skinner et al. 2006) and has also been shown to play a role in apical endocytosis (Shmuel, Santy et al. 2006). CASP-facilitated assembly of

ARNO (or other cytohesins) with SNX27 could support a novel trafficking and/or signaling pathway at the endosomal compartment and plasma membrane in hematopoietic cells (Figure 5.2).

Altogether, CASP's localization has emerged as dynamic and varied. Endogenous

CASP localization has been shown to be both endosomal and cytoplasmic, depending on the cell-type examined, and typically presents in tubulovesicular structures. CASP can translocate to and from the plasma membrane and/or cell cortex upon stimulation in a variety of contexts (e.g. EGF, PMA, cell adhesion, immune synapse formation). Further elucidation of CASP's role in specific endosomal and membrane-related trafficking/signaling events will help characterize its function within the various cell types of the hematopoietic lineage.

105 Figure 5.2 A model implicating CASP in putative novel polarized hematopoietic endosomal sorting events. A CASP-mediated link between members of the cytohesin/ARNO family and SNX27 gives this novel complex a dynamic affinity for phosphoinositides enriched both at the plasma membrane (via the PH domain of cytohesins) and endosomal compartment (via the PX domain of SNX27). Likely dependent on the cytohesin/ARNO implicated, vesicle initiation may take place on either membrane (e.g. with cytohesin at the plasma membrane, or with ARNO at endosomal compartments). While shown from the T cell side, this complex may well support a parallel mechanism on the APC (antigen presenting cell) side. SE: sorting endosome; TGN: trans Golgi network;,PI: phosphoinositol.

106 5.6 FUNCTIONAL ASSESSMENT IN THE CONTEXT OF IMMUNITY: KNOCKOUTS AND RNAI

Evaluation of the role of C ASP in the immune system has shed some light on the biological function of this protein. Recently, two mouse models of C ASP knockouts have been produced and described (Coppola, Barrick et al. 2006; Watford, Li et al. 2006).

Although both studies show minimal effects in the development of an apparently normal immune system, with unaffected general cellularity and cell populations, effects were noted under stress conditions. One of the knockouts showed limited cell migration to inflammation sites in an aseptic peritonitis model. Also, reduced lymph node enlargement and larger tumors were found when knockouts were infected with the

Moloney murine sarcoma/leukemia virus (Coppola, Barrick et al. 2006). In contrast, under normal conditions the other CASP knockout showed no dramatic effect on migration, but rather a slight deficit in competitive stem cell bone marrow repopulation assays (Watford, Li et al. 2006). This study however, focused primarily on myeloid cell migration as opposed to lymphocytes, and a small in vitro effect on T cell migration in response to chemokine CXCL12 was observed (Watford, Li et al. 2006). This study also noted limited effects on overall lymphoid cell populations in knockouts, but used mice aged 6-12 weeks (Watford, Li et al. 2006) as compared to 6 months in the previous study where reduced numbers of lymphocytes were found in lymph nodes and fewer leukocytes in the blood (Coppola, Barrick et al. 2006). Taken together, these results point to a role for CASP in lymphocyte activation and migration to distinctively inflamed sites and/or tumor cell cytolysis.

Considering the overall effect noted on immune cell migration in stress conditions, several possibilities present themselves. CASP deficiency may affect one or a

107 number of aspects of the immune response, including: lymphocyte interaction with vascular endothelial cells during rolling and extravasation in homing to tumor and infection sites and/or lymphoid organs; efficient migration and presentation of antigen by

DCs to lymphocytes; and CTL interaction with cells targeted for destruction. All of these scenarios highlight a critical element of the immune response: efficient cell-cell interaction and communication for activation and transmigration (Ley, Laudanna et al.

2007).

5.6.1 ROLE IN IMMUNE CELL POLARIZATION

Cells of the immune system are continually responding and adjusting to their microenvironment. Whether it is in response to secreted, soluble effectors or other cell types encountered, it is critical for immune cells to be capable of defining a direction for movement and orientation, allowing proper interaction with relevant cell membranes

(Krummel and Macara 2006). Immune cells achieve this polarization through dramatic and dynamic intracellular organization of molecular components and organelles (Barreiro, de la Fuente et al. 2007). The microtublue organizing center (MTOC) and Golgi complex are reoriented toward the apical membrane to facilitate intracellular traffic in the direction of the leading edge or membrane contact site (i.e. immunological synapse) (Stinchcombe,

Majorovits et al. 2006). A constant flow of receptors and adhesion-related proteins must be delivered and maintained at the apical end and this supply also depends heavily on molecular recycling pathways to continually replenish the membrane with these essential components (Lee, Dinner et al. 2003; Krummel and Macara 2006).

108 Specialized immune cell secretory mechanisms also rely on polarization. NK cells and CTLs require polarization for the proper delivery of granule components involved cell-mediated cytotoxicity such as perforin and granzymes (Clark, Stinchcombe et al. 2003; Trambas and Griffiths 2003; Stinchcombe, Majorovits et al. 2006;

Stinchcombe and Griffiths 2007). Upon recognition of a virus-infected or tumorigenic cell, a highly organized cytoplasmic event follows within the killer cell, producing a spatially defined contact site, referred to as the immunological synapse (IS) (Stinchcombe and Griffiths 2007; Patino-Lopez, Dong et al. 2008). Central (cSMAC) and peripheral

(pSMAC) zones are defined at the IS with lytic granules delivered centrally and endosomal components orchestrating the maintenance of the peripheral contact site

(Orange, Harris et al. 2003).

Polarization of organelles requires many trafficking- and signaling-related adaptor proteins that can rapidly alter intracellular traffic and properly coordinate fusion at appropriate membranes (Krummel and Macara 2006; Gomi, Mori et al. 2007; Menager,

Menasche et al. 2007). An emerging subject of interest in molecular immunology is deciphering the role played by proteins of the endocytic pathway in polarized events such as active maintenance (establishment to dissipation) of the immunological synapse in both T cells and DCs, as well as other polarized events such as migration (Krummel and

Macara 2006). Recycling of receptors involved in both activation and adhesion are central to maintaining polarity both as lymphocytes navigate their microenvironment and in communication with antigen presenting cells. Several endocytic proteins including

Rab35, Epi64C (Patino-Lopez, Dong et al. 2008), Spinophilin (Bloom, Unternaehrer et al. 2008), Rab27, and Muncl3-4 (Menager, Menasche et al. 2007) have emerged as key

109 components in the maintenance of polarity and/or in secretion of receptors, including

TcR, as well as cytotoxic granule contents (Hong 2005; Pivot-Pajot, Varoqueaux et al.

2008). It is also becoming clear that novel and uncharacterized rapid recycling pathways play a role in trafficking receptors such as the NK inhibitory receptor CD94/NKG2A which is required on the cell surface to suppress NK cell activation in response to normal cells (Masilamani, Narayanan et al. 2008). Our lab has reported on the active polarization of endosomal SNX27 during lymphocyte migration and tumor cell engagement, indicating that this CASP binding partner plays a role during several polarized events in lymphocytes (MacNeil and Pohajdak 2007). The highly limited and regulated expression of CASP in immune cells, coupled with its role in reducing cell adhesion and interaction with the polarizing endocytic protein SNX27, suggests a possible role for CASP in novel endocytic pathways associated with immune cell polarity and activation.

5.6.2 ROLE IN CELL MIGRATION

Immune cell migration is a critical and complex polarized event (Sanchez-Madrid and del Pozo 1999; Barreiro, de la Fuente et al. 2007; Randolph, Ochando et al. 2008).

Cells must traffic many stimulus-laden microenvironments en route to their destination and will encounter and engage with scores of cells along the way, making the process very dynamic and rapidly controlled at the molecular level. Distinct phophoinositol enrichment takes place at the leading and trailing edge through PI3K and PTEN activation, (Funamoto, Meili et al. 2002) helping define each membrane site and facilitate

110 the asymmetric recruitment of proteins actively supporting migration.(Ludford-Menting,

Oliaro et al. 2005).

Specialized receptors, including integrins, are central to all aspects of immune cell migration as well as interaction and communication with antigen presenting cells

(Barreiro, de la Fuente et al. 2007). Several studies have shown that CASP plays a role in mediating detachment of cell-cell interactions through a role in reducing integrin-related adhesion associated with cytohesin-1 and also in DC-T cell priming assays (Boehm,

Hofer et al. 2003; Hofer, Pfeil et al. 2006). Interestingly, migration of CASP knockout lymphocytes in response to chemoattractants ex vivo seems unaffected in classic micro-

Boyden chamber experiments, yet under similar scenarios in vivo, CASP knockout lymphocytes show reduced migration to targeted sites (Coppola, Barrick et al. 2006).

The contrasting results for ex vs. in vivo contexts highlights the role of cell-cell interactions in rapid and effective cell migration. Interestingly, as Coppola et al.

(Coppola, Barrick et al. 2006) noted, knockout models of an integrin (LFA-1), functionally related to cytohesin-1, shows an identical effect, as well as inhibited transendothelial migration (Andrew, Spellberg et al. 1998; Coppola, Barrick et al. 2006).

Polarization of SNX27 in actively migrating lymphocytes, and possibly DCs, may enhance CASP movement to the apical membrane and help promote cell detachment to efficiently regulate quick and effective cell-cell interactions in the several scenarios mentioned above. A reduction in the efficiency of these interactions, leading to consistently longer and less productive cell-cell interactions in each case, would cumulatively lead to delayed and reduced migration of immune cells to appropriate tissue sites as noted in knockout studies (Figure 5.3). This is also supported by RNAi studies in

111 DCs where reduced CASP expression leads to longer interactions through reduced detachment during presentation to lymphocytes (Hofer, Pfeil et al. 2006). In vivo this would wreak havoc on the process of screening naive lymphocytes for efficient activation and proliferation (Figure 5.3). In one CASP knockout study, DCs were investigated for a deficit in presentation ability via adoptive transfer of peptide pre-loaded DCs and corresponding transgenic TcR CD4+ T cells, followed by isolation of T cells for proliferation and cytokine production analysis (Watford, Li et al. 2006). No effect was observed here although it is possible that the model does not reflect a natural scenario where cell-cell interaction efficiency would play a much greater role.

Combined, these knockout results suggest that other proteins may be able to substitute for CASP. Interestingly Coppola et al. (Coppola, Barrick et al. 2006) suggest

GRASP as a potential functionally overlapping molecule while Watford et al. (Watford,

Li et al. 2006) suggested that GRSP1 may be able to rescue CASP function in knockouts.

Both suggestions are well supported based on overlapping protein structure and functional role in the cell. The latter is very intriguing because GRSP1 can interact directly with the cytohesin/ARNO family, is expressed in lymphatic tissues, and contains a FERM domain (Klarlund, Holik et al. 2001; Watford, Li et al. 2006). SNX27 contains two subunits of the FERM domain suggesting that a GRSP1 rescue of CASP function may link the required FERM domain functions to cytohesins directly, circumventing an adaptor molecule (CASP) as shown in our proposed model (Figure 5.2). GRASP, meanwhile, is also a strong candidate for overlapping function with CASP and will be further discussed in the next section. Finally, it is also interesting to note that basal CASP

112 expression in mice (knockout model organism) was found to be markedly lower than in humans, particularly in the thymus (~7 fold) and lymph node (-20 fold), as shown in real time PCR experiments (Watford, Li et al. 2006). Thus a species difference in the function and/or requirement of CASP between mice and humans should be considered.

113 © ©

©••

Legend engage

.j disengage impeded Tcell • migrate to target site

Figure 5.3 A model depicting the role of CASP expression in antigen presentation by DCs. (a) In normal conditions, antigen presentation at the lymph node involves interactions between DCs and T cells, which include engagement, presentation, and disengagement, followed by proliferation and migration of activated T cells, (b) In the absence of CASP, active disengagement is impeded, resulting in a bottleneck effect where ultimately fewer T cells are being primed, activated, and exiting the lymph node and blood stream, as evidenced in several distinct studies (Coppola, Barrick et al. 2006; Hofer, Pfeil et al. 2006; Watford, Li et al. 2006) examining CASP knockouts from both the T cell and dendritic cell side.

114 5.7 ADAPTOR PROTEIN FAMILY: CASP AND GRASP

While CASP expression appears limited to cells of the hematopoietic lineage, a protein with remarkable structural similarity was originally detected and characterized in neuronal cells. GRASP (GRP1 -Associated Scaffolding Protein; aka Tamalin) has precisely the same domain/motif distribution to CASP: an N-terminal PDZ domain, a central coiled-coil motif, and a carboxy terminal PDZbm (Figure 5.1). Like CASP,

GRASP also binds members of the cytohesin/ARNO family through its coiled-coil motif

(Nevrivy, Peterson et al. 2000; Kitano, Kimura et al. 2002), but those proteins interacting with the PDZ domain and PDZbm of GRASP are neuron specific proteins, such as mGluRs (Kitano, Kimura et al. 2002) and TrkC (Esteban, Yoon et al. 2006) among others

(Kitano, Yamazaki et al. 2003; Kimura, Kitano et al. 2004), This contrasting tissue expression profile and similar domain architecture combined with specialized interacting partners for GRASP draws attention to the possibility that CASP and GRASP are cell- type specialized members of a small adaptor protein family whose functions potentially mirror one another in their respective cellular environments. In fact, we have recently provided a broad evaluation of specific molecular similarities and differences between

CASP and GRASP, as well as an analysis of the evolutionary relatedness of these two proteins (MacNeil, McEachern et al. 2008). We conclude that CASP and GRASP most likely originated from a gene duplication event in early vertebrates and became spatially subfunctionalized during the evolution of vertebrate adaptive immunity and increased brain complexity (MacNeil, McEachern et al. 2008).

Somewhat surprisingly, and in contrast to previous reports of neuronal specificity, preliminary reports indicate detectable GRASP expression in human lymphocytes

115 (Coppola, Barrick et al. 2006). The CASP/GRASP class I PDZbm (consensus S/T-X-O) is highly conserved, not only between respective proteins in mammals, but across all vertebrate classes (MacNeil, McEachern et al. 2008). This degree of similarity indicates a conserved functional motif dating prior to the gene duplication event that produced

CASP and GRASP, and suggests a common binding affinity at this motif between proteins. Given that we have previously shown that SNX27 interacts with the PDZbm of

CASP (MacNeil, Mansour et al. 2007), a highly similar PDZbm, coupled with its known interaction with members of the cytohesin/ARNO family (Nevrivy, Peterson et al. 2000;

Kitano, Kimura et al. 2002), makes GRASP a very strong candidate for redundancy with

CASP in cells of the immune system, as suggested by the mild phenotype in knockout studies (Figure 5.1).

5.8 PROGRESS TO DATE AND OUTLOOK

The past fifteen years of CASP research have revealed many interesting characteristics: from its limited and inducible expression, protein interactions, dynamic localizations and functions, to its evolutionary origin and potential redundancy in the immune system. Based on all findings to date, CASP surely plays the role of assembling protein complexes in various hematopoietic cell types. Given its interactions with the cytohesin/ARNO family and SNX27, CASP participates in vesicular/endocytic trafficking and/or signaling events, and represents a prospective node where these fundamental intracellular events may converge and integrate. CASP is responsive to cell- cell interaction and adhesion events, as well as having a role cell migration. These functions are important properties of all hematopoietic-derived cells from macrophages

116 and dendritic cells to T and NK cells. We propose that CASP's intracellular activities confer a rapidly responsive ability for hematopoietic cells to efficiently interact with their environment both temporally and spatially. Whether considering communication with other immune cells or vascular endothelial cells, efficient engagement and communication, followed by active disengagement, allows immune cells to effectively respond to their environment in the framework of immunity. One might functionally consider CASP as a 'touch & go' protein in the context of the immune system, since

CASP seems to contribute to the process of efficiently interacting with the cellular environment and then allowing the process to continuously carry on until the appropriate tissue or cell receptor has been reached. This proposal is consistent with knockout models where migration to sites of inflammation or infection, and efficient antigen presentation, are inhibited. For lymphocytes, consistently prolonged cell-cell interactions would delay migration to targeted tissue sites. For DCs, the 'tattle-tails' of the immune system, less efficient migration to the lymph nodes and subsequent T cell priming would have the same effect.

Of the visible remaining pieces of the puzzle, discovering and characterizing binding partner(s) for the PDZ domain should rank highly. The PDZ domain of CASP's twin, GRASP, binds several proteins that have been crucial in relating GRASP to its role in the nervous system; the same may hold true for CASP. A diverse array of results has provided many clues into CASP function. Moving forward with these collectively in hand should greatly enhance our ability to gain a more comprehensive understanding of how this small adaptor protein functions to promote an efficient and active immune system.

117 5.9 ACKNOWLEDGEMENT

Funding in support of this work was from the Natural Sciences and Engineering Research

Council of Canada (NSERC).

118 5.10 REFERENCES

Andrew, D. P., J. P. Spellberg, H. Takimoto, R. Schmits, T. W. Mak and M. M. Zukowski (1998). Transendothelial migration and trafficking of leukocytes in LFA-1 -deficient mice. Eur J Immunol 28(6): 1959-69.

Barreiro, O., H. de la Fuente, M. Mittelbrunn and F. Sanchez-Madrid (2007). Functional insights on the polarized redistribution of leukocyte integrins and their ligands during leukocyte migration and immune interactions. Immunol Rev 218: 147-64.

Bloom, O., J. J. Unternaehrer, A. Jiang, J. S. Shin, L. Delamarre, P. Allen and I. Mellman (2008). Spinophilin participates in information transfer at immunological synapses. J Cell Biol 181(2): 203-11.

Boehm, T., S. Hofer, P. Winklehner, B. Kellersch, C. Geiger, A. Trockenbacher, S. Never, H. Fiegl, S. Ebner, L. Ivarsson, R. Schneider, E. Kremmer, C. Heufler and W. Kolanus (2003). Attenuation of cell adhesion in lymphocytes is regulated by CYTIP, a protein which mediates signal complex sequestration. Embo J 22(5): 1014-24.

Burden, J. J., X. M. Sun, A. B. Garcia and A. K. Soutar (2004). Sorting motifs in the intracellular domain of the low density lipoprotein receptor interact with a novel domain of sorting nexin-17. J Biol Chem 279(16): 16237-45.

Carlton, J., M. Bujny, A. Rutherford and P. Cullen (2005). Sorting nexins-unifying trends and new perspectives. Traffic 6(2): 75-82.

Casanova, J. E. (2007). Regulation of Arf activation: the Sec7 family of guanine nucleotide exchange factors. Traffic 8(11): 1476-85.

Chardin, P., S. Paris, B. Antonny, S. Robineau, S. Beraud-Dufour, C. L. Jackson and M. Chabre (1996). A human exchange factor for ARF contains Sec7- and pleckstrin- homology domains. Nature 384(6608): 481-4.

Chen, Q., A. Coffey, S. G. Bourgoin and M. Gadina (2006). Cytohesin binder and regulator augments T cell receptor-induced nuclear factor of activated T Cells. AP- 1 activation through regulation of the JNK pathway. J Biol Chem 281(29): 19985- 94.

119 Choi, J. H., W. P. Hong, M. J. Kim, J. H. Kim, S. H. Ryu and P. G. Sun (2004). Sorting nexin 16 regulates EGF receptor trafficking by phosphatidylinositol-3 -phosphate interaction with the Phox domain. J Cell Sci 117(Pt 18): 4209-18.

Clark, R. H., J. C. Stinchcombe, A. Day, E. Blott, S. Booth, G. Bossi, T. Hamblin, E. G. Davies and G. M. Griffiths (2003). Adaptor protein 3-dependent microtubule- mediated movement of lytic granules to the immunological synapse. Nat Immunol 4(11): 1111-20.

Cohen, L. A., A. Honda, P. Varnai, F. D. Brown, T. Balla and J. G. Donaldson (2007). Active Arf6 recruits ARNO/cytohesin GEFs to the PM by binding their PH domains. Mol Biol Cell 18(6): 2244-53.

Cozier, G. E., J. Carlton, A. H. McGregor, P. A. Gleeson, R. D. Teasdale, H. Mellor and P. J. Cullen (2002). The phox homology (PX) domain-dependent, 3- phosphoinositide-mediated association of sorting nexin-1 with an early sorting endosomal compartment is required for its ability to regulate epidermal growth factor receptor degradation. J Biol Chem 277(50): 48730-6.

Cullen, P. J. (2008). Endosomal sorting and signalling: an emerging role for sorting nexins. Nat Rev Mol Cell Biol 9(7): 574-582.

Czech, M. P. (2003). Dynamics of phosphoinositides in membrane retrieval and insertion. Annu Rev Physiol 65: 791-815.

Czubayko, M., P. Knauth, T. Schluter, V. Florian and R. Bohnensack (2006). Sorting nexin 17, a non-self-assembling and a PtdIns(3)P high class affinity protein, interacts with the cerebral cavernous malformation related protein KRIT1. Biochem Biophys Res Commun 345(3): 1264-72.

D'Souza-Schorey, C. and P. Chavrier (2006). ARF proteins: roles in membrane traffic and beyond. Nat Rev Mol Cell Biol 7(5): 347-58.

120 Dixon, B., B. Sanely, L. Liu and B. Pohajdak (1993). Cloning a cDNA from human NK/T cells which codes for an unusual leucine zipper containing protein. Biochim BiophysActa 1216(2): 321-4.

Fullekrug, J., B. Sonnichsen, U. Schafer, P. Nguyen Van, H. D. Soling and G. Mieskes (1997). Characterization of brefeldin A induced vesicular structures containing cycling proteins of the intermediate compartment/cis-Golgi network. FEBS Lett 404(1): 75-81.

Funamoto, S., R. Meili, S. Lee, L. Parry and R. A. Firtel (2002). Spatial and temporal regulation of 3-phosphoinositides by PI 3-kinase and PTEN mediates chemotaxis. Cell 109(5): 611-23.

Fuss, B., T. Becker, I. Zinke and M. Hoch (2006). The cytohesin Steppke is essential for insulin signalling in Drosophila. Nature 444(7121): 945-8.

Geiger, C, W. Nagel, T. Boehm, Y. van Kooyk, C. G. Figdor, E. Kremmer, N. Hogg, L. Zeitlmann, H. Dierks, K. S. Weber and W. Kolanus (2000). Cytohesin-1 regulates beta-2 integrin-mediated adhesion through both ARF-GEF function and interaction with LFA-1. Embo J 19(11): 2525-36.

Gomi, H., K. Mori, S. Itohara and T. Izumi (2007). Rab27b is expressed in a wide range of exocytic cells and involved in the delivery of secretory granules near the plasma membrane. Mol Biol Cell 18(11): 4377-86.

Hafner, M., A. Schmitz, I. Grune, S. G. Srivatsan, B. Paul, W. Kolanus, T. Quast, E. Kremmer, I. Bauer and M. Famulok (2006). Inhibition of cytohesins by SecinFB leads to hepatic insulin resistance. Nature 444(7121): 941-4.

Hanson, B. J. and W. Hong (2003). Evidence for a role of SNX16 in regulating traffic between the early and later endosomal compartments. J Biol Chem 278(36): 34617-30.

121 Hawadle, M. A., N. Folarin, R. Martin and T. R. Jackson (2002). Cytohesins and centaurins control subcellular trafficking of macromolecular signaling complexes: regulation by phosphoinositides and ADP-ribosylation factors. Biol Res 35(2): 247-65.

Hofmann, I., A. Thompson, C. M. Sanderson and S. Munro (2007). The Arl4 family of small G proteins can recruit the cytohesin Arf6 exchange factors to the plasma membrane. Curr Biol 17(8): 711-6.

Hong, W. (2005). Cytotoxic T lymphocyte exocytosis: bring on the SNAREs! Trends Cell Biol 15(12): 644-50.

Huang, Y. H., D. Li, A. Winoto and E. A. Robey (2004). Distinct transcriptional programs in thymocytes responding to T cell receptor, Notch, and positive selection signals. Proc Natl Acad Sci USA 101(14): 4936-41.

Jackson, C. L. and J. E. Casanova (2000). Turning on ARF: the Sec7 family of guanine- nucleotide-exchange factors. Trends Cell Biol 10(2): 60-7.

Kim, H. S. (1999). Assignment of the human B3-1 gene (PSCDBP) to chromosome 2 band ql 1.2 by radiation hybrid mapping. Cytogenet Cell Genet 84(1-2): 95.

122 Kimura, K., J. Kitano, Y. Nakajima and S. Nakanishi (2004). Hyperpolarization- activated, cyclic nucleotide-gated HCN2 cation channel forms a protein assembly with multiple neuronal scaffold proteins in distinct modes of protein-protein interaction. Genes Cells 9(7): 631-40.

Klarlund, J. K., J. Holik, A. Chawla, J. G. Park, J. Buxton and M. P. Czech (2001). Signaling complexes of the FERM domain-containing protein GRSP1 bound to ARF exchange factor GRP1. J Biol Chem 276(43): 40065-70.

Knauth, P., T. Schluter, M. Czubayko, C. Kirsch, V. Florian, S. Schreckenberger, H. Hahn and R. Bohnensack (2005). Functions of sorting nexin 17 domains and recognition motif for P-selectin trafficking. JMol Biol 347(4): 813-25.

Kolanus, W. (2007). Guanine nucleotide exchange factors of the cytohesin family and their roles in signal transduction. Immunol Rev 218: 102-13.

Kolanus, W., W. Nagel, B. Schiller, L. Zeitlmann, S. Godar, H. Stockinger and B. Seed (1996). Alpha L beta 2 integrin/LFA-1 binding to ICAM-1 induced by cytohesin- 1, a cytoplasmic regulatory molecule. Cell 86(2): 233-42.

Krummel, M. F. and I. Macara (2006). Maintenance and modulation of T cell polarity. Nat Immunol 7(11): 1143-9.

Langille, S. E., V. Patki, J. K. Klarlund, J. M. Buxton, J. J. Holik, A. Chawla, S. Corvera and M. P. Czech (1999). ADP-ribosylation factor 6 as a target of guanine nucleotide exchange factor GRP1. J Biol Chem 274(38): 27099-104.

Laroche, G., P. M. Giguere, E. Dupre, G. Dupuis and J. L. Parent (2007). The N-terminal coiled-coil domain of the cytohesin/ARNO family of guanine nucleotide exchange factors interacts with Galphaq. Mol Cell Biochem 306(1-2): 141-52.

123 Lee, H. S., R. M. Bellin, D. L. Walker, B. Patel, P. Powers, H. Liu, B. Garcia-Alvarez, J. M. de Pereda, R. C. Liddington, N. Volkmann, D. Hanein, D. R. Critchley and R. M. Robson (2004). Characterization of an actin-binding site within the talin FERM domain. J Mol Biol 343(3): 771-84.

Lee, S. Y., M. Mansour and B. Pohajdak (2000). B2-1, a Sec7- and pleckstrin homology domain-containing protein, localizes to the Golgi complex. Exp Cell Res 256(2): 515-21.

Lee, S. Y. and B. Pohajdak (2000). N-terminal targeting of guanine nucleotide exchange factors (GEF) for ADP ribosylation factors (ARF) to the Golgi. J Cell Sci 113: 1883-9.

Ley, K., C. Laudanna, M. I. Cybulsky and S. Nourshargh (2007). Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol 7(9): 678-89.

Li, C. C., T. C. Chiang, T. S. Wu, G. Pacheco-Rodriguez, J. Moss and F. J. Lee (2007). ARL4D recruits cytohesin-2/ARNO to modulate actin remodeling. Mol Biol Cell 18(11): 4420-37.

Liu, L. and B. Pohajdak (1992). Cloning and sequencing of a human cDNA from cytolytic NK/T cells with homology to yeast SEC7. Biochim Biophys Acta 1132(1): 75-8.

Lunn, M. L., R. Nassirpour, C. Arrabit, J. Tan, I. McLeod, C. M. Arias, P. E. Sawchenko, J. R. Yates, 3rd and P. A. Slesinger (2007). A unique sorting nexin regulates trafficking of potassium channels via a PDZ domain interaction. Nat Neurosci 10(10): 1249-59.

124 Macian, F., C. Garcia-Rodriguez and A. Rao (2000). Gene expression elicited by NFAT in the presence or absence of cooperative recruitment of Fos and Jun. Embo J 19(17): 4783-95.

MacNeil, A. J., M. Mansour and B. Pohajdak (2007). Sorting nexin 27 interacts with the Cytohesin associated scaffolding protein (CASP) in lymphocytes. Biochem Biophys Res Commun 359(4): 848-853.

MacNeil, A. J., L. A. McEachern and B. Pohajdak (2008). Gene duplication in early vertebrates results in tissue-specific subfunctionalized adaptor proteins: CASP and GRASP. JMol Evol 67(2): 168-178.

MacNeil, A. J. and B. Pohajdak (2007). Polarization of endosomal SNX27 in migrating and tumor-engaged Natural Killer cells. Biochem Biophys Res Commun 361(1): 146-50.

Mansour, M. (2002). Genomic and Functional Characterization of CASP. Department of Biology. Halifax, Nova Scotia, Dalhousie University. Ph.D.: 128.

Masilamani, M., S. Narayanan, M. Prieto, F. Borrego and J. E. Coligan (2008). Uncommon endocytic and trafficking pathway of the natural killer cell CD94/NKG2A inhibitory receptor. Traffic 9(6): 1019-34.

Menager, M. M., G. Menasche, M. Romao, P. Knapnougel, C. H. Ho, M. Garfa, G. Raposo, J. Feldmann, A. Fischer and G. de Saint Basile (2007). Secretory cytotoxic granule maturation and exocytosis require the effector protein hMuncl3-4. Nat Immunol 8(3): 257-67.

Mick, V. E., T. K. Starr, T. M. McCaughtry, L. K. McNeil and K. A. Hogquist (2004). The regulated expression of a diverse set of genes during thymocyte positive selection in vivo. J Immunol 173(9): 5434-44.

Montoya, M. C, D. Sancho, M. Vicente-Manzanares and F. Sanchez-Madrid (2002). Cell adhesion and polarity during immune interactions. Immunol Rev 186: 68-82.

125 Nagel, W., L. Zeitlmann, P. Schilcher, C. Geiger, J. Kolanus and W. Kolanus (1998). Phosphoinositide 3-OH kinase activates the beta2 integrin adhesion pathway and induces membrane recruitment of cytohesin-1. J Biol Chem 273(24): 14853-61.

Orange, J. S., K. E. Harris, M. M. Andzelm, M. M. Valter, R. S. Geha and J. L. Strominger (2003). The mature activating natural killer cell immunologic synapse is formed in distinct stages. Proc Natl Acad Sci USA 100(24): 14151-6.

Patino-Lopez, G., X. Dong, K. Ben-Aissa, K. M. Bernot, T. Itoh, M. Fukuda, M. J. Krahlak, L. E. Samelson and S. Shaw (2008). RAB35 and its GAP EPI64C in T- cells regulate receptor recycling and immunological synapse formation. J Biol Chem 283(26): 18323-18330.

Perez, O. D., D. Mitchell, G. C. Jager, S. South, C. Murriel, J. McBride, L. A. Herzenberg, S. Kinoshita and G. P. Nolan (2003). Leukocyte functional antigen 1 lowers T cell activation thresholds and signaling through cytohesin-1 and Jun- activating binding protein 1. Nat Immunol 4(11): 1083-92.

Pivot-Pajot, C, F. Varoqueaux, G. de Saint Basile and S. G. Bourgoin (2008). Muncl3-4 regulates granule secretion in human neutrophils. J Immunol 180(10): 6786-97.

Randolph, G. J., J. Ochando and S. Partida-Sanchez (2008). Migration of dendritic cell subsets and their precursors. Annu Rev Immunol 26: 293-316.

Rincon, E., T. Santos, A. Avila-Flores, J. P. Albar, V. Lalioti, C. Lei, W. Hong and I. Merida (2007). Proteomic identification of sorting nexin 27 as a diacylglycerol kinase zeta -associated protein: new diacylglycerol kinase roles in endocytic recycling. Mol Cell Proteomics 6(6): 1073-1087.

Rogge, L., E. Bianchi, M. Biffi, E. Bono, S. Y. Chang, H. Alexander, C. Santini, G. Ferrari, L. Sinigaglia, M. Seiler, M. Neeb, J. Mous, F. Sinigaglia and U. Certa (2000). Transcript imaging of the development of human T helper cells using oligonucleotide arrays. Nat Genet 25(1): 96-101.

126 Sanchez-Madrid, F. and M. A. del Pozo (1999). Leukocyte polarization in cell migration and immune interactions. Embo J 18(3): 501-11.

Santy, L. C, S. R. Frank, J. C. Hatfield and J. E. Casanova (1999). Regulation of ARNO nucleotide exchange by a PH domain electrostatic switch. Curr Biol 9(20): 1173- 6.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta 1761(8): 878-96.

Stanasila, L., L. Abuin, D. Diviani and S. Cotecchia (2006). Ezrin directly interacts with the alphalb-adrenergic receptor and plays a role in receptor recycling. J Biol Chem 281(7): 4354-63.

Sfiffler, M. A., J. R. Chen, V. P. Grantcharova, Y. Lei, D. Fuchs, J. E. Allen, L. A. Zaslavskaia and G. MacBeath (2007). PDZ domain binding selectivity is optimized across the mouse proteome. Science 317(5836): 364-9.

Stinchcombe, J. C. and G. M. Griffiths (2007). Secretory mechanisms in cell-mediated cytotoxicity. Annu Rev Cell Dev Biol 23: 495-517.

Stinchcombe, J. C, E. Majorovits, G. Bossi, S. Fuller and G. M. Griffiths (2006). Centrosome polarization delivers secretory granules to the immunological synapse. Nature 443(7110): 462-5.

Terawaki, S., R. Maesaki and T. Hakoshima (2006). Structural basis for NHERF recognition by ERM proteins. Structure 14(4): 777-89.

Trambas, C. M. and G. M. Griffiths (2003). Delivering the kiss of death. Nat Immunol 4(5): 399-403.

127 van Kerkhof, P., J. Lee, L. McCormick, E. Tetrault, W. Lu, M. Schoenfish, V. Oorschot, G. J. Strous, J. Klumperman and G. Bu (2005). Sorting nexin 17 facilitates LRP recycling in the early endosome. Embo J 24(16): 2851-61.

Varnai, P., T. Bondeva, P. Tamas, B. Toth, L. Buday, L. Hunyady and T. Balla (2005). Selective cellular effects of overexpressed pleckstrin-homology domains that recognize PtdIns(3,4,5)P3 suggest their interaction with protein binding partners. J Cell Sci 118(Pt 20): 4879-88.

Venkateswarlu, K. (2003). Interaction protein for cytohesin exchange factors 1 (IPCEF1) binds cytohesin 2 and modifies its activity. J Biol Chem 278(44): 43460-9.

Venkateswarlu, K., F. Gunn-Moore, P. B. Oatey, J. M. Tavare and P. J. Cullen (1998). Nerve growth factor- and epidermal growth factor-stimulated translocation of the ADP-ribosylation factor-exchange factor GRP1 to the plasma membrane of PC 12 cells requires activation of phosphatidylinositol 3-kinase and the GRP1 pleckstrin homology domain. Biochem J 335 (Pt 1): 139-46.

Venkateswarlu, K., F. Gunn-Moore, J. M. Tavare and P. J. Cullen (1999). EGF-and NGF- stimulated translocation of cytohesin-1 to the plasma membrane of PC 12 cells requires PI 3-kinase activation and a functional cytohesin-1 PH domain. J Cell Sci 112: 1957-65.

Woodrow, M., N. A. Clipstone and D. Cantrell (1993). p21ras and calcineurin synergize to regulate the nuclear factor of activated T cells. J Exp Med 178(5): 1517-22.

Worby, C. A. and J. E. Dixon (2002). Sorting out the cellular functions of sorting nexins. Nat Rev Mol Cell Biol 3(12): 919-31.

128 5.11 LINKING PARAGRAPH

The following chapter is a summary of supporting and preliminary results that were either not included in particular manuscripts or not yet fully examined, but provide support and interesting discussion points related to this thesis. Topics covered include additional work on the biological significance of the CASP-SNX27 interaction and further protein interactions related to CASP and SNX27.

129 CHAPTER 6: SUPPORTING AND PRELIMINARY RESULTS AND DISCUSSION

6.1 SNX27 RECRUITMENT OF CASP TO ENDOSOMES

6.1.1 CASPAPDZBM HAS ALTERED INTRACELLULAR DISTRIBUTION

The following subsection is a follow-up of the work presented in chapter two.

Studies in a non-hematopoietic cell type, HEK 293, demonstrated that SNX27 localized to the early endosomes regardless of CASP expression, and upon transfection, N- terminally tagged CASP was recruited to these same structures with considerable overlap

(MacNeil, Mansour et al. 2007). Using targeted deletion studies, we have since shown that the CASP PDZbm is required for SNX27 recruitment of transfected CASP to endosomal structures, demonstrating a functional significance for the interaction.

Knockout of the CASP PDZbm, responsible for the interaction with SNX27, redistributed

CASP away from the endosomal network and into a diffuse cytoplasmic distribution

(Figure 6.1). These data indicate that SNX27 can recruit CASP to lymphocyte endosomes and that this recruitment is specifically dependent on the PDZ-PDZbm interaction.

6.1.2 MATERIALS AND METHODS

The methods for this section are as described in section 2.3.5. Briefly, transfection, immunocytochemistry, confocal imaging, and CASP constructs are as previously described. DNA constructs were as described in section 2.3.2, however,

CASP and CASPAPDZbm were ligated in pJ3H, which resulted in an HA tag replacing

GST (MacNeil, Mansour et al. 2007).

130 CASPAPDZbm

Figure 6.1 SNX27 recruits transfected CASP to the endosomal compartment, dependent on the PDZ-PDZbm interaction. (A-C) Transfected HA-tagged CASP co- localizes with endogenous SNX27 in HEK 293 cells. (D-F) Upon targeted deletion of the CASP PDZbm, transfected CASPAPDZbm is redistributed to the cytoplasm and does not co-localize with SNX27.

131 6.2 ADDITIONAL INTERACTIONS: CASP IS CLEAVED BY GRANZYME B IN VITRO

6.2.1 CASP AS A TARGET FOR GRB AND FURTHER INTERACTIONS

We have also detected a further interaction partner for CASP. Cytotoxic cells

(NK and CTLs) contain granules that, upon encountering and engaging a target tumor or virus-infected cell, can polarize to the immunological synapse (IS) and release a series of proteins including perforin and several types of granzymes (Trambas and Griffiths 2003).

These proteins then orchestrate the destruction of the targeted cell by initiating a caspase cascade and disrupting several intracellular processes (Adrain, Murphy et al. 2005;

Adrain, Duriez et al. 2006; Bots and Medema 2006; Goping, Sawchuk et al. 2006; Pipkin and Lieberman 2007). One of the main contributors to this mechanism is granzyme B

(GrB). We have detected a specific cleavage of several recombinant CASP proteins that is particular to lysates of cytotoxic cells (NK-92, YT, YTS) and not other cell types

(Jurkat, K562, Raji, CCRF-CEM). We have determined that GrB mediates proteolytic cleavage of CASP through preliminary inhibition studies and also using purified GrB protein (Figure 6.2). Using a bioinformatics tool (GraBCas) (Backes, Kuentzer et al.

2005), the predicted and only GrB cleavage site for CASP is CVSED | SS , consistent with the location of the cleavage site originally detected (Figure 6.3). This interaction facilitates the removal of 10 kDa from the carboxy-terminus of CASP, eliminating

CASP's interaction with SNX27. Physiologically, the effect of this cleavage may be similar to the dramatic redistribution of CASP as above in where CASPAPDZbm is redistributed to the cytoplasm, eliminating its interaction with, and recruitment to endosomes by, SNX27 (Figure 6.1). Targeting of CASP by GrB is an intriguing result and may indicate that disruption of CASP function is important in apoptosis of

132 hematopoietic cells and/or destruction of lymphomas by cytotoxic cells. Also noteworthy, peptides corresponding to GrB were identified in mass spectrometry analysis of C ASP binding partners, lending further support for a transient interaction with CASP.

In the course of identifying SNX27 as an interaction partner for CASP (chapter 2), mass spectrometry analysis also revealed several other proteins, including ezrin and hsp90 (heat shock protein 90), which may be involved in the higher architecture of

CASP-associated complexes. Ezrin is a member of the ERM (ezrin radixin moesin) family and is associated with the cell cortex and regulates actin contacts with the plasma membrane as well as participating in the regulation of signalling molecules (Niggli and

Rossy 2008). These proteins have been implicated in cell migration, organization of the plasma membrane, phagocytosis, and apoptosis (Niggli and Rossy 2008). These functions fit very neatly with the proposed role of CASP and SNX27 and warrant future investigation. As discussed in chapter 5, CASP has been shown to be associated with the actin-rich cell cortex and this association is dependent on the on the PDZ domain of

CASP, which is thus far uncharacterized. Hsp90 is a chaperone protein associated with stabilizing and/or protecting large molecular complexes (Wandinger, Richter et al. 2008).

Ezrin and hsp90 have not been thoroughly examined for an association with the higher architecture of CASP-related interactions.

6.2.2 MATERIALS AND METHODS

Cells and lysates are as described in section 2.3.1. For in vitro protease activity,

50ug of cleared (killer cells: YT or NK-92, and other cells: Jurkat, Raji, or K562) total lysate, or 1 unit of purified GrB in PBS-TX was incubated with 0.5-2ug of recombinant

133 GST-CASP bound to glutathione beads for one hour. Beads were washed with PBS-TX five times. Remaining bound protein was electrophoesed and stained as described in section 2.3.3.

134 Protease source: -V & f G° KDa

Figure 6.2 Recombinant CASP is cleaved by granzyme B in vitro. CASP protein bound to glutathione beads has approximately 10 kDa removed from the carboxy terminus when incubated with granzyme-containing lysates (YT cells) and with purified GrB protein. No protease activity is observed in control cells types (e.g. K562). Single arrow indicates full length recombinant CASP. Double arrow indicates recombinant CASP with the carboxy-terminus removed. Bands between the two arrows are truncated recombinant CASP from the original purification, not products of the reaction.

135 CASP CVSED | SS PDZ CC 268 359

Figure 6.3 Predicted CASP proteolytic cleavage site by granzyme B (GrB). Bioinformatics tool GraBCas predicts a significant (p<0.05) site for CASP at amino acid 268, CVSED | SS. This is consistent with experimental findings using recombinant CASP as a target for ganzyme B-containing lysates or purified GrB.

136 6.3 GRASP INTERACTS WITH SNX27 IN VITRO

The following sub-section is a preliminary investigation into the possibility of

GRASP serving a redundant role with CASP in immunity. This work was partially carried out and described by an honours student working in our laboratory (Abidali 2008).

As mentioned previously, CASP knockout mice have a mild phenotype under normal conditions, with a deficit in lymphocyte migration and tumour cytolysis when challenged with physiological stress. This phenotype suggests that other proteins may be capable of substituting for CASP. The most obvious candidate in this case, is the CASP paralogue,

GRASP. As described in chapter four and five, CASP and GRASP are very similar in structure, and both bind members of the cytohesin family through coiled-coil mediated interactions. The problem with this scenario is that GRASP has been described as a neuronal protein and had not previously been detected in hematopoietic cells. Interesting, we, and others (Coppola, Barrick et al. 2006), have preliminarily detected GRASP expression in lymphocytes. Notably, while we do not presently know the full sequence, the transcript appears to be a truncated splice variant of GRASP that may have, by virtue of its sequence, evaded early expression profiling via northern analysis (Kitano, Kimura et al. 2002).

To examine potential functional overlap, we have produced recombinant GRASP protein and conducted GST pulldown experiments using recombinant SNX27 as well as lymphocyte cell lysates as a source of binding partners. Interestingly, GRASP was found to interact with both recombinant and endogenous SNX27 protein in vitro. This was not an altogether surprising result, as CASP and GRASP have very similar class I PDZbms which suggest similar binding properties (Figure 4.3). Nevertheless, this lends further

137 support to a potential functional overlap between CASP and GRASP and may contribute to the mild phenotype of CASP knockout mice. CASP and GRASP are capable of binding both cytohesins and SNX27. The GRASP-SNX27 interaction in lymphocytes will require further investigation in vivo. Confirmation that the GRASP variant expressed in lymphocytes contains the carboxy-terminal PDZbm, which is essential for this interaction, will be critical in assessing the significance of this interaction. Also noteworthy, the GRASP-SNX27 interaction may have significance in the nervous system as both proteins are highly expressed in these tissues and overlap considerably in the hippocampus (Kitano, Kimura et al. 2002; Lunn, Nassirpour et al. 2007).

138 6.4 LINKING PARAGRAPH

The following chapter is a summary and discussion of the work presented herein.

I outline the significance of my discoveries and findings in the broader context of research on these proteins. I also place my published findings in perspective with more recent preliminary and supporting results that shed new light and offer new avenues of investigation with these molecules.

139 CHAPTER 7: CONCLUSION

The work presented and discussed here is the first to report on new binding partners for CASP since the original discovery of its interaction with cytohesin-1 in 2002

(Mansour, Lee et al. 2002; Tang, Cheng et al. 2002). New interaction partners lead to new avenues of investigation, new hypotheses, and shed light on the potential role of a protein in cellular events. The discovery of SNX27's interaction with CASP is functionally intriguing because it brings together a novel protein of the immune system with a relatively ubiquitously expressed, but unique member of the sorting nexin family.

As shown in chapter 2 (Figure 2.3), cytohesin-1 can also participate in these interactions, forming a complex with dynamic capabilities in Ptdlns binding. This molecular feature seems highly advantageous for a complex of proteins that associates with the endocytic trafficking system. Also, considering the role and functions of the cells in which this complex might exist, dynamic trafficking machinery would seem essential. For example, altering a cell's polarization to define direction requires considerable intracellular movement and shifting of trafficking and signalling pathways. Lymphocytes and other hematopoietic cells require considerable mobility to fulfill their role in recirculating to various tissues and in specialized functional roles, like antigen presentation. We propose

CASP and its associated binding partners to be mediators of novel endocytic trafficking and/or signalling pathways in these specialized cells (Figure 5.2). A potential association with other proteins including ezrin and hsp90 also supports this functional role, with ezrin providing a link to the cytoskeleton and hsp90 potentially acting as a chaperone and stabilizing the complex. These proteins have not yet been examined experimentally, but

140 suggest potentially interesting hypotheses regarding the higher architecture of the cytohesin-CASP-SNX27 complex.

This work has also provided the first association of CASP to an organelle system.

Though CASP's intracellular localization is clearly dynamic and varied depending on the cell type in question, most descriptions have concluded that CASP is cytoplasmic or associated with the plasma membrane. This work has also detailed the molecular mechanism and critical domains/motifs responsible for CASP's association with endosomes. However, these studies were conducted in non-hematopoietic cells, and seemed to exaggerate the effect of the CASP-SNX27 interaction, compared to an endogenous scenario where the association appears less intense (Figure 2.4).

Endogenously, it appears CASP has a varied localization with loose associations to one compartment or another (e.g. endosomes, membrane), which may be accentuated during particular cellular events including cell adhesion and migration.

Interestingly, based on in vitro studies, it appears probable that CASP is a proteolytic target for granzyme B in vivo. The specific cleavage of CASP that we have observed would disrupt CASP interactions with SNX27. The potential physiological effect of this disruption can be represented in the altered distribution of CASP in figure

6.1. Removal of the PDZbm would, in theory, have a similar effect to GrB cleavage.

This would also separate potential cytohesin/ARNO functions from SNX27. As discussed previously, targets downstream of granzymes and the caspase cascade generally disrupt critical functions for the cell as basic as tubulin structure (Adrain, Duriez et al.

2006; Goping, Sawchuk et al. 2006; Bovenschen, de Koning et al. 2008), and targeting of

141 CASP by GrB suggests that disruption of CASP-SNX27 functions may promote cell death.

The further surprising findings that GRASP may yet play a role in immune cells through a redundant pathway with CASP is an intriguing prospect. Interestingly, a

GRASP knockout model has recently been produced and described (Ogawa, Miyakawa et al. 2007). The authors only report on various deficits in drug challenges related to the nervous system, where GRASP is thought to be exclusively expressed. An exciting possibility is the creation of a double knockout mouse model of CASP and GRASP.

Knockout of both may provide a more revealing phenotype and more clearly suggest a biological role for this family of adapter proteins in immunity.

GRASP expression in hematopoietic cells also suggests that the subfunctionalized

(immune vs. nervous) nature of these two proteins is not as black and white as first anticipated. Chapter 4 reports on successful work done to characterize the origins and conservation of the CASP/GRASP adaptor family. The conservation in structure at the

PDZbm in vertebrate CASPs and GRASPs was a fascinating result. It suggests that the

CASP/GRASP-SNX27 interaction may be highly conserved across all vertebrate species with the intriguing exception of CASP proteins in fishes (Figure 4.3). This degree of conservation is commonly associated with functionally important molecular events and may underscore the significance of this interaction. This manuscript was well-received at the Journal of Molecular Evolution. As one reviewer suggested, this form of study could apply to any gene/protein family and should have broad applicability as a general approach to studying gene/protein family evolution.

142 There have been a number of limiting factors in working with CASP that have restricted research progress with this protein. Some of these factors are simply natural properties of CASP, such as its role as a small adaptor protein likely buried in large complex assemblies and its generally low basal expression, both making endogenous detection difficult. Others, such as its relative insolubility, and very poor immunogen properties limit its effectiveness in biochemical applications. CASP has been a notoriously difficult immunogen when attempting to raise quality antibodies that work well in biochemical assays. Several groups, including ours, have attempted this with limited success. Polyclonal and monoclonal antibodies have been raised against CASP and while some work moderately well in one application, such as western blotting, immunoprecipitation, or immunocytochemistry, they completely fail to work in others.

Despite these difficulties, which are fairly common in this area of research, CASP has garnered much attention from numerous groups of researchers eager to discover what role this novel protein plays in immunity.

CASP has been a particularly difficult protein to assign a particular role in the cell. Many researchers have studied CASP from various experimental approaches. The work I have presented in this thesis has provided novel insight on CASP and SNX27 function and it will pave the way for considerable additional study on the role of these proteins in trafficking and signalling processes specific to hematopoietic cells. My work has culminated in publishing a comprehensive review article in a journal with wide readership (chapter 5) that should greatly enhance the direction of future research on these proteins in many laboratories around the world. I am confident that my

143 contributions to this field will produce new future investigations and provide useful information to the scientific community at large. References

Abidali, A. (2008). GRASP interaction with sorting nexin 27 suggests possible rescue of CASP in knockout models. Department of Biology. Halifax, Dalhousie University. B.Sc. Honours: 33.

Adrain, C, P. J. Duriez, G. Brumatti, P. Delivani and S. J. Martin (2006). The cytotoxic lymphocyte protease, granzyme B, targets the cytoskeleton and perturbs microtubule polymerization dynamics. J Biol Chem 281(12): 8118-25.

Adrain, C, B. M. Murphy and S. J. Martin (2005). Molecular ordering of the caspase activation cascade initiated by the cytotoxic T lymphocyte/natural killer (CTL/NK) protease granzyme B. J Biol Chem 280(6): 4663-73.

Backes, C, J. Kuentzer, H. P. Lenhof, N. Comtesse and E. Meese (2005). GraBCas: a bioinformatics tool for score-based prediction of Caspase- and Granzyme B- cleavage sites in protein sequences. Nucleic Acids Res 33(Web Server issue): W208-13.

Bonifacino, J. S. and R. Rojas (2006). Retrograde transport from endosomes to the trans- Golgi network. Nat Rev Mol Cell Biol 7(8): 568-579.

145 Bots, M. and J. P. Medema (2006). Granzymes at a glance. J Cell Sci 119(Pt 24): 5011-4.

Bovenschen, N., P. J. de Koning, R. Quadir, R. Broekhuizen, J. M. Damen, C. J. Froelich, M. Slijper and J. A. Kummer (2008). NK cell protease granzyme M targets alpha- tubulin and disorganizes the microtubule network. J Immunol 180(12): 8184-91.

Carlton, J., M. Bujny, A. Rutherford and P. Cullen (2005). Sorting nexins—unifying trends and new perspectives. Traffic 6(2): 75-82.

Casanova, J. E. (2007). Regulation of Arf activation: the Sec7 family of guanine nucleotide exchange factors. Traffic 8(11): 1476-85.

Choi, J. H., W. P. Hong, M. J. Kim, J. H. Kim, S. H. Ryu and P. G. Suh (2004). Sorting nexin 16 regulates EGF receptor trafficking by phosphatidylinositol-3-phosphate interaction with the Phox domain. J Cell Sci 117(Pt 18): 4209-18.

146 Cohen, L. A., A. Honda, P. Varnai, F. D. Brown, T. Balla and J. G. Donaldson (2007). Active Arf6 recruits ARNO/cytohesin GEFs to the PM by binding their PH domains. Mol Biol Cell 18(6): 2244-53.

Cooper, M. D. and M. N. Alder (2006). The evolution of adaptive immune systems. Cell 124(4): 815-22.

Cullen, P. J. (2008). Endosomal sorting and signalling: an emerging role for sorting nexins. Nat Rev Mol Cell Biol 9(7): 574-582.

Czech, M. P. (2003). Dynamics of phosphoinositides in membrane retrieval and insertion. Annu Rev Physiol 65: 791-815.

D'Souza-Schorey, C. and P. Chavrier (2006). ARF proteins: roles in membrane traffic and beyond. Nat Rev Mol Cell Biol 7(5): 347-58.

Dixon, B., B. Sahely, L. Liu and B. Pohajdak (1993). Cloning a cDNA from human NK/T cells which codes for an unusual leucine zipper containing protein. Biochim Biophys Acta 1216(2): 321-4.

147 Force, A., M. Lynch, F. B. Pickett, A. Amores, Y. L. Yan and J. Postlethwait (1999). Preservation of duplicate genes by complementary, degenerative mutations. Genetics 151(4): 1531-45.

Funamoto, S., R. Meili, S. Lee, L. Parry and R. A. Firtel (2002). Spatial and temporal regulation of 3-phosphoinositides by PI 3-kinase and PTEN mediates chemotaxis. Cell 109(5): 611-23.

Fuss, B., T. Becker, I. Zinke and M. Hoch (2006). The cytohesin Steppke is essential for insulin signalling in Drosophila. Nature 444(7121): 945-8.

Goping, I. S., T. Sawchuk, D. A. Underhill and R. C. Bleackley (2006). Identification of {alpha}-tubulin as a granzyme B substrate during CTL-mediated apoptosis. J Cell SW119(Pt 5): 858-65.

Hanson, B. J. and W. Hong (2003). Evidence for a role of SNX16 in regulating traffic between the early and later endosomal compartments. J Biol Chem 278(36): 34617-30.

148 Hawadle, M. A., N. Folarin, R. Martin and T. R. Jackson (2002). Cytohesins and centaurins control subcellular trafficking of macromolecular signaling complexes: regulation by phosphoinositides and ADP-ribosylation factors. Biol Res 35(2): 247-65.

Hofmann, I., A. Thompson, C. M. Sanderson and S. Munro (2007). The Arl4 family of small G proteins can recruit the cytohesin Arf6 exchange factors to the plasma membrane. Curr Biol 17(8): 711-6.

Holland, P. W., J. Garcia-Fernandez, N. A. Williams and A. Sidow (1994). Gene duplications and the origins of vertebrate development. Dev Suppl: 125-33.

Hong, W. (2005). Cytotoxic T lymphocyte exocytosis: bring on the SNAREs! Trends Cell Biol 15(12): 644-50.

Hughes, A. L. and H. Piontkivska (2005). DNA repeat arrays in chicken and human genomes and the adaptive evolution of avian genome size. BMC Evol Biol 5(1): 12.

Hung, A. Y. and M. Sheng (2002). PDZ domains: structural modules for protein complex assembly. J Biol Chem 277(8): 5699-702.

Jackson, C. L. and J. E. Casanova (2000). Turning on ARF: the Sec7 family of guanine- nucleotide-exchange factors. Trends Cell Biol 10(2): 60-7.

149 Jaillon, O., J. M. Aury, F. Brunet, J. L. Petit, N. Stange-Thomann, E. Mauceli, L. Bouneau, C. Fischer, C. Ozouf-Costaz, A. Bernot, S. Nicaud, D. Jaffe, S. Fisher, G. Lutfalla, C. Dossat, B. Segurens, C. Dasilva, M. Salanoubat, M. Levy, N. Boudet, S. Castellano, V. Anthouard, C. Jubin, V. Castelli, M. Katinka, B. Vacherie, C. Biemont, Z. Skalli, L. Cattolico, J. Poulain, V. De Berardinis, C. Cruaud, S. Duprat, P. Brottier, J. P. Coutanceau, J. Gouzy, G. Parra, G. Lardier, C. Chappie, K. J. McKernan, P. McEwan, S. Bosak, M. Kellis, J. N. Volff, R. Guigo, M. C. Zody, J. Mesirov, K. Lindblad-Toh, B. Birren, C. Nusbaum, D. Kahn, M. Robinson-Rechavi, V. Laudet, V. Schachter, F. Quetier, W. Saurin, C. Scarpelli, P. Wincker, E. S. Lander, J. Weissenbach and H. Roest Crollius (2004). Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature 431(7011): 946-57.

Jelen, F., A. Oleksy, K. Smietana and J. Otlewski (2003). PDZ domains - common players in the cell signaling. Acta Biochim Pol 50(4): 985-1017.

Jemth, P. and S. Gianni (2007). PDZ domains: folding and binding. Biochemistry 46(30): 8701-8.

Kent, W. J. (2002). BLAT-the BLAST-like alignment tool. Genome Res 12(4): 656-64.

Kim, H. S. (1999). Assignment of the human B3-1 gene (PSCDBP) to chromosome 2 band ql 1.2 by radiation hybrid mapping. Cytogenet Cell Genet 84(1-2): 95.

150 Kitano, J., K. Kimura, Y. Yamazaki, T. Soda, R. Shigemoto, Y. Nakajima and S. Nakanishi (2002). Tamalin, a PDZ domain-containing protein, links a protein complex formation of group 1 metabotropic glutamate receptors and the guanine nucleotide exchange factor cytohesins. JNeurosci 22(4): 1280-9.

Kolanus, W. (2007). Guanine nucleotide exchange factors of the cytohesin family and their roles in signal transduction. Immunol Rev 218: 102-13.

Krummel, M. F. and I. Macara (2006). Maintenance and modulation of T cell polarity. Nat Immunol 7(11): 1143-9.

Lee, H. S., R. M. Bellin, D. L. Walker, B. Patel, P. Powers, H. Liu, B. Garcia-Alvarez, J. M. de Pereda, R. C. Liddington, N. Volkmann, D. Hanein, D. R. Critchley and R. M. Robson (2004). Characterization of an actin-binding site within the talin FERM domain. JMol Biol 343(3): 771-84.

151 Lee, K. H., A. R. Dinner, C. Tu, G. Campi, S. Raychaudhuri, R. Varma, T. N. Sims, W. R. Burack, H. Wu, J. Wang, O. Kanagawa, M. Markiewicz, P. M. Allen, M. L. Dustin, A. K. Chakraborty and A. S. Shaw (2003). The immunological synapse balances T cell receptor signaling and degradation. Science 302(5648): 1218-22.

Lee, S. Y., M. Mansour and B. Pohajdak (2000). B2-1, a Sec7- and pleckstrin homology domain-containing protein, localizes to the Golgi complex. Exp Cell Res 256(2): 515-21.

Lee, S. Y. and B. Pohajdak (2000). N-terminal targeting of guanine nucleotide exchange factors (GEF) for ADP ribosylation factors (ARF) to the Golgi. J Cell Sci 113: 1883-9.

Ley, K., C. Laudanna, M. I. Cybulsky and S. Nourshargh (2007). Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol 7(9): 678-89.

Li, C. C., T. C. Chiang, T. S. Wu, G. Pacheco-Rodriguez, J. Moss and F. J. Lee (2007). ARL4D recruits cytohesin-2/ARNO to modulate actin remodeling. Mol Biol Cell 18(11): 4420-37.

Liu, L. and B. Pohajdak (1992). Cloning and sequencing of a human cDNA from cytolytic NK/T cells with homology to yeast SEC7. Biochim Biophys Acta 1132(1): 75-8.

Lynch, M. and J. S. Conery (2000). The evolutionary fate and consequences of duplicate genes. Science 290(5494): 1151-5.

152 Lynch, M. and A. Force (2000). The probability of duplicate gene preservation by subfunctionalization. Genetics 154(1): 459-73.

Macian, F., C. Garcia-Rodriguez and A. Rao (2000). Gene expression elicited by NFAT in the presence or absence of cooperative recruitment of Fos and Jun. Embo J 19(17): 4783-95.

MacNeil, A. J., M. Mansour and B. Pohajdak (2007). Sorting nexin 27 interacts with the Cytohesin associated scaffolding protein (CASP) in lymphocytes. Biochem Biophys Res Commun 359(4): 848-853.

MacNeil, A. J., L. A. McEachern and B. Pohajdak (2008). Gene duplication in early vertebrates results in tissue-specific subfunctionalized adaptor proteins: CASP and GRASP. JMol Evol 67(2): 168-178.

MacNeil, A. J. and B. Pohajdak (2007). Polarization of endosomal SNX27 in migrating and tumor-engaged Natural Killer cells. Biochem Biophys Res Commun 361(1): 146-50.

Mansour, M. (2002). Genomic and Functional Characterization of CASP. Department of Biology. Halifax, Nova Scotia, Dalhousie University. Ph.D.: 128.

153 Mick, V. E., T. K. Starr, T. M. McCaughtry, L. K. McNeil and K. A. Hogquist (2004). The regulated expression of a diverse set of genes during thymocyte positive selection in vivo. J Immunol 173(9): 5434-44.

Montoya, M. C, D. Sancho, M. Vicente-Manzanares and F. Sanchez-Madrid (2002). Cell adhesion and polarity during immune interactions. Immunol Rev 186: 68-82.

Nagel, W., L. Zeitlmann, P. Schilcher, C. Geiger, J. Kolanus and W. Kolanus (1998). Phosphoinositide 3-OH kinase activates the beta2 integrin adhesion pathway and induces membrane recruitment of cytohesin-1. J Biol Chem 273(24): 14853-61.

Niggli, V. and J. Rossy (2008). Ezrin/radixin/moesin: versatile controllers of signaling molecules and of the cortical cytoskeleton. Int J Biochem Cell Biol 40(3): 344-9.

Nourry, C., S. G. Grant and J. P. Borg (2003). PDZ domain proteins: plug and play! Sci STKE 2003(179): RE7.

Nowak, M. A., M. C. Boerlijst, J. Cooke and J. M. Smith (1997). Evolution of genetic redundancy. Nature 388(6638): 167-71.

Osorio, J. and S. Retaux (2008). The lamprey in evolutionary studies. Dev Genes Evol 218(5): 221-35.

Pacheco-Rodriguez, G., J. Moss and M. Vaughan (2005). Cytohesin-1: structure, function, and ARF activation. Methods Enzymol 404: 184-95.

154 Page, R. D. (1996). Tree View: an application to display phylogenetic trees on personal computers. Comput Appl Biosci 12(4): 357-8.

Pancer, Z. and M. D. Cooper (2006). The evolution of adaptive immunity. Annu Rev Immunol 24: 497-518.

Patino-Lopez, G., X. Dong, K. Ben-Aissa, K. M. Bernot, T. Itoh, M. Fukuda, M. J. Kruhlak, L. E. Samelson and S. Shaw (2008). RAB35 and its GAP EPI64C in T- cells regulate receptor recycling and immunological synapse formation. J Biol Chem 283(26): 18323-18330.

Pipkin, M. E. and J. Lieberman (2007). Delivering the kiss of death: progress on understanding how perforin works. Curr Opin Immunol 19(3): 301-8.

Pivot-Pajot, C., F. Varoqueaux, G. de Saint Basile and S. G. Bourgoin (2008). Muncl3-4 regulates granule secretion in human neutrophils. J Immunol 180(10): 6786-97.

Randolph, G. J., J. Ochando and S. Partida-Sanchez (2008). Migration of dendritic cell subsets and their precursors. Annu Rev Immunol 26: 293-316.

155 Sanchez-Madrid, F. and M. A. del Pozo (1999). Leukocyte polarization in cell migration and immune interactions. Embo J 18(3): 501-11.

Santy, L. C, S. R. Frank, J. C. Hatfield and J. E. Casanova (1999). Regulation of ARNO nucleotide exchange by a PH domain electrostatic switch. Curr Biol 9(20): 1173- 6.

Santy, L. C, K. S. Ravichandran and J. E. Casanova (2005). The DOCK180/Elmo complex couples ARNO-mediated Arf6 activation to the downstream activation of Racl. Curr Biol 15(19): 1749-54.

Sasaki, A. T. and R. A. Firtel (2005). Finding the way: directional sensing and cell polarization through Ras signalling. Novartis Found Symp 269: 73-87; discussion 87-91,223-30.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta.

Seet, L. F. and W. Hong (2006). The Phox (PX) domain proteins and membrane traffic. Biochim Biophys Acta 1761(8): 878-96.

Shmuel, ML, L. C. Santy, S. Frank, D. Avrahami, J. E. Casanova and Y. Altschuler (2006). ARNO through its coiled-coil domain regulates endocytosis at the apical surface of polarized epithelial cells. J Biol Chem 281(19): 13300-8.

Sidow, A. (1996). Gen(om)e duplications in the evolution of early vertebrates. Curr Opin Genet Dev 6(6): 715-22.

Stamatakis, A., T. Ludwig and H. Meier (2005). RAxML-III: a fast program for maximum likelihood-based inference of large phylogenetic trees. Bioinformatics 21(4): 456-63.

Stanasila, L., L. Abuin, D. Diviani and S. Cotecchia (2006). Ezrin directly interacts with the alphalb-adrenergic receptor and plays a role in receptor recycling. J Biol Chem 281(7): 4354-63.

Stiffler, M. A., J. R. Chen, V. P. Grantcharova, Y. Lei, D. Fuchs, J. E. Allen, L. A. Zaslavskaia and G. MacBeath (2007). PDZ domain binding selectivity is optimized across the mouse proteome. Science 317(5836): 364-9.

156 Stinchcombe, J. C. and G. M. Griffiths (2007). Secretory mechanisms in cell-mediated cytotoxicity. Annu Rev Cell Dev Biol 23: 495-517.

Stinchcombe, J. C, E. Majorovits, G. Bossi, S. Fuller and G. M. Griffiths (2006). Centrosome polarization delivers secretory granules to the immunological synapse. Nature 443(7110): 462-5.

Terawaki, S., R. Maesaki and T. Hakoshima (2006). Structural basis for NHERF recognition by ERM proteins. Structure 14(4): 777-89.

Trambas, C. M. and G. M. Griffiths (2003). Delivering the kiss of death. Nat Immunol 4(5): 399-403. van Kerkhof, P., J. Lee, L. McCormick, E. Tetrault, W. Lu, M. Schoenfish, V. Oorschot, G. J. Strous, J. Klumperman and G. Bu (2005). Sorting nexin 17 facilitates LRP recycling in the early endosome. Embo J 24(16): 2851-61.

Venkateswarlu, K. (2003). Interaction protein for cytohesin exchange factors 1 (IPCEF1) binds cytohesin 2 and modifies its activity. J Biol Chem 278(44): 43460-9.

157 Venkateswarlu, K., F. Gunn-Moore, J. M. Tavare and P. J. Cullen (1999). EGF-and NGF- stimulated translocation of cytohesin-1 to the plasma membrane of PC 12 cells requires PI 3-kinase activation and a functional cytohesin-1 PH domain. J Cell Sci 112:1957-65.

Wandinger, S. K., K. Richter and J. Buchner (2008). The Hsp90 chaperone machinery. J Biol Chem 283(27): 18473-7.

Woodrow, M., N. A. Clipstone and D. Cantrell (1993). p21ras and calcineurin synergize to regulate the nuclear factor of activated T cells. J Exp Med 178(5): 1517-22.

Worby, C. A. and J. E. Dixon (2002). Sorting out the cellular functions of sorting nexins. Nat Rev Mol Cell Biol 3(12): 919-31.

158 APPENDIX A

DEVELOPMENT OF SNX27 ANTISERUM

We have also developed and characterized our own SNX27 antisera. Peptides specific to the carboxy terminus of each variant (SNX27a and b) and recombinant protein were each used as antigen to raise antibodies in rabbits with the help of Dr. Marc

Mansour. These antisera are useful in several biochemical assays including immunopreciptation of endogenous SNX27 (Figure A.l). One, anti-SNX27a, is particularly useful in immunocytochemistry as it detects endosomal SNX27 (Figure A.2).

Detection is blocked by pre-treatment of fixed cells with the peptide used to generate the antiserum. SNX27b antiserum was shown to cross-react with tubulin protein as both have similar carboxy termini. None are particularly useful in western blotting.

159 IP: anti- -6 <^ #

SNX27(rec) • SfajMB

Figure A.l New antibodies for SNX27 immunoprecipitate endogenous protein. Antisera raised against 2 distinct 15 amino acid carboxy-terminal peptides (SNX27a -DVAT and SNX27b-KEEY) and a recombinant SNX27 protein spanning the N-terminal half (28kDa) of SNX27 were used in IP reactions in lysates of YT, Raji, and K562 cells. Western blots were carried out with an independent anti-SNX27 antibody (anti-Mrtl, Japan) as described in section 2.3.3. IPs were as described in section 2.3.4.

160 SNX27 EEA1

Figure A.2 SNX27a antiserum detects endogenous, endosomal SNX27 in YT cells. Immunocytochemistry was as described in section 2.3.5 and demonstrates colocalization with early endosomal marker EEA1.

161