Functional Studies of VANGL Mutations Associated with Neural Tube Defects

Functional studies of VANGL mutations associated with neural tube defects.

Alexandra Iliescu

Department of Biochemistry McGill University Montreal, Quebec, Canada

February 2015

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

Abstract

Loop-tail (Lp) mice show a very severe neural tube defect (craniorachischisis) caused by mutations in the Vangl2 gene (D255E, R259L, S464N). Mammalian VANGL1 and VANGLl2 are membrane proteins that play critical roles in development such as establishing planar cell polarity (PCP) in epithelial layers and convergent extension (CE) movements during neural tube closure. In this thesis, the molecular mechanism of VANGL proteins was investigated. Chapters 2 and 6 explore structure-function relationships, while Chapters 3, 4 and 5 study the molecular basis of loss-of-function of mutations either Lp-associated or identified in human cases of neural tube defects (NTDs). In Chapter 2, we used epitope tagging and immunofluorescence to establish the secondary structure of VANGL proteins, including the number, position, and polarity of transmembrane domains. These studies indicate that VANGL proteins have a four-transmembrane domain structure and that both the amino and large carboxy termini portions of the protein are located intracellularly. Work performed in Chapters 3 and 4 show that all three Lp-associated mutations (D255E, R259L, D464N) share a common loss-of-function mechanism. While WT VANGL proteins are expressed at the plasma membrane of transfected MDCK cells, their site for biological function, Lp-associated mutations are retained intracellularly in the endoplasmic reticulum, have reduced half-life and are rapidly degraded in a proteasome-dependent fashion. These studies provided a biochemical framework for the study of recently identified mutations in VANGL1 and VANGL2 in sporadic or familial cases of NTDs. Chapter 5 focuses on two arginine residues R181 and R274, that are highly conserved in VANGL protein relatives, and that are found independently mutated in VANGL1 (R181Q and R274Q) and VANGL2 (R177H and R270H) in human cases of NTDs. Compared to WT, the R181Q and R274Q mutations also displayed impaired targeting to the plasma membrane, showed impaired stability and a large reduction in half-life. In Chapter 6 a set of linear deletions in the amino and carboxyl termini of VANGL1 were created to delineate the sequence determinants that are required for targeting these proteins to the plasma membrane. This work identified Vangl plasma membrane motifs to be restricted to the carboxy terminus, including the PDZ binding motif.

Résumé

La souris loop-tail (Lp) meurt d’une anomalie très grave du tube neural (craniorachischisis) causée par des mutations dans le gène Vangl2 (D255E, R259L, S464N). Chez les mammifères, il existe deux gènes; Vangl1 et Vangl2 qui codent pour des protéines membranaires. Celles-ci jouent un rôle critique dans le développement, contrôlant l'établissement de la polarité cellulaire planaire (PCP) chez les cellules épithéliales ainsi que le processus d’extension convergente lors de la fermeture du tube neural. Dans cette thèse, le mécanisme moléculaire des protéines VANGL a été étudié. Les chapitres 2 et 6 enquêtent sur les relations entre la structure et la fonction de la protéine, tandis que les chapitres 3, 4 et 5 étudient le mécanisme moléculaire associé à la perte de fonction chez les mutations Lp ou bien chez celles identifiées chez des patients souffrant d’anomalies de tubes neuraux (ATN). Au chapitre 2, nous avons établi la structure secondaire des protéines VANGL, y compris le nombre, la position et la polarité des domaines transmembranaires. Ces études indiquent que ces protéines possèdent quatre domaines transmembranaires et que les deux extrémités amino et carboxy terminales sont situées au niveau intracellulaire. Le travail effectué dans les chapitres 3 et 4 montre que les trois mutations Lp (D255E, R259L, D464N) partagent un mécanisme de perte de fonction commun. Tandis que les protéines VANGL WT sont exprimées à la membrane cellulaire des cellules MDCK transfectées, les mutations Lp sont retenues dans le réticulum endoplasmique, elles possèdent une demi-vie réduite et sont rapidement dégradées par le protéasome. Ces études ont fourni un cadre de tests biochimiques pour analyser les mutations récemment trouvées dans VANGL1 et VANGL2 dans des cas sporadiques ou familiaux d’anomalies de tube neural. Le chapitre 5 se concentre sur deux arginines; R181 et R274, qui sont hautement conservées chez les homologues des protéines VANGL, et qui sont aussi mutées de façon indépendante dans VANGL1 (R181Q et R274Q) et VANGLl2 (R177H et R270H) dans des cas humains d’ATN. Comparativement au WT, les mutations R181Q et R274Q ne sont pas ciblées à la membrane plasmique, elles affichent une stabilité moindre ainsi qu’une réduction importante de la demi-vie. Dans le chapitre 6, une série de délétions linéaires à l'extrémité amino-et carboxy- terminales de VANGL1 ont été créés afin de déterminer les signaux dans la séquence nécessaires au ciblage de ces protéines à la membrane cellulaire. Ces travaux ont identifié que ces motifs se situent à l'extrémité carboxy-terminale, y compris au domaine de liaison PDZ.

Preface

The work described in Chapters 2, 3, 4, 5 and 6 of this thesis are published as follows:

Chapter 2: Iliescu A, Gravel M, Horth C, Apuzzo S, Gros P. (2011) Transmembrane topology of mammalian planar cell polarity protein VANGL1. Biochemistry. 50(12):2274-82.

Chapter 3: Gravel M, Iliescu A, Horth C, Apuzzo S, Gros P. (2010) Molecular and cellular mechanisms underlying neural tube defects in the loop-tail mutant mouse. Biochemistry. 49(16):3445-55.

Chapter 4: Iliescu A, Gravel M, Horth C, Kibar Z, Gros P. (2011) Loss of membrane targeting of VANGL proteins causes neural tube defects. Biochemistry. 50(5):795-804.

Chapter 5: Iliescu A, Gravel M, Horth C, Gros P. (2014) Independent mutations at Arg181 and Arg274 of VANGL proteins that are associated with neural tube defects in humans decrease protein stability and impair membrane targeting. Biochemistry. (Epub ahead of print)

Chapter 6: Iliescu A, Gros P. (2014) The intracellular carboxy terminal domain of VANGL proteins contains plasma membrane targeting elements. Protein Science. 23(4):337-43.

Contribution of Authors

Chapter 2: The work described in this chapter was mainly performed by myself. Prior to my arrival in the lab, Sergio Appuzo created all six c-Myc/HA-tagged VANGL1 constructs used in this study. In order to facilitate the detection of positively transfected clones, Dr. Michel Gravel fused the constructs in-frame to GFP. I performed hydrophathy profiling, stably transfected and expressed the constructs in MDCK cells, characterized the subcellular localization of the recombinant proteins and quantified cell surface expression. Cynthia Horth provided valuable microscopy advice and helped perform some of the image acquisition. All the figures in this manuscript are based on experiments that I performed. I constructed all the figures and wrote the first draft of the manuscript.

Chapter 3: The work described in this chapter was produced in collaboration with Dr. Michel Gravel and Cynthia Horth. I performed the subcellular localization and membrane targeting studies. I participated, along with Dr. Gravel in constructing the figures and the writing of the manuscript. Dr. Elena Torban did a critical reading of the manuscript. Dr. Gravel helped in the creation of the mutant constructs.

Chapter 4: The work described in this chapter was mainly performed by myself and in collaboration with Dr. Michel Gravel and Cynthia Horth. Dr. Gravel helped in the creation of the mutant constructs. I performed the subcellular localization and membrane targeting studies. Dr. Michel Gravel carried out the Pulse-Chase experiments (figure 6) and Cynthia Horth performed the coexpression of WT and Lp mutations in MDCK cells (figure 9). I constructed all the figures and wrote the first draft of the manuscript. Dr. Zoha Kibar and Dr. Elena Torban did a critical reading of the manuscript.

Chapter 5: The work described in this chapter was mainly performed by myself. Prior to my arrival on the project, Dr. Zoha Kibar initially sequenced the cohort of human patients and identified all three 5 mutations (R181Q, L202F and R274Q). I created the mutants in collaboration with Dr. Gravel and I expressed the constructs, characterized their proper membrane targeting, subcellular localization, protein stability, proteasomal dependent degradation and determined the half-life of the proteins. Michel Gravel performed the pulse-chase of the R181Q mutant and Cynthia Horth helped select positive clones and perform some of the immunofluorescence and image acquisition. I constructed all the figures and wrote the first draft of the manuscript.

Chapter 6: The work described in this chapter is essentially my own. Dr. Michel Gravel provided valuable technical advice on the strategy for creating the deletion constructs and Jean-Daniel Castonguay, an undergraduate student that I supervised, helped me with the molecular cloning steps. I created all the deletion constructs, stably expressed them in MDCK cells and characterized their subcellular localization. All the figures in this manuscript are based on experiments that I performed. I constructed all the figures and wrote the first draft of the manuscript. Dr. Michel Gravel did a critical reading of the manuscript.

My supervisor, Dr. Philippe Gros examined results, provided feedback and guidance on this project through my graduate studies. He also edited all the final manuscripts included in this thesis.

Table of Contents

Résumé ...... 3 Preface ...... 4 Contribution of Authors ...... 5 Table of Contents ...... 7 List of Figures ...... 11 List of Tables...... 14 Acknowledgements ...... 15 Objectives and Rationale ...... 16 Chapter 1: Introduction and literature review ...... 17 1.1 Neurulation ...... 18 1.1.1 Primary neurulation ...... 18 1.1.2 Secondary neurulation ...... 20 1.2 Neural tube defects (NTDs) ...... 21 1.2.1 Open NTDs ...... 21 1.2.2 Closed NTDs ...... 23 1.2.3 Analogy to humans...... 23 1.3 Etiology of NTDs ...... 23 1.3.1 Environmental determinants ...... 23 1.3.2 Genetic determinants ...... 24 1.4 Loop-tail mouse ...... 24 1.4.1 Ltap/Vangl2 ...... 26 1.4.1.1 Lp-associated independent alleles ...... 26 1.4.2 Additional phenotypes...... 27 1.5 VANGL1 and VANGL2 ...... 28 1.5.1 Gene description ...... 28 1.5.2 Protein description...... 29 1.6 Study of Vangl genes in different species...... 29 1.6.1 Insight from flies ...... 29

1.6.1.1 Establishment of Planar Cell Polarity (PCP) ...... 30 1.6.2 Insight from vertebrates ...... 34 1.6.2.1 Frog (Xenopus laevis) ...... 35 1.6.2.2 Zebrafish (Danio rerio) ...... 35 1.6.3 Insight from mammals ...... 35 1.6.3.1 VANGL1 and VANGL2 mouse mutants ...... 36 1.6.3.2 Additional PCP mouse mutants ...... 39 1.6.4 Insight from VANGL1 and VANGL2 human mutations ...... 39 1.7 Cellular and molecular pathogenesis of Vangl genes in vertebrate and mammals ...... 44 1.7.1 In vivo asymmetric distribution ...... 44 1.7.2 Genetic interactions ...... 44 1.7.3 Physical interactions ...... 44 1.7.3.1 ER-Golgi transport ...... 45 1.7.3.2 Export out of trans-Golgi network (TGN) ...... 45 1.7.4 Phosphorylation ...... 45 1.7.5 Mode of inheritance of Lp mutations ...... 46 1.8 Conclusion ...... 46 Chapter 2: Transmembrane topology of the mammalian planar cell polarity protein VANGL1 . 48 2.1 Abstract ...... 50 2.2 Introduction ...... 51 2.3 Experimental Procedure ...... 54 2.4 Results ...... 58 2.5 Discussion ...... 69 2.6 Acknowledgments ...... 73 2.7 Supplementary Figures...... 74 Chapter 3: Molecular and cellular mechanisms underlying neural tube defects in the loop-tail mutant mouse ...... 76 3. 1 Abstract ...... 78 3. 2 Introduction ...... 79 3.3 Experimental Procedures ...... 81 3.4 Results ...... 86

3.5 Discussion ...... 103 3.6 Acknowledgements ...... 107 3.7 Supplementary Figures...... 108 Chapter 4: Loss of membrane targeting of VANGL proteins causes neural tube defects ...... 110 4.1 Abstract ...... 111 4.2 Introduction ...... 112 4.3 Experimental Procedures ...... 115 4.4 Results ...... 120 4.5 Discussion ...... 137 4.6 Acknowledgments ...... 141 4.7 Supplementary Figures...... 142 Chapter 5: Independent mutations at Arg181 and Arg274 of VANGL proteins that are associated with neural tube defects in humans decrease protein stability and impair membrane targeting. 144 5.1 Abstract ...... 146 5.2 Introduction ...... 147 5.3 Experimental Procedures ...... 150 5.4 Results ...... 154 5.5 Discussion ...... 163 5.6 Acknowledgements ...... 166 5.7 Supplementary Figures...... 167 Chapter 6: The intracellular carboxyl terminal domain of VANGL proteins contains plasma membrane targeting signals ...... 168 6.1 Abstract ...... 170 6.2 Introduction ...... 171 6.3 Experimental Procedures ...... 173 6.4 Results ...... 175 6.6 Acknowledgements ...... 183 Chapter 7: Summary and future perspectives ...... 184 7.1 Summary ...... 185 7.1.1 Subcellular targeting and trafficking of VANGL proteins...... 186 7.1.1.1 Sequence motifs for membrane targeting ...... 186

7.1.1.2 SEC24B dependent membrane trafficking ...... 187 7.1.1.3 Phosphorylation ...... 188 7.1.2 Future perspectives ...... 188 7.1.2.1 Biochemical tests ...... 188 7.1.2.2 Yet to be discovered genetic and environmental factors ...... 189 7.1.3 Final conclusions ...... 191 Original Contributions to Knowledge ...... 192 References ...... 193

List of Figures

Chapter 1 1. 1: Convergent Extension (CE) movements...... 19 1. 2: Primary Neurulation...... 20 1. 3: Secondary Neurulation...... 20 1. 4: Closure Sites...... 22 1. 5: Phenotype of Lp/+ and of Lp/Lp mice as compared to WT...... 25 1. 6: PCP in the Drosophila wing...... 31 1. 7: Proposed model of PCP in Drosophila...... 33 1. 8: Mutational map of VANGL1 and VANGL2 in neural tube defects...... 42

Chapter 2 2. 1: Construction of VANGL1 proteins containing hemagglutinin (HA) epitope tags...... 60 2. 2: Expression of HA epitope-tagged VANGL1 constructs in MDCK cells...... 62 2. 3: Cellular localization HA epitope-tagged VANGL1 proteins transfected in MDCK cells. ... 64 2. 4: Immunofluorescence analysis of HA epitope-tagged VANGL1 proteins in intact and permeabilized cells...... 66 2. 5: Detection of HA epitope-tagged proteins by surface labeling...... 68 2. 6: Topological model and structural features of the VANGL protein family...... 71

2. S 1: TMHMM probability plot and comparison with TOPPRED2 model...... 74 2. S 2: Immunofluorescence analysis of HA epitope-tagged VANGL1 proteins without the GFP in intact and permeabilized cells...... 75

Chapter 3 3. 1: Expression of WT and D255E VANGL1 variant in the membrane fraction of transfected MDCK cells...... 88 3. 2: Cellular localization of WT and D255E VANGL1 proteins in transfected MDCK cells. .... 90

3. 3: Cellular localization of the D255E mutant VANGL1 variant in stably transfected MDCK cells; Z-line image analysis...... 92 3. 4: Analysis of plasma membrane targeting and surface expression of WT and D255E VANGL1 variants in MDCK cells...... 94 3. 5: Stability of GFP- VANGL1 WT and D255E proteins in MDCK cells...... 96 3. 6: Pulse-chase studies of GFP- VANGL1 WT and D255E proteins expressed in MDCK cells...... 98 3. 7: Determination of subcellular localization of VANGL1 D255E by double immunofluorescence...... 100 3. 8: Proteasome-mediated degradation of WT and D255E VANGL1 expressed in MDCK cells...... 102

3. S 1: Immunolocalization of GFP- VANGL1 WT and D255E in MDCK cells...... 108 3. S 2: Degradation and UPR response of WT and D255E VANGL1 proteins expressed in MDCK cells...... 109

Chapter 4

4. 1: Schematic representation of the GFP-VANGL1 protein...... 121 4. 2: Expression of WT, R259L and S464N VANGL1 variants in the membrane fraction of transfected MDCK cells...... 123 4. 3: Cellular localization of WT, R259L and S464N VANGL1 variants in transfected MDCK cells...... 125 4. 4: Subcellular localization of the WT, R259L and S464N VANGL1 variants in stably transfected MDCK cells; Z-line image analysis...... 127 4. 5: Analysis of plasma membrane targeting and surface expression of WT, R259L and S464N VANGL1 variants in MDCK cells...... 129 4. 6: Pulse-chase studies of GFP- VANGL1 WT, R259L and S464N proteins expressed in MDCK cells...... 131 4. 7: Determination of subcellular localization of R259L and S464N variants by double immunofluorescence...... 133 4. 8: Cellular Degradation of WT, R259L and S464N VANGL1 variants in MDCK cells...... 135

4. 9: Coexpression and cellular localization of WT and Lp-associated variants (D255E, R259L and S464N) in MDCK cells...... 137

4. S 1: Expression of Cherry-VANGL1 WT protein in MDCK cells stably expressing GFP- VANGL1 WT, D255E, R259L and S464N proteins...... 142 4. S 2: Subcellular localization of the WT, D255E, R259L and S464N VANGL1 variants in stable doubly transfected MDCK cells coexpressing Cherry-VANGL1 WT; Z-line image analysis...... 143

Chapter 5 5. 1: Pedigree and schematic representation of human mutations (R181Q and R274Q) found in NTDs patients...... 155 5. 2: Cellular localization, surface expression and quantification of WT, R181Q and R274Q h VANGL1 mutations in stably transfected MDCK cells...... 158 5. 3: Subcellular localization of WT, R181Q and R274Q h VANGL1 in stably transfected MDCK cells...... 160 5. 4: Cellular stability, half –life, and degradation studies of WT, R181Q and R274Q hVANGL1 in stably transfected MDCK cells...... 162

5. S 1: Schematic representation, multiple sequence alignment and cellular localization of the L202F variant...... 167

Chapter 6 6. 1: Schematic representation and immunoblot expression of WT and deletion recombinant protein constructs...... 176 6. 2: Immunofluorescence for WT and deletion constructs of plasma membrane targeting and of cell surface expression in intact and permeabilized cells...... 178 6. 3: Subcellular localization of WT and deletion constructs...... 180

List of Tables

Chapter 1 1. Table 1: Phenotype of VANGL2 and VANGL1 mutant mice. Adapted from Torban et al. 2012...... 38 1. Table 2: Human mutations and clinical description...... 43

Chapter 2

2. Table 1: Oligonucleotides Used for Epitope Insertion by Site-Directed Mutagenesis...... 55

Acknowledgements

First, I would like to thank Dr. Philippe Gros for being a great mentor during my graduate studies; it really has been an immense privilege working for him. Coming to the lab on my first day, I could have never imagined the influence Philippe would have on me; he helped shape me scientifically and academically, but also life and career wise. I would like to thank him for his patience, his guidance and his advice on my project both when results looked exciting and promising, but especially also when they failed. In this past year, I am also extremely grateful for the career and life advice Philippe gave me, his support has truly been invaluable. I would also like to thank Dr. Michel Gravel and Cynthia Horth, as their work on the Vangl project has contributed enormously in making this thesis a reality. Their hard work, their help, and their sense of humour made working in the Vangl trio great fun every day. I also cannot thank enough Irena, Vicki and Silayuv! We started out as coworkers but we leave as amazing friends. You are incredibly smart, talented and beautiful young women and have made grad school both outside and inside the lab unforgettable. Also, I would like to give a big thank you to Susan and Normand and to all the past and present members of the Gros lab for their friendships and for making the lab a great environment to work in. On a more personal note, I would like to thank Ana Manescu and Zoe Mamalingas for your friendship, our endless talks and great hearts, you both are extremely important in my life. I would also like to thank my little brother Daniel for always coming to the rescue of my numerous computer problems. Finally, I dedicate this thesis to my parents Ana and Serban and to my husband Stefan. I would like to thank my parents for all their sacrifices and for always pushing me to be better and for always believing that I could achieve anything I set my mind to. Last, but not least I want to thank my husband, the love of my life, but also the person who drove me countless Sunday nights at the lab just to check on my cells, never once complaining. You are truly one in a billion and I am so incredible lucky to share my life with you. I would like to also acknowledge the Canadian institute of Health Research (CIHR) for providing financial support for this project.

Objectives and Rationale

In 2001, our laboratory identified Vangl2 as the gene mutated in the Lp/Lp mouse afflicted with the extremely severe neural tube defect (NTD) craniorachischisis. In vertebrates, there is a second Vangl gene which codes for a very similar protein (over 75% sequence similarity). Both Vangl1 and Vangl2 are extremely conserved in evolution and have been found to play critical roles in establishing planar cell polarity (PCP) and convergent extension (CE) movements during embryogenesis and neural tube closure. However, how VANGL proteins interact with other PCP proteins or how they mature, get trafficked and targeted to the plasma membrane, their site of biological function, remains poorly understood.

At the start of my thesis in September 2007, the exciting discovery, in VANGL1 of the first three human mutations found in both familial and sporadic cases of NTDs, pushed forth the work of also identifying the molecular mechanism causing the NTD phenotype in both mice and humans. The specific aim of my thesis was to gain insight into the structure-function relationship in the VANGL family of proteins in order to provide critical information about the normal protein and how it is altered during NTDs.

By using a set of biochemical and molecular biology assays established in vitro in stably transfected MDCK epithelial cells, work described here was instrumental in gaining insight into the etiology of NTDs. In Chapter 2, the elucidation of the structure of VANGL proteins at the plasma membrane was essential, in the absence of a 3D crystallography model, as not to rely solely on primary amino acid sequence. In chapters 3 and 4, the characterization of the molecular basis of loss-of-function, including loss of membrane targeting, in the three independent Lp-associated alleles (D255E, R259L and S464N) provided an experimental framework to study mutations identified in cohorts of human patients with NTDs. Chapter 5 characterized the pathological mechanism of two VANGL mutations (R181Q and R274Q) associated with neural tube defects in humans and identified a molecular mechanism identical to that associated with Lp. Finally, Chapter 6 identified sequence determinants in the C-terminal tail of VANGL proteins responsible for membrane targeting.

Chapter 1: Introduction and literature review

1.1 Neurulation

Neurulation is the process by which the brain and spinal cord form during embryonic development and culminates in the formation of the primordial structure of the central nervous system, the neural tube. In humans, this happen very early on in pregnancy, between 3 and 6 weeks gestation. In mice, the most studied animal model, neural tube formation occurs between embryonic days 8 and 10 (extensively reviewed in 1) Neurulation is traditionally divided into primary (which forms the brain and most of the spinal cord) and secondary (which is responsible for forming the most caudal portions of the spinal cord).

1.1.1 Primary neurulation

The neural tube originates from the neural plate, which itself arises as a thickening of surface ectoderm at the dorsal aspect of a developing vertebrate embryo. Prior to neurulation, the neural plate is flat, wide and relatively short. However, at the onset of neurulation, it undergoes an extensive re-shaping via a process known as convergent extension (CE). In this process, neuroectodermal cells of the neural plate form protrusions that are stabilized predominantly in the mediolateral direction. These oriented protrusions enable collective movement of neuroepithelial cells mediolaterally toward the midline, but also generate traction enabling these cells to intercalate with one another. As a result of CE, the neural plate is extended in the anterior-posterior axis (extension) and concomitantly becomes narrower in its mediolateral axis (convergence).

Two neural folds arise at the lateral sides of the neural plate. The folds elevate, bend toward each other, bridge the distance between them, meet and fuse with one another, and finally separate from the overlying ectoderm to form a closed neural tube. Almost the entire neural axis from the most rostral cranial region down to the sacral level of spinal cord is formed via this process. (Figure 1.1 and 1.2)

1. 1: Convergent Extension (CE) movements. Simplified representation of Convergent Extension (CE) during neural tube closure. (A) CE movements during morphogenesis lead to tissue elongation in the anteroposterior axis. (B) CE process facilitates neural tube closure by bringing the two neural folds closer together. Adapted from Niswander et al. 2013 2

1.1.2 Secondary neurulation

The sacral and coccygeal regions are generated through a process of secondary neurulation. The multipotent tail bud cells (derived from the embryo’s primitive streak) expand and form a rod- like neuroectodermal cell mass beneath the dorsal ectoderm of the tail bud. Within this mass, a central canal is formed via cavitation. The caudal portion of the neural tube merges with a more anterior part (generated through primary neurulation) to form a continuous neural tube with a single lumen. (Figure 1.3)

1. 2: Primary Neurulation. Schematic representation of primary neurulation Adapted from Niswander et al. 2013.2

1. 3: Secondary Neurulation. Schematic representation of secondary neurulation Adapted from Niswander et al. 2013. 2

1.2 Neural tube defects (NTDs)

Aberrations in the formation of the neural tube result in neural tube defects (NTDs) such as spina bifida and anencephaly, conditions in which the central canal of the malformed brain or spinal cord remains open to the environment. Spina bifida and anencephaly are very frequent in humans (1/1000 live births) and are second only to congenital heart defects as a cause of perinatal mortality 8. The mouse is an excellent animal model for the study of neural tube formation, because of the accessibility of the developing embryo, the abundance of mutants affecting neurulation, and the ease to manipulate environmental and genetic influences on this process 3,4.

1.2.1 Open NTDs

Open NTDs are characterized by the exposure of the brain and/or spinal cord to the environment. In mice, closure of the neural tube is initiated sequentially at 3 discreet closure sites, and failure of closure at any of these sites leads to various “open” neural tube defects (NTDs)1. The closure site I is at the hindbrain-cervical boundary, from which neural tube fusion spreads in rostral and caudal directions. Failure of closure I leads to the most severe NTD, craniorachischisis, a condition characterized by a completely open neural tube from the midbrain-hindbrain junction to the most caudal tail end region. Failure of the caudal spread of fusion from closure site I causes spina bifida (meningomyelocele). The separate closure site II is located at the forebrain-midbrain boundary, and the closure site III is found at the most rostral aspect of forebrain. Distortion of closure at sites II and III leads to a variety of cranial NTDs including, exencephaly (in mice) or anencephaly (in humans). (Figure 1.4)

1. 4: Closure Sites. Schematic representation of the pattern of anterior neural tube closure. The closure initiation points are shown (closure 1-3), and neural tube defects (anencephaly, spina bifida, craniorachischisis) associated with failure to initiate or propagate closure at these sites are indicated. Adapted from Torban et al. 5

1.2.2 Closed NTDs

Closed NTDs are less serious skin covered disorders that do not involve the brain and are a result of abnormal secondary neurulation. The emerging spinal cord cannot separate from adjoining tissues in the tail bud and cause caudal regression syndrome, abnormal tethering of spinal cord and other so-called “closed” NTDs 1.

1.2.3 Analogy to humans

It is still debated whether, in human fetuses, fusion of the neural tube is initiated from a single closure site or from multiple sites. However, by analogy with the mechanism of neural tube closure in mice, the latest consensus appears to favor a “multi-sited” closure model. The number of the closure sites in human neural tube is unclear, but is estimated to be from 3 to 5 6. Whatever the actual number, a multiple closure site model is required to explain all the known types of human NTDs. Since the mechanisms for neural tube closure appear to be evolutionarily conserved, genetic and biochemical studies in mice and other vertebrates, have provided insight into the pathological mechanisms of NTDs in humans.

1.3 Etiology of NTDs

The cellular and molecular mechanisms contributing to neurulation are largely unknown, and multiple factors, both genetic and environmental, have been implicated in the etiology of NTDs. 7,8.

1.3.1 Environmental determinants

Environmental evidence includes maternal intake of folic acid which has been shown in epidemiological studies to prevent 50 to 70% of NTDs. Additionally, other non-genetic factors include parental age, socioeconomic status, maternal gestational diabetes, maternal obesity,

23 hypothermia during early pregnancy, exposure to pesticides and the anticonvulsant drug valproic acid (VPA). 9,10.

1.3.2 Genetic determinants

Genetic factors, on the other hand, include an increased risk in twins and in families with one affected child. 11. Additionally, while 70% of NTDs are isolated, non-syndrome diseases, they have also been observed to occur with some genetic conditions such as trisomonies 13 and 18, chromosomal rearrangements and Meckle-Gruber Syndrome 12. In particular, the genetic determinants of NTDs have been difficult to study in humans due to genetic heterogeneity and incomplete penetrance. On the other hand, genetic lesions altering the normal process of neurulation and leading to NTDs have been extensively studied in mice, fish, frogs, and chicken, while some of the basic pathways involved were initially identified in the fly. In effect, a large number of either spontaneously arising or experimentally induced mouse NTDs mutants have been described 4. The identification and characterization of mutations causing NTDs in mice provide an entry point for not only studying the role of these genes, proteins, and also cellular mechanisms of neurulation, but also for identifying genes and pathways involved in the pathogenesis of NTDs in humans. In particular, genes of the planar cell polarity (PCP) pathway have been strongly implicated.

1.4 Loop-tail mouse

A large number of mouse mutants harboring different types of NTDs have been described over the years 4. Amongst them, loop-tail (Lp), a semi-dominant mutation on chromosome 1, was originally described by Strong in 1949 13. The Lp/+ heterozygotes display a characteristic “looped” or severely kinked tail (Figure 1.5) and show wobbly head movements. The Lp/Lp homozygotes die in utero of a severe NTD, craniorachischisis, that arises from a failure to initiate closure at the closure site I, and the neural tube remains completely open from midbrain-hindbrain juncture to tail 14.

A B

1. 5: Phenotype of Lp/+ and of Lp/Lp mice as compared to WT. (A) Lp/+ heterozygotes are recognizable by their characteristic “looped” tail, while (B) Lp/Lp homozygotes are affected with the NTD craniorachischisis and die in utero. Adapted from Torban et al. 5

1.4.1 Ltap/Vangl2

Using positional cloning and functional complementation in mutant mice, the gene mutated in Lp mouse was identified and given the provisional appellation Ltap (“loop-tail associated protein”), which was subsequently changed to Vangl2 (Van Gogh-like 2) 15,16. Ltap/Vangl2 mRNA and protein are expressed throughout the neuroepithelium of the forming neural tube (NT) prior to (E7.5), during (E7.5-8.5) and after (E9.5) NT closure 16-18. Ltap/Vangl2 protein is also developmentally expressed in additional tissues 19, and indeed, Lp/Lp embryos display defects in the inner ear (organization of hair cells of the cochlea), the heart (outflow tract defects), and the musculoskeletal system (fused ribs). This suggests that Ltap/Vangl2 is not only required for successful NT formation but is instrumental for multiple additional aspects of embryonic development 18,20,21.

1.4.1.1 Lp-associated independent alleles

Three independent Lp alleles have been described: one spontaneously arising (LpJAX), and two induced by chemical mutagenesis (Lpm1Jus, Lpm2Jus). Although homozygosity for each of the 3 alleles is embryonic lethal and associated with craniorachischisis, the Lpm2Jus allele is either less severe or has reduced penetrance, and displays a recessive mode of inheritance as only a small fraction of Lpm2Jus/+ mice display a looped tail. These mutations: S464N (LpJAX), D255E (Lpm1Jus), and R259L(Lpm2Jus) are specific for Lp mice and affect amino acid residues that are highly conserved throughout evolution. In various embryonic tissues positive for Vangl2 mRNA expression, the wild type protein is expressed at the plasma membrane 19, however, the Lp- associated Vangl2 variant S464N was expressed at low level and was undetectable at the plasma membrane of corresponding Lp/Lp embryos 19. Systematic biochemical characterization of the three Lp-associated protein variants in transfected cells has identified severe loss-of-function exhibited as loss of plasma membrane and apico-lateral targeting in polarized MDCK epithelial kidney cells, trapping in the endoplasmic reticulum, reduced half-life, and increased proteasome- dependent degradation, all possibly secondary to a basic misfolding defect 22-24 5.

It is anticipated that such behavior would disrupt the formation of membrane-bound PCP signaling complexes and, hence, ablate the activity of Vangl2 in PCP signaling, CE movement and tissue patterning. Finally, the pathogenic nature of the Vangl2D255E and Vangl2S464N has been investigated by measuring their effect on CE in zebrafish 25. Morpholino-mediated silencing of the trilobite (Vangl2) gene in zebrafish causes typical CE defects: shortened body axis, reduced curvature of the embryo, and widened somites. Co-injection of human VANGL1 RNA partially rescues the trilobite morpholino-induced defect. VANGL1 constructs bearing the Lp-associated variants D255E and S464N failed to rescue the CE phenotype, suggesting that these alleles represent loss-of-function mutations 25.

1.4.2 Additional phenotypes

The careful morphological and cytological analyses of anomalies detected in Lp mice have shown that, in addition to NT closure and cardiogenesis, Vangl2 is required for patterning in many additional tissues and organs. Vangl2 (and Celsr1) is expressed in restricted spatial domains of epithelial cells in developing lung. Lp-/- mice show defects in lung branching morphogenesis manifested as lung hypoplasia, reduction in number and width of airways, thickening of the interstitial mesenchyme, defective sacular formation, and impaired branching in response to FGF10 26. VANGL2 protein is expressed in developing skin (embryonic epidermis), and homozygous Lp embryos have hair follicles that are misangled, with altered anterior-posterior polarization of interfollicular epithelium 27. VANGL2 is expressed at the lateral edges of the apical pole of uterine epithelial cells, and Lp/Lp homozygotes display loss of columnar appearance of uterine epithelium. As a result, the Lp/Lp as well as Lp/+ heterozygotes both show gross anatomical defects of reproduction system and the Lp heterozygotes are often sterile 28. VANGLl2 is expressed in the developing gut, including epithelial cells of the developing stomach and duodenum 19, and Lp/Lp E13.5 embryos display a smaller fore-stomach with a loss of apico-basal polarity and oriented cell division 29. Early studies detected Vangl mRNA expression in the developing limbs in the E10.5-13.5 embryos 18. Recent studies identified altered shapes and dimensions of early limb buds in Lp/Lp embryos, lack of chondrogenic condensation due to increased cell death in distal limbs, and digits and limb defects including loss of phalanges and nails 30. In the developing eye at E16.5, VANGL2 protein is expressed in the cornea and in

27 migrating peridermal cells at the time of eyelid closure, immediately prior to keratinization of the fused eyelids 19; Lp/Lp mice show a defect in eyelid closure 14,18.

VANGL2 is also required for asymmetric cell division and specification of different cell types (reviewed in 31). For example, examination of Lp/Lp brain shows a decrease in the size of the neocortex that is caused by the premature differentiation of the neuronal progenitors at the expense of later born neurons 32. These and additional studies in vitro demonstrated a role for VANGL2 in suppressing progenitor differentiation and promoting cell fate diversity. VANGL2 appeared to control spindle orientation in neuronal precursors, thereby promoting asymmetric cell division during neuronal fate specification 32. Furthermore, VANGL2 is also required for tangential migration of branchiomotor neurons 33-35, anterior-posterior patterning of the axons of the monoaminergic neurons in brainstem 36, and an establishment of the retinal axon trajectories 37. Elimination of Vangl2 in Lp mice or in other animal models causes alterations in these neuronal populations.

1.5 VANGL1 and VANGL2

1.5.1 Gene description

The human VANGL2 is composed of 8 exons that cover 28kb (positions 160,370b- 160,297Mb) on the 1p23.3 region of Chromosome 1. The VANGL2 mRNA (1.8 kb) is expressed at low level in most adult human tissues. Embryonic expression of the VANGL2 gene has not been studied in humans. Vertebrates have a second Vangl gene. The human VANGL1 gene (Vang-like gene 1) is also known as Van Gogh-like 1, Strabismus 2, Loop-tail protein 2 homolog. VANGL1 is composed of 8 exons that cover 64kb (positions 116,185Mb-116,242Mb) on the 1p13.1 region of Chromosome 1 (extensively reviewed in 5). The VANGL1 mRNA (1.8 kb) is expressed at low level in adult human tissues but is enriched in brain. Although, embryonic expression of VANGL1 mRNA in humans has not been studied either, much is known about mouse Vangl1 and Vangl2 mRNA and protein expression during development.

1.5.2 Protein description

Vangl2 gene encodes a 521-amino-acid-residue highly hydrophobic integral membrane protein composed of four transmembrane domains 16,22, two solvent accessible extra-cytoplasmic loops, and intracellular amino and carboxy termini 38. The large carboxy-terminal cytoplasmic domain includes a PDZ-binding motif (PBM) at its extremity.

Vangl1 gene encodes a protein sharing ~75% similarity with VANGL2 protein 18. In the developing NT, Vangl1 mRNA and protein are expressed in the most ventral part of NT (floorplate) and notochord in a pattern complementary to that of Vangl2 18. Embryos homozygote for a targeted deletion of Vangl1 (Vangl1gt/gt) 39 or doubly heterozygote for Vangl2 and Vangl1 mutations (Vangl2Lp/+:Vangl1gt/+) 18 display craniorachischisis at low and high frequency, respectively. These results point at a genetic interaction between Vangl1 and Vangl2 and indicate that both proteins play a critical role in NT formation.

1.6 Study of Vangl genes in different species

VANGL proteins are highly conserved in evolution with relatives found in flies (Vangl/Stbm), fish (trilobite/Vangl2), and frogs (Xstbm) 18 35,40,41. The study of these relatives in model organisms has shed considerable light into the role of Vangl genes during development, especially during normal neurulation and in the pathogenesis of human NTDs. Insight comes from parallel studies in several experimental models:

1) The study of Van Gogh and other PCP genes in the fly Drosophila melanogaster 2) Experiments in the frog Xenopus laevis (xstbm) and in the zebrafish Vangl2 (trilobite) mutants 3) Studies of Vangl2 (loop-tail) and Vangl1 mutant mice 4) In vitro characterization of VANGL1 and VANGL2 mutant variants associated with NTDs in mice and humans.

1.6.1 Insight from flies

In flies, mutations in the Vang (Van Gogh) /Stbm (strabismus) gene were initially identified as impairing the planar cell polarization (PCP) of several epithelial structures 42. PCP refers to the uniform orientation of a group of cells within the plane of the epithelium. In flies, this axis of polarity is visible in the orientation of different structures including the eye unit (ommatidium) and in the uniform orientation of certain appendages such as the hair (trichome) on wing cells and of sensory bristles on the thorax and abdomen. To exemplify, in the wing, each hexagonal wing cell produces a single hair protruding from the most distal vertex of the cell; all hairs are uniformly oriented in a distal direction. Vang mutant wing cells display a single hair per cell, however, the hair is found in the center of the cell (as opposed to the most distal point) and its orientation is randomized. (Figure 1.6) Genetic screens have identified several so-called “core” PCP genes in which mutations cause defects in polarity in these structures: frizzled (fz), disheveled (dsh), prickle (pk), Van Gogh (Vang), flamingo (fmi), and diego (dgo) 42; extensively reviewed by 43,44.

1.6.1.1 Establishment of Planar Cell Polarity (PCP)

Before PCP becomes visible, PCP components are localized in the sub apical regions. The mechanism by which these proteins establish PCP in flies is poorly understood, but one of the first events involves their asymmetric distribution and selective enrichment to the membrane.

A B

1. 6: PCP in the Drosophila wing. (A) Images of hair bristles in WT and mutant PCP wing. (B) Asymmetric distribution at the membrane of PCP proteins. Adapted from Vladar et al. 2009.45

Fz, a 7 transmembrane domain protein and Vang a 4 transmembrane domain protein recruit cytoplasmic proteins Dsh and Pk respectively, which culminates in the appearance of two polarized membrane-bound complexes: Fz/Dsh at the distal side and Vang/Pk at the proximal side of wing epithelial cells. In flies, Vang can bind Dsh, Pk and the cytoplasmic ankyrin-repeat protein Dgo, while Dsh is a modular protein consisting of DIX, PDZ and DEP domains that can bind Pk, Vang, and Dgo. This asymmetry in distribution of the core PCP proteins establishes critical positional information. Fmi, a 7-TM membrane protein with a long extracellular domain encoding nine cadherin domains (they enable homotypic interactions), becomes enriched at both the proximal and distal side, interacting with both Dsh/Fz and Vang/Pk complexes, thus stabilizing and propagating the PCP signal through a “domino” effect 43,44. This asymmetry is additionally stabilized by a couple of antagonistic reactions, namely Pk which inhibits Dsh from accumulating at the proximal side and Dgo which in turn inhibits Pk from accumulating at the distal side. 46. Genetic screens in flies have also identified cadherin proteins fat (Ft), dachsous (Ds) and the Golgi kinase four-jointed (Fj) as additional general PCP genes 47-49 that act to stabilize and propagate the PCP signal. These proteins are expressed under a gradient and are part of a so called ‘global’ module which generates a global polarity that assures tissues develop harmoniously within the body axis. There has been some controversy as to the way the ‘global module’ directs information to the ‘core module’ in order to provide information for specific polarization events. 50. While some studies in the Drosophila eye have suggested the ‘global module’ acts directionally on the ‘core module’, more recent genetic mosaic studies in the Drosophila abdomen suggest both modules acts in parallel to provide critical information to effector proteins. (Figure 1.7)

Global Module

Core Module

1. 7: Proposed model of PCP in Drosophila. Proposed model of PCP in Drosophila. ‘Global’ module provides either directional or parallel information to ‘core’ PCP proteins. The ‘core’ module provides the output of polarized events by asymmetrically distributing ‘core’ protein complexes to the plasma membrane. Adapted from Vladar et al 2009. 45

Proteins such as fuzzy (Fy), fritz (Frtz) and inturned (In) have also been identified as wing- specific PCP “effector” genes in flies 42,51. Downstream effectors mediate tissue specific polarity events by activating cytoskeletal pathways and cellular rearrangements. These pathways include small GTPase of the Rho subfamily (Rho, Rac and cdc42), the Daam1 formin homology protein, c-Jun kinase and protein kinase C (PKC) 8,10,22.

In conclusion, detailed analysis of Drosophila Vang and other PCP mutants has revealed a few central features of the PCP pathway: a) at the cellular level, the PCP signaling controls cytoskeleton rearrangement; b) PCP signaling results in asymmetric localization of PCP protein complexes to distinct membrane domains on either side of the cell in the planar plane; c) asymmetric Vang localization depends on the normal function of other PCP proteins; d) asymmetric distribution of PCP proteins is initially weak, but is amplified via intercellular protein interactions.

Many of the molecular evens underlying polarization in the fly wing are conserved in vertebrate tissues. In effect, these studies of Vang protein in fly wings have provided the foundation to understanding the function of Vang in vertebrate relatives, especially in its regulation of cell shape and cell movements during the formation of the neural tube. 46.

1.6.2 Insight from vertebrates

In vertebrates, PCP proteins also participate in morphogenic activity designated convergent extension (CE) movements. CE plays a critical role in several aspects of embryogenesis and this process has been particularly well studied in Xenopus laevis. CE is the process in which a tissue or epithelium narrows in one axis and lengthens in a perpendicular axis, allowing tissue remodeling 52-54. This process is responsible for reshaping a short and wide tissue or epithelium into a thin and long one, which is why CE plays a critical role during gastrulation, neurulation and organogenesis. During CE, cells form mediolateral protrusions that generate traction used for mediolateral intercalation. CE movements are responsible for the narrowing and lengthening of the neural plate during NT closure.

1.6.2.1 Frog (Xenopus laevis)

It was originally discovered that silencing of the PCP gene Dishevelled in X. laevis dorsal mesenchyme and neuroepithelium causes defects in gastrulation and neurulation 54. In these pioneering studies, the authors established that loss of PCP function in vertebrate cells causes a failure to stabilize actin-based cell protrusions in the mediolateral direction. The loss of polarized protrusive activity in mesenchymal and neuroepithelial cells affected CE movements necessary for gastrulation and neurulation. Importantly, a specific knockdown of Xstbm (frog homologue of mammalian Vangl gene) in Xenopus similarly inhibited convergent extension, disturbing normal gastrulation and actin-induced elongation of Keller explants 55. These early studies provided the first indication that CE was controlled by the PCP signaling pathway and established a direct role for Vangl proteins in regulation of actin cytoskeleton and CE in vertebrates.

1.6.2.2 Zebrafish (Danio rerio)

The zebrafish mutant trilobite (tri, relative of mammalian Vangl2) exhibits defective gastrulation movements, manifested as a shorter embryonic axis, reduced dorsal flexure of axial curvature, and abnormal posterior migration of hindbrain neurons 35. Jessen et al demonstrated that during gastrulation, trilobite regulates mediolateral cell polarity enabling cell intercalation and polarized dorsal migration. Ciruna et al also showed that trilobite mutants with maternally depleted Vangl2 transcript (maternal-zygotic tri zebrafish) display a more severe neurulation defect associated with abnormal intercalation of recently divided neuroepithelial cells into neural tube (neurocoel in zebrafish) and an increase in cell proliferation. Surprisingly, blocking cell division in tri mutants rescued neural tube morphogenesis 56. These observations revealed a potential role for Vangl2 and the PCP pathway in control of neural cell proliferation that affects neural tube closure.

1.6.3 Insight from mammals

The PCP pathway is conserved in mammals and has been the focus of intense investigation in the past few years. These studies have shown that PCP genes play a critical role in the

35 development and maintenance of broad range of tissues and organs. In mammals, there exist 2 Vangl genes, 10 Frizzled (Fz), 4 Prickle (Pk), 3 Dishvelled (Dvl), 3 Flamingo (Celsr), and 1 Diego (Inversin) 43,44, 4 Fat 57 ;, Inturned and fuzzy relatives have also been described and studied in vertebrates 58-60. Additionally, a number of PCP genes, whose mutations cause defects in planar polarity only in vertebrates, has been identified including Ptk7 61,62, Scribble 63and Sec24b 64,65. The presence of multiple paralogous PCP genes that are expressed at a variable degree of tissue and cell specificity and often display functional redundancy have complicated the functional characterization of individual PCP genes. In addition, embryonic lethality caused by inactivation of some PCP genes (Vangl2, Celsr1) has until recently precluded a study of their role in the structure, function and maintenance of adult tissues and possible association with pathologies in the adults.

1.6.3.1 VANGL1 and VANGL2 mouse mutants

Identification of Vangl2 as the gene responsible for the craniorachischisis phenotype in the Lp mouse was the first evidence linking the PCP pathway to neural tube defects. The NTD phenotype in Lp mice was originally described as “an excessive breadth to the floorplate and notochord”. Indeed, analysis of the transverse sections through neurulating Lp/Lp embryos revealed a substantial lateral expansion of the floor plate (the most ventral part of the neural tube), and a wider irregular diameter to the underlying notochord; no change in local proliferation or apoptosis was associated with the expansion of these structures 17. By using a traceable stain and a fluorescent GFP reporter injected into the embryonic node or midline of Lp embryos, Ybot- Gonzales et al directly showed that Lp mutants exhibit defective convergent extension (CE) of dorsal mesoderm and neuroectoderm prior to neurulation 66. Defective CE in the neural plate results in expansion of floor plate width. The neural tube folds of Lp embryos form properly, however, they are unable to bridge the wider floor plate distance between them and, thus, are unable to fuse at closure site I. As a result, neural tube in Lp mice remains open along its entire length from midbrain-hindbrain junction to the most caudal part of the embryo resulting in craniorachischisis. Lp/Lp embryos also show pronounced defects in cardiogenesis (cardiac outflow track defects), as well as classical disruption of PCP manifested by misorientation of stereocilliary bundles of the sensory hair cells of the organ of Corti. These studies established that the Vangl2

36 gene plays an important role in PCP signaling, CE movements, and is required for neural tube closure in mammals.

The role of the second mammalian Vangl gene, Vangl1, in embryonic development and neural tube formation was also investigated in mice. In the developing neural tube, Vangl1 mRNA and protein are expressed in the most ventral part of the neural tube (floorplate) and in the notochord in a pattern complementary to that of Vangl2, as well as in the most dorsal cells of the neural tube (roofplate) at E8.5-9.5 where Vangl1 expression overlaps with that of Vangl2 18. Embryos homozygous for a targeted deletion of Vangl1 (Vangl1gt/gt) display craniorachischisis at low frequency; double Vangl2Lp/+:Vangl1gt/+ heterozygotes 18or double Vangl2Lp/Lp:Vangl1gt/gt homozygotes display craniorachischisis at high frequency 67 indicating that the two genes interact genetically and that both are critical for neural tube development. Vangl1 and Vangl2 display similar biochemical properties, and recent studies have also identified a physical interaction between Vangl1 and Vangl2 proteins in the mouse 68. It was inferred that craniorachischisis in Vangl1gt/gt is a consequence of defective CE. (Table 1)

1. Table 1: Phenotype of VANGL2 and VANGL1 mutant mice. Adapted from Torban et al. 2012.

1.6.3.2 Additional PCP mouse mutants

At the cellular level, VANGL and other mammalian PCP proteins developmentally regulate the structure, positioning and polarity of cellular appendages in different cell types. They are also involved in the asymmetric morphology, polarization, oriented cell division, migration and re-distribution of many cell populations. Consequently, mutations in PCP gene have been shown to affect biogenesis, structure and function of a surprisingly large number of tissues and organs 31,43,44,69. For example, the uniform orientation of stereo-cilliary bundles on outer and inner hair cells of the cochlear duct (organ of Corti) is a classical manifestation of PCP in mammals.

Mutations in Vangl1 18 and Vangl2 20, but also in other PCP genes such as Dvl1/Dvl2 70, Dvl3 71, Fz3/Fz6 72, Fat4 57and Celsr1 73 disrupt PCP of the stereo-ciliary bundles to a varying degree. Homozygosity for mutations in Vangl1 39, Vangl2 16, Dvl1/Dvl272, Dvl2/Dvl3 71, Celsr1 73, Fz3/Fz6 72, Scribble 63, Ptk7 61and Sec24b 64 cause NTDs of different severity, including the characteristic very severe craniorachischisis. Likewise, severe NTDs (and the PCP defects) are often detected in mutant mice doubly heterozygous for mutations in Vangl2 (Vangl2Lp/+) and other PCP genes from the above-mentioned list, highlighting the key role of Vangl2 in this process. CE movements are also essential for cardiogenesis; mutations in Vangl2 and Fz1/Fz2 cause cardiac defects in the form of aberrant right subclavian artery+ and double outlet right ventricle 18,74.

1.6.4 Insight from VANGL1 and VANGL2 human mutations

Following the discovery that Vangl1 and Vangl2 mutations are associated with the severe NTD craniorachischisis in mice, the VANGL1 and VANGL2 genes were investigated for sequence alterations in familial and sporadic cases of neural tube defects in humans (reviewed in 5). The first study investigated 144 patients with neural tube defects including 137 Italian patients with nonsyndromic spinal dysraphisms, and 7 fetuses with craniorachischisis. All mothers from Italian patients lacked periconceptional folate dietary supplementation in this study, and 13 of these patients had a positive family history of clinically documented NTDs, whereas all other cases were sporadic 120. Control group of this study consisted of 106 ethnically and geographically matched individuals. A de novo mutation (V239I) was identified in the VANGL1 gene of a 10-year old girl

39 who had a severe form of caudal regression with lipomyeloschisis, anorectal malformation, hydromelia and tethered spinal cord. Her brother also carried the V239I variant and had a milder form of the disease, dermal sinus, while the mother (who also carried the mutation) was normal. A second missense mutation (R274Q) was identified in a 19-year old girl who had meninogomyelocele, hydrocephalus, and congenital club foot. Her mother and maternal aunt had vertebral schisis. R274Q was present in the daughter and mother, while the aunt could not be analyzed. Finally, a M328T variant was identified in a 21-year old woman with myelomeningocele, hydrocephalus, tethered spinal cord, club foot, lumbosacral scoliosis, and sacrococcygeal kyphosis. The three variants (V239I, R274Q, M328T) were heterozygous, patient- specific and not detected in controls. The three affected amino acids map to the carboxy-terminal cytoplasmic domain of VANGL1 protein; all 3 mutations show a very high degree of sequence conservation across different species. Functionally, the V239I variant was shown to impair interaction of VANGL1 with PCP proteins DVL1, DVL2, and DVL3 (yeast two-hybrid system), in a manner similar to that observed for the Lp-associated Vangl2 variants D255E and S464N that were used as positive controls 120. Finally, complementation studies in zebrafish embryos, aimed at testing the ability of wild type and mutant VANGL1 variants to suppress the convergent extension defect of the trilobite morpholino model, demonstrated that V239I and M328T behave as loss-of-function mutations 25. Together, these studies led the authors to conclude that VANGL1 is a risk factor in humans for neural tube defects.

The integrity of the VANGL2 gene was also investigated in 163 stillborn or miscarried fetuses displaying various types of severe NTDs including craniorachischisis, cranial NTDs (anencephaly, encephalocele, holoprosencephaly, iniencephaly) as well as spinal NTDs (nonsyndromic spina bifida and spina bifida with hydrocephalus) 75. In these studies a group of 508 unrelated healthy ethnically-matched infants were used as controls. A total of three patient-specific variants were identified. VANGL2 variant R353C was identified in a male fetus (21 weeks gestation) with anencephaly with occipital and cervical spina bifida. VANGL2 variant F437S was identified in a male fetus (24 weeks gestation) with anencephaly while another variant, S84F, was detected in a female fetus with holoprosencephaly (22 weeks gestation). All VANGL2 variants were heterozygotes, and affected residues that showed high degree of cross-species conservation. Functionally, variants R353C and F437S impaired interaction of VANGL2 with DVL1, DVL2, and

DVL3 in a yeast two-hybrid system, in a manner similar to that observed for the Lp-associated Vangl2 variants D255E and S464N that were used as positive controls 75. Furthermore, the S83L variant maps to a serine/threonine cluster in the amino terminus of the protein which is a subject to regulatory phosphorylation by WNT proteins 5. These two studies elegantly demonstrated that heterozygosity for loss-of-function alleles at VANGL1 and VANGL2 are associated with different types of neural tube defects in humans.

Additional sequencing of larger cohorts led to the identification of other VANGL1 and VANGL2 variants associated with NTDs. However, the functionality of these variants has yet to be fully evaluated. Screening of a large cohort of 673 non-syndromic sporadic cases of NTDs (cranial, open and closed spinal dysraphisms) in patients of various ethnic origins identified additional independent patient-specific VANGL1 (S83L, F153S, R181Q, L202F, and A404S) 76 and VANGL2 (R135W, R177H, L242V, T247M, R270H, and R482H) mutations 77. In these studies, putative pathogenic mutations were identified on the basis of (a) absence of the variant in ethnically matched controls and in a ~1000 additional normal controls, (b) evolutionary conservation of the residue affected, and (c) non-conservative nature of the variant. Of notable interest are Arginine at positions R181 and R274 (VANGL1 numbering) which were found independently mutated to Histidine (VANGL2R177H) or Glutamine (VANGL1R181Q) for R181 and mutated to H (VANGL2R270H) or Q (VANGL1R274Q) for R274 in unrelated patients, suggesting a critical role for these two invariant Arginine residues in the function of Vangl proteins. Both the R177H and the R181Q were recently shown to behave as loss-of-function in transfected MDCK cells, including impaired membrane expression, reduced stability and increased proteasome dependent degradation 157. Recently, screening 144 patients with NTDs of Eastern European descent identified three new mutations in VANGL1, namely R173H, R186H and G205R (also present in healthy younger sister) 78. As of today, a total of 20 naturally occurring VANGL protein variants associated with clinical cases of NTDs in humans have been discovered. (Figure 1.8 and Table2)

VANGL1 (524 aa) VANGL2 (521 aa)

R135W Out

In V239I R173H M328T R186H L242V R353C S83L A404S S84F R181Q* R177H* T247M F437S D255E S464N R259L R482H R274Q* R270H*

1. 8: Mutational map of VANGL1 and VANGL2 in neural tube defects. Mutational map of VANGL1 and VANGL2 in neural tube defects. Schematic representation of human VANGL1 and VANGL2 proteins, including membrane polarity (in, out), and major structural features (four transmembrane domains, positions of amino and carboxy termini and PDZ binding motif). Amino acid variants associated with neural tube defects in humans (red) and in loop-tail mice (green) are identified. Two homologous arginine residues conserved at the same position in VANGL1 and VANGL2 were found to be independently mutated in different patients and are identified by asterisks R181Q/R177H, and R274Q/R270H (VANGL1/VANGL2 numbering). Adapted from Torban et al 2014. 155

1. Table 2: Human mutations and clinical description.

Mutation Gene Clinical Data Funtionally tested

S83L VANGL1 76 Two female and one male of Hispanic origin with familial form N of closed NTD (tethered cord).

S84F VANGL2 75 Female foetus at 22 week’s gestation with holoprosencephaly. N

R135W VANGL2 77 Male with lumbo-sacral myelomenigocele (open NTD). N

F153S VANGL1 76 2 year old female of Tunisian origin with closed NTD (tethered N cord). Variant detected in father, but not in mother.

R173H VANGL1 78 Slovakian male (born in 1987) with sporadic NTD N

R177H VANGL2 77 Female with diastematomyelia (closed NTD). Maps at the same N location as mutation R181Q in VANGL1.

R181Q VANGL1 76 8 year old male of Italian origin with familial form of open NTD Y (myelomeningocele). Variant present in mother, absent in father and unaffected sister and distant maternal ancestor also afflicted with myelomeningocele. Maps at the same location as mutation R177H in hVANGL2.

R186H VANGL1 78 Slovakian female (born in 2003) with sporadic spinal lipoma and N tethered cord.

L202F VANGL1 76 Located in invariant ‘WLF’ motif and detected in 9 year old N female of Italian origin with open NTD (myelomeningocele).

G205R VANGL1 78 9 year old Slovakian male with sporadic lumbo-sacral N meningomyelocele. Variant present in healthy younger sister, but absent in mother.

V239I VANGL1 120 Located in invariant ‘VLLE’ motif and detected in 10 year old Y female of Italian origin with closed NTD. Familial case with variant found in both mother (no clinical signs) and in brother (milder form of disease), but absent in father, maternal aunt and maternal grandparents.

L242V VANGL2 77 Two unrelated Caucasian females of American and Italian origin N with myelomeningocele (open NTD) and with lumbar myelocystocele (closed NTD) respectively. In Italian patient, variant also found in mother, but absent in father and unaffected sister.

T247M VANGL2 77 Caucasian American male with tethered cord (closed NTD). N

1.7 Cellular and molecular pathogenesis of Vangl genes in vertebrate and mammals

At the molecular level, the PCP signaling pathway in vertebrates share structural and mechanistic similarities with its fly counterpart, including the formation of asymmetrically distributed membrane PCP complexes in polarized cells and tissues.

1.7.1 In vivo asymmetric distribution

Indeed, Vangl1 39,67, Vangl2 27,67,79-81 , Fz3 72,80,81, Fz6 27,72, Dvl2 41, Dvl3 71, Pk2 39,82, and Celsr1 27have all been described as asymmetrically distributed in vivo in polarized and/or ciliated cells of different tissues. These include structures such as the cochlea, epidermis, ependyma, neuroepithelium, and the posterior node in developing embryos.

1.7.2 Genetic interactions

Importantly, a unique and critical feature of Vangl2 is that its function represents a key rate-limiting and dosage-sensitive step in the PCP signaling. Heterozygosity for Vangl2 mutation (Lp/+) sensitizes the PCP pathway and causes either exacerbation of mild PCP phenotypes in animals homozygotes for mutations in other PCP genes or an appearance of full-fledged PCP defects (inner ear, NT, heart) in otherwise normal heterozygotes bearing one mutated allele of other PCP genes including Vangl118,39, Dvl1/2/3 71,72, Fz1/2 72,74, Celsr173, Scribble 63, Ptk7 61, and Sec24b 64 These observations establish the critical and dosage-sensitive functional role of Vangl2 in molecular processes underlying PCP signaling in many cell types.

1.7.3 Physical interactions

In addition, Vangl2 has been shown to physically interact in vitro (pull- down/immunoblotting, immunoprecipitation, yeast two-hybrid system) with other components of the PCP machinery including Dvl1/2/3 8,22, Celsr1 27, and Scribble 80,83, and Lp-associated inactivating Vangl2 mutations disrupt these interactions. Furthermore, mutations in mammalian

PCP proteins Scribble 80, Celsr1 27, and Celsr2 69 impair membrane targeting of Vangl2 in vivo and/or physical interaction with Vangl2 in vitro. These observations have led to a basic model for vertebrate PCP signaling 69 involving asymmetrically positioned Vangl/Pk vs. Fz/Dvl membrane complexes that provide polarity information to cells and cell layers. This polarity information would then be propagated to adjacent cells via homotypic interactions between Celsr proteins expressed by neighboring cells (different variations of this model may be active in different cell types). PCP signaling further requires an interaction between the PDZ-binding domain of Vangl2 and the PDZ-domain of Scribble 80.

1.7.3.1 ER-Golgi transport

SEC24B protein is needed for maturation of VANGL2 positive COPII vesicles (ER-Golgi transport vesicles) and controls membrane targeting of VANGL2 to membrane PCP complexes 64,65. Sec24b mouse mutants display craniorachischisis and inner ear defects, and Lp-associated VANGL2 mutant proteins cannot enter COPII vesicles for SEC24B- dependent maturation. In addition, there is an aberrant localization of VANGL2 in SEC24B mutant embryos in vivo.

1.7.3.2 Export out of trans-Golgi network (TGN)

Additionally, it was recently reported that transport of VANGL2 out of the TGN is dependent on two proteins, ARFRP1 (a GTP-binding protein) and AP-1 (a clathrin adaptor protein). Specifically, ARFRP1 exposes a binding side on Ap-1, which is needed to physically interact with YYXXF, a sorting motif in the C-terminal tail of VANGL2 and allow for its transport to the surface of polarized epithelial cells. 84

1.7.4 Phosphorylation

Early work by Kalabis, Rosenberg, and Podolsky (2006) 85 showed that VANGL1 is phosphorylated in response to intestinal trefoil factor (ITF)/TFF3, which reduces VANGL1 plasma membrane localization and may affect cell motility. More recently, Gao et al. 86 observed that VANGL2 interacts with Ror2 and is phosphorylated both in vivo and in vitro. In addition,

VANGL2 phosphorylation is stimulated by WNT ligands. A Ser/Thr cluster at VANGL2 amino terminus is targeted for phosphorylation, and the importance of this event is revealed by the presence of hVANGL1 (S83L) and hVANGL2 (S84F) variants at this site in human NTDs patients. Finally, coexpression of Lp-associated variant VANGL2S464N reduced phosphorylation of the WT protein expressed in the same cell. These results have provided a model in which VANGL2 phosphorylation and associated role in PCP are regulated by WNT ligands 86.

1.7.5 Mode of inheritance of Lp mutations

Mechanistically, Lp variants have been proposed to behave either as partially penetrant with negative codominance or as haploid insufficiency in a gene dosage-dependent pathway. In favor of the former are the observations that (a) several Vangl2-associated PCP phenotypes appear more severe for Lp alleles (Vangl2D255E, Vangl2S464N) than for null allele (Vangl2TM/TM) 68; (b) VANGL1 and VANGL2 appear to physically interact and, in cotransfection experiments, Lp alleles disrupt VANGL1/VANGL2 interactions, as well as trafficking and membrane targeting of VANGL1 and VANGL2 68, and decrease of the posttranslational modification of WT protein 67,86. In support of haploid insufficiency in a gene dosage-dependent pathway are the observations that (a) experimental overexpression or silencing of core PCP genes causes the same phenotype in different animal models tested; (b) all experimentally induced 67,86or naturally occurring Vangl2 mutations 16,17,87described so far show varying degrees of the same phenotype (looped tail, inner ear defects) in heterozygotes and homozygotes in vivo; (c) Lp-associated VANGL2 protein variants are expressed at lower levels in vivo 19,80,88 and display reduced stability and shorter half- life when tested in vitro 23,24,77; (d) colocalization studies by double immunofluorescence and confocal microscopy in transfected MDCK cells show that expression of VANGL2D255E has no effect on membrane targeting of WT VANGL2 23. Finally, genetic background has a strong influence on penetrance and expressivity of Lp-associated phenotypes 13, which complicates analysis of mode of inheritance of Vangl2 mutations and the associated interpretation of experimental results obtained in vivo mechanism for VANGL protein function.

1.8 Conclusion

The pivotal role of VANGL proteins in neural tube formation was initially established 10 years ago by the discovery that mutations in Vangl2 are responsible for the severe neural tube defect craniorachischisis in the well-known mouse mutant loop-tail 16. The careful examination of morphology and dynamic developmental processes altered in Lp mice have demonstrated that Vangl2 plays a similarly important role in patterning and development of a many tissues and organs. The genetic interaction studies in mice heterozygous for Vangl2Lp/+ and for mutations in other genes have linked Vangl2 and the uniquely Vangl2-sensitive planar cell polarity pathway to a number of additional and unsuspected phenomena such as asymmetric cell division, cell migration, and function of key cellular appendages. Additionally, studies have demonstrated that heterozygosity for loss-of-function alleles at VANGL1 and VANGL2 are associated with different types of neural tube defects in humans. Taken together, the observations above suggest that specific Vangl1 and Vangl2 missense mutations may affect protein stability, trafficking, subcellular localization of VANGL proteins, as well as interaction with other PCP proteins. This may disrupt formation and function of PCP signaling complexes in embryonic neuroepithelial cells causing defects in CE movements. In this thesis, structure-function studies were performed to ascertain the possible mechanistic basis of Vangl genes loss of functions in neurulation and NTDs. Part of this introduction, including several figures, were taken from two reviews which I co- authored. 5, 155

Chapter 2: Transmembrane topology of the mammalian planar cell polarity protein VANGL1

Preface to Chapter 2

The identification, in 2001, of Vangl2 as the gene mutated in the loop-tail (Lp) mouse suffering from craniorachischisis was the first study to link the planar cell polarity (PCP) pathway to neural tube closure in mammals. It was established from earlier work in the Drosophila model that VANGLl proteins along with several other ‘core PCP’ proteins transmit polarity cues by forming asymmetrically distributed membrane bound complexes. Structure-function studies are essential to better understand the molecular mechanisms of the VANGL family of proteins.

Chapter 2 describes a systematic effort to determine the secondary structure of VANGL proteins at the plasma membrane. In the absence of a high resolution 3D X-ray crystallography model, an epitope tag approach was used to characterize membrane topology including number, position and polarity of individual transmembrane domains of VANGL proteins. Hydropathy plots performed by several predictive programs (including TOPPRED and TMHMM) predict a 3 or 4 TM domain protein, with results identifying TM4 as uncertain. This makes the intra- vs. extracellular assignment of the amino and carboxy termini difficult to determine, and so several models have to be considered, including a) intracellular N- and C- terminus b) extracellular N-terminus and intracellular C-terminus c) extracellular N- and C-terminus. Our topological mapping proposed a 4TM domain protein with both the amino and carboxy termini located intracellularly.

2.1 Abstract

VANGL1 and VANGL2 are membrane proteins that play an important role in neurogenesis, and Vangl1/Vangl2 mutations cause neural tube defects in mice and humans. At the cellular level, VANGL proteins regulate the establishment of planar cell polarity (PCP), a process requiring membrane assembly of asymmetrically distributed multi-protein complexes that transmit polarity information to neighbouring cells. The membrane topology of VANGL proteins and the protein segments required for structural and functional aspects of multi-protein membrane PCP complexes is unknown. We have used epitope tagging and immunofluorescence to establish the secondary structure of VANGL proteins, including the number, position and polarity of transmembrane domains. Antigenic hemagglutinin A (HA) peptides (YYDVPDYS) were inserted in predicted intra- or extracellular loops of VANGL1 at positions 18, 64, 139, 178, 213 and 314, and individual mutant variants were stably expressed at the membrane of MDCK polarized cells. The membrane topology of the exofacial HA tag was determined by a combination of immunofluorescence in intact (extracellular epitopes) and in permeabilized cells (intracellular epitopes), and by surface labeling. Results indicate that VANGL proteins have a 4 transmembrane domain structure with the N-terminal portion (HA18, HA64), and the large C-terminal portion (HA314) of the protein being intracellular. Topology mapping and hydropathy profiling show that the loop delineated by TMD1-2 (HA139) and TMD3-4 (HA213) are extracellular, while the segment separating predicted TMD2-3 (HA178) is intracellular. This secondary structure reveals a compact membrane-associated portion with very short predicted extra- and intracellular loops, while the protein harbours a large intracellular domain.

2.2 Introduction

Neural tube defects (NTDs) are a group of common congenital malformations in humans that arise from a partial or complete failure of the neural tube to close during embryogenesis. Complex genetic and environmental factors are known to be important etiological factors in the emergence of NTDs in humans 89. The complex genetic component has been difficult to decipher in humans, but has been studied in naturally occurring and experimentally-induced mouse mutants 90,91. The loop-tail (Lp) mouse is an accepted model for the study of NTDs 1,13. Heterozygotes Lp mice are normal except for a characteristic “looped” tail, while homozygote Lp/Lp embryos die in utero of a severe form of NTD, craniorachischisis, in which the neural tube remains open from hindbrain to the most caudal extremity of the developing embryo. Three Lp alleles have been described so far (Lp, Lpm1jus and Lpm2jus), and we have shown that the NTD in these 3 allelic variants is caused by independent loss-of-function mutations (D255E, R259L and S464N) in the Vangl2 gene 16,92. In mammals, there are two Vangl genes, Vangl1 and Vangl2. These two genes encode proteins that are expressed during embryogenesis and show complementary patterns of expression in the developing neural tube. Their expression persist post-nataly in certain tissues, but Vangl2 is expressed abundantly and broadly in the dorsal and ventral portions of the neural tube, prior to, during and after closure (E7.5-E11.5), while Vangl1 expression is mostly restricted to the ventral portion of the neural tube and to the notochord 19,93. Vangl1 and Vangl2 genetically interact during neural tube formation and animals doubly heterozygote for Vangl1 and Vangl2 mutations (Vangl1+/-:Vangl2+/-) show craniorachischisis 18. Importantly, we 8,76,77and others 75have subsequently identified loss-of-function mutations in the human VANGL1 and human VANGL2 genes in familial and sporadic cases of neural tube defects in humans.

Mammalian Vangl genes are relatives of the Drosophila Van Gogh/Strabismus (Vang/Stbm) gene, a member of the so-called planar cell polarity (PCP) gene family. In flies, this family also includes Frizzled (Fz), Diego (Dgo), Flamingo (Fmi), Prickle (Pk) and Dishevelled (Dvl). Mutations in these genes alter establishment of planar cell polarity which causes a number of abnormalities in the fly, including the orientation of the eye (omatidiae), and hair bristles on the wings and legs 94,95. In the fly, the redistribution and asymmetric positioning of membrane bound Dvl/Fz and Vang/Pk complexes in epithelial cells is believed to propagate a polarity signal

51 essential for PCP organization of cell layers 80,96,97. The PCP pathway has been conserved in mammals, and a clear example is the precise orientation of the stereocilliary bundles of the neurosensory epithelium of the inner ear. In the mouse, inactivating mutations in PCP genes Vangl1 or Vangl2, Dsh (Dvl), and Fmi (Celsr1) cause disruption of these structures in the inner ear 98. In addition, vertebrate relatives of fly PCP genes regulate convergent extension movements during embryogenesis, including the narrowing and lengthening of the neural plate required for neural tube closure, and mutations in Vangl2 (Lp) 16, Vangl1/Vangl2 18, Dvl1/2 70, Dvl2/3 71, Celsr1 73, and Fz3/672,99cause craniorachischisis.

Mammalian VANGL proteins are composed of 521-526 amino acids. Subcellular localization studies by immunofluorescence and cell fractionation studies have shown that VANGL proteins are integral membrane proteins, both in primary tissues in vivo and in transfected cultured cells in vitro 23,64,80. Plasma membrane (PM) targeting of VANGL proteins is essential for function and we have shown that Lp-associated pathological mutations in Vangl2 impair PM localization of the protein 23. VANGL membrane targeting is also required for interaction with other PCP proteins, including Dvl 22. The secondary structure of VANGL proteins including its membrane topology has not been established and this information is required to understand how VANGL proteins interact with other PCP proteins in order to assemble membrane bound polarity signaling complexes. Multiple sequence alignments and hydropathy profiling have predicted at least three and possibly four TM domains in the amino terminal half of the protein, with the carboxy terminal portion coding for a large hydrophilic domain. This highly conserved carboxy half of the protein is responsible for physical interaction with cytoplasmic Dvl proteins (Dvl1, Dvl2, Dvl3) 22, possibly though a PDZ binding motif (ETSV tetrapeptide) located at its carboxy terminal extremity. This strongly suggests that the carboxyl half of Vangl proteins is intracellular.

In the present study, we have used epitope tagging to establish the secondary structure of VANGL proteins, including the number, position and polarity of transmembrane domains. The membrane topology of the hemagglutinin HA tags inserted in different locations of the protein was determined by a combination of immunofluorescence in intact (extracellular epitopes) and in permeabilized cells (intracellular epitopes), and was verified by surface labeling. Results from

52 these studies were unambiguous and established a four TM domains topology for VANGL proteins with the amino and carboxy termini located on the intracellular side of the membrane.

2.3 Experimental Procedure

Material and Antibodies Geneticin (G418) and penicillin/streptomycin were purchased from Invitrogen (Carlsbad, CA). Restriction endonucleases were from New England Biolabs (Ipswich, MA), and Taq DNA polymerase was from Invitrogen. The mouse monoclonal antibodies directed against the influenza hemagglutinin epitope (HA.11) was purchased from Covance (Berkeley, CA). The mouse monoclonal antibody recognizing Na, K-ATPase (alpha) was from Santa Cruz Biotechnology (Santa Cruz, CA). Cy3-conjugated goat anti-mouse antibody and peroxidase-coupled goat anti- mouse antiserum were from Jackson ImmunoResearch Laboratories (West Grove, PA).

Plasmids and Site-Directed Mutagenesis The human VANGL1 cDNA was used as a template for the construction of all mutants. The cDNA was amplified from total human RNA by RT-PCR and cloned into the pCS2+ plasmid vector. PCR-mediated mutagenesis with overlapping oligonucleotides was used to insert a short antigenic peptide epitope (EQKLISEEDL) from the human c-Myc protein at the N-terminus of VANGL1 as we have previously described, followed by cloning into the pCB6 mammalian expression vector. For the transmembrane domain topology mapping analysis, hemagglutinin (HA) epitopes (YPYDVPDYA) were inserted in pre-determined positions in the protein by recombinant PCR, and using mutagenic oligonucleotide primers listed in Table 1. To enable potential insertion of additional contiguous HA epitope(s) at these positions, a Nhe I restriction site was engineered at the 3’-end of each HA sequence, accounting for the additional Ser residue found at the end of the HA tag. To facilitate identification of transfected cells expressing various recombinant hVANGL1 proteins, c-Myc/HA-tagged hVANGL1 cDNAs were fused in-frame to the green fluorescent protein (GFP) by cloning into peGFP-C1 vector (Clontech, Mountain View, CA). The integrity of all VANGL1 cDNA constructs was verified by nucleotide sequencing.

2. Table 1: Oligonucleotides Used for Epitope Insertion by Site-Directed Mutagenesis.

Cell Culture, Transfection and Western Blotting Madin-Darby canine kidney (MDCK) epithelial cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 µg/ml streptomycin (37°C, 5% CO2). To generate stable transfectants, MDCK cells were transfected with c-Myc/HA-tagged VANGL1 constructs subcloned in pEGFP-C1 using Lipofectamine Plus Reagent as per the manufacturer’s instructions (Invitrogen). Selection for stably transfected clones was carried out in growth medium containing 0.3 mg/ml G418 for 10-14 days. Expression of recombinant c-Myc/HA-tagged GFP-VANGL1 was initially identified by fluorescence microscopy on live cells, and confirmed by Western blotting analysis. Total cell lysates were prepared in RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1% SDS supplemented with protease inhibitors), cleared by passage through a 25G needle and by centrifugation at 13000 g for 10 minutes (4oC). Whole cell extracts (50 µg) were separated by electrophoresis using 7.5% SDS-polyacrylamide gels, followed by electroblotting and incubation with the monoclonal anti-HA antibody (HA.11) (used at 1:1000). Immune complexes were revealed with a horseradish peroxidase conjugated goat anti-mouse antibody (1:10000) and visualized by enhanced chemiluminescence (SuperSignal West Pico kit, Thermo Scientific Rockford, IL).

Immunofluorescence and Confocal Microscopy Stably transfected MDCK cell clones expressing individual HA-tagged VANGL1 proteins were seeded at high density onto glass coverslips in a 24-well plate. Forty-eight hours later, cells were washed twice with ice-cold PBS, fixed for 15min in 4% paraformaldehyde (PFA) in PBS and then permeabilized with 0.1% Triton X-100 in PBS for 15 min. After blocking for non-specific binding with 5% goat serum and 0.1% bovine serum albumin (BSA) in PBS for 1 h, the cells were incubated with the desired primary antibody [mouse anti-Na, K-ATPase (1:50); mouse anti- HA (1:200)] for 1 h, and washed three times with 0.1% BSA in PBS. Finally, the coverslips were incubated for 1 h with the goat anti-mouse-Cy3 secondary antibody (1:1000) and washed three times with 0.1% BSA in PBS. All the incubations were done at room temperature, and the antibodies were all diluted in blocking solution. To detect cell surface expression of the HA epitope under nonpermeabilized conditions, MDCK cells expressing HA-tagged VANGL1 variants were incubated with the mouse anti-HA (1:200) antibody in DMEM containing 2% nonfat

56 milk for 2 hours at 37ºC. After several washes with PBS, cells were fixed for 15 min with 4% PFA in PBS, and incubated with goat anti-mouse-Cy3 antibody (1:1000) for 1 h. For immunofluorescence, coverslips were rinsed twice with PBS and once in water and mounted by using Permafluor Aqueous Mounting Medium (Thermo Scientific, Fremont, CA). Confocal microscopy was carried out using a Zeiss LSM5 Pascal laser scanning confocal microscope. All image analyses were performed using the LSM5 Image software. To maximize image quality, a Median filter 3x3 was applied to the images using Image-Pro software.

Estimation of Cell Surface Expression by Enzyme-linked Immunosorbent Assay The cell surface expression of HA-tagged VANGL1 proteins was estimated using an enzyme-linked immunosorbent assay (ELISA), as described previously 23with the following modifications. Briefly, MDCK cells stably expressing HA-tagged VANGL1 variants were seeded at confluence in 24-well plates and grown for 4 days. Cells were then washed with PBS and incubated in Ca++-free DMEM for 20 min, prior to incubation with 10 mM EGTA for 5 min. To evaluate the expression of HA-tagged VANGL1 proteins at the cell surface, cells were incubated ++ with mouse anti-HA antibody (1:200 in Ca free DMEM) for 2 h (37°C, 5% CO2). Cells were washed, fixed with 4% PFA in PBS for 15 min, and incubated with HRP conjugated goat anti- mouse antibody (1:4000 in 5% nonfat milk/PBS for 1 h). To evaluate total HA-tagged VANGL1 variants, cells were fixed after the EGTA treatment, permeabilized with 0.1% Triton X-100 in PBS for 30 minutes and blocked in 5% nonfat milk/PBS. Cells were then incubated with anti-HA antibody (1:200 in blocking solution) for 1 h, washed, and incubated with HRP conjugated goat anti-mouse antibody (1:4000 in blocking solution). Peroxidase activity was quantified colorimetrically using the HRP substrate [0.4 mg/ml O-phenylenediamine dihydrochloride (OPD), Sigma-Aldrich] according to the manufacturer’s instructions. The reaction was stopped by adding 3N HCl, and absorbance readings (490 nm) were taken in an ELISA plate reader, and background absorbance reading from non-specific binding of primary antibody to vector-transfected cells was subtracted for each sample. Cell surface readings were normalized to total HA-tagged GFP- VANGL1 value for each cell clone and were expressed as a percentage.

2.4 Results

Epitope-Tagged mutant VANGL1 constructs Hydropathy profiling and analysis of hydrophobic segments of the predicted amino acid sequence of the human VANGL1 protein using the TOPPRED software package is shown in Figure 1. According to this analysis, hVANGL1 shows four predicted transmembrane domains in the amino terminal half of the protein. Using a predicted length of 20 amino acids, and minimizing the number of charges to be included in the transmembrane portion, the 4 TM domains are predicted to span segments 114–134 (TM1), 152–172 (TM2), 186–206 (TM3), and 222–242 (TM4). In addition, we also used the TMHMM predictive program and both the probability plot and comparison with TOPPRED are shown in Supplemental Figure S1. The amino acid sequence of all 4 predicted TMs is highly conserved amongst Vangl relatives from mouse, human, zebrafish, worm and flies with 8/20 (TM1), 5/20 (TM2), 14/20 (TM3) and 17/20 (TM4) highly conserved positions (defined as >7 of 9 sequences identical at that position), suggesting that these domains play an important conserved role in the function of these Vangl relatives. However, TM4 consistently gives lower scores in predictive programs than the other 3 TM domains. In addition, there is a predicted N-linked glycosylation site (N-X-S/T) in the amino terminus of the protein that is well conserved amongst Vangl relatives, thereby arguing against a 4 TM model with the amino and carboxy termini of the protein being intra-cytoplasmic.

To determine the membrane topology of VANGL1, including the number and position of individual TM domains, as well as the cytoplasmic vs. extracytoplasmic location of the amino and carboxy termini, we used an epitope tagging method 100. In this method, antigenic hemagglutinin HA epitope tags (YPYDVPDYAS) are inserted in strategic positions of the protein in individual mutants, followed by expression of the HA tagged protein at the plasma membrane of transfected cells and determination of the polarity of the exofacial tag by immunofluorescence in intact (extracellular) and permeabilized cells (intracellular). This method provides topological data on proteins that are properly expressed at the plasma membrane in a functional state. HA epitopes were inserted at two positions of the amino terminus (positions 18, 64; NT constructs), and at position 314 in the carboxy terminal portion (CT construct). Three additional HA tags were inserted at positions 139, 178, and 213. These position correspond to peak hydrophilic segments

(from hydropathy profiling; Figure 1A) in the otherwise hydrophobic moiety of the protein, and therefore possibly corresponding to short intra- or extra-cytoplasmic loops delineated by TM1- TM2 (139), TM2-TM3 (178), and TM3-TM4 (213) intervals (Figure 1).

N-term C-term

5 4 2 6

3 5 B

NH2- GFP 2 4 6

COOH

2. 1: Construction of VANGL1 proteins containing hemagglutinin (HA) epitope tags. The sites for insertion of YPYDVDYA hemagglutinin (HA) epitopes in individual VANGL1 proteins are indicated (black arrowheads), along with numerical designation of the construct (NT, 1-6, CT). (A) These have been superimposed on the hydropathy plot of VANGL1 protein produced by TOPPRED2 software package using a 20-residue long sliding window (core window, 12 residues; wedge windows, 4 residues). The default parameters suggested for predicting the presence of TMDs in eukaryotic proteins were used; such as the GES cut-off values (lower 0.6; putative; upper, 1.0; certain). (B) In the bottom panel is a 2D model of the VANGL protein with the inserted HA tags at various positions.

Expression of Epitope-Tagged VANGL1 Mutants in Transfected MDCK Kidney Cells All tagged VANGL1 cDNAs constructs were transfected into MDCK cells and the expression of the resulting HA-tagged VANGL1 proteins was initially detected by GFP fluorescence in intact cells and subsequently verified by immunoblotting. Representative immunoblots using an anti-HA monoclonal antibody are shown for single cell clones expressing individual mutant VANGL1 proteins (Figure 2). The MDCK cell line was chosen as a recipient in transfection experiments since it is derived from a cell lineage (kidney tubular epithelium) that normally expresses VANGL proteins in vivo. Briefly, we could obtain stable transfectants expressing robust levels of VANGL1 for constructs 18, 64, 139 and 213, while lower levels of expression were consistently noted in cell clones expressing constructs 178 and 314 (Figure 2). The recombinant proteins were detected as immunoreactive bands of ~80kDs in agreement with the predicted size of the VANGL1 protein backbone (~55KDa) fused to GFP (27kDa)

HA-Vangl1

tubulin

2. 2: Expression of HA epitope-tagged VANGL1 constructs in MDCK cells. MDCK cells were stably transfected with GFP-hVANGL1 cDNA constructs modified by the addition of individual HA epitopes (HA18, HA64, HA139, HA178, HA213 and HA314) and inserted in the mammalian expression vector pCB6 (immonoreactive band of ~80 kDa). Total cell lysates (50 μg per lane) were prepared, resolved by electrophoresis on a 7.5% SDS-polyacrylamide gel, and analysed by Western blotting using a mouse monoclonal anti-HA.11 epitope antibody.

The subcellular localization of GFP-HA VANGL1 proteins was analyzed by confocal microscopy. As opposed to the cytoplasmic fluorescence signal detected in MDCK cells expressing only control GFP (Figure 3), GFP- VANGL1 MDCK transfectants showed strong localization of the GFP fluorescence at the membrane (Figure 3; middle panels), producing a typical mesh-like signal in confluent cell monolayers. In addition, the GFP-VANGL1 signal was found to fully colocalize with Na, K-ATPase, a specific marker for basolateral membrane of MDCK cells (Figure 3; middle panels). Taken together, results in Figures 3 show that all GFP-tagged VANGL1 constructs are expressed at the basolateral membrane of MDCK cells, indicating that none of the HA tag insertions affect the maturation or membrane targeting of the protein.

GFP Na,K-ATPase Merge

MDCK

HA 18

HA 64

HA 139

HA 178

HA 213

HA 314

2. 3: Cellular localization HA epitope-tagged VANGL1 proteins transfected in MDCK cells. Transfected MDCK cells stably expressing GFP alone or the different GFP- VANGL1 constructs (green) containing single HA epitope tags (HA18, HA64, HA139, HA178, HA213 and HA314) were grown to confluence, fixed, permeabilized, and stained for Na, K-ATPase followed by Cy3- conjugated secondary antibody (red), and examined by confocal microscopy. The merged images show that the majority of GFP-tagged VANGL1-HA constructs are associated with the plasma membrane and colocalize with Na, K-ATPase (yellow). Images are representatives of at least three independent experiments.

Membrane Topology of the Inserted HA-Epitope Tags The intra- or extra-cytoplasmic localization of the HA tags was determined for each construct by immunofluorescence analysis of the corresponding MDCK transfectant using the anti- HA monoclonal antibody (Figure 4 and Supplemental Figure S2). Immunofluorescence was performed in parallel on intact cells to detect extra-cytoplasmic tags and in cells permeabilized with detergent to detect tags present on the intra-cytoplasmic face of the membrane. For all MDCK transfectants, GFP fluorescence images were also acquired in order to identify the position of cells examined by immunofluorescence with the anti-HA antibody. Strong fluorescent signals were observed in both intact and permeabilized cells for constructs 139 (TMD 1-2) and 213 (TMD 3−4), suggesting that these corresponding segments of VANGL1 are extra-cytoplasmic. Conversely, strong fluorescence signals were obtained only under permeabilized conditions for cells transfected with constructs 18 (N-terminus), 64 (N-terminus), 178 (TMD 2-3) and 314 (C- terminus), suggesting that the corresponding tagged segments of VANGL1 are intra-cytoplasmic. For all of the constructs, the polarity of the HA tags was validated by immunofluorescence in multiple independently transfected cell clones.

Permeabilized Intact

GFP anti-HA GFP anti-HA HA 18 A B C D out in D N- GFP C HA 64 E F G HH out in N- GFP C HA 139 I J K L out in N- GFP C HA 178 M N O PP out in N- GFP C HA 213 Q R S TT out in N- GFP C HA 314 U V W XX out in N- GFP C

2. 4: Immunofluorescence analysis of HA epitope-tagged VANGL1 proteins in intact and permeabilized cells. MDCK cells stably expressing individual HA-tagged VANGL1 proteins were analyzed by immunofluorescence. A schematic representation of the VANGL1 proteins with the position (HA18, HA64, HA139, HA178, HA213 and HA314) of the inserted HA tag (red arrows) is shown on the left. Immunofluorescence was carried out with the mouse monoclonal anti-HA epitope antibody (HA.11) on cells either untreated (intact cells: right two columns) or permeabilized with 0.1% Triton X-100 (permeabilized; left two columns). Cells were then incubated with Cy3- conjugated secondary goat anti-mouse antibody and images were acquired by confocal microscopy. The HA epitopes in MDCK cells expressing HA139 (I-L) and HA213 (Q-T) were detected in intact and permeabilized cells (extra-cytoplasmic), while the HA18 (A-D), HA64 (E- H), HA178 (M-P), and HA314 (U-X) were detected only in permeabilized cells (intra- cytoplasmic). GFP fluorescent images (columns 1 and 4) were used as controls. Images are representative of at least three independent experiments.

Validation of HA-Epitope Localization in VANGL1 The topology of inserted HA tags established by immunofluorescence was validated by an enzymatic method, which additionally provides an estimative measure of the amount of exofacial tag expressed at the surface of the cell. For this, MDCK transfectants expressing individual HA- tagged VANGL1 constructs were fixed and incubated with primary anti-HA antibody with (total protein expression) or without (surface protein expression) permeabilization with detergent. The cells were then incubated with a secondary horse radish peroxidase (HRP) conjugated antibody and the amount of bound secondary antibody was quantified colorimetrically using HRP substrate o-phenylenediamine dihydrochloride (OPD). Results in figure 5 show the amount of surface expression of each VANGL1-HA construct, presented as a percentage of the total amount of protein expressed in the corresponding clone and measured under permeabilized conditions. MDCK cells transfected with constructs 139 and 213 were found to have a significantly greater proportion of accessible HA-tagged VANGL1 at their surface (87% and 41%, respectively) than constructs 18, 64, 178 and 314 (which had between 3% and 11%). These results are in agreement with immunofluorescence experiments and indicate that the HA tags in constructs 139 and 213 are expressed on the extra-cytoplasmic side of the membrane, while the HA tags in constructs 18, 64, 178 and 314 are expressed on the intra-cytoplasmic side of the membrane.

100

40 expression

20 % of cell surface % of cell surface 0

2. 5: Detection of HA epitope-tagged proteins by surface labeling. MDCK cells stably expressing individual HA-tagged VANGL1 proteins were incubated with anti- HA antibody with or without prior permeabilization with 0.1% Triton X-100 treatment, followed by incubation with HRP-coupled secondary antibody. The amount of bound secondary antibody was determined by a colorimetric assay using the HRP substrate o-phenylenediamine dihydrochloride (OPD) measured at 490 nm. The presence of HA-tagged VANGL1 expressed at the cell surface (in intact cells) is shown as a fraction of total protein expression (in permeabilized cells) normalized for non-specific binding of the secondary antibody alone in each individual clone. Error bars represent standard errors of the means of at least four independent experiments (in quadruplicates).

2.5 Discussion

To date, VANGL1 and VANGL2 are the only genes in which loss-of-function mutations have been associated with neural tube defects in humans. The Vangl gene was initially identified in genetic screens in flies for mutations that affect the structure of certain appendages on the wing and legs (hair bristles) and the orientation of the eye unit (omatidiae) structure. This indicates that Vangl plays an important role in the orientation of cells in the plane of epithelial structures, a basic process known as planar cell polarity (PCP). In mammals; Vangl also regulates the process of convergent extension (CE) movements. CE is critical to the formation of many embryonic structures including the neural tube and the heart. Little is known about the molecular mechanism of action of VANGL proteins in the establishment of PCP or in the control of CE movements. However, recruitment of VANGL proteins to the plasma membrane is absolutely required for function and we have shown that Vangl1/2 mutations affecting PCP of inner ear structures and causing craniorachischisis in loop-tail (Lp) mice abrogate membrane expression of the protein 19,23. Likewise, VANGL proteins can interact with cytoplasmic DVL 22and we have observed that Lp- associated Vangl2 mutations in mice, and neural tube defects associated VANGL1 mutations in humans abrogate interaction with DVL proteins. However, the protein domains involved in protein-protein interactions in these membrane complexes, or in the recruitment of extracellular adhesion molecules to propagate positional information, or in the transmission of intracellular signals are unknown.

As a first step to address these issues, we have determined the membrane topology of human VANGL1. For this, we inserted HA epitopes at discrete positions of the membrane associated portion of the protein. The modified proteins were then expressed in transfected cells and the membrane topology of the inserted tag was determined by immunofluorescence and by surface labelling. For these studies, we chose the kidney polarized epithelial cell line MDCK. VANGL proteins localize at the baso and lateral side of several epithelial cells in vivo, including bronchial tree, intestinal crypt, sweat glands and kidney tubules 19. MDCK is derived from kidney and we have shown that the VANGL proteins expressed in these cells are targeted to the basal and lateral membrane 23. Thus, topology data generated from proteins expressed at the basolateral membrane of MDCK cells are likely to reflect native membrane topology of VANGL proteins in

69 normal tissues. HA epitopes were inserted in the amino terminus at 2 different positions; 18 and 64 (NT constructs), and at position 314 in the carboxy terminal portion (CT construct). Three additional HA tags were inserted at positions 139, 178, and 213 which represented peaks of hydrophilicity in the otherwise hydrophobic membrane associated portion of the protein (based on hydropathy analysis shown in Fig. 1). All constructs could be expressed at robust levels in the membrane of MDCK transfectants, suggesting that none of the inserted tags disrupted protein folding and/or membrane targeting. These results suggest that the short predicted intra-cytoplasmic or extra-cytoplasmic loops in the membrane regions are solvent exposed and indeed accessible for antibody binding to the tag. Results from epitope mapping studies (immunofluorescence and surface labeling) were unambiguous and demonstrated a 4 transmembrane domain structure for VANGL1 with intracellular amino and carboxy termini, including a large intra-cytoplasmic domain in the C-terminal half downstream of TM4 (Figure 6).

2. 6: Topological model and structural features of the VANGL protein family. The position of the four transmembrane domains of the protein as predicted by hydropathy profiling and as established here by epitope mapping is presented. Individual predicted intracellular and extracellular segments and their position in the primary amino acid sequence is shown. Amino acid residues defining sequence landmarks and signature motifs are depicted in different colors, including amino acid variants in Vangl1/Vangl2 associated with neural tube defects in mice and humans (orange, pink) and where loss-of-function has been demonstrated (pink), the ETSV sequence corresponding to a PDZ binding motif (green), as well as consensus tyrosine based and di-leucine motifs associated with protein sorting and membrane targeting (blue) [see text for details]. Finally, the polarity of the protein and membrane domains with respect to the membrane (light blue) is indicated (in, out, lumen).

This topology data and associated model established here experimentally for the human VANGL1 are most likely applicable to the rest of the VANGL protein family. Indeed, not only do members of the VANGL family share a high degree of sequence similarity, but functional complementation studies in zebrafish mutants (trilobite) have also shown conservation of function (directing convergent extension movements) between the fish and human proteins 25. The membrane topology and associated secondary structure established here put in perspective a number of predicted structural and functional features of the protein. First, they demonstrate a cytoplasmic location for the C-terminal moiety of the protein, including the PDZ-binding motif ETSV. This agrees with the previous observations made in intact proteins and with protein subdomains showing interactions of this C-terminal VANGL1/2 segment with the DIX sub-domain of cytoplasmic PCP proteins DVL1, DVL2, and DVL3 22. The ETSV tetrapeptide has been shown to promote these interactions 101,102. Secondly, this topology places membrane targeting and/or internalization motifs for clathrin mediated endocytosis (YSFG7-10, YSYY10-13, LL56-57) in the N- terminus and (YKDF284-287, NPNL291-294) in the C-terminus of the protein, where they would be expected to be functional. However, this topological model also suggests that the predicted asparagine (N)-linked glycosylation sites NDS59-61 (N-terminus), NST314-316 and NNS317-319 (C- terminus) are unlikely to be functionally modified since they map to intra-cytoplasmic portions of the protein (Figure 6).

There are three available loop-tail (Lp) alleles for which loss-of-function mutations at Vangl2 have been reported (D255E, R259L, S464N). In humans, 8 VANGL2 mutations (S84F, R135W, R177H, L242V, T247M, R353C, F437S) have been reported in neural tube defects, and so far two of them have been shown to be a partial (R353C) or a complete (F437S) loss-of-function expressed as loss of the VANGL/DVL interaction. On the other hand, 8 VANGL1 mutations (S83L, F153S, R181Q, L202F, V239I, R274Q, M328T, A404S) have been detected in sporadic or familial cases of NTDs, and two (M328T, V239I) are known to be complete loss-of-function as shown by impairment of VANGL/DVL interaction and loss of functional complementation in zebrafish 25. Although the molecular mechanism underlying the defect of the other mutations remain unknown, a superimposition of the site of disease-associated VANGL variants on the structural model established points out interesting structure-function relationships (Figure 6). We note that of the 19 pathology-associated variants, 5 of them map to the membrane associated

72 portion of the proteins including the loss-of-function VANGL1V239I allele detected as a de novo mutation in a familial case of NTDs. Strikingly, we note the clustering of 5 additional such variants to a short 30 amino acids segment immediately downstream TM4, including two Lp-associated mutations Vangl2D255E and Vangl2R259L. The clustering of L242F, T247M and R274Q to this segment indirectly suggests that these variants may also be pathological. The functional role of this mutation-sensitive segment is unknown, but biochemical analyses of D255E R259L and S464N 23 shows that it is likely to be important for protein folding and membrane targeting. Finally, the remaining variants are distributed in the C-terminal cytoplasmic domain of the protein. Such variants may interfere with recruitment of other PCP proteins such as DVL or PK to VANGL and these hypotheses can now be tested experimentally.

2.6 Acknowledgments

Image acquisition, data analysis and image processing were done on equipment and with the assistance of the McGill Life Science Complex Imaging Facility which is founded by the Canadian Foundation for Innovation.

2.7 Supplementary Figures

HA position TOPPRED TMHMM N-terminal 18, 64 1-114 1-110

Loop 1-2 139 135-152 134-147

Loop 2-3 178 173-186 171-181

Loop 3-4 213 207-221 205-218

C-terminal 314 242-524 242-524

2. S 1: TMHMM probability plot and comparison with TOPPRED2 model. (A) TMHMM model prediction of hVANGL1 transmembrane helices. (B) Table indicating the position of the inserted HA tags in the hVANGL1 protein and a comparison of predicted position of loops between TOPPRED2 and TMHMM software tools.

2. S 2: Immunofluorescence analysis of HA epitope-tagged VANGL1 proteins without the GFP in intact and permeabilized cells. MDCK cells stably expressing individual wild type VANGL1 proteins modified by the addition of an HA tag at position 64 and 139. A schematic representation of the two transfected VANGL1 constructs is shown. Immunofluorescence was carried out as described in Figure 4, using the mouse monoclonal anti-HA epitope antibody (HA.11) on cells either untreated (intact cells: right column) or permeabilized with 0.1% Triton X-100 (permeabilized; left column). In agreement with results from Figure 4, the HA epitopes at position 64 and 139 are found to localize to the intra and extra-cytoplasmic side, respectively. This indicates that insertion of the GFP tag in constructs used in Figure 4 does not affect the membrane targeting or membrane topology of the VANGL protein. Images are representative of at least three independent experiments.

Chapter 3: Molecular and cellular mechanisms underlying neural tube defects in the loop-tail mutant mouse

Preface to Chapters 3 and 4

The loop-tail mouse was first described in 1949, when Strong noticed several mice with a ‘kinked’ tail in his ‘A’ strain. When homozygous, these mice died in utero of craniorachischisis, a phenotype that took fifty years to attribute to three independent mutations in the Vangl2 gene. S464N (Lp) is the the naturally occurring allele, while D255E (Lpm1Jus) and R259L (Lpm2Jus) are chemically ENU induced. In vivo, the S464N and D255E mutations behave identically and are transmitted in a semi-dominant fashion, while the R259L is a hypomorphic allele and shows a weaker spina bifida phenotype.

Work described in chapter 2 provides a structural basis for investigations done in chapter 3 and 4. While topology mapping allowed us to position all three Lp-associated alleles intracellularly, characterizing the consequence of D255E, R259E and S464N by using a series of biochemical assays provides additional critical comprehension in the structure-function relationship of VANGL proteins. In order to functionally characterize the Lp-associated mutations, an in vitro system stably expressing independently WT or mutant constructs was set up in the polarized epithelial MDCK cell line. Results showing impaired plasma membrane targeting, altered subcellular localization, decreased stability and much smaller half-life provide an explanation as to how normal VANGL protein function is altered in NTDs.

3. 1 Abstract

Loop-tail (Lp) mice show a very severe neural tube defect (craniorachischisis) caused by mutations in the Vangl2 gene (D255E, S464N). Mammalian VANGL1 and VANGL2 are membrane proteins that play critical roles in development such as establishment of planar cell polarity (PCP) in epithelial layers and convergent extension movements during neurogenesis and cardiogenesis. VANGL proteins are thought to assemble with other PCP proteins (DVL, PK) to form membrane-bound PCP signaling complexes that provide polarity information to the cell. In the present study, we show that VANGL1 is expressed exclusively at the plasma membrane of transfected MDCK cells, where it is targeted to the basolateral membrane. Experiments with an inserted exofacial HA epitope indicate that the segment delimited by the predicted transmembrane domains 1 and 2 is exposed to the extracellular milieu. Comparative studies of the Lp-associated pathogenic mutation D255E indicate that the targeting of the mutant variant at the plasma membrane is greatly reduced; the mutant variant is predominantly retained intracellularly in endoplasmic reticulum (ER) vesicles colocalizing with the ER marker calreticulin. In addition, the D255E variant shows drastically reduced stability with a half-life of ~ 2hrs, compared to > 9hrs for its wild type counterpart and is rapidly degraded in a proteasome-dependent and MG132 sensitive pathway. These findings highlight a critical role for D255 for normal folding and processing of VANGL proteins, with highly conservative substitutions not tolerated at that site. Our study provide an experimental framework for the analysis of human VANGL mutations recently identified in familial and sporadic cases of spina bifida.

3. 2 Introduction

Neural tube defects (NTDs) are a heterogeneous group of pathologies that occur when the neural tube fails to close properly. NTDs constitute one of the most common congenital abnormalities in humans, occurring in 1 per 1000 live births 103. Multiple factors, both genetic and environmental have been implicated in the etiology of NTDs, underlying the causal heterogeneity of the disorder 89,103. The study of mouse mutants affected with NTDs has enabled the identification of molecular pathways involved in the complex process of neurulation 89. loop-tail (Lp) is a semi- dominant mutation where Lp/+ heterozygotes display urogenital defects and a characteristic “looped” tail 13, while Lp/Lp homozygotes display craniorachischisis and die in utero shortly before or at birth. Craniorachischisis is characterized by a completely open neural tube, from the hindbrain region to the most caudal extremity of the embryo, while the forebrain and the midbrain regions are closed properly 16,17,92. We have identified Vangl2 as the gene mutated in the loop-tail (Lp) mouse model of NTDs 16,17. Vangl2 RNA is expressed embryonically in several tissues, including robust levels in the neural tube immediately prior to, during and after closure 16,17,19,92. Vangl2 encodes a membrane protein comprising four putative transmembrane domains and a large intracellular domain with a PDZ-domain-binding motif (PBM) at its carboxy terminus 22. Recently, we have shown that the VANGL2 protein is expressed at the plasma membrane of neuroepithelial cells of E10.5 of mouse embryos and is also present at the basolateral membrane of several types of epithelial cells 19.

Vangl genes are highly conserved in evolution with members in flies (Vang/Stbm), fish (trilobite/Vangl2) and frogs (Xstbm) 35,40,41,104. In Drosophila the Vang/Stbm relative is required for establishing planar cell polarity (PCP) in the developing eye, wing and leg95,105. This process is regulated by membrane-associated signaling complexes composed of Vang/Stbm, the plasma membrane receptor Frizzled (Fz), the cytoplasmic proteins Dishevelled (Dsh/Dvl) and Prickle (Pk), and the atypical cadherin Flamingo/Starry night (Fmi/Stan) and Diego (Dgo) 106-112. Establishment of PCP involves the redistribution of these proteins; cytoplasmic Dvl and Pk are recruited to the membrane to form asymmetrically distributed membrane complexes with Fz and Stbm/Vang, respectively 113. Vangl proteins have been shown to bind both Dvl and Pk 96,114-116. In the absence of Vang, Dsh and Pk are mislocalized and PCP signaling is impaired 96. Fz and Vang

79 also participate in intercellular PCP signal relay 80. In addition, vertebrate relatives of fly PCP genes have been shown to participate in convergent extension (CE) movements during embryogenesis 54. CE is the process in which layers of cells intercalate (converge) and become longer (extension) contributing to a variety of morphogenetic processes. CE movements are responsible for the narrowing and lengthening of the neural plate during neural tube closure 35,54,117,118. Interestingly, mutations in mammalian counterparts of the Drosophila PCP genes Vangl, Fz, Dvl and Celsr1 cause severe NTDs 16,17,20,72,73.

In vertebrates, a second Vangl gene, VANGL1, has been described 22,119 VANGL1 and VANGL2 proteins are highly similar, including identical major predicted secondary structure features. In mice, Vangl1 is also expressed in the developing neural tube (ventral neural tube, notochord). VANGL1 genetically interacts with VANGL2 and mouse embryos doubly heterozygous for Vangl1/2 mutations (Vangl1+/-: Vangl2+/-) display craniorachischisis 18. Moreover, several mutations in sporadic (M328T) and familial (V239I, R274Q) cases of NTDs were identified in the human VANGL1 gene, including a de novo mutation (V239I) appearing in a familial setting 76,120. The structural similarity shared by VANGL proteins 22, their genetic interaction during neural tube development 18, together with additional functional complementation data in other model organisms 93,120suggest that Vangl1 and Vangl2 have identical tissue-specific biochemical activities.

So far, two Lp alleles have been described, the naturally occurring Lp and the chemically induced Lpm1Jus 16,17. Sequence analysis revealed the presence of independent Vangl2 mutations in both alleles, namely D255E (Lpm1Jus) and S464N (Lp), both of which map to the predicted cytoplasmic domain of the protein 16,17,92. The similar phenotypes in vivo of embryos heterozygous and homozygous for each mutant suggest that both mutations behave as loss-of-function in a gene dosage dependent pathway. The mechanistic and molecular basis for the loss-of-function in these proteins in vivo and in vitro remains largely unknown and was investigated in the present study. For this, we stably expressed the wild type protein and the D255E Lp-associated mutant variant in Madin-Darby canine kidney (MDCK) cells and studied the effect of D255E on protein expression, stability, maturation, subcellular localization and cell surface expression.

3.3 Experimental Procedures

Material and antibodies Cycloheximide was purchased from Sigma-Aldrich (St-Louis, MO). MG132 was from Calbiochem (San Diego, CA). Geneticin (G418), penicillin and streptomycin were obtained from Invitrogen (Carlsbad, CA). All the restriction enzymes were from New England Biolabs (Ipswich, MA). Taq DNA polymerase was from Invitrogen (Carlsbad, CA). The mouse monoclonal antibodies directed against the influenza hemagglutinin epitope (HA.11) and the c-Myc epitope were purchased from Covance (Berkeley, CA). The mouse monoclonal antibody recognizing Na, K-ATPase (alpha) was from Santa Cruz Biotechnology (Santa Cruz, CA). The mouse monoclonal anti-BiP antibody was from BD Biosciences (San Jose, CA). The rabbit polyclonal antibodies against calnexin and PDI (Protein Disulfide Isomerase) were from Stressgen (Ann Arbor, Mi). The rabbit polyclonal antibody against calreticulin was obtained from Affinity BioReagents (Golden, CO). Cy3-conjugated goat anti-mouse and anti-rabbit antibodies and peroxidase-coupled goat anti- mouse antibody were from Jackson ImmunoResearch Laboratories (West Grove, PA).

Plasmids and constructs The human Vangl1 cDNA was amplified from total human RNA by RT-PCR and cloned into the pCS2+ vector. The D255E mutation was introduced in the human Vangl1 cDNA by PCR overlap extension mutagenesis, followed by subcloning into pCS2+. PCR was used to insert a short antigenic epitope (EQKLISEEDL) from the human c-Myc protein at the N-terminus of VANGL1 wild type (WT) and VANGL1 D255E sequences, using oligonucleotides 5’- CAAGAAGAATTCATGGAGCAGAAGCTAATCTCTGAGGAGGATCTGGATACCGAATC CACTTATTC-3’ (forward Vangl1 primer with the c-Myc epitope sequence) and 5’- CTCTTGTCTAGATTAAACGGATGTCTCAGACTG-3’ (reverse Vangl1 primer). The recombinant c-Myc tagged cDNAs were re-constructed using unique EcoRI and XbaI sites present in the primers, followed by cloning into the corresponding sites of the pCB6 expression vector. For the cell surface quantification of VANGL1 protein expression, an hemagglutinin (HA) epitope (YPYDVPDYA) was inserted immediately after amino acid position 139 of VANGL1 WT and VANGL1 D255E, using PCR mutagenesis with primers: 5’- TACCCATACGATGTGCCAGACTACGCTAGCGATGAGCTGGAGCCTTGTGGCACAATT

TGT-3’ (forward VANGL1-HA139) and 5’- GCTAGCGTAGTCTGGCACATCGTATGGGTACCTCCACAGGATCGGAGGTAAAAGGA TGAA-3’ (reverse VANGL1-HA139), with HA coding sequences in bold. To enable insertion of additional adjacent HA epitope(s), a NheI restriction site was inserted at the 3’-end of the HA sequence, accounting for the additional Ser residue found at the end of the HA tag. To facilitate identification of transfected cells expressing recombinant VANGL1 proteins, c-Myc tagged Vangl1 cDNAs were fused in-frame to the green fluorescent protein (GFP) by cloning the various Vangl1 EcoRI-XbaI PCR fragments into peGFP-C1 vector (Clontech, Mountain View, CA) modified to accommodate the in-frame insertion of the EcoRI-XbaI Vangl1 wild type, Vangl1 HA139, Vangl1 D255E and Vangl1 HA139+D255E PCR fragments.

Cell Culture and Transfection Madin-Darby canine kidney (MDCK) epithelial cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin and 100 µg/ml streptomycin at 37°C in a 5% CO2 incubator. For stable transfection, MDCK cells were transfected with VANGL1 constructs subcloned in peGFP-C1 using Lipofectamine Plus Reagent (Invitrogen). Selection for stably transfected clones was in growth medium containing 0.4 mg/ml G418 for 10-14 days. Expression of recombinant GFP- VANGL1 variants was initially identified by fluorescence microscopy on live cells and confirmed by Western blotting analysis. Total cell lysates were prepared in RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100 supplemented with protease inhibitors), cleared by passage through a 25G needle and centrifugation at 13000 rpm for 10 minutes (4oC). Cell lysates (50 µg) were separated by electrophoresis using 7.5% SDS-polyacrylamide gels, followed by electroblotting and incubation with the monoclonal anti-c-Myc antibody 9E10 (used at 1:500). Immune complexes were revealed with a horseradish peroxidase conjugated goat anti- mouse antibody (1:10000) and visualized by enhanced chemiluminescence (SuperSignal West Pico kit, Thermo Scientific Rockford, IL).

Crude membrane preparation Crude membrane fractions from stably transfected MDCK cells expressing GFP-VANGL1 WT and D255E were prepared as described 121. Cells seeded in three 150-mm dishes were grown

82 to confluence; they were then washed once in ice-cold phosphate buffered saline (PBS) and once in cold NTE buffer (10 mM Tris-HCl pH 7.5, 100 mM NaCl, 10 mM EDTA) and promptly harvested by scrapping in Tris-Mg buffer (10 mM Tris-HCl pH 7.5, 1 mM MgCl2 with protease inhibitors) followed by homogenization using a Dounce homogenizer (50 strokes on ice). Unbroken cells and nuclei were removed by centrifugation at (500g for 5 min) and a crude membrane fraction was prepared by centrifugation of the supernatant fraction at 50000g for 30 min. The pellet containing the membrane proteins was re-suspended in NTE buffer supplemented with protease inhibitors. The supernatant containing the cytosolic proteins was also harvested for analysis.

Quantification of Cell Surface Expression by Enzyme-linked Immunosorbent Assay Quantification of cell surface expression of HA-tagged GFP-VANGL1 proteins was done by enzyme-linked immunosorbent assay as described previously 100. Briefly, MDCK cells stably transfected with HA-tagged VANGL1 WT or D255E were grown to confluence in 24-well plates and fixed with 4% paraformaldehyde for 20 minutes. Cells were blocked in 5% nonfat milk in PBS for 30 minutes, incubated with mouse anti-HA Ab (1:200 dilution) for 1 hour, washed and incubated with HRP conjugated goat anti-mouse Ab (1:4000 dilution) for 1 hour. For quantification of total HA-tagged GFP-VANGL1, cells were permeabilized with 0.1% Triton X- 100 in PBS for 30 minutes prior to incubation with anti-HA Ab. Peroxidase activity was quantified colorimetrically using the HRP substrate [0.4 mg/ml O-phenylenediamine dihydrochloride (OPD), Sigma-Aldrich] according to the manufacturer’s instructions. Absorbance readings (492 nm) were taken in an ELISA plate reader and background absorbance reading from nonspecific binding of primary antibody to vector-transfected cells were subtracted for each sample. Cell surface reading were normalized to total HA-tagged GFP-VANGL1 value for each cell clone and were expressed as a percentage.

Metabolic Labeling, Pulse-Chase Study and Immunoprecipitation Confluent MDCK cells stably expressing GFP-VANGL1 WT and D255E proteins seeded on 60 mm dishes were pre-incubated 90 min at 37°C in methionine- and cysteine-free DMEM media containing 10% dialyzed FBS (labeling media). Thereafter, cells were pulsed-labeled for 60 min at 37°C with 2 ml of labeling media containing 100 μCi of [35S] methionine-cysteine (Perkin-

Elmer, Boston, MA). Cells were then washed twice with PBS and incubated in 2 ml of chase medium (standard DMEM containing 10% FBS, 15 µg/ml methionine and 15 µg/ml cysteine) for up to 8 h. Cells were lysed in 400 μl of RIPA buffer. For immunoprecipitation, equal amounts of cell lysates were adjusted to 1 mg/ml in RIPA buffer and incubated with the mouse anti-Myc antibody 9E10 (2.5 µg) overnight at 4°C, followed by incubation with protein A/G-sepharose (GE Healthcare, Piscataway, NJ) for 4 h. Beads were washed three times in RIPA buffer and proteins were eluted with 50µl of 2X sample buffer. The radiolabeled proteins were separated by electrophoresis using 7.5% SDS polyacrylamide gels. The gels were then fixed, impregnated with Amplify (GE Healthcare, Piscataway, NJ), dried and exposed to film. Densitometric analysis was performed by using NIH ImageJ software (NIH, Bethesda, MD).

Determination of Relative Protein Stability using Cycloheximide Chase To determine the stability of both GFP-VANGL1 WT and D255E proteins, MDCK cells expressing these proteins were treated with cycloheximide (20 µg/ml). At pre-determined time points, cells were harvested, lysed with RIPA buffer and subjected to SDS-PAGE and immunoblot analysis. Signal intensities were quantitated using NIH ImageJ software, with actin or tubulin used as internal controls.

Immunofluorescence and confocal microscopy To examine subcellular localization (permeabilized conditions), MDCK cells stably expressing different VANGL1 constructs were seeded at high density onto glass coverslips in a 24-well plate. Twenty-four hours later, cells were washed twice with ice-cold PBS, fixed for 15- 20 min in 4% paraformaldehyde (PFA) in PBS and then permeabilized with 0.5% Triton X-100 in PBS for 15 min. After blocking non-specific binding with 5% goat serum and 0.1% bovine serum albumin (BSA) in PBS for 1 h, the cells were incubated with the desired primary antibody [mouse anti-Na, K-ATPase (1:100); mouse anti-HA (1:200); rabbit anti-calreticulin (1:100)] for 1 h and washed three times with 0.1% BSA in PBS. Finally, the coverslips were incubated for 1 h with the appropriate secondary antibody [goat anti-mouse-Cy3 (1:1000); goat anti-rabbit-Cy3 (1:1000)] and washed three times with 0.1% BSA in PBS. All the incubations were done at room temperature and the antibodies were all diluted in blocking solution. To detect cell surface expression of the exofacial HA epitope engineered in VANGL1 proteins (nonpermeabilized conditions), MDCK

84 cells expressing HA-tagged GFP-VANGL1 WT and D255E were incubated with the mouse anti- HA (1:200) in DMEM containing 2% nonfat milk for 2 hours at 37ºC. After several washes with PBS, cells were fixed for 15 min with 4% PFA in PBS and incubated with goat anti-mouse-Cy3 antibody (1:1000) for 1 h. For immunofluorescence, coverslips were rinsed once in water and mounted by using Permafluor Aqueous Mounting Medium (Thermo Scientific, Fremont, CA). Confocal microscopy was done using a Zeiss LSM5 Pascal laser scanning confocal microscope. All image analyses were performed using the LSM5 Image software. To maximize image quality, a Median filter 3x3 was applied to the images using Image-Pro software. To remove out-of-focus fluorescence, the Z stack sections were also deconvoluted using 10 iterations with the AutoQuant software.

3.4 Results

Expression and subcellular localization of VANGL1 WT and D255E in MDCK cells. Targeting of the VANGL1 protein to the cell membrane is thought to be critical for biological function in vivo, including the formation of membrane associated multi-subunit PCP signaling complexes 55,96,122. To investigate the possible effect of the Lp-associated pathogenic mutation D255E on this and other aspects of VANGL1 protein structure and function, we stably expressed wild type (WT) and D255E VANGL1 protein in MDCK cells. This cell type was chosen, since it is derived from a lineage that normally expresses VANGL proteins in vivo during kidney tubulogenesis 19and is therefore likely to possess the cellular machinery required for proper expression, maturation, membrane insertion and subcellular targeting of the protein. To facilitate detection in transfected cells, the WT Vangl1 cDNAs were first modified by in-frame addition of GFP at the amino terminus and also with the insertion of the antigenic c-Myc for fluorescence localization and immunoblotting, respectively (Figure 1A). Several stably transfected MDCK clones were identified by GFP fluorescence and VANGL1 protein expression was confirmed by immunoblotting analysis (data not shown).

Total cell extracts and crude membrane fractions were prepared from two independent MDCK VANGL1 WT transfectants and were analyzed by immunoblotting (anti-c-Myc antibody; Figure 1B). A specific immunoreactive species of ~80 kDa was detected in total cell extracts, in agreement with the predicted molecular mass of the GFP-tagged protein (Figure 1B; top panel). This species was greatly enriched in crude membrane extracts but was absent from cytoplasmic fractions. These results suggest that the GFP- VANGL1 WT protein is associated with membrane in the stable MDCK cell clones.

The VANGL1 D255E mutant fused to GFP was also stably expressed in MDCK cells and the effect of the mutation on various structural and functional properties of VANGL1 was investigated. Several D255E expressing MDCK cell clones were initially identified by GFP fluorescence and total cell extracts, crude membrane and cytosolic fractions were analyzed by immunoblotting (Figure 1B; bottom panel). As for WT VANGL1, D255E was found to be highly enriched in the membrane fraction of these transfectants, suggesting that D255E does not impair

86 membrane association of the VANGL1 protein. On the other hand, we systematically observed in multiple transfected clones a lower overall level expression of the D255E variant compared to WT (Figure 1B, Supplemental Figure S2 and data not shown).

A B

3. 1: Expression of WT and D255E VANGL1 variant in the membrane fraction of transfected MDCK cells. (A) Schematic representation of the GFP- VANGL1 protein including positions of the hemagglutinin (HA) tag inserted at position 139 (predicted TM1-TM2 connecting loop), the c- Myc tag inserted in the amino-terminal segment, GFP at the N-terminus and the D255E mutation in the intracellular C-terminal region. (B) Independent MDCK cell clones stably expressing either WT or D255E mutant variant (reconstructed in human GFP- VANGL1 protein) were isolated. Total cell extracts (CE), membrane-enriched fractions (M) and soluble cytoplasmic extracts (C) were prepared from these cells and equal amounts of protein (50 μg) were analyzed by immunoblotting with the mouse monoclonal antibody 9E10 directed against the c-Myc tag inserted in-frame near the amino terminus of the recombinant proteins.

The subcellular localization of GFP- VANGL1 WT was analyzed by confocal microscopy. As opposed to the cytoplasmic fluorescence signal detected in MDCK cells expressing control GFP (Figure 2; top panels), GFP- VANGL1 MDCK transfectants showed strong localization of the GFP fluorescence at the membrane (Figure 2; middle panels), producing a typical mesh-like signal in confluent cell monolayers. In addition, the GFP- VANGL1 signal was found to fully colocalize with Na, K-ATPase, a specific marker for basolateral membrane of MDCK cells (Figure 2; middle panels). Taken together, results in Figure 1B and 2 show that GFP-tagged VANGL1 is mainly expressed at the plasma membrane of MDCK cell clones.

To detect a possible effect of the D255E mutation on the subcellular distribution and plasma membrane (PM) targeting of VANGL1, the cellular localization of VANGL1 D255E in MDCK cells was also analyzed by confocal microscopy and compared to that of the WT protein (Figure 2; bottom panels). By contrast to WT VANGL1 that shows a clear PM-associated signal, D255E signal did not appear concentrated at the PM, but rather showed an intracellular punctate pattern suggestive of localization to an intracellular endomembrane compartment.

3. 2: Cellular localization of WT and D255E VANGL1 proteins in transfected MDCK cells. Transfected MDCK cells stably expressing GFP alone, GFP- VANGL1 WT and GFP- VANGL1 D255E (green) were grown to confluence, fixed, permeabilized and stained for Na, K-ATPase followed by Cy3-conjugated secondary antibody (red), and examined by confocal microscopy. The merge images show that while the majority of GFP- VANGL1 WT signal is associated to the plasma membrane and overlaps with Na, K-ATPase (yellow), the GFP- VANGL1 D255E staining is mostly intracellular with little if any colocalization with Na, K-ATPase. Images are representative of at least 3 independent experiments of each type.

In addition, D255E did not colocalize with the PM marker Na, K-ATPase. The seemingly distinct subcellular localization of the WT and D255E proteins was confirmed by examination of cell sections (Z stacks) (Figure 3). These studies showed that the WT VANGL1 is expressed predominantly at the basal and lateral membranes of polarized MDCK cells (Figure 3A – top panel and Supplemental Figure S1), where it shows complete colocalization with the basolateral membrane marker Na, K-ATPase (Figure 3A – middle and bottom panels). However, a similar analysis of D255E expressing cells showed that the mutant variant is not detected at the basolateral membrane and does not colocalize with Na, K-ATPase; the D255E signal is largely limited to an intracellular compartment (Figure 3B and Supplemental Figure S1). Together, these results strongly suggest that the pathogenic D255E mutation interferes with normal subcellular targeting of the protein.

3. 3: Cellular localization of the D255E mutant VANGL1 variant in stably transfected MDCK cells; Z-line image analysis. Transfected MDCK cells stably expressing either GFP- VANGL1 WT (A) or the D255E mutant variant (B) were processed as described in the legend to Figure 2. Cells were analyzed by confocal microscopy to visualize the expression of GFP- VANGL1 (green) and Na, K-ATPase (red). Representative X-Z sections are shown. The merge images show that while GFP- VANGL1 WT colocalizes with Na, K-ATPase at the basolateral membrane (yellow) of the polarized cells, the D255E mutant does not and is mainly localized to an intracellular compartment.

Surface expression of VANGL1 proteins in MDCK cells. Results in Figure 3 show that in polarized MDCK cells, WT VANGL1 is expressed at the basolateral membrane. It is anticipated that in younger, non-polarized cultures of MDCK cells, WT VANGL1 will be broadly present at the cell surface. To detect possible surface expression of WT VANGL1 in non-polarized MDCK cells, we engineered an exofacial HA antigenic epitope at position 139 located in the first predicted extracellular loop of VANGL1 (delineated by the predicted TM1-TM2 segments). The mutant D255E was similarly modified by HA tag addition and analyzed in parallel. The polarity of the HA tag with respect to PM was determined by immunofluorescence with an anti-HA antibody used in intact cells (extracellular) and in cells permeabilized with Triton X-100 (intracellular) (Figure 4), as we have previously described 100,121 A strong fluorescence signal was detected in both permeabilized and nonpermeabilized MDCK clones expressing WT VANGL1 (Figure 4A); this signal was absent from control untransfected MDCK cells similarly exposed to the anti-HA antibody (data not shown). These results indicate that the inserted HA tag is surface accessible and strongly suggest that a) WT VANGL1 is expressed at the cell surface of non-polarized MDCK cells and b) that the protein segment delineated by predicted TM1-TM2 is indeed extracytoplasmic. By contrast, MDCK clones expressing the D255E variant displayed bright intracellular fluorescence in permeabilized cells (Figure 4A bottom – left), but showed almost no fluorescence under nonpermeabilized conditions (Figure 4A bottom – right), suggesting that D255E is not expressed at the cell surface of MDCK cells. The presence/absence of immunoreactive HA tag at the cell surface of MDCK cells expressing WT or D255E proteins was also investigated by an ELISA based assay 100. For this, corresponding MDCK transfectants were fixed and incubated with anti-HA antibody with or without prior permeabilization. The amount of bound anti-HA antibody was quantified using a secondary antibody coupled to horseradish peroxidase. Cells expressing WT VANGL1 -HA analyzed in this manner were found to express 93 ± 6% of total immunoreactive VANGL1 at the cell surface (Figure 4B), while only a small fraction of D255E was detected at the cell surface (14% ± 1%). Together, results in Figures 2-4 show that the D255E mutation disrupts normal subcellular localization of VANGL1 and interferes with normal PM targeting of the protein.

3. 4: Analysis of plasma membrane targeting and surface expression of WT and D255E VANGL1 variants in MDCK cells. (A) MDCK cells stably expressing either HA-tagged GFP- VANGL1 WT or D255E were analyzed by immunofluorescence to detect surface expression of an HA antigenic epitope inserted at position 139 in the predicted TM1-TM2 connecting loop. Immunofluorescence was carried out on either intact cells (non-permeabilized, right panel) or in cells pre-treated with 0.1% Triton-X-100 (permeabilized; left panel), using a mouse anti-HA antibody. Cells were incubated with Cy3- conjugated secondary antibody and images were acquired by confocal microscopy. Images are representative of at least 3 independent experiments of each type. (B) In other experiments, the same cells exposed to the anti-HA antibody were incubated with an HRP-coupled secondary anti- mouse antibody and the amount of the primary antibody present was determined for both conditions (permeabilized and non-permeabilized) by a colorimetric reaction using an HRP substrate quantitated by spectrometry. The amount of HA-tagged GFP- VANGL1 WT and D255E proteins expressed at the cell surface (detected in nonpermeabilized cells) is shown as a fraction (%) of total protein expression (measured in permeabilized cells).

Stability of the D255E variant in MDCK cells. The overall lower level of the D255E variant detected in transfected MDCK cells (compared to WT; Figure 1B, Supplemental Figure S2 and data not shown), together with the noted defect in subcellular targeting of the variant prompted us to analyze a possible effect of the mutation on protein half-life and stability. For this, MDCK transfectants expressing WT or D255E variants were treated with cycloheximide to block protein synthesis, and the fate of the mature protein was monitored over time by immunoblotting analysis of corresponding cell lysates harvested at pre-determined time-points (Figure 5).

3. 5: Stability of GFP- VANGL1 WT and D255E proteins in MDCK cells. (A) MDCK cells stably expressing either GFP- VANGL1 WT or D255E were grown to confluence and treated with cycloheximide (20 μg/ml) for the indicated time. Cell lysates (50 μg) were analyzed by immunoblotting using the anti-c-Myc antibody 9E10. Actin was used as an internal loading control and was detected using a rabbit polyclonal antiserum. To obtain comparable signal intensity at T0, the GFP- VANGL1 blots were exposed for longer times. (B) Estimation of the rate of disappearance of GFP- VANGL1 WT and D255E. The Vangl1 signal from multiple immunoblots (mean ± SE of 4-9 experiments) was quantified (pixel density of each band) and is expressed as a fraction (%) of the untreated sample which is set at 100%.

Using this approach, we determined that the half-life of WT VANGL1 expressed in MDCK cells was ~ 9 hours, while that of the D255E variant was ~ 2 hours. Eight hours of incubation with cycloheximide resulted in almost complete disappearance (~ 95%) of D255E, while >50% of the WT protein was still detectable in cells similarly treated. The stability of WT and D255E VANGL1 proteins was also investigated by pulse-chase experiments. MDCK clones were labeled with 35S- methionine/cysteine for 60 min (pulse) and the fate of the radiolabeled proteins was followed for 8 hrs by immunoprecipitation using extracts from pulsed cells grown in isotope-free medium (chase). Representative autoradiograms and quantification from 3 independent experiments are shown in Figure 6A and B, respectively. In these experiments, the half-life of WT VANGL1 in MDCK cells was estimated to be >13 hours, while the half-life of D255E was found to be significantly shorter than ~4 hours. Results from these experiments agree with that monitoring protein stability following exposure to cycloheximide and strongly suggest that the D255E mutation causes a strong decrease in protein stability.

3. 6: Pulse-chase studies of GFP- VANGL1 WT and D255E proteins expressed in MDCK cells. (A) MDCK cells stably expressing either GFP- VANGL1 WT or D255E were metabolically labeled by a 60-minute pulse of 35S-Met/Cys. The cells were transferred to radioactivity-free complete medium, followed by different chase periods. Cell lysates were prepared and subjected to VANGL1 immunoprecipitation followed by gel electrophoresis and autoradiography. To obtain comparable signal intensity at T0, the GFP- VANGL1 D255E blots were exposed for longer times. (B) Estimation of the rate of disappearance of GFP- VANGL1 WT and D255E labeled proteins. The amount of radiolabeled VANGL1 WT and D255E proteins was deduced from scanning the autoradiograms as described in the legend to Figure 5. The graph represents the mean of three experiments ± SE.

ER retention of VANGL1 D255E protein. The reduced half-life, the absence of PM targeting and the localization of the D255E variant to a subcellular endomembrane compartment suggested that the D255E may have a defect in maturation associated with retention of the protein in the endoplasmic reticulum (ER). This possibility was tested by immunofluorescence using an antibody directed against the ER marker calreticulin. MDCK cells expressing WT VANGL1 showed absence of overlap between the PM- associated GFP fluorescence (GFP- VANGL1) and the intracellular punctate ER staining produced by calreticulin (Figure 7 – top panels). In contrast, there was significant colocalization of D255E and calreticulin, suggesting that D255E is indeed present predominantly in the ER endomembrane compartment (Figure 7 – middle and bottom panels).

3. 7: Determination of subcellular localization of VANGL1 D255E by double immunofluorescence. MDCK cells stably expressing either GFP- VANGL1 WT or the D255E (clone 1 and clone 2) were fixed and immunostained for the endoplasmic reticulum marker calreticulin (red). As opposed to GFP- VANGL1 WT which is expressed at the plasma membrane, the mutant GFP- VANGL1 D255E shows intracellular staining that significantly overlaps with calreticulin. Insets are magnifications of the indicated areas.

100

Retention of inappropriately folded proteins in the ER, including membrane proteins, ultimately leads to their degradation by the proteasome 123. To analyze proteasome-mediated proteolytic degradation of VANGL1, MDCK cells expressing WT or D255E were exposed to cycloheximide (CHX) (up to 8 hrs) either in the presence or absence of the proteasome inhibitor MG132 (Figure 8 and Supplemental Figure S2). Cell extracts were prepared and analyzed by immunoblotting, and the amount of immuno-reactive VANGL1 protein was quantified. As previously shown in Figure 5, CHX treatment of WT and D255E expressing cells had a differential effect on the level of VANGL1 protein remaining after 8 hours of treatment, with a >90% decrease in D255E variant, compared to only a small reduction in the amount of WT protein. On the other hand, MG132- mediated proteasome inhibition of CHX treated cells resulted in a dramatic increase in the amount of detectable D255E variant (~60% of untreated controls) in these cell extracts. MG132/CHX treatment also resulted in a modest increase of detectable WT protein (when compared to CHX treatment alone). Furthermore, the addition of chloroquine, a lysosome inhibitor, did not prevent the rapid degradation of D255E (Supplemental Figure S2). Together, these results suggest that D255E is retained in the ER, where it is targeted for degradation in a proteasome-dependent manner.

101

3. 8: Proteasome-mediated degradation of WT and D255E VANGL1 expressed in MDCK cells. (A) MDCK cells stably expressing either GFP- VANGL1 WT or D255E were grown to confluence and treated with cycloheximide (20 μg/ml) for various time periods, either in the absence or presence of the proteasome inhibitor MG132 (5 μg/ml). Cell lysates (50 μg) were subjected to gel electrophoresis and immunoblotting for VANGL1 using the anti-c-Myc antibody. Tubulin was used as an internal loading control (not shown). To obtain comparable signal intensity at T0, the GFP- VANGL1 D255E blots were exposed for longer times. (B) The effect of MG132 on the stability of the WT and D255E mutant after 8 hrs of treatment was quantitated by densitometry scanning as described in the legend of Figure 5. Each value represents mean ± SE of 5-6 independent experiments.

102

The presence of misfolded proteins in the ER may induce the unfolded protein response (UPR), which involves the upregulation of the expression of genes encoding ER-resident chaperones, e.g. BiP 124. If D255E was recognized by the cell as a misfolded protein it should trigger the UPR when expressed in cells. To analyze UPR activation, we measured the expression of ER chaperones (BiP, calnexin and PDI) by western blotting in MDCK cells expressing WT and D255E. No differences were seen in the expression of BiP, calnexin and PDI between the different cell lines (Supplemental Figure S2), indicating that UPR is not activated in MDCK cells expressing D255E VANGL1 protein.

3.5 Discussion

Vangl1 and Vangl2 are mammalian homologs of the Drosophila Van Gogh (vang) or Strabismus (Stbm) proteins 40,104. The two proteins share ~70% sequence similarity, including a 4 TM domains predicted structure with intracellular amino and carboxy termini. In Drosophila, Vang/Stbm belongs to the PCP gene family 95,105,125,126, and is involved in the coordination of polarization of cells within the epithelial plane. In vertebrates, a clear example of PCP is the alignment and orientation of stereocilia in neurosensory cells of the cochlea (organ of Corti) 20,72,73,80,98,125and mutations in Vangl1 18and Vangl2 19,20 disrupt polarity of these stereociliary bundles 127. Moreover, studies in Xenopus and subsequently in other vertebrates, showed that alterations in the Vangl genes cause defects in convergent extension movements during embryogenesis 20,63,70,128. In mammals, CE movements play a critical role in a number of developmental pathways such as neural tube closure, kidney tubulogenesis, as well as biogenesis of the cardiovascular system 19,66,71,129. Indeed, mice carrying mutations in Vangl1 18 and Vangl2 16,17,92 show severe craniorachischisis, as well as other malformations, most notably in the heart and large blood vessels 18,71,130. Finally, studies in normal mouse embryos by our lab 19 and others 57,80 have shown that VANGL proteins are expressed at the cell membrane in a number of epithelial cell lineages throughout embryonic development. Together, these results indicate that proper biogenesis of Vangl proteins, including targeting to the membrane, plays a critical role in the establishment of PCP and in CE movements in mammals. The biochemical characteristics of VANGL proteins remain largely unknown. Likewise, their molecular mechanism of action both during normal embryogenesis and in the case of pathogenic

103 mutations (Lp) associated with NTDs, are poorly understood. For these studies, human VANGL1 was used as a molecular backbone; this was to anchor potential future characterization of the rare VANGL1 variants recently identified in human cases of spina bifida 76,120, in the biochemical framework described here. Although the pathogenic D255E mutation studied here emerged in Vangl2, much data indicate that VANGL1and VANGL2 proteins are functionally equivalent and that the choice of the molecular backbone should not impact the biochemical characterization of pathogenic mutations. Indeed, the two proteins are structurally highly similar and expression of the zebrafish Vangl1 93 or human VANGL1 25 in zebrafish tri/Vangl2 mutants can correct the CE defect of these mutant animals. In addition, loss-of-function mutations in mouse Vangl1 and Vangl2 show the same PCP defect in vivo 18. Finally, the Lp-associated D255E mutation abrogates VANGL interaction with DVL1, DVL2 and DVL3, when reconstructed in either VANGL1 or VANGL2 22. The biochemical characterization of Vangl1 reported here thus provides an experimental framework to study the molecular basis of loss-of-function mutations in the VANGL proteins.

We have stably expressed WT VANGL1 protein (fused to GFP) in MDCK cells and observed that VANGL is mainly expressed at the plasma membrane. Furthermore, using an exofacial HA antigenic epitope inserted between the predicted TM1 and TM2, we show that this portion of the protein is exposed to the extracellular milieu, where it is available for possible interactions with other proteins. The extracytoplasmic localization of the TM1-TM2 segment demonstrated here also supports the topological model predicted by hydropathy profiling that we proposed previously 22. In polarized cultures of MDCK cells, VANGL1 is expressed at the basolateral membrane, where it colocalizes with Na, K-ATPase. This restricted basolateral expression is similar to the subcellular localization that we observed for VANGL1 and VANGL2 in the inner ear of mouse embryos 18. Moreover, we also demonstrated that VANGL2 was detected mostly at the cell membrane in several tubular structures of mouse embryos, including developing kidney, oesophagus and skin 19. This comparable pattern of expression indicates that MDCK cells represent an appropriate host to study the biochemical properties of VANGL proteins, including maturation, processing and membrane targeting. Comparative analysis of the NTD-associated D255E revealed that the mutation has multiple and complex effects on the biochemical properties of the protein. First, the expression of the D255E

104 variant is greatly reduced at the plasma membrane or at the basolateral membrane of polarized MDCK, but is majorly found in an endomembrane compartment that shows overlapping staining with the ER marker calreticulin. This suggests a major defect in maturation and targeting of the D255E variant. Concomitant with altered subcellular localization, we observed that the D255E mutation causes a dramatic reduction in protein stability, both at steady state (reduced expression in transfectants) and in pulse-chase studies where a 2hr half-life was noted compared to > 9hrs for the WT protein. Finally, reduced stability of the D255E variant is also accompanied by rapid, MG132 sensitive, proteasome-dependent degradation of the mutant. Together, these results suggest that the D255E mutation causes misfolding of the protein which is associated with a failure to mature and in the accumulation of the mutant variant in the ER. The D255E variant appears unstable and is rapidly targeted for degradation by the proteasome. The overall net effect is that D255E is predominantly not expressed at the plasma membrane, thereby strongly suggesting that it cannot participate in the formation of PCP signaling complexes. While the current manuscript was under review, another study published by Merte et al. supports our claim that the D255E mutation interferes with the targeting of the protein at the cell surface. They showed that the VANGL2 D255E protein fails to become recruited by COPII vesicles and leads to its retention in the ER 64. The properties of the D255E mutant including instability, ER retention and failure to reach the site of biological function (plasma membrane) are likely to result in a complete loss-of- function variant. Such behavior and associated null phenotype has been previously described for pathogenic mutations in a number of other membrane proteins, for example Pgp (ABCB1) 131, CFTR (ABCC7) 132 and Slc11a2 (Nramp2) 133. Although we favor a model in which D255E causes a general defect in protein folding, it is also possible that D255E affects a targeting signal required for protein sorting and trafficking to the plasma membrane. Sequence analysis reveals the presence of putative basolateral targeting motifs in the N-terminus (YSGF10, YSYY13) and downstream

TM4 (YKDF287) of VANG1. Furthermore, a dileucine motif, known to be required for endocytosis, but sometimes also for basolateral sorting, is found in the N-terminus of VANGL1 (LL57). However, D255E does not affect any of these signatures.

In the absence of high resolution three-dimensional structure of VANGL proteins, one can only speculate on the effect of the D255E on VANGL protein structure, including the associated folding/maturation defect of this variant. D255 maps to the large C-terminal intracellular domain

105 of the protein, approximately 10 residues downstream TM4 and is absolutely conserved (invariant) in all VANGL proteins sequenced to date, from flies to humans. Although D255 is not part of any recognizable signature or structural motif, its high degree of conservation suggests an important role. Moreover, the nature of the substitution at D255 that causes loss-of-function in vivo is highly conservative, replacing an aspartate by a glutamate; this substitution preserves the negatively charged carboxylate, while introducing a modest change in the size of the carbon side chain. The fact that a highly conservative change at that position has such a profound effect on protein structure and function also suggest a critical role for this residue and domain in protein folding and stability, including possible interaction with other proteins required for maturation.

The complex biochemical defect of the D255E variant characterized herein explains a number of in vivo observations on the behavior of the mutant. Using immunofluorescence on sections of neural tubes from E10.5 embryos, we have detected reduced overall expression level and decreased apparent plasma membrane expression of D255E in neuroepithelial cells of Lp/Lp embryos compared to WT 19. Likewise, in hair cells of the cochlea, membrane staining of both VANGL2 and associated DVL2 is disrupted in Lp/Lp (D255E) embryos 70, the latter possibly through impaired binding to VANGL2. These findings are consistent with the severe targeting defect of D255E detected here in MDCK cells. In addition, we have shown that the C-terminal portion (251-526) of VANGL1 and VANGL2 physically interacts with the N-terminal portion (containing the DIX and PDZ domains) of the three Dvl proteins (Dvl1, 2 and 3) and that the D255E mutation abrogates this interaction 22. The major effect of the D255E mutation on protein stability detected in MDCK cells (including possible misfolding and retention in the ER) would also explain the loss of interaction with Dvl proteins detected by yeast two hybrid analysis 22.

Lp/+ heterozygous mouse exhibits a mild looped tail phenotype, whereas Lp/Lp homozygotes exhibit both a looped tail and craniorachischisis 13,16,17. This co-dominant phenotype is the same for both known alleles at Lp, namely Lp (S464N) and Lpm1Jus (D255E)16. The codominance nature of the Lp defect can be explained either by a) a partial or complete loss-of- function of the mutant gene/protein in a gene dosage dependent pathway or b) a gain of function associated with a dominant negative effect. Our results strongly suggest that D255E is a loss-of- function mutation. We believe that VANGL proteins form part of a gene dosage dependent

106 pathway critical for neural tube formation. This is also supported by the observations that mice doubly heterozygote for loss-of-function at Vangl1 and Vangl2 (Vangl1-/+:Vangl2Lp/+) show the same severe craniorachischisis NTD as Vangl2Lp/Lp homozygotes 18. However, one cannot formally exclude the possibility that the D255E exerts a dominant negative effect on the WT protein. In such a scenario, D255E would interfere with normal maturation or targeting of the WT protein produced in the same cells. This hypothesis is supported by the report that VANGL proteins can form multimers 134. Additional studies will be required to clarify this point.

Taken together, our finding of improper targeting of the D255E mutation associated with NTDs in mice suggests that this conserved residue in both VANGL1 and VANGL2 plays an important role in correct protein folding/targeting. Consequently, the absence of the VANGL proteins at the plasma membrane disrupts the PCP pathway and causes major developmental defects, such as neural tube defect. Identifying protein determinants responsible for targeting of VANGL proteins to the plasma membrane and elucidating the mechanistic basis of loss-of- function in mutant variants associated with NTDs in mice and human may help understanding the mechanism by which VANGL1 and VANGL2 proteins control cell polarity and CE movements.

3.6 Acknowledgements

We thank Drs. Elena Torban, Alan Underhill, Doug Epstein and Sergio Grinstein for their thoughtful comments on the manuscript. Image acquisition, data analysis and image processing were done on equipment and with the assistance of the McGill Life Sciences Complex Imaging Facility which is funded by the Canadian Foundation for Innovation.

107

3.7 Supplementary Figures

3. S 1: Immunolocalization of GFP- VANGL1 WT and D255E in MDCK cells. MDCK cells stably expressing either GFP- VANGL1 WT or D255E were grown to confluence and stained for Na, K-ATPase. The immunofluorescence images obtained by Z-stack confocal microscopy showed that GFP- VANGL1 WT is expressed at the basolateral membrane, whereas GFP- VANGL1 D255E is expressed mostly intracellularly. Green corresponds to the localization of GFP- VANGL1 WT or D255E, and red shows the distribution of Na, K-ATPase. X-Z and Y- Z correspond to a lateral view of the cells; X-Y shows a central section of the cells.

108

3. S 2: Degradation and UPR response of WT and D255E VANGL1 proteins expressed in MDCK cells. (A) Degradation of WT and D255E VANGL1 proteins expressed in MDCK cells. Stably transfected MDCK cells expressing either GFP- VANGL1 WT or D255E were grown to confluence and treated with cycloheximide (20 μg/ml) for 4 hrs, either in the absence or presence of the proteasome inhibitor MG132 (5 μg/ml) or the lysosome inhibitor chloroquine (100 μM). Cell lysates (50 μg) were subjected to gel electrophoresis and immunoblotting for VANGL1 using the anti-c-Myc antibody. Tubulin was used as an internal loading control. The rapid degradation of GFP- VANGL1 D255E was prevented by MG132, but not by chloroquine, indicating that the D255E protein is mainly degraded by proteasome. These data are representative of two such experiments performed. (B) ER retention of GFP- VANGL1 D255E does not induce UPR. Confluent non-transfected MDCK cells or stably transfected MDCK cells expressing GFP- VANGL1 WT and D255E were lysed, and western blots performed on cell lysates (50µg) using the anti-c-Myc antibody to detect VANGL1, and antibodies directed against UPR markers (anti- BiP, anti-calnexin and anti-PDI). Tubulin was used as an internal loading control. The expression pattern of the UPR markers is identical in all cells, indicating that the ER stress response is not induced by the accumulation of D255E in the ER. These data are representative of two such experiments performed.

109

Chapter 4: Loss of membrane targeting of VANGL proteins causes neural tube defects

110

4.1 Abstract

In the mouse, the loop-tail mutation (Lp) causes a very severe neural tube defect, which is caused by mutations in the Vangl2 gene. In mammals, Vangl1 and Vangl2 code for integral membrane proteins that assemble into asymmetrically distributed membrane complexes that establish planar cell polarity in epithelial cells, and that regulate convergent extension movements during embryogenesis. To date, VANGL are the only genes, in which mutations cause neural tube defects in humans. Three independently arising Lp alleles have been described for Vangl2: D255E, S464N, and R259L. Here we report a common mechanism for both the naturally occurring Lp (S464N) and for a novel ENU-induced mutation Lpm2Jus(R259L). We show that the S464N and R259L variants stably expressed in polarized MDCK kidney cells fail to reach the plasma membrane, their site for biological function. The mutant variants are retained intracellularly in the endoplasmic reticulum, colocalizing with ER chaperone calreticulin. Furthermore, the mutants also show a dramatically reduced half-life of ~3 hours, compared to ~22 hours for the wild type protein, and are rapidly degraded in a proteasome-dependent and MG132-sensitive fashion. Coexpressing individually the three known allelic Lp variants with the wild type protein does not influence the localization of the WT at the plasma membrane, suggesting that the co-dominant nature of the Lp trait in vivo is due to haploid insufficiency caused by a partial loss-of-function in a gene dosage dependent pathway, as opposed to a dominant negative phenotype. Our study provides a biochemical framework for the study of recently identified mutations in VANGL1 and VANGL2 in sporadic or familial cases of neural tube defects.

111

4.2 Introduction

Neural tube defects (NTDs) are the second most common congenital malformations in humans affecting 1-2 infants per 1000 live births 1. In humans, the etiology of NTDs is poorly understood, but is thought to involve interplay between complex genetic determinants and ill- defined environmental factors. This complexity has precluded the systematic identification of major gene effects associated with NTDs in humans. On the other hand, the study of naturally occurring or experimentally induced mouse mutants bearing NTDs (over 200 described to date) 91 has been a rich source of candidates whose relevance to the human condition can be readily tested in cohorts of sporadic or familial cases of NTDs 75,76,120. The loop-tail (Lp) mouse mutant is an established model for the study of NTDs in humans; Lp/+ heterozygote mice are viable but present a strongly kinked (looped) tail, while Lp/Lp homozygotes are not viable and suffer from a very severe NTD called craniorachischisis, characterized by a completely open neural tube, from the hindbrain to the most caudal region of the embryo 16,17,92

We have previously shown that the NTD phenotype of Lp mice is caused by mutations in the Vangl2 gene 16. Vangl2 was originally identified in Drosophila as Vang/Stbm (VanGogh, Strabismus), a so-called “core PCP gene”, part of a group of genes that play a critical role in the establishment of planar cell polarity (PCP) of a number of structures in the fly (including the eye, the wing, and several cellular appendages) 94,135. Vang is an integral membrane protein that directly or indirectly interacts with other PCP proteins such as membrane bound Frizzled (Fz), Diego (Dgo), Flamingo (Fmi), cytoplasmic Prickle (Pk) and Dishevelled (Dvl); The asymmetric distribution of membrane bound Dvl/Fz and Vang/Pk complex in epithelial cells is believed to propagate a polarity signal essential for PCP organization of cell layers 97. In absence of Vang, Dsh and Pk are mislocalized and PCP signaling is impaired 96. PCP genes and proteins have been highly conserved throughout evolution 120. In vertebrates, a clear example of PCP is the precise orientation of the stereocilliary bundles of the neurosensory epithelium of the organ of Corti (inner ear) 80,98. Mouse mutants bearing mutations in vertebrate relatives of Vang/Stbm (Vangl1/Vangl2) 18,22, Dsh (Dvl) 71, and Fmi (Celsr1) 73 show disruption of these structures of the inner ear. In addition, vertebrate relatives of fly PCP genes regulate convergent extension (CE) movements during embryogenesis 125. CE movements contribute to several morphogenetic processes during

112 embryogenesis, including the narrowing and lengthening of the neural plate critical to neural tube closure 54,136. The severe NTD phenotype (craniorachischisis) we originally described in Vangl2 (Lp) 16,17 and Vangl1/Vangl2 mutants 18 has since been reported for mutations in other mouse PCP genes such as Dvl1/2 72, Dvl2/3 71, Celsr1 73, and Fz3/6 72.

In mice and humans, there are two VANGL proteins, VANGL1 and VANGL2, that share over 80% sequence similarity. Both encode tetra-spanning integral membrane proteins comprising of a PDZ-binding motif (PBM) in a large C-terminal intra-cytoplasmic tail that promotes interactions with PDZ containing proteins (such as Dvl). VANGL1 and VANGL2 show a dynamic pattern of mRNA and protein expression during neural tube closure: VANGL1 is expressed primarily in ventral structures including the notochord, while VANGL2 is abundant on the dorsal side, with both expressed in the floor plate 19. VANGL1and VANGL2 are also expressed in a number of additional tubular structures during embryogenesis 19. Partial inactivation of Vangl1/Vangl2 in double heterozygotes causes a NTD as severe as that seen in Vangl2Lp/Lp mutants, highlighting the critical role of both genes and proteins during neural tube formation and associated NTDs 18.

We have recently identified mutations in the human VANGL1 gene in familial (V239I, R274Q) and sporadic cases (M328T) of neural tube defects 8,76. Some of these variants (V239I, M328T) were shown to abrogate interaction of VANGL proteins with DVL and impair convergent extension movements in a complementation assay in zebrafish 8,25. In addition, independent mutations in VANGL2 (S84F, R353C and F437S) have been identified in still-born human foetuses afflicted by severe NTDs 75; Two of these mutations have been shown to abrogate (F437S) or significantly reduce (R353C) interaction of VANGL2 protein with members of the Dvl family. These results have identified VANGL1 and VANGL2 as critical genetic determinants in the etiology of NTDs in humans. Finally, we have identified three independent Lp alleles caused by independent mutations at Vangl2, namely Lpm1Jus (D255E) 23, Lp (S464N) and a novel Lpm2Jus (R259L) 87, all of which map to the long cytoplasmic domain of VANGL2.

The biochemical mechanism of action of VANGL proteins in establishing PCP and mediating convergent extension movements during embryogenesis in vertebrates remain poorly

113 understood. Likewise, the discrete molecular determinants required for assembly of membrane- bound Vangl-dependent signaling complexes, and how such complexes may become non- functional in mice and humans bearing NTDs is unknown. Recently, we have established the biochemical characteristics of VANGL proteins in transfected MDCK cells, and used several assays to identify the molecular basis for loss-of-function in the Lpm1Jus Vangl2 allele (D225E) 23. We have now characterized two additional Lp-associated loss-of-function variants S464N (Lp) and R259L (Lpm2Jus). The results show that in all cases, loss of function is caused by impaired targeting of the protein to plasma membrane, which is associated with reduced half-life, and retention of the mutant variants in the endoplasmic reticulum.

114

4.3 Experimental Procedures

Material and Antibodies Cycloheximide was purchased from Sigma-Aldrich (St-Louis, MO). MG132 was from Calbiochem (San Diego, CA). Geneticin (G418), hygromycin, penicillin and streptomycin were obtained from Invitrogen (Carlsbad, CA). All restriction enzymes were from New England Biolabs (Ipswich, MA). Taq DNA polymerase was from Invitrogen. The mouse monoclonal antibodies directed against the influenza hemagglutinin epitope (HA.11) and the c-Myc epitope were purchased from Covance (Berkeley, CA). The mouse monoclonal antibody recognizing Na, K-ATPase (alpha) was from Santa Cruz Biotechnology (Santa Cruz, CA). The rabbit polyclonal antibody against calreticulin was obtained from Affinity BioReagents (Golden, CO). Cy3- conjugated goat anti-mouse and anti-rabbit antibodies and peroxidase-coupled goat anti-mouse antibody were from Jackson ImmunoResearch Laboratories (West Grove, PA). Alexa-Fluor 647- conjugated goat anti-mouse was from Invitrogen.

Plasmids and Constructs The R259L and S464N mutations were introduced in the human VANGL1 cDNA by PCR overlap extension mutagenesis, as described previously 23. The following nucleotides were used 5’-ACCGATGGCGAGTCCCTCTTCTACAGCCTGGGACAC-3’ (forward Vangl1-R259L primer), 5’-GTGTCCCAGGCTGTAGAAGAGGGACTCGCCATCGGT-3’ (reverse Vangl1- R259L primer), 5’-CAGTGGAGGCTTGTCAATGATGAGGCTGTGACTAAT-3’ (forward Vangl1-S464N primer), 5’-ATTAGTCACAGCCTCATCATTGACAAGCCTCCACTG-3’ (reverse Vangl1-S464N primer). Briefly, a short antigenic epitope (EQKLISEEDL) from the human c-Myc protein was introduced at the N-terminus of VANGL1 sequence. For quantification of protein cell surface expression, a hemagglutinin (HA) epitope (YPYDVPDYA) was inserted immediately after amino acid position 139 of each variant. To facilitate identification of transfected cells expressing recombinant proteins, c-Myc tagged VANGL1 was fused in-frame to the green fluorescent protein (GFP) by cloning into a modified peGFP-C1 vector (Clontech, Mountain View, CA). To study the effect of coexpression of wild type (WT) and mutant variants in the same cells, the c-Myc tagged VANGL1 cDNA without the HA epitope was fused to a

115 fluorescent Cherry proteins (from pmCherry; Clonetech), and the resulting Cherry-fusion was subcloned into pcDNA3.1/Hygro(+) vector (Invitrogen) for expression in transfected cells. Cell Culture and Transfection. Madin-Darby canine kidney (MDCK) epithelial cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 µg/ml streptomycin at 37°C in a 5% CO2 incubator. For stable transfection, MDCK cells were transfected with cDNA constructs subcloned in peGFP-C1 using Lipofectamine Plus Reagent (Invitrogen). Selection for stably transfected clones was in growth medium containing 0.3 mg/ml G418 for 10-14 days. Expression of recombinant GFP-VANGL1 R259L and S464N was initially identified by fluorescence microscopy on live cells, and confirmed by Western blotting analysis of cell extracts. Total cell lysates were prepared in RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1% SDS supplemented with protease inhibitors), cleared by passage through a 25G needle and centrifugation at 13 000 rpm for 10 minutes (4oC). Cell lysates (50 µg) were separated by electrophoresis using 7.5% SDS-polyacrylamide gels, followed by electroblotting and incubation with the monoclonal anti-c-Myc antibody 9E10 (used at 1:500). Immune complexes were revealed with a horseradish peroxidase conjugated goat anti-mouse antibody (1:10000) and visualized by enhanced chemiluminescence (SuperSignal West Pico kit, Thermo Scientific Rockford, IL). For coexpression studies, stable MDCK cells expressing WT, D255E, R259L, or S464N variants were transfected as described above with Cherry-VANGL1 WT construct. Selection for stably transfected clones was in medium containing 0.3 mg/ml G418 and 0.3 mg/ml hygromycin for 10-14 days. Several clones coexpressing Cherry-VANGL1 WT and the GFP-VANGL1 variants were identified by fluorescence microscopy on live cells.

Crude Membrane Preparation Crude membrane fractions from MDCK transfected cell clones were prepared as described 23. Cells were seeded in three 150-mm dishes and grown to confluency. After two washes in ice- cold phosphate buffered saline (PBS) and once in cold NTE buffer (10 mM Tris-HCl pH 7.5, 100 mM NaCl, 10 mM EDTA), cells were promptly harvested by scrapping in Tris-Mg buffer (10 mM

Tris-HCl pH 7.5, 1 mM MgCl2 with proteinase inhibitors) followed by homogenization using a Dounce homogenizer. Unbroken cells and nuclei were removed by centrifugation (500g for 5 min), and a crude membrane fraction was prepared by centrifugation of the supernatant fraction at 50

116

000g for 30 min. The final pellet containing the enriched membrane fraction was re-suspended in NTE buffer supplemented with protease inhibitors.

Quantification of Cell Surface Expression by Enzyme-Linked Immunosorbent Assay. Quantification of cell surface expression of HA-tagged VANGL proteins in transfected MDCK cells was carried out by enzyme-linked immunosorbent assay as described previously 100 with the following modifications. MDCK cells stably expressing HA-tagged WT, R259L, and S464N variants were seeded at near confluency in 24-well tissue culture plates, and grown for 4 days. Cells were then washed with PBS and incubated in Ca++-free DMEM for 20 min, prior to incubation with EGTA 10 mM for 5 min. To evaluate the expression of HA-tagged proteins at the cell surface, cells were incubated with mouse anti-HA antibody (1:200 in Ca++-free DMEM) for 2 h at 37ºC/5% CO2. Cells were washed, and fixed with fresh 4% paraformaldehyde (PFA) in PBS for 15 min., and incubated with HRP conjugated goat anti-mouse antibody (1:4000 in 5% skim milk) for 1 h. To evaluate the total amount of HA-tagged WT and mutant variants present, cells were fixed after the EGTA treatment, permeabilized with 0.1% Triton X-100 in PBS for 30 minutes, and blocked in 5% skim milk in PBS. Cells were then incubated with anti-HA antibody (1:200 in blocking solution) for 1 h, washed, and incubated with HRP conjugated goat anti-mouse antibody (1:4000 in blocking solution). Peroxidase activity was quantified colorimetrically using the HRP substrate [0.4 mg/ml O-phenylenediamine dihydrochloride (OPD), Sigma-Aldrich] according to the manufacturer’s instructions. After incubating the cells for 30 min in the dark, the reaction was stopped using 50µl of 3M HCl per well containing 200μl of reaction and absorbance readings (490 nm) were taken in an ELISA plate reader. Background absorbance reading from nonspecific binding of primary antibody to vector-transfected cells were subtracted for each sample and cell surface expression was normalized to total HA-tagged protein expression for each cell clone and was expressed as a percentage.

Pulse-Chase Assays and Determination of Relative Protein Stability MDCK cells stably expressing WT, R259L, and S464N proteins were pre-incubated 90 min at 37°C in methionine- and cysteine-free DMEM medium containing 10% dialyzed FBS (labelling media). Thereafter, cells were pulsed-labeled for 60 min at 37°C with 2 ml of labelling medium containing 100 μCi of [35S] methionine cysteine (Perkin-Elmer, Boston, MA), as

117 described previously 23. Cells were then washed twice with PBS, and incubated in 2 ml of chase medium (DMEM containing 10% FBS, 15 µg/ml methionine and 15 µg/ml cysteine) and incubation was continued for various times and up to 8h. Cells were then lysed in 400 μL of RIPA buffer. For immunoprecipitation, equal amounts of cell lysates (adjusted to 1 mg/ml in RIPA buffer) were incubated with the mouse anti-Myc antibody 9E10 (2.5 µg) overnight at 4°C, followed by incubation with protein A/G-sepharose (GE Healthcare, Piscataway, NJ) for 4h. Beads were washed three times in RIPA buffer and proteins were eluted with 50µl of 2X sample buffer. The radiolabelled proteins were separated by electrophoresis using 7.5% SDS polyacrylamide gels. The gels were then fixed, impregnated with Amplify (GE Healthcare, Piscataway, NJ), dried and exposed to film. Densitometric analysis was performed by using NIH ImageJ software (NIH, Bethesda, MD). In certain experiments aiming to measure protein stability, MDCK cells expressing WT, R259L and S464N variants were treated with cycloheximide (20 µg/ml) for 4h to inhibit protein synthesis. When used, the proteasome inhibitor MG132 (5 µg/ml) and chloroquine (100 µM) were added to the growth media supplemented with cycloheximide. Cells treated in this manner were lysed with RIPA buffer and protein stability was monitored by Western blotting.

Immunofluorescence and Confocal Microscopy For subcellular localization studies, MDCK cells stably expressing WT or mutant variants were seeded at high density onto glass coverslips in a 24-well plate. Twenty-four hours later, cells were washed twice with ice-cold PBS, fixed for 15-20 min in 4% PFA in PBS, and then permeabilized using 0.5% Triton X-100 in PBS for 15 min (permeabilized conditions); Cells were blocked with 5% goat serum in PBS for 1 h, followed by incubation with primary antibody [mouse anti-Na, K-ATPase (1:50); mouse anti-HA (1:200); rabbit anti-calreticulin (1:100)] for 1 h, and washing three times with 0.1% BSA in PBS. Finally, the coverslips were incubated for 1 h with the appropriate secondary antibody [goat anti-mouse-Cy3 (1:1000); goat anti-rabbit-Cy3 (1:1000); Alexa-fluor 647 conjugated goat anti-mouse (1:1000)], and washed three times with PBS. All the incubations were done at room temperature, and the antibodies were all diluted in blocking solution. To detect cell surface expression of the exofacial HA epitope engineered in VANGL proteins (nonpermeabilized conditions), MDCK cells expressing HA-tagged WT, R259L and S464N variants were incubated with the mouse anti-HA (1:200) in DMEM containing 2% nonfat

118 milk for 2 hrs at 37ºC. After several washes with PBS, cells were fixed for 15 min with 4% PFA in PBS, and incubated with goat anti-mouse-Cy3 antibody (1:1000) for 1 h. For immunofluorescence, coverslips were rinsed once in water and mounted using Permafluor Aqueous Mounting Medium (Thermo Scientific, Fremont, CA). Confocal microscopy was performed using a Zeiss LSM5 Pascal laser scanning confocal microscope. All image analyses were performed using the LSM5 Image software. To maximize image quality, a Median filter 3x3 was applied to the images using Image-Pro software. To remove out-of-focus fluorescence, the Z stack sections were also deconvoluted using 10 iterations with AutoQuant software.

119

4.4 Results

Expression and Subcellular Localization of WT, R259L and S464N variants in MDCK cells The Lp-associated D255E, R259L and S464N mutations affect highly conserved residues in the large intracellular cytoplasmic segment of VANGL proteins (Figure 1A and B). In addition, the R259L mutation maps close to the mouse D255E (Lpm1Jus allele) 23, and the V239I de novo loss-of-function VANGL1 mutation detected in a human case of familial NTDs 120, suggesting that this protein segment is critical for a yet to be defined aspect of VANGL proteins structure and function. To investigate the effect of the R259L and S464N mutations on the biochemical properties of VANGL proteins, the two variants were expressed as GFP fusion proteins in MDCK cells. This cell type was chosen, since it is derived from a lineage that normally expresses VANGL proteins in vivo during kidney tubulogenesis 19 and is therefore likely to possess the cellular machinery required for proper expression, maturation, membrane insertion and subcellular targeting of the protein.

120

A HA139

c-Myc R259L NH2- GFP S464N COOH B

D255E R259L S464N

VANGL2 VVRSTDGASRFY TDGASRFYNVGH QWTLVSEEPVTN VANGL1 VVRSTDGESRFY TDGESRFYSLGH QWRLVSDEAVTN mVangl2 VVRSTDGASRFY TDGASRFYNVGH QWTLVSEEPVTN mVangl1 VVRSTDGESRFY TDGESRFYSLGH QWRLISEEAVTN rVangl2 VVRSTDGASRFY TDGASRFYNVGH QWTLVSEEPVTN xVangl2 VVRSTDGASRFY TDGASRFYNIGH QWTLVSEEPVTN zVangl2 VVRTTDGASRFY TDGASRFYNVGH QWTLVSEEPVTA dStbm/Vang IIRSPDGVSRSY PDGVSRSYMLGQ SWSLICDEIVSR ceStbm IVRDPDGEMHTL PDGEMHTLNIGA KWSVICDEAVSS

4. 1: Schematic representation of the GFP-VANGL1 protein. Schematic representation of the GFP- VANGL1 protein. (A) The positions of the hemagglutinin (HA) tag inserted at position 139 (predicted TM1-TM2 connecting loop), the c-Myc tag inserted in the amino-terminal segment, the GFP fused at the N-terminus and the R259L and S464N mutations in the intracellular C-terminal region are shown. (B) Sequence conservation across species of the R259 and S464 amino acids mutated in Lp mice. Species analysed; human (VANGL2 and VANGL1), mouse (mVangl2 and mVangl1), rat (rVangl2), xenopus (xVangl2), zebrafish (zVangl2), drosophila (dStbm/Vang) and caenorhabditis elegans (ceStbm).

121

Several transfected MDCK cell clones stably expressing either WT, R259L and S464N variants were initially identified by GFP fluorescence and recombinant protein expression was confirmed by Western blot analysis (Figure 2) of total cell extracts and crude membrane fractions from independent MDCK clones. In these cells, the protein was expressed as a specific HA- immunoreactive species of ~80 kDa present in total cell extracts and greatly enriched in crude membrane extracts from individual MDCK clones tested, suggesting that the R259L and S464N mutations do not impair membrane association of the protein. However, in multiple independent transfected clones analyzed, we noted a lower overall expression level of the R259L and S464N variants compared to WT (Figure 2; data not shown).

122

clone 1 clone 2

CE C M CE C M kDa 83 WT

83 R259L

83 S464N

4. 2: Expression of WT, R259L and S464N VANGL1 variants in the membrane fraction of transfected MDCK cells. Independent MDCK cell clones stably expressing either wild type (WT), R259L and 464N mutant variants were isolated. Total cell extracts (CE), membrane-enriched fractions (M) and soluble cytoplasmic extracts (C) were prepared from these cells and equal amounts of protein (50 μg) were analyzed by Western blotting analysis with the mouse monoclonal antibody 9E10 directed against the c-Myc tag inserted in-frame near the amino terminus of the recombinant proteins.

123

A possible effect of the R259L and S464N mutations on subcellular distribution and plasma membrane (PM) targeting was next investigated by immunofluorescence and confocal microscopy (Figure 3). MDCK transfectants expressing the WT protein displayed a strong PM localization (Figure 3; top panels), producing a typical mesh-like signal in confluent monolayers, and showing colocalization with Na, K-ATPase, a specific marker for basolateral membrane of MDCK cells (Figure 3; top center panel). By contrast, R259L and S464N signals did not appear concentrated at the PM, but rather showed an intracellular punctate pattern suggestive of localization to an intracellular endomembrane compartment (Figure 3; middle and bottom panels).

124

GFP Na,K-ATPase merge

R259L

S464N

4. 3: Cellular localization of WT, R259L and S464N VANGL1 variants in transfected MDCK cells. Transfected MDCK cells stably expressing GFP- VANGL1 -WT, R259L and S464N (green) were grown to confluence, fixed, permeabilized and stained for Na, K-ATPase followed by Cy3- conjugated secondary antibody (red), and examined by confocal microscopy. The merge images show that while the majority of WT signal is associated to the plasma membrane and overlaps with Na, K-ATPase (yellow), the R259L and S464N staining is mostly intracellular. Images are representative of at least 3 independent experiments of each type.

125

In addition, R259L and S464N did not colocalize with the PM marker Na, K-ATPase. The seemingly distinct subcellular localization of the WT and Lp-associated variant proteins was confirmed by examination of cell sections (Z stacks) (Figure 4). Indeed, while the WT protein was expressed predominantly at the basal and lateral membranes of polarized MDCK cells (Figure 4 – left panels), colocalizing with Na, K-ATPase (Figure 3 – left middle panel), R259L and S464N were not present at the basolateral membrane, but were found in an intracellular compartment negative for Na, K-ATPase (Figure 4; center and right panels). Together, these results strongly suggest that the pathogenic, Lp-associated mutations, R259L and S464N mutations interfere with normal subcellular targeting of the protein.

126

WT R259L S464N

GFP-Vangl1

Na,K-ATPase

merge

4. 4: Subcellular localization of the WT, R259L and S464N VANGL1 variants in stably transfected MDCK cells; Z-line image analysis. Transfected MDCK cells stably expressing either GFP- VANGL1 WT, R259L, S464N mutant variants were processed as described in the legend to Figure 3. Cells were analyzed by confocal microscopy to visualize the expression of GFP-Vangl1 (green) and Na, K-ATPase (red). Representative X-Z sections are shown. The merge images show that while GFP-Vangl1 WT colocalizes with Na, K-ATPase at the basolateral membrane (yellow) of the polarized cells, while the R259L and S464N mutants are absent at that site, and are mainly localized to an intracellular compartment.

127

Surface expression of WT, R259L and S464N variants in MDCK cells We previously used immunofluorescence with an exofacial HA epitope (inserted at position 139 in the extracellular loop delineated by TM1-TM2 segments; see Figure 1A) to demonstrate cell surface expression of WT VANGL1 protein 23. To further investigate a possible effect of the R259L and S464N mutations on PM targeting, we similarly monitored cell surface expression of these two variants in confluent MDCK cells. The polarity of the HA tag with respect to PM was determined by immunofluorescence with an anti-HA antibody used in intact cells (extracellular) and in cells permeabilized with Triton X-100 (intracellular) (Figure 5) 23. A strong fluorescence signal was detected in both permeabilized and nonpermeabilized MDCK clones expressing the WT protein (Figure 5A; left panels); this signal was absent from control untransfected MDCK cells similarly exposed to the anti-HA antibody (data not shown). By contrast, MDCK clones expressing the R259L and S464N variants displayed bright intracellular fluorescence under permeabilization conditions (Figure 5A; top center/right panels), but showed no fluorescence under nonpermeabilized conditions (Figure 5A; bottom center/right panels), suggesting that R259L and S464N are not expressed at the PM. The presence/absence of immunoreactive HA tag at the surface of cells expressing WT, R259L and S464N proteins was also investigated by an ELISA based assay 100. For this, corresponding MDCK transfectants were incubated with anti-HA antibody with or without prior permeabilization. The amount of bound anti-HA antibody was quantified using a secondary antibody coupled to horseradish peroxidase. Results show that while 78 ± 6% of total immunoreactive WT protein was found at the cell surface (Figure 5B), only a modest fraction of R259L (35% ± 6%) and S464N (29% ± 4%) was detected at that site. Together, results in Figures 3-5 show that the two Lp-associated mutations, R259L and S464N, disrupt normal subcellular localization and interfere with normal PM targeting.

128

A WT R259L S464N

Permeabilized

Non permeabilized

B % expression % surfaceof cell

WT R259L S464N

4. 5: Analysis of plasma membrane targeting and surface expression of WT, R259L and S464N VANGL1 variants in MDCK cells. (A) MDCK cells stably expressing HA-tagged GFP- VANGL1 WT, R259L and S464N were analyzed by immunofluorescence to detect surface expression of an HA antigenic epitope inserted at position 139 in the predicted TM1-TM2 connecting loop. Immunofluorescence was carried out on either intact cells (non-permeabilized) or in cells pre-treated with 0.1% Triton-X-100 (permeabilized), using a mouse anti-HA antibody. Cells were incubated with Cy3-conjugated secondary antibody and images were acquired by confocal microscopy. Images are representative of at least five independent experiments of each type. (B) In other experiments, the same cells exposed to the anti-HA antibody were incubated with an HRP-coupled secondary anti-mouse antibody and the amount of the primary antibody bound was determined for both conditions (permeabilized and non-permeabilized) by a colorimetric reaction using an HRP substrate quantitated by spectrometry. The amount of HA-tagged WT, R259L and S464N proteins expressed at the cell surface (detected in nonpermeabilized cells) is shown as a fraction (%) of total protein expression (measured in permeabilized cells).

129

Stability of the R259L and S464N variants in MDCK cells The overall lower level of the R259L and S464N variants detected in transfected MDCK cells (compared to WT; Figure 2 and data not shown), together with the noted defect in subcellular targeting of the variant prompted us to analyze a possible effect of the mutations on protein half-life and stability as measured by pulse-chase experiments. MDCK clones expressing WT, R259L and S464N variants were labeled with 35S-methionine/cysteine for 60 min (pulse) and the fate of the radiolabeled proteins was followed for 8 hrs by immunoprecipitation using extracts from pulsed cells grown in isotope-free medium (chase). Representative autoradiograms and quantification from 4 independent experiments are shown in Figure 6A and B, respectively. In these experiments, the half-life of WT protein in MDCK cells was estimated at greater than 20 hours, while the half-life of R259L and S464N were found to be shorter than ~3 hours. Results from the pulse-chase experiments strongly suggest that both R259L and S464N mutations cause a strong decrease in protein stability.

130

A 0 2 4 8 hours

R259L

S464N

Remaining(%) fraction

Chase (hours)

4. 6: Pulse-chase studies of GFP- VANGL1 WT, R259L and S464N proteins expressed in MDCK cells. (A) MDCK cells stably expressing GFP- VANGL1 WT, R259L or S464N were metabolically labeled by a 60-minute pulse of 35S-Met/Cys. The cells were transferred to radioactivity-free medium, followed by different chase periods. Cell lysates were prepared and subjected to immunoprecipitation followed by gel electrophoresis and autoradiography. To obtain comparable signal intensity at T0, the R259L and S464N blots were exposed for longer times. (B) Estimation of the rate of disappearance of WT, R259L and S464N labeled proteins. The amount of radiolabeled WT, R259L and S464N proteins was quantified (pixel density of each band) from scanning the autoradiograms and expressed as a fraction (%) of the sample at T0 which is set at 100%. The graph represents the mean of four experiments ± SE.

131

Retention of the R259L and S464N variants in the endoplasmic reticulum The reduced half-life, the absence of PM targeting and the localization of the R259L and S464N variants to a subcellular endomembrane compartment suggested that they may have a defect in maturation associated with retention of the protein in the endoplasmic reticulum (ER). This possibility was tested by immunofluorescence using an antibody directed against the ER marker calreticulin. Results show that while the WT protein displayed no colocalization with the intracellular punctate ER staining produced by calreticulin (Figure 7 - top panels), there was significant colocalization of R259L and S464N and calreticulin, suggesting that both variants are present in the ER endomembrane compartment (Figure 7 – middle and bottom panels).

132

GFP-Vangl1 calreticulin merge

R259L

S464N

4. 7: Determination of subcellular localization of R259L and S464N variants by double immunofluorescence. MDCK cells stably expressing WT, R259L or S464N were fixed and immunostained for the endoplasmic reticulum marker calreticulin (red). As opposed to GFP- VANGL1 WT which is expressed at the plasma membrane, the R259L and S464N mutants show intracellular staining that significantly overlaps with calreticulin. Insets are magnifications of the indicated areas.

133

Retention of inappropriately folded proteins in the ER, including membrane proteins, ultimately leads to degradation via the proteasome 123. To analyze the behavior of the R259L or S464N variants with respect to proteasome-mediated degradation, MDCK clones expressing the corresponding proteins were exposed to cycloheximide (CHX) either in the presence or absence of the proteasome inhibitor MG132 (Figure 8). Cell extracts were prepared and analyzed by immunoblotting (Figure 8A), and the signal was quantified (Figure 8B). CHX treatment of WT, R259L and S464N expressing cells had a differential effect on the level of protein remaining after 4 h of treatment, with a ~80% decrease in R259L and S464N variants, compared to only a small reduction in the amount of WT protein. Addition of the proteasome inhibitor MG132 to cells treated with CHX resulted in a dramatic increase in the amount of detectable R259L and S464N variants (~70% of untreated controls) in these cell extracts. MG132/CHX treatment also resulted in a modest increase of detectable WT protein (when compared to CHX treatment alone). Furthermore, the addition of chloroquine, a lysosome inhibitor, did not prevent the rapid degradation of both R259L and S464N (Figure 8). Together, these results suggest that R259L and S464N proteins are retained in the ER, where they are targeted for degradation in a proteasome- dependent manner.

134

WT R259L S464N

CHX - + + + - + + + - + + + MG132 - - + - - - + - - - + - Chloroquine - - - + - - - + - - - + Vangl1

tubulin

100

90 80 ns ns 70 60 NT 50 CHX 40 CHX+MG132 *** 30 *** CHX+CQ **

20 ** Remaining(%) fraction 10 0 WT R259L S464N

4. 8: Cellular Degradation of WT, R259L and S464N VANGL1 variants in MDCK cells. (A) Stably transfected MDCK cells expressing either GFP- VANGL1 WT, R259L or S464N were grown to confluence and treated with cycloheximide (20 μg/ml) for 4 h, either in the absence or presence of the proteasome inhibitor MG132 (5 μg/ml) or the lysosome inhibitor chloroquine (100 μM). Cell lysates (50 μg) were subjected to gel electrophoresis and Western blot analysis using the anti-c-Myc antibody. Tubulin was used as an internal loading control and was detected using a mouse monoclonal antiserum. The rapid degradation of R259L or S464N variants was prevented by MG132, but not by chloroquine, indicating that the R259L and S464N proteins are mainly degraded by the proteasome. To obtain comparable signal intensity, the R259L and S464N blots were exposed for longer times than those containing WT protein. (B) The effect of MG132 and chloroquine on the stability of WT, R259L and S464N mutant proteins after 4 hrs of treatment was quantitated by densitometry scanning as described in the legend of Figure 6. Each value represents mean ± SE of 4 independent experiments. Statistical significance of inter-group difference was determined by Student’s t-test, where ** and *** were found to be significant (P<0.01) and (P<0.001) respectively; NS, not significant.

135

Coexpression and cellular localization of Vangl WT and mutant proteins in MDCK cells For the D255E and S464N variants, heterozygote mice exhibit a mild looped tail phenotype, whereas homozygotes exhibit both looped tail and craniorachischisis 16,17. This codominance phenotype is the same for both D255E and S464N variants, while the phenotype of R259L appears somewhat less severe 16,17,87. Codominance could be explained by (a) a partial or complete loss-of-function of the mutant gene/protein in a gene dosage pathway (haploid insufficiency) or (b) a gain of function (dominant negative effect) in which the mutant protein interferes with the normal maturation/targeting of the WT protein produced in the same cell. Arguments favoring haploid insufficiency 18 and dominant negative effects 67 to explain the codominance inheritance of Lp have been published. To study the effect of the three known Lp- associated variants D255E, R259L and S464N on targeting of the WT and protein and vice versa, a Cherry-WT protein was stably coexpressed in independent MDCK cell clones expressing either WT, D255E, R259L and S464N (all fused to GFP). Several stably double-transfected MDCK clones were identified by fluorescence, and Cherry-VANGL1 and GFP-VANGL1 protein coexpression was confirmed by immunoblotting (Supplementary Figure S1) and fluorescence activated cell sorting (data not shown). These experiments also indicated similar levels of coexpression of the Cherry and GFP-tagged proteins in the cell clones analyzed by microscopy. The cellular localization of Cherry- and various GFP-tagged proteins (WT, D255E, R259L and S464N) in MDCK cells was analyzed by confocal microscopy (Figure 9 and Supplementary Figure 2). As expected, the coexpressed Cherry- and GFP-tagged WT proteins were both found at the PM, where they colocalized with the Na, K-ATPase PM marker. Coexpression in the same cells of Cherry-WT with either D255E, R259L or S464N (GFP-tagged) did not affect the PM staining of the Cherry-WT protein. In each case, the GFP-tagged D255E, R259L and S464N proteins displayed mostly an ER-like staining, while the Cherry-WT colocalized with Na, K-ATPase marker, suggesting that expression of the mutant variant did not affect maturation and PM targeting of the WT protein. Expression of the various mutant variants of VANGL1 also did not affect the distribution and subcellular localization of WT VANGL1 (fused to Cherry), as determined by Z stack analysis (Supplementary Figure 2) and surface biotinylation (data not shown). These results support the contention that D255E, R259L and S464N mutations represent a loss-of-function of the mutated protein in a gene dosage dependent pathway (haploid insufficiency), and argue against a dominant negative effect of the three mutations.

136

WT D255E R259L S464N

GFP-Vangl1

Cherry-Vangl1 (WT)

merge

Na,K-ATPase (647 nm)

merge all

4. 9: Coexpression and cellular localization of WT and Lp-associated variants (D255E, R259L and S464N) in MDCK cells. Stably transfected MDCK cells coexpressing Cherry- VANGL1 WT (red) with either GFP-tagged (green) WT, or R259L or S464N or D255E were grown to confluency, fixed, permeabilized and stained for Na, K-ATPase followed by Alexa-Fluoro 647-conjugated secondary antibody (violet), and examined by confocal microscopy. The merged images show that while the majority of GFP- VANGL1 WT signal is associated to the plasma membrane and overlaps with Cherry- VANGL1 WT (yellow), the GFP-tagged D255E, R259L and S464N staining is mostly intracellular, and does not influence the staining for Cherry- VANGL1 which remains associated to the plasma membrane. Images are representative of at least three independent experiments of each type with two different clones.

137

4.5 Discussion

VANGL proteins are integral membrane proteins that play a critical role in two pathways that control cell and tissue morphogenesis during development: planar cell polarity (PCP) and convergent extension movements (CE) 8,22. Although the precise mechanism of action of VANGL proteins in PCP and CE in vertebrates remains elusive, independent mutations in Vangl genes have been associated with neural tube defects in mice and humans 16,17,75,76,120. In mice, homozygote mutations in Vangl2 (Lp) 16,17and double heterozygosity for loss-of-function mutations at Vangl1 and Vangl2 18, cause craniorachischisis. In the mouse, there are three independent Lp alleles at Vangl2, the spontaneously arising S464N (Lp)17and two chemically induced D255E (Lpm1Jus) 17 and R259L (Lpm2Jus); 87. Herein, we have studied the molecular basis of the loss-of-function variants S464N and R259L, after expression of these proteins in polarized MDCK kidney epithelial cells. In our study, human VANGL1 was used as a molecular backbone. Although the pathogenic R259L and S464N mutations emerged in VANGL2, much data indicate that VANGL1 and VANGL2 proteins are functionally equivalent 18,25,93. Thus, the choice of the molecular backbone should not impact the biochemical characterization of pathogenic mutations, as we reported previously 23.

We found that both mutations impaired plasma membrane targeting and surface expression of the protein (absence of colocalization with Na, K-ATPase), with both mutant variants being retained in the calreticulin positive endoplasmic reticulum. The S464N and R259L proteins also show a dramatic reduction in half-life (from 22 hrs for the WT protein to 3 hrs for both mutants) and stability in MDCK, with increased degradation by a proteasome-dependent and MG132- sensitive fashion. These characteristics are similar to those we previously reported for the D255E Lp-allele 23 and suggest a common mechanism for the loss-of-function of the three pathogenic Lp- associated variants, namely absence of expression of the protein at the plasma membrane. The mutations appear to cause a maturation defect with retention of the mutant variants in the ER. Our findings are in agreement with the previous demonstration that PCP signaling requires membrane targeting and asymmetric distribution of complexes containing VANGL proteins.

138

Our combined analysis of the Lp-associated Vangl2 variants D255E, S464N and R259L also point to a critical structural/functional role of the large predicted intra-cytoplasmic domain of the protein, as all three mutations map within this segment 22. Likewise, the three VANGL1 variants (V239I, R274Q, and M328T) we detected in the heterozygote state in sporadic and familial cases of neural tube defects in humans also map to the cytoplasmic domain of the protein 120. Finally, two mutations in this portion of the VANGL2 protein (R353C and F437S) were recently identified in foetuses bearing severe neural tube defects 75. Although the clustering of pathogenic mutations in this part of the protein clearly suggests a critical role for this segment, the variable nature of the mutations being either subtle (V239I, D255E, S464N) or non-conservative (R259L, R353C), together with the absence of any obvious structural/functional motifs or signatures being interrupted by these variants does not suggest an obvious mechanistic basis for the mutations effects. The C-terminal domain is highly conserved between the different members of the Vangl/Stbm family (e.g. 72% sequence identity and 85% sequence conservation between mouse and fish Vang1), and several residues affected by mutations are either invariant (V239, D255, F437) or highly conserved (R259, R274, M328, R353, S464) in the VANGL protein family. One of the mutations, R274Q alters a key residue in a D-box consensus sequence of ubiquitination (RXXLXXXXE/N/D); however, this motif is not preserved beyond the mouse and human VANGL1 sequence, and its relevance to a conserved mechanism of action of VANGL proteins remains unknown. On the other hand, VANGL proteins have recently been suggested to interact with SEC24B, a cargo sorting member of COPII vesicle 64. COPII complex is part of the ER to Golgi transport, implicated in selectively trafficking VANGL2. In addition, studies in mice with either inactive (ENU-induced) or knock-out alleles at sec24b display craniorachischisis, CE and PCP defects 65. It was also additionally demonstrated that the two Lp-associated mutations D255E and S464N fail to sort into COPII vesicles and remain trapped in the ER 64. These observations are compatible with the biochemical properties, including lack of PM association, reduced half-life and retention in the ER of the three Lp mutations we present in the current study. Although retention of inappropriately folded proteins in the ER may induce the UPR (unfolded protein response) in which the expression of ER-resident chaperones (calnexin, BIP, PDI) is unregulated, we note that expression of the R259L and S464N variants fail to induce the UPR in MDCK cells (data not shown).

139

In vivo, two Lp mutations, D255E and S464N, are inherited in a co-dominant fashion with heterozygotes showing the mild looped-tail phenotype, while Lp/Lp homozygotes show the very severe craniorachischisis and die in utero. The codominance characteristics of Lp can be explained by haploid insufficiency caused by a partial loss-of-function in a gene dosage dependent pathway; alternatively, it may result from a negative dominance caused by a gain of function of the mutant variant which alters the activity of the normal protein encoded by the intact wild type allele present in the heterozygotes. The observation that two allelic variants (D255E, S464N) at Lp a) display exactly the same phenotype in the heterozygote state in vivo, and b) are associated with impaired maturation/processing and absence of membrane expression in vitro, strongly suggest that these mutations cause a loss-of-function in a gene dosage dependent pathway. Nevertheless, others have recently suggested that Lp alleles behave as dominant negative 67, with the mutant variant altering normal maturing, processing, targeting or function of the a coexpressed wild type protein. Results shown in Figure 9 show that coexpression of individual Lp mutants with the wild type protein in MDCK cells (mimicking the situation of Lp/+ heterozygotes) fails to influence the subcellular localization and PM targeting of the wild type protein. These results do not support a dominant negative effect of Lp mutations and favour a loss-of-function mechanism. However, we cannot formally exclude the possibility that mutant VANGL variants might sequester (or liberate, and hence mistarget) a different but essential interaction partner present in limiting quantities. On the other hand, the R259L variant shows a low frequency of “looped tail” animals in heterozygotes, while R259L homozygotes display very severe craniorachischisis 87. The reduced frequency of “looped tail” animals in R259L /+ heterozygotes may suggest that R259L is less detrimental to protein function in vivo (as suggested by reduced but not abrogated surface expression of this mutant). It may also be caused by possible genetic background effects (in the original ENU mutagenized mouse stock) that modulate penetrance or expressivity of an otherwise loss-of- function variant.

140

In conclusion, our results suggest that VANGL proteins are part of a very sensitive gene dosage dependent pathway during the development of the neural tube and that Lp alleles (D255E, R259L, S464N) behave as loss-of-function mutations. Furthermore, we show that both the S464N and the novel Lp alleles R259L are unable to reach the plasma membrane, potentially affecting the delicate balance in the PCP pathway and leading to NTDs. Taken together, our biochemical analyses, of mutant Lp variants in MDCK cells provide an experimental framework for the future analysis of VANGL1/VANGL2 mutations detected in familial and sporadic cases of NTDs.

4.6 Acknowledgments

Image acquisition, data analysis, and image processing were done on equipment and with the assistance of the McGill Life Science Complex Imaging Facility which is founded by the Canadian Foundation for Innovation. We also thank Nassima Fodil-Cornu, Ajitha Thanabalasuriar and Irena Radovanovic for their help and technical advice.

141

4.7 Supplementary Figures

Cherry-Vangl1 WT

GFP-Vangl1 Cherry-Vangl1

Actin

1 2 3 4 5 6

4. S 1: Expression of Cherry-VANGL1 WT protein in MDCK cells stably expressing GFP- VANGL1 WT, D255E, R259L and S464N proteins. Total cell extracts were prepared from doubly-transfected MDCK cell clones expressing Cherry- VANGL1 WT in presence of either GFP-VANGL1 WT or mutant variants (D255E, R259L and S464N). Equal amounts of protein (50 μg) were separated by electrophoresis using 7.5% SDS- polyacrylamide gels and analyzed by immunoblotting with the mouse monoclonal antibody 9E10 directed against the c-Myc tag inserted in-frame near the amino terminus of Cherry and GFP recombinant VANGL1 proteins. Both GFP- and Cherry-VANGL1 proteins can be distinguished on gel: the GFP-VANGL1 proteins (90.6kDa) containing an internal HA epitope migrate slower than the Cherry-VANGL1 protein (89.1kDa). Our results show that in doubly transfected MDCK cells, the level of expression of Cherry-VANGL1 is comparable to GFP-VANGL1 variants. Actin was used as an internal loading control. Lane 1 – cells expressing GFP-VANGL1 WT; Lane 2 – cells expressing Cherry-VANGL1 WT alone; Lanes 3-6 – cells coexpressing Cherry-VANGL1 WT with GFP-VANGL1 WT, D255E, R259L, and S464N.

142

Cherry-Vangl1 WT

GFP-WT GFP-D255E GFP- R259L GFP-S464N

A+B

B+C

4. S 2: Subcellular localization of the WT, D255E, R259L and S464N VANGL1 variants in stable doubly transfected MDCK cells coexpressing Cherry-VANGL1 WT; Z-line image analysis. Transfected MDCK cells stably expressing GFP-VANGL1 WT, D255E, R259L or S464N mutant variants (green, Panel A) and coexpressing Cherry-VANGL1 WT (red, Panel B) were grown to confluence, fixed, permeabilized and stained for Na, K-ATPase followed by Alexafluor 647- conjugated secondary antibody (purple, Panel C). Cells were analyzed by confocal microscopy to visualize the expression of GFP-VANGL1 variants (green), Cherry-VANGL1 WT (red) and Na, K-ATPase (purple). Representative x-z- sections are shown. The merged images show that Cherry- VANGL1 WT colocalizes with Na, K-ATPase at the basolateral membrane (pink) in all doubly transfected cells including both GFP-VANGL1 WT or GFP-VANGL1 mutant variants.

143

Chapter 5: Independent mutations at Arg181 and Arg274 of VANGL proteins that are associated with neural tube defects in humans decrease protein stability and impair membrane targeting

144

Preface to Chapter 5

Following the discovery that mutations in Vangl genes are causing the NTD in the loop- tail mouse, cohorts of human patient were also sequenced. Since 2007, five publications have identified 20 mutations in sporadic and familial cases of NTDs in both VANGL1 and VANGL2. These mutations are non-conservative substitutions located residues that are evolutionary conserved and absent in ethnically matched control patients. Chapter 5 concentrates on R181Q and R274Q, two mutations identified in a familial setting in the VANGL1 gene in an 8 and 19 year old Italian patients respectively. Interestingly, two additional patients were also found to be mutated at those same but homologous positions in the VANGL2 gene. Work done in chapter 5 utilized the series of biochemical tests established in chapters 3 and 4 (to monitor membrane targeting, subcellular distribution, half-life and stability) to also fully characterize the loss-of- function mechanism of these human mutations.

145

5.1 Abstract

In vertebrates, VANGL proteins play important roles during embryogenesis including establishing planar polarity and coordinating convergent extension movements. In mice, homozygosity for mutations in the Vangl1 and Vangl2 genes or combined heterozygosity for Vangl1/Vangl2 mutations causes the very severe neural tube defect (NTD) craniorachischisis. Recently, a number of patient-specific VANGL1 and VANGL2 protein mutations have been identified in familial and sporadic cases of mild and severe forms of NTDs. The biochemical nature of pathological effects in these mutations remains unknown. Of interest are two arginine residues, R181 and R274, that are highly conserved in VANGL protein homologues, and found to be independently mutated in VANGL1 (R181Q and R274Q) and VANGL2 (R177H and R270H) in human cases of NTDs. The cellular and biochemical properties of R181Q and R274Q were established in transfected MDCK kidney epithelial cells and compared to those of wild-type (WT) VANGL1. Compared to that of WT, these mutations displayed impaired targeting to the plasma membrane and were instead detected in an intracellular endomembrane compartment that was positive for the endoplasmic reticulum. R181Q and R274Q showed impaired stability with significant reductions in measured half-lives from >20h for WT protein to 9h and 5h, respectively. These mutations have a cellular and biochemical phenotype that is indistinguishable from that of Vangl mutations known to cause craniorachischisis in mice (Lp). These results strongly suggest that R181 and R274 play critical roles in VANGL protein function and that their mutations cause neural tube defects in humans.

146

5.2 Introduction

The neural tube is the primordial embryonic structure that gives rise to the brain and spinal cord. Failure of the neural tube to close properly results in neural tube defects (NTDs), a group of heterogeneous birth defects that are common in developed countries (0.5-2 per 1000 pregnancies) 7,8,137. Anencephaly, craniorachischisis, and myelomeningocele (spina bifida), represent the most severe forms of NTDs in which the neural tube remains partially or completely open 7. The etiology of NTDs is complex and involves the interplay between genetic and environmental factors 5Population clustering, family (increased risk in siblings and first degree relatives), and twin studies have established a genetic component to NTDs 7,91. On the other hand, epidemiological studies have pointed at a strong environmental component in the etiology of NTDs: while hypothermia, diabetes, psychological and emotional stress are associated with an increased risk of NTDs, perinatal folate supplementation has a strong protective effect 91. Genetic mutations that cause NTDs in humans have been difficult to identify due to the polygenic nature of this heterogeneous trait, and the incomplete penetrance of genetic effects. However, the parallel characterization of mouse NTD mutants has provided a rich source of candidate genes whose relevance to human NTDs can be established in case control studies 138

In mice, positional cloning 16,92and functional complementation studies 15 have shown that mutations in Vangl2 cause the NTD phenotype in the loop-tail (Lp) mutant. Homozygous Lp/Lp embryos die in utero and display craniorachischisis, a very severe NTD in which the neural tube remains completely open from hindbrain to the most caudal portion of the embryo. Individual loss- of-function Vangl2 mutations have been identified in independent Lp mutant stocks, Lp (S464N), Lpm1Jus(D255E), and Lpm2Jus(R259L). 16,17,87 In vertebrates, there are two Vangl genes, Vangl1 and Vangl2. The two encoded mRNAs and proteins are expressed during embryogenesis and show complementary expression patterns in the developing neural tube. Vangl2 is expressed abundantly and broadly in the dorsal and ventral portions of the neural tube, prior to, during and after closure (E7.5-E11.5), while Vangl1 expression is mostly restricted to the ventral portion of the neural tube and to the notochord 16,18,19. In addition, homozygosity for a loss-of-function Vangl1 mutation 18 and double heterozygosity for Vangl1gt/+/Vangl2Lp/+ 18both give rise to craniorachischisis, suggesting a genetic interaction between both genes, and further implicating both Vangl1 and

147

Vangl2 as risk factors for NTDs. Finally, recent evidence suggests that Vangl1 and Vangl2 proteins may physically interact in certain cell types 68.

Vangl proteins play a critical role in different aspects of cell and tissue patterning that has been conserved in evolution 5. In flies, the Vangl homologue (Van Gogh; Stbm) provides positional cues and is required for polarization of certain epithelial layers and associated appendages 95,105. This process is regulated by a set of planar cell polarity (PCP) genes that include Vang, Frizzled (Fz), Dishevelled (Dvl), Prickle (Pk), Flamingo (Fmi), and Diego (Dgo) 43. At the molecular level, PCP signaling in flies is associated with asymmetric positioning of membrane-bound Dvl/Fz and Vang/Pk complexes on opposite sides of adjacent cells, a distribution believed to propagate the polarity signal 113. In vertebrates, the role of Vangl proteins in PCP signaling preserved (e.g., polarization of ciliary bundles on sensory hair cells), and Vangl and other core PCP proteins (Dvl, Fz and Pk) are required for convergent extension (CE) movements. This process, in which a group of cells intercalates in one direction and elongates in the perpendicular direction, fundamentally regulates many aspects of tissue patterning during embryogenesis, including neural tube closure, cardiogenesis, and the formation of various epithelial structures 54. Mice bearing homozygous mutations in Vangl2 or compound heterozygous for Vang2Lp/+ and other PCP gene mutations display multiple developmental defects including craniorachischisis.5

VANGL1 and VANGL2 proteins are integral membrane proteins (521-526 amino acids) composed of four TM domains and a long C-terminal intracellular domain. The plasma membrane association of VANGL proteins has been verified in primary tissues in vivo and in transfected cultured cells in vitro, 19,23,24 and Lp-associated pathological mutations in mice cause a loss of VANGL membrane targeting. 19,23,24 The intracellular cytoplasmic domain of VANGL proteins contains a PDZ binding motif, and interacts with other PCP proteins such as DVL and SEC24B, and Lp-associated mutations (D255E and S464N) in the C-terminal domain impair these interactions 22,65. The export of Vangl proteins from the trans-Golgi network (TGN) to the PM was also found to be mediated by the interaction between the Phe in the YXXF(280-283) motif and it’s binding with AP-1, a clathrin adaptor complex. 84 Additionally, VANGL2 has been shown to be phosphorylated in response to WNT ligands, suggesting a possible regulation of VANGL2- dependent PCP signaling by morphogenic gradients 85,86

148

Mutations in VANGL1 and VANGL2, all present in a heterozygous state, have been recently identified in human patients with NTDs. VANGL1 mutations have been identified in a cohort of patients with sporadic (M328T) and familial (V239I and R274Q) cases of myelomeningocele (spina bifida). 120 independently, mutations in human VANGL2 (S84F, R353C and F437S) were reported in stillborn fetuses with anencephaly and holoprosencephaly. 75 Complementation studies in zebrafish trilobite mutants (Vangl homologues), and protein-protein interaction assays have established that disease associated VANGL1 mutations V239I and M328T and VANGL2 mutation F347S are functionally inactive.75,120 Screening of additional cohorts of nonsyndromic sporadic cases of NTDs (cranial, open and closed spinal dysraphisms) in patients from different ethnic origins identified 14 additional independent patient-specific VANGL1 and VANGL2 mutations.76- 78 In these studies, putative pathogenic mutations were identified on the basis of (a) the absence of the mutation in ethnically matched controls, (b) evolutionary conservation of the affected residue, and c) the nonconservative nature of the mutation. The possible effect of these mutations on VANGL protein function has yet to be determined. Of notable interest are two Arginine residues at position 181 (R181) in the first intracellular loop linking TM2 and TM3 and at position 274 (R274) in the C-terminal intracellular half of the protein (VANGL1 numbering). R181 is independently mutated (at the homologous position) to histidine in VANGL2 (VANGL2R177H) and to glutamine in VANGL1 (VANGL1R181Q) in two unrelated patients with NTDs. Likewise, R274 is independently mutated (at the homologous position) to histidine in VANGL2 (VANGL2R270H) and to glutamine in VANGL1 (VANGL1R274Q) in two unrelated patients with NTDs. R181 and R274 are extremely conserved in the VANGL protein family from flies to humans. Their independent mutations to nonconserved amino acids in independent patients with NTDs suggest not only that these mutations are pathological but also that R181 and R274 play a critical role in the function of the protein.

In this study, we have studied the function of R181 and R274 by characterizing the effects of mutations at these residues on the biochemical properties of the proteins.

149

5.3 Experimental Procedures

Material and Antibodies Restriction enzymes were obtained from New England Biolabs (Ipswich, MA), and Taq DNA polymerase was purchased from Invitrogen (Carlsbad, CA). The antibodies against the influenza hemagglutinin epitope (HA.11) (mouse monoclonal) and the c-Myc epitope (9e10) (mouse monoclonal) were purchased from Covance (Berkeley, CA). The antibody against Na, K-ATPase (α1 subunit) (mouse monoclonal) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody against calreticulin (rabbit polyclonal) was purchased from Affinity BioReagents (Golden, CO). Cy3-conjugated goat anti-mouse and anti-rabbit antibodies and the peroxidase-coupled goat anti-mouse antibody were from Jackson ImmunoResearch Laboratories (West Grove, PA). The reagents cycloheximide and MG132 was purchased from Sigma-Aldrich (St-Louis, MO) and from Calbiochem (San Diego, CA) respectively. Geneticin (G418) and streptomycin were obtained from Invitrogen (Carlsbad, CA).

Plasmids and Constructs Polymerase chain reaction overlap extension mutagenesis was used to introduce the R181Q, R274Q, and L202F mutations into the human VANGL1 cDNA, as previously described. 23 All four constructs (VANGL1WT, VANGL1R181Q, VANGL1R274Q, and VANGL1L202F) contained a human c-Myc protein epitope tag (EQKLISEEDL) inserted at the N-terminus, a hemagglutinin (HA) epitope (YPYDVPDYA) inserted in the first extracellular loop at position 139 (used for cell surface detection and quantification of VANGL1 protein expression), and green fluorescent protein (GFP) (used to facilitate detection of positive stable clones) at the N-terminus by cloning the various VANGL1 cDNAs into a modified peGFP-C1 vector (Clontech, Mountain View, CA).

150 on live cells, and confirmed by Western blotting analysis of cell extracts. Total cell lysates were prepared in RIPA buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA and 1% Triton X-100 supplemented with protease inhibitors], left on ice for 30 min and spun down at 13000g for 10 min at 4ºC. 50 µg of protein from cell lysates were loaded onto 7.5% sodium dodecyl sulfate-polyacrylamide (SDS- PAGE) gels, followed by electroblotting and incubation with monoclonal anti-HA antibody (HA.11) used at 1:1000 dilution, followed by incubation with a secondary horseradish peroxidase conjugated goat anti- mouse antibody (1:5000), and visualization with enhanced chemiluminescence (SuperSignal West Femto Chemiluminescent Substrate Kit, Thermo Scientific Rockford, IL).

Quantification of Cell Surface Expression by an Enzyme Linked Immunosorbent Assay. Quantification of cell surface expression of HA-tagged VANGL1 (WT or mutant) proteins was performed using an enzyme-linked immunosorbent assay (ELISA) based assay described previously 23 in which MDCK cells stably transfected with HA-tagged VANGL1WT, VANGL1R181Q or VANGL1R274Q were grown to confluency in 24-well plates for 4 days. The cells were washed with PBS and incubated for 30 min in Ca2+-free DMEM, prior to incubation for 5 min with 10 mM EGTA. To quantify cell surface expression of HA-tagged VANGL1 proteins, cells were incubated with the mouse anti-HA antibody for 2 2+ h (1:200 in Ca -free DMEM) (37 °C, 5% CO2), washed with PBS, fixed for 15 min with 4% paraformaldehyde in PBS, and incubated for 1 h with the HRP-conjugated goat anti-mouse antibody (1:4000 in a 5% nonfat milk/PBS mixture). To quantify total protein expression of HA-tagged VANGL1 proteins, cells were fixed immediately after the EGTA treatment, permeabilized for 30 min with 0.1% Triton X-100 in PBS, blocked for 30 min in 5% nonfat milk/PBS mixture, incubated with the anti-HA antibody (1:200 in blocking solution) for 1 h, washed with PBS, and incubated with the HRP-conjugated goat anti-mouse antibody (1:4000 in blocking solution). Both cell surface expression and total cell expression were quantified colorimetrically using the HRP substrate [0.4 mg/mL O-phenylenediamine dihydrochloride (OPD) (Sigma-Aldrich)] according to the manufacturer’s instructions. The reaction was stopped via addition of 3 N HCl; absorbance readings (492 nm) were taken in an ELISA plate reader, and the background absorbance reading from the nonspecific binding of the primary antibody to vector- transfected cells was subtracted for each sample. Cell surface readings were normalized to the total HA- tagged GFP-VANGL1 value for each cell clone and are expressed as a percentage. Metabolic Labeling, Pulse-Chase Study and Immunoprecipitation.

151

MDCK cells stably expressing VANGL1 WT, VANGL1 R181Q or VANGL1 R274Q proteins were grown to confluency for 48 h in 60 mm plastic dishes. Cells were then incubated for 90 min at 37°C in methionine- and cysteine-free DMEM containing 10% dialyzed FBS (labeling medium). This was followed by a 60 min incubation at 37°C with 2 mL of labeling medium containing 100 μCi of [35S] methionine-cysteine (PerkinElmer, Boston, MA). Cells were washed twice with PBS and incubated for different periods of time (up to 16 h) in 2 mL of chase medium (standard DMEM containing 10% FBS, 15 µg/mL methionine and 15 µg/mL cysteine). Total cell lysates were prepared in 400 μL of RIPA buffer, and equal amounts of cell lysates were subjected to immunoprecipitation using a mouse anti-cMyc antibody 9E10 (2.5 µg) in RIPA buffer and incubated overnight at 4°C. The following day, proteins were incubated for 4 h at 4°C with protein A/G-Sepharose (GE Healthcare, Piscataway, NJ), followed by three washes in RIPA buffer, and eluted with 50µL of 3x sample buffer. Radiolabeled proteins were separated by electrophoresis on 7.5% SDS-PAGE gels. Gels were fixed for 30 min in 40% MeOH/10% acetic acid mixture, soaked for 15-30min in Amplify (GE Healthcare), dried, and exposed to film. Band intensity was quantified using ImageJ (National Institute Health, Bethesda, MD).

Determination of Relative Protein Stability. To determine the stability of wild-type and mutant VANGL1 protein mutations, MDCK cells stably expressing these proteins were grown to confluency 48 h, and treated for 6 h with cycloheximide (20 µg/mL) in the presence or absence of MG132 (5 μg/mL). Total cell lysates were prepared with RIPA buffer, and 50μg of protein was loaded on a 7.5% SDS-PAGE gel, followed by immunoblot analysis using the anti-HA antibody. Band intensity of Western blot gels were quantified using data from three independent experiments and quantitated using ImageJ, with β-actin used as an internal loading control.

Immunofluorescence and Confocal microscopy. MDCK cells stably expressing different VANGL1 proteins were examined for protein expression under both permeabilized (total cell expression) and nonpermeabilized conditions (cell surface expression), and at confluency on glass coverslips in a 12-well plate. Cells were washed twice with PBS, fixed for 15 min in 4% PFA in PBS, permeabilized with 0.5% Triton X-100 in PBS for 15 min, blocked in 5% goat serum in PBS for 1 h, incubated with the desired primary antibody [mouse anti-Na, K-ATPase (1:100), mouse anti-HA (1:200), rabbit anti-calreticulin (1:100)] for 1 h, and washed three times with 0.1% BSA in PBS. Finally, the coverslips were incubated for 1 h with the appropriate secondary antibody

152

[goat anti-mouse-Cy3 (1:1000) and goat anti-rabbit-Cy3 (1:1000), and washed three times with 0.1% BSA in PBS. All the incubations were conducted at room temperature, and the antibodies were all diluted in a blocking solution. To detect cell surface expression of the exofacial HA epitope engineered in VANGL1 proteins (nonpermeabilized conditions), MDCK cells expressing GFP- VANGL1 WT , R181Q, or R274Q were incubated with the mouse anti-HA antibody (1:200) in DMEM containing 2% nonfat milk for 2 h at 37ºC. After being washed several times with PBS, cells were fixed for 15 min with 4% PFA in PBS and incubated with the goat anti-mouse-Cy3 antibody (1:1000) for 1 h. For immunofluorescence, coverslips were rinsed once in water and mounted with Permafluor Aqueous Mounting Medium (Thermo Scientific, Fremont, CA). Confocal microscopy was performed using a Zeiss LSM5 Pascal laser scanning confocal microscope. All image analyses were performed using the LSM5 Image software.

153

5.4 Results

R181Q and R274Q mutations associated with human neural tube defects. More than 20 human VANGL1 and VANGL2 mutations have been found to be associated with human neural tube defects (NTDs). 8,75-78 The mechanistic basis for the loss of VANGL protein function in these human disease-associated mutations has not been studied. Two arginine residues at positions 181 (R181) and 274 (R274) (VANGL1 nomenclature) attracted our attention for several reasons. First, R181 and R274 are independently mutated (a) to glutamine in VANGL1 (R181Q and R274Q) in two unrelated NTDs patients and mutated (b) to histidine in VANGL2 (R177H and R270H) at the homologous position in two other unrelated NTDs patients. Second, both R181 and R274 are extremely conserved in the VANGL family, R274 being invariant and R181 being highly conserved, with both mapping to protein subdomains that also show a high degree of sequence conservation among family homologues and orthologs (Figure 1). Finally, the charged positive side chain of arginine is lost in both substitutions. VANGL1R274Q was identified in a familial case of a NTD, where the proband is a 19 year old female with open NTD with both the mother (who is also a carrier) and aunt suffering from vertebral schisis, a milder form of NTD. 120 R274Q maps to the C-terminal intracellular domain of VANGL2, and the invariant arginine forms part of a putative D-box ubiquitination motif (RXXLXXXD/E/X) 137 (Figure 1B and C). VANGL1R181Q was detected in an 8 year old Italian male with myelomeningocele; the boy’s mother also carries the R181Q mutation but is phenotypically normal, suggesting incomplete penetrance of the genetic effect (Figure 1A). 76 R181Q (and the corresponding R177H mutation in VANGL2) maps to the short 16 amino acid intracellular domain linking TM domains 2 and 3, which is the site of two additional disease-specific VANGL2 mutations (R173H and R186H) associated with NTDs in an Eastern European cohort. 78

Together, these results suggest that R181 and R274 play important and evolutionarily conserved structural or functional roles in the VANGL protein family and suggest that loss of either arginine (in independent NTDs patients) is detrimental to function, ultimately leading to NTDs. Therefore, we have studied the potential molecular defect in the R181Q and R274Q mutants following their reconstruction in a human VANGL1 backbone and expression in transfected cells.

154

A B

HA 139 R181Q R274Q

* *

R181Q GFP R274Q NH2

COOH

C pArg181Gln pArg274Gln

Human Vangl1 RADMPRVFVF GHLSIQRAALV Human Vangl2 KASLP RVFVL GHLSIQRVAVW Mouse Vangl1 RADVP RVFVF GHLSIQRAAVV Mouse Vangl2 KASLPRVFVL GHLSIQRVAVW Zebrafish Vangl1 RASLPGIAVF GQLSIQRAALV Zebrafish Vangl2 RSTLPRFFVF GHLSIQRAAVW Xenopus Vangl2 KAFFPRVFVF GHLSIQRVAVW Drosophila Stbm SATMPRIFLY GQLSIQRAAVW C.Elegans Stbm MADMPRLYFV GAGSIQRAATE

5. 1: Pedigree and schematic representation of human mutations (R181Q and R274Q) found in NTDs patients. (A) Familial pedigree showing the R181Q mutation in a patient with myelomeningocele, a mutation also present in the mother (*asterisk) but absent in the father and unaffected sister, a case with a positive familial history of myelomeningocele in a distant maternal ancestor. Familial pedigree showing the R274Q mutation in a patient with an open NTD, a mutation present in the mother (*asterisk) but absent in the father, a case with a history of familial NTDs (mother and aunt have mild NTD vertebral schisis) represented by the cross-hatched circles. (B) Schematic representation of the secondary structure of GFP-tagged h VANGL1 protein with the locations of mutants. The position of the HA tag located extracellularly at position 139 (used for surface expression and surface expression of VANGL1 protein) and of both mutants located intracellularly between TM1 and TM2. (C) Sequence conservation of mutations across species.

155

Expression and cellular localization of R181Q and R274Q VANGL1 mutations in stably transfected MDCK cells. Targeting of VANGL proteins to the plasma membrane is essential for biological function and PCP signaling. In mouse embryos, the VANGL2 protein is targeted to the basolateral membrane of several epithelial cells and tubular structures. Lp mouse mutant embryos with craniorachischisis show a reduced level of VANGL2 protein expression and loss of membrane targeting. 19 To investigate the effect of the R181Q and R274Q mutations on the biochemical properties of VANGL proteins, the two mutations were expressed as GFP fusion proteins in MDCK cells. This cell type was chosen because it is derived from a lineage (kidney tubules) that normally expresses VANGL proteins and is likely to possess the cellular machinery required for the proper expression, maturation, and subcellular targeting of the protein. We have previously shown that WT VANGL1 protein is targeted to the basolateral membrane of MDCK cells and that this membrane targeting is lost in Lp-associated Vangl2 mutations (D255E, R259L and S464N) that cause craniorachischisis in mice. 23,24

MDCK cells were transfected with plasmids (see Experimental Procedures) expressing either WT or mutant VANGL1 mutations R181Q and R274Q (as GFP and HA tagged fusions), followed by the isolation of single clones stably expressing each protein. These were initially selected by epifluorescence (GFP positive), followed by immunoblotting of crude membrane extracts (data not shown) with an anti-HA antibody directed against an exofacial HA epitope tag inserted in the extracellular domain defined by the TM1-TM2 interval (Figure 1). A possible effect of the R181Q and R274Q mutations on plasma membrane targeting of the protein was then investigated in these clones by double immunofluorescence. WT VANGL1 shows strong expression at the plasma membrane: it produces a typical “mesh like” staining pattern that shows strong colocalization with the plasma membrane marker Na, K-ATPase (Figure 2A, top panel). On the other hand, R181Q and R274Q do not appear to be concentrated at the plasma membrane, but rather show a diffuse cytoplasmic signal and do not colocalize with Na, K-ATPase (Figure 2A, middle and bottom panels), suggesting that both mutations alter the normal targeting of VANGL1 to the plasma membrane. Normal plasma membrane targeting of VANGL1 in MDCK cells results in exposure of the TM1-TM2 domain of the protein to the extracellular milieu, where it can be recognized by antibodies directed against an HA epitope tag inserted at that site. 23 The polarity

156 and degree of exposure of this exofacial HA to the extracellular milieu were monitored in WT, R181Q, and R274Q expressing cells by immunofluorescence with an anti-HA antibody in intact cells versus cells permeabilized with 0.5%Triton X100. Results in Figure 2B (top panel) show that the WT VANGL1 protein is detected in both intact and permeabilized cells, where it shows punctate cell surface staining (permeabilized). By contrast, R181Q and R274Q mutations are detected only under permeabilized conditions, suggesting that both of these human mutations are not properly targeted to the plasma membrane. In addition, cell surface expression of the HA tag was further quantified by an ELISA using a secondary antibody coupled to horseradish peroxidase and applied to WT, R181Q and R274Q expressing cells under permeabilized (total) and nonpermeabilized (cell surface) conditions (Figure 2C). The amount of HA-bound antibody was quantified in stably expressing MDCK cells. In cells expressing the WT protein, 70% of the HA- VANGL1 associated signal was at the cell surface and only 29% of R181Q and 22% of R274Q (Figure 2C). Taken together, these results suggest that the NTD-specific human VANGL1 mutations R181Q and R274Q interfered with proper plasma membrane targeting of the protein and reduced the level of cell surface expression.

157

GFP Na, K-ATPase Merge A

Vangl1 WT

Vangl1 R181Q

Vangl1 R274Q

B Perm. Intact C

Vangl1

WT 100.00

80.00

60.00 Vangl1

R181Q 40.00

20.00 % cell surface expression surface cell %

Vangl1 0.00 R274Q WT R181Q R274Q

5. 2: Cellular localization, surface expression and quantification of WT, R181Q and R274Q h VANGL1 mutations in stably transfected MDCK cells. (A) Stably transfected MDCK cells expressing GFP-tagged hVANGL1 WT, hVANGL1 R181Q or hVANGL1 R274Q (green) were grown to confluency on coverslips, stained with plasma membrane marker Na, K-ATPase (red) and analyzed by confocal microscopy. The merged images show the WT protein colocalizing with Na, K-ATPase (yellow), while intracellular staining was observed for the R181Q and R274Q mutants. Images are representative of three independent experiments. (B) Cells expressing the three constructs (including an extracellular HA tag at position 139) were analyzed by confocal microscopy in either nonpermeabilized or permeabilized (0.5% Triton X100) conditions followed by incubation with a mouse anti-HA antibody. Images are representative of at least three independent experiments. (C) Amount of surface expression quantified in a different experiment using the same cells (in either intact or permeabilized condition) by following incubation with the HA antibody with an HRP-coupled secondary antibody and colorimetrically quantifying an HRP substrate (OPD) by spectrometry. The amount of HA-tagged WT, R181Q, or R274Q protein expressed at the cell surface of MDCK cells (intact) was shown as a percentage of total protein expression (permeabilized condition).

158

Subcellular localization of R181Q and R274Q VANGL1 mutations in stably transfected MDCK cells. Subcellular localization of the R181Q and R274Q mutations was also investigated by confocal microscopy followed by examination of serial cell sections (Z-stacks). This analysis (Figure 3A) showed that the WT VANGL1 protein is expressed predominantly in the lateral membrane of polarized MDCK cells, where it colocalizes with Na, K-ATPase. This localization is largely lost in the R181Q and R274Q mutants, which instead display a more diffuse intracellular staining associated with the periphery of the cells (Figure 3A). The altered plasma membrane recruitment and the more pronounced intracellular staining noted for R181Q and R274Q suggested these mutations may be targeted to an intracellular endomembrane compartment (Figures 2A, 2B and 3A), possibly the endoplasmic reticulum (ER). Retention and accumulation of pathological Lp-associated mutations in the ER have been described 23, and caused by aberrant processing by SEC24B into COPII vesicles, for ER to Golgi transport. 64,65 Therefore, we investigated this possibility by double immunofluorescence with the ER marker calreticulin (Figure 3B). While WT VANGL1 does not colocalize with calreticulin (Figure 3B, top panel), the R181Q and R274Q mutations show overlapping staining with calreticulin [yellow-orange color in merged images, Figure 3B]. These results indicate that, like Lp-specific pathological mutations causing a NTD in vivo 5, the human R181Q and R274Q mutations detected in familial cases of NTDs show impaired plasma membrane targeting and appear to be retained in the ER.

159

GFP Na, K-ATPase Merge

R181Q

R274Q

B GFP Calreticulin Merge

R181Q

R274Q

5. 3: Subcellular localization of WT, R181Q and R274Q h VANGL1 in stably transfected MDCK cells. (A) Cells expressing GFP-tagged hVANGL1 mutations (green) were grown to confluency, stained with Na, K-ATPase (red) and analyzed by Z-line image analysis on the confocal microscope. Representative X-Z sections are shown. The merged images shown that both R181Q and R274Q localize to an intracellular compartment and do not colocalize with Na, K-ATPase (yellow). (B) Cells expressing GFP-tagged hVANGL1 mutations (green) were grown to confluency and stained with ER marker calreticulin (red), and subcellular localization was analyzed by confocal microscopy. Merged images show that R181Q and R274Q mutants are trapped in the ER, colocalizing with calreticulin (yellow and boxed magnification), while the WT protein is expressed at the plasma membrane. Images are representative of three independent experiments.

160

Stability of the R181Q and R274Q hVangl1 mutations in transfected MDCK cells. The retention of membrane protein mutations in the ER is frequently associated with decreased protein stability and an increased level of degradation 123. The stability of the R181Q and R274Q mutations was investigated in pulse-chase studies, and the half-life of WT and mutant proteins was determined. Cells were metabolically labeled with [35S]-methionine/cysteine for 60 min, followed by a chase period in radioisotope-free medium (1-16 h), and cell extracts were analyzed by immunoprecipitation with an antibody directed against a c-Myc epitope tag inserted at the N-terminus of the proteins. A representative autoradiogram is shown in Figure 4A, while quantification of the average of several experiments was used to calculate half-lives of each protein (Figure 4B). These studies show that the half-life of WT VANGL1 in MDCK cells is ~16hrs, as we have previously documented 23. On the other hand, the half-lives of the R181Q and R274Q mutants were significantly shorter at around 7 and 5 h, respectively (Figure 4B). The stability of the WT and mutants was also monitored in cells following inhibition of protein synthesis (cycloheximide, CHX), and their possible degradation via the proteasome (MG132) was also investigated (Figure 4C, D). CHX treatment of WT had an only modest effect on the level of protein, with an ~25% reduction in WT protein, compared to ~60and ~70% reductions of R181Q and R274Q, respectively (Figure 4D). On the other hand, addition of MG132 to CHX treated cells resulted in a large increase in the level of detectable mutant protein expression, with a 60% increase for both mutants compared to only a modest 17% increase for the WT protein, clearly indicating mutant proteins are degraded in a proteosomally dependent manner. These results suggest that R181Q and R274Q mutations are pathogenic; they prevent plasma membrane insertion and cause retention in the ER, which is associated with reduced protein stability and an increased level of degradation via the proteasome.

161

A B

WT R181Q R274Q 140 0 3 6 16 hours 120 WT 100

R181Q 80 60 R274Q 40

20 Remaining fraction (%) Remaining fraction 0 0 5 10 15 20 Chase (hours) C D

NT CHX CHX+MG

140 WT R181Q R274Q 120 CHX 100

MG132 80

hVANGL1 60 40 actin 20

0 Remaining Fraction (%) WT R181Q R274Q 5. 4: Cellular stability, half –life, and degradation studies of WT, R181Q and R274Q hVANGL1 in stably transfected MDCK cells. (A) Cells expressing GFP-tagged hVANGL1 mutations were metabolically labeled by a 60 min pulse of [35S]-Met/Cys, followed by incubation in a radioisotope free medium and chased for different periods of time (up to 16 h). Cells were lysed and immunoprecipitated, followed by gel electrophoresis and autoradiography. (B) Quantification of the remaining protein was conducted on the scans of the autoradiograms using Image J. The disappearance of the protein is shown as a fraction (percent) of the total protein at time T0, which is set at 100%. The graph shows the mean of three experiments ± the standard deviation. (C) Cells expressing GFP-tagged hVANGL1 mutations were grown to confluence and treated with cycloheximide (20 μg/mL) for 6 h in the presence or absence of proteosomal inhibitor MG132 (5 μg/mL). Cells were lysed, and 50 μg of protein was separated by gel electrophoresis. Actin was used as an internal loading control. (D) Remaining fractions of WT, R181Q, and R74Q.

162

5.5 Discussion

Early studies in Drosophila identified a role for Van Gogh proteins (Vangl homologues) in providing planar polarity (PCP) information to epithelial cells, including planar polarization of cellular appendages. 43 Subsequent studies in Xenopus demonstrated that Vangl and other PCP proteins are also required for convergent extension (CE) movements during tissue patterning, including formation of the neural tube. 41 Positional cloning of Vangl2 as the gene mutated in the Lp mutant mice demonstrated that in mammals VANGL proteins were involved in both planar cell polarity (orientation of the stereocilliary bundles of the neurosensory epithelium of the organ of Corti) and convergent extension movements (neurogenesis and cardiogenesis). 5,16 Subsequently, mutations in other PCP genes were found to similarly cause neural tube defects in mice: this includes homozygosity for loss-of-function alleles in Celsr1, Dvl1/Dvl2, Dvl2/Dvl3, Fzd3/6, Ptk7, Scribble, Ror2, and Sec24b, as well as combined heterozygosity for mutations at Vangl2Lp/+ and other PCP genes. 5 In humans, the study of familial and sporadic cases of NTDs identified unique, disease-associated mutations in VANGL176,78,120 and VANGL2 75,120 proteins, further highlighting a very important role for mammalian VANGL proteins in PCP signaling, CE movements and particularly formation of the neural tube. The 20 mutations identified in human VANGL1 and VANGL2 in cohorts of familial or sporadic NTDs cases have been flagged as pathological on the basis of the following criteria. (a) They are patient-specific and are not present in ethnically matched unaffected controls. (b) They affect residues that are highly conserved among Vangl homologues. (c) They represent nonconserved substitutions that are likely to affect local protein structure. (d) In some instances they behave as loss-of-function in complementation studies in model organisms such as yeast (interaction with DVL family members) 22,75and zebrafish (complementation of CE phenotypes induced by silencing of the fish trilobite ortholog). 25,120 However, the mechanistic basis for the impaired function of these human mutations has not yet been investigated.

To start to characterize the biochemical basis of altered function in Vangl mutations found in human NTDs patients, we elected to study the R181 and R274 residues. R181 and R274 are independently mutated to glutamine in VANGL1 (R181Q and R274Q) in two unrelated NTDs patients and mutated to histidine in VANGL2 (R177H and R270H) at the homologous position in

163 two other unrelated NTDs patients. In addition, both R181 and R274 are extremely conserved in the VANGL family, R274 being invariant and R181 being highly conserved, with both mapping to protein subdomains that also show a high degree of sequence conservation (Figure 1). Finally, the charged positive side chain of arginine is lost in both substitutions found in independent NTDs patients. Together, these findings suggested that R181 and R274 play critical and conserved structural and functional roles in VANGL protein functions, and are particularly mutation sensitive, and that the loss of either of these arginines results in altered function. Hence, we compared the biochemical features of R181Q and R274Q to those of the WT VANGL1 protein following expression of each mutation (reconstructed in the VANGL1 protein template) in MDCK epithelial kidney cells. Our results indicate that normal membrane targeting of these two mutations is impaired, and rather than being recruited to the plasma membrane (colocalization with Na, K- ATPase at the basolateral side of polarized epithelial cells), R181Q and R274Q are found in an intracellular endomembrane compartment that shows overlap staining with the ER marker calreticulin. The two mutations show reduced stability and 2 and 4 fold reductions in protein half- life, respectively in MDCK cells (> 20h for WT, ~ 9hrs for R181Q and ~ 5hrs for R274Q). Compared to WT, both mutants were found to be rapidly degraded in a proteasome-dependent and MG132-sensitive fashion. The behavior of the two human mutations is essentially identical to what we have recently demonstrated for the three available Vangl2 alleles associated with the severe neural tube defect in the Lp mouse, i.e., D255E (Lpm1Jus), R259L(Lpm2Jus) and S464N (Lp). 23,24 The effect of the R181Q and R274Q mutations on membrane targeting and protein stability is specific and is not seen in other VANGL1 and VANGL2 mutants associated with neural tube defects. For example, an L202F mutation was identified in a 9 year old female of Italian origin with an open NTD (myelomeningocele). This mutant maps to the ‘WLF’ motif that is invariant in the VANGL protein family. The L202F mutant is properly targeted to the plasma membrane and is not retained intracelluarly, in contrast to the R181Q and R274Q mutations (Figure 1 if the Supporting Information). The mechanistic basis for the loss-of-function in L202F remains to be established but may involve impairment of interaction with VANGL binding partners such as members of the DVL protein family. Together, these findings clearly establish that the human R181Q and R274Q mutations detected in familial cases of NTDs (a) have impaired function, (b) are phenotypically indistinguishable from the S464N allele characteristic of the Lp mutation, and (c) are likely pathological in vivo contributing to the etiology of NTDs in these patients. By the

164 same token, it is very likely that the R177H and R270H substitutions detected in NTDs cohorts at the homologous positions of human VANGL2 are also pathological. Finally, our findings provide a set of simple biochemical assays in vitro in which the impact of the VANGL1 and VANGL2 mutations detected in clinical specimens on protein function can be formally assessed.

It is important to note that the VANGL1R181Q, VANGL1R274Q, VANGL2R177H and VANGL2R270H mutations have been detected as heterozygous mutations in clinical cases of NTDs. 76,77,120 This suggests that these allelic mutations at VANGL1 and VANGL2 behave either as haplo- insufficient in a gene-dosage dependent pathway or as partially penetrant with negative codominance. This situation is very similar to the current debate regarding the mode of inheritance of pathological Lp-associated Vangl mutations in mice, which have also been alternatively assigned as dominant negative or loss-of-function in gene dosage dependent pathways 5. Favouring a dominant negative effect are the observations in mice that (a) several Vangl2-associated PCP phenotypes appear more severe for Lp alleles (Vangl2D255E and Vangl2S464N) than for a null allele 68 and (b) VANGL1 and VANGL2 appear to physically interact and, in cotransfection experiments, Lp alleles disrupt VANGL1-VANGL2 interactions, as well as trafficking and membrane targeting of VANGL1 and VANGL2, 68 and decrease the level of post-translational modification of the WT protein. 67,86 Favoring haploid insufficiency in a gene dosage dependent pathway are the reports that (a) experimental overexpression or silencing of core PCP genes causes the same phenotype in different animal models tested, (b) all experimentally-induced or naturally occurring Vangl2 mutations described so far show varying degrees of the same phenotype in mice (looped tail, inner ear defects) in heterozygotes and homozygotes in vivo, 5 (c) Lp-associated VANGL2 protein mutations are expressed at lower levels in vivo and display reduced stabilities and shorter half- lives when tested in vitro, 16,17,87 (d) colocalization studies by double immunofluorescence and confocal microscopy in transfected MDCK cells show that expression of VANGL2 D255E has no effect on membrane targeting of WT VANGL2. 23

In conclusion, our findings identify R181 and R274 as critical arginines that are essential for the normal function of VANGL proteins. Their independent substitution with glutamine or histidine in multiple independent patients with NTDs causes a biochemical phenotype that is indistinguishable from those of known mutations in Vangl2 that are associated with the NTD in

165 vivo in the Lp mouse. Hence, our study strongly suggest that R181Q and R274Q in VANGL1 are pathological and cause neural tube defects in humans.

5.6 Acknowledgements

Image acquisition, data analysis, and image processing were conducted on equipment and with the assistance of the McGill Life Science Complex Imaging Facility, which was funded by the Canadian Foundation for Innovation.

166

5.7 Supplementary Figures A HA 139 L202F pLeu202Phe

Human Vangl1 VVSYWLFYGVR Human Vangl2 VVSYW LFYGVR Mouse Vangl1 VVSYW LFYGVR Mouse Vangl2 VISYWLFYGVR Zebrafish Vangl1 LLSYWLFFGVR Zebrafish Vangl2 VASYWLFYGVR Xenopus Vangl2 VVSYWLFYGVR R181Q GFP R274Q Drosophila Stbm TFAYWLFYIVQ C.Elegans Stbm LFAFWLFYIVR NH2

COOH B GFP Na, K-ATPase Merge

L202F

C GFP Calreticulin Merge

L202F

5. S 1: Schematic representation, multiple sequence alignment and cellular localization of the L202F variant. (A) Schematic representation of secondary structure of GFP-tagged hVANGL1 protein, including location of L202F and sequence conservation of the mutated residue across species. (B) Stably transfected MDCK cells expressing GFP-tagged L202F protein (green) were grown to confluency on coverslips, stained for plasma membrane marker Na, K-ATPase (red) and analysed by confocal microscopy. The merged image show protein colocalizing with Na, K-ATPase (yellow). (C) The same cells were stained for ER marker calreticulin (red). Merged image shows that the L202F mutation does not colocalize with calreticulin and is expressed at the plasma membrane.

167

Chapter 6: The intracellular carboxyl terminal domain of VANGL proteins contains plasma membrane targeting signals

168

Preface to Chapter 6

Chapter 6 of this thesis further explores structure-function relationships in the VANGL family of proteins by gaining a better insight into protein determinants responsible for plasma membrane targeting. Many in vivo and in vitro studies (including work previously presented in this thesis) have shown that targeting VANGL protein to the membrane is essential for its function, including interaction with binding partners. By examining the sequence of VANGL proteins, several motifs previously identified to behave as membrane or internalization signals during clathrin mediated endocytosis were also found in the amino and carboxy termini. These include YXXϕ, (ϕ is a hydrophobic amino acid), NPXY/ϕ and di-leucine LL and they were tested by creating progressively larger deletion at the N- and C-terminal tails.

169

6.1 Abstract

VANGL1 and VANGL2 are integral membrane proteins that play a critical role in establishing planar cell polarity (PCP) in epithelial cells and are required for convergent extension (CE) movements during embryogenesis. Their proper targeting to the plasma membrane (PM) is required for function. We created discrete deletions at the amino and carboxy termini of VANGL1 and monitored the effect of the mutations on PM targeting in MDCK cells. Our results show that the VANGL1 amino terminus lacks PM targeting determinants, and these are restricted to the carboxy terminus, including the predicted PDZBM motif at the C-terminus.

170

6.2 Introduction

Neural tube defects (NTDs) are a group of heterogeneous birth defects caused by failure of the neural tube (precursor of the central nervous system) to close properly during embryogenesis 9. These malformations are common (1/1000 birth) and often cause infant mortality 139. Epidemiological studies indicate that NTDs are caused by complex genetic determinants in combination with environmental factors, both poorly understood 138. We have shown that the Vangl1 and Vangl2 genes are essential for the development of the neural tube, and that mutations in these genes cause NTDs in mice and humans 5.

Vangl genes are vertebrate relatives of the fly Vang/Stbm, a gene required for polarization of epithelial cell layers and their associated cellular appendages, a process called “planar cell polarity” (PCP)5. Core PCP genes and proteins include Vangl/Stbm (van gogh-like, strabismus), Pk (prickle), Dgo (diego), Fz (frizzled), Dsh (dishevelled) and Fmi (flamingo) 140. These assemble into membrane-bound multiprotein complexes that partition to the proximal (Vangl/Pk/Dgo) and distal side (Dsh/Fmi/Fz) of cells during PCP. The asymmetrical distribution of these complexes in adjacent cells propagates polarity signals to organize cells during PCP. 140. The mechanism underlying recruitment, assembly and signaling by these membrane complexes remain unknown.

Mammalian VANGL1 and VANGL2 are integral membrane proteins that share high sequence similarity (~75%), and that are expressed at the plasma membrane (PM) of different tissues during embryogenesis 19,22,23. Polarity mapping using VANGL1 recombinant proteins bearing exofacial epitope tags expressed at the PM of transfected cells, demonstrated that VANGL proteins contain four transmembrane domains, with intracellular amino and carboxyl termini. PM recruitment of VANGL proteins is essential for function during PCP 23,24. In the normal embryos in vivo, VANGL proteins are present at the basolateral side of epithelial cells of several tissues. This staining is lost in mutant embryos that harbor NTDs caused by defective VANGL proteins 19,23. PM recruitment of VANGL proteins is also required for formation of complexes with other PCP proteins 22-24. While sequence determinants required for VANGL membrane association are poorly characterized, several motifs responsible for targeting in other membrane proteins are found in

171

VANGL proteins, including tyrosine based motifs YXXϕ and NPXY/ϕ (where ϕ is hydrophobic) and di-leucine motifs known to act as internalization signals for clathrin mediated endocytosis 141. Specifically, YSGY at the amino acid position (pst.) 7-10, YSYY (pst 10-14) and a di-leucine motifs (pst. 56) are found in the amino terminus, while YKDF (pst. 284-287), NPNL (pst. 290- 294) are found at the carboxyl terminus. Finally, VANGL proteins contain a PBM (PDZ binding motif), that is essential for interaction and recruitment of several PDZ containing partners including PCP proteins Scribble and Dvl1, Dvl2 and Dvl3 22,83. Here, we created linear deletions of the amino and carboxyl termini of VANGL1 and determine the effect of the deletions on PM targeting.

172

6.3 Experimental Procedures

Material and Antibodies Geneticin (G418) and penicillin/streptomycin were purchased from Invitrogen (Carlsbad, CA). Restriction endonucleases were from New England Biolabs (Ipswich, MA), and Taq DNA polymerase was from Invitrogen. The mouse monoclonal antibodies directed against the influenza hemagglutinin epitope (HA.11) were purchased from Covance (Berkeley, CA). The mouse monoclonal antibody recognizing Na, K-ATPase (α) was from Santa Cruz Biotechnology (Santa Cruz, CA). Cy3-conjugated goat anti-mouse antibody and peroxidase-coupled goat anti-mouse antiserum were from Jackson ImmunoResearch Laboratories (West Grove, PA). The rabbit polyclonal antibody against calreticulin was obtained from Affinity BioReagents (Golden, CO).

Plasmids and Site-Directed Mutagenesis The human VANGL1 cDNA was used for constructing all deletions. The cDNA was amplified from total human RNA by reverse transcriptase polymerase chain reaction (RT-PCR) and cloned into the pCS2+ plasmid vector. PCR-mediated mutagenesis with overlapping oligonucleotides was used to insert a short antigenic peptide epitope (EQKLISEEDL) from the human c-Myc protein at the N-terminus of VANGL1 and a small hemagglutinin (HA) epitopes (YPYDVPDYA) in the extra-cellular loop between TM1-2 at position 139, followed by cloning into the pCB6 mammalian expression vector, as we have previously described 23. To facilitate identification of transfected cells expressing recombinant hVANGL1 proteins, c-Myc/HA-tagged hVANGL1 cDNAs were fused in-frame to the green fluorescent protein (GFP) by cloning into the peGFP-C1 vector (Clontech, Mountain View, CA). The deletions were introduced in hVANGL1 cDNA by PCR overlap extension mutagenesis, using the following primers; N1 (∆ 1-15) (Forward: ATCGAATTCCATGGAGCAGAAGCTAA; Reverse: CTTTGTAGTAATTTTCTAGGACCACC), N2 (∆ 1-95) (F: ATCGAATTCATGGAGCAGAAGCTAAT; R: CTTTGTAGTAATTTTCTAGGACCACC), C1 (∆ 520-524) (F: TACTACAAAGATTTCACCATCTATAAC; R: ATCTCTAGACTGTAAGCGAAGGACAA), C2 (∆ 354-524) (F: CTATTGGCTTTTTTACGGGGTCCGCAT; R: GTCATCTCTAGATACTCGCCGTTCAT), C3 (∆ 284-524) (F: TTTACCTCCGATCCTGTGGAGGTACCC; R:

173

GTCATCTCTAGAGTAATTTTCTAGGA), C4 (∆ 249-524) (F: TTTACCTCCGATCCTGTGGAGGTACC; R: GTCATCTCTAGAGGGCTGCAGCTGCC). The integrity of all VANGL1 cDNA constructs was verified by nucleotide sequencing. Cell Culture, Transfection, and Western Blotting Madin-Darby canine kidney (MDCK) epithelial cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/mL penicillin, and

100 μg/mL streptomycin (37 °C, 5% CO2). To generate stable transfectants, MDCK cells were transfected with c-Myc/HA-tagged VANGL1 WT or linear deletion constructs (subcloned in peGFP-C1) using Lipofectamine Plus Reagent as described in the manufacturer’s instructions (Invitrogen). Total cell lysates, western blot analysis, immunofluorescence (HA, Na, K-ATPase, calreticulin) and cell surface expression were performed as we have previously described. 24

174

6.4 Results

Recombinant proteins were constructed using a human VANGL1 cDNA modified by addition of a GFP protein and a cMyc epitope tag at the amino terminus, and by the in-frame insertion of a hemagglutinin (HA) epitope tag at position 139 in the first extracellular loop delineated by transmembrane domains 1 and 2 (Figure 1A). Six deletions overlapping different segments of the amino (N1, ∆1-13; N2, ∆1-95) and carboxyl termini (C1, ∆520-524; C2, ∆354- 524; C3, ∆284-524; C4, ∆249-524) of Vangl1 were created by PCR mutagenesis, cloned in an expression vector and transfected into MDCK kidney cells, in which transfected VANGL1 is expressed at the basal and lateral membranes 23. At the amino terminus, the putative membrane targeting function of sequence motifs YSGY (pst.7-10), YSYY (pst.10-14) and di-leucine LL (pst. 56-57) can be assessed using constructs N1 and N2, while at the carboxy terminus, a role of the PDZ binding motif ETSV (pst. 520-524) can be tested in construct C1. The function of motifs YKDF (pst. 284-287), and NPNL (pst. 290-294) can be examined in constructs C3 and C4 (Figure 1). The level of expression of the various recombinant VANGL proteins was measured by immunoblotting (anti-HA antibody) and shows low level of all variants in transfected MDCK cells (Fig. 1B)

175

pst. 1 pst. 524 WT GFP hVANGL1 (WT) 87 kDa

14 524 N1 GFP ∆1-13 (N1) 85 kDa

96 524 N2 GFP ∆1-95 (N2) 76 kDa

1 519 C1 GFP ∆520-524 (C1) 86 kDa 1 353 C2 GFP GFP ∆354-524 (C2) 68 kDa 1 283 Myc C3 GFP hVangl1 ∆284-524 (C3) 60 kDa TM 1 248 HA C4 GFP PBM ∆249-524 (C4) 56 kDa

+ - N1 N2 + - C1 + - C2 + - C3 + - C4

6. 1: Schematic representation and immunoblot expression of WT and deletion recombinant protein constructs. (A) Schematic representation of the WT VANGL1 protein and of 6 deletion constructs, including position (pst.) corresponding to amino acids residue number, sequence features (GFP, Myc and HA), and expected molecular mass of corresponding truncated proteins. (B) Expression of the recombinant VANGL proteins was monitored by immunoblotting of total cell extracts from stably transfected MDCK cells, and using an anti-HA monoclonal antibody directed against an HA epitope tag inserted in frame in all VANGL constructs. Positive control (+): WT protein and negative control (-): whole cellular extracts of untransfected MDCK cells.

176

The effect of the deletions on PM targeting of VANGL1 was assessed by double immunofluorescence using the Na, K-ATPase as a PM marker 23. Recombinant VANGL1 proteins lacking N-terminal segments ∆1-13 (N1) and ∆1-95 (N2) are correctly targeted to the PM, and colocalize with Na, K-ATPase similarly to the WT VANGL1 (Figure 2). Conversely, VANGL1 proteins lacking any portion of the C-terminal domain failed to reach the PM (Figure 2). Indeed recombinant proteins C1 (∆520-524), C2 (∆354-524), C3 (∆284-524), and C4 (∆249-524) did not colocalize with Na, K-ATPase, but rather displayed a punctate intracellular staining. These results suggest that sequence elements in the carboxyl terminus of the protein are essential for PM association of VANGL1 (Figure 2).

177

GFP Na,K-ATPase Merge Perm Non-Perm

hVANGL1 (WT)

α-HA α-HA

∆1-13 (N1)

∆1-95 (N2)

∆520-524 (C1)

∆354-524 (C2)

∆284-524 (C3)

∆249-524 (C4)

6. 2: Immunofluorescence for WT and deletion constructs of plasma membrane targeting and of cell surface expression in intact and permeabilized cells. MDCK cells stably transfected with either GFP-tagged hVANGL1 WT (green) or GFP-tagged hVANGL1 amino or carboxy terminal truncated constructs were grown on cover slips and either (A) stained with the PM marker Na, K-ATPase (red) or (B) incubated with an anti-HA antibody (each construct possessed an extracellular HA tag at position 139) under either nonpermeabilized or permeabilized (0.5% TritonX-100) condition, followed by incubation with a Cy3-conjugated secondary antibody. Images were acquired by confocal microscopy.

178

To further validate presence or absence of VANGL1 deletion variants at the PM, we monitored surface expression and accessibility of an exofacial HA epitope tag inserted in the first extracellular loop of VANGL1 (pst. 139, predicted TM1-TM2 interval). Intact cells producing a fluorescent signal with an anti-HA monoclonal antibody detect cell surface expression, while permeabilized conditions also reveal intracytoplasmic protein expression (Figure 2). Under these conditions, wild type VANGL1 1 protein is inserted at the PM, exposing the HA tag to the extracellular milieu, resulting in immunofluorescence detection under both permeabilized and nonpermeabilized conditions. While amino-terminal deletion constructs N1 and N2 produce staining similar to the WT protein, C1-C4 are only detected under permeabilized conditions, verifying that there is little if any expression of these proteins at the PM. Under these conditions, mutant variants C1-C4, but not WT or NI-N2 constructs, show overlapping staining with the ER marker calreticulin, suggesting inappropriate ER retention of these mutants (Figure 3). Results reveal that a) the N-terminal of VANGL1 is not essential for PM targeting, and b) that the C- terminal domain is required for PM targeting of the protein in MDCK cells, including the short PDZ binding motif (ETSV; C1, ∆520-524), whose deletion abrogates PM targeting.

179

GFP Calreticulin Merge

hVANGL1 (WT)

∆1-13 (N1)

∆1-95 (N2)

∆520-524 (C1)

∆354-524 (C2)

∆284-524 (C3)

∆249-524 (C4)

6. 3: Subcellular localization of WT and deletion constructs. MDCK cells stably transfected with either GFP-tagged hVANGL1 WT (green) or GFP-tagged hVANGL1 amino or carboxy terminal truncated constructs were immunostained with the ER marker calreticulin (red) to determine subcellular localization. The magnification views point to overlapping staining of the ER marker with mistargeted Vangl constructs. Images were acquired by confocal microscopy.

180

6.5 Discussion

Previous work in vivo and in transfected mammalian cell lines has demonstrated that PM targeting of VANGL proteins and asymmetric partitioning of VANGL -containing PCP complexes to polar opposites of epithelial cells are critical for biological function 5. For example, Vangl mutations in Lp mice (D255E, S464N) abrogate this partitioning to PM domains and cause PCP defects, including NTDs 22-24. Additionally, Lp-associated VANGL2 variants (D255E, S464N) fail to interact with SEC24B, a cargo-selecting component of COPII vesicles responsible for ER-Golgi transport, resulting in ER retention of VANGL proteins. This, combined with the observation that Sec24b mutant mice exhibit craniorachischisis further emphasized the critical importance of VANGL proteins trafficking in regulating PCP and ultimately neural tube closure in mammals. Here, we used deletion analysis to locate sequence determinants responsible for VANGL membrane targeting.

Both amino terminal deletions, including ∆1-13 (N1) which eliminates both tyrosine based motifs YSGY(7-10) and YSYY(10-14) and ∆1-95 (N2) which additionally removes the di-leucine motif LL(56) did not seem to affect PM targeting. Indeed, both truncated constructs showed basolateral staining in MDCK cells, colocalization with Na, K-ATPase membrane marker and showed staining with an HA antibody under nonpermeabilized conditions, suggesting that the amino terminus of VANGL proteins does not contain major determinants for PM membrane targeting. Interestingly, previous study of the Vangl zebrafish relative trilobite showed that a 13 amino-acid addition at position 21 in the N-terminus of this protein causes a loss-of-function in vivo 35. In addition, the N-terminal segment of VANGL1/VANGL2 proteins harbors pathological variants (S83L and S84F) associated with sporadic cases of NTDs 75,76; in particular, S84F maps to a Ser/Thr cluster that is phosphorylated in response to a Wnt5 signaling gradient required for PCP 86. Hence, the amino terminal segment of VANGL proteins does not harbour critical determinants for PM targeting, but is nevertheless required for other functional aspects.

Conversely, none of the C-terminal deletions ∆520-524 (C1), ∆354-524 (C2), ∆284-524 (C3), ∆249-524 (C4) were recruited to the PM, and they displayed intracellular staining (Figure 2), reflecting retention in the ER, as demonstrated by overlapping staining with calreticulin (Figure

181

3). The ~280 amino acid residues carboxy terminal intracellular domain is highly conserved across species (> 80% sequence similarity), and all three Lp mutant alleles (D255E, R259L and S464N) and the majority of human variants associated with NTDs (VANGL1: R274Q, M328T and A404S and VANGL2: L242V, T247M, R270H, R353C, F437S and R482H) map to this domain. Furthermore, a recent study deleting the Vangl segment spanning amino acid positions 280-291

(VANGL2 numbering) impaired membrane targeting. More specifically, the Phe in the YXXF(280-

283) motif was found to be a critical amino acid responsible for the binding to the µ subunit of the clathrin adaptor complex AP-1, thus mediating the export of VANGL proteins from the trans- Golgi network (TGN) to the PM 84. Finally, the recent description of a new Lp-allele, Vangl2m1Yzcm (displaying craniorachischisis) showing a stop codon at position 449, thus lacking the PDZ binding motif, additionally highlights the importance of this domain for PCP signaling and neural tube closure 142.

Our data suggests that deleting the PDZ binding motif in the VANGL1 protein is sufficient for intracellular retention and impaired PM targeting. Our findings do not indicate that the PDZ binding motif functions directly as a membrane targeting motif, but that its action could be indirect, for example, by permitting interactions with other proteins which are relevant for membrane targeting. This motif is generally thought to mediate interactions with other PDZ containing proteins. These include interaction with the PDZ and DIX domains of Dvl1, Dvl2 and Dvl3, which is abrogated by Lp mutations (D255E, S464N) and by NTDs specific human variants (R353C, F437S)22,75. Additionally, the VANGL2 PDZ binding motif recruits the PDZ domain of Scribble, and is disrupted by the Lp variant S464N and by absence of the PBM itself 83. More recently, interaction with a member of the nexin family, the PDZ containing protein SNX-27, which promotes the recycling of transmembrane proteins from endosomes to the PM has also been demonstrated 143. Additionally, this motif was also shown to affect the localization of VANGL2 at post synaptic sites in the rat brain, by directly mediating interaction between VANGL2 and the third PDZ motif of PSD-95143. These findings together with the results reported here strongly suggest that the PDZ binding motif of VANGL proteins is required for interactions with a number of PCP proteins, which is essential for targeting PCP complexes to the PM.

182

6.6 Acknowledgements

Image acquisition, data analysis, and image processing were conducted on equipment and with the assistance of the McGill Life Science Complex Imaging Facility, which was founded by the Canadian Foundation for Innovation.

183

Chapter 7: Summary and future perspectives

184

7.1 Summary

The work presented in this thesis studied the structure-function relationship and subcellular targeting of VANGL proteins. In chapter 2, the secondary structure of VANGL proteins was determined using an epitope mapping method. Small HA tags were inserted at individual predicted intracellular and extracellular sites of VANGL1, followed by IF in intact or permeabilized stably transfected polarized MDCK cells. This technique has the advantage of establishing the membrane topology in intact cells of the properly targeted protein with respect to the plasma membrane. These studies demonstrated a 4-transmembrane domain model, with intracellular N- and C- terminus. Additionally, this topology model also positioned the PBM intracellularly. In chapter 3 and 4, the effect of Lp-associated mutations (D255E, R259L and S464N) was studied by developing a set of biochemical assays to monitor membrane targeting, subcellular localization, protein stability and half-life. This work determined Lp mutations cause loss of protein function by inhibiting proper insertion in the plasma membrane, decreasing protein half-life and by targeting mutant protein for proteasome mediated degradation more rapidly. These results establish the molecular mechanism by which Lp-associated mutations interfere with normal protein function, such as the ability of VANGL1 proteins to provide polarity information during embryogenesis and cause neural tube defects. The biochemical assays established for the studies in chapter 3 and 4 were then used to characterize the molecular basis of loss-of-function for patient- specific VANGL1 and VANGL2 mutations identified in sporadic and familial cases of NTDs. The work described in chapter 5 concentrates on two highly conserved residues R181 and R274. R181 is independently mutated to H (VANGL2R177H) and Q (VANGL1R181Q) in two patients, and R274 is mutated to H (VANGL2R270H) and Q (VANGL1R274Q) in two other patients, indicating an essential role for these two arginines. These in vitro studies established a common molecular mechanism for Lp-associated mutations and these human mutations, namely impaired membrane targeting, retention in the ER, reduced protein stability and diminished half-life. These experiments helped elucidate the pathological mechanism of VANGL1 and VANGL2 mutations associated with NTDs in humans. VANGL proteins have been found to be expressed at the lateral and basolateral side in vivo in several epithelial cells and in vitro in stably transfected polarized MDCK epithelial cells. Additionally, work done in chapters 3, 4 and 5 demonstrates that both Lp-associated mutations and several patient specific VANGL1 and VANGL2 mutations identified in human NTDs abrogate

185 proper membrane insertion. In chapter 6, protein determinants responsible for membrane targeting of VANGL proteins were tested. Several motifs, namely: YXXϕ (ϕ is hydrophobic), NPXY and di-leucine (LL), found in both N- and C-terminal segments of the VANGL proteins, were previously identified to behave as membrane and internalization motifs for clathrin-mediated endocytosis. Progressive deletions in the amino and carboxy regions of VANGL proteins found that while the N-terminal tail lacks plasma membrane targeting motifs, these sequence determinants were restricted to the C-terminal tail and to the PDZ binding motif. In addition to the role of the PBM in interacting with several PCP proteins, these studies establish a critical role for this short motif in properly targeting PCP complexes to the plasma membrane.

7.1.1 Subcellular targeting and trafficking of VANGL proteins

While it is clear that membrane localization of VANGL proteins is essential for their function, the molecular network and associated proteins involved in this process remains poorly understood. Comprehension in vertebrates of how VANGL proteins are transported to the membrane and which proteins they interact with is critical to understand how this tetraspanin family of proteins establishes and propagates PCP signals, including coordination of CE and neural tube closure. Some of the mechanisms devised by cells to target proteins to specific cellular compartments include; sequences in the protein itself, binding to various interacting partners or scaffolding proteins and certain post-translational modifications.

In chapter 2, establishing the secondary structure at the membrane of VANGL proteins was an essential first step to better characterise their functions. In the absence of a high resolution 3D X-ray crystallography tertiary structure, topology mapping provides a much needed upgrade from the linear amino acid sequence and positions various structural determinants either intra- or extra- cytoplasmic. As a result, the C-terminal tail of VANGL proteins, including the short protein interaction domain (PDZ) was found to be located intracellularly.

7.1.1.1 Sequence motifs for membrane targeting

186

Last year, a study published by Scheckman’s group showed known basolateral sorting motif YXXϕ (X; any amino acid and ϕ; bulky hydrophobic amino acid) as essential for VANGL2 membrane localization. In transfected cells, removal of this motif retained Vangl2 protein in the trans-golgi network (TGN).In particular, transport of VANGL proteins out of the TGN was found to be dependent on ARFRP1 (a GTP-binding protein) and on AP-1 (a clathrin adaptor complex), as both knockdowns of ARFRP1 and of AP-1 independently also lead to improper targeting of VANGL2. In the model proposed by the authors, Arfp1 exposes a binding site on AP-1 that is then able to specifically recognize the YYXXF sorting motif located in the C-terminal tail of VANGL proteins, thus allowing for the capture of Vangl2 into a transport vesicle. 84

A recent study also showed a VANGL2 construct lacking amino acid 298-382 (PKBM: corresponding to a previously identified PK binding domain) to be improperly targeted to cell bodies in transfected neurons. This suggests an essential role for this short C-terminal signal for correct localization of VANGL proteins in synapses. In addition, VANGL2∆PKBM greatly weakened a newly identified physical interaction between VANGL2 and adherens junction molecule N-cadherin in synapse rich brain extracts. Interestingly, neither Lp-associated mutations (D255E, S464N) nor a VANGL2 construct lacking the PDZ-binding motif had any effect on this interaction. 144 In vivo, in the mouse cochlea, N-cadherin has been found to colocalize with VANGL2 at the membrane of inner hair cells. 80 Additionally, in Lp/Lp embryos, the expression pattern of N-cadherin has been found to be altered further suggesting a role for this adhesion molecule in the PCP signaling cascade. In addition to N-cadherin, another adherens junction protein, RACK1, has also been found to play an important role for VANGL membrane localization. More specifically, the VANGL2 PBM was found essential for interaction with RACK1. In VANGL2 overexpressing cells, RACK1 is mistargeted, while knockdown of RACK1 alters the membrane distribution of Vangl2.145 In zebrafish, RACK1 has been found essential during gastrulation and neurulation 146 and in vivo in Lp/Lp embryos, normal distribution of RACK1 was disturbed. These studies suggest a potentially important role for the cell adhesion molecules in Vangl2 mediated PCP events.

7.1.1.2 SEC24B dependent membrane trafficking

187

SEC24B is a component of coat protein complex II (COPII) vesicle system responsible for the export and targeting of the majority of properly folded proteins out of the ER. SEC24B selectively sorts VANGL2 while Lp-associated mutations (D255E, S464N) remain trapped in the ER. 64 Homozygous Sec24b mutant mice exhibit craniorachischisis, while genetic interaction with the Lp allele in double heterozygote mice exacerbates the spina bifida phenotype. 65 Recently, four SEC24B mutations were identified in a Chinese cohort of 163 foetuses with NTDs, including three (F227S, R1248N, A1251G) that affected the subcellular localization of VANGL2 in transfected cells. 147

7.1.1.3 Phosphorylation

Phosphorylation of VANGL proteins has been found to be an important regulator of their subcellular targeting. In 2006, a study found VANGL proteins to be Ser/Thr phosphorylated in intestinal epithelial cell lines in response to stimulation by intestinal trefoil factor (ITF/TFF3). Increased phosphorylation as a result of ITF stimulation resulted in a decrease of VANGL1 at the plasma membrane. 85 This was followed several years later with findings that phosphorylation of Vang-1 (C. Elegans ortholog) by ELG-15 (FGF like receptor tyrosine kinase) is critical for asymmetric localization of VANGL proteins. During morphogenesis in C. Elegans, Vang-1 is not located apically, but found restrained at the basolateral side of intestinal cells. On the other hand, in ELG-15 embryos, the distribution pattern of Vang-1 was found at the apical junctions suggesting phosphorylation as an important mechanism of asymmetric membrane targeting.148

7.1.2 Future perspectives

7.1.2.1 Biochemical tests

VANGL1 and VANGL2 have been found to physically interact with a number of PCP proteins, a fundamental step for membrane targeting. One method to test if human mutations identified in sporadic or familial cases of NTDs affect binding with a known protein partner is to co-transfect individually all human mutations and potential binding partners in epithelial cells. These include DVL1, 2, 3, SCRIBBLE, PRICKLE, N-CADHERIN etc. and interaction with VANGL proteins could be monitored by immunoblotting cell extracts following co-

188 immuprecipitation. Additionally, all mutations can be functionally tested in zebrafish complementation assays. Vangl2 knockdown in zebrafish (trilobite mutants) using anti-sense morpholinos causes CE defects. These are able to be rescued by co-injecting human Vangl1. The ability of Lp-associated mutations (D255E, S464N) and of several human mutations identified in VANGL1 (V239I, M328T, R274Q) to rescue the CE phenotype has previously been studied.25 All human mutations can be additionally assessed in a zebrafish model quantifying the convergent extension phenotype (body length, body angle, and CE of somites)

As previously discussed, phosphorylation of VANGL proteins also plays an important role in targeting VANGL proteins to the plasma membrane. In eukaryotic cells, it occurs on serine, threonine, tyrosine and histidine residues. Gao et al. 86 found Vangl2 to be phosphorylated on two Ser/Thr clusters in the N-terminus. Additional phosphorylation signals can be found in Vangl1 at T251, T288 and S487. These can be tested by first labelling MDCK cells stably expressing Vangl1 protein with P32, followed by immunoprecipitation, SDS-PAGE and autoradiography. This will assess normal Vangl1 phosphorylation in MDCK cells. If VANGL proteins are phosphorylated in these cells, these signals can be further tested by either deleting these phosphorylation clusters or by mutating threonine or serine to alanine (inhibiting phosphorylation) and re-assessing the phosphorylation status in MDCK cells.

7.1.2.2 Yet to be discovered genetic and environmental factors

In the mouse, loop-tail mutations cause craniorachischisis in the homozygous state. On the other hand, all predicted pathogenic mutations identified in humans cases of NTDs have been found as heterozygotes. This can be explained by the presence of either yet to be identified additional genetic and/or environmental factors.

7.1.2.2.1 Genetic determinants

It is hypothesised that up to 70% of NTDs can be attributed to genetic factors. Currently, there are over 240 genes that have been identified to cause NTDs in mice, 4 with many more to be discovered. Out of those, genes involved in the PCP pathways have been predominantly associated

189 with the most severe NTD, craniorachischisis. 138Interestingly, the pathogenic mutations identified in sporadic and familial cases of humans NTDs include not only craniorachischisis, but also myelomeningocele, anencephaly and other different types of NTDs. 139The prevailing hypothesis is that in human patients, there must be interaction between two or more heterozygous gene mutations. More precisely, it is probably the different combinations between mutations that will determine the tendency towards spina bifida or craniorachischisis. For PCP genes, digenic and compound heterozygote mutations have been found in a total of nine patients. Two different studies identified five CELSR1 compound heterozygous mutations, 149,150 while De Marco et al. pinpointed digenic combinations in four additional patients (VANGL1/DVL3, CELSR1/DVL2 and two independent VANGL2/DVL2 patients). 151 Recently, the variability in the severity of the NTDs phenotypes observed in humans was also shown in mutant mice. On a uniform C3H/HeH background, Murdoch et al. created every possible combination between Celsr1, Vangl2 and Scribble (both double and triple heterozygote mice). Individually homozygous, these three genes have the craniorachischisis phenotype 100% percent of the time, but in different combinations, they produced a range of NTDs phenotypes similar to that seen in humans (different in penetrance and in severity) 156. Since the genotype-phenotype in mice and in humans does not always correspond, it is critical to further identify additional genetic determinants.

7.1.2.2.2 Environmental determinants

Folate deficiency has been the most studied environmental factor predisposing to NTDs. While preconception dietary folate supplementation has greatly reduced the incidence of NTDs around the world, about 30% of pregnancies do not respond to folic acid (FA).1 Several mouse models, including loop-tail have been found to be resistant to FA. Additionally, loop-tail mice have not been found to harbour any defects in the folate metabolism. 152Deficiencies in a number of additional nutriments have also been observed in NTDs pregnancies, these include, zinc, vitamin B12 and inositol. In particular, inositol supplementation has been found to prevent the NTD phenotype in the Grhl3 (curly tail) mutant mouse. 153 In mice, mutation identified in inositol metabolism can also lead to NTDs.4 Supplementation of loop-tail mouse diet with these micronutrients could be tested and resulting pups analyzed for an increase or a decrease in the NTDs phenotype.

190

7.1.3 Final conclusions

Future studies on Vangl genes are essential to fully comprehend their role in neural tube formation and NTDs. In recent years, epigenetic changes have been found strongly associated to diseases such as cancer and some have suggested with NTDs as well. A newly published study tracked the changes in DNA-methylation of several genes in children with NTDs and found a borderline association with Vangl genes. 154In the future, it will be especially important to further study changes such as methylation and histone modifications in these genes associated with NTDs. Additionally, recent advances in genomic medicine will allow going from single gene analysis to searching for deleterious mutations in the whole exome or the whole genome of patients with NTD. Mutations in Vangl genes account only for a subset of NTDs and it is hypothesised that no single gene mutation will ever be uncovered to explain the cause of this multifactorial disease. Whole genome profiling will be able to uncover previously unidentified genes and pathways involved in the pathogenesis of NTDs. Importantly, it will be essential to also further validate all the variants identified in sporadic and familial cases of NTDs by developing additional in vitro assays and also by testing likely pathogenic ones in animal models.

191

Original Contributions to Knowledge

1. Established the secondary structure of VANGL proteins, including the number, position and polarity of transmembrane domains.

2. Developed in vitro biochemical assays in stably transfected polarized epithelial MDCK cells to characterize functional aspects of VANGL proteins, including polarized expression, membrane targeting, subcellular localization, stability and half-life.

3. Determined the molecular mechanism for the basis of loss of VANGL protein function for the three independent Lp-associated alleles (D255E, R259L, S464N)

4. Characterized the functional properties of two mutations (R181Q and R274Q) identified in VANGL1 in human patients with NTDs.

5. Determined that sequence determinants responsible for targeting VANGL proteins to the membrane are not found in the amino terminus, but are instead restricted to the carboxy terminus, including the predicted PDZBM motif at the C-terminus.

192

References

(1) Copp, A. J.; Greene, N. D. Genetics and development of neural tube defects. The Journal of pathology 2010, 220, 217-230. (2) Niswander, C. P. a. L. Neural-Tube Defects. Neural Circuit Development and Function in the Brain: Comprehensive Developmental Neuroscience 2013, 3, 503-519. (3) Wallingford, J. B. Neural tube closure and neural tube defects: studies in animal models reveal known knowns and known unknowns. American journal of medical genetics. Part C, Seminars in medical genetics 2005, 135C, 59-68. (4) Harris, M. J.; Juriloff, D. M. An update to the list of mouse mutants with neural tube closure defects and advances toward a complete genetic perspective of neural tube closure. Birth defects research. Part A, Clinical and molecular teratology 2010, 88, 653-669. (5) Torban, E.; Iliescu, A.; Gros, P. An expanding role of Vangl proteins in embryonic development. Current topics in developmental biology 2012, 101, 237-261. (6) Van Allen, M. I.; Kalousek, D. K.; Chernoff, G. F.; Juriloff, D.; Harris, M.; McGillivray, B. C.; Yong, S. L.; Langlois, S.; MacLeod, P. M.; Chitayat, D.; et al. Evidence for multi-site closure of the neural tube in humans. American journal of medical genetics 1993, 47, 723-743. (7) Greene, N. D.; Stanier, P.; Copp, A. J. Genetics of human neural tube defects. Human molecular genetics 2009, 18, R113-129. (8) Kibar, Z.; Capra, V.; Gros, P. Toward understanding the genetic basis of neural tube defects. Clinical genetics 2007, 71, 295-310. (9) Wallingford, J. B.; Niswander, L. A.; Shaw, G. M.; Finnell, R. H. The continuing challenge of understanding, preventing, and treating neural tube defects. Science 2013, 339, 1222002. (10) Wu, G.; Huang, X.; Hua, Y.; Mu, D. Roles of planar cell polarity pathways in the development of neural [correction of neutral] tube defects. Journal of biomedical science 2011, 18, 66. (11) Massarwa, R.; Ray, H. J.; Niswander, L. Morphogenetic movements in the neural plate and neural tube: mouse. Wiley interdisciplinary reviews. Developmental biology 2014, 3, 59-68. (12) De Marco, P.; Merello, E.; Cama, A.; Kibar, Z.; Capra, V. Human neural tube defects: genetic causes and prevention. BioFactors 2011, 37, 261-268.

193

(13) Hollander, L. S. a. W. Heredity loop-tail in the house mouse accompaniedby interpeforate vagina and with lethal craniorachischisis when homozygous. Journal of Heredity 1949, 40, 329-334. (14) Greene, N. D.; Gerrelli, D.; Van Straaten, H. W.; Copp, A. J. Abnormalities of floor plate, notochord and somite differentiation in the loop-tail (Lp) mouse: a model of severe neural tube defects. Mechanisms of development 1998, 73, 59-72. (15) Kibar, Z.; Gauthier, S.; Lee, S. H.; Vidal, S.; Gros, P. Rescue of the neural tube defect of loop-tail mice by a BAC clone containing the Ltap gene. Genomics 2003, 82, 397-400. (16) Kibar, Z.; Vogan, K. J.; Groulx, N.; Justice, M. J.; Underhill, D. A.; Gros, P. Ltap, a mammalian homolog of Drosophila Strabismus/Van Gogh, is altered in the mouse neural tube mutant Loop-tail. Nature genetics 2001, 28, 251-255. (17) Kibar, Z.; Underhill, D. A.; Canonne-Hergaux, F.; Gauthier, S.; Justice, M. J.; Gros, P. Identification of a new chemically induced allele (Lp(m1Jus)) at the loop-tail locus: morphology, histology, and genetic mapping. Genomics 2001, 72, 331-337. (18) Torban, E.; Patenaude, A. M.; Leclerc, S.; Rakowiecki, S.; Gauthier, S.; Andelfinger, G.; Epstein, D. J.; Gros, P. Genetic interaction between members of the Vangl family causes neural tube defects in mice. Proceedings of the National Academy of Sciences of the United States of America 2008, 105, 3449-3454. (19) Torban, E.; Wang, H. J.; Patenaude, A. M.; Riccomagno, M.; Daniels, E.; Epstein, D.; Gros, P. Tissue, cellular and sub-cellular localization of the Vangl2 protein during embryonic development: effect of the Lp mutation. Gene expression patterns : GEP 2007, 7, 346-354. (20) Montcouquiol, M.; Rachel, R. A.; Lanford, P. J.; Copeland, N. G.; Jenkins, N. A.; Kelley, M. W. Identification of Vangl2 and Scrb1 as planar polarity genes in mammals. Nature 2003, 423, 173-177. (21) Phillips, H. M.; Murdoch, J. N.; Chaudhry, B.; Copp, A. J.; Henderson, D. J. Vangl2 acts via RhoA signaling to regulate polarized cell movements during development of the proximal outflow tract. Circulation research 2005, 96, 292-299. (22) Torban, E.; Wang, H. J.; Groulx, N.; Gros, P. Independent mutations in mouse Vangl2 that cause neural tube defects in looptail mice impair interaction with members of the Dishevelled family. The Journal of biological chemistry 2004, 279, 52703-52713.

194

(23) Gravel, M.; Iliescu, A.; Horth, C.; Apuzzo, S.; Gros, P. Molecular and cellular mechanisms underlying neural tube defects in the loop-tail mutant mouse. Biochemistry 2010, 49, 3445-3455. (24) Iliescu, A.; Gravel, M.; Horth, C.; Kibar, Z.; Gros, P. Loss of membrane targeting of Vangl proteins causes neural tube defects. Biochemistry 2011, 50, 795-804. (25) Reynolds, A.; McDearmid, J. R.; Lachance, S.; De Marco, P.; Merello, E.; Capra, V.; Gros, P.; Drapeau, P.; Kibar, Z. VANGL1 rare variants associated with neural tube defects affect convergent extension in zebrafish. Mechanisms of development 2010, 127, 385-392. (26) Yates, L. L.; Papakrivopoulou, J.; Long, D. A.; Goggolidou, P.; Connolly, J. O.; Woolf, A. S.; Dean, C. H. The planar cell polarity gene Vangl2 is required for mammalian kidney- branching morphogenesis and glomerular maturation. Human molecular genetics 2010, 19, 4663-4676. (27) Devenport, D.; Fuchs, E. Planar polarization in embryonic epidermis orchestrates global asymmetric morphogenesis of hair follicles. Nature cell biology 2008, 10, 1257-1268. (28) Vandenberg, A. L.; Sassoon, D. A. Non-canonical Wnt signaling regulates cell polarity in female reproductive tract development via van gogh-like 2. Development 2009, 136, 1559-1570. (29) Matsuyama, M.; Aizawa, S.; Shimono, A. Sfrp controls apicobasal polarity and oriented cell division in developing gut epithelium. PLoS genetics 2009, 5, e1000427. (30) Wang, B.; Sinha, T.; Jiao, K.; Serra, R.; Wang, J. Disruption of PCP signaling causes limb morphogenesis and skeletal defects and may underlie Robinow syndrome and brachydactyly type B. Human molecular genetics 2011, 20, 271-285. (31) Wansleeben, C.; Meijlink, F. The planar cell polarity pathway in vertebrate development. Developmental dynamics : an official publication of the American Association of Anatomists 2011, 240, 616-626. (32) Lake, B. B.; Sokol, S. Y. Strabismus regulates asymmetric cell divisions and cell fate determination in the mouse brain. The Journal of cell biology 2009, 185, 59-66. (33) Bingham, S.; Higashijima, S.; Okamoto, H.; Chandrasekhar, A. The Zebrafish trilobite gene is essential for tangential migration of branchiomotor neurons. Developmental biology 2002, 242, 149-160.

195

(34) Bingham, S. M.; Sittaramane, V.; Mapp, O.; Patil, S.; Prince, V. E.; Chandrasekhar, A. Multiple mechanisms mediate motor neuron migration in the zebrafish hindbrain. Developmental neurobiology 2010, 70, 87-99. (35) Jessen, J. R.; Topczewski, J.; Bingham, S.; Sepich, D. S.; Marlow, F.; Chandrasekhar, A.; Solnica-Krezel, L. Zebrafish trilobite identifies new roles for Strabismus in gastrulation and neuronal movements. Nature cell biology 2002, 4, 610-615. (36) Fenstermaker, A. G.; Prasad, A. A.; Bechara, A.; Adolfs, Y.; Tissir, F.; Goffinet, A.; Zou, Y.; Pasterkamp, R. J. Wnt/planar cell polarity signaling controls the anterior-posterior organization of monoaminergic axons in the brainstem. The Journal of neuroscience : the official journal of the Society for Neuroscience 2010, 30, 16053-16064. (37) Rachel, R. A.; Murdoch, J. N.; Beermann, F.; Copp, A. J.; Mason, C. A. Retinal axon misrouting at the optic chiasm in mice with neural tube closure defects. Genesis 2000, 27, 32-47. (38) Iliescu, A.; Gravel, M.; Horth, C.; Apuzzo, S.; Gros, P. Transmembrane topology of mammalian planar cell polarity protein Vangl1. Biochemistry 2011, 50, 2274-2282. (39) Antic, D.; Stubbs, J. L.; Suyama, K.; Kintner, C.; Scott, M. P.; Axelrod, J. D. Planar cell polarity enables posterior localization of nodal cilia and left-right axis determination during mouse and Xenopus embryogenesis. PloS one 2010, 5, e8999. (40) Wolff, T.; Rubin, G. M. Strabismus, a novel gene that regulates tissue polarity and cell fate decisions in Drosophila. Development 1998, 125, 1149-1159. (41) Darken, R. S.; Scola, A. M.; Rakeman, A. S.; Das, G.; Mlodzik, M.; Wilson, P. A. The planar polarity gene strabismus regulates convergent extension movements in Xenopus. The EMBO journal 2002, 21, 976-985. (42) Gubb, D.; Garcia-Bellido, A. A genetic analysis of the determination of cuticular polarity during development in Drosophila melanogaster. Journal of embryology and experimental morphology 1982, 68, 37-57. (43) McNeill, H. Planar cell polarity: keeping hairs straight is not so simple. Cold Spring Harbor perspectives in biology 2010, 2, a003376. (44) Simons, M.; Mlodzik, M. Planar cell polarity signaling: from fly development to human disease. Annual review of genetics 2008, 42, 517-540. (45) Vladar, E. K.; Antic, D.; Axelrod, J. D. Planar cell polarity signaling: the developing cell's compass. Cold Spring Harbor perspectives in biology 2009, 1, a002964.

196

(46) Singh, J.; Mlodzik, M. Planar cell polarity signaling: coordination of cellular orientation across tissues. Wiley interdisciplinary reviews. Developmental biology 2012, 1, 479-499. (47) Rawls, A. S.; Guinto, J. B.; Wolff, T. The cadherins fat and dachsous regulate dorsal/ventral signaling in the Drosophila eye. Current biology : CB 2002, 12, 1021-1026. (48) Zeidler, M. P.; Perrimon, N.; Strutt, D. I. The four-jointed gene is required in the Drosophila eye for ommatidial polarity specification. Current biology : CB 1999, 9, 1363-1372. (49) Yang, C. H.; Axelrod, J. D.; Simon, M. A. Regulation of Frizzled by fat-like cadherins during planar polarity signaling in the Drosophila compound eye. Cell 2002, 108, 675-688. (50) Bayly, R.; Axelrod, J. D. Pointing in the right direction: new developments in the field of planar cell polarity. Nature reviews. Genetics 2011, 12, 385-391. (51) Collier, S.; Gubb, D. Drosophila tissue polarity requires the cell-autonomous activity of the fuzzy gene, which encodes a novel transmembrane protein. Development 1997, 124, 4029- 4037. (52) Keller, R. E. Vital dye mapping of the gastrula and neurula of Xenopus laevis. I. Prospective areas and morphogenetic movements of the superficial layer. Developmental biology 1975, 42, 222-241. (53) Keller, R. E. Vital dye mapping of the gastrula and neurula of Xenopus laevis. II. Prospective areas and morphogenetic movements of the deep layer. Developmental biology 1976, 51, 118-137. (54) Wallingford, J. B.; Harland, R. M. Neural tube closure requires Dishevelled-dependent convergent extension of the midline. Development 2002, 129, 5815-5825. (55) Park, M.; Moon, R. T. The planar cell-polarity gene stbm regulates cell behaviour and cell fate in vertebrate embryos. Nature cell biology 2002, 4, 20-25. (56) Ciruna, B.; Jenny, A.; Lee, D.; Mlodzik, M.; Schier, A. F. Planar cell polarity signalling couples cell division and morphogenesis during neurulation. Nature 2006, 439, 220-224. (57) Saburi, S.; Hester, I.; Fischer, E.; Pontoglio, M.; Eremina, V.; Gessler, M.; Quaggin, S. E.; Harrison, R.; Mount, R.; McNeill, H. Loss of Fat4 disrupts PCP signaling and oriented cell division and leads to cystic kidney disease. Nature genetics 2008, 40, 1010-1015. (58) Heydeck, W.; Zeng, H.; Liu, A. Planar cell polarity effector gene Fuzzy regulates cilia formation and Hedgehog signal transduction in mouse. Developmental dynamics : an official publication of the American Association of Anatomists 2009, 238, 3035-3042.

197

(59) Gray, R. S.; Abitua, P. B.; Wlodarczyk, B. J.; Szabo-Rogers, H. L.; Blanchard, O.; Lee, I.; Weiss, G. S.; Liu, K. J.; Marcotte, E. M.; Wallingford, J. B.; Finnell, R. H. The planar cell polarity effector Fuz is essential for targeted membrane trafficking, ciliogenesis and mouse embryonic development. Nature cell biology 2009, 11, 1225-1232. (60) Park, T. J.; Haigo, S. L.; Wallingford, J. B. Ciliogenesis defects in embryos lacking inturned or fuzzy function are associated with failure of planar cell polarity and Hedgehog signaling. Nature genetics 2006, 38, 303-311. (61) Lu, X.; Borchers, A. G.; Jolicoeur, C.; Rayburn, H.; Baker, J. C.; Tessier-Lavigne, M. PTK7/CCK-4 is a novel regulator of planar cell polarity in vertebrates. Nature 2004, 430, 93-98. (62) Yen, W. W.; Williams, M.; Periasamy, A.; Conaway, M.; Burdsal, C.; Keller, R.; Lu, X.; Sutherland, A. PTK7 is essential for polarized cell motility and convergent extension during mouse gastrulation. Development 2009, 136, 2039-2048. (63) Murdoch, J. N.; Henderson, D. J.; Doudney, K.; Gaston-Massuet, C.; Phillips, H. M.; Paternotte, C.; Arkell, R.; Stanier, P.; Copp, A. J. Disruption of scribble (Scrb1) causes severe neural tube defects in the circletail mouse. Human molecular genetics 2003, 12, 87-98. (64) Merte, J.; Jensen, D.; Wright, K.; Sarsfield, S.; Wang, Y.; Schekman, R.; Ginty, D. D. Sec24b selectively sorts Vangl2 to regulate planar cell polarity during neural tube closure. Nature cell biology 2010, 12, 41-46; sup pp 41-48. (65) Wansleeben, C.; Feitsma, H.; Montcouquiol, M.; Kroon, C.; Cuppen, E.; Meijlink, F. Planar cell polarity defects and defective Vangl2 trafficking in mutants for the COPII gene Sec24b. Development 2010, 137, 1067-1073. (66) Ybot-Gonzalez, P.; Savery, D.; Gerrelli, D.; Signore, M.; Mitchell, C. E.; Faux, C. H.; Greene, N. D.; Copp, A. J. Convergent extension, planar-cell-polarity signalling and initiation of mouse neural tube closure. Development 2007, 134, 789-799. (67) Song, H.; Hu, J.; Chen, W.; Elliott, G.; Andre, P.; Gao, B.; Yang, Y. Planar cell polarity breaks bilateral symmetry by controlling ciliary positioning. Nature 2010, 466, 378-382. (68) Yin, H.; Copley, C. O.; Goodrich, L. V.; Deans, M. R. Comparison of phenotypes between different vangl2 mutants demonstrates dominant effects of the Looptail mutation during hair cell development. PloS one 2012, 7, e31988. (69) Tissir, F.; Goffinet, A. M. Planar cell polarity signaling in neural development. Current opinion in neurobiology 2010, 20, 572-577.

198

(70) Wang, J.; Mark, S.; Zhang, X.; Qian, D.; Yoo, S. J.; Radde-Gallwitz, K.; Zhang, Y.; Lin, X.; Collazo, A.; Wynshaw-Boris, A.; Chen, P. Regulation of polarized extension and planar cell polarity in the cochlea by the vertebrate PCP pathway. Nature genetics 2005, 37, 980-985. (71) Etheridge, S. L.; Ray, S.; Li, S.; Hamblet, N. S.; Lijam, N.; Tsang, M.; Greer, J.; Kardos, N.; Wang, J.; Sussman, D. J.; Chen, P.; Wynshaw-Boris, A. Murine dishevelled 3 functions in redundant pathways with dishevelled 1 and 2 in normal cardiac outflow tract, cochlea, and neural tube development. PLoS genetics 2008, 4, e1000259. (72) Wang, J.; Hamblet, N. S.; Mark, S.; Dickinson, M. E.; Brinkman, B. C.; Segil, N.; Fraser, S. E.; Chen, P.; Wallingford, J. B.; Wynshaw-Boris, A. Dishevelled genes mediate a conserved mammalian PCP pathway to regulate convergent extension during neurulation. Development 2006, 133, 1767-1778. (73) Curtin, J. A.; Quint, E.; Tsipouri, V.; Arkell, R. M.; Cattanach, B.; Copp, A. J.; Henderson, D. J.; Spurr, N.; Stanier, P.; Fisher, E. M.; Nolan, P. M.; Steel, K. P.; Brown, S. D.; Gray, I. C.; Murdoch, J. N. Mutation of Celsr1 disrupts planar polarity of inner ear hair cells and causes severe neural tube defects in the mouse. Current biology : CB 2003, 13, 1129-1133. (74) Yu, H.; Smallwood, P. M.; Wang, Y.; Vidaltamayo, R.; Reed, R.; Nathans, J. Frizzled 1 and frizzled 2 genes function in palate, ventricular septum and neural tube closure: general implications for tissue fusion processes. Development 2010, 137, 3707-3717. (75) Lei, Y. P.; Zhang, T.; Li, H.; Wu, B. L.; Jin, L.; Wang, H. Y. VANGL2 mutations in human cranial neural-tube defects. The New England journal of medicine 2010, 362, 2232-2235. (76) Kibar, Z.; Bosoi, C. M.; Kooistra, M.; Salem, S.; Finnell, R. H.; De Marco, P.; Merello, E.; Bassuk, A. G.; Capra, V.; Gros, P. Novel mutations in VANGL1 in neural tube defects. Human mutation 2009, 30, E706-715. (77) Kibar, Z.; Salem, S.; Bosoi, C. M.; Pauwels, E.; De Marco, P.; Merello, E.; Bassuk, A. G.; Capra, V.; Gros, P. Contribution of VANGL2 mutations to isolated neural tube defects. Clinical genetics 2011, 80, 76-82. (78) Bartsch, O.; Kirmes, I.; Thiede, A.; Lechno, S.; Gocan, H.; Florian, I. S.; Haaf, T.; Zechner, U.; Sabova, L.; Horn, F. Novel VANGL1 Gene Mutations in 144 Slovakian, Romanian and German Patients with Neural Tube Defects. Molecular syndromology 2012, 3, 76-81. (79) Guirao, B.; Meunier, A.; Mortaud, S.; Aguilar, A.; Corsi, J. M.; Strehl, L.; Hirota, Y.; Desoeuvre, A.; Boutin, C.; Han, Y. G.; Mirzadeh, Z.; Cremer, H.; Montcouquiol, M.; Sawamoto,

199

K.; Spassky, N. Coupling between hydrodynamic forces and planar cell polarity orients mammalian motile cilia. Nature cell biology 2010, 12, 341-350. (80) Montcouquiol, M.; Sans, N.; Huss, D.; Kach, J.; Dickman, J. D.; Forge, A.; Rachel, R. A.; Copeland, N. G.; Jenkins, N. A.; Bogani, D.; Murdoch, J.; Warchol, M. E.; Wenthold, R. J.; Kelley, M. W. Asymmetric localization of Vangl2 and Fz3 indicate novel mechanisms for planar cell polarity in mammals. The Journal of neuroscience : the official journal of the Society for Neuroscience 2006, 26, 5265-5275. (81) Jones, C.; Chen, P. Primary cilia in planar cell polarity regulation of the inner ear. Current topics in developmental biology 2008, 85, 197-224. (82) Borovina, A.; Superina, S.; Voskas, D.; Ciruna, B. Vangl2 directs the posterior tilting and asymmetric localization of motile primary cilia. Nature cell biology 2010, 12, 407-412. (83) Kallay, L. M.; McNickle, A.; Brennwald, P. J.; Hubbard, A. L.; Braiterman, L. T. Scribble associates with two polarity proteins, Lgl2 and Vangl2, via distinct molecular domains. Journal of cellular biochemistry 2006, 99, 647-664. (84) Guo, Y.; Zanetti, G.; Schekman, R. A novel GTP-binding protein-adaptor protein complex responsible for export of Vangl2 from the trans Golgi network. eLife 2013, 2, e00160. (85) Kalabis, J.; Rosenberg, I.; Podolsky, D. K. Vangl1 protein acts as a downstream effector of intestinal trefoil factor (ITF)/TFF3 signaling and regulates wound healing of intestinal epithelium. The Journal of biological chemistry 2006, 281, 6434-6441. (86) Gao, B.; Song, H.; Bishop, K.; Elliot, G.; Garrett, L.; English, M. A.; Andre, P.; Robinson, J.; Sood, R.; Minami, Y.; Economides, A. N.; Yang, Y. Wnt signaling gradients establish planar cell polarity by inducing Vangl2 phosphorylation through Ror2. Developmental cell 2011, 20, 163-176. (87) Guyot, M. C.; Bosoi, C. M.; Kharfallah, F.; Reynolds, A.; Drapeau, P.; Justice, M.; Gros, P.; Kibar, Z. A novel hypomorphic Looptail allele at the planar cell polarity Vangl2 gene. Developmental dynamics : an official publication of the American Association of Anatomists 2011, 240, 839-849. (88) Shafer, B.; Onishi, K.; Lo, C.; Colakoglu, G.; Zou, Y. Vangl2 promotes Wnt/planar cell polarity-like signaling by antagonizing Dvl1-mediated feedback inhibition in growth cone guidance. Developmental cell 2011, 20, 177-191.

200

(89) Copp, A. J.; Greene, N. D.; Murdoch, J. N. The genetic basis of mammalian neurulation. Nature reviews. Genetics 2003, 4, 784-793. (90) Harris, M. J.; Juriloff, D. M. Mouse mutants with neural tube closure defects and their role in understanding human neural tube defects. Birth defects research. Part A, Clinical and molecular teratology 2007, 79, 187-210. (91) Bassuk, A. G.; Kibar, Z. Genetic basis of neural tube defects. Seminars in pediatric neurology 2009, 16, 101-110. (92) Murdoch, J. N.; Doudney, K.; Paternotte, C.; Copp, A. J.; Stanier, P. Severe neural tube defects in the loop-tail mouse result from mutation of Lpp1, a novel gene involved in floor plate specification. Human molecular genetics 2001, 10, 2593-2601. (93) Jessen, J. R.; Solnica-Krezel, L. Identification and developmental expression pattern of van gogh-like 1, a second zebrafish strabismus homologue. Gene expression patterns : GEP 2004, 4, 339-344. (94) Roszko, I.; Sawada, A.; Solnica-Krezel, L. Regulation of convergence and extension movements during vertebrate gastrulation by the Wnt/PCP pathway. Seminars in cell & developmental biology 2009, 20, 986-997. (95) Fanto, M.; McNeill, H. Planar polarity from flies to vertebrates. Journal of cell science 2004, 117, 527-533. (96) Bastock, R.; Strutt, H.; Strutt, D. Strabismus is asymmetrically localised and binds to Prickle and Dishevelled during Drosophila planar polarity patterning. Development 2003, 130, 3007-3014. (97) Strutt, H.; Strutt, D. Asymmetric localisation of planar polarity proteins: Mechanisms and consequences. Seminars in cell & developmental biology 2009, 20, 957-963. (98) Chacon-Heszele, M. F.; Chen, P. Mouse models for dissecting vertebrate planar cell polarity signaling in the inner ear. Brain research 2009, 1277, 130-140. (99) Guo, N.; Hawkins, C.; Nathans, J. Frizzled6 controls hair patterning in mice. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, 9277-9281. (100) Czachorowski, M.; Lam-Yuk-Tseung, S.; Cellier, M.; Gros, P. Transmembrane topology of the mammalian Slc11a2 iron transporter. Biochemistry 2009, 48, 8422-8434.

201

(101) Songyang, Z.; Fanning, A. S.; Fu, C.; Xu, J.; Marfatia, S. M.; Chishti, A. H.; Crompton, A.; Chan, A. C.; Anderson, J. M.; Cantley, L. C. Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 1997, 275, 73-77. (102) Saras, J.; Heldin, C. H. PDZ domains bind carboxy-terminal sequences of target proteins. Trends in biochemical sciences 1996, 21, 455-458. (103) Frey, L.; Hauser, W. A. Epidemiology of neural tube defects. Epilepsia 2003, 44 Suppl 3, 4-13. (104) Taylor, J.; Abramova, N.; Charlton, J.; Adler, P. N. Van Gogh: a new Drosophila tissue polarity gene. Genetics 1998, 150, 199-210. (105) Klein, T. J.; Mlodzik, M. Planar cell polarization: an emerging model points in the right direction. Annual review of cell and developmental biology 2005, 21, 155-176. (106) Chae, J.; Kim, M. J.; Goo, J. H.; Collier, S.; Gubb, D.; Charlton, J.; Adler, P. N.; Park, W. J. The Drosophila tissue polarity gene starry night encodes a member of the protocadherin family. Development 1999, 126, 5421-5429. (107) Feiguin, F.; Hannus, M.; Mlodzik, M.; Eaton, S. The ankyrin repeat protein Diego mediates Frizzled-dependent planar polarization. Developmental cell 2001, 1, 93-101. (108) Gubb, D.; Green, C.; Huen, D.; Coulson, D.; Johnson, G.; Tree, D.; Collier, S.; Roote, J. The balance between isoforms of the prickle LIM domain protein is critical for planar polarity in Drosophila imaginal discs. Genes & development 1999, 13, 2315-2327. (109) Jenny, A.; Reynolds-Kenneally, J.; Das, G.; Burnett, M.; Mlodzik, M. Diego and Prickle regulate Frizzled planar cell polarity signalling by competing for Dishevelled binding. Nature cell biology 2005, 7, 691-697. (110) Klingensmith, J.; Nusse, R.; Perrimon, N. The Drosophila segment polarity gene dishevelled encodes a novel protein required for response to the wingless signal. Genes & development 1994, 8, 118-130. (111) Theisen, H.; Purcell, J.; Bennett, M.; Kansagara, D.; Syed, A.; Marsh, J. L. dishevelled is required during wingless signaling to establish both cell polarity and cell identity. Development 1994, 120, 347-360. (112) Usui, T.; Shima, Y.; Shimada, Y.; Hirano, S.; Burgess, R. W.; Schwarz, T. L.; Takeichi, M.; Uemura, T. Flamingo, a seven-pass transmembrane cadherin, regulates planar cell polarity under the control of Frizzled. Cell 1999, 98, 585-595.

202

(113) Strutt, D. I. The asymmetric subcellular localisation of components of the planar polarity pathway. Seminars in cell & developmental biology 2002, 13, 225-231. (114) Axelrod, J. D. Unipolar membrane association of Dishevelled mediates Frizzled planar cell polarity signaling. Genes & development 2001, 15, 1182-1187. (115) Das, G.; Jenny, A.; Klein, T. J.; Eaton, S.; Mlodzik, M. Diego interacts with Prickle and Strabismus/Van Gogh to localize planar cell polarity complexes. Development 2004, 131, 4467- 4476. (116) Tree, D. R.; Shulman, J. M.; Rousset, R.; Scott, M. P.; Gubb, D.; Axelrod, J. D. Prickle mediates feedback amplification to generate asymmetric planar cell polarity signaling. Cell 2002, 109, 371-381. (117) Carreira-Barbosa, F.; Concha, M. L.; Takeuchi, M.; Ueno, N.; Wilson, S. W.; Tada, M. Prickle 1 regulates cell movements during gastrulation and neuronal migration in zebrafish. Development 2003, 130, 4037-4046. (118) Goto, T.; Keller, R. The planar cell polarity gene strabismus regulates convergence and extension and neural fold closure in Xenopus. Developmental biology 2002, 247, 165-181. (119) Katoh, Y.; Katoh, M. Comparative genomics on Vangl1 and Vangl2 genes. International journal of oncology 2005, 26, 1435-1440. (120) Kibar, Z.; Torban, E.; McDearmid, J. R.; Reynolds, A.; Berghout, J.; Mathieu, M.; Kirillova, I.; De Marco, P.; Merello, E.; Hayes, J. M.; Wallingford, J. B.; Drapeau, P.; Capra, V.; Gros, P. Mutations in VANGL1 associated with neural-tube defects. The New England journal of medicine 2007, 356, 1432-1437. (121) Kast, C.; Gros, P. Epitope insertion favors a six transmembrane domain model for the carboxy-terminal portion of the multidrug resistance-associated protein. Biochemistry 1998, 37, 2305-2313. (122) Jenny, A.; Mlodzik, M. Planar cell polarity signaling: a common mechanism for cellular polarization. The Mount Sinai journal of medicine, New York 2006, 73, 738-750. (123) Ellgaard, L.; Helenius, A. Quality control in the endoplasmic reticulum. Nature reviews. Molecular cell biology 2003, 4, 181-191. (124) Austin, R. C. The unfolded protein response in health and disease. Antioxidants & redox signaling 2009, 11, 2279-2287.

203

(125) Wang, Y.; Nathans, J. Tissue/planar cell polarity in vertebrates: new insights and new questions. Development 2007, 134, 647-658. (126) Zallen, J. A. Planar polarity and tissue morphogenesis. Cell 2007, 129, 1051-1063. (127) Jones, C.; Chen, P. Planar cell polarity signaling in vertebrates. BioEssays : news and reviews in molecular, cellular and developmental biology 2007, 29, 120-132. (128) Park, T. J.; Gray, R. S.; Sato, A.; Habas, R.; Wallingford, J. B. Subcellular localization and signaling properties of dishevelled in developing vertebrate embryos. Current biology : CB 2005, 15, 1039-1044. (129) McNeill, H. Planar cell polarity and the kidney. Journal of the American Society of Nephrology : JASN 2009, 20, 2104-2111. (130) Phillips, H. M.; Rhee, H. J.; Murdoch, J. N.; Hildreth, V.; Peat, J. D.; Anderson, R. H.; Copp, A. J.; Chaudhry, B.; Henderson, D. J. Disruption of planar cell polarity signaling results in congenital heart defects and cardiomyopathy attributable to early cardiomyocyte disorganization. Circulation research 2007, 101, 137-145. (131) Ambudkar, S. V.; Kim, I. W.; Sauna, Z. E. The power of the pump: mechanisms of action of P-glycoprotein (ABCB1). European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 2006, 27, 392-400. (132) Cheung, J. C.; Deber, C. M. Misfolding of the cystic fibrosis transmembrane conductance regulator and disease. Biochemistry 2008, 47, 1465-1473. (133) Lam-Yuk-Tseung, S.; Camaschella, C.; Iolascon, A.; Gros, P. A novel R416C mutation in human DMT1 (SLC11A2) displays pleiotropic effects on function and causes microcytic anemia and hepatic iron overload. Blood cells, molecules & diseases 2006, 36, 347-354. (134) Bellaiche, Y.; Beaudoin-Massiani, O.; Stuttem, I.; Schweisguth, F. The planar cell polarity protein Strabismus promotes Pins anterior localization during asymmetric division of sensory organ precursor cells in Drosophila. Development 2004, 131, 469-478. (135) Goodrich, L. V. The plane facts of PCP in the CNS. Neuron 2008, 60, 9-16. (136) Keller, R. Shaping the vertebrate body plan by polarized embryonic cell movements. Science 2002, 298, 1950-1954. (137) Choi, E.; Dial, J. M.; Jeong, D. E.; Hall, M. C. Unique D box and KEN box sequences limit ubiquitination of Acm1 and promote pseudosubstrate inhibition of the anaphase-promoting complex. The Journal of biological chemistry 2008, 283, 23701-23710.

204

(138) Juriloff, D. M.; Harris, M. J. A consideration of the evidence that genetic defects in planar cell polarity contribute to the etiology of human neural tube defects. Birth defects research. Part A, Clinical and molecular teratology 2012, 94, 824-840. (139) Copp, A. J.; Stanier, P.; Greene, N. D. Neural tube defects: recent advances, unsolved questions, and controversies. Lancet neurology 2013, 12, 799-810. (140) Goodrich, L. V.; Strutt, D. Principles of planar polarity in animal development. Development 2011, 138, 1877-1892. (141) Wu, X.; Zhao, X.; Puertollano, R.; Bonifacino, J. S.; Eisenberg, E.; Greene, L. E. Adaptor and clathrin exchange at the plasma membrane and trans-Golgi network. Molecular biology of the cell 2003, 14, 516-528. (142) Chen, B.; Mao, H. H.; Chen, L.; Zhang, F. L.; Li, K.; Xue, Z. F. Loop-tail phenotype in heterozygous mice and neural tube defects in homozygous mice result from a nonsense mutation in the Vangl2 gene. Genetics and molecular research : GMR 2013, 12, 3157-3165. (143) Belotti, E.; Polanowska, J.; Daulat, A. M.; Audebert, S.; Thome, V.; Lissitzky, J. C.; Lembo, F.; Blibek, K.; Omi, S.; Lenfant, N.; Gangar, A.; Montcouquiol, M.; Santoni, M. J.; Sebbagh, M.; Aurrand-Lions, M.; Angers, S.; Kodjabachian, L.; Reboul, J.; Borg, J. P. The human PDZome: a gateway to PSD95-Disc large-zonula occludens (PDZ)-mediated functions. Molecular & cellular proteomics : MCP 2013, 12, 2587-2603. (144) Nagaoka, T.; Ohashi, R.; Inutsuka, A.; Sakai, S.; Fujisawa, N.; Yokoyama, M.; Huang, Y. H.; Igarashi, M.; Kishi, M. The Wnt/planar cell polarity pathway component Vangl2 induces synapse formation through direct control of N-cadherin. Cell reports 2014, 6, 916-927. (145) Lindqvist, M.; Horn, Z.; Bryja, V.; Schulte, G.; Papachristou, P.; Ajima, R.; Dyberg, C.; Arenas, E.; Yamaguchi, T. P.; Lagercrantz, H.; Ringstedt, T. Vang-like protein 2 and Rac1 interact to regulate adherens junctions. Journal of cell science 2010, 123, 472-483. (146) Oh, D.; Han, S.; Seo, J.; Lee, J. R.; Choi, J.; Groffen, J.; Kim, K.; Cho, Y. S.; Choi, H. S.; Shin, H.; Woo, J.; Won, H.; Park, S. K.; Kim, S. Y.; Jo, J.; Whitcomb, D. J.; Cho, K.; Kim, H.; Bae, Y. C.; Heisterkamp, N.; Choi, S. Y.; Kim, E. Regulation of synaptic Rac1 activity, long- term potentiation maintenance, and learning and memory by BCR and ABR Rac GTPase- activating proteins. The Journal of neuroscience : the official journal of the Society for Neuroscience 2010, 30, 14134-14144.

205

(147) Yang, X. Y.; Zhou, X. Y.; Wang, Q. Q.; Li, H.; Chen, Y.; Lei, Y. P.; Ma, X. H.; Kong, P.; Shi, Y.; Jin, L.; Zhang, T.; Wang, H. Y. Mutations in the COPII vesicle component gene SEC24B are associated with human neural tube defects. Human mutation 2013, 34, 1094-1101. (148) Hoffmann, M.; Segbert, C.; Helbig, G.; Bossinger, O. Intestinal tube formation in Caenorhabditis elegans requires vang-1 and egl-15 signaling. Developmental biology 2010, 339, 268-279. (149) Allache, R.; De Marco, P.; Merello, E.; Capra, V.; Kibar, Z. Role of the planar cell polarity gene CELSR1 in neural tube defects and caudal agenesis. Birth defects research. Part A, Clinical and molecular teratology 2012, 94, 176-181. (150) Robinson, A.; Escuin, S.; Doudney, K.; Vekemans, M.; Stevenson, R. E.; Greene, N. D.; Copp, A. J.; Stanier, P. Mutations in the planar cell polarity genes CELSR1 and SCRIB are associated with the severe neural tube defect craniorachischisis. Human mutation 2012, 33, 440- 447. (151) De Marco, P.; Merello, E.; Consales, A.; Piatelli, G.; Cama, A.; Kibar, Z.; Capra, V. Genetic analysis of disheveled 2 and disheveled 3 in human neural tube defects. Journal of molecular neuroscience : MN 2013, 49, 582-588. (152) Copp, A. J. Prevention of neural tube defects: vitamins, enzymes and genes. Current opinion in neurology 1998, 11, 97-102. (153) Greene, N. D.; Copp, A. J. Inositol prevents folate-resistant neural tube defects in the mouse. Nature medicine 1997, 3, 60-66. (154) Stolk, L.; Bouwland-Both, M. I.; van Mill, N. H.; Verbiest, M. M.; Eilers, P. H.; Zhu, H.; Suarez, L.; Uitterlinden, A. G.; Steegers-Theunissen, R. P. Epigenetic profiles in children with a neural tube defect; a case-control study in two populations. PloS one 2013, 8, e78462. (155) Torban, E.; Iliescu, A.; Gros, P. Role of Vangl genes in neural tube formation and neural tube defects. Inborn Errors of Development 2014, Edition 3 (in print) (156) Murdoch JN, Damrau C, Paudyal A, Bogani D, Wells S, Greene ND, Stanier P, Copp AJ. Genetic interactions between planar cell polarity genes cause diverse neural tube defects in mice. Disease Models & Mechanisms 2014, 10, 1153-1163. (157) Iliescu A, Gravel M, Horth C, Gros P. Independent mutations at Arg181 and Arg274 of Vangl proteins that are associated with neural tube defects in humans decrease protein stability and impair membrane targeting. Biochemistry 2014, 32, 5356-5364.

206