Molecular Dissection of the Functional Specificity of Glycophosphatidylinositol Anchors

Molecular dissection of the functional specificity of glycophosphatidylinositol anchors

By Thomas B. Nicholson Department of Biochemistry McGill University, Montreal

September 2007

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy

Carcinoembryonic antigen (CEA) is a cell surface protein attached to the membrane by a glycophosphatidylinositol (GPI) anchor, a common modification of cell surface proteins. CEA is overexpressed in many human cancers, and plays a role in tumor progression through its ability to activate certain integrins, thereby blocking cellular differentiation, inhibiting anoikis, and disrupting normal tissue architecture. Previous work established that the CEA GPI anchor contains important and specific information directing these functions, which served as the basis for an investigation of the underlying mechanisms involved. The ability of the GPI anchor to determine protein function was examined using a chimera that consisted of the CEA GPI anchor attached to neural cell adhesion molecule (NCAM) self-adhesive external domains; this chimera, NCB, possessed CEA-like, rather than NCAM-like, functions. The CEA anchor targets the protein to specific domains on the cell surface, resulting in an association with specific signaling molecules. This targeting was employed to modify CEA function, as the presence of a protein with non-functional external domains but the same anchor led to a complete and specific loss of biological function of the CEA-like protein. GPI anchor addition is determined by a specific carboxy-terminal signal sequence, which we hypothesized contained information directing the addition of a particular GPI anchor with functional specificity. To identify this signal, chimeras were generated exchanging amino acids in this signal sequence between CEA and NCAM, a protein with different functional properties. A stretch of 6 amino acids within the signal sequence was found to be necessary and sufficient for the addition of the CEA-specific anchor. Since this region is well conserved, but not identical, in the CEA family members CC6 and CC7, we examined whether these proteins were also attached to the same GPI anchors. Surprisingly, while the anchors of these proteins are functionally equivalent to that of CEA, they are not completely identical. This work therefore explores the molecular basis for functional specificity of GPI anchors, demonstrating how specificity of GPI-anchored proteins is determined and the resulting functional consequences, while offering a novel method to inhibit the function of proteins with this type of anchorage.

Résumé de Thèse

L’antigène carcinoembryonnaire humain (CEA) est un membre d’une famille de protéines de surface cellulaire fixées à la membrane par un ancrage glycophosphatidylinositol (GPI), une modification commune des protéines de la membrane plasmique. CEA est surexprimé dans plusieurs cancers humains et joue un rôle dans la progression tumorale par sa capacité à activer certaines intégrines, bloquant de ce fait la différentiation cellulaire et l’anoikis, et perturbant l’architecture tissulaire normale. Plusieurs recherches antérieures ont établi que l’ancre GPI de CEA contient de l’information importante et spécifique dirigeant ces fonctions. Ces travaux ont servi de base pour l’étude biologique et moléculaire des mécanismes impliqués. Afin de démontrer les fonctions biologiques de CEA, nous avons étudié les capacités de l’ancre GPI à modifier la fonction des protéines en utilisant une protéine hybride (NCB) composée de l’ancre GPI de CEA attachée aux domaines externes auto-adhésifs de la molécule d’adhésion cellulaire neuronale (NCAM). La protéine chimérique NCB possède des fonctions similaires à CEA. L’ancre de CEA cible la protéine à des domaines spécifiques de la surface cellulaire, ce qui mène à l’association de la protéine avec des éléments de signalisation spécifiques. Ce ciblage a été utilisé pour modifier la fonction de CEA, car la présence d’une protéine dont les domaines extracellulaires sont non fonctionnels, mais dont l’ancre est la même, a causé la perte complète et spécifique des fonctions biologiques de NCB. L’ajout d’ancres GPI est déterminé par une séquence spécifique située à l’extrémité carboxyle-terminale. Nous avons présumé que cette séquence contiendrait l’information nécessaire à l’addition sélective d’une ancre GPI de fonction spécifique. Afin d’identifier ce signal, des protéines hybrides ont été produites en échangeant des acides aminés entre CEA et NCAM, deux protéines de fonctions distinctes. La caractérisation de ces hybrides a démontré que l’ajout de l’ancre spécifique de CEA est déterminé par une séquence nécessaire et suffisante de 6 acides aminés. Puisque cette région est bien conservée, mais non identique, chez les membres CC6 et CC7 de la famille CEA, nous avons mené des recherches pour déterminer si ces protéines sont également fixées à la même ancre GPI. Quoique les ancres de ces protéines soient équivalentes à celle de CEA en termes de fonction, elles ne sont pas complètement identiques. Cette thèse présente de nouvelles informations sur la spécificité fonctionnelle des ancres GPI, en démontrant comment leur spécificité est déterminée.

Ce travail discute les conséquences fonctionnelles des ancres GPI et présente une nouvelle méthode pour altérer la fonction de protéines associées à ce type d’ancrage.

iii

Table of Contents

Abstract i

Résumé de Thèse ii

Table of contents iv

List of figures vii

List of Tables x

Acknowledgements xi

Abbreviations xiii

Chapter 1: The specificity of the GPI anchor of CEA 1

1. General Introduction 2 2. The Carcinoembryonic antigen family 3 2.1 Expression of CEA family members 3 2.2 Structure of CEA family proteins 5 2.3 The CEA family in other organisms 7 3. In vitro functions of CEA family Members 8 3.1 Intercellular adhesion 8 3.2 Cellular differentiation 10 3.3 Apoptosis 11 3.4 Tissue architecture 12 3.5 Signaling 13 3.6 Importance of the membrane anchor of CEA family members 14 3.7 Other functions 15 4. The CEA family in cancer 17 4.1 Expression pattern in cancers 17 4.2 Clinical relevance of CEA 19

4.3 CEA as a cancer target 20 5. CEA: in vivo studies 21 5.1 Tissue implantation studies 21 5.2 CEA transgenic mice 21 6. The plasma membrane 23 6.1 Plasma membrane composition 23 6.2 Membrane rafts 24 6.3 Raft heterogeneity 27 6.4 Functions of rafts 28 6.5 Other membrane domains 29 7. GPI-anchored proteins 32 7.1 GPI anchors 32 7.2 GPI anchor signal sequence 33 7.3 GPI anchor addition to preproteins 36 7.4 GPI anchor structure 37 7.5 Functional consequences of GPI anchorage 39 7.6 GPI anchor heterogeneity 40 7.7 GPI-anchored proteins and disease 41 8. Scope of the current work 42

Publications status of the research chapters presented in this thesis 44 Contribution of authors 44

Preface to Chapter 2 45

Chapter 2: Specific inhibition of GPI-anchored protein function by homing and self-association of specific GPI anchors 46

Addendum to Chapter 2 79

Preface to Chapter 3 81

Chapter 3: Identification of a novel functional specificity signal within the GPI anchor signal sequence of carcinoembryonic antigen 82

Preface to Chapter 4 103

Chapter 4: Exploring the biological properties of the GPI anchors of CEACAM6 and CEACAM7 104

Chapter 5: General discussion 128

Preface to the Research Appendix: 147

Research Appendix: The GPI anchor of CEA mediates external membrane incorporation (“painting”) 148

Original contributions to knowledge 160 Bibliography 161 Appendices 196

List of Figures

Chapter 1

Figure 1 Structure of the human CEA family 6 Figure 2 The composition of membrane rafts 25 Figure 3 Type of membrane domains formed in plasma membrane 30 Figure 4 The GPI anchor signal sequence 35 Figure 5 GPI anchor biosynthesis 37 Figure 6 The structure of the GPI anchor 38

Chapter 2

Figure 1 Surface expression of CEA and NCAM proteins on L6 myoblasts 56 Figure 2 ∆NCEA exists in close proximity to NCB but not to NCAM 57 Figure 3 ∆NCEA restores differentiation to NCB- expressing cells 61 Figure 4 Increased ECM binding by NCB transfectants is lost in the presence of ∆NCEA 63 Figure 5 NCB membrane raft association is unaltered in the presence of ∆NCEA 65 Figure 6 ∆NCEA interferes with NCB-mediated intercellular adhesion 69 Figure 7 ∆NCEA increases the size of NCB-containing rafts 71 Figure 8 Antibody crosslinking restores integrin activation 73

vii

Addendum to Chapter 2

Figure 1 Effect of targeting the GPI anchor on anoikis resistance 80

Chapter 3

Figure 1 Reducing the CEA-derived sequence in NCB 91 Figure 2 Replacing 5 amino acid stretches in the GPI anchor signal sequence of CEA 94 Figure 3 Introducing shorter NCAM sequences into the CEA sequence of 1C 95 Figure 4 Inserting 5 amino acid sequences into NCAM is insufficient to create a protein with CEA-like properties 97 Figure 5 The upstream proline is required to give CEA-like properties 98 Figure 6 Scrambling the amino acid sequence in the identified key region 100

Chapter 4

Figure 1 Effect of ∆NCEA on the differentiation of L6 transfectants expressing CEA 113 Figure 2 Differentiation under co-culture conditions of the co-expressing transfectants 115 Figure 3 Characterization of the NCAM chimeras with the GPI anchors of CC6 and CC7 117 Figure 4 Integrin modulation by N6 and N7 chimeras 119 Figure 5 The effect of N6 and N7 on differentiation 121

viii

General discussion

Figure 1 Schematic of shank-defective and shank-less anchors 142 Figure 2 Using shankless anchors to modify GPI-anchored protein function 144

Appendix

Figure 1 Painting L6 cells with whole cell lysates 154 Figure 2 CEA incorporation is more efficient for cells in suspension 155 Figure 3 BSA specifically increases CEA painting 157

List of Tables

Chapter 1

Table 1 Revised nomenclature for the CEA family 3 Table 2 Expression pattern of the human CEA family Members 4 Table 3 Raft targeting mechanisms 27 Table 4 Examples of GPI-anchored proteins 33

Chapter 2

Table 1 Differentiation of L6 transfectants 60

Chapter 3

Table 1 The primers employed to generate the chimeras used in this study 87

Chapter 4

Table 1 Primers used in this study 108 Table 2 Surface expression of CC6 and CC7 chimeras 116

Acknowledgements

Through 6 years of graduate school, countless individuals have provided me with support and encouragement. That almost guarantees that people will not be recognized in this section, but rest assured that all your help was greatly appreciated. First and foremost, I must recognize the contributions of my supervisor, Cliff Stanners. From the challenging project he proposed (which, I must admit, was met with some trepidation at first) and the independence he provided me in the lab, to the wine tastings and the Christmas parties (I’ll never hear carols the same way again), the experience of being in 901 has affected me immensely (hopefully, in a good way…). Many members of the lab also played important parts of my success. Notably, Alex, my original TA in the lab, for putting up with all of my “newby” questions at the start of my PhD, and for making our time in the lab VERY entertaining (“you’re having too much fun”). Good luck with Theranova, dude. Luisa, for being T.W.O and always protecting our interests, while entertaining and teaching as we continued our journey to somewhere “warmer.” Anthony, for keeping me company during the mornings, and for sharing his words of wisdom – “bird-watching” is an excellent pastime, but always remember the mice! Other members of the lab, including Carlos, Pilar and Juan Carlos, as well as the NT gang, also helped with scientific and non-scientific interaction. Of course, I must also thank my students, Brandon Bernard, Jovy Rosario and Denise Cook, for their contributions to my research and, most importantly, for teaching me the value of patience.

Other people in the Biochemistry department also made significant contributions to my progress and success. I must thank Dr. Gordon Shore and Dr. John Silvius for their excellent advice and suggestions as the members of my RAC. Maureen Caron for all her help, in spite of how scary she was that first time I talked with her way back in summer 1998. I also need to acknowledge Ken McDonald, for his help with flow cytometry, and Jacynthe Laliberté, for teaching me confocal microscopy. While there are too many other people that helped me to acknowledge in this short space (Gregor, Tamiko, Kurt, Lucky Meatsauce, Sappy, Joop, etc etc) I would like to specifically say thanks to Natasha (AKA The Thing) for all the help and advice over the years.

Finally, many people outside of McMed also helped me out greatly. My family for their support and interest in my research, even when it was “all Greek” to them. Edith, for accepting my long hours and supporting my research, and always being there with a smile when I needed her (even when it was only from a distance). And, of course, the EC (Grasshopper, Yellow Bear, Senorita Salsa, Morgue Mama, Cute Hermit, N!, AA and all the guests) for their friendship as we ate our way around Montreal and generally amused ourselves (Pit anyone?).

Thanks everyone, and best of luck.

xii

Abbreviations

BAC – bacterial artificial chromosome BGP – biliary glycoprotein BSA – bovine serum albumin CC/CEACAM – carcinoembryonic antigen related cell adhesion molecule CEA – carcinoembryonic antigen CGM – CEA gene family member CHO – Chinese hamster ovary DAF – decay-accelerating factor ECM – extracellular matrix ER – endoplasmic reticulum Fn – fibronectin FRET – fluorescence resonance energy transfer GI – gastrointestinal GPI – glycophosphatidylinositol HIV – human immunodeficiency virus IG – immunoglobulin IP – immunoprecipitation ITAM – immunoreceptor tyrosine-based activating motif ITIM – immunoreceptor tyrosine-based inhibitory motif mAb – monoclonal antibody NCA – non-specific cross-reacting antigen NCAM – neural cell adhesion molecule NK cell – natural killer cell PIPLC – bacterial phosphatidylinositol phospholipase C PSG – pregnancy specific glycoprotein

xiii

TM – transmembrane uPAR – receptor for urokinase-type plasminogen activator VEGF – vascular endothelial growth factor Vn – vitronectin

xiv

Chapter 1:

The specificity of the GPI anchor of carcinoembryonic antigen

1 1. General Introduction

Cancer arises from the disruption of normal cellular processes, in part because of the selective up- and down-regulation of numerous proteins, including oncogenes and tumor suppressors. Human carcinoembryonic antigen (CEA) was initially discovered in a screen for tumor-specific antigens over-expressed by human colon carcinomas (Gold and Freedman, 1965). In the approximately four decades that have followed, significant basic and clinical research has been performed on this protein, and, as they have been discovered, the other members of its family. The Stanners laboratory has been instrumental in determining the biological properties of CEA, elucidating several mechanisms that may explain the over-expression of this protein in many human cancers. These properties include the ability of CEA to block cellular differentiation (Eidelman et al., 1993) and the apoptotic process of anoikis (Ordonez et al., 2000), and to disrupt tissue architecture (Ilantzis et al., 2002), through the activation of specific integrin signaling receptors (Camacho-Leal et al., 2007; Ordonez et al., 2007). CEA is a glycophosphatidylinositol (GPI)-anchored protein, a modification that has been noted for many cell surface proteins since its characterization in the 1980s. Originally, the reason for this alternative method of plasma membrane anchorage was poorly understood, in particular because these proteins do not have a direct connection to the cytoplasm for signaling. However, recent results demonstrate that GPI anchorage alters the cell surface distribution of these proteins, such that they now associate with membrane rafts (Ikezawa, 2002). These rafts function as platforms on the cell surface promoting the association of GPI-anchored proteins with signaling molecules on the interior of the cell. The functional relevance of GPI anchorage was clarified by structure-function studies from the Stanners lab that revealed a critical role for the anchor in determining the function of CEA (Screaton et al., 2000). This work also demonstrated that functionally-specific anchors exist, as the GPI anchor of neural cell adhesion molecule (NCAM) was unable to mediate the same effects. The work presented in this thesis explores this specificity from several perspectives, offering novel insights into the specificity and functional consequences of the addition of the CEA GPI anchor.

2 2. The carcinoembryonic antigen family

2.1 Expression of CEA family members The CEA gene family, in humans, consists of 34 genes and pseudogenes located on the long arm of chromosome 19, at position 19q13.2 (Hammarstrom, 1999). This family can be further divided into 3 subgroups: 17 CEA genes, of which 12 are expressed; 11 pregnancy-specific glycoprotein (PSG) genes, with 9 expressed; and 6 unexpressed pseudogenes (Hammarstrom, 1999; Horst and Wagener, 2004). The members of this family were originally discovered by several laboratories, resulting in confusing, non-uniform nomenclature for these proteins; the abbreviations used in this thesis are consistent with the most recently updated nomenclature, as defined in Table 1 (Beauchemin et al., 1999; Kuespert et al., 2006).

Table 1: Revised Nomenclature for the human CEA family Number of Preferred name Alternate names CD Splice Variants CEA CD66e 1 CEACAM6 (CC6) NCA-90, NCA-50/90 1 CEACAM7 (CC7) CGM2 CD66c 2 CEACAM8 (CC8) NCA-95, CGM6 CD66b 1 CEACAM1 (CC1) BGP, TM-CEA CD66a 12 CEACAM3 (CC3) CGM1a CD66d 3 CEACAM4 (CC4) CGM7 1 (Beauchemin et al., 1999; Kuespert et al., 2006)

In the human fetus, CEA is expressed primarily in the developing gastrointestinal (GI) tract, and slightly in the lung, beginning at the 9th week of gestation (Gadler et al., 1980; Nap et al., 1988; Wagener et al., 1983). In adults, CEA is present, at detectable protein levels, in the colon, stomach, small intestine, tongue, esophagus, sweat glands, cervix, prostate and urinary bladder (Table 2; Frangsmyr et al., 1999; Kinugasa et al., 1998; Kodera et al., 1993; Metze et al., 1996; Nap et al., 1988). CEACAM6 (CC6, formerly NCA) shows broader expression, as it is present in the same tissues as CEA as well as in breast, pancreas, bone marrow, spleen, granulocytes, monocytes and in lung (Frangsmyr et al., 1999; Kuijpers et al., 1992; Scholzel et al., 2000). CEACAM7 (CC7, formerly CGM2) expression is highly restricted, with detectable expression only in the colon (Frangsmyr et al., 1999; Scholzel et al., 2000). The expression of

3 CEACAM1 (CC1, formerly BGP) is widespread in the human body, as it is present in the esophagus, stomach, duodenum, jejunum, ileum, colon, pancreas, liver, gall bladder, kidney, bladder, prostate, cervix, testis, squamous epithelial cells, endometrium, glandular epithelial cells, sweat and sebaceous glands, granulocytes, leukocytes, and dendritic cells (Frangsmyr et al., 1995; Horst and Wagener, 2004; Kammerer et al., 1998; Kammerer et al., 2001; Lauke et al., 2004; Prall et al., 1996). The cytoplasmic domain of CC1 exists in short and long forms; while both are expressed in the same tissues, with the exception of breast, endothelia and T cells, different ratios of the two forms are noted between various tissues (Baum et al., 1996; Turbide et al., 1997). For CEACAM3, CEACAM4 and CEACAM8 (CC3, CC4, CC8; formerly CGM1,

Table 2. Expression Pattern of the human CEA family members GPI-anchored TM domain CEA CC6 CC7 CC8 CC1 CC3 CC4 Tissue N T N T N T N T Colon + ↑ + ↑ + ↓ - + ↓↑ - - Small intestine + - - + Stomach + ↑ + ↑ - ↑ - Esophagus + ↑ + - + Salivary gland + - Tongue + + - Liver - ↑ - ↑ - + ↓ Pancreas unc ↑ + ↑ + ↑ + Gallbladder - ↑ + - + Kidney - - - + ↓ Lung unc ↑ + ↑ - - ↑ Urinary Bladder + ↑ - - + Breast unc ↑ + ↑ - + ↓ Cervix + ↑ + ↑ - + Ovary - ↑ ↑ - Endometrium - ↑ - ↑ - - + ↓ - - Testis - - - + Prostate + ↑ + - + ↓ Sweat gland + + - + Sebaceous gland - - - + Lymphocytes - - ↑ - + ↑ - Granulocytes - + - + + + + Monocytes - + - Legend: (+) positive expression; (-) negative expression; (unc) expression uncertain; blank, not determined; (↑) increased expression in tumors; (↓) decreased tumor expression; (N) normal tissue; (T) tumors. For references, see text.

4 CGM7 and CGM6, respectively), expression is exclusively in granulocytes (Kuijpers et al., 1992; Kuroki et al., 1992). The PSGs are present at high levels in the placenta during the first trimester (Rebstock et al., 1993; Zhou et al., 1997), as well as in fetal liver and adult testis, myeloid cells, salivary glands, and hematopoietic cells (Thompson et al., 1991). Recent work has described five previously uncharacterized members of the human CEA family, CC16, CC18, CC19, CC20 and CC21 (Clark et al., 2003; Zebhauser et al., 2005). While the expression pattern of these proteins has not been examined in humans, the mRNAs for these proteins (except CC21, whose expression pattern is unknown) are present primarily in the GI tract of mice, with CC19 is also found in tongue and skin. It is notable that four CEA family members, CEA, CC6, CC7 and CC1, are expressed in the colon (Frangsmyr et al., 1999), offering the possibility of interactions between these proteins (see section 3.1). In this tissue, expression of all four proteins increases as the cells migrate up the crypt and undergo differentiation (Hammarstrom, 1999). Thus, the expression of all proteins is maximal in the upper third of the crypt and the luminal surface of the colon, specifically on the apical surface of the colonocytes, while only CEA and CC6 are present (at much lower levels) closer to the base of the crypts. CEA and CC6 are also expressed in the goblet cells of the colon, unlike CC1 and CC7.

2.2 Structure of CEA family proteins The CEA family is a subfamily of the immunoglobulin (Ig) superfamily (Paxton et al., 1987), with a structure that consists of a variable immunoglobulin (IgV)-like domain at the protein’s extreme N-terminus, followed by, depending on the family member, 0-6 I-set immunoglobulin (IgI) domains (denoted A1, B1, A2, B2, A2, B3; see Figure 1). The IgI domains are often paired, consisting of one type A domain, 93 amino acids in length, and one type B domain, which are 85 amino acids long (Hammarstrom, 1999). The CEA family members are all cell surface molecules that are anchored to the plasma membrane, with the exception of the secreted CC16. The structure of CC1 varies depending on alternative splicing, resulting in multiple isoforms

Figure 1. Structure of the human CEA family with different intracellular domains, having either 12-14 amino acids (the short cytoplasmic domain) or 72 amino acids (the long cytoplasmic domain) (Barnett et al., 1993; McCuaig et al., 1993; Najjar et al., 1993). CEA family proteins are highly glycosylated with 28 potential N- linked glycosylation sites, specified by the amino acid sequence N-X-S/T, located in CEA, such that approximately half of CEA’s molecular weight arises from these sugars. In humans, CEA family members are attached to the cellular membrane by either a transmembrane (TM) domain or a GPI anchor. The TM members are CC1, CC3 and CC4, while CEA, CC6, CC7 and CC8 are modified with GPI anchors (Hammarstrom, 1999; Thompson et al., 1991). The nature of membrane anchorage has important consequences for the biological functions of these proteins (Screaton et al., 2000), which is explored in depth in section 3.6.

6 2.3 The CEA family in other organisms The CEA family contains members in various other mammalian species. Using human cDNAs as probes, other family members were initially identified in both mice and rats (Beauchemin et al., 1989; Kodelja et al., 1989). There are, presently, 22 known murine CEA family members, of which only a CC1 ortholog has been positively identified (Zebhauser et al., 2005). Of these 22 genes, two TM (CC1 and CC2) and three secreted (CC9-11) CEA members are expressed, as well as eight PSG members (Zimmermann, 1998). In rats, there are one TM (CC1) and four secreted (CC9-12) members, in addition to eight PSG proteins (Zimmermann, 1998). Significantly, GPI-anchored members of the CEA family have not been described in either mice or rats. In primates, the CEA family consists of both GPI-anchored and TM members, as well as the PSGs (Tobi et al., 2000; Tobi et al., 1994; Zhou and Hammarstrom, 2001; Zhou et al., 2001). As rodents appear to lack GPI-anchored CEA family members (Zimmermann, 1998), it is thought that the CEA family is undergoing rapid evolution (Stanners et al., 1992). This is supported by experiments demonstrating that minimal changes are required in the amino acid sequence of CC1 to generate a GPI-anchored protein (Naghibalhossaini and Stanners, 2004), suggesting that the GPI-anchored family members arose after the divergence of primates from rodents. By comparing the sequences of CEA family proteins from 13 primate species, it was determined that GPI anchorage evolved separately in old- and new-world monkeys, implying the existence of an adaptive advantage for this modification due to its convergent evolution (Naghibalhossaini et al., 2007). To confirm that these primate family members were GPI-anchored, the sequence from a CEA family protein from Callicebus Molloch monkeys, which theoretically determined anchor addition, was attached to the external domain of CC1. The resulting chimera was GPI anchored, confirming that primates do express GPI-anchored CEA family members, and showed CEA-like biological functions (Naghibalhossaini et al., 2007). Thus, GPI anchorage is a fairly new modification in the CEA family, and has important functional consequences (see section 3.6). Further studies have also identified CEA family members in bats (Naghibalhossaini et al., 2007) and cows (Kammerer et al., 2004a), suggesting that this family is widespread among mammals.

7 3. In vitro functions of CEA Family members

3.1 Intercellular adhesion Cell-cell communication is critical for the survival of multicellular organisms, and is mediated, in large part, by interactions between adhesion molecules expressed on the cell surface. Because it is a cell surface protein belonging to the Ig superfamily, CEA was hypothesized to function as an intercellular adhesion molecule, similar to other members of the Ig family (Williams and Barclay, 1988). When this was examined, it was determined that CEA mediates, in vitro, calcium-independent homotypic adhesion when expressed by Chinese hamster ovary (CHO) or human colorectal carcinoma cells (Benchimol et al., 1989). This occurs through direct protein- protein interactions, as CEA expression by cells also increases binding to purified and immobilized CEA molecules (Jessup et al., 1993; Levin and Griffin, 1991). Cells expressing CEA do not interact with those expressing another Ig superfamily member, NCAM, indicating the specificity of CEA-mediated adhesion (Zhou et al., 1993a). Intercellular adhesion between apposed CEA molecules occurs by double reciprocal, antiparallel bonding between the N- terminal domain of one CEA molecule and the internal A3B3 domains of the other CEA molecule (Zhou et al., 1993a). Deleting 75 of the 108 amino acids composing the N-terminal domain of CEA (∆NCEA) completely abrogates adhesive function, confirming the importance of this region for adhesion, and has significant effects on biological function, as described in subsequent sections (Eidelman et al., 1993). Monoclonal antibodies with epitopes within the N- terminal domain of CEA can also be used to block this cellular aggregation (Zhou et al., 1993b). Finer mapping of the regions required for the adhesive function, by genetically mutating residues in the N-terminal domain, identified 3 critical subdomains: 30GYSWYK, 42NRQII and 80QNDTG (Taheri et al., 2000). The epitope of one of the antibodies (A20) that blocks CEA-mediated adhesion bridges the first two of the identified domains, while treatment with peptide fragments representing these regions is also capable of blocking cellular aggregation, confirming the importance of these subdomains (Taheri et al., 2000; Zhou et al., 1993a). Despite its high level of glycosylation, the carbohydrate side-chains of CEA are not required for adhesion, as expression of CEA in CHO glycosylation (Lec) mutants alters the kinetics of adhesion but does not completely block this function (Charbonneau and Stanners, 1999). CEA-mediated adhesion may also have an important parallel component, occurring between molecules on the surface of the

8 same cell (Taheri et al., 2003) and possibly involving the N-terminal domain, which, when expressed in solution free of the other Ig domains, can oligomerize (Krop-Watorek et al., 1998). All other members of the CEA family that have been examined also mediate intercellular adhesion. GPI-anchored CC6 mediates Ca+2-independent intercellular adhesion similar to CEA (Oikawa et al., 1989; Rojas et al., 1996; Zhou et al., 1990), while TM CC1 also functions as an intercellular adhesion molecule (Rojas et al., 1996; Rojas et al., 1990; Watt et al., 1994), although in a different manner from CEA and CC6. CC1 adhesion is dependent on the presence of Ca+2 ions and incubation at physiological temperature, requirements that are not seen for CEA (Rojas et al., 1996; Rojas et al., 1990). Furthermore, CC1 mediates adhesion through direct N- terminal domain interactions, unlike the antiparallel mechanism of CEA (Teixeira et al., 1994). CEA family members also mediate heterotypic intercellular adhesion, both with other members of the CEA family and with various other cell surface proteins. As such, CEA, CC6 and CC1 can all interact with each other (Oikawa et al., 1989; Oikawa et al., 1992; Zhou et al., 1990), although the functional consequences of homo- versus heterotypic interaction are unknown. Similarly, CC7 can mediate adhesion with CEA (Zhai and Stanners, manuscript in preparation), and CC8, which does not mediate homotypic intercellular adhesion, undergoes heterotypic adhesion with CC6 (Oikawa et al., 1991) through the N domains of these proteins (Oikawa et al., 2000). Thus, in tissues expressing multiple CEA family members, such as the colon, there may be a complex interplay between the various proteins, resulting in different signaling events depending on the family members involved. Heterotypic interactions between CEA and non-family members may also play important physiological roles. CEA interacts with an 80 kDa receptor, a membrane-anchored homolog of heterogeneous nuclear protein M4 (hnRNP M4), on the surface of liver Kupffer cells (Bajenova et al., 2003; Bajenova et al., 2001; Gangopadhyay et al., 1996b), resulting in the induction of cytokine expression (Gangopadhyay et al., 1996a). hnRNP M4 is also expressed by colon cancer cells, where it interacts with CEA and signals in a CEA-dependent manner (Laguinge et al., 2005). CEA also interacts with specific galectins (notably, -1, -3 and -4), proteins that bind to carbohydrates, through its N-linked carbohydrate side chains; in particular, CEA and galectin-4 colocalize on the surface of colon cells and adhesion by galectin-4 is modulated by CEA (Ideo et al., 2005; Ohannesian et al., 1994; Ohannesian et al., 1995). Thus, the CEA family undergoes a complex set of homo- and heterotypic interactions, greatly influencing cell-cell communication.

9 3.2 Cellular differentiation Cells progress along differentiation pathways, beginning with stem cells, to form the tissues that compose multicellular organisms. Blocking this maturation process is thought to be one mechanism that contributes to cancer progression, by creating a pool of proliferation-competent cells that can undergo further mutations to generate a tumor. The Stanners laboratory hypothesized that this could be an effect mediated by CEA, and demonstrated that ectopic expression of CEA blocks the terminal differentiation of L6 rat myoblasts (Eidelman et al., 1993). When L6 cells expressing CEA reach confluency and are placed in medium that promotes differentiation by reducing growth factors, these transfectants enter a reversible Go-like state rather than differentiate (Screaton et al., 1997). This property has also been attributed to CC6 (Rojas et al., 1996) and CC7 (Zhai and Stanners, manuscript in preparation), which are also GPI- anchored proteins, while the expression of TM CC1 has no effect (Rojas et al., 1996). The ability of ectopic CEA to block differentiation has been extended to several other cell types: C2C12 mouse myoblasts (Screaton et al., 1997), neurite extension of mouse P19 carcinoma cells induced by retinoic acid (Malette and Stanners, manuscript in preparation) and adipogenic differentiation of both C3H10T1/2 and 3T3L1 cells (DeMarte and Stanners, manuscript in preparation). In the adult human colon, CEA and CC6 are expressed primarily by the differentiated cells of this tissue, resulting in the use of these proteins as differentiation markers (Ilantzis et al., 1997; Jothy et al., 1993; Scholzel et al., 2000). However, examination of colorectal carcinomas indicated that CEA is expressed at higher levels in less differentiated tumors (Ilantzis et al., 1997), and overexpression of CEA/CC6 in various colorectal cancer cell lines blocks the differentiation of these cells (Ilantzis et al., 2002). The ability of CEA to block the differentiation of these diverse cell lines has important implications for human malignancies. Approximately 50% of all human cancers over-express CEA; the ability of CEA to function as a pan-differentiation inhibitor, regardless of cellular background, suggests that even in tissues that do not normally express CEA, this protein can play an important role in malignancy progression. Expression of the non-adhesive deletion mutant ∆NCEA (see section 3.1) does not block L6 differentiation, suggesting that the adhesive function is required for CEA’s biological functions (Eidelman et al., 1993). Co-culturing CEA- and ∆NCEA-expressing myoblasts demonstrated that the expression of CEA on the surface of one cell can activate ∆NCEA on the surface of apposed cells (due to the interaction of the N-terminal domain of CEA with the

10 internal domains of ∆NCEA), indicating that anti-parallel adhesion is sufficient to mediate the blockage of differentiation (Taheri et al., 2003). This study also identified specific CEA point mutants, within the subdomain 80QNDTG, that retain the ability to mediate cellular aggregation but no longer block differentiation (Taheri et al., 2003), suggesting that while anti-parallel interactions were unaffected, parallel interactions were blocked, and that this is sufficient to affect biological function. Thus, the adhesive function is a combination of parallel and anti- parallel interactions, both of which are important for the ability to block differentiation. As explored in section 3.6, further examination of the requirements for the CEA block of differentiation demonstrated a requirement for the GPI anchor as well (Screaton et al., 2000).

3.3 Apoptosis Biological systems maintain proper tissue homeostasis through a balance of cellular growth and death; these death processes result from the initiation of programmed cell death, or apoptosis. As an example, the colon is composed of crypts, with the structure maintained by a finely tuned balance between proliferation and apoptosis, which, if disrupted, can lead to disease and neoplasia (Edelblum et al., 2006; Ramachandran et al., 2000). When cells lose contact with their normal extracellular matrix (ECM), a particular form of cell death, termed anoikis, occurs (Frisch and Ruoslahti, 1997; Frisch and Screaton, 2001). This is of particular relevance in the colon, where cells are continually sloughed off into the colonic lumen, resulting in the initiation of anoikis. Importantly, anoikis inhibits cells from growing in inappropriate locations, so disrupting this process can contribute to malignancy and tumor metastasis (Ruoslahti and Reed, 1994). Cells interact with the ECM through a family of proteins known as integrins, receptors that have their functions modulated by CEA and CC6 (see section 3.5). In fact, over-expression of CEA or CC6 in a variety of cell lines blocks the induction of anoikis when cells are cultured under conditions that do not permit attachment (Duxbury et al., 2004d; Ordonez et al., 2000). Decreasing the expression of CEA/CC6, by RNA interference or ribozyme targeting, is sufficient to restore anoikis in these cells, demonstrating a direct role for these proteins in inhibiting this process (Duxbury et al., 2004d; Soeth et al., 2001; Wirth et al., 2002). The role of CC1 in cell death is somewhat less conclusive, as certain studies show that it blocks apoptosis in granulocytes (Singer et al., 2005), others indicate no effect in various cell lines, including L6 myoblasts (Ordonez et al., 2000), and one study demonstrated that CC1 accelerates apoptosis in

11 mammary carcinoma cells (Kirshner et al., 2003a). While the role of CC1 in the induction of apoptosis is unclear, and may be dependent on cell type, it is interesting to note that CC1 can serve as a substrate for caspase-mediated cleavage after apoptosis commences, increasing its adhesive ability, although the biological result of this increased adhesion is unknown (Houde et al., 2003). Ultimately, however, the disruption of anoikis provides another possible means by which CEA/CC6 can promote tumor growth.

3.4 Tissue architecture The colonic epithelium is composed of crypts that are continually losing cells into the lumen of the colon, which need to be replaced. This occurs through the division of stem cells located at the bottom one-third of the crypt, providing replacement cells that differentiate as they migrate up the crypt walls (Burgess, 1998; Karam, 1999). Cell adhesion molecules play a critical role in transducing the signals that promote this differentiation (Burgess, 1998). Cellular interaction with the ECM is critical for the proper formation of tissue architecture and the activation of quality control mechanisms, a process that is likely to be relevant in other organ systems (Kenny and Bissell, 2003; Nelson and Bissell, 2006; Park et al., 2006a; Wang et al., 2002a; Zahir and Weaver, 2004). In breast cancer cells , therefore, factors that alter cell adhesion may interfere with this process, resulting in cells that retain inappropriate proliferation potential and are susceptible to further transformation. Certain colon cancer cell lines, such as SW-1222 and especially Caco-2, express CEA/CC6 at levels similar to normal cells and are capable of forming colon crypt-like structures in culture, so these cell lines were employed to study the effects of CEA on the formation of proper tissue architecture. Over-expression of CEA and CC6 in these human colon carcinoma cells, at comparable levels to those seen in many human colorectal carcinomas, significantly abrogates tissue architecture formation (Ilantzis et al., 2002). Furthermore, human malignancies show an inverse correlation between CEA expression and differentiation, where less differentiated tumors have increased expression levels of CEA and CC6 (Ilantzis et al., 1997; Kodera et al., 1993). The ability of CEA to block the proper formation of tissue architecture may represent another mechanism by which it contributes to malignancy, by allowing cells to grow in an unorganized manner.

12 3.5 Signaling CEA does not have a TM or cytoplasmic domain, so it requires cooperating molecules to signal to the inside of the cell. GPI-anchored proteins are targeted to membrane rafts due to the presence of this anchor (see section 7), microdomains that are highly enriched in signaling molecules, providing the proteins necessary for the propagation of signals initiated by GPI- anchored proteins (Foster et al., 2003; von Haller et al., 2001). Integrins are TM proteins that bind to ECM molecules and localize to rafts upon activation (Baron et al., 2003; Hogg et al., 2002; Krauss and Altevogt, 1999; Leitinger and Hogg, 2002); they also play a role in determining raft properties (Bodin et al., 2005; Gaus et al., 2006; Pankov et al., 2005). Each integrin receptor is composed of one α chain (of which there are 18 different types) and one β chain (with 8 different proteins), leading to 24 known combinations; depending on the combination, a particular integrin receptor will have specific ligands and different downstream signaling properties (Hynes, 2002). The biological outcomes of the activation of certain integrins are very similar to those of CEA, including affecting cellular differentiation and apoptosis; the diversity of integrins means that other processes, including angiogenesis and migration, can also be modulated (Hood and Cheresh, 2002; Pasqualini et al., 1996; Stupack and Cheresh, 2002; Stupack and Cheresh, 2004; Watt, 2002). Several Ig superfamily proteins interact with integrins; for example, the urokinase-type plasminogen activator receptor uPAR associates with and regulates β1 integrins (Monaghan et al., 2004; Wei et al., 2005; Wei et al., 2001; Xue et al., 1997). Due to the similarity between integrin activation and the biological effects of CEA/CC6 expression, it was hypothesized that CEA and CC6’s signaling could occur through integrins. This was examined in rodent, canine and human colorectal cell lines, with the demonstration that the integrin receptor α5β1 is activated upon expression of CEA and CC6 (Ordonez et al., 2007). Additionally, both CEA and CC6 are capable of activating the integrin αvβ3(Camacho-Leal et al., 2007; Duxbury et al., 2004b; Ordonez et al., 2007). In neutrophils, artificial clustering of CEA family members results in increased integrin-ligand interaction and activation of Src family kinases (Nair and Zingde, 2001). Further characterization of CEA-initiated molecular signaling was performed in rat basophilic leukemia cells, where CEA colocalizes with Lck, a Src-family protein tyrosine kinase, and initiates tyrosine phosphorylation (Draber and Stanners, unpublished data). Recent work, performed both in vitro, in L6 myoblasts, and in vivo, in a CEA-transgenic

13 mouse model, has demonstrated that CEA activation results in recruitment of ILK, Akt and MAPK to the plasma membrane and phosphorylation of Akt and MAPK (Camacho-Leal et al., 2007; Chan, Camacho-Leal and Stanners, manuscript submitted). In pancreatic adenocarcinoma cells, signaling by CC6 involves activation of the non-receptor tyrosine kinase FAK, in a c-Src and caveolin dependent manner (Duxbury et al., 2004c). The signaling events that occur upon CC1 activation have been more thoroughly characterized, with the demonstration of a key role in initiating and transducing numerous pathways. The CC1 long cytoplasmic tail is phosphorylated following activation of the epidermal growth factor receptor or the insulin receptor (Abou-Rjaily et al., 2004; Afar et al., 1992; Najjar et al., 1993; Skubitz et al., 1992), while the short cytoplasmic domain can be phosphorylated by protein kinase C (Edlund et al., 1998). CC1 contains two immunoreceptor tyrosine-based inhibitory motifs (ITIMs) in its long cytoplasmic tail, the phosphorylation of which, by Src kinases (Brummer et al., 1995), results in CC1-mediated growth inhibition (Beauchemin et al., 1997; Boulton and Gray-Owen, 2002; Huber et al., 1999). CC1 also associates with various integrins, including the β3 and β1 chains, increasing cellular adhesion and potentially enhancing invasion (Brummer et al., 2001; Muenzner et al., 2005). Many other proteins interact with the cytoplasmic domains of CC1: for the long isoform, these include actin and tropomyosin (Sadekova et al., 2000; Schumann et al., 2001), paxillin (Ebrahimnejad et al., 2000), and filamin A (Klaile et al., 2005); the short isoform interacts with actin and annexin II (Kirshner et al., 2003b; Schumann et al., 2001). Thus, CC1 is involved in signaling through multiple pathways, explaining how it has a role in several cellular processes (see section 3.7). While the signaling abilities of other CEA family members have not been extensively characterized, CC3 contains a immunoreceptor tyrosine-based activating motif (ITAM) that is phosphorylated by Src-family kinases and results in stimulation of Rac and PI3-kinase (Booth et al., 2003; Schmitter et al., 2004), while the recently identified CC20 appears to contain an ITIM motif (Zebhauser et al., 2005).

3.6 Importance of the membrane anchor of CEA family members As enumerated in section 3.2 and 3.3, radical differences, in terms of effects on differentiation and anoikis of L6 myoblasts, are seen when comparing the functions of the GPI-anchored (CEA, CC6) and TM (CC1) members of the CEA family. This suggested that the different methods of

14 membrane anchorage could have a role in determining the function of CEA family proteins. To examine this possibility, chimeric proteins were generated exchanging the GPI anchor of CEA and the TM anchor of CC1, resulting in a protein with the extracellular portion of CEA attached to the TM and short intracellular domains of CC1, and a GPI-anchored protein with the extracellular domain of CC1. When these proteins were examined for effects on L6 differentiation, the GPI-anchored CC1 protein blocked differentiation, while the TM CEA-like protein did not affect differentiation (Screaton et al., 2000). This confirmed that the method of anchorage determined protein function, but it was unclear whether this was a general property of all GPI anchors or specific to the CEA anchor. This was examined by creating chimeras that exchanged the GPI anchors between CEA and GPI-anchored NCAM. NCAM accelerates differentiation of both L6 and C2C12 myoblasts (Dickson et al., 1990; Screaton et al., 2000), so it is functionally different from CEA despite being an adhesive competent protein attached to a GPI anchor, similar to CEA. When the NCAM anchor was replaced with the anchor of CEA, the resulting protein had NCAM external domains (and, importantly, no CEA-derived amino acids in its mature form) attached to the CEA GPI anchor. The resulting protein blocked differentiation, similar to CEA, demonstrating that the biological specificity resided in the CEA GPI anchor (Screaton et al., 2000). From this work, it was suggested that CEA’s biological functions result from a combination of its GPI anchor attached to an adhesive extracellular domain, as non- adhesive mutants are no longer functional (see section 3.2). Thus, not all GPI anchors are functionally equivalent, with the CEA anchor containing important information specifying protein function.

3.7 Other functions Due to the altered expression of CEA family members in many malignancies, studies have primarily focused on possible effects of these proteins in cancer. However, growing evidence indicates that the proteins of this family also have numerous roles in the healthy adult. CEA family members are recognized by many bacterial strains associated with the gut mucosa, including E. coli, Salmonella, Neisseriae meningococci, H. influenzae, N. gonorrhoeae and Moraxella catarrhalis (Hill et al., 2001; Hill and Virji, 2003; Leusch et al., 1991; Sauter et al., 1993; Virji et al., 2000; Virji et al., 1996), leading to the hypothesis that these proteins help regulate bacterial colonization of the colon (Hammarstrom and Baranov, 2001). The interaction

15 of E. coli and Salmonella with human CEA family members appears to be mediated by the carbohydrates on these proteins (Leusch et al., 1991; Sauter et al., 1993), while other bacteria employ direct interactions between cell surface glycoproteins and CEA family members (Berger et al., 2004; de Vries et al., 1998; Dveksler et al., 1993; Hill et al., 2001; Hill and Virji, 2003). There are other effects of bacterial interaction, as binding of N. gonorrhoeae to CEA family members leads to the upregulation of the protein CD105, resulting in increased ECM binding by host cells, as the bacteria attempt to counteract the host defense mechanism of exfoliation of infected cells (Muenzner et al., 2005). The mouse hepatitis virus receptor has also been found to be CC1, so certain viruses may employ the CEA family to infect cells (Dveksler et al., 1991). The CEA family has other roles in host immunity beyond serving as bacterial and viral receptors. Activation of regulatory CD8+ T cells by intestinal epithelial cells is regulated by a complex of proteins on epithelial cells that includes CEA, which serves as a CD8 ligand (Campbell et al., 2002). CC1 is expressed on T cells, and can serve either inhibitory functions, when the ITIM-containing long isoform is expressed, or excitatory functions, when the short isoform is expressed (Chen et al., 2004). CC1 is also involved in the regulation of B cell activation (Greicius et al., 2003), dendritic cell maturation and activation (Kammerer et al., 2001), and in protecting cells from killing by NK cells through homophilic or heterophilic (with CEA) interactions (Markel et al., 2002; Markel et al., 2006; Stern et al., 2005). In addition to its role in immunity, CC1 stimulates angiogenesis, the formation of mature blood vessels, as it is expressed in several locations of the vascular system, including at the fetal- maternal interface, on stenotic aortic valve tissue, and in the microvessels of many human tumors and wounded tissues (Chalajour et al., 2004; Kilic et al., 2005; Prall et al., 1996). CC1 promotes angiogenesis by acting in synergy with VEGF, while treating cells with VEGF upregulates CC1 expression, suggesting that these proteins work together for their effects (Ergun et al., 2000; Wagener and Ergun, 2000). The role of this protein in angiogenesis has been confirmed in CC1 knockout mice, which have defects in their vasculature and are impaired when forming new blood vessels (Horst et al., 2006). Recently, a role for CC1 in tumor blood vessel growth has been established, as CC1 is downregulated in tumor cells and upregulated in the corresponding blood vessels, leading to angiogenesis (Oliveira-Ferrer et al., 2004; Tilki et al., 2006). CC1 can also be phosphorylated by the insulin receptor, suggesting it has a role in glucose homeostasis (Najjar et al., 1995). Using mice with a dominant-negative CC1 expressed in their livers, it was

16 demonstrated that CC1 plays a key role in receptor-mediated endocytosis and degradation of insulin (Soni et al., 2000). In addition, mice lacking a functional CC1 show chronic hyperinsulinemia due to deficient insulin clearance, as well as altered lipid metabolism, increased visceral adiposity and glucose intolerance (Dai et al., 2004; Park et al., 2006b; Poy et al., 2002). Thus, the CEA family is involved in a diverse set of functions in the healthy adult body, and is not restricted to a role in fetal colon development and cancer.

4. The CEA family in cancer

4.1 Expression pattern in cancers The in vitro properties of CEA suggest several mechanisms by which this protein contributes to malignancy, and studies examining CEA and CC6 have demonstrated that these proteins are over-expressed in up to 70% of all human cancers. CEA levels can increase as much as 20-fold in colorectal cancers, while CC6 levels have been observed to increase up to 70 times the levels seen in normal colon tissue (Ilantzis et al., 1997). CEA is primarily expressed in healthy adult humans in the GI tract (section 2.1), but cancers from many locations in the body show CEA upregulation (Table 2). These include cancers originating in the lung, breast, and prostate, urinary bladder cancer, medullary thyroid tumors, female reproductive tract (ovarian, cervical and endometrial) carcinomas, as well as cancers of the GI tract (esophageal, gastric, pancreatic, hepatocellular, gallbladder and colorectal carcinomas) (Ballesta et al., 1995; Barroso and Alpert, 1983; Bhatnagar et al., 2002; Blumenthal et al., 2007; Bojunga et al., 2001; Chevinsky, 1991; Comin et al., 2001; Genega et al., 2000; Hammarstrom, 1999; Harlozinska et al., 1984; Kanthan et al., 2000; Kijima et al., 2000; Kim et al., 1992; Kinugasa et al., 1998; Lloyd et al., 1983; Logsdon et al., 2003; Shi et al., 1994; Ugrinska et al., 2002). CC6, which is more widely expressed in adult humans, is also upregulated in many cancers (Table 2), including lung, breast, pancreatic, hepatic, gastric and colorectal carcinomas, cancer of the female reproductive tract, and acute lymphoblastic leukemia and acute myelocytic leukemia (Allard et al., 1994; Blumenthal et al., 2007; Boccuni et al., 1998; Hanenberg et al., 1994; Jantscheff et al., 2003; Kuroki et al., 1999). Therefore, both CEA and CC6 are consistently upregulated in human cancers of diverse origin. While these proteins are normally restricted to the apical surface of colon cells, in cancer cells they are expressed on the entire cell surface, meaning localization as

17 well as expression is affected (Hammarstrom, 1999). This may result in association with signaling molecules that the protein does not interact with in healthy tissues, further altering its properties. In particular, integrins are normally present on the basolateral surface of colon cells, in contrast to CEA, so altering CEA localization may promote CEA-integrin interactions, resulting in novel CEA-initiated signaling. CC7, another GPI-anchored CEA family member, is downregulated in colon cancer (Boccuni et al., 1998; Hammarstrom, 1999; Harlozinska et al., 1984; Kinugasa et al., 1998; Logsdon et al., 2003; Nollau et al., 1997a; Ohlsson et al., 2006; Sugita et al., 1999; Thompson et al., 1997a; Thompson et al., 1994), but is upregulated in pancreatic and gastric cancers (Kinugasa et al., 1998; Yoshida et al., 2003), indicating that its role in cancer may be tissue specific, and is not as clearly defined. The in vitro properties of CC1, conversely, suggest that it is a tumor suppressor, and as such would likely be downregulated by cancer cells. In rodents, both the mouse and rat versions of CC1 function as tumor suppressors in prostate and colon cancers, as restoring expression of this protein reduces tumorigenicity (Hsieh et al., 1995; Kunath et al., 1995). Humans have a more complicated situation (Table 2), as CC1 expression can be either upregulated or downregulated, depending on the tissue of origin of the malignancy, and, in some cases, the particular study examining its expression. In colorectal cancers, CC1 has been found to be downregulated in some studies (Neumaier et al., 1993; Nittka et al., 2004; Nollau et al., 1997a; Nollau et al., 1997b), while others have found CC1 to be upregulated with levels that correlate with tumor stage (Jantscheff et al., 2003; Yeatman et al., 1997). CEA expression in colon cancer varies depending on the differentiation state of the tumor (Ilantzis et al., 1997), and a similar variability for CC1 may explain these contradictory results, as CC1 is only expressed by well- differentiated colon cells (Hammarstrom, 1999). Tumors that downregulate CC1 include hepatic, breast, endometrial, prostate and renal cell cancers (Bamberger et al., 1998; Kammerer et al., 2004b; Luo et al., 1999; Riethdorf et al., 1997; Tanaka et al., 1997). Other malignancies upregulate CC1, including lung cancer, melanomas, multiple myeloma, acute lymphoblastic leukemia and chronic myelocytic leukemia (Carrasco et al., 2000; Ohwada et al., 1994; Satoh et al., 2002; Sienel et al., 2003; Thies et al., 2002). Thus, the members of the CEA family show altered expression in the majority of cancers, suggesting that they play an important role in carcinogenesis.

18 4.2 Clinical relevance of CEA CEA was originally identified as a molecule that is expressed in colon cancers but not the corresponding normal tissue (Gold and Freedman, 1965). While more recent investigations have concluded that this distinction is not so clear-cut, due to the presence of CEA in certain healthy adult tissues (see section 2.1), CEA retains its clinical value. It was initially theorized that serum CEA levels could be used to detect the presence of cryptic malignancies, as tumors that produce high levels of CEA could shed some of this protein into the blood stream. However, it was determined that serum CEA levels were inadequate to accurately predict the presence of cancer of the liver, colon or breast, particularly at earlier, more treatable, stages, in part because tumor and serum CEA levels are not always quantitatively correlated (Fletcher, 1986; Macnab et al., 1978; Melia et al., 1981; Moertel et al., 1978; Stacker et al., 1988). This is in part because serum CEA levels increase with advancing colon cancer stage, as only 3% of Dukes’ A patients show increased CEA levels, while 65% of Dukes’ D patients show elevated serum CEA (Wanebo et al., 1978). A further complication is that plasma CEA levels can be changed, without the presence of a malignancy, upon exposure to various environmental toxins, including cigarette smoke, and in the presence of nonmalignant diseases such as emphysema and colitis (Alexander et al., 1976; El Far et al., 2006; Fujishima et al., 1995; Rule et al., 1972; Stevens and Mackay, 1973). Thus, while CEA was originally isolated as a tumor specific antigen, its utility as a general screening antigen is quite low. However, as other studies have examined the relationship between CEA and cancer prognosis and diagnosis, CEA has gained renewed clinical importance (Goldstein and Mitchell, 2005). When colorectal cancer patients were divided based on low or high CEA expression, it was noted that elevated serum CEA at presentation is a prognostic factor for poor disease-free survival (Diez et al., 2000; Gasser et al., 2007). Similarly, preoperative CEA levels are predictive for poor outcome in women with stage 1 non-small cell lung cancer (Hsu et al., 2007). Patients with elevated serum CEA at presentation, with levels that do not decrease following surgical resection of the primary tumor, are likely to have occult metastases (Bhattacharjya et al., 2006; Goldstein and Mitchell, 2005; Veingerl, 2001). CEA levels often decrease following surgical resection of colorectal cancers; a return of elevated CEA levels is often the first sign of recurrent disease or metastases, and can be used to initiate other treatments (Behr et al., 1997; Juweid et al., 1996; Minton et al., 1985; Wanebo et al., 1989). Furthermore, the presence of CEA mRNA in the blood following surgery can be used as a negative prognostic

19 factor for both colorectal and biliary-pancreatic cancers (Mataki et al., 2004; Sadahiro et al., 2007). Levels of CEA in the urine are more useful than serum levels for detecting urinary bladder cancer, and the amount of CEA can be correlated with tumor stage (Saied et al., 2007). Thus, while CEA is not the ideal tumor marker that it was originally envisaged to be it still has significant value for predicting the outcome and treatment of cancer patients.

4.3 CEA as a cancer target Cancers have traditionally been treated using one, or a combination of several, of three therapeutic approaches: surgery, radiation and chemotherapy. The type of treatment is chosen based on the location and extent of the tumor, but all three methods have significant drawbacks; for example, the inability of some traditional chemotherapeutic drugs to distinguish between normal and cancerous cells can lead to the indiscriminant killing of too many normal cells in the body, resulting in a multitude of side-effects. As the molecular alterations present in malignancies are elucidated in greater detail, the targeting of various tumor-specific proteins can be employed to specifically kill cancer cells and eliminate many of the side-effects. Because CEA and CC6 are cell-surface proteins that are consistently over-expressed in human cancers, they are attractive cancer targets. In particular, for CEA, with such a restricted expression pattern in healthy tissues (Table 2), the side-effects for any CEA-directed treatments should be greatly reduced. CEA-specific antibodies have been used simply as a targeting mechanism for cancer cells, which are then killed by a radioactive molecule or other cytocidal agent coupled to the antibody. This therapy is particularly useful for the treatment of medullary thyroid carcinoma, resulting in a slight increase in average survival (Chatal et al., 2006; Stein and Goldenberg, 2004). CEA antibodies are also used to detect occult tumor metastases and residual primary tumor cells following surgical resection (Sharkey et al., 2005). Antibodies directed against CEA can be used to retarget the immune system, and in particular cytotoxic T cells and NK cells, to CEA-bearing tumors that have previously evaded immune surveillance (Berinstein, 2002; Kuroki et al., 2005; Kuroki et al., 2004). Furthermore, CEA-based vaccines are currently being evaluated in clinical trials, where preliminary results have demonstrated an induction of T cell responses towards CEA, offering a method to target the immune system and destroy cancer cells (Huang and Kaufman, 2002; Marshall, 2003). While these antibody-based methods have been designed to use CEA as a homing signal for other treatments, the theoretical possibility also

20 exists of directly targeting CEA and inhibiting its biological functions. This should have the effect of restoring the processes of differentiation and apoptosis to these cells, inhibiting their division potential. In vitro, for example, treating CEA-expressing L6 cells with peptides and monovalent fragments of monoclonal antibodies that inhibit its self-binding function restores differentiation to these cells (Taheri et al., 2003). Thus, numerous methods to target cancer cells via CEA are currently being developed, several of which may ultimately improve the health and survival of cancer patients.

5. CEA: in vivo studies

5.1 Tissue implantation studies Based on the in vitro results demonstrating that CEA blocks differentiation and anoikis (section 3), combined with the prevalent over-expression of this protein in many cancers (section 4.1), our group originally hypothesized that CEA plays a role in tumor initiation and/or progression (Benchimol et al., 1989). Various studies have been undertaken to examine this possibility, using both cell lines and, in particular, transgenic mice models. Implanting cell lines ectopically expressing CEA into nude mice increases the tumorigenicity of these lines; for example, studies using the rat L6 myoblast system indicated that CEA expression significantly decreases the latency of tumor formation (Screaton et al., 1997). Human colon cancer cell lines, normally expressing moderate CEA levels but transfected to over-express CEA and CC6 at levels similar to those seen on tumor cells, also had increased tumorigenicity in nude mice (Ilantzis et al., 2002). Thus, it appears that CEA plays a role in cancer in vivo, but confirmation required transgenic mouse models.

5.2 CEA transgenic mice In order to conclusively determine the contribution of CEA to cancer initiation and progression, several transgenic mice lines have been generated. Mice contain only TM members of the CEA family (section 2.3), so introduction of the human CEA gene is required to study its effects in this organism. The first CEA-transgenic mouse expressed CEA under the control of the SV40 promoter, resulting in non-specific production of CEA in all tissues examined. In the epithelial cells of these mice, CEA expression was polarized to the luminal surface, similar to humans

21 (Hasegawa et al., 1991). A more representative expression pattern was required to examine the role of CEA in cancer, so transgenic mice were generated using a cosmid clone that contained the CEA gene and 5 kb of flanking human genomic sequences, to provide the proximal natural transcriptional activation sequences (Eades-Perner et al., 1994). The spatiotemporal expression pattern of CEA in these mice was essentially identical to that of humans, with the protein primarily expressed by the GI tract, demonstrating that the regulatory factors determining CEA expression are conserved between humans and mice. However, one caveat was observed: CEA expression did not show the normal gradient of very low to high expression as colonocytes migrated up the crypts. When these mice were examined for tumors, it was found that the incidence of all cancers was similar between the transgenic and control mice, suggesting that CEA alone does not increase cancer rates. Hypothesizing that CEA cooperated with other oncogenes for its effects, these mice were mated with several tumor models: ApcMin/+ mice, which develop intestinal tumors; SPC-Tag mice, which develop lung tumors; and MMTV-Neu mice, which have increased rates of breast cancer. While CEA was produced by certain tumors, notably those originating in the intestine, the overall tumor incidence and disease progression was comparable between the CEA transgenic and control mice (Thompson et al., 1997b), suggesting that either mice lacked certain elements required for CEA function or that CEA was not involved in carcinogenesis. CEA family members undergo heterotypic interactions (see section 3.1), so the lack of phenotype in the CEA-only transgenic mice potentially indicated a requirement for these interactions for function. Therefore, transgenic mice were generated containing a bacterial artificial chromosome (BAC) with a 187 kb insert that included the genes for CEA, CC3, CC6 and CC7, as well as significant amounts (>20 kb) of flanking human sequences (Chan and Stanners, 2004). These mice also showed spatiotemporal CEA expression almost identical to that seen in humans, including increased CEA levels as colonocytes moved up the crypts, but as in previous models this did not increase spontaneous colorectal tumor formation. BAC-transgenic mice did, however, have double the number of tumors of their wild-type littermates following treatment with azoxymethane, a carcinogen that induces colon tumors (Chan et al., 2006). These tumors upregulated both CEA and CC6, mirroring the situation seen in human colorectal cancer. These mice were generated in a mouse strain (FVB) that spontaneously forms tumors of the lungs, and the CEABAC transgenics had an increased incidence of these tumors compared to

22 controls (Chan and Stanners, manuscript in preparation). The most significant demonstration of the importance of CEA in tumorigenesis came when CEABAC-homozygous mice were generated, which expressed CEA and CC6 in the colon at levels that are similar to those seen in human colon carcinomas. At these elevated expression levels differentiation and apoptosis were both blocked, and the architecture of the colon was completely disrupted, with the resulting morphology showing extreme hyperplasia and dysplasia including regions of neoplastic architecture comparable to that of human colonic serrated adenomas (Chan, Camacho-Leal and Stanners, manuscript submitted). This occurred despite a lack of mutations in the classical pathway for generating colon cancer, confirming that aberrant CEA expression alone can generate dysplasia. Thus, this transgenic mouse model demonstrated conclusively that the GPI- anchored family members of the CEA family can play an instrumental role in the initiation and progression of cancers.

6. The Plasma membrane

6.1 Plasma membrane composition The plasma membrane is composed of both lipids and proteins and surrounds the cell, forming a semi-permeable barrier between the cytoplasm and the extracellular milieu. The membrane is formed of a bilayer of amphipathic lipids, where the hydrophobic tails are buried within the bilayer and the polar headgroups are exposed to the aqueous environment. The fluid mosaic model of Singer and Nicolson attempted to explain the structure and function of the plasma membrane, by postulating that the membrane consists of proteins inserted into a uniform lipid barrier, where lipids and proteins undergo rapid diffusion (Singer and Nicolson, 1972). While this model remains valid in certain situations, it is now apparent that it is too simplistic to explain many phenomena that occur in the membrane. Other characteristics of the membrane include the fact that membrane components migrate within the two-dimensional membrane but not, normally, between the leaflets of the bilayer (Frye and Edidin, 1970; Rothman and Lenard, 1977), and the thickness of the membrane varies between regions (Borochov and Shinitzky, 1976). The two leaflets of the membrane have very different lipid profiles, as the cytoplasmic fraction contains, primarily, phosphatidylserine and phosphatidylethanolamine, and the exoplasmic leaflet contains phosphatidyl-choline and glycosphingolipids (Bretscher, 1973;

23 Storch and Kleinfeld, 1985). Interestingly, the components of a particular leaflet are not homogeneously distributed within the membrane, so heterogeneity also exists in this sense (van Meer and Simons, 1982; Vereb et al., 2003). The plasma membrane contains sterols (notably, cholesterol), which are a major component of the lipid bilayer (Hao et al., 2002; Lange et al., 1989; van Meer, 1989). In addition, the fatty acid components of the lipids in the membrane can be saturated (i.e. containing no double bonds) or unsaturated (with one or more double bonds). The double bonds impose restrictions on the conformation of these chains, such that saturated chains can pack more closely, possibly forming specific domains in the membrane (Koumanov et al., 2004). Ultimately, the great variety of lipids that exists within the plasma membrane results in significant heterogeneity in the cellular membrane, providing intriguing possibilities for the spatial and temporal organization of various cell surface processes.

6.2 Membrane rafts The heterogeneity of the plasma membrane generated the hypothesis that specific lipids, especially sphingolipids, with highly saturated carbon chains, could associate to form specific domains, or rafts, in the membrane (Simons and Ikonen, 1997). Cholesterol also plays a major role in the formation and subsequent stability of these domains (Silvius, 2003), although the existence of these rafts has recently become controversial (Munro, 2003). Evidence for the existence of these domains has been provided by various techniques, including chemical crosslinking, fluorescence polarization, single particle tracking, two-photon microscopy and fluorescence resonance energy transfer (FRET) (Dietrich et al., 2002; Friedrichson and Kurzchalia, 1998; Gaus et al., 2003; Harder et al., 1998; Pralle et al., 2000; Sharma et al., 2004; Varma and Mayor, 1998). However, a large volume of literature questions the presence of these domains in the membrane, with several FRET studies being unable to demonstrate their existence (Glebov and Nichols, 2004; Kenworthy and Edidin, 1998; Kenworthy et al., 2000; Mayor et al., 1994; Munro, 2003). Domains form spontaneously in model membranes that have a composition similar to that of the plasma membrane, lending credence to their existence (Dietrich et al., 2001a; Edidin, 2003; Feigenson and Buboltz, 2001; Silvius, 2003). A significant source of confusion regarding rafts arises from their initial identification as domains that are resistant to extraction by cold, non-ionic detergents (Brown and Rose, 1992; Lichtenberg et al.,

Figure 2. The composition of membrane rafts.

2005), as recent work has suggested that both low temperatures and disruption of membranes with detergents significantly alter membrane structure, potentially resulting in biologically non- relevant lipid aggregation (Heerklotz, 2002; Heerklotz et al., 2003). In partial response to this criticism, it has been determined that rafts exist when extracted with detergent at physiological temperatures (Braccia et al., 2003; Kim et al., 2004), and that raft-like domains can be isolated by non-detergent treatment (Macdonald and Pike, 2005). However, these experiments still required the disruption of the plasma membrane, which may have unknown effects on lipid distribution and properties. Further indirect evidence has arisen from the study of viruses, as

25 many viruses are thought to use rafts for entry into or exit out of cells (Ono and Freed, 2005). In particular, the lipids that compose the HIV envelope, which are taken from the host cell membrane upon virus budding, contain a lipid composition extremely similar to that proposed for rafts (Brugger et al., 2006). The size of rafts remains a further point of contention, as studies suggest that they may be smaller than 5 nm and contain at most 4 proteins (Sharma et al., 2004), whereas others have suggested that rafts may be larger than 50 nm (Simons and Ikonen, 1997; Simons and Toomre, 2000). This discrepancy may be a result of the use of different model systems, as work with model membranes has demonstrated that the size of rafts is related to the composition of the membrane, so different cell lines under variable conditions would likely give different results (de Almeida et al., 2005). The current consensus regarding rafts is that they are small, heterogeneous domains in the membrane that can be stabilized to form larger aggregates by either protein- protein or protein-lipid interactions (Figure 2; Hancock, 2006; Harris and Siu, 2002; Pike, 2006). Rafts contain a specific subset of membrane proteins, which are targeted to these domains either because they have a lipid modification or by information present in their TM domain (Brown and Rose, 1992; Melkonian et al., 1999; Yamabhai and Anderson, 2002). The localization of proteins to specific domains is thought to arise from the formation of an initial shell of lipids around the portion of the protein that is embedded within the bilayer, which then targets the protein to specific rafts (Anderson and Jacobson, 2002). The size of rafts may depend upon protein activation state, as rafts go from, at rest, small, unstable entities to large signaling platforms following receptor activation (Hancock, 2006; Harder et al., 1998; Harris and Siu, 2002; Kusumi et al., 2004). This clustering has been examined in model systems, where a significant increase in overall domain size is observed following crosslinking of GM1, a lipid component normally present in rafts (Hammond et al., 2005). Pulse electron paramagnetic resonance spectroscopy has demonstrated that rafts have three forms: small, unstable resting- state rafts; stable “core” rafts induced by resident protein oligomerization; and signaling rafts, which result from the coalescence of core rafts (Subczynski and Kusumi, 2003). Thus, rafts exist in the membrane in several states, which are significantly influenced by the proteins they are associated with.

26 6.3 Raft heterogeneity It has recently become apparent that multiple raft subtypes exist within the plasma membrane of a single cell (Pike, 2004). This was initially suggested from the observation that the gangliosides GM1 and GD3, which are both components of rafts, segregate to different regions of the membrane, demonstrating that different lipids exist in non-identical domains (Fujita et al., 2007; Gomez-Mouton et al., 2001; Gomez-Mouton et al., 2004; Vyas et al., 2001). Different raft subsets can be isolated by immunoprecipitation, where the precipitation of one raft-resident protein does not co-isolate all other raft proteins, demonstrating that rafts are also heterogeneous in terms of protein composition (George et al., 2006). Isolation of rafts from rat cerebellar granule cells with detergent at 37ºC demonstrated the presence of at least two types of raft, one + + containing the GABAA receptor and the other the Na , K - ATPase (Dalskov et al., 2005). Extracellular proteins are directly targeted to rafts (Table 3), either by having a lipid anchor (see section 7 and table 3) or through information contained in their TM domain (Arni et al., 1998; Brown and Rose, 1992; Scheiffele et al., 1997). For example, the influenza virus protein haemmagglutinin is targeted to rafts, through residues in its TM domain that are in contact with the exoplasmic leaflet (Scheiffele et al., 1997). Certain proteins are targeted to rafts through domains that are not associated with the membrane but present either in the cytosol or outside the cell, such as a cysteine-rich domain in the epidermal growth factor receptor (Crossthwaite et al., 2005; Yamabhai and Anderson, 2002). Other proteins contain a short amino acid region, known

Table 3: Raft Targeting Mechanisms Type Examples Lipid-based GPI anchor myristoylation palmitoylation Protein-based transmembrane residues transmembrane domain length sorbin homology domain (SoHo) membrane proximal regions Pike, 2004

27 as a sorbin homology (SoHo) domain, that mediates interaction of the protein with another protein, flotillin, which determines raft association (Kimura et al., 2001). Therefore, certain proteins contain information within their amino acid sequence specifying the occupancy of particular domains, leading to heterogeneous rafts. GPI-anchored proteins have been examined for similar effects, with the initial demonstration that the domains inhabited by thymocyte differentiation antigen 1 (Thy-1) are heterogeneous and differ depending on Thy-1 activation state (Surviladze et al., 1998). In addition, the domains inhabited by the Prion protein in neuronal cells can be isolated from other membrane domains (Botto et al., 2004). While proteins appear to be specifically targeted to specific domains, there may also be a certain competition for localization to particular rafts, such that an equivalent domain may be very different in different cell types, depending on the proteins each cell expresses (Dietrich et al., 2001b). Thus, significant evidence exists demonstrating that rafts are non-homogeneous in the plasma membrane, in terms of both lipids and proteins. As explored in Chapter 2, GPI anchors can contain signals that direct their localization to specific plasma membrane domains, as the GPI anchor of CEA is able to target proteins to specific membrane domains regardless of the external protein domains that are attached to it.

6.4 Functions of Rafts Proteomic analyses of rafts demonstrate that they are highly concentrated in intracellular signaling molecules, suggesting that external leaflet rafts are coupled to a complement of cytoplasmic signaling proteins (Foster et al., 2003; von Haller et al., 2001). Modeling of the plasma membrane demonstrated a critical role for rafts in promoting and inhibiting specific protein-protein interactions, by selectively concentrating some proteins and excluding others (Nicolau et al., 2006). The targeting of numerous proteins, belonging to multiple functional classes, to rafts has led to the attribution of many properties to these domains. Rafts play a role in intracellular sorting and exocytotic trafficking (Ikonen, 2001), the infective cycle of certain viruses in cells (Bavari et al., 2002; Laliberte et al., 2006; Nayak and Barman, 2002; Takeda et al., 2003), in invasion by pathological microbes (Abraham et al., 2005; Lafont and van der Goot, 2005; Rosenberger et al., 2000) and in signaling (Kabouridis, 2006; Zajchowski and Robbins, 2002). In T cells, clustering of specific membrane proteins results in the co-clustering of various other membrane raft residents, leading to the activation of cell signaling (Janes et al., 1999;

28 Nguyen et al., 2005). This receptor signaling is initiated by a signaling complex that occupies a subset of rafts, with the signals resulting from activation coordinated by the aggregation of different raft domains with specific signaling proteins (He et al., 2005b; Schade and Levine, 2002). Aggregation is thought to stabilize the raft domains and recruit signaling molecules; for instance; clustering of rafts in platelets results in the recruitment of the tyrosine kinase c-Src (Heijnen et al., 2003). The activation of signaling by the T cell receptor results in significant reorganization of the rafts, which affects the protein composition of these domains by recruiting and excluding various proteins (Bini et al., 2003). Due to the raft localization of certain key cell death molecules, such as the Bcl-2 family member Bad, Fas, and caspase-3 and -8, these domains are involved in the control of apoptosis (Aouad et al., 2004; Garcia et al., 2003). Defective raft-based signaling is implicated in cancer; for example, in B cell lineage non- Hodgkin’s lymphomas, NF-κB is constitutively activated due to the presence of a “signalosome” anchored to rafts, resulting in aberrant cellular proliferation (Pham et al., 2002). Thus, many cell surface processes are dependent on raft function.

6.5 Other membrane domains Other domains, distinct from membrane rafts, also exist within the plasma membrane (Figure 3). The protein caveolin serves as a marker for caveolae, invaginations in the plasma membrane that are distinct from, and show different signaling mechanisms from, membrane rafts (Parton, 1996; Sowa et al., 2001). Under certain lysis conditions, caveolae can be isolated separately from rafts, demonstrating that these two domains are not identical (Arvanitis et al., 2005). Ectopic expression of the protein caveolin-1 is sufficient, on its own, to induce the formation of caveolae, indicating that the formation of these domains is mediated by protein-lipid interactions (Lipardi et al., 1998; Vogel et al., 1998). Caveolae play a functional role in endocytosis and signal transduction (Anderson, 1998), and contain various proteins, including GPI-anchored proteins and both receptor and cytosolic tyrosine kinases (Mayor et al., 1994; Mineo et al., 1996). Recent work has suggested that caveolae can be functionally divided into two separate subdomains, the “neck” and the “bulb,” each being distinct regions with different protein components (Foti et al., 2007). Other domains on the cell surface arise through protein-protein interactions (Douglass and Vale, 2005; McCabe and Berthiaume, 2001). These membrane domains include tetraspan

Figure 3. Types of domains formed in plasma membranes. (A) Membrane rafts are present in the exoplasmic leaflet of the plasma membrane, which consist primarily of cholesterol and sphingolipids. (B) The inner leaflet of the plasma membrane is also thought to contain rafts, which concentrate various lapidated signaling molecules. (C) The rafts in the two membranes may become coupled, leading to the association of certain extracellular proteins, such as those with GPI anchors, with signaling molecules on the interior of the cell. (D) Tetraspanins, proteins with four transmembrane domains, have been found to oligomerize in the membrane and create domains that recruit particular proteins, through protein-protein interactions. (E) The protein caveolin alters the structure of plasma membranes, generating a flask-like structure called caveolae that contain a specific subset of membrane lipids. Caveolae are involved in signaling, endocytosis, and bacterial and viral uptake.

30 microdomains, which enrich MHC class II molecules containing certain peptide antigens (Kropshofer et al., 2002). Tetraspanins have specific protein domains that result in protein homo- and heterodimerization, leading to aggregation on the cell surface and the formation of large complexes within the membrane (Kovalenko et al., 2004). Tetraspanins, such as EWI-2, also interact with, and modulate, the functions of other cell surface molecules, including integrins (Berditchevski and Odintsova, 1999; Kolesnikova et al., 2004; Stipp et al., 2003). Integrins can, in turn, modulate the clustering and hetero- and homo- association of tetraspanins, altering the proportion of homo- and hetero-association, leading to the formation of distinct domains (Yang et al., 2006). Thus, tetraspanin domains arise through specific protein-protein interactions, resulting in different complexes of various cell surface proteins with a wide range of biological properties. On the interior of the cell, it has become apparent, through studies with the isoforms of Ras, that certain lipid-modified signaling molecules inhabit distinct, non-overlapping regions on the cytoplasmic surface of the membrane (Lommerse et al., 2005; Niv et al., 2002; Plowman et al., 2005; Prior et al., 2003). The inner cell leaflet domains inhabited by H-Ras, but not those of K- Ras, are sensitive to cholesterol depletion, indicating a difference in lipid composition of these domains (Murakoshi et al., 2004; Niv et al., 2002; Prior et al., 2003). This particular raft association is mediated by the lipid anchor of H-Ras, in particular by the palmitoyl group attached to Cys184, which is not present in K-Ras (Rotblat et al., 2004; Roy et al., 2005). In a direct examination of the effect of external rafts on internal proteins, clustering a cell-surface, raft-associated protein, influenza haemmagglutinin (and thus, its associated rafts) affects the lateral diffusion of H-Ras but not K-Ras, suggesting that internal proteins also become clustered (Eisenberg et al., 2006). Interestingly, proteins that are acylated (with myristoyl and/or palmitoyl moieties), but not those that are prenylated (with farnesyl or geranylgeranyl structures), are the major cytosolic proteins that are targeted to rafts, showing that specific targeting also occurs for proteins associated with this leaflet (Melkonian et al., 1999; Zacharias et al., 2002). Under certain solubilization conditions, inner membrane lipids can be co-isolated with external domain raft lipids, suggesting a linkage between these two domains (Pike et al., 2005). This may arise from the extended length of the acyl chains of the sphingolipids that compose the external rafts, which could interact with internal leaflet lipids (Simons and Ikonen, 1997; Simons and Toomre, 2000). Thus, specific targeting of extracellular and cytosolic proteins to

31 complimentary domains in the membrane results in a complex with defined characteristics. Further stabilization of these cytoplasmic rafts results from protein interactions, as, for example, the proteins Reggie-1/2 serve as scaffolds for domain formation (Stuermer and Plattner, 2005). The heterogeneity of both leaflets of the plasma membrane, therefore, appears to be a critical component for the generation of specific signaling pathways, which transfer extracellular signals to the interior of the cell.

7. GPI-anchored proteins

7.1 GPI anchors During the 1970s, it was demonstrated that a bacterial enzyme, PI-PLC, could release a subset of cell surface proteins from the membrane, suggesting that not all surface proteins were attached to the plasma membrane via the classical TM protein domain (Ikezawa et al., 1976; Low and Finean, 1977; Low and Finean, 1978). However, it wasn’t until the mid-1980s that Ferguson et al. reported that the coat protein of Trypanosoma brucei, which is also sensitive to PI-PLC, was anchored to the membrane by a modification composed of carbohydrates and lipids (Ferguson et al., 1985a; Ferguson et al., 1985b). Since the discovery of the first protein with this type of modification, GPI-anchored proteins (as they are now known) have been identified in all eukaryotes examined, but not prokaryotes (Ferguson and Williams, 1988; Ikezawa, 2002). These GPI-anchored proteins belong to numerous functional classes (Table 4), including adhesion molecules, hydrolytic enzymes, and mammalian and protozoal antigens, as well as other proteins with diverse properties (Low, 1989). The recent demonstration of a GPI-anchored protein that resides in the Golgi apparatus, and plays a role in maintaining Golgi structure, adds another dimension to the properties of these proteins (Li et al., 2007). GPI-anchored proteins also associate with, and are major protein components of, membrane rafts by virtue of their lipid anchors (Brown and Rose, 1992; Schnitzer et al., 1995), and signal through tyrosine kinases, among other proteins (Stefanova et al., 1991).

32 7.2 GPI anchor signal sequence GPI anchors are only added to a subset of surface proteins, leading to the hypothesis that a specific signal within these proteins determines anchor addition. As addition of the GPI anchor occurs in the lumen of the ER, an N-terminal sequence specifying the translocation of the protein across the ER membrane is also required. Studies with placental alkaline phosphatase demonstrated that a portion of the protein was cleaved upon addition of the anchor, suggesting that the signal sequence is removed upon anchor addition (Takami et al., 1988). The initial identification of the signal came with the observation that the addition of an anchor to decay accelerating factor is mediated by its extreme 3’ terminus, as this stretch of amino acids, when

Table 4: Examples of GPI-anchored proteins

Functional Class Examples Enzymes Alkaline phosphatase Acetylcholinesterase 5'-nucleotidase Renal dipeptidase (MDP) Carboxypeptidase M ADP-ribosyltransferase Receptors CD14 CD16 CD48 Folate-binding protein Urokinase receptor Mammalian antigens CEA Thy-1 DAF (CD55) Ly6 family Qa-2 CD24 Protozoal antigens Giardia GP49 Paramecium surface antigens Trypanosoma VSG Toxoplasma surface antigens Others NCAM LFA-3 (CD58) Prion protein Low, 1989; Ikezawa, 2002

33 fused to another protein, is sufficient to give GPI anchorage (Caras et al., 1987; Tykocinski et al., 1988). This led to the characterization of the signal for anchor addition, a signal that has low homology, but similar overall physical properties, between proteins. The site of anchor addition [known as the omega (ω) site] requires small amino acids, in particular, glycine, alanine, serine, cysteine, aspartate or asparagine, as substitution with large, bulky amino acids inhibits processing (Micanovic et al., 1990; Moran et al., 1991). Immediately downstream of this site, the amino acid after the omega site can be any residue except proline, while the amino acid at ω+2 has similar requirements for small amino acids as the ω site (Gerber et al., 1992; Kodukula et al., 1993). Thus, the location of anchor addition consists of a small amino acid domain, which is absolutely required for anchor addition (Aceto et al., 1999; Engle et al., 1995). The signal continues, downstream of this domain, with a series of hydrophilic residues and terminates with a hydrophobic stretch (see Figure 4; Coyne et al., 1993; Low, 1989). Interestingly, comparison of the GPI anchor signal sequence and the N-terminal leader peptide indicated similar requirements for both stretches, suggesting that a similar recognition mechanism exists for both (Yan et al., 1998). The remaining requirements for efficient GPI anchor addition are all determined by overall characteristics of the signal. The length of the signal sequence of placental alkaline phosphatase is of critical importance, as deleting large amounts of the signal, particularly within the hydrophobic region, interferes with processing (Berger et al., 1988; Kodukula et al., 1992). The requirement for a particular length extends to the hydrophobic and hydrophilic regions of the bovine 5’-nucleotidase signal sequence, (Furukawa et al., 1994; Furukawa et al., 1997) and the hydrophilic domains of Thy-1 (Beghdadi-Rais et al., 1993) and acetylcholinesterase (Bucht et al., 1999), demonstrating that efficient anchor addition is also dependent on the presence of a signal of sufficient length. The signal sequence does not need to be folded to be recognized and processed, indicating that the signal resides in the primary amino acid sequence, and not a conformation adopted by this stretch (Spurway et al., 2001). Substitutions within the hydrophobic domain of the folate receptor, notably the replacement of a particular alanine with a proline or aspartate, or introduction of a charged residue in the hydrophobic domain of placental alkaline phosphatase, greatly reduces the efficiency of anchor addition, signifying that a

Figure 4. The GPI anchor signal sequence

minimum level of hydrophobicity in this region is required for efficient anchor addition (Lowe, 1992; Yan and Ratnam, 1995; Yan et al., 1998). The entire signal sequence, including the hydrophobic domain, is translocated into the lumen of the ER, which may explain the requirement for a minimum length of this region, as the hydrophobic domain may serve to properly position the ω site for anchor addition (Dalley and Bulleid, 2003; Wang et al., 1999). The GPI anchor transamidase is also able to recognize the GPI anchor signal sequence when it is located in the interior of a protein, suggesting that the transamidase specifically recognizes the signal sequence, rather than any set of small amino acids followed by a hydrophobic stretch (Caras, 1991). When the GPI-anchor signal sequences of 155 proteins were compared, the consensus signal for anchor attachment was elucidated: an unstructured set of 11 amino acids located upstream of the ω site; four small amino acids located at ω-1 to ω+2; a spacer sequence, extending from ω+3 to ω+8, which is relatively hydrophilic except for hydrophobic residues at ω+4 and ω+5; and a hydrophobic stretch beginning at ω+9 (Eisenhaber et al., 1998). While this represents the consensus, significant variability in the actual amino acids present exists between various proteins. This sequence variability may affect anchoring efficiency, as the efficiency of anchor addition differs depending on the protein destined to be anchored (Chen et al., 2001). The GPI anchor signal sequence from CEA directs the addition of a specific functional anchor, demonstrating that these sequences also specify the addition of a particular anchor (Screaton et al., 2000). This is explored in Chapter 3, where we identify the amino acids that specify the addition of a particular GPI anchor, and demonstrate that a sequence of 6 amino acids in the hydrophilic spacer determines the addition of a functionally-specific anchor.

7.3 GPI anchor addition to preproteins The GPI anchor precursor is synthesized in the endoplasmic reticulum (ER) in a reaction sequence involving at least 19 proteins (Eisenhaber et al., 2003). The protein fated to be anchored is translocated into the lumen of the ER, where the signal sequence is cleaved and the preassembled GPI anchor is added (Figure 5). This transamidation reaction is catalyzed by a complex that recognizes both the GPI anchor precursor and the preprotein destined to receive the anchor, and occurs within 2 minutes of translocation (Bangs et al., 1985; He et al., 1987). The transamidase complex, as currently characterized, is composed of 5 proteins: hGAA1p, GPI8, PIG-S, PIG-T, and PIG-U. hGAA1p is involved in the recognition of the GPI anchor precursor, as deletion of its C-terminus leads to a complex that interacts with the protein to be anchored but not with the anchor precursor (Chen et al., 2003; Vainauskas and Menon, 2004). The catalytic activity resides in GPI8, which not only shows homology to a family of plant endopeptidases, but also forms a covalent bond with the preprotein, as a reaction intermediate, after signal cleavage (Benghezal et al., 1996; Spurway et al., 2001). PIG-S and PIG-T defective cells are unable to form the carbonyl intermediates required for anchor addition, while PIG-T also serves to stabilize the expression of hGAA1p and GPI8 (Ohishi et al., 2001), through the formation of a disulfide bridge with GPI8 (Ohishi et al., 2003). The other four proteins form a non-functional complex when PIG-U is not present, demonstrating its importance. It is thought that PIG-U serves to recognize either the preassembled anchor or the signal sequence, but its specific activity is currently unknown (Hong et al., 2003). However, the precise role of each of these proteins during the anchor addition reaction remains to be determined. Following transamidation, the anchor is often remodeled, notably with the removal of a palmitoyl chain attached to the inositol ring (section 7.4) by an ER-resident protein PGAP1 (Tanaka et al., 2004). The removal of this chain from the GPI anchor is a cell-specific event, as the same protein expressed in different cells will show different levels of acylation of the anchor (Chen et al., 1998).The lipids are also remodeled in the Golgi apparatus by the proteins PGAP2 and PGAP3, resulting in an exchange of the unsaturated lipid chains that are originally added to the precursor with saturated lipids that mediate raft association (Maeda et al., 2007; Tashima et al., 2006).

Figure 5. GPI anchor biosynthesis

7.4 GPI anchor structure The GPI anchor was originally described as a lipid and carbohydrate moiety, and many subsequent studies have determined the precise molecular structure of these anchors in various organisms. The first structure determined was that of the GPI anchor of the variant surface glycoprotein from Trypanosoma brucei, which had an invariant core structure with heterogeneous carbohydrate substitutions along its length (Ferguson et al., 1988). Many proteins, such as the rat brain Thy-1, have a similar core anchor structure, suggesting that this core is conserved among different proteins and different species (Homans et al., 1988). This conserved central structure consists of a linear chain of 6-O-(ethanolamine-PO4)-α-Manp-(1→2)-α-Manp-

(1→6)-α-Manp-(1→4)-α-GlcNH2p-(1→6)-myo-inositol-1-PO4 (Figure 6; Englund, 1993; Thomas et al., 1990). This core is similar in proteins derived from Saccharomyces cerevisiae, Aspergillus fumigatus, Leishmania major and Dictyostelium discoideum, demonstrating its conservation in organisms other than mammals (Fankhauser et al., 1993; Fontaine et al., 2003; Haynes et al., 1993; Schneider et al., 1990). Thus, in most anchors currently analyzed, the core structure remains identical, but significant differences exist in the side chain structures. In certain

37 cell types, the inositol ring in the mature anchor structure is acylated, a hydrophobic alteration that is hypothesized to increase the stability of the anchor’s association with the plasma membrane (Roberts et al., 1988). The structures of GPI anchors can now be efficiently characterized using mass spectrometry (Omaetxebarria et al., 2006), which has been used to determine the structures of the anchors from various mammalian proteins: bovine 5’-nucleotidase (Taguchi et al., 1994; Taguchi et al., 1999), human CD59 (Meri et al., 1996; Nakano et al., 1994), bovine and human acetylcholinesterase (Deeg et al., 1992; Haas et al., 1996; Roberts et al., 1988), and the human prion protein (Baldwin, 2005). A comparison of the anchor structures of renal membrane dipeptidase isolated from humans and pigs demonstrated that the same anchor variants were added to this protein in both species, suggesting that there is information in the protein sequence determining the addition of a particular class of anchor (Brewis et al., 1995). Furthermore, the anchors of both bovine alkaline phosphatase and mouse NCAM demonstrate heterogeneity in the

Figure 6. The structure of the GPI anchor

38 side chains attached to the respective anchors, even when isolated from the same cells (Armesto et al., 1996; Mukasa et al., 1995). The hydrocarbon chains that are attached to the anchor do not appear to have any specific requirements, beyond the observation that they are generally saturated (McConville and Ferguson, 1993), although the saturation of these chains does affect the association of these proteins with membrane rafts (Benting et al., 1999; Maeda et al., 2007). Thus, while GPI anchors contain the same core structure, significant structural variability exists, leading to the hypothesis that not all anchors are equivalent.

7.5 Functional consequences of GPI anchorage Cell surface proteins are attached to the membrane by either a TM domain or a GPI anchor, with the mode of membrane anchorage significantly affecting function. This may be due, in large part, to the localization of GPI-anchored proteins to membrane rafts containing a different subset of signaling molecules compared to the rest of the plasma membrane. Once a GPI anchor is attached to a protein, in the ER, these proteins are sorted to the apical surface of cells, so the anchor determines, at least in polarized cells, protein localization (Paladino et al., 2006; Polishchuk et al., 2004). To examine the consequences of GPI anchorage, chimeric proteins have been generated attaching GPI anchors to the external domains of proteins that normally have a TM domain, or exchanging anchors between two GPI-anchored proteins. This altered anchorage has profound effects on function, as a GPI-anchored form of CD4 associates with different signaling molecules compared to the wild-type TM molecule (Solomon et al., 1998). The association of GPI anchors and signaling molecules was confirmed when a chimera of GFP attached to the GPI anchor of CD59 recapitulated, following antibody clustering to mimic activation, the signaling events initiated by CD59 (Hiscox et al., 2002). NCAM is expressed as both TM and GPI-anchored isoforms, and these proteins show different effects on myoblast fusion, with the TM variants enhancing fusion to a much greater extent (Peck and Walsh, 1993). In neural growth cones, these NCAM variants both localize to rafts, but these rafts are non- identical and contain the signaling proteins PKC, GAP-43 and fyn in different activation states (He and Meiri, 2002). Attaching the GPI anchor of CEA to the external domain of NCAM results in a protein with CEA-like functions, demonstrating that different GPI anchors can also have strikingly different effects on function, and that not all anchors are functionally equivalent

39 (Screaton et al., 2000). Thus, having a GPI anchor or a TM domain can result in significant alterations of the biological properties of a protein, as can different GPI anchors.

7.6 GPI anchor heterogeneity The structures of various GPI anchors show microheterogeneity in the side chains attached to a conserved core structure (see section 7.4), although a correlation between these structural differences and function at the cell surface has not been established. When the folate receptor is expressed in different cells, the anchor has different structures (in certain cells, it is acylated on the inositol ring), differences that correlate with its ability to bind ligand, demonstrating that the anchor can affect the activity of the extracellular domain (Wang et al., 1996). On the surface of a single cell, different GPI-anchored proteins, such as Thy-1 and the prion protein, or the folate receptor and placental alkaline phosphatase, are localized to different regions of the plasma membrane (Madore et al., 1999; Wang et al., 2002b). GPI-anchored, raft-associated Thy-1 does not colocalize with the raft-resident protein LAT or the raft lipid GM1 in the resting state, but after activation it colocalizes with LAT, demonstrating both raft heterogeneity at resting state and that activation can result in coalescence of multiple, different, rafts (Wilson et al., 2004). Additionally, isoforms of Thy-1 are distributed differently on the cell surface and activate separate signaling events after clustering (Heneberg et al., 2006). The possibility exists that the different cell surface localizations of these proteins reflects the affinity of different GPI anchor structures for a particular subset of membrane rafts (section 6.3). Interestingly, the rafts inhabited by the prion protein and Thy-1 have very different lipid profiles, demonstrating that rafts with different lipid compositions exist and concentrate specific proteins (Brugger et al., 2004). Work presented in Chapter 2 of this thesis demonstrates that the GPI anchor of CEA contains information targeting the protein to specific regions of the cell surface, such that regardless of their extracellular domain, proteins with the CEA anchor inhabit the same membrane regions. While the molecular mechanism that defines this specificity remains unknown, this specific targeting of proteins is likely to be the source of the signaling specificity seen for different GPI- anchored proteins.

40 7.7 GPI-anchored proteins and disease CEA and CC6 are upregulated in many human cancers, and various other GPI-anchored proteins have also been found to have roles in cancer. In fact, the subunits of the GPI transamidase complex are upregulated in several cancers, and over-expression of PIG-T or hGAA1p is sufficient to induce tumorigenesis in breast cancer cells (Ho et al., 2006; Wu et al., 2006). This may result in increased processing efficiency, and higher surface expression, of GPI-anchored proteins, several of which appear to function as oncogenes. The expression of many GPI- anchored proteins is altered in tumors: for example, Thy-1 is over-expressed in melanomas, promoting invasion and metastasis, but serves as a tumor suppressor in ovarian cancer (Abeysinghe et al., 2003; Rege and Hagood, 2006a; Rege and Hagood, 2006b); CD24 is over- expressed in many neural cancers (Poncet et al., 1996); downregulation of glypican-3 occurs in ovarian and breast cancers, as well as in mesotheliomas (Lin et al., 1999; Murthy et al., 2000; Xiang et al., 2001); uPAR is involved in invasion and metastasis of lung and GI tract cancers (Andreasen et al., 1997; Laufs et al., 2006; Moller, 1993), and decay-accelerating factor (DAF) is upregulated in colorectal cancers (Koretz et al., 1992). Thus, GPI-anchored proteins show diverse effects on cancers, serving roles that promote or suppress tumor growth. Other diseases also result from the inappropriate expression of GPI-anchored proteins. For example, the prion protein is GPI-anchored (Stahl et al., 1987) and localizes to membrane rafts (Naslavsky et al., 1997). These proteins are responsible for a category of diseases known as transmissible spongiform encephalopathies, which arise when the prion adopts an altered conformation, generating the infectious isoform (Prusiner, 1991). The interaction of the infectious isoform with the normal isoform occurs in rafts, resulting in the conversion of the normal isoform into an infectious particle (Naslavsky et al., 1997; Vey et al., 1996). The conversion to the infectious particle requires the GPI anchor, as a TM-anchored prion protein is resistant to conversion (Kaneko et al., 1997). Another disease, paroxysmal nocturnal haemoglobinuria, results from a defect in GPI anchor biosynthesis, such that red blood cells, leukocytes and platelets do not express GPI-anchored proteins (such as DAF), leading to complement attack and lysis of these cells (Boccuni et al., 2000; Hall et al., 2002). It seems logical, because of the wide variety of proteins that have this type of anchor, that a diverse set of diseases will be characterized that involve GPI-anchored proteins.

41 Because of their lipid portion, GPI-anchored proteins can reincorporate into the plasma membrane from the extracellular medium, in a process commonly termed “painting” (Ilangumaran et al., 1996). These added proteins regain their lipid raft association with time, which correlates with their ability to mediate signaling (Premkumar et al., 2001; van den Berg et al., 1995a). This offers a method to epigenetically modify cell function without the need for DNA transfection, which may eventually provide a more feasible treatment technique for patients lacking certain cell surface proteins. Treating erythrocytes with DAF is sufficient to alleviate the problems seen in patients with paroxysmal nocturnal haemoglobinuria (Medof et al., 1985), while it has been suggested that immunotherapy directed against tumors could be enhanced through the painting of GPI-anchored costimulatory molecules (McHugh et al., 1999). We demonstrate, in Chapter 2, that the GPI anchor alone contains the information required for protein targeting and specific functional inactivation. We have also established that the GPI anchor of CEA is sufficient for membrane incorporation (see Research Appendix), so this may one day offer a novel method for treating cancers over-expressing this protein.

8. Scope of the current work Based on the observation that the biological properties of CEA and CC6 are substantially different from those of CC1, our group hypothesized that the method of membrane anchorage plays a critical role in determining protein function. Work by Screaton et al. verified this hypothesis by demonstrating that attaching the GPI anchor of CEA to the adhesion-competent external domains of TM CC1 and an NCAM variant that is GPI-anchored results in proteins that now function similarly to CEA (Screaton et al., 2000). This, however, also raised many new questions, including: how does altering the anchor so radically affect protein function? If both the anchor and the adhesive property are required for CEA function, can this anchor specificity be targeted to disrupt its function, similar to previous reports that targeted the adhesive ability of CEA (Taheri et al., 2003)? Does a second signal exist within the GPI anchor signal sequence of CEA to specify the addition of a particular functional anchor? Other GPI-anchored proteins have demonstrated the ability to be “painted” on the surface of cells; does this property extend to the anchor of CEA? What are the physical differences between the anchors of CEA and NCAM that explain their functional differences? The work presented in this thesis was undertaken in an effort to find the answers to some of these questions.

42 The specificity of anchors suggested a novel mechanism by which protein function could be modulated, and we examined this through studies co-expressing various combinations of GPI- anchored proteins. We demonstrate, in Chapter 2, that the CEA GPI anchor determines cell surface distribution of the protein (Nicholson and Stanners, 2006). This offered the potential for a novel method of modulating protein function, and we show that targeting the GPI anchor of CEA with “shank-defective” GPI anchors of the same type is sufficient to inhibit protein function. Secondly, we sought to understand the signal that specifies the addition of a particular functional anchor through further structure-function studies of the signal sequence for anchor addition. We hypothesized that a second signal existed within the anchor addition signal from CEA, which determines the specificity of addition of a particular functional GPI anchor. As shown in Chapter 3, we employed CEA/NCAM chimeras to characterize a novel signal sequence that is necessary and sufficient to determine the addition of the CEA GPI anchor (Nicholson and Stanners, 2007). In Chapter 4, we show that the anchors of CC6 and CC7 mediate similar biological effects to that of CEA, but these anchors are not identical to that of CEA (Nicholson and Stanners, manuscript in preparation). Additionally, as described in the Research Appendix, we sought to determine if the anchor of CEA had the ability to incorporate into biological membranes when administered exogenously, similar to other GPI-anchored proteins. We determined the conditions required for efficient insertion of CEA, confirming that its anchor contains the structures required for this function. This work, therefore, provides several novel observations regarding the functional specificity of GPI anchors. We demonstrate that a specific signal of 6 amino acids exists within the signal sequence that determines the addition of a specific GPI anchor, that this GPI anchor then determines cell surface localization of the mature protein, and that the addition of “shank- defective” GPI anchors of the same type can be employed to specifically block the function of the protein. This has, therefore, increased the understanding of the functional relevance of GPI anchorage, which should open up many new fields of study for the proteins with this type of modification, and may ultimately lead to novel in vitro and in vivo methods for specifically modulating GPI-anchored protein function.

43 Publication status of the research chapters presented in this thesis

Chapter 2 was published as: Specific inhibition of GPI-anchored protein function by homing and self-association of specific GPI anchors, T.B. Nicholson and C.P. Stanners. Reproduced from The Journal of Cell Biology, 2006, 175: 647-659. Copyright 2006 Rockefeller University Press.

Chapter 3 was published as: Identification of a novel functional specificity signal within the GPI anchor signal sequence of carcinoembryonic antigen, T.B. Nicholson and C.P. Stanners. Reproduced from The Journal of Cell Biology, 2007, 177: 211-218. Copyright 2007 Rockefeller University Press.

Chapter 4 consists of the manuscript: Exploring the biological properties of the GPI anchors of CEACAM6 and CEACAM7, T. B. Nicholson and C. P. Stanners, manuscript in preparation.

Research Appendix consists of the manuscript: The GPI anchor of CEA mediates membrane incorporation, T. B. Nicholson and C. P. Stanners, manuscript in preparation.

Contribution of Authors

All data and figures presented in this thesis represent my own work.

44 Preface to Chapter 2 This chapter was undertaken in order to examine in detail the contribution of the GPI anchor of CEA to protein function. Previous studies from our laboratory, employing chimeras attaching the GPI anchor of CEA to the external domains of proteins with different biological properties, demonstrated that the GPI anchor of CEA contains information that determines protein function (Screaton et al., 2000). This work suggested that the specific GPI anchor of CEA, in conjunction with non-specific but self-binding external domains, mediates CEA’s effects. Other data from our laboratory had demonstrated that blocking the adhesive function, using peptides or antibodies directed against certain key regions for adhesion, could inhibit function (Taheri et al., 2003). Therefore, the study presented in Chapter 2 of this thesis was undertaken in order to answer two questions: firstly, what is the mechanism by which the GPI anchor plays such a major role in protein function; and secondly, could targeting the anchor result in a similar loss- of-function as seen when the adhesive ability is disrupted. It is demonstrated that the GPI anchor determines the cell surface localization of the protein, and this targeting can be exploited to inhibit protein function.

Chapter 2

Specific inhibition of GPI-anchored protein function by homing and self association of specific GPI anchors

Abstract The functional specificity conferred by glycophosphatidyl inositol (GPI) anchors on certain membrane proteins may arise from their occupancy of specific membrane micro-domains. We show here that membrane proteins with non-interactive external domains attached to the same carcinoembryonic antigen (CEA) GPI anchor, but not to unrelated NCAM GPI anchors, colocalize on the cell surface, confirming that the GPI anchor mediates association with specific membrane domains and providing a mechanism for specific signaling. This directed targeting was exploited by co-expressing an external domain-defective protein with a functional protein, both with the CEA GPI anchor. The result was a complete loss of signaling capabilities (through integrin-ECM interaction) and cellular effect (differentiation blockage) of the active protein, which involved an alteration of the size of the microdomains occupied by the active protein. This work clarifies how the GPI anchor can determine protein function, while offering a novel method for its modulation.

47 Introduction Many cell surface proteins are attached to the membrane by a GPI anchor, which consists of a conserved central structure (Low, 1989) with variable carbohydrate and lipid peripheral components (Homans et al., 1988). GPI anchors can determine protein functional specificity, as switching a transmembrane (TM) domain for a GPI anchor can result in novel function due to association with new signaling elements located in a shared membrane micro-domain (Shenoy- Scaria et al., 1993; Shenoy-Scaria et al., 1992). Membrane rafts, originally defined by their insolubility in cold, non-ionic detergents such as Triton X-100 (Simons and Ikonen, 1997), are small, heterogeneous aggregations of cholesterol and sphingolipids on the cell surface (Pike, 2004; Pralle et al., 2000) that concentrate GPI anchored proteins but also contain other proteins. Although the existence of membrane rafts in vivo has been questioned (Munro, 2003), recent studies using a variety of methods have provided evidence for raft-like membrane micro-domains (Dietrich et al., 2002; Friedrichson and Kurzchalia, 1998; Gaus et al., 2003; Pralle et al., 2000; Sharma et al., 2004; Varma and Mayor, 1998). Such micro-domains may act as signaling scaffolds, determining the identity of a subset of signaling elements, as proteomic analyses have found a high concentration of such proteins in purified rafts (Foster et al., 2003; von Haller et al., 2001), with GPI anchored proteins involved in activating this signaling (Robinson, 1997; Solomon et al., 1998). The existence of heterogeneous raft populations has been inferred from studies showing that different GPI- anchored proteins exist in separate rafts (Li et al., 2003; Madore et al., 1999; Wang et al., 2002b). External rafts with different proteins may each have a defined set of associated cytoplasmic proteins, whereby aggregation of GPI-anchored proteins by external domain self- binding or by multivalent ligand binding could cluster specific rafts, resulting in downstream signaling (Harris and Siu, 2002). Carcinoembryonic antigen (CEA), and the closely-related CEACAM6, are GPI-anchored cell-surface glycoproteins that block cellular differentiation (Eidelman et al., 1993) and inhibit the apoptotic process of anoikis (Duxbury et al., 2004d; Ordonez et al., 2000), effects that appear to be due to the activation of specific integrins (Duxbury et al., 2004a; Ordonez et al., 2007). CEA is up-regulated in many human malignancies (Hinoda et al., 1991; Ilantzis et al., 1997), implying a similar role in human cancer, while the TM-anchored CEACAM1 (CC1) may act as a tumor suppressor (Kunath et al., 1995; Luo et al., 1997).

48 Most CEA family members mediate intercellular adhesion by anti-parallel self-binding (Zhou et al., 1993a) which, together with parallel binding on the same cell surface (Taheri et al., 2003), may result in clustering of rafts containing CEA (Benchimol et al., 1989). Deletion of the last two thirds of the CEA N-terminal domain (∆NCEA) abrogates its adhesive ability, which leads to a loss of differentiation blocking activity (Eidelman et al., 1993). The method of membrane anchorage determines CEA family member activity, as genetically fusing the GPI anchor of CEA to CC1’s external domain creates a differentiation-blocking molecule, while a chimera consisting of the external domain of CEA attached to the TM domain of CC1 does not block differentiation (Screaton et al., 2000). The fact that GPI-anchored NCAM does not block differentiation but can be converted to a differentiation-blocking molecule, denoted NCB (previously “NC blunt”), by swapping its GPI anchor for that of CEA, suggests that the CEA GPI anchor harbors the specificity for the differentiation-blocking function and that the external domains merely function to cluster the molecules, and thus the associated rafts (Screaton et al., 2000). Based on the above model, it should be possible to inhibit the biological functions of CEA (and by implication that of any GPI-anchored molecule whose function is regulated by a similar mechanism) by interfering with clustering. This has been achieved for CEA by mutating regions in its N-terminal external domain responsible for self-binding or by the addition of peptides or monovalent monoclonal antibodies that target these regions (Taheri et al., 2003). We test here a second strategy which exploits the specificity of the CEA GPI anchor: if “shank- defective” or “shankless” CEA GPI anchors that were incapable of self association and clustering were introduced, they could occupy the same rafts as CEA and thus possibly interfere with its clustering. We show that non-functional ∆NCEA inhabits the same membrane microdomains as NCB, since both have the same GPI anchor, but not those of NCAM, and is capable of completely inhibiting NCB’s CEA-like differentiation-blocking activity.

49 Materials and Methods Constructs and Antibodies ∆NCEA is a CEA deletion mutant that has the last 75 amino acids of the N domain deleted, such that it is no longer biologically active (Zhou et al., 1993a). The NCAM splice variant used in this study, p125, is a human GPI-anchored NCAM isoform containing the muscle specific domain (MSD) (Barton et al., 1988). NC blunt (“NCB”) is a chimera of the NCAM p125 external domain genetically fused to the CEA GPI anchor signal sequence (Screaton et al., 2000). The mAbs J22 and D14 bind to internal CEA domains (Zhou et al., 1993a), whereas the epitope of D13 is in the portion of the CEA N domain that is deleted in ∆NCEA and rabbit polyclonal anti- CEA binds to all CEA external protein domains. The mAb 123C3 (Santa Cruz Biotechnology, Inc.) recognizes human NCAM, whereas antibodies H-293, H-104 and M-106 (Santa Cruz Biotechnology, Inc.) recognize the α2, α5 and β1 integrins, respectively. The mAb 47A (De Giovanni et al., 1993) binds to myosin. C20 is a goat polyclonal anti-Fn antibody (Santa Cruz Biotechnology, Inc.).

Cell culture and differentiation assay Cells were grown attached to tissue culture plastic surfaces (Nunc) as previously described (Screaton et al., 2000). In brief, CHO-derived LR-73 fibroblasts were grown in α-MEM with 10% FBS. Rat L6 myoblasts were grown in DME containing 10% FBS (GM), and were subcultured prior to reaching confluency to avoid selecting for non-fusing variants. Cell concentrations were determined using a particle counter (Beckman Coulter). For myoblast differentiation, 104 L6 cells/cm2 were seeded in 60 mm dishes. After 3 days, the media was switched to D-MEM with 2% horse serum (DM). 4-7 days later, cultures were assessed for differentiation by hematoxylin (Sigma-Aldrich) staining and microscopic examination (Screaton et al., 1997), or by lysing and assessing myosin levels by western blotting.

Transfections 100 mm dishes were seeded with 2 or 4 x 105 cells/plate for LR or L6, respectively. 24 hours later, cells were cotransfected, by calcium phosphate co-precipitation, with 5 µg of cDNA, 0.5 µg of pSV2(neo) and 10 µg of carrier DNA isolated from LR-73 cells. Double transfections were performed in the same manner, either by cotransfecting both cDNAs at once (for LR cells) with

50 pSV2(neo) or with 0.5 µg of pBabe(puro) to supertransfect L6 transfectants; transfectants were isolated by selection with 400-600 µg/ml neomycin (G418; Invitrogen) or 1 µg/ml puromycin (Sigma-Aldrich). After 10-14 days, resistant clones were pooled and sorted for high expression by FACS using mAbs J22 or 123C3. Although pooled populations of many clones were used, two independent transfections of L6 cells were performed in order to ensure no clonal variation occurred, with identical assay results. Note that the L6 (NCB) cells were pooled colonies resistant to both G418 and puromycin, and that while data from L6 and LR-73 parental cells is shown, no difference between these cells and pooled, G418 resistant clones transfected with the pSV2(neo) alone has been noted (Ordonez et al., 2007).

FACS analysis Cells were collected with PBS-citrate containing 4 mM EDTA (PBSCE) for NCAM-expressing transfectants (because of the sensitivity of the NCAM external domain to trypsinization; Screaton et al., 2000), or 0.063% trypsin in PBS-citrate, for CEA transfectants. 2.5 x 105 cells were resuspended in ice-cold PBS with 2% FBS (PBSF). Cells were incubated for 30 minutes with mAb at a dilution of 1:50-1:100, washed with PBSF, and incubated with FITC-conjugated goat anti-mouse antibody (Jackson ImmunoResearch Laboratories) diluted 1:100. After a further 30 minutes, cells were pelleted, resuspended in PBSF, and analyzed using a FACScanTM instrument (Becton Dickinson).

Binding to purified ECM components Cells were seeded at 104 cells per cm2 on day 0. On day 2, subconfluent cultures were collected with PBSCE, resuspended in GM, and incubated at 37˚C for 30 minutes. Cells were washed with serum-free media, and resuspended in serum-free media at 4 x 105 cells/ml. 100 µl/well of this suspension was added to wells from a 96-well plate coated with Fn, Vn or collagen I (CHEMICON International, Inc.) and incubated for 1 hour at 37˚C. Wells were washed with PBS containing Mg2+, and adherent cells stained with crystal violet. Wells were washed again

with PBS, and the bound stain solubilized with 0.05M NaH2PO4 pH 4.5 plus 25% ethanol. Staining was quantified with a microplate reader (Bio-Tek Instruments) at 570 nm. Statistical significance was determined using a t test. (http://www.physics.csbsju.edu/stats/t- test_bulk_form.html).

Immunoblotting Proteins in cellular lysates were resolved by SDS-PAGE and transferred electrophoretically to a 0.45 µm PVDF membrane (Millipore). Immunoblotting was performed as previously described (Screaton et al., 2000), with antibody binding detected using ECL Plus reagent (GE Healthcare).

Triton X-100 Solubility and Isopycnic Sucrose Gradient Ultracentrifugation Triton X-100 solubility was determined as described (Screaton et al., 2000). In brief, subconfluent cell cultures were rinsed with PBS, collected with PBSCE and rendered single cell suspensions by passing through a 27-gauge needle. 107 cells/ml were resuspended in ice-cold lysis buffer containing 1% Triton X-100 and the protease inhibitors aprotonin (Roche), leupeptin (Roche) and PMSF (Sigma-Aldrich). Lysates were syringed with a 27-gauge needle, incubated on ice for 15 minutes, and centrifuged at 13 500 g for 20 minutes at 4˚C. Soluble fractions were removed, and the pellets resuspended in 0.9 volumes of lysis buffer and 0.1 volume 10% SDS. The relative amounts of soluble vs. pellet protein were determined by immunoblotting. For sucrose gradient ultracentrifugation, two T175 flasks were seeded with 104 cells per cm2. Two days later, cells were collected with PBSCE, and lysed with 1 ml of 1% Brij-98 (Sigma-Aldrich) in sucrose gradient buffer (10 mM Tris pH 8.0, 140 mM NaCl) containing aprotonin, leupeptin, and PMSF for 30 minutes at 4˚C. 1 ml of ice-cold 80% sucrose was added to this lysate, and overlayed successively with 2 mls of 35% sucrose and 1 ml of 5% sucrose. Lysates were centrifuged, with a Beckman SW55 rotor, for 19 hours at 45 000 RPM at 4˚C. 400 µl fractions were collected from the top of the gradient and equal volumes of each fraction assessed by immunoblotting.

Velocity Sedimentation L6 cells from 2-4 T175 flasks were collected with PBSCE, pooled, and lysed with 500 µl of 1% Brj-98 in sucrose gradient buffer for 30 minutes on ice. The lysate was then added on top of 11 ml of 12.5% sucrose, and centrifuged for 1 hour at 12 300 RPM (approximately 18 700 x g) in an SW41 rotor (Beckman Coulter). Twenty five 460 µl fractions were collected from the top, and assayed by immunoblotting. In certain cases, cells were pretreated with methyl-β-cyclodextrin (Sigma-Aldrich) for 15 minutes at 37˚C prior to Brj-98 lysis, to disrupt membrane rafts.

Chemical Crosslinking and Immunoprecipitation 4 x 105 L6 cells were seeded in three 100 mm dishes for each transfectant. 2 days later, cells were washed with PBS, and incubated, with gentle rocking at 4˚C, with either 1 ml of 1 mM DTSSP (Pierce Chemical Co.) in PBS or with PBS alone. After 1 hour, unconjugated DTSSP was neutralized with 100 mM Tris, pH 7.4. Cells were lysed with 400 µl per plate of 60 mM n- Octyl β-D glucopyranoside (Sigma-Aldrich) in lysis buffer containing protease inhibitors. Lysates were pooled and syringed to reduce viscosity. 1.2 ml of each lysate was precleared by rotation with 75 µl of Protein A/G Plus-Agarose beads (Santa Cruz Biotechnology, Inc.) for 3 hours at 4˚C. Precleared lysates were then diluted with an equal amount of lysis buffer, and divided into three aliquots, receiving no antibody, 5 µg 123C3 or 5 µg J22. Samples were rotated overnight at 4˚C, and then 75 µl of Protein A/G Plus-Agarose beads was added. 3 hours later, the beads were washed 5 times with lysis buffer and resuspended in 75 µl 1x Laemmli sample buffer for analysis by western blotting

Immunofluorescence and Confocal Microscopy L6 transfectants were seeded in 8-well Lab-Tek Permanox chamber slides (Nunc) at a density of 104 cells/well. 2 days later, cells were washed with PBSF and then incubated with primary antibodies 123C3 (at a dilution of 1:100 in PBSF) and Rabbit polyclonal anti-CEA (1:2000 dilution) for 30 minutes at RT. Cells were washed with PBSF, then incubated at RT for 30 minutes, in the dark, with a 1:250 dilution of both Cy2-conjugated goat anti-rabbit and rhodamine-conjugated goat anti-mouse secondary antibodies. Cells were then washed twice with PBSF, and fixed by incubation with 4% formaldehyde for 10 minutes at 4˚C followed by 100% methanol for 20 minutes at 4˚C. Samples were then mounted using fluorescent mounting medium (DakoCytomation). Localization of stained proteins was observed using a LSM 510 Axiovert 100M confocal microscope with a Plan-Achromat x 63/1.4 NA oil differential interfence contrast objective (both Carl Zeiss MicroImaging, Inc.).

Adhesion and aggregate size assays Adhesion assays were performed as described (Zhou et al., 1993a). In brief, 106 LR cells were seeded in 80 cm2 tissue culture flasks (Nunc), and collected 2 days later by incubation with

53 PBSCE. 3 x 106 cells were resuspended in 3 ml α-MEM containing 0.8% FBS and 10 µg/ml DNase I (Roche), syringed with a 27-gauge needle to obtain single cell suspensions, and allowed to aggregate at 37˚C with stirring at 100 rpm using a small magnetic stirring bar. Aliquots were removed at the indicated times, and the cells counted with a hemocytometer to determine the percent single cells. For the aggregate assay, cells were prepared as for adhesion assays, but the number of cells present in each of about 50 multicellular aggregates was scored after 1 hour in suspension.

Soluble fibronectin binding assay On day 0, 104 cells per well were seeded in a 96-well plate. Two days later, cells were washed with PBSF, and then incubated for 30 minutes at 37˚C in D-MEM with 50 µl of 10 µg/ml of human Fn (BD Biosciences), along with, where indicated, 5 µg/ml primary mAb and 30 µg/ml donkey anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories) in order to further cross-link CEA or NCAM constructs. Cells were then washed three times with PBSF, and fixed with 4% formaldehyde. Bound Fn was determined by incubation with anti-Fn Ab, C20, at a dilution of 1:100 in 3% BSA (Sigma-Aldrich) in PBS (PBSB) for 90 minutes at RT, having blocked non-specific binding by incubation for 1 hour at RT with PBSB. Cells were washed with PBSB, and incubated with HRP-conjugated rabbit anti-goat secondary antibody (Jackson ImmunoResearch Laboratories) at a dilution of 1:2500 in PBSB for 1 hour. After incubation with a hydrogen peroxide solution containing ABTS (Sigma-Aldrich), bound Fn was determined with a microplate reader at 405 nm with a reference wavelength of 490 nm.

54 Results Design and construction of experiments To test the hypothesis that the functional specificity of GPI anchors could be exploited to specifically inhibit the activity of GPI-anchored proteins, cells expressing a functional GPI- anchored protein were super-transfected with a “shank-defective” molecule with the same GPI anchor, and assessed for effects on function. The former functioning molecule was NCB, which has NCAM self-binding external domains linked to the CEA GPI anchor (Screaton et al., 2000); the defective molecule was ∆NCEA, which has the same GPI anchor but external domains that are defective in self-binding (Eidelman et al., 1993; Fig. 1A). Since ∆NCEA cannot bind to the external NCAM domain of NCB (Zhou et al., 1990), this combination allowed a study focused on potential interaction between their GPI anchors. ∆NCEA was stably cotransfected into NCB transfectants of rat L6 myoblasts, which are blocked for myogenic differentiation due to the expression of NCB. ∆NCEA was present on the cell surface of the double transfectants at slightly higher levels than NCB, as seen by FACS (Fig. 1B) and western blot (unpublished data). As a control for specificity of effects, double transfectants stably expressing molecules with different GPI anchors were used, i.e., ∆NCEA or CEA, with CEA GPI anchors, and NCAM, with the NCAM GPI anchor. Similar expression levels were also obtained for these transfectants (Fig. 1B).

∆NCEA and NCB exist in close proximity CEA and NCAM appear to exist in separate membrane regions, potentially explaining their opposite biological effects (Screaton et al., 2000). If the GPI anchor alone determines cell surface localization, then molecules with the same GPI anchor should exist in close proximity, while those with different anchors should not. Thus, NCAM and NCB would be expected to have different cell surface distributions, with NCB showing a distribution similar to that of CEA. To test this hypothesis, we examined whether ∆NCEA existed in close proximity on the cell surface to NCB but not to NCAM, using confocal microscopy to examine the cell surface localization of these proteins. The relative surface distribution of NCAM and NCB compared to ∆NCEA was determined after indirect immunofluorescent staining. ∆NCEA showed substantial, although incomplete, colocalization with NCB, while ∆NCEA and NCAM showed essentially no

Figure 1. Surface expression of CEA and NCAM proteins on L6 myoblasts. (A) Schematic representation of proteins used in this study, with their ability to mediate intercellular adhesion and to inhibit myogenic differentiation shown. (B) FACS profiles, after staining with mAbs D14 (anti-CEA) or 123C3 (anti-NCAM) and FITC-conjugated secondary antibody, demonstrate cell surface levels of indicated proteins with fluorescence means in brackets. In co-expressing cells, D14 staining is shown by shaded profiles.

57 Figure 2. ΔNCEA exists in close proximity to NCB but not to NCAM. (A) ∆NCEA co-localizes with NCB but not NCAM. Indirect immunofluorescence was visualized by confocal microscopy, using mouse anti-NCAM and rabbit anti-CEA primary mAbs, followed by Cy2-conjugated goat anti-rabbit (left panel) and rhodamine-conjugated goat anti-mouse secondary Abs (center panel). Merged figures (right panel) demonstrate significant ∆NCEA colocalization with NCB, but not NCAM, as shown by the yellow regions. (B) Crosslinking with DTSSP resulted in a similar pattern of NCB or NCAM high molecular weight bands [with monomer (1x), dimer (2x), trimer (3x) being the predominant bands] on a non- reducing acrylamide gel (top), with similar protein levels seen on a reducing (Red.) gel (bottom). (C) IP of these samples was then undertaken. IP in the absence of DTSSP cross-linking showed specific antibody binding, as no coIP occurred. (D) IP of DTSSP cross-linked samples demonstrated coIP of ΔNCEA with NCB and vice-versa but not with NCAM, demonstrating a specific, close association of ΔNCEA and NCB. IB: immunoblot.

58 co-localization (Fig. 2A). Due to the fact that the incubations were performed at room temperature, the antibodies used for detection may have caused partial clustering of the proteins. This, however, should not affect the heterophylic association in question, as clustering of rafts containing both proteins should not change the final amount of colocalization seen. Indeed, fixing the cells prior to antibody incubation, to avoid clustering, resulted in very similar patterns of colocalization to what is shown in Fig. 2A (data not shown). This therefore suggests that the GPI anchor of CEA is sufficient to determine cell surface localization of a protein. In order to verify these results, L6 co-transfectants were treated at 4˚C (to limit protein diffusion) with the chemical cross-linker DTSSP. Non-reducing western blots demonstrated similar cross-linking patterns for NCB and NCAM, consisting of dimers, trimers and higher molecular weight complexes, both alone and in the presence of ∆NCEA (Fig. 2B). To determine the cellular distribution of ∆NCEA relative to NCAM and NCB, immunoprecipitation (IP) studies of extracts from cross-linked cells expressing similar amounts of these proteins were carried out. The crosslinking approach was undertaken, rather than using detergent lysis, due to the potential effects of detergents on membrane raft structure. IP of extracts from untreated cells did not result in any co-IP (Fig. 2C), confirming the expected antibody specificity. However, IP with an anti-CEA mAb of extracts of DTSSP-treated co-transfectants resulted in the co-IP of a considerable amount of NCB but, importantly, not of NCAM (Fig. 2D). Similarly, IP with an anti-NCAM mAb of extracts of crosslinked co-transfectants showed co-IP of ∆NCEA only in the case of NCB but not NCAM (Fig. 2D). The low proportion of co-immunoprecipitated protein can likely be explained by the lack of interaction between the external NCAM and CEA protein domains, the requirement for close (<12 Å) apposition to be crosslinked, the presence of large levels of monomeric proteins even after crosslinking (Fig. 2B), and the incomplete colocalization seen by confocal microscopy (Fig. 2A). These results demonstrate that proteins with GPI anchors of the same type can exist in close proximity, providing a rationalization for specific interference with protein function.

∆NCEA restores differentiation to NCB-expressing cells Having demonstrated specific co-localization of ∆NCEA and NCB on the cell surface, the effect of this defective protein on NCB’s ability to block differentiation was examined. NCB levels in the ∆NCEA co-expressing L6 transfectants were actually higher than those in NCB-only

59 transfectants, thus validating comparisons between NCB alone and in the presence of ∆NCEA (Fig. 1B; FACS means of 124 vs. 63, respectively). NCB completely blocked differentiation, while co-expression of ∆NCEA with NCB resulted in an almost complete restoration of differentiation, with a fusion index of 78% of that seen for ∆NCEA alone (Fig. 3A and Table 1; P<0.0001). As a control, co-expressing NCAM had no effect on the differentiation block imposed by CEA (Fig. 3A), despite the differentiation-enhancing effects of NCAM (Dickson et al., 1990). To confirm this result, up-regulation of myosin, a biochemical differentiation marker, was examined. ∆NCEA induced myosin production in two independent populations of NCB- expressing cells, as shown by western blot, while NCB alone showed no myosin expression (Fig. 3B), confirming the previous results. Due to the length of the differentiation assay (10 days total), it was possible that a loss of NCB expression caused the differentiation restoration in these co-transfectants. However, no decrease in NCB levels was seen in differentiated ∆NCEA+NCB cultures, as western blots showed higher expression levels in both cell populations than the NCB-alone transfectants for which differentiation was blocked (Fig. 3C). Thus, ∆NCEA expression interfered markedly with the differentiation-blocking function of NCB, presumably via their common feature, the GPI anchor.

Table I. Differentiation of L6 Transfectants

Cell Line % Fusion*

L6 parental 71 L6 (CEA) 0 L6 (ΔNCEA) 64 L6 (NCB) 0 L6 (ΔNCEA + NCB) 50 L6 (CEA + NCAM) 0

* Fusion index was measured as the number of cells containing 3 or more nuclei divided by total nuclei in the field and expressed as a percentage. Values are the mean of 3 different experiments, with 3 fields scored from each experiment.

61 Figure 3. ΔNCEA restores differentiation to NCB-expressing cells. (A) Photomicrographs of L6 transfectants tested for myogenic differentiation after a 7 day incubation in DM. Parental and ΔNCEA cells fused readily, while NCB and CEA completely blocked differentiation; co-expression of ΔNCEA with NCB restored differentiation. (B) Upregulation of the differentiation marker myosin, as determined by Western blot with mAb 47A on 10 μg of total lysate, was seen only in parental, ΔNCEA and two separate populations of ΔNCEA + NCB. No myosin expression was seen in exponentially growing (exp) cultures of NCB or ΔNCEA + NCB transfectants. (C) NCB expression in differentiated cultures; Western blots of 10 μg of total cellular lysate demonstrated that NCB expression remained higher in co- expressing cells than in NCB-alone transfectant cells after 7 days in DM.

Figure 4. Increased ECM binding by NCB transfectants is lost in the presence of ΔNCEA. (A) L6 (CEA) and L6 (NCB) transfectants showed increased binding, compared to parental L6 cells, to immobilized fibronectin and vitronectin, whereas the presence of ΔNCEA in NCB transfectants abrogated this increase. No difference in collagen I binding was seen between parental and transfectant lines. Mean values, +/- SD, from 3 independent experiments are shown (*: p<0.004). (B) LR transfectants showed a similar pattern of binding to fibronectin, with an increase seen for NCB but not for ΔNCEA + NCB transfectants (*:p<0.003). (C) No significant alteration in integrin β1 (top) or α5 (bottom) expression was seen by Western blots on 5 μg of lysate from each L6 cell line.

63 Effects of CEA-like proteins on binding to ECM ∆NCEA releases NCB’s block of differentiation, suggesting that it is interfering with downstream signaling by NCB. CEA signaling has been found to involve activation of the integrin α5β1 in rat myoblast and human colonic cell lines (Ordonez et al., 2007) and the integrin αvβ3 in neuronal cells (Mallette and Stanners, unpublished data). We assessed NCB signaling by incubating single cell suspensions, prepared from exponential cultures, with plates coated with the ECM components fibronectin (Fn), vitronectin (Vn) and collagen I. Either CEA or NCB expression increased binding to both Fn and Vn relative to L6 parental cells (Fig. 4A; P<0.004). Cells expressing ∆NCEA+NCB (and ∆NCEA alone) showed no such increase, demonstrating a complete loss, in the presence of ∆NCEA, of the NCB-mediated increase in ECM binding (P<0.0001). As a control, no difference in binding to collagen I was seen between any of these cell lines (Fig. 4A). In addition, LR-73 (LR) transfectants were tested for binding to Fn, and these cells showed a similar loss of NCB-mediated effects upon co-expression of ∆NCEA (Fig. 4B; P<0.01). The total cell levels of α5 and β1 integrins in the L6 transfectants were assessed by western blot (Fig. 4C) and cell surface levels by FACS (α5 only, unpublished data) and showed only minor differences between transfectants, confirming that the ability of cells to adhere to Fn, rather than changes in integrin surface expression level, was the source of the observed difference (Ordonez et al., 2007); Camacho-Leal and Stanners, in press). Thus, the ability of ∆NCEA to interfere with the NCB-mediated differentiation blockage is correlated with interference of enhanced integrin-ECM interaction promoted by NCB.

NCB raft association is unaltered by the presence of ∆NCEA Signaling by GPI-anchored proteins requires intact membrane rafts (Stulnig et al., 1997). One possible mechanism for the effects of ∆NCEA on NCB functional properties could be by expulsion of NCB from rafts. When this possibility was examined, however, NCB remained primarily insoluble in cold Triton X-100 after co-expression of ∆NCEA in either L6 or LR cells (Fig. 5A and 5B, respectively). Complete cellular lysis was demonstrated by the fact that the integrin α5 chain, an integral membrane protein, was localized in the soluble fractions. As confirmation, isopycnic sucrose gradient ultracentrifugation, where raft associated proteins migrate to the lower density regions of the gradient, was performed on cold Brj-98 lysates of L6 transfectants. Again, no obvious difference was noted between NCB alone and that co-expressed

65 Figure 5. NCB membrane raft association is unaltered in the presence of ΔNCEA. TX-100 solubility assays of L6 (A) and LR (B) cells at 4ºC showed no alteration in membrane raft association of NCB +/- ΔNCEA transfectants with the majority of the protein in the insoluble (P), i.e. raft associated, fraction. As a control for efficiency of lysis, the integrin α5 chain showed essentially no insolubility, with all protein found in the soluble (S) fraction for both L6 and LR parental and transfectant cell lines. Isopycnic sucrose density gradient ultracentrifugation showed similar membrane raft association, as seen by flotation at lower density fractions of NCB in both L6 (NCB) (C) and L6 (ΔNCEA +NCB) (D) transfectants. Integrin α5 had partial raft association, while integrin α2 was not found in lower density fractions. The raft marker GM1 was found exclusively in lower density fractions, while ΔNCEA showed similar distribution to NCB.

66 with ∆NCEA (Fig. 5C and 5D), as almost all of the NCB was present in the low density fractions in both cases. Under these conditions, the α5 integrin chain showed partial raft association for both transfectants, demonstrating that an alteration of α5 localization was not responsible for the lack of NCB function. The distribution of ΔNCEA was also found to be essentially the same as that of NCB, as expected for two GPI anchored proteins. As controls, the α2 integrin chain was found solely in higher density fractions, while the raft lipid GM1 was entirely in the low density fractions. Thus, NCB retained membrane raft association in the presence of ∆NCEA, so that this could not explain the loss of NCB function.

Effect of expression of ∆NCEA on NCB-mediated intercellular adhesion The proteins of the CEA family mediate intercellular adhesion, as does NCAM (Soroka et al., 2003), by external domain self-binding. Such self-binding is required for the differentiation- inhibitory activity of CEA, presumably to effect raft clustering (Eidelman et al., 1993; Taheri et al., 2003). If ∆NCEA interfered with NCB clustering, one might predict a reduction in the ability of NCB to mediate intercellular adhesion. NCAM was used as a control, as it inhabits different rafts from ∆NCEA (Fig. 2C). NCAM and NCB were expressed at very similar levels, with and without ∆NCEA, on the surface of LR cells, thus allowing for quantitative comparisons in adhesion between populations (Fig. 6A). A significant reduction in the strength of NCB- mediated adhesion occurred in the presence of ∆NCEA, as shown by a reproducible decrease of about 20% in the number of aggregated cells in suspension after 2 hours (P<0.001), a difference that was not seen for NCAM-mediated adhesion (Fig. 6B). This was accompanied by a decrease in the size of aggregates in NCB-expressing cells as a result of ∆NCEA co-expression (P<0.0001), which again was not seen for NCAM (Fig. 6C). Thus, introducing the same functional GPI anchor with a defective shank led to a specific reduction in the strength of intercellular adhesion by NCB. Effective intercellular adhesion by GPI-anchored proteins is believed to involve the formation of large, zipper-like structures through the aggregation of multiple proteins and rafts, creating stabilized platforms (Harris and Siu, 2002). The ability of ∆NCEA to interfere with NCB-mediated adhesion is thus consistent with models invoking interference with NCB clustering.

67 ∆NCEA alters the size of NCB-containing rafts One mechanism whereby ∆NCEA could interfere with NCB clustering is by altering the structure of the rafts it is associated with. Therefore, the size of the rafts that NCB occupied was approximated by lysing the cells under identical conditions to those used for isopycnic separation on sucrose density gradients, then separating the lysate by velocity sedimentation through a uniform 12.5% sucrose solution. Under these conditions, NCB was almost entirely raft associated (Fig. 5); therefore, this technique should provide a measure of the size of the rafts inhabited by these proteins. The fractions, which were collected from the top (fraction #1), were assessed by western blot for protein localization, using equal volumes of each fraction. NCB was found to be shifted to fractions farther from the top when ∆NCEA was co-expressed, indicating that it was present in larger complexes under these conditions (Fig. 7A). The distribution of NCAM, on the other hand, was found to be similar whether ∆NCEA was present or not (Fig. 7B), demonstrating that the size of the NCAM complexes was not altered in the presence of ∆NCEA. The distribution of ∆NCEA was very similar to that of NCB in shifting to larger complexes when co-expressed with NCB, while remaining in smaller complexes when co- expressed with NCAM (data not shown). A significant (P<0.05) difference in NCB distribution (Fig. 7C) relative to NCAM distribution (Fig. 7D) upon co-expression with ∆NCEA was demonstrated by densitometric analysis of three independent experiments. This suggests that the presence of ∆NCEA specifically alters the rafts containing NCB. In order to confirm that this was a raft-specific effect, cells were pretreated with methyl-β-cyclodextrin (MβCD) in order to sequester cholesterol and disrupt raft structure. Initially, sucrose gradient ultracentrifugation was performed on lysates of these treated cells, in order to confirm the disruption of the rafts. The distribution of NCB demonstrated that this treatment partially disrupted the rafts, as a portion of the NCB was now present in higher density fractions (Fig. 7E; compare to Fig. 5). When these samples were tested for the size of the complexes that NCB was localized to, it was found that NCB, both alone and co-expressed with ∆NCEA, remained in the first few fractions after velocity sedimentation (Fig. 7F). Thus, treatment with MβCD abrogated the difference seen for NCB complex size following ∆NCEA co-expression, confirming that the difference seen was a raft-mediated effect. While cellular lysis with detergents at low temperatures can affect raft structure (Shogomori and Brown, 2003), the fact that the ∆NCEA-dependent sedimentation difference is seen for NCB but not for NCAM transfectants suggests that it represents a valid

69 Figure 6. ΔNCEA interferes with NCB-mediated intercellular adhesion. (A) LR surface expression (fluorescence means in brackets) of NCAM-like proteins (unshaded profiles) and ΔNCEA (shaded profiles), after staining with mAbs 123C3 and D14, respectively. (B) Adhesion assays of LR (NCB) and LR (ΔNCEA + NCB) (right) and LR (NCAM) and LR (ΔNCEA + NCAM) (left). Note the 20% decrease (*: p< 0.001) in NCB-mediated adhesion in the presence of ΔNCEA, which is not seen in NCAM vs NCAM + ∆NCEA transfectants. Values represent the mean +/- SEM for 4 independent experiments. (C) Aggregate assays of cells from (B), showing that the size of multicellular aggregates created by NCB was significantly reduced in the presence of ΔNCEA. The number of cells per multicellular aggregate was determined, after 1 hour of incubation, in 4 separate fields per experiment. Data represent the mean +/- SEM of 3 independent experiments (*: p<0.0001).

Figure 7. ∆NCEA increases the size of NCB-containing rafts. Cellular lysates were assessed for membrane raft size by velocity centrifugation through a 12.5% sucrose column. Fractions collected from the top were probed for protein localization; note that only the first 13, of 25, fractions are shown, as no NCB/NCAM protein was located in lower fractions. (A) Fractions of L6 (NCB) and L6 (∆NCEA+NCB) were analyzed for NCB, integrin α5 and GM1 distribution by Western blot. (B) Similar analysis of L6 (NCAM) and L6 (∆NCEA+NCAM) lysates was performed. Densitometric quantitation of NCB (C) and NCAM (D) localization was then performed for these fractions. Values represent mean +/- SEM of 3 independent experiments (*: p<0.05). (E) Effect of raft disruption with MβCD on sucrose gradient flotation of NCB. Note the shift of NCB distribution into higher density fractions, when compared to Figure 5. (F) Treatment with MβCD results in a complete loss of the effect of ∆NCEA on NCB complex size. Fractions of cells pretreated with MβCD were collected as in (A), and probed for protein localization. Raft disruption caused NCB to be found solely in the highest fractions, whether ∆NCEA was present or not.

71 increase in raft size. This would indicate a dilution of the NCB concentration in membrane rafts, as an increase in the size of a raft containing the same number of NCB molecules would cause a relative concentration decrease. This would thus reduce the incidence of cis-interactions between the proteins, necessary for clustering, explaining the decrease in intercellular adhesion (Fig. 6B) and the loss of biological function (Figs. 3A and 4A).

Antibody cross-linking restores function of NCB inhibited by ∆NCEA As the mechanism of inhibition of NCB function by ∆NCEA appears to involve interference with clustering, NCB function should be restored by artificial clustering with antibodies. Antibody cross-linking of cell surface proteins induces signaling events, including restoring the defective differentiation-blocking function of ∆NCEA (Taheri et al., 2003) through integrin activation manifested by increased cellular binding of fibronectin (Camacho-Leal et al., 2007). ∆NCEA and NCB co-expressed with ∆NCEA both appear to be non-functional due to defects in protein clustering, so clustering of NCB with antibodies should have a similar effect to what has previously been seen for ∆NCEA. To test if NCB retained the potential to modulate ECM binding in spite of the deactivating effects of co-expressed ∆NCEA, cells in monolayer culture were treated with mAbs directed against the NCAM external domains of NCB, along with secondary antibodies to enhance clustering, and binding of soluble Fn was measured. Several mAbs were used, including J22, which binds to internal CEA domains and as such remains capable of clustering ∆NCEA; D13, a control mAb which has an epitope in the region deleted from ∆NCEA, and 123C3, which binds to the NCAM external domains of both NCAM 125 and NCB. As expected, crosslinking ∆NCEA with J22 but not with D13 resulted in a significant increase in bound Fn (Fig. 8). Similarly, crosslinking of NCB, alone and in the presence of ∆NCEA, increased bound Fn (Fig. 8). Crosslinking NCAM, which does not normally modulate integrins, with the NCAM-specific antibody did not lead to an increase in bound Fn levels, demonstrating the specificity of this effect. The lack of difference in Fn binding between NCB and parental cells, unlike that seen in Fig. 4, in monolayer culture is likely due to the intact ECM surrounding the cells in this assay. This would provide the ligands for integrins that have previously been activated, so that these integrins would not bind to the Fn added to the culture medium. Thus NCB, in the presence of ∆NCEA, remained capable of altering Fn interaction

Figure 8. Antibody crosslinking restores integrin activation. Clustering induced by anti-CEA mAB J22 and anti-NCAM mAb 123C3, plus secondary anti-mouse Abs, of ∆NCEA and NCB, respectively, leads to an increase in binding of fibronectin from the culture media by ΔNCEA- and NCB-expressing cells. A non-binding antibody, D13, as well as crosslinking of NCAM did not lead to such an increase. Values shown are means +/- SD for a representative experiment (*: p<0.005).

73 following antibody crosslinking, which is consistent with the hypothesis that a defect in NCB clustering is created upon introduction of ∆NCEA.

74 Discussion The GPI anchor of CEA contains specific information determining protein function (Screaton et al., 2000) when attached to an adhesive extracellular protein domain. We investigated the hypothesis that interference with the GPI anchor could cause specific inhibition of function, using a defective CEA GPI anchor-bearing molecule, ∆NCEA, co-expressed with NCB, a chimera of the NCAM external domain attached to the CEA GPI anchor. Co-expression of ∆NCEA with NCB resulted in a complete loss of the latter’s biological function.

Determination of specific protein function by GPI anchors TM- and GPI-anchored NCAM isoforms exhibit different effects on myoblast fusion (Peck and Walsh, 1993). Genetically attaching the CEA GPI anchor to the external domains of CC1 and NCAM, two adhesive cell surface molecules that do not inhibit cellular differentiation, results in chimeras that block differentiation, demonstrating functional anchor specificity (Screaton et al., 2000). Anchors may determine associated signaling proteins, as GPI-anchored CD59 can signal through kinases after having been exogenously administered to cells (van den Berg et al., 1995b). Attaching this anchor to GFP also recreates the same signaling events upon antibody cross-linking (Hiscox et al., 2002). Membrane rafts concentrate a wide variety of different proteins on both sides of the plasma membrane; thus, GPI anchor-mediated targeting of a protein to rafts containing specific signaling molecules, in conjunction with clustering through external domain interactions, would explain how the CEA anchor can determine specific protein function.

Existence of distinct membrane raft domains In order to have functionally distinct membrane rafts, these domains should contain different molecules; indeed, the segregation of gangliosides GM1 and GD3 into different domains has been observed (Gomez-Mouton et al., 2001; Gomez-Mouton et al., 2004; Vyas et al., 2001). This may help determine signaling specificity, as a close and specific association occurs between some gangliosides and certain signaling molecules (Iwabuchi et al., 1998; Kasahara et al., 1997). Similarly, GPI anchors are heterogeneous, as differences in both hydrophobic and hydrophilic GPI anchor regions have been documented (Armesto et al., 1996; Ferguson et al., 1988). These raft and anchor variations may be sufficient to create GPI anchor specific membrane domains with different protein repertoires. In support of this idea, human folate receptor and placental AP

75 do not exist in close proximity (Wang et al., 2002b), while Thy-1 and prion protein associate with different domains (Madore et al., 1999) that show major differences in lipid composition (Brugger et al., 2004). CEA and NCAM do not co-patch on the cell surface (Screaton et al., 2000), which was confirmed by IP and co-localization results presented here (Fig. 2). Replacing the GPI anchor signal sequence of NCAM for that of CEA, however, produces a chimera (NCB) partially located in close proximity to ∆NCEA (Fig. 2). This incomplete co-localization is likely a result of the small size of membrane rafts prior to activation, as it has been suggested that under resting conditions, GPI anchored proteins are primarily found as monomers, with a proportion found in small clusters (Sharma et al., 2004). NCB affects cell-Fn interaction similarly to CEA (Fig. 4A) but unlike NCAM, suggesting that this altered localization affects signaling. Different ras isoforms inhabit separate cytoplasmic leaflet compartments (Prior et al., 2003), showing that differences in protein distribution on the cell surface may be mirrored on the cell interior. We suggest that the carboxy-terminal primary amino acid sequence can direct the addition of a specific type of GPI anchor, which determines both membrane localization and function through associated signaling elements.

Role of membrane raft clustering in signaling Intense study of membrane raft dynamics has been undertaken in T cells, where these domains have a key role in signaling upon activation of the T cell receptor. Co-stimulation of T cells by CD28 is caused by raft redistribution, which alters internal tyrosine phosphorylation patterns (Viola et al., 1999), while antibody crosslinking of raft lipids replicates the membrane rearrangements that occur upon receptor activation (Janes et al., 1999). CEA mediates intercellular adhesion by antiparallel binding (Zhou et al., 1993a) and differentiation inhibition by both antiparallel and parallel binding (Taheri et al., 2003), both of which may cause clustering of the rafts inhabited by CEA. The discovery of CEA mutants that retain intercellular adhesive ability but do not block differentiation indicates the key role, in conjunction with anti-parallel adhesion, of parallel interactions in CEA’s biological function (Taheri et al., 2003). NCAM- mediated adhesion involves intermolecular cis- and trans-interactions (Rao et al., 1992; Soroka et al., 2003), so at a threshold cell surface density, CEA and NCAM could create large cell surface raft aggregates leading to signaling (Harris and Siu, 2002). CC7, a weakly adhesive member of the CEA family, poorly activates integrin α5β1; however, attaching its anchor to the stronger

76 adhesive external domain of CC1 results in increased integrin activation (Zhai and Stanners, unpublished data). Previous work has demonstrated that treatment of CEA-expressing cells with fragment antigen binding antibody fragments (Fabs), to reduce protein clustering, is sufficient to release the CEA block of differentiation (Taheri et al., 2003). Similarly, reducing initial cis- clustering, through the addition of molecules with defective non-adhesive extracellular domains attached to the same GPI anchor, could also interfere with effective protein and raft clustering, causing substantial effects on downstream signaling. ∆NCEA co-expression resulted in a reproducible decrease in intercellular adhesive strength of NCB-expressing cells (Fig. 6B) and an observed loss of cell-Fn interaction (Fig. 4A). ∆NCEA expression increased the size of the membrane rafts containing NCB (Fig. 7A), which would dilute the NCB molecules within the raft compartment. Thus, a larger raft with the same number of NCB molecules would lead to less efficient cis-clustering, decreasing the resulting adhesive strength. Consistent with this model, mimicking the clustering seen upon adhesion by antibody-mediated cross-linking of the NCB molecules restored its effects on integrin-Fn interaction (Fig. 8). Therefore, it would appear that treatment with Fabs or non-functional GPI anchors leads to similar effects on clustering and protein activity.

Modulation of raft lipid content The function of membrane proteins can be altered by modulating membrane lipid content. Depletion of cellular cholesterol inhibits signaling from membrane raft domains (Incardona and Eaton, 2000), while administration of exogenous gangliosides displaces GPI-anchored proteins from membrane rafts (Simons et al., 1999). Treatment of T cells with polyunsaturated fatty acids (PUFAs) interferes with tyrosine kinase activation and calcium release upon stimulation (Stulnig et al., 1998), which appears similar to this study, where signaling by NCB is inhibited via its GPI anchor. The key difference lies in the GPI anchor specificity, since exogenous administration of lipids likely causes a global alteration in raft structure, while GPI anchors contain information which target specific subdomains. Thus, ∆NCEA is not found in close proximity to NCAM, but changing the anchor of NCAM for that of CEA alters the localization of this protein such that it now colocalizes with ∆NCEA (Fig. 2). This colocalization explains how ∆NCEA, which shows no effect on differentiation, can restore differentiation to NCB-expressing cells, while NCAM, which accelerates differentiation, does not restore fusion to CEA-expressing cells (Fig. 3A). It is

77 intriguing to note that ∆NCEA expression has such a significant effect on NCB function despite the fact that the proteins are found on the cell surface at similar levels. This suggests that once a threshold number of proteins have been inhibited, the biological function is completely lost, as any remaining functional proteins may be incapable to cluster sufficiently to lead to effective signaling activation. We therefore suggest that the GPI anchor plays a key role in protein function by directing localization to a specific subset of membrane rafts, which determines the associated signaling molecules. Exploiting this biological specificity by competition with functionally specific GPI anchors attached to non-functional external protein domains results in a complete loss of biological activity. This therefore confirms the key biological role of the GPI anchor, and suggests a novel method for the manipulation GPI-anchored proteins.

78 Addendum to Chapter 2: Effect of targeting the GPI anchor on anoikis resistance

This addendum contains additional information that was not included in the version of Chapter 2 that was published in The Journal of Cell Biology. The data presented in Chapter 2 demonstrates that co-expression of ΔNCEA in cells expressing NCB results in a complete loss of NCB function, as demonstrated by a loss of integrin modulation (Fig. 4) and a restoration of differentiation (Fig. 3). Another important property of CEA and NCB is the ability to block anoikis, a process by which cells that lose their anchorage to their extracellular matrix die by apoptosis (Ordonez et al., 2000); as such, this is a tumorigenic property, since it allows cells to survive after breaking away from their normal tissue. The effect of ΔNCEA on this effect of NCB was therefore examined. Cells in the late phase of cell death lose the integrity of their membrane, such that DNA staining dyes, such as propidium iodide, have access to the nucleus, which can be quantitated by FACS analysis (Nicoletti et al., 1991). Thus, 1.5 ml of a cell suspension (2 X 105 cells/ml) in GM was added to ultra low cell attachment hydrogel plates (Corning) for 48 hours. The cells were then collected and stained with propidium iodide (2 µg/ml) prior to FACS analysis (as described in the Materials and Methods section of Chapter 2). FACS profiles of L6 transfectants treated in this manner demonstrated a substantial decrease in the amount of strongly stained cells for L6 (NCB), compared to parental and NCAM controls (Addendum figure, top). However, L6 (ΔNCEA + NCB) cells had a profile very similar to that of L6 (NCAM). The amount of cell death in each cell line was then approximated by determining the percentage of cells that showed increased staining above the background (which was determined using cells that had been permitted to adhere; data not shown) (Addendum figure, bottom). This analysis indicated that approximately half of the control parental (48%) and NCAM (44%) cells showed cell death, while NCB transfectants showed significantly (P<0.0001) less cell death (18%). Co-expressing ΔNCEA with NCB resulted in a complete loss of this blockage (P<0.0001), as 50% of these cells were observed to be in the late stages of cell death. Therefore, this data demonstrates that the interference of ΔNCEA on NCB’s biological properties can be extended to anoikis.

Figure 1. Effect of targeting the GPI anchor on anoikis resistance. FACS profiles of propidium iodide- stained L6 cells demonstrate a decrease in staining for L6 cells expressing NCB, an effect that was released by the presence of ∆NCEA (top). Quantitation of these results confirmed that this decrease occurred only for NCB-alone expressing cells, but not for the co-expressing cells (bottom). Data represent the mean +/- SEM for 4 independent experiments.

80 Preface to Chapter 3 The work presented in Chapter 2 indicated that the GPI anchor signal sequence, despite being cleaved in the ER, determines the surface localization and function of the mature protein. This suggested that this signal sequence did not simply specify the addition of an anchor, but also determined the particular functional anchor to be added. We therefore hypothesized that a second specificity signal existed within this stretch of amino acids, and generated chimeras between CEA and NCAM to test for the presence of this signal. The results presented in Chapter 3 identify a novel, short stretch of amino acids that are necessary and sufficient to specify the addition of the CEA-specific GPI anchor.

Chapter 3:

Identification of a novel functional specificity signal within the GPI anchor signal sequence of carcinoembryonic antigen

82 Abstract Exchanging the glycophosphatidylinositol (GPI) anchor signal sequence of neural cell adhesion molecule (NCAM) for the signal sequence of carcinoembryonic antigen (CEA) generates a mature protein with NCAM external domains but CEA-like tumorigenic activity. We hypothesized that this resulted from the presence of a functional specificity signal within this sequence and generated CEA/NCAM chimeras to identify this signal. Replacing the residues (GLSAG) 6-10 amino acids downstream of the CEA anchor addition site with the corresponding NCAM residues resulted in GPI-anchored proteins lacking the CEA-like biological functions of integrin modulation and differentiation blockage. Transferring this region from CEA into NCAM in conjunction with the upstream proline (PGLSAG) was sufficient to specify the addition of the CEA anchor. Therefore, this study identifies a novel specificity signal consisting of six amino acids located within the GPI anchor attachment signal, which is necessary and sufficient to specify the addition of a particular functional GPI anchor and, thereby, the ultimate function of the mature protein.

83 Introduction Glycophosphatidylinositol (GPI) anchorage is a common feature of surface proteins that leads to membrane raft localization. The observation of different functional GPI anchors (Nicholson and Stanners, 2006; Screaton et al., 2000), as well as the fact that different rafts show markedly different lipid and protein profiles (Brugger et al., 2004; Madore et al., 1999; Wang et al., 2002c), implies the existence of a heterogeneous set of anchors and matching rafts. Anchor addition is determined by the GPI anchor signal sequence, which consists of a set of small amino acids at the site of anchor addition (the ω site), followed by a hydrophilic spacer and ending in a hydrophobic stretch (Low, 1989). Cleavage of this signal sequence occurs in the ER prior to the addition of an anchor with conserved central components (Low, 1989) but with variable peripheral moieties (Homans et al., 1988). Carcinoembryonic antigen (CEA) is a GPI-anchored protein, whereas the closely-related CEACAM1 (CC1) contains a transmembrane (TM) domain. In vitro, both proteins mediate intercellular adhesion (Benchimol et al., 1989; Rojas et al., 1990), but CEA, not CC1, blocks cellular differentiation (Eidelman et al., 1993) and inhibits the apoptotic process of anoikis (Ordonez et al., 2000; Soeth et al., 2001). Exchanging the membrane anchors of these proteins results in a TM version of CEA that does not show CEA-like activity and a GPI-anchored CC1- like protein that now exhibits CEA-like properties, demonstrating the importance of the membrane anchor (Screaton et al., 2000). Replacing the GPI anchor signal sequence of neural cell adhesion molecule (NCAM) for that of CEA results in a functionally CEA-like protein (NCB), demonstrating the existence of functionally specific anchors whose addition is determined by a particular signal sequence (Screaton et al., 2000). The sole role of the external domains is to mediate self-binding and, thereby, concentration-dependent clustering (Taheri et al., 2003; Camaco-Leal et al., 2007), as external domain mutations that disrupt the self-binding of CEA abrogate CEA function, while irrelevant self-binding external domains (such as that of NCAM in NCB) suffice for function. Because the amino acid sequence of various GPI anchor signal sequences radically affects protein function, we hypothesized that a signal existed within the GPI anchor signal sequence specifying the addition of a particular functional GPI anchor. Chimeras were generated by exchanging fragments of the CEA and NCAM GPI anchor signal sequences and were tested for CEA-like biological properties. We identify a specificity signal, consisting of five amino

84 acids, that is necessary and sufficient in conjunction with an upstream proline for the addition of the CEA-specific GPI anchor.

85 Materials and Methods Constructs All chimeras were generated by PCR overlap extension. The ω site of NCAM remains unknown and was assigned to be A736 on the basis of sequence alignment with chicken NCAM, in which other potential anchor addition sites were not conserved between humans and chickens and, as such, are unlikely to serve as the ω site (Screaton et al., 2000). Constructs were generated using NCAM, NCB or C1-C cDNA [Note that the C1-C used in this study did not contain the I to F point mutation described in the original study (Screaton et al., 2000), and was called 1C to allow for differentiation between the two proteins] and the primers indicated in Table 1. Initial PCR reactions involved separate extensions using the CC1 or NCAM sense (S) primer with the corresponding antisense (AS) chimera primer and the sense chimera primer with the antisense CEA or NCAM primer. These fragments were joined by overlap PCR using the CC1 or NCAM sense primer and the CEA or NCAM antisense primer. The resulting NCAM-like fragments were inserted into the EcoRI sites of NCAM (at positions 1616 and 2794) in the p91023b expression vector. The C1-C chimeras replaced the corresponding sequence in C1-C using the internal CC1 BamHI digestion site (at position 971) and the BamHI digestion site located in the polylinker region of the vector pEGFP-C2 (used as a cloning vector; BD Biosciences) and were subcloned into p91023b using flanking EcoRI sites.

Cell Culture, FACS analysis and Transfection CHO-derived LR-73 fibroblasts and rat L6 myoblasts were cultured as previously described (Nicholson and Stanners, 2006). Cell surface protein expression was determined by FACS analysis, using the mouse mAbs J22, which recognizes CEA and CC1 (Zhou et al., 1993a), and 123C3 (Santa Cruz Biotechnology, Inc.), which recognizes the NCAM external domain. Transfection and sorting for high expression was performed as previously described (Nicholson and Stanners, 2006).

86 Table 1. Primers employed to generate the chimeras used in this study

Primers Primer Sequence NCAM S 5’-GGGCAGGAGTCCTTGGAATTCATCCTTGTTCAAGCAGACACCC-3’ AS 5’-CCACACAGAATTCTTGCTCAGC-3’ CC1 S 5’-GTGAATAATAGTGGATCCTATACCTGCCACGCC-3’ CEA AS 5’-CCATGCGGATCCGAATTCTGGAGCGACCACATAGGG-3’ Chimera Primers Primer Sequence NCBΔ5 S 5’- GCAACCTTGGGAGGCAATCCTGGTCTCTCAGCTGGG-3’ AS 5’-CCCAGCTGAGAGACCAGGATTGCCTCCCAAGGTTGC-3’ NCBΔ10 S 5’-AATTCTGCATCCTACACCGGAGCCACTGTCGGCATC-3’ AS 5’-GATGCCGACAGTGGCTCCGGTGTAGGATGCAGAATT-3’ NCBΔ15 S 5’-ACCTTTGTCTCATTGCTTATCATGATTGGAGTGCTG-3’ AS 5’-CAGCACTCCAATCATGATAAGCAATGAGACAAAGGT-3’ NCBΔ20 S 5’-CTTTTCTCTGCAGTGACTCTGGTTGGAGTTGCTCTG-3’ AS 5’-CAGAGCAACTCCAACCAGAGTCACTGCAGAGAAAAG-3’ 1CN5 S 5’-ATCACAGTCTCTGCAACCTTGGGAGGCAATGGTCTCTCCGCGGGG-3’ AS 5’-CCCCGCGGAGAGACCATTGCCTCCCAAGGTTGCAGAGACTGTGAT-3’ 1CN10 S 5’-TCTGGTACCTCTCCTTCTGCATCCTACACCGCCACTGTCGGCATC-3’ AS 5’-GATGCCGACAGTGGCGGTGTAGGATGCAGAAGGAGAGGTACCAGA-3’ 1CN15 S 5’-GGTCTCTCAGCTGGGTTTGTCTCATTGCTTATGATTGGAGTACTG-3’ AS 5’-CAGTACTCCAATCATAAGCAATGAGACAAACCCAGCTGAGAGACC-3’ NC5 S 5’-TCTGGTACCTCTCCTTCTGCATCCTACACC-3’ AS 5’-AGGAGAGGTACCAGATGCTGGGATGGCTG-3’ NC10 S 5’-GGTCTCTCCGCGGGGTTTGTCTCATTGCTT-3’ AS 5’-CCCCGCGGAGAGACCATTGCCTCCCAAGG-3’ NC15 S 5’-GCCACTGTCGGAATCTTTTCTGCAGTGACT-3’ AS 5’-GATTCCGACAGTGGCGGTGTAGGATGCAG-3’ NΔ15C S 5’-CCTGGTCTCTCAGCTGGGGCCACTGTCGGCATCTTCTCTGCAGTGACTC-3’ AS 5’-CCCAGCTGAGAGACCAGGAGAAGTTCCAGATGCTGGGATGGCTGTGGGC-3’ NΔ110C S 5’-TCTGGAACTTCTCCTGGTCTCTCAGCTGGGTTTGTCTCATTGCTTTTC-3’ AS 5’-CCCAGCTGAGAGACCAGGAGAAGTTCCAGATGCTGGGATGGCTGTGGGC-3’ NΔ615C S 5’-GGTCTCTCAGCTGGGGCCACTGTCGGCATCTTCTCTGCAGTGACTC-3’ AS 5’-GATGCCGACAGTGGCCCCAGCTGAGAGACCATTGCCTCCCAAGG-3’ NC7 S 5’-CCTGGTCTCTCAGCTGGGGCCGTCTCATTGCTTTTCTCTGC-3’ AS 5’-GGCCCCAGCTGAGAGACCAGGGCCTCCCAAGGTTGC-3’ NC9 S 5’-TCTCCTGGTCTCTCAGCTGGGGCCACTTCATTGCTTTTCTCTGC-3’ AS 5’-AGTGGCCCCAGCTGAGAGACCAGGAGATCCCAAGGTTGCTGG-3’ NC11 S 5’-ACTTCTCCTGGTCTCTCAGCTGGGGCCACTGTCTTGCTTTTCTCTGCAG-3’ AS 5’-GACAGTGGCCCCAGCTGAGAGACCAGGAGAAGTCAAGGTTGCTGGGATG-3’ NC13 S 5’-TCTCCTGGTCTCTCAGCTGGGGCCACTGTCGGCCTTTTCTCTGCAGTG-3’ AS 5’- AGTGGCCCCAGCTGAGAGACCAGGAGAAGTTCCGGTTGCTGGGATGGC-3’ NCB-K2 S 5’-GCTTCAGGTGGGCTCGCCACTGTCGGCATC-3’ AS 5’-GAGCCCACCTGAAGCAGGAGAAGTTCCAGA-3’ NCB-S2 S 5’-TCAGGTCTCGGGGCTGCCACTGTCGGCATC-3’ AS 5’-AGCCCCGAGACCTGAAGGAGAAGTTCCAGA-3’ 1CN2-1 S 5'-CTCCTTCTGCATCAGCTGGGGCCACTGTCGGC-3’ AS 5'-CCCCAGCTGATGCAGAAGGAGAAGTTCCAG-3’ 1CN2-2 S 5'-CTCCTGGTCTCTCATACACCGCCACTGTCGGCATC-3’ AS 5'-GACAGTGGCGGTGTATGAGAGACCAGGAGAAGTTCC-3’ 1CN3 S 5'-CTCCTGGTGCATCATACGGGGCCACTGTCGGCATC-3’ AS 5'-GACAGTGGCCCCGTATGATGCACCAGGAGAAGTTCC-3’ 1CN4-1 S 5'-CTCCTTCTGCATCATACGGGGCCACTGTCGGCATC-3’ AS 5'-GACAGTGGCCCCGTATGATGCAGAAGGAGAAGTTCCAG-3’ 1CN4-2 S 5'-CTCCTGGTGCATCATACACCGCCACTGTCGGCATC-3’ AS 5’-GACAGTGGCGGTGTATGATGCACCAGGAGAAGTTCC-3’

87 Differentiation assays L6 myoblasts were seeded at 104 cells/cm2 in 60 mm dishes in medium containing 10% FBS on day 0. 3 days later, the culture medium was changed to 2% horse serum. Myogenic differentiation was assayed 5 days after changing media either by staining with hematoxylin (Sigma-Aldrich) to assess fusion into multinuclear myotubes by light microscopy (Screaton et al., 2000), or by lysing cells and performing western blots for myosin expression using mouse monoclonal mAb 47A (De Giovanni et al., 1993). Photomicrographs of representative fields of stained cells were obtained at room temperature using a Nikon Eclipse E800 microscope and a 10X NA 0.30 Ph1 ∞/0.17 objective. Images were acquired with a Nikon DXM1200 digital camera and ACT-1 image acquisition software (Nikon). The fusion index was determined by counting the number of nuclei present in fused myotubes (taken as cells with three or more nuclei) and comparing this to the total number of nuclei in the field.

TX-100 Solubility and PI-PLC Sensitivity Assays Assays were performed essentially as described previously (Screaton et al., 2000). For TX-100 solubility, cells were collected with PBSCE, and resuspended in cold 1% TX-100 with protease inhibitors. Cells were syringed through a 27-gauge needle, incubated on ice for 15 minutes, and centrifuged at 13 500 g for 15 minutes. The supernatant fraction was removed, and the pellet was resuspended in the same volume as the supernatant. Partitioning between pellet and soluble fractions was assessed by immunoblotting; the integrin α5 should be found in the supernatant (soluble) fraction and was used as a lysis control, with detection by a rabbit polyclonal anti-α5 (H-104; Santa Cruz Biotechnology, Inc.). For PI-PLC sensitivity, monolayer cultures were incubated with 0.1 U bacterial PI-PLC (Sigma-Aldrich) in a 1:1 solution of DMEM/PBS containing 0.2% BSA for 45 minutes at 37°C. Treated and control untreated cultures were then washed with PBS, rendered single cell suspensions by light (0.063%) trypsin treatment, and processed for FACS analysis. Percent sensitivity was determined as the percent decrease in mean fluorescence value (relative units) in the treated sample compared with the untreated control.

Adhesion assays Adhesion assays were performed as previously described (Zhou et al., 1993a). Cells were removed from culture flasks by light trypsin treatment (for CC1 external domain chimeras,

88 which are insensitive to trypsin) or PBS Citrate + 4 mM EDTA (PBSCE; for NCAM external domain chimeras, which are cleaved by trypsin) and resuspended at a concentration of 106 cells/ml in α-MEM containing 0.8% FBS and 10 µg/ml DNase I (Roche). Single cell suspensions were obtained by syringing through a 27-gauge needle and were allowed to aggegrate at 37˚C with stirring at 100 rpm using a magnetic stirring bar. Aliquots were removed at the indicated times and the percentage of single cells was determined by a hemocytometer.

Fn Binding Assay Assays were performed essentially as previously described (Nicholson and Stanners, 2006). Cells were resuspended at a concentration of 4 x 105 cells/ml (for LR) or 2 x 105 cells/ml (for L6). 100 μl of this suspension was added to Fn-coated plates (CHEMICON International, Inc.) and incubated for 1 hour at 37°C. Adherent cells were stained with crystal violet, and the optical density determined with a plate reader (Bio-Tek Instruments) at 570 nm. Note that in certain cases for L6 cells, depending on the particular experiment, integrin activation resulted in decreased cellular binding to Fn, likely due to the previously described integrin activation- dependent formation of a cocoon of polymerized Fn around the cells (Ordonez et al., 2007). However, the relative difference between parental cells and activated transfectant cells remained in the inverse sense and so data is presented, for ease of interpretation, as an increase in all cases. Statistical significance was determined using the Student’s t-test (http://www.physics.csbsju.edu/stats/t-test_bulk_form.html).

Immunoblotting Cellular lysates were resolved by SDS-PAGE, and then transferred to a 0.45 μm PVDF membrane (Millipore). Antibody binding was detected using the ECL Plus chemiluminescent reagent (GE Healthcare).

89 Results and Discussion Although the primary sequences of the GPI anchor signal sequences of CEA and CEACAM6 are very similar, mirroring their identical tumorigenic functions, that of NCAM differs greatly (Fig. 1A). The signal sequence from CEA is capable on its own of specifying the addition of the CEA anchor, so chimeras were generated reducing (in five-amino acid increments) the CEA-derived sequence in NCB to localize the sequence responsible for this effect (Fig. 1B). These chimeras were tested for biological activity in the CHO-derived LR-73 and the rat myoblast L6 cell lines, although NCBΔ20 was not expressed in L6 transfectants (Fig. 1B). The sensitivity of these proteins to PI-PLC and their insolubility in cold Triton X-100, with most of each protein present in the insoluble fraction, confirmed GPI anchorage (Fig. 1C; Screaton et al., 2000). LR transfectants showed strong intercellular adhesive ability (Fig. 1D), indicative of the retention of self-binding activity of their external domains (Eidelman et al., 1993; Taheri et al., 2003), although NCBΔ20 adhered somewhat less, likely due to its lower expression level. CEA and NCB but not NCAM alter the activity of the integrin α5β1 (Nicholson and Stanners, 2006; Ordonez et al., 2007) and block differentiation (Eidelman et al., 1993; Screaton et al., 2000), which are characteristics used to determine which chimeras retained the functional activities conferred by the CEA GPI anchor. LR transfectants were tested for binding to the major α5β1 ligand, fibronectin (Fn; Fig. 1E), with NCB expression significantly increasing binding compared to NCAM (P<0.01). Replacing the first 5 CEA-derived amino acids of NCB with the equivalent NCAM residues had no effect on CEA-like function (P<0.005, versus NCAM), but replacing 10 residues resulted in a protein (NCB∆10) that no longer altered binding. In L6 myoblasts, NCBΔ5 had a similar effect on binding to Fn compared to NCB (Fig. 1F; P<0.005), while NCBΔ10 transfectants lost this ability completely. NCB expression completely blocks L6 morphological differentiation, while NCAM transfectants differentiate readily to form large multinucleated myotubes (Screaton et al., 2000). NCBΔ5 expression also blocked differentiation, while NCBΔ10-expressing cells fused substantially, comparable to the parental and NCAM controls (Fig. 1G; all fused between 75% and 80%). NCAM and NCBΔ10 transfectants upregulated myosin, a biochemical differentiation marker, but NCB and NCBΔ5 transfectants did not (Fig. 1H). Thus, the addition of the CEA-specific GPI anchor is determined

Figure 1. Reducing the CEA-derived sequence in NCB. (A) The amino acid sequences of the GPI anchor signal sequences of CEA, CEACAM6, CEACAM7, NCAM and the NCAM-CEA chimera NCB. SAD, small amino acid domain; ω, GPI anchor attachment site. The underlined sequence represents the key residues determined in this study. (B) Representation of the CEA-NCAM chimeras, with the GPI anchor attachment site (ω) denoted. FACS means in arbitrary units for cell surface levels of NCAM and chimeric proteins are indicated. (C) The chimeric constructs are GPI anchored on LR transfectants; western blots of TX-100 solubility assays (top) demonstrated that the majority of each chimera localized in the pellet (P) fraction. The integrin α5 was a lysis and gel loading control. The proteins were also PI- PLC sensitive (bottom), with the percent decrease in surface levels given. (D) LR transfectants showed increased intercellular adhesive ability compared to the parental cell line. (E) Binding of LR parental and transfectant cells to immobilized Fn. Values represent relative binding, compared to parental, +/- SEM for 3 independent experiments (*: P<0.01). (F) Binding of L6 transfectant cells to Fn. Values represent means relative to parental +/- SEM for 3 independent experiments (*:P<0.01). (G) Morphological differentiation of L6 parental and transfected myoblasts. Inset values indicate fusion index, expressed as a percent. (H) Differentiated cultures were lysed, and 5 μg was probed for the biochemical differentiation marker myosin.

92 by the residues that differ from NCBΔ5 to NCBΔ10 (GLSAG) as their substitution with the corresponding NCAM amino acids (SASYT) abrogated CEA-like biological function.5 amino acid stretches in the CEA GPI anchor signal sequence were next replaced with the corresponding NCAM residues (Fig. 2A). CC1-CEA (1C) chimeras were used because previous attempts to attach the NCAM anchor to CEA failed (Screaton et al., 2000), while the CC1-NCAM chimera (1N) is partially processed (Fig. 2B), as seen previously in certain CC1 mutants (Naghibalhossaini and Stanners, 2004). These proteins were expressed at high levels, with the exception of 1N (Fig. 2A), and were GPI anchored (Fig. 2B) and mediated strong intercellular adhesion (Fig. 2C). In LR cells, expression of 1C caused a significant increase (P<0.001) in cellular binding to Fn compared to parental cells (Fig. 2D). Replacing the first 5 CEA amino acids (1CN5) or amino acids 11-15 (1CN15) downstream of the ω site resulted in proteins that were still active (P<0.001); however, replacing amino acids 6-10 (1CN10) completely abrogated the increased binding. In L6 cells, 1C, 1CN5 and 1CN15 expression significantly increased binding to Fn (Fig. 2E; P<0.001), whereas 1CN10 transfectants bound the same as parental. Although the expression of 1CN10 was lower than 1C in L6 cells, 1CN5 and 1CN15 still showed increased binding despite having expression levels similar to 1CN10 (Fig. 2A). Expression of 1C, 1CN5 and 1CN15, but not 1CN10, also strongly inhibited L6 morphological (Fig. 2F) and biochemical (Fig. 2G) differentiation. Replacing amino acids stretches shorter than 5 residues in this region had no effect on binding of LR transfectants to Fn (Fig. 3D). These results therefore confirm the importance of the residues GLSAG in determining the addition of the CEA anchor.

Figure 2. Replacing 5 amino acid stretches in the GPI anchor signal sequence of CEA. (A) Representation of the replacement of CEA amino acids with the corresponding NCAM residues. Expression of the chimeras is given in terms of FACS means. (B) The CC1 chimeras were GPI anchored, as shown by insolubility in cold TX-100 [top; localization to the pellet (P) fraction]. Note that the TM CC1 was found in the soluble (S) fraction. This was confirmed by PI-PLC treatment where, except for CC1, surface levels decreased following treatment (bottom). (C) LR transfectants aggregated in suspension, demonstrating that the chimeras mediate intercellular adhesion. (D) Binding of LR cells to Fn. Values represent the mean relative binding compared to parental cells, +/- SEM, for 4 independent experiments (*: P<0.001). (E) L6 transfectants showed similar effects on Fn binding as in LR. Values represent the mean +/- SEM for 5 independent experiments, relative to parental cells (*: P<0.0001). (F) These effects of these chimeras on morphological differentiation of L6 cells. Percent fusion is given in the inset. (G) Probing 5 μg of differentiated cell lysates by immunoblot for myosin upregulation.

Figure 3. Introducing shorter NCAM sequences into the CEA sequence of 1C. (A) Representation of the amino acids in 1C that were replaced with the corresponding amino acids from NCAM, with the FACS means in LR cells indicated. (B) All chimeric constructs were GPI anchored, as shown by TX-100 insolubility (Top) and PI-PLC sensitivity (bottom). (C) All proteins retained the ability to mediate intercellular adhesion, similar to the 1C construct. (D) Effect of these chimeras on LR Fn binding. Values represent relative mean +/- SEM, for 6 independent experiments (*: P<0.0001).

95 It was next examined whether this sequence was sufficient to give CEA-like biological properties. CEA amino acids were inserted into NCAM at positions 1-5, 6-10 or 11-15, downstream of the omega site, with NC10 containing the GLSAG sequence (Fig. 4A), resulting in GPI-anchored proteins (Fig. 4B). While the proteins mediated intercellular adhesion (Fig. 4C), chimera expression did not result in increased Fn binding in either LR or L6 cells (Fig. 4D and Fig. 4E). Differentiation of L6 cells was not blocked, since morphological (Fig. 4F) and biochemical (Fig. 4G) differentiation was observed in these transfectants. Thus, inserting the sequence GLSAG into NCAM was insufficient to give CEA-like biological activities, suggesting a requirement for further CEA residues. Therefore, larger amounts of CEA-derived sequence were inserted into NCAM to determine the minimum sequence sufficient for specifying the addition of the CEA GPI anchor (Fig. 5A). All chimeras contained the GLSAG sequence, with variable amounts of upstream and/or downstream CEA sequence, and the resulting proteins were GPI anchored (Fig. 5B) and mediated intercellular adhesion (Fig. 5C). When examined for effects on LR binding to Fn, adding 5 upstream CEA amino acids (NΔ110C), but not 5 downstream (NΔ615C), resulted in increased binding (Fig. 5D; P<0.0001). Simply adding one CEA residue on each side of GLSAG (PGLSAGA; NC7) also produced increased cellular binding, suggesting, with the NΔ110C result, that the upstream proline was required for CEA anchor addition. This was examined directly by generating constructs containing only the upstream proline (NC6P) or the downstream alanine (NC6A) (Fig. 5A). In LR cells, NC6P expression increased binding to Fn, while NC6A had no effect, confirming the importance of the proline (Fig. 5D; P<0.0001). These results were recapitulated in L6 cells, where only transfectants of NΔ110C, NC7 and NC6P showed a significant difference in binding to Fn compared to the parental cell line (Fig. 5E; P<0.0001). All chimeras containing the sequence PGLSAG blocked differentiation, while those lacking the proline fused similarly to NCAM transfectants (Fig. 5F). Thus, inserting the sequence PGLSAG into the GPI anchor signal sequence of NCAM is sufficient to generate a protein with

Figure 4. Inserting 5 amino acid sequences into NCAM is insufficient to create a protein with CEA- like properties. (A) 5 amino acid CEA sequences were inserted into the corresponding regions of NCAM. FACS means for LR and L6 tranfectants are shown. (B) Insolubility, as shown by localization to the pellet (P) fraction, of all chimeras after cold TX-100 lysis demonstrated GPI anchorage (top). Cell surface levels of each protein also decreased following PI-PLC treatment (bottom). (C) Adhesion assay demonstrating that these proteins mediated intercellular adhesion. (D) Binding to Fn by LR transfectants. Values represent the mean, relative to parental, +/- SEM from 4 independent experiments (*: P<0.002). (E) L6 transfectant binding to immobilized Fn. Values, relative to parental cells, represent the mean +/- SEM for 4 independent experiments (*: P<0.002). (F) Effect of chimera expression on L6 morphological differentiation; fusion index is also provided. (G) Immunoblot on 5 μg of lysate from differentiated cells, to examine biochemical differentiation through myosin upregulation.

Figure 5. The upstream proline is required to give CEA-like properties. (A) NCAM chimeras that were used to localize the CEA-specific signal. Mean FACS surface expression is indicated for LR and L6 transfectants. (B) The chimeric constructs were GPI anchored, as shown by cold TX-100 insolubility (top) and PI-PLC sensitivity (bottom). (C) LR transfectants demonstrated that all proteins had the ability to mediate intercellular adhesion. (D) Binding by LR transfectants to immobilized Fn, where values represent the mean +/- SEM for at least 4 independent experiments (*:P<0.0001). (E) L6 cells showed similar effects on Fn binding. Values, relative to parental cells, represent the mean +/- SEM for at least 4 independent experiments (*:P<0.0001). (F) Morphological differentiation of L6 transfectants. The fusion index, provided as a percent, is given in the inset.

98 CEA-like biological properties, demonstrating a requirement for presence of the proline. This proline is, in CEA and all of these NCAM chimeras, a part of a G(X)XP sequence (Fig. 1A). This consensus sequence has been shown to result in a kink in TM helices (Cordes et al., 2002), and can be suggested to serve a similar function in this GPI anchor signal sequence. The resulting altered structure may be important to determine the addition of a certain functional anchor. It should be noted, however, that the lack of the proline can be overcome if sufficient downstream CEA sequence is included (see NCB∆5, Fig. 1). The sequence GLSAG was also randomly scrambled to give sequences of ASGGL (denoted NCB-K) and SGLGA (NCB-S) (Fig. 6A). It was hypothesized that a complete loss of biological function would be observed if there was a requirement for a particular amino acid sequence or a particular amino acid at a given position, while at least partial function should be retained if the signal resulted from a general characteristic of this stretch. These proteins were GPI anchored (Fig. 6B) and promoted intercellular adhesion (Fig. 6C). LR transfectants of both chimeras significantly increased binding to Fn compared to the NCAM cell line (Fig. 6D; P<0.001). However, both scrambled transfectants bound less than NCB, particularly NCB-S (P<0.002). L6 transfectants showed altered binding compared to the parental cell line (Fig. 6E; P<0.003 for NCB and P<0.05 for NCB-K and NCB-S), although transfectants of NCB-K and NCB-S again bound significantly (P<0.05) less compared to the NCB transfectants. L6 (NCB-K) and L6 (NCB-S) cells differentiated substantially less than the parental or NCAM controls, but did not show completely blocked morphological fusion (Fig. 6F). Myosin upregulation was seen in the NCB-K and NCB-S cell lines, contrary to NCB, although at lower levels than the NCAM transfectant (Fig. 6G). Thus, scrambling this region results in an incomplete loss of function, indicating that the primary source of the signal is the overall region’s characteristics, though this signal is maximized by the sequence PGLSAG. The signal for the addition of a GPI anchor consists of a set of small amino acids followed by a spacer and a hydrophobic region (Fig. 1A) (Coyne et al., 1993). Work on the bovine liver 5’-nucleotidase has demonstrated the requirement for particular lengths of both the spacer and the hydrophobic domain for proper processing (Furukawa et al., 1994; Furukawa et al., 1997). Differences in the efficiency of GPI anchor addition for various signal

Figure 6. Scrambling the amino acid sequence in the identified key region. (A) Schematic of the constructs containing scrambled amino acids at positions 6-10, with the amino acid sequence shown underneath. Relative cell surface expression is shown. (B) GPI anchorage of these chimeric proteins was demonstrated by insolubility in cold TX-100 (top) and sensitivity to PI-PLC treatment (bottom). (C) The scrambled constructs mediated intercellular adhesion in LR cells. (D) Binding to Fn by LR transfectants of the scrambled constructs. Values represent the mean +/- SEM for 4 independent experiments (*: P<0.0005). (E) Effect of expression of the scrambled constructs on L6 Fn binding. Values represent the relative mean +/- SEM for 3 independent experiments (*: P<0.004; **: P<0.03). (F) Morphological differentiation of L6 transfectants. Fusion index is given in the inset. (G) Biochemical differentiation, in terms of myosin production, as demonstrated by western blot on 5 μg of cellular lysate.

100 sequences suggest that these stretches are not processed identically (Chen et al., 2001). All previous studies, however, have been concerned with the efficiency of anchor addition; the current work is the first to demonstrate that the specificity of anchor addition is the result of a second signal within this sequence. We have previously demonstrated that the CEA GPI anchor signal sequence determines function and localization to a specific membrane raft, despite being cleaved in the ER (Nicholson and Stanners, 2006; Screaton et al., 2000). This study was designed to establish the residues that are critical for this specification, with the demonstration that the amino acids PGLSAG in the hydrophilic region of the GPI signal sequence are necessary and sufficient for this effect. The identified sequence from CEA is fairly well conserved in CEACAM6 (PVLSAV) and CEACAM7 (PDLSAG) (Fig. 1A), proteins that show similar biological effects to CEA (Rojas et al., 1996) (Zhai and Stanners, manuscript in preparation), suggesting that it may have a similar role in determining the function of these proteins. Since GPI anchors from 5 other proteins did not show any of these effects when bound to the CC1 external domain (Zhai and Stanners, unpublished data), this amino acid set is quite specific. This work has identified a novel signal within the GPI anchor signal sequence of CEA, which determines protein functionality. Studies using various GPI-anchored protein comparisons, such as CEA and NCAM (Nicholson and Stanners, 2006), Thy-1 and the prion protein (Madore et al., 1999), and folate receptor and placental alkaline phosphatase (Wang et al., 2002c), have suggested that different GPI-anchored proteins exist in different microdomains on the cell surface. This distribution is important, since switching the anchor, and subsequent distribution, of NCAM to that of CEA is sufficient to radically alter its function (Nicholson and Stanners, 2006; Screaton et al., 2000). This occurs through association with particular rafts and their specific signaling elements, which determine the downstream tumorigenic effects of CEA (Camacho-Leal et al., 2007). Examining the GPI anchor signal sequences of various other proteins should further elucidate the specificity and importance of this signal in determining specific anchor addition and, ultimately, protein function. It will also be important to characterize how this region is capable of determining the addition of a particular anchor. It is possible that this stretch of amino acids interacts directly with the GPI anchor precursor prior to the transamidation binding reaction, and only combinations that match structurally proceed enzymatically. Alternatively, the transamidase complex, which is composed of 5 different subunits, may play a direct role in this effect. One component, hGaa1p, has been suggested to

101 recognize the signal sequence (Chen et al., 2003), while another, Gpi8p, functions as the enzymatic subunit (Ohishi et al., 2000). Either of these, or one of the three complex components (PIG-S, PIG-T and PIG-U) with currently unknown function, could serve to bring together specific signal sequences and GPI anchors. This study demonstrates that a specific 6 amino acid stretch of the GPI anchor signal sequence determines the addition of a particular functional anchor which in turn can determine the ultimate function of the protein.

102 Preface to Chapter 4 Following the demonstration, in Chapter 3, of a specific signal within the GPI anchor signal sequence of CEA that directs the addition of the CEA-specific anchor, we attempted to determine how similar the functions mediated by the GPI anchors of CC6 and CC7 are to those of the CEA anchor. This was accomplished by generating chimeric constructs with the NCAM external domain attached to the GPI anchors of CC6 and CC7, and testing for biological functions. In addition, the ability of a non-functional CEA mutant to block the functions of these chimeras, similar to the effect described in Chapter 2, was also tested, to examine the anchor’s similarities. We demonstrate that, while the anchors of CC6 and CC7 mediated similar biological effects, they are not identical to that of CEA, incorporating another level of complexity to the concept of GPI anchor specificity.

103

Chapter 4

Exploring the biological properties of the GPI anchors of CEACAM6 and CEACAM7

104 Abstract

CEA, CC6 and CC7 are GPI-anchored proteins in the CEA family whose expression blocks cellular differentiation and alters integrin function. These functions, at least for CEA, are specified by its GPI anchor, provided it is attached to an adhesive external domain. We demonstrate in this study that the anchors of CC6 and CC7 also contain the necessary information to determine protein function. We have previously demonstrated that targeting the GPI anchor of a CEA-like chimera is sufficient to abrogate its biological functions. We confirm this finding by demonstrating that CEA itself can have its functions inhibited by a non-functional mutant with the same anchor, provided external domain interactions are reduced. However, while the anchors of CEA, CC6 and CC7 appear to be functionally equivalent, the CEA anchor cannot be used to inhibit these other proteins, suggesting that they are not identical. Thus, the GPI anchors of the CEA family play critical roles in determining function, but the anchor of CEA is not identical to that of CC6 or CC7. This has important implications for the study of GPI anchors, as it suggests that functionally equivalent anchors are not necessarily structurally identical.

105

Introduction Carcinoembryonic antigen (CEA) is a cell-surface protein belonging to the CEA gene family that is composed of 7 external immunoglobulin (Ig) domains attached to the plasma membrane by a glycophosphatidylinositol (GPI) anchor (Hammarstrom, 1999; Thompson et al., 1991). While CEA and the closely-related, GPI-anchored CEACAM6 (CC6) are upregulated in many cancers (Blumenthal et al., 2007), the transmembrane (TM) CEACAM1 (CC1) is often downregulated and has been suggested to function as a tumor suppressor (Hsieh et al., 1995; Kunath et al., 1995). CC7, another GPI-anchored family member, is downregulated in colon cancer (Scholzel et al., 2000; Thompson et al., 1997a; Thompson et al., 1994), but upregulated in gastric and pancreatic cancers (Kinugasa et al., 1998; Yoshida et al., 2003). CEA and CC6 block differentiation and the apoptotic process of anoikis, unlike CC1 (Duxbury et al., 2004d; Eidelman et al., 1993; Ordonez et al., 2000); CC7 also blocks differentiation but its effects on anoikis remain unclear (Zhai and Stanners, manuscript in preparation). The CEA GPI anchor contains specific information that determines protein function, as attaching the GPI anchor of CEA to the external domains of CC1 results in a protein with CEA- like properties (Screaton et al., 2000). In addition, attaching the CEA anchor to the external domains of neural cell adhesion molecule (NCAM), a GPI-anchored protein with different biological properties from CEA, also generates a protein with CEA-like properties (Screaton et al., 2000). Therefore, in combination with studies demonstrating the importance of the adhesive function for CEA’s properties (Eidelman et al., 1993; Taheri et al., 2003), CEA’s function has been established to require a self-adhesive external domain attached to its GPI anchor. The external domains are non-specific for function, provided they mediate this self-adhesive effect, while the GPI anchor harbors the biological specificity As such, blocking the adhesive function of CEA is sufficient to abrogate its properties (Taheri et al., 2003), while targeting the anchor has similar functional consequences (Nicholson and Stanners, 2006). We have previously demonstrated that a short sequence of amino acids within the GPI anchor signal sequence of CEA directs the addition of the CEA-specific functional anchor (Nicholson and Stanners, 2007). This sequence, PGLSAG, has similar corresponding residues in CC6 (PVLSAV) and CC7 (PDLSAG), which could be hypothesized to lead to the addition of similar, if not identical, GPI anchors.

106 The plasma membrane is composed of numerous heterogeneous domains, including membrane rafts, aggregates composed of sphingolipids and cholesterol that were initially characterized based on their insolubility in cold non-ionic detergents (Simons and Ikonen, 1997). Different GPI-anchored proteins and raft-resident lipids show non-overlapping distributions in the membrane, demonstrating that rafts consist of a heterogeneous population within the membrane (Fujita et al., 2007; Gomez-Mouton et al., 2001; Gomez-Mouton et al., 2004; Li et al., 2003; Madore et al., 1999; Vyas et al., 2001; Wang et al., 2002b). Proteins with GPI anchors are specifically targeted, by their anchors, to a subset of rafts (Nicholson and Stanners, 2006), which is of particular importance because rafts concentrate many signaling molecules (Foster et al., 2003; von Haller et al., 2001). GPI-anchored proteins are not directly associated with the cytoplasm, which makes their targeting to specific domains that contain a specific subset of signaling molecules to transduce their signals critically important. The initiation of signaling cascades by these proteins, at least for those with adhesive function, results from oligomerization via their external domains, which leads to co-aggregation of their associated rafts, stabilizing the rafts to form large signaling platforms (Harris and Siu, 2002). In this study, we have examined the functional similarities between the anchors of CEA, CC6 and CC7. We confirm that CEA can be inhibited by co-expression of ΔNCEA, as has previously been reported for an NCAM-CEA chimera, validating this effect (Nicholson and Stanners, 2006). Chimeras were then generated with the NCAM external domain attached to the GPI anchors of CC6 and CC7. These chimeric proteins had similar biological properties to an NCAM-CEA chimera, confirming that the anchors of CC6 and CC7 are critical determinants of function. However, ΔNCEA had no observable effect on the function of these CC6 and CC7 chimeras, unlike for the CEA chimera, suggesting that the anchors of these proteins are not identical to that of CEA, despite their functional similarities. This therefore indicates that similar functional anchors are not targeted to the same domains, suggesting that a wide range of anchors and domains exists.

107

Materials and Methods Constructs The cDNAs used in this study were: full-length CEA (Beauchemin et al., 1987); the CEA deletion mutant ΔNCEA, which lacks the last 75 amino acids of the N-terminal domain of CEA (Eidelman et al., 1993); the GPI-anchored CEA family member CC6 (Zhou et al., 1990); CC7, another GPI-anchored CEA family member (Thompson et al., 1994); and the GPI-anchored isoform of NCAM, p125 (Screaton et al., 2000).

Table 1: Primers used in this study Primer Sequence 5'-GGG CAG GAG TCC TTG GAA TTC ATC CTT NCAM Forward GTT CAA GCA GAC ACC C-3’

CC6 Reverse 5'-GAT GAG AAT TCT TGT TTT GAC ATC TTG GG-3’ N-6 Junction, Forward 5'-ACC TCG GCC CAG CCC GTC ACG ATG ATC ACA-3’ N-6 Junction, Reverse 5'-TGT GAT CAT CGT GAC GGG CTG GC CGA GGT-3’

CC7 Reverse 5'-CGT GGA GCG GCC GCG AAT TCG AAG CTC GGT ACC-3’ N-7 Junction, Forward 5'-ACC TCG GCC CAG CCC CTG AAT GTC CGC TAT-3’ N-7 Junction, Reverse 5'-ATA GCG GAC ATT CAG GGG CTG GGC CGA GGT-3’

To study the properties of the GPI anchors of CC6 and CC7, the signal sequences of these proteins were genetically attached to the external domain of NCAM (Fig. 3), such that no CC6/CC7 amino acids were present following addition of the anchor. Chimeras were generated by first performing separate PCRs using the NCAM Forward primer and the respective Junction Reverse primer with the NCAM cDNA template; and the Junction Forward primer and the corresponding Reverse primer, with either CC6 or CC7 cDNA. The resulting products were then joined by overlap PCR (Screaton et al., 2000). The chimeric cDNA fragments were cloned into the expression vector p91023b (Screaton et al., 2000) containing NCAM cDNA using EcoRI sites flanking the PCR products and sites located in NCAM at positions 1616 and 2794.

108

Cell Culture and FACS analysis The cell lines used in this study were the CHO-derived LR-73 fibroblasts (Pollard and Stanners, 1979) and the rat myoblast cell line L6 (Yaffe, 1968), which were cultured in a humidified

atmosphere containing 5% CO2, as previously described (Nicholson and Stanners, 2006). Cells were transfected with cDNA by calcium phosphate coprecipitation, and positive transfectants were selected as previously described (Nicholson and Stanners, 2006). The resulting colonies were pooled, and sorted using a FACSAria instrument (Becton Dickinson) for high expression using the CEA-specific antibody J22 (Zhou et al., 1993b), which recognizes both CEA and ΔNCEA, or the NCAM-specific antibody 123C3 (Santa Cruz Biotechnology, Inc.).

Differentiation Assays For differentiation assays involving homogeneous cell cultures, L6 cells were seeded at 104 cells/cm2 in 60-mm dishes in medium containing 10% FBS. 3 days later, after reaching confluency, the medium was changed to DME containing 2% horse serum (DM). 5-7 days later, cells were stained with hematoxylin (Sigma-Aldrich) as previously described (Screaton et al., 1997). For co-culture differentiation assays (Taheri et al., 2003), 3 x 105 cells of each type were mixed and seeded in a 35-mm dish, resulting in a confluent culture immediately upon attachment. 24 hours later, the medium was switched to DM, and cells were stained after 7 days of incubation. The extent of differentiation was quantitated by light microscopy, where the percentage of nuclei in fused cells (containing 3 or more nuclei) was compared to the total number of nuclei in the field, to generate the fusion index. The data represent 3 separate experiments, with 2 fields counted per experiment. Statistical significance was determined by t test (http://www.physics.csbsju.edu/stats/t-test_bulk_form.html).

Binding to purified fibronectin Cells were seeded at a concentration of 104 cells/ml and collected, 2 days later, with PBS-Citrate containing 4 mM EDTA (PBSCE). These cells were incubated in medium, containing 10% FBS, for 30 minutes at 37ºC, before being washed and resuspended in serum-free medium at a concentration of 4 X 105 cells/ml. 100 µl of this suspension was added to wells from a 96-well plate coated with Fn (CHEMICON International, Inc.), and incubated for 1 hour at 37ºC. Wells

109 were washed twice with PBS containing Mg2+ to remove non-adherent cells, and then stained with crystal violet. Excess stain was removed by washing again with PBS+Mg2+, followed by

solubilizing the bound stain with a solution of 0.05 M NaH2PO4, pH 4.5 and 25% ethanol. The optical density of the resulting solution was determined with a microplate reader (Bio-Tek Instruments) at a wavelength of 570 nm. Note that in certain cases, the effects of CEA-family proteins on integrins resulted in decreased binding to Fn; the data presented in this study represents the absolute difference and is shown as an increase, for ease of interpretation.

TX-100 insolubility and PI-PLC sensitivity assays Subconfluent cell cultures were collected with PBSCE and passed through a 27-gauge needle to provide single-cell suspensions. Cells were pelleted and resuspended, at a concentration of 107 cells/ml, in 1% TX-100 in ice-cold lysis buffer [20 mM Tris, pH 8.0, 150 mM NaCl, 2.5 mM EDTA containing the protease inhibitors aprotonin (Roche), leupeptin (Roche) and PMSF (Sigma-Aldrich)]. Cells were disrupted by syringing with a 27-gauge needle, incubated on ice for 15 minutes, and then centrifuged for 15 minutes at 13 500 g at 4ºC. The soluble fraction was removed, and the pellet resuspended in an equal volume of lysis buffer containing 1% SDS. To confirm GPI anchorage of these proteins, LR transfectants were seeded at a concentration of 7 X 103 cells/cm2 in 24-well plates. 2 days later, adherent cells were washed twice with PBS, and incubated with 0.1 U bacterial phosphatidylinositol phospholipase C (PI- PLC; Sigma-Aldrich) in a 1:1 mixture of α-MEM:PBS containing 0.2% BSA. This was incubated at 37ºC for 40 minutes, and then the cells were collected with trypsin and analyzed by FACS. Sensitivity to PI-PLC was determined by comparing FACS profiles of the PI-PLC treated cells to profiles of cells treated without PI-PLC.

Adhesion assays Subconfluent cultures of LR transfectants were collected with PBSCE, in order to avoid cleaving the NCAM external domain with trypsin (Screaton et al., 2000). Cells were then resuspended, at a concentration of 1 X 106 cells/ml, in 3 ml α-MEM containing 0.8% FBS and 10 µg/ml DNase I (Roche). In order to ensure the samples were single cell suspensions, these solutions were syringed with a 27-gauge needle, and then allowed to aggregate with stirring (100 rpm, with a magnetic stirring bar) at 37ºC. At the indicated times, aliquots were removed and counted with a

110 hemocytometer to compare the number of cells in multicellular clumps, including doublets, versus single cells.

111

Results Co-expression of CEA and ∆NCEA on L6 Cells The expression of ∆NCEA on the surface of L6 myoblasts expressing NCB, a chimera with CEA-like biological properties, results in a complete loss of NCB function (Nicholson and Stanners, 2006). We tested whether CEA itself could be inhibited by ∆NCEA (Fig. 1A) by supertransfecting L6 (CEA) cells with ∆NCEA (Fig. 1B). Due to the similarities between CEA and ∆NCEA, FACS analysis of the cotransfected cells shows CEA-alone staining, using mAbs with epitopes in the deleted region, and CEA+∆NCEA total staining (Fig. 1B). The presence of ∆NCEA was indicated by the substantial increase in total CEA+∆NCEA staining (FACS mean: 182), compared to CEA alone (FACS mean: 55). As a control, CEA-expressing cells were cotransfected with NCAM, a protein with different cell surface localization and biological properties from CEA (Nicholson and Stanners, 2006; Screaton et al., 2000), with similar expression levels of both proteins (respective means of 50 and 80; Fig. 1B).

Effect of ∆NCEA expression on CEA function Ectopic expression of CEA results in a complete block of L6 differentiation, even after prolonged exposure to conditions that promote fusion (Eidelman et al., 1993). To examine the effect of ΔNCEA on CEA function, the co-expressing L6 cells were tested for the ability to differentiate (Fig. 1C). CEA expression, as previously described (Eidelman et al., 1993), completely blocked the differentiation of L6 transfectants (Fig. 1D). Co-expression of ∆NCEA with CEA did not restore fusion, as these cells remained completely blocked for differentiation, while the co-expression of CEA and NCAM also did not increase differentiation (Nicholson and Stanners, 2006). Previous work demonstrated that the presence of CEA on the surface of one cell is sufficient to activate ∆NCEA molecules on apposed cells (Fig. 1C; Taheri et al., 2003), presumably by anti-parallel binding between the N domain of CEA and the A3 domain of

112

Figure 1. Effect of ∆NCEA on the differentiation of L6 transfectants expressing CEA. (A) Schematic of the GPI-anchored proteins used in this study. ΔNCEA is a non-functional CEA deletion. (B) Surface expression of the proteins in L6 cells, as determined by FACS using antibodies directed against the indicated protein, followed by secondary antibodies conjugated to FITC. The epitope of the A20 antibody lies in the N-terminal domain of CEA, so it does not label ΔNCEA. The antibody J22, however, recognizes the internal domains of CEA, and can bind to both CEA and ΔNCEA, so for the L6 (CEA/ΔNCEA) co-expressing cells, A20 values represent CEA alone, while J22 values are the combination of CEA and ΔNCEA. (C) Model of the differentiation assays performed. Expression of CEA, but not ΔNCEA, blocks L6 myoblast fusion. CEA on the surface of one cell is able to activate the ΔNCEA on the surface of apposed cells, as demonstrated by co-culture experiments. (D) Pictures of differentiated cultures of L6 transfectants. While L6 (neo) transfectants fused at high levels, L6 (CEA) cells remained single cells. Co-expression of ΔNCEA or NCAM with CEA did not have an appreciable effect on function, as both transfectant cell lines did not show increased fusion.

113 ∆NCEA (Zhou et al., 1993a). This could explain why ∆NCEA expression had no effect on CEA function, unlike its effect on NCB, and led to attempts to reduce these external domain interactions.

Effect of ∆NCEA on CEA function under co-culture conditions To reduce the transactivation of ∆NCEA by CEA, the co-expressing cells were mixed with an equal number of L6 (neo) cells, and seeded so that the cells were confluent immediately upon attaching to the culture surface, ensuring the cells maintained a random distribution (Fig. 2A). The presence of the neo cells should result in a substantially decreased interaction between CEA molecules on apposed surfaces, reducing the adhesion-mediated effect of CEA on ∆NCEA. Under these conditions, the CEA/∆NCEA co-expressing cell culture differentiated significantly more (P<0.005) than the CEA/NCAM co-culture (Fig. 2B). When quantitated, the co-culture of L6 (CEA/∆NCEA) and L6 (neo) cells had a fusion index of 45%, while L6 (CEA/NCAM) and L6 (neo) co-culture fused 23% (Fig. 2C), indicating that ∆NCEA is capable of interfering with the biological function of CEA. This confirms that CEA function can be inhibited by targeting its GPI anchor, provided the interaction between the external protein domains attached to these anchors is limited.

114

Figure 2. Differentiation under co-culture conditions of the co-expressing transfectants. (A) Model of the co-culture differentiation assays performed. The L6 co-expressing cells were mixed with an equal number of L6 (neo) cells, in order to reduce interactions between molecules on the surfaces of apposed cells. (B) Photomicrographs of the differentiated L6 cells. The co-culture L6 (neo) + L6 (CEA/NCAM) situation resulted in a low amount of fusion, while the combination of L6 (neo) + L6 (CEA/ΔNCEA) fused substantially more. (C) Quantitation of the co-culture fusion assays, demonstrating that a significant increase in fusion was seen for the co-culture samples containing L6 (CEA/ΔNCEA) cells, compared to the L6 (CEA/NCAM) control. Data represent the mean +/- SEM of 3 indepent experiments (*: P<0.005).

115 NCAM chimeras with the GPI anchors of CC6 and CC7 To examine the relationship between the GPI anchors of CEA, CC6 and CC7, the anchors of CC6 and CC7 were attached to the external domain of NCAM (generating N6 and N7, respectively), to eliminate variations due to the external domain (Fig. 3A). These proteins were expressed in both LR and L6 cells at high levels, as determined by FACS (Table 2). N6 and N7 were also co-expressed with ΔNCEA in both LR and L6 cells, to examine the similarities between these GPI anchors.

Table 2: Surface expression of CC6 and CC7 chimeras Surface Expression cDNA Cell Line (relative FACS means) α-NCAM α-CEA neo LR 5 5 L6 4 4 NCB LR 356 N/A L6 504 N/A NCAM LR 267 N/A L6 307 N/A N6 LR 234 N/A L6 108 N/A N7 LR 161 N/A L6 94 N/A ΔNCEA + NCB LR 172 533 L6 292 1486 ΔNCEA + N6 LR 280 1085 L6 116 1555 ΔNCEA + N7 LR 177 851 L6 83 1400

The chimeric constructs are GPI anchored To confirm that these chimeras were GPI anchored, two complementary methods were employed to examine the method of anchorage of N6 and N7. When lysed with cold TX-100, membrane rafts and their associated proteins, including those with GPI anchors, are present in the insoluble fraction, while most TM proteins are fully solubilized (Screaton et al., 2000). N6 and N7

116

Figure 3. Characterization of the NCAM chimeras with the GPI anchors of CC6 and CC7. (A) Schematic representation of the GPI-anchored proteins involved in this study. The N6 chimera has the external domains of NCAM attached to the GPI anchor of CC6, while N7 consists of the same external domains attached to the anchor of CC7. (MSD: muscle specific domain). (B) Insolubility in cold non- ionic detergents of the chimeras. L6 transfectants were lysed with ice-cold 1% TX-100, and then separated into insoluble (or pellet; P) and soluble (S) fractions. As seen in the top immunoblot, both N6 and N7 were primarily localized to the pellet fraction, similar to NCB and NCAM controls. As a control for efficient membrane disruption, the TM integrin α5, is mainly present in the soluble fraction. (C) PI- PLC treatment confirms GPI anchorage of N6 and N7. N6 and N7 both show large decreases, as determined by FACS, in cell surface levels following PI-PLC treatment, confirming their GPI anchorage. (D) Intercellular adhesion mediated by N6 and N7. LR transfectants were rendered single cells and allowed to aggregate in suspension, with aliquots taken at the indicated times to determine the percent of single cells remaining. N6 and N7 transfectants both aggregated substantially more than the negative control, although the adhesion mediated by N7 was somewhat weaker.

117 primarily to the pellet, or insoluble, fraction, while the α5 integrin, a TM protein, was almost entirely present in the soluble fraction (Fig. 3B), showing the specific association of N6 and N7 were both localized with the detergent resistant fraction. The GPI anchorage of these proteins was confirmed using bacterial PI-PLC, an enzyme that cleaves the GPI anchor, releasing the proteins from the cell surface. For both N6 and N7, a decrease of at least 60% of the cell surface levels was seen by FACS following treatment with PI-PLC (Fig. 3C), while the TM CC1 control decreased minimally. These results therefore confirmed that these proteins were GPI anchored, allowing their biological properties to be examined.

The chimeric constructs retain homotypic adhesion function

The CEA-specific GPI anchor must be attached to adhesion-competent protein external domains for function (Eidelman et al., 1993). The NCAM extracellular domain is able to mediate cellular aggregation (Rao et al., 1992), and can be substituted for the CEA external domain without any loss of CEA function (Screaton et al., 2000). LR (neo) control transfectants showed minimal aggregation, as the percentage of single cells remained above 90% even after 2 hours in suspension (Fig. 3D). As previously described (Nicholson and Stanners, 2006), NCAM and NCB both mediate strong intercellular adhesion, with approximately 30% single cells remaining after 2 hours. For N6 expressing cells, a decrease similar to that of NCB and NCAM was also noted (Fig. 3D; 29% single cells after 2 hours), indicating that this protein mediates strong adhesion. N7-mediated adhesion was not as efficient, as approximately 50% of the cells remained as single cells after 2 hours (compared to 20-30% for the other transfectants). While this is likely due to the lower expression levels of N7 (Table 2), the aggregation observed should be sufficient for function. Thus, both chimeras retained the ability to mediate intercellular adhesion, and can be examined for biological properties.

Effects on cell-ECM interaction Expression of CEA, CC6 and CC7, but not NCAM, results in altered interaction with the ECM molecules fibronectin (Fn) and vitronectin (Vn), through the modulation of the integrins α5β1 and αvβ3 (Camacho-Leal et al., 2007; Ordonez et al., 2007). When LR transfectants were tested for interaction with Fn, the NCB-expressing cells showed a significant (p<0.0001) increase in binding compared to the neo control (Fig. 4A). Expression of the N6 and N7 chimeras caused a

118

Figure 4. Integrin modulation by N6 and N7 chimeras. (A) Binding of LR transfectants to immobilized Fn. Expression of NCB, N6 and N7 resulted in a significant increase in cellular binding to Fn, compared to neo and NCAM controls (*: P < 0.0001). Co-expression of ΔNCEA with NCB, but not with N6 and N7, resulted in a complete loss of this effect. Data represent the mean +/- SEM for at least 3 independent experiments. (B) Fn binding by L6 transfectants recapitulates the LR results. Expression of NCB, N6 and N7 all significantly affected cellular binding to immobilized Fn compared to the parental control (*: P < 0.0001). While L6 (ΔNCEA + NCB) showed binding similar to the parental cells, and unlike L6 (NCB) cells, L6 (ΔNCEA + N6) and L6 (ΔNCEA + N7) showed no effect of ΔNCEA on N6 or N7 function. Data represent the mean +/- SEM for at least 3 independent experiments.

119 similar increase in binding to Fn, confirming that the anchors from these proteins have comparable effects on function as the anchor of CEA. In L6 cells a similar alteration of cell-Fn interaction was seen for transfectants of NCB, N6 and N7 (Fig. 4B; p<0.0001). Therefore, the anchors of CEA, CC6 and CC7 have similar effects on integrin activity, suggesting that the anchors of these proteins are functionally similar.

The GPI anchor of CEA can be targeted using a non-functional protein with the same anchor, resulting in a complete loss of biological function, including a loss of integrin modulation (Nicholson and Stanners, 2006). Thus, co-expressing the non-functional CEA mutant ΔNCEA with NCB results in a complete loss of NCB function, such that co-transfectants bind to Fn similar to NCAM controls (Fig. 4A and Fig. 4B) (Nicholson and Stanners, 2006). This effect was used to examine whether the CC6 and CC7 anchors are the same as that of CEA. When ∆NCEA was co-expressed with N6 and N7 in LR cells, no difference was noted compared to the cells only expressing the chimera (Fig. 4A). Similarly, L6 cells co-expressing ΔNCEA with either N6 or N7 showed no effect on the altered Fn interaction resulting from N6 or N7 expression (Fig. 4B). Thus, while the anchors from CC6 and CC7 mediate similar Fn-binding effects to that of the CEA anchor, these anchors are not identical, as ΔNCEA is unable to inhibit the function of N6 or N7.

Effects of N6 and N7 on differentiation

The GPI-anchored members of the CEA family, CEA (Eidelman et al., 1993), CC6 (Rojas et al., 1996) and CC7 (Zhai and Stanners, manuscript in preparation) all block the terminal differentiation of L6 myoblasts. Therefore, we examined the effect of N6 and N7 on differentiation, with N6 and N7 expression significantly reducing differentiation compared to the parental or NCAM controls (Fig. 5A). Thus, the role of the anchors of these proteins in determining function is similar to the anchor of CEA. However, as seen for the cell-ECM data, co-expression of ∆NCEA with N6 or N7 did not increase the differentiation of these cells (Fig. 5B). This confirmed that the anchors of CC6 and CC7 are not identical to that of CEA, despite their functional similarities.

120

Figure 5. The effect of N6 and N7 on differentiation. (A) Photomicrographs of representative fields of L6 cells cultured in low serum conditions to induce differentiation. NCB, N6 and N7 expression results in a significant decrease in differentiation of L6 myoblasts, compared to parental and NCAM control transfectants. The co-expression of ΔNCEA with N6 and N7 did not result in any restoration of differentiation, confirming that ΔNCEA does not affect N6 and N7 function. (B) Quantitation of the fusion index for the differentiated L6 cells. Data represents the mean +/- SD for three independent experiments. Values written on the graph represent the percent fusion.

121 Discussion

In this study, we expand on previous reports (Nicholson and Stanners, 2006; Screaton et al., 2000) that demonstrated that the GPI anchor of CEA contains specific information directing the localization and functional specificity of the molecule. We have examined the similarities that exist between the anchors of CEA and the closely-related proteins CC6 and CC7, three proteins with comparable biological properties. This was accomplished by generating chimeras with the external domain of NCAM attached to the GPI anchors of CC6 and CC7, similar to the chimera (NCB) that demonstrated the importance of the CEA anchor in determining protein function (Screaton et al., 2000). These chimeras were tested for CEA-like properties, notably integrin modulation and differentiation blockage, while also examining the effect of ΔNCEA on their function. Both were shown to be able to modulate integrin activity and block differentiation but, unlike CEA, could not be inhibited in these activities by ∆NCEA.

Inhibiting the function of CEA by targeting its GPI anchor We initially sought to confirm that the intact CEA protein, and not just a chimera with the CEA anchor but different external domains, can have its functions inhibited by targeting its anchor. Previous work had demonstrated that the biological functions of CEA require a self-adhesive external domain attached to the CEA-specific GPI anchor (Eidelman et al., 1993; Screaton et al., 2000). Specific 5 amino acid subdomains within the external domain are required for the adhesive function of CEA (Taheri et al., 2000), and antibodies or peptides targeting these subdomains are able to block the biological properties of CEA (Taheri et al., 2003). As both the adhesive function and the anchor are required for CEA function, we hypothesized that targeting the anchor could result in a similar effect. Our initial studies employed an NCAM-CEA chimera (NCB) to eliminate the possibility of external domain interactions with ΔNCEA, and demonstrated that this non-functional CEA mutant could inhibit the function of NCB (Nicholson and Stanners, 2006). NCB was used, instead of CEA, due to previous reports of anti-parallel interaction between CEA and ΔNCEA (Zhou et al., 1993a), leading to activation of ΔNCEA (Taheri et al., 2003). The possibility existed that these external domain interactions could overcome the anchor effect, a concern that was validated in this study, as co-expression of ΔNCEA and CEA did not restore differentiation (Fig. 1). However, it was also possible that the

122 NCB protein did not completely recapitulate CEA’s properties, due to its different external domain, so the ability to inhibit CEA as well as NCB by this mechanism required demonstration. Under conditions that minimized the interactions between the external domains of CEA and ΔNCEA, an inhibition of the function of CEA was noted (Fig. 2), confirming that targeting the GPI anchor can be employed to inhibit protein function. GPI-anchored proteins, by virtue of their lipid anchor, are able to spontaneously insert from the external medium into the plasma membrane and regain their biological function (Ilangumaran et al., 1996). The information targeting the protein to specific domains resides solely in the GPI anchor, at least for CEA (Nicholson and Stanners, 2006). Because the localization depends on the anchor, and not the external protein domains, it should be possible to treat cells, from the external medium, with the CEA GPI anchor attached to non-functional external domains and inhibit CEA function. The successful external application of GPI-anchored proteins to the surface of cells has already been documented and is known as “painting” (Ilangumaran et al., 1996; Medof et al., 1996; Premkumar et al., 2001). If this hypothesis is accurate, it would offer a novel possibility for targeting CEA to revert its tumorigenic effects, and could be used as a method to modulate cancer cells towards the acquisition of a more normal phenotype. Studies in the Stanners laboratory have in fact demonstrated that the administration of specific de-clustering agents both in vitro and in vivo can achieve this (Ilantzis and Stanners, unpublished data). It remains to be seen whether CEA GPI anchor-based agents could achieve the same.

The biological properties of the CC6 and CC7 anchors The CEA family, in humans, contains several members that are attached to the plasma membrane by a GPI anchor, including CEA, CC6 and CC7. Previous studies have demonstrated that CC6 (Rojas et al., 1996) and CC7 (Zhai and Stanners, manuscript in preparation) have the ability to block differentiation, similar to CEA. As the GPI anchor signal sequences of these proteins are highly homologous and the differentiation blocking function of CEA resides in its GPI anchor (Screaton et al., 2000), it has been hypothesized that the anchors of these proteins may be identical. Therefore, we examined whether the anchors of CC6 and CC7 also determine protein function. The anchors of CC6 and CC7 were attached to the external domain of NCAM, resulting in chimeric proteins with the ability to modulate integrin function (Fig. 4) and block differentiation (Fig. 5), confirming that the anchors of these other CEA family members also

123 specify similar functions to those of CEA. Interestingly, a protein with the anchor of a CEA family protein from a non-human primate also blocks differentiation (Naghibalhossaini et al., 2007), suggesting that the GPI anchors from all CEA family proteins have similar functional properties. It is intriguing to note that the GPI-anchored CEA family members evolved twice from different mutational packages, suggesting that a functional advantage exists for this modification (Naghibalhossaini et al., 2007). Much of the information found in the GPI anchor signal sequence appears to be present in the ancestor of all these GPI-anchored proteins, CC1, as minimal mutations in its TM domain result in a GPI-anchored protein with CEA-like functions (Naghibalhossaini and Stanners, 2004). Therefore, as all of the GPI-anchored family members evolved from a primordial TM protein (Hammarstrom et al., 1998; Zimmermann, 1998), it is reasonable to suggest that they would have similar functional properties.

The GPI anchors of CC6 and CC7 are not identical to that of CEA The functional similarity between the anchors of these CEA family members resulted in the hypothesis that the anchors should be identical structurally, and target the same membrane domains. This hypothesis was tested using co-expression studies that had been developed previously (Nicholson and Stanners, 2006), where the expression of ΔNCEA inhibited the function of NCB, a protein with the same anchor. If the anchors of CC6 and CC7 are identical to that of CEA, then ΔNCEA should also inhibit the function of the N6 and N7 chimeras (Nicholson and Stanners, 2006). However, co-expression of ∆NCEA with N6 and N7 did not alter the properties of these proteins, as cell-Fn binding remained increased, compared to controls (Fig. 4), and there was no restoration of differentiation (Fig. 5). This implies that, despite their functional similarities, the anchors of CC6 and CC7 are not identical to that of CEA, which has been suggested by other studies of these proteins. Thus, while the differentiation block resulting from CEA expression is not sensitive to the presence of insulin, the effect of CC6 on differentiation can be partially overcome when this hormone is present (Ordonez and Stanners, unpublished data). Additionally, while CEA and CC6 expression significantly inhibits the induction of the cell death pathway, anoikis, the effect of CC7 on this process is not as clear (Zhai and Stanners, manuscript in preparation). These slightly different effects also suggest that the anchors of these proteins are not identical, despite their similar abilities to block differentiation and alter integrin binding to ECM. To further this investigation, the exact

124 localization of these proteins, to different domains, will need to be explored using microscopy or immunoprecipitation studies.

Heterogeneity of GPI anchors Recent work involving GPI-anchored proteins has begun to establish the heterogeneity of these moieties. Thus, while all anchors consist of a conserved core structure, significant variability occurs in the attached side chains, as the carbohydrates present along this core differ among proteins. For example, the anchors of the protein NCAM, when isolated from a homogeneous cell preparation, consist of a structurally heterogeneous set of anchors (Mukasa et al., 1995). Presently, no correlation between these structural differences and protein function has been accomplished. The lipid moieties composing the anchor can also be modified, as most anchors, when added to proteins, contain a palmitoyl chain attached to the inositol, which is then removed in the majority, but not all, of these proteins (Roberts et al., 1988). When this extra lipid is not removed from the anchor, the result is a more stable association of the protein with the membrane, and a protein that is resistant to cleavage by the enzyme PI-PLC. Recent reports have also demonstrated that the lipid chains of GPI anchors are modified in the Golgi by the enzymes PGAP2 and PGAP3 (Maeda et al., 2007; Tashima et al., 2006). The addition of a particular functional anchor is determined by a short stretch of amino acids located within the GPI anchor signal sequence (Nicholson and Stanners, 2007). It was hypothesized that the sequences in this region from CC6 (PVLSAV) and CC7 (PDLSAG) were sufficiently similar to that of CEA (PGLSAG) to result in the addition of the same GPI anchor. However, it appears that these slight differences result in the addition of functionally similar, but not identical, anchors. Whether this is the result of different processing of the anchor before or after addition is unknown; regardless of the timing, the different effects of ΔNCEA on NCB versus N6 and N7 would argue that the difference is specified by the GPI anchor signal sequence. Recent work from our laboratory has examined the properties of several other GPI anchors, including those from Thy-1, alkaline phosphatase, DAF, and uPAR (Zhai, Nicholson and Stanners, manuscript in preparation). The characterization of these anchors, which were all attached to the external domain of CC1 for the sake of comparison, demonstrated that none of the chimeras blocked differentiation. This suggests, with the caveat that many more anchors need to be examined, that the properties of the CEA family anchors are rather unique, as none of

125 these other anchors are able to show CEA-like properties. This also indicates that anchor- mediated targeting, as shown in Figure 2, should specifically interfere with CEA’s function, and not that of other GPI-anchored proteins, including other CEA family members, in cells expressing several proteins with GPI anchors. This could be an advantage in both basic and clinical applications of the GPI anchor-specific inhibition phenomenon.

Heterogeneity of membrane rafts To exploit the heterogeneity of GPI anchors, the plasma membrane must also have a corresponding set of heterogeneous rafts on the cell surface. Rafts are primarily composed of saturated lipids and cholesterol, but the composition of different domains appears quite diverse. For example, the raft resident gangliosides GM1 and GD3 are present in different rafts (Fujita et al., 2007), while various GPI-anchored proteins, such as CEA and NCAM, do not inhabit the same membrane regions (Nicholson and Stanners, 2006; Screaton et al., 2000). Therefore, the localization of a protein to a specific raft plays a substantial role in determining its resulting function, as altering the distribution of NCAM, by changing its anchor, is sufficient to completely alter function (Screaton et al., 2000). The mechanism by which specific rafts form and attract specific proteins remains uncharacterized and promises to represent an exciting area for future study. It is intriguing that the rafts that CEA, CC6 and CC7 localize to are functionally similar, suggesting that the function of various non-identical rafts overlaps. The reason for this redundancy is unknown, although it would likely provide another mechanism to fine-tune the signaling cascades resulting from the activation of each protein. CEA, CC6 and CC7 are expressed, in the adult colon, in the apical brush border of terminally differentiated colonocytes (Hammarstrom et al., 1998; Ilantzis et al., 1997), while integrins are primarily present on the basolateral surface (Agrez and Bates, 1994). This compartmentalization means that CEA and these signaling partners are separated, in different membrane regions, in differentiated colonocytes; however, overexpression of CEA in undifferentiated, non-polarized cells at the base of the crypt, where it is not normally expressed, may result in aberrant association of this protein with signaling molecules like integrins, contributing to cancer development. This could lead to different functions for CEA in healthy differentiated cells, such as its hypothesized role in immunity (Hammarstrom and Baranov, 2001), while promoting cancer progression in other cells.

126 This is one example of raft heterogeneity based on cellular context, as in cancer cells the rafts can bring together CEA and integrins, potentially altering protein function. The results presented in this study demonstrate that the GPI-anchored members of the CEA family all have similar functional properties, which is mediated by their anchors, but that they do not inhabit the same membrane regions. This implies that while rafts are heterogeneous in the plasma membrane, different rafts can have similar functions, possibly offering a mechanism to fine-tune the signaling properties of related proteins, through their targeting to related but distinct rafts. This, for example, could result in different signals if just CEA is activated, compared to cells that have both CEA and CC6 activated. Further exploration and confirmation of this observation will clarify the role of this extra level of specificity in the function of cell surface proteins.

127

Chapter 5

General discussion

128 GPI anchorage is a relatively common modification seen for approximately 10% of all eukaryotic cell surface proteins. Recent evidence has demonstrated that these anchors are not all functionally equivalent, as different GPI-anchored proteins reside in separate regions of the membrane, and the specific anchor added to the external domains greatly influences function (see Chapter 1). The existence of different functional anchors, which has only recently been demonstrated experimentally, has important consequences for all membrane proteins with this type of modification. The initial studies on GPI-anchored proteins identified the signal sequence that mediates the addition of the anchor, as well as the enzymatic steps required for the biosynthesis and transamidation of this anchor. With the recent demonstration of functionally unique GPI anchors, new avenues of research have been opened, where the effect of this anchor on the cell surface function of the mature protein can now be explored. CEA is a GPI-anchored protein that is upregulated in a substantial percentage of human cancers. The Stanners laboratory has been instrumental in characterizing the tumorigenic properties of this protein, including its ability to block both differentiation and the apoptotic process of anoikis through the activation of specific integrin receptors (Camacho-Leal et al., 2007; Eidelman et al., 1993; Ordonez et al., 2000; Ordonez et al., 2007). CEA’s biological properties require two separate structural elements, a self-adhesive external domain that is attached to the CEA-specific GPI anchor (Eidelman et al., 1993; Screaton et al., 2000). Thus, the specific properties of the CEA GPI anchor can be elucidated by attaching the anchor to unrelated external domains, provided they have a self-adhesive function. The functional properties that are specified by the CEA anchor are now well-established, so the presence and activity of the CEA GPI anchor can be readily assayed by testing for effects on integrin function and cellular differentiation. We have exploited this fact, in the studies presented in this thesis, to demonstrate that: the GPI anchor of CEA is responsible for targeting the protein to specific membrane domains, a property that can then be exploited to modify protein function (Chapter 2); a novel signal exists within the GPI anchor signal sequence of CEA, which is responsible for the addition of the CEA-specific anchor (Chapter 3); the GPI anchors of CC6 and CC7 are not identical to that of CEA, despite their similar biological properties (Chapter 4); and, the GPI anchor of CEA contains the information required for incorporation of the protein into the membrane from the extracellular medium (Research Appendix).

129 The specificity of GPI anchors Experiments performed by Screaton et al. demonstrated that exchanging the GPI anchor of NCAM, a protein that accelerates differentiation, for that of CEA results in a chimeric protein, NCB, that blocks differentiation, indicating that GPI anchors contain specific information determining protein function (Screaton et al., 2000). One possible explanation, given the direct association of this anchor with the plasma membrane, is that the anchor affects the localization of the protein. Indeed, CEA and NCAM, which have very different biological properties, do not co- cluster on the cellular surface. Various other combinations of GPI-anchored proteins, including the folate receptor and placental alkaline phosphatase (Wang et al., 2002b), Thy-1 and the prion protein (Madore et al., 1999), and DAF and the prion protein (Li et al., 2003), have non- overlapping distributions, suggesting that this is a fairly widespread phenomenon. The role of the anchor was directly tested by examining the relative surface distribution of NCAM and NCB. NCB, but not NCAM, colocalizes with ΔNCEA on the cell surface, indicating that proteins with the same anchor are targeted to the same regions of the membrane (Chapter 2). Therefore, it is the anchor, and not the external protein domains, that determines the localization of the protein on the cell surface. The targeting of GPI-anchored proteins, which do not have a direct connection with the cytoplasm, to particular regions of the membrane likely determines the signaling molecules associated with the protein. For example, NCB is able to activate specific integrins, similar to CEA, a property that is not seen for NCAM (Chapter 2). How this targeting occurs remains an open question, although it can be hypothesized that the anchor may specify the interaction of the protein with a particular subset of membrane rafts that harbor a particular subset of signaling molecules (see later).

The role of clustering in protein function While the properties of CEA are determined by the CEA-specific GPI anchor, its biological properties require the attachment of a homophilic adhesion-competent external domain to the anchor. The external domain of CEA mediates double reciprocal anti-parallel adhesion between molecules on the surface of apposed cells (Zhou et al., 1993a); deletions or mutations that abrogate the adhesive abilities of CEA result in a loss of biological function (Eidelman et al., 1993; Taheri et al., 2003). CEA point mutants can be isolated that retain the ability to mediate intercellular adhesion but no longer block differentiation (Taheri et al., 2003), but all non-

130 functional CEA deletion mutants tested can have their function restored by crosslinking with antibodies (Camacho-Leal et al., 2007; Taheri et al., 2003). These results suggest that a major role of the CEA external domain involves parallel interactions between molecules on the surface of the same cell. Thus, the external domains of CEA are required to mediate both parallel (on the surface of one cell) and anti-parallel (between two apposed cells) inter-molecular adhesion. In Chapter 2, we demonstrate a complete loss of the biological properties of CEA that is accompanied by a partial loss of the intercellular adhesion function, implying a reduction in parallel and anti-parallel CEA-CEA binding. This effect can be overcome by artificial clustering with antibodies, indicating the critical importance that this parallel clustering has for the initiation of signaling. Parallel interactions are promoted, in part, by membrane rafts, which concentrate GPI- anchored proteins through specific associations with the anchor, resulting in an increased local concentration of these proteins (Harris and Siu, 2002). This encourages a combination of parallel and anti-parallel protein interactions, leading to protein clustering and the concomitant clustering of the associated rafts. Clustering rafts serves to stabilize these domains as large complexes, resulting in the formation of platforms that lead to the initiation and transduction of intracellular signaling (Subczynski and Kusumi, 2003). This explains how the NCAM extracellular domain can produce the same effects as CEA, when attached to the CEA anchor, as signaling results from the clustering of the specific rafts targeted by the anchor. It remains to be determined the extent of clustering that is required for the initiation of signaling. The demonstration that CEA expression must be greater than a minimum threshold to block differentiation (Screaton and Stanners, unpublished data) would suggest that there is a minimum amount of clustering needed to initiate signaling. Future work could attach the GPI anchor of CEA to the external domain of a protein where the level of clustering can be readily controlled. For example, using the external domain of a receptor that dimerizes upon ligand binding would indicate if dimerization is sufficient for CEA’s effects, or if a higher level of molecular clustering is needed.

Heterotypic interactions The proteins of the CEA family can undergo heterotypic interactions (Oikawa et al., 1989; Oikawa et al., 1992; Zhou et al., 1990), which offers the possibility of novel interactions in

131 tissues expressing multiple family members, such as the colon (Frangsmyr et al., 1999). The functional consequences of homo- versus heterotypic adhesion by these proteins remain to be determined. It is possible that CEA-initiated signaling varies depending on whether it is interacting homotypically with another CEA molecule, or if it is interacting with a different CEA family member. This is of particular interest because CC6 and CC7 do not appear to be targeted to the same domains as CEA (Chapter 4). Thus, adhesion involving, for example, CEA and CC6 may result in the aggregation of multiple related, but not identical, rafts, leading to the propagation of a different signal from that occurring after homotypic adhesion by CEA. This homo- versus hetero-aggregation of rafts, if it is found to result in appreciably different signaling, would have implications for the function of many raft-resident proteins. For example, activation of the T cell receptor involves the aggregation of multiple lipid rafts (He et al., 2005a). If the aggregation of different combinations of rafts occurs depending on the external stimuli, the result could be the fine-tuning of the signal from this complex (Janes et al., 2000; Magee et al., 2002). In fact, it has been determined that rafts play an important role in certain immunological disorders. Patients with systemic lupus erythematosus show higher overall levels of rafts in the plasma membrane, and these rafts concentrate a different set of proteins, alterations that result in abnormal T cell responses (Krishnan et al., 2004). This provides a basis for the suggestion that protein function is determined by the raft that it is associated with, and that changing the properties of these rafts can have important consequences for protein function. The specific aggregation of rafts may provide a general mechanism by which signaling specificity can be modulated, offering a method by which the properties of a protein can vary depending on cellular context.

The existence of membrane rafts GPI-anchored proteins are thought to localize to rafts in the plasma membrane, domains that serve to concentrate these proteins and bring them in close proximity to specific signaling molecules. Recently, the existence and basic characteristics of rafts have been questioned, which has implications for the study of GPI-anchored proteins (Munro, 2003). Numerous experiments have provided both positive and negative evidence for the presence of rafts (see Chapter 1), resulting in a lack of consensus regarding their properties. Due to mounting evidence (see

132 Chapter 1), the current consensus is that rafts do exist, as small heterogeneous entities that can be stabilized to form larger domains (Hancock, 2006; Harris and Siu, 2002; Pike, 2006). Even in studies providing positive evidence for the existence of rafts, significant variability is observed in the experimentally predicted sizes of these domains. For example, using chemical crosslinking, the size of rafts was hypothesized to be slightly greater than 4 nm (Friedrichson and Kurzchalia, 1998), while other studies using microscopic techniques have estimated that rafts are as large as several hundred nanometers (Dietrich et al., 2002; Schutz et al., 2000). These discrepancies may result from the use of different cell types, as work with model membranes has demonstrated that the size of the domain formed depends on the lipid composition of the membrane (de Almeida et al., 2005). Thus, cells that naturally have different compositions of their plasma membrane may have rafts that differ in size. This has been confirmed using the protein VP4, which associates with rafts that have different physical properties depending on the cell line it is expressed in (Delmas et al., 2007). Furthermore, the activation state of GPI-anchored proteins determines raft shape and size (Subczynski and Kusumi, 2003), so studying cells where these proteins are in different activation states may also provide divergent values for raft size. This could also be one factor in studies that have been unable to find evidence for rafts, as the small, non-stable rafts that represent resting state domains may be difficult to distinguish from the bulk plasma membrane. Thus, while more in-depth studies are needed to clarify the specific properties of rafts, it seems likely that these domains exist and are important for GPI-anchored protein function.

Membrane raft heterogeneity The non-overlapping distribution of different GPI-anchored proteins suggests that these anchors target proteins to specific rafts on the cell surface, which was confirmed by results presented in Chapter 2. Rafts have been studied extensively in recent years to identify their role(s) in the biological function of cells, and they have been implicated in many processes (see Chapter 1). These domains are not homogeneous on the cell surface, but consist of multiple types within the membrane of one cell. GPI-anchored proteins show specific targeting to a subset of membrane domains, resulting in a matched set of heterogeneous rafts and proteins in the membrane. For example, the rafts inhabited by the prion protein and Thy-1, proteins that have non-identical distributions on the cell surface (Madore et al., 1999), have very different lipid compositions

133 (Brugger et al., 2004). In particular, the prion protein is targeted to domains that contain lipids with a higher degree of unsaturation, and with higher levels of the lipid hexosylceramide (Brugger et al., 2004). This targeting to specific domains determines the ultimate function of the protein, and can be exploited to modify protein function. In Chapter 2, we demonstrate that the expression of ΔNCEA specifically inhibits the function of NCB, while NCAM, which differs from NCB only in its GPI anchor, is not affected by ∆NCEA and has no effect on the properties of CEA. It is interesting to note that the presence of ΔNCEA results in an increase in the size of the rafts that NCB is associated with. However, no apparent difference was noted in the size of the domains containing GM1, a lipid that is commonly employed as a raft marker, demonstrating raft heterogeneity and suggesting that the CEA-specific rafts do not contain this lipid. It will be important to determine how the rafts inhabited by CEA and NCAM differ in their composition, as the localization of a protein to these rafts has significant consequences on function. Targeting a self-adhesive protein to the CEA-specific raft, by its GPI anchor, generates a protein that blocks differentiation, while targeting the same external domains to the NCAM- specific rafts, by changing its anchor, leads to a protein without this property (Screaton et al., 2000). Identifying the lipid and protein differences between the rafts inhabited by CEA and NCAM will help explain the effects these domains have on function. The observation that the properties conferred by the CC6 and CC7 anchors cannot be inhibited by the co-expression of ∆NCEA implies that the anchors of these proteins do not target exactly the same domains as the CEA anchor (Chapter 4). However, many of the biological functions of CEA, including integrin activation and differentiation blockage, are seen for both CC6 (Camacho-Leal et al., 2007; Rojas et al., 1996) and CC7 (Zhai and Stanners, manuscript in preparation). This would suggest that the domains inhabited by these proteins are functionally related, but not identical, increasing the number of rafts that can theoretically be found within the plasma membrane. It would be of particular interest to compare the composition of the rafts inhabited by CEA, CC6 and CC7, to determine how their components differ, and how this compares to less closely-related proteins. Extending this type of analysis to other GPI-anchored proteins should make it possible to have a better understanding of the compositional differences of functionally-distinct rafts. This could also be used to determine whether every raft-resident TM protein is targeted to all domains, or if there is specificity in their targeting as well, a possibility that has yet to be examined. For these studies, the isolation of these rafts would likely be performed using the immunoaffinity isolation

134 protocol employed to study the rafts inhabited by Thy-1 and the prion protein (Brugger et al., 2004), which was found to result in a reproducible raft population for each protein that could be studied, by mass spectrometry, to determine their composition. A further question that arises is how do lipids associate to form specific domains? The current model of raft formation suggests that the lipids that compose these domains contain saturated hydrocarbon chains that permit tight packing, resulting in the formation of a liquid- ordered region that is separate from the bulk membrane (van der Goot and Harder, 2001). How this results in the formation of distinct domains, however, remains to be determined. Likely explanations include the possibility that there are interactions between the head groups of these lipids, or that there may be a requirement for proteins to organize these domains. It is also possible that the formation of a specific raft is dependent on the presence of all of the components. For example, while all the other molecules that compose the CEA-specific raft may be present in many cell types, this particular domain may actually only form in cells that express CEA. If so, this could explain the results presented in this thesis using the CEA family (Chapter 4), suggesting that each GPI-anchored protein could have its own unique associated raft. By comparing the rafts inhabited by numerous GPI-anchored proteins, a better understanding of raft formation and heterogeneity should result. It may be that rafts can be classified into related groups, with associated functions, similar to the way the anchors of the CEA family members can be grouped functionally, despite their apparent differences.

Targeting to different rafts GPI-anchored proteins are known to sort to membrane rafts on the cell surface (Brown and Rose, 1992; Schnitzer et al., 1995), with the results presented in Chapter 2 indicating that this targeting is specific to a subset of raft domains. This has important consequences for the function of GPI- anchored proteins, which lack TM or cytoplasmic domains and as such require other proteins to transduce their signal to the interior of the cell. Therefore, the rafts containing GPI-anchored proteins must also be inhabited by specific signaling molecules for the correct downstream signaling by these proteins. A mechanism must exist by which these other proteins also localize to specific rafts, or all GPI-anchored proteins would have the same biological functions. TM proteins contain information in both their membrane-spanning and extracellular domains that specifies their overall localization to rafts, although whether this leads to association with

135 specific domains is unknown. A membrane-proximal domain, which was first identified in an HIV surface protein but which exists in many proteins, has the ability to bind specifically to the carbohydrate components of lipids (Fantini, 2003; Mahfoud et al., 2002). The exact amino acid composition of this domain may regulate the specific interaction of the protein with a particular raft, as different amino acids may direct an association with a defined lipid class. Furthermore, cytoplasmic proteins with specific lipid modifications are localized to raft-like domains that are linked to the rafts present in the exoplasmic leaflet of the membrane. This targeting depends on the type of modification, as the addition of acylated lipids promotes raft association while prenylation does not (Lucero and Robbins, 2004). The exact mechanism that leads to the specific targeting of these three types of proteins to the same membrane region, resulting in the initiation of a particular signaling event, is only beginning to be examined, with much information still to be determined. Ectopic expression of CEA, in many cells lines derived from several organisms, consistently results in the activation of the integrin α5β1 (Ordonez et al., 2007). Thus, the targeting mechanism for both proteins appears to be conserved in mammals, suggesting that protein function is often determined in this manner. CEA and the integrin α5β1 colocalize on the surface of the cell, most likely by inhabiting the same membrane rafts (Camacho-Leal et al., 2007; Ordonez et al., 2007). The conservation of targeting for both integrins and GPI-anchored proteins was confirmed in a transgenic mouse expressing several CEA family members, where the expression of CEA resulted in a specific increase in the colonocyte cell surface level of integrin α5β1 (Chan, Camacho-Leal, and Stanners, manuscript submitted). Initiation of CEA signaling results in the recruitment to detergent-insoluble domains of several proteins, including ILK and AKT (Camacho-Leal et al., 2007). The mechanism by which these proteins alter their membrane localization in an activation-dependent manner is currently unknown and may provide insights into how proteins localize to specific domains. It may be that any type of targeting, provided it leads to an association with the desired membrane domain, is sufficient to give the correct function. To examine this possibility, it would be interesting to see if attaching the external domain of CEA to the TM and cytoplasmic domains of the integrin α5β1 could recapitulate CEA’s activities. This chimera could, theoretically, be sufficient on its own to initiate signaling, with the adhesive ability of the CEA external domain clustering the rafts and

136 the cytoplasmic integrin domain providing direct association with cytoplasmic signaling molecules.

GPI anchor signal sequence The signal sequence specifying the addition of a GPI anchor is located at the C-terminus of the protein, and consists of a set of small amino acids, followed by a hydrophilic region and ending in a stretch of hydrophobic residues (Eisenhaber et al., 1998). The requirements for efficient GPI anchorage have been extensively studied, resulting in a detailed understanding of the residues that are, and are not, amenable to anchor addition. However, it has been noted that the actual sequences of these signals are very divergent between proteins, a fact that has not been carefully studied for functional significance. This is because the GPI anchor signal sequence is cleaved in the ER, but the mature protein exerts its effects on the cell surface, so the possibility that the signal sequence affects protein function would seem to be remote. However, our laboratory demonstrated that exchanging the GPI anchor signal sequence of NCAM for that of CEA was sufficient to complete alter protein function, indicating that the signal sequences of these proteins have major consequences for protein function (Screaton et al., 2000). Because the signal sequence is cleaved upon addition of the anchor, the signal must specify protein function in the ER, likely through the addition of a particular GPI anchor. We have accomplished an initial characterization of this signal, as presented in Chapter 3, where we demonstrate that a stretch of 6 amino acids within the CEA GPI anchor signal sequence is required for the addition of the CEA-specific anchor. This sequence is also sufficient, when inserted into the corresponding position of the NCAM signal sequence, to specify the addition of the CEA GPI anchor. Since we show that scrambling this region does not result in a complete loss of function, the exact identity of this signal remains unknown. Further CEA-NCAM chimeras, as well as the examination of the sequences of other proteins, will help clarify the precise nature of this signal. In Chapter 4, we demonstrate that while the anchor signal sequences of CC6 and CC7 are similar to that of CEA, the anchors that are added are not identical. There are small differences in the critical 6 amino acid region that was identified in Chapter 3, so the addition of specific anchors appears to be very dependent on the amino acid composition. This is of particular interest because the resulting proteins have similar functional properties despite their different anchors, so different but functionally related anchors must exist.

137 A possible hypothesis regarding the nature of the signal is that this stretch adopts a specific structure, which is complimentary to the structure of the anchor precursor. Only anchors with side-chains that “fit” the conformation of this region would be efficiently added to the proteins. The requirement for the proline immediately upstream of this sequence, as shown in Chapter 3, would seem to support the notion of the requirement for a particular conformation. As suggested in Chapter 3, this proline may be part of a motif that adopts a specific conformation, strengthening the argument that the structure is important for determining the addition of a particular anchor. The other likely possibility is that the addition of the specific anchor is mediated by one of the components of the GPI anchor transamidase complex. This complex is currently thought to consist of 5 subunits (hGAA1p, GPI8, PIG-S, PIG-T and PIG-U), all of which are required for anchor addition. The exact function of each of these components remains unknown, although each is required for the proper function of the complex. An in-depth examination of the role of each protein is difficult, as removing any of the subunits leads to a complete loss of enzyme activity. However, understanding the role of each member of the complex should demonstrate which, if any, subunit is involved in determining specific anchor addition. This question could be approached by determining the structure of the transamidase complex with the GPI anchor precursor and the signal sequence present, and identifying the specific interactions. Alternatively, an in vitro system for GPI anchorage has been developed (Chen et al., 1996), which may allow for manipulation of the transamidase process to elucidate the method for determining specific anchor addition. While the GPI anchor can be remodeled after addition to the protein, primarily by proteins located in the Golgi apparatus (Maeda et al., 2007; Tashima et al., 2006), it is hard to reconcile how the signal sequence, which is cleaved in the ER, could direct this process. As mentioned above, it may be that a currently unidentified enzyme exists in the ER, most likely associated with the transamidase complex, which modifies the anchor structure immediately upon transamidation of the protein. It has been noted that the anchor often has a third lipid chain removed from the inositol ring in the ER, although this occurs in the majority of anchors and is not specific to particular proteins. Alternatively, multiple GPI anchor precursors may be generated in the ER, with a “mix-and-match” process happening where complementary signal sequences and anchors come together. This also therefore requires that each individual GPI anchor precursor is generated and present in the ER prior to anchor addition. It has recently been

138 demonstrated that Ras plays a role in the regulation of the first step of GPI anchor synthesis, indicating that different signaling pathways could have input in the synthesis of various GPI anchors (Sobering et al., 2004). This leads to the possibility that different functional anchors may be present in a cell depending on its biological context, which offers another method for specifying protein function.

GPI anchor structure Significant differences in protein properties occur depending on the GPI anchor signal sequence that is found at the C-terminus of the primary amino acid sequence (Screaton et al., 2000). However, as this sequence is cleaved in the ER, we hypothesize that different structural features of the anchor, determined by this signal sequence, are the ultimate source of this specificity. The core structure of the anchor is conserved in all structures that have been analyzed to date, regardless of the organism (Englund, 1993), although many variations in the side chains have also been noted. The common type of modification that is seen for most anchors involves the addition of carbohydrate residues along the core stretch. How these different sugars could alter protein function is currently unknown, and to date no correlation has been made between the physical structure of a GPI anchor and the functional properties of the protein. It has now become standard, and relatively straightforward, to determine the structure of GPI anchors using mass spectrometry (Omaetxebarria et al., 2006). It will be important to determine the structural differences between the CEA and NCAM anchors, in order to clarify the features that might specify CEA function. How a particular structure then specifies the membrane localization of the protein is difficult to picture. Perhaps these structures have an affinity for the components of the rafts that CEA is targeted to, based either on the charge or the size of the molecules. This will need to be studied, likely through the use of chemically synthesized GPI anchor moieties with known and specific substitutions. These can then be examined for association with specific raft domains, to determine the targeting signal.

How many different GPI anchors exist? The observation that the GPI anchors of CEA and NCAM are not functionally equivalent (Screaton et al., 2000) leads to the question of how many different GPI anchors exist. This has been difficult to address, largely because it is difficult to separate the role of the dissimilar

139 external protein domains from that of the GPI anchor. For example, while our laboratory has established the critical role of the GPI anchor in determining the function of CEA (Screaton et al., 2000), we have also noted the requirement for self-adhesive external domains for the anchor’s effects (Eidelman et al., 1993; Taheri et al., 2003). Similar requirements for specific external domain functions may be important for other proteins. Therefore, a legitimate comparison of the functional properties of various GPI anchors requires that each anchor be attached to the same external domain, eliminating the effects of the different external protein domains. This type of study, on a small scale, has been undertaken in our laboratory, where the GPI anchors of several proteins that have altered expression in cancer (including alkaline phosphatase, DAF, Thy-1 and uPAR) have been genetically attached to the external domain of CC1 (Zhai, Nicholson and Stanners, manuscript in preparation). The preliminary characterization of these proteins involved testing for CEA-like properties, notably the ability to block differentiation, with the demonstration that this property appears restricted to the CEA family. While this suggests that the CEA-family anchor is relatively unique, it does not answer the question of how many anchor classes exist. Further experiments will be required with a greater number of GPI anchors and more thoroughly characterized functional effects before the various GPI anchors can be categorized. As well, once the properties of a particular anchor are determined, it will be interesting to determine if this protein’s anchor is identical in other species. In the CEA family, the anchor from a GPI-anchored protein from the primate Callicebus molloch was demonstrated to block differentiation and activate integrins similar to CEA (Naghibalhossaini et al., 2007). Therefore, CEA family proteins from various species appear to have similar functions, which are mediated by the GPI anchor, but the universality of this effect requires the examination of other proteins. The possibility that this is the case for other proteins is strengthened by the fact that the anchors of porcine and human renal membrane dipeptidase consist of the same 3 anchor variants (Brewis et al., 1995). It will also be important to determine if proteins from similar functional classes have similar functional properties. In Chapter 4, we demonstrate that the anchors of CC6 and CC7 are able to mediate similar biological effects to that of CEA. Thus, both anchors, when attached to the NCAM external domain, result in a chimera with the ability to block of differentiation and modulate integrin α5β1 function. However, despite being functionally similar to that of CEA,

140 these anchors are not identical to CEA’s GPI anchor. Therefore, this adds a further level of complexity to the concept of GPI anchor specificity, as different anchors appear to be able to mediate similar effects. How common this property is in other protein families will be interesting to analyze, in order to understand how many different anchors exist. This will also allow a further examination of the signal that is described in Chapter 3, by comparing the similarities between signal sequences and the resulting functional properties of the GPI-anchored protein. It would be important to generate CC6 and CC7 constructs with non-adhesive external domains (similar to ΔNCEA) to determine if the N6 and N7 proteins can be specifically inhibited by a defective protein with the same anchor. This would confirm that the anchors of CC6 and CC7 are not identical to that of CEA, rather than being unaffected by the shank-defective anchor effect described in Chapter 2. If specificity is demonstrated, i.e., that functional inhibition of NCB, N6 and N7, requires matching shank-defective anchors, their respective 6 amino acid sequences could be inserted in the anchor signal sequence of NCAM to see if this specificity is conferred on NCAM, as shown in Chapter 3 for the CEA sequence.

Targeting GPI anchors to alter protein function The role of the GPI anchor in determining localization of the protein to specific membrane domains can be exploited to inhibit protein function, as described in Chapter 2. This leads to the possibility that targeting the GPI anchor of CEA could be a novel method for interfering with the propagation of cancer cells over-expressing this protein. While this is a novel concept for GPI anchors, treatment with lipids to modulate cellular function, and disease symptoms, has previously been attempted. Modulating plasma membrane components, by either adding gangliosides, ceramide or polyunsaturated fatty acids from the external medium or by cholesterol depletion, inhibits raft mediated signaling (Megha and London, 2004; Simons et al., 1999; Stulnig et al., 1998; Stulnig et al., 1997). This offers a general mechanism whereby protein function could be modified by the addition of specific lipids. Some of these studies have been validated by results in experimental organisms and in patients. In vivo, treatment with gangliosides has specific effects, in particular in the recovery of laboratory animals from brain injury (Fong et al., 1995; Hadjiconstantinou and Neff, 1998; Tan et al., 1993). A particular use of gangliosides lies in the treatment of Parkinson’s disease, where primate models of this disease show improved symptoms following ganglioside administration (Schneider et al., 1992). In

141

Figure 1. Schematic of shank-defective or shank-less anchors. (A) GPI-anchored proteins with intact external domains. (B) Types of modifications that can be employed to give shank-defective external domains. Deletions (left) or point mutations (right; as shown by the asterisk) can be introduced into the external domain, based on the knowledge of the external domain requirements for function, rendering the protein non-functional. Note that these alterations would require a complete loss of function to serve as dominant-negative modulators of GPI-anchored protein function; (C) In order to avoid these unwanted interactions, major deletions (left) or a completely different external domain (right) can be attached to the anchor. For true shank-less anchors, the GPI anchor will be generated lacking external protein domains completely (center).

142 humans, treatment with gangliosides also results in an improvement in the symptoms of Parkinson’s, stroke and spinal cord injury (Geisler, 1993; Geisler et al., 1991; Lenzi et al., 1994; Schneider et al., 1998). GPI anchors consist of a distinct class of lipid, so we hypothesize that cellular function can also be modulated through the application of these lipids. Proteins with GPI anchors have the ability to incorporate into the plasma membrane of cells when applied externally due to this lipid modification (Ilangumaran et al., 1996), which we confirmed to also be a property of the CEA GPI anchor (see Research Appendix). Painted proteins eventually regain raft association and signaling properties, indicating that they are able to target their “native” domains (Premkumar et al., 2001). Therefore, the CEA GPI anchor contains the information necessary for its incorporation into the membrane and targeting to specific rafts once incorporated. As demonstrated in Chapter 2, targeting CEA-specific rafts with non-functional GPI-anchored proteins can inhibit the function of an active CEA-like molecule. However, external domain interactions need to be minimized (Chapter 4) for this effect to be manifest. By using either CEA mutants that are incapable of self-association or the CEA specific GPI anchor completely lacking external protein domains (Figure 1), it may therefore be possible to reverse the effects of CEA overexpression by tumor cells. These molecules would be able to incorporate into the membrane and target to the proper domains. The result could be the specific inhibition of CEA function in cells overexpressing this protein, which in adult humans consists primarily of many different types of cancer cells (Figure 2). This would inhibit the biological effects of CEA, such as the block of differentiation and anoikis, and may offer a novel paradigm for cancer treatment in the future.

143

144

Figure 2. Using shankless anchors to modify GPI-anchored protein function. Model of GPI-anchored protein association on the cell surface. On the cell surface, GPI-anchored proteins and membrane rafts co-exist, with specific proteins being targeted to specific rafts depending on their anchor (A). This targeting leads to a localized increase in concentration of the proteins (B), which allows them to undergo parallel interactions (C). This interaction on the cell surface serves as a priming mechanism, allowing the proteins to mediate intercellular adhesion and initiate downstream signaling. As the GPI anchor is the moiety that targets proteins to rafts, even proteins with mutated external domains (as signified by the triangle) are targeted to their specific domains (C). This results in rafts that contain both the functional and non-functional proteins, which show an increased size (D). This results in a decreased ability of the functional proteins to undergo parallel interactions (E), resulting in less clustering of these proteins (F) and a lack of the initiation of downstream signaling.

145 Final comments Despite having been first identified more than 30 years ago, much remains to be determined regarding GPI anchors. The initial focus on these anchors rested on determining the proteins that contain this modification, characterizing the signal that mediates their attachment to proteins, and elucidating the biosynthetic pathway for their generation. However, recent studies have begun to explore the functional consequences of this alteration. In the studies presented in this thesis, we offer novel insights into the specificity of this anchor, and the role it plays in determining the function of CEA. With the information provided in this thesis, future experiments should allow for a broader examination of the effect of the GPI anchor on the function of various cell surface proteins. This should also include greater knowledge about the rafts that these proteins are targeted to. Ultimately, the role the GPI anchor has in determining protein and cellular function has been underappreciated until just recently, with the work presented in this thesis shining new light on the molecular basis for the functional specificity of this important structure.

146 Preface to the Research Appendix In the previous chapters, we examined the source and consequences of the functional specificity of the GPI anchor of CEA, a modification that is a critical determinant of the function of this protein. This knowledge may eventually be useful for treating patients, as approximately 50% of all human cancers over-express this protein. Because our laboratory has established that CEA plays an instrumental role in cancer progression, inhibiting the functions of this protein may improve the outlook of patients with tumors that have upregulated CEA expression. In Chapter 2, we demonstrated that non-functional GPI-anchored proteins could inhibit the function of CEA when co-expressed by the same cells. A similar process could be used in patients, provided a method for delivering the non-functional anchor can be devised. Interestingly, certain GPI- anchored proteins have been shown to incorporate into the plasma membrane from the external medium, which could eventually be used to deliver non-functional GPI anchors to cells over- expressing CEA. In this Research Appendix, we present an initial demonstration confirming that the GPI anchor of CEA possesses the requisite structure for membrane insertion.

147

Research Appendix

The GPI anchor of CEA mediates external membrane incorporation (“painting”)

148 Abstract GPI-anchored proteins can incorporate, from the external medium, into the plasma membrane of cells, in a process known as “painting”. We tested whether the GPI anchor of CEA contains sufficient information to mediate this effect, and demonstrate that L6 cells can be efficiently painted with purified CEA. The amount of CEA incorporated can be increased by painting cells in suspension, rather than monolayer, and through the addition of BSA to the painting medium. Therefore, this epigenetic method of modifying cells is a property of CEA, and may eventually be employed to alter cellular function.

149 Introduction Carcinoembryonic antigen (CEA) is a cell-surface protein that is attached to the plasma membrane by a glycophosphatidylinositol (GPI) anchor. This protein is over-expressed in about 50% of all human cancers, suggesting that it may play a role in cancer progression (Blumenthal et al., 2007). In vitro, CEA blocks differentiation and the apoptotic process of anoikis and disrupts tissue architecture (Eidelman et al., 1993; Ilantzis et al., 2002; Ordonez et al., 2000), which could explain its frequent upregulation in cancer. Because CEA is consistently over- expressed by cancer cells, it is an attractive target for cancer therapies, and numerous methods that exploit this protein to inhibit tumor growth are currently being explored (Berinstein, 2002; Chatal et al., 2006; Kuroki et al., 2005; Kuroki et al., 2004; Marshall, 2003; Stein and Goldenberg, 2004). Due to the lipid components of the GPI anchor, proteins with this modification can be isolated from one cell and reincorporated into the membrane of other cells exogenously (Ilangumaran et al., 1996). The transferred proteins gradually regain raft association and signaling abilities, following incorporation (Premkumar et al., 2001). This protein insertion, which is commonly referred to as “painting,” allows for epigenetic modulation of cellular function; for example, resistance to complement-mediated attack can be restored to cells genetically lacking the GPI-anchored protein DAF by painting them with this protein (Medof et al., 1984; Medof et al., 1985). Painting naturally occurs in vivo, as GPI-anchored proteins can be transferred from erythrocytes to the endothelium (Kooyman et al., 1995). The GPI anchor of CEA contains information that targets the protein to specific domains on the cell surface, which can be exploited to inhibit its function (Nicholson and Stanners, 2006). We hypothesize that treating cells, from the external medium, with CEA GPI anchors attached to non-functional external domains can also inhibit function. As an initial step towards verifying this hypothesis, we confirmed that the GPI anchor of CEA has a structure that can mediate membrane incorporation by painting. We also examined various conditions for their effect on the efficiency of CEA incorporation.

150 Materials and Methods Cell Culture The cell lines used in this study were CHO-derived LR-73 fibroblasts and rat L6 myoblasts, 0 which were cultured at 37 C in a humidified atmosphere containing 5% CO2. These cells had previously been transfected with the cDNA for full-length CEA (Benchimol et al., 1989; Eidelman et al., 1993), with expression confirmed using the CEA-specific antibody J22 (Zhou et al., 1993b). Their detailed properties are described in Chapter 2 of this Thesis.

Isolation of CEA CEA transfectant cells were seeded in four 100-mm dishes (Nunc) at a density of 104 cells/cm2. After 2 days, cells were collected with PBS-Citrate containing 4 mM EDTA, and lysed on ice with 60 mM n-Octyl-β-D glucopyranoside (Sigma-Aldrich) containing the protease inhibitors aprotonin (Roche), leupeptin (Roche) and PMSF (Sigma-Aldrich). Following centrifugation at 4°C to eliminate unsolubilized material, these whole cell lysates were dialyzed against serum- free DME and then used for painting. In certain cases, the whole cell lysate was treated with PI- PLC (Sigma-Aldrich) for 45 minutes at 37°C, prior to dialysis, to remove the GPI anchor. For purification of CEA, the mAb D13 (Zhou et al., 1993b) was coupled to Protein G-Sepharose beads (Santa Cruz Biotechnologies, Inc.) using dimethylpimelimidate (Sigma-Aldrich). The lysate was added to these beads and incubated at 4ºC overnight. The beads were then washed three times with 10mM phosphate, pH 6.8, and the CEA eluted from the beads by incubating for 1 minute at room temperature with 0.2 M glycine, pH 2.5, which was immediately neutralized with 1 M phosphate, pH 8.0. This was then dialyzed overnight against (serum-free) DME and promptly added to cells.

Incorporation of purified CEA L6 cells were collected with trypsin (Sigma-Aldrich; 0.063% in PBS-citrate) to generate a single- cell suspension. Either 1.5 x 105 of these suspended cells or attached cells in a sub-confluent 35- mm dish were then incubated in DMEM with the dialyzed CEA sample at 30°C for 45 minutes. The effect of essentially lipid-free BSA (Sigma-Aldrich) on incorporation was also examined by addition of 10 µg/ml of BSA to the incubation medium. Cells were then washed with PBS to

151 remove unincorporated proteins, and analyzed by FACS with J22, as previously described (Nicholson and Stanners, 2006), to determine incorporation.

152 Results Externally applied CEA can incorporate into biological membranes Certain purified GPI-anchored proteins can reincorporate into the cellular membrane on the basis of their GPI anchors (Ilangumaran et al., 1996), so we sought to determine whether the CEA anchor has this property. L6 parental cells were incubated, for one hour, with dialyzed cell lysates generated from both L6 (CEA) and LR (CEA) transfectants (Fig. 1). Under these conditions, CEA incorporation, as determined by FACS, was noted for cells treated with the L6 lysate, with an increase in fluorescence from a background value of 3 to a mean of 14. The LR lysate, which came from cells expressing higher levels of CEA (data not shown), resulted in a greater fluorescence signal after painting L6 cells, with a mean of 31. Painting cells with a CEA lysate that had been treated with PI-PLC, to remove the anchor of CEA, did not result in significant incorporation, as seen by FACS, indicating the importance of the anchor of this association (Fig. 1). This data demonstrates that CEA is able to incorporate into cellular membranes.

Efficiency of incorporation for monolayer versus suspension cells The conditions that promoted maximal membrane uptake were next examined. It was first tested whether insertion was more efficient for L6 cells in monolayer culture, or for trypsinized cells in suspension. These cells were painted with a solution of CEA that had been enriched by immunoprecipitation with the CEA-specific antibody D13, to remove contaminants such as membrane lipids that may interfere with painting efficiency (Medof et al., 1984; Nagarajan et al., 1995). Under these conditions, painting with the same preparation of CEA resulted in appreciably more incorporation for the cells in suspension, with a FACS mean of 33, compared to the monolayer culture, which had a mean of 17. This indicates that painting of CEA in solution is more efficient, likely because of better access of the anchored protein to the plasma membrane of the cells.

153

Figure 1. Painting L6 cells with whole cell lysates. L6 parental cells (top; “untreated”) were incubated with the dialyzed lysates of L6 (CEA) (second from top) and LR (CEA) (third from top). Incorporation of CEA from these lysates was determined by FACS. The L6 lysate increased the mean fluorescence compared to the untreated control, while the LR lysate gave an even greater increase, indicating that CEA can be painted. Treatment with PI-PLC, which removed the GPI anchor, resulted in a complete block of this effect, confirming that it is anchor specific.

154

Figure 2. CEA incorporation is more efficient for cells in suspension. CEA purified by immunoprecipitation was added to L6 cells either in monolayer or suspension and incubated for 45 minutes. FACS analysis demonstrated substantially greater incorporation for the single cell suspension (unfilled profiles). The filled profiles represent cells that were treated with DMEM containing no purified proteins.

155 BSA improves CEA incorporation A molecule that appears to improve painting efficiency is bovine serum albumin (BSA) (Zheng et al., 2001), so the effect of this macromolecule on CEA painting was examined. The presence of BSA resulted in a significant shift in the FACS peak of the parental cells to a mean of 95 (middle). To confirm the specificity of this effect, the mAb J22, which has no lipid modification and as such should not associate with the membrane, was used. Under these conditions, incubation with J22 resulted in only a minor increase in fluorescence (bottom). Therefore, BSA increases the efficiency of painting of CEA, reaching levels that are sufficient to alter cellular properties when achieved by transfection.

156

Figure 3. BSA specifically increases CEA painting. The addition of BSA results in a significant and specific increase in incorporation of CEA (middle). To confirm the specificity of this effect, the mAb J22 was incubated with cells in the presence of BSA, resulting in a minimal increase in the fluorescence mean (bottom). Thus, BSA can be used to specifically increase the efficiency of painting.

157 Discussion We demonstrate in this study that the GPI anchor of CEA contains the information required for incorporation into the plasma membrane. This offers the possibility of altering the function of cells through the addition of this anchor attached to a non-functional protein domain.

Modulating membrane lipid content Applying lipids to cells, from the external medium, modulates biological function both in vitro and in vivo (see Chapter 5). While current work on this type of treatment has concentrated on ameliorating neurological disorders, it should be possible in the future to use lipids on various other diseases. This will require the identification of both a target for the treatment, and a lipid molecule that can be used as the therapy. One ideal target for this type of treatment would appear to be proteins attached to GPI anchors, lipid modifications that have been demonstrated by our lab to contain specific information that determines protein function (Screaton et al., 2000), and can incorporate into the plasma membrane from the external medium (Ilangumaran et al., 1996). These two properties, in addition to the observation that many GPI-anchored proteins are over- expressed in cancer, make targeting this type of protein an interesting possibility for future medical therapies.

Altering CEA function with shankless anchors The GPI anchor of CEA contains specific information directing localization to specific rafts, which can be exploited resulting in a complete loss of function (Nicholson and Stanners, 2006). This offers not only the possibility of modulating protein function in vitro, but may eventually be of use in a clinical setting. We suggest that treatment with the GPI anchor from CEA, with either with greatly reduced or completely lacking external protein domains, could affect the function of CEA, offering a novel cancer therapy, by restoring to cancer cells over-expressing CEA the functions that CEA blocks, including differentiation and anoikis. Painting has been used, recently, to cover the surface of cancer cells with costimulator molecules, resulting in antitumor immunity in mice (Chen et al., 2000; Zheng et al., 2001). Once the structure of the GPI anchor of CEA is known, in particular the specific modifications that direct function, it may be possibly to generate protein-free “sub-anchors” that consist of structures containing these modifications. Cells could then be treated with these molecules, to inhibit the function of CEA and restore

158 normal biological function. Ultimately, painting may one day be clinically relevant, with the possibility of it being employed to target the tumorigenic functions of CEA.

159 Original Contributions to Knowledge

1. The demonstration that the GPI anchor of CEA determines cell surface localization of the protein that it is attached to, as the NCAM-CEA chimera NCB with the CEA anchor, but not NCAM itself with the NCAM GPI anchor, colocalizes with ∆NCEA, a mutant with the CEA anchor. This localization determines the signaling molecules that the protein is associated with, thereby determining function. 2. The novel observation that non-functional GPI-anchored proteins can inhibit the function of a protein with the same anchor. This results from an increase in the size of the membrane rafts that the protein inhabits, which interferes with its ability to cluster by self-binding and, as such, to initiate downstream signaling. 3. A novel signal that determines the addition of the CEA-specific GPI anchor was identified. This consists of 6 amino acids (PGLSAG) located within the hydrophilic spacer of the signal sequence, at positions +5 to +10 downstream of the omega (anchor addition) site. This sequence is necessary and sufficient to specify the addition of the CEA anchor. 4. The GPI anchors of CC6 and CC7 specify similar functional properties (integrin modulation and differentiation block) as that of CEA. However, ∆NCEA is unable to inhibit the function of these proteins, demonstrating that the anchors of CC6 and CC7 are not identical to that of CEA. This suggests that functionally equivalent GPI anchors are not necessarily identical. 5. The confirmation that the GPI anchor of CEA contains sufficient information to incorporate the protein into the plasma membrane, in a process known as “protein painting.” This may offer therapeutic potential one day, by taking advantage of the ability of “shank defective” GPI anchors to inhibit the effects of functional proteins.

160

Bibliography

161 Abeysinghe, H.R., Q. Cao, J. Xu, S. Pollock, Y. Veyberman, N.L. Guckert, P. Keng, and N. Wang. 2003. THY1 expression is associated with tumor suppression of human ovarian cancer. Cancer Genet Cytogenet. 143:125-32. Abou-Rjaily, G.A., S.J. Lee, D. May, Q.Y. Al-Share, A.M. Deangelis, R.J. Ruch, M. Neumaier, H. Kalthoff, S.H. Lin, and S.M. Najjar. 2004. CEACAM1 modulates epidermal growth factor receptor--mediated cell proliferation. J Clin Invest. 114:944-52. Abraham, S.N., M.J. Duncan, G. Li, and D. Zaas. 2005. Bacterial penetration of the mucosal barrier by targeting lipid rafts. J Investig Med. 53:318-21. Aceto, J., T. Kieber-Emmons, and D.B. Cines. 1999. Carboxy-terminal processing of the urokinase receptor: implications for substrate recognition and glycosylphosphatidylinositol anchor addition. Biochemistry. 38:992-1001. Afar, D.E., C.P. Stanners, and J.C. Bell. 1992. Tyrosine phosphorylation of biliary glycoprotein, a cell adhesion molecule related to carcinoembryonic antigen. Biochim Biophys Acta. 1134:46-52. Agrez, M.V., and R.C. Bates. 1994. Colorectal cancer and the integrin family of cell adhesion receptors: current status and future directions. Eur J Cancer. 30A:2166-70. Alexander, J.C., N.A. Silverman, and P.B. Chretien. 1976. Effect of age and cigarette smoking on carcinoembryonic antigen levels. Jama. 235:1975-9. Allard, W.J., I.E. Neaman, J.J. Elting, T.R. Barnett, H. Yoshimura, H.A. Fritsche, and K.K. Yeung. 1994. Nonspecific cross-reacting antigen 50/90 is elevated in patients with breast, lung, and colon cancer. Cancer Res. 54:1227-34. Anderson, R.G. 1998. The caveolae membrane system. Annu Rev Biochem. 67:199-225. Anderson, R.G., and K. Jacobson. 2002. A role for lipid shells in targeting proteins to caveolae, rafts, and other lipid domains. Science. 296:1821-5. Andreasen, P.A., L. Kjoller, L. Christensen, and M.J. Duffy. 1997. The urokinase-type plasminogen activator system in cancer metastasis: a review. Int J Cancer. 72:1-22. Aouad, S.M., L.Y. Cohen, E. Sharif-Askari, E.K. Haddad, A. Alam, and R.P. Sekaly. 2004. Caspase-3 is a component of Fas death-inducing signaling complex in lipid rafts and its activity is required for complete caspase-8 activation during Fas-mediated cell death. J Immunol. 172:2316-23. Armesto, J., E. Hannappel, K. Leopold, W. Fischer, R. Bublitz, L. Langer, G.A. Cumme, and A. Horn. 1996. Microheterogeneity of the hydrophobic and hydrophilic part of the glycosylphosphatidylinositol anchor of alkaline phosphatase from calf intestine. Eur J Biochem. 238:259-69. Arni, S., S.A. Keilbaugh, A.G. Ostermeyer, and D.A. Brown. 1998. Association of GAP-43 with detergent-resistant membranes requires two palmitoylated cysteine residues. J Biol Chem. 273:28478-85. Arvanitis, D.N., W. Min, Y. Gong, Y.M. Heng, and J.M. Boggs. 2005. Two types of detergent- insoluble, glycosphingolipid/cholesterol-rich membrane domains from isolated myelin. J Neurochem. 94:1696-710. Bajenova, O., E. Stolper, S. Gapon, N. Sundina, R. Zimmer, and P. Thomas. 2003. Surface expression of heterogeneous nuclear RNA binding protein M4 on Kupffer cell relates to its function as a carcinoembryonic antigen receptor. Exp Cell Res. 291:228-41. Bajenova, O.V., R. Zimmer, E. Stolper, J. Salisbury-Rowswell, A. Nanji, and P. Thomas. 2001. Heterogeneous RNA-binding protein M4 is a receptor for carcinoembryonic antigen in Kupffer cells. J Biol Chem. 276:31067-73.

162 Baldwin, M.A. 2005. Analysis of glycosylphosphatidylinositol protein anchors: the prion protein. Methods Enzymol. 405:172-87. Ballesta, A.M., R. Molina, X. Filella, J. Jo, and N. Gimenez. 1995. Carcinoembryonic antigen in staging and follow-up of patients with solid tumors. Tumour Biol. 16:32-41. Bamberger, A.M., L. Riethdorf, P. Nollau, M. Naumann, I. Erdmann, J. Gotze, J. Brummer, H.M. Schulte, C. Wagener, and T. Loning. 1998. Dysregulated expression of CD66a (BGP, C-CAM), an adhesion molecule of the CEA family, in endometrial cancer. Am J Pathol. 152:1401-6. Bangs, J.D., D. Hereld, J.L. Krakow, G.W. Hart, and P.T. Englund. 1985. Rapid processing of the carboxyl terminus of a trypanosome variant surface glycoprotein. Proc Natl Acad Sci U S A. 82:3207-11. Barnett, T.R., L. Drake, and W. Pickle, 2nd. 1993. Human biliary glycoprotein gene: characterization of a family of novel alternatively spliced RNAs and their expressed proteins. Mol Cell Biol. 13:1273-82. Baron, W., L. Decker, H. Colognato, and C. ffrench-Constant. 2003. Regulation of integrin growth factor interactions in oligodendrocytes by lipid raft microdomains. Curr Biol. 13:151-5. Barroso, A., and E. Alpert. 1983. Diagnostic procedures in the evaluation of hepatic diseases. Functional evaluation: carcinoembryonic antigen (CEA) and alphafetoprotein (AFP) as tumor markers of the liver. Lab Res Methods Biol Med. 7:153-7. Barton, C.H., G. Dickson, H.J. Gower, L.H. Rowett, W. Putt, V. Elsom, S.E. Moore, C. Goridis, and F.S. Walsh. 1988. Complete sequence and in vitro expression of a tissue-specific phosphatidylinositol-linked N-CAM isoform from skeletal muscle. Development. 104:165-73. Baum, O., S. Troll, and D.C. Hixson. 1996. The long and the short isoform of cell-CAM 105 show variant-specific modifications in adult rat organs. Biochem Biophys Res Commun. 227:775-81. Bavari, S., C.M. Bosio, E. Wiegand, G. Ruthel, A.B. Will, T.W. Geisbert, M. Hevey, C. Schmaljohn, A. Schmaljohn, and M.J. Aman. 2002. Lipid raft microdomains: a gateway for compartmentalized trafficking of Ebola and Marburg viruses. J Exp Med. 195:593- 602. Beauchemin, N., S. Benchimol, D. Cournoyer, A. Fuks, and C.P. Stanners. 1987. Isolation and characterization of full-length functional cDNA clones for human carcinoembryonic antigen. Mol Cell Biol. 7:3221-30. Beauchemin, N., P. Draber, G. Dveksler, P. Gold, S. Gray-Owen, F. Grunert, S. Hammarstrom, K.V. Holmes, A. Karlsson, M. Kuroki, S.H. Lin, L. Lucka, S.M. Najjar, M. Neumaier, B. Obrink, J.E. Shively, K.M. Skubitz, C.P. Stanners, P. Thomas, J.A. Thompson, M. Virji, S. von Kleist, C. Wagener, S. Watt, and W. Zimmermann. 1999. Redefined nomenclature for members of the carcinoembryonic antigen family. Exp Cell Res. 252:243-9. Beauchemin, N., T. Kunath, J. Robitaille, B. Chow, C. Turbide, E. Daniels, and A. Veillette. 1997. Association of biliary glycoprotein with protein tyrosine phosphatase SHP-1 in malignant colon epithelial cells. Oncogene. 14:783-90. Beauchemin, N., C. Turbide, D. Afar, J. Bell, M. Raymond, C.P. Stanners, and A. Fuks. 1989. A mouse analogue of the human carcinoembryonic antigen. Cancer Res. 49:2017-21. Beghdadi-Rais, C., M. Schreyer, M. Rousseaux, P. Borel, R.J. Eisenberg, G.H. Cohen, C. Bron, and N. Fasel. 1993. Carboxyl terminus structural requirements for glycosyl-

163 phosphatidylinositol anchor addition to cell surface proteins. J Cell Sci. 105 ( Pt 3):831- 40. Behr, T.M., S. Gratz, P.M. Markus, R.M. Dunn, M. Hufner, A. Schauer, M. Fischer, D.L. Munz, H. Becker, and W. Becker. 1997. Anti-carcinoembryonic antigen antibodies versus somatostatin analogs in the detection of metastatic medullary thyroid carcinoma: are carcinoembryonic antigen and somatostatin receptor expression prognostic factors? Cancer. 80:2436-57. Benchimol, S., A. Fuks, S. Jothy, N. Beauchemin, K. Shirota, and C.P. Stanners. 1989. Carcinoembryonic antigen, a human tumor marker, functions as an intercellular adhesion molecule. Cell. 57:327-34. Benghezal, M., A. Benachour, S. Rusconi, M. Aebi, and A. Conzelmann. 1996. Yeast Gpi8p is essential for GPI anchor attachment onto proteins. Embo J. 15:6575-83. Benting, J., A. Rietveld, I. Ansorge, and K. Simons. 1999. Acyl and alkyl chain length of GPI- anchors is critical for raft association in vitro. FEBS Lett. 462:47-50. Berditchevski, F., and E. Odintsova. 1999. Characterization of integrin-tetraspanin adhesion complexes: role of tetraspanins in integrin signaling. J Cell Biol. 146:477-92. Berger, C.N., O. Billker, T.F. Meyer, A.L. Servin, and I. Kansau. 2004. Differential recognition of members of the carcinoembryonic antigen family by Afa/Dr adhesins of diffusely adhering Escherichia coli (Afa/Dr DAEC). Mol Microbiol. 52:963-83. Berger, J., A.D. Howard, L. Brink, L. Gerber, J. Hauber, B.R. Cullen, and S. Udenfriend. 1988. COOH-terminal requirements for the correct processing of a phosphatidylinositol-glycan anchored membrane protein. J Biol Chem. 263:10016-21. Berinstein, N.L. 2002. Carcinoembryonic antigen as a target for therapeutic anticancer vaccines: a review. J Clin Oncol. 20:2197-207. Bhatnagar, J., W. Heroman, M. Murphy, and G.E. Austin. 2002. Immunohistochemical detection of carcinoembryonic antigen in esophageal carcinomas: a comparison with other gastrointestinal neoplasms. Anticancer Res. 22:1849-57. Bhattacharjya, S., R. Aggarwal, and B.R. Davidson. 2006. Intensive follow-up after liver resection for colorectal liver metastases: results of combined serial tumour marker estimations and computed tomography of the chest and abdomen - a prospective study. Br J Cancer. 95:21-6. Bini, L., S. Pacini, S. Liberatori, S. Valensin, M. Pellegrini, R. Raggiaschi, V. Pallini, and C.T. Baldari. 2003. Extensive temporally regulated reorganization of the lipid raft proteome following T-cell antigen receptor triggering. Biochem J. 369:301-9. Blumenthal, R.D., E. Leon, H.J. Hansen, and D.M. Goldenberg. 2007. Expression patterns of CEACAM5 and CEACAM6 in primary and metastatic cancers. BMC Cancer. 7:2. Boccuni, P., L. Del Vecchio, R. Di Noto, and B. Rotoli. 2000. Glycosyl phosphatidylinositol (GPI)-anchored molecules and the pathogenesis of paroxysmal nocturnal hemoglobinuria. Crit Rev Oncol Hematol. 33:25-43. Boccuni, P., R. Di Noto, C. Lo Pardo, M.R. Villa, F. Ferrara, B. Rotoli, and L. Del Vecchio. 1998. CD66c antigen expression is myeloid restricted in normal bone marrow but is a common feature of CD10+ early-B-cell malignancies. Tissue Antigens. 52:1-8. Bodin, S., C. Soulet, H. Tronchere, P. Sie, C. Gachet, M. Plantavid, and B. Payrastre. 2005. Integrin-dependent interaction of lipid rafts with the actin cytoskeleton in activated human platelets. J Cell Sci. 118:759-69.

164 Bojunga, J., C. Dragan, P.M. Schumm-Draeger, K.H. Usadel, and K. Kusterer. 2001. Circulating calcitonin and carcinoembryonic antigen m-RNA detected by RT-PCR as tumour markers in medullary thyroid carcinoma. Br J Cancer. 85:1546-50. Booth, J.W., D. Telio, E.H. Liao, S.E. McCaw, T. Matsuo, S. Grinstein, and S.D. Gray-Owen. 2003. Phosphatidylinositol 3-kinases in carcinoembryonic antigen-related cellular adhesion molecule-mediated internalization of Neisseria gonorrhoeae. J Biol Chem. 278:14037-45. Borochov, H., and M. Shinitzky. 1976. Vertical displacement of membrane proteins mediated by changes in microviscosity. Proc Natl Acad Sci U S A. 73:4526-30. Botto, L., M. Masserini, A. Cassetti, and P. Palestini. 2004. Immunoseparation of Prion protein- enriched domains from other detergent-resistant membrane fractions, isolated from neuronal cells. FEBS Lett. 557:143-7. Boulton, I.C., and S.D. Gray-Owen. 2002. Neisserial binding to CEACAM1 arrests the activation and proliferation of CD4+ T lymphocytes. Nat Immunol. 3:229-36. Braccia, A., M. Villani, L. Immerdal, L.L. Niels-Christiansen, B.T. Nystrom, G.H. Hansen, and E.M. Danielsen. 2003. Microvillar membrane microdomains exist at physiological temperature. Role of galectin-4 as lipid raft stabilizer revealed by "superrafts". J Biol Chem. 278:15679-84. Bretscher, M.S. 1973. Membrane structure: some general principles. Science. 181:622-9. Brewis, I.A., M.A. Ferguson, A. Mehlert, A.J. Turner, and N.M. Hooper. 1995. Structures of the glycosyl-phosphatidylinositol anchors of porcine and human renal membrane dipeptidase. Comprehensive structural studies on the porcine anchor and interspecies comparison of the glycan core structures. J Biol Chem. 270:22946-56. Brown, D.A., and J.K. Rose. 1992. Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface. Cell. 68:533-44. Brugger, B., B. Glass, P. Haberkant, I. Leibrecht, F.T. Wieland, and H.G. Krausslich. 2006. The HIV lipidome: a raft with an unusual composition. Proc Natl Acad Sci U S A. 103:2641- 6. Brugger, B., C. Graham, I. Leibrecht, E. Mombelli, A. Jen, F. Wieland, and R. Morris. 2004. The membrane domains occupied by glycosylphosphatidylinositol-anchored prion protein and Thy-1 differ in lipid composition. J Biol Chem. 279:7530-6. Brummer, J., A. Ebrahimnejad, R. Flayeh, U. Schumacher, T. Loning, A.M. Bamberger, and C. Wagener. 2001. cis Interaction of the cell adhesion molecule CEACAM1 with integrin beta(3). Am J Pathol. 159:537-46. Brummer, J., M. Neumaier, C. Gopfert, and C. Wagener. 1995. Association of pp60c-src with biliary glycoprotein (CD66a), an adhesion molecule of the carcinoembryonic antigen family downregulated in colorectal carcinomas. Oncogene. 11:1649-55. Bucht, G., P. Wikstrom, and K. Hjalmarsson. 1999. Optimising the signal peptide for glycosyl phosphatidylinositol modification of human acetylcholinesterase using mutational analysis and peptide-quantitative structure-activity relationships. Biochim Biophys Acta. 1431:471-82. Burgess, A.W. 1998. Growth control mechanisms in normal and transformed intestinal cells. Philos Trans R Soc Lond B Biol Sci. 353:903-9. Camacho-Leal, P., A.B. Zhai, and C.P. Stanners. 2007. A co-clustering model involving alpha5beta1 integrin for the biological effects of GPI-anchored human carcinoembryonic antigen (CEA). J Cell Physiol. 211:791-802.

165 Campbell, N.A., M.S. Park, L.S. Toy, X.Y. Yio, L. Devine, P. Kavathas, and L. Mayer. 2002. A non-class I MHC intestinal epithelial surface glycoprotein, gp180, binds to CD8. Clin Immunol. 102:267-74. Caras, I.W. 1991. An internally positioned signal can direct attachment of a glycophospholipid membrane anchor. J Cell Biol. 113:77-85. Caras, I.W., G.N. Weddell, M.A. Davitz, V. Nussenzweig, and D.W. Martin, Jr. 1987. Signal for attachment of a phospholipid membrane anchor in decay accelerating factor. Science. 238:1280-3. Carrasco, M., L. Munoz, M. Bellido, S. Bernat, E. Rubiol, J. Ubeda, J. Sierra, and J.F. Nomdedeu. 2000. CD66 expression in acute leukaemia. Ann Hematol. 79:299-303. Chalajour, F., H. Treede, A. Ebrahimnejad, H. Lauke, H. Reichenspurner, and S. Ergun. 2004. Angiogenic activation of valvular endothelial cells in aortic valve stenosis. Exp Cell Res. 298:455-64. Chan, C.H., D. Cook, and C.P. Stanners. 2006. Increased colon tumor susceptibility in azoxymethane treated CEABAC transgenic mice. Carcinogenesis. 27:1909-16. Chan, C.H., and C.P. Stanners. 2004. Novel mouse model for carcinoembryonic antigen-based therapy. Mol Ther. 9:775-85. Charbonneau, J., and C.P. Stanners. 1999. Role of carbohydrate structures in CEA-mediated intercellular adhesion. Cell Adhes Commun. 7:233-44. Chatal, J.F., L. Campion, F. Kraeber-Bodere, S. Bardet, J.P. Vuillez, B. Charbonnel, V. Rohmer, C.H. Chang, R.M. Sharkey, D.M. Goldenberg, and J. Barbet. 2006. Survival improvement in patients with medullary thyroid carcinoma who undergo pretargeted anti- carcinoembryonic-antigen radioimmunotherapy: a collaborative study with the French Endocrine Tumor Group. J Clin Oncol. 24:1705-11. Chen, A., G. Zheng, and M.L. Tykocinski. 2000. Hierarchical costimulator thresholds for distinct immune responses: application of a novel two-step Fc fusion protein transfer method. J Immunol. 164:705-11. Chen, D., H. Iijima, T. Nagaishi, A. Nakajima, S. Russell, R. Raychowdhury, V. Morales, C.E. Rudd, N. Utku, and R.S. Blumberg. 2004. Carcinoembryonic antigen-related cellular adhesion molecule 1 isoforms alternatively inhibit and costimulate human T cell function. J Immunol. 172:3535-43. Chen, R., V. Anderson, Y. Hiroi, and M.E. Medof. 2003. Proprotein interaction with the GPI transamidase. J Cell Biochem. 88:1025-37. Chen, R., J.J. Knez, W.C. Merrick, and M.E. Medof. 2001. Comparative efficiencies of C- terminal signals of native glycophosphatidylinositol (GPI)-anchored proproteins in conferring GPI-anchoring. J Cell Biochem. 84:68-83. Chen, R., S. Udenfriend, G.M. Prince, S.E. Maxwell, S. Ramalingam, L.D. Gerber, J. Knez, and M.E. Medof. 1996. A defect in glycosylphosphatidylinositol (GPI) transamidase activity in mutant K cells is responsible for their inability to display GPI surface proteins. Proc Natl Acad Sci U S A. 93:2280-4. Chen, R., E.I. Walter, G. Parker, J.P. Lapurga, J.L. Millan, Y. Ikehara, S. Udenfriend, and M.E. Medof. 1998. Mammalian glycophosphatidylinositol anchor transfer to proteins and posttransfer deacylation. Proc Natl Acad Sci U S A. 95:9512-7. Chevinsky, A.H. 1991. CEA in tumors of other than colorectal origin. Semin Surg Oncol. 7:162- 6.

166 Clark, H.F., A.L. Gurney, E. Abaya, K. Baker, D. Baldwin, J. Brush, J. Chen, B. Chow, C. Chui, C. Crowley, B. Currell, B. Deuel, P. Dowd, D. Eaton, J. Foster, C. Grimaldi, Q. Gu, P.E. Hass, S. Heldens, A. Huang, H.S. Kim, L. Klimowski, Y. Jin, S. Johnson, J. Lee, L. Lewis, D. Liao, M. Mark, E. Robbie, C. Sanchez, J. Schoenfeld, S. Seshagiri, L. Simmons, J. Singh, V. Smith, J. Stinson, A. Vagts, R. Vandlen, C. Watanabe, D. Wieand, K. Woods, M.H. Xie, D. Yansura, S. Yi, G. Yu, J. Yuan, M. Zhang, Z. Zhang, A. Goddard, W.I. Wood, P. Godowski, and A. Gray. 2003. The secreted protein discovery initiative (SPDI), a large-scale effort to identify novel human secreted and transmembrane proteins: a bioinformatics assessment. Genome Res. 13:2265-70. Comin, C.E., L. Novelli, V. Boddi, M. Paglierani, and S. Dini. 2001. Calretinin, thrombomodulin, CEA, and CD15: a useful combination of immunohistochemical markers for differentiating pleural epithelial mesothelioma from peripheral pulmonary adenocarcinoma. Hum Pathol. 32:529-36. Cordes, F.S., J.N. Bright, and M.S. Sansom. 2002. Proline-induced distortions of transmembrane helices. J Mol Biol. 323:951-60. Coyne, K.E., A. Crisci, and D.M. Lublin. 1993. Construction of synthetic signals for glycosylphosphatidylinositol anchor attachment. Analysis of amino acid sequence requirements for anchoring. J Biol Chem. 268:6689-93. Crossthwaite, A.J., T. Seebacher, N. Masada, A. Ciruela, K. Dufraux, J.E. Schultz, and D.M. Cooper. 2005. The cytosolic domains of Ca2+-sensitive adenylyl cyclases dictate their targeting to plasma membrane lipid rafts. J Biol Chem. 280:6380-91. Dai, T., G.A. Abou-Rjaily, Q.Y. Al-Share, Y. Yang, M.A. Fernstrom, A.M. Deangelis, A.D. Lee, L. Sweetman, A. Amato, M. Pasquali, G.D. Lopaschuk, S.K. Erickson, and S.M. Najjar. 2004. Interaction between altered insulin and lipid metabolism in CEACAM1-inactive transgenic mice. J Biol Chem. 279:45155-61. Dalley, J.A., and N.J. Bulleid. 2003. The endoplasmic reticulum (ER) translocon can differentiate between hydrophobic sequences allowing signals for glycosylphosphatidylinositol anchor addition to be fully translocated into the ER lumen. J Biol Chem. 278:51749-57. Dalskov, S.M., L. Immerdal, L.L. Niels-Christiansen, G.H. Hansen, A. Schousboe, and E.M. Danielsen. 2005. Lipid raft localization of GABA A receptor and Na+, K+-ATPase in discrete microdomain clusters in rat cerebellar granule cells. Neurochem Int. 46:489-99. de Almeida, R.F., L.M. Loura, A. Fedorov, and M. Prieto. 2005. Lipid rafts have different sizes depending on membrane composition: a time-resolved fluorescence resonance energy transfer study. J Mol Biol. 346:1109-20. De Giovanni, C., P.L. Lollini, R. Dolcetti, L. Landuzzi, G. Nicoletti, E. D'Andrea, K. Scotland, and P. Nanni. 1993. Uncoupling of growth inhibition and differentiation in dexamethasone-treated human rhabdomyosarcoma cells. Br J Cancer. 67:674-9. de Vries, F.P., R. Cole, J. Dankert, M. Frosch, and J.P. van Putten. 1998. Neisseria meningitidis producing the Opc adhesin binds epithelial cell proteoglycan receptors. Mol Microbiol. 27:1203-12. Deeg, M.A., D.R. Humphrey, S.H. Yang, T.R. Ferguson, V.N. Reinhold, and T.L. Rosenberry. 1992. Glycan components in the glycoinositol phospholipid anchor of human erythrocyte acetylcholinesterase. Novel fragments produced by trifluoroacetic acid. J Biol Chem. 267:18573-80.

167 Delmas, O., M. Breton, C. Sapin, A. Le Bivic, O. Colard, and G. Trugnan. 2007. Heterogeneity of Raft-type membrane microdomains associated with VP4, the rotavirus spike protein, in Caco-2 and MA 104 cells. J Virol. 81:1610-8. Dickson, G., D. Peck, S.E. Moore, C.H. Barton, and F.S. Walsh. 1990. Enhanced myogenesis in NCAM-transfected mouse myoblasts. Nature. 344:348-51. Dietrich, C., L.A. Bagatolli, Z.N. Volovyk, N.L. Thompson, M. Levi, K. Jacobson, and E. Gratton. 2001a. Lipid rafts reconstituted in model membranes. Biophys J. 80:1417-28. Dietrich, C., Z.N. Volovyk, M. Levi, N.L. Thompson, and K. Jacobson. 2001b. Partitioning of Thy-1, GM1, and cross-linked phospholipid analogs into lipid rafts reconstituted in supported model membrane monolayers. Proc Natl Acad Sci U S A. 98:10642-7. Dietrich, C., B. Yang, T. Fujiwara, A. Kusumi, and K. Jacobson. 2002. Relationship of lipid rafts to transient confinement zones detected by single particle tracking. Biophys J. 82:274-84. Diez, M., M. Pollan, J.M. Muguerza, M.J. Gaspar, A.M. Duce, M.J. Alvarez, T. Ratia, P. Hernandez, A. Ruiz, and J. Granell. 2000. Time-dependency of the prognostic effect of carcinoembryonic antigen and p53 protein in colorectal adenocarcinoma. Cancer. 88:35- 41. Douglass, A.D., and R.D. Vale. 2005. Single-molecule microscopy reveals plasma membrane microdomains created by protein-protein networks that exclude or trap signaling molecules in T cells. Cell. 121:937-50. Duxbury, M.S., H. Ito, S.W. Ashley, and E.E. Whang. 2004a. c-Src-dependent cross-talk between CEACAM6 and alpha(v)beta(3) integrin enhances pancreatic adenocarcinoma cell adhesion to extracellular matrix components. Biochem Biophys Res Commun. 317:133-41. Duxbury, M.S., H. Ito, S.W. Ashley, and E.E. Whang. 2004b. c-Src-dependent cross-talk between CEACAM6 and alphavbeta3 integrin enhances pancreatic adenocarcinoma cell adhesion to extracellular matrix components. Biochem Biophys Res Commun. 317:133- 41. Duxbury, M.S., H. Ito, S.W. Ashley, and E.E. Whang. 2004c. CEACAM6 cross-linking induces caveolin-1-dependent, Src-mediated focal adhesion kinase phosphorylation in BxPC3 pancreatic adenocarcinoma cells. J Biol Chem. 279:23176-82. Duxbury, M.S., H. Ito, M.J. Zinner, S.W. Ashley, and E.E. Whang. 2004d. CEACAM6 gene silencing impairs anoikis resistance and in vivo metastatic ability of pancreatic adenocarcinoma cells. Oncogene. 23:465-73. Dveksler, G.S., C.W. Dieffenbach, C.B. Cardellichio, K. McCuaig, M.N. Pensiero, G.S. Jiang, N. Beauchemin, and K.V. Holmes. 1993. Several members of the mouse carcinoembryonic antigen-related glycoprotein family are functional receptors for the coronavirus mouse hepatitis virus-A59. J Virol. 67:1-8. Dveksler, G.S., M.N. Pensiero, C.B. Cardellichio, R.K. Williams, G.S. Jiang, K.V. Holmes, and C.W. Dieffenbach. 1991. Cloning of the mouse hepatitis virus (MHV) receptor: expression in human and hamster cell lines confers susceptibility to MHV. J Virol. 65:6881-91. Eades-Perner, A.M., H. van der Putten, A. Hirth, J. Thompson, M. Neumaier, S. von Kleist, and W. Zimmermann. 1994. Mice transgenic for the human carcinoembryonic antigen gene maintain its spatiotemporal expression pattern. Cancer Res. 54:4169-76.

168 Ebrahimnejad, A., R. Flayeh, G. Unteregger, C. Wagener, and J. Brummer. 2000. Cell adhesion molecule CEACAM1 associates with paxillin in granulocytes and epithelial and endothelial cells. Exp Cell Res. 260:365-73. Edelblum, K.L., F. Yan, T. Yamaoka, and D.B. Polk. 2006. Regulation of apoptosis during homeostasis and disease in the intestinal epithelium. Inflamm Bowel Dis. 12:413-24. Edidin, M. 2003. The state of lipid rafts: from model membranes to cells. Annu Rev Biophys Biomol Struct. 32:257-83. Edlund, M., K. Wikstrom, R. Toomik, P. Ek, and B. Obrink. 1998. Characterization of protein kinase C-mediated phosphorylation of the short cytoplasmic domain isoform of C-CAM. FEBS Lett. 425:166-70. Eidelman, F.J., A. Fuks, L. DeMarte, M. Taheri, and C.P. Stanners. 1993. Human carcinoembryonic antigen, an intercellular adhesion molecule, blocks fusion and differentiation of rat myoblasts. J Cell Biol. 123:467-75. Eisenberg, S., D.E. Shvartsman, M. Ehrlich, and Y.I. Henis. 2006. Clustering of raft-associated proteins in the external membrane leaflet modulates internal leaflet H-ras diffusion and signaling. Mol Cell Biol. 26:7190-200. Eisenhaber, B., P. Bork, and F. Eisenhaber. 1998. Sequence properties of GPI-anchored proteins near the omega-site: constraints for the polypeptide binding site of the putative transamidase. Protein Eng. 11:1155-61. Eisenhaber, B., S. Maurer-Stroh, M. Novatchkova, G. Schneider, and F. Eisenhaber. 2003. Enzymes and auxiliary factors for GPI lipid anchor biosynthesis and post-translational transfer to proteins. Bioessays. 25:367-85. El Far, M., M. El Naggar, O.A. Elkhawaga, R. Yahya, A. Allam, and A. Khalifa. 2006. Carcinoembryonic antigen, alpha-fetoprotein, and prostate-specific antigen in the sera of industrial workers exposed to phenol, formaldehyde, urea, and mixed vapors. Inhal Toxicol. 18:1041-6. Engle, M.J., A. Mahmood, and D.H. Alpers. 1995. Two rat intestinal alkaline phosphatase isoforms with different carboxyl-terminal peptides are both membrane-bound by a glycan phosphatidylinositol linkage. J Biol Chem. 270:11935-40. Englund, P.T. 1993. The structure and biosynthesis of glycosyl phosphatidylinositol protein anchors. Annu Rev Biochem. 62:121-38. Ergun, S., N. Kilik, G. Ziegeler, A. Hansen, P. Nollau, J. Gotze, J.H. Wurmbach, A. Horst, J. Weil, M. Fernando, and C. Wagener. 2000. CEA-related cell adhesion molecule 1: a potent angiogenic factor and a major effector of vascular endothelial growth factor. Mol Cell. 5:311-20. Fankhauser, C., S.W. Homans, J.E. Thomas-Oates, M.J. McConville, C. Desponds, A. Conzelmann, and M.A. Ferguson. 1993. Structures of glycosylphosphatidylinositol membrane anchors from Saccharomyces cerevisiae. J Biol Chem. 268:26365-74. Fantini, J. 2003. How sphingolipids bind and shape proteins: molecular basis of lipid-protein interactions in lipid shells, rafts and related biomembrane domains. Cell Mol Life Sci. 60:1027-32. Feigenson, G.W., and J.T. Buboltz. 2001. Ternary phase diagram of dipalmitoyl-PC/dilauroyl- PC/cholesterol: nanoscopic domain formation driven by cholesterol. Biophys J. 80:2775- 88.

169 Ferguson, M.A., K. Haldar, and G.A. Cross. 1985a. Trypanosoma brucei variant surface glycoprotein has a sn-1,2-dimyristyl glycerol membrane anchor at its COOH terminus. J Biol Chem. 260:4963-8. Ferguson, M.A., S.W. Homans, R.A. Dwek, and T.W. Rademacher. 1988. Glycosyl- phosphatidylinositol moiety that anchors Trypanosoma brucei variant surface glycoprotein to the membrane. Science. 239:753-9. Ferguson, M.A., M.G. Low, and G.A. Cross. 1985b. Glycosyl-sn-1,2- dimyristylphosphatidylinositol is covalently linked to Trypanosoma brucei variant surface glycoprotein. J Biol Chem. 260:14547-55. Ferguson, M.A., and A.F. Williams. 1988. Cell-surface anchoring of proteins via glycosylphosphatidylinositol structures. Annu Rev Biochem. 57:285-320. Fletcher, R.H. 1986. Carcinoembryonic antigen. Ann Intern Med. 104:66-73. Fong, T.G., N.H. Neff, and M. Hadjiconstantinou. 1995. Systemic administration of GM1 ganglioside increases choline acetyltransferase activity in the brain of aged rats. Exp Neurol. 132:157-61. Fontaine, T., T. Magnin, A. Melhert, D. Lamont, J.P. Latge, and M.A. Ferguson. 2003. Structures of the glycosylphosphatidylinositol membrane anchors from Aspergillus fumigatus membrane proteins. Glycobiology. 13:169-77. Foster, L.J., C.L. De Hoog, and M. Mann. 2003. Unbiased quantitative proteomics of lipid rafts reveals high specificity for signaling factors. Proc Natl Acad Sci U S A. 100:5813-8. Foti, M., G. Porcheron, M. Fournier, C. Maeder, and J.-L. Carpentier. 2007. The neck of caveolae is a distinct plasma membrane subdomain that concentrates insulin receptors in 3T3-L1 adipocytes. PNAS. 104:1242-1247. Frangsmyr, L., V. Baranov, and S. Hammarstrom. 1999. Four carcinoembryonic antigen subfamily members, CEA, NCA, BGP and CGM2, selectively expressed in the normal human colonic epithelium, are integral components of the fuzzy coat. Tumour Biol. 20:277-92. Frangsmyr, L., V. Baranov, F. Prall, M.M. Yeung, C. Wagener, and S. Hammarstrom. 1995. Cell- and region-specific expression of biliary glycoprotein and its messenger RNA in normal human colonic mucosa. Cancer Res. 55:2963-7. Friedrichson, T., and T.V. Kurzchalia. 1998. Microdomains of GPI-anchored proteins in living cells revealed by crosslinking. Nature. 394:802-5. Frisch, S.M., and E. Ruoslahti. 1997. Integrins and anoikis. Curr Opin Cell Biol. 9:701-6. Frisch, S.M., and R.A. Screaton. 2001. Anoikis mechanisms. Curr Opin Cell Biol. 13:555-62. Frye, L.D., and M. Edidin. 1970. The rapid intermixing of cell surface antigens after formation of mouse-human heterokaryons. J Cell Sci. 7:319-35. Fujishima, T., Y. Honda, N. Shijubo, H. Takahashi, and S. Abe. 1995. Increased carcinoembryonic antigen concentrations in sera and bronchoalveolar lavage fluids of patients with pulmonary alveolar proteinosis. Respiration. 62:317-21. Fujita, A., J. Cheng, M. Hirakawa, K. Furukawa, S. Kusunoki, and T. Fujimoto. 2007. Gangliosides GM1 and GM3 in the Living Cell Membrane Form Clusters Susceptible to Cholesterol Depletion and Chilling. Mol. Biol. Cell:E07-01-0071. Furukawa, Y., H. Tamura, and H. Ikezawa. 1994. Mutational analysis of the COOH-terminal hydrophobic domain of bovine liver 5'-nucleotidase as a signal for glycosylphosphatidylinositol (GPI) anchor attachment. Biochim Biophys Acta. 1190:273- 8.

170 Furukawa, Y., K. Tsukamoto, and H. Ikezawa. 1997. Mutational analysis of the C-terminal signal peptide of bovine liver 5'-nucleotidase for GPI anchoring: a study on the significance of the hydrophilic spacer region. Biochim Biophys Acta. 1328:185-96. Gadler, H., K. Bremme, and B. Wahren. 1980. Distribution of tumor-associated antigen CEA and cross-reacting NCA in fetal organs. Int J Cancer. 25:91-4. Gangopadhyay, A., O. Bajenova, T.M. Kelly, and P. Thomas. 1996a. Carcinoembryonic antigen induces cytokine expression in Kuppfer cells: implications for hepatic metastasis from colorectal cancer. Cancer Res. 56:4805-10. Gangopadhyay, A., D.A. Lazure, T.M. Kelly, and P. Thomas. 1996b. Purification and analysis of an 80-kDa carcinoembryonic antigen-binding protein from Kupffer cells. Arch Biochem Biophys. 328:151-7. Garcia, A., X. Cayla, A. Fleischer, J. Guergnon, F. Alvarez-Franco Canas, M.P. Rebollo, F. Roncal, and A. Rebollo. 2003. Rafts: a simple way to control apoptosis by subcellular redistribution. Biochimie. 85:727-31. Gasser, M., C. Gerstlauer, M. Grimm, M. Bueter, T. Lebedeva, J. Lutz, U. Maeder, C. Ribas, C. Ribas, E. Nichiporuk, A. Thalheimer, U. Heemann, A. Thiede, D. Meyer, and A.M. Waaga-Gasser. 2007. Comparative Analysis of Predictive Biomarkers for Therapeutical Strategies in Colorectal Cancer. Ann Surg Oncol. Gaus, K., E. Gratton, E.P. Kable, A.S. Jones, I. Gelissen, L. Kritharides, and W. Jessup. 2003. Visualizing lipid structure and raft domains in living cells with two-photon microscopy. Proc Natl Acad Sci U S A. 100:15554-9. Gaus, K., S. Le Lay, N. Balasubramanian, and M.A. Schwartz. 2006. Integrin-mediated adhesion regulates membrane order. J Cell Biol. 174:725-34. Geisler, F.H. 1993. GM-1 ganglioside and motor recovery following human spinal cord injury. J Emerg Med. 11 Suppl 1:49-55. Geisler, F.H., F.C. Dorsey, and W.P. Coleman. 1991. Recovery of motor function after spinal- cord injury--a randomized, placebo-controlled trial with GM-1 ganglioside. N Engl J Med. 324:1829-38. Genega, E.M., B. Hutchinson, V.E. Reuter, and P.B. Gaudin. 2000. Immunophenotype of high- grade prostatic adenocarcinoma and urothelial carcinoma. Mod Pathol. 13:1186-91. George, S., M.D. Nelson, N. Dollahon, and A. Bamezai. 2006. A novel approach to examining compositional heterogeneity of detergent-resistant lipid rafts. Immunol Cell Biol. 84:192- 202. Gerber, L.D., K. Kodukula, and S. Udenfriend. 1992. Phosphatidylinositol glycan (PI-G) anchored membrane proteins. Amino acid requirements adjacent to the site of cleavage and PI-G attachment in the COOH-terminal signal peptide. J Biol Chem. 267:12168-73. Glebov, O.O., and B.J. Nichols. 2004. Distribution of lipid raft markers in live cells. Biochem Soc Trans. 32:673-5. Gold, P., and S.O. Freedman. 1965. Demonstration of tumor-specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques. J. Exp. Med. 121:439-462. Goldstein, M.J., and E.P. Mitchell. 2005. Carcinoembryonic antigen in the staging and follow-up of patients with colorectal cancer. Cancer Invest. 23:338-51. Gomez-Mouton, C., J.L. Abad, E. Mira, R.A. Lacalle, E. Gallardo, S. Jimenez-Baranda, I. Illa, A. Bernad, S. Manes, and A.C. Martinez. 2001. Segregation of leading-edge and uropod

171 components into specific lipid rafts during T cell polarization. Proc Natl Acad Sci U S A. 98:9642-7. Gomez-Mouton, C., R.A. Lacalle, E. Mira, S. Jimenez-Baranda, D.F. Barber, A.C. Carrera, A.C. Martinez, and S. Manes. 2004. Dynamic redistribution of raft domains as an organizing platform for signaling during cell chemotaxis. J Cell Biol. 164:759-68. Greicius, G., E. Severinson, N. Beauchemin, B. Obrink, and B.B. Singer. 2003. CEACAM1 is a potent regulator of B cell receptor complex-induced activation. J Leukoc Biol. 74:126-34. Haas, R., B.C. Jackson, B. Reinhold, J.D. Foster, and T.L. Rosenberry. 1996. Glycoinositol phospholipid anchor and protein C-terminus of bovine erythrocyte acetylcholinesterase: analysis by mass spectrometry and by protein and DNA sequencing. Biochem J. 314 ( Pt 3):817-25. Hadjiconstantinou, M., and N.H. Neff. 1998. GM1 ganglioside: in vivo and in vitro trophic actions on central neurotransmitter systems. J Neurochem. 70:1335-45. Hall, C., S.J. Richards, and P. Hillmen. 2002. The glycosylphosphatidylinositol anchor and paroxysmal nocturnal haemoglobinuria/aplasia model. Acta Haematol. 108:219-30. Hammarstrom, S. 1999. The carcinoembryonic antigen (CEA) family: structures, suggested functions and expression in normal and malignant tissues. Semin Cancer Biol. 9:67-81. Hammarstrom, S., and V. Baranov. 2001. Is there a role for CEA in innate immunity in the colon? Trends Microbiol. 9:119-25. Hammarstrom, S., A. Olsen, S. Teglund, and V. Baranov. 1998. The nature and expression of the human CEA family. In Cell Adhesion and Communication Mediated by the CEA Family: Basic and Clinical Perspectives. Vol. 5. C.P. Stanners, editor. Hardwood Academic Publishers, Amsterdam. 1-30. Hammond, A.T., F.A. Heberle, T. Baumgart, D. Holowka, B. Baird, and G.W. Feigenson. 2005. Crosslinking a lipid raft component triggers liquid ordered-liquid disordered phase separation in model plasma membranes. Proc Natl Acad Sci U S A. 102:6320-5. Hancock, J.F. 2006. Lipid rafts: contentious only from simplistic standpoints. Nat Rev Mol Cell Biol. 7:456-62. Hanenberg, H., M. Baumann, I. Quentin, G. Nagel, H. Grosse-Wilde, S. von Kleist, U. Gobel, S. Burdach, and F. Grunert. 1994. Expression of the CEA gene family members NCA-50/90 and NCA-160 (CD66) in childhood acute lymphoblastic leukemias (ALLs) and in cell lines of B-cell origin. Leukemia. 8:2127-33. Hao, M., S.X. Lin, O.J. Karylowski, D. Wustner, T.E. McGraw, and F.R. Maxfield. 2002. Vesicular and non-vesicular sterol transport in living cells. The endocytic recycling compartment is a major sterol storage organelle. J Biol Chem. 277:609-17. Harder, T., P. Scheiffele, P. Verkade, and K. Simons. 1998. Lipid domain structure of the plasma membrane revealed by patching of membrane components. J Cell Biol. 141:929-42. Harlozinska, A., J. Kula, B. Stepinska, M. Jelen, and R. Richter. 1984. Tumor-associated antigens in female genital tract cancers. Tumori. 70:281-9. Harris, T.J., and C.H. Siu. 2002. Reciprocal raft-receptor interactions and the assembly of adhesion complexes. Bioessays. 24:996-1003. Hasegawa, T., K. Isobe, Y. Tsuchiya, S. Oikawa, H. Nakazato, H. Ikezawa, I. Nakashima, and K. Shimokata. 1991. Establishment and characterisation of human carcinoembryonic antigen transgenic mice. Br J Cancer. 64:710-4. Haynes, P.A., A.A. Gooley, M.A. Ferguson, J.W. Redmond, and K.L. Williams. 1993. Post- translational modifications of the Dictyostelium discoideum glycoprotein PsA.

172 Glycosylphosphatidylinositol membrane anchor and composition of O-linked oligosaccharides. Eur J Biochem. 216:729-37. He, H.-T., A. Lellouch, and D. Marguet. 2005a. Lipid rafts and the initiation of T cell receptor signaling. Seminars in Immunology. 17:23-33. He, H.T., J. Finne, and C. Goridis. 1987. Biosynthesis, membrane association, and release of N- CAM-120, a phosphatidylinositol-linked form of the neural cell adhesion molecule. J Cell Biol. 105:2489-500. He, H.T., A. Lellouch, and D. Marguet. 2005b. Lipid rafts and the initiation of T cell receptor signaling. Semin Immunol. 17:23-33. He, Q., and K.F. Meiri. 2002. Isolation and characterization of detergent-resistant microdomains responsive to NCAM-mediated signaling from growth cones. Mol Cell Neurosci. 19:18- 31. Heerklotz, H. 2002. Triton promotes domain formation in lipid raft mixtures. Biophys J. 83:2693-701. Heerklotz, H., H. Szadkowska, T. Anderson, and J. Seelig. 2003. The sensitivity of lipid domains to small perturbations demonstrated by the effect of Triton. J Mol Biol. 329:793-9. Heijnen, H.F., M. Van Lier, S. Waaijenborg, Y. Ohno-Iwashita, A.A. Waheed, M. Inomata, G. Gorter, W. Mobius, J.W. Akkerman, and J.W. Slot. 2003. Concentration of rafts in platelet filopodia correlates with recruitment of c-Src and CD63 to these domains. J Thromb Haemost. 1:1161-73. Heneberg, P., P. Lebduska, L. Draberova, J. Korb, and P. Draber. 2006. Topography of plasma membrane microdomains and its consequences for mast cell signaling. Eur J Immunol. 36:2795-806. Hill, D.J., M.A. Toleman, D.J. Evans, S. Villullas, L. Van Alphen, and M. Virji. 2001. The variable P5 proteins of typeable and non-typeable Haemophilus influenzae target human CEACAM1. Mol Microbiol. 39:850-62. Hill, D.J., and M. Virji. 2003. A novel cell-binding mechanism of Moraxella catarrhalis ubiquitous surface protein UspA: specific targeting of the N-domain of carcinoembryonic antigen-related cell adhesion molecules by UspA1. Mol Microbiol. 48:117-29. Hinoda, Y., H. Takahashi, T. Higashide, T. Nakano, Y. Arimura, M. Yoshimoto, M. Tsujisaki, K. Imai, and A. Yachi. 1991. Correlated expression of mRNAs of carcinoembryonic antigen and nonspecific cross-reacting antigen genes in malignant and nonmalignant tissues of the colon. Jpn J Clin Oncol. 21:75-81. Hiscox, S., M.B. Hallett, B.P. Morgan, and C.W. van den Berg. 2002. GPI-anchored GFP signals Ca2+ but is homogeneously distributed on the cell surface. Biochem Biophys Res Commun. 293:714-21. Ho, J.C., S.T. Cheung, M. Patil, X. Chen, and S.T. Fan. 2006. Increased expression of glycosylphosphatidylinositol anchor attachment protein 1 (GPAA1) is associated with gene amplification in hepatocellular carcinoma. Int J Cancer. 119:1330-7. Hogg, N., R. Henderson, B. Leitinger, A. McDowall, J. Porter, and P. Stanley. 2002. Mechanisms contributing to the activity of integrins on leukocytes. Immunol Rev. 186:164-71. Homans, S.W., M.A. Ferguson, R.A. Dwek, T.W. Rademacher, R. Anand, and A.F. Williams. 1988. Complete structure of the glycosyl phosphatidylinositol membrane anchor of rat brain Thy-1 glycoprotein. Nature. 333:269-72.

173 Hong, Y., K. Ohishi, J.Y. Kang, S. Tanaka, N. Inoue, J. Nishimura, Y. Maeda, and T. Kinoshita. 2003. Human PIG-U and yeast Cdc91p are the fifth subunit of GPI transamidase that attaches GPI-anchors to proteins. Mol Biol Cell. 14:1780-9. Hood, J.D., and D.A. Cheresh. 2002. Role of integrins in cell invasion and migration. Nat Rev Cancer. 2:91-100. Horst, A.K., W.D. Ito, J. Dabelstein, U. Schumacher, H. Sander, C. Turbide, J. Brummer, T. Meinertz, N. Beauchemin, and C. Wagener. 2006. Carcinoembryonic antigen-related cell adhesion molecule 1 modulates vascular remodeling in vitro and in vivo. J Clin Invest. 116:1596-605. Horst, A.K., and C. Wagener. 2004. CEA-related CAMs. In Handbook of Experimental Pharmacology, Cell Adhesion. Vol. 165. J. Behrens and W.J. Nelson, editors. Springer- Verlag, Berlin Heidelberg. 284-341. Houde, C., S. Roy, N. Leung, D.W. Nicholson, and N. Beauchemin. 2003. The cell adhesion molecule CEACAM1-L is a substrate of caspase-3-mediated cleavage in apoptotic mouse intestinal cells. J Biol Chem. 278:16929-35. Hsieh, J.T., W. Luo, W. Song, Y. Wang, D.I. Kleinerman, N.T. Van, and S.H. Lin. 1995. Tumor suppressive role of an androgen-regulated epithelial cell adhesion molecule (C-CAM) in prostate carcinoma cell revealed by sense and antisense approaches. Cancer Res. 55:190- 7. Hsu, W.H., C.S. Huang, H.S. Hsu, W.J. Huang, H.C. Lee, B.S. Huang, and M.H. Huang. 2007. Preoperative serum carcinoembryonic antigen level is a prognostic factor in women with early non-small-cell lung cancer. Ann Thorac Surg. 83:419-24. Huang, E.H., and H.L. Kaufman. 2002. CEA-based vaccines. Expert Rev Vaccines. 1:49-63. Huber, M., L. Izzi, P. Grondin, C. Houde, T. Kunath, A. Veillette, and N. Beauchemin. 1999. The carboxyl-terminal region of biliary glycoprotein controls its tyrosine phosphorylation and association with protein-tyrosine phosphatases SHP-1 and SHP-2 in epithelial cells. J Biol Chem. 274:335-44. Hynes, R.O. 2002. Integrins: bidirectional, allosteric signaling machines. Cell. 110:673-87. Ideo, H., A. Seko, and K. Yamashita. 2005. Galectin-4 binds to sulfated glycosphingolipids and carcinoembryonic antigen in patches on the cell surface of human colon adenocarcinoma cells. J Biol Chem. 280:4730-7. Ikezawa, H. 2002. Glycosylphosphatidylinositol (GPI)-anchored proteins. Biol Pharm Bull. 25:409-17. Ikezawa, H., M. Yamanegi, R. Taguchi, T. Miyashita, and T. Ohyabu. 1976. Studies on phosphatidylinositol phosphodiesterase (phospholipase C type) of Bacillus cereus. I. purification, properties and phosphatase-releasing activity. Biochim Biophys Acta. 450:154-64. Ikonen, E. 2001. Roles of lipid rafts in membrane transport. Curr Opin Cell Biol. 13:470-7. Ilangumaran, S., P.J. Robinson, and D.C. Hoessli. 1996. Transfer of exogenous glycosylphosphatidylinositol (GPI)-linked molecules to plasma membranes. Trends Cell Biol. 6:163-7. Ilantzis, C., L. DeMarte, R.A. Screaton, and C.P. Stanners. 2002. Deregulated expression of the human tumor marker CEA and CEA family member CEACAM6 disrupts tissue architecture and blocks colonocyte differentiation. Neoplasia. 4:151-63. Ilantzis, C., S. Jothy, L.C. Alpert, P. Draber, and C.P. Stanners. 1997. Cell-surface levels of human carcinoembryonic antigen are inversely correlated with colonocyte differentiation in colon carcinogenesis. Lab Invest. 76:703-16.

174 Incardona, J.P., and S. Eaton. 2000. Cholesterol in signal transduction. Curr Opin Cell Biol. 12:193-203. Iwabuchi, K., S. Yamamura, A. Prinetti, K. Handa, and S. Hakomori. 1998. GM3-enriched microdomain involved in cell adhesion and signal transduction through carbohydrate- carbohydrate interaction in mouse melanoma B16 cells. J Biol Chem. 273:9130-8. Janes, P.W., S.C. Ley, and A.I. Magee. 1999. Aggregation of lipid rafts accompanies signaling via the T cell antigen receptor. J Cell Biol. 147:447-61. Janes, P.W., S.C. Ley, A.I. Magee, and P.S. Kabouridis. 2000. The role of lipid rafts in T cell antigen receptor (TCR) signalling. Semin Immunol. 12:23-34. Jantscheff, P., L. Terracciano, A. Lowy, K. Glatz-Krieger, F. Grunert, B. Micheel, J. Brummer, U. Laffer, U. Metzger, R. Herrmann, and C. Rochlitz. 2003. Expression of CEACAM6 in resectable colorectal cancer: a factor of independent prognostic significance. J Clin Oncol. 21:3638-46. Jessup, J.M., J.C. Kim, P. Thomas, S. Ishii, R. Ford, J.E. Shively, H. Durbin, C.P. Stanners, A. Fuks, H. Zhou, and et al. 1993. Adhesion to carcinoembryonic antigen by human colorectal carcinoma cells involves at least two epitopes. Int J Cancer. 55:262-8. Jothy, S., S.Y. Yuan, and K. Shirota. 1993. Transcription of carcinoembryonic antigen in normal colon and colon carcinoma. In situ hybridization study and implication for a new in vivo functional model. Am J Pathol. 143:250-7. Juweid, M., R.M. Sharkey, T. Behr, L.C. Swayne, A.D. Rubin, T. Herskovic, D. Hanley, A. Markowitz, R. Dunn, J. Siegel, T. Kamal, and D.M. Goldenberg. 1996. Improved detection of medullary thyroid cancer with radiolabeled antibodies to carcinoembryonic antigen. J Clin Oncol. 14:1209-17. Kabouridis, P.S. 2006. Lipid rafts in T cell receptor signalling. Mol Membr Biol. 23:49-57. Kammerer, R., S. Hahn, B.B. Singer, J.S. Luo, and S. von Kleist. 1998. Biliary glycoprotein (CD66a), a cell adhesion molecule of the immunoglobulin superfamily, on human lymphocytes: structure, expression and involvement in T cell activation. Eur J Immunol. 28:3664-74. Kammerer, R., T. Popp, B.B. Singer, J. Schlender, and W. Zimmermann. 2004a. Identification of allelic variants of the bovine immune regulatory molecule CEACAM1 implies a pathogen-driven evolution. Gene. 339:99-109. Kammerer, R., R. Riesenberg, C. Weiler, J. Lohrmann, J. Schleypen, and W. Zimmermann. 2004b. The tumour suppressor gene CEACAM1 is completely but reversibly downregulated in renal cell carcinoma. J Pathol. 204:258-67. Kammerer, R., D. Stober, B.B. Singer, B. Obrink, and J. Reimann. 2001. Carcinoembryonic antigen-related cell adhesion molecule 1 on murine dendritic cells is a potent regulator of T cell stimulation. J Immunol. 166:6537-44. Kaneko, K., M. Vey, M. Scott, S. Pilkuhn, F.E. Cohen, and S.B. Prusiner. 1997. COOH-terminal sequence of the cellular prion protein directs subcellular trafficking and controls conversion into the scrapie isoform. Proc Natl Acad Sci U S A. 94:2333-8. Kanthan, R., J.M. Radhi, and S.C. Kanthan. 2000. Gallbladder carcinomas: an immunoprognostic evaluation of P53, Bcl-2, CEA and alpha-fetoprotein. Can J Gastroenterol. 14:181-4. Karam, S.M. 1999. Lineage commitment and maturation of epithelial cells in the gut. Front Biosci. 4:D286-98.

175 Kasahara, K., Y. Watanabe, T. Yamamoto, and Y. Sanai. 1997. Association of Src family tyrosine kinase Lyn with ganglioside GD3 in rat brain. Possible regulation of Lyn by glycosphingolipid in caveolae-like domains. J Biol Chem. 272:29947-53. Kenny, P.A., and M.J. Bissell. 2003. Tumor reversion: correction of malignant behavior by microenvironmental cues. Int J Cancer. 107:688-95. Kenworthy, A.K., and M. Edidin. 1998. Distribution of a glycosylphosphatidylinositol-anchored protein at the apical surface of MDCK cells examined at a resolution of <100 A using imaging fluorescence resonance energy transfer. J Cell Biol. 142:69-84. Kenworthy, A.K., N. Petranova, and M. Edidin. 2000. High-resolution FRET microscopy of cholera toxin B-subunit and GPI-anchored proteins in cell plasma membranes. Mol Biol Cell. 11:1645-55. Kijima, H., G. Oshiba, T. Kenmochi, Y. Kise, H. Tanaka, O. Chino, H. Shimada, Y. Ueyama, M. Tanaka, and H. Makuuchi. 2000. Stromal CEA immunoreactivity is correlated with lymphatic invasion of human esophageal carcinoma. Int J Oncol. 16:677-82. Kilic, N., L. Oliveira-Ferrer, J.H. Wurmbach, S. Loges, F. Chalajour, S. Neshat-Vahid, J. Weil, M. Fernando, and S. Ergun. 2005. Pro-angiogenic signaling by the endothelial presence of CEACAM1. J Biol Chem. 280:2361-9. Kim, J., F.J. Kaye, J.G. Henslee, J.E. Shively, J.G. Park, S.L. Lai, R.I. Linnoila, J.L. Mulshine, and A.F. Gazdar. 1992. Expression of carcinoembryonic antigen and related genes in lung and gastrointestinal cancers. Int J Cancer. 52:718-25. Kim, K.B., S.I. Kim, H.J. Choo, J.H. Kim, and Y.G. Ko. 2004. Two-dimensional electrophoretic analysis reveals that lipid rafts are intact at physiological temperature. Proteomics. 4:3527-35. Kimura, A., C.A. Baumann, S.H. Chiang, and A.R. Saltiel. 2001. The sorbin homology domain: a motif for the targeting of proteins to lipid rafts. Proc Natl Acad Sci U S A. 98:9098-103. Kinugasa, T., M. Kuroki, H. Takeo, Y. Matsuo, K. Ohshima, Y. Yamashita, T. Shirakusa, and Y. Matsuoka. 1998. Expression of four CEA family antigens (CEA, NCA, BGP and CGM2) in normal and cancerous gastric epithelial cells: up-regulation of BGP and CGM2 in carcinomas. Int J Cancer. 76:148-53. Kirshner, J., C.J. Chen, P. Liu, J. Huang, and J.E. Shively. 2003a. CEACAM1-4S, a cell-cell adhesion molecule, mediates apoptosis and reverts mammary carcinoma cells to a normal morphogenic phenotype in a 3D culture. Proc Natl Acad Sci U S A. 100:521-6. Kirshner, J., D. Schumann, and J.E. Shively. 2003b. CEACAM1, a cell-cell adhesion molecule, directly associates with annexin II in a three-dimensional model of mammary morphogenesis. J Biol Chem. 278:50338-45. Klaile, E., M.M. Muller, C. Kannicht, B.B. Singer, and L. Lucka. 2005. CEACAM1 functionally interacts with filamin A and exerts a dual role in the regulation of cell migration. J Cell Sci. 118:5513-24. Kodelja, V., K. Lucas, S. Barnert, S. von Kleist, J.A. Thompson, and W. Zimmermann. 1989. Identification of a carcinoembryonic antigen gene family in the rat. Analysis of the N- terminal domains reveals immunoglobulin-like, hypervariable regions. J Biol Chem. 264:6906-12. Kodera, Y., K. Isobe, M. Yamauchi, T. Satta, T. Hasegawa, S. Oikawa, K. Kondoh, S. Akiyama, K. Itoh, I. Nakashima, and et al. 1993. Expression of carcinoembryonic antigen (CEA) and nonspecific crossreacting antigen (NCA) in gastrointestinal cancer; the correlation with degree of differentiation. Br J Cancer. 68:130-6.

176 Kodukula, K., D. Cines, R. Amthauer, L. Gerber, and S. Udenfriend. 1992. Biosynthesis of phosphatidylinositol-glycan (PI-G)-anchored membrane proteins in cell-free systems: cleavage of the nascent protein and addition of the PI-G moiety depend on the size of the COOH-terminal signal peptide. Proc Natl Acad Sci U S A. 89:1350-3. Kodukula, K., L.D. Gerber, R. Amthauer, L. Brink, and S. Udenfriend. 1993. Biosynthesis of glycosylphosphatidylinositol (GPI)-anchored membrane proteins in intact cells: specific amino acid requirements adjacent to the site of cleavage and GPI attachment. J Cell Biol. 120:657-64. Kolesnikova, T.V., C.S. Stipp, R.M. Rao, W.S. Lane, F.W. Luscinskas, and M.E. Hemler. 2004. EWI-2 modulates lymphocyte integrin alpha4beta1 functions. Blood. 103:3013-9. Kooyman, D.L., G.W. Byrne, S. McClellan, D. Nielsen, M. Tone, H. Waldmann, T.M. Coffman, K.R. McCurry, J.L. Platt, and J.S. Logan. 1995. In vivo transfer of GPI-linked complement restriction factors from erythrocytes to the endothelium. Science. 269:89-92. Koretz, K., S. Bruderlein, C. Henne, and P. Moller. 1992. Decay-accelerating factor (DAF, CD55) in normal colorectal mucosa, adenomas and carcinomas. Br J Cancer. 66:810-4. Koumanov, K.S., C. Wolf, and P.J. Quinn. 2004. Lipid composition of membrane domains. Subcell Biochem. 37:153-63. Kovalenko, O.V., X. Yang, T.V. Kolesnikova, and M.E. Hemler. 2004. Evidence for specific tetraspanin homodimers: inhibition of palmitoylation makes cysteine residues available for cross-linking. Biochem J. 377:407-17. Krauss, K., and P. Altevogt. 1999. Integrin leukocyte function-associated antigen-1-mediated cell binding can be activated by clustering of membrane rafts. J Biol Chem. 274:36921-7. Krishnan, S., M.P. Nambiar, V.G. Warke, C.U. Fisher, J. Mitchell, N. Delaney, and G.C. Tsokos. 2004. Alterations in lipid raft composition and dynamics contribute to abnormal T cell responses in systemic lupus erythematosus. J Immunol. 172:7821-31. Krop-Watorek, A., S. Oikawa, Y. Oyama, and H. Nakazato. 1998. Oligomerization of N- terminal domain of carcinoembryonic antigen (CEA) expressed in Escherichia coli. Biochem Biophys Res Commun. 242:79-83. Kropshofer, H., S. Spindeldreher, T.A. Rohn, N. Platania, C. Grygar, N. Daniel, A. Wolpl, H. Langen, V. Horejsi, and A.B. Vogt. 2002. Tetraspan microdomains distinct from lipid rafts enrich select peptide-MHC class II complexes. Nat Immunol. 3:61-8. Kuespert, K., S. Pils, and C.R. Hauck. 2006. CEACAMs: their role in physiology and pathophysiology. Curr Opin Cell Biol. 18:565-71. Kuijpers, T.W., M. Hoogerwerf, L.J. van der Laan, G. Nagel, C.E. van der Schoot, F. Grunert, and D. Roos. 1992. CD66 nonspecific cross-reacting antigens are involved in neutrophil adherence to cytokine-activated endothelial cells. J Cell Biol. 118:457-66. Kunath, T., C. Ordonez-Garcia, C. Turbide, and N. Beauchemin. 1995. Inhibition of colonic tumor cell growth by biliary glycoprotein. Oncogene. 11:2375-82. Kuroki, M., K. Hachimine, J. Huang, H. Shibaguchi, T. Kinugasa, S. Maekawa, and M. Kuroki. 2005. Re-targeting of cytotoxic T lymphocytes and/or natural killer cells to CEA- expressing tumor cells with anti-CEA antibody activity. Anticancer Res. 25:3725-32. Kuroki, M., M. Kuroki, H. Shibaguchi, A. Badran, K. Hachimine, J. Zhang, and T. Kinugasa. 2004. Strategies to endow cytotoxic T lymphocytes or natural killer cells with antibody activity against carcinoembryonic antigen. Tumour Biol. 25:208-16. Kuroki, M., Y. Matsuo, T. Kinugasa, and Y. Matsuoka. 1992. Three different NCA species, CGM6/CD67, NCA-95, and NCA-90, are comprised in the major 90 to 100-kDa band of

177 granulocyte NCA detectable upon SDS-polyacrylamide gel electrophoresis. Biochem Biophys Res Commun. 182:501-6. Kuroki, M., H. Matsushita, H. Matsumoto, Y. Hirose, T. Senba, and T. Yamamoto. 1999. Nonspecific cross-reacting antigen-50/90 (NCA-50/90) as a new tumor marker. Anticancer Res. 19:5599-606. Kusumi, A., I. Koyama-Honda, and K. Suzuki. 2004. Molecular dynamics and interactions for creation of stimulation-induced stabilized rafts from small unstable steady-state rafts. Traffic. 5:213-30. Lafont, F., and F.G. van der Goot. 2005. Bacterial invasion via lipid rafts. Cell Microbiol. 7:613- 20. Laguinge, L., O. Bajenova, E. Bowden, J. Sayyah, P. Thomas, and H. Juhl. 2005. Surface expression and CEA binding of hnRNP M4 protein in HT29 colon cancer cells. Anticancer Res. 25:23-31. Laliberte, J.P., L.W. McGinnes, M.E. Peeples, and T.G. Morrison. 2006. Integrity of membrane lipid rafts is necessary for the ordered assembly and release of infectious Newcastle disease virus particles. J Virol. 80:10652-62. Lange, Y., M.H. Swaisgood, B.V. Ramos, and T.L. Steck. 1989. Plasma membranes contain half the phospholipid and 90% of the cholesterol and sphingomyelin in cultured human fibroblasts. J Biol Chem. 264:3786-93. Laufs, S., J. Schumacher, and H. Allgayer. 2006. Urokinase-receptor (u-PAR): an essential player in multiple games of cancer: a review on its role in tumor progression, invasion, metastasis, proliferation/dormancy, clinical outcome and minimal residual disease. Cell Cycle. 5:1760-71. Lauke, H., N. Kilic, R. Bozorgzad, M. Fernando, S. Neshat-Vahid, T. Pottek, and S. Ergun. 2004. Expression of carcinoembryonic antigen-related cell adhesion molecule-1 (CEACAM1) in normal human Sertoli cells and its up-regulation in impaired spermatogenesis. Mol Hum Reprod. 10:247-52. Leitinger, B., and N. Hogg. 2002. The involvement of lipid rafts in the regulation of integrin function. J Cell Sci. 115:963-72. Lenzi, G.L., F. Grigoletto, M. Gent, R.S. Roberts, M.D. Walker, J.D. Easton, A. Carolei, F.C. Dorsey, W.A. Rocca, R. Bruno, and et al. 1994. Early treatment of stroke with monosialoganglioside GM-1. Efficacy and safety results of the Early Stroke Trial. Stroke. 25:1552-8. Leusch, H.G., Z. Drzeniek, Z. Markos-Pusztai, and C. Wagener. 1991. Binding of Escherichia coli and Salmonella strains to members of the carcinoembryonic antigen family: differential binding inhibition by aromatic alpha-glycosides of mannose. Infect Immun. 59:2051-7. Levin, L.V., and T.W. Griffin. 1991. Specific adhesion of carcinoembryonic antigen-bearing colorectal cancer cells to immobilized carcinoembryonic antigen. Cancer Lett. 60:143-52. Li, R., T. Liu, F. Yoshihiro, M. Tary-Lehmann, M. Obrenovich, H. Kuekrek, S.C. Kang, T. Pan, B.S. Wong, M.E. Medof, and M.S. Sy. 2003. On the same cell type GPI-anchored normal cellular prion and DAF protein exhibit different biological properties. Biochem Biophys Res Commun. 303:446-51. Li, X., D. Kaloyanova, M. van Eijk, R. Eerland, G. van der Goot, V. Oorschot, J. Klumperman, F. Lottspeich, V. Starkuviene, F.T. Wieland, and J.B. Helms. 2007. Involvement of a

178 Golgi-resident GPI-anchored Protein in Maintenance of the Golgi Structure. Mol Biol Cell. Lichtenberg, D., F.M. Goni, and H. Heerklotz. 2005. Detergent-resistant membranes should not be identified with membrane rafts. Trends Biochem Sci. 30:430-6. Lin, H., R. Huber, D. Schlessinger, and P.J. Morin. 1999. Frequent silencing of the GPC3 gene in ovarian cancer cell lines. Cancer Res. 59:807-10. Lipardi, C., R. Mora, V. Colomer, S. Paladino, L. Nitsch, E. Rodriguez-Boulan, and C. Zurzolo. 1998. Caveolin transfection results in caveolae formation but not apical sorting of glycosylphosphatidylinositol (GPI)-anchored proteins in epithelial cells. J Cell Biol. 140:617-26. Lloyd, R.V., J.C. Sisson, and P.J. Marangos. 1983. Calcitonin, carcinoembryonic antigen and neuron-specific enolase in medullary thyroid carcinoma. Cancer. 51:2234-9. Logsdon, C.D., D.M. Simeone, C. Binkley, T. Arumugam, J.K. Greenson, T.J. Giordano, D.E. Misek, R. Kuick, and S. Hanash. 2003. Molecular profiling of pancreatic adenocarcinoma and chronic pancreatitis identifies multiple genes differentially regulated in pancreatic cancer. Cancer Res. 63:2649-57. Lommerse, P.H., B.E. Snaar-Jagalska, H.P. Spaink, and T. Schmidt. 2005. Single-molecule diffusion measurements of H-Ras at the plasma membrane of live cells reveal microdomain localization upon activation. J Cell Sci. 118:1799-809. Low, M.G. 1989. Glycosyl-phosphatidylinositol: a versatile anchor for cell surface proteins. Faseb J. 3:1600-8. Low, M.G., and J.B. Finean. 1977. Release of alkaline phosphatase from membranes by a phosphatidylinositol-specific phospholipase C. Biochem J. 167:281-4. Low, M.G., and J.B. Finean. 1978. Specific release of plasma membrane enzymes by a phosphatidylinositol-specific phospholipase C. Biochim Biophys Acta. 508:565-70. Lowe, M.E. 1992. Site-specific mutations in the COOH-terminus of placental alkaline phosphatase: a single amino acid change converts a phosphatidylinositol-glycan-anchored protein to a secreted protein. J Cell Biol. 116:799-807. Lucero, H.A., and P.W. Robbins. 2004. Lipid rafts-protein association and the regulation of protein activity. Arch Biochem Biophys. 426:208-24. Luo, W., M. Tapolsky, K. Earley, C.G. Wood, D.R. Wilson, C.J. Logothetis, and S.H. Lin. 1999. Tumor-suppressive activity of CD66a in prostate cancer. Cancer Gene Ther. 6:313-21. Luo, W., C.G. Wood, K. Earley, M.C. Hung, and S.H. Lin. 1997. Suppression of tumorigenicity of breast cancer cells by an epithelial cell adhesion molecule (C-CAM1): the adhesion and growth suppression are mediated by different domains. Oncogene. 14:1697-704. Macdonald, J.L., and L.J. Pike. 2005. A simplified method for the preparation of detergent-free lipid rafts. J Lipid Res. 46:1061-7. Macnab, G.M., J.M. Urbanowicz, and M.C. Kew. 1978. Carcinoembryonic antigen in hepatocellular cancer. Br J Cancer. 38:51-4. Madore, N., K.L. Smith, C.H. Graham, A. Jen, K. Brady, S. Hall, and R. Morris. 1999. Functionally different GPI proteins are organized in different domains on the neuronal surface. Embo J. 18:6917-26. Maeda, Y., Y. Tashima, T. Houjou, M. Fujita, O.T. Yoko, Y. Jigami, R. Taguchi, and T. Kinoshita. 2007. Fatty Acid Remodeling of GPI-anchored Proteins Is Required for Their Raft Association. Mol Biol Cell. 18: 1497-506.

179 Magee, T., N. Pirinen, J. Adler, S.N. Pagakis, and I. Parmryd. 2002. Lipid rafts: cell surface platforms for T cell signaling. Biol Res. 35:127-31. Mahfoud, R., N. Garmy, M. Maresca, N. Yahi, A. Puigserver, and J. Fantini. 2002. Identification of a common sphingolipid-binding domain in Alzheimer, prion, and HIV-1 proteins. J Biol Chem. 277:11292-6. Markel, G., N. Lieberman, G. Katz, T.I. Arnon, M. Lotem, O. Drize, R.S. Blumberg, E. Bar- Haim, R. Mader, L. Eisenbach, and O. Mandelboim. 2002. CD66a interactions between human melanoma and NK cells: a novel class I MHC-independent inhibitory mechanism of cytotoxicity. J Immunol. 168:2803-10. Markel, G., R. Seidman, N. Stern, T. Cohen-Sinai, O. Izhaki, G. Katz, M. Besser, A.J. Treves, R.S. Blumberg, R. Loewenthal, O. Mandelboim, A. Orenstein, and J. Schachter. 2006. Inhibition of human tumor-infiltrating lymphocyte effector functions by the homophilic carcinoembryonic cell adhesion molecule 1 interactions. J Immunol. 177:6062-71. Marshall, J. 2003. Carcinoembryonic antigen-based vaccines. Semin Oncol. 30:30-6. Mataki, Y., S. Takao, K. Maemura, S. Mori, H. Shinchi, S. Natsugoe, and T. Aikou. 2004. Carcinoembryonic antigen messenger RNA expression using nested reverse transcription- PCR in the peripheral blood during follow-up period of patients who underwent curative surgery for biliary-pancreatic cancer: longitudinal analyses. Clin Cancer Res. 10:3807- 14. Mayor, S., K.G. Rothberg, and F.R. Maxfield. 1994. Sequestration of GPI-anchored proteins in caveolae triggered by cross-linking. Science. 264:1948-51. McCabe, J.B., and L.G. Berthiaume. 2001. N-terminal protein acylation confers localization to cholesterol, sphingolipid-enriched membranes but not to lipid rafts/caveolae. Mol Biol Cell. 12:3601-17. McConville, M.J., and M.A. Ferguson. 1993. The structure, biosynthesis and function of glycosylated phosphatidylinositols in the parasitic protozoa and higher eukaryotes. Biochem J. 294 ( Pt 2):305-24. McCuaig, K., M. Rosenberg, P. Nedellec, C. Turbide, and N. Beauchemin. 1993. Expression of the Bgp gene and characterization of mouse colon biliary glycoprotein isoforms. Gene. 127:173-83. McHugh, R.S., S. Nagarajan, Y.C. Wang, K.W. Sell, and P. Selvaraj. 1999. Protein transfer of glycosyl-phosphatidylinositol-B7-1 into tumor cell membranes: a novel approach to tumor immunotherapy. Cancer Res. 59:2433-7. Medof, M.E., T. Kinoshita, and V. Nussenzweig. 1984. Inhibition of complement activation on the surface of cells after incorporation of decay-accelerating factor (DAF) into their membranes. J Exp Med. 160:1558-78. Medof, M.E., T. Kinoshita, R. Silber, and V. Nussenzweig. 1985. Amelioration of lytic abnormalities of paroxysmal nocturnal hemoglobinuria with decay-accelerating factor. Proc Natl Acad Sci U S A. 82:2980-4. Medof, M.E., S. Nagarajan, and M.L. Tykocinski. 1996. Cell-surface engineering with GPI- anchored proteins. Faseb J. 10:574-86. Megha, and E. London. 2004. Ceramide selectively displaces cholesterol from ordered lipid domains (rafts): implications for lipid raft structure and function. J Biol Chem. 279:9997- 10004.

180 Melia, W.M., P.J. Johnson, S. Carter, A. Munro-Neville, and R. Williams. 1981. Plasma carcinoembryonic antigen in the diagnosis and management of patients with hepatocellular carcinoma. Cancer. 48:1004-8. Melkonian, K.A., A.G. Ostermeyer, J.Z. Chen, M.G. Roth, and D.A. Brown. 1999. Role of lipid modifications in targeting proteins to detergent-resistant membrane rafts. Many raft proteins are acylated, while few are prenylated. J Biol Chem. 274:3910-7. Meri, S., T. Lehto, C.W. Sutton, J. Tyynela, and M. Baumann. 1996. Structural composition and functional characterization of soluble CD59: heterogeneity of the oligosaccharide and glycophosphoinositol (GPI) anchor revealed by laser-desorption mass spectrometric analysis. Biochem J. 316 ( Pt 3):923-35. Metze, D., R. Bhardwaj, U. Amann, A.M. Eades-Perner, M. Neumaier, C. Wagener, P. Jantscheff, F. Grunert, and T.A. Luger. 1996. Glycoproteins of the carcinoembryonic antigen (CEA) family are expressed in sweat and sebaceous glands of human fetal and adult skin. J Invest Dermatol. 106:64-9. Micanovic, R., K. Kodukula, L.D. Gerber, and S. Udenfriend. 1990. Selectivity at the cleavage/attachment site of phosphatidylinositol-glycan anchored membrane proteins is enzymatically determined. Proc Natl Acad Sci U S A. 87:7939-43. Mineo, C., G.L. James, E.J. Smart, and R.G. Anderson. 1996. Localization of epidermal growth factor-stimulated Ras/Raf-1 interaction to caveolae membrane. J Biol Chem. 271:11930- 5. Minton, J.P., J.L. Hoehn, D.M. Gerber, J.S. Horsley, D.P. Connolly, F. Salwan, W.S. Fletcher, A.B. Cruz, Jr., F.G. Gatchell, M. Oviedo, and et al. 1985. Results of a 400-patient carcinoembryonic antigen second-look colorectal cancer study. Cancer. 55:1284-90. Moertel, C.G., A.J. Schutt, and V.L. Go. 1978. Carcinoembryonic antigen test for recurrent colorectal carcinoma. Inadequacy for early detection. Jama. 239:1065-6. Moller, L.B. 1993. Structure and function of the urokinase receptor. Blood Coagul Fibrinolysis. 4:293-303. Monaghan, E., V. Gueorguiev, C. Wilkins-Port, and P.J. McKeown-Longo. 2004. The receptor for urokinase-type plasminogen activator regulates fibronectin matrix assembly in human skin fibroblasts. J Biol Chem. 279:1400-7. Moran, P., H. Raab, W.J. Kohr, and I.W. Caras. 1991. Glycophospholipid membrane anchor attachment. Molecular analysis of the cleavage/attachment site. J Biol Chem. 266:1250-7. Muenzner, P., M. Rohde, S. Kneitz, and C.R. Hauck. 2005. CEACAM engagement by human pathogens enhances cell adhesion and counteracts bacteria-induced detachment of epithelial cells. J Cell Biol. 170:825-36. Mukasa, R., M. Umeda, T. Endo, A. Kobata, and K. Inoue. 1995. Characterization of glycosylphosphatidylinositol (GPI)-anchored NCAM on mouse skeletal muscle cell line C2C12: the structure of the GPI glycan and release during myogenesis. Arch Biochem Biophys. 318:182-90. Munro, S. 2003. Lipid rafts: elusive or illusive? Cell. 115:377-88. Murakoshi, H., R. Iino, T. Kobayashi, T. Fujiwara, C. Ohshima, A. Yoshimura, and A. Kusumi. 2004. Single-molecule imaging analysis of Ras activation in living cells. Proc Natl Acad Sci U S A. 101:7317-22. Murthy, S.S., T. Shen, A. De Rienzo, W.C. Lee, P.C. Ferriola, S.C. Jhanwar, B.T. Mossman, J. Filmus, and J.R. Testa. 2000. Expression of GPC3, an X-linked recessive overgrowth gene, is silenced in malignant mesothelioma. Oncogene. 19:410-6.

181 Nagarajan, S., M. Anderson, S.N. Ahmed, K.W. Sell, and P. Selvaraj. 1995. Purification and optimization of functional reconstitution on the surface of leukemic cell lines of GPI- anchored Fc gamma receptor III. J Immunol Methods. 184:241-51. Naghibalhossaini, F., and C.P. Stanners. 2004. Minimal mutations are required to effect a radical change in function in CEA family members of the Ig superfamily. J Cell Sci. 117:761-9. Naghibalhossaini, F., A.D. Yoder, M. Tobi, and C.P. Stanners. 2007. Evolution of a Tumorigenic Property Conferred by Glycophosphatidyl-Inositol Membrane Anchors of Carcinoembryonic Antigen Gene Family Members during the Primate Radiation. Mol Biol Cell. 18:1366-1374. Nair, K.S., and S.M. Zingde. 2001. Adhesion of neutrophils to fibronectin: role of the cd66 antigens. Cell Immunol. 208:96-106. Najjar, S.M., D. Accili, N. Philippe, J. Jernberg, R. Margolis, and S.I. Taylor. 1993. pp120/ecto- ATPase, an endogenous substrate of the insulin receptor tyrosine kinase, is expressed as two variably spliced isoforms. J Biol Chem. 268:1201-6. Najjar, S.M., N. Philippe, Y. Suzuki, G.A. Ignacio, P. Formisano, D. Accili, and S.I. Taylor. 1995. Insulin-stimulated phosphorylation of recombinant pp120/HA4, an endogenous substrate of the insulin receptor tyrosine kinase. Biochemistry. 34:9341-9. Nakano, Y., K. Noda, T. Endo, A. Kobata, and M. Tomita. 1994. Structural study on the glycosyl-phosphatidylinositol anchor and the asparagine-linked sugar chain of a soluble form of CD59 in human urine. Arch Biochem Biophys. 311:117-26. Nap, M., K. Mollgard, P. Burtin, and G.J. Fleuren. 1988. Immunohistochemistry of carcinoembryonic antigen in the embryo, fetus and adult. Tumour Biol. 9:145-53. Naslavsky, N., R. Stein, A. Yanai, G. Friedlander, and A. Taraboulos. 1997. Characterization of detergent-insoluble complexes containing the cellular prion protein and its scrapie isoform. J Biol Chem. 272:6324-31. Nayak, D.P., and S. Barman. 2002. Role of lipid rafts in virus assembly and budding. Adv Virus Res. 58:1-28. Nelson, C.M., and M.J. Bissell. 2006. Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer. Annu Rev Cell Dev Biol. 22:287-309. Neumaier, M., S. Paululat, A. Chan, P. Matthaes, and C. Wagener. 1993. Biliary glycoprotein, a potential human cell adhesion molecule, is down-regulated in colorectal carcinomas. Proc Natl Acad Sci U S A. 90:10744-8. Nguyen, D.H., B. Giri, G. Collins, and D.D. Taub. 2005. Dynamic reorganization of chemokine receptors, cholesterol, lipid rafts, and adhesion molecules to sites of CD4 engagement. Exp Cell Res. 304:559-69. Nicholson, T.B., and C.P. Stanners. 2006. Specific inhibition of GPI-anchored protein function by homing and self-association of specific GPI anchors. J Cell Biol. 175:647-59. Nicholson, T.B., and C.P. Stanners. 2007. Identification of a novel functional specificity signal within the GPI anchor signal sequence of carcinoembryonic antigen. J. Cell Biol. 177:211-218. Nicolau, D.V., Jr., K. Burrage, R.G. Parton, and J.F. Hancock. 2006. Identifying optimal lipid raft characteristics required to promote nanoscale protein-protein interactions on the plasma membrane. Mol Cell Biol. 26:313-23.

182 Nicoletti, I., G. Migliorati, M.C. Pagliacci, F. Grignani, and C. Riccardi. 1991. A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry. J Immunol Methods. 139:271-9. Nittka, S., J. Gunther, C. Ebisch, A. Erbersdobler, and M. Neumaier. 2004. The human tumor suppressor CEACAM1 modulates apoptosis and is implicated in early colorectal tumorigenesis. Oncogene. 23:9306-13. Niv, H., O. Gutman, Y. Kloog, and Y.I. Henis. 2002. Activated K-Ras and H-Ras display different interactions with saturable nonraft sites at the surface of live cells. J Cell Biol. 157:865-72. Nollau, P., F. Prall, U. Helmchen, C. Wagener, and M. Neumaier. 1997a. Dysregulation of carcinoembryonic antigen group members CGM2, CD66a (biliary glycoprotein), and nonspecific cross-reacting antigen in colorectal carcinomas. Comparative analysis by northern blot and in situ hybridization. Am J Pathol. 151:521-30. Nollau, P., H. Scheller, M. Kona-Horstmann, S. Rohde, F. Hagenmuller, C. Wagener, and M. Neumaier. 1997b. Expression of CD66a (human C-CAM) and other members of the carcinoembryonic antigen gene family of adhesion molecules in human colorectal adenomas. Cancer Res. 57:2354-7. Ohannesian, D.W., D. Lotan, and R. Lotan. 1994. Concomitant increases in galectin-1 and its glycoconjugate ligands (carcinoembryonic antigen, lamp-1, and lamp-2) in cultured human colon carcinoma cells by sodium butyrate. Cancer Res. 54:5992-6000. Ohannesian, D.W., D. Lotan, P. Thomas, J.M. Jessup, M. Fukuda, H.J. Gabius, and R. Lotan. 1995. Carcinoembryonic antigen and other glycoconjugates act as ligands for galectin-3 in human colon carcinoma cells. Cancer Res. 55:2191-9. Ohishi, K., N. Inoue, and T. Kinoshita. 2001. PIG-S and PIG-T, essential for GPI anchor attachment to proteins, form a complex with GAA1 and GPI8. Embo J. 20:4088-98. Ohishi, K., N. Inoue, Y. Maeda, J. Takeda, H. Riezman, and T. Kinoshita. 2000. Gaa1p and gpi8p are components of a glycosylphosphatidylinositol (GPI) transamidase that mediates attachment of GPI to proteins. Mol Biol Cell. 11:1523-33. Ohishi, K., K. Nagamune, Y. Maeda, and T. Kinoshita. 2003. Two subunits of glycosylphosphatidylinositol transamidase, GPI8 and PIG-T, form a functionally important intermolecular disulfide bridge. J Biol Chem. 278:13959-67. Ohlsson, L., M.L. Hammarstrom, A. Israelsson, L. Naslund, A. Oberg, G. Lindmark, and S. Hammarstrom. 2006. Biomarker selection for detection of occult tumour cells in lymph nodes of colorectal cancer patients using real-time quantitative RT-PCR. Br J Cancer. 95:218-25. Ohwada, A., H. Takahashi, I. Nagaoka, and S. Kira. 1994. Biliary glycoprotein mRNA expression is increased in primary lung cancer, especially in squamous cell carcinoma. Am J Respir Cell Mol Biol. 11:214-20. Oikawa, S., C. Inuzuka, M. Kuroki, F. Arakawa, Y. Matsuoka, G. Kosaki, and H. Nakazato. 1991. A specific heterotypic cell adhesion activity between members of carcinoembryonic antigen family, W272 and NCA, is mediated by N-domains. J Biol Chem. 266:7995-8001. Oikawa, S., C. Inuzuka, M. Kuroki, Y. Matsuoka, G. Kosaki, and H. Nakazato. 1989. Cell adhesion activity of non-specific cross-reacting antigen (NCA) and carcinoembryonic antigen (CEA) expressed on CHO cell surface: homophilic and heterophilic adhesion. Biochem Biophys Res Commun. 164:39-45.

183 Oikawa, S., M. Kuroki, Y. Matsuoka, G. Kosaki, and H. Nakazato. 1992. Homotypic and heterotypic Ca(++)-independent cell adhesion activities of biliary glycoprotein, a member of carcinoembryonic antigen family, expressed on CHO cell surface. Biochem Biophys Res Commun. 186:881-7. Oikawa, S., M. Sugiyama, M. Kuroki, M. Kuroki, and H. Nakazato. 2000. Extracellular N- domain alone can mediate specific heterophilic adhesion between members of the carcinoembryonic antigen family, CEACAM6 and CEACAM8. Biochem Biophys Res Commun. 278:564-8. Oliveira-Ferrer, L., D. Tilki, G. Ziegeler, J. Hauschild, S. Loges, S. Irmak, E. Kilic, H. Huland, M. Friedrich, and S. Ergun. 2004. Dual role of carcinoembryonic antigen-related cell adhesion molecule 1 in angiogenesis and invasion of human urinary bladder cancer. Cancer Res. 64:8932-8. Omaetxebarria, M.J., P. Hagglund, F. Elortza, N.M. Hooper, J.M. Arizmendi, and O.N. Jensen. 2006. Isolation and characterization of glycosylphosphatidylinositol-anchored peptides by hydrophilic interaction chromatography and MALDI tandem mass spectrometry. Anal Chem. 78:3335-41. Ono, A., and E.O. Freed. 2005. Role of lipid rafts in virus replication. Adv Virus Res. 64:311-58. Ordonez, C., R.A. Screaton, C. Ilantzis, and C.P. Stanners. 2000. Human carcinoembryonic antigen functions as a general inhibitor of anoikis. Cancer Res. 60:3419-24. Ordonez, C., A.B. Zhai, P. Camacho-Leal, L. Demarte, M.M. Fan, and C.P. Stanners. 2007. GPI- anchored CEA family glycoproteins CEA and CEACAM6 mediate their biological effects through enhanced integrin alpha5beta1-fibronectin interaction&. J Cell Physiol. 210:757-65. Paladino, S., T. Pocard, M.A. Catino, and C. Zurzolo. 2006. GPI-anchored proteins are directly targeted to the apical surface in fully polarized MDCK cells. J Cell Biol. 172:1023-34. Pankov, R., T. Markovska, R. Hazarosova, P. Antonov, L. Ivanova, and A. Momchilova. 2005. Cholesterol distribution in plasma membranes of beta1 integrin-expressing and beta1 integrin-deficient fibroblasts. Arch Biochem Biophys. 442:160-8. Park, C.C., H. Zhang, M. Pallavicini, J.W. Gray, F. Baehner, C.J. Park, and M.J. Bissell. 2006a. Beta1 integrin inhibitory antibody induces apoptosis of breast cancer cells, inhibits growth, and distinguishes malignant from normal phenotype in three dimensional cultures and in vivo. Cancer Res. 66:1526-35. Park, S.Y., Y.R. Cho, H.J. Kim, E.G. Hong, T. Higashimori, S.J. Lee, I.J. Goldberg, G.I. Shulman, S.M. Najjar, and J.K. Kim. 2006b. Mechanism of glucose intolerance in mice with dominant negative mutation of CEACAM1. Am J Physiol Endocrinol Metab. 291:E517-24. Parton, R.G. 1996. Caveolae and caveolins. Curr Opin Cell Biol. 8:542-8. Pasqualini, R., S. Bourdoulous, E. Koivunen, V.L. Woods, Jr., and E. Ruoslahti. 1996. A polymeric form of fibronectin has antimetastatic effects against multiple tumor types. Nat Med. 2:1197-203. Paxton, R.J., G. Mooser, H. Pande, T.D. Lee, and J.E. Shively. 1987. Sequence analysis of carcinoembryonic antigen: identification of glycosylation sites and homology with the immunoglobulin supergene family. Proc Natl Acad Sci U S A. 84:920-4. Peck, D., and F.S. Walsh. 1993. Differential effects of over-expressed neural cell adhesion molecule isoforms on myoblast fusion. J Cell Biol. 123:1587-95.

184 Pham, L.V., A.T. Tamayo, L.C. Yoshimura, P. Lo, N. Terry, P.S. Reid, and R.J. Ford. 2002. A CD40 Signalosome anchored in lipid rafts leads to constitutive activation of NF-kappaB and autonomous cell growth in B cell lymphomas. Immunity. 16:37-50. Pike, L.J. 2004. Lipid rafts: heterogeneity on the high seas. Biochem J. 378:281-92. Pike, L.J. 2006. Rafts defined: a report on the Keystone Symposium on Lipid Rafts and Cell Function. J Lipid Res. 47:1597-8. Pike, L.J., X. Han, and R.W. Gross. 2005. Epidermal growth factor receptors are localized to lipid rafts that contain a balance of inner and outer leaflet lipids: a shotgun lipidomics study. J Biol Chem. 280:26796-804. Plowman, S.J., C. Muncke, R.G. Parton, and J.F. Hancock. 2005. H-ras, K-ras, and inner plasma membrane raft proteins operate in nanoclusters with differential dependence on the actin cytoskeleton. Proc Natl Acad Sci U S A. 102:15500-5. Polishchuk, R., A. Di Pentima, and J. Lippincott-Schwartz. 2004. Delivery of raft-associated, GPI-anchored proteins to the apical surface of polarized MDCK cells by a transcytotic pathway. Nat Cell Biol. 6:297-307. Pollard, J.W., and C.P. Stanners. 1979. Characterization of cell lines showing growth control isolated from both the wild type and a leucyl-tRNA synthetase mutant of Chinese hamster ovary cells. J Cell Physiol. 98:571-85. Poncet, C., V. Frances, R. Gristina, C. Scheiner, J.F. Pellissier, and D. Figarella-Branger. 1996. CD24, a glycosylphosphatidylinositol-anchored molecules is transiently expressed during the development of human central nervous system and is a marker of human neural cell lineage tumors. Acta Neuropathol (Berl). 91:400-8. Poy, M.N., Y. Yang, K. Rezaei, M.A. Fernstrom, A.D. Lee, Y. Kido, S.K. Erickson, and S.M. Najjar. 2002. CEACAM1 regulates insulin clearance in liver. Nat Genet. 30:270-6. Prall, F., P. Nollau, M. Neumaier, H.D. Haubeck, Z. Drzeniek, U. Helmchen, T. Loning, and C. Wagener. 1996. CD66a (BGP), an adhesion molecule of the carcinoembryonic antigen family, is expressed in epithelium, endothelium, and myeloid cells in a wide range of normal human tissues. J Histochem Cytochem. 44:35-41. Pralle, A., P. Keller, E.L. Florin, K. Simons, and J.K. Horber. 2000. Sphingolipid-cholesterol rafts diffuse as small entities in the plasma membrane of mammalian cells. J Cell Biol. 148:997-1008. Premkumar, D.R., Y. Fukuoka, D. Sevlever, E. Brunschwig, T.L. Rosenberry, M.L. Tykocinski, and M.E. Medof. 2001. Properties of exogenously added GPI-anchored proteins following their incorporation into cells. J Cell Biochem. 82:234-45. Prior, I.A., C. Muncke, R.G. Parton, and J.F. Hancock. 2003. Direct visualization of Ras proteins in spatially distinct cell surface microdomains. J Cell Biol. 160:165-70. Prusiner, S.B. 1991. Molecular biology of prion diseases. Science. 252:1515-22. Ramachandran, A., M. Madesh, and K.A. Balasubramanian. 2000. Apoptosis in the intestinal epithelium: its relevance in normal and pathophysiological conditions. J Gastroenterol Hepatol. 15:109-20. Rao, Y., X.F. Wu, J. Gariepy, U. Rutishauser, and C.H. Siu. 1992. Identification of a peptide sequence involved in homophilic binding in the neural cell adhesion molecule NCAM. J Cell Biol. 118:937-49. Rebstock, S., K. Lucas, M. Weiss, J. Thompson, and W. Zimmermann. 1993. Spatiotemporal expression of pregnancy-specific glycoprotein gene rnCGM1 in rat placenta. Dev Dyn. 198:171-81.

185 Rege, T.A., and J.S. Hagood. 2006a. Thy-1 as a regulator of cell-cell and cell-matrix interactions in axon regeneration, apoptosis, adhesion, migration, cancer, and fibrosis. Faseb J. 20:1045-54. Rege, T.A., and J.S. Hagood. 2006b. Thy-1, a versatile modulator of signaling affecting cellular adhesion, proliferation, survival, and cytokine/growth factor responses. Biochim Biophys Acta. 1763:991-9. Riethdorf, L., B.W. Lisboa, U. Henkel, M. Naumann, C. Wagener, and T. Loning. 1997. Differential expression of CD66a (BGP), a cell adhesion molecule of the carcinoembryonic antigen family, in benign, premalignant, and malignant lesions of the human mammary gland. J Histochem Cytochem. 45:957-63. Roberts, W.L., J.J. Myher, A. Kuksis, M.G. Low, and T.L. Rosenberry. 1988. Lipid analysis of the glycoinositol phospholipid membrane anchor of human erythrocyte acetylcholinesterase. Palmitoylation of inositol results in resistance to phosphatidylinositol-specific phospholipase C. J Biol Chem. 263:18766-75. Robinson, P.J. 1997. Signal transduction via GPI-anchored membrane proteins. Adv Exp Med Biol. 419:365-70. Rojas, M., L. DeMarte, R.A. Screaton, and C.P. Stanners. 1996. Radical differences in functions of closely related members of the human carcinoembryonic antigen gene family. Cell Growth Differ. 7:655-62. Rojas, M., A. Fuks, and C.P. Stanners. 1990. Biliary glycoprotein, a member of the immunoglobulin supergene family, functions in vitro as a Ca2(+)-dependent intercellular adhesion molecule. Cell Growth Differ. 1:527-33. Rosenberger, C.M., J.H. Brumell, and B.B. Finlay. 2000. Microbial pathogenesis: lipid rafts as pathogen portals. Curr Biol. 10:R823-5. Rotblat, B., I.A. Prior, C. Muncke, R.G. Parton, Y. Kloog, Y.I. Henis, and J.F. Hancock. 2004. Three separable domains regulate GTP-dependent association of H-ras with the plasma membrane. Mol Cell Biol. 24:6799-810. Rothman, J.E., and J. Lenard. 1977. Membrane asymmetry. Science. 195:743-53. Roy, S., S. Plowman, B. Rotblat, I.A. Prior, C. Muncke, S. Grainger, R.G. Parton, Y.I. Henis, Y. Kloog, and J.F. Hancock. 2005. Individual palmitoyl residues serve distinct roles in H-ras trafficking, microlocalization, and signaling. Mol Cell Biol. 25:6722-33. Rule, A.H., E. Straus, J. Vandevoorde, and H.D. Janowitz. 1972. Tumor-associated (CEA- reacting) antigen in patients with inflammatory bowel disease. N Engl J Med. 287:24-6. Ruoslahti, E., and J.C. Reed. 1994. Anchorage dependence, integrins, and apoptosis. Cell. 77:477-8. Sadahiro, S., T. Suzuki, Y. Maeda, S. Yurimoto, S. Yasuda, H. Makuuchi, A. Kamijo, and C. Murayama. 2007. Detection of Carcinoembryonic Antigen Messenger RNA-Expressing Cells in Peripheral Blood 7 Days After Curative Surgery is a Novel Prognostic Factor in Colorectal Cancer. Ann Surg Oncol. Sadekova, S., N. Lamarche-Vane, X. Li, and N. Beauchemin. 2000. The CEACAM1-L glycoprotein associates with the actin cytoskeleton and localizes to cell-cell contact through activation of Rho-like GTPases. Mol Biol Cell. 11:65-77. Saied, G.M., W.H. El-Metenawy, M.S. Elwan, and N.R. Dessouki. 2007. Urine carcinoembryonic antigen levels are more useful than serum levels for early detection of Bilharzial and non-Bilharzial urinary bladder carcinoma: Observations of 43 Egyptian cases. World J Surg Oncol. 5:4.

186 Satoh, Y., T. Hayashi, T. Takahashi, F. Itoh, M. Adachi, M. Fukui, M. Kuroki, M. Kuroki, K. Imai, and Y. Hinoda. 2002. Expression of CD66a in multiple myeloma. J Clin Lab Anal. 16:79-85. Sauter, S.L., S.M. Rutherfurd, C. Wagener, J.E. Shively, and S.A. Hefta. 1993. Identification of the specific oligosaccharide sites recognized by type 1 fimbriae from Escherichia coli on nonspecific cross-reacting antigen, a CD66 cluster granulocyte glycoprotein. J Biol Chem. 268:15510-6. Schade, A.E., and A.D. Levine. 2002. Lipid raft heterogeneity in human peripheral blood T lymphoblasts: a mechanism for regulating the initiation of TCR signal transduction. J Immunol. 168:2233-9. Scheiffele, P., M.G. Roth, and K. Simons. 1997. Interaction of influenza virus haemagglutinin with sphingolipid-cholesterol membrane domains via its transmembrane domain. Embo J. 16:5501-8. Schmitter, T., F. Agerer, L. Peterson, P. Munzner, and C.R. Hauck. 2004. Granulocyte CEACAM3 is a phagocytic receptor of the innate immune system that mediates recognition and elimination of human-specific pathogens. J Exp Med. 199:35-46. Schneider, J.S., A. Pope, K. Simpson, J. Taggart, M.G. Smith, and L. DiStefano. 1992. Recovery from experimental parkinsonism in primates with GM1 ganglioside treatment. Science. 256:843-6. Schneider, J.S., D.P. Roeltgen, E.L. Mancall, J. Chapas-Crilly, D.S. Rothblat, and G.T. Tatarian. 1998. Parkinson's disease: improved function with GM1 ganglioside treatment in a randomized placebo-controlled study. Neurology. 50:1630-6. Schneider, P., M.A. Ferguson, M.J. McConville, A. Mehlert, S.W. Homans, and C. Bordier. 1990. Structure of the glycosyl-phosphatidylinositol membrane anchor of the Leishmania major promastigote surface protease. J Biol Chem. 265:16955-64. Schnitzer, J.E., D.P. McIntosh, A.M. Dvorak, J. Liu, and P. Oh. 1995. Separation of caveolae from associated microdomains of GPI-anchored proteins. Science. 269:1435-9. Scholzel, S., W. Zimmermann, G. Schwarzkopf, F. Grunert, B. Rogaczewski, and J. Thompson. 2000. Carcinoembryonic antigen family members CEACAM6 and CEACAM7 are differentially expressed in normal tissues and oppositely deregulated in hyperplastic colorectal polyps and early adenomas. Am J Pathol. 156:595-605. Schumann, D., C.J. Chen, B. Kaplan, and J.E. Shively. 2001. Carcinoembryonic antigen cell adhesion molecule 1 directly associates with cytoskeleton proteins actin and tropomyosin. J Biol Chem. 276:47421-33. Schutz, G.J., G. Kada, V.P. Pastushenko, and H. Schindler. 2000. Properties of lipid microdomains in a muscle cell membrane visualized by single molecule microscopy. Embo J. 19:892-901. Screaton, R.A., L. DeMarte, P. Draber, and C.P. Stanners. 2000. The specificity for the differentiation blocking activity of carcinoembryonic antigen resides in its glycophosphatidyl-inositol anchor. J Cell Biol. 150:613-26. Screaton, R.A., L.Z. Penn, and C.P. Stanners. 1997. Carcinoembryonic antigen, a human tumor marker, cooperates with Myc and Bcl-2 in cellular transformation. J Cell Biol. 137:939- 52. Sharkey, R.M., T.M. Cardillo, E.A. Rossi, C.H. Chang, H. Karacay, W.J. McBride, H.J. Hansen, I.D. Horak, and D.M. Goldenberg. 2005. Signal amplification in molecular imaging by pretargeting a multivalent, bispecific antibody. Nat Med. 11:1250-5.

187 Sharma, P., R. Varma, R.C. Sarasij, Ira, K. Gousset, G. Krishnamoorthy, M. Rao, and S. Mayor. 2004. Nanoscale Organization of Multiple GPI-Anchored Proteins in Living Cell Membranes. Cell. 116:577-89. Shenoy-Scaria, A.M., L.K. Gauen, J. Kwong, A.S. Shaw, and D.M. Lublin. 1993. Palmitylation of an amino-terminal cysteine motif of protein tyrosine kinases p56lck and p59fyn mediates interaction with glycosyl-phosphatidylinositol-anchored proteins. Mol Cell Biol. 13:6385-92. Shenoy-Scaria, A.M., J. Kwong, T. Fujita, M.W. Olszowy, A.S. Shaw, and D.M. Lublin. 1992. Signal transduction through decay-accelerating factor. Interaction of glycosylphosphatidylinositol anchor and protein tyrosine kinases p56lck and p59fyn 1. J Immunol. 149:3535-41. Shi, Z.R., D. Tacha, and S.H. Itzkowitz. 1994. Monoclonal antibody COL-1 reacts with restricted epitopes on carcinoembryonic antigen: an immunohistochemical study. J Histochem Cytochem. 42:1215-9. Shogomori, H., and D.A. Brown. 2003. Use of detergents to study membrane rafts: the good, the bad, and the ugly. Biol Chem. 384:1259-63. Sienel, W., S. Dango, U. Woelfle, A. Morresi-Hauf, C. Wagener, J. Brummer, W. Mutschler, B. Passlick, and K. Pantel. 2003. Elevated expression of carcinoembryonic antigen-related cell adhesion molecule 1 promotes progression of non-small cell lung cancer. Clin Cancer Res. 9:2260-6. Silvius, J.R. 2003. Role of cholesterol in lipid raft formation: lessons from lipid model systems. Biochim Biophys Acta. 1610:174-83. Simons, K., and E. Ikonen. 1997. Functional rafts in cell membranes. Nature. 387:569-72. Simons, K., and D. Toomre. 2000. Lipid rafts and signal transduction. Nat Rev Mol Cell Biol. 1:31-9. Simons, M., T. Friedrichson, J.B. Schulz, M. Pitto, M. Masserini, and T.V. Kurzchalia. 1999. Exogenous administration of gangliosides displaces GPI-anchored proteins from lipid microdomains in living cells. Mol Biol Cell. 10:3187-96. Singer, B.B., E. Klaile, I. Scheffrahn, M.M. Muller, R. Kammerer, W. Reutter, B. Obrink, and L. Lucka. 2005. CEACAM1 (CD66a) mediates delay of spontaneous and Fas ligand- induced apoptosis in granulocytes. Eur J Immunol. 35:1949-59. Singer, S.J., and G.L. Nicolson. 1972. The fluid mosaic model of the structure of cell membranes. Science. 175:720-31. Skubitz, K.M., T.P. Ducker, and S.A. Goueli. 1992. CD66 monoclonal antibodies recognize a phosphotyrosine-containing protein bearing a carcinoembryonic antigen cross-reacting antigen on the surface of human neutrophils. J Immunol. 148:852-60. Sobering, A.K., R. Watanabe, M.J. Romeo, B.C. Yan, C.A. Specht, P. Orlean, H. Riezman, and D.E. Levin. 2004. Yeast Ras regulates the complex that catalyzes the first step in GPI- anchor biosynthesis at the ER. Cell. 117:637-48. Soeth, E., T. Wirth, H.J. List, S. Kumbhani, A. Petersen, M. Neumaier, F. Czubayko, and H. Juhl. 2001. Controlled ribozyme targeting demonstrates an antiapoptotic effect of carcinoembryonic antigen in HT29 colon cancer cells. Clin Cancer Res. 7:2022-30. Solomon, K.R., M.A. Mallory, K.A. Hanify, and R.W. Finberg. 1998. The nature of membrane anchorage determines kinase association and detergent solubility of CD4. Biochem Biophys Res Commun. 242:423-8.

188 Soni, P., M. Lakkis, M.N. Poy, M.A. Fernstrom, and S.M. Najjar. 2000. The differential effects of pp120 (Ceacam 1) on the mitogenic action of insulin and insulin-like growth factor 1 are regulated by the nonconserved tyrosine 1316 in the insulin receptor. Mol Cell Biol. 20:3896-905. Soroka, V., K. Kolkova, J.S. Kastrup, K. Diederichs, J. Breed, V.V. Kiselyov, F.M. Poulsen, I.K. Larsen, W. Welte, V. Berezin, E. Bock, and C. Kasper. 2003. Structure and interactions of NCAM Ig1-2-3 suggest a novel zipper mechanism for homophilic adhesion. Structure (Camb). 11:1291-301. Sowa, G., M. Pypaert, and W.C. Sessa. 2001. Distinction between signaling mechanisms in lipid rafts vs. caveolae. Proc Natl Acad Sci U S A. 98:14072-7. Spurway, T.D., J.A. Dalley, S. High, and N.J. Bulleid. 2001. Early events in glycosylphosphatidylinositol anchor addition. substrate proteins associate with the transamidase subunit gpi8p. J Biol Chem. 276:15975-82. Stacker, S.A., N.P. Sacks, J. Golder, J.J. Tjandra, C.H. Thompson, A. Smithyman, and I.F. McKenzie. 1988. Evaluation of MSA as a serum marker in breast cancer: a comparison with CEA. Br J Cancer. 57:298-303. Stahl, N., D.R. Borchelt, K. Hsiao, and S.B. Prusiner. 1987. Scrapie prion protein contains a phosphatidylinositol glycolipid. Cell. 51:229-40. Stanners, C.P., M. Rojas, H. Zhou, A. Fuks, and N. Beauchemin. 1992. The CEA family: a system in transitional evolution? Int J Biol Markers. 7:137-42. Stefanova, I., V. Horejsi, I.J. Ansotegui, W. Knapp, and H. Stockinger. 1991. GPI-anchored cell- surface molecules complexed to protein tyrosine kinases. Science. 254:1016-9. Stein, R., and D.M. Goldenberg. 2004. A humanized monoclonal antibody to carcinoembryonic antigen, labetuzumab, inhibits tumor growth and sensitizes human medullary thyroid cancer xenografts to dacarbazine chemotherapy. Mol Cancer Ther. 3:1559-64. Stern, N., G. Markel, T.I. Arnon, R. Gruda, H. Wong, S.D. Gray-Owen, and O. Mandelboim. 2005. Carcinoembryonic antigen (CEA) inhibits NK killing via interaction with CEA- related cell adhesion molecule 1. J Immunol. 174:6692-701. Stevens, D.P., and I.R. Mackay. 1973. Increased carcinoembryonic antigen in heavy cigarette smokers. Lancet. 2:1238-9. Stipp, C.S., T.V. Kolesnikova, and M.E. Hemler. 2003. EWI-2 regulates alpha3beta1 integrin- dependent cell functions on laminin-5. J Cell Biol. 163:1167-77. Storch, J., and A.M. Kleinfeld. 1985. The lipid structure of biological membranes. Trends Biochem Sci. 10:418-421. Stuermer, C.A., and H. Plattner. 2005. The 'lipid raft' microdomain proteins reggie-1 and reggie- 2 (flotillins) are scaffolds for protein interaction and signalling. Biochem Soc Symp:109- 18. Stulnig, T.M., M. Berger, T. Sigmund, D. Raederstorff, H. Stockinger, and W. Waldhausl. 1998. Polyunsaturated fatty acids inhibit T cell signal transduction by modification of detergent-insoluble membrane domains. J Cell Biol. 143:637-44. Stulnig, T.M., M. Berger, T. Sigmund, H. Stockinger, V. Horejsi, and W. Waldhausl. 1997. Signal transduction via glycosyl phosphatidylinositol-anchored proteins in T cells is inhibited by lowering cellular cholesterol. J Biol Chem. 272:19242-7. Stupack, D.G., and D.A. Cheresh. 2002. Get a ligand, get a life: integrins, signaling and cell survival. J Cell Sci. 115:3729-38.

189 Stupack, D.G., and D.A. Cheresh. 2004. Integrins and angiogenesis. Curr Top Dev Biol. 64:207- 38. Subczynski, W.K., and A. Kusumi. 2003. Dynamics of raft molecules in the cell and artificial membranes: approaches by pulse EPR spin labeling and single molecule optical microscopy. Biochim Biophys Acta. 1610:231-43. Sugita, K., T. Mori, S. Yokota, M. Kuroki, T.O. Koyama, T. Inukai, K. Iijima, K. Goi, T. Tezuka, S. Kojika, K. Shiraishi, M. Nakamura, N. Miyamoto, N. Karakida, K. Kagami, and S. Nakazawa. 1999. The KOR-SA3544 antigen predominantly expressed on the surface of Philadelphia chromosome-positive acute lymphoblastic leukemia cells is nonspecific cross-reacting antigen-50/90 (CD66c) and invariably expressed in cytoplasm of human leukemia cells. Leukemia. 13:779-85. Surviladze, Z., L. Draberova, L. Kubinova, and P. Draber. 1998. Functional heterogeneity of Thy-1 membrane microdomains in rat basophilic leukemia cells. Eur J Immunol. 28:1847-58. Taguchi, R., N. Hamakawa, M. Harada-Nishida, T. Fukui, K. Nojima, and H. Ikezawa. 1994. Microheterogeneity in glycosylphosphatidylinositol anchor structures of bovine liver 5'- nucleotidase. Biochemistry. 33:1017-22. Taguchi, R., N. Hamakawa, N. Maekawa, and H. Ikezawa. 1999. Application of electrospray ionization MS/MS and matrix-assisted laser desorption/ionization-time of flight mass spectrometry to structural analysis of the glycosyl-phosphatidylinositol-anchored protein. J Biochem (Tokyo). 126:421-9. Taheri, M., H.U. Saragovi, and C.P. Stanners. 2003. The adhesion and differentiation-inhibitory activities of the immunoglobulin superfamily member, carcinoembryonic antigen, can be independently blocked. J Biol Chem. 278:14632-9. Taheri, M., U. Saragovi, A. Fuks, J. Makkerh, J. Mort, and C.P. Stanners. 2000. Self recognition in the Ig superfamily. Identification of precise subdomains in carcinoembryonic antigen required for intercellular adhesion. J Biol Chem. 275:26935-43. Takami, N., S. Ogata, K. Oda, Y. Misumi, and Y. Ikehara. 1988. Biosynthesis of placental alkaline phosphatase and its post-translational modification by glycophospholipid for membrane-anchoring. J Biol Chem. 263:3016-21. Takeda, M., G.P. Leser, C.J. Russell, and R.A. Lamb. 2003. Influenza virus hemagglutinin concentrates in lipid raft microdomains for efficient viral fusion. Proc Natl Acad Sci U S A. 100:14610-7. Tan, W.K., C.E. Williams, A.J. Gunn, E.C. Mallard, and P.D. Gluckman. 1993. Pretreatment with monosialoganglioside GM1 protects the brain of fetal sheep against hypoxic- ischemic injury without causing systemic compromise. Pediatr Res. 34:18-22. Tanaka, K., Y. Hinoda, H. Takahashi, H. Sakamoto, Y. Nakajima, and K. Imai. 1997. Decreased expression of biliary glycoprotein in hepatocellular carcinomas. Int J Cancer. 74:15-9. Tanaka, S., Y. Maeda, Y. Tashima, and T. Kinoshita. 2004. Inositol deacylation of glycosylphosphatidylinositol-anchored proteins is mediated by mammalian PGAP1 and yeast Bst1p. J Biol Chem. 279:14256-63. Tashima, Y., R. Taguchi, C. Murata, H. Ashida, T. Kinoshita, and Y. Maeda. 2006. PGAP2 is essential for correct processing and stable expression of GPI-anchored proteins. Mol Biol Cell. 17:1410-20.

190 Teixeira, A.M., J. Fawcett, D.L. Simmons, and S.M. Watt. 1994. The N-domain of the biliary glycoprotein (BGP) adhesion molecule mediates homotypic binding: domain interactions and epitope analysis of BGPc. Blood. 84:211-9. Thies, A., I. Moll, J. Berger, C. Wagener, J. Brummer, H.J. Schulze, G. Brunner, and U. Schumacher. 2002. CEACAM1 expression in cutaneous malignant melanoma predicts the development of metastatic disease. J Clin Oncol. 20:2530-6. Thomas, J.R., R.A. Dwek, and T.W. Rademacher. 1990. Structure, biosynthesis, and function of glycosylphosphatidylinositols. Biochemistry. 29:5413-22. Thompson, J., M. Seitz, E. Chastre, M. Ditter, C. Aldrian, C. Gespach, and W. Zimmermann. 1997a. Down-regulation of carcinoembryonic antigen family member 2 expression is an early event in colorectal tumorigenesis. Cancer Res. 57:1776-84. Thompson, J., W. Zimmermann, P. Nollau, M. Neumaier, J. Weber-Arden, H. Schrewe, I. Craig, and T. Willcocks. 1994. CGM2, a member of the carcinoembryonic antigen gene family is down-regulated in colorectal carcinomas. J Biol Chem. 269:32924-31. Thompson, J.A., A.M. Eades-Perner, M. Ditter, W.J. Muller, and W. Zimmermann. 1997b. Expression of transgenic carcinoembryonic antigen (CEA) in tumor-prone mice: an animal model for CEA-directed tumor immunotherapy. Int J Cancer. 72:197-202. Thompson, J.A., F. Grunert, and W. Zimmermann. 1991. Carcinoembryonic antigen gene family: molecular biology and clinical perspectives. J Clin Lab Anal. 5:344-66. Tilki, D., S. Irmak, L. Oliveira-Ferrer, J. Hauschild, K. Miethe, H. Atakaya, P. Hammerer, M.G. Friedrich, G. Schuch, R. Galalae, C.G. Stief, E. Kilic, H. Huland, and S. Ergun. 2006. CEA-related cell adhesion molecule-1 is involved in angiogenic switch in prostate cancer. Oncogene. 25:4965-74. Tobi, M., S. Chintalapani, K. Kithier, and N. Clapp. 2000. Carcinoembryonic antigen family of adhesion molecules in the cotton top tamarin (Saguinus oedipus). Cancer Lett. 157:45-50. Tobi, M., M. Memon, K. Kithier, and N. Clapp. 1994. A putative CEA moiety is shared by the cotton-top tamarin (Saguinus oedipus) and humans. Cancer Lett. 77:7-13. Turbide, C., T. Kunath, E. Daniels, and N. Beauchemin. 1997. Optimal ratios of biliary glycoprotein isoforms required for inhibition of colonic tumor cell growth. Cancer Res. 57:2781-8. Tykocinski, M.L., H.K. Shu, D.J. Ayers, E.I. Walter, R.R. Getty, R.K. Groger, C.A. Hauer, and M.E. Medof. 1988. Glycolipid reanchoring of T-lymphocyte surface antigen CD8 using the 3' end sequence of decay-accelerating factor's mRNA. Proc Natl Acad Sci U S A. 85:3555-9. Ugrinska, A., E. Bombardieri, M.P. Stokkel, F. Crippa, and E.K. Pauwels. 2002. Circulating tumor markers and nuclear medicine imaging modalities: breast, prostate and ovarian cancer. Q J Nucl Med. 46:88-104. Vainauskas, S., and A.K. Menon. 2004. A conserved proline in the last transmembrane segment of Gaa1 is required for glycosylphosphatidylinositol (GPI) recognition by GPI transamidase. J Biol Chem. 279:6540-5. van den Berg, C.W., T. Cinek, M.B. Hallett, V. Horejsi, and B.P. Morgan. 1995a. Exogenous CD59 incorporated into U937 cells through its glycosyl phosphatidylinositol anchor becomes associated with signalling molecules in a time dependent manner. Biochem Soc Trans. 23:269S.

191 van den Berg, C.W., T. Cinek, M.B. Hallett, V. Horejsi, and B.P. Morgan. 1995b. Exogenous glycosyl phosphatidylinositol-anchored CD59 associates with kinases in membrane clusters on U937 cells and becomes Ca(2+)-signaling competent. J Cell Biol. 131:669-77. van der Goot, F.G., and T. Harder. 2001. Raft membrane domains: from a liquid-ordered membrane phase to a site of pathogen attack. Semin Immunol. 13:89-97. van Meer, G. 1989. Lipid traffic in animal cells. Annu Rev Cell Biol. 5:247-75. van Meer, G., and K. Simons. 1982. Viruses budding from either the apical or the basolateral plasma membrane domain of MDCK cells have unique phospholipid compositions. Embo J. 1:847-52. Varma, R., and S. Mayor. 1998. GPI-anchored proteins are organized in submicron domains at the cell surface. Nature. 394:798-801. Veingerl, B. 2001. Serum carcinoembryonic antigen levels in patients operated for colorectal carcinoma. Wien Klin Wochenschr. 113 Suppl 3:32-8. Vereb, G., J. Szollosi, J. Matko, P. Nagy, T. Farkas, L. Vigh, L. Matyus, T.A. Waldmann, and S. Damjanovich. 2003. Dynamic, yet structured: The cell membrane three decades after the Singer-Nicolson model. Proc Natl Acad Sci U S A. 100:8053-8. Vey, M., S. Pilkuhn, H. Wille, R. Nixon, S.J. DeArmond, E.J. Smart, R.G. Anderson, A. Taraboulos, and S.B. Prusiner. 1996. Subcellular colocalization of the cellular and scrapie prion proteins in caveolae-like membranous domains. Proc Natl Acad Sci U S A. 93:14945-9. Viola, A., S. Schroeder, Y. Sakakibara, and A. Lanzavecchia. 1999. T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science. 283:680-2. Virji, M., D. Evans, J. Griffith, D. Hill, L. Serino, A. Hadfield, and S.M. Watt. 2000. Carcinoembryonic antigens are targeted by diverse strains of typable and non-typable Haemophilus influenzae. Mol Microbiol. 36:784-95. Virji, M., K. Makepeace, D.J. Ferguson, and S.M. Watt. 1996. Carcinoembryonic antigens (CD66) on epithelial cells and neutrophils are receptors for Opa proteins of pathogenic neisseriae. Mol Microbiol. 22:941-50. Vogel, U., K. Sandvig, and B. van Deurs. 1998. Expression of caveolin-1 and polarized formation of invaginated caveolae in Caco-2 and MDCK II cells. J Cell Sci. 111 ( Pt 6):825-32. von Haller, P.D., S. Donohoe, D.R. Goodlett, R. Aebersold, and J.D. Watts. 2001. Mass spectrometric characterization of proteins extracted from Jurkat T cell detergent-resistant membrane domains. Proteomics. 1:1010-21. Vyas, K.A., H.V. Patel, A.A. Vyas, and R.L. Schnaar. 2001. Segregation of gangliosides GM1 and GD3 on cell membranes, isolated membrane rafts, and defined supported lipid monolayers. Biol Chem. 382:241-50. Wagener, C., and S. Ergun. 2000. Angiogenic properties of the carcinoembryonic antigen-related cell adhesion molecule 1. Exp Cell Res. 261:19-24. Wagener, C., F. Hain, H.J. Fodisch, and H. Breuer. 1983. Localisation of carcinoembryonic antigen in embryonic and fetal human tissues. Histochemistry. 78:1-9. Wanebo, H.J., M. Llaneras, T. Martin, and D. Kaiser. 1989. Prospective monitoring trial for carcinoma of colon and rectum after surgical resection. Surg Gynecol Obstet. 169:479-87. Wanebo, H.J., B. Rao, C.M. Pinsky, R.G. Hoffman, M. Stearns, M.K. Schwartz, and H.F. Oettgen. 1978. Preoperative carcinoembryonic antigen level as a prognostic indicator in colorectal cancer. N Engl J Med. 299:448-51.

192 Wang, F., R.K. Hansen, D. Radisky, T. Yoneda, M.H. Barcellos-Hoff, O.W. Petersen, E.A. Turley, and M.J. Bissell. 2002a. Phenotypic reversion or death of cancer cells by altering signaling pathways in three-dimensional contexts. J Natl Cancer Inst. 94:1494-503. Wang, J., W. Gunning, K.M. Kelley, and M. Ratnam. 2002b. Evidence for segregation of heterologous GPI-anchored proteins into separate lipid rafts within the plasma membrane. J Membr Biol. 189:35-43. Wang, J., K. Maziarz, and M. Ratnam. 1999. Recognition of the carboxyl-terminal signal for GPI modification requires translocation of its hydrophobic domain across the ER membrane. J Mol Biol. 286:1303-10. Wang, X., G. Jansen, J. Fan, W.J. Kohler, J.F. Ross, J. Schornagel, and M. Ratnam. 1996. Variant GPI structure in relation to membrane-associated functions of a murine folate receptor. Biochemistry. 35:16305-12. Wang, X.Q., P. Sun, and A.S. Paller. 2002c. Ganglioside induces caveolin-1 redistribution and interaction with the epidermal growth factor receptor. J Biol Chem. 277:47028-34. Watt, F.M. 2002. Role of integrins in regulating epidermal adhesion, growth and differentiation. Embo J. 21:3919-26. Watt, S.M., J. Fawcett, S.J. Murdoch, A.M. Teixeira, S.E. Gschmeissner, N.M. Hajibagheri, and D.L. Simmons. 1994. CD66 identifies the biliary glycoprotein (BGP) adhesion molecule: cloning, expression, and adhesion functions of the BGPc splice variant. Blood. 84:200- 10. Wei, Y., R.P. Czekay, L. Robillard, M.C. Kugler, F. Zhang, K.K. Kim, J.P. Xiong, M.J. Humphries, and H.A. Chapman. 2005. Regulation of alpha5beta1 integrin conformation and function by urokinase receptor binding. J Cell Biol. 168:501-11. Wei, Y., J.A. Eble, Z. Wang, J.A. Kreidberg, and H.A. Chapman. 2001. Urokinase receptors promote beta1 integrin function through interactions with integrin alpha3beta1. Mol Biol Cell. 12:2975-86. Williams, A.F., and A.N. Barclay. 1988. The immunoglobulin superfamily--domains for cell surface recognition. Annu Rev Immunol. 6:381-405. Wilson, B.S., S.L. Steinberg, K. Liederman, J.R. Pfeiffer, Z. Surviladze, J. Zhang, L.E. Samelson, L.H. Yang, P.G. Kotula, and J.M. Oliver. 2004. Markers for detergent- resistant lipid rafts occupy distinct and dynamic domains in native membranes. Mol Biol Cell. 15:2580-92. Wirth, T., E. Soeth, F. Czubayko, and H. Juhl. 2002. Inhibition of endogenous carcinoembryonic antigen (CEA) increases the apoptotic rate of colon cancer cells and inhibits metastatic tumor growth. Clin Exp Metastasis. 19:155-60. Wu, G., Z. Guo, A. Chatterjee, X. Huang, E. Rubin, F. Wu, E. Mambo, X. Chang, M. Osada, M. Sook Kim, C. Moon, J.A. Califano, E.A. Ratovitski, S.M. Gollin, S. Sukumar, D. Sidransky, and B. Trink. 2006. Overexpression of glycosylphosphatidylinositol (GPI) transamidase subunits phosphatidylinositol glycan class T and/or GPI anchor attachment 1 induces tumorigenesis and contributes to invasion in human breast cancer. Cancer Res. 66:9829-36. Xiang, Y.Y., V. Ladeda, and J. Filmus. 2001. Glypican-3 expression is silenced in human breast cancer. Oncogene. 20:7408-12. Xue, W., I. Mizukami, R.F. Todd, 3rd, and H.R. Petty. 1997. Urokinase-type plasminogen activator receptors associate with beta1 and beta3 integrins of fibrosarcoma cells: dependence on extracellular matrix components. Cancer Res. 57:1682-9.

193 Yaffe, D. 1968. Retention of differentiation potentialities during prolonged cultivation of myogenic cells. Proc Natl Acad Sci U S A. 61:477-83. Yamabhai, M., and R.G. Anderson. 2002. Second cysteine-rich region of epidermal growth factor receptor contains targeting information for caveolae/rafts. J Biol Chem. 277:24843- 6. Yan, W., and M. Ratnam. 1995. Preferred sites of glycosylphosphatidylinositol modification in folate receptors and constraints in the primary structure of the hydrophobic portion of the signal. Biochemistry. 34:14594-600. Yan, W., F. Shen, B. Dillon, and M. Ratnam. 1998. The hydrophobic domains in the carboxyl- terminal signal for GPI modification and in the amino-terminal leader peptide have similar structural requirements. J Mol Biol. 275:25-33. Yang, X.H., O.V. Kovalenko, T.V. Kolesnikova, M.M. Andzelm, E. Rubinstein, J.L. Strominger, and M.E. Hemler. 2006. Contrasting effects of EWI proteins, integrins, and protein palmitoylation on cell surface CD9 organization. J Biol Chem. 281:12976-85. Yeatman, T.J., W. Mao, and R.C. Karl. 1997. Biliary glycoprotein is overexpressed in human colon cancer cells with high metastatic potential. J Gastrointest Surg. 1:292-8. Yoshida, K., S. Ueno, T. Iwao, S. Yamasaki, A. Tsuchida, K. Ohmine, R. Ohki, Y.L. Choi, K. Koinuma, T. Wada, J. Ota, Y. Yamashita, K. Chayama, K. Sato, and H. Mano. 2003. Screening of genes specifically activated in the pancreatic juice ductal cells from the patients with pancreatic ductal carcinoma. Cancer Sci. 94:263-70. Zacharias, D.A., J.D. Violin, A.C. Newton, and R.Y. Tsien. 2002. Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells. Science. 296:913-6. Zahir, N., and V.M. Weaver. 2004. Death in the third dimension: apoptosis regulation and tissue architecture. Curr Opin Genet Dev. 14:71-80. Zajchowski, L.D., and S.M. Robbins. 2002. Lipid rafts and little caves. Compartmentalized signalling in membrane microdomains. Eur J Biochem. 269:737-52. Zebhauser, R., R. Kammerer, A. Eisenried, A. McLellan, T. Moore, and W. Zimmermann. 2005. Identification of a novel group of evolutionarily conserved members within the rapidly diverging murine Cea family. Genomics. 86:566-80. Zheng, G., A. Chen, R.E. Sterner, P.J. Zhang, T. Pan, N. Kiyatkin, and M.L. Tykocinski. 2001. Induction of antitumor immunity via intratumoral tetra-costimulator protein transfer. Cancer Res. 61:8127-34. Zhou, G.Q., V. Baranov, W. Zimmermann, F. Grunert, B. Erhard, L. Mincheva-Nilsson, S. Hammarstrom, and J. Thompson. 1997. Highly specific monoclonal antibody demonstrates that pregnancy-specific glycoprotein (PSG) is limited to syncytiotrophoblast in human early and term placenta. Placenta. 18:491-501. Zhou, G.Q., and S. Hammarstrom. 2001. Pregnancy-specific glycoprotein (PSG) in baboon (Papio hamadryas): family size, domain structure, and prediction of a functional region in primate PSGs. Biol Reprod. 64:90-9. Zhou, G.Q., Y. Zhang, and S. Hammarstrom. 2001. The carcinoembryonic antigen (CEA) gene family in non-human primates. Gene. 264:105-12. Zhou, H., A. Fuks, G. Alcaraz, T.J. Bolling, and C.P. Stanners. 1993a. Homophilic adhesion between Ig superfamily carcinoembryonic antigen molecules involves double reciprocal bonds. J Cell Biol. 122:951-60. Zhou, H., A. Fuks, and C.P. Stanners. 1990. Specificity of intercellular adhesion mediated by various members of the immunoglobulin supergene family. Cell Growth Differ. 1:209-15.

194 Zhou, H., C.P. Stanners, and A. Fuks. 1993b. Specificity of anti-carcinoembryonic antigen monoclonal antibodies and their effects on CEA-mediated adhesion. Cancer Res. 53:3817-22. Zimmermann, W. 1998. The nature and expression of the rodent CEA families: evolutionary considerations. In Cell Adhesion and Communication Mediated by the CEA Family. Vol. 5. C.P. Stanners, editor. Harwood Academic Publishers, Amsterdam. 31-55.

195

Appendices

196