Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The bisecting N-acetylglucosamine of N-glycans appear dispensable for development and reproduction Priatel, John Jacob 1997

You are currently on our download blacklist and unable to view media. You will be unbanned within an hour.
To un-ban yourself please visit the following link and solve the reCAPTCHA, we will then redirect you back here.

Item Metadata


831-ubc_1997-196402.pdf [ 12.71MB ]
JSON: 831-1.0088070.json
JSON-LD: 831-1.0088070-ld.json
RDF/XML (Pretty): 831-1.0088070-rdf.xml
RDF/JSON: 831-1.0088070-rdf.json
Turtle: 831-1.0088070-turtle.txt
N-Triples: 831-1.0088070-rdf-ntriples.txt
Original Record: 831-1.0088070-source.json
Full Text

Full Text

THE BISECTING N-ACETYLGLUCOSAMINE OF N-GLYCANS APPEARS DISPENSABLE FOR DEVELOPMENT AND REPRODUCTION By: John Jacob Priatel B.Sc , The University of British Columbia, 1989 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Genetics Programme We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A March 1997 © John Jacob Priatel, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study; I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Greyertcs The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT The biosynthesis of complex asparagine (N)-linked oligosaccharides i n vertebrates proceeds wi th the linkage of N-acetylglucosamine (GlcNAc) to the core mannose residues. UDP-N-acetylglucosamine:(3-D-mannoside (31-4 N-acetylglucosaminyltransferase HI (GlcNAc-TIII, E.C. catalyzes the addit ion of G l c N A c to the mannose that is itself (31-4 linked to under lying N-acetylglucosamine. GlcNAc-TIII thereby produces what is known as a "bisecting" G l c N A c linkage which can be found on various hybrid and complex N-glycans. GlcNAc-TIII can also play a regulatory role in N-glycan biosynthesis as addition of the bisecting G l c N A c eliminates the potential for a-mannosidase-II, GlcNAc-TI I , G l c N A c - T I V , G l c N A c - T V , and core al-6-fucosyltransferase to act. To investigate the physiologic relevance of GlcNAc-TIII function and bisected N-glycans, the mouse gene encoding GlcNAc-TIII (Mgat3) was cloned, characterized, and inactivated using Cre/loxP site-directed recombination. The Mgat3 gene is highly conserved i n comparison to the rat and human homologs and is normally expressed at h i g h levels in brain and kidney tissue. Using fluorescence in situ hybridization (FISH), the Mgat3 gene was regionally mapped to chromosome 15E11, near the Scn8a sodium channel gene at 15F1. Fol lowing homologous recombination in embryonic stem cells and Cre mediated gene deletion, Mgaf3-deficient mice were produced that lacked GlcNAc-TIII activity and E 4 - P H A visualized GlcNAc-bisected N - l i n k e d oligosaccharides. Such GlcNAc-TIII nu l l mice were found to be viable and reproduced normally. Moreover, such mice exhibited normal cellularity and morphology among organs including brain and kidney. N o alterations were I l l apparent in circulating leukocytes, red cells or in serum metabolite levels that reflect kidney function. Al though we find that GlcNAc-TIII and the bisecting G l c N A c of N-glycans appear dispensable for normal development and reproduction in the mouse, it may have a role in physiologic responses to pathogenic and environmental stimuli. Moreover, it may be important for some function still to be determined. iv TABLE OF CONTENTS ABSTRACT T A B L E OF CONTENTS LIST OF FIGURES LIST OF TABLES ABBREVIATIONS P R E F A C E Thesis Format x List of Publications arising from the work of this thesis x Contributions of Co-Authors to the Work Presented in this Thesis x i A C K N O W L E D G E M E N T xii CHAPTER 1. Introduction 1.1 Introduction and Thesis Objectives 1 1.2 Glycoconjugates and Vertebrate Development 3 1.2.1 The Glycosyltransferase 4 1.2.2 Roles of Protein Glycosylation 5 (a) Oligosaccharide Function Varies wi th the Protein 6 (b) Structural and Protective Roles 7 (c) Trai torous/Masking Functions 8 (d) Cel l -Cel l and Cell-Matrix Interaction 8 1.2.3 Lectins-Carbohydrate Binding Proteins 10 (a) C-Type Lectins 11 (b) I-Type Lectins 11 (c) P-Type Lectins 12 (d) Galectins 12 1.2.4 O-Glycans 13 (a) M u c i n Glycoproteins 13 (b) Proteoglycans 14 (c) O-GlcNAc-Type Glycans 14 1.2.5 N-Glycans 15 (a) N-Glycan Biosynthesis 16 (b) Formation of the Dolichol-Linked Precursor Sugar 18 (c) Transfer of the Oligosaccharide to Protein 19 (d) Tr imming of the Protein-Bound Oligosaccharide 20 (e) Formation of Hybr id and Complex N-Glycans 21 (f) Microheterogeneity 23 1.2.6 Control of N - L i n k e d Branching 23 (a) Factors Influencing Glycan Formation 23 (b) N-Acetylglucosaminyltransferase I 24 (c) Order of Act ion and Competition for Substrate 25 (d) Peptide and Its Influence on the Oligosaccharide 25 (e) Control by the Endomembrane 27 (f) Diversity of N-glycans 27 (g) Poly-N-Acetyllactosamines 28 1.2.7 N-Glycans and Disease 30 (a) Altered Branching and Malignancy 30 (b) Genetic Diseases Involving N-Glycans 32 (c) Complex N-Glycans are Required for Development 33 1.2.8 The Bisecting N-Aceytlglucosamine of N-Glycans 35 (a) Cloning of the Rat N-Acetylglucosaminyltransferase III 37 (b) H i g h GlcNAc-TIII Activity is Associated wi th Some Malignancies 38 (c) The Effects of the Bisecting N-Acetylglucosamine on Oligosaccharide Structure and Lectin Binding 39 1.3 Manipula t ing the Mouse Genome 41 1.3.1 Introduction 41 1.3.2 Embryonic Stem Cells 41 1.3.3 Homologous Recombination 42 1.3.4 The P I Bacteriophage and Cre Recombinase 45 (a) Cre and the PI phage lifecycle 46 (b) Cre and the loxP site 47 (c) The Use of Cre in Heterologous Systems 48 (d) Site-Specific Recombination in Transgenic Animals 49 (e) Cre and The Induction of Chromosomal Rearrangements 53 (f) Cre and Site-Specific Insertion 55 (g) Other Cre Delivery Systems 57 1.4 Thesis Objectives 57 CHAPTER 2. Materials and Methods 2.1 Genomic D N A isolation and analysis of the mouse Mgat3 gene 58 2.2 D N A Sequencing 59 2.3 FISH Detection and Image Analysis 59 2.4 R N A Blot Analysis 60 2.5 Targeting Vector Construction 61 2.6 Homologous Recombination in ES Cells 61 2.7 Transient Cre Transfection in ES Cells 62 2.8 Generation of Chimaeric and Mutant Mice 62 2.9 D N A Isolation 63 2.10 In Vitro Differentiation of Embryonic Stem Cel l Clones 63 2.11 GlcNAc-TIII Enzyme Assay 66 2.12 E 4 - P H A / L 4 - P H A Lectin Blotting 67 2.13 Hematology 68 2.14 Serum Chemistry 68 2.15 Flow Cytometry 69 vi CHAPTER 3. Mgat3 Cloning and the Generation of a Mgat3 Mutation in Embryonic Stem Cells 71 3.1 Introduction 3.2 Results 72 3.2.1 Genomic Library Screening wi th a Rat c D N A probe 72 3.2.2 Mgat3 Genomic Insert is in the Germline Configuration 73 3.2.3 Mgat3 is High ly Conserved and Has a Single Protein Encoding Exon 75 3.2.4 Chromosomal Localization of Mouse Mgat3 78 3.2.5 Brain and Kidney Show Highest Levels of Expression A m o n g N o r m a l Tissues Examined 80 3.2.6 The pflox Vector 81 3.2.7 Construction of the Mgat3 Targeting Vector 83 3.2.8 P C R Detection of Homologous Recombination 84 3.2.9 Confirmation of Homlogous Recombination by Southern Blotting 86 3.2.10 Three of Eight Recombinants Retained A l l Three loxP sites 88 3.2.11 Transient Cre Transfection in Embryonic Stem Cells Results in Two Types of Recombination 89 3.2.12 The Mgat3 Mutation in Embryonic Stem Cells is Associated wi th a Loss in GlcNAc-TIII Act ivi ty 92 3.2.13 The Generation of Chimaeric Animals and the Transmission of the Mutated Allele 93 3.3 Discussion 93 CHAPTER 4. Bisecting N-Acetylglucosamine of N-glycans Appears Dispensable For Development and Reproduction 4.1 Introduction 4.2 Results 4.2.1 Mgat3 knockouts are Viable and of Normal Weight 83 4.2.2 N u l l Mice Lack Hybridization to a Mgat3 Coding Sequence Probe 4.2.3 Mutant Mice are Fertile and Transmit the Mutated Allele at a Predicted Mendelian Frequency 4.2.4 Disruption of the Mgat3 gene is Associated wi th a Deficiency in GlcNAc-TIII Activi ty 4.2.5 Lack of GlcNAc-TIII Activity Correlates with a Depletion of Bisected N-Glycans 4.2.6 Serum Metabolite Analyses 4.2.7 Hematological Analyses 4.3 Discussion BIBLIOGRAPHY 96 96 97 97 99 99 102 104 105 106 111 vii LIST OF FIGURES CHAPTER 1 Figure 1.1 The three major types of asparagine-lined 15 oligosaccharides Figure 1.2 Branching of mammalian N-glycans 17 Figure 1.3 The synthesis of the oligosaccharide precursor 18 Figure 1.4 The antennae formation of N-glycans 22 Figure 1.5 Synthesis of Lewis blood group antigens 29 Figure 1.6 Maintenance of PI during lysogeny 46 Figure 1.7 The loxP site 47 Figure 1.8 Cre-mediated D N A deletion 50 Figure 1.9 Site-specific integration 55 CHAPTER 2 Figure 2.1 Embryoid body formation 66 CHAPTER 3 Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 Figure 3.9 Figure 3.10 Figure 3.11 Figure 3.12 CHAPTER 4 Figure 4.1 Heterozygous and homozygous mutations in the Mgat3 allele in intact mice 98 Figure 4.2 Mgat3 mutation is associated with a loss of GlcNAc-TIII activity 101 Figure 4.3 Loss of GlcNAc-TIII activity correlates wi th a depletion of E 4 - P H A lectin binding 103 Mgat3 insert is in germline configuration 74 Nucleotide and amino acid sequence of the mouse Mgat3 gene 75 Comparison of putative mouse, rat and human GlcNAc-TIII sequences 78 Regional chromosomal localization of the mouse Mgat3 gene by FISH 79 Expression of Mgat3 gene among normal mouse tissues 81 The pflox vector 82 Mouse Mgat3 genomic structure and targeting vector production 84 P C R detection of homologous recombination 85 Southern confirmation of homologous recombination 87 Structure and presence of loxP sites 88 The p M C - C r e Expression Vector 90 Cre-mediated recombination in embryonic stem cells 91 viii LIST OF TABLES CHAPTER 4 Table 1 Transmission of the Mgat3A allele 99 Table 2 Renal panel 104 Table 3 Peripheral Blood Hematology 105 ix ABBREVIATIONS A R S Autonomous replicating sequence A s n Asparagine B H K cells Baby hamster kidney cells C M L - B C Chronic Myelogenous Leukemia-Blast Crisis C o n A Concavalin Agglutinin C Immunoglobulin Constant Domain C R D Carbohydrate Recognition Domain D N A Deoxyribonucleic Acid E - P H A Erythrocyte-Phytohemagglutinin ES Cells Embryonic Stem Cells F A C S Fluorescence activated cell sorter FISH fluorescence in situ hybridization Fuc Fucose Gal Galactose G a l N A c N-acetylgalactosamine GlcNAc N-acetylglucosamine G l c N A c - T N-acetylglucosaminyltransferase IgSF Immunoglobulin Superfamily L - P H A Leukocyte-Phytohemagglutinin M a n Mannose M A G Myelin Associated Glycoprotein M M Multiple Myeloma P A G E Polyacrylamide gel electrophoresis PBS Phosphate buffered saline PCR polymerase chain reaction PE Phycoerythrin R E R rough endoplasmic reticulum R N A Ribonucleic Acid SDS Sodium dodecyl sulphate Ser Serine SSEA-1 Stage-Specific Embryonic Antigen-1 T h r Threonine V Immunoglobulin Variable Domain PREFACE Thesis Format I have chosen to write my thesis in the following format. The thesis is organized into four chapters. Chapter One has an overview and thesis objectives section. Chapter One also contains an introduction to the field of glycosylation, review of N - l i n k e d branching, the development of embryonic stem cell technology and applications of Cre recombinase in transgenic mice. These elements were used to create the Mgat3 mutation. Chapter Two contains the general materials and methods. Chapters Three and Four include an introduction, a results and discussion sections. Chapter Three discusses the cloning, characterization and the inactivation of the mouse Mgat3 gene. Chapter Four reviews the evidence demonstrating the fidelity of the generated mutation. This chapter also includes some analyses testing for deficits associated w i t h the mutation. The references are listed at the end in the bibliography section. Publications Arising From The Work Of This Thesis ABSTRACT: Priatel, J . J . , Sarkar, M., Schachter and Marth, J. D. (1996) Isolation, Characterization of the Mouse Mgat3 Gene: The Bisecting N-Acetylglucosamine in Asparagine-linked Oligosaccharides Appears Dispensable for Viability and Reproduction. Glycobiology 6, 750. PEER-REVIEWED: Priatel, J . J . , Sarkar, M., Schachter and Marth, J. D. (1997) Isolation, Characterization of the Mouse Mgat3 Gene: The Bisecting N-Acetylglucosamine in Asparagine-linked Oligosaccharides Appears Dispensable for Viability and Reproduction. Glycobiology 7, 45-56. xi Contributions of Co-Authors to the Work Presented in this Thesis M u c h thanks is given to Ani ta Gertz-Borowski for lending her great technical expertise. After microinjecting the targeted ES cells into host blastocysts, she implanted the embryos into pseudo-pregnant mice. The chimaeric mice generated were necessary for introducing the Mgat3 mutation into the germline. I wou ld like to give special thanks to Mohan Sarkar and Harry Schachter for their work on the GlcNAc-TIII enzyme analyses and lectin blotting. I am also indebted to the U C S D Medical Center (Hillcrest) for performing the peripheral blood hematology (Table 2) and the renal panel (Table 3). Chromosomal localization studies were performed by Zong M e i Zhang and Teresa Scheidl. They labelled the genomic insert, isolated for targeting vector construction, wi th biotin and utilized FISH analyses. X l l ACKNOWLEDGEMENTS I am very grateful to my supervisor, Dr. Jamey D. Marth, for giving me the opportunity of being a graduate student in his laboratory and for providing an excellent scientific environment. Furthermore, I wish to thank Jamey for his advice and the support he provided during difficult times. I extend my sincere appreciation to Kenneth Harder, Daniel C h u i and A n i t a Gertz-Borowski for their friendship and support. Michael Durn in and Kurt Marek for their superb technical assistance. Financial support during these studies was provided by Roman Matthew Babicki and Howard Hughes Medical Institute Fellowships. I wou ld like to thank my family and friends for their support during the work of this thesis. None of my accomplishments would have been possible without the love, support, guidance and encouragement given to me by my parents, Ivan and Ange l ina , and my brother Andrew. I would also like to express special thanks to my wife, Erica, for her constant support and encouragement and for being there during the tough times of the thesis. Introduction 1 CHAPTER 1 Introduction 1.1 Overview and Thesis Objectives The role of N-glycans in mammalian development and physiology has remained elusive even though many of the enzymes and substrates involved i n this catalysis have been wel l studied. Knowledge of the biochemistry has al lowed many of the steps to be recapitulated in vitro. Rapid changes occur to the repertoire of asparagine (N)-linked oligosaccharides during embryogenesis, differentiation and malignant transformation. The enormous energy investment to synthesize such structures as wel l as their elaborate complexity signifies that they are v i ta l . Moreover, comparison of mammalian homologs of a given gene has resulted in the observation that putative glycosylation sites are often highly conserved, thereby suggesting that oligosaccharides are biologically-active. Glycosylation provides proteins additional structural and functional variation. Furthermore, the var ia t ion in the type of linkage (a & (3), the carbon atoms involved in the linkage (1-2, 1-3, 1-4, 1-6, etc.), the monosaccharide residues used, as wel l as the possibility of branching and additional modifications make carbohydrates suitable information carriers. Since carbohydrates extend far from the surface of the cell, scientists have long suspected that these molecules serve as ligands for lectins, mediating functions i n cell-cell adhesion and recognition. Despite many advances in the understanding of glycoconjugate synthesis, the knowledge of their biology has lagged. The cloning of many glycosyltransferases Introduction 2 in recent years allows for genetic approaches to be used in the field of glycosylation. In addition, the discovery of embryonic stem (ES) cells permits the investigation of gene function in vivo. ES cells have been a boon for two unique talents they possess, the ability to undergo homologous recombination at relatively h i g h frequencies and the potential to give rise to the germline upon their re-introduction into a host blastocyst, which allows one to generate specific gene mutations and study the consequences of such losses in a whole animal. A s glycosyltransferase loss may result in embryonic lethality and/or pleiotropic changes, it wou ld be desirable to create a conditional mutation, thereby, granting the ability to disrupt the gene in a tissue or cell-subset specific manner. The Cre/loxP site-specific recombination system can be employed for gene targeting to meet the above-mentioned criterion. In this system, expression of the Cre recombinase in a cell or tissue determines whether the /oxP-flanked gene (exon) of interest is erased. Previous work has shown that complex N- l inked oligosaccharide biosynthesis is required early in embryonic development (Ioffe and Stanley, 1994; Metzler et al., 1994) although the min imal structure necessary for ontogeny is not yet defined. This regulated biosynthetic pathway often generates oligosaccharide structures bearing multiple antennae, suggesting the possibility that unique biological information is contained along each branch. The biosynthesis of complex asparagine (N)-linked oligosaccharides in vertebrates proceeds wi th the linkage of N-acetylglucosamine (GlcNAc) to the core mannose residues. U D P - N -acetylglucosamine:p-D-mannoside (31-4 N-acetylglucosaminyltransf erase III (GlcNAc-TIII, E.C. catalyzes the addition of G l c N A c to the mannose that is Introduction 3 itself (31-4 linked to underlying N-acetylglucosamine. GlcNAc-TIII thereby produces what is known as a "bisecting" G l c N A c linkage which can be found on various hybrid and complex N-glycans. GlcNAc-TIII may also play a regulatory role in N -glycan biosynthesis as addition of the bisecting G l c N A c eliminates the potential for cc-mannosidase-II, G lcNAc-TIL G lcNAc-TIV , G l c N A c - T V , and core a l -6-fucosyltransferase to act. Al though the bisecting G l c N A c has not been shown to interact wi th an animal lectin, its presence is known to have a marked effect on the binding of exogenous lectins to oligosaccharide structures. To investigate the physiologic relevance of GlcNAc-TIII function and bisected N-glycans, the mouse gene encoding GlcNAc-TIII (Mgat3) was cloned, characterized, and inactivated using Cre /loxP site-directed recombination. 1.2 Glycoconjugates and Vertebrate Development A myriad of oligosaccharide structures envelops eukaryotic cells forming a dense covering. The majority of these structures are covalently l inked to some other macromolecule, either a l ip id or protein. Unl ike the linearity of protein and D N A , the branching of these information carriers confers on them an extra degree of variation. The information encoded by these glycoconjugates is thought to be important for ontogeny as many glycosyltransferases are expressed in a tissue- and developmentally-restricted fashion. This regulation is illustrated by specific histochemical staining observed wi th different lectins and carbohydrate-recognizing Introduction 4 antibodies. Furthermore, oligosaccharide complexity is seen to increase w i t h vertebrate phylogeny suggesting involvement in ontogeny and morphogenesis. Al though glycosylation appears to be crucial for in vivo, it appears to be inconsequential for mammalian cell lines. Cel l lines defective in var ious glycosyltransferases proliferate normally and exhibit no phenotypic differences from w i l d type cells (Stanley, 1984; Stanley, 1989; Esko, 1991). For instance, the der ivat ion of a C H O cell line deficient in GlcNAc-TI (Stanley, 1992), an enzyme required for the formation of complex N-glycans, resulted in no observable phenotype as compared to the parental cells. These findings suggest that complex N-glycans are not important for the housekeeping functions of a single cell but are required for the viabili ty of the multicellular organism. O n the other hand, embryos lacking this gene died by embryonic day 10 (Ioffe and Stanley, 1994; Metzler et al, 1994). These experiments underlie the importance of studying protein glycosylation in vivo. 1.3 The Glycosyltransferase Since oligosaccharides are non-linear, it is not possible to have a template-driven system analogous to protein and D N A synthesis. Their complexity demands an assembly line-style of manufacturing. The endomembrane, referring to the endoplasmic reticulum and Golgi apparatus, fulfills this criterion wi th its membrane-bound glycosidases and glycosyltransferases acting in a sequential and competitive fashion (Schachter, 1991a). A substantial part of the m a m m a l i a n genome, greater than 250 genes, codes for enzymes participating in the biosynthesis of glycans (Varki and Marth, 1995). Furthermore, this number does not include the Introduction 5 great quantity of proteins (lectins) which recognize oligosaccharide structures. The formation of oligosaccharides do not result from the random action of glycosyltransferases. Since glycosyltransferases reside to specific sub-regions of the Golgi apparatus, they act in a specific order and time during the movement of the oligosaccharide through the lumen of the endomembrane. In addition, many glycosyltransferases, except for the enzymes that bui ld the core structures, are expressed in a tissue-specific manner. Glycosyltransferases catalyze the following reaction: nucleoti de-sugar + R-OH • R-O-sugar + nucleotide where R is either a 'free saccharide or linked to a protein (or l ipid)-bound oligosaccharide and O H is a free hydroxy 1 group. These enzymes show remarkable specificity for both the acceptor and the nucleotide donor. Al though there are exceptions, the conventional rule is that there is a single glycosyltransferase for each glycosidic linkage. Apart from a few families (eg. sialyltransferases), glycosyltransferases show virtually no sequence homology to each other (Schachter, 1994). However, they are all type II integral membrane proteins and do share some structural domains. These include a short cytoplasmic amino terminal domain, a transmembrane (anchor) domain, a proline-rich neck region and the carboxy-terminal catalytic domain. The neck, also called stem, region resides between the transmembrane and catalytic domains. Introduction 6 1.4 Roles of Protein Glycosylation The question regarding the biological roles for protein-bound oligosaccharides and why they display so much complexity have been baffling. Realizing their distal position from the cell, scientists have long speculated the involvement of carbohydrate structures in arbitrating cell-matrix and cell-cell adhesion. M u c h study now suggests a broader spectrum of functions for protein glycosylation (reviewed in Vark i , 1993). Some glycoconjugates are critical to the organism while others appear to be meaningless. A great deal of this research has been focused on indiv idual proteins and how one or more of its characteristics are affected by these modifications. To date, protein-bound oligosaccharides have been postulated to play the following roles: (1) an important structural role, (2) mediate protein folding and conformation, (3) enhance stability of the protein (ie. protect from proteases), (4) regulate the activity/function of the resident protein, (5) provide receptors for invading organisms, (6) mask receptors for invading microorganisms, (7) contribute ligands for lectin-like receptor molecules involved in protein trafficking and (8) provide ligands for either cell-cell or cell-matrix adhesion and recognition. 1.4.1 Oligosaccharide Function Varies with the Protein Asparagine-linked oligosaccharides have been the most widely studied glycan, at least in part, due to the ease of manipulating them. These strategies include the enzymatic/chemical removal, drugs which block processing reactions and the elimination of glycosylation sites via site-directed mutagenesis. The results Introduction 7 of these investigations indicate that the repercussions of altering glycosylation can be highly variable. The effects on some proteins can be undetectable, some can be quite dramatic whereas others may exhibit an intermediate loss of function. For example, glycosylation confers different properties on the two cytokines, erythropoietin and interferon y. Changes in erythropoietin's glycosylation have many effects including failure to secrete, decreased stability and decreased biological activity (Dube et al, 1988; Takeuchi et al, 1989; Fukuda et al, 1989; Takeuchi et al., 1990; Imai et al, 1990; Fukuda et al, 1990; Tsuda et al, 1990; Narh i et al, 1991; Wasley et al, 1991; Yamaguchi et al, 1991; Delorme et al, 1992). O n the other hand, interferon y7s antiviral activity and target cell specificity appear to be unaffected by altered glycosylation (Kelker et al, 1983) In addition, in vitro studies have o n occasion given opposite results from those observed in vivo. Therefore, the consequences of the above-mentioned studies illustrate that the effects of glycosylation on a protein are difficult, if not impossible, to predict. 1.4.2 Structural and Protective Roles One of the best known functions of protein-bound oligosaccharides is to mediate protein folding, thereby enabling the peptide to adopt the correct conformation. Illustrating this fact, a number of deglycosylated proteins are insoluble, fail to fold properly and/or fail to exit the ER. Whi le carbohydrates may be important for the structure of some proteins, they can also confer a protective function. For instance, the mucins, carbohydrate-rich proteins found in abundance on epithelial cells, are thought to blanket vulnerable parts of the polypeptide from Introduction 8 the attack of proteases, acid and invading pathogens. In an apparent contradiction to the above, many proteins appear to fold and function properly in the absence of glycosylation. Thus, this brings up a frequently encountered paradox i n glycobiology: oligosaccharides attached to various proteins may have wi ld ly different functions. It is believed that the structural, protective and stabilizing attributes of oligosaccharides should not have required the amount of complexity that is seen i n mammals. Nature would not be expected to devise such complex and energy-consuming patterns when simple more efficient ones would have sufficed. 1.4.3 Pathogen/Host Evolution and Oligosaccharide Variation In addition to the mediation of many vital processes, oligosaccharides are also widely known to bear ligands participating in pathogen/host adhesion. This attachment, a critical element of infection, is mediated by receptors on the surfaces of viruses, bacteria and parasites which recognize highly specific carbohydrate structures. Due to the specificity of the receptor, the host may protect itself by masking the treasonous epitope. The mask could include the addition of another monosaccharide residue or modification of an existing one. Likewise, the receptor would be predicted to evolve as the pathogens wi th the original receptor are selected against. Thus, some oligosaccharide variation may have resulted from host/pathogen relations. Hence, such variation may be protective but in the absence of pathogen have no consequence. Introduction 9 1.4.4 Cell-Cell and Cell-Matrix Interaction O n the extracellular surface, a vast array of carbohydrates can be presented. Different carbohydrate structures can be attached to a particular protein. Each of these glycosylation variants, or "glycoforms", possesses the identical polypeptide backbone. Therefore, glycosylation provides proteins additional functional and structural variation. The formation of different glycoforms allows a single protein to display multiple ligands and each of these ligands could potentially interact w i t h a different lectin receptor. Alternatively, different proteins harbouring the same oligosaccharide structure could present ligands to the same receptor. Addi t iona l ly , the temporal and spatial expression of a carbohydrate structure may dictate what potential receptors are available to interact with. Moreover, the same structure could have multiple roles and could be solely dependent on the nature of the receptor that it encounters. Oligosaccharide participation in cell-cell and cell-matrix interaction, either protein-carbohydrate or carbohydrate-carbohydrate, interaction has been demonstrated. A n example of the former, the selectin molecules have been shown to mediate the adhesion of leukocytes to endothelial cells. The leukocyte-bound L-selectin can recognize either sialyl Lewis x or sialyl Lewis 3 carbohydrate ligands o n the endothelial cell. A n example of the latter, stage-specific embryonic antigen 1 (SSEA-1), a Lewis x oligosaccharide, is transiently expressed on embryo cells. A t the beginning of the 16-cell stage, the antigen first appears and its homotypic adhesion plays an important role for the compaction of the embryo (Fenderson et al, 1984; Bird and Kimber, 1984; Eggens et al, 1989). Introduction 10 To decipher much of this three-dimensional oligosaccharide information, there must exist molecules which can recognize them. Lectins can fulfill this requirement by binding to a unique or select subset of specific structures. 1.5 Lectins-Carbohydrate Binding Proteins Lectins refer to a group of protein molecules which bind to sugar molecules on the surface of cells. The plant lectins, phytohemagglutinins, were first studied for their unusual ability to bind and clump red blood cells. Later findings indicated that lectins as a whole can bind to a myriad of carbohydrate motifs as wel l as many other types of cells in tissue and develop mentally-restricted manners. Genes encoding these molecules must have evolved quite early during phylogeny since they are found in bacteria, plants, invertebrates and vertebrates. Some hypotheses suggest that lectins are the predecessors of antibody molecules and play important roles in the immune system of plants and invertebrates by coating and/or causing aggregation of foreign organisms, thereby enhancing phagocytosis. Furthermore, lectins sparked the interest of immunologists when they were shown to induce the proliferation of T and B lymphocytes. A s many of the biochemical and morphological changes resembled antigen-stimulated immune reactions, lectins were used to study lymphocyte signaling and proliferation. Lectins were favourable over antigen treatment as they induced polyclonal activation. Based on sequence similarity of their carbohydrate recognition domains (CRD), there are four main groups of mammalian lectins, the I-Type, the P-Type, the C-Type lectins and the Galectins. Addit ional ly, the discovery of new lectins as w e l l Introduction 11 as the fact that some lectins remain to be categorized bodes wel l for the existence of even more groups. 1.5.1 C Type Lectins The C-Type lectins presently represent the largest and most diverse group of mammalian lectins (Drickamer, 1988). Common to all members, calcium ions are required for saccharide binding. The best known example of this class is the endocytotic asialoglycoprotein receptor, a molecule involved in the clearance of serum glycoproteins lacking terminal sialic acid on their complex oligosaccharides (Speiss, 1990). Other wel l known members of this group, the selectins, are cell adhesion molecules and have been wel l documented in leukocyte-endothelia interaction. 1.5.2 / Type Lectins The established members of this new family, includes CD22, CD33, Sialoadhesin and M A G , recognize sialic acid containing structures (Powell and Vark i , 1995). A l l are cell surface glycoproteins and have large cytosolic domains which bear multiple potential and established phosphorylation sites. This family of lectins has recently been named I-type due to the discovery that they are members of the immunoglobulin superfamily (IgSF). These molecules have a characteristic V j -C 2 n element in their extracellular domains. The V and C domains are so named according to homology to the domains of the immunoglobulin. Introduction 12 1.5.3 P-Type Lectins This group of lectins is commonly referred to as the mannose-6-phosphate receptors. The two different receptors, a cation-independent and a cation-dependent, are both Type I membrane proteins (Kornfeld and Mel lmann , 1989; M a et al., 1991). These receptors mediate the trafficking of lysosomal enzymes from the fratts-Golgi to the lysosomes. In addition, the cation dependent receptor can target the internalization of extracellular hydrolases from the surrounding medium. 1.5.4 Galectins In addition to the homologous C R D , constituents of this group have an affinity for (3-galactoside sugars although the specificity of each is unique (Barondes et al., 1994). Originally referred to as S-Type lectins, the nomenclature was descriptive of the fact that some members were stabilized by the presence of thiols. Al though these soluble lectins lack an apparent secretion signal, they are found extracellularly as wel l as residing in the cytoplasm. The biological roles of these molecules are not understood. Taking into account the bi- and multi-valency of some family members, they may act by cross-linking carbohydrate moieties on the surface of cells or in the extracellular matrix. Introduction 13 1.6 0-Glycans A major class of protein glycosylation involves the attachment of the carbohydrate to the hydroxyl groups of a given polypeptide. This class is made up of a number subgroups based on the type of amino acid, the monosaccharide and the linkage making up the protein-sugar junction. The mucin-type, O-GlcNAc- type and the proteo-glycans are the three most common subgroups found in animal cells. 1.6.1 Mucin Glycoproteins Tremendous diversity is displayed by the mucin-type glycoproteins w h i c h were first identified on the mucus substance mucin and thus deriving the name mucin-type for this subgroup. Mucin-type structures are initiated by the linkage of the reducing end of G a l N A c to serine or threonine of the polypeptide (Sadler, 1984). U n l i k e N-glycans, the mucin-type glycoproteins do not have a common m u l t i -saccharide core but only share this single monosaccharide unit. Four different core structures are usually observed and each of these can then be extended. The epithelial surfaces of the respiratory and gastrointestinal tract are very rich in these structures and it is hypothesized that they are critical for providing a protective barrier against proteases and other irritants. Addi t iona l roles postulated for mucin-type glycans include acting as ligands for endogenous lectins, thereby, mediating important recognition events. Likewise, mucin binding to bacterial lectins, molecules necessary for bacterial attachment, may stimulate the clearance of these pathogens from the body. Introduction 14 1.6.2 Proteoglycans Proteoglycans are known to play an critical part in maintaining tissue architecture and integrity (Roden, 1980). These structures show less structural variation than N- and Mucin-type glycans and are characterized by bearing one or more bound glycosaminoglycan chains. Glycosaminoglycans (GAGs) are a class of polysaccharides which display high molecular weight and strong negative charge. The core protein of proteoglycans can harbour from one to over a hundred of these highly acidic polysaccharides imparting potent biochemical properties. These G A G s are usually covalently linked to the protein backbone through a xylose-serine linkage. 1.6.3 O-GlcNAc-Type Glycans O - G l c N A c modification is simply the addition of a single G l c N A c residue to serine or threonine (Hart et ah, 1989). To date, this modification has only been associated wi th nuclear and cytoplasmic-based proteins. Due to its highly dynamic nature, O-GlcNAc-yla t ion has prompted comparisons to protein phosphorylat ion and is hypothesized to play a regulatory role. Recent evidence has indicated that phosphorylation and O-GlcNAc-ylation may occur at the same site wi th in the identical protein. This occurrence implies that the enzymes mediating these two events may be in competition wi th one another. Introduction 15 1.7 W-Glycans The oligosaccharides bound to proteins are classified into two groups, N -glycans and O-glycans, bearing great structural variation. In O-glycans, sugars are attached to the hydroxyl groups of serine and threonine, whereas in N-glycans, N -acetylglucosamine (GlcNAc) is linked to amide group of asparagine residues. N -linked oligosaccharides can be subdivided into three main types: high mannose, hybrid and complex. In common among these three types is the pentasaccharide Manal-6(Manal-3)Man(31-4GlcNAc(3l-4GlcNAc-Asn. This structure is often called the "invariant core" and/or "trimannosyl core" and its derivation was revealed once the biosynthetic mechanisms were understood. Knowledge of the processing is important because it unifies all the structures to a set of successive enzymatic steps. Figure 1.1 The three major types of asparagine-linked ol igosaccharides. The shaded area indicates the invariant core structure of A/-glycans. The sugars are symbolized as follows: open circles, mannose; black squares, /V-acetylglucosamine (GlcNAc); black circles, galactose; open triangle, fucose; diamonds, sialic acid. The striped black box represents the bisecting GlcNAc and its addition is catalyzed by GlcNAc-TIII. The "±" indicates that this sugar may or may not be incorporated into the structure. The three consecutive arrows signify the processing reactions that separate these structures. Introduction 16 H i g h mannose-type oligosaccharides refer to structures which have two to six mannose residues ligated to the internal core mannoses. O n the other hand, complex oligosaccharides share the internal core mannoses, the outer two a-mannoses of this structure having N-acetyllactosamine (The disaccharide of G l c N A c and galactose) attached. These N-acetyllactosamines can be further elongated by the addition of sialic acid or additional N-acetyllactosamines. Hybrid-type oligosaccharides are exactly what the name implies, they have more than three mannose residues and also have N-acetyllactosamine chain bound to the od-3 l inked mannose. Hybrid molecules also usually contain a "bisecting" N -acetylglucosamine linked pi-4 to the (3-linked mannose residue. 1.7.1 N-Glycan Biosynthesis In contrast to O-glycan synthesis which initiates and progresses one saccharide moiety at a time, N-glycan formation begins in the lumen of the rough endoplasmic reticulum wi th the en bloc transfer of a l ipid-l inked oligosaccharide to asparagine cotranslationally (Kornfeld and Kornfeld, 1985; Fukuda, 1992). The l i p id functions as a carrier and is recycled following oligosaccharide transfer. The asparagine-attached oligosaccharide is then trimmed giving rise to high mannose structures. After processing, high mannose structures are modified to yield hybrid and complex oligosaccharides. The biosynthesis of complex N- l inked branching proceeds wi th the linkage of GlcNAc(s) to the core mannose residues. In turn, these Introduction \ 7 bound G l c N A c branches are extended by the sequential addition of monosaccharides. The following gives a brief synopsis of this process. Monoantennary Biantennary Triantennary 2,4-branched Triantennary 2,6-branched Tetrantennary Figure 1.2 Branching of mammalian A/-glycans. The shaded area indicates the invariant core structure of /V-glycans. The sugars are symbolized as follows: open circles, mannose; black squares, /v-acetylglucosamine; open triangle, fucose. The "±" indicates that this sugar may or may not be incorporated into the structure. The represents the extension of these branches by monosaccharide additions. Introduction 18 1.7.2 Formation of the Dolichol-Linked Precursor Sugar Forming of sugar/ l ipid precursor begins on the cytosolic side of the rough endoplasmic reticulum (RER) and is completely autonomous from the protein that w i l l be glycosylated subsequently. Synthesis of the l ipid-l inked oligosaccharide is initiated by the transfer of GlcNAc-l-phosphate to dolichol pyrophosphate (Dol-P) yielding GlcNAc-P-P-Dol . This opening step can be blocked by the drug tunicamycin. After the attachment of this G l c N A c , the remaining core sugars which are presented by the nucleotide donors, U D P - G l c N A c and G D P - M a n , become fastened by stepwise addition eventually producing M a n 5 G l c N A c 2 - P - P - D o l . A a2 i A a3 i o o o o poo 0 O O A a3Or • • • o • • • UDP UMP J UDP UDP ^ GDP GDP dol dol ^ dol dol p-j^ * p p p p p I I I i i dol dol do do! dol 1 | AAacetylglucosamine Mannose Glucose oligosaccharide transferred from dolichol-PP to growing polypeptide Figure 1.3 The synthesis of the oligosaccharide precursor. Sugars are added on one at a time to the dolichol carrier molecule. The finished lipid-linked oligosaccharide is transferred en bloc to the nascent peptide by oligosaccharyltransferase. The dolichol product of this reaction is subsequently recycled. This heptasaccharide-lipid intermediate, M a n 5 G l c N A c 2 - P - P - D o l , is then translocated to the lumen of the RER where it is subjected to further processing. Lumen-based additions begin wi th the transfer of four oc-Mans via Dol -P-Man. The Introduction 19 completed l ip id oligosaccharide, G l c 3 M a n 9 G l c N A c 2 - P - P - D o l , is synthesized by the joining of three oc-glucosyl residues from Dol-P-Glc. Some studies suggest that glucosylation of the precursor is necessary to protect it from degradation prior to protein attachment (Hoflack, 1981) while others suggest it may be important for transfer to the protein (Murphy and Spiro, 1981). The DolP-Man-contributed-mannoses appear to have no effect on protein glycosylation (Spiro et al., 1979; Staneloni et al., 1981). Al though G l c 3 M a n 9 G l c N A c 2 - P - P - D o l is the full grown l ip id oligosaccharide found in normal cells, a G l c 3 M a n 5 G l c N A c 2 - P - P - D o l structure, a glucosylated vers ion of the heptasaccharide-lipid mentioned above, is the largest one found in a Dol-P-Man-deficient lymphoma cell line (Chapman et al., 1979) and glucose starved Chinese hamster ovary cells (Rearick, 1981). Taking into account the fact that this l ipid-l inked decasaccharide can also act as a donor substrate in protein glycosylation, it may play a role in the biosynthesis of N-glycans. This surrogate route has been termed the "alternate pathway" albeit its relative prominence is unknown. 1.7.3 Transfer of the Oligosaccharide to Protein The assembled precursor molecule is forwarded cotranslationally to asparagine residues wi th the consensus sequence Asn-X-Ser /Thr where X is any amino acid except proline or aspartic acid (Marshall, 1972). As not all sites sharing this sequence are glycosylated, a number of other factors are thought to influence the enzyme oligosaccharyltransferase and its transfer of the sugar moiety to the glycosylation site. These include the availability of glucosylated precursors, the Introduction 20 amount of oligosaccharyltransferase activity, the number of consensus sequences i n a protein and their conformational accessibility. This latter factor may prove to be most important as studies have shown the preferential glycosylation of certain sites wi th in a protein. Furthermore, since the oligosaccharide transfer occurs cotranslationally, the nascent peptide is in the process of folding and, thus, it may or may not be in a favourable conformation for the oligosaccharyltransferase. 1.7.4 Trimming of the Protein-Bound Oligosaccharide Initial processing begins wi th the removal of the three glucose residues by two RER-membrane-bound-glucosidases, cd,2 glucosidase and ocl,3 glucosidase (Grinna and Robbins, 1978; Elting et al., 1980). Prior to transiting to the cis-Golgi, an R E R mannosidase cleaves off at least one ot-mannosyl residue (Bischoff and Kornfeld, 1983). Addi t ional t r imming results in the conversion of the oligosaccharide to M a n 5 G l c N A c 2 by Golgi a l ,2 mannosidase I (Tabas and Kornfeld , 1979). N-acetylglucosaminyltransferase I (GlcNAc-TI) modifies the high mannose structure to a hybrid glycan as it traverses the medial-Golgi compartment (Harpaz and Schachter, 1980). This is a decisive step as it is required for the synthesis of either hybrid or complex N-glycans. Cells deficient in GlcNAc-TI are blocked in this pathway and can only generate high mannose structures (Li and Kornfeld, 1978; Robertson et al, 1978; Tabas et al, 1978). Lysosomal enzymes endure identical reactions as plasma membrane and secretory proteins prior to the addition of G l c N A c by GlcNAc-TI . Subsequently, a Introduction 21 specific mannose phosphorylation of these glycoproteins targets prelysosome transport (Varki, 1992). This topic w i l l not be discussed further here. 1.7.5 Formation of Hybrid and Complex N-Glycans After the addition of G l c N A c to the Mana l ,3Manf3 l ,4GlcNAc arm, one of two different reactions can occur. The bisecting G l c N A c can be added by G l c N A c -TIII, thereby, shunting the route toward hybrid structures (Narasimhan, 1982; A l l e n et al. 1984) or alternatively, Golg i oc-mannosidase II can remove the two te rmina l mannoses on the Manal ,6Man(31,4GlcNAc arm forming a monoantennary glycan (Gleeson and Schachter, 1983). After the action of a-mannosidase II, a second G l c N A c can be added by GlcNAc-TII resulting in the conversion of the monoantennary glycan to a complex biantennary structure (Harpaz and Schachter, 1980; Oppenheimer and H i l l , 1981; Oppenheimer et al., 1981). Addi t iona l branches can be generated by GlcNAc-TIV (Gleeson and Schachter, 1983; A l l e n et al., 1984) and G l c N A c - T V (Cummings et al, 1982; Brockhausen et al, 1988) which add the (31,4 and (31,6 GlcNAc- l inked residues respectively. Before the glycoprotein leaves the medial-Golgi, its branching has already been decided by the medial-Go\gi G l c N A c -transf erases. Introduction 22 CK), /V-acetylglucosamine ^ O Mannose QVJ <3 Glucose <-<-<-00''' O-O O-B-B-P-P-doi (*2 R3 « 3 r c 2 CK) ,Q O-O O-B-B-Asn O <M-<K>d 4 O' O-B-B-Asn O 1 Tl a o TIM o, • p-B-B-Asn • O O-B-B-Asn Q_ p-B-B-Asn -O i Til O-B-B-Asn te> P-B-B-Asn l-O B-O TV ^ \ T I V • / Q O « « A B ^ p-B-B-Asn " p-B-B-Asn ™vO B-O B ' w T\\\S T I V \ ^ T V \ T I I I I B6 O O B-O v Mi4 Vp-B-B-Asn b-B-B-Asn B . O-B-B-Asn l-O " " A s n " " O |TIII - J 4 O-B-B- Asn Figure 1.4 The antennae formation of A/-glycans. After the transfer of the l ip id-linked oligosaccharide to asparagine cotranslationally, trimming reactions in the Golgi and ER yield a substrate for GlcNAc-TI. The action of GlcNAc-TII-V determined the number of antennae formed. The sugars are symbolized as follows: open circles, mannose; black squares, N-acetylglucosamine; shaded triangles, glucose. The GlcNAc-Transferases I-V are denoted by T l -TV. Introduction 23 After migration to the trans-Golgi, monosaccharides are added one at a time to complete the elongation of the antennas. Possible additions include galactose, fucose, sialic acid, G l c N A c , G a l N A c and others. Moreover, the added sugar residues can undergo modifications such as phosphorylation, sulfation and acetylation. 1.7.6 Microheterogeneity The study of a number of purified proteins has revealed that an identical protein can carry a variety of oligosaccharides even when manufactured in the same cell (Schachter, 1985). Whi le this variance is quite restricted, the reason for its occurrence, termed "microheterogeneity", and whether it has functional significance is presently unknown. Since oligosaccharide synthesis does not fol low a template, microheterogeneity may result from some randomness inherent to such a system. O n the other hand, variants of a glycoprotein, called "glycoforms", produced in two distinct cell types might arise from differential glycosyltransferase gene expression and may be biologically relevant. 1.8 Control of Af-Linked Branching 1.8.1 Factors Influencing Glycan Formation Asparagine-bound oligosaccharides are subdivided into three main groups: high mannose, hybrid and complex. These differences arise from a variable amount of processing to the trimannosyl core by downstream-acting glycosidases and glycosyltransferases. Some of the factors that influence the formation of branches, Introduction 24 also called antennae, from this core structure include the relative activity and expression of glycosyltransferases within a cell, the route that is taken by the nascent peptide and the conformation of the protein backbone. In addition, the order i n which glycosyltransferases act is very important due to their strict substrate specificities (Schachter, 1991). Often, one glycosyltransferase requires the prior action of another for it to function. 1.8.2 N-Acetylglucosaminyltransferase I Since these oligosaccharides are non-linear, it is not possible to have a template-driven system analogous to protein and D N A synthesis. Therefore, different mechanisms must be regulating the assembly of N-glycans. One of the main checkpoints is controlled by the action N-acetylglucosaminyltransferases I -V (GlcNAc-TI - TV) which initiate the construction of branches extending outward from the core. Performing at a pivotal point, G lcNAc-TI executes the transition of high mannose N-glycans to hybrid and complex N-glycans by forming the inaugural branch (Schachter, 1983; Schachter 1991). The addition of this G l c N A c is pivotal due to the fact that GlcNAc-TII - T V as wel l as oc-mannosidase II cannot act in its absence. Ce l l lines deficient in GlcNAc-TI are blocked i n N-glycan synthesis resulting in the buildup of high mannose structures. Without this branch initiating step, the chain elongating reactions inserting Gal , Fuc and sialic acid cannot take place. Introduction 25 1.8.3 Order of Action and Competition for Substrate After formation of a hybrid structure, a second pivotal step in the biosynthesis of N-glycans is reached. The decision to make hybrid or complex oligosaccharides is a fork in the biosynthetic route. Taking into account that either GlcNAc-TIII or oc-mannosidase II can act at this point in the pathway, a competi t ion for the common substrate occurs. If GlcNAc-TIII works next in the procession, it transfers the bisecting G l c N A c and thus shifts synthesis towards hybrid glycans since both oc-mannosidase II and GlcNAc-TII cannot act on this product. Al though the presence of the bisecting G l c N A c terminates further branching in the medial-Golgi , it does not prevent the elongation of initiated branches in the trans-Golgi. O n the other hand, if oc-mannosidase II removes the two terminal mannoses prior to GlcNAc-TIII action, then, GlcNAc-TII is capable of transforming the hybrid glycan into a complex structure. Addi t ional complexity can be achieved by G l c N A c - T I V and G l c N A c - T V which catalyze the attachment of two more branches to the trimannosyl core. G l c N A c - T I V can transfer a G l c N A c to hybrid structures whereas G l c N A c - T V requires the prior action of GlcNAc-TII for it to catalyze its reaction. 1.8.4 Peptide and Its Influence on the Oligosaccharide Another element thought to wield the formation of N-glycans is termed "site directed processing". This element, the effects the polypeptide backbone exerts on oligosaccharide processing, was first insinuated by the evidence showing that ind iv idua l glycosylation sites tend to have characteristic oligosaccharides. The Introduction 26 polypeptide could affect either the conformation of the glycan or an interaction w i t h a glycosyltransferase or glycosidase and is demonstrated by the following example. A study on the oligosaccharides of human IgG which epitomizes this term was carried out by Savvidou et ah, 1984. These colleagues investigated the carbohydrate makeup at Asn-107 on the Light chain as wel l as Asn-297 on the Heavy chain. Al though these molecules were manufactured in the same cell, the oligosaccharide composition on the Light chain was made up of entirely bisected complex N-glycans whereas the Heavy chain consisted of mostly (73%) non-bisected N-glycans. Moreover, the two glycans could not have been buried inside the i m m u n o g l o b u l i n since they were accessible to the downstream processing enzymes, a galactosyl- and a sialyl-transferase. Therefore, this outcome raises the query of why GlcNAc-TIII wou ld act more efficiently at one site over another. The prevailing interpretation is based on the subsequent postulates: (i) oligosaccharides are free in solution and can exist in multiple conformations, (ii) processing enzymes (like GlcNAc-TIII) are capable of acting on only a subset of these conformations and (iii) the influence of the polypeptide on the N-glycan restricts the possible conformations the oligosaccharide can occupy (reviewed in Schachter, 1991). Consequently, it is suggested that the oligosaccharide at Asn-297 is held in such a conformation that it is a poor substrate for GlcNAc-TIII . Other studies have also lent support to above-mentioned hypotheses i to i i i (Srivastava et ah, 1988; L i n d h and Hindsgaul , 1991). Introduction 27 1.8.5 Control by the Endomembrane The environment of the endomembrane, the endoplasmic reticulum and Golgi apparatus, plays a major role in N-glycan biosynthesis by controlling the availability of substrates, nucleotide-sugar donors and co-factors (Kornfeld and Kornfeld, 1985; Schachter, 1986; Schachter 1991). Furthermore, the endomembrane system, the assembly line of oligosaccharide synthesis, governs the access of processing enzymes to the growing carbohydrate and the length of its resident stay. 1.8.6 Diversity of N-glycans N-glycan diversity is principally generated by two means: (i) the degree of branching and (ii) the variation in chain termination structures. These two components of N-glycan diversity are intimately related since certain terminal structures are preferentially formed on particular antennae. For example, the a2,6 sialyltransferase shows a strong preference for Gal residues l inked to the (31,2 G l c N A c - a l , 3 M a n arm, the GlcNAc-TI catalyzed branch (Joziasse et al., 1987). Therefore, the finding, termed "branch specificity", suggests that unique information is carried along each branch rather than the branches being functionally redundant. Moreover, the branch initiating GlcNAc-transferases play a large part i n determining terminal structures. After the number of branches has been settled and the molecule migrates to the frans-Golgi, the first addition is usually a Gal which is (31,4 l inked to the underlying G l c N A c . This disaccharide sequence, Gal linked to G l c N A c , is c o m m o n Introduction 28 to complex N-glycans is often referred to as an N-acetyllactosamine. The product can now be either terminated by the addition of sialic acid or be a substrate for further extension. 1.8.7 Poly-N-Acetyllactosamines Although complex N-glycans typically have one N-acetyllactosamine per antennae, there are notable exceptions which may have many multiples of this repeating disaccharide unit. The multiple repeating units, termed polylactosamines or polylactosaminoglycans, show branch specificity towards the Manocl,6 arm, particularly the G l c N A c - T V catalyzed branch. Customarily, polylactosamines tend to be modified more often than the single disaccharide N-acetyllactosamine. The polymer is synthesized by the successive action of a (31,3-N-acetylglucosaminyltransferase and the |3l,4 galactosyltransferase. Polylactosamine chains can serve as the base architecture for many cell-type specific modifications. Their large size makes them accessible ligands for lectin receptors on neighbouring cells. Therefore, polylactosamines are speculated to be important for cell-cell interaction and adhesion. Polyllactosamines, also found on glycolipids and O-glycans, were originally described on the N-glycans of human erythrocytes. Erythrocyte-specific chains are responsible for the blood group antigen I associated wi th adult cells. Progression through the first year of life converts the red blood cells from harbouring i to I . This conversion is a result of linear chains becoming branched by the expression of a Introduction 29 pi,6-N-acetylglucosaminytransferase transferring G l c N A c to the Gal residues of the chain. Further modification of these side branches gives rise to the erythroid-specific A B O blood group antigens. In contrast to red blood cells which possess branched polylactosamines, granulocytes have linear chains which are often fucosylated at their termini by an ocl,3-N-fucosyltransferase. The myeloid-restricted reaction produces a structure termed Lewis X , abbreviated L e x . If sialylation takes place prior to fucosylation, the structure formed is called sialyl-Le x . R R cc1,3Fucosyltransferase P 4 ^ D — P4 - «3 • • R - i r m • r r o LeX Sialyl-L£ Figure 1.5 The synthesis of the Lewis blood group antigens. The Lewis X structure is formed if fucosylation occurs prior to sialylation. The terminal positions of polylactosamines are preferred substrates for these reactions. The sugars are symbolized as follows: black squares, GlcNAc; black circles, galactose; diamonds, sialic acid. This terminal moiety is of particular hematological interest since it is a ligand for a group of endogenous lectins, the selectins (Springer, 1994; McEver et al, 1995). The E (Endothelia)-, L (Leukocyte)- and P (Platelet)-selectins, a family of molecules sharing a highly homologous carbohydrate recognition domain, have been shown to be critical in leukocyte-endothelial adhesion and extravasation (Mayadas et al, 1993; Arbones et al, 1994; Labow et al, 1994; Frenette et al, 1996). Furthermore, the Introduction 30 studies of mice deficient in a l ,3 fucosyltransferase Fuc-TVII, a transferase required for L e x and sialyl-Le x synthesis, revealed an essential role for this structure i n neutrophil extravasation and the leukocyte homing (Maly et al., 1996). Work on the Fuc TVII and E- and P- selectin knockouts have demonstrated that polylactosamines and the modifications they carry are necessary for leukocyte recruitment and trafficking in diseased and healthy states. 1.9 iV-Glycans and Disease 1.9.1 Altered Branching and Malignancy One of the most common changes associated wi th malignancy is the increased size of Asn-l inked oligosaccharides. Regardless of the type of transforming agent, the three principal changes, augmented branching, polylactosamine synthesis and sialylation, are frequently observed (Alhadeff, 1989). Studies looking at B H K (baby hamster kidney) cells and their polyoma transformants divulged that only G l c N A c - T V activity, among GlcNAc-TI - T V , was elevated following transfection (Yamashita et al., 1984). Due to the specificity of polylactosamines for this branch, it was thus postulated that increased G l c N A c - T V was the reason for the increase in polylactosamines present in the highly metastatic tumour cells. Other groups working from another angle util ized glycosylation inhibitors to study the role of complex N-glycans in metastasis. Two inhibitors, tunicamycin and swainsonine, were both shown to retard tumour metastasis w h e n administered to animals (Irimura et al., 1981; Humphries et al., 1986). A l t h o u g h Introduction 31 tunicamycin completely blocks N-glycan synthesis, swainsonine blocks processing of hybrid glycans to complex oligosaccharides. Therefore, the swainsonine result suggests that the terminal structure on the M a n a l , 6 arm enhances metastatic ability. To investigate how closely N-glycan structure and malignancy are l inked , Jim Dennis and colleagues selected glycosylation mutants of a metastatic t umour cell line (Dennis et al., 1987). They found that an increased amount of tetrantennary N-glycans and poly-N-acetyllactosamines are associated wi th highly metastatic clones whereas decreased amounts correlated with low metastatic potential. In addition, two of the mutants, both having reduced G l c N A c - T V activity, exhibited a dramatic loss in metastatic potential. Hence, the significance of high G l c N A c - T V activity in malignancy may be to provide the substrate necessary for increased polylactosamine synthesis. This hypothesis is attractive especially since polylactosamines show branch specificity for the GlcNAc-TV-mediated antennae. A number of investigations have noticed that G l c N A c - T V activity was elevated in fibroblast lines which had been transformed by activated oncogenes from the ras-signalling pathway (Yamashita et al., 1985; Dennis et al., 1987, 1989; L u and Chaney, 1993; Palcic et ah, 1990). To further study this (31,6 branch in cellular transformation, Demetriou et al. transfected a G l c N A c - T V expression vector into an immortalized lung epithelial cell line (Demetriou et al., 1995). Ampl i f i ed G l c N A c -T V expression resulted in the acquisition of transformation-associated characteristics. These included a reduction in cell-stratum adhesion and contact inhibit ion, increased susceptibility to apoptosis and increased tumourigenicity i n nude mice. Introduction 32 Collectively these investigations suggest an association between polylactosamines and metastatic potential but why would such an association exist? One appealing model is based on the fact that tumour metastasis resembles leukocyte adhesion at a site of inflammation. A s mentioned earlier, polylactosamines are a suitable backbone for the synthesis of L e x and sialyl-Le x determinants. Dur ing circulation in the blood, tumour cells harbouring these structures could form large platelet/tumour cell aggregates via P-selectin-mediated adhesion and become trapped in small capillaries. Activat ion of the surrounding endothelium would result in E-selectin expression, the establishment of steadfast attachment to the tumour/platelet mass and ensuing extravasation. If this adherence is found to be selectin-dependent, it may pave the way for s ia lyl-Le x analogues being used pharmaceutically. In tune wi th this model is the fact that the greater the amount of s ialyl-Le x structure on the tumour cell surface, the poorer the patient's prognosis (Nakamon, S. et al., 1993). 1.9.2 Genetic Diseases Involving N-Glycans In theory, glycosyltransferase gene dysfunction may cause disease either by inaction (recessive mutation) or by competing wi th or interfering wi th other glycosyltransferases (dominant mutation). A t this point in time, the number of known diseases resulting from defects in glycosylation is quite small. The reasons for this paucity could be many. Firstly, many of the glycosyltransferase genes have only been recently cloned, therefore, the number of associated diseases may grow substantially. Secondly, many carbohydrate determinants may be vital for Introduction 33 development, so mutation of a given gly cosy transferase gene may result i n embryonic lethality. Other explanations include the inability of clinical and research laboratories to diagnose such diseases. Presently, there are five genetic diseases associated wi th N-glycans: Inclusion Cel l (I-Cell) Disease (Varki, 1992), Leukocyte Adhes ion Deficiency Type II (LAD-2) (Philips et al., 1995), Congenital Dyserythropoiesis Type II (HEMPAS) (Fukuda, 1990) and Carbohydrate-Deficient Glycoprotein Syndromes (CDGS) Types I (Powell et al., 1994) and II (Charuk et al., 1995). I-Cell disease resembles most glycosylation disorders in that it is autosomal recessive. The G l c N A c phosphotransferase which is defective in I-Cell disease catalyzes a modification necessary for proper intracellular trafficking of lysosomal enzymes. L A D - 2 results from a defect in the synthesis of fucosylated oligosaccharides. Patients suffer from developmental defects and recurrent bacterial infections. H E M P A S (hereditary erythroblastic raultinuclearity wi th posit ive acidified serum) is distinguished by a loss of polylactosamines on erythrocyte band 3 and 4.5 glycoproteins although the primary defect has not been determined. Patients suffer from a mi ld life-long anemia. C D G S Types I and II are multisystemic diseases displaying neurological defects. Recent work on C D G S Type II has identified point mutations (single amino acid changes) in Mgatl from two unrelated patients. These results illustrate the importance of complex N-glycans in n o r m a l neurological development. Introduction 34 1.9.3 Complex N-Glycans are Required for Development Due to their terminal position on the exterior of the cell, carbohydrate moieties are thought to be critical for cell-cell recognition and adhesion. Dur ing mammal ian development, cell surface-bound-oligosaccharides display great variation and are highly dynamic. Al though this effort by the embryo implies importance, ultimate confirmation did not come unti l a study by A z i m Surani (Surani, 1979). Mouse 2-cell embryos were cultured in the presence or absence of tunicamycin, a drug which completely blocks Asn-l inked oligosaccharide synthesis. Control embryos developed to form blastocysts whereas the experimental group failed to undergo compaction or form blastocysts. The data demonstrates that these oligosaccharides are required for early development but does not answer what types of structures are important, To further define the structures necessary for ontogeny, two different groups used gene targeting technology to create a mouse mutation i n the Mgatl, the gene coding for GlcNAc-TI (Ioffe and Stanley, 1994; Metzler et al., 1994). Mice homozygous for this mutation die at mid-gestation. A t embryonic day 10 (E10), multiple pathologies are apparent. These include the failure of the neural tube to close, impaired vascularization and loss of left-right asymmetry. Since GlcNAc-TI is obligatory for the conversion of high mannose structures to hybrid and complex N-glycans, this result indicates that either hybrid, complex or both types of structures are required for development. The outcome of this experiment is quite interesting since animals deficient in GlcNAc-TI were capable of undergoing early organogenesis in the apparent absence of hybrid and complex N - l i n k e d oligosaccharides. Mgatl-mx\\ embryos may carry some residual G l c N A c - T I either Introduction 35 from oocyte formation or transport from maternal stores. If it can be proven that Mgatl-nu\\ animals never express G l c N A c - T L it w i l l demonstrate that compaction and blastocyst formation can occur without these oligosaccharides. Moreover, since the systemic knockout of Mgatl results in embryonic lethality, it is not possible to study the loss of this gene in the adult animal. Therefore, this result emphasizes the need to use the Cre /loxP system to engineer conditional mutations. Such a conditional mutation would allow one to investigate the effects of Mgatl loss in a particular tissue of an adult animal. Creation of mice deficient in GlcNAc-TII has now been achieved (R. Campbell et ah, unpublished observations). Since manyMga£2-nul l animals die neonatally and display a heart defect, it implies that complex N-glycans play a key role in mammalian development. Furthermore, the fact that branches initiated from both the M a n a l , 3 and Manocl,6 arms appear necessary for n o r m a l development suggests that each arm contains unique and important information. 1.10 The Bisecting A/-AcetyIgIucosamine of A^-Glycans A n early step in complex and hybrid N-glycan biosynthesis may be initiated by GlcNAc-TIII which adds a G l c N A c monosaccharide in (31-4 linkage to the (31-4 linked mannose of the processed mannose core (reviewed in Schachter et al., 1983). Al though the function of the resulting "bisecting" G l c N A c is not presently k n o w n , this modification can inhibit completely the action of other enzymes in subsequent N-glycan biosynthesis (oc-mannosidase-II, GlcNAc-TII , G l c N A c - T I V , G l c N A c - T V , Introduction 36 core al-6-fucosy transferase) and inhibit partially U D P - G a l : G l c N A c (31-4 galactosyltransferase (Schachter, 1986), suggesting a regulatory role in the formation and function of complex and hybrid N-glycans. Whi le bisected N - l i n k e d oligosaccharides have not thus far been shown to specifically interact w i t h endogenous lectin receptors, altered exogenous lectin binding has been reported (Cummings and Kornfeld, 1982; Narasimhan et al, 1986) as wel l as decreased accessibility of cell surface Gal residues to Gal-binding lectins in a Chinese hamster ovary (CHO) glycosylation variant line expressing a dominant mutation in G l c N A c -TIII (Campbell and Stanley, 1984; Stanley et al, 1991). GlcNAc-TIII activity was first described in the hen oviduct (Narasimhan, 1982) and has subsequently been found in various systems including the rat l ive r during hepatocarcinogenesis (Narasimhan et al., 1988b; Nishikawa et ah, 1988a; Pascale, 1989), rat kidney (Nishikawa et al, 1988b), rat brain (Nishikawa et al, 1988b), human B lymphocytes (Narasimhan et al, 1988a), HL60 cells (Koenderman et al, 1987), Novikoff ascites tumor cells (Koenderman et al, 1989) and CaCO-2 cells (Brockhausen et al, 1991). The gene encoding GlcNAc-TIII has been isolated from vertebrate genomes including the rat, human (MGAT3) , and mouse (Mgat3) (Nishikawa et al, 1992; Ihara et al, 1993; Bhaumik et al, 1995). Expression of Mgat3 R N A appears high in mouse brain and kidney tissue, while increased expression following gene transfer in cell lines has been reported to suppress cellular susceptibility to N K cell cytotoxic mechanisms, promote spleen colonizat ion, suppress lung metastatic activity of the B16 mouse melanoma and suppress expression of hepatitis B virus (Yoshimura et al, 1995a, 1996; Miyosh i et al, 1995). Introduction 37 In addition, high levels of GlcNAc-TIII activity and bisecting N-glycans have been reported in cells derived from patients wi th chronic myelogenous leukemia in blast crisis (Yoshimura et ah, 1995b,1995c). Alterations in N-glycans associated w i t h metastasis frequently involve altered branching at the trimannosyl core, for example by G l c N A c - T V mediated (31-6 G l c N A c addition (Dennis et al., 1987; Yousefi et ah, 1991; Saitoh et ah, 1992). As these enzymes can compete for c o m m o n substrates, a functional interplay in oncogenesis between GlcNAc-TIII and G l c N A c -T V thus appears possible and may be further defined by altering oligosaccharide production in the context of tumourigenic physiology. 1.10.1 Cloning of the Rat N-Acetylglucosaminyltransferase III UDP-N-acetylglucosamine:(3-D-mannoside (3-1,4-N acetylglucosaminytransferase III (Mgat3), coding for GlcNAc-TIII , was purified from rat kidney by Atsushi Nishikawa et ah (Nishikawa et ah, 1992). The purified protein was subjected to trypsin digestion and the resultant peptides were sequenced. A m i n o acid sequence was used to design oligonucleotide primers for PCR. Screening of a rat kidney c D N A library with the P C R derived probe succeeded in the retrieval of 4 positives phage clones. A compilation of the sequencing informat ion from these clones revealed an open reading frame of 1608 base pairs coding for a 536 amino acid protein. The Mgat3 c D N A was cloned into a mammalian expression vector and transfected into COS-1 cells. COS-1 cells transfected wi th Mgat3 Introduction 38 expression vector exhibited 500-3600 fold increase in GlcNAc-TIII activity as compared to cells containing a control vector alone. The Mgat3 gene has a domain structure common to other glycosyltransferases which include a short amino terminal cytoplasmic tail , transmembrane domain, a neck region and a long terminal catalytic region. The coding sequence has three putative glycosylation sites and shares no homology to other cloned glycosyltransferases. However, it does share a small motif, 10/12 consecutive amino acids, wi th human integrin P 4 subunit that resides near the catalytic domain of GlcNAc-TIII . The significance of this relation is not known. 1.10.2 High GlcNAc-TIII Activity is Strongly Associated with Some Malignancies The observation that the glycan attached to y-glutamyltranspeptidase from rat and human hepatomas contained bisecting G l c N A c while the same glycoprotein derived from normal tissues did not, first suggested that N-acetylglucosaminyltransferase III may be induced by liver carcinogenesis. To investigate the closeness of this relationship, Narasimhan et al. used an experimental model of rat hepatocarcinogenesis (Narasimhan et al., 1988). Fisher rats were administered a single dose of a carcinogen, followed by a partial hepatectomy 18 hours later and then fed a diet supplemented wi th 1% orotic acid for 32-40 weeks. Preneoplastic hepatic nodules had a greatly increased GlcNAc-TIII activity (0.78-2.18 n m o l / h / m g of protein) as compared to regenerating liver (24 h Introduction 39 after partial hepatectomy) and control liver (0.02-0.03 n m o l / h / m g of protein) w h i c h had negligible activity. Addi t ional studies wi th different rat hepatocarcinogenesis models concurred wi th this first result (Nishikawa et ah, 1988; Pascale et ah, 1989). Therefore, the increase in GlcNAc-TIII activity is independent of the type of mode l system used and appears to be connected wi th the hepatic lesions. Recently, levels of GlcNAc-TIII activity were investigated in hematological malignancies (Yoshimura et ah, 1995b and 1995c). Patients wi th chronic myelogenous leukemia in blast crisis (CML-BC) and patients wi th multiple myeloma (MM) displayed elevated GlcNAc-TIII whereas other hematological malignancies, including C M L in chronic phase, had insignificant activity. Al though the connection between high G l c N A c -TIII activity and malignancy in the above examples is not known, the bisecting G l c N A c has a profound effect on oligosaccharide structure and may alter cell-cell or cell-matrix interaction. Such effects could be responsible for some of the membrane changes seen in oncogenesis. 1.10.4 The Effect of the Bisecting N-Acetylglucosamine on Oligosaccharide Structure and Lectin Binding The effects of the bisecting G l c N A c addition to the trimannosyl core of N-glycans are manifold. The presence of the bisecting G l c N A c is known to alter the conformation of the core residues (Brisson and Carver, 1983). Such conformational changes to the core can affect its recognition by other glycosyltransferases. A s the presence of the bisecting G l c N A c can block the action of oc-mannosidase II, G l c N A c -Introduction 40 T i l , G l c N A c - T I V and G l c N A c - T V , the addition of this sugar by GlcNAc-TIII can influence the type of N- l inked structure synthesized. Moreover, changes i n oligosaccharide structure induced by GlcNAc-TIII addition could modify binding to endogenous lectins. Al though the bisecting G l c N A c has not been shown to interact wi th an endogenous lectin yet, it is known to affect the binding of exogenous lectins. E 4 -P H A , or erythroid-phytohemagglutinin, has a high affinity for bisected oligosaccharides (Cummings and Kornfeld, 1982; Green and Baenzinger, 1987; Kobata and Yamashita, 1989, 1993). This lectin is of great importance for our Mgat3-deficient mice as it can be used to probe tissues for bisected structures. Contrary to the effects on E 4 - P H A binding, the bisecting G l c N A c hinders the interaction of N-glycans wi th the lectins Concavalin A (Narasimhan et ah, 1986) and Datura Stramonium Agglutinin (Yamashita etal.,1987). 1.3 Manipulating the Mouse Genome 1.3.1 Introduction It has not been long since the thought of introducing specific gene mutations into the mouse germline would have been considered only a w i l d fantasy. However, a number of technical breakthroughs, particularly P C R and ES cell discovery and pluripotential maintenance, have turned this fantasy into reality (Palmiter and Brinster, 1986; Capecchi, 1989; Wagner, 1990). As some of these mutations resulted in complex and/or embryonic lethal phenotypes, gene targeting limitations became apparent. Addit ionally, a mutation causing embryonic lethality Introduction 41 makes it impossible to study the consequences of gene loss in tissues of an adult animal. Secondly, if, for example, a "specialized cell" of a tissue fails to develop i n the gene knockout animal, it may be difficult to determine if the deficit lies w i t h i n the "specialized cell" or its environment or both. Such scenarios beckon for the ability to perform cell- or tissue-specific gene inactivation. The application of the Cre/loxP recombination system in transgenic mice fulfills this desire (Gu et al., 1994a). In this system, the mutation is generated only when the Cre recombinase is applied. 1.3.2 Embryonic Stem Cells The discovery of embryonic stem (ES) cells has revolutionized the field of mouse genetics by allowing scientists to introduce specific gene mutations into the mouse germline (Robertson et al., 1986; Capecchi, 1989). ES cells, cells derived from the outgrowth of the inner cell mass of a 3-4 day old (most commonly "129" strain) blastocyst, are unique with respect to their totipotency, an ability to differentiate into multiple cell types whether this comes about through microinjection into a host blastocyst or induction via in vitro culture conditions (Robertson, 1987). These cells have been a boon to scientists since they can be genetically altered and selected for i n vitro and subsequently, their re-introduction into a host blastocyst can produce a mouse strain harbouring the mutation. Therefore, ES cells allow one to generate specific gene mutations and study the consequences of such losses i n a whole animal. Al though nul l mutations have been most commonly made, this Introduction 42 technology could be used to generate virtually any type of desired muta t ion (Ramirez-Solis et al, 1993). 1.3.3 Homologous Recombination Since ES cells had been shown to be capable of taking up exogenous D N A , being microinjected into a host blastocyst to form a chimaeric animal and subsequently transmitting the incorporated D N A into the mouse germline, there was great interest in achieving homologous recombination in these cells. Gene targeting, or the homologous recombination of D N A residing on the chromosome wi th newly transfected D N A sequences, would enable one to guide incoming D N A to specific chromosomal loci, thus, allowing specific gene mutations to be introduced into the mouse germline (Haber, 1992; Capecchi, 1989). K i r k Thomas and Mario Capecchi were the first to show that such an approach was possible in ES cells (Thomas and Capecchi, 1987). The altered locus, the hypoxanthine phosphoribosyl transferase (HPRT), was chosen since the gene resided on the X chromosome (only one copy to mutate in male ES cells) and that cells mutated for this gene could be selected for. Al though this experiment demonstrated that homologous recombination in ES cells is possible, it d id not answer the question of how applicable this technology would be to other genes where the mutation could not be selected. Early in the gene targeting era, scientists battled wi th l o w frequencies of homologous recombination and wondered whether some regions of the genome were amenable to targeting. Addit ional ly, the advent of P C R was of Introduction 43 great importance as it allowed one to screen pools of ES cell clones for a putative recombination event. Fol lowing many experiments, scientists started to realize the significance of using isogenic D N A to bui ld their target vectors. When the incoming D N A is derived from the same strain as the ES cell to be transfected, it is said to be isogenic. As polymorphisms exist between different mouse strains (particularly in non-coding regions), non-isogenic D N A constructs, vectors derived from different strains as the ES cell line, are thought to homologously recombine less efficiently since they have shorter lengths of perfect homology (Riele et al., 1992; van Deursen et al., 1992; van Deursen and Wieringa, 1992; Ramirez-Solis et al., 1993). A s nearly all ES cells are derived from the 129 strain of mouse, one w o u l d screen a 129 mouse genomic library for vector construction. Frequencies of gene targeting wi th isogenic D N A are now so high that many scientists have abandoned the negative selection strategy (Ramirez-Solis et al., 1993). These frequencies are often 1/10 G418 resistant colonies and seldom lower than 1 in a hundred (Riele et al., 1992; Hasty et al., 1994; laboratory observations). Conventional gene targeting vectors mutate the gene by either insertion of a selectable marker in an early or critical exon or by replacement of a genomic fragment containing coding sequence wi th the selectable marker. Aside from selection, the marker also functions to stop translation of the endogenous message via the placement of a stop codon and its poly A signal may also terminate transcription of the locus if it is pointed in the same direction. The design usually includes flanking the marker cassette wi th one long arm and a short arm of genomic D N A . The short arm is util ized for P C R screening. The long arm is greater than 2 kilobases whereas the short arm is 0.5 to 2 Introduction 44 kilobases in length (Hasty et ah, 1991; Ramirez-Solis et ah, 1993). If Southern blotting is used for detection, then two long arms of sequence are applied. To use the Cre /loxP recombination system to create a conditional mutation, one selects a critical element of the desired gene to flank with loxP sites which is in turn placed between flanking genomic DNA (Marth, 1996). Cre action potentiates the mutation by deleting the loxP flanked sequence. As most genes contain multiple exons and span large distances, it is not usually practical to flank the entire protein encoding sequence with loxP sites. Therefore, it is necessary to choose the gene (exon) to be flanked wisely. To create a null mutation, it is advantageous to flank a 5' exon. Alternatively, an exon coding for a domain required for the protein's function can be used. Even more desirable, the deletion of this exon disrupts the reading frame of the gene and, thereby, inserting a nonsense mutation (stop codon) into the downstream message. 1.3.4 The PI Bacteriophage and Cre Recombinase PI is a temperate bacteriophage which has a genome of approximately 90 kilobases of double-stranded DNA (Sternberg and Hoess, 1983). About 7-12% of its terminal D N A is cyclically permuted as it acts as a general transducing phage, snatching up adjacent E. coli DNA. The great majority of the phage's genome is involved in vegetative growth although there are several regions which encode gene products participating in the establishment and maintenance of the prophage. During lysogeny, the PI bacteriophage's vegetative functions are repressed and it exists in E. coli as a single copy plasmid. Interestingly, PI has a linear genetic map Introduction 45 whereas other viruses wi th a similar D N A organization have a circular genetic map. This linearity of the PI genetic map suggests either a large segment of D N A is without genetic markers or that the plasmid contains a locus which undergoes frequent recombination, a so-called "hot spot". The latter hypothesis turns out to be correct: the P I phage contains a recombination recognition sequence, loxP (locus of X-ing-over), and a recombinase enzyme, Cre. Cre (causes recombination) is a member of the integrase family of recombinases which includes FLP and R of yeast. These members demonstrate remarkable similarity with regard to the types of reaction they carry out, the structure of their recognition sequences and in mechanism of action (Kilby et al., 1993). Recombination can be intermolecular or intramolecular and can be on a supercoiled, relaxed circle or linear D N A . When two loxP sites are in direct repeat on a circular molecule, Cre-mediated recombination removes the intervening D N A and generates two circular molecules. Cre is also capable of negotiating the reverse reaction, i.e. recombining sites on two separate molecules to produce a plasmid dimer. Dur ing intramolecular recombination on a linear molecule, the in tervening D N A is excised when two loxP sites are in direct repeat wi th one another (head to tail), whereas the D N A is inverted when the loxP sites are in reverse orientation (head to head; tail to tail) with respect to one another. The proficiency of Cre to delete regions of D N A from the genome has empowered geneticists wi th a molecular switch, an ability to turn on or off a gene in a conditional manner. Introduction 46 (a) Cre and the PI phage life cycle Single copy replicons such as sex factors, antibiotic resistance elements and plasmid prophage are replicated and partitioned to progeny wi th high fidelity. Due to their single copy nature, a highly efficient mechanism must be in place to secure its existence among a dividing bacterial population. The PI bacteriophage has a buil t - in safeguard, the Cre/loxP system, to ensure its survival . If recombination took place after replication of a single copy plasmid, the resulting dimer wou ld not be partitioned equally to daughter cells at cell division. The outcome of the aforementioned scenario is a high loss of plasmid from a growing population. The Pl-encoded Cre recombinase functions to ensure stable maintenance of the PI plasmid by resolving dimers into plasmid monomers, thereby ensuring transmission to the next generation (Figure 1.6). Cre-mediat ed Dimer Recombination Monomers Figure 1.6 Maintenance of the P1 during lysogeny. The Cre recombinase functions by resolving plasmid dimers to plasmid monomers. Introduction Al A PI phage wi th a defective Cre /loxP system is much less stably maintained than the w i l d type P I (Austin et al, 1981). (b) Cre and the loxP site A variety of bacterial, yeast and mammalian recombinase enzymes exist which cleave D N A at specific target sequences and then ligate it to the cleaved D N A at a second site. The complexity of these D N A rearrangement reactions differs substantially wi th respect to requirements for co-factors, supplementary proteins and the recombinase recognition sequences. Three of these site-specific recombinases have been used to manipulate D N A in heterologous cellular systems (Kilby et al, 1993). Two, FLP and R, are encoded by the yeasts Saccharomyces cerevisiae and Zygosaccharomyces rouxii (Broach et al, 1982; A r a k i et flZ.,1985) whereas the third, Cre, is from the PI bacteriophage. Cre and F L P are attractive to molecular biologists since they have been shown to be sufficient in themselves to catalyze recombination between specific target sites (Abremski and Hoess, 1984; Senecoff and Fox, 1986). The target sequences of the Cre protein is 34 base pairs i n length (Figure 1.7). Introduction 48 loxP site ATAACTTCGTATA ATGTATGC TATACGAAGTTAT |^ 1 3bp —8bp — • l ^ 1 3bp • ! Figure 1.7 The loxP site. This recognition sequence is composed of two 13-mer inverted repeats separated by an 8 base-pair core. Due to the 13-mer inverted repeats, the core element gives the toxP site its directionality. Also, note the "ATG" triplet found in the core as it has relevance with regard to designing a conditional mutation. The loxP is composed of two inverted 13 base pair repeats separated by the centrally located spacer of 8 base pairs. This core sequence confers the directionality of the loxP site (Hoess et al., 1986). Owing to the length of this recognition sequence, it is unlikely to occur randomly in eukaryotic genomes (A34 - 3 X 102 0). ( c ) The Use of Cre in Heterologous Systems Brian Sauer performed an experiment in the yeast Saccharomyces cerevisiae to determine whether the Cre recombinase of the PI coliphage could enter a eukaryotic nucleus, recognize histone-associated D N A which is packaged into nucleosomes and execute site-specific recombination (Brain Sauer, 1987). The design included the Cre gene, placed under the regulatory control of the G A L 1 promoter, and the target, the L E U 2 gene flanked by directly repeated loxP sites was used to complement a leu- yeast strain. Growth on glucose suppresses the G A L 1 promoter and as a consequence, the Cre gene is switched off and no recombination Introduction 49 takes place. Induction of the G A L 1 promoter by growth on galactose results in the specific excision of the LEU2 gene and loss of the LEU2-containing plasmid product from the cell (as it does not include an A R S , an autonomously replicating sequence), thereby demonstrating that Cre is not impeded by chromatin structure. By 24 hours, 98% of the induced cells required exogenous leucine for growth indicating that Cre action is highly efficient. Two of these leu- auxotrophs had the loci of recombination cloned and subjected to sequence analyses. In both cases, a precise recombination event had taken place wi th a single intact loxP site remaining. After the successes wi th the Cre recombinase system in the yeast, Brian Sauer and Nancy Henderson went on to address whether Cre could function in a mammalian cell (Sauer and Henderson, 1988). A stable mouse cell line was constructed that contains the Cre gene under the control of the C d 2 + inducible metallothionein I promoter. When D N A substrates, such as a marker gene flanked by loxP signals, were introduced into these cells, they were shown to undergo Cre-mediated site-specific recombination. These data suggest many possible uses for the Cre /loxP system in manipulating the mammalian genome, from inducing translocations to making conditional mutations in transgenic mice. ( d ) Site-Specific Recombination in Transgenic Animals After studies had shown Cre capable of mediating recombination i n heterologous systems of yeast and a mammalian cell line (Sauer, 1987; Sauer and Henderson, 1988), it sparked interest in whether Cre could act wi th in tissues of an Introduction 50 animal. Two groups generated Cre transgenic mice wi th the aim of performing chromosomal D N A recombination in vivo. One group generated a mouse strain carrying a lens-specific aA-crystal l in (maA) promoter separated from the SV40 T antigen (TAg) coding sequence by a stop sequence (Stop) (Lakso et al., 1992). Previous work demonstrated that mice bearing this transgene without the stop signal developed malignant eye tumours (Mahon et al., 1987). According to the experimental design, the silenced SV40 T antigen could only become activated if Cre deleted the loxP flanked stop sequence. Histological analyses of mice harboring both mocA-Stop-TAg and mocA-cre transgenes revealed morphological changes characteristic of proliferating lens tumours. Furthermore, the expected size banding by P C R and Southern analyses confirmed the genomic structure after recombination. Despite proving Cre's functioning in vivo, an assessment of its efficiency could not be made since transformation of a single cell should suffice for oncogenesis. Introduction 51 Lost — • Figure 1.8 Cre-mediated DNA deletion. The excision reaction is thought to be favoured over re-integration as it is a unimolecular reaction. Excision by Cre causes the circularization of the intervening DNA. This intervening sequence is lost from mammalian cells as it does not contain an origin of replication. The second group devised an alternative scheme for detecting Cre-mediated recombination. Instead of using Cre recombination to activate expression of a silenced transgene, the (3-galactosidase reporter gene was flanked by loxP signals Introduction 52 (loxP-$-gal-loxP) and would be ablated by Cre action (Orban et al, 1992). Once again, two mouse lines were constructed, one carrying the Cre gene driven by the lymphoid cell kinase (lck) thymus-specific promoter and the other bearing the marker, loxP-^-ged-loxP, also under control by lck promoter. Double transgenic mice had undergone tissue-specific recombination as only thymocytes had lost the loxP flanked (3-galactosidase gene. Moreover, using densitometry, the recombination event was found to be highly efficient and heritable as >95% of splenic T cells had lost hybridization to the (3-galactosidase gene. These two experiments have shown Cre to be operable on all n ine independent lines of transgenic mice bearing loxP targets, thereby suggesting that the majority of the mammalian genome may be accessible to Cre function. The above-mentioned experiments have served as the foundation for many present and future experiments, not only limited to the activation or deletion of specific genes. (e) Cre and the Induction of Chromosomal Rearrangements Although homologous recombination in murine ES cells has now made it possible to introduce desired gene mutations into the germline at w i l l , the technology to make large scale alterations to the mammalian genome has only recently become available. Chromosomal rearrangements, the major cause of inherited human disease and fetal loss, can have dramatic effects on the expression levels of rearranged loci (Epstein, 1986). Such rearrangements include translocations and deletions, events that are often responsible for the loss of Introduction 53 heterozygosity and are highly associated wi th neoplasia. A method for recombining mammalian chromosomes at predetermined sites would be an extraordinary tool for studying the activation of oncogenes by translocations, the relationship between chromosomal position and gene expression and the mechanisms of genomic imprinting. Initial work on Cre-mediated chromosomal recombination in yeast and plants systems boded wel l for its possible success in mammalian cells (Sauer, 1992; Q i n et al, 1993). Two groups both working wi th mouse ES cells have now succeeded i n engineering mammalian chromosomes (Smith et al, 1995; Ramirez-Solis et al, 1995). Owing to the distance between the loxP sites, the frequency of the desired recombination event was anticipated to be low enough to require a positive selection strategy. A selection scheme was based on the reconstitution of a hypoxanthine phosphoribosyltransferase (HPRT) minigene in an Hprt' ES cell l ine . Sequential gene targeting guided the loxP sites to two disparate loci, chromosomes 12 and 15, wi th in the genome. Each of the two targeting constructs contained a single loxP site wi th a selectable marker and either the 5' or 3' ends of the Hprt minigene. Due to the reversibility of the reaction, Cre was transiently transfected into ES cells bearing the inserted loxP sites and selected in H A T medium. Cells capable of growth in H A T medium had undergone the desired recombination event as a result of the reconstruction of the Hprt minigene by Cre-mediated recombination. Smith et al, util ized the above-mentioned experimental design to program a translocation similar to one found in human lymphomas, a juxtaposition of the heavy chain immunoglobul in promoter in front of the c-myc Introduction 54 proto-oncogene (Smith et al., 1995). Such translocations may compromise the developmental potential of the ES cell clone, thereby preventing the transmission of the alteration into the mouse germline. In addition, translocations induced in ES cells wou ld produce mice wi th a constitutional genetic abnormality. This is i n contrast to rearrangements associated wi th cancer, where lesions are often somatic and limited to the tumourous tissue. Rather than recapitulating a translocation present in a human malignancy, A l a n Bradley's laboratory focused their efforts on engineering a chromosome wi th a defined deletion (Ramirez et al., 1995). The region eliminated is syntenic to a region on human chromosome 17q which is believed to contain tumour suppresser genes from "loss of heterozygosity" studies. Two sequential gene targetings were carried out to deliver loxP sites to different locations on chromosome 11 giving rise to two types of clones, one in which both loxP sites are on one chromosome (Type 1) and the other which has a loxP site on each chromosome 11 (Type 2). The first clone, bearing the doubly targeted chromosome, and the second clone wi th two single targeted chromosomes yielded two types of Cre recombination: Type 1, an intrachromosomal deletion and Type 2, a balanced translocation. The ES cell harbouring the balanced translocation was capable of transmission into the germline, either passing on a chromosome 11 wi th a deletion or having a duplication for the loxP flanked sequence. The largest deletion spanned a region of 3-4 c M , approximately 10% of chromosome 11. The aforementioned deficiency results in segmental haploidy at a specified region wi th in the genome and can n o w serve as an invaluable tool for the screening of recessive mutations. Investigators Introduction 55 can study the phenotypes in mice housing the deletion or in vitro by studying cell lines derived from these mice. (f) Cre and Site-Specific Insertion Another powerful application of the Cre recombinase is to mediate site-specific integration, the inverse of loxP flanked D N A excision. It may be particularly useful when trying to introduce D N A into a cell which undergoes non-homologous recombination (random integration) at low frequency. Other potential uses include expression analyses of numerous constructs which have integrated into a single chromosomal position, thus eliminating position effects on expression. A n additional advantage over gene targeting is that it does not require the cloning of flanking sequence homologous to the destination locus. O • V » » Figure 1.9 Site-specif ic integrat ion. Pairing of the loxP (arrowheads) sites precedes the integration reaction. As this reaction results in two intact loxP sites, this reaction is fully reversible. Introduction 56 When Cre excises loxP flanked D N A , the single loxP site which remains can serve as a substrate for further recombination (see Figure 1.9). Incoming D N A which bears a loxP site can be targeted by Cre to the loxP site which has been previously inserted into the genome. Due to the reversibility of this reaction, the presence of Cre must be brief to prevent re-excision. Temporally limited Cre expression can be achieved by either lipofection wi th the purified Cre recombinase protein or by co-transfection of the loxP targeting vector wi th a Cre expression vector. Work to date has succeeded in delivering a loxP containing plasmid to previously inserted loxP sites in the yeast and mammalian genomes (Sauer and Henderson, 1990; Baubonis and Sauer, 1993). Moreover, heterologous D N A has been inserted into a loxP bearing alphaherpesvirus and the recombinant virus used as a shuttle (Sauer et ah, 1987b). (g) Other Cre Delivery Systems Alternative strategies for delivering Cre have been employed. The goal of these efforts has been to achieve greater control of Cre expression and, thus, increasing the precision of gene inactivation. Ralf K u h n and colleagues have presented a method for the inducible inactivation of a target gene (Kuhn et al., 1995). The previously discussed applications relied on a developmentally-active endogenous promoter to drive Cre expression. The drawback of such an approach is that the disruption of the target gene occurs early in development which may either (1) allow for the animal to compensate or (2) induce secondary effects and, hence, Introduction 57 make the phenotype difficult to interpret. The inducible system takes advantage of the Mxl promoter which is normally silent in healthy mice (Hug et al, 1988; Arnheiter et al., 1990). This promoter is activated by vira l infections and can direct high levels of transcription via interferon a or (3 administration. In this experiment, Mxl-Cre transgenic mice were crossed wi th a loxP flanked target bearing strain. Animals harbouring both alleles were treated wi th interferon. After a few days, Cre-mediated D N A deletion was complete in liver, nearly complete i n lymphocytes and partial in other tissues examined (Kuhn et al, 1995). 1.4 Thesis Objectives A t present it is clear that complex N- l inked oligosaccharide biosynthesis is required early in embryonic development (Ioffe and Stanley, 1994; Metzler et al, 1994). This highly regulated pathway often generates oligosaccharide structures bearing multiple antennae, suggesting the possibility that unique biological information is contained along each branch, information that may be perceived following the ablation of specific branching/diversification steps as controlled by specific glycosyltransferases and glycosidases. It is also possible that complex N-glycans may function in part through multi-valent interactions as a result of m u l t i -anntennary structures harboring identically-modified branch termini. To begin to investigate these possibilities including the biological role of the bisecting G l c N A c , we have created mice that lack a functional Mgat3 gene and are devoid of G l c N A c -Introduction 58 T i l l activity and bisecting N-glycans using the Cre-loxP site-directed mutagenesis approach (Marth, 1996). Materials and Methods CHAPTER 2 M a t e r i a l s and Methods 59 2.1 Genomic DNA Isolation and Analysis of the Mouse Mgat3 Gene A 1.6 kilobase GlcNAc-TIII-encoding D N A probe was amplified from 5 nanograms of rat genomic D N A using primers designed wi th the rat c D N A sequence (Nishikawa et ah, 1992). The primers were as follows: 5 ' T C C T A T G T C A C C T T C C C G A G A G A A C T G G C C T C C C T C A G C C C T A C C C T C A T A T C C A G C T T C 3 ' and 5 ' G C C C T C C G T T G T A T C C A A C T T G C C 3 ' . The P C R product was purified on a 1% agarose gel, verified by restriction analyses and cloned into pUC19. The cloned insert was isolated, 3 2P-labeled, and used as a probe for genomic library screening using methods and under conditions previously described (Marth et al., 1985). The genomic library which had undergone one round of amplification was derived from 129/SvJ female mouse liver D N A (Stratagene,CA). After hybridization, duplicate duralose - U V membranes (Stratagene, C A ) were washed two times at 55° C in 0.1 X SSC/0.1% SDS for 10 minutes and autoradiographed for 24 hours on Kodak X A R - 5 fi lm. Two positive clones were identified from screening 300,000 plaques o n Escherichia coli X L l - B l u e MRA(P2) (Stratagene, C A ) and both were subsequently purified to homogeneity. Phage D N A was purified from plate lysates, digested w i t h Not I, cloned into Bluescript, and subjected to restriction enzyme analyses. A 1.8 kilobase Bam H I mouse genomic D N A fragment hybridizing to the rat c D N A probe Materials and Methods 60 was subcloned into pUC19, producing pMgat3, and propagated in E. coli DH-5cc for large scale purification. D N A sequencing was initiated using M13 primers and employing dideoxy fluorescent nucleotides. Subsequent sequencing was carried out using sequence-internal oligonucleotides following synthesis and p r iming , approximately every 300 bases on both D N A strands. 2.2 DNA Sequencing Sequencing was carried out using 500 ng of template D N A plus 3.2 pmoles of relevant primer in a 12 | i L volume. 8 | i L of A B I P R I S M Dye Terminator Cycle Sequencing Ready Reaction Ki t (with Ampl iTaq DNAPolymerase , FS) was added to give a final volume of 20 uX. M J Research D N A Engine Peltier Thermal Cycler 200 was operated for the cycle sequencing. The reaction parameters included a hot start at 96° C, denaturation at 96 °C for 10 s, annealing temperature for 5 s and elongation at 60 °C for 4 min, repeated for 25 cycles, followed by cooling to 4°C. Dye nucleotide cleanup was accomplished by spin column purification using MicroSpin G50 Columns from P H A R M A C I A . Gels were run on an A B I 373A autosequencer under the following conditions: 6% acrylamide, 8.3 M urea, I X TBE gel. I X TBE as buffer system. Before loading, samples are given 3-4 uL of loading buffer (lpart 50 m M E D T A p H 8.0 in 30 m g / m L Blue Dextran solution to 5 parts deionized formamide). Materials and Methods 61 2.3 FISH Detection and Image Analysis The regional assignment of the Mgat3 allele was determined by fluorescence in situ hybridization (FISH) to normal mouse lymphocyte chromosomes counterstained wi th propid ium iodide and 4' / 6-diamidin-2-phenylindol-dihydrochloride (DAPI) following published methods (Lichter et al., 1990; Boyle et al., 1992). Biotinylated probe (entire Not I-Not I mouse genomic Mgat3 clone, see Figure 4) was prepared by nick translation and detected wi th avidin-fluorescein isothiocyanate (FITC), followed by biotinylated anti-avidin antibody and avidin-FITC. Images of metaphase preparations were captured by a thermoelectrically cooled charge coupled camera (Photometries, Tucson, A Z ) . Separate images of D A P I banded chromosomes (Heng and Tsui, 1993) and of FITC targeted chromosomes were obtained and merged electronically using image analysis software (courtesy of T i m Rand and David W a r d , Yale University, N e w Haven, CT) and pseudo colored blue (DAPI) and yellow (FITC) as described (Boyle et al., 1992). The band assignment was determined by measuring the fractional chromosome length and by analyzing the banding pattern generated by the D A P I counterstained image (Francke, 1994; ISCN, 1978). The chromosomal localization was verified by double color FISH using a gene probe known to map to chromosome 15 (Kohrman et al., 1995; Mock et al., 1994) kindly provided by Dr. M . Meisler, University of Michigan. Mgat3 probe was labeled wi th digoxygenin (DIG) and detected wi th mouse anti-DIG antibody followed by DIG-anti-mouse antibody and rhodamine-anti-DIG. The chromosome 15 probe (med /Scn8a) was biotinylated and detected wi th FITC-avidin as described above. Materials and Methods 62 2.4 RNA Blot Analysis R N A was prepared by the method of Chi rgwin et al. (1979) and subjected to formaldehyde denaturing agarose gel electrophoresis. R N A was transferred onto nitrocellulose (Schleicher & Schuell) and hybridized to a random-primed radiolabeled D N A probe wi th Klenow. The Mgat3 probe consisted of a 1.8 kilobase Bam H I fragment which encompassed the entire GlcNAc-TIII protein-coding sequence. Fol lowing hybridization and washing conditions as above, the filter was exposed to Kodak X A R - 5 film at -80° for 72 hours. Methods were as described previously (Marth etal, 1985). 2.5 Targeting Vector Construction From a phage 129/SvJ mouse genomic library, a purified clone containing a centrally located Mgat3 was used to generate a targeting vector as follows: A 1.8 kilobase Bam HI/Sma I fragment which contained the single Mgat3 protein-encoding exon was cloned into the Bam H I site of the pflox vector. The flanking 2.0 kilobase Bam H I fragment and an 8.5 kilobase Sma I fragment were subsequently cloned into the Xba I and Xho I sites of pflox, respectively. The targeting D N A construct was made linear by Not I digestion and purified by agarose gel electrophoresis. 2.6 Homologous Recombination in ES Cells Ten micrograms of targeting vector D N A was introduced into the R l ES cell line via electroporation. ES cells were plated on gelatin-coated culture plates and Materials and Methods 63 selected for 10 days wi th medium containing 150 m g / m L of G418 (Life Technologies, Grand Island, N Y ) . Homologous recombinants were initially detected by polymerase chain reaction (PCR) using a thymidine kinase promoter and Mgat3 allele specific primers Tk303: 5' T G C A A A A C C A C A C T G C T C G A T C C G 3' and GnTffl: 5' C T T C A T T T A G A G G G A G A G G G G A G A A A T T A A C T T G G 3' respectively. Putative P C R positive clones were subjected to Southern Blotting as described (Marth et al., 1985) to confirm homologous recombination had occurred. The genomic probe used was a 0.6 kb fragment which resides adjacent to the targeted Mgat3 sequence and was isolated from the phage clone by Not I (phage cloning site) and Kpn I digestion. The 0.23 kb loxP probe was'derived from pl015-lox 2 by Not I digestion and contains two loxP sites flanked by polylinker sequence, isolated previously from plox 2 as a 0.18 kb Hind III-Eco RI fragment (Orban et al., 1992) prior to blunt-end ligation into Bam H I site of pl015-ASph I. 2.7 Transient Cre Transfection in ES Cells In ES cell clone 1H8 bearing the targeted Mgat3 allele and all three loxP sites, ten micrograms of the Cre expression plasmid was transfected via electroporation. Cells were plated at low dilution and four days later the transfectants were selected for resistance to ganciclovir (2 x 10-6 M) for five days. Fol lowing selection, D N A from ganciclovir resistant clones was isolated and tested for recombination by Southern blot analyses. Cre-mediated deletion resulted in two types of recombination w h i c h were initially detected by use of a loxP probe and later verified by using a genomic probe. Materials and Methods 64 2.8 Generation of Chimaeric and Mutant Mice Chimaeric mice were generated by microinjection of 8-10 ES cells into twenty 3.5 day C57BL/6 blastocyst stage embryos and implanted into the uteri of pseudo-pregnant outbred albino foster mother recipients. Nine neonates mice were assessed for chimerism by the presence of agouti coat color. A l l nine were male chimaeric mice displaying greater than 90% agouti coat color were mated to C57BL/6 females. Tail D N A of agouti progeny, an indicator of germline transmission of the ES cell genotype was analyzed for the presence of the mutate allele. Heterozygous mice were intercrossed to generate mice homozygous for Mgat3 mutation. 2.9 DNA Isolation Cells and tissues were incubated for five hours or overnight in 50 m M Tris-H C I , p H 8.0; 50 m M E D T A , p H 8.0; 0.5% SDS, and 100 u g / m L proteinase K (Boehringer Mannheim). D N A was purified by one phenol:chloroform extraction, one chloroform extraction and ethanol precipitation. The pellet was washed wi th 70% ethanol and resuspended in TE ( lOmM Tris p H 8.0; I m M E D T A p H 8.0) buffer. 2.10 In V i t r o Differentiation of Embryonic Stem Cell Clones In gene targeting experiments, it is possible to produce multiple clones (more than the number that can be desirably micro-injected). In these instances, it would be beneficial to be able to pre-screen ES cell clones and to select only those capable of Materials and Methods 65 differentiation for microinjection. This may be of particular importance when using Cre /loxP site-directed recombination which requires going through an extra round of selection and thereby extending the time of an ES cell clone in culture. The protocol used for the in vitro differentiation is adapted from Robertson, Teratocarcinomas and embryonic stem cells: a practical approach. 1987. Embryonic stem cells are induced to form simple embryoid bodies by the following procedure. The medium used throughout is D M E M supplemented wi th either 5 or 10% FCS, 2 m M L-glutamine and I m M 2-mercaptoethanol. 1. Grow ES cells on gelatinized dishes for one to two passages to remove /reduce the number of fibroblast feeder cells. Plate cells at high density as a single cell suspension (approximately 2 x 106 cells for a 6 cm plate or 6 x 106 for a 10 cm plate. 2. Incubate cells for 2-3 days unti l they are approaching confluency and the boundaries between colonies have disappeared. Wash the cells wi th PBS prior to adding 1 m l .25% trypsin I m M E D T A for a 6 cm dish or 3 m l for a 10 cm dish. Gently rock the plate unti l 20- 30 % of the cells have flaked off the dish. Immediately neutralize the trypsin wi th excess medium. Gently rock the dish to remove the majority of the flaking cells. 3. Use a wide bore pipette to transfer the cellular aggregates to bacteriological dishes at a 1/4 to 1/5 dilution. Great care must be taken not to break up the cell aggregates. A confluent 6cm dish would be dispensed to 4-5 equal sized bacteriological dishes. Aggregates which are not diluted in this way w i l l adhere to each other forming long chains and result in poor differentiation. Materials and Methods 66 4. The majority of cell aggregates formed should not be able to adhere and w i l l m o l d into three-dimensional structures wi th in 24 hours. M e d i u m should be changed every two days by collecting the aggregates in a conical, al lowing them to settle, aspirating off the supernatant, resuspending in fresh medium and transferring to new bacteriological dishes. After 48-72 hours, the majority of aggregates should have formed an endoderm layer and thus are called simple embryoid bodies. I have observed that G418 selected clones show great variation wi th respect to the proportion of cell aggregates that form embryoid bodies (Figure 2.1). For example, one clone may form two embryoid bodies out of a hundred aggregates whereas a second clone may form seventy embryoid bodies out of a hundred cell aggregates. Clones wi th low developmental potential show many aggregates which seem never to initiate differentiation and/or die i n suspension culture. Materials and Methods 67 Figure 2.1 Embryoid body formation. (A) Confluent ES cells are gently lifted from the plate and placed in a bacterial dish for suspension culture. (B) After 36-48 hours post differentiation induction, most aggregates of the parental ES cell line will form simple embryoid bodies. This ability is variable among different clones and appears to dissipate with increasing passage, prior to differentiation. Materials and Methods 68 Clones wi th high developmental potential have a high percentage of cell aggregates form simple embryoid bodies with wel l defined endoderm and ectoderm and go on to form cystic embryoid bodies. 2.11 GlcNAc-TIII Enzyme Assays Tissues were homogenized by 20 strokes using a Dounce homogenizer wi th 25 m M 2-(N-morpholino) ethanesulphonic acid (MES) buffer p H 6.5 containing 1% Triton X-100. GlcNAc-TIII was assayed using a synthetic acceptor substrate, G l c N Ac( pl-2) [6-O-methyl-Man] (al-6) {GlcNAc((3l-2) [4-O-methyl-Man] (ocl-3) }Man(3-0-(CH 2 ) 8 COO(CH 3 ) (Khan et al., 1994). This substrate cannot be acted on by G l c N A c - T I , GlcNAc-TII , G l c N A c - T I V nor G l c N A c - T V and is a highly specific substrate for GlcNAc-TIII . The assay was performed in a total volume of 20 U.L containing 0.3 m M acceptor substrate, 62.5 m M G l c N A c (to inhibit N-acetylglucosaminidases), 3 m M A M P (to inhibit breakdown of UDP- [ 3 H]GlcNAc by pyrophosphatase), 10 m M MnC12, 0.125% Triton X-100, 0.1 M M E S p H 6.5, 1.25 m M U D P 10 m M M n C l 2 , 0.125% Triton X -100, 0.1 M M E S p H 6.5, 1.25 m M UDP-[ 3 H]-GlcNAc (10,000 dpm/nmol . ) and 2.5 uE crude tissue homogenate. After incubating the tubes at 37 °C for 2 h, the mixtures were diluted wi th 0.5 m L of water and loaded on to Sep-Pak C 1 8 cartridges. The cartridges were washed wi th water (40-60 mL) to remove unreacted radiolabeled donor and buffer components. The bound radiolabeled product was eluted wi th 3 m L of methanol and quantitated by l iquid scintillation counting in L K B 1209 Rackbeta instruments after addition of 15 m L of scintillation fluid. Materials and Methods 69 2.12 E 4 -PHA/L 4 -PHA Lectin Blotting Tissues from wild-type (+/+), heterozygous (+/A), and homozygous-null (A/A) mice were homogenized as described above for the GlcNAc-TIII enzyme assays. The homogenates were made 0.1 N in HC1 and heated at 80°C for 60 min . followed by neutralization wi th dilute N a O H to remove terminal sialic acid and fucose residues which interfere wi th E 4 - P H A binding (Kobata and Yamashita, 1989, 1993). Aliquots (30 jug protein) were subjected to electrophoresis in 12% SDS-polyacrylamide minigels (Laemmli, 1970). Proteins were electrophoretically transferred to a P V D F membrane (Towbin et al., 1979) and the membranes were blocked wi th 0.25% gelatin, 10% ethanolamine, 0.1 M Tris-HCI (pH 9.0), for 60 min . (Olmsted, 1981). The membranes were incubated in the absence of lectin or with biotinylated lectins at 0.2 u.g/ml (either E 4 - P H A , Phaseolus vulgaris leukoagglutinin or L 4 - P H A , Phaseolus vulgaris erythroagglutinin, Seikagaku, Japan) for 24 hrs at room temperature. Incubations and washes were conducted using the buffer system developed by Olmsted (Olmsted, 1981) (0.25% gelatin, 0.15 M NaCI, 5 m M E D T A , 0.05% Nonidet P-40, 0.05 M Tris-HCI (pH 7.5)). Bound lectins and biotinylated molecular-mass standard proteins were detected using Vectastain ABC™ (Vector Labs). Peroxidase activity was visualized using the chemiluminescent ECL™ reagent (Amersham) and X-ray film (Kodak). 2.13 Hematology Mice were anaesthetized wi th methoxyfluorane and bled from the tail v e i n . Blood was collected in EDTA-coated polypropylene tubes (Becton Dickenson). Materials and Methods 70 Automated differentials were determined by a C E L L - D Y N 3500 (UCSD Medica l Center, Hillcrest). Blood cell morphology was visualized by preparing a blood f i lm o n a slide. After air drying, slides were immersed in Wright Giemsa stain (Sigma Diagnostics) for one minute, placed in phosphate buffer (pH 7.2) for five minutes, rinsed in deionized water. Smears were examined under oi l immersion. 2.14 Serum Chemistry Mice were anaesthetized wi th 0.3 m L of 2.5% Aver t in , blood was drawn via a heart puncture and allowed to clot for 30 minutes at room temperature. Samples were subjected to centrifugation and supernatant was transferred to a new tube. Serum was examined by a Kodak EktaChem700 Analyzer (UCSD Medical Center, Hillcrest). 2.15 Flow Cytometry Bone marrow cells were derived from femurs extracted from adult mice and flushed wi th 3 m L of fluorescence-activated cell scanner (FACS) buffer (2% fetal bovine serum in phosphate buffered saline, PBS) using a 25 gauge needle. Splenocytes and thymocytes were harvested as single cell suspensions by mincing splenic and thymic tissue through fine wire-mesh screens respectively (Cooke et al., 1991). Cells were incubated with antibody at a density 5 x 10 6 /mL in a lOOmL total volume. A l l incubations and washes were performed on ice wi th F A C S buffer. F A C S analyses was carried out using a F A C S C A N Flow Cytometer and CELLQues t Materials and Methods Software (Becton Dickenson, Mountainview, C A ) . Dual labeling was performed wi th phycoerythrin (PE)-conjugated anti-CD4 antibody and fluoresceinated (FITC)-conjugated anti-CD8 antibody (Becton Dickenson). 72 CHAPTER 3 Mgat3 C l o n i n g and the G e n e r a t i o n of a Mgat3 M u t a t i o n in E m b r y o n i c Stem Cells 3.1 Introduction GlcNAc-TIII adds the bisecting G l c N A c to the (31,4 l inked mannose of the tri-mannose core (Narasimhan, 1982). A s the presence of this residue alters the conformation of other core sugars and blocks the ability of other transferases to initiate subsequent branches (Brisson and Carver, 1983; Schachter, 1991), it may play a key regulatory role in N-glycan biosythesis. A t present the biological significance of this modification is unknown. The cloning of the rat c D N A coding for G l c N A c -TIII (Mgat3) allows genetic approaches towards investigating the in vivo roles of this enzyme (Nishikawa et al., 1992). Conventional gene targeting techniques result i n a systemic genetic lesion. Such a deficit may result in embryonic lethality and/or complex pleiotropic effects, which can preclude the study of postdevelopment gene function. Furthermore, observed phenotypes can be difficult to discern whether they originated from defects in the affected cell or in the affected cell's envi ronment . To maximize the utility of our work, we employed the Cre/loxP recombination system to generate an Mgat3 mutation. This system enables one to selectively delete the gene of interest only in cells where the Cre recombinase is applied. Successful employment of the Cre /loxP system requires one to choose a suitable exon(s) which when deleted w i l l create a nu l l mutation. Therefore, it is desirable to either flank a 5' exon that when excised disrupts the reading frame or an exon coding for a critical domain of the protein. Addi t ional ly , for the condit ional 73 mutation, it is crucial that the placement of the loxP sequences do not interfere w i t h the normal expression of the gene. After screening a 129 S v / J genomic library, two phage isolates were cloned into Bluescript and subjected to restriction enzyme mapping. In addition to confirming the identity of the hybridizing element, sequence analyses revealed that the entire protein encoding sequence of Mgat3 was enclosed in a single exon. To generate the Mgat3 mutation, we designed a targeting vector which flanked the single protein encoding exon wi th loxP signals. Gene-targeted clones were transfected wi th the Cre recombinase and selected wi th ganciclovir. Clones heterozygous for the mutation were isolated and shown to exhibit approximately 50% of wild-type GlcNAc-TIII activity. 3.2 Results 3.2.1 Genomic Library Screening with a Rat cDNA probe The first step towards making a mutation in the mouse Mgat3 gene was to create a probe that would be highly specific. Two primers based on the rat Mgat3 c D N A were designed to amplify a 1.5 kilobase fragment harbouring greater than 90% of the protein encoding sequence. To enhance probe specificity, the short leader and transmembrane domain sequences were not included in the amplified region. Us ing P C R and rat genomic D N A as a template, the product was amplified, purified, cloned into pUC19 and verified by restriction enzyme analyses. The fact that it was possible to amplify the predicted size fragment from rat genomic D N A strongly suggested that Mgat3 protein coding sequence was contained wi th in a single exon. The insert was 74 liberated from the plasmid via Eco R I / Bam H I (in polylinker) digestion , 3 2P-labeled and used as a probe to screen a 129Sv/J genomic library. A m o n g 300,000 plaques, two phage clones were found to hybridize to the probe wi th high stringency. Fo l lowing purification, phage D N A was prepared, digested wi th Not I and cloned into Bluescript and subjected to restriction enzyme analyses. The two clones called ST-13 and ST-20, approximately 16 and 17 kilobases in size respectively, were shown to overlap and each possessed a single 1.8 kilobase Bam H I fragment which cross-hybridized wi th the rat probe. 3.2.2 Mgat3 Genomic Insert is in the Germline Configuration Southern analyses of the resulting plasmids revealed that a single 1.8 kilobase Bam H I band hybridized to the rat c D N A probe (Figure 3.1). To identify the mouse Mgat3 coding sequence, the fragment was subcloned into pUC19 and sequence analyses was initiated using M13 primers. The sequence not only confirmed the presence of Mgat3 coding D N A but also detected the translational start (ATG) and stop (TAG) signals of the gene (Figure 3.2). In view of the fact that our objective was to use the Mgat3 genomic D N A for construction of a targeting vector, it was necessary to confirm that the insert was in the germline state. Either the assembly of the phage library or replication and maintenance in E.coli can result in deletions and rearrangements of the exogenous D N A . Since such modifications could have serious inhibitory consequences on the frequency of homologous recombination events in ES cells, it was essential to verify the condition of plasmid-derived Mgat3 75 genomic D N A . The status of this £. co/f-propagated D N A was determined by digesting ES cell and plasmid-derived genomic D N A with restriction enzymes and Southern blotting. Germline Configuration Phage Insert 129 Genomic DNA Figure 3 .1 . Mgat3 insert is in the germline conf igurat ion. Analyses by Southern blot were used to determine that the isolated phage insert was in the expected 129 configuration. Plasmid and ES cell DNAs were cut with restriction enzymes and hybridized with 3 2 P- labe led Mgat3 coding sequence. Both DNAs revealed an identical banding pattern suggesting that the phage insert was in the native configuration. 76 Comparison of the banding pattern of these two D N A s after probing wi th labeled Mgat3 D N A authenticated the germline configuration of the cloned D N A (Fig. 3.1). 3.2.3 Mgat3 is Highly Conserved and Has a Single Protein-Encoding Exon Since the M13-primed sequencing runs yielded only approximately 400 base pairs of sequence from either end, additional primers for internal sequencing were synthesized. These oligonucleotides were used for pr iming at intervals approximately every 300 base pairs, on both D N A strands. 5 ' -GGATCCTCGGGCTGCTCTCCCTGACTTCTTGTTCTCTCCATCTCCTGCAGG 51 1 M K M R R Y K L F L M F C M A G L C L I ATGAAGATGAGACGCTACAAGCTCTTTCTCATGTTCTGTATGGCTGGCCTGTGCCTCATA 111 21 S F L H F F K T L S Y V T F P R E L A S TCCTTCCTGCACTTCTTTAAGACCTTATCCTATGTCACCTTCCCGAGAGAACTGGCCTCC 171 41 L S P N L V S S F F W N N A P V T P Q A CTCAGCCCTAACCTCGTATCCAGCTTCTTCTGGAACAATGCCCCTGTCACTCCCCAGGCC 231 61 S P E P G G P D L L R T P L Y S H S P L AGTCCGGAGCCGGGTGGCCCCGACCTATTGCGGACACCCCTCTACTCCCACTCTCCCCTG 291 81 L Q P L S P S K A T E E L H R V D F V L CTCCAGCCACTGTCCCCGAGCAAGGCCACAGAGGAACTGCACCGGGTGGACTTCGTGTTG 351 101 P E D T T E Y F V R T K A G G V C F K P CCGGAGGACACCACGGAGTATTTTGTGCGCACCAAAGCTGGTGGTGTGTGCTTCAAACCA 411 121 G T R M L E K P S P G R T E E K P E V S GGTACCAGGATGCTGGAGAAACCTTCGCCAGGGCGGACAGAGGAGAAGCCCGAAGTGTCT 471 141 E G S S A R G P A R R P M R H V L S T R GAGGGCTCCTCAGCCCGGGGACCTGCTCGGAGGCCCATGAGGCACGTGTTGAGTACGCGG 531 161 E R L G S R G T R R K W V E C V C L P G GAGCGCCTGGGCAGCCGGGGCACTAGGCGCAAGTGGGTTGAGTGTGTGTGCCTGCCAGGC 591 181 W H G P S C G V P T V V Q Y S N L P T K TGGCACGGGCCCAGTTGCGGGGTGCCCACGGTGGTGCAGTATTCCAACCTGCCCACCAAG 651 77 201 E R L V P R E V P R R V I N A I N I N H GAACGCCTGGTACCCAGGGAGGTACCGAGGCGGGTTATCAACGCCATCAACATCAACCAC 711 221 E F D L L D V R F H E L G D V V D A F V GAGTTCGACCTGCTGGATGTGCGCTTCCATGAGCTGGGAGATGTTGTGGACGCCTTCGTG 771 241 V C E S N F T A Y G E P R P L K F R E M GTCTGTGAATCTAATTTCACCGCCTACGGGGAGCCTCGGCCGCTCAAGTTCCGAGAGATG 831 261 L T N G T F E Y I R H K V L Y V F L D H CTGACCAATGGCACCTTCGAGTACATCCGCCACAAGGTGCTCTATGTCTTCCTGGACCAT 891 281 F P P G G R Q D G W I A D D Y L R T F L TTCCCACCTGGTGGCCGTCAGGACGGCTGGATTGCGGATGACTACCTGCGCACCTTCCTC 951 301 T Q D G V S R L R N- L R P D D V F I I D ACCCAGGATGGCGTCTCCCGCCTGCGCAACCTGCGGCCCGATGACGTCTTTATCATCGAC 1011 321 D A D E I P A R D G V L F L K L Y D G W GATGCGGACGAGATCCCTGCGCGTGATGGTGTGCTGTTCCTCAAACTCTACGATGGCTGG 1071 341 T E P F A F H M R K S L Y G F F W K Q P ACAGAGCCCTTCGCCTTCCACATGCGGAAGTCCCTGTATGGTTTCTTCTGGAAGCAGCCG 1131 361 G T L E V V S G C T M D M L Q A V Y G L GGCACACTGGAGGTGGTGTCAGGCTGCACCATGGACATGCTGCAGGCCGTGTATGGGCTG 1191 381 D G I R L R R R Q Y Y T M P N F R Q Y E GATGGCATCCGCCTGCGCCGCCGCCAGTACTACACCATGCCCAACTTCCGGCAGTATGAG 1251 401 N R T G H I L V Q W S L G S P L H F A G AACCGCACCGGCCACATCCTAGTGCAGTGGTCTCTCGGCAGCCCCCTGCACTTCGCGGGC 1311 421 W H C S W C F T P E G I Y F K L V S A Q TGGCATTGCTCCTGGTGCTTCACACCCGAGGGCATCTACTTTAAACTCGTGTCAGCCCAG 1371 441 N G D F P R W G D Y E D K R D L N Y I R AATGGCGACTTCCCCCGCTGGGGTGACTATGAGGACAAGAGGGACCTCAATTACATCCGC 1431 461 S L I R T G G W F D G T Q Q E Y P P A D AGCTTGATCCGCACTGGGGGATGGTTCGACGGAACGCAGCAGGAGTACCCTCCTGCGGAC 1491 481 P S E H M Y A P K Y L L K N Y D Q F R Y CCCAGTGAGCACATGTATGCTCCTAAATACCTGCTCAAGAACTATGACCAGTTCCGCTAC 1551 501 L L E N P Y R E P K S T V E G G R Q N Q TTGCTGGAAAATCCCTACCGGGAGCCCAAGAGCACTGTAGAGGGTGGGCGCCAGAACCAG 1611 521 G S D G R P S A V R G K L D T V E G * GGCTCAGATGGAAGGCCATCTGCTGTCAGGGGCAAGTTGGATACAGTGGAGGGCTAGGGC 1668 TGTGCACTTTCACAGGGCTGGGTAGGCTGAAATAATGGCTAAGCCAGTGCTATCTTAGGC 1728 CTCCTCCTTATCCCGGG GCACTTGAGAGAGCCAGGATCC-3' 1771 Figure 3.2. Nucleotide and amino acid sequence of the mouse Mgat3 gene. Single letter symbols are used for the amino acid sequence. The Bam H I sites used to clone the fragment are underlined while the initiating methionine is in bold. The Sma I site used to build the targeting vector is italicized. 78 Previous work has shown that Mgat3 is highly conserved and exists in the mammalian genome as a single allelic copy (Nishikawa et al., 1992; Ihara et al., 1993). The early indication that the protein encoding sequence was contained wi th in a single exon proved to be correct. D N A sequencing revealed that the mouse Mgat3 gene (Genbank: U66844) we cloned encodes an uninterrupted 538 amino acid GlcNAc-TIII and displays greater than 90% identity to the rat and human homologs (Figure 3.3). Furthermore, the mouse gene would be expected to have an additional 2 amino acids in comparison wi th the rat since translation by eukaryotic ribosomes would be expected to initiate at the first 5' A U G in the sequence context . . . A N N A U G N . . . . Unexpectedly, comparisons between Mgat3 isolated here and that published (Bhaumik et al., 1995) revealed 21 nucleotide differences that generates 7 amino acid changes between the two mouse sequences. 79 1 I D * 80 mouse MKMRRYKLFL MFCMAGLCLI SFLHFFKTLS YVTFPRELAS LSPNLVSSFF WNNAPVTPQA SPEPGGPDLL RTPLYSHSPL rat I. . . D human 81 T * * S 160 mouse LQPLSPSKAT EELHRVDFVL PEDTTEYFVR TKAGGVCFKP GTRMLEKPSP GRTEEKPEVS EGSSARGPAR RPMRHVLSTR rat TK.A . . . .V A. human ....P. ...A L K...R.P. ..P GAN... .. ..P.YL..A. 161 * 240 mouse ERLGSRGTRR KWVECVCLPG WHGPSCGVPT WQYSNLPTK ERLVPREVPR RVINAININH EFDLLDVRFH ELGDWDAFV rat . . . .G human ..T.G..A V 241 D * 320 mouse VCESNFTAYG EPRPLKFREM LTNGTFEYIR HKVLYVFLDH FPPGGRQDGW IADDYLRTFL TQDGVSRLRN LRPDDVFIID rat human 321* * * 400 mouse DADE I PAR DG VLFLKLYDGW TEPFAFHMRK SLYGFFWKQP GTLEWSGCT MDMLQAVYGL DGIRLRRRQY YTMPNFRQYE rat I human T V 401 * * * * 480 mouse NRTGHILVQW SLGSPLHFAG WHCSWCFTPE GIYFKLVSAQ NGDFPRWGDY EDKRDLNYIR SLIRTGGWFD GTQQEYPPAD rat human G 481 * S A 538 mouse PSEHMYAPKY LLKNYDQFRY LLENPYREPK STVEGGRQNQ GSDGRPSAVR GKLDTVEG rat R S T.. human R.H. . .D. . .Q. .R . . AA. . WRHR .PE...P.-R ...-EA.V Figure 3.3. Comparison of putative mouse, rat, and human GlcNAc-TI I I sequences. Identities between mouse and rat or human amino acids are denoted (.). Gaps are present in the human sequence (-). Differences between the mouse GlcNAc-TIII sequence displayed here and that published (Bhaumik et al., 1995) are indicated as amino acid changes or silent substitutions (*) above the mouse sequence. 3.2.4 Chromosomal Localization of Mouse Mgat3 To derive a probe, the 16 kilobase mouse genomic insert, ST-13 (from above), was released from Bluescript via Not I digestion and purified by agarose gel electrophoresis. Wi th the probe isolated, fluorescence in situ hybridization (FISH) to normal mouse chromosomes was undertaken. Regional assignment of Mgat3 to mouse chromosome 15 at position E l l was determined following analyses of 20 wel l spread metaphases as presented in Figure 3.4A. A 80 * • 41 • Figure 3.4. Regional chromosomal localization of the Mgat3 gene by fluorescence in situ hybridization. (A) A representative metaphase preparation is shown in which Mgat3 probe hybridization was detected on both homologues of chromosome 1 5 as described in Materials and methods. (B) Double color FISH show-ing DIG-labeled Mgat3 detected with FITC-avidin (yellow, as in A) and biotinylated chromosome 1 5 marker probe (med/ScnSa) detected with rhodamine-anti-DIG (red) as described in Chapter 2. Mgat3 is localized adjacent and telomeric to med/Scn8a. 81 Positive hybridization signals were noted in all 20 metaphases and were visual ized on both homologs in 85% of these metaphase spreads (17/20). Al though the signal was distinctly visualized, some background hybridization was observed, perhaps as a result of repetitive sequence present wi th in the 16 kilobase genomic Mgat3 probe. Chromosome assignment was initially determined by banding karyotype (see Chapter 2) and confirmed by a chromosome 15-specific probe med /Scn8a that maps to 15F1 (See Figure 3.4B). 3.2.5 Brain and Kidney Show Highest Levels of Expression Among Normal Tissues Examined N o r m a l expression of Mgat3 R N A among various mouse tissues was determined using a mouse 1.8 kilobase Mgat3 probe containing the entire G l c N A c -TIII coding sequence. Results indicated that highest steady-state R N A expression levels occur in brain and kidney, followed by colon, small intestine, lung, thymus, stomach, and ovary (Figure 3.5). The size of the transcript is about 4.8 kilobases, only slightly larger than the 28S ribosomal R N A . This expression pattern is consistent wi th the results of similar studies on various tissue samples (Nishikawa et al, 1992; Bhaumik et al, 1995). Tissues that demonstrate Mgat3 expression and that conserve this expression in phylogeny may require GlcNAc-TIII during development and i n normal function. 82 ^ 2 8 S < 1 8 S Figure 5. Expression of Mgat3 RNA among normal mouse t issues. Five micrograms of total cellular RNA was analyzed from each tissue sample. A 1.8 kilobase Bam HI Mgat3 genomic fragment containing the entire GlcNAc-TIII protein-encoding region was used as a probe. The bottom panel represents the ethidium-stained profile of RNA levels analyzed. 3.2.6 The pflox Vector Due to the possible pleiotropic nature of effects resulting from the loss of the gene, the Cre / l oxP recombination system was employed for Mgat3 mutagenesis. This recombination system enables one to create a conditional mutation: the muta t ion only occurs in the presence of the Cre protein. Work ing towards such a goal, our laboratory produced a new vector, pflox, which bears three different cloning sites for 83 the integration of adjacent pieces of genomic D N A , only one of which is flanked by loxP sites. After pflox construction, the fidelity of the loxP sites was investigated by D N A sequencing. The data confirmed the presence of the three intact loxP sites as wel l as sequences necessary for detecting homologous recombination by P C R (Figure 3.6). pflox Vector N t X b # l r M r % B X N t \ i A iK neo i . i i / B Xba I GTCGAAC TCTAGA GGATCAGCTTGGGCTGCAGGTCGAGGGACCTA ATAACTTCGTATA GCATACATTATACGAAGTTATATT AAGGGTTCCGGATCGAGCAGTGTGGTTTTGCA C Bam HI TA ATAACTTCGTATAGCATACATTATACGAAGTTAT ATTAAGGGTTCC GGAICCCC GGAGCTTGGGCTGCAGGTCGAGGGACCTA ATAACTTCGTATAGCATACATTATACGA AGTTATATT AAGGGTTCCGGATCGATCCCCGGGCGAGCTCGAATTGATCCCCGGGTAC Xho I CGGGCCCCCC CJCGAG GTCG Figure 6. The pflox vector. (A) Diagram of the pflox vector. The unique cloning sites are B: Bam HI, X: Xba I, X: Xho I. The large arrowheads represent the loxP sites (not drawn to scale)and the small half arrow indicates the position of the vector-based primer (tk303) used for PCR detection of homologous recombination. The finished targeted vector can be liberated from the plasmid via the flanking Not I (Nt) sites. Sequencing results shown in (B) and (C) indicate that the loxP sites (dark full arrow) in pflox vector are intact. The light half arrow shows the complementary sequence of primer tk303. 84 Using the pflox backbone for gene targeting, homologous recombination places two selectable markers, thymidine kinase (tk) and neomycin phosphotransferase (neo), flanked by loxP signals into the genomic locus. The new allele now serves as a substrate for Cre-mediated deletion yielding two possible recombinants w h e n ganciclovir selection is applied (Cells bearing the tk gene are selected against by the drug ganciclovir). The Type 1 deletion, the systemic mutation, ensues from the recombination between the two outermost loxP sites whereas the Type 2 deletion results in an allele wi th the exon being flanked by loxP sequence. Taking into account that promoter elements from selectable markers have been known to alter gene expression of adjacent genes, the Type 2 event, which has the neo and tk genes removed, is used for the conditional mutation. The vector also contains f lanking Not I sites which allows for removal of plasmid sequence prior to transfection. 3.2.7 Construction of the Mgat3 Targeting Vector From a phage 129/SvJ mouse genomic library, an Mgat3 containing clone was used to generate a targeting vector as follows (also see Figure 3.7). A 1.8 kilobase Bam HI/Sma I fragment which contained the single Mgat3 protein coding exon was cloned into the Bam H I site of the pflox vector. The flanking 2.0 kilobase Bam H I fragment and an 8.5 kilobase Sma I fragment were subsequently cloned into the Xba I and Xho I sites of pflox, respectively. The loxP signals contain an " A T G " when read in the direction of the arrowhead (see Figure 1.7), therefore, the insertion of Mgat3 was purposely directed to be in the opposite orientation so as to not introduce an 85 upstream translational start signal. This is important for the conditional mutation as upstream " A T G " s have been shown to reduce translation efficiency (Kozak, 1991). Genomic Clone 129/SvJ pflox Vector Xb A BAX tk neo i—i 1 kb - PCR Primers Figure 3.7. Mouse Mgat3 genomic structure and targeting vector production. The mouse genomic Mgat3 gene as isolated is shown with restriction enzyme sites mapped (open arrow depicts position and transcriptional orientation of GlcNAc-TIII protein-encoding sequence). The pflox vector was used as depicted to generate an Mgat3 targeting vector and contains two selectable markers tk and neo (shaded and black arrow, respectively). 34 base-pair loxP sites (not to scale) are depicted as black arrowheads. Restriction enzyme sites: A, Apa I; B, Bam HI; K, Kpn I; N, Nde I; No, Not I; S, Sma I. The finished targeting vector was digested wi th Not I and purified by agarose gel electrophoresis. Subsequently, the " R l " ES cell line (Nagy et al, 1993) was transfected by electroporation wi th the isolated vector (see Chapter 2 for Methods). Cells were transferred to plates coated wi th gelatin and selected wi th G418 (Life Technologies, Grand Island, N Y ) . 3.2.8 PCR Detection of Homologous Recombination Homologous recombinants were detected by P C R ut i l iz ing one internal primer, Tk303, and an external primer, GnTIII, which resides outside the targeted sequence. Employing a control vector that had been transfected into ES cells, the P C R reaction was first optimized using buffers wi th varying p H and M g 2 + concentrations 86 (Invitrogen). The sensitivity of this assay was determined by mixing transfected cells wi th parental cells. Amplif icat ion of a 2.0 kilobase fragment was diagnostic of a positive result. Ten days following the electroporation, ES cell colonies were picked and D N A was prepared for P C R (See Chapter 2). Pools of twelve clones each were screened by PCR. Each of the ten positive pools was shown to harbour a single positive clone. Ten positive clones were detected out of 384 G418 resistant clones screened (See Figure 3.8). Each of these clones was expanded for cryopreservation and large scale D N A preparation. 2C1 2C2 2C3 2C4 2C5 2C6 2C7 2C8 2C9 2C10 2C11 2C12 ttgf m l m • 2F1 2F2 2F3 2F4 2F5 2F6 2F7 2F8 2F9 2F10 2F11 2F12 i W m • f m I <-2kb <2kb Figure 3.8. PCR detection of homologous recombinat ion. DNA from clones of positive pools was prepared and subjected to PCR, with primers Tk303 and GnTIII, for the identification of positive clones. Products were run out on a 1% agarose gel, stained with ethidium bromide and photographed. Two positive clones, 2C9 and 2F7, were expanded for Southern analyses. 87 3.2.9 Confirmation of Homologous Recombination by Southern Blotting After D N A preparation, putative targeted clones were analyzed by Southern blotting. Selected restriction digests resulted in distinct bands being observed for the w i l d type and targeted alleles (see Figure 3.9). The genomic element used for probing the blots resided outside the targeted sequences, resulting in easy discr iminat ion between random integration and homologous recombination events. Furthermore, an internal probe was used to check for the presence of multiple integration events (data not shown). Eight PCR-positive clones were shown to have undergone homologous recombination by Southern analyses. In each case, one Mgat3 allele exhibited the expected structural alteration (Mgat3mkneo1) in comparison to germline 129 D N A (Figure 3.9B and data not shown). Hybridization wi th the Mgat3 probe suggested that the G418-resistant clones arose from a single copy of the vector. 88 A Mgat3 WT-129 g e n o m i c p r o b e N K i i Targeting Vector B Mgat3 B A S B N -lr—J—>l—L (B) A (S )AB N L—•—^J-U l_ neo B & & # & & e # Mga13F[tkne°] MgatSWT-129 g e n o m i c p r o b e / N d e I D i g e s t Figure 3.9. Southern confirmation of homologous recombination. (A) Homologous recombination with the wild-type (WT) Mgat3 allele in embryonic stem (ES) cells generated the /Wgaf3 F [ , k n e° l allele. The position of the probe used to determine homologous recombination of targeted Mgat3 allelic structure is depicted. (B) Homologous recombination at the Mgat3 locus in embryonic stem (ES) cells. PCR positive clones were analyzed by Southern blot. The presence of a targeted Mgat3 allele is observed by Southern blot analyses in seven G418-resistant PCR-positive ES cell clones, and in comparison to wild-type 129 DNA. These studies revealed that each clone had one wild type band (8.5 kilobase band) plus an additional 1 2 kilobase band, expected for a homologous recombinant. Restriction enzyme sites: A, Apa I; B, Bam HI; K, Kpn II; N, Nde I; S, Sma I. The bold line in A indicates sequence common to the Mgat3 targeting vector. 89 3.2.10 Three of Eight Recombinants Retained All Three loxP sites After homologous recombination had been substantiated, it was imperative to determine if all three loxP sites had been maintained, since they were required for producing the systemic and conditional Mgat3 mutations. To ascertain this information, ES cell D N A was digested wi th Apa I and analyzed by Southern blotting. Hybridization wi th a loxP probe allowed for the visualization of each loxP site as a discrete band. <P s£> rr> <A x£ <~v> <?> 0 ^ 12 — 5' loxP site Internal loxP site < - 3' loxP site 1 -loxPprobe/Apa I D i g e s t Figure 3.10. Structure and presence of loxP s i tes. DNA from homologous recombinant ES cell clones was digested with Apa I and analyzed by Southern. Use of a loxP-specific probe indicates that two of seven Mgaretargeted clones retain all three loxP sites. Five of eight clones examined in this way d id not contain the 3' loxP site (Figure 3.10 and data not shown), likely the result of homologous recombination wi th in the loxP flanked genomic Mgat3 sequence. Subsequently, clone 1H8 which preserved all of the Cre recombination signals was used for deriving the Mgat3 mutation. 90 3.2.11 Transient Cre Transfection in Embryonic Stem Cells Results in Two Types of Recombination The targeted Mgat3 allele (Figure 3.12A, Mgat3FItkneo1) was then a substrate for Cre recombinase activity in producing two types of recombined Mgat3 alleles. Recombination between distal loxP sites generated a nu l l allele (Mgat3A) by deletion of al l intervening D N A , while recombination between loxP sites flanking the tk and neo cistrons generated a conditional mutation (Mgat3F) wi th loxP sites flanking the GlcNAc-TIII protein encoding sequence (Figure 3.12A). This presumptive conditional mutation achieved in ES cells would then allow the production of mice bearing a functional Mgat3 allele that could be deleted by transgenic Cre recombinase expression; this is significant because inactivation of both Mgat3 alleles might lead to multi-systemic phenotypes and early developmental lethality in mice ( Marth , 1996). TKpr pMC-Cre Pro • Lys • Lys• Lys • Arg• Lys A/al r- ACCATG CCCAAGAAGAAGAGGAAGGTG Cre pA 1 Z J Figure 3 .11 . The PMC-Cre expression vector. The plasmid was a gift from Klaus Rajewsky (University of Cologne). To increase the efficiency of translation in mammalian cells, Cre has an adenine placed at the -3 position. Additionally, the Cre gene in the pMC-Cre vector has been further modified with the addition of a nuclear localization signal from SV40 large T antigen (Gu et al., 1994b). To produce ES cell subclones that contained the Type I (systemic-null) and Type II (conditional-null) Mgat3 mutations, ES clone 1H8 was subjected to transient Cre expression by electroporation of pMC-Cre (Figure 3.11) and subsequent selection in the presence of ganciclovir. Subclones resistant to ganciclovir were isolated and analyzed by genomic Southern blotting. 91 genomic probe loxP probe Mgat3 probe Mgat3 Fpkneo] A N K I I B J _ 4 •12kb-A (B)A (S )ABN S tk neo Mgat3 +Cre +Ganciclovir Mgat3A Type 2 Deletion ^_Mgat3F [tkneo] 1 2 ^ ^-Mgat3F ^Mgat3A Mgat3F['kne°] Mgat3F Mgat3WT-l29 Mgal3A /oxPprobe/Nde I Digest genomic probe/Nde I Digest Figure 3 . 1 2 . Cre-mediated recombination in embryonic stem ce l l s . (A) The Mgat3m"eo] allele is used as a substrate for subsequent recombination by Cre recombinase. Using the 1H8 /Wgaf3-targeted ES clone, Cre recombination and ganciclovir selection results in two types of recombination events with subclone 1 exhibiting a Type II recombination and subclones 2-4 bearing a Type I deletion. In each allele, the size of fragments detected by the genomic probe are noted in light gray and flanked by arrows. The Type II deletion produces a genomic fragment similar in size to the wild-type allele (C) but which bears two loxP sites where the wild-type allele does not (B). As compared to wild type, the decrease in size of the Type I band (B & C) corresponds to the loss of Mgat3 coding sequence. Five micrograms of ES cell DNA was analyzed in the above studies. Restriction enzyme sites: A, Apa I; B, Bam HI; K, Kpn I; N, Nde I; S, Sma I. The bold line indicates sequence common to the Mgat3 targeting vector. 92 Using a loxP or genomic probe, sixteen of eighteen ganciclovir resistant clones screened were found to have undergone a Type I recombination while the remaining two exhibited the Type II deletion. (Figure 3.12B & C and data not shown), confirming that the targeted Mgat3 allele exhibited the expected structure fol lowing Cre recombination. The loxP flanked D N A is presumably lost from the cell as it does not contain sequence required for replication and maintenance. 3.2.12 The Mgat3 Mutation in Embryonic Stem Cells is Associated with a Loss in GlcNAc-TIII Activity R l parental ES cells and 1H8 subclones 1 and 2 were analyzed for GlcNAc-TIII activity in vitro. Comparison of GlcNAc-TIII enzyme activities among extracts of R l ES cells {Mgat3WTmT), Type I clones (Mgat3WT/A), and Type II clones (Mgat3WT/F) yielded 100%, 37%, and 97% respectively (data not shown). The observation that the heterozygous ES cells displayed an approximate 50% loss in GlcNAc-TIII activity strongly suggested that there was a single gene associated wi th this function, at least in the ES cell. Moreover, the Type II clones closely resembled the parental R l cells in GlcNAc-TIII activity, thereby, implying that the loxP sequences are not inhibitory for transcription or translation. Fol lowing these results and cells, we initiated the production of mice bearing this nul l mutation in the Mgat3 gene, using these cells. 93 3.2.13 The Generation of Chimaeric Animals and the Transmission of the Mutated Allele From the twenty C57BL/6 blastocysts injected wi th the ES cell 1H8 subclone #2 and implanted, nine male agouti mice were produced. The maleness and agouti coat color were indicators of strong chimerism as the R l ES cells are male and confer agouti color. Chimaeric mice were generated initially from Type I (Mgat3WT/Mgat3A) cells to determine whether systemic loss of Mgat3 function would produce phenotypic results indicating a developmental role for the bisecting G l c N A c in N-glycans. A m o n g matings wi th C57BL/6 females, six of the nine chimaeric animals transmitted the ES cell genotype to its offspring as judged by agouti coat color. Ta i l D N A was isolated from agouti animals and used to determine whether they inherited the w i l d type or Mgat3A allele. Mice heterozygous for Mgat3 appeared normal and were intercrossed to generate embryos and mice that were homozygous for the Mgat3A allele. A l l genotypes were determined by Southern blotting of genomic D N A derived from tail samples. 3.3 Discussion A s the init ial step in creating a Mgat3 mutation, a 129Sv/J mouse l iver genomic library was screened wi th a rat c D N A probe to identify GlcNAc-TIII coding sequence. After the cloning of Mgat3 hybridizing elements and Southern analyses, a single 1.8 kilobase Bam H I fragment was identified to contain the entire protein encoding sequence. Interestingly, Mgatl and Mgat2 harbour their protein encoding 94 sequence wi th in a single exon as wel l (Pownall et al, 1992; Campbell and Mar th , unpublished observations). Sequence analyses of both strands revealed the presence of a 1614 base pair open reading frame coding for a 538 amino acid protein which displays 97.8 % and 92.2 % identity wi th the rat and human homologs, respectively. A s compared w i t h the previously reported rat protein, Figure 3.3 presents the mouse protein wi th an extra two amino acids since translation by eukaryotic ribosomes would be expected to initiate virtually 100% of the time at the first 5' A U G in the sequence context . . A N N A U G N . . . Interestingly, these two codons are conserved between all three homologs. Al though initiation of translation in the mouse may wel l provide two additional amino acids (M-K-) at the N-terminus (Figure 3.3 and Bhaumik et al., 1995), this addition may not occur as frequently in the rat as the first 5' A U G identified is not in a sequence context that would be recognized at high frequency by eukaryotic ribosomes (a pyrimidine is present at position -3, Kozak, 1991). Whether a similar situation exists in human MGAT3 GlcNAc-TIII is not clear at present as the reported 5' nucleotide sequence does not extend sufficiently into that region (Ihara et al., 1993). Unexpectedly, additional comparisons between the Mgat3 isolated herein and that published revealed 21 nucleotide differences that generate 7 amino acid changes between the two mouse sequences. Interestingly, in all but three instances (positions 159, 243 and 536), the amino acid differences between the two mouse sequences nevertheless maintain identity wi th either the rat or human GlcNAc-TIII sequences. This result was quite unexpected since both sequences were derived from a 129 mouse genomic library, so these changes may reflect divergence between Mgat3 95 sequences wi th in this outbred mouse strain. Previous chromosomal localization studies of mouse and human homologues (Ihara et al, 1993; Bhaumik et ah, 1995) are fully consistent wi th this study which maps the Mgat3 gene to mouse chromosome 15 at E l l in a region likely syntenic wi th the human genome and homolog at chromosome 22q.l3.1. The Mgat3 expression profile at the R N A level is conserved among mul t ip le vertebrate cell types wi th high levels present in kidney of mouse and rat (Figure 3.5; Nish ikawa et al., 1992; Bhaumik et al, 1995). Addit ional ly, Mgat3 expression was also found at high levels in mouse brain, and to a lesser amount in other mouse tissues surveyed herein and previously. Like some other glycosyltransferases cloned and characterized to date (Schachter, 1994), the putative GlcNAc-TIII amino acid sequence exists uninterrupted by introns and within an R N A transcript much larger than necessary to encode the enzyme. It is possible that 5' and 3' Mgat3 untranslated sequences may reside in other exons and may play some role yet to be disclosed i n enzyme production, although regulation of Mgat3 expression and function may be completely accomplished by mechanisms involv ing transcriptional control, intracellular localization and N-glycan substrate availability. 96 CHAPTER 4 Bisecting N-Acetylglucosamine of N-glycans Appears Dispensable F o r Development and Reproduction 4.1 Introduction The action of N-acetylglucosaminyltransferases control the degree of branching arising from the trimannosyl core prior to their movement to the trans-Golgi . GlcNAc-TIII catalyzes the addition of G l c N A c to the p-mannosyl residue of the chitobiose core of complex and hybrid N-glycans. A s this addition inhibits other GlcNAc-transferases from acting, it influences the amount of branching. These initiated branches are then extended by the resident glycosyltransferases that are expressed by the cell. Some of these branch-extending enzymes show branch specificity, therefore, the action of the GlcNAc-transferases impact on the type of terminal structures that are synthesized (Joziasse et al., 1987; van den Eijden et al., 1988). To pursue the developmental/physiological roles of GlcNAc-TIII , mice deficient in Mgat3 were generated. These mice are viable, completely lack detectable GlcNAc-TIII activity and are deficient in E 4 - P H A visualized bisected N-glycans. However, the possibility of a second GlcNAc-TIII isozyme still remains but our results do not support such a finding. Al though our preliminary studies have not uncovered a phenotype, we believe that a function for GlcNAc-TIII may still exist. Necessary studies w i l l include analyzing the oligosaccharide alterations found i n mice lacking GlcNAc-TIII . 97 4.2 Results 4.2.1 Mgat3 knockouts are Viable and of Normal Weight With a genomic Mgat3 probe and Nde I-digested D N A as used previously, the presence of all three genotypes was observed among offspring of heterozygous matings (Figure 4.1A). A n Nde I restriction fragment length polymorphism was found between 129 and C57BL/6 mouse strains, thus producing an Mgat3m'BL/6 derived fragment of approximately 4 kilobases (kb) in comparison to the 8.5 kb fragment observed in wild-type 129 D N A (compare wi th results in Figure 3.9). In comparison to litter-mate controls, Mgat3-mx\\ mice were of normal size and weight. 4.2.2 Null Mice Lack Hybridization to a Mgat3 Coding Sequence Probe Using the entire GlcNAc-TIII coding region as a probe, Cre recombination in ES cells was found to have resulted in both excision and degradation of Mgat3 > D N A flanked by loxP sites as no hybridization signal was found in genomic D N A samples from Mgat3A/Mgat3A mice (Figure 4.IB). Offspring from parents bearing the heterozygous Mgat3wr/Mgat3A genotype were analyzed for Mgat3 allelic structure. Results indicated that homozygous deletion of the Mgat3 gene was not lethal i n embryonic development (Table 1). 98 1 — Genomic Probe/Nde I Digest 1 — M g a t 3 Probe/Nde I Digest Figure 4 .1 . Heterozygous and homozygous mutations at the Mgat3 allele in intact mice. (A) Analyses of tail DNA isolated from animals generated from matings between mice heterozygous for the Mgat3A allele. The wild-type Mgat3 allele is derived from the C57BL/6 background and is smaller than the 129-derived Mgat3 allele using A/del in Southern blot analysis, due to a restriction fragment length polymorphism between 129 and C57BL/6 strains. (B) Southern blot analyses using the Mgat3 coding sequence as a probe confirms a complete absence of hybridization in Mgat3A/Mgat3A DNA samples as a result of Cre recombination and degradation of excised DNA. 99 Furthermore, since nu l l animals display no banding wi th the Mgat3 probe, it indicates that there are no closely related species. Therefore, if a second enzyme with GlcNAc-TIII activity exists, it cannot share much D N A identity wi th Mgat3. 4.2.3 Mutant Mice are Fertile and Transmit the Mutated Allele at a Predicted Mendelian Frequency Moreover, homozygous Mgat3A / Mgat3A mice appeared normal (see below) and were subsequently found to be fertile in crosses wi th either inbred C57BL/6 or Mgat3AIMgat3A mice. From these crosses, offspring were generated of both sexes and wi th frequencies of Mgat3 genotypes representing a Mendel ian distribution (Table 1 and data not shown). Table 1. Transmission of the Mgat3A allele. Mgat3A Mouse Production Genotype and number of offspring Parental Genotypes Mg^/Mg^™7 Mgat3i/Mgat3WT Mgat3i/Mgat3i Mgat3i/Mgat3WT Cf x Mgat3i/Mgat3wrQ Mgat3"Mgat3wrCff. Mgat3m/Mgat3mQ Mgat3WT/Mgat3m& x Mgat3i/Mgat3mQ Mgat3i/Mgat3i Cf x Mgat3i/Mgat3i Q 100 179 45 64 55 57 103 44 4.2.4 Disruption of the Mgat3 gene is Associated with a Deficiency in GlcNAc-TIII Activity To confirm that the Mgat3A allele resulted in loss of GlcNAc-TIII enzyme activity, a synthetic acceptor substrate was used in vitro to measure enzyme activity in extracts from brain and kidney tissues of mice bearing Mgat3WT or Mgat3A alleles 100 (see Materials and Methods). Approximately 50% loss of GlcNAc-TIII activity was found in extracts of brain and kidney derived from heterozygous animals while loss of all significant GlcNAc-TIII activity was observed in extracts from Mgat3A/Mgat3A mice (Figure 4.2). 101 These results confirm that the deletion generated in the Mgat3 allele (Mgat3A) is a nu l l mutation and inactivates all measurable GlcNAc-TIII enzyme activity. 0.6 > u 0 X N • IH 0 H g 0.5 0.4 0.3 0.2 0.1 .02 +/+ +/A A / A +/+ + / A A / A Brain Kidney Figure 4.2. Mgat3 mutation is associated with a reduction of GlcNAc-TI I I activity. Loss of GlcNAc-TIII activity in brain and kidney tissue extracts bearing a mutation in the Mgat3 allele. Heterozygous samples exhibit approximately 50% reduction in GlcNAc-TIII activity towards a synthetic acceptor substrate, while homozygous Mgat3A samples lacked significant GlcNAc-TIII activity. In addition, GlcNAc-TIII activities were also determined for thymus, spleen and liver tissues (data not shown). In all cases, tissues derived from n u l l 102 animals lacked detectable GlcNAc-TIII activities suggesting that the Mgat3 gene had been eliminated. Although this lack of activity in the mutant implies a single gene encoding this function, it could be argued that a second isozyme may not recognize this synthetic substrate. Accordingly, it is imperative to compare the carbohydrate structures present in the control versus the mutant mice. 4.2.5 Lack of GlcNAc-TIII Activity Correlates with a Depletion of Bisected N-Glycans Whether loss of all GlcNAc-TIII activity would , as expected, result in loss of bisecting G l c N A c residues in N-glycans was addressed using the lectin E 4 - P H A which has been found to bind specifically to bisected N-glycans when employed under certain experimental conditions (Kobata and Yamashita, 1989, 1993). Us ing extracts from kidney harboring the three Mgat3 genotypes (Mgat3WT/Mgat3WT, Mgat3WT /Mgat3A, Mgat3&/Mgat3&), loss of E 4 - P H A binding was observed to correlate wi th the presence of the nu l l genotype (Figure 4.3). In addition, we looked for structural changes wi th the lectin L 4 - P H A . Though L 4 - P H A binding is not affected by the bisecting G l c N A c per se, the G l c N A c - T V glycosyltransferase which adds the L 4 - P H A reactive-pl,6 linked G l c N A c to the trimannosyl core is blocked by the bisecting G l c N A c . Consequently, we might expect to find an elevation of L 4 - P H A -reactive oligosaccharides in Mgat3 mutant mice. However, L 4 - P H A lectin blotting of kidney tissue extracts was not grossly affected by the Mgat3 mutation (Figure 4.3). 103 No Lectin + E4-PHA i I I 1 A/A +/A +/+ A/A +/A +/+ No Lectin + L4-PHA 1 I I 1 A/A +/A +/+ A/A +/A +/+ 205 — 97 * fl^r ~~  ^  31 Figure 4 .3 . Depletion of E,-PHA lectin binding correlates with a loss G l c N A c -TIII ac t iv i ty . Loss of E 4 -PHA lectin binding in kidney tissue derived from Mgat3 homozygous-null mice. Acid-treated kidney homogenates (30 ug protein) from wild-type (+/+), heterozygous ( + / A ) , and homozygous-null (A/A) mice were subjected to electrophoresis in 12% SDS-polyacrylamide minigels followed by transfer to a PVDF membrane. The membranes were blocked and incubated in the absence of lectin (no lectin) or with biotinylated lectins at 0.2 ng/ml (either E 4 -PHA or L 4 -PHA). Bound lectins and biotinylated molecular-mass standard proteins were detected using Vectastain ABC™ and ECL™. The positions of protein standards are in kilodaltons. Proteins detected in the absence of lectin presumably contain biotin which reacts with the avidin reagent (Vectastain ABC™). L 4 -PHA detects glycoproteins containing non-bisected tri- and tetra-antennary A/-glycans (Cummings and Kornfeld, 1982); L 4-PHA-positive bands are seen between 31 and 66 kDa. E 4 -PHA has a high affinity for bisected oligosaccharides (Kobata and Yamashita, 1989, 1993); E 4 - P H A -positive bands are visible in the wild-type extract between 66 and 205 kDa. These are completely absent in the Mgat3-nu\\ mouse extract and are reduced in the heterozygous extract. The band at 43 kDa in the Mgat3-nu\\ mouse extract is probably due to weak reactivity of E 4 -PHA with anon-bisected A/-glycan-containing glycoprotein (Kobata and Yamashita, 1 989, 1993) detected by L 4 -PHA. 104 In conclusion, these results suggest that the Mgat3 allele inactivated in this study appears solely responsible for the production of GlcNAc-TIII activity and the formation of bisecting G l c N A c residues in N-glycans. 4.2.6 Serum Metabolite Analyses A s the kidney appears to have the highest GlcNAc-TIII activity of n o r m a l tissues, measures of kidney function were sought. Serum samples were therefore analyzed at the U C S D Medical Center, which has a renal panel used to detect kidney dysfunction. Five samples, derived from animals approximately three months of age, from control and mutant mice, were analyzed. The values were averaged and the data are presented in Table 2. A t this age, the kidneys of mutant animals seemed to be performing satisfactorily as determined by this assay. Table 2. Renal Panel Wild Type Mgat3A/ Mgattf Bicarbonate (mM) 14.2 + 2.4 14.2 ± 2 . 9 Chloride (mM) 118.5 ± 2 . 3 118.6 ± 3.7 Sodium (mM) 151 .0±2 .7 150.3 ± 4.6 Potassium (mM) 5.4 ±0.7 6.2 ±1.0 Glucose (mM) 9.3 + 1.0 9.3 ±1.3 Blood Urea Nitrogen (mM) .8.6 + 1.4 9.0 ±2.9 Phosphate (mM) 2.9 ±0.5 2.8 ±0.3 Calcium (mM) 2.3 ±0.2 2.2 ±0.2 Creatinine (uM) 24 + 6 29 ±10 Total Bilirubin (|j.M) 6.8 + 0.7 8.2 ± 3.8 Direct Bilirubin (uM) 3.4 ±1.4 6.2 ± 3.9 Albumin (g/L) 12±1.4 11.±1.0 Total Protein (g/L) 4 4 ± 2 43.6 ±1 .7 ALT [alanine aminotransferase] (IU/L) 2 6 ± 5 27 ±13 AST [aspartate aminotransferase] (IU/L) 177 ± 5 9 188 ± 7 5 Alkaline Phosphatase (IU/L) 76±34 65 ±14 n=5 for wild type and Mgat3 null mice 105 4.2.7 Hematological Analyses To investigate possible hematologic abnormalities in mutant mice, peripheral blood was examined by Wright-Giemsa staining of blood smears. Neutrophils, lymphocytes and red blood cells from mutant animals displayed normal size and cellular morphology. In addition, automated differentials obtained via a C E L L - D Y N 3500 determined that the number and proportion of circulating leukocytes resembled controls (Table 3). Table 3. Peripheral Blood Hematology Wild Type Mqat3A/ Mgat^ White Blood Cells (cells/uL) 6500±1900 7500 ± 2300 Neutrophils 1000 ± 290 1000 ± 2 3 0 Lymphocytes 5100 ±2200 5800 ±2200 Red Blood Cells (x103/uL) 8730 + 410 8780 ± 370 Hemaglobin (g/L) 148 + 20 146 ± 3 9 Hematocrit (%) 45.8 ± 1.5 46.3 ±2 .0 Mean Cell Volume (fl_) 52.5 + 1.6 52.3 ± 1.4 Mean Corpuscular Hemaglobin (pg) 16.9 + 0.8 16.7 ±0 .9 Mean Corpuscular Hemaglobin Concentration (g/L) 323 +1.0 319 + 1.1 Red Cell Distribution Width (%) 16.7 ± 1.1 16±1 Platelets (x 103/uL) 882±133 871 ± 75 Mean Platelet Volume (fL) 4.3 + 0.2 4.5 ± 0.2 n=10 for wild type mice and n=12 for Mgat3 null mice. Furthermore, we decided to look at lymphocytes in our nul l animals since G l c N A c -TIII activity is developmentally regulated in these cells (Narasimhan et al., 1988). H i g h levels of activity are found in B lymphocytes whereas undetectable levels are found in T lymphocytes. For examination of the thymic and splenic compartments, F A C S (fluorescein-activated cell sorter) analyses were employed. Developmental profiles of control and mutant mice were judged by C D 4 / C D 8 staining. The four thymic subpopulations, CD4/CD8", C D 4 + / C D 8 + , CD4VCD8" and C D 4 / C D 8 + , were comparable wi th respect to the percentage of cells in each population. The 106 colonization of the spleen by CD4+ and CD8+ T cells and the numbers of B cells, distinguished by B220 and surface immunoglobul in labeling, present appeared norma l . 4.3 Discussion The biosynthetic pathway leading to complex and hybrid N-glycan production is conserved in mammalian organisms and has recently been found to be necessary for embryonic development (Ioffe and Stanley, 1994; Metzler et al., 1994). Addi t ional experiments to further restrict the ability of mammalian embryos to generate a diverse repertoire of N-glycans have thus been predicted to further i l luminate the identity and functions of N- l inked oligosaccharides participating i n normal and aberrant physiology (Marth, 1994). GlcNAc-TIII can act relatively early in the biosynthesis of complex and hybrid N- l inked oligosaccharides, immediately after the addition of a G l c N A c residue to the Manocl-3 arm of the core by G l c N A c - T I . The resulting bisecting G l c N A c residue prevents subsequent action by oc-mannosidase II, GlcNAc-TII , G lcNAc-TIV , G l c N A c T V , and core od-6-fucosyltransferase thereby l imit ing N-l inked oligosaccharide biosynthesis to bisected hybrid forms wi th a single Manocl-3-linked antenna which may be extended i n various ways (Schachter et ah, 1983; Schachter, 1986). Therefore, production of a Mga£3-null mouse wi th loss of GlcNAc-TIII activity and lack of bisecting G l c N A c residues may result in additional N-glycan branching wi th multiple antennae formation and decreased abundance of hybrid structures in cells which express a-107 mannosidase II. Regardless of the outcome, we reasoned that a mouse lacking GlcNAc-TIII would be useful in further determining the physiologic roles of complex and bisected hybrid N- l inked oligosaccharides and may ultimately be a necessary reagent for further studies. Studies accomplished herein have succeeded in generating mice lacking GlcNAc-TIII activity and bisecting G l c N A c residues in N-glycans. We suspected that lack of GlcNAc-TIII and the bisecting G l c N A c might promote increased branching of N- l inked oligosaccharides, by G l c N A c - T V for example. To investigate this possibility, kidney glycoproteins were analyzed via L 4 - P H A lectin blotting to detect the L 4 - P H A reactive G l c N A c - T V branched oligosaccharides. Our results show a very modest elevation in L 4 - P H A staining associated wi th the loss of Mgat3. A d d i t i o n a l structural analysis w i l l be needed to confirm that the deficiency observed reflects a total loss of bisecting G l c N A c residues. The modest increase may reflect a l o w G l c N A c - T V activity in normal kidney or that GlcNAc-TIII and G l c N A c - T V may have preferences for different substrates. Our data can not exclude the existence of some remaining bisecting G l c N A c residues resulting from a second GlcNAc-TIII isoenzyme not encoded by the Mgat3 gene. Southern genomic blotting did not reveal any Mgat3 cross-hybridizing elements. Moreover, the mutation generated herein deleted the entire protein encoding sequence wi th a corresponding loss of GlcNAc-TIII activity and depletion of E 4 - P H A lectin binding among homozygous-nu l l samples. Therefore, at this time, we do not have sufficient reason to i nvoke the existence of a second gene encoding GlcNAc-TIII activity. 108 Mice devoid of GlcNAc-TIII appeared normal (Table 1 and data not shown). Whi le mice lacking a functional Mgat3 allele did not display overt phenotypic consequences, additional studies were undertaken to determine whether some tissues and physiologic systems known to express GlcNAc-TIII were normal. They were similar in weight to wild-type offspring, matured equally wel l , and reproduced normally. Addit ional ly, tissues that normally exhibit the highest levels of G l c N A c -TIII, including brain and kidney, were similar in wet-weight mass in mice lacking a functional Mgat3 allele. Moreover, tissues from mutant mice were histologically analyzed and appeared to lack gross alterations in structure (data not shown). The identity and characteristics of circulating leukocytes and red blood cells were n o r m a l in the absence of a functional Mgat3 gene. Addit ional ly, serum metabolite levels used to assess kidney function were also unaffected. Therefore, at the present time, and in a relatively stress-free environment, mice lacking a functional Mgat3 gene and bisected N-glycans appear to develop, function and reproduce normal ly . Furthermore, behavioral alterations in Mgat3-mx\\ mice have not been observed thus far, wi th the oldest of such animals presently reaching the age of 1 year. Interestingly, another laboratory has also created an Mgat3 mutation via neo insertion but have observed different findings. Their mutant animals suffer from a number of abnormalities including being underweight and having a staggered type of locomotion (Bhaumik and Stanley, unpublished observations). Picking up the animals results in unusual behaviour such as leg clasping and curling up into a ball. This apparent contradiction could have multiple explanations including the nature of the mutation and/or the strain background. Addi t ional ly , an observed 109 phenotype could result from disrupting genes adjacent to the target. A s conventional targeting inserts a drug selection cassette wi th its own promoter, translational start/stop signals and poly A sequence, it is possible that nearby genes could be affected. Such consequences would necessarily be l inked to the target and therefore confound the interpretation of the phenotype (reviewed in Olson et ah, 1996). Al though our results do not indicate a physiologic role for GlcNAc-TIII and bisecting G l c N A c residues in N-l inked oligosaccharides, we do not view this data as evidence that a function for the bisecting G l c N A c modification does not exist. Addi t iona l experiments wi th Mgat3-mx\\ mice can be focused on the potential roles for GlcNAc-TIII action in other processes, such as cell metastasis, tumourigenesis, and homing (Narasimhan et al., 1988b; Nishikawa et al., 1988a; Pascale et al., 1989; Yoshimura et al., 1995a,b,c, 1996; Miyosh i et al., 1995). For instance, high levels of GlcNAc-TIII have been strongly associated wi th hepatocarcinogenesis but the significance of this finding is unknown (Narasimhan et al, 1988b; Nish ikawa et al, 1988a; Pascale et al., 1989). Since bisected structures have been shown to confer a protective effect from natural killer cell cytotoxicity (Yoshimura et al, 1996), it may allow proliferating cells to evade immune system attack. To determine the strength of this association, control and Mgat3 nul l animals could participate in experimental models of liver carcinogenesis to look at effects on tumour incidence. Furthermore, immunoglobul in structures which contain the bisecting G l c N A c may depend u p o n that modification for structural stability and functional integrity in Fc binding 110 (Opdenakker et al., 1993). Studies of interest include looking at antibody binding to Fc receptors and ability of these antibodies to activate complement. Hematopoietic reconstitution and analyses of tumourigenesis in Mgat3-nu\\ mice bearing various oncogenic lesions are also relevant experiments to undertake, and may now include the production of cell lines bearing either the Mgat3A/Mgat3A or Mgat3F/Mgat3F genotype, the latter allowing for the deletion of the Mgat3 gene in vitro by Cre recombinase expression systems. Furthermore, cell lines derived fromMgat3F/Mgat3F could be used for studying the role of the bisecting G l c N A c o n expressed proteins as they could be easily compared before and after Cre application (expression of other glycosyltransferases would be expected to remain unchanged). Even though a significant percentage of genetically-manipulated mice bearing n u l l alleles derived by gene-targeting techniques have been subsequently found to be normal in the laboratory environment, the high degree of evolut ionary conservation of genes such as Mgat3 throughout mammalian evolution may w e l l be indicative of yet to be established selective advantages by enabling an appropriate response to a changing stressful environment. O n the other hand, an apparent absence of a phenotype may also be attributed to alternative oligosaccharides and biological mechanisms that compensate for the missing enzyme. Further studies wi th this nu l l animal system should provide insight into the biological role(s) of GlcNAc-TIII . I l l B i b l i o g r a p h y Abremski ,K. and Hoess,R. (1984) Bacteriophage PI site-specific recombination. Purification and properties of the Cre recombinase protein. /. Biol. Chem. 259, 1509-1514. Abremski,K., Hoess,R. and Sternberg,N. (1983) Studies on the properties of P I site-specific recombination: evidence for topologically unlinked products following recombination. Cell, 32, 1301-1311. Al len , S. D. , Tsai, D. and Schachter, H . (1984) Control of glycoprotein synthesis. The in vitro synthesis by hen oviduct membrane preparations of hybrid asparagine-linked oligosaccharides containing 5 mannose residues. /. Biol. Chem. 259, 6984-6990. Alhadeff, J. A . (1989)A1A CRC Crit. Rev. Oncol. Hemat. 9, 37-107. A r a k i , H . , Jearnpipatkul, A . , Tatsumi, H . , Sakurai, T., Ushio, K . , Muta , T. and Oshima, Y . (1985) Molecular and functional organization of yeast plasmid p S R l . /. Mol. Biol. 182, 191-203. Arak i ,K. , A r a k L M . , Miyazaki,J.-I. and Vassalli,P. (1995) Site-specific recombination of a transgene in fertilized eggs by transient expression of Cre recombinase. Proc. Natl. Acad. Sci. USA, 92, 160-164. Arbones, M . L . , Ord , D . C , Ley, K. , Ratech, H . , Maynard-Curry, C , Often, G. , Capon, D. J., and Tedder, T. F. (1994) Lymphocyte homing and leukocyte rolling and migration are impaired in L-selectin-deficient mice. Immunity 1, 247-260. Arnheiter, H . , Skuntz, S., Notebom, M . , Chang, S. and Meier, E. (1990) Transgenic mice wi th intracellular immunity to influenza virus. Cell 62, 51-61. Aust in , S., Ziese, M . and Sternberg, N . (1981) A novel role for site-specific recombination in maintenance of bacterial replicons. Cell, 25, 729-736. Barondes,S.H., Cooper,D.N.W., Gi t t ,M.A. and Leffler,H. (1994) Galectins. Structure and function of a large family of animal lectins. /. Biol. Chem., 269, 20807-20810. Baubonis,W. and Sauer,B. (1993) Genomic targeting wi th purified Cre recombinase. Nucleic Acids Res., 9, 2025-2029. 112 Bhaumik, M . , Seldin, M.F . , and Stanley, P. (1995) Cloning and chromosomal mapping of the mouse Mgat3 gene encoding N-acetylglucosaminyltransferase III. Gene 164, 295-300. B i r d J . M . and Kimber,S.J. (1984) Oligosaccharides containing fucose linked a(l-3) and a(l-4) to N-acetylglucosamine cause decompaction of mouse morulae. Dev. Biol., 104,449-460. Bischoff, J. and Kornfeld, S. (1983) Evidence for an alpha-mannosidase in endoplasmic reticulum of rat liver. /. Biol. Chem. 258, 7907-7914. Boyle, A . L . , Feltquite, D . M . , Dracopoli, N . C . , Housman, D.E. , and Ward, D .C . (1992) Rapid physical mapping of cloned D N A on banded mouse chromosomes by fluorescence in situ hybridization. Genomics 12, 106-115. Brisson, J., and Carver, J.P. (1983) The relation of three dimensional structure to biosynthesis in the N- l inked oligosaccharides. Can. }. Biochem. Cell Biol., 61, 1067-1078. Broach, J. R., Guarascio, V . R. and Jayaram, M . (1982) Recombination wi th in the yeast plasmid 2mu circle is site-specific. Cell 29, 227-234. Brockhausen et al. (1988) Biochem. Cell Biol. 66, 1134-1151. Brockhausen, I., Romero, P .A. , and Herscovics, A . (1991) Glycosyltransferase changes upon differentiation of CaCo-2 human colonic adenocarcinoma cells. Cancer Res. 51, 3136-3142. Burgess, D .L . , Kohrman, D . C , Gait, J., Plummer, N . W . , Jones, J .M. , Spear, B., and Meisler, M . H . (1995) Mutation of a new sodium channel gene, Scn8a, in the mouse mutant 'motor endplate disease'. Nature Genetics 10, 461-465. Campbell , C. and Stanley, P. (1984) A dominant mutation to ricin resistance in Chinese hamster ovary cells induces UDP-GlcNAc:glycopeptide [3-4-N-acetylglucosaminyltransferase III activity. /. Biol. Chem. 259, 13370-13378. Capecchi,M.R. (1989) Alter ing the genome by homologous recombination. Science, 244,1288-1292. Chapman, A . , L i , E. and Kornfeld, S. (1979) The biosynthesis of the major l ip id-linked oligosaccharide of Chinese hamster ovary cells occurs by the ordered addition of mannose residues. /. Biol. Chem. 254, 10243-10249. CharukJ .H . , TanJ. , Bernardini,M., Haddad,S., Reithmeier,R.A., JaekenJ. and Schachter,H. (1995) Carbohydrate-deficient glycoprotein syndrome type II. A n autosomal recessive N-acetylglucosaminyltransferase II deficiency different from 113 typical hereditary erythroblastic multinuclearity, wi th a positive acidified-serum lysis test ( H E M P A S ) . Eur. J. Biochem., 230, 797-805. Cummings,R.D. and Kornfeld,S. (1982) Characterization of the structural determinants required for the high affinity interaction of asparagine-linked oligosaccharides wi th immobil ized Phaseolus vulgaris leukoagglutinating and erythroagglutinating lectins. /. Biol. Chem., 257, 11230-11234. Cummings, R.D. , and Kornfeld, S. (1982) Characterization of the structural determinants required for the high affinity interaction of asparagine-linked oligosaccharides wi th immobil ized Phaseolus vulgaris leukoagglutinating and erythroagglutinating lectins. /. Biol. Chem. 257, 11230-11234. Delorme, E., Lorenzini, T., Giffin, J., Mart in, F., Jacobsen, F., Boone, T. and Elliot, S. (1992). Role of glycosylation on the secretion and biological activity of erythropoietin. Biochemstry 31, 9871-9876. Demetriou,M., Nabi,I.R., Coppol ino,M. , Dedhar,S. and DennisJ .W. (1995) Reduced contact-inhibition and substratum adhesion in epithelial cells expressing G l c N A c -transferase V . /. Cell Biol, 130, 383-392. DennisJ.W., Laferte,S., Waghorne,C, Breitman,M.L. and Kerbel,R.S. (1987) (31-6 branching of Asn-l inked oligosaccharides is directly associated wi th metastasis. Science, 236, 582-585. Dennis, J. W., Kosh, K . , Bryce, D. M . and Breitman, M . L . (1989) Oncogenes conferring metastatic potential induce increased branching of Asn-l inked oligosaccharides to rat2 fibroblasts.Oncogene 4, 853-860. van Deursen, J. and Wieringa, B. (1992) Targeting of the creatine kinase M gene in embryonic stem cells using isogenic and non-isogenic vectors. Nucleic Acids Res. 20, 3815-3820. Drickamer, K . (1988) Two distinct classes of carbohydrate-recognition domains in animal lectins./ Biol. Chem. 263, 9557-9560. Drickamer,K. (1993) Biology of animal lectins. Ann. Rev. Cell Biol, 9, 237-264. Dube, S., Fisher, J. W. and Powell , J. S. (1988) Glycosylation at specific sites of erythropoietin is essential for the biosynthesis, secretion and biological function. /. Biol. Chem. 263, 17516-17517. Easton, E.W., Bolscher, J .G.M. , and van den Eijnden, D . H . (1991) Enzymatic amplification involving glycosyltransferases forms the basis for the increased size of asparagine-linked glycans at the surface of NIH3T3 cells expressing the N -ras proto-oncogene. /. Biol. Chem. 266, 21674-21680. 114 Eggens,L, Fenderson,B., ToyokunLT., Dean,B., Stroud,M. and Hakomori,S. (1989) Specific interaction between Le x and Le x determinants. /. Biol. Chem., 264, 9476-9484. Elting, J. J., Chen, W. W. and Lennarz, W. J. (1980) Characterization of a glucosidase involved in an initial step in the processing of oligosaccharide chains. /. Biol. Chem. 255, 2325-2330. Epstein, C. J. The consequences of chromosomal imbalance: principles, mechanism and models. Cambridge University Press, 1986. EskoJ .D. (1991) Genetic analysis of proteoglycan structure, function and metabolism. Curr. Opinion Cell Biol, 3, 805-816. Feizi,T. and Chi lds ,R.A. (1987) Carbohydrates as antigenic determinants of glycoproteins. Biochem. J., 245, 1-11. Feizi,T. (1985) Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are onco-developmental antigens. Nature, 314, 53-57. Fenderson,B.A., Sehavi,U. and Hakomori,S. (1984) A multivalent lacto-N-furopentaose III-lysyllysine conjugate decompacts preimplantation mouse embryos, while the free oligosaccharide is ineffective. /. Exp. Med., 160, 1591-1596. Francke U . (1994) Digitized and differentially shaded human chromosome ideograms for genomic applications. Cytogenet and Cell Genet 65, 206-219. Frenette, P.S., Mayadas, T. N . , Rayburn, H . , Hynes, R. O. and D . D. (1996) Susceptibility to infection and altered hematopoiesis in mice deficient in both P- and E-selectins. Cell, 84, 563-574. Fukuda ,M.N. (1990) H E M P A S disease: genetic defect of glycosylation. Glycobiology, 1, 9-15. Fukuda, M . , Sasaki, H . and Fukuda, M . N . (1989) Structure and role of carbohydrate in human erythropoietin. Adv. Exp. Med. Bio. 271, 53-67. Fukuda, M . N . , Sasaki, H . , Lopez, L . and Fukuda, M . (1989) Survival of recombinant erythropoietin in the circulation: the role of carbohydrates. Blood 73, 84-89. Gleeson, P. A . and Schachter, H . (1983) Control of glycoprotein synthesis. /. Biol. Chem. 258, 6162-6173. Grinna, L . S. and Robbins, P. W. (1978) Glycoprotein synthesis. Rat liver microsomal glucosidases which process oliogsaccharides. /. Biol. Chem. 254, 8814-8818. 115 G u , H v MarthJ .D. , Orban,P.C, Mossmann,H. and Rajewsky,K. (1994a) Deletion of a D N A polymerase p gene segment in T cells using cell type-specific gene targeting. Science, 265, 103-106. G u , H . , Zou , Y-R. and Rajewsky, K . (1994b) Independent control of the immunoglobulin switch recombination at individual switch regions evidenced through Cre-ZoxP-mediated gene targeting. Cell 73, 1155-1164. Haber, J. E. (1992) Exploring pathways of homologous recombination. Cur. Opin. Cell Biol. 4, 401-412. Harpaz ,N. and Schachter,H. (1980) Control of glycoprotein synthesis. Bovine colostrum UDP-N-acetylglucosamine:a-D-mannoside p2-N-acetylglucosaminyltransferase I. Separation from UDP-N-acetylglucosamine:a-D-mannoside p2-N-acetylglucosaminyltransferase II, partial purification, and substrate specificity. /. Biol. Chem., 255, 4885-4893. Harpaz ,N. and Schachter,H. (1980) Control of glycoprotein synthesis. Processing of asparagine-linked oligosaccharides by one or more rat liver golgi a-D-mannosidases dependent on the prior action of UDP-N-acetylglucosamine:a-D-mannoside p2-N-acetylglucosaminyltransferase I. /. Biol. Chem., 255, 4894-4902. Hart ,G.W., Haltiwanger,R.S., Hol t ,G.D. and Ke l ly ,W.G. (1989) Glycosylation in the nucleus and cytoplasm. Annu. Rev. Biochem., 58, 841-874. Hasty, P., Crist, M . , Grompe, M . and Bradley, A . (1994) Efficiency of insertion versus replacement vector targeting varies at different chromosomal loci. Mol. Cell. Biol. 14,8385-8390. Hasty, P., Crist, Rivera-Perez, J. and Bradley, A . (1991) The length of homolgy required for gene targeting in embryonic stem cells. Mol. Cell. Biol. 11, 5586-5591. Heng, H . and Tsui, L - C . (1993) Modes of D A P I banding and simultaneous in situ hybridizat ion. Chromosoma, 102, 325-332. Hennet,T., Hagen,F.K., Tabak,L.A. and MarthJ .D. (1995) T-cell-specific deletion of a polypeptide N-acetylgalactosaminyl- transferase gene by site-directed recombination. Proc. Natl. Acad. Sci. USA., 92, 12070-12074. Hoess, R. H . , Abremski, K. , Irwin, S., Kendall , M . and Mack, A . (1990) D N A specificity of the Cre recombinase resides in the 25 kDa carboxyl domain of the protein. /. Mol. Biol, 216, 873-882. Hoess,R.H., Wierzbicki ,A. and Abremski,K. (1986) The role of the loxP spacer region in P I site-specific recombination. Nucleic Acids Res., 14, 2287-2300. 116 Hoess,R.H. and Abremski,K. (1985) Mechanism of strand cleavage and exhcnage in the Cxe-lox site-specific recombination system. /. Mol. Biol, 181, 351-362. Hoflack, B., Cacan, R. and Verbert, A . (1981) Dolichol pathway in lymphocytes from rat spleen. Influence of the glucosylation on the cleavage of dolichyldiphosphate oligosaccharides into phosphooligosaccharides. Eur. }. Biochem. 117, 285-290. Humphries, M . J., Matsumoto, K . , White, S.L. and Olden, K . (1986) Inhibition of experimental metastasis by casanospermine in mice: blockage of two distinct stages of tumor colonization by oligosaccharide processing inhibitors. Cancer Res. 46, 5215-5222. H u g , H . , Costas, M . , Staeheli, P., Aebi , M . and Weissmann, C.(1988) Organization of the murine M x gene and characterizatio of its interferon- and virus-inducible promoter. Molecular and Cellular Biology 8 , 3065-3079. Ihara, Y . , Nishikawa, A . , Tohma, T., Soejima, H . , Ni ikawa, N . , and Taniguchi, N . (1993) c D N A cloning, expression, and chromosomal localization of human N -acetylglucosaminyltransferase III (GnT-III). /. Biochem. 113, 692-698. Imai, N . , Higuchi , M . , Kawamura, A . , Tomonoh, K . , Oh-Eda, M . , Fujiwara, M . , Shimonaka, Y . and Ochi, N.(1990) Physiochemical and biological characterization of asialoerythropoietin. Suppressive effects of sialic acid in the expression of biological activity of human erythropoietin in vitro. Eur. }. Biochem. 194, 457-462. Ioffe,E. and Stanley,P. (1994) Mice lacking N-acetylglucosaminyltransferase I activity die at midgestation, revealing an essential role for complex or hybrid N- l inked carbohydrates. Proc. Natl. Acad. Sci. USA, 91 , 728-732. Irimura, T., Gonzalez, R. and Nicolson, G.L. (1981) Effects of tunicamycin on B16 metastatic melanoma cell surface glycoproteins and blood-borne arrest and survival properties. Cancer Res. 41 , 3411-3418. I S C N (1978) A n International System for Human Cytogenetic Nomenclature. Cytogenet ad Cell Genet 2 1 , 309-404. Jeannotte,L., Ru iz J .C . and Robertson,E.J. (1991) L o w level of Hox l .3 gene expression does not preclude the use of promoterless vectors to generate a targeted gene disruption. Mol. Cell. Biol, 11 , 5578-5585. Joziasse, D . H . , Schiphorst, W. E. C. M . , Van den Eijnden, D . H . , Van Kuik , J. A . , V a n Halbeek, H . and Vliegenthart, J. F. G . (1987) Branch specificity of bovine colostrom CMP-sia l ic acid: Galpi ,4GlcNAcRa2,6 sialyltransferase. /. Biol. Chem. 262, 2025-2033. 117 Kelker, H . C , Y ip , Y . K . , Anderson, P. and Vilcek, J. (1983) Effects of glycosidase treatment on the physicochemical properties and biological activity of human interferon-gamma. /. Biol. Chem. 258, 8010-8013. Khan, S.H., Compston, C . A . , Palcic, M . M . , and Hindesgaul, O. (1994) Synthesis of a di-O-methylated pentasaccharide for use in the assay of N -acetylglucosaminyltransferase III activity. Carbohydrate Res. 262, 283-295. Ki lby ,N.J . , Snaith,M.R. and M u r r a y J . A . H . (1993) Site-specific recombinases: tools for genome engineering. Trends Genet., 9, 413-421. Kobata A . and Yamashita, K . (1989) Affinity chromatography of oligosaccharides on E4-phytohemagglutinin-agarose column. Methods Enzymol. 179, 46-54. Kobata, A . and Yamashita, K . (1993) Fractionation of oligosaccharides by serial affinity chromatography wi th use of immobilized lectin columns. In Fukuda, M . and Kobara, A . (eds.), Glycobiology - A Practical Approach, Oxford University Press, Oxford, U K , 103-125. Koenderman, A . H . L . , Wijermans, P.W., and van den Eijnden, D . H . (1987) Changes in the expression of N-acetylglucosaminyltransferase III, IV, V associated with the differentiation of HL-60 cells. FEBS Lett. 222:42-46. Koenderman, A . H . L . , Koppen, P.L., Koeleman, C . A . M . , and Van den Eijnden, D . H . (1989) N-acetylglucosaminyltansferase III, IV and V activities in Novikoff ascites tumour cells, mouse lymphoma cells and hen oviduct. European Journal of Biochemistry 181, 651-655. Kohrman, D.C. , Plummer, N . W . , Schuster, T., Jones, J .M. , Jang, W. , Burgess, D .L . , Gait, J., Spear, B.T., Meisler, M . H . (1995) Insertional mutation of the motor endplate disease (med) locus on mouse chromosome 15. Genomics 26, 171-177. Kornfeld,R. and Kornfeld,S. (1985) Assembly of asparagine-linked oligosaccharides. Ann. Rev. Biochem., 54, 631-664. Kornfeld, S. and Mel lmann, I. (1989) Ann. Rev. Cell Biol. 5, 483. Kozak, M . (1991) Structural features in eukaryotic m R N A s that modulate the initiation of translation. /. Biol. Chem. 266, 19867-19870. Kuhn,R., Schwenk,F., Aguet ,M. and Rajewsky,K. (1995) Inducible gene targeting in mice. Science, 269, 1427-1429. Laemmli , U . K . (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T 4 . Nature 227, 680-685. 118 Lakso,M. , Sauer,B., Mosinger,B.,Jr., Lee,E.J., Manning,R.W., Yu ,S .H. , Mulde r ,K .L . and Westphal,H. (1992) Targeted oncogene activation by site-specific recombination in transgenic mice. Proc. Natl. Acad. Sci. U. S. A., 89, 6232-6236. Labow, M . A . , Norton, C. R., Rumberger, J. M . , Lombard-Gillooly, K . M . , Shuster, D . J., Hubbard, J., Bertko, R., Knaack, P. A . , Terry, R. W., Harbison, M . L . , Hontgen, F., Stewart, C. L . , Mclntyre, K . W., W i l l , P. C , Burns, D . K . and Wolitsky, B. A . (1994) Characterization of E-selectin deficient mice: demonstration of overlapping function of the endothelial selectins. Immunity 1, 709-720. Lichter, P., Tang C.J., Cal l , K. , Hermanson, G. , Evans, G .A. , et al. (1990) H igh -resolution mapping of human chromosome 11 by in situ hybridization wi th cosmid clones. Science 247, 64-69. L indh , I. and Hindsgaul, O. (1991) / Am. Chem. Soc, 113,216-223. L i , E . and Kornfeld, S. (1978) Structure of the altered oligosaccharide present in glycoproteins from a clone of Chinese hamster ovary cells deficient in N -acetylglucosaminyltransferase activity. /. Biol. Chem. 253, 6426-6431. L u , Y . and Chaney, W. (1993) Induction of N-acetylgucosaminyltransferase V by elevated expression of activated or proto-Ha-ras oncogenes. Mol. Cell. Biochem. 122, 85-92. M a , Z . M . , Grubb, J .H. and Sly, W.S. (1991) Cloning, sequencing, and functional characterization of the murine 46-kDa mannose 6-phosphate receptor. /. Biol. Chem. 266, 10589-10595. Mack, A . , Sauer, B., Abremski, K . and Hoess, R. (1992)Stoichiometry of the Cre recombinase bound to the lox recombining site. Nucleic Acids Research, 20, 4451-4455. Mahon, K . A . , Chapelinsky, A . B., Khi l lan, J. S., Overbeek, P. A . , Piatigorsky, J. and Westphal, H . (1987) Oncogenesis of the lens in transgenic mice. Science 235, 1622-1628. Maly ,P . , Thall ,A., Petryniak,B., Rogers,C.E., Smith,P.L., Marks ,R.M. , Kelly,R.J. , Gersten,K.M., Cheng,G., Saunders,T.L. and et al., (1996) The a (1, 3) fucosyltransferase Fuc-TVII controls leukocyte trafficking through an essential role in L - , E-, and P-selectin ligand biosynthesis. Cell, 86, 643-653. Mar thJ .D. (1994) W i l l the transgenic mouse serve as a Rosetta Stone to glycoconjugate function? Glycoconjugate J., 11, 3-8. Mar thJ .D. (1996) Recent advances in gene mutagenesis by site-directed recombination. /. Clin. Invest., 97, 1999-2002. 119 MarthJ .D. , Peet,R., Krebs,E.G. and Perlmutter,R.M. (1985) A lymphocyte-specific protein-tyrosine kinase gene is rearranged and overexpressed in the murine T cell lymphoma L S T R A . Cell, 43, 393-404. Marshal l , R. D . (1972) Glycoproteins. Ann. Rev. Biochem. 41, 673-702. Mayadas, T. N . , Johnson, R. C., Rayburn, H . , Hynes, R. O. and Wagner, D . D . (1993) Leukocyte rolling and extravasation are severely compromised in P-selectin deficient mice.CeZZ 74, 541-554. McEver,R.P. , Moore ,K.L. and Cummings,R.D. (1995) Leukocyte trafficking mediated by selectin-carbohydrate interactions. /. Biol. Chem., 270, 11025-11028. Metzler ,M. , Gertz,A., Sarkar,M., Schachter,H., SchraderJ.W. and Mar thJ .D. (1994) Complex asparagine-linked oligosaccharides are required for morphogenic events during post-implantation development. EMBO }., 13, 2056-2065. Miyosh i , E. , Ihara, Y . , Hayashi, N . , Fusamoto, H . , Kamada, T., and Taniguchi, N . (1995) Transfection of N-acetylglucosaminyltransferase III gene suppresses expression of hepatitis B virus in a human hepatoma cell line, HB611. /. Biol. Chem. 270, 28311-28315. Mock, B .A. , Eppig, J.T., Neumann, P.E., and Hupp i , K . E . (1994) Mouse Chromosome 15 Committee Report. Mammalian Genome 5, S217-S228. Murphy , L . A . and Spiro, R. G . (1981) Transfer of glucose to oligosaccharide-lipid intermediates by thyroid microsomal enzymes and its relationship to the N -glycosylation of proteins. /. Biol. Chem. 256, 7487-7494. Nagy,A. , RossantJ., Nagy,R., Abramow-Newerly,W. and RoderJ.C. (1993) Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc. Natl. Acad. Sci. USA, 90, 8424-8428. Nakamori , S., Kameyama, M . , Imaoka, S., Furukawa, H . , Ishikawa, O., Sasaki, Y . , Kabuto, T., Iwanaga, T., Matsushita, Y . and Irimura, T. (1993) Increased expression of sialyl Lewisx antigen correlates wi th poor survival in patients wi th colorectal carcinoma: clinicopathological and immunohistochemical study. Cancer Res. 53, 3632-3637. Narasimhan,S. (1982) Control of glycoprotein synthesis. UDP-GlcNAc:glycopept ide (34-N-acetylglucosaminyltransferase II, an enzyme in hen oviduct which adds G l c N A c in pl-4 linkage to the (3-linked mannose of the trimannosyl core of N-glycosyl oligosaccharides. /. Biol. Chem., 257, 10235-10242. 120 Narasimhan, S., Lee, J.W.W., Cheung, R.K. , Gelfand, E.W., and Schachter, H . (1988a) (3-1,4-N-acetylglucosaminyltransferase III activity in human B and T lymphoctye lines and in tonsillar T and B lymphocytes. Biochem. Cell Biol. 66, 889-900. Narasimhan,S., FreedJ.C. and Schachter,H. (1986) The effect of a "bisecting" N -acetylglucosaminyl group on the binding of biantennary, complex oligosaccharides to concanavalin A , Phaseolus vulgaris erythroagglutinin (E-PHA) , and Ricinus communis agglutinin (RCA-120) immobil ized on agarose. Carbohydr. Res., 149, 65-83. Narasimhan, S., Schachter, H . , and Rajalakshmi, S. (1988b) Expression of N -acetylglucosaminyltransferase III in hepatic nodules during rat liver carcinogenesis promoted by orotic acid. Journal of Biological Chemistry , 263, 1273-1281. Narh i , L .O . , Arakawa, T., A o k i , K . H . , Elmore, R., Rohde, M.F . , Boone, T. and Strickland, T.W. (1991) The effect of carbohydrate on the structure and stability of erythropoietin. /. Biol. Chem. 266, 23022-23026. Nicolson, G .L . (1989) Curr. Opin. Cell Biol, 1, 1009. Nishikawa, A . , Fujii, S., Sugiyama, T., Hayashi, N . and Taniguchi, N . (1988a) H i g h expression of an N-acetylglucosaminyltransferase III in 3'-methyl DAB-induced hepatoma and ascites hepatoma. Biochem. Biophys. Res. Commun., 152, 107-112. Nishikawa, A . , Ihara, Y. , Hatakeyama, M . , Kangawa, K. , and Taniguchi, N . (1992) Purification, c D N A cloning, and expression of UDP-N-acetylglucosamine: (J -D-mannoside p-l,4N-acetylglucosaminyltransferase III from rat kidney. /. Biol. Chem. 267, 18199-18204. Nishikawa, A . , Shigeru, F., Toshihiro, S., and Taniguchi, N . (1988b) A method for the determination of N-Acetylglucosaminyltransferase III activity in rat tissues involving H P L C . Analytical Biochemistry 170, 349-354. Nishikawa,Y., Pegg,W., Paulsen,H. and Schachter,H. (1988) Control of glycoprotein synthesis. Purification and characterization of rabbit liver U D P - N -acetylglucosamine: oc-3-D-mannoside (3-1,2-N- acetylglucosaminyltransferase I. /. Biol. Chem. 263, 8270-8281. 0'Gorman,S., Fox,D.T. and Wahl , G . M . (1991) Recombinase-mediated gene activation and site-specific integration in mammalian cells. Science 251, 1351-1355. Olmsted, J.B. (1981) Affinity purification of antibodies from diazotized paper blots of heterogenous protein samples. /. Biol Chem. 256, 11955-11957. 121 Olson, E . N . , Arno ld , H . H . , Rigby, P.W. and Wold , B.J. (1996) K n o w your neighbours: three phenotypes in nul l mutants of the myogenic b H L H gene MRF4 . Cell 85, 1-4. Opdenakker, G. , Rudd, P. M . , Ponting, C P . and Dwek, R. A . (1993) Concepts and principles of glycobiology. FASEB }. 7, 1330-1337. Oppenheimer, C. L . and H i l l , R. L . (1981) Purification and characterization of a rabbit liver a 1,3 mannoside p 1,2 N-acetylglucosaminyltransferase. /. Biol. Chem. 256, 799-804. Oppenheimer,C.L., Eckhardt,A.E. and H i l l , R . L . (1981) The nonidentity of porcine N -acetylglucosaminyltransferases I and II. /. Biol. Chem., 256, 11477-11482. Orban,P.C, Chui ,D. and MarthJ .D. (1992) Tissue- and site-specific D N A recombination in transgenic mice. Proc. Natl. Acad. Sci. U. S. A., 89, 6861-6865. Palcic, M . M . , Ripka, J., Kaur, K . J., Shoreibah, M . , Hindsgaul, O. (1990) Regulation of N-acetylglucosaminyltransferase V activity. Kinetic comparisons of parental, Rous sarcoma virus-transformed B H K and L-phytohemagglutinin-resistant B H K cells using synthetic substrates and an inhibitory substrate analog. /. Biol. Chem. 265, 6759-6769. Pascale, R., Narasimhan, S., and Rajalakshmi, S. (1989) Expression of N -acetylglucosaminyltransferase III in hepatic nodules generated by different models of rat liver carcinogenesis. Carcinogenesis 10, 961-964. Palmiter, R. and Brinster, R. L . (1986) Germline transformation of mice. Ann. Rev. Genet. 20, 465-499. Philips, M . L . , Schwartz, B. R., Etzioni, A . , Bayer, R., Ochs, H . D. , Paulson, J. C . and Harlan, J. M . (1995) Neutrophil adhesion in leukocyte adhesion deficiency syndrome type-2. /. Clin. Invest. 96, 2898-2906. Powell , L . D. , Paneerselvam, K. , Vi j , R., Diaz, S., Manz i , A . , Buist, N , Freeze, H . and Vark i , V . (1994) Carbohydrate-deficient glycoprotein syndrome - not an N- l inked processing defect, but an abnormality in l ipid-linked oligosaccharide biosynthesis? /. Clin. Vest. 94, 1901-1909. PowelLL.D. and V a r k L A . (1995) I-type lectins. /. Biol. Chem., 270, 14243-14246. Pownall , S., Kozak, C. A . , Schappert, K . , Sarkar, M . , H u l l , E., Schachter, H . and Marth, J. D . (1992) Molecular cloning and characterization of the mouse U D P - N -acetylglucosamine:a-3-D-mannoside p-l,2-N-acetylglucosaminyltransferase I gene. Genomics 12, 699-704. 122 Q i n , M . , Bayley ,C, Stockton,T. and Ow,D.W. (1994) Cre recombinase-mediated site-specific recombination between plant chromosomes. Proc. Natl. Acad. Sci. USA, 91, 1706-1710. Ramirez-Solis,R., Davis, A . C. and Bradley,A. (1993) Gene targeting in mouse embryonic stem cells. Meth. In. Enzymol. 225, 855-878. Ramirez-Solis,R v L iu ,P . and Bradley,A. (1995) Chromosome engineering in mice. Nature, 378, 720-724. Rearick, J. I., Chapman, A . and Kornfeld, S. (1981) Glucose starvation alters l ip id-l inked oligosaccharide biosynthesis in Chinese hamster ovary cells. /. Biol. Chem. 256, 6255-6231. Riele, H . T., Maandag, E. R. and Berns, A . (1992) Highly efficient gene targeting in embryonic stem.cells through homologous recombination wi th isogenic D N A constructs. Proc. Natl. Acad. Sci. USA 89, 5128-5132. Robertson, M . A . , Etchison, J.R., Robertson, J.S., Summers, D.F. and Stanley, P. (1978) Specific changes in the oligosaccharide moieties of V S V grown in different lectin-resistant C H O cells. Cell 13, 515-526. Robertson,E. J., Bradley,A., Kuehn ,M. and Evans,M. (1986) Germ-line transmission of genes introduced into cultured pluripotential cells by retroviral vector. Nature, 323, 445-448. Robertson, E. J. (1987) Teratocarcinomas and embryonic stem cells: a practical approach. IRL Press Limited, Oxford and Washington, D . C. Roden, L . (1980) The biochemistry of glycoproteins and proteoglycans, 267. Plenum Press, N e w York. Sadler, J. E . (1984) Biosynthesis of glycoproteins: formation of O-linked oligosaccharides. In Biology of Carbohydrates, V o l . 2, 199. Wiley, N e w York. Sai toh,O v W a n g , W . C , Lotan,R. and Fukuda,M. (1992) Differential glycosylation and cell surface expression of lysosomal membrane glycoproteins in sublines of a human colon cancer exhibiting distinct metastatic potentials. /. Biol. Chem., 267, 5700-5711. Sauer,B. (1987) Functional expression of the cre-lox site-specific recombination system in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol., 7, 2087-2096. Sauer,B. and Henderson,N. (1989) Cre-stimulated recombination at ZoxP-containing D N A sequences placed into the mammalian genome. Nucleic Acids Res., 17, 147-161. 123 Sauer,B., Whealy ,M. , Robbins,A. and Enquist,L. (1987b) Site-specific insertion of D N A into a pseudorabies virus vector. Proc. Natl. Acad. Sci. USA, 84, 9108-9112. Sauer,B. and Henderson,N. (1988) The cyclization of linear D N A in Escherichia coli by site-specific recombination. Gene, 70, 331-341. Sauer, B., and Henderson, N . (1988) Site-specific D N A recombination in mammalian cells by Cre recombinase of bacteriphage P I . Proc. Natl. Acad. Sci. USA, 85, 5166-5170. Sauer,B. and Henderson,N. (1990) Targeted insertion of exogenous D N A into the eukaryotic genome by the Cre recombinase. New Biol., 2, 441-449. Sauer, B. (1992) Identification of cryptic lox sites in the yeast genome by selection for Cre-mediated chromosome translocations that confer multiple drug resistance. /. Mol. Biol. 223, 911-928. Savvidou, G. , Kle in , M . , Grey, A . A . , Dorrington, K . J. and Carver, J. P. (1984) Possible role for peptide-oligosaccharide processing at asparagine-107 of the light chain and asparagine-297 of the heavy chain in a monoclonal I g G l kappa. Biochemistry 23, 3736-3740. Schachter,H. (1986) Biosynthetic controls that determine the branching and microheterogeneity of protein-bound oligosaccharides. Biochem. Cell Biol., 64, 163-181. Schachter, H . (1991) Enzymes associated wi th glycosylation.Currenf Opinion in Structural Biology, 1, 755-765. Schachter,H. (1991) The 'yellow brick road' to branched complex N-glycans. Glycobiology, 1, 453-461. Schachter, H . (1994) Molecular cloning of glycosyltransferase genes. In Fukuda, M . and Hindsgaul, O. (eds.), Molecular Glycobiology, Oxford University Press, Oxford, U K , 88-162. Schachter, H . , Narasimhan, S., Gleeson, P., and Vella, G . (1983) Control of branching during the biosynthesis of asparagine-linked oligosaccharides. Can. J. Biochem. Cell Biol. 61, 1049-1066. Schwenk, F., Baron, U . and Rajewsky, K . (1995) A cre-transgenic mouse strain for the ubiquitous deletion of loxP flanked gene segments including deletion in germ cells. Nucleic Acids Research, 23, 5080-5081. 124 Senecoff, J. F., and Cox, M . M . (1986) Directionality in F L P protein-promoted site-specific recombination is mediated by D N A - D N A pairing. /. Biol. Chem. 261, 7380-7386. Smith,A.J., De Sousa,M.A., Kwabi-Addo,B. , Heppell-Parton,A., Impey,H. and Rabbitts,P. (1995) A site-directed chromosomal translocation induced in embryonic stem cells by Cre-loxP recombination. Nature Genetics, 4, 376-385. Spiro, M . J., Spiro, R .G. and Bhoyroo, V . D . (1979) Glycosylation of proteins by oligosaccharide-lipids. Studies on a thyroid enzyme involved in oligosaccharide transfer and the role of glucose in this reaction. /. Biol. Chem. 254, 7668-7674. Springer, T. A . (1994) Traffic signals for lymphocyte recirculation and leucocyte emigration: the multistep paradigm. Cell 76, 301-314. Srivastava, O.P. et al. (1988) Carbohydrate Res., 179, 137-161. Staneloni, R. et al. (1981) Plant Physiol. 68, 1175-1179. Stanley, P. (1984) Glycosylation mutants of animal cells. Ann. Rev. Genet. 18, 525-552. Stanley, P. (1989) Chinese hamster ovary cell mutants wi th multiple glycosylation defects for production of glycoproteins wi th minimal carbohydrate heterogeneity. Mol. Cell. Biol. 9, 377-383. Stanley,P. (1992) Glycosylation engineering. Glycobiology, 2, 99-107. Stanley, P., Sundaram, S., and Sallustio, S. (1991) A subclass of cell surface carbohydrates revealed by a C H O mutant wi th two glycosylation mutations. Glycobiology 1, 307-314. Stanley,P., Narasimhan,S., Siminovitch,L. and Schachter,H. (1975) Chinese hamster ovary cells selected for resistance to the cytotoxicity of phytohemagglutinin are deficient in a UDP-N-acetylglucosamine— glycoprotein N-acetylglucosaminyltransferase activity. Proc. Natl. Acad. Sci. U. S. A., 72, 3323-3327. Sternberg, N . A n d Hoess, R. (1983) The molecular genetics of the bacteriophage P I . Ann. Rev. Genet., 17, 123-154. Sternberg,N. and Hamilton,D. (1981) Bacteriophage PI site-specific recombination. I. Recombination between loxP and bacterial chromosome. /. Mol. Biol, 150, 487-507. Sternberg,N. and Hamilton,D. (1981) Bacteriophage PI site-specific recombination. I. Recombination between loxP sites. /. Mol. Biol, 150, 467-486. 125 S u r a n L M . A . H . (1979) Glycoprotein synthesis and inhibition of glycosylation by tunicamcyin in preimplantation mouse embryos: compaction and trophoblast adhesion. Cell, 18, 217-227. Tabas,I. and Kornfeld,S. (1978) The synthesis of complex-type oligosaccharides. III. Identification of an apha-D-mannosidase activity involved in a late stage of processing of complex-type oligosaccharides. /. Biol. Chem., 253, 7779-7786. Tabas et al. (1978) Processing of high mannose oligosaccharides to form complex type oligosaccharides on the newly synthesizedpolypeptides of the vesicular stomatitus virus G protein and the IgG heavy chain. /. Biol. Chem. 253, 716-722. Tabas, I. and Kornfeld, S. (1979) Purification and characterization of a rat liver Golg i alpha-mannosidase capable of processing asparagine-linked oligosaccharides. /. Biol. Chem. 254, 11655-11663. Takeuchi, M . , Inoue, N . , Strickland, T.W., Kubota, M . , Wada, M . , Shimizu, R., Hoshi , S., Kozutsumi, H . , Takasaki, S. and Kobata, A . (1989) Relationship between sugar chain structure and biological activity of recombinant human erythropoietin produced in Chinese hamster ovary cells. Proc. Natl. Acad. Sci. USA, 86, 7819-7822. Takeuchi, M . , Takasaki, S., Shimada, M . and Kobata, A . (1990) Roles of sugar chains in the in vitro biological activity of human erythropoietin produced in recombinant Chinese hamster ovary cells. /. Biol. Chem. 265, 12127-12130. Thomas, K . R. and Capecchi, M . R. (1987) Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51, 503-512. Towbin, H . , Staehelin, T., and Gordon, J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedures and some applications. Proc. Natl. Acad. Sci. USA 76, 4350-4354. Tsuda, E., Kawanishi, G. , Ueda, M . , Masuda, S. and Sasaki, R. (1990) The role of carbohydrate in recombinant human erythropoietin. Eur. J. Biochem. 188, 405-411. van DeursenJ., Fornerod,M., van Rees,B. and Grosveld,G. (1995) Cre-mediated site-specific translocation between nonhomologous mouse chromosomes. Proc. Natl. Acad. Sci. USA, 92, 7376-7380. Van den Eijden, D . N . , Koenderman, A . H . L . and Schiphorst, W. E. C. M . (1988) Partial purification and properties of an UDP-GlcNAc:N-acetyllactosaminide 131,3 N-acetylglucosaminyltransferase from Novikoff tumour cell ascites f luid. /. Biol. Chem. 263, 12461-12647. 126 Vark i , A . (1992) Role of oligosaccharides in intracellular trafficking of mammalian glycoproteins. In Cell surface carbohydrates and cell development. C R C Press, Boca Raton, Florida, 25-70. Vark i ,A . (1993) Biological roles of oligosaccharides: all of the theories are correct. Glycobiology, 3, 97-130. Vark i ,A . and Mar thJ .D. (1995) Oligosaccharides in vertebrate development. Seminars in Developmental Biology, 6, 127-138. Wagner, E. (1990) O n transferring genes into stem cells and mice. EMBO f. 9, 3025-3032. Wasley, L . C , Timony, G. , Murtha, P., Stoudemire, J., Domer, A . J., Caro, J., Krieger, M . and Kaufman, R. J. (1991) The importance of N- and O-linked oligosaccharides for the biosynthesis and in vitro and in vivo biologic activities of erythropoietin. Blood 77, 2624-2632. Wierzbicki , A . , Kendall , M . , Abremski, K . and Hoess, R. (1987) A mutational analysis of the bacteriophage PI recombinase Cre. /. Mol. Biol., 195, 785-794. Wiles, M . V . (1993) Embryonic stem cell differrentiation in vitro. Methods in Enzymology, 225, 900-918. Yamaguchi, K . , A k a i , K. , Kawanishi, G. , Ueda, M . , and Masuda, S. and Sasaki, R. (1991) Effects of site-directed removal of N-glycosylation sites in human erythropoietin on its production and biological properties. /. Biol. Chem. 266, 20434-20439. Yamashita, K. , Ohkura, T., Tachibana, Y. , Takasaki, S. and Kobata, A . (1984) Comparative study of the oligosaccharides released from baby hamster kidney cells and their polyoma transformant by hydrazinolysis. /. Biol. Chem. 259, 10834-10840. Yamashita, K . , Tachibana, Y , Ohkura, T. and Kobata, A . (1985) Enzymatic basis for the structural changes of asparagine-linked sugar chains of membrane glycoproteins of baby hamster kidney cells induced by polyoma transformation. /. Biol. Chem. 260, 3963-3969. Yamashita, K . , Totani, K . , Ohkura, T., Takasaki, S., Goldstein, J., and Kobata, A . (1987) Carbohydrate binding properties of complex-type oliogsaccharides on immobi l ized Datura stramonium lectin. /. Biol. Chem. 262, 1602-1607. Yoshimura, M . , Nishikawa, A . , Ihara, Y . , Taniguchi, S., and Taniguchi, N . (1995a) Suppression of lung metastasis of B16 mouse melanoma by N-acetylglucosaminyl-transferase III gene transfection. Proc. Natl. Acad. Sci. USA 92, 8754-8758. 127 Yoshimura, M v Ihara, Y. , and Taniguchi, N . (1995b) Changes of (3-1,4-N-acetylglucosaminyltransferase III (GnT-III) in patients wi th leukemia. Glycoconjugate }. 12, 234-240. Yoshimura, M . , Nishikawa, A . , Ihara, Y. , Nishiura, T., Nakao, H . , Kanayama, Y . , Matuzawa, Y. , and Taniguchi, N . (1995c) H i g h expression of U D P - N -acetylglycosamine: (3-D mannoside p-l,4-N-acetylglucosaminyltransferase III (GnT-III) in chronic myelogenous leukemia in blast crisis. Int. }. Cancer 60, 443-449. Yoshimura, M . , Ihara, Y. , Ohnishi, A . , Ijuhin, N . , Nishiura, T., Kanakura, Y . , Matsuzawa, Y . , and Taniguchi, N . (1996) Bisecting N-acetylglucosamine on K562 cells suppresses natural killer cytotoxicity and promotes spleen colonization. Cancer Res. 56, 412-418. Yousefi, S., Higgins, E., Daoling, Z. , Pollex-Kruger, A . , Hindsgaul, O., and Dennis, J. (1991) Increased U D P - G l c N A c : G a l (3l-3GalNAc-R (GlcNAc to Ga lNAc) p- l ,6 -N-acetylglucosaminyltransferase activity in metastatic murine tumour cell lines. /. Biol. Chem. 266, 1772-1782. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items