Implications of alpha- glycosylation in the proper assembly of the dystroglycan associated protein complex and the polarized distribution of the inwardly rectifying potassium channel, Kir4.1, and the water permeable channel, AQP4, in perivascular glia

by

Jennifer Mary Elizabeth Rurak

B.Sc. Honours Physiology, University of British Columbia, 2004

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

in

The Faculty of Graduate Studies

(Neuroscience)

THE UNIVERSITY OF BRITISH COLUMBIA

March 2007

© Jennifer Mary Elizabeth Rurak, 2007 ABSTRACT

The dystroglycan protein complex spans the plasma membrane and provides a link between the cytoskeleton and the extracellular matrix (ECM). Defective O-linked glycosylation of a-dystroglycan (a-DG) within the complex severs this link leading to a subset of muscular dystrophies known as the dystroglycanopathies. These are characterized by muscle weakness and degeneration as well as by brain and ocular defects of unknown etiology. In brain and retina, a-DG and ECM molecules are enriched at astrocytic domains abutting blood vessels where they may be involved in localizing the inwardly rectifying potassium channel, Kir4.1, and aquaporin channel, AQP4, to astrocytic endfeet. To investigate the role of ECM ligand-binding to glycosylated sites on a-DG in the polarized distribution of these channels in vivo, we used the Largemyd mouse, an accepted animal model for dystroglycanopathies. We found that Kir4.1 and AQP4 are lost from astrocytic endfeet in brain while significant labeling for these channels is still detected at analogous cell domains in retina. Furthermore, while both a- and b\- are lost from perivascular astrocytes in brain of the Largemyd mouse, labeling for pi- is still evident at parallel domains in retina. These findings show that while ligand-binding to glycosylated sites on a-DG in concert with a- and pVsyntrophins is crucial for the polarized distribution of Kir4.1 and AQP4 to functional domains in brain, distinct mechanisms may contribute to their localization in retina.

ii TABLE OF CONTENTS

ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES vi LIST OF ABBREVIATIONS vii ACKNOWLEDGEMENTS viii DEDICATION ix CO-AUTHORSHIP STATEMENT x 1 INTRODUCTION 1 1.1 Muscular Dystrophy 1 1.2 The Dystroglycan-Associated Protein Complex 3 1.3 5 1.4 Utrophin 7 1.5 The Syntrophins 9 1.6 11 1.7 The - Complex 12 1.8 The 13 1.9 O-linked glycosylation of a-dystroglycan 15 1.10 Alpha-Dystroglycanopathy Related Muscular Dystrophies 19 1.11 The Dystroglycan Associated Protein Complex in the Central Nervous System 26 1.12 Inwardly Rectifying Potassium Channels (Kir) 31 1.13 Kir4.1 in brain and retina 32 1.14 Aquaporins 36 1.15 Aquaporin 4 in brain and retina 37 1.16 The Largemyd mouse 39 1.17 Project Outline and Hypotheses..'. 42 1.18 Works Cited 44 2 FIRST MANUSCRIPT CHAPTER 58 2.1 Alpha-dystroglycan hypoglycosylation disrupts the polarized distribution of potassium ion and water permeable channels in perivascular astrocytes of the Largemyd mouse 58 2.2 Materials and Methods 62 2.2.1 Tissues 62 2.2.2 Antibodies 64 2.2.3 Immunofluorescence 65 2.2.4 Immunoblotting 65 2.2.5 RNA extraction and RT-PCR 66 2.3 Results 67 2.3.1 Abnormalities in the localization of basement membrane proteins in the Largemyd brain 67 2.3.2 Loss of perivascular localization of AQP4 and Kir4.1 in the Largemyd brain 71 2.3.3 Loss of perivascular localization of a- and p,-syntrophins in the Largemyd brain 76 2.3.4 Expression of Large 1 and Large2 mRNAs in retina and impact of Large 1 mutation on Kir4.1 and AQP4 localization in the Largemyd retina 78 myd 2.3.5 Loss of perivascular localization of a-syntrophin but not prsyntrophin in the Large retina 85 2.4 Discussion 87 2.4.1 AQP4 and Kir4.1 redistribution in brain 87 2.4.2 AQP4 and Kir4.1 distribution in retina 90 2.5 Acknowledgements 92 2.6 Works Cited 93 3 DISCUSSION 98 3.1 Implication of the DAP complex in ion channel localization 98 3.2 Stabilization of AQP4 and Kir4.1 via a-syntrophin 103 3.3 Impact of ct-dystroglycan glycosylation on AQP4 and Kir4.1 expression 105 3.4 Possible physiological implications of AQP4 and Kir4.1 mislocalization 107 3.5 Disruptions in the cortical surface basal lamina 109 3.6 Expression of Large transcripts and glycosylation of a-dystroglycan in retina 111 3.7 Spatial differences in the DAP complex-induced anchoring of AQP4 and Kir4.1 112 3.8 Disruption of the inner limiting membrane and aberrant retinal lamination 117 3.9 Conclusions 117 3.10 Future Directions 118 3.10.1 Involvement of a-DG glycosylation in the laminin-induced clustering of Kir4.1 and AQP4 118 3.10.2 Effect of the Largemyd mutation on other retinal DAP complex proteins 119 3.10.3 Expression of integrins in the retina 120 3.10.4 Pathological implications of AQP4 and Kir4.1 mislocalization in Largemyd brain 121 3.11 Works Cited 123 4 APPENDICES 131 4.1 Appendix A-Supplemental Data 131

iv LIST OF TABLES

Table 1.1 Summary of sugar-peptide bonds in eukaryotes 16 Table 1.2 Summary of a-DG Hypoglycosylation Related Muscular Dystrophies 24 Table 2.1 Summary of Experimental Animals: Sex, Genotype and Age 63

v LIST OF FIGURES

Figure 1.1 Schematic Representation of Molecular Interaction Within the DAP Complex Figure 1.2 O-Mannosyl Glycan Synthesis of a-Dystroglycan Figure 1.3 Schematic of Perivascular Astrocytic Domains Figure 1.4 Schematic of Mammalian Retinal Lamination Figure 1.5 Schematic model of spatial pottasium buffering Figure 2.1 Disruption of the glia limitans and accumulation of agrin and perlecan puncta in the Largemyd brain Figure 2.2 AQP4 and Kir4.1 fail to localize to perivascular astrocyte endfeet in the Largemyd brain Figure 2.3 Total Kir4.1 and AQP4 expression levels are unchanged in the Largemyd brain Figure 2.4 a- and pVsyntrophins fail to localize to perivascular astrocyte endfeet in the Largemyd brain Figure 2.5 Expression profile of Large 1 and Large2 Figure 2.6 a-dystroglycan is hypoglycosylated in the Largemyd retina Figure 2.7 Disruption of the lamination and the inner limiting membrane in the Largemyd retina Figure 2.8 Perivascular AQP4 and Kir4.1 are maintained in the Largemyd retina Figure 2.9 Pi- but not a-syntrophin is still expressed at perivascular astrocytes endfeet in the Largemyd retina Figure 4.1 AQP4 and Kir4.1 fail to localize to the glia limitans in the Largemyd brain. 1 Figure 4.2 Radial Miiller cells appear discontinuous and tortuous in the Largemyd retina. 1 LIST OF ABBREVIATIONS

AAV - Adeno Associated Virus LGMD - Limb Girdle MD AchR - Acetylcholine Receptor mAb - Monoclonal Antibody AQP - Aquaporin Man - Mannose Ara - Arabinose MAST - Microtubule Associated Ser/ Asn - Asparagine Thr Kinase CMD - Congenital Muscular Dystrophy MD - Muscular Dystrophy CNS - Central Nervous System MDC - Congenital MD CSF - Cerebrospinal Fluid Mdx - Dystrophin Deficient Mouse DAP - Dystroglycan Associated Protein MEB - Muscle Eye Brain Disease

DB -Dystrobrevin Nav- Voltage Gated Sodium Channel DG - Dystroglycan (a or P) NeuAc - Neuraminic Acid DMD -Duchenne Muscular Dystrophy NMJ - Dol-P-X - Dolichol-Phosphate- nNOS - Neuronal Nitric Oxide Synthase Oligosaccarides OPL - Outer Plexiform Layer Dp - Dystrophin PBS - Phosphate Buffered Saline ECM - Extracellular Matrix PH - Plekstrin Homology Domain EHS - Engelbreth-Holm-Swarm POMGnTl - Protein O-mannose pi,2- ENU - N-ethylnitrosourea N-acetylglucosaminyltransferase ER - Endoplasmic Reticulum POMT - protein O-mannosyltransferase FCMD - Fukuyama CMD (1 or 2) FKRP - Fukutin Related Protein RMP - Resting Membrane Potential Fuc - Fucose SAST - Syntrophin Associated Ser/ Fugu - Takifugu rubripes (puffer fish) Thr Kinase GABA - Gamma aminobutyric Acid Ser - Serine Gal - Galactose SG - Sarcoglycan GalNAc - N-acetyl- galactosamine Sia - Sialic Acid Glc - Glucose SNV sequence - GlcNAc - N-acetylglucosamine SSC - Sarcoglycan Sarcospan Complex HSPG - Heparan Sulfate Proteoglycan SSV sequence- Hyl - Hydroxylysine SU - Syntrophin Unique Domain Hyp - Hydroxyproline Syn - Syntrophin IPL - Inner Plexiform Layer Thr - Threonine + [K ]0 - Extracellular potassium Tyr - Tyrosine concentration w/v - weight per volume Kir - Inwardly Rectifying Potassium W WS - Walker Warburg Syndrome Channel Xyl - Xylose

vii ACKNOWLEDGEMENTS

This thesis, from its nascence through to its completion, has been shaped and molded by several instrumental and talented researchers.

Many, many thanks to:

Dr. Hakima Moukhles, my primary investigator, whose insight and patience gave rise to this project, whose support helped it puzzle together into a linear idea and whose valuable criticism gave rise to its completion.

My motley crew of lab mates: Daniel (R2) Tham, Geoffroy (where are you from again?) Noel, Leona Lui and Natalie Tarn. You laughed when I laughed and you laughed when I cried.

Dr. S. Carbonetto and Nadia Melan, University of Montreal, whose generous donation of Largemyd brain and eye tissue allowed this project to be realized.

Dr. S. Froehner, University of Washington, for generously providing antibodies directed against syntrophins.

vin DEDICATION

For my mother and father Scientists through and through

and for Braedon, the orchard of my eye. CO-AUTHORSHIP STATEMENT

All components of this manuscript were produced by its primary author, Jennifer Rurak, with the following exceptions:

The immunofluoresence experiments and imaging in Figures 2.1 and 2.2 were carried out by Leona Lui.

The RT-PCR data in Figure 2.5 were generated by Bharat Joshi.

All experiments were carried out under the supervision and guidance of Dr. Hakima Moukhles.

x 1 INTRODUCTION

1.1 Muscular Dystrophy

The term muscular dystrophy embodies a myriad of heterogeneous, inheritable diseases characterized by progressive muscle weakness and deterioration. One of the most common, and therefore most thoroughly studied, forms is Duchenne Muscular

Dystrophy (DMD) which, like all forms of the disease, is a devastating illness with a large physical and socioeconomic burden to patients and caretakers alike. Being an X- linked inherited defect, this particular form affects approximately 1 in 3500 live born males and presents symptomatically around the third year of life. Muscle weakness and wasting is non-neurological in origin and is first evident in the pelvic and shoulder girdles, radiating gradually toward the extremities with muscles of fine movement, such as those of the hand, being the last to fail. Sufferers are often confined to a wheelchair before the end of their second decade and death typically ensues during their thirties, most frequently from cardiac failure or respiratory complications such as pneumonia due to intercostal muscle and diaphragm insufficiency (Emery 2001; Emery 2002). There is currently no cure for muscular dystrophy and therapeutic approaches are typically focused on respiratory support and management of cardiac complications (Emery 2002).

DMD represents one facet of a large number of distinct categories of muscular dystrophies. Interestingly, despite the variability in molecular deficits, presentation and severity among the diseases, there remains at least one unifying pathologic finding: All

1 forms exhibit a chronic degeneration and regeneration of as seen on muscle biopsy which is associated with fibrosis of the and fatty infiltration. This is known as pseudo-hypertrophy and is most prominent in the muscles of the leg.

Histologically, muscles display scattered bundles of necrosis, fibres of variable size and shape and centralized nuclei. In addition, several forms of the disease are also confounded by cognitive and ocular defects of varying severity that, unlike their muscular counterparts, are stable in nature (Emery 2001). These range from mild learning disabilities to severe mental retardation (Emery 2001). Furthermore, epilepsy and a reduction of the electroretinogram b-wave have both been described, insinuating electrophysiological abnormalities in the brain and retina as well (Emery 2001).

Histological analysis reveals aberrant neuronal migration and insufficient myelin ensheathing throughout the central nervous system (Emery 2002).

In 1987, the DMD product, named dystrophin, was cloned and mapped within the short arm of the X at position Xp21 (Koenig et al., 1987).

Recently, positional cloning techniques have led to the identification of nearly 30 , mutations in which are associated with various forms of muscular dystrophy (reviewed by Cohn and Campbell, 2000) thus sparking a huge interest in understanding the underlying pathophysiology of these diseases in general.

2 1.2 The Dystroglycan-Associated Protein Complex

In , dystrophin is a component of a large multiprotein complex known as the dystroglycan associated protein complex (DAP). The dystroglycans (DG) consist of two subunits, a-DG and P-DG, that are located at the most extracellular aspect of the complex with P-DG spanning the plasma membrane and interacting with the extracellular a-DG. Intracellularly, P-DG interacts with the subplasmalemmal dystrophin (Fig. 1.1).

Several other proteins are associated with the complex and their expression varies between tissues, species and developmental stages. These include the transmembrane sarcospan and , and the intracellular utrophin, syntrophins, dystrobrevins, , caveolin-3, Grb2 and nNOS (reviewed by Ehmsen et al., 2002). It is now generally accepted that the complex serves to form a physical bridge connecting the cell's cytoskeleton to the encompassing basal lamina by way of dystrophin binding to

(Way et al., 1992) and a-DG binding to extracellular matrix proteins (ECM) such as laminin, perlecan and agrin (Ibraghimov-Beskrovnaya et al., 1992; Bowe et al., 1994;

Talts et al., 1999). In muscle, it is thought that this mechanical link offers a degree of protection to the muscles by evenly distributing the forces generated from repeated cycles of contraction and relaxation throughout the via dispersion across the basal lamina, thus preventing the from focal damage (Zubrzycka-Gaarn et al.,

1988; Petrofetal., 1993).

3 Basal Lamina 1 r

T2

fyntrophins Cys-Rich 'Rod Domain Domain

F-actin

Figure 1.1 Schematic Representation of Molecular Interactions Within the DAP Complex P-DG spans the plasma membrane and interacts with the extracellular a-DG. Intracellular^ the PPXY motif of P-DG binds to the WW domain of dystrophin's C- terminus. Several other proteins are associated with the DAP as outlined in the figure. Laminin binds to the O-linked glycosylated epitopes of a-DG by way of globular domains 4 and 5 in the C-terminal a-chain. Dystrophin binds F-actin through its amino and rod-like domains. Arrow highlights the O-linked glycosylated moities of a-DG.

4 In DMD, a missense mutation in dystrophin prevents its interaction with P-DG thus breaching the continuity of the complex as a whole. However, mutations in a number of different components of the DAP complex have been correlated with various forms of muscular dystrophy, as have post-translational disruptions of a-DG and mutations within proteins of the ECM itself (reviewed by Kanagawa and Toda, 2006). The resultant loss of the trans-sarcolemmal connection between the cytoskeleton and the ECM is believed to be the primary deficit underlying the pathogenesis of the disease in muscle. The dystrophic sarcolemma is friable and repeated cycles of degeneration and regeneration in the face of contractile stresses eventually result in muscle necrosis, infiltration of fatty and connective tissues and muscle wasting. In order to appreciate the molecular details of these diseases, a thorough understanding of the DAP complex is required (Emery 2001).

1.3 Dystrophin

The large, 427kDa protein dystrophin is located at the cytoplasmic face of the sarcolemma (Fig. 1.1). Mutations within the dystrophin gene lead to some forms of muscular dystrophy including DMD (Emery 2001). It contains 3685 amino acids, shares homology with spectrin and a-actinin and is divided into four discreet functional domains. These are the amino terminal domain, a large rod-like domain containing 24 triple helical spectrin-like repeats interspersed with 4 presumed hinge domains, a cystein- rich domain and a carboxyl terminal (Fig. 1.1) that shares homology with the closely related utrophin and dystrobrevins.

5 Dystrophin binds F-actin through its amino and rod-like domains and binds P-DG through its WW, cystein rich and carboxyl terminal domains making this protein the primary intracellular anchor of the DAP complex (Fig. 1.1; reviewed by Winder et al.,

1995; Huang et al., 2000; Barresi and Campbell, 2006). The dystrophin gene contains three promoters that regulate its full length expression as well as four internal promoters that prompt the expression of serially truncated proteins in tissues other than muscle.

These are Dp260, Dpl40, Dpi 16 and Dp71 which are expressed in the retina, the brain, the peripheral nervous system and variably throughout other non-muscle tissues respectively (Hugnot et al., 1992; Byers et al., 1993; D'Souza et al., 1995; Lidov et al.,

1995). The mdx mouse is the leading animal model for the study of DMD as it completely lacks full length dystrophin and presents with degenerative changes similar to those seen in DMD albeit to a milder extent (Tanabe et al., 1986). Examination of the mdx muscle further reveals that loss of dystrophin compromises the stability of the entire complex as reflected in a notable reduction of several other DAP complex components including a-DG, P-DG, a-Sarcoglycan, y-Sarcoglycan and syntrophin (Ohlendieck and

Campbell, 1991). There exist four other mouse models for muscular dystrophy. These are the mdx2cv"5cv mice that were generated by N-ethylnitrosourea (ENU) chemical mutagenesis and which, in addition to full length dystrophin, lack the shorter isoforms of dystrophin to a variable extent (Chapman et al., 1989; Danko et al., 1992; Howard et al.,

1998). Again, they all present with myopathies analogous to DMD yet their phenotypes are less severe. The basis of the variability in the manifestation of dystrophin deficiencies between individuals and species is unknown (lm et al., 1996).

6 Viral-mediated gene delivery has been addressed as a therapeutic approach for those muscular dystrophies caused by mutations in the dystrophin gene, however, the adeno-associated virus (AAV), commonly used as a delivery vector, has a limited packaging allowance of 4.5 kb which precludes its use in the delivery of the 14 kb cDNA for dystrophin (Yue et al., 2006). Not all regions of dystrophin are equally important for its function and, as such, functionally truncated mutants lacking regions of the rod-like domain have been generated and proven effective in the treatment of animal models of

DMD (Gregorevic et al., 2004). This technology is still being considered as a potential therapeutic approach for these diseases in humans.

1.4 Utrophin

Utrophin shares considerable homology with dystrophin making it the focus of much interest as a possible therapeutic approach in the treatment of dystrophin deficient forms of muscular dystrophy. It contains 3433 amino acids and has a molecular mass of

395 kDa (Tinsley et al., 1992). The F-actin binding amino terminal domain, along with the cystein rich and C-terminal domain, shares an 80% homology with the analogous domains of dystrophin (Tinsley et al., 1992). The rod-like domain is the least conserved with only 35% homology between dystrophin and utrophin (Tinsley et al., 1992). In addition to the between these two proteins, there is also similarity in their binding partners in muscle (Matsamura et al., 1992) including actin such that their

7 respective affinities for the same binding partners makes them functionally interchangeable proteins (Rybakova et al., 2002). Furthermore, utrophin has two reported full length isoforms as well as three described short gene products representing C- terminal isoforms, much like dystrophin, that are expressed in non-muscle tissues

(Dennis et al., 1996). Utrophin is found ubiquitously yet its subcellular distribution is not uniform. In skeletal muscle, where dystrophin is richly localized at the sarcolemma as part of the DAP complex, utrophin is expressed in intramuscular vessels and nerves as well as at the crests of the neuromuscular junction (NMJ) folds (reviewed by Perkins and

Davies, 2002). In brain, where dystrophin is expressed predominantly at the astrocytic endfeet surrounding blood vessels, utrophin is found as part of the DAP complex in the vascular endothelium (Knuesel et al., 2000). Furthermore, while both proteins are found within neurons, they are not known to co-localize within the same cells. In short, the expression patterns of the two proteins throughout the body are almost entirely reciprocal suggesting complimentary functions between the two (Knuesel et al., 2000). The utrophin null mouse shows only mild phenotypic changes that are primarily characterized by a reduction of acetylcholine receptors (AChR) and postsynaptic folds at the NMJ with no signs of dystrophy (Deconinck et al., 1997a). Given the similarly mild phenotype of the mdx mice, this suggests that dystrophin and utrophin are capable of compensating for the selective loss of the other. In support of this, the utrophin-dystrophin double knock out mouse has a severe myopathy that is lethal within 20 weeks of birth and overlaps more consistently with the phenotypes of human DMD (Deconinck et al., 1997b; Grady et al.,

1997). Interestingly, it has been shown that overexpression of utrophin rescues the

8 dystrophin deficient mdx mouse phenotype with no reported toxicity indicating that utrophin might be a potential therapeutic resource for DMD (Tinsley et al., 1998).

1.5 The Syntrophins

Syntrophins are modular adaptor proteins that bind directly to components of the

DAP complex, namely dystrophin, dystrobrevin and utrophin, where they serve to localize signaling molecules to specific cell domains. To date five syntrophins have been characterized, (a-, Pi-, P2- and yi- and Y2-syn), all of which are similar in structure

(Adams et al., 2000). Syntrophins contain two pleckstrin homology (PHI and PH2) domains, the first of which is split and contains a PDZ domain, and a syntrophin unique

(SU) domain at the carboxyl terminus that is conserved among all isoforms suggesting that these proteins are capable of interacting simultaneously with multiple other proteins

(Blake et al., 2002). The SU in combination with the PH2 domain is required for syntrophin's interaction with dystrobrevin, utrophin and dystrophin (Kachinsky et al.,

1999) while the PDZ domain binds numerous signaling molecules including nNOS

(Brenman et al., 1996), microtubule-associated serine/threonine kinase (MAST) and syntrophin-associated serine/threonine kinase (SAST) (Lumeng et al., 1999) as well as a collection of ion channels including alD adrenergic receptors (Chen et al., 2006), the

voltage-gated sodium channels Nav1.4 and 1.5 (Schultz et al., 1998), the inwardly rectifying potassium channels Kir2.1 (Leonoudakis et al., 2004) and Kir4.1 (Connors et al., 2004) and the water channel aquaporin-4 (AQP4) (Neely et al., 2001). While all of the syntrophin isoforms are capable of binding these proteins in vitro, it has not been

9 proven that they do so in vivo. In fact, there is evidence to indicate that each syntrophin has a preferred set of PDZ binding ligands (Adams et al., 2001).

Despite their structural similarities, each syntrophin boasts a unique distribution pattern, a-syn is predominantly expressed in skeletal muscle where it is enriched at the postsynaptic membrane of the NMJ (Adams et al., 2000) and in (Williams et al., 2006). A lower expression level of this isoform is found in both brain and kidney

(Alessi et al., 2006). Pi-syn, on the other hand, is ubiquitously expressed in a multitude of tissues with the highest levels found in skeletal and smooth muscles as well as liver and kidney while P2-syn is enriched in testis, brain, liver, kidney, intestinal and, to a lesser extent, in skeletal muscle (Alessi et al., 2006). yi-syn is found predominantly in a subset of neurons throughout the brain including hippocampal pyramidal cells, cortical neurons and cerebellar Purkinje neurons while y2-syn is predominantly localized to skeletal muscle. However, while the a- and P-syntrophins are

found at the sarcolemma of muscle, y2-syn is associated with the and does not overlap with the expression of a or P isoforms (Alessi et al., 2006).

Mice mutant for a-syn have no overt signs of muscular dystrophy and show normal of the muscle (Adams et al., 2000). At the molecular level, they show a notable reduction in utrophin as well as AchRs and postsynaptic folds at the NMJ

(Adams et al., 2000). Interestingly, the similarities between the NMJ phenotypes of the a- syn null and utrophin null mice suggest that the defects could be secondary to the loss of utrophin rather than directly related to the loss of syntrophin (Adams et al., 2000). The a-

10 syn null mouse also exhibits a severe disruption of the polarized distribution of AQP4 in brain (Amiry-Moghaddam et al., 2003; Amiry-Moghaddam et al., 2004) and retina

(Puwarawuttipanit et al., 2006). This will be discussed in detail in Chapter 3.

1.6 Dystrobrevin

Dystrobrevin (DB) is a dystrophin-related and -associated protein that is part of the DAP complex. Two separate genes encode the family of the mammalian DBs. The a-

DB gene encodes three major isoforms, a-DB-1, -2 and -3 (Ambrose et al., 1997) which are further multiplied by the use of three different promoters allowing for variable tissue specific expression (Holzfeind et al., 1999). All a-DB isoforms are present at the sarcolemma of normal muscle as well as concentrated at the NMJ (Nawrotzki et al.,

1998). a-DB is also found throughout the brain (Grady et al., 2006). Another gene encoding P-DB is expressed in non-muscle tissues including brain (Blake et al., 1998;

Blake et al., 1999) and retina where it colocalizes with Dp260 (Blank et al., 2002). Both a- and P-DB colocalize with and bind to dystrophin, utrophin and members of the syntrophin family (Blake et al., 1996; Blake et al., 1999). In the mdx mouse, DB is lost from the extrajunctional regions but maintained at the NMJ suggesting that its anchoring mechanism at this site differs from that existing extrajunctionally, perhaps binding to utrophin instead of dystrophin (Blake et al., 2002).

Animal models lacking a-DB 1 and 2 show a mild dystrophic phenotype one month after birth that is characterized by muscle fibrosis, necrosis and centralized nuclei

11 (Grady et al., 1999). Interestingly, in contrast to the subcellular deficits characteristic of the mdx mouse, the DAP appears to remain intact in the muscles of a-DB null animals as reflected by the maintenance of various components of the complex at the sarcolemma

(Grady et al., 1999). The loss of membrane bound nNOS despite the maintenance of syntrophins suggests that the dystrophies associated with DB loss are the result of disruptions in signaling pathways rather than disassembly of the axis of the complex

(Grady et al., 1999). Furthermore, the a- and P-DB null mouse exhibits sensorimotor defects suggesting abnormalities in the organization or function of the central nervous system (CNS) (Grady et al., 2006).

1.7 The Sarcoglycan-Sarcospan Complex

In skeletal muscle, the sarcoglycan-sarcospan complex (SSC) is composed of four transmembrane glycoproteins, namely, a-, P-, y- and 8-sarcoglycan (SG) and a tetraspan protein, sarcospan (Ozawa et al., 2005). It is believed that the SSC is important in anchoring and/or stabilizing the DAP to the cell membrane in muscle (Ozawa et al.,

2005). It is not clear how the SGs interact with each other or with the DAP although there is evidence to suggest that y-SG interacts with dystrophin (Vainzof et al., 1999).

However, another study provides evidence that the SSC is associated with the DAP complex via an interaction between 5-SG and dystroglycan (Chan et al., 1998). In any event, loss of any one of the SGs abolishes the expression of the others at the sarcolemma and mutations in a-, P-, 8-, and y-SG cause the recessive limb girdle muscular dystrophies

12 2D, 2E, 2C and 2F respectively, underscoring the importance of this mysterious complex in muscle integrity and function (Bushby 1999 ).

1.8 The Dystroglycans

The dystroglycans originate as a single propeptide that is post-translationally cleaved to yield the a- and P-DG subunits (Holt 2000). P-DG is a 43 kDa protein with a single transmembrane domain, one possible N-linked glycosylation site (Barresi and

Campbell, 2006) and a cytoplasmic C-terminus of 121 residues that interacts via the

PPXY motif with the WW domain of dystrophin (Fig. 1.1; Huang et al., 2000; Barresi and Campbell, 2006). Extracellularly, P-DG interacts non-covalently with the extracellular protein a-DG thus stabilizing it at the plasma membrane (Sciandra et al.,

2001). a-DG contains three possible N-linked glycosylation sites, one on the N terminus and two on the carboxyl terminus (Barresi and Campbell, 2006) which are separated by the central, Serine/Threonine -rich, mucin-like domain (arrow in Fig. 1.1). This region undergoes extensive post-translational O-linked glycosylation (Brancaccio et al., 1995;

Brancaccio et al., 1997). Its protein sequence predicts a molecular mass of approximately

40 kDa (Barresi and Campbell, 2006) however the size of a-DG, as detected using western blots, varies substantially between tissues, developmental stages and species reaching a mass as high as 156 kDa in skeletal muscle (Ibraghimov-Beskrovnaya et al.,

1992). Selectively removing the N-linked glycans from a-DG reduces its molecular mass by only 4 kDa (Ervasti and Campbell, 1991) arguing that the variability in detected mass can be attributed to tissue- and species-specific levels of O-linked glycosylation of the

13 central mucin-like domain. a-DG binds several ECM proteins in muscle including agrin, perlecan and laminin (Ibraghimov-Beskrovnaya et al., 1992; Bowe et al., 1994; Talts et

al., 1999). The molecular interaction between laminin and a-DG is the best characterized

of the three and is represented schematically in Figure 1.1. Laminins are composed of

three distinct chains referred to as a, P and y and currently 5a, 4p and 3y chains have been

identified that can combine to give rise to fifteen different laminin isoforms (Hallmann et

al., 2005). The C-terminal ends of the a-chains contain globular domains composed of

five similar modules (LG1-5) (Fig. 1.1, Talts et al., 1999). a-DG is known to bind with

high affinity to LG4 and LG5 of laminin 1 (aiPiyi) and laminin 2 (a2Piyi) (Talts et al.,

1999) and recent studies have shown that this interaction is heavily dependent on the O-

linked glycosylation status of the mucin-like domain of a-DG (Ervasti and Campbell,

1993). DG is widely expressed and has been implicated in playing roles in cell migration, basement membrane assembly and signaling in addition to its role in muscle integrity

(reviewed by Winder 2001; Higginson and Winder, 2005).

Deletion of the DG gene (Dagl) in mice leads to embryonic lethality (Williamson

et al., 1997; Cote et al., 1999). However, its selective deletion from skeletal muscle

results in a mild dystrophy and loss of other components of the DAP from the

sarcolemma (Cohn et al., 2002) indicating that DG plays a pivotal role in the integrity of

the entire complex. While no mutations in dystroglycan have been linked to any human

disease (Barresi and Campbell, 2006), it is becoming increasingly apparent that several

forms of muscular dystrophy can be traced to defects in its post-translational O-linked

14 glycosylation that disrupt its interaction with ECM molecules (Ervasti and Campbell,

1993).

1.9 O-linked glycosylation of a-dystroglycan

The ability of a-DG to interact with and bind proteins of the ECM is heavily reliant on a dense region of O-linked glycosylation that is added to the protein as a post- translational modification (Michele et al., 2002). Proteins of all known living organisms are modified in many different ways in order to modulate their activity, stability, localization and oligomerization with other proteins (reviewed by Geyer and Geyer,

2006; reviewed by Hanks and Hunter, 1995). Glycosylation represents one of the most complex and energetically taxing forms of post translational modification and is, at the same time, one of the most fascinating and poorly understood (Lis and Sharon, 1993).

Glycosylated moieties can be broadly divided into two distinct categories being either N- glycosidically linked to the amino acid asparagine or O-glycosidically linked to the hydroxy amino acids threonine, serine, hydroxyproline, hydroxylysine and tyrosine (Lis and Sharon, 1993). All N-linked glycans are attached to the y-amido group of asparagine by way of the sugar N-acetylglucosamine (GlcNAc) and the diversity of these chains results from variable degrees of branching and differential composition of the sugars. O- linked glycans exist as extensions from several different hydroxy amino acids that link though numerous different sugars (Lis and Sharon, 1993). This is summarized in Table

1.1 (Lehle et al., 2006).

15 Table 1.1 Summary of sugar-peptide bonds in eukaryotes

N-Glycan O-Glycan

Asn - GlcNAc Ser/Thr- Man

- GlcNAc

- GalNAc

- Fuc

- Glc

- Gal

Hyp - Gal

- Ara

- GlcNAc

Ser - Xyl

Hyl - Gal

Tyr - Glc

Asn=asparagine, Ser=serine, Thr=threonine, Hyl = hydroxylysine, Hyp=hydroxyproline, Tyr=tyrosine, GlcNAc=N-acetylglucosamine, GalNAc= N-acetyl- galactosamine, Glc=glucose, Gal=galactose, Rha= rhamnose, Xyl=xylose, Ara=arabinose, Man=mannose, Fuc=fucose. The O-glycans expressed by the mucin domain of a-DG are attached to serine/threonine residues by way of the sugar mannose (Chiba et al., 1997). This particular type of sugar chain is referred to as O-mannosylation and, along with N- glycans, is distinct from all other forms of glycosylation in two ways: (1) The initiation of glycosylation occurs in the Endoplasmic Reticulum (ER) (as opposed to the Golgi apparatus) and (2) lipid-activated sugars, such as dolichol-phosphate-oligosaccarides

(Dol-P-X) or non-sugar nucleotides act as precursors providing the linkage to the protein

(reviewed by Lehle et al.., 2006). Moreover, defects in O-mannosylation and N- glycosylation, by virtue of the lack of individual , manifest in known glycosylation-related congenital diseases, distinguishing them even further from other glycosylation strategies (reviewed by Lehle et al., 2006). Interestingly, O-mannosylation of proteins was long thought to be unique to fungal species (reviewed by Lehle et al..,

2006). It was in 1969 that this protein modification was first described in Baker's yeast

{S. cerevisiae) (Sendantreu and Northcote, 1969) and it wasn't until 30 years later that its existence was revealed in mammals (Yamada et al., 1996). Most recently, O- mannosylation has been found in one bacterial species {Mycobacterium tuberculosis)

(VanderVen et al., 2005). It has not been detected in nematodes (C. elegans) nor in plants

(A. thaliana, O. sativa) (reviewed by Lehle et al., 2006).

a-DG contains an O-mannosidically attached linear tetrasaccharide defined by:

a2-3 P1-4 Rl-2 NeuAc • Gal — • GlcNAc — • Man

17 with Dol-P-Mannose acting as the mannosyl donor (NeuAc=neuraminic acid) (Chiba et al., 1997). The subscripts indicate the bond conformation between sugars. The addition of the first mannose occurs in the ER lumen by a protein O-mannosyl transferase (POMT)

(Haselbeck and Tanner, 1983; Fig. 1.2). The stepwise extension of the sugar chain then takes place in the Golgi apparatus where, in mammals, the GlcNAc residue is transferred from UDP-GlcNAc to mannose by POMGnTl, a protein O-mannose beta 1,2 acetylglucosaminyl transferase (Yoshida et al., 2001). The galactosyl and sialyl transferases involved in the subsequent extension of the chain remain to be identified but fukutin and FKRP are putative candidates at present (reviewed by Lehle et al., 2006).

The ability of a-DG to bind laminin by way of this O-linked sugar chain is believed to be dependent on sialylated residues on a-DG. This idea is based on experiments using sialidase treatment of a-DG which results in a loss of binding affinity of a-DG for laminin (Yamada et al., 1996). Recently, data has immerged to warrant the supposition that these sialylated residues exist on a novel O-mannosyl type oligosaccharide of the type:

Sia "2~3 »Gal pM •GlcNAc 131-2 »Man-Ser/Thr where Sia refers to Sialic Acid. Furthermore it has been suggested that the sialyl-N- acetyllactosamine residue of this sugar chain specifically confers the interaction of laminin with a-DG (Chiba et al., 1997).

18 In any event, complexities of sugar chemistry aside, it is clear that defects in individual glycosyl transferases prevent the extension of the a-DG bound O-mannosyl oligosaccharides and that this in turn circumvents the ability of a-DG to form a functional interaction with laminin. If the residues directly involved in the laminin-a-DG link are at the distal aspects of the chain, defects in any one of the upstream glycosyltransferases involved in the extension of the chain would disrupt the affinity of a-DG for laminin.

This is reflected in the extensive list of congenital muscular dystrophies believed to be a direct result of a-DG hypoglycosylation. This is discussed in further detail in section

1.10.

1.10 Alpha-Dystroglycanopathy Related Muscular Dystrophies

The preceding sections have been intended to deliver a comprehensive outline of the major players in the structure and function of the DAP complex. As mentioned above, the ability of the complex to effectively bind extracellular matrix proteins and complete its trans-membrane bridge is heavily dependent on a dense region of O-linked glycosylation in the mucin-like domain of the extracellular protein a-DG (Ervasti and

Campbell, 1993). Disruption of these glycosylated moieties prevents the binding of a-DG to ECM proteins (Michele et al., 2002; Kanagawa et al., 2005) and this is associated with a unique subset of human muscular dystrophies (Muntoni et al., 2002) earning them the secondary name of dystroglycanopathies (Toda et al., 2003). Recently, several of these diseases have been traced to mutations within specific genes encoding putative

19 glycosyltransferases involved in the O-linked glycosylation of a-DG. For instance,

Fukuyama Congenital Muscular Dystrophy (FCMD) has been traced to mutations in the fukutin gene (Kobayashi et al., 1998). Likewise, the genes responsible for Muscle Eye

Brain Disease (MEB) and Walker-Warburg Syndrome (WWS) are POMGnTl (Yoshida et al., 2001) and POMT1, respectively (Beltran-Valero De Bernabe et al., 2002). All of these human diseases are characterized by a hypoglycosylation of a-DG and a concomitant loss of laminin binding (Hayashi et al., 2001; Kano et al., 2002; Michele et al., 2002; Kim et al., 2004). Aberrant glycosylation of a-DG can be detected by western blotting using antibodies specifically directed against glycosylated moieties of the protein

(IIH6C4 and V1A4-1, Upstate, NY) in conjunction with antibodies that detect the core protein such as GT20ADG (Michele et al., 2002). Furthermore, all of these diseases present as a progressive myopathy with variable associated non-progressive defects in ocular and brain anatomy and function (Hayashi et al., 2001; Kano et al., 2002; Kim et al., 2004). These are typically characterized by neuronal migration abnormalities within the brain and retina, epilepsy and variable cognitive defects ranging from mild learning disabilities to severe mental retardation (Muntoni et al., 2002).

FCMD is the most commonly occurring form of congenital muscular dystrophy in

Japan (Fukuyama et al., 1981). It is characterized by a severe dystrophy, mental retardation, epilepsy and ocular abnormalities as well as recently reported aberrant neuromuscular junctions (Fukuyama et al., 1981; Taniguchi et al., 2006). It is caused by a

3 kb retrotransposal insertion in the 3' non-coding region of the fukutin gene leading to truncation of the protein and loss of its function (Kobayashi et al., 1998). While the

20 specific function of fukutin is not known, a severe reduction of a-DG glycosylation and consequent loss of ligand binding caused by its mutation (Michele et al., 2002), as well as its homology to known enzymes involved in the modification of glycolipids and proteins

(Avarind and Koonin, 1999), suggest a role in the post-translational glycosylation of a-

DG. The targeted homozygous disruption of the fukutin gene in mice results in embryonic lethality highlighting its importance in developmental processes (Takeda et al., 2003). This research dilemma has been overcome through the generation of viable, yet dystrophic, fukutin chimeric mice whose muscle histology reveals massive necrosis, regenerating fibres with centralized nuclei and fatty and connective tissue infiltration

(Takeda et al., 2003). Furthermore, the brains of fukutin chimera reveal severe disruption in the normal cortical and cerebellar lamination and fusion of the cerebral hemispheres.

Similarly, severe ocular abnormalities exist in these mice including disruption of retinal lamination and vascular infiltration (Takeda et al., 2003). Together these findings underscore the essential role of this protein in the maintenance of muscle integrity, cortical histogenesis and proper ocular development (Takeda et al., 2003).

Based on sequence homology, a fukutin homologue was identified and termed fukutin related protein (FKRP) (Brockington et al., 2001). Mutations in FKRP are now known to underlie both Congenital Muscular Dystrophy Type IC (CMD1C) and Limb

Girdle Muscular Dystrophy 21 (LGMD2I) (Torelli et al., 2005). While the exact function of FKRP is unknown, sequence analysis and cellular localization suggest that it is a (Esapa et al., 2002). This is substantiated by the fact that both

21 CMD1C and LGMD21 show a reduction in a-DG glycosylation and are characterized by myopathy as well as neuronal migration and ocular abnormalities (Brown et al., 2004).

Similarly MEB is characterized by a hypoglycosylation of a-DG, myopathy and brain and eye malformations (Santavuori et al., 1989). This form of dystrophy is found predominantly in Finland (Santavuori et al., 1989), and is caused by mutations in the

POMGnTl gene that encodes a UDP-GlcNAc:Mana-Opl,2-N- acetylglucoaminyltransferase assumed to catalyze the GlcNAcpi,2Man linkage in the a-

DG sugar chain (Yoshida et al., 2002). POMT1 which underlies WWS, shares considerable homology with protein O-mannosyltransferases in yeast known to catalyze the transfer of a mannosyl residue from dolichyl phosphate mannose to Serine/Threonine residues on proteins, predicting POMT1 as encoding a protein involved in the first step of the O-mannosyl glycan synthesis of a-DG (Beltran-Valero de Bernabe et al., 2002). As with FCMD, CMD1C and MEB, WWS exhibits a severe myopathy accompanied by structural and functional defects in the brain and retina. WWS is the most severe congenital muscular dystrophy often resulting in death within the first year of life (Vajsar and Schachter, 2006). Engineered POMT1 null mice are embryonic lethals (Wilier et al.,

2004).

Together these diseases serve to underscore the importance of the proper glycosylation of a-DG in muscle integrity. Dystroglycan effectively acts as a bridge, spanning the plasma membrane and connecting the basal lamina to the intracellular components of the DAP complex. The common brain and ocular phenotypes suggest a

22 role of the DAP complex in the integrity and function of the CNS and that its disruption in these tissues could underlie their respective pathologies. A summary of these diseases, their associated genes and clinical features are summarized in Table 1.2. A schematic diagram of the O-linked glycosylation of a-DG requiring these putative glycosyltransferases is illustrated in Figure 1.2.

23 Table 1.2 Summary of a-DG Hypoglycosylation Related Muscular Dystrophies

Condition Gene Protein Function Clinical Features a-DG

Fukuyama Fukutin Putative -Severe muscle weakness Not Congenital glycosyltransferase -Mental Retardation glycosylated* Muscular -Epilepsy Dystrophy -Neuronal Migration (FCMD) Disorders -Ocular Abnormalities

Muscle- POMGnTI O-mannosyl glycan -Severe muscle weakness Not Eye-Brain synthesis -Mental Retardation glycosylated* Disease -Epilepsy (MEB) -Neuronal Migration Disorders -Ocular Abnormalities

Walker- POMT1 Putative O- -Severe muscle weakness Not Warburg mannosyl- -Death in infancy glycosylated* Syndrome transferase -Neuronal Migration (WWS) Disorders -Ocular Abnormalities Congenital Fukutin- Putative -Muscle weakness Not Muscular Related glycosyltransferase -Neuronal Migration glycosylated* Dystrophy Protein Disorders 1C (FRP) -Ocular Abnormalities (CMD1C)

Largemyd Large Putative -Severe muscle weakness Not mouse and glycosyltransferase -Neuronal Migration glycosylated* Congenital Disorders Muscular -Ocular Abnormalities Dystrophy 1D

* As determined by immunohistochemistry using an antibody towards glycosylated epitopes POMGnT 1 =Protein-0-manose-P— 1,2-N-acetylglucosaminyl transferase, POMT=protein O-mannosyl transferase

24 FKRP: Fukutin: LARGE: LGMD2I, FCMD MDC1D MDC1C ? • 1

Neu5Ac Gal GlcNAc Man a-DG Mucin-like / Domain POMGnTI: MEB POMT1/2: WWS

Figure 1.2 O-Mannosyl Glycan Synthesis of a-Dystroglycan Based on sequence homology with other known O-mannosyltransferases, POMT1 is believed to catalyze the first step in the O-glycosylation of a-DG. The second step is then catalyzed by the enzyme POMGnTI. Fukutin and FKRP are also involved in this glycosylation however their specific roles are not yet known. A mutation in any of these genes results in congenital muscular dystrophy. Overexpression of the putative glycosyltransferase (Large) has some rescue benefits in certain forms of muscular dystrophy as discussed in section 1.15. MDC=congenital muscular dystrophy.

25 1.11 The Dystroglycan Associated Protein Complex in the Central Nervous System

The most immediate detriment affiliated with muscular dystrophy pertains to muscle weakness and pseudohypertrophy (Emery 2002). However, as mentioned previously, many MDs, including those associated with O-linked a-DG hypoglycosylation, present concomitantly with defects in CNS structure and function, the etiologies of which are poorly understood (Emery 2001). Interestingly, the expression of dystroglycan is by no means limited to muscle. In fact, among other tissues^4t is densely localized to the CNS where it forms complexes in both neuronal and glial cells including astrocytes and Miiller cells (Blake et al., 1999; Moukhles et al., 2000; Moukhles and

Carbonetto, 2001). This suggests that defects within the complex in these tissues may underlie the CNS phenotypes.

Dystroglycan is widely distributed throughout most regions of the brain yet its expression is restricted to specific subtypes of neurons demonstrating that its expression is highly selective (Zaccaria et al., 2001). It is found in pyramidal neurons and a subset of interneurons in the cerebral cortex, in Purkinje cells of the cerebellum and in neurons in the CA1-CA3 regions of the hippocampus. It is also localized to neurons in the basal ganglia, thalamus, hypothalamus and brain stem (Zaccaria et al., 2001). The precise role of the DAP complex within neurons is not currently known but DG, along with Dp, colocalizes with GABAA receptors at post-synaptic sites of inhibitory synapses of the cerebral cortex, hippocampus and cerebellum (Levi Craig et al., 2002; Knuesel et al.,

26 1999) and in the mdx mouse, the synaptic clustering of GABAA receptors is impaired

(Knuesel et al., 1999). This has been correlated with a reduction of inhibitory transmission in Purkinje cells (Anderson et al., 2003). Furthermore, mutant mice with a brain specific deletion of DG present with a reduction in high-frequency stimulation induced long term potentiation at CA3-CA1 synapses of the hippocampus (Moore et al.,

2002) and human DMD patients with associated cognitive defects present with a loss of specific populations of neurons that typically express dystrophin as well as small ischemic infarcts throughout the cerebral cortex (reviewed by Haenggi and Fritschy,

2006). In the retina, a DAP complex containing dystrophin, dystroglycan, a-syn and a-

DB is localized to the outer plexiform layer (OPL) where photoreceptor terminals synapse onto bipolar cells and horizontal cells and a similar complex is found localized to the inner plexiform layer (IPL) where bipolar cells form synapses with amacrine and ganglion cells (reviewed by Haenggi and Fritschy 2006). Together these data suggest that the neuronal expression of the DAP complex is involved in plasticity and modulation at a subset of synapses as well as in protection against hypertrophy or neuronal death secondary to ischemic damage within the CNS.

In addition to the synaptic expression of components of the DAP complex in both brain and retina, of particular interest to the present investigation, a major DAP complex is highly enriched in glial cell processes at the boundaries between neural and fluid compartments in the brain and retina (Fig. 1.3). For instance, astrocytic endfeet in brain, and Miiller and astrocyte endfeet in retina, abut blood vessels where they boast a dense and highly polarized distribution of a DAP complex containing DG, Dp, and a- and 0-

27 syn. A similar complex is found in astrocytic endfeet that exist subpially lining the subarachnoid space and the ventricles of the brain and at analogous domains in glia whose endfeet abut the inner limiting membrane of the retina apposed to the vitreous body. This localization pattern suggests a possible role in water homeostasis and electrolyte balance within these tissues (Ueda et al., 1998; Zaccaria et al., 2001).

Substantiating this possibility further is the coincident expression pattern of both the water channel aquaporin 4 (AQP4) and the inwardly rectifying potassium channel 4.1

(Kir4.1) that both distribute to the same astrocytic domains as the DAP complex (Nielsen et al., 1997; Nagelhus et al., 1999; Higashi et al., 2001). The codistribution of these proteins, along with other data outlined below, suggests that their localization and function could be interrelated and that disruption in their expression profiles could contribute to the manifestation of both the morphological and electrophysiological CNS defects in MD. The reader is referred to Fig. 1.4 for a schematic representation of retinal lamination.

28 Figure 1.3 Schematic of Perivascular Astrocytic Domains The DAP complex, AQP4 and Kir4.1 all co-distribute at the perivascular domains of glial cell endfeet in the brain and retina.

29 Retinal Pigment •Epithelium (RPE) Outer Rod Nuclear Cone Layer (ONL)

Outer Plexiform Horizontal Layer (OPL) Cell SA Bipolar Cell o Inner Nuclear e Layer (WL) "j o

Inner Plexiform Amacrine Cell Layer (JPL) Ganglion Cell Ganglion Cell- Layer ' Optic Nerve Fibres

Figure 1.4 Schematic of Mammalian Retinal Lamination Photoreceptors (Rods and Cones) are embedded in retinal pigment epithelium (RPE) at the posterior aspect of the retina. The cell bodies of these cells comprise the outer nuclear layer of the retina. Photoreceptor cells form synapses with both bipolar and horizontal cells in the outer plexiform layer. The soma of bipolar and horizontal cells comprise the inner nuclear layer. Bipolar cells synapses onto ganglion cells in the inner plexiform layer. The cell bodies of ganglion and amacrine cells comprise the inner plexiform layer at the most anterior aspect of the retina. Ganglionic axons travel together to form the fibres of the optic nerve.

30 1.12 Inwardly Rectifying Potassium Channels (Kir)

The close spatial relationship between the DAP complex and Kir4.1 in brain and retina suggests a possible functional association between the two and could offer clues into the pathologies associated with the CNS in MD. Kir channels are loosely grouped into seven subfamilies (Kir 1.0-7.0) based on molecular and electrophysiological characteristics (Hille 1992). They assemble as tetramers of either homomeric or heteromeric Kir subunits which further increases their functional diversity. They are defined as a family by their preference to carry current in the inward direction, steep voltage dependence, strict dependence on external potassium ion (K+) concentration and their strong modulation by intracellular factors and second messengers (Hille 1992).

These channels have fascinating kinetics as they carry current at resting membrane potentials (RMP) negative to the equilibrium potential for K+ but with more positive potentials, the current flow is blocked in a voltage-dependent manner by intracellular polyamines or Mg2+ (Ruppersberg 2000). Since an increase in extracellular K+

+ concentration ([K ]0) is physiologically equivalent to a reduction in RMP, Kir channel

+ current increases with increases in [K ]0 thus maintaining the membrane potential near the equilibrium potential for K+. This is a defining characteristic of glia and Kir channels are predominantly responsible for the negative RMP of these cells as well as their high

K+ permeability (Sontheimer 1994). Based on their biophysical properties, the Kirs are divided into five subtypes including (1) classic inward rectifiers, (2) G-protein activated

K+ channels, (3) ATP-sensitive K+ channels, (4) ATP-dependent K+channels and (5) channels with miscellaneous characteristics including the weakly rectifying channels.

31 Their ability to form heteromeric tetramers further increases the diversity of these channels and to date they have been shown to be critical in a plethora of different physiological processes including acetylcholine-induced slowing of the heart rate, control of RMP and membrane excitability, ion and water transport, acid-base homeostasis, and hormone secretion in heart, smooth muscle, kidneys, stomach, pancreas and CNS

(Reviewed by Butt and Kalsi, 2006). Glia have been shown to express representatives of most, if not all, of the Kir subtypes giving clues as to the various functions that these channels have in these cells including maintenance of a highly negative RMP, K+ and H+ transport, membrane excitability, cell volume regulation and cell viability (Reviewed by

Butt and Kalsi, 2006).

1.13 Kir4.1 in brain and retina

In glia, the Kir4.1 subtype is consistently found to have a widespread and specific expression profile within both brain and retina implicating this channel in their physiological characteristics (Butt and Kalsi, 2006). This is substantiated by the severe reduction in K+ currents and depolarization in Muller cells of the Kir4.1 knock-out mouse, suggesting that Kir4.1 is a major mediator of K+ fluxes in these cells (Kofuji et al., 2000). In astrocytes, weakly rectifying homomeric Kir4.1 channels are found enriched in the processes that wrap blood vessels, otherwise known as perivascular foot processes (Fig. 1.5; Higashi et al., 2001), with an analogous distribution in retinal Muller cell endfeet at the interface between the inner limiting membrane and the vitreous body

32 (Kofuji et al., 2000). Astrocytic processes surrounding synapses, on the other hand, are rich in strongly rectifying heteromeric Kir4.1/Kir5.1 (Hibino et al., 2004). This expression profile predicts a specialized role for Kir4.1 in spatial potassium buffering in which K+ released by active neurons is taken up by heteromeric Kir channels and then extruded into the vascular system by homomeric Kir4.1 (Fig. 1.5). This process functions to prevent large fluxes in extracellular K+ concentration in the face of extended neural discharge. Neuronal activity results in local increases in extracellular K+ which, if left unresolved, would compromise electrical excitability. Potassium buffering refers to a process in which excess K+ is siphoned from the extracellular milieu into the bloodstream by way of brain astrocytes and retinal Muller cells (Kofuji and Newman, 2004).

33 Blood Vessel

Synapse

Figure 1.5 Schematic model of spatial pottasium buffering (1) K+ released from neural activity at Nodes of Ranvier and synapses is taken up into astrocytes via strongly rectifying heteromeric Kir4.1/5.1 channels. (2) K+ is released from astrocytes via weakly rectifying homomeric Kir4.1 channels at perivascular + endfeet in areas of low [K ]D. (3) The astrocytic syncytium allows for the movement of K+ taken up by one astrocyte to be shared between connecting astrocytes to + facilitate its extrusion into regions of low [K ]D through homomeric Kir4.1 by perivascular astrocytic endfeet (2 and 4).

34 For the purposes of the present study, the polarized distribution of Kir4.1 to perivascular astrocytic domains is of particular interest because of its overlap with that of dystroglycan and associated proteins (Higashi et al., 2001; Guadagno and Moukhles,

2004). Several lines of evidence suggest that the localization of Kir4.1 to these specialized domains is dependent on the integrity of the dystroglycan complex. For instance, targeted inactivation of Dp71 results in a loss of Kir4.1 polarization characterized by a uniform redistribution to other glial domains within the retina (Dalloz et al., 2003). Furthermore, in mdx3cv mice lacking all isoforms of dystrophin, Kir4.1 is also mislocalized away from Miiller cell endfeet causing an alteration in K+ currents throughout these glial cells (Connors and Kofuji, 2002). Likewise, in brain and retina,

Kir4.1 associates with the DAP complex as indicated by recent co-immunoprecipitation experiments showing an interaction between Kir4.1 and several components of the complex (Leonoudakis et al., 2004; Connors et al., 2004). Furthermore, Kir4.1 contains a

C-terminus SNV sequence that binds to the PDZ motif of the DAP complex protein a- syntrophin, thereby stabilizing the channel at the plasma membrane (Leonoudakis et al.,

2004; Connors et al., 2004).

All of these data provide evidence for the involvement of the DAP complex in the proper localization and thus function of Kir4.1 at glial cell endfeet in both brain and retina. Of further interest is that AQP4 presents the same polarized distribution pattern as

Kir4.1 and the DAP complex suggesting a possible common mechanism involved in the targeting of these channels as well as their function (Nagelhus et al., 1999).

35 1.14 Aquaporins

Aquaporins (AQP) are small (-30 kDa), highly hydrophilic, integral membrane proteins that assemble as tetramers. Each monomer consists of six membrane spanning a- helical domains with cytoplasmically oriented amino and carboxyl tails and each containing a distinct water pore (King et al., 2004). To date, eleven mammalian AQPs have been identified and many others from plants, yeast, bacteria, amphibians and other smaller organisms (King et al., 2004). Mammalian AQPs can be broadly divided into two main classes being either permeable to water alone or water and glycerol (King et al.,

2004). This is, however, an oversimplification as certain AQPs have been shown to carry larger neutral solutes (AQP9) and, at low pH, even chloride (AQP6) (Verkman 2002) demonstrating other biophysical roles. The mammalian AQPs are expressed in specific tissues but it should be carefully noted that not all cell types contain these channels. For instance, there are no known AQPs expressed in neurons (Agre and Kozono, 2003). The specific sites of AQP expression predict roles in renal water reabsorption, cerebrospinal fluid dynamics, aqueous fluid dynamics and salivary gland secretion as well as multiple other processes (Papadopoulus et al., 2002). Since selective antagonists of AQP function are not available for in vivo studies, several transgenic mice deficient for specific AQPs have been generated and offer indispensable insight into their respective functions. For example, mice lacking AQP1, 2 or 3 have defective urine concentrating ability (Ma et al.,

1998; Ma et al., 2000; Yang et al., 2001) while those lacking AQP5 have compromised excretion from salivary (Ma et al., 1999) and airway submucosal glands (Song and

Verkman, 2001). Collectively, these studies of AQP null mice suggest that these

36 channels mediate rapid, iso-osmolar fluid transport as well as water fluxes driven by established osmotic gradients.

1.15 Aquaporin 4 in brain and retina

Both AQP1 and AQP4 are expressed in the brain. AQP1 expression is restricted to the ventricles facing the surface of the choroid plexus suggesting that it plays a role in cerebrospinal fluid (CSF) secretion (Boassa et al., 2006). AQP4, on the other hand, is localized to astrocyte foot processes adjacent to blood vessels as well as at ependymal and pial surfaces in contact with CSF (Frigeri et al., 1995; Nielsen et al., 1997) an expression profile highly reminiscent of both the DAP complex and Kir4.1. Likewise, in retina, AQP4 is enriched at Muller cell endfeet abutting blood vessels (Dalloz et al.,

2003). The mdx mouse, lacking dystrophin, shows a pronounced perturbation of AQP4 away from astrocytic endfeet in the brain (Frigeri et al., 2001) suggesting a dependence of AQP4 on the DAP complex for proper localization. Similarly, in retina, targeted inactivation of Dp71 leads to a loss of AQP4 from Muller cell endfeet (Dalloz et al.,

2003). As with Kir4.1, AQP4 is thought to associate with the complex through a PDZ mediated interaction with a-syntrophin. Indeed, like Kir4.1, AQP4 contains a PDZ- binding consensus and deletion of a-syntrophin leads to a mislocalization of AQP4 away from perivascular domains (Amiry-Moghaddam et al., 2004; Puwarawuttipanit et al.,

2006). Interestingly, the deletion of the PDZ binding motif of AQP4 (SSV) induces an increase in its degradation rate rather than a redistribution arguing that the interaction of

37 this region of AQP4 with the PDZ domain of a-syntrophin is involved in its stability rather than its localization (Neely et al., 2001). Furthermore, in astrocyte as well as

Muller cell cultures, both AQP4 and Kir4.1 co-aggregate with members of the DAP complex following laminin treatment, suggesting that the polarization of these channels is dependent on the ability of the DAP complex to interact with components of the basement membrane (Guadagno and Moukhles, 2004; Noel et al., 2005).

Not only do Kir4.1 and AQP4 share a spatial relationship, it appears that the two channels have a functional interdependence as well. Spatial buffering of K+ ions by

Kir4.1 generates osmotic gradients (Dietzel et al., 1980) and neuronal activation leads to a decrease in adjacent extracellular space suggesting water movement (Holthoff and

Witte, 1996). These data together with the co-distribution of Kir4.1 and AQP4 channels to the same subcellular domains have led to the proposal that Kir4.1 -mediated K+ ion buffering is associated with a concomitant flux of water through AQP4 (Holthoff and

Witte, 2000; Niermann et al., 2001 ). Recently, further evidence substantiating this showed that redistribution of AQP4 away from perivascular glial endfeet is accompanied by a reduction in potassium buffering capability in a-syntrophin null mice (Amiry-

Moghaddam et al., 2003). Taken together these data provide compelling evidence that the integrity of the DAP complex is essential for the proper distribution and coordinated function of both Kir4.1 and AQP4 within the brain and retina and give reason to suggest that the disruption of the complex in muscular dystrophy could compromise the normal activity of these channels. This, in turn, could ultimately encumber normal neuronal, and hence brain and retina, function.

38 1.16 The Largemyd mouse

As discussed above, the ability of the DAP complex to effectively bind ECM proteins such as laminin by way of a-DG is essential to the integrity of the complex as a whole and thus its role in the protection of muscle fibres (Zubrzycka-Gaarn et al., 1988;

Petrof et al., 1993). Furthermore, this interaction is dependent on a dense region of O- linked glycosylation within the mucin-like domain of the a-DG protein and several forms of human MD have been linked to glycosylation deficits at this site (Muntoni et al.,

2002). These include FCMD, MEB, WWS and CMD Type IC (see Table 1.2), all of which present with a progressive myopathy and defects in the CNS. The Largemyd mouse is an animal model with a spontaneous mutation in the Large gene used to study dystroglycanopathy associated MDs. It presents with a diffuse, progressive myopathy and shortened lifespan, as well as structural defects in the brain and retina (Holzfeind et al.,

2002). Furthermore, this mouse exhibits a hypoglycosylation of a-DG and a concomitant loss of binding to ECM proteins suggesting that LARGE is a glycosyltransferase involved in the O-glycan synthesis of a-DG (Michele et al., 2002). Interestingly, the human CMD Type ID (CMDID) has also been traced to mutations in the Large gene

(Longman et al., 2003).

The Large gene, named for its considerable size (650kb) is predicted to encode a protein carrying a cytoplasmic N-terminal, a transmembrane region, a coiled-coil motif and two catalytic domains (Grewal et al., 2001). It shares homology with both a family of known bacterial glycosyltransferases as well as with the human glycosyltransferase UDP-

GlcNAc:GalBl,3-A^-acetyl-glucosaminyltransferase (Coutinho et al., 2003). This, in

39 combination with the hypoglycosylation of a-DG associated with Large mutations, suggests that its protein product is a glycosyltransferase involved in the O-glycan synthesis of a-DG. This is further substantiated by the recent finding that overexpression of Large in cell lines derived from patients with FCMD, MEB or WWS results in an increased glycosylation of a-DG (Fig. 1.2) and a rescue of laminin binding activity providing evidence for the possible therapeutic use of this gene in the treatment of these diseases (Baressi et al., 2004).

In the Largemyd mouse, the Large gene carries a deletion mutation of more than lOOkb resulting in the removal of exons 5, 6 and 7. This leads to a 379-bp deletion within the coding region which causes a frameshift within the mRNA. This in turn creates a premature stop signal resulting in a truncated, non functional protein (Grewal and Hewitt,

2002). Shortly after the identification of the Large 1 gene, BLAST searches revealed a closely related gene referred to as Large2 (Grewal et al., 2005). The original Large, commonly referred to as Large 1, is widely expressed with the highest levels found in brain, heart, skeletal muscle, stomach, bladder, and pancreas. Large2, on the other hand, exhibits a more restricted expression profile with highest levels found in the placenta, adult and fetal kidney, pancreas, prostate and salivary glands with minimal levels found in either brain or skeletal muscle (Grewal et al., 2005). In mouse, Large2 also exists in an alternative splice form, the significance of which is unclear (Grewal et al., 2005). No alternatively spliced products for Large 1 have been found in either mouse or human.

Homologues of Large have been identified in mammals, fish and chicken genomes

(Grewal et al., 2005). As mentioned above, null mutants for both fukutin (Takeda et al.,

40 2003) and POMT1 (Wilier et al., 2004) are embryonic lethal yet homozygous Large mutants are viable and express a comparatively mild phenotype (Grewal et al., 2001).

Furthermore, in Large 1 mutants, tissues expressing high levels of Large2, such as kidney, maintain their normal level of a-DG glycosylation. This suggests that Large2 could compensate for deficiencies in Largel (Grewal et al., 2005).

Analyses of muscles from the Largemyd mouse show degenerative clusters of muscle fibres characterized by variable size and shape as well as centralized nuclei while examination of the sarcolemma suggests a destabilization of the DAP complex as determined by a reduction of dystrophin at the intercostameric regions (Reed et al.,

2004) . Histological assessment of the CNS system reveals aberrant neuronal migration, especially in the cortex and cerebellum as well as defects in retinal organization and activity (Holzfeind et al., 2002; Lee et al., 2005). All of these defects are highly mnemonic of FCMD, WWS, MEB and CMD ID. Interestingly, heterozygote littermates are asymptomatic, showing no signs of dystrophy or shortened life span (Grewal and

Hewitt, 2002), making them ideal candidates for experimental controls. Based on the uncanny similarities between the Largemyd mouse and the human diseases in which I am interested, I have used this mouse as the animal model for all experiments included in this project.

41 1.17 Project Outline and Hypotheses

As outlined above, the primary deficits characteristic of MD pertain to skeletal muscle and the underlying causes of these pathologies involve discontinuities in the transmembrane DAP complex and an associated disruption of the sarcolemmal integrity.

However, many forms of MD are also associated with cognitive and ocular defects of unknown etiology. The expression of the DAP complex in both brain and retina, suggests its possible involvement in their respective pathologies. The integrity of the DAP complex requires an interaction between a-DG and proteins of the basement membrane and this is dependent on a dense region of O-linked glycosylation of a-DG. Loss of this glycosylation encumbers the DAP complex's ability to bind the basement membrane resulting in the instability of the complex as a whole. This is associated with a subset of human forms of MD including FCMD, MEB, WWS and CMD ID as well as the Largemyd mouse. In the CNS, the DAP complex codistributes with both AQP4 and Kir4.1 and several studies suggest that the integrity of the DAP complex could be necessary for the proper polarized distribution of these channels. Because these channels play a critical role in the process of potassium buffering, their mislocalization could result in loss of proper electrolyte homeostasis leading to compromised neuronal activity.

/ propose that the loss of a-DG O-linked glycosylation and

concomitant loss of basal lamina binding in the Largemyd mouse

and associated human forms of muscular dystrophy disrupts the

normal, polarized distribution of the water channel aquaporin 4

and the inwardly rectifying potassium channel 4.1 in the CNS.

42 Mislocalization of these channels in the brain and retina could

compromise normal neuronal activity through a failure to

regulate ion and water homeostasis.

In the present study I have examined the distribution and expression of these channels in brain and retina of the Largemyd mouse and used littermate or wild type mice as controls. Using immunofluorescence and immunoblot techniques I show that in the

Largemyd brain both AQP4 and Kir4.1 are lost from perivascular astrocytic domains. This substantiates the involvement of the a-DG O-linked glycosylation dependent interaction between the DAP complex and the basal lamina in the proper distribution of these channels. However, in retina, I show that despite the hypoglycosylation of a-DG in this tissue, Kir4.1 and AQP4 are retained at perivascular astrocytic domains. These data reveal a differential anchoring mechanism for AQP4 and Kir4.1 between brain and retina.

43 1.18 Works Cited

Adams ME, Kramarcy N, Krall SP, Rossi SG, Rotundo RL, Sealock R, Froehner SC (2000) Absence of alpha-syntrophin leads to structurally aberrant neuromuscular synapses deficient in utrophin. J Cell Biol 150:1385-1398.

Adams ME, Mueller HA, Froehner SC (2001) In vivo requirement of the alpha- syntrophin PDZ domain for the sarcolemmal localization of nNOS and aquaporin- 4. J Cell Biol 155:113-122.

Agre P and Konozo D (2003) Aquaporin water channels: molecular mechanisms for human diseases. FEBS Letts 555:72-78.

Alessi A, Bragg AD, Percival JM, Yoo J, Albrecht DE, Froehner SC, Adams ME (2006) gamma-Syntrophin scaffolding is spatially and functionally distinct from that of the alpha/beta syntrophins. Exp Cell Res 12:3084-3095.

Ambrose HJ, Blake DJ, Nawrotzki RA, Davies KE (1997) Genomic organization of the mouse dystrobrevin gene: comparative analysis with the dystrophin gene. Genomics 39:359-369.

Amiry-Moghaddam M, Williamson A, Palomba M, Eid T, de Lanerolle NC, Nagelhus EA, Adams ME, Froehner SC, Agre P, Ottersen OP (2003) Delayed K+ clearance associated with aquaporin-4 mislocalization: phenotypic defects in brains of alpha-syntrophin-null mice. PNAS 100:13615-13620.

Amiry-Moghaddam M, Xue R, Haug FM, Neely JD, Bhardwaj A, Agre P, Adams ME, Froehber SC, Mori S, Ottersen OP (2004) Alpha-syntrophin deletion removes the perivascular but not endothelial pool of aquaporin-4 at the blood-brain barrier and delays the development of brain edema in an experimental model of acute hyponatremia. FSEB 18:542-544.

Anderson JL, Head SI and Morley JW (2003) Altered inhibitory input to Purkinje cells of dystrophin-deficient mice. Brain Res. 982: 280-283.

Avarind L and Koonin EV (1999) The fukutin protein family- predicted enzymes modifying cell surface molecules. Curr Biol 9: R836-R837.

Barresi R and Campbell KP (2006) Dystroglycan: from biosynthesis to pathogenesis of human disease. J Cell Sci 119:199-207.

44 Beltran-Valero De Bernabe D, Currier S, Steinbrecher A, Celli J, Van Beusekom E, Van Der Zwaag B, Kayserili H, Merlini L, Chitayat D, Dobyns WB, Cormand B, Lehesjoki AE, Cruces J, Voit T, Walsh CA, Van Bokhoven H, Brunner HG (2002) Mutations in the O-Mannosyltransferase Gene POMT1 Give Rise to the Severe Neuronal Migration Disorder Walker-Warburg Syndrome. Am J Hum Genet 71:1033-1043.

Blake DJ, Hawkes R, Benson MA, Beesley PW (1999) Different dystrophin-like complexes are expressed in neurons and glia. J Cell Biol 147:645-658.

Blake DJ, Nawrotzki R, Loh NY, Gorecki DC, Davies KE (1998) Beta-dystrobrevin, a member of the dystrophin-related protein family. PNAS 95:241-246.

Blake DJ, Nawrotzki R, Peters MF, Froehner SC, Davies KE (1996) Isoform diversity of dystrobrevin, the murine 87-kDa postsynaptic protein. J Biol Chem 271:7802- 7810.

Blake DJ, Weir A, Newey SE, Davies KE (2002) Function and genetics of dystrophin and dystrophin-related proteins in muscle. Physiol Rev 82:291-329.

Blank M, Blake DJ, Kroger S (2002) Molecular diversity of the dystrophin-like protein complex in the developing and adult avian retina. Neurosci 111:259-273.

Bowe MA, Deyst KA, Leszyk JD, Fallon JR (1994) Identification and purification of an agrin receptor from Torpedo postsynaptic membranes: a heteromeric complex related to the dystroglycans. Neuron 12:1173-1180.

Brancaccio A, Schulthess T, Gesemann M, Engel J (1995) Electron microscopic evidence for a mucin-like region in chick muscle alpha-dystroglycan. FEBS Lett 368:139- 142.

Brancaccio A, Schulthess T, Gesemann M, Engel J (1997) The N-terminal region of alpha-dystroglycan is an autonomous globular domain. Eur J Biochem 246:166- 172.

Brenman JE, Chao DS, Gee SH, McGee AW, Craveb SE, Santillano DR, Wu Z, Huang F, Xia H, Peters MF, Froehner SC, Bredt DS (1996) Interaction of nitric oxide synthase with the postsynaptic density protein PSD-95 and alpha 1-syntrophin mediated by PDZ domains. Cell_84:757-767.

Brockington M, Blake DJ, Prandini P, Brown SC, Torelli S, Benson MA, Ponting CP, Estournet B, Romero NB, Mercuri E, Voit T, Sewry CA, Guicheney P, Muntoni F (2001) Mutations in the fukutin-related protein gene (FKRP) cause a form of congenital muscular dystrophy with secondary laminin alpha2 deficiency and abnormal glycosylation of alpha-dystroglycan. Am J Hum Genet 69:1198-1209.

45 Brown SC, Torelli S, Brockington M, Yuva Y, Jimenez C, Feng L, Anderson L, Ugo I, Kroger S, Bushby K , Voit T, Sewry C, Muntoni F (2004) Abnormalities in a- Dystroglycan Expression in MDC1C and LGMD2I Muscular Dystrophies Am J Pathol 164:727-737.

Bushby KM (1999) Making sense of the limb girdle muscular dystrophies. Brain 122:1403-1420.

Butt AM and Kalsi A (2006) Inwardly rectifying potassium channels (Kir) in central nervous system glia: a special role for Kir4.1 in glial functions. J Cell Mol Med 10: 33-44.

Byers TJ, Lidov HG, Kunkel LM (1993) An alternative dystrophin transcript specific to peripheral nerve. Nat Genet 4:77-81.

Chapman VM, Miller DR, Armstrong D, Caskey CT (1989) Recovery of induced mutations for X chromosome-linked muscular dystrophy in mice. PNAS 86:1292- 1296.

Chen Z, Hague C, Hall RA, Minnerman KP (2006) Syntrophins regulate alphalD- adrenergic receptors through a PDZ domain-mediated interaction. J Biol Chem 281: 12414-12420.

Chiba A, Matsumura K, Yamada H, Inazu T, Shimizu T, Kusunoki S, Kanazawa I, Kobata A and Endo T (1997) Structures of sialylated O-linked oligosaccharides of bovine peripheral nerve alpha-dystroglycan. The role of a novel O-mannosyl-type oligosaccharide in the binding of alpha-dystroglycan with laminin. J Biol Chem 272: 2156-2162.

Cohn RD and Campbell KP (2000) The molecular basis of muscular dystrophies. Muscle Nerve 23:1456-1471.

Cohn RD, Henry MD, Michele DE, Barresi R, Saito F, Moore SA, Flanagan JD, Skwarchuk MW, Robbins ME, Mendell JR, Williamson RA, Campbell KP (2002) Disruption of DAG1 in differentiated skeletal muscle reveals a role for dystroglycan in muscle regeneration. Cell 110:639-648.

Connors NC, Adams ME, Froehner SC, Kofuji P (2004) The potassium channel Kir4.1 associates with the dystrophin-glycoprotein complex via a-syntrophin in glia. J Biol Chem 279:28387-28392.

Connors NC and Kofuji P (2002) Dystrophin Dp71 is critical for the clustered localization of potassium channels in retinal glial cells. J Neurosci 22:4321-4327.

46 Cote PD, Moukhles H, Lindenbaum M, Carbonetto S (1999) Chimaeric mice deficient in dystroglycans develop muscular dystrophy and have disrupted myoneural synapses. Nat Gen 23: 338-342.

Coutinho PM, Deleury E, Davies GJ, Henrissat B (2003) The evolving hierarchal family classification for glycosyltransferases. J Mol Biol 328: 307-317.

D'Souza VN, Nguyen TM, Morris GE, Karges W, Pillers DA, Ray PN (1995) A novel dystrophin isoform is required for normal retinal electrophysiology. Hum Mol Genet 4:837-42.

Dalloz G, Sarig R, Fort P, Yaffe D, Bordais A, Pannicke T, Grosche J, Mornet D, Reichenbach A, Sahel J, Nudel U, Rendon A (2003) Targeted inactivation of dystrophin gene product Dp71: phenotypic impact in mouse retina. Hum Mol Genet 12:1543-1554.

Danko I, Chapman V, Wolff JA (1992) The frequency of revertants in mdx mouse genetic models for Duchenne muscular dystrophy. Pediatr Res 32:128-131.

Deconinck AE, Potter AC, Tinsley JM, Wood SJ, Vater R, Young C, Metzinger L, Vincent A, Slater CR, Davies KE (1997a) Postsynaptic abnormalities at the neuromuscular junctions of utrophin-deficient mice. J Cell Biol 136:883-894.

Deconinck AE, Rafael JA, Skinner JA, Brown SC, Potter AC, Metzinger L, Watt DJ, Dickson JG, Tinsley JM and Davies KE (1997b) Utrophin-Dystrophin-Deficient Mice as a Model for Duchenne Muscular Dystrophy. Cell 90:717-727.

Dennis CL, Tinsley JM, Deconinck AE, Davies KE (1996) Molecular and functional analysis of the utrophin promoter. Nuleic Acids Res 24:1646-1652.

Dietzel I, Heinemann U, Hofmeier G and Lux HD (1980) Transient changes in the size of the extracellular space in the sensorimotor cortex of cats in relation to stimulus- induced changes in potassium concentration. Exp Brain Res 40: 432-439.

Ehmsen J, Poon E, Davies K (2002) The dystrophin-associated protein complex. J Cell Sci 115:2801-2803.

Emery A (2001) Duchenne muscular dystrophy or Meryon's disease. In The muscular dystrophies. Edited by Alan E.H. Emery. Oxford University Press pp. 55-68.

Emery A (2002) The muscular dystrophies. QJM 98:729-736.

Ervasti JM and Campbell KP (1991) Membrane organization of the dystrophin- glycoprotein complex. Cell 66:1121-1131.

47 Ervasti JM and Campbell KP (1993) A role for the dystrophin-glycoprotein complex as a transmembrane linker between laminin and actin. J Cell Biol 122:809-823.

Esapa CT, Benson MA, Schroder JE, Martin-Rendon E, Brockington M, Brown SC, Muntoni F, Kroger S, Blake DJ (2002) Functional requirements for fukutin- related protein in the Golgi apparatus. Hum Mol Genet 11:3319-3331.

Frigeri A, Gropper MA, Umenishi F, Kawashima M, Brown D, Verkman AS (1995) Localization of MIWC and GLJP water channel homologs in neuromuscular, epithelial and glandular tissues. J Cell Sci 108:2993-3002.

Frigeri A, Nicchia GP, Nico B, Quondamatteo F, Herken R, Roncali L, Svelto M (2001) Aquaporin-4 deficiency in skeletal muscle and brain of dystrophic mdx mice. FASEB J 15:90-98.

Fukuyama Y, Osawa M, Suzuki H (1981) Congenital muscular dystrophy of the Fukuyama type - clinical, genetic and pathological considerations. Brain Dev 3: 1-29.

Geyer H and Geyer R (2006) Strategies for analysis of glycoprotein glycosylation. Biochim and Biophys Acta 1764: 1853-1869.

Grady RM, Teng H, Nichol MC, Cunningham JC, Wilkinson RS, Sanes JR (1997) Skeletal and cardiac myopathies in mice lacking utrophin and dystrophin: a model for Duchenne muscular dystrophy. Cell 90:729-738.

Grady RM, Wozniak DF, Ohlemiller KK, Sanes JR (2006) Cerebellar synaptic defects and abnormal motor behavior in mice lacking alpha- and beta-dystrobrevin. J Neurosci 26:2841-2851.

Gregorevic P, Blankinship MJ, Allen JM, Crawford RW, Meuse L, Miller DG, Russell DW, Chamberlain JS (2004) Systemic delivery of genes to striated muscles using adeno-associated viral vectors. Nat Med 10:828-834.

Grewal PK and Hewitt JE (2002) Mutation of Large, which encodes a putative glycosyltransferase, in an animal model of muscular dystrophy. Biochim Biophys Acta 1573:216-224.

Grewal PK, Holzfeind PJ, Bittner RE, Hewitt JE (2001) Mutant glycosyltransferase and altered glycosylation of cc-dystroglycan in the myodystrophy mouse. Nat. Genet 28: 151-154.

Grewal PK, McLaughlan JM, Moore CJ, Browning CA, Hewitt JE (2005) Characterization of the LARGE family of putative glycosyltransferases associated with dystroglycanopathies. Glycobiol 15:912-923.

48 Guadagno E and Moukhles H (2004) Laminin-induced aggregation of the inwardly rectifying potassium channel, Kir4.1, and the water-permeable channel, AQP4, via a dystroglycan-containing complex in astrocytes. Glia 47:138-149.

Haenggi T and Fritschy JM (2006) Role of dystrophin and utrophin for the assembly and function of the dystrophin glycoprotein complex in non-muscle tissue. Cell. Mol. Life. Sci. 63: 1614-1631.

Hallmann J, Horn N, Selg M, Wendler O, Pausch F, Sorokin LM (2005) Expression and function of laminins in the embryonic and mature vasculature. Phys Rev 85:979- 1000.

Hanks SK and Hunter T (1995) Protein kinases 6. The eukaryotic protein kinase superfamily: kinase (catalytic) domain structure and classification. FASEB J 9: 576-596.

Haselbeck A and Tanner W (1983) O-glycosylation in Saccharomyces cerevisiae is initiated at the endoplasmic reticulum. FEBS Lett 158: 335-338.

Hayashi YK, Ogawa M, Tagawa K, Noguchi S, Ishihara T, Nonaka I, Arahata K (2001) Selective deficiency of alpha-dystroglycan in Fukuyama-type congenital muscular dystrophy. Neurology 57:115-121.

Hibino H, Fujita A, Iwai K, Yanada M and Kurachi Y (2004). Differential assembly of inwardly rectifying K+ channel subunits, Kir4.1 and Kir5.1, in brain astrocytes. J Biol Chem 279: 44065-44073.

Higashi K, Fujita A, Inanobe A, Tanemoto M, Doi K, Kubo T, Kurachi Y (2001) An inwardly rectifying K(+) channel, Kir4.1, expressed in astrocytes surrounds synapses and blood vessels in brain. Am J Physiol Cell Physiol 281:C922-C931.

Higginson JR and Winder SJ (2005) Dysroglycan: a multifunctional adapter protein. Biochem Soc Trans 33: 1254-1255.

Hille B (1992) Ionic channels in excitable membranes. 2nd ed. Sutherland, Sinauer. p.127-130.

Holt KH, Crosbie RH, Venzke DP, Campbell KP (2000) Biosynthesis of dystroglycan: Processing of a precursor peptide. FEBS Lett 468:79-83.

Holthoff K and Witte OW (1996) Intrinsic optical signals in rat neocortical slices measured with near-infrared dark-field microscopy reveal changes in extracellular space. J Neurosci 16: 2740-2749.

Holthoff K and Witte OW (2000) Directed spatial potassium redistribution in rat neocortex. Glia 29:288-292.

49 Holzfeind PJ, Ambrose HJ, Newey SE, Nawrotzki RA, Blakse DJ, Davies KE (1999) Tissue-selective expression of alpha-dystrobrevin is determined by multiple promoters. J Biol Chem 274:6250-6258.

Holzfeind PJ, Grewal PK, Reitsamer HA, Kechvar J, Lassmann H, Hoeger H, Hewitt JE, Bittner RE (2002) Skeletal, cardiac and tongue muscle pathology, defective retinal transmission, and neuronal migration defects in the Large(myd) mouse defines a natural model for glycosylation-deficient muscle-eye-brain disorders. Hum Mol Genet 11:2673-2687.

Howard PL, Dally GY, Wong MH, Ho A, Weleber RG, Pillers DM, Ray PN (199.8) Localization of dystrophin isoform Dp71 to the inner limiting membrane of the retina suggests a unique functional contribution of Dp71 in the retina. Hum Mol Genet 9:1385-1391.

Huang X, Poy F, Zhang R, Joachimiak A, Sudol M, Eck MJ (2000) Structure of a WW domain containing fragment of dystrophin in complex with (3-dystroglycan. Nat Struct Biol 7:634-638.

Hugnot JP, Gilgenkrantz H, Vincent N, Chafey P, Morris GE, Monaco AP, Berwald- Netter Y, Koulakoff A, Kaplan JC, Kahn A, Chelly J (1992) Distal Transcript of the Dystrophin Gene Initiated from an Alternative First Exon and Encoding a 75- kDa Protein Widely Distributed in Nonmuscle Tissues. PNAS 89:7506-7510.

Ibraghimov-Beskrovnaya O, Ervasti JM, Leveille CJ, Slaughter CA, Sernett S.W, Campbell KP (1992) Primary structure of dystrophin-associated glycoproteins linking dystrophin to the extracellular matrix. Nature 355:696-702.

Im WB, Phelps SF, Copen EH, Adams EG, Slightom JL, Chamberlain JS (1996) Differential expression of dystrophin isoforms in strains of mdx mice with different mutations. Hum Mol Genet 5:1149-1153.

Kachinsky AM, Froehner SC, Milgram SL (1999) A PDZ-containing scaffold related to the dystrophin complex at the basolateral membrane of epithelial cells. J Cell Biol 145:391-402.

Kanagawa M and Toda T (2006) The genetic and molecular basis of muscular dystrophy: roles of cell-matrix linkage in the pathogenesis. J Hum Gen 51:915-926.

Kanagawa M, Michele DE, Satz JS, Barresi R, Kusano H, Sasaki T, Timpl R, Henry MD, Campbell KP (2005) Disruption of perlecan binding and matrix assembly by post- translational or genetic disruption of dystroglycan function. FEBS Lett 579:4792- 4796.

50 Kano H, Kobayashi K, Herrman R, Tachikawa M, Manya H, Nishino I, Nonaka I, Straub V, Talim B, Voit T, Topaloglu H, Endo T, Yoshikawa H, Toda T (2002) Deficiency of alpha-dystroglycan in muscle-eye-brain disease. Biochem Biophys Res Commun 291:1283-1286.

Kim DS, Hayashi YK, Matsumoto H, Ogawa M, Nogushi S, Murakami N, Sakuta R, Mochizuki M, Michele DE, Campbell KP, Nonaka I, Nishino I (2004) POMT1 mutation results in defective glycosylation and loss of laminin-binding activity in alpha-DG. Neurology 62:1009-1011.

King LS, Kozono D, Agre P (2004) From structure to disease: the evolving tale of aquaporin biology. Nat Rev Mol Cell Biol 5:687-698.

Knuesel I, Bornhauser BC, Zuellig RA, Heller F, Schaub MC, Fritchy JM (2000) Differential expression of utrophin and dystrophin in CNS neurons: an in situ hybridization and immunohistochemical study. J Comp Neurol 422:594-611.

Knuesel I, Mastrocola M, Zuellig RA, Bornhauser B, Schaub MC and Fritschy JM (1999)

Short communication: altered synaptic clustering of GABAA receptors in mice lacking dystrophin (mdx). Eur. J. Neurosci. 11: 4457-4462.

Kobayashi K, Nakahori Y, Miyake M, Matsumura K, Kondo-Iida E, nomura Y, Segawa M, Yoshioka M, Saito K, Osawa M, Hamano K, Sakakihara Y, Nonaka I, Nakagome Y, Kanazawa I, Nakamura Y, Tokunaga K, Toda T (1998) An ancient retrotransposal insertion causes Fukuyama-type congenital muscular dystrophy. Nature 394:388-392.

Koenig M, Hoffman EP, Bertelson CJ, Monaco AP, Feener C, Kunkel LM (1987) Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organisation of the DMD gene in normal and affected individuals. Cell 50:509-517.

Kofuji P, Cellen P, Zahs KR, Surbeck LW, Lester HA, Newman EA (2000) Genetic inactivation of an inwardly rectifying potassium channel (Kir4.1 subunit) in mice: phenotypic impact in retina. J Neurosci 20:5733-5740.

Kofuji P and Newman EA (2004) Potassium buffering in the central nervous system. Neurosci 129:1045-1056.

Lee Y, Kameya S, Cox GA, Hsu J, Hicks W, Maddatu TP, Smith RS, Naggert JK, Peachey NS and Nishina PM (2005) Occular abnormalities in the Large(myd) and Large(vls) mice, spontaneous models for muscle, eye and brain diseases. Mol Cell Neurosci 30:160-172.

51 Lehle L, Tanner W and Strahl S (2006) Protein glycosylation, conserved from yeast to man: a model organism helps elucidate congenital human diseases. Angew Chem Int Ed 45: 6802-6818.

Leonoudakis D, Conti LR, Anderson S, Radeke CM, McGuire LM, Adams ME, Froehner SC, Yates JR 3rd and Vanderberg CA (2004) Protein trafficking and anchoring complexes revealed by proteomic analysis of inward rectifier potassium channel (Kir2.x)-associated proteins. J Biol Chem 279:22331-22346.

Levi S, Grady RM, Henry MD, Campbell KP, Sanes JR and Craig AM (2002) Dystroglycan is selectively associated with inhibitory GABAergic synapses but is dispensable for their differentiation. J. Neurosci. 22: 4272-4285.

Lidov HG, Selig S, Kunkel LM (1995) Dpl40: a novel 140 kDa CNS transcript from the dystrophin locus. Hum Mol Genet. 4:329-335.

Lis H and Sharon N (1993) Protein glycosylation. Structural and functional aspects. Eur J Biochem 218: 1-27.

Longman C, Brockington M, Torelli S, Jimenez-Mallebrera C, Kennedy C, Khalil N, Feng L, Saran RK, Voit T, Merlini L, Sewry CA, Brown SC, Muntoni F (2003) Mutations in the human LARGE gene cause MDC1D, a novel form of congenital muscular dystrophy with severe mental retardation and abnormal glycosylation of alpha-dystroglycan. Hum Mol Genet 12:2853-2861.

Lumeng G, Phelps S, Crawford GE, Walden PD, Barald K, Chamberlain JS (1999) Interactions between beta 2-syntrophin and a family of microtubule-associated serine/threonine kinases. Nat Neurosci 2:611-617.

Ma T, Song Y, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS (1999) Defective secretion of saliva in transgenic mice lacking aquaporin-5 water channels. J Biol Chem 274:20071-20074.

Ma T, Song Y, Yang B, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS (2000) Nephrogenic diabetes insipidus in mice lacking aquaporin-3 water channels. PNAS 97:4386-4391.

Ma T, Yang B, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS (1998) Severely impaired urinary concentrating ability in transgenic mice lacking aquaporin-1 water channels. J Biol Chem 273:4296-4299.

Matsumura K, Ervasti JM, Ohlendieck K, Kahl SD, Campbell KP (1992) Association of dystrophin-related protein with dystrophin-associated proteins in mdx mouse muscle. Nature 360:588-591.

52 Michele DE, Barresi R, Kanagawa M, Saito F, Cohn RD, Satz JS, Dollar J, Nishino I, Kelley Rl, Somer H, Straub V, Mathews KD, Moore SA, Campbell KP (2002) Post-translational disruption of dystroglycan-ligand interactions in congenital muscular dystrophies. Nature 418:417-422.

Moore SA, Saito F, Chen J, Michele DE, Henry MD, Meesing A, Cohn RD, Ross-Barta SE, Westra S, Williamson RA, Hoshi T and Campbell KP (2002) Deletion of brain dystroglycan recapitulates aspects of congenital muscular dystrophy. Nature 418: 422-425.

Moukhles H and Carbonetto S (2001) Dystroglycan contributes to the formation of multiple dystrophin-like complexes in brain. J Neurochem 78:824-834.

Moukhles H, Roque R, Carbonetto S (2000) alpha-dystroglycan isoforms are differentially distributed in adult rat retina. J Comp Neurol 420:182-194.

Muntoni F, Brockington M, Blake DJ, Torelli S, Brown SC (2002) Defective glycosylation in muscular dystrophy. Lancet 360:1419-1421.

Nagelhus EA, Horio Y, Inanobe A, Fujita A, Haug FM, Nielsen S, Kurachi Y, Ottersen OP (1999) Immunogold evidence suggests that coupling of K+ siphoning and water transport in rat retinal Muller cells is mediated by a coenrichment of Kir4.1 and AQP4 in specific membrane domains. Glia 26:47-54.

Nawrotzki R, Loh NY, Ruegg MA, Davies KE, Blake DJ (1998) Characterization of alpha-dystrobrevin in muscle. J Cell Sci 111:2595-2605.

Neely JD, Amiry-Moghaddam M, Ottersen OP, Froehner SC, Agre P, Adams ME (2001) Syntrophin-dependent expression and localization of Aquaporin-4 water channel protein. PNAS 98:14108-14113.

Nielsen S, Nagelhus EA, Amiry-Moghaddam M, Bourque C, Agre P and Ottersen OP (1997) Specialized membrane domains for water transport in glial cells: high- resolution immunogold cytochemistry of aquaporin-4 in rat brain. J Neurosci 17:171-180.

Niermann H, Amiry-Moghaddam M, Holthoff K, Witte OW, Ottersen OP (2001) A novel role of vasopressin in the brain: modulation of activity-dependent water flux in the neocortex. J Neurosci 21:3045-3051.

Noel G, Belda M, Guadagno E, Micoud J, Klocker, Moukhles H (2005) Dystroglycan and Kir4.1 coclustering in retinal Muller glia is regulated by laminin-1 and requires the PDZ-ligand domain of Kir4.1. J Neurochem 94:691-702.

Ohlendieck K and Campbell KP (1991) Dystrophin-associated proteins are greatly reduced in skeletal muscle from mdx mice. J Cell Biol 115:1685-1694.

53 Ozawa E, Mizuno Y, Hagiwara Y, Sasaoka T, Yoshida M (2005) Molecular and cell biology of the sarcoglycan complex. Muscle Nerve 32:563-576.

Papadopoulus MC, Krishna S, Verkman AS (2002) Aquaporin Water Channels and Brain Edema. The Mount Sinai Journal of Medicine 69:242-248.

Perkins KJ and Davies KE (2002) The role of utrophin in the potential therapy of Duchenne muscular dystrophy. Neuromuscul Disord 12:S78-S89.

Petrof BJ, Shrager JB, Stedman HH, Kelly AM, Sweeney HL (1993) Dystrophin protects the sarcolemma from stresses developed during . PNAS 90: 3710-3714.

Puwarawuttipanit W, Bragg AD, Frydenlund DS, Mylonakou MN, Nagelhus EA, Peters MF, Kotchabhakdi N, Adams ME, Froehner SC, Haug FM, Ottersen OP, Amiry- Moghaddam M (2006) Differential effect of alpha-syntrophin knockout on aquaporin-4 and Kir4.1 expression in retinal macroglial cells in mice. Neurosci 137:165-175.

Reed PW, Mathews KD, Mills KA, Bloch RJ (2004) The sarcolemma in the Large(myd) mouse. Muscle Nerve 30:585-595.

Ruppersberg JP (2000) Intracellular regulation of inward rectifier K+ channels. Pflugers Arch 441:1-11.

Rybakova IN, Patel JR, Davies KE, Yurchenco PD, Ervasti JM (2002) Utrophin binds laterally along actin filaments and can couple costameric actin with sarcolemma when overexpressed in dystrophin-deficient muscle. Mol Cell Biol 13:1512-1521.

Santavuori P, Somer H, Sainio K, Rapola J, Kruus S, Nikitin T, Ketonen L, Leisti J. (1989) Muscle-eye-brain disease (MEB). Brain Dev 11:147-153.

Schultz J, HoffMuller U, Krause G, Ashurst J, Macias MJ, Schmieder P, Schneider- Mergener J, Oschkinat H (1998) Specific interactions between the syntrophin PDZ domain and voltage-gated sodium channels. Nat Struct Biol 5:19-24.

Sciandra F, Schneider M, Giardina B, Baumgartner S, Petrucci TC, Brancaccio A (2001) Identification of the p-dystroglycan binding epitope within the C-terminal region of a-dystroglycan. Eur J Biochem 268:4590-4597.

Sentandreu R and Northcote DH (1969) The characterization of oligosaccharides attached to threonine and serine in a mannan glycopeptide obtained from the cell wall of yeast. Carbohydr Res 10: 584-585.

54 Skwarchuk MW, Robbins ME, Mendell JR, Williamson RA, Campbell KP (2002) Disruption of DAG1 in differentiated skeletal muscle reveals a role for dystroglycan in muscle regeneration. Cell 110:639-648.

Song Y and Verkman AS (2001) Aquaporin-5 dependent fluid secretion in airway submucosal glands. J Biol Chem 276:41288-41292.

Sontheimer H (1994) Voltage-dependent ion channels in glial cells. Glia 11:156-172.

Takeda S, Kondo M, Sasaki J, Kurahashi H, Kano H, Arai K, Misaki K, Fukui T, Kobayashi K, Tachikawa M, Imamura M, Nakamura Y, Shimizu T, Murakami T, Sunada Y, Fujikado T, Matsumura K, Terashima T, Toda T (2003) Fukutin is required for maintenance of muscle integrity, cortical histiogenesis and normal eye development. Hum Mol Genet 12:1449-1459.

Talts JF, Andac Z, Gohring W, Brancaccio A, Timpl R (1999) Binding of the G domains of laminin alpha 1 and alpha2 chains and perlecan to heparin, sulfatides, alpha- dystroglycan and several extracellular matrix proteins. Embo J 18:863-870.

Tanabe Y, Esaki K, Nomura T (1986) Skeletal muscle pathology in X chromosome- linked muscular dystrophy (mdx) mouse. Acta Neuropathologica 69:91-95.

Taniguchi M, Kurahashi H, Noguchi S, Fukudome T, Okinaga T, Tsukahara T, Tajima Y, Ozono K, Nishino I, Nonaka I, Toda T (2006) Aberrant neuromuscular junctions and delayed terminal muscle fiber maturation in a-dystroglycanopathies. Hum Mol Gen 15:1279- 1289.

Tinsley J, Deconinck N, Fisher R, Kahn D, Phelps S, Gillis JM, Davies K (1998) Expression of full-length utrophin prevents muscular dystrophy in mdx mice. Nat Med 4:1441-1444.

Tinsley JM, Blake DJ, Roche A, Faribrother U, Riss J, Byth BC, Knight AE, Kendrick- Jones J, Suthers GK, Love DR, Edwards YH, Davies KE (1992) Primary structure of dystrophin-related protein. Nat 360:591-593.

Torelli S, Brown SC, Brockington M, Dolatshad NF, Jimenez C, Skordis L, Feng LH, Merlini L, Jones DH, Romero N, Wewer U, Voit T, Sewry CA, Noguchi S, Nishino I, Muntoni F (2005) Sub-cellular localisation of fukutin related protein in different cell lines and in the muscle of patients with MDC1C and LGMD2I. Neuromuscul Disord 15:836-843.

Toda T, Kobayashi K, Takeda S, Sasaki J, Kurahashi H, Kano H, Tachikawa M, Wang F, Nagai Y, Taniguchi K, Taniguchi M, Sunada Y, Terashima T, Endo T, Matsumura K. 2003. Fukuyama-type congenital muscular dystrophy (FCMD) and alpha-dystroglycanopathy. Congenit Anom (Kyoto) 43:97-104.

55 Ueda H., Gohdo T, Ohno S (1998) Beta-dystroglycan localization in the photoreceptor and Miiller cells in the rat retina revealed by immunoelectron microscopy. J Histochem Cytochem 46:185-191.

Vainzof M, Moreira ES, Ferraz G, Passos-Bueno MR, Marie SK, Zatz M (1999) Further evidence for the organisation of the four sarcoglycans proteins within the dystrophin-glycoprotein complex. Eur J Hum Genet 7:251-254.

Vajsar J and Schachter H (2006) Walker-Warburg Syndrome. Orphanet J Rare Dis 1:29- 33.

VanderVen BC, Harder JD, Crick DC and Belisle JT (2005) Export-mediated assembly of mycobacterial glycoproteins parallels eukaryotic pathways. Science 309: 941- 943.

Verkman AS (2002) Physiological importance of aquaporin water channels. Annals of Medicine 34:192-200.

Way M, Pope B, Cross RA, Kendrick-Jones J, Weeds AG (1992) Expression of the N- terminal domain of dystrophin in E. coli and demonstration of binding to F-actin. FEBS Lett 301:243-245.

Wilier T, Prados B, Falcon-Perez JM, Renner-Miiller I, Przemeck GKH, Lommel M, Coloma A, Valero MC, Hrabe de Angelis M, Tanner W, Wolf E, Strahl S, Cruces J (2004) Targeted disruption of the Walker-Warburg syndrome gene Pomtl in mouse results in embryonic lethality. PNAS 101:14126-14131.

Williams JC, Armesilla AL, Mohamed TMA, Hagarty CL, Mclntyre FH, Schomburg S, Zaki AO, Oceandy D, Cartwright EJ, Buch MH, Emerson M, Neyses L (2006) The sarcolemmal calcium pump, a-1 syntrophin, and neuronal nitric-oxide synthase are parts of a macromolecular protein complex. J Biol Chem 281: 23341-23348.

Williamson RA, Henry MD, Daniels KJ, Hrstka RF, Lee JC, Sunada Y, Ibraghimov- Beskrovnaya O, Campbell KP (1997) Dystroglycan is essential for early embryonic development: disruption of Reichert's membrane in Dag 1-null mice. Hum Mol Genet 6:831-841.

Winder SJ (2001) The complexities of dystroglycan. Trends Biochem Sci 26:118-124.

Winder SJ, Gibson TJ, Kendrick-Jones J (1995) Dystrophin and Utrophin: the missing links. FEBS Letts 369:27-33.

56 Yamada H, Chiba A, Endo T, Kobata A, Anderson LV, Hori H, Fukuta-Ohi H, Kanaza I, Campbell KP, Shimizu T and Matsumura K (1996) Characterization of dp6troglycan-laminin interaction in peripheral nerve. J Neurochem 66: 1518- 1524.

Yang B, Gillespie A, Carlson EJ, Epstein C, Verkman AS (2001) Neonatal mortality in an aquaporin-2 knock-in mouse model of recessive nephrogenic diabetes insipidus. J Biol Chem 276:2775-2779.

Yoshida A, Kobayashi K, Manya H, Taniguchi K, Kano H, Mizuno M, Inazu T, Mitsuhashi H, Takahashi S, Takeuchi M, Herrmann R, Straub V, Talim B, Voit T, Topaloglu H, Toda T, Endo T (2002) Muscular dystrophy and neuronal migration disorder caused by mutations in a glycosyltransferase, POMGnTI. Dev Cell 1:717-724.

Yue Y, Liu M, Duan D (2006) C-Terminal-Truncated Microdystrophin Recruits Dystrobrevin and Syntrophin to the Dystrophin-Associated Glycoprotein Complex and Reduces Muscular Dystrophy in Symptomatic Utrophin/Dystrophin Double-Knockout Mice. Mol Therapy 14:79-87.

Zaccaria ML, Di Tommaso F, Brancaccio A, Paggi P, Petrucci TC (2001) Dystroglycan distribution in adult mouse brain: a light and electron microscopy study. Neurosci 104:311-324.

Zubrzycka-Gaarn EE, Bulman DE, Karpati G, Burghes AH, Belfall B, Klamut HJ, Talbot J, Hodges RS, Ray PN, Worton RG (1988) The Duchenne muscular dystrophy gene product is localized in sarcolemma of human skeletal muscle. Nature 333:466-469.

57 2 FIRST MANUSCRIPT CHAPTER1

2.1 Alpha-dystroglycan hypoglycosylation disrupts the polarized distribution of potassium ion and water permeable channels in perivascular astrocytes of the Largemyd mouse

The dystroglycan associated protein (DAP) complex is a group of interacting proteins that link the cytoskeleton to the extracellular matrix (ECM). In muscle, where the complex has been most thoroughly examined, it is believed to maintain the structural integrity of muscle fibres by protecting myotubes from the shear stress arising from repeated cycles of contraction and relaxation (Zubrzycka-Gaarn et al., 1988; Petrof et al.,

1993). The axis of the complex is composed of the transmembrane protein (3- dystroglycan (p-DG) whose cytoplasmic domain interacts with the subplasmalemmal dystrophin (Jung et al., 1995) while its extracellular domain interacts with a-dystroglycan

(a-DG) (Ibraghimov-Beskrovnaya et al. 1992). Dystrophin in turn binds to F-actin of the cytoskeleton (Way et al., 1992) while a-DG binds to laminin (Ibraghimov-Beskrovnaya et al., 1992) and other ligands including agrin, perlecan and neurexin (Bowe et al., 1994;

Talts et al., 1999; Kanagawa et al., 2005). Several other proteins also comprise the DAP and their expression varies depending on cell or tissue type. In muscle, these include at least 6 transmembrane proteins (a-, p-, 8-, y-, 8- sarcoglycan and sarcospan) and numerous cytoplasmic proteins including syntrophin, dystrobrevin, syncoilin, caveolin-3,

Grb2 and nNOS (Ehmsen et al., 2002). Mutations of various members of the DAP

' A version of this chapter has been submitted for publication. Rurak J., Lui L., Bharat J. and Moukhles H. (2007) Alpha-dystroglycan hypoglycosylation disrupts the polarized distribution of potassium ion and water permeable channels in perivascular astrocytes of the Largemyd mouse. J. Neurochem.

58 complex underlie the pathogenesis of many muscular dystrophies, a class of congenital disorders characterized primarily by progressive muscle weakness and degeneration

(Ibraghimov-Beskrovnaya et al., 1992; Campbell 1995).

The ability of a-DG to bind ECM ligands such as laminin is dependent on a dense region of O-linked glycosylation in its mucin like domain (Ervasti and Campbell, 1993).

Several forms of muscular dystrophy, including Fukuyama congenital muscular dystrophy (FCMD), Muscle-Eye Brain disease (MEB), Walker-Warburg Syndrome

(WWS) and congenital muscular dystrophy type ID (MDC1D), have all been associated with a reduction in a-DG glycosylation and a concomitant loss of laminin binding

(Michele et al., 2002; Longman et al., 2003; Kim et al., 2004). Each of these disorders has been linked to a mutation in a single gene encoding a glycosyltransferase that may be involved in the O-linked glycosylation of a-DG (Kobayashi et al., 1998; Yoshida et al.,

2001; Beltran-Valero de Bernabe et al., 2002; Longman et al., 2003). Interestingly, in addition to their muscle phenotypes, these diseases have defects in brain and ocular development, the etiologies of which are poorly understood (Hayashi et al., 2001; Kano et al., 2002; Longman et al., 2003; Kim et al., 2004).

Dystroglycan and many members of the DAP complex are expressed in the central nervous systems where they form complexes in both neuronal and glial cells including astrocytes and retinal Muller cells (Blake et al., 1999; Moukhles et al., 2000; Moukhles and Carbonetto, 2001). Of particular interest, DG is enriched at boundaries between neural tissues and fluid compartments, namely at glial cell endfeet abutting blood vessels

59 of the brain and retina and at those facing the pia mater of the brain as well as at the inner limiting membrane of the retina apposed to the vitreous body (Ueda et al., 1998; Zaccaria et al., 2001). These domains are also enriched for both the inwardly rectifying potassium channel, Kir4.1, and the water permeable channel aquaporin 4, AQP4 (Nielsen et al.,

1997; Nagelhus et al., 1999; Higashi et al., 2001). Based on the highly polarized distribution of the Kir4.1 channel and studies using Kir4.1 knock out mice (Kofuji et al.,

2000), this channel is critical for the regulation of extracellular K+ ions in neural tissues

(Brew et al., 1986; Higashi et al., 2001). This process, called potassium buffering, consists of funneling excess extracellular K+ ions from regions of high neuronal activity to blood vessels via astrocytic cells (Newman and Reichenbach, 1996) and generates osmotic gradients (Dietzel et al., 1980). In addition, neuronal activation leads to a decrease in adjacent extracellular space suggesting water movement (Holthoff and Witte,

1996). These data together with the co-distribution of Kir4.1 and AQP4 channels to the same subcellular domains led to the suggestion that Kir4.1-mediated K+ ions buffering is associated with a concomitant flux of water through AQP4 (Holthoff and Witte, 2000;

Niermann et al., 2001). Recently, further evidence substantiating this showed that a redistribution of AQP4 away from perivascular glial endfeet was accompanied by a delayed potassium clearance in a-syntrophin null mice (Amiry-Moghaddam et al., 2003).

Multiple lines of evidence suggest that the integrity of the DAP complex is essential for the proper localization and function of Kir4.1 and AQP4. Indeed, mutations in the dystrophin gene result in a dramatic reduction of the expression of AQP4 and Kir4.1 at perivascular astrocytic endfeet (Frigeri et al., 2001; Vajda et al., 2002; Dalloz et al., 2003;

60 Nico et al., 2003; Connors and Kofuji, 2002; Nicchia et al., 2004). Furthermore, the deletion of a-syntrophin also results in a reduction of perivascular AQP4 expression

(Neely et al., 2001). Together, these data indicate that the polarized distribution of both

Kir4.1 and AQP4 is dependent on the DAP complex. In addition, in cultured astrocytes and retinal Muller cells, laminin induces the co-clustering of Kir4.1 and AQP4 along with the DAP complex (Guadagno and Moukfiles, 2004; Noel et al., 2005), suggesting that the ability of the DAP to bind ECM ligands is essential for the localization and stability of these channels.

In the present study, we set out to characterize the polarized distribution of Kir4.1,

AQP4 and components of the DAP complex in brain and retina of the Largemyd mouse.

This mouse model of glycosylation-deficient muscular dystrophy carries a spontaneous mutation of the Large 1 gene and presents with a progressive myopathy (Grewal and

Hewitt, 2002). Consistent with FCMD, MEB, WWS and MDC1D patients, this mouse also displays cerebral abnormalities characterized by aberrant neuronal migration throughout the brain and a severe hypoglycosylation of a-DG with a concomitant loss of laminin binding (Michele et al., 2002). We used immunofluorescence and immunoblot analyses to examine the alterations in the polarized distribution and expression of Kir4.1,

AQP4 and other components of the DAP complex in brain and retina of the Largemyd mouse. We hypothesized that the reduction of O-linked glycosylation in a-DG and resultant loss of laminin binding will compromise the polarized distribution pattern of

Kir4.1 and AQP4. In brain, our data show a loss of Kir4.1 and AQP4 channels as well as a- and pi-syntrophin isoforms from perivascular astrocytic endfeet. This loss reflects a

61 mislocalization of the channels, since immunoblot data reveal that their total expression levels remain unchanged in the Largemyd compared to control mice. However, in retina the perivascular distribution of Kir4.1, AQP4 and pi-syntrophin is retained. These data reveal that the mechanisms involved in the polarized distribution of these channels at glial cell endfeet differ between brain and retina.

2.2 Materials and Methods

2.2.1 Tissues

Homozygous and heterozygous Largemyd mice as well as wild type littermate mice used in this study were genotyped using PCR tail DNA. Once the genotype was determined the mice were deeply anesthetized and tissues of interest were dissected on ice. Brains and eyes embedded in Tissue-Tek (O.C.T. compound; Pelco International,

Redding, CA, USA) and fresh frozen in liquid nitrogen-cooled isopentane (-80°C) were generously provided by Dr. S. Carbonetto (McGill University, Montreal). Upon receipt, the tissues were immediately transferred to -80°C where they were maintained for later analysis. 12-15 |im horizontal brain and sagittal eye cryostat sections were mounted on

Superfrost/Plus glass slides (Fisher, Ottawa, ON, Canada) and processed for subsequent immunofluorescence. The genotype, sex and age of all animals used in this study are outlined in Table 2.1 below. Data presented in this manuscript reflect representative findings characteristic all animals studied.

62 Table 2.1 Summary of Experimental Animals: Sex, Genotype and Age

Mouse Number Sex Genotype Age

3.1a Male Large"7" null 10 months

4.2e Male Large+/+ wt 10 months

12.2f Male Large"7" null 3 months

12.2b Female Large+/+ wt 3 months

12.2h Male Large+/+ wt 3 months

3.2g Female Large"'" null 11.5 months

3.2e Female Large+/" 11.5 months heterozygous 2.2.2 Antibodies

Rabbit antisera to Kir4.1 and AQP4 were raised against residues 356-375 of rat

Kir4.1 and against rat GST AQP4 corresponding to residues 249-323, respectively

(Alomone Laboratories, Jerusalem, Israel). Rabbit antiserum to laminin was raised against purified mouse Engelbreth-Holm-Swarm (EHS) Sarcoma laminin and recognizes laminin al, pi and yl chains (a generous gift from Dr S. Carbonetto, McGill University,

Montreal, QC, Canada); rabbit antiserum to agrin was raised against purified agrin and recognizes all agrin isoforms (a generous gift from Dr M. Ferns, University of California,

Davis, CA) and rabbit antisera to a-syntrophin (SYN 17) and pi-syntrophin (SYN37) were raised against residues 191-206 of a-syntrophin and 220-240 of mouse pi-

syntrophin, plus an NH2-terminal cysteine (a generous gift from Dr. S. Froehner,

University of Washington, Seattle, WA). Mouse monoclonal antibody (mAb) to a-DG,

IIH6C4, was raised against rabbit skeletal muscle membrane (Upstate Cell Signaling

Solutions, Lake Placid, NY); mouse mAb to P-DG, 43DAG1/8D5, was raised against the

15 of the last 16 amino acids at the C-terminus of the human dystroglycan sequence

(Novocastra Laboratories, Newcastle-upon-tyne, UK); mouse mAb to P-actin was raised against a synthetic p-actin N-terminal peptide (Sigma-Aldrich, St. Louis, MO, USA). Rat mAb to heparan sulfate proteoglycan (HSPG) perlecan, A7L6, was raised against HSPG from mouse EHS tumor and reacts with perlecan domain 4 (Chemicon International, CA,

USA).

64 2.2.3 Immunofluorescence

Brain and retina sections were fixed by immersion in 4% (w/v) paraformaldehyde in 0.1 M phosphate buffer for 20 min followed by rinsing in phosphate-buffered saline

(PBS) 3x15 min. Sections were then incubated for 1 h at room temperature (20-22°C) in a solution containing 3% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA) and 0.2% Triton X-100. Double immunolabeling was performed by incubating the sections at room temperature for 1 h in the presence of primary antibodies against either

Kir4.1 (1:100), AQP4 (1:200), a-syntrophin (5 ug/ml 1:500), pi-syntrophin (5 ug/ml

1:100), agrin (1/700), laminin (1/1500) or a-DG (1/50) and perlecan (1/50).

Subsequently, they were rinsed with PBS (3x15 min) and incubated with Alexa Fluor

488 goat anti-rabbit IgG and Alexa Fluor 568 goat anti-rat or Alexa Fluor 488 goat anti- mouse IgM and Alexa Fluor 568 goat anti-rat for 1 h (1/200; Molecular Probes, Eugene,

ON, USA). Finally the sections were thoroughly rinsed with PBS and mounted on glass coverslips using Prolong Gold Antifade Reagent with or without DAPI (Invitrogen,

Burlington, ON, Canada). To confirm the specificity of the labeling, control sections were treated equivalently in the absence of primary antibodies.

2.2.4 Immunoblotting

Brain regions from Largemyd mice and littermate wild type controls containing frontal cerebral cortex and hippocampus were homogenized using a Dounce homogenizer in cold PBS containing 1 x complete protease inhibitor cocktail (Boehringer Mannheim,

Mannheim, Germany). The homogenates were centrifuged at 200xg for 5 min and the

65 pellet was solubilized in extraction buffer (25 mM Tris pH 7.4, 25 mM glycine and 150 mM NaCl) containing 1% Triton X-100, 1 x complete protease inhibitor cocktail and 5 mM EDTA. The samples were left on ice for 15 min with occasional vortexing followed by a 10 min centrifugation at 16 OOOxg. Extracted proteins were denatured by boiling the samples for 9 min in reducing sample buffer and then loaded on a 10% sodium dodecyl sulfate-polyacrylamide electrophoresis gels. The gels were electrotransferred to nitrocellulose membranes (Bio-Rad, Mississauga, ON, Canada) and the blots were probed with rabbit antisera to Kir4.1 (1/400) or AQP4 (1/1000) or with mouse mAbs to

P-DG (1/200) or P-actin (1/25000). Bound antibodies were detected using horseradish peroxidase-conjugated goat anti-rabbit IgG or goat anti-mouse IgG (1/2000; Jackson

ImmunoResearch, West Grove, PA, USA). Signals were visualized on Bioflex econo films (Interscience, Markham, ON, Canada) using chemiluminescence (Amersham

Biosciences, Buckinghamshire, UK).

2.2.5 RNA extraction and RT-PCR

Total RNA was extracted from 50 ug of previously snap-frozen mouse tissues in liquid nitrogen using micro-homogenizing device (VWR, Mississauga, ON, Canada) and

TRIZOL® reagent (Invitrogen) following the supplied protocol. The quality and integrity of RNA was assessed on formaldehyde agarose gel and concentration was determined using NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies, USA).

66 Total cDNA was prepared using Oligo dT(25) (custom made, Invitrogen), 1 ug of total

RNA from each sample and Superscript™ III Reverse Transcriptase (RT; Invitrogen) following the supplied protocol. The presence of Large 1, Large! and Large2 alternative spliced transcripts was assessed on 2 (xl of synthesized total cDNA of each tissue by PCR using the set of primers described by Grewal et al (2005) and Platinum®-Taq DNA polymerase (Invitrogen). A set of P-actin primers (Forward: 5'TCT ACG AGG GCT

ATG CTC TCC3', Reserse: 5'GGA TGC CAC AGG ATT CCA TAC3'), no-RNA and no-RT were used as controls.

2.3 Results

2.3.1 Abnormalities in the localization of basement membrane proteins in the Largemyd brain

Previous data reported a disruption of the glia limitans which leads to abnormal neuronal migration in the Largemyd brain (Michele et al., 2002). This disruption is reflected in a reduction of laminin (Fig. 2. IB) and perlecan (Fig. 2.1G) (Michele et al.,

2002; Kanagawa et al., 2005) as well as agrin labeling (Fig. 2.IF) in the basement membrane surrounding the surface of the brain. However, no apparent changes in laminin

(Fig. 2.IB), agrin (Fig. 2.IF), or perlecan (Fig. 2.1G) immunolabeling were seen at the vascular basement membrane in the Largemyd mouse compared to the wild type control

(Fig. 2.1 A, I and J). Interestingly, scattered puncta labeled for laminin and perlecan (Fig.

2.1M) have been reported in the neuropil of the Largemyd brain (Kanagawa et al., 2005).

67 Here we show that these puncta also label for agrin (Fig. 2.1L) and that they are particularly abundant in the frontal cerebral cortex where a large number of them appear as complex geometrical structures when examined at high magnification (Fig. 2.10).

These data indicate that despite a hypoglycosylation of a-DG in all regions and sites within the Largemyd brain (Michele et al., 2002), the disruption of basement membrane proteins localization is restricted to the glia limitans. This suggests that mechanisms implicating receptors other than the fully glycosylated a-DG such as integrins may be involved in the assembly of the vascular basement membrane.

68

Figure 2.1 Disruption of the glia limitans and accumulation of agrin and perlecan puncta in the Largemyd brain. Horizontal brain sections containing cerebral cortex and cerebellum from wild type (A) and Largemyd mice (B) were immunolabeled for laminin. Compared to the wild type brain (arrows), the glia limitans of the Largemyd brain shows reduced laminin immunolabeling (arrows). Horizontal brain sections containing cortex from wild type (C- E; I-K) and Largemyd mice (F-H; L-O) double immunolabeled for Agrin (C, F, I, L) and Perlecan (D, G, J, M) show reduced immunolabeling at the glia limitans (arrows in F and G) and presence of scattered puncta positive for both these proteins in the parenchyma of the cerebral cortex of the Largemyd mouse (L-O). Note that the perivascular basement membrane from both wild type (A, I and J) and Largemyd mice (B, F and G) is equally immunolabeled for laminin, agrin and perlecan. The merged images show the codistribution between agrin and perlecan (E, H, K and N). A high magnification of the boxed area in L is illustrated in O. Scale bar, 1 mm for A and B, 30 um for C-K and 10 pm for L-N.

70 2.3.2 Loss of perivascular localization of AQP4 and Kir4.1 in the Largemy brain

Previous studies have shown that Kir4.1 and AQP4 are concentrated in astrocytic endfeet around blood vessels both in brain and retina (Nielsen et al., 1997; Nagelhus et al., 1999; Poopalasundaram et al., 2000; Higashi et al., 2001; Neely et al., 2001; Connors and Kofuji, 2002; Guadagno and Moukhles, 2004) where they are co-distributed with a-

DG and P-DG (Guadagno and Moukhles, 2004; Noel et al., 2005). Laminin, a ligand that binds glycosylated moieties of a-DG, is highly expressed in vascular basement membrane (Tian et al., 1996; Fig. 2.1 A) and recent studies by Kanagawa et al., (2004) showed that its binding to a-DG occurs within the first half of the mucin-like domain and relies on the O-glycosylation of this domain by the glycosyltransferase Large 1. A disruption of a-DG binding to basement membrane ligands such as laminin, agrin and neurexin has been reported both in brain and muscle of the Largemyd mouse (Grewal et al., 2001; Michele et al., 2002). Similar results have been reported in patients with MEB,

FCMD and MCD1C (Michele et al., 2002; Longman et al., 2003). In addition, the targeting of the a-DG core protein, P-DG and dystrophin to astrocytic endfeet around blood vessels is impaired in the Largemyd brain indicating that a-DG targets proteins to functional sites through its interaction with basement membrane proteins (Michele et al.,

2002). In previous studies we have shown that the co-clustering of Kir4.1 and AQP4 with a-DG in astrocytes and Muller cells is induced by laminin (Guadagno and Moukhles,

2004; Noel et al., 2005). In light of these data, we asked whether the perivascular targeting of these channels is impaired in the Largemyd brain and assessed their mistargeting by immunofluorescence. We found that AQP4 was no longer expressed at

71 perivascular glia in all brain regions examined (Fig. 2.2B). Subsequently, we used perlecan as a marker for vascular basement membrane and found that while perlecan- positive (Fig. 2.2D and J) blood vessels were immunolabeled for AQP4 (Fig. 2.2C) and

Kir4.1 (Fig. 2.21) in wild type control brain, they were devoid of labeling for both types of channels in the Largemyd brain (Fig. 2.2F-H and L-N). As reflected by the perlecan vascular labeling, the number of blood vessels in the Largemyd brain (Fig. 2.2G and M) is comparable to that in the wild type control (Fig. 2.2D and J) indicating that the loss of labeling for the channels is not a result of a reduction in the density of blood vessels but of their mistargeting. At the glia limitans, a significant reduction in the labeling for both

AQP4 (Fig. 2.2B) and Kir4.1 was seen (Appendices Fig. 4.1). The immunoblot data show that total expression levels of both Kir4.1 and AQP4 channels are similar in the Largemyd and control brains (Fig. 2.3), supporting the hypothesis that a-DG hypoglycosylation and the consequent loss of ligand binding lead to Kir4.1 and AQP4 mistargeting to perivascular astrocytic endfeet.

72

Figure 2.2 AQP4 and Kir4.1 fail to localize to perivascular astrocyte endfeet in the Largemyd brain. Horizontal brain section containing cerebral cortex and cerebellum from wild type (A) and Largemyd mice (B) were immunolabeled for AQP4. Note that AQP4 immunolabeling is lost from perivascular sites in all brain regions of the Largemyd mouse. Horizontal brain sections containing cortex from wild type (C, D; I, J) and Largemyd (F, G; L, M) mice were double immunolabeled for AQP4 (C, F) and perlecan (D, G) or Kir4.1 (I, L) and perlecan (J, M). The perivascular labeling for AQP4 (F) and Kir4.1 (L) is undetectable in Largemyd compared to the wild type control brain (C, I). The merged images show the codistribution between AQP4 and perlecan (E) and Kir4.1 and perlecan (K). Scale bar, 1mm for A and B and 30 pm for C-N.

74 Figure 2.3 Total Kir4.1 and AQP4 expression levels are unchanged in the Largemy brain. Homogenates from Largemyd and wild type brain were immunoblotted sequentially for Kir4.1, AQP4, P-DG and P-actin. Immunoblot analysis showed specific reactions with antibodies to Kir4.1 (monomer, 42kDa; dimer, 84kDa and the trimer, over lOOkDa), AQP4 (32kDa) and P-DG (43kDa). P-Actin labeling shows equal protein loading in both lanes.

75 2.3.3 Loss of perivascular localization of a- and Pi-syntrophins in the Large' brain

We have previously shown that a-DG-ligand binding is involved in syntrophin clustering in Muller glial cell cultures (Noel et al., 2005). Other studies have demonstrated that a PDZ-mediated interaction with a-syntrophin is important in localizing and stabilizing AQP4 in specific membrane domains of perivascular glia

(Neely et al., 2001). Recent biochemical studies have shown that in brain astrocytes,

Kir4.1 co-purifies with DG and members of the DG-containing complex through a direct interaction of its PDZ ligand domain (-SNV) with the PDZ domain of a-syntrophin

(Connors et al., 2004; Leonoudakis et al., 2004). Furthermore, we have shown that a deletion of the -SNV C-terminal sequence in Kir4.1 prevents its laminin-induced co- clustering with a-DG in Muller glia, indicating that the PDZ-domain-mediated interaction is crucial for Kir4.1 clustering (Noel et al., 2005). To investigate the importance of a-DG glycosylation in targeting cytoplasmic components of the complex as well as their possible implication in the proper targeting of Kir4.1 and AQP4 to perivascular astrocytes, we examined the expression pattern of both a- and pi-syntrophins in the

Largemyd brain. We found that unlike in the wild type control brain (Fig. 2.4A and G), these syntrophin isoforms were no longer concentrated at the perivascular astrocytic endfeet in the Largemyd brain (Fig. 2.4D and J). Given that syntrophin interacts with

Kir4.1 and AQP4 (Connors et al., 2004; Leonoudakis et al., 2004), these data suggest that the mistargeting of the channels may be a consequence of syntrophin mistargeting.

76 Figure 2.4 a- and Pi-syntrophins fail to localize to perivascular astrocyte endfeet in the Largemyd brain. Horizontal brain sections from wild type (A, B; G, H) and Largemyd mice (D, E; J, K) were double immunolabeled for a-syntrophin (A, D) and perlecan (B, E) or pi- syntrophin (G, J) and perlecan (H, K). The merged images show the codistribution between a-syntrophin and perlecan (C) or Pi-syntrophin and perlecan (I). Scale bar, 30 um.

77 2.3.4 Expression of Largel and Large2 mRNAs in retina and impact of Largel mutation on Kir4.1 and AQP4 localization in the Largemy retina

In retina, Kir4.1 and AQP4 are concentrated both in perivascular glia and Miiller cell endfeet where they play a crucial role in K+ ion buffering and therefore participate in the retinal electrical activity. In previous studies we have shown that Kir4.1 and AQP4 undergo clustering upon laminin treatment suggesting that their polarized distribution is mediated through a-DG interaction with laminin (Guadagno and Moukhles, 2004; Noel et al., 2005). To investigate, in vivo, the role of a-DG glycosylation in this process we assessed the distribution of these channels in the Largemyd retina by immunofluorescence.

We first determined, by RT-PCR, the expression of Largel, Large2 and Large2 alternative splice variants in mouse retina in parallel with brain regions and other non neural tissues where the expression of these transcripts has already been established

(Grewal et al., 2005). Here we show that retina not only expresses Largel but also Large2 and its alternative transcripts (Fig. 2.5). Interestingly, unlike cortex and hippocampus that express only Largel, the cerebellum expresses the same Large transcripts as in the retina

(Fig. 2.5). Despite the expression of Large2 and Large2 alternative transcripts in retina, a-DG is hypoglycosylated, as shown by lack of perivascular immunofluorescence when using the IIH6C4 antibody directed against the glycosylated moieties of the protein (Fig.

2.6D). These data indicate that Large2 and Large2 alternative transcripts do not compensate for lack of Largel, and that Largel is the primary glycosyltransferase associated with a-DG glycosylation in retina. Furthermore, the majority of the laminin labeling is lost at the inner limiting membrane of the Largemyd retina (Fig. 2.71), suggesting that Largel-mediated glycosylation of a-DG plays a critical role in the proper

78 assembly of the laminin-rich basement membrane at this site. In contrast, laminin labeling associated with the vascular basement membrane is maintained in the Largemyd retina (Fig. 2.7F).

In parallel, we performed histological studies on eyes from 3 (Fig. 2.7D) and 10 month old Largemyd mice (Fig. 2.7G). Retinas from both age groups were detached from the retinal pigment epithelium and a profound disruption in the layering was observed

(Fig. 2.7E and H), consistent with data previously described by Lee et al., (2005). In addition, many rosette-like structures and bundles of axons were present throughout the retina (Fig. 2.7E and H). Radial Miiller cell processes spanning the ganglion cell- and inner plexiform layers in wild type control retina appear discontinuous and tortuous in

Largemyd retina as reflected by vimentin immunolabeling (Appendices Fig. 4.2). These phenotypes were as severe in 3 months as in 10 months old mutant mice arguing against their progressive nature.

Subsequently, we analyzed Kir4.1 and AQP4 localization by immunofluorescence and found that although there is a reduction in the labeling for these channels around blood vessels in the Largemyd compared to wild type retina, they remain substantially expressed at these sites (Fig. 2.8B, D). These data are in contrast with those in the

Largemyd brain where no detectable labeling for either Kir4.1 or AQP4 was found (Fig.

2.2F and L), showing that the molecular mechanisms underlying the polarized distribution of these channels in retina may be different from those in brain.

79 1000 — 750 — »• Largel 500 — ; iSIIII 250 —

1000 — 750 — 500 — • m Largel 250 —

1000_ m- « Spliced forms 250 — * of Large2

m= :j-Actin 250 —

Figure 2.5 Expression profile of Largel and Large2. RT-PCR of mouse Largel and Large2 mRNAs from retina, various brain regions, diaphragm, liver, kidney, lung, testis and liver. Note that like in the kidney and testis, retina as well as cerebellum express not only Largel but also Large2 and the Large2 alternative spliced forms.

80 Figure 2.6 a-dystroglycan is hypoglycosylated in the Largemyd retina. Retina sections from wild type (A-C) and Largemyd (D-F) mice were double immunolabeled using the IIH6 antibody against carbohydrate moieties of a-DG (A, D) and perlecan (B, E). Perlecan positive blood vessels are devoid of a-DG labeling in the Largemyd (asterisks, D-F) compared to the wild type control (asterisks, A-C). High magnifications of the boxed areas in A, B and D, E are illustrated as merged images in C and F. Scale bar, 25 pm. Anterior to posterior retina = bottom to top in all images

81 Figure 2.7 Disruption of the lamination and the inner limiting membrane in the Largemyd retina. Cryostat sections from wild type (A-C) and Largemyd eyes (D-I) were stained with haematoxylin and eosin. Retinas of both 3 months (D, E) and 10 months old mice (G, H) were detached from the retinal pigment epithelium (D and G) and their lamination was severely disrupted. In addition, axon bundles (arrows in E) and rosette-like structures (arrows in H) were seen. Laminin immunolabeling shows a profound disruption of the inner limiting membrane (ILM; arrows in I) and no apparent changes in the perivascular basement membrane in the Largemyd (asterisks in F) compared to wild type retina (asterisks in C). Scale bar, 250 pm for A, D and G and 25 pm for B, E, H, C, F and I. Nuclear layers stain deep purple with H&E. Anterior to posterior retina = left to right in all images. Inner nuclear layer: INL; outer nuclear layer: ONL.

82 Figure 2.8 Perivascular AQP4 and Kir4.1 are maintained in the Largemyd retina. Retina sections from wild type (A and C) and Largemyd mice (B and D) were immunolabeled for Kir4.1 (A and B) or for AQP4 (C and D). Note that a significant Kir4.1 (arrows, B) and AQP4 (arrows, D) labeling is maintained at perivascular astrocyte endfeet in the Largemyd retina. Scale bar, 25 pm. Anterior to posterior retina = top to bottom in all images. Inner nuclear layer: ENL; outer nuclear layer: ONL.

83 2.3.5 Loss of perivascular localization of a-syntrophin but not Pi-syntrophin in the Largemyd retina

Data from Connors and Kofuji, (2006) show that Kir4.1 and AQP4 bind to a- syntrophin in retina. However while this association is critical for AQP4 localization in brain, it is not in retina where AQP4 localization was only marginally decreased in perivascular astrocytes and Muller cell endfeet as assessed by immunoelectron microscopy in the a-syntrophin null mouse (Puwarawuttipanit et al., 2006). Interestingly,

Kir4.1 localization was unaffected both in brain and retina of these mice (Amiry-

Moghaddam et al., 2003; Frydenlund et al., 2004; Puwarawuttipanit et al., 2006). In the present study, we show that Kir4.1 and AQP4 polarized distribution to astrocyte endfeet was profoundly disrupted in brain but not in retina of the Largemyd mouse. In brain, this was accompanied by loss of perivascular labeling of both a- and pi-syntrophins. To see whether a similar response is found in the retina, we investigated the localization of both these syntrophins in the Largemyd and wild type control mice. Similarly to brain (Fig.

2.4D), a-syntrophin labeling in retina was no longer detected at perivascular glia (Fig.

2.9E) whereas pi-syntrophin labeling was still readily detected at these sites (Fig. 2.9M).

These data show that while a-syntrophin targeting to perivascular domains depends on a-

DG-ligand interaction, that of pi-syntrophin is independent of such an interaction.

Furthermore, persistence of the perivascular localization of Kir4.1 and AQP4 channels together with that of pi-syntrophin raises the possibility that in retina this syntrophin isoform may be involved in targeting these channels through a mechanism independent of both a-syntrophin and the fully glycosylated a-DG isoform.

84 Figure 2.9 pi- but not a-syntrophin is still expressed at perivascular astrocytes endfeet in the Largemyd retina. Retina sections from wild type (A-D; I-L) and Largemyd mice (E-H; M-P) were double immunolabeled for a-syntrophin (A, E) and perlecan (B, F) or Pl-syntrophin (I, M) and perlecan (J, N). Unlike a-syntrophin (E), pi-syntrophin (M) is still localized at perivascular astrocytes endfeet of the Largemyd retina. High magnifications of the boxed areas in B, F, J and N are illustrated as merged images in D, H, L and P. These confirm the codistribution between a-syntrophin and perlecan (D) and pi-syntrophin and perlecan (L, P). Scale bar, 25 pm. Anterior to posterior retina = bottom to top is all images.

85 2.4 Discussion

2.4.1 AQP4 and Kir4.1 redistribution in brain

The proper localization and coordinated activity of glial Kir4.1 and AQP4 is essential to normal neuronal function and is believed to underlie potassium ion homeostasis and water flux in brain (Brew et al., 1986; Higashi et al., 2001). Several studies suggest that the DAP complex is involved in the stability of Kir4.1 and AQP4 channels at astrocytic domains active in electrolyte balance and fluid movement (Neely et al., 2001; Connors and Kofuji, 2002; Vajda et al., 2002; Guadagno and Moukhles, 2004). Here we have used immunofluorescence to investigate the distribution pattern of these channels in the brain of the Largemyd mouse, a model of glycosylation-deficient forms of muscular dystrophy

(Michele et al., 2002). In agreement with previous findings, we show that AQP4 fails to localize at perivascular domains of the Largemyd brain (Michele et al., 2002).

Furthermore, Kir4.1 is also lost from these sites supporting our hypothesis that the inability of the complex to bind extracellular ligands due to a-DG hypoglycosylation compromises the distribution of both AQP4 and Kir4.1 channels. These findings are consistent with previous studies using animal models characterized by disruptions of members of the DAP complex other than DG. In fact, dystrophin mutant mice exhibit a dramatic reduction in perivascular expression of both Kir4.1 and AQP4 with no apparent disruption in the formation or density of the blood vessels themselves (Frigeri et al.,

2001; Connors and Kofuji, 2002; Vajda et al., 2002; Dalloz et al., 2003; Nicchia et al.,

2004). The mislocalization of channels caused by mutations of members of the DAP complex is not unique to Kir4.1 and AQP4 in the central nervous system. In fact, muscles

86 of dystrophic mice and patients present an abnormal aggregation of acetylcholine receptors at the neuromuscular junction (Levedakou et al., 2005). Similarly, a selective deletion of Schwann cells DG results in reduced voltage-gated sodium channels aggregation at the nodes of Ranvier (Saito et al., 2003). These findings, together with our data, provide compelling evidence that an unbreached continuity between the DAP complex and the ECM is essential to the localization and stability of both Kir4.1 and

AQP4 to perivascular domains of astrocytic endfeet and that any compromise, be it cytoplasmic or extracellular, causes their mislocalization.

There exists evidence that both Kir4.1 and AQP4 associate with the DAP complex via the intracellular protein a-syntrophin. Both channels contain a consensus C-terminal SXV motif that binds to the PDZ domain of a-syntrophin (Neely et al., 2001; Connors et al.,

2004) and the a-syntrophin null mouse show loss of AQP4 from perivascular astrocytic membranes in brain (Amiry-Moghaddam et al., 2003). Interestingly, we show here that a- syntrophin (Fig 2.4) and AQP4 (Fig 2.2) are no longer localized to perivascular astrocytic endfeet in the brain of the Largemyd mouse. These data not only substantiate the hypothesis that a-syntrophin provides a link between the DAP complex and this channel but they also show that severance of the link between the complex and the ECM adversely affects the stability of a-syntrophin at perivascular sites. In contrast with muscle, where most members of the DAP complex remain localized at the sarcolemma, dystrophin, P-DG (Michele et al., 2002) as well as a-syntrophin (Fig. 2.4D) fail to localize at astrocytic endfeet in the brain of the Largemyd mouse. This difference between brain and muscle is surprising especially given that the interaction between a-DG and the

87 ECM is impaired in both tissues. Nevertheless, it reveals that the functional implications of such an interaction are spatially distinct.

While the immunofluorescence data show lack of Kir4.1 and AQP4 labeling at perivascular astrocytes in the Largemyd brain, immunoblot analyses show no reduction in their total expression levels (Fig. 2.3) revealing a redistribution of these channels to other astrocytic domains. This is in agreement with data reported for Kir4.1 in retina of mdx3cv and Dp71 null mice (Frigeri et al., 2001; Connors and Kofuji, 2002; Dalloz et al., 2003) and for AQP4 in the brain of a-syntrophin null mouse (Neely et al., 2001). Interestingly, other components of the DAP complex including P-DG and dystrophin fail to localize at perivacular astrocytes despite normal total expression levels (Fig. 2.3; Michele et al.,

2002). These data combined with our current findings provide evidence that the distribution of Kir4.1, AQP4 and components of the DAP complex to specialized membrane domains such as astrocyte endfeet relies on a coordinated interaction between all of these proteins.

The physiological impact of the redistribution of AQP4 in a-syntrophin null mice has been correlated with delayed K+ ion clearance in brain indicating that water flux through perivascular AQP4 is required for efficient removal of K+ after neuronal activation

(Amiry-Moghaddam et al., 2003). Likewise, the Kir4.1 knock out mouse exhibits reduced potassium currents and potassium buffering capacity (Kofuji et al., 2000). In models of pathogenic brain edema induced by acute water intoxication and focal ischemic stroke, AQP4 deletion has been shown to significantly reduce the progression of

88 cytotoxic edema (Manley et al., 2000). In light of these data and our findings showing a redistribution of Kir4.1 and AQP4, it is reasonable to speculate that the Largemyd mouse may present with similar physiological effects under specific experimental conditions.

2.4.2 AQP4 and Kir4.1 distribution in retina

Similarly to brain, retinal glial cells express Kir4.1, AQP4 and the DAP complex in a highly polarized fashion so as to facilitate potassium buffering (Kofuji and Connors,

2003). In fact most of what is known about the involvement of these channels in potassium siphoning has come to light from studies using retinal Muller glia. At least two glycoforms of a-DG are expressed in retina (Moukhles et al., 2000), however the expression of the Large glycosyltransferase isoforms that may be associated with a-DG glycosylation has not been documented. We therefore determined the expression of Large isoforms by RT-PCR and found that in addition to Large 1, Large2 and its alternative transcripts are also expressed in retina. Immunolabeling for a-DG using the IIH6 antibody showed that this protein is no longer expressed at perivascular glia, indicating that despite the expression of other isoforms of Large, none are capable of compensating for the loss of Large 1. Subsequently, we investigated the distribution of Kir4.1 and AQP4 in retina by immunofluorescence. In contrast with our findings in brain, our data revealed a significant labeling for these channels at perivascular glia in the Largemyd retina.

Furthermore, while both a- and pi-syntrophins fail to localize at astrocyte endfeet in the

Largemyd brain, staining for pi-syntrophin comparable to that in wild type retina is found in the Largemyd retina.

89 The differential effects of the Largemy mutation on the distribution of Kir4.1 and

AQP4 channels between brain and retina provides compelling evidence that the underlying mechanisms involved in their targeting are different between these tissues.

The presence of perivascular pi-syntrophin in the Largemyd retina raises the possibility that this syntrophin isoform may compensate for a-syntrophin loss at perivascular astrocytes and therefore target the channels appropriately at these sites. This is in accordance with the data of Puwarawuttipanit et al., (2006) showing that the entire pool of perivascular Kir4.1 and 30% of AQP4 are still expressed in the retina of the a- syntrophin null mouse. This indicates that the localization of most of AQP4 but not

Kir4.1 depends on a-syntrophin (Puwarawuttipanit et al., 2006). Of particular interest,

(31 -syntrophin is still expressed in retina of the a-syntrophin null mouse, suggesting that it may be involved in localizing Kir4.1 as well as a pool of AQP4 (Puwarawuttipanit et al.,

2006).

The primary objective of this study was to determine the impact of the Largemyd mutation on the perivascular localization of Kir4.1 and AQP4 in brain and retina. In brain, we show that the hypoglycosylation of a-DG secondary to mutation of Largel is associated with loss of both AQP4 and Kir4.1 as well as a- and Pi-syntrophins from perivascular domains. In retina however, significant labeling for both channel types as well as pi-syntrophin is present at perivascular glia. These data show that Kir4.1 and

AQP4 may be anchored and stabilized at glial cell endfeet via a pi-syntrophin dependent and a-DG independent mechanism and reveal the existence of a novel pi-syntrophin-

90 containing complex lacking DG. While the components of this complex remain to be identified, our findings point to a spatial variability in the composition of the syntrophin- containing complex within different neural tissues that may have distinct functional implications.

2.5 Acknowledgements

The authors thank Dr. S. Froehner for providing antibodies to syntrophins and Dr S.

Carbonetto and Nadia Melian for providing the Largemyd brains and eyes. This work was supported by grants from the Canadian Institutes of Health Research (CIHR), Muscular

Dystrophy Canada (MDC), Amyotrophic Lateral Sclerosis (ALS) Society of Canada and the Natural Sciences and Engineering Research Council of Canada (NSERC) to H.M.

91 2.6 Works Cited

Amiry-Moghaddam M., Williamson A., Palomba M., Eid T., de Lanerolle N. C, Nagelhus E. A., Adams M. E., Froehner S. C, Agre P. and Ottersen O. P. (2003) Delayed K+ clearance associated with aquaporin-4 mislocalization: phenotypic defects in brains of alpha-syntrophin-null mice. Proc Natl Acad Sci USA 100:13615- 13620.

Beltran-Valero de Bernabe D., Currier S., Steinbrecher A., Celli J., van Beusekom E., van der Zwaag B., Kayserili H., Merlini L., Chitayat D., Dobyns W. B., Cormand B., Lehesjoki A. E., Cruces J., Voit T., Walsh C. A., van Bokhoven H. and Brunner H. G. (2002) Mutations in the O-mannosyltransferase gene POMT1 give rise to the severe neuronal migration disorder Walker-Warburg syndrome. Am J Hum Genet 71:1033 1043.

Bloom T. (1995) Muscular dystrophy. Brain, as well as brawn? Curr Biol 5:338-341.

Bowe M. A., Deyst K. A., Leszyk J. D. and Fallon J. R. (1994) Identification and purification of an agrin receptor from Torpedo postsynaptic membranes: a heteromeric complex related to the dystroglycans. Neuron 12:1173-1180.

Brew H., Gray P. T., Mobbs P. and Attwell D. (1986) Endfeet of retinal glial cells have higher densities of ion channels that mediate K+ buffering. Nature 324:466-468.

Campbell K. P. (1995) Three muscular dystrophies: loss of cytoskeleton-extracellular matrix linkage. Cell 80:675-679.

Connors N. C. and Kofuji P. (2002) Dystrophin Dp71 is critical for the clustered localization of potassium channels in retinal glial cells. J Neurosci 22:4321-4327.

Connors N. C. and Kofuji P. (2006) Potassium channel Kir4.1 macromolecular complex in retinal glial cells. Glia 53:124-31.

Dalloz C, Sarig R., Fort P., Yaffe D., Bordais A., Pannicke T., Grosche J., Mornet D., Reichenbach A., Sahel J., Nudel U. and Rendon A. (2003) Targeted inactivation of dystrophin gene product Dp71: phenotypic impact in mouse retina. Hum Mol Genet 12:1543-1554.

Dietzel I., Heinemann U., Hofmeier G. and Lux H. D. (1980) Transient changes in the size of the extracellular space in the sensorimotor cortex of cats in relation to stimulus-induced changes in potassium concentration. Exp Brain Res 40:432-439.

Ehmsen J., Poon E. and Davies K. (2002) The dystrophin-associated protein complex. J Cell Sci 115:2801-2803.

92 Ervasti J. M. and Campbell K. P. (1993) A role for the dystrophin-glycoprotein complex as a transmembrane linker between laminin and actin. J Cell Biol 122:809-823.

Frigeri A., Nicchia G. P., Nico B., Quondamatteo F., Herken R., Roncali L. and Svelto M. (2001) Aquaporin-4 deficiency in skeletal muscle and brain of dystrophic mdx mice. FASEB J 15:90-98.

Grewal P. K. and Hewitt J. E. (2002) Mutation of Large, which encodes a putative glycosyltransferase, in an animal model of muscular dystrophy. Biochim Biophys Acta 1573:216-224.

Grewal P. K., Holzfeind P. J., Bittner R. E. and Hewitt J. E. (2001) Mutant glycosyltransferase and altered glycosylation of alpha-dystroglycan in the myodystrophy mouse. Nat Genet 28:151-154.

Guadagno E. and Moukhles H. (2004) Laminin-induced aggregation of the inwardly rectifying potassium channel, Kir4.1, and the water-permeable channel, AQP4, via a dystroglycan-containing complex in astrocytes. Glia 47:138-149.

Hayashi Y. K., Ogawa M., Tagawa K., Noguchi S., Ishihara T., Nonaka I. and Arahata K. (2001) Selective deficiency of alpha-dystroglycan in Fukuyama-type congenital muscular dystrophy. Neurology 57:115-121.

Higashi K., Fujita A., Inanobe A., Tanemoto M., Doi K., Kubo T. and Kurachi Y. (2001) An inwardly rectifying K(+) channel, Kir4.1, expressed in astrocytes surrounds synapses and blood vessels in brain. Am J Physiol Cell Physiol 281:C922-931.

Holthoff K. and Witte O. W. (1996) Intrinsic optical signals in rat neocortical slices measured with near-infrared dark-field microscopy reveal changes in extracellular space. J Neurosci 16:2740-2749.

Holthoff K. and Witte O. W. (2000) Directed spatial potassium redistribution in rat neocortex. Glia 29:288-292.

Ibraghimov-Beskrovnaya O., Ervasti J. M., Leveille C. J., Slaughter C. A., Sernett S. W. and Campbell K. P. (1992) Primary structure of dystrophin-associated glycoproteins linking dystrophin to the extracellular matrix. Nature 355:696-702.

Jung D., Yang B., Meyer J., Chamberlain J. S. and Campbell K. P. (1995) Identification and characterization of the dystrophin anchoring site on beta-dystroglycan. J Biol Chem 270: 27305-27310.

Kanagawa M., Michele D. E., Satz J. S., Barresi R., Kusano H., Sasaki T., Timpl R., Henry M. D. and Campbell K. P. (2005) Disruption of perlecan binding and matrix assembly by post-translational or genetic disruption of dystroglycan function. FEBS Lett 579:4792-4796.

93 Kano H., Kobayashi K., Herrmann R., Tachikawa M., Manya H., Nishino I., Nonaka I., Straub V., Talim B., Voit T., Topaloglu H., Endo T., Yoshikawa H. and Toda T. (2002) Deficiency of alpha-dystroglycan in muscle-eye-brain disease. Biochem Biophys Res Commun 291:1283-1286.

Kim D. S., Hayashi Y. K., Matsumoto H., Ogawa M., Noguchi S., Murakami N., Sakuta R., Mochizuki M., Michele D. E., Campbell K. P., Nonaka I. and Nishino I. (2004) POMT1 mutation results in defective glycosylation and loss of laminin-binding activity in alpha-DG. Neurology 62:1009-1011.

Kobayashi K., Nakahori Y., Miyake M., Matsumura K., Kondo-Iida E., Nomura Y., Segawa M., Yoshioka M., Saito K., Osawa M., Hamano K., Sakakihara Y., Nonaka I., Nakagome Y., Kanazawa I., Nakamura Y., Tokunaga K. and Toda T. (1998) An ancient retrotransposal insertion causes Fukuyama-type congenital muscular dystrophy. Nature 394:388-392.

Kofuji P., Ceelen P., Zahs K. R., Surbeck L. W., Lester H. A. and Newman E. A. (2000) Genetic inactivation of an inwardly rectifying potassium channel (Kir4.1 subunit) in mice: phenotypic impact in retina. J Neurosci 20:5733-5740.

Kofugi P and Connors NC (2003) Molecular substrates of potassium spatial buffering in glial cells. Mol Neurobiol 28:195-208.

Lee Y, Kameya S, Cox GA, Hsu J, Hicks W, Maddatu TP, Smith RS, Naggert JK, Peachey NS and Nishina PM (2005) Ocular abnormalities in Large(myd) and Large(vls) mice, spontaneous models for muscle, eye, and brain diseases. Mol Cell Neurosci 30:160-72.

Longman C, Brockington M., Torelli S., Jimenez-Mallebrera C, Kennedy C, Khalil N., Feng L., Saran R. K., Voit T., Merlini L., Sewry C. A., Brown S. C. and Muntoni F. (2003) Mutations in the human LARGE gene cause MDC1D, a novel form of congenital muscular dystrophy with severe mental retardation and abnormal glycosylation of alpha-dystroglycan. Hum Mol Genet 12:2853-2861.

Michele D. E., Barresi R., Kanagawa M., Saito F., Cohn R. D., Satz J. S., Dollar J., Nishino I., Kelley R. I., Somer H., Straub V., Mathews K. D., Moore S. A. and Campbell K. P. (2002) Post-translational disruption of dystroglycan-ligand interactions in congenital muscular dystrophies. Nature 418:417-422.

Nagelhus E. A., Horio Y., Inanobe A., Fujita A., Haug F. M., Nielsen S., Kurachi Y. and Ottersen O. P. (1999) Immunogold evidence suggests that coupling of K+ siphoning and water transport in rat retinal Muller cells is mediated by a coenrichment of Kir4.1 and AQP4 in specific membrane domains. Glia 26:47-54.

94 Neely J. D., Amiry-Moghaddam M., Ottersen O. P., Froehner S. C, Agre P. and Adams M. E. (2001) Syntrophin-dependent expression and localization of Aquaporin-4 water channel protein. Proc Natl Acad Sci USA 98:14108-14113.

Newman E. and Reichenbach A. (1996) The Muller cell: a functional element of the retina. Trends Neurosci 19:307-312.

Nicchia G. P., Nico F3., Camassa L. M., Mola M. G., Loh N., Dermietzel R., Spray D. C., Svelto M. and Frigeri A. (2004) The role of aquaporin-4 in the blood-brain barrier development and integrity: studies in animal and cell culture models. Neuroscience 129: 935-945.

Nielsen S., Nagelhus E. A., Amiry-Moghaddam M., Bourque C, Agre P. and Ottersen O. P. (1997) Specialized membrane domains for water transport in glial cells: high- resolution immunogold cytochemistry of aquaporin-4 in rat brain. J Neurosci 17:171- 180.

Niermann H., Amiry-Moghaddam M., Holthoff K., Witte O. W. and Ottersen O. P. (2001) A novel role of vasopressin in the brain: modulation of activity-dependent water flux in the neocortex. J Neurosci 21:3045-3051.

Noel G., Belda M., Guadagno E., Micoud J., Klocker N. and Moukhles H. (2005) Dystroglycan and Kir4.1 coclustering in retinal Muller glia is regulated by laminin-1 and requires the PDZ-ligand domain of Kir4.1. J Neurochem 94:691-702.

Petrof B. J., Shrager J. B., Stedman H. H., Kelly A. M. and Sweeney H. L. (1993) Dystrophin protects the sarcolemma from stresses developed during muscle contraction. Proc Natl Acad Sci USA 90:3710-3714.

Poopalasundaram S., Knott C, Shamotienko O. G., Foran P. G., Dolly J. O., Ghiani C. A., Gallo V. and Wilkin G. P. (2000) Glial heterogeneity in expression of the inwardly rectifying K(+) channel, Kir4.1, in adult rat CNS. Glia 30:362-372.

Puwarawuttipanit W, Bragg AD, Frydenlund DS, Mylonakou MN, Nagelhus EA, Peters MF, Kotchabhakdi N, Adams ME, Froehner SC, Huang FM, Ottersen OP and Amiry- Moghaddam M (2006) Differential effect of alpha-syntrophin knockout on aquaporin-4 and Kir4.1 expression in retinal macroglial cells in mice. Neuroscience 137:165-75.

Talts J. F., Andac Z., Gohring W., Brancaccio A. and Timpl R. (1999) Binding of the G domains of laminin alphal and alpha2 chains and perlecan to heparin, sulfatides, alpha-dystroglycan and several extracellular matrix proteins. EMBO J 18:863-870.

Ueda H., Gohdo T. and Ohno S. (1998) Beta-dystroglycan localization in the photoreceptor and Muller cells in the rat retina revealed by immunoelectron microscopy. J Histochem Cytochem 46:185-191.

95 Vajda Z., Pedersen M., Fuchtbauer E. M., Wertz K., Stodkilde-Jorgensen H., Sulyok E., Doczi T., Neely J. D., Agre P., Frokiaer J. and Nielsen S. (2002) Delayed onset of brain edema and mislocalization of aquaporin-4 in dystrophin-null transgenic mice. Proc Natl Acad Sci USA 99:13131-13136.

Way M., Pope B., Cross R. A., Kendrick-Jones J. and Weeds A. G. (1992) Expression of the N-terminal domain of dystrophin in E. coli and demonstration of binding to F- actin. FEBS Lett 301:243-245.

Yoshida A., Kobayashi K., Manya H., Taniguchi K., Kano H., Mizuno M., Inazu T., Mitsuhashi H., Takahashi S., Takeuchi M., Herrmann R., Straub V., Talim B., Voit T., Topaloglu H., Toda T. and Endo T. (2001) Muscular dystrophy and neuronal migration disorder caused by mutations in a glycosyltransferase, POMGnTl. Dev Cell 1:717-724.

Zaccaria M. L., Di Tommaso F., Brancaccio A., Paggi P. and Petrucci T. C. (2001) Dystroglycan distribution in adult mouse brain: a light and electron microscopy study. Neuroscience 104:311-324.

Zubrzycka-Gaarn E. E., Bulman D. E., Karpati G., Burghes A. H., Belfall B., Klamut H. J., Talbot J., Hodges R. S., Ray P. N. and Worton R. G. (1988) The Duchenne muscular dystrophy gene product is localized in sarcolemma of human skeletal muscle. Nature 333: 466-469.

96 3 DISCUSSION

3.1 Implication of the DAP complex in ion channel localization

The coordinated activity of Kir4.1 and AQP4 in neural tissues is heavily reliant on their proper localization to glial endfeet abutting blood vessels. Redistribution of these channels away from these domains encumbers the process of spatial K+ buffering thus compromising brain and retinal function by virtue of electrolyte imbalances in regions of high neuronal activity (Holthoff and Witte, 2000; Nierman et al., 2001; Amiry-

Moghaddam et al., 2003). There exists a substantial body of evidence to suggest that the proper patterning of these channels is dependent on the integrity of the DAP complex

(Frigeri et al., 2001; Dalloz et al., 2003; Amiry-Moghaddam et al., 2004), defects in which underlie muscular dystrophy (Emery 2001). This intimates a possible involvement of the improper localization of these channels in the cognitive and ocular defects that are associated with certain dystrophies. The Largemyd mouse is an animal model representing a subset of muscular dystrophies collectively referred to as the dystroglycanopathies in which the ligand binding of the DAP complex is disrupted as a result of a hypoglycosylation of a-DG (Michele et al., 2002). This group includes the human forms of muscular dystrophy FCMD, MEB, WWS and CMD1C (Michele et al., 2002). The high incidence of cognitive and ocular dysfunctions within these diseases has prompted the present investigation into the patterning of both AQP4 and Kir4.1, as well as representative members of the DAP complex, in the brain and retina of this mouse in order to ascertain if they exhibit any alteration associated with a-DG hypoglycosylation.

97 Immune-labeling techniques were used to examine the distribution of these channels in tissue sections from Largemyd and littermate control mice. In accordance with previously published data, this study shows that the distribution of AQP4 is severely disrupted in the Largemyd brain (Michele et al., 2002; Fig 2.2C and F), failing to localize to perivascular astrocytic domains. In addition to this, I show that Kir4.1 is also lost from these sites (Fig 2.21 and L), and that both AQP4 and Kir4.1 labeling is greatly reduced at the glial-limiting membrane formed by astrocyte end-feet in the submeningeal region

(Appendices Fig 4.1). Co-labeling with vascular markers such as perlecan, laminin and agrin show that the perivascular loss of channel labeling is not secondary to loss of vascular density. These data support my hypothesis that the inability of the complex to bind extracellular ligands by virtue of a-DG hypoglycosylation compromises the distribution of these two channels. These findings are consistent with previous studies using animal models characterized by disruptions of members of the DAP complex other than DG. For instance, in the mdx mouse lacking dystrophin, AQP4 expression at pericapillary and submeningeal domains is greatly reduced throughout the brain, as are both AQP4 and Kir4.1 in the retina in spite of maintained vascular integrity (Frigeri et al.,

2001; Connors and Kofuji, 2002; Vajda et al., 2002; Dalloz et al., 2003). Furthermore, in the presence of laminin, AQP4 and Kir4.1 co-aggregate with DG as well as other components of the DAP complex in cultured astrocytes (Guadagno et al., 2004) substantiating the dependence of AQP4 and Kir4.1 distribution on the ability of the complex to bind laminin by way of a-DG.

98 Data supporting the involvement of the DAP complex in ion channel distribution and/or stability is not limited to neural tissues, nor to Kir4.1 and AQP4. Recently, it was

shown using GST-fusion protein constructs, that the cardiac sodium channel Nav1.5 interacts with dystrophin, most probably through ai- or Pi-syntrophin (Gavillet et al.,

2006). Western blot analysis of heart lysates from the dystrophin deficient mdx5cv mouse

revealed a 50% reduction in the expression of Nav1.5 that was deemed to not be a consequence of a reduction in the mRNA level (Gavillet et al., 2006). These data indicate

that the DAP complex is involved in anchoring Nav1.5 in cardiomyocyte membranes.

Interestingly, the reduced expression of these channels in the heart is associated with detectable deficits in cardiac conduction (Gavillet et al., 2006).

In peripheral nerves, a DAP complex composed of DG, Dpi 16 and utrophin is present at the nodes of Ranvier (Occhi et al., 2005). It was revealed that selective ablation of DG from Schwann cells results in a decrease of voltage gated sodium channels,

including Nav1.2 and 1.6, at these sites and an aberrant redistribution of these channels to paranodal domains (Saito et al., 2003). More recently, it was shown that both laminin 2

a ( 2piYi) (laminin nomenclature reviewed in Aumailley et al., 2005) and 10 (a5|3iyi) are abundant at the nodes of Ranvier (Occhi et al., 2005). Serendipitously, the dystrophic

71 71

mouse, which carries a mutation of the a2 chain in laminin 2 , presents with a similar, albeit milder, phenotype to the Schwann cell DG deficient mice (Occhi et al., 71 71

2005). The dfJ/dfJ mice also exhibit an increase in the expression of laminin 1 (aiPiyi) and mutation of the yi chain, common to laminins 1, 2 and 10, results in the same abnormalities found in the DG deficient mouse (Occhi et al., 2005). These defects in channel distribution are associated with slowed neuronal conduction velocity (Saito et al.,

99 2003; Occhi et al., 2005). Together these data strongly suggest that the proper targeted

distribution of Navs at nodal regions of axons is heavily dependent on the interaction between DG and laminins. Furthermore, they suggest that in conditions of laminin 2 disruption, a compensatory upregulation of laminin 1 is able to partially rescue the phenotype.

The muscle specific Nav (SkMl) is particularly concentrated at the trough of neuromuscular junction folds (Colledge and Froehner, 1998) where it colocalizes with

dystrophin (Sealock et al., 1991). ar, Pi- and p2- syntrophin are also all expressed in junctional troughs (Kramarcy and Sealock, 2000) and bind to the C-terminus of SkMl in vitro (Gee et al., 1998). SkMl expression decreases rapidly away from the endplate but extrajunctional channels are present (Lupa et al., 1993) as are ai- and pi-syntrophins

(Peters et al., 1997). In the mdx mouse, junctional syntrophin expression in unaltered despite the loss of dystrophin (Peters et al., 1997) which could account for the normal expression of SkMl in mdx NMJs (Wood and Slater, 1998). However, extrajunctional syntrophin expression is highly reduced in the mdx mouse (Peters et al., 1997) and this is associated with a concomitant reduction in extrajunctional SkMl with no change in mRNA levels (Ribaux et al., 2001) which is highly indicative of a compromise in SkMl anchoring mechanisms associated with disruption of the DAP complex. Similarly,

+ spermatozoa express Dp71 as well as SkMl and the voltage gated K channel Kvl.l in a polarized fashion (Hernandez-Gonzalez et al., 2001). Analysis of these cells from mdx3cv mice, which lack all dystrophin isoforms, reveals a marked mislocalization of both of these channels indicating the involvement of the DAP complex in their proper distribution (Hernandez-Gonzalez et al., 2005).

100 Finally, in cultured C2C12 myotubes, application of agrin, a known basal lamina protein and binding partner of a-DG (Bowe et al., 1994), induces acetylcholine receptors

(AchR) clustering (Gee et al., 1994; Ferns et al., 1996). Interestingly, myotubes cultured from mdx mice fail to intiate this agrin induced AchR clustering (Kong and Anderson,

1999) as do antisense derivatives of the C2 mouse muscle cell line with a reduced expression of a-DG (Jacobson et al., 1998). Together these data provide evidence of a

DAP complex dependent mechanism, most likely reliant on the interaction between agrin and a-DG, underlying this channel clustering. In vivo, AchRs are concentrated at the crest of postsynaptic folds at the NMJ where they are coenriched with a utrophin containing DAP complex (Deconinck et al., 1997a). The mdx mouse, lacking dystrophin, the utrophin null mouse and the double knock out mdx/utr"/_ mouse all present with a reduction in junctional AchRs (Lyons and Slater, 1991; Deconinck et al., 1997a;

Deconinck et al., 1997b). This reduction, however, is only characteristic of regenerating muscle fibres, as identified by the centralized organization of nuclei, and is accompanied by a gross simplification of the junctional architecture defined by a reduction in the number and depth of postsynaptic folds which, in turn, compromises the polarized expression of the AchR at the junctional crests (Lyons and Slater, 1991; Deconinck et al.,

1997a; Deconinck et al., 1997b). Similar results are also true for the a-syn null mouse which shows a 65% reduction in NMJ AchRs and a concomitant reduction in postsynaptic fold number (Adams et al., 2000), the chimeric mouse deficient for DG

(Cote et al., 1999; Jacobson et al., 2001), the a-DB null mouse (Grady et al., 2001), the

Largemyd mouse and human FCMD (Taniguchi et al., 2006). The disruption of the NMJ architecture amounts to a fairly confounding variable and makes it difficult to delineate

101 the specific involvement of the DAP complex in the localization of these channels in vivo. There is, however, little doubt that the stability of the AchRs at their proper membrane domains and the integrity of the DAP complex are somehow interrelated.

In summary, the amalgamation of these data provide evidence that the transmembrane bridge provided by the DAP complex is involved in anchoring ion channels to specific membrane domains throughout the body. The relationship between the DAP complex and ion channel stability lends support to the present hypothesis that the unbreached continuity between the complex and the ECM, by way of O-linked glycosylated moieties of a-DG, is essential to the stability of both Kir4.1 and AQP4 to their proper perivascular domains of astrocytic endfeet. Furthermore, the above mentioned examples substantiate that the mislocalization of these channels, secondary to

DAP complex defects, could result in physiologically relevant disruptions in electrical activity.

3.2 Stabilization of AQP4 and Kir4.1 via a-syntrophin

Both Kir4.1 and AQP4 contain a consensus C-terminal SXV motif that binds to the PDZ domain of a-syn, as indicated by co-immunoprecipitation experiments, (Neely et al., 2001; Connors et al., 2004; Connors and Kofuji, 2006) offering the possibility that these channels could interact with the DAP complex thus conferring their membrane stability. In support of this, a-syn null mice exhibit a redistribution of AQP4 (Neely et al.,

2001; Amiry-Moghaddam et al., 2004; Amiry-Moghaddam et al., 2003) away from

102 perivascular sites in the brain. Similarly, in the retina of the a-syn null mice, AQP4 expression at perivascular Muller cell endfeet is reduced by 70% (Puwarawuttipanit et al.,

2006). Furthermore, this PDZ-dependent interaction is crucial for the laminin-induced clustering of Kir4.1 in Muller cell cultures (Noel et al., 2005). AQP4 is also expressed at the sarcolemma of fast twitch muscle fibres where it is anchored through a PDZ- dependent interaction with a-syn (Adams et al., 2001).

Here I present data showing that in the brain of the Largemyd mouse, a-syn expression is greatly reduced from perivascular astrocytic endfeet (Fig 2.4A and D) analogous to the findings for both AQP4 and Kir4.1 channels in this tissue (Fig 2.2).

Hypoglycosylation of a-DG in this mouse leads to a severe disruption of the complex and a failure of a-DG, P-DG, dystrophin and a-syn to concentrate to perivascular astrocytic endfeet (Fig. 2.4; Michele et al., 2002). These data, in combination with those outlined above, provide strong evidence that the DAP complex acts as a scaffold involved in the polarized distribution of Kir4.1 and AQP4 via a direct interaction with syntrophin. Based on the findings of the present study and their congruence with the data mentioned above, it is tempting to speculate that the hypoglycosylated state of a-DG and the resultant loss of ligand binding destabilize the complex as a whole leading to a loss of AQP4 and

Kir4.1 stability secondary to a-syn disruption.

103 3.3 Impact of a-dystroglycan glycosylation on AQP4 and Kir4.1 expression

While the immunofluorescence data reveal a loss of both AQP4 and Kir4.1 from perivascular endfeet, immunoblot data reveal that their total expression levels in the

Largemyd brain do not differ from littermate controls (Fig 2.3). This indicates that loss of perivascular labeling reflects a redistribution of the channels to other astrocytic domains, perhaps in a more uniform rather than highly polarized pattern. This is in accordance with previously published data reporting that despite the loss of perivascular expression of

AQP4 in brain of the mdx mouse, AQP4 levels are unchanged when compared to the wild type mouse (Vajda et al., 2002). Similarly, in brain of the a-syn null mouse, AQP4 levels are comparable to those of control mice despite a marked reduction in its perivascular astrocyte labeling (Amiry-Moghaddam et al., 2003; Neely et al., 2001). In fact, immunogold analysis of AQP4 in the a-syn null brain revealed that its subcellular distribution is reversed compared to control brains. In other words, AQP4 expression at non-perivascular domains was increased, offsetting its loss at the astrocytic endfeet

(Neely et al., 2001; Amiry-Moghaddam et al., 2003).

Alternatively, the loss of the immunolabeling for these channels may reflect an absolute loss of protein expression. In this case, the association between a-syn and AQP4 would confer its stability and disruption of this interaction would result in the removal from, or prevent the insertion of, AQP4 in the membrane. Based on the immunoblot data from the present investigation this does not appear to be the case in the Largemyd brain.

However, when a truncated form of AQP4, lacking the SSV motif (ASSV), is expressed in MDCK cells that express a-syn and other DAP components, it is targeted to basolateral

104 domains exactly as is wild type AQP4 (Neely et al., 2001). This shows that the SSV motif of AQP4 is not required for its proper membrane localization in these cells.

Interestingly, when the expression of ASSV AQP4 in stably transfected HEK293 cells was examined, immunoblots revealed an overall reduction of the truncated AQP4 expression compared to the full length protein in spite of similar levels of the encoding mRNA. Furthermore, [35S] Methionine pulse chase labeling of the transfected cells showed that the half life of the full length AQP4 was three times longer than that of the truncated AQP4 indicating that the SSV motif expressed at the extreme C terminus of the channel confers its stability, rather than its localization within the membrane (Neely et al., 2001). This is in stark contrast to the results obtained from the brain of both the a-syn

(Neely et al., 2001; Amiry-Moghaddam et al., 2003) and Largemyd brains. However, these findings are congruent with those obtained from the skeletal muscle of both a-syn null

(Neely et al., 2001) and mdx mouse (Frigeri et al., 2001) in which loss of syntrophin and dystrophin respectively results in a significant reduction of the total expression levels of

APQ4. The difference in AQP4 expression between muscle (Neely et al., 2001; Frigeri et al., 2001), cell culture systems (Neely et al., 2001) and brain (Neely et al., 2001; Amiry-

Moghaddam et al., 2003) may be due to a difference in the composition of the DAP complex which may have an impact on the affinity of AQP4 for the complex.

In any case, the data presented here, in combination with studies in the a-syn null brain (Neely et al., 2001; Amiry-Moghaddam et al., 2003) suggest that the integrity of the DAP complex in this tissue is essential to the proper compartmentalization of AQP4

105 and Kir4.1, as well as the proteins of the complex itself, to the appropriate membrane domains. Any breach within the continuity of the protein axis of the DAP results in a collapse of the complex, one consequence of which is the loss of the accurate localization of AQP4 and Kir4.1 to specific astrocytic domains. Interestingly, a mere hypoglycosylation of a-DG and the concurrent disruption of basal lamina binding appear to be sufficient to bring about the loss of polarization of these channels in the brain.

3.4 Possible physiological implications of AQP4 and Kir4.1 mislocalization

The physiological impact of AQP4 redistribution in the a-syn null mouse is correlated with delayed K+ ion clearance in brain indicating that water flux through perivascular AQP4 is required for efficient removal of excess K+ after neuronal activation

(Amiry-Moghaddam et al., 2003). The aberrant accumulation of extracellular potassium serves to depolarize surrounding neurons giving rise to hyper-excitability which, in the a- syn null mouse, manifests as an increased susceptibility to febrile induced seizures having a larger magnitude and longer duration when compared to control animals

(Amiry-Moghaddam et al., 2003). This provides evidence that the mislocalization of glial

AQP4 is sufficient to compromise neuronal function. Corroborating this data, in Mesial

Temporal Lobe Epilepsy (MTLE) characterized by hyper-excitable hippocampi giving rise to seizures, AQP4 expression is significantly reduced in perivascular membrane domains of astrocytes as is Dp71, suggesting a correlation between seizure susceptibility and AQP4 mislocalization secondary to disruption of the DAP complex (Eid et al., 2005).

Of particular interest, between 50% and 80% of patients with FCMD suffer from febrile

106 convulsions and epileptic seizures (Yoshioka and Higuchi, 2005). Similarly, WWS patients have an increased incidence of seizure activity (Vajsar and Schachter, 2006).

With respect to the redistribution of Kir4.1 in the Largemyd brain, genetic linkage studies have identified correlations between missense mutations in the Kir4.1 gene and increased seizure susceptibility in humans (Buono et al., 2004). These mutations are not believed to affect the biophysical properties of the channel (Shang et al., 2005) suggesting the defect could be associated with faulty trafficking or stability within specific membrane domains.

The most parsimonious explanation that can be derived from the data above is that the structural integrity of the DAP complex, or more specifically, the ability of the complex to bind proteins of the basal lamina via a-DG O-linked glycosylation, confers a specialized organization of AQP4 and Kir4.1 to precise astrocytic domains to allow for their coordinated activity. A disruption in this organization would compromise potassium clearance and eventuate in seizure susceptibility. In light of this, it is reasonable to speculate that the Largemyd mouse would also suffer from an increased propensity towards hyperthermia induced seizure behaviour under the proper conditions.

The deletion of AQP4 from brain confers no abnormalities in brain structure, vascular architecture, nor in intracranial compliance, pressure or blood brain barrier function (Manley et al., 2000; Papadopoulos et al., 2004). However, in models of pathogenic brain edema induced by acute water intoxication and focal ischemic stroke,

AQP4 deletion has been shown to significantly reduce the progression of cytotoxic edema resulting in an improved neurological outcome (Manley et al., 2000).

Interestingly, opposite effects are observed in models of vasogenic edema where intracranial pressure and brain water content are higher than normal (Papadopoulos et al.,

107 2004). Likewise, in models of obstructive hydrocephalus, progression of the pathology was accelerated (Bloch et al., 2006). The involvement of AQP4 in osmotically driven water transport as well as excess brain water absorption is apparent yet the implications of this in human health and pathology are still unclear. It would be of interest to examine the coping abilities of the Largemyd mouse in experimentally induced conditions of cytotoxic and vasogenic edema as well as kaolin induced hydrocephalus (Khan et al.,

2006) in order to ascertain the impact of AQP4 mislocalization on these pathologies and their possible correlation with human dystroglycanopathies.

3.5 Disruptions in the cortical surface basal lamina

In addition to the mislocalization of AQP4 and Kir4.1 in the brain of the Largemyd mouse, I show here that the basal lamina of the cortical surface, which covers the pial aspect of the cortices continuously in the control animals, is severely disrupted as evidenced by large discontinuities in laminin staining (Fig 2.1). This has been previously reported in the brain of this mouse with respect to both laminin and perlecan (Michele et al., 2002; Kanagawa et al., 2005) and here I show that agrin (Fig 2.1), exhibits a similar discontinuous pattern. Similarly, the fukutin deficient mouse exhibits a discontinuity in the basal lamina apparent as early as E14 (Chiyonobu et al., 2005). This is associated with neuronal migration errors characterized by abnormal cortical lamination and ectopic clusters of neurons invading into the subarachnoid space (Chiyonobu et al., 2005). This mouse, however, shows no detriments in the initiation of neuronal migration since neurons are found throughout the cerebral cortex, indicating that the aberrant progression

108 is secondary to defects at the Glia-Limitans-Basal-Lamina interface (Chiyonobu et al.,

2005).

During the development of the cerebral cortex, radial glia span the neuroepithelium with their endfeet anchored to the pial basement membrane. These glial cells serve as a scaffold along which neurons migrate toward the developing cortical plate, which at later developmental stages becomes divided into six laminated layers

(reviewed by Montanaro and Carbonetto, 2003). Defects in each of the major events in this development give rise to unique defects which have been previously characterized

(reviewed by Gleeson and Walsh, 2000). The discontinuities in the pial basement membrane in the Largemyd brain most likely prevent the proper attachment of radial glia thus compromising neuronal migration and giving rise to misplaced neurons and aberrant cortical layering. Furthermore, focal voids in the basement membrane could result in the over extension of radial glia and the presence of neurons beyond the pial border

(reviewed by Montanaro and Carbonetto, 2003) which would account for the ectopic neurons in the subarachnoid space of the Fukutin deficient mouse.

The similarities in anatomical brain defects between the fukutin deficient chimera

(Chiyonobu et al., 2005), the Largemyd mouse and human dystrophies such as FCMD,

MEB and WWS (Michele et al., 2002) suggest that disruption of the cortical surface basal lamina induces severe abnormalities in neuronal migration that cause a disorganization of the cortical lamination. This is further underscored by the fact that targeted deletion of dystroglycan from brain in mutant mice gives rise to histological

109 abnormalities throughout the brain that are highly congruent with those reported in both human and animal forms of dystroglycan (Moore et al., 2002). Interestingly, the vascular basement membrane remains intact in all of the pathologies and animal models mentioned above implying a differential involvement of the DAP complex in the maintenance of basal lamina integrity between the vascular basement membrane and the glia limitans.

3.6 Expression of Large transcripts and glycosylation of a-dystroglycan in retina

Patients with FCMD, MEB and "WWS suffer from ocular anomalies as detected in the electroretinogram (Cormand et al., 1999; Dobyns et al., 2005; Fukuyama et al., 1981;) and, as with the brain, neuronal activity challenges potassium homeostasis in the retina as well (Connors and Kofuji, 2006). Retinal Muller cells are known to express both AQP4 and Kir4.1 in a highly polarized fashion that is coincident with the expression of the DAP complex, adhering to glial domains abutting blood vessels and the inner limiting membrane apposed to the vitreous body (Ishii et al., 1997; Kofuji et al., 2002; Nagelhus et al., 1998). Here they participate in a coordinated effort to clear excess potassium from regions of high neuronal activity analogous to systems in the brain. In fact, most of what is currently known about the phenomenon of potassium siphoning has come to light from studies using Muller cells.

110 As discussed in the introduction, the Large protein exists in two different isoforms, conveniently referred to as Large 1 and Large2, with the Large2 gene also giving rise to several variable transcripts (Grewal et al., 2005). This investigation shows that all of these known Large isoforms are expressed in the retina (Fig. 2.5). However, immunolabeling with the IIH6 antibody directed against glycosylated moieties of a-DG reflects lack of fully glycosylated a-DG at perivacular glia in retina (Fig 2.6). These data argue that unlike in the Largemyd kidney (Grewal et al., 2005), Large 2 and corresponding transcripts do not compensate for the lack of Large 1 in the Largemyd retina. This tissue- dependent compensation phenomenon mediated by Large 2 remains an unresolved question at present.

3.7 Spatial differences in the DAP complex-induced anchoring of AQP4 and Kir4.1

Given that in the brain, hypoglycosylation of a-DG was associated with a mislocalization of AQP4 and Kir4.1 away from perivascular astrocytic domains, I reasoned that a similar effect would be found in retina. However, despite the large amount of congruency in the subcellular organization of these channels in glial cells between brain and retina, as well as their close association with representative members of the DAP complex in retina, AQP4 and Kir4.1 labeling was not completely lost from perivascular astrocytic domains in the Largemyd retina (Fig 2.8). This simple difference sheds light on some fundamental differences in the anchoring mechanisms of AQP4 and

Kir4.1 utilized by these two tissues and supports the possibility that there are tissue- specific strategies through which the DAP acts as a transcellular pore anchor.

Ill In the brain of the a-syn null mouse, AQP4 is reduced at perivascular domains while Kir4.1 is largely unaffected (Neely et al., 2001; Amiry-Moghaddam et al., 2003).

Similarly, in the retina of this mouse, AQP4 expression at perivascular astrocyte processes shows a 70% decrease while that of Kir4.1 is not notably altered

(Puwarawuttipanit et al., 2006). These data suggest that AQP4's expression is heavily dependent on a-syn while Kir4.1 is reliant on other anchoring mechanisms. Of interest is that in this mouse, the expression of pi-syn is unaffected suggesting the differential involvement of syntrophins in the anchoring of these two channels (Puwarawuttipanit et al., 2006). In the present investigation, I show that AQP4 and Kir4.1 expression is still detectable at perivascular sites in the retina of the Largemyd mouse (Fig 2.8) in spite of a loss of a-syn (Fig 2.9). This is in contrast with the data found in the brain of the Largemyd mouse where neither type of channels is detectable perivascularly. Intriguingly, Pi-syn which is also lost from brain, is still detectable in the Largemyd retina. This suggests that

Pi-syn may be involved in anchoring a pool of AQP4 and Kir4.1 at perivascular astrocytes in retina.

In support of this possibility, the sarcolemma expresses a complement of AQP4 that co-distributes with members of the DAP complex and, in the a-syn null mouse, the expression of this channel is reduced. Interestingly, Pi-syn is still expressed much like in the retina of both the a-syn null and Largemyd mouse arguing against the ability of this syntrophin to protect against AQP4 destabilization in muscle. This is another example substantiating the possibility that tissues are unique in the composition and stability of

112 their DAP complexes as well as their channel anchoring mechanisms. What is exceptionally intriguing, is that in the a-syn null retina, a pool of AQP4 reflecting 30% of normal expression remains located at perivascular astrocytic domains (Puwarawuttipanit et al., 2006). This suggests the possibility of two separate AQP4 pools, one that is dependent on a-syn for its proper distribution and one that is reliant on another binding partner. The persistence of Pi-syn in both the Largemyd and a-syn null retinas makes it a tempting candidate for the stabilization of this secondary pool as well as for Kir4.1. In the current study it is not known what percentage of AQP4 or Kir4.1 is represented by the remaining detectable pools and immunogold electromicroscopy will be required to delineate this. In any event, the maintenance of Pi-syn in the retina but not in brain, raises the possibility that a protein present in retina prevents Pi-syn mislocalization. Assuming that AQP4 and Kir4.1 interact with Pi-syn, this would explain the pool of these channels still detected in the Largemyd retina.

The differences in the anchoring mechanisms could reflect a different complex entirely, perhaps relying on another transmembrane protein capable of binding to proteins of the basal lamina. A tempting candidate would be the integrin receptors. Integrins are transmembrane proteins composed of an aP-heterodimer that bind matrix proteins

(reviewed by Hynes 1992). In muscle, a6pi and a7pi are exclusive laminin receptors

(Mayer 2003) and integrins aiPi, asPi, a6Pi and a6P4 are all expressed in the CNS microvasculature and are known to share laminin as their primary basal lamina ligand

(reviewed by del Zoppo and Milner, 2006). Furthermore, mutations in a7-integrin underlie certain forms of human muscular dystrophy (Hayashi et al., 1998) and its

113 expression is increased in both DMD and mdx muscle (Hodges et al., 1997). Of particular interest is that oi7p\ integrin is upregulated in the DG deficient chimeric mouse muscle

(Cote et al., 2003), suggesting that it could be compensating for the loss of DG and its role as a basal lamina receptor. Moreover, double knock out mice models lacking both dystrophin and 017-integrin (mdx/a7"/") present with a severe myopathy leading to death at puberty, which is in drastic contrast to the comparatively mild phenotype of the mdx mouse. This corroborates the possible existence of a heretofore unidentified compensatory mechanism (Rooney et al., 2006). Perhaps there is an integrin complex containing Pi-syn and independent of DG in retina that may be implicated in localizing a pool of AQP4 and Kir4.1 in perivascular astrocytes. This hypothesis is supported by the fact that (1) dystrophin is associated with a Pi-integrin containing complex in PC 12 cells

(Cerna et al., 2006) and (2) dystrophin interacts with pi-syn in skeletal muscle (Peters et al., 1997). Furthermore, Pi-syntrophin, like all other syntrophins, contains a PDZ binding domain (Blake et al., 2002) which could conceivably interact directly with the SSV and

SNV motifs of AQP4 and Kir4.1, respectively (Connors et al., 2004; Neely et al., 2001).

With the data outlined above, it is tempting to hypothesize that an axPi-integrin receptor complex containing dystrophin and Pi-syn may be expressed in at least a subset of perivascular astrocytes in retina. This putative complex could be involved in anchoring a pool of AQP4 and Kir4.1 in normal retina or as a compensatory mechanism for the lack of a-DG interaction with the ECM in the Largemyd retina. Interestingly, targeted disruption of Dp71 leads to an altered distribution of both Kir4.1 and AQP4 away from perivascular sites in retina despite the maintenance of both P-DG and ai-syn at these domains (Dalloz et al., 2003). The expression of Pi-syn in this retina, however, remains

114 to be revealed. These data not only assert that proper expression of a-syn is insufficient to protect against AQP4 and Kir4.1 mislocalization, they suggest the involvement the conjectural integrin based complex outlined above. If Pi-syn is the predominant syntrophin involved in the proper localization of these channels and if it is anchored to its complex, be it DG or integrin based, by way of dystrophin then disruption of dystrophin as in the mdx3cv mouse would be expected to destabilize pi-syn and ultimately disrupt the localization of these channels. Of course these allegations are strictly hypothetical at present, but they provide a basis from which to further pursue the unraveling of the complexities of the differential anchoring mechanisms of these channels. This will be discussed further in section 3.10.3.

Together these data suggest possible overlapping roles for DG and specific integrins and raise the question of whether this could explain the differential effect of the

Largemyd mutation between brain and retina. In order to pursue this further, a thorough characterization of the integrin isoforms expressed in retina would need to be carried out.

In any event, it is clear that in retina the anchoring mechanism for these channels is at least partially a-DG independent and suggest the existence of a Pi-syn containing complex lacking a-DG or, at the very least, different anchoring strategies between the retina and brain. Obviously, the findings of the present investigation leave room for a multitude of exciting studies in both brain and retina which will need to be carried out in order to elucidate the molecular intricacies underlying these differences.

115 3.8 Disruption of the inner limiting membrane and aberrant retinal lamination

Interestingly, analogous to the findings in the brain of the Largemyd mouse, the inner limiting membrane (ILM) of the retina was found to be discontinuous in comparison to the littermate controls (Fig 2.7). As discussed above, in brain, the discontinuities found in the glia limitans are believed to be an underlying cause for the migrational errors found throughout the brain of the Largemyd and fukutin deficient mice

(Chiyonobu et al., 2005). Likewise, it is possible that the disruption of the ILM in the retina is a contributing factor to the neuronal migration errors evidenced by the severe disorganization of the retinal lamination. Interestingly, as with the brain, the basement membrane surrounding blood vessels remains unperturbed indicating the partial conservation of certain characteristics of the DAP complex between the retina and brain.

3.9 Conclusions

The primary objective of this study was to determine the impact of the Largemyd mutation on the perivascular localization of Kir4.1 and AQP4 in brain and retina. In brain, I show that the hypoglycosylation of a-DG, secondary to mutation of Largel, is associated with loss of both AQP4 and Kir4.1 channels as well as of a- and Pi- syntrophins from perivascular astrocytic domains. In retina however detectable pools of both channel types as well as pi-syntrophin remain expressed at perivascular astrocytes.

These data show that Kir4.1 and AQP4 may be anchored and stabilized at astrocytic endfeet via a pi-syn dependent and a-DG-ECM independent mechanism and reveal the

116 existence of a novel pi-syn-containing complex lacking DG. The components of this hypothetical complex remain to be identified. In any event, these findings point to a spatial variability in the composition of the syntrophin-containing complex within different neural tissues that may have distinct functional implications.

3.10 Future Directions

3.10.1 Involvement of a-DG glycosylation in the laminin-induced clustering of Kir4.1 andAQP4

It has been demonstrated that application of laminin to cultured astrocytes is sufficient to induce the clustering of AQP4 and Kir4.1 (Guadagno et al., 2004).

Furthermore, these clusters are closely associated with members of the DAP complex that cluster in the presence of laminin (Guadagno et al., 2004). These data suggest that the laminin initiated clustering of these channels could be dependent on the known interaction between laminin and a-DG which itself is heavily reliant on the O-linked glycosylation status of a-DG (Michele et al., 2002). Based on these data, I propose to carry out laminin induced clustering assays on primary astrocyte cultures harvested from the Largemyd mouse brain. These studies would offer valuable insight into the molecular interactions that underlie the clustering of these channels in astrocyte cultures as well as their polarized organization in vivo. If the aggregation of these channels is indeed dependent on the interaction between laminin and a-DG, then the laminin treatment would fail to induce their aggregation in the glycosylation deficient Largemyd astrocytes cultures. This, in addition to the in vivo data presented in this thesis on the Largemyd brain,

117 will provide further evidence of the direct involvement of a-DG glycosylation in the proper localization of these channels to perivascular domains of astrocytes.

The main focus of this thesis was on brain and retinal perivascular astrocytes of the Largemyd mouse and future experiments on Muller glial cultures will be essential to investigate whether the targeting of Kir4.1 and AQP4 to their endfeet is also dependent on a-DG glycosylation. Previous data from my lab have shown that Kir4.1 clustering in

Muller cells was induced by laminin application and that this was dependent on a PDZ- mediated interaction (Noel et al., 2005). The polarized distribution of these channels to

Muller cell endfeet is crucial for K+ buffering which has been correlated with the production of the b wave, one of the components of the electroretinogram (Newman and

Odette, 1984). Of particular interest, the electroretinogram of the Largemyd mouse revealed a deficiency in the b-wave amplitude (Holzfeind et al., 2002) and the study I propose on Muller cell cultures from the Largemyd mouse may reveal the molecular basis underlying such a deficit.

3.10.2 Effect of the Largemy mutation on other retinal DAP complex proteins

In order to further characterize the dissimilarities in the effects of the Largemyd mutation between the brain and retina, a more detailed investigation on the effects that this mutation has on the components of the other retinal DAP complex proteins would be

118 of value. The study would be carried out in a manner analogous to the one presented in the present thesis. Tissue sections of retina from both Largemyd and littermate controls would be labeled against various DAP complex proteins previously uncharacterized in the Largemyd retina. In particular, the expression of dystrophin in response to Largemyd mutation is well characterized in brain where it fails to concentrate to astrocytic endfeet

(Michele et al., 2002). The effect of the Largemyd mutation on the expression of dystrophin in retina has not yet been demonstrated. Likewise, in wild type brain, ai-

Dystrobrevin (DB), a dystrophin related and associated protein, localizes to perivascular astrocytic domains (Blake et al., 1999). In retina, DBs are also found at perivascular glial cell domains in wild type mice (Ueda et al., 2000) yet the expression of these proteins in the Largemyd brain and retina, remains to be ascertained. Together these studies could offer valuable insights into the differential effect of the Largemyd mutation on the distribution of AQP4 and Kir4.1 in brain and retina.

3.10.3 Expression of integrins in the retina

As discussed in detail in section 3.7, integrin receptors, specifically ot^Pi and

CC7P1, act as laminin receptors in skeletal muscle and are suspected to act in compensation for DG disruption in DG chimeric mice (Cote et al., 2003). Similarly, in brain oiipi, (X3P1, oi6Pi and a$4 integrins are all expressed in the CNS microvasculature and are known to share laminin as their primary basal lamina ligand (reviewed by del Zoppo and Milner,

2006). To the best of our knowledge, no thorough investigation of the expression of these

119 integrin receptors in the retina has been carried out. In view of this, it would be of interest to analyze their expression profiles by immunofluorescence in the retina of both Largemyd and littermate control mice. Expression of these integrin isoforms at glial cell endfeet would serve to substantiate the hypothesis put forward in section 3.7 that proposes that the retinal distribution of AQP4 and Kir4.1 could be dependent on a novel, conjectural and uncharacterized complex containing integrins in place of DG. An analogous study in brain could serve to complement these findings as well.

3.10.4 Pathological implications of AQP4 and Kir4.1 mislocalization in Large"" brain

AQP4 mislocalization is associated with disturbances in potassium buffering, hyperthermia induced seizures (Amiry-Moghaddam et al., 2003) and altered water homeostasis in condition of osmotic stress (Manley et al., 2000; Papadopoulos et al.,

2004; Bloch et al., 2006). In order to ascertain whether the Largemyd mutation is associated with similar deficits, Largemyd mice will be tested for the incidence of febrile seizures, as well as their duration and severity by immersing them in a warm water bath as described by Amiry-Moghaddam et al., (2003). The response of these mice to cytotoxic and vasogenic edema can be determined following acute water intoxication

(Manley et al., 2000) and melanoma tumour implantation, respectively (Verkman et al.,

2006). Kaolin-induced hyrocephalus is also an accepted model (Khan et al., 2006) and could be induced in the Largemyd mouse to examine their outcome in view of the AQP4 mislocalization. These studies could offer insight into the water handling abilities of these

120 ' animals when submitted to physiological stresses and may shed light on the way humans

with dystroglycanopathy type muscular diseases handle water imbalances secondary to

stroke or head trauma.

This study has revealed many findings seminal to the understanding of the brain

and retinal pathologies secondary to a-DG hypoglycosylated muscular dystrophies and it

has simultaneously unveiled a large number of questions that remain to be elucidated.

Investigations into the molecular deficits that underlie the cognitive and ocular

abnormalities associated with muscular dystrophies present great possibilities through

which a comprehensive understanding of the subcellular etiologies underlying this

devastating group of congenital diseases can be delineated. This, in turn, could give rise

to novel diagnostics, therapies and even preventative strategies for muscular dystrophy in

the future.

121 3.11 Works Cited

Adams ME, Kramarcy N, Krall SP, Rossi SG, Rotundo RL, Sealock R, Froehner SC (2000) Absence of alpha-syntrophin leads to structurally aberrant neuromuscular synapses deficient in utrophin. J Cell Biol 150:1385-1398.

Adams ME, Mueller HA, Froehner SC (2001) In vivo requirement of the alpha- syntrophin PDZ domain for the sarcolemmal localization of nNOS and aquaporin-4. J Cell Biol 155:113-122.

Amiry-Moghaddam M, Williamson A, Palomba M, Eid T, de Lanerolle NC, Nagelhus EA, Adams ME, Froehner SC, Agre P, Ottersen OP (2003) Delayed K+ clearance associated with aquaporin-4 mislocalization: phenotypic defects in brains of alpha- syntrophin-null mice. PNAS 100:13615-13620.

Amiry-Moghaddam M, Xue R, Haug FM, Neely JD, Bhardwaj A, Agre P, Adams ME, Froehber SC, Mori S, Ottersen OP (2004) Alpha-syntrophin deletion removes the perivascular but not endothelial pool of aquaporin-4 at the blood-brain barrier and delays the development of brain edema in an experimental model of acute hyponatremia. FSEB 18:542-544.

Aumailley M, Bruckner-Tuderman L, Carter WG, Deutzmann R, Edgar D, Ekblom P, Engel J, Eva Engvall E, Hohenester E, Jones JCR, Kleinman HK, Marinkovich MP, Martin GR, Mayer U, Meneguzzi GMiyazaki K, Patarroyo M, Paulsson M, Quaranta V, Sanes JR, Sasaki T, Sekiguchi K, Sorokin LM, Talts JF, Tryggvason K, Uitto J, Virtanen I, von der Mark K, Wewer UM, Yamada Y, Yurchenco PD (2005) A simplified laminin nomenclature . Matrix Biol 24: 326-332.

Blake DJ, Hawkes R, Benson MA, Beesley PW (1999) Different dystrophin-like complexes are expressed in neurons and glia. J Cell Biol 147:645-658.

Blake DJ, Weir A, Newey SE, Davies KE (2002) Function and genetics of dystrophin and dystrophin-related proteins in muscle. Physiol Rev 82:291-329.

Bloch O, Auguste KI, Manley GT, Verkman AS (2006) Accelerated progression of kaolin-induced hydrocephalus in aquaporin-4-deficient mice. J Cereb Blood Flow Metab26: 1527-1537.

Bowe MA, Deyst KA, Leszyk JD, Fallon JR (1994) Identification and purification of an agrin receptor from Torpedo postsynaptic membranes: a heteromeric complex related to the dystroglycans. Neuron 12:1173-1180.

122 Buono RJ, Lohoff FW, Sander T, Sperling MR, O'Connor MJ, Dlugos DJ, Ryan SG, Golden GT, Zhao H, Scattergood TM, Berrettini WH, Ferraro TN (2004) Association between variation in the human KCNJ10 potassium ion channel gene and seizure susceptibility. Epilepsy Res 58:175-183.

Cerna J, Cerecedo D, Ortega A, Garcia-Sierra F, Centeno F, Garrido E, Mornet D, Cisneros B (2006) Dystrophin Dp71f Associates with the pi-Integrin Adhesion Complex to Modulate PC12 Cell Adhesion. J Mol Biol 352:954-965.

Chiyonobu T, Sasaki J, Nagai Y, Takeda S, Funakoshi H, Nakamura T, Sugimoto T, Toda T (2005) Effect of fukutin deficiency in the developing mouse brain. Neuromuscul Disord 15:416- 426.

Colledge M and Froehner SC (1998) Signals mediating ion channel clustering at the neuromuscular junction. Curr Opin Neurobiol 8:357-363.

Connors NC and Kofuji P (2002) Dystrophin Dp71 is critical for the clustered localization of potassium channels in retinal glial cells. J Neurosci 22:4321-4327.

Connors NC and Kofuji P (2006) Potassium channel Kir4.1 macromolecular complex in retinal glial cells. Glia 52:124-131.

Connors NC, Adams ME, Froehner SC, Kofuji P (2004) The Potassium Channel Kir4.1 Associates with the Dystrophin-Glycoprotein Complex via a-Syntrophin in Glia. J Biol Chem 279:28387-28392.

Cormand B, Avela K, Pihko H, Santavuori P, Talim B, Topaloglu H, de la Chapelle A, Lehesjoki AE (1999) Assignment of the muscle-eye-brain disease gene to Ip32-p34 by linkage analysis and homozygosity mapping. Am J Hum Genet 64:126-135.

Cote PD, Moukhles K, Carbonetto S (2003) Dystroglycan is not required for localization of dystrophin, syntrophin, and neuronal nitric-oxide synthase at the sarcolemma but regulates integrin alpha 7B expression and caveolin-3 distribution. J Biol Chem 277:4672-4679.

Cote PD, Moukhles H, Lindenbaum M, Carbonetto S (1999) Chimaeric mice deficient in dystroglycans develop muscular dystrophy and have disrupted myoneural synapses. Nat Gen 23: 338-342.

Dalloz G, Sarig R, Fort P, Yaffe D, Bordais A, Pannicke T, Grosche J, Mornet D, Reichenbach A, Sahel J, Nudel U, Rendon A (2003) Targeted inactivation of dystrophin gene product Dp71: phenotypic impact in mouse retina. Hum Mol Genet 12:1543-1554.

123 Deconinck AE, Potter AC, Tinsley JM, Wood SJ, Vater R, Young C, Metzinger L, Vincent A, Slater CR, Davies KE (1997a) Postsynaptic abnormalities at the neuromuscular junctions of utrophin-deficient mice. J Cell Biol 136:883-894.

Deconinck AE, Rafael JA, Skinner JA, Brown SC, Potter AC, Metzinger L, Watt DJ, Dickson JG, Tinsley JM and Davies KE (1997b) Utrophin-Dystrophin-Deficient Mice as a Model for Duchenne Muscular Dystrophy. Cell 90:717-727. del Zoppo GJ and Milner R (2006) Integrin-matrix interactions in the cerebral micro vasculature. Arterioscler Thromb Vase Biol 26:1966-1975.

Dobyns WB, Paagon RA, Armstrong D, Curry CJR, Greenberg F, Grix A, Holmes LB, Laxova R, Michels VV, Robinow R, Zimmerman RL, Opitz JM, Reynolds JF (2005) Diagnostic criteria of walker-warburg syndrome. Am J Med Gen 32:195- 210.

Eid T, Lee TW, Thomas MJ, Amiry-Moghaddam M, Bjornsen LP, Spencer DD, Agre P, Ottersen OP, de Lanerolle NC (2005) Loss of perivascular aquaporin 4 may underlie deficient water and K+ homeostasis in the human epileptogenic hippocampus. PNAS 102:1193-1198.

Emery A (2001) Duchenne muscular dystrophy or Meryon's disease. In The muscular dystrophies. Edited by Alan E.H. Emery. Oxford University Press pp. 55-68.

Ferns M, Deiner M, Hall Z (1996) Agrin-induced acetylcholine receptor clustering in mammalian muscle requires tyrosine phosphorylation. J Cell Biol 132:937-944.

Frigeri A, Nicchia GP, Nico B, Quondamatteo F, Herken R, Roncali L, Svelto M (2001) Aquaporin-4 deficiency in skeletal muscle and brain of dystrophic mdx mice. FASEB J 15:90-98.

Fukuyama Y, Osawa M, Suzuki H (1981) Congenital muscular dystrophy of the Fukuyama type - clinical, genetic and pathological considerations. Brain Dev 3: 1- 29.

Gavillet B, Rougier JS, Domenighetti A A, Behar R, Boixel C, Ruchat P, Lehr HA, Pedrazzini T, Abriel H (2006) Cardiac sodium channel Navl.5 is regulated by a multiprotein complex composed of syntrophins and dystrophin. Circ Res 99:407- 414.

Gee SH, Medhavan R, Levinson SR, Caldwell JH, Sealock R, Froehner SC (1998) Interaction of muscle and brain sodium channels with multiple members of the syntrophin family of dystrophin-associated proteins. J Neurosci 18: 128-137.

124 Gee SH, Montanaro F, Lindenbaum MH and Carbonetto S (1994) Dystroglycan-alpha, a dystrophin-associated glycoprotein, is a functional agrin receptor. Cell 77: 675-686.

Gleeson JG and Walsh CA (2000) Neuronal migration disorders: from genetic disorders to developmental mechanisms. Trends Neurosci 23:352-359.

Grady RM, Zhou H, Cunningham JM, Henry MD, Campbell KP, Sanes JR (2001) Maturation and maintenance of the neuromuscular synapse: genetic evidence for roles of the dystrophin—glycoprotein complex. Neuron 25:279-293.

Grewal PK, McLaughlan JM, Moore CJ, Browning CA, Hewitt JE (2005) Characterization of the LARGE family of putative glycosyltransferases associated with dystroglycanopathies. Glycobiol 15:912-923.

Guadagno E and Moukhles H (2004) Laminin-induced aggregation of the inwardly rectifying potassium channel, Kir4.1, and the water-permeable channel, AQP4, via a dystroglycan-containing complex in astrocytes. Glia 47:138-149.

Hayashi YK, Chou FL, Engvall E, Ogawa M, Matsuda C, Hirabayashi S, Yokochi K, Ziober BL, Kramer RH. Kaufman SJ, Ozawa E, Goto Y, Nonaka I, Tsukahara T, Wang JZ, Hoffman EP, Arahata K (1998) Mutations in the integrin alpha7 gene cause congenital myopathy. Nat Genet 19:94-97.

Hernandez-Gonzalez EO, Martinez-Rojas D, Mornet D, Rendon A, Mujica A (2001) Comparative distribution of short dystrophin superfamily products in various guinea pig spermatozoa domains. Eur J Cell Biol 80:792-798.

Hernandez-Gonzalez EO, Mornet D, Rendon A, Martinez-Rojas D (2005) Absence of Dp71 in mdx3cv mouse spermatozoa alters flagellar morphology and the distribution of ion channels and nNOS. J Cell Sci 118:137-145.

Hodges BL, Hayashi YK, Nonaka I, Wang W, Arahata K, Kaufman SJ (1997) Altered expression of the alpha7betal integrin in human and murine muscular dystrophies. J Cell Sci 110:2873-2881.

Holthoff K and Witte OW (2000) Directed spatial potassium redistribution in rat neocortex. Glia 29:288-292.

Holzfeind PJ, Grewal PK, Reitsamer HA, Kechvar J, Lassmann H, Hoeger H, Hewitt JE, Bittner RE (2002) Skeletal, cardiac and tongue muscle pathology, defective retinal transmission, and neuronal migration defects in the Large(myd) mouse defines a natural model for glycosylation-deficient muscle-eye-brain disorders. Hum Mol Genet 11:2673-2687.

Hynes RO (1992) Integrins: Versatility, modulation, and signaling in cell adhesion. Cell 69:11-25.

125 Ishii M, Horio Y, Tada Y, Hibino H, Inanobe A, Ito M, Yamada M, Gotow T, Uchiyama Y and Kurachi Y (1997) Expression and clustered distribution of an inwardly rectifying potassium channel, KAB-2/Kir4.1, on mammalian retinal Muller cell membrane: their regulation by insulin and laminin signals. J Neurosci 17: 7725- 7735.

Jacobson C, Cote PD, Rossif SG, Rotundo RL, Carbonetto S (2001) The Dystroglycan Complex Is Necessary for Stabilization of Acetylcholine Receptor Clusters at Neuromuscular Junctions and Formation of the Synaptic Basement Membrane. JCB 152:435-450.

Jacobson C, Montanaro F, Lindenbaum M, Carbonetto S, Ferns M (1998) Alpha- Dystroglycan functions in acetylcholine receptor aggregation but is not a coreceptor for agrin-MuSK signaling. J Neurosci 18:6340-6348.

Kanagawa M, Michele DE, Satz JS, Barresi R, Kusano H, Sasaki T, Timpl R, Henry MD, Campbell KP (2005) Disruption of perlecan binding and matrix assembly by post- translational or genetic disruption of dystroglycan function. FEBS Lett 579:4792- 4796.

Khan OH, Enno TL, Del Bigio MR (2006) Brain damage in neonatal rats following kaolin induction of hydrocephalus. Exp Neurol 200:311-320.

Kofugi P, Biedermann B, Sidderharthan V, Raap M, Iandiey I, Milenkovic I, Thomzig A, Veh RW, Bringmann A and Reichenbach A (2002) Kir potassium channel subunit expression in retinal glial cells: implications for spatial potassium buffering. Glia 39: 292-303.

Kong J and Anderson JE (1999) Dystrophin is required for organizing large acetylcholine receptor aggregates. Brain Res 839:298-304.

Kramarcy NR and Sealock R (2000) Syntrophin Isoforms at the Neuromuscular Junction: Developmental Time Course and Differential Localization. Mol Cell Neurosci 15: 262-274.

Lupa MT, Krzemien DM, Schaller KL, Caldwell JH (1993) Aggregation of sodium channels during development and maturation of the neuromuscular junction. J Neurosci 13:1326-1336.

Lyons PR and Slater CR (1991) Structure and function of the neuromuscular junction in young adult mdx mice. J Neurocytol 20: 969-981

Manley GT, Fujimura M, Ma T, Noshita N, Filiz F, Bollen AW, Chan P, Verkman AS (2000) Aquaporin 4 deletion in mice reduces brain edema after acute water intoxication and ischemic stroke. Nat Med 6:159-163.

126 Mayer U (2003) Integrins: Redundant or Important Players in Skeletal Muscle? J Biol Chem 278:14587-14590.

Michele DE, Barresi R, Kanagawa M, Saito F, Cohn RD, Satz JS, Dollar J, Nishino I, Kelley RI, Somer H, Straub V, Mathews KD, Moore SA, Campbell KP (2002) Post-translational disruption of dystroglycan-ligand interactions in congenital muscular dystrophies. Nature 418:417-422.

Montanerro F and Carbonetto S (2003) Targeting dystroglycan in the brain. Neuron 37: 193-196.

Moore SA, Saito F, Chen J, Michele DE, Henry MD, Meesing A, Cohn RD, Ross-Barta SE, Westra S, Williamson RA, Hoshi T and Campbell KP (2002) Deletion of brain dystroglycan recapitulates aspects of congenital muscular dystrophy. Nature 418: 422-425.

Nagelhus EA, Veruki ML, Torp R, Haug FM, Laake JH, Nielsen S, Agre P and Ottersen OP (1998) Aquaporin-4 water channel protein in the rat retina and optic nerve: polarized expression in Muller cells and fibrous astrocytes. J Neurosci 18: 2506- 2519.

Neely JD, Amiry-Moghaddam M, Ottersen OP, Froehner SC, Agre P, Adams ME (2001) Syntrophin-dependent expression and localization of Aquaporin-4 water channel protein. PNAS 98:14108-14113.

Niermann H, Amiry-Moghaddam M, Holthoff K, Witte OW, Ottersen OP (2001) A novel role of vasopressin in the brain: modulation of activity-dependent water flux in the neocortex. J Neurosci 21:3045-3051.

Newman EA and Odette LL (1984) Model of electroretinogram b-wave generation: a test of the potassium hypothesis. J Neurophysiol 51: 164-182.

Noel G, Belda M, Guadagno E, Micoud J, Klocker, Moukhles H (2005) Dystroglycan and Kir4.1 coclustering in retinal Muller glia is regulated by laminin-1 and requires the PDZ-ligand domain of Kir4.1. J Neurochem 94:691-702.

Occhi S, Zambroni D, Del Cairo U, Amadio S, Sirkowski EE, Scherer SS, Campbell KP, Moore SA, Chen ZL, Strickland S, Di Muzio A, Uncini A, Wrabetz L, Feltri ML (2005) Both laminin and Schwann cell dystroglycan are necessary for proper clustering of sodium channels at nodes of Ranvier. J Neurosci 25:9418-9427.

Papadopoulus MC, Manley GT, Krishna S, Verkman AS (2004) Aquaporin 4 facilitates reabsorption of excess fluid in vasogenic brain edema. FASEB J 18:1291-1293.

Peters MF, Adams ME, Froehner SC (1997) Differential association of syntrophin pairs with the dystrophin complex. J Cell Biol 138:81-93.

127 Puwarawuttipanit W, Bragg AD, Frydenlund DS, Mylonakou MN, Nagelhus EA, Peters MF, Kotchabhakdi N, Adams ME, Froehner SC, Haug FM, Ottersen OP, Amiry- Moghaddam M (2006) Differential effect of alpha-syntrophin knockout on aquaporin-4 and Kir4.1 expression in retinal macroglial cells in mice. Neurosci 137: 165-175.

Ribaux P, Bleicher F, Couble ML, Amsellem J, Cohen SA, Berthier C, Blaineau S (2001) Voltage-gated sodium channel (SkMl) content in dystrophin deficient muscle. Pflugers Arch 441:746-755.

Rooney JE, Welser JV, Dechert MA, Flintoff-Dye NL, Kaufman SJ, Burkin DJ (2006) Severe muscular dystrophy in mice that lack dystrophin and alpha7 integrin. J Cell Sci 119:2185-2195.

Saito F, Moore SA, Barresi R, Henry MD, Messing A, Ross-Barta SE, Cohn RD, Williamson RA, Sluka KA, Sherman DL, Brophy PJ, Schmelzer JD, Low PA, Wrabetz L, Feltri ML, Campbell KP (2003) Unique Role of Dystroglycan in Peripheral Nerve Myelination, Nodal Structure, and Sodium Channel Stabilization. Neuron 38:747-758.

Sealock R, Butler MH, Kramarcy NR, Gao KX, Murnane AA, Douville K, Froehner SC (1991) Localization of dystrophin relative to acetylcholine receptor domains in electric tissue and adult and cultured skeletal muscle. J Cell Biol 113:1133-1144.

Shang L, Lucchese CJ, Haider S, Tucker SJ (2005) Functional characterisation of missense variations in the Kir4.1 potassium channel (KCNJ10) associated with seizure susceptibility. Mol Brain Res 139:178-183.

Taniguchi M, Kurahashi H, Noguchi S, Fukudome T, Okinaga T, Tsukahara T, Tajima Y, Ozono K, Nishino I, Nonaka I, Toda T (2006) Aberrant neuromuscular junctions and delayed terminal muscle fiber maturation in oc-dystroglycanopathies. Hum Mol Gen 15:1279- 1289.

Ueda H, Baba T, Kashiwagi K, Iijma H, Ohno S (2000) Dystrobrevin localization in photoreceptor axon terminals and at blood ocular barrier sites. Inv Opth Vis Sci 49: 3908-3914.

Vajda Z, Pedersen M, Fuchtbauer EM, Wertz K, Stodkilde-Jorgensen H, Sulyok E, Doczi T, Neely JD, Agre P, Frokiaer J, Nielsen S (2002) Delayed onset of brain edema and mislocalization of aquaporin-4 in dystrophin-null transgenic mice. PNAS 99:13131- 13136.

Vajsar J and Schachter H (2006) Walker-Warburg Syndrome. Orphanet J Rare Dis 1:29- 33.

128 Verkman As, Binder DK, Block O, Auguste K, Papadopoulos MC (2006) Three distinct roles of aquaporin-4 in brain function revealed by knockout mice. Biochim Biophys Acta 1758:1085-1093.

Wood SJ and Slater CR (1998) P-Spectrin is colocalized with both voltage-gated sodium channels and ankyrinG at the adult rat neuromuscular junction. J Cell Biol 140:675- 684.

Yoshioka M and Higuchi Y (2005) Long-term prognosis of epilepsies and related seizure disorders in Fukuyama-type congenital muscular dystrophy. J Child Neurol 20:385- 391.

129 4 APPENDICES

4.1 Appendix A-Supplemental Data

130 Figure 4.1 AQP4 and Kir4.1 fail to localize to the glia limitans in the Largemy brain. Horizontal brain sections containing cerebral cortex from wild type (A, B, C; G, H, I) and Largemyd (D, E, F; J, K, L) mice were double immunolabeled for AQP4 (A, D) and perlecan (B, E) or Kir4.1 (G, J) and perlecan (H, K). Notice the absence of labeling for Kir4.1 (J) and AQP4 (D) at the glia limitans in the Largemyd. Arrows demarcate the presence (H) and absence (K) of perlecan labeling at the glia limitans. The merged images show the co-distribution between AQP4 and perlecan (C) or Kir4.1 and perlecan (I). Scale bar, 25 pm for A-F, 50 pm for G-L.

131 Figure 4.2 Radial Muller cells appear discontinuous and tortuous in the Largemyd retina. Retina sections from wild type (A,B) and Largemyd (D, E) mice were double immunolabeled for vimentin (A, D) and Kir4.1 (B, E). Muller cell processes spanning the ganglion cell and inner plexiform layers appear radially aligned in the wild type (A) retina. This organization is disrupted and replaced by tortuous, discontinuous processes in the Largemyd retina. Note the persistence of Kir4.1 at Muller cell processes and perivascular domains (asterix E) in the Largemyd. Merged images show the localization of Kir4.1 to radial Muller cells (C, F) and perivascular glia (arrows). Scale bar, 25 um.

132