The Regulation of Lunatic Fringe During Somitogenesis

Home , HES7 gene, LFNG, MESP2, PAX1, PAX9

THE REGULATION OF LUNATIC FRINGE DURING SOMITOGENESIS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Emily T. Shifley

*****

The Ohio State University 2009

Dissertation Committee: Approved by Professor Susan Cole, Advisor

Professor Christine Beattie ______Professor Mark Seeger Advisor Graduate Program in Molecular Genetics Professor Michael Weinstein

ABSTRACT

Somitogenesis is the morphological hallmark of vertebrate segmentation.

Somites bud from the presomitic mesoderm (PSM) in a sequential, periodic fashion and

give rise to the rib cage, vertebrae, and dermis and muscles of the back. The regulation

of somitogenesis is complex. In the posterior region of the PSM, a segmentation clock

operates to organize cohorts of cells into presomites, while in the anterior region of the

PSM the presomites are patterned into rostral and caudal compartments (R/C patterning).

Both of these stages of somitogenesis are controlled, at least in part, by the Notch

pathway and Lunatic fringe (Lfng), a glycosyltransferase that modifies the Notch receptor. To dissect the roles played by Lfng during somitogenesis, we created a novel allele that lacks cyclic Lfng expression within the segmentation clock, but that maintains expression during R/C somite patterning (Lfng∆FCE1). Lfng∆FCE1/∆FCE1 mice have severe

defects in their anterior vertebrae and rib cages, but relatively normal sacral and tail

vertebrae, unlike Lfng knockouts. Segmentation clock function is differentially affected

by the ∆FCE1 deletion; during anterior somitogenesis the expression patterns of many clock genes are disrupted, while during posterior somitogenesis, certain clock components have recovered. R/C patterning occurs relatively normally in Lfng∆FCE1/∆FCE1 embryos, likely contributing to the partial phenotype rescue, and confirming that Lfng

ii plays separate roles in the two regions of the PSM. These results reveal that the

oscillatory regulation of the Notch pathway plays an important role in the segmentation

clock during the development of the anterior skeleton.

As part of the segmentation clock, Lfng mRNA is periodically transcribed and

LFNG protein levels have also been observed to cycle. Lfng mRNA and LFNG protein

molecules must therefore have rapid turnover rates in the PSM. We hypothesize that

cyclic Lfng transcription is coupled with signals in the Lfng 3’UTR that confer a short

half-life on Lfng mRNA. To test this, we examined the expression patterns of transgenes

containing conserved sections of the Lfng 3’UTR and determined which sections

cooperatively confer a short RNA half-life in the PSM. LFNG protein acts in the Golgi,

but is also cleaved and released into the extracellular space. We hypothesized that this

cleavage/secretion contributes to short LFNG intracellular half-life, facilitating its rapid

oscillations. To test this, we localized N-terminal protein sequences that control the

secretory behavior of the fringe proteins and found that LFNG processing is promoted by

furin-like protein convertases. Mutations that alter LFNG processing increase its intracellular half-life in vitro without affecting the specificity of its function in the Notch pathway. To determine the importance of LFNG cleavage/secretion in vivo, we created a novel allele that tethers LFNG to the Golgi (LfngRL). LfngRL/+ mice show significant

segmentation and patterning defects suggesting that the short intracellular half-life of

LFNG is important for somitogenesis. Thus, the cyclic activity of Lfng in the

segmentation clock is achieved through multiple mechanisms, including tight regulation

of mRNA and protein levels. Overall, we find that Lfng plays an important role in spatially and temporally regulating Notch signaling during vertebrate segmentation.

iii

DEDICATION

This work is dedicated to my parents, to my siblings, and to my husband for all of their love and support.

ACKNOWLEDGMENTS

I would like to acknowledge my advisor, Dr. Susan Cole, for all of her guidance

and encouragement during my graduate studies. I am grateful for the opportunity to have

been a part of her research and truly appreciate all of the mentoring and training she has

given me. I could not have asked for a more supportive advisor.

I would also like to thank my committee members, Dr. Beattie, Dr. Chamberlin,

Dr. Seeger, and Dr. Weinstein for their time and guidance during my graduate studies. I

am especially grateful to Dr. Weinstein for helping with knock-in mice.

I would like to acknowledge current and previous members of the lab for their contributions to this work and their support. Kellie contributed some of the data included

in this thesis, as indicated, and was a wonderful co-worker and friend. To the other Cole

lab graduate students, Ariadna, Maurisa, and Dustin, thank you for making the lab a

supportive and fun place to work, I wish you all the best. To Dawn and Jorge, thank you

for your contributions to the GFP transgenic mice and all of the work you do in the lab.

To all of the undergraduates who have come through the Cole lab, thank you for your

enthusiasm.

v I would like to thank my classmates for their camaraderie as we have worked through graduate school together. I have truly enjoyed our friendship and time spent together.

I also wish to thank the editors of Birth Defects Research (Part C), Development, and Biochimica et Biophysica Acta: Molecular Cell Research for the use of the previously published text and figures in this dissertation.

VITA

September 29, 1981………………..….Born –Cincinnati, OH.

2003……………………………………Bachelor of Science. Ohio University.

2003 – 2004………………………...….University Fellow The Ohio State University.

2004 – 2009…………………………....Research and Teaching Assistant The Ohio State University.

PUBLICATIONS

1. E.T. Shifley and S.E. Cole. (2008) Lunatic fringe protein processing by proprotein convertases may contribute to the short protein half-life in the segmentation clock. Biochim Biophys Acta, Mol Cell Res., 1783.2384-2390.

2. E.T. Shifley, K.M. VanHorn, A. Pérez-Balaguer, J.D. Franklin, M. Weinstein, and S.E. Cole. (2008). Oscillatory Lunatic Fringe activity is crucial for segmentation of anterior but not posterior skeleton. Development, 135.899-908.

3. E.T. Shifley and S.E. Cole. (2007). The vertebrate segmentation clock and its role in skeletal birth defects. Birth Defects Part C, 81.121-33.

FIELDS OF STUDY

Major Field: Molecular Genetics

vii

TABLE OF CONTENTS

Page

Abstract……………………………………………………………………………... ii

Dedication………………………………………………………………………...... iv

Acknowledgements………………………………………………………………… v

Vita…………………………………………………………………………………. vii

List of Tables……………………………………………………………………….. xii

List of Figures……………………………………………………………………… xiii

List of Abbreviations……………………………………………………………….. xvi

Chapters:

1. Introduction…………………………………………………………………... 1 1.1 Introduction…………………………………………………………….… 1 1.2 Somitogenesis………………………………………………………….… 2 1.3 The Segmentation Clock…………………………………………………. 5 1.3.1 Cycling genes……………………………………………………… 6 1.3.2 Segmentation Clock Variations…………………………………… 9 1.3.3 Recent Clock Models……………………………………………… 10 1.4 The Wavefront…………………………………………………………… 12 1.5 The Notch Signaling Pathway…………………………………………… 13 1.5.1 Notch signaling plays multiple roles during somitogenesis………. 15 1.5.2 Notch pathway knockouts…………………………………………. 16 1.6 Human Disease…………………………………………………………... 18 1.6.1 Spondylocostal Dysostosis………………………………………... 18 1.7 The regulation of Lfng during somitogenesis……………………………. 23

viii 2. Lunatic Fringe is important for segmentation clock function during primary not secondary body formation………...……………………………………... 32 2.1 Introduction………………………………………………………………. 32 2.2 Materials and Methods………………………………………………….... 34 2.2.1 Targeted deletion of FCE1………………………………………… 34 2.2.2 Genotyping……………………………………………………….... 34 2.2.3 Whole mount in situ hybridization……………………………….... 35 2.2.4 Skeletal preparations………………………………………………. 36 2.2.5 Whole mount immunohistochemistry……………………………... 36 2.3 Results……………………………………………………………………. 36 2.3.1 Deletion of FCE1 from the Lfng locus perturbs clock-linked Lfng RNA expression……………………………………………………….. 36 2.3.2 The loss of Lfng expression in the segmentation clock perturbs normal skeletal development…………………………………………….. 37 2.3.3 Rostral-caudal somite patterning is partially rescued in Lfng∆FCE1/∆FCE1 embryos………………………………………………….. 39 2.3.4 Somites differentiate properly in Lfng∆FCE1/∆FCE1 embryos………... 41 2.3.5 The loss of cyclic Lfng expression in the posterior PSM perturbs oscillatory NOTCH1 activity……………………………………………. 41 2.3.6 Expression of oscillatory genes is differentially affected during primary and secondary body formation in Lfng∆FCE1/∆FCE1 embryos…….. 42 2.3.7 Wnt targets continue to cycle, and Brachyury expression is not disrupted in Lfng∆FCE1/∆FCE1 embryos……………………………………. 44 2.4 Discussion………………………………………………………………... 44 2.4.1 The FCE1 enhancer is necessary for cyclic expression of Lfng in Region I of the PSM………...…………………………………………... 44 2.4.2 Oscillatory Lfng expression and Notch signaling are critical for the proper segmentation during primary, but not secondary, body formation………………………………………………………………… 45 2.4.3 Lfng plays separable roles in the segmentation clock and R/C patterning………………………………………………………………... 47 2.4.4 Differential segmentation clock regulation at distinct levels of the axial skeleton?…………………………………………………………... 48 2.4.5 R/C patterning of anterior somites may affect ongoing segmentation during secondary body formation………………………… 51

3. The lunatic fringe 3’UTR contributes to short RNA half-life in the Presomitic mesoderm………………………………………………………… 66 3.1 Introduction………………………………………………………………. 66 3.2 Materials and Methods…………..………………………………………. 69 3.2.1 Transgene Construction…………………………………………… 69 3.2.2 Luciferase Reporter Assay………………………………………… 70 3.2.3 Whole Mount in situ Analysis…………………………………….. 70 3.3 Results……………………………………………………………………. 70 3.3.1 Localizing Lfng 3’UTR signals affecting RNA stability…………. 70

ix 3.3.2 Sections of the Lfng 3’UTR have differential effects on RNA stability in the mouse PSM………….…………………………………... 71 3.3.2.1 Fragment E has a partial effect on RNA stability………….. 72 3.3.2.2 Sequences in Fragment B+D confer an Lfng-like RNA half- life on GFP…………………………………………………………. 73 3.3.2.3 Fragment B appears to have an effect on RNA stability in the PSM…………………………………………………………….. 74 3.4 Discussion………………………………………………………………... 74

4. Lunatic fringe protein processing by proprotein converstases contributes to the short protein half-life in the segmentation clock…………………………. 85 4.1 Introduction………………………………………………………………. 85 4.2 Materials and Methods………………………………………………….... 87 4.2.1 LFNG mutants……………………………………………………... 87 4.2.2 Alkaline phosphatase assays………………………………………. 88 4.2.3 Immunoflourescence………………………………………………. 88 4.2.4 Notch1 signaling assay…………………………………………….. 89 4.2.5 Western blot analysis……………………………………………… 89 4.2.6 Cycloheximide treatment………………………………………….. 90 4.2.7 Whole Mount in situ hybridization………………………………... 91 4.2.8 Targeted replacement of LFNG pre/pro region…………………… 91 4.2.9 Genotyping……………………………………………………….... 92 4.2.10 Skeletal preparations……...……………………………………… 92 4.3 Results……………………………………………………………………. 92 4.3.1 SPC1/furin, SPC6A and SPC6B promote LFNG processing……... 92 4.3.2 N-terminal sequences regulate the secretory behavior of fringe family proteins…………………………………………………………... 93 4.3.3 SPC family proteases recognize two dibasic cleavage sites in LFNG, but protein processing is not required for secretion……………... 94 4.3.4 Mutation of SPC processing sites increases the intracellular half- life of the LFNG protein…………………...……………………………. 96 4.3.5 Alterations of LFNG intracellular half-life do not affect the specificity of its function in the Notch signaling pathway………………. 96 4.3.6 The expression pattern of Spc6 suggests a role in clearance of LFNG from maturing somites………………………………………….... 97 4.3.7 LfngRL/+ mice show significant segmentation defects……...……… 98 4.3.8 Somitogenesis is highly disrupted in LfngRL/+ embryos…………… 100 4.4 Discussion………………………………………………………………... 100 4.4.1 LFNG processing by furin/SPC proteases is not necessary for its secretion or modification of Notch receptors……………………………. 101 4.4.2 LFNG intracellular half-life is critical for proper somitogenesis….. 102 4.4.3 SPC/furin cleavage of LFNG may be important for multiple aspects of somitogenesis...………………………………………………. 104 4.4.4 An extracellular role for LFNG?…………………………………... 106

5. Conclusion……………………………………………………………………. 124 5.1 Overview…………………………………………………………………. 124 5.2 Lfng is important for segmentation clock function during primary, but not secondary body formation………..………………………………………. 125 5.3 The Lfng 3’UTR contributes to short RNA half-life in the PSM…….…... 126 5.4 LFNG protein processing by proprotein convertases contributes to the short protein half-life in the segmentation clock……………………………... 127

List of References…………………………………………………………………… 129

LIST OF TABLES

Table Page

4.1 Primer sequences for mutagenesis of LFNG………....…………………. 107

xii

LIST OF FIGURES

Figure Page

1.1 Mouse Somitogenesis………………………………...…………………. 26

1.2 The Clock and Wavefront model……………………………………….. 28

1.3 Schematic of a clock gene feedback loop………………………………. 29

1.4 Overview of the Notch Signaling pathway……………………………... 30

1.5 A schematic of Notch signaling in Regions I and II of the PSM……….. 31

2.1 Deletion of the FCE1 enhancer…………………………………………. 53

2.2 Deletion of the FCE1 enhancer from the endogenous locus alters Lfng expression in the posterior PSM………………………………………... 54

2.3 Lfng expression is localized to the proper somite compartment in Lfng∆FCE1/∆FCE1 embryos…………………………………………………. 55

2.4 The Lfng ∆FCE1 allele interferes with normal skeletal development during primary body formation…………………………………………. 56

2.5 R/C patterning in Lfng∆FCE1/∆FCE1 embryos…………………………….. 57

2.6 Somites differentiate properly in Lfng∆FCE1/∆FCE1 embryos……………... 59

2.7 Notch1 signaling is altered in Lfng∆FCE1/∆FCE1 embryos………………… 60

2.8 Hes7 transcription is affected in Lfng∆FCE1/∆FCE1 embryos……………… 61

2.9 Oscillatory gene expression is differentially perturbed in Lfng∆FCE1/∆FCE1 embryos…………………………………………………………………. 63

xiii 2.10 Hes expression is disrupted in Lfng∆FCE1/∆FCE1 embryos………………... 64

2.11 The Lfng∆FCE1 allele does not prevent cyclic Wnt in the PSM or mesoderm induction in the tail bud……………………………………... 65

3.1 Lfng 3’UTR is necessary for RNA turn-over in the PSM………………. 78

3.2 Conservation of Lfng 3’UTR……………………………………………. 79

3.3 Conserved regions of Lfng 3’UTR contribute to RNA turn-over in vitro. 80

3.4 Sections of Lfng 3’UTR are sufficient to reveal cyclic RNA expression in the PSM………………………………………………………………. 81

3.5 Different sections of the Lfng 3’UTR are differentially able to promote an Lfng-like RNA expression pattern in the PSM………………………. 83

3.6 Fragment B+D reveals a cyclic expression pattern similar to Lfng…….. 84

4.1 Alignment of pre/pro region of fringe genes……………………………. 108

4.2 LFNG is processed by SPC proconvertases…………………………….. 109

4.3 The N-terminus of fringe proteins controls their secretory behavior…… 110

4.4 LFNG protein is cleaved at two sites, but neither cleavage is necessary for secretion……………………………………………………………. 111

4.5 LFNG variant proteins localize to the Golgi…………………………… 113

4.6 Mutations affecting LFNG processing result in an increased intracellular protein half-life…………………………………………………………. 115

4.7 LFNG variant proteins modify Notch signaling with an LFNG-like activity……………………………………………………………………. 116

4.8 SPC6 expression patterns suggest a role in clearing LFNG from maturing somites…………………………………………………………. 117

4.9 Knock-in of the R/LFNG allele………………………………………….. 118

4.10 LfngRL/+ mice have shortened tails and bodies…………………………… 119

4.11 Skeletal derivatives of somites are highly disorganized in LfngRL/+ mice.. 120

xiv 4.12 LfngRL/+ embryos are not forming epithelial somites properly…………… 122

4.13 Somite patterning is disrupted in LfngRL/+ embryos……………………… 123

LIST OF ABBREVIATIONS

AVS Abnormal Vertebral Segmentation

CSL CBF1, Suppressor of Hairless, LAG1 protein

dGFP Destabilized GFP

Dll1 Deltalike1

Dll3 Deltalike3 dpc days postcoitus

DSL Delta/Serrate/LAG-2 Notch ligands

EGF Epidermal Growth Factor

FCE1 Fringe Clock Element 1

FGF Fibroblast Growth Factor

Hes Hairy/enhancer of split

Lfng Lunatic Fringe gene or mRNA

LFNG Lunatic Fringe protein

MAML Mastermind like

NICD Notch intracellular domain

PSM Presomitic Mesoderm

R/C Rostral/Caudal

SCDO Spondylocostal Dysostosis

xvi UTR Untranslated Region

xvii

CHAPTER 11

INTRODUCTION

1.1 Introduction

The development of complex, multicellular organisms is a carefully regulated

process. There are an amazing number of mechanisms that coordinate development as a

single cell gives rise to a diverse array of tissue and organ systems. A surprisingly small

number of genetic signaling pathways control much of embryonic development, being

used multiple times in various contexts. It is important to study these processes because

when genes that help control development are disrupted, birth defects can arise.

Additionally, cancer cells have often inappropriately activated developmental signaling

pathways that aid in their growth. This study focuses on the role of a gene called Lunatic

Fringe (Lfng), formally known as O-fucosylpeptide 3-beta-N-

acetylglucosaminyltransferase, in regulating the Notch signaling pathway. Lfng acts to

spatially and temporally regulate Notch signaling during the developmental process of

somitogenesis. I find that tight regulation of Lfng activity is key for proper Notch signaling during vertebrate development.

1 Portions of this chapter were originally published in Birth Defects Research Part C, Volume 81, Pages 121-33. 1 1.2 Somitogenesis

The process of somitogenesis has been the subject of several excellent reviews

(see, for example, Christ et al., 1998; Gossler and Hrabe de Angelis, 1998). Somites are transient embryonic structures that give rise to the most obvious examples of segmentation in vertebrates: the vertebrae and intervertebral discs of the backbone, the rib cage, and the dermis and striated muscle of the back. Somites bud from the overtly unsegmented paraxial mesoderm that flanks the neural tube at the caudal end of an

embryo called the presomitic mesoderm (PSM) (or the segmental plate in chick embryos)

(Figure 1.1 A). During somitogenesis, pairs of somites bud from the anterior region of

the PSM at a regular, species-specific rate while mesodermal tissue is replenished in the

caudal tip of the PSM. Early in development, cells are recruited to the PSM via the primitive streak, but later, around 9.5 days postcoitus (dpc.) in the mouse, a structure called the tailbud forms and all subsequent PSM cells arise from this structure (Tam,

1981). The PSM experiences a constant turnover as groups of cells move up together to form somites and new cells are added at the caudal end. This process is highly regulated, but changes somewhat during development. Early in development mouse somites form every hour, while later in development the rate of somite formation slows to once every

2–3 hr. The overall size of the PSM varies over two-fold through development, expanding from 0.4 mm early in development to 1.5 mm at the 20–30 somite stage and then decreasing in size to 0.5 mm later in development (Tam, 1981). Despite these changes in size, explant analysis demonstrates that throughout development, the mouse

PSM contains five to six presumptive somites at any given time during somitogenesis: the size of somites produced simply varies along with the size of the PSM (Tam, 1986).

2 The process of somitogenesis is extraordinarily robust. Dissected PSM fragments continue to segment in culture without any of the tissues that normally surround the PSM present (see, for example, (Packard, 1976)), and only the fibronectin network of the ectoderm is necessary to form morphological somites from isolated PSM (Rifes et al.,

2007). Additionally, the segmentation of the PSM is regulated at the whole tissue level.

For instance, when fragments of the PSM are dissected out of a chick embryo and then transplanted back in an inverted orientation, somitogenesis continues in the rostral to caudal orientation of the original PSM (Christ et al., 1974; Menkes, 1977).

Somitogenesis is therefore an autonomous process within the PSM; once cells enter this tissue they take part in a highly regulated and patterned process.

The process of somitogenesis is quite complex: it requires several sequential

events that appear to occur in different regions of the PSM (defined in Saga and Takeda,

2001) (Figure 1.1 B). In the posterior PSM (Region I) a cohort of cells is coordinated

and organized into a presomite. This process is regulated, at least in part, by the

segmentation clock that functions within Region I. Groups of mesenchymal cells move

up the PSM together, experience repetitive pulses of signaling from the segmentation

clock, and eventually form morphologically identifiable structures called somitomeres in

the anterior part of the PSM. The borders formed between these somitomeres are

maintained as the borders between adjacent somites once they mature (Tam, 1986).

Within the anterior region of the PSM (Region II), the presomites are patterned to

produce a reiterated pattern of anterior and posterior compartments. Differential gene

expression defines the anterior and posterior compartments of the developing presomite,

patterning the cells in these compartments even before the formation of morphologically

3 distinct epithelial somites. Finally, intersomitic boundaries are formed between adjacent

somites when the somitomeres undergo a mesenchymal to epithelial transition to form

mature, epithelial somites.

Differentiation continues after somite formation. A group of cells from the

mature somite de-epithelialize and become mesenchymal. The mesenchymal sclerotome

gives rise to the vertebral column: groups of mesenchymal cells migrate ventromedially

to form vertebral bodies and intervertebral discs, dorsomedially to form pedicles of the

neural arches, and ventrolaterally in the thoracic region to form the ribs. The remaining

epithelial cells of the somite form the dermomyotome, which gives rise both to the

myoblasts and fibroblasts of the dorsal dermis. Differentiated portions of the myotome

give rise to epaxial back muscles, hypaxial muscles of the body wall, and skeletal

muscles of the limbs. Interestingly, a resegmentation event occurs in the differentiating

somite as well. The caudal sclerotome of one somite and the rostral sclerotome of the

adjacent somite form a single vertebral body; thus the borders between vertebrae do not

correspond to the original borders between somites (Figure 1.1 C) (reviewed in Christ et

al., 1998). This resegmentation event allows a single muscle group (arising from one

somite) to interact with two vertebrae, facilitating intervertebral motion (Huang et al.,

1996; Huang et al., 2000). Thus, the correct development and patterning of somites is

essential for the proper formation of skeletal and muscular derivates. Somite patterning

also influences the symmetric branching of blood vessels between somites and the

migration of peripheral nervous system motor neuron axons and neural crest cells that cross only at the anterior halves of somites, and later form ganglia (Keynes and Stern,

4 1984; Rickmann et al., 1985). Interference with any aspect of somitogenesis can perturb

the final structure of the axial skeleton and the associated segmental structures.

1.3 The Segmentation Clock

Somites arise from the PSM in a controlled, periodic fashion at a rate that is

species specific, but which may vary over the course of development. In mouse embryos,

one pair of somites is formed every two hours during much of development; however, as

discussed above, this rate is not constant throughout somitogenesis (Tam, 1981). While

the somites are being formed in this periodic fashion they are also being formed in strict

formed between landmarks such as the limb buds is kept constant within a species. The

regular, periodic formation of somites suggested the possibility of a timing mechanism

that could be involved in regulating somitogenesis, although this mechanism must be

flexible enough to allow the rate of somite formation to evolve over the course of

somitogenesis. Several experiments performed in the 1970’s supported this idea. For

instance, in Xenopus, surgical manipulation was used to reduce the size of the early embryo by up to 60%. Despite the reduction in the pool of mesodermal cells, the correct

number of somites was formed, but somite sizes were reduced proportionally (Cooke,

1975). Similar results were found with the Amputated mouse mutant, which forms the

correct number of somites in the trunk region although the embryo length is reduced by

50% (Flint et al., 1978). These experiments and others led to the prediction that

somitogenesis might be regulated via a clock mechanism.

The clock and wavefront model was the first description of a hypothetical clock

that might regulate somitogenesis. Cooke and Zeeman proposed that a segmentation

5 clock operates in the PSM underlying the inherent periodicity of somite formation

(Cooke and Zeeman, 1976). This model proposes the existence of an oscillatory activity

in the PSM, with a period matching the rate of somite formation. Neighboring cells

oscillate in synchrony, and the period of these cycles governs and controls periodic

somite formation. In order to explain the periodic pattern of somites, this model also

invokes a wavefront that moves down the PSM in an anterior to posterior direction,

providing position and size information as somite formation progresses. The interaction

of the segmentation clock with the wavefront acts to produce the correct number of

properly sized somites during the course of somitogenesis (Cooke and Zeeman, 1976)

(Figure 1.2). Several other models have been proposed suggesting that the timing of segmentation is regulated by a clock (see, for example, (Kerszberg and Wolpert, 2000;

Meinhardt, 1986; Schnell and Maini, 2000). Although these models differ in their specifics, they all predict the existence of a clock activity in the PSM that oscillates between two states, with a periodicity that matches the rate of somite formation.

Molecular evidence supporting the existence of the segmentation clock has since been found in a number of species.

1.3.1 Cycling genes

The first molecular evidence for the existence of the segmentation clock was described in the expression patterns of c-hairy1 mRNA in chick embryos (Palmeirim et al., 1997). c-hairy1 encodes a transcription factor homologous to the Drosophila

Hairy/enhancer of split (Hes) genes and is a downstream target of the Notch signaling pathway. In stage-matched embryos, several different patterns of c-hairy1 expression are seen. These dynamic expression patterns actually reflect cyclic expression of the mRNA,

6 and the period of the c-hairy1 mRNA oscillations matches the period of somite formation

in the embryo (Palmeirim et al., 1997). The expression of c-hairy1 then stabilizes in the

anterior PSM (Region II) where it is expressed only in the caudal compartment of the

epithelial somites, suggesting it is also involved in the R/C patterning of somites.

Although c-hairy1 expression is visualized as a ‘‘wave’’ traversing the PSM from posterior to anterior, PSM resection experiments demonstrate that the expression pattern is not propagated from cell to cell and there is no communication required between different regions of the PSM. Instead, the wave-like expression pattern is due to the fact that cells across the PSM oscillate slightly out of phase. Thus, the expression pattern described for c-hairy1 closely matched the predicted expression pattern of a clock-linked

gene. Shortly after the cyclic expression pattern of c-hairy1 was described in the PSM,

another Notch pathway member, Lunatic fringe (Lfng), was also found to oscillate in the

PSM of the chick and mouse (Aulehla and Johnson, 1999; Forsberg et al., 1998; Jiang et

al., 2000; McGrew et al., 1998). Similar to the pattern seen in c-hairy1, Lfng mRNA

levels oscillate with a periodicity matching the rate of somite formation. The oscillatory

expression of Lfng is controlled at the transcriptional level (Cole et al., 2002; Morales et

al., 2002), and both Lfng mRNA and protein levels cycle in the PSM (Dale et al., 2003).

In the years since these discoveries, many Notch pathway genes have been shown to

cycle in the PSM, including Hes1, Hes7, Hey1-3, and Nrarp in the mouse (Bessho et al.,

2001a; Dequeant et al., 2006; Jouve et al., 2000; Leimeister et al., 2000; Leimeister et al.,

1999); c-hairy2 and cHey2 in the chick (Jouve et al., 2000; Leimeister et al., 2000); her1,

her7, and deltaC in zebrafish (Holley et al., 2000; Jiang et al., 2000; Oates and Ho, 2002;

Sawada et al., 2000); her7 and hey1 in medaka (Elmasri et al., 2004); and esr9 and esr10

7 in Xenopus (Li et al., 2003). More recently, immunohistochemistry against the activated

Notch intracellular domain (NICD) has been used to demonstrate that although Notch1

mRNA and protein are expressed throughout the PSM, Notch activity levels oscillate in

Region I (Huppert et al., 2005; Morimoto et al., 2005). Thus, as cells move up the PSM, they experience repetitive pulses of oscillatory mRNAs and Notch activity. These data suggest that at least some aspects of the segmentation clock may be conserved among mammals, birds, amphibians, and fish, and that the Notch pathway plays an important role in the segmentation clock.

Gene expression also links the Wnt signaling pathway to the segmentation clock.

Axin2 mRNA levels oscillate in the PSM with an identical periodicity to genes in the

Notch pathway, but in the reverse phase, such that when Lfng RNA levels are low, Axin2

RNA levels are high (Aulehla et al., 2003). Similar links to the Wnt pathway were found in the expression of Dact1, encoding a Dishevelled-binding regulator that oscillates in phase with Axin2 (Suriben et al., 2006). Interestingly, the Wnt antagonist Nkd1 was shown to oscillate in phase with Notch pathway genes, and to be transcriptionally regulated via HES7, suggesting that this gene may represent a point of communication between the Wnt and Notch pathway (Ishikawa et al., 2004). More recently, Dequeant et al. utilized microarrays to identify over 30 new genes that are cyclically expressed in the

PSM (Dequeant et al., 2006). Many of these oscillating genes belong to the Notch pathway, the Wnt pathway, or the FGF pathway. Similar to the results discussed above, oscillatory genes could be generally grouped into two expression patterns, with most genes associated with the Wnt pathway oscillating out of phase with genes in the Notch and FGF pathways. The authors predict the existence of between 50 and 100 oscillatory

8 genes, with many belonging to these three traditional growth factor/ receptor signaling

pathways. Furthermore, they propose that interaction among these pathways is important

in the mechanism of the segmentation clock and may contribute to the robust nature of

somitogenesis. Many potential areas of cross-talk are currently being identified, for

instance, the FGF inhibitor Dusp4 was found to be regulated by the Notch target Hes7

(Niwa et al., 2007), the Notch target Nrarp can both inhibit Notch and potentiate Wnt

signaling, and the Notch target Tbx6 coordinates with Notch and Wnt members to control

gene expression in the PSM (reviewed in Shifley and Cole, 2007).

1.3.2 Segmentation Clock Variations

In Xenopus embryos, somitogenesis is the process by which myotomal cells segregate out of the PSM together and rotate 90° before forming morphological somites.

Despite these differences from other vertebrates, the Notch signaling pathway appears to be used in Xenopus as well. It has been proposed that Notch signaling is first required in the Xenopus tailbud domain to establish segment size, and then plays a second role to maintain R/C identities in developing somitomeres through a combination of positive and negative feedback loops (Jen et al., 1999). Somitogenesis was recently analyzed in the cornsnake; it was found to have a very rapid segmentation clock which generates a large number of somites (Gomez et al., 2008). Although Axin2 was not observed to cycle in the snake PSM, Lfng was expressed as many bands of varying widths suggesting an oscillator is present, but may be regulated in a different way than other vertebrates. In zebrafish, the PSM displays cyclic Notch activity similar to mice and chick, although it is thought to be maintained by a negative feedback loop controlled by Her genes and the

Notch ligand DeltaC (reviewed in Mara and Holley, 2007). Neither Lfng nor members of

9 the Wnt or FGF pathway have been found to cycle in zebrafish. Additionally, recent experimental evidence suggests that the Notch pathway oscillations may in fact be required to synchronize neighboring cells in the zebrafish PSM rather than acting as a master oscillator (reviewed in Ozbudak and Lewis, 2008). Thus, although various model organisms show cyclic Notch activity in the PSM, different pathway members are cycling in different species and the pathway appears to be utilized for a variety of functions.

1.3.3 Recent Clock Models

Researchers have used mathematical modeling to predict how cyclic genes might sustain the cyclic activity of the segmentation clock. For instance, Goldbeter and

Pourquie modeled negative feedback loops for the Notch pathway based on Lfng inhibition of Notch receptors; for the Wnt pathway based on Axin2 inhibiting β-catenin; and for the FGF pathway based on Dusp6 inhibiting ERK or Ras (Goldbeter and

Pourquie, 2008). Each pathway can be modeled to maintain its own oscillations through negative feedback loops under conditions of a constant stimulus from the clock. Notch,

Wnt, and FGF could either oscillate independently provided their periods are similar enough, or there are a number of ways the pathways might interact to synchronize with one another, as discussed above, and these connections can also be incorporated into clock models. Momiji et al. preformed mathematical modeling specifically of Hes1 oscillations, which can be observed in tissue culture cells (Hirata et al., 2002) and are based on a negative feedback loop where HES1 inhibits its own transcription (Momiji and Monk, 2008). By incorporating a number of variables including transcriptional and translational time delays, transport of Hes1 mRNA, HES1 dimerization, Hes1 mRNA stability, and HES1 protein stability (which is influenced by levels of Stat3), researchers

10 were able to write a model that correlates well with experimental data (Figure 1.3). The

importance of incorporating each of these variables suggests that RNA and protein

turnover of cyclic clock genes would need to be tightly regulated during segmentation

clock function.

Another ongoing issue in segmentation clock research is identifying the master

regulator or pacemaker that sets the period of the clock. It is thought that the Notch

pathway is downstream of a master regulator because experimentally removing Notch

activity does not perturb the formation of the first several somites. The recent

identification of new signaling pathways cycling in the mouse PSM that are not

invariably perturbed in Notch mutants suggested that perhaps the Wnt or FGF pathway

could be the master regulator. However, in recent experiments where β-catenin is

stabilized in the PSM, cyclic Notch activation (measured by Lfng and Hes7 expression) continues, albeit more cycles of Notch expression are observed than normal because the maturation of the somites is delayed, expanding the PSM clock region (Aulehla et al.,

2008; Dunty et al., 2008). Additionally, the loss of cyclic Lfng observed in FGF mutants can be rescued by constitutive β-catenin, suggesting that neither the Notch, Wnt, or FGF pathways can be defined as the master regulators of the segmentation clock because disrupting any one pathway does not always prevent cyclic activity of members of another, supposedly downstream pathway (reviewed in Ozbudak and Lewis, 2008). This leaves open the possibility that all three pathways are able to independently oscillate, but are entrained to one another at specific points defining the segmentation clock, or that there is an as yet unidentified master pacemaker for somitogenesis.

11 1.4 The Wavefront

The segmentation clock is proposed to interact with positional information (the

‘‘wavefront’’) to specify segments in the anterior PSM. The nature of the molecules

specifying this positional and orientation information within the PSM has been a matter

of considerable interest. An FGF8 gradient transverses the PSM with the highest

concentration at the caudal end of the PSM where it is thought to keep the cells in the

PSM in an immature, undifferentiated state (Dubrulle et al., 2001; Dubrulle and Pourquie,

2004; Sawada et al., 2001). Manipulating the activity of FGF8 with inhibitors or beads

soaked in FGF8 and placed along the PSM alters the size of mature somites that bud off

the PSM, suggesting that the maintenance of the appropriate concentration of FGF plays

an important role in the maturation and formation of boundaries between somites

(Dubrulle et al., 2001). The FGF gradient is opposed and possibly maintained by a

retinoic acid gradient that is high in the anterior PSM (Diez del Corral et al., 2003;

Moreno and Kintner, 2004). Interestingly, recent studies also demonstrate that retinoic

acid signaling acts to coordinate somite formation across the left/right axis

(Vermot and Pourquie, 2005). Recently, a Wnt3a gradient has also been suggested as a

possible gradient in the PSM, perhaps working with the FGF gradient, linking the

wavefront with clock, and activating the Notch pathway to form borders between somites at the appropriate concentration of Wnt3a (Aulehla and Herrmann, 2004; Aulehla et al.,

2003). Indeed, nuclear β-catenin is observed as a gradient in the PSM and if β-catenin is

constitutively activated, the determination wavefront is shifted anterior suggesting

somites cannot differentiate under high levels of Wnt signaling (Aulehla et al., 2008;

Dunty et al., 2008). It is theorized that the gradients create a specific domain of

12 bistability that the segmentation clock interacts with in order to transition cells from

oscillations to differentiation (Goldbeter et al., 2007).

Thus, the early models predicting a segmentation clock interacting with a morphogen gradient operating in the PSM have been validated by experimental evidence, although the exact mechanisms of each of these components is still being elucidated in model systems. Furthermore, given the fact that both the FGF and Wnt pathways appear to be linked to both the clock and the gradient it interacts with, further complexity in the mechanisms controlling somitogenesis seem likely.

1.5 The Notch signaling pathway

The Notch genes encode single-pass transmembrane receptors with tandem epidermal growth factor (EGF) domains in the extracellular domain where Notch ligands bind. Notch and its ligands exist as gene families: in mice there are 4 Notch receptors and 5 DSL (Delta/Serrate/LAG-2) ligands. In the PSM, the most relevant pathway members appear to be Notch1, Deltalike1 (Dll1), and Deltalike-3 (Dll3), although

Jagged1 and Notch2 are also expressed in the PSM and may contribute to the process of somitogenesis.

The Notch receptor is modified in the Golgi before being presented to the cell surface. A furin-like protease cleaves Notch (S1 cleavage) creating a mature heterodimer of the intracellular/transmembrane domain and the extracellular domain. Additionally, certain EGF repeats in the extracellular domain may be modified in the Golgi by the addition of sugar moieties by O-fucosyl transferase and fringe proteins. After Notch is presented to the cell surface, it can bind to a DSL ligand expressed on the surface of a neighboring cell. A series of cleavage events is initiated after a Notch receptor

13 successfully binds to its ligand, including an S2 cleavage by ADAM metalloprotease and

an S3 cleavage by γ-secretase, a complex containing presenilin. Cleavage of the Notch receptor releases the intracellular domain (NICD) which then translocates into the nucleus and associates with the promoters of target genes, coactivating transcription with

CSL (CBF1, Suppressor of Hairless, LAG1 proteins, called RBPJk in mice) and

Mastermindlike (MAML) family members. NICD activates a variety of target genes including several Hes-related transcriptional repressors, many of which cycle in the PSM

(reviewed in Weinmaster and Kintner, 2003) (Figure 1.3).

Several Notch modulating genes are expressed in the PSM including Lfng, the

focus of this study. Lfng encodes a glycosyltransferase that modifies the extracellular

domain of the Notch receptor in the Golgi by adding N-acetylglucosamine residues to O-

linked fucose sugars on the EGF repeats of Notch (Dale et al., 2003). Depending on the

cellular context, which Notch receptor, which DSL ligand, and which fringe gene are

involved, these modifications can either enhance Notch signaling or repress Notch

signaling (Bruckner et al., 2000; Moloney et al., 2000; Panin et al., 2002). Both Lfng mRNA and protein levels cycle in the PSM during somitogenesis (Dale et al., 2003).

Therefore, similar to the clock model proposed for Hes1 oscillations (Momiji and Monk,

2008), we would expect that the half-life of Lfng mRNA and protein would play important roles in maintaining Lfng and Notch pathway oscillations, which need to be tightly regulated for the rapid rate of somitogenesis. Indeed, increasing the protein half- life of a transcriptional repressor of Lfng (HES7) by only 8 minutes has been shown to significantly disrupt somitogenesis (Hirata et al., 2004). Lfng is not only expressed in

Region I of the PSM where the segmentation clock operates, but also in Region II where

14 R/C patterning occurs, suggesting Lfng may play multiple roles during somitogenesis.

Indeed, researchers have defined genomic sequences sufficient to direct cyclic expression

of Lfng in the PSM, and demonstrated that independent Lfng cis-acting regulatory regions drive stable RNA expression in the rostral compartment of the developing somites in the anterior PSM (Cole et al., 2002; Morales et al., 2002). This suggests that Lfng may play separate roles in Region I and II of the PSM.

1.5.1 Notch signaling plays multiple roles during somitogenesis

In Region I of the PSM, Notch activity levels oscillate, suggesting its function in this region is linked to the clock (Huppert et al., 2005; Morimoto et al., 2005). Some models suggest that this oscillatory activation may be achieved partially through the transitory inhibition of Notch signaling via its glycosylation by LFNG in the Golgi, and transcriptional feedback loops involving Hes7 (Bessho et al., 2003) (Figure 1.4). This oscillatory mechanism, however, clearly receives input from other members of the Notch pathway and from other signaling pathways, including Wnt and FGF as discussed earlier.

In addition to its role in the segmentation clock, Notch signaling is important in the patterning of the presumptive somites in Region II of the PSM. In the presumptive caudal compartment of the developing somites, Notch upregulates Dll1 expression while in the presumptive rostral compartment, Mesp2 acts to downregulate Dll1 expression

allowing Dll3 and Lfng to be expressed and to inhibit Notch activity (Morimoto et al.,

2005; Takahashi et al., 2000) (Figure 1.4). Additionally, knockouts of Notch pathway

members often exhibit segmentation defects both in the morphology and the patterning of

the somites suggesting Notch signaling plays multiple roles in somitogenesis.

15 1.5.2 Notch pathway knockouts

Evidence from multiple model organisms supports the importance of the Notch

signaling pathway during vertebrate somitogenesis, however I will focus my discussion

of the phenotypes caused by mutations in the Notch pathway on the mouse, the model

organism used in my studies. It is quite clear from the phenotypes caused by mutations in the Notch pathway that the proper temporal and spatial regulation of Notch signaling is

of key importance in the process of segmentation. Targeted deletion of the Notch1 gene

causes delayed, disorganized somitogenesis (Conlon et al., 1995), similar to, but slightly

less severe than, the segmentation defects caused by deletion of the Notch coactivator

RBPJk (Oka et al., 1995). A potential role for the Notch2 gene in somitogenesis is

proposed due to the fact that Notch1/Notch2 double-null embryos have more severe phenotypes than Notch1 null embryos, with axial truncation and no somite formation seen after the cervical somites (Huppert et al., 2005). Mutation of the Notch ligand Dll1 causes a failure in the formation of epithelial somites, and a complete loss of R/C somite patterns, with rostralization of the paraxial mesodermal derivatives (Hrabe de Angelis et al., 1997). Thus, mutations that are predicted to reduce or abolish Notch signaling in the

PSM cause segmentation defects in developing mouse embryos, and these defects suggest that both clock function and R/C somite patterning have been disturbed.

Mutations affecting several oscillatory Notch pathway genes also cause segmentation defects. In contrast to the mutations in core Notch members, these mutations tend to be viable, suggesting a more specific developmental role in the process of somitogenesis. Lfng null animals exhibit severe segmentation defects in their vertebrae, ribs, and tails. In these animals, the vertebrae are malformed and there are

16 missing vertebrae, especially in the tail, which is kinked and truncated. Additionally, ribs are frequently bifurcated or fused, and sometimes detached from the vertebral column.

These phenotypes are not fatal, but some severe mutants die after birth, perhaps due to respiratory problems relating to the malformation of the rib cages. Although some allele- specific differences are seen, Lfng null embryos in general have disorganized somitogenesis, a failure to form normal boundaries between somites, and improper R/ C somite patterning (Evrard et al., 1998; Zhang and Gridley, 1998). Furthermore, in the absence of LFNG protein activity, Notch activity levels in the PSM cease to oscillate, and instead are observed as a gradient within the PSM, with highest levels seen in the anterior

PSM (Morimoto et al., 2005).

Targeted deletion of the Hes7 gene causes similar defects to the Lfng null mutation (Bessho et al., 2001b). The role of HES7 protein as an important transcriptional repressor in the segmentation clock is confirmed by the findings that in Hes7 null

embryos both Hes7 and Lfng expression cease to cycle, and are instead ubiquitous

throughout the PSM (Bessho et al., 2003).

Interestingly, the deletion or mutation of the divergent-ligand Dll3 causes

phenotypes similar to those seen in the Lfng null animals. Dll3 null animals are viable

and frequently fertile, but the bodies and tails of these mice are highly truncated due to

missing vertebrae. The vertebral arches are highly disorganized and there are rib

abnormalities including absences and fusions. Boundary formation between somites is

dysregulated, R/C somite patterning fails, and the cyclic expression of clock-linked genes

is disrupted in Dll3 null embryos (Dunwoodie et al., 2002; Kusumi et al., 2004; Kusumi

et al., 1998).

17 The relative importance of proper R/C somite patterning in the process of

vertebrate segmentation is made clear by the targeted deletion of Mesp2. This protein is

expressed exclusively in the anterior PSM, where it defines the presumptive rostral

somite compartment (Saga et al., 1997). MESP2 acts to downregulate Notch1 activity in

this compartment, at least in part through its activation of stable Lfng transcription

(Morimoto et al., 2005). The Mesp2 knockout exhibits disordered segmentation and demonstrates a loss of anterior specification of the somites, the phenotype is similar in severity to Lfng and Dll3 knockouts (Saga et al., 1997). Here, only an indirect effect on clock function can be proposed, and the null phenotype can be largely attributed to dysregulated R/C somite patterning and boundary formation.

Each of the mouse models discussed here demonstrates the importance of the

Notch pathway for proper somitogenesis and these findings have important implications

for several human genetic diseases.

1.6 Human Disease

Several mutations that are predicted to affect segmentation clock function and/or

R/C somite patterning have been linked to human disease. As expected, these mutations

result in disorganization of the axial skeleton, presumably due to the improper regulation

of somitogenesis during embryonic development. In some cases, the phenotypic similarities between the human disease and mutations in model organisms are quite striking (reviewed in Dunwoodie, 2009).

1.6.1 Spondylocostal Dysostosis

Spondylocostal Dysostosis (SCDO) is a group of disorders characterized by short trunk dwarfism with vertebral segmentation defects along the entire spinal column, often

18 accompanied by rib malformations. SCDO is characterized by at least 10 affected

vertebrae as opposed to the more common Abnormal Vertebral Segmentation (AVS)

birth defect which often manifests with a single, defective vertebrae. In many cases,

SCDO is caused by the mutation of genes in the Notch signaling pathway (Bulman et al.,

2000; Sparrow et al., 2006; Sparrow et al., 2008; Whittock et al., 2004b). Different

subtypes of SCDO disorders are inherited in an autosomal recessive or autosomal

dominant fashion (the dominant form is more rare), and an example of pseudodominant

inheritance has also been shown (Whittock et al., 2004a). The radiological features of

SCDO include malformed vertebrae such as hemivertebrae, or butterfly vertebrae.

Additionally, the ossification centers do not always cross the midline and there may be a

reduced or increased number of vertebrae. Individuals have low birth weights and short

statures throughout adulthood due to a shortened trunk (see, for example, Roberts et al.,

1988; Turnpenny et al., 2003; Whittock et al., 2004b). The trunk may also be

asymmetrically shaped and there is often a protrusion of the abdomen. The spine is

curved and often exhibits scoliosis or kyphoscoliosis. Although there are familial cases of SCDO, the disorder more often is seen arising spontaneously in individuals (for a review of SCDO cases see Mortier et al., 1996). Takikawa et al. examined individuals with spontaneous SCDO and found that individuals with the spontaneous form usually have four or more vertebral abnormalities, always have abnormalities in the thoracic region, and also have an average of 11 rib anomalies (Takikawa et al., 2006). They identified common vertebral abnormalities found in the spontaneous SCDO individuals, including butterfly vertebrae (seen at all levels of the spine), hemivertebrae, complete block vertebrae, and unilateral bars.

19 To date, four subtypes of autosomal recessive, familial SCDO have been

associated with mutations that are predicted to affect Notch pathway function in either

the segmentation clock, R/C somite patterning in Region II of the PSM, or both. In many cases, affected individuals exhibit slightly different phenotypes depending on the nature of the gene mutation. SCDO type 1 (OMIM 277300) is caused by mutations in the human DLL3 gene (Bulman et al., 2000). At least 24 different mutations in DLL3 have been identified in SCDO1 families; these mutations are found throughout the gene affecting the EGF repeats, the transmembrane domain, or the intracellular domain of

DLL3, often prematurely terminating the protein sequence (Bonafe et al., 2003; Bulman et al., 2000; Turnpenny et al., 2003; Whittock et al., 2004a). Individuals with SCDO1

have shortened trunks as discussed above, fused or misaligned ribs, and a characteristic

‘‘pebble beach sign’’ of hemivertebrae throughout their vertebral column; the vertebrae

have smooth circular shapes and are not segmented properly. The mobility of the spine is

limited in some affected individuals with multiple scoliosis (Whittock et al., 2004a). The

phenotype of SCDO1 shows a high degree of correspondence with various mouse

mutants of Dll3 suggesting that the role of Dll3 in the Notch pathway during

somitogenesis is conserved from mice to humans (Dunwoodie et al., 2002; Kusumi et al.,

1998). Although Dll3 mRNA expression has not been shown to be cyclic in the posterior

PSM, the loss of Dll3 expression in this region does affect the cyclic expression of other

genes linked to the clock (Kusumi et al., 2004). In addition, the compartmentalized

expression of Dll3 in Region II of the PSM suggests that it may play roles in R/C

patterning of the presomites. The phenotypes seen in SCDO1 may arise from the effects

of DLL3 mutations on clock function, somite patterning, or both.

20 SCDO type 2 (OMIM 608681) is caused by mutations in the human gene

encoding MESP2 (Whittock et al., 2004b). Individuals affected with SCDO2 have a

similar but slightly milder phenotype compared to individuals with SCDO1. A shortened

trunk and neck, and abnormal segmentation of the vertebrae of the backbone have been

described; however, the thoracic group of vertebrae is the most malformed while other

regions of the spine are less affected. Additionally, the vertebrae have more angular,

irregular patterns compared to the ‘‘pebble beach sign’’ of SCDO1. The causal mutation

in SCDO2 was identified as a frameshift mutation that presumably results in the loss of

the carboxy terminus of the MESP2 protein in the affected individuals. It is apparent that

the human MESP2 is important for somitogenesis, again mirroring the mouse Mesp2

knockout. However, the mouse model is not a perfect phenocopy of the human SCDO2,

as it produces quite severe segmentation defects along the length of the vertebral column

(Saga et al., 1997). More recently, it has been suggested that the combined dose of

MESP activity from the Mesp1 and Mesp2 genes is important in the proper regulation of somitogenesis, and that subtle alterations of MESP2 activity levels can lead to localized vertebral defects in the anterior skeleton (Morimoto et al., 2006). The differences seen between SCDO2 and the Mesp2 null mouse may reflect differing requirements for

MESP1 and MESP2 in human compared to mouse, or may indicate that the SCDO2

mutation produces a hypomorphic allele. Given that the expression of Mesp2 is confined

to Region II of the PSM, it may be that the phenotypes caused by Mesp mutations arise

due to dysregulated Notch activity during R/C somite patterning. However, in Mesp2

null embryos, Notch1 and Notch2 expression are reduced or absent in Region I of the

PSM. This suggests that mutations in MESP2 may have an indirect effect on clock

21 function, in addition to affecting the well-understood functions of MESP2 in somite

patterning.

SCDO type 3 (SCDO3, OMIM 609813) is caused by a mutation in the human

LFNG gene (Sparrow et al., 2006). The phenotype of this form is more severe than

SCDO1 and SCDO2. The affected individual has a shortened spine with scoliosis in the cervical and thoracic regions. There are examples of vertebral abnormalities all along the

spine, similar to SCDO1, but the vertebrae are more misshapen and angular, similar to

the malformed vertebrae of SCDO2. The affected individual also has long, slender

fingers and camptodactyly of the left index finger, which may have arisen secondary to

the spinal column defects. The missense mutation of Lfng results in the substitution of a

highly conserved phenylalanine residue near the enzymatic site (Sparrow et al., 2006).

This mutation prevents the proper localization of LFNG to the Golgi and prevents its

ability to modulate Notch signaling through the modification of the Notch receptor. The

severe phenotype of the individual with SCDO3 correlates very well to the severe

phenotype of Lfng knockout mice, again highlighting the importance of the Notch

pathway in the segmentation clock during both mouse and human somitogenesis (Evrard

et al., 1998; Zhang and Gridley, 1998). Researchers have suggested that Lfng may have

distinct and separable roles in the segmentation clock and R/C somite patterning (Cole et

al., 2002; Morales et al., 2002), and the phenotypes of Lfng null mice and SCDO3

patients may have arisen due to dysregulation of either or both of these processes.

SCDO type 4 (SCDO4, OMIM 608681) is caused by a mutation in the human

HES7 gene (Sparrow et al., 2008). All regions of the affected individual’s spine are

affected with angular, disorganized vertebrae, although similar to SCDO2, the thoracic

22 region is highly disrupted. The ribs are not aligned properly, fused in many places, and

show crowded origins. Additionally, the individual was diagnosed with hydrocephalus,

myelomeningocele, and a stenotic anus. The causative mutation alters a highly conserved

basic residue in the DNA binding domain of HES7 which impairs its ability to repress transcription in vitro. Again, the phenotype observed in SCDO4 correlates very well with that of the Hes7 knockout mice which have highly disorganized vertebrae and rib cages (Bessho et al., 2001b).

It is clear, therefore, that the proper regulation and function of the segmentation clock underlies the establishment of the segmental vertebrate body plan, and that mutations affecting the roles of the Notch signaling pathway in this clock are a cause of human disorders characterized by disorganization of the axial skeleton.

1.7 The regulation of Lfng during somitogenesis

The work in this dissertation characterizes the role of Lfng during somitogenesis and examines the regulation of cyclic Lfng activity in the PSM. The analysis of the segmentation clock mechanism has been complicated by the fact that Notch signaling appears to have multiple, distinct roles during somitogenesis. Notch signaling is likely involved in both the establishment of segmental identity through the segmentation clock as well as in the determination and maintenance of R/C identity of presumptive somites.

In Chapter 2, to dissect the roles played by Notch pathway member Lfng during somitogenesis, we created a novel allele that lacks cyclic Lfng expression within the segmentation clock, but that maintains Lfng expression during R/C patterning

(LfngΔ FCE1). We analyzed LfngΔ FCE1/ΔFCE1 mice for vertebral defects and LfngΔ FCE1/Δ FCE1 embryos for effects on various clock components as well as mature somite patterning

23 markers. We find that LfngΔ FCE1/Δ FCE1 mice have segmentation defects specifically in

their anterior skeleton. The cyclic expression pattern of some clock genes is disrupted,

depending on the stage of somitogenesis, while R/C patterning genes show relatively

normal expression patterns in LfngΔ FCE1/Δ FCE1 embryos. These studies show that cyclic

Lfng is necessary for proper segmentation of the anterior, but not posterior skeleton.

As part of the segmentation clock, both Lfng mRNA and protein levels cycle in

the PSM during somitogenesis (Dale et al., 2003). We hypothesize that cyclic Lfng transcription is coupled with signals in the Lfng 3’UTR that confer a short half-life on

Lfng mRNA. In Chapter 3, we examine the expression patterns of transgenes constructed with various conserved sections of Lfng 3’UTR to dissect the functions of these conserved regions in Lfng RNA stability. Our results suggest that several regions of the

Lfng 3’UTR coordinately contribute to a short mRNA half-life. In Chapter 4, we examine the mechanism by which LFNG protein intracellular half-life is controlled. We hypothesize that the cleavage and secretion of the LFNG protein influences its intracellular half-life. We therefore examine the cleavage of LFNG by furin-like proteases and find that the cleavage of LFNG is not necessary for its secretion but does influence its intracellular half-life. To determine the importance of LFNG cleavage and secretion in vivo, we created a novel allele that tethers LFNG to the Golgi (R/LFNG).

Mice heterozygous for the R/LFNG allele show significant segmentation and patterning defects suggesting that the short intracellular half-life of LFNG is important for somitogenesis. Thus, the cyclic activity of Lfng in the segmentation clock is achieved through multiple mechanisms, including tight regulation of mRNA and protein levels.

24 The results of this thesis reveal that Lfng plays an important role in spatially and temporally regulating Notch signaling during anterior vertebrate segmentation.

Figure 1.1: Mouse somitogenesis. A. Picture of a mouse tail showing the overtly unsegmented Presomitic Mesoderm (PSM) and mature, epithelial somites which have formed on either side of the neural tube. B. Schematic representation of the PSM. The somites (represented by the circles) are formed on either side of the neural tube (black line). The presomitic mesoderm (PSM) is the overtly unsegmented tissue posterior to the mature somites which can functionally be divided into separate regions: Region I is where the segmentation clock operates, Region II is where the pre-somites (S-II, S-I, S0) are patterned with Rostral/Caudal specific gene expression (grey boxes). C. Schematic of the resegmentation event which occurs after somites have matured: the posterior half of a somite and the anterior half of its neighboring somite to give rise to a single vertebral body.

Presomitic Mesoderm Mature Somites A

B Region I Region II

Segmentation Clock R/C Patterning

S-II S-I S0 SI C

vertebrae

PSM

Figure 1.1

Wavefront

Clock ON OFF ON OFF ON

Figure 1.2: The Clock and Wavefront model. In the Clock and Wavefront model, the clock oscillates in Region 1 of the PSM (ON and OFF) with the same period as somite formation, and is controlled by multiple signaling pathways. The wavefront (horizontal line) moves down the PSM from the anterior. Only when the clock is on and the wavefront has moved past the proper number of cells is there formation of presomites which will be patterned in Region II (boxes) and differentiate into mature, epithelial somites (circles).

Hes1

Transcriptional delay

RNA RNA Degradation

Translational delay

Protein Protein Degradation

Figure 1.3: Schematic of a clock gene feedback loop. The Notch target HES1 inhibits its own transcription causing a negative feedback loop. When researchers modeled this loop, they incorporated transcriptional and translational time delays, transport of Hes1 mRNA, HES1 dimerization, Hes1 mRNA stability, and HES1 protein stability in order to produce an accurate model (Momiji and Monk, 2008).

DSL Ligand DSL Ligand

Notch Receptor Notch Receptor

Presenillin Presenillin ICN

FRINGE FRINGE Cytosol Cytosol Golgi Golgi

ICN

bHLH bHLH CSL CSL MAML MAML

Nucleus Nucleus

Figure 1.4: Overview of the Notch Signaling pathway. The Notch receptor is modified by Lunatic Fringe (FRINGE) in the Golgi through the addition of sugar moieties (small pentagons). Notch receptor on the cell surface binds to DSL ligands, upon which protein cleavages (one requiring Presenilin) release the active intracellular domain (ICN). ICN translocates to the nucleus and forms a complex containing the CSL transcription factor and Mastermindlike proteins (MAML), which directly activates the transcription of target genes including bHLH proteins such as c-hairy and the hes genes (reviewed in (Weinmaster and Kintner, 2003)).

Lfng

S-IIS-I S0 SI

Region I Region II Segmentation Clock Rostro-caudal patterning

Notch Dll1 Mesp2

Lfng Notch Hes7 Lfng Notch

Figure 1.5: A schematic of Notch signaling in Regions I and II of the PSM. In the posterior PSM (Region 1), Lfng cycles as a part of a feedback loop where Notch turns on Hes7 and Lfng, Lfng inhibits Notch activity, and Hes7 inhibits its own transcription as well as that of Lfng (Bessho et al., 2003). In the anterior PSM (Region 2), Mesp2 turns on Lfng which then inhibits Notch creating a border between Notch activated and repressed cells corresponding to the boarder between the caudal half of one presomite and the rostral half of the next presomite (Takahashi et al., 2000).

CHAPTER 22

LUNATIC FRINGE IS IMPORTANT FOR SEGMENTATION CLOCK FUNCTION DURING PRIMARY, NOT SECONDARY BODY FORMATION

2.1 Introduction

There are a number of factors to consider when studying the regulation of somitogenesis. As mentioned above, early in development cells enter the presomitic mesoderm via the primitive streak. Later in development (after 10.0 dpc) the tailbud forms, and mesodermal cells now arise from this structure. Although this distinction between primary body formation (giving rise to cervical, thoracic and lumbar vertebrae) and secondary body formation (giving rise to post-anal structures) was originally proposed in 1925 (Holmdahl, 1925), it remains unclear how, and to what extent, the genetic regulation of somitogenesis between these two processes may vary (reviewed in

Handrigan, 2003).

The analysis of Notch signaling in the segmentation clock is further complicated by the fact that this pathway appears to play multiple roles during somitogenesis. As discussed above, many Notch pathway members and downstream targets exhibit cyclic expression in Region I of the PSM and then become restricted to either the caudal or

2 Portions of this chapter were originally published in Development, Volume 135, Pages 899-908. 32 rostral compartments of the pre-somites in Region II. Notch signaling appears to play

multiple roles in somitogenesis; the phenotypes associated with Notch knockouts suggest

multiple aspects of somitogenesis have been disrupted. Researchers have defined

genomic sequences sufficient to direct cyclic expression of Notch pathway member Lfng

in the PSM, and demonstrated that independent Lfng cis-acting regulatory regions drive stable RNA expression in the rostral compartment of the developing somites in the anterior PSM (Cole et al., 2002; Morales et al., 2002). Deletion of a conserved regulatory

element termed Fringe Clock Element 1 (FCE1) from Lfng reporter transgenes eliminates cyclic expression in the caudal PSM, while maintaining expression in the anterior PSM, reflecting the distinct roles of Lfng in the segmentation clock and in R/C patterning of the developing somites. Thus, the complex phenotypes of Lfng-/- mice may arise from

disruption of both of these roles, with variations in somite size perhaps resulting from impaired clock function, while the apparent mingling of somite compartments might be exacerbated by altered R/C patterning.

To dissect the functions of the Notch pathway during segmentation, we perturbed only one of the roles of Notch signaling, by disrupting oscillatory Lfng expression in

Region I of the PSM, while sparing its expression in Region II of the PSM. We find that

the clock and patterning roles of Lfng during somitogenesis are functionally separable.

Strikingly, the loss of oscillatory Lfng expression and Notch1 activity in Region I of the

PSM has more severe effects during the segmentation of the thoracic and lumbar skeleton

than the sacral and tail skeleton. This suggests that oscillatory Notch1 activation in the

segmentation clock is much more important during primary body formation than during

secondary body formation. In contrast, the specific localization of Notch activity to the

33 presumptive caudal compartment of the pre-somite in Region II of the PSM is important

throughout development.

2.2 Materials and Methods

2.2.1 Targeted deletion of FCE1

164 bp encompassing FCE1 and minimal flanking sequences were deleted from

the Lfng fragment extending from the 5’XhoI site in the 5’ flank to the HindIII site in

intron 1 and replaced with an EcoRV site (final allele:

ggactttttccttgtcctGATATCaccaccatatcccactcc, upper case = EcoRV). This deleted

fragment was the 5’flanking sequence for a floxed cassette containing a neo /testis cre

cassette (Bunting et al., 1999). 3’ flanking sequences extended from the HindIII site to

the XhoI site in intron 1. Linearized vector was electroporated into TC1 cells (Deng et al.,

1996) and G418 resistant colonies were screened by Southern Blot analysis. Two

independent ES cell lines were injected by the OSUCCC Transgenic/ES Core Facility,

and transmitted through the germline. Results from the lines were identical and are

combined. Mice were maintained on a mixed 129/SvxC57BL/6J background. LfngtmRjo1/+ mice (R. Johnson), were maintained on a mixed 129SVxC57BL6J background, or crossed one generation with FVB mice to increase the recovery of adult LfngtmRjo1/tmRjo1

mice (referred to as Lfng-/- here). Mice were maintained in an SPF facility under the care

of the Ohio State University ILACUC.

2.2.2 Genotyping

Genomic DNA was prepared from tail clips via proteinase K saltout or from yolk sac fragments via the HOTSHOT procedure (Truett et al., 2000). Animals were genotyped by PCR. LfngtmRjo1 primers FNG322 (5’-GAGCACCAGGAGACAAGCC-3’) 34 FNG325 (5’-AGAGTTCCTGAAGCGAGAG-3’) and PGK3 (5’-

CTTGTGTAGCGCCAAGTGC-3’) amplify a 170bp wt product and a 200 bp mutant

product. LfngΔFCE1 primers SC284 (5’-TTTGGTGGGAATGGATTAGC-3’) and SC285

(5’-CTGGTCCATTTGCTCTGAGG-3’) produce a 340 bp wildtype band and a 182 bp

mutant band while SC286 (5’-TTGGGTCTATCTGGGAAACG-3’) and SC287 (5’-

GCGACTCATCCAGACACAGA-3’) produce a 149 bp wildtype band and a 250 bp

mutant band.

2.2.3 Whole mount in situ hybridization

Embryos were collected from timed pregnancies with noon of the day of plug

identification designated as 0.5 dpc RNA in situ hybridization using digoxigenin-labeled

probes was performed essentially as described (Riddle et al., 1993), however: before incubation with AP linked antibody embryos were blocked in a mixture of MABT +20% sheep serum +2% Boehringer blocking reagent, and all post-antibody washes were

performed in MABT. Hes7 cDNA probes extend from the internal SmaI site to the stop

codon. Brachyury cDNA probes extend from PstI to EcoRI. Hes7 intron probe was

amplified using the primers 5'-GCTAGAGGCCATAGCTGGTG and 5'-

CTGTGACCAGCGGGAAAG. Dll1 intron probe was amplified using primers 5'-

GTTGGCAGTGGGAAGAAGG and

5'-TGTGTTGTGCCAATGAAGGT. Nrarp probe was amplified using the primers 5'-

GCGTGGTTATGGGAGAAAGA and 5'-TTCCTCCCACACTGGTTCAT. The Hesr1

and Hes5 probes are comprised of the complete coding regions of cDNAs. Other probes were Lfng (Johnston et al., 1997), Mesp2 (Saga et al., 1997), Uncx4.1 (Mansouri et al.,

1997), Mox2 (Candia et al., 1992), Axin2 (Huang and Gorman, 1990), Tbx18 (Kraus et

35 al., 2001), and Pax9 (a gift from R. Johnston). Double label in situ experiments were

performed as described previously (Cole et al., 2002).

2.2.4 Skeletal preparations

Skeletal preparations of neonates or 18.5 d.p.c embryos were performed

essentially as described (Kessel and Gruss, 1991).

2.2.5 Whole mount immunohistochemistry

Embryos were fixed in fresh 4% PFA in PBS, then washed in PBS. After

overnight incubation at 4°C in PBS containing 0.1% hydrogen peroxide, 1% Triton X-

100, and 10% fetal calf serum (TS-PBS) embryos were transferred into 10 mM Na

Citrate pH 6.0, 0.1% Tween-20 (CT), boiled in CT for 10 minutes and then transferred back to PBS. After washing in TS-PBS, embryos were incubated for five days in primary

Cleaved Notch1 (Val1744) antibody (Cell Signaling Technology) diluted 1:250 in TS-

PBS. After washing in TS-PBS embryos were incubated overnight in AP-conjugated secondary antibody diluted 1:500 in TS-PBS. After extensive washing in MABT, embryos were transferred to NTMT and stained with BCIP/NBT as described (Riddle et al., 1993).

2.3 Results

2.3.1 Deletion of FCE1 from the Lfng locus perturbs clock-linked Lfng RNA expression

To specifically disrupt Lfng expression in the segmentation clock, we deleted

FCE1 from the endogenous Lfng locus resulting in the allele LfngΔFCE1 (Figure 2.1). We

hypothesized that this mutation would disrupt expression of Lfng in the caudal PSM

36 (Region I), where the clock is active, while preserving Lfng expression in the anterior

PSM (Region II), where R/C somite patterning is initiated.

Lfng expression is perturbed in the PSM of LfngΔFCE1/Δ FCE1 mutant embryos

(Figure 2.2). In wild-type embryos, three distinct phases of Lfng expression are seen in the PSM, reflecting cyclic expression. In contrast, LfngΔ FCE1/Δ FCE1 embryos express Lfng

RNA in a single band in the anterior PSM, with no expression observed in the caudal

PSM where the clock is active at any embryonic stage tested. Although the anterior band

of Lfng expression in LfngΔ FCE1/Δ FCE1 embryos is weaker than the anterior-most band of

Lfng expression in wild-type embryos, these results demonstrate that the deletion of the

FCE1 enhancer prevents oscillatory expression of Lfng in Region I of the PSM, while

sparing some level of expression in Region II. In addition, we find that Lfng expression in Region II of the PSM is largely confined to the presumptive rostral compartment of somite S-1 (Figure 2.3), indicating that the endogenous Lfng expression pattern in the anterior PSM is preserved in the LfngΔ FCE1 allele.

2.3.2 The loss of Lfng expression in the segmentation clock perturbs normal skeletal

development

Although the Lfng-/- genotype is reported to be viable, we find that on a mixed

129/SvxC57BL/6J background, only rare animals survive postnatally, and homozygous

males are infertile. In contrast, homozygous LfngΔ FCE1/Δ FCE1 animals survive to adulthood

at Mendelian ratios, and homozygous animals of both sexes are fertile. Lfng∆FCE1/∆FCE1

animals have segmentation defects including shortened body and variably kinked tails

(Figure 2.4 A). In the anterior skeleton, both Lfng-/- and LfngΔ FCE1/Δ FCE1 animals are

severely affected. Multiple rib fusions and bifurcations as well as severely disorganized 37 vertebrae are observed (Figure 2.4B, panels a-f). When defects in the thoracic region of the skeleton are quantified, we find similar levels of disorganization in LfngΔ FCE1/Δ FCE1

animals and Lfng-/- animals (data not shown).

In the more posterior skeleton, however, LfngΔ FCE1/Δ FCE1 animals are much less

affected than Lfng-/- animals (Figure 2.4 B, panels g-i). In the thoracic and lumbar region

of the skeleton, vertebral condensations in both LfngΔ FCE1/Δ FCE1 and Lfng-/- animals are

irregular and misaligned. Strikingly, this pattern is altered at the lumbo-sacral junction.

In the sacral region of LfngΔ FCE1/Δ FCE1 animals, normal vertebral condensations are seen in all animals, and the tail vertebrae appear relatively normal, though variable kinks in the tail are seen ranging from mild (0-1 kinks in 40% of adult mice) to moderate (2-5 kinks in 60% of adult mice). In contrast, in Lfng-/- animals, vertebral condensations are

abnormal throughout the sacral region and the tail is invariably truncated, a phenotype

never seen in LfngΔ FCE1/ΔFCE1 animals. Thus we find that the loss of oscillatory Lfng

expression in Region I of the PSM causes pronounced defects in the axial skeleton, but

these defects are much more pronounced in the thoracic and lumbar regions, while the

sacral and more caudal regions of the skeleton are less affected in comparison to the null

allele. Somites that will contribute to the thoracic and lumbar regions are produced

during primary body formation (Gossler, 2002). In Lfng∆FCE1/∆FCE1 embryos these

somites are irregularly sized and spaced with frequent fusions between neighboring

somites, while somites formed during secondary body formation are relatively evenly sized and spaced (data not shown). After 10.0 dpc, we observe normal sized somites with

proper separation. In fact, the lumbo-sacral junction, the point where somite and skeletal

morphology largely recovers in LfngΔ FCE1/Δ FCE1 animals, represents the transition point 38 between primary and secondary body formation, suggesting that oscillatory Lfng plays a

minor role at best in secondary body formation.

2.3.3 Rostral-caudal somite patterning is partially rescued in Lfng∆FCE1/∆FCE1 embryos

Lfng-/- embryos have severe defects in R/C somite patterning (Evrard et al., 1998;

Zhang and Gridley, 1998). To address whether the Lfng expression in Region II of the

PSM of LfngΔFCE1/Δ FCE1 embryos could rescue R/C patterning, we examined compartment

formation in the anterior PSM and in the mature somites of Lfng mutant mice. We

examined R/C patterning in Region II of the PSM by assessing the expression of Mesp2.

Mesp2 defines the presumptive rostral compartment of somite S-1, and interacts with

Lfng and Notch1 signaling during the process of R/C patterning (Morimoto et al., 2005;

Takahashi et al., 2000). During both primary and secondary body formation, we find that

Mesp2 is expressed in a single band of varying width in the anterior PSM of both wild-

type and LfngΔFCE1/Δ FCE1 embryos reflecting the early expression and subsequent

refinement of Mesp2 in the presumptive rostral compartment. However, in

LfngΔFCE1/Δ FCE1 embryos we frequently see a less distinct rostral border, regardless of the

stage of somitogenesis (Figure 2.5 A). These results demonstrate that the rostral

compartment is being defined in the presomites in Region II of LfngΔFCE1/ΔFCE1 embryos

throughout somitogenesis but may suggest that this earliest marker of patterning is mildly

disrupted. As this disruption is seen throughout somitogenesis, it may be due to the

reduced dosage of Lfng in the anterior PSM, rather than to differences in primary and

secondary body formation.

39 We then examined patterning of the mature somites. Uncx4.1, marking the caudal compartment of epithelial and mature somites, is expressed in clear compartments in all somites of wild-type embryos during primary body formation (Figure 2.5, panels a, b). In

Lfng-/- embryos little compartmentalization of somites is seen, with rostral and caudal

cells appearing mixed in a "salt and pepper" pattern (Figure 2.5, panels i, j, Evrard et al.,

1998; Zhang and Gridley, 1998). In LfngΔ FCE1/Δ FCE1 embryos at this stage Uncx4.1 expression in newly formed somites is largely compartmentalized, with stronger expression in the more caudal region of somites S1 and S2. Clearer compartmentalization is observed in more anterior somites, but compartments are frequently irregularly spaced (Figure 2.5, panels e, f). Compartmentalization of the mature somites in the thoracic region of LfngΔFCE1/Δ FCE1 embryos is more distinct by 10.5

dpc, with clear bands of Uncx4.1 visible in the sclerotome. Again, compartments of

Uncx4.1 expression are frequently irregularly spaced or shaped, presumably reflecting the

irregularities in somite size and shape observed morphologically in the thoracic region of

the embryo (Figure 2.5, panel g). In contrast, the Uncx4.1 signal in the thoracic region of

Lfng-/- embryos fails to compartmentalize, maintaining an unsegmented pattern (Figure

2.5, panel k). The somites formed during secondary body formation in LfngΔFCE1/Δ FCE1

embryos are clearly compartmentalized with regular rostral and caudal segmentation,

while in Lfng-/- embryos R/C patterning continues to be abnormal (compare Figure 2.5 B

panels h and l).

Thus, although LfngΔFCE1/Δ FCE1 embryos produce irregular somites during primary body formation, the retention of Lfng expression in the anterior PSM supports relatively

normal R/C patterning, and somites formed during secondary body formation undergo 40 normal R/C patterning. This supports the idea that the role of Lfng in R/C somite

patterning is distinct and separable from its functions in the segmentation clock.

2.3.4 Somites differentiate properly in Lfng∆FCE1/∆FCE1 embryos

After somites leave the PSM, they continue to differentiate with groups of cells

becoming mesenchymal sclerotome and others differentiating into dermomyotome.

Mox1 is expressed in all somite cell types, but becomes localized to the mesenchyme and

is strongly expressed in the posterior (Candia et al., 1992). We examined the expression

of Mox1 in LfngΔFCE1/Δ FCE1 embryos and found that it is expressed in the proper somite compartment, although its expression is sometimes irregular, reflecting irregularly sized somites in the thoracic region (Figure 2.6 A, panels d,e). Tbx18 is expressed in the anterior lateral sclerotome where it helps maintain the R/C polarity initially established in

Region II of the PSM in mature somites (Bussen et al., 2004; Kraus et al., 2001). Pax9, on the other hand, is expressed in the posterior, ventrolateral sclerotome (Neubuser et al.,

1995). Both markers of sclerotome show normal expression patterns in LfngΔFCE1/Δ FCE1

embryos (Figure 2.6 B,C) suggesting that the maturation and differentiation of somites is not disrupted by the ∆FCE1 allele.

2.3.5 The loss of cyclic Lfng expression in the posterior PSM perturbs oscillatory

NOTCH1 activity

Several groups have suggested that oscillatory expression of Lfng is involved in interlocking feedback loops that regulate oscillatory Notch1 activation in the PSM. To examine the effects of the Lfng∆FCE1 allele on Notch1 signaling, we visualized Notch

activation using an antibody specific for the Notch1 Intracellular Domain (NICD). Notch

signaling levels oscillate in the PSM of wild-type embryos during primary and secondary 41 body formation (Fig. 5 and Morimoto et al., 2005), with different patterns of NICD

staining found in different embryos. In contrast, in both Lfng∆FCE1/∆FCE1 (Figure 2.7) and

Lfng-/- (Figure 2.7 and Morimoto et al., 2005) embryos a gradient of NICD is seen in the

PSM, reflecting ubiquitous, non-oscillatory Notch signaling throughout the PSM. This

confirms that the Lfng∆FCE1 allele inhibits oscillatory Notch signaling in Region I of the

PSM during both primary and secondary body formation, and indicates that oscillatory

Notch activation in Region I of the PSM is largely dispensable for segmentation during

secondary body formation.

2.3.6 Expression of oscillatory genes is differentially affected during primary and secondary body formation in Lfng∆FCE1/∆FCE1 embryos

To assess the effects of the loss of cyclic Lfng expression on the transcription of

other segmentation clock genes, we first examined the expression of Hes7 in

Lfng∆FCE1/∆FCE1 embryos. Hes7 has been proposed to play a role in the segmentation

clock mechanism as part of the feedback loops regulating oscillatory Lfng transcription and Notch1 activation (reviewed in Rida et al., 2004). One report has suggested that

Hes7 expression is ubiquitous in the Lfngtm1Grid/tm1Grid null background at 9.5 dpc (Chen et

al., 2005), while more recent results suggest the Hes7 expression is affected but still

dynamic in the absence of Lfng (Niwa et al., 2007). During primary body formation, we

utilized a probe specific for Hes7 intronic sequences to show that Hes7 RNA is

transcribed in a stable, ubiquitous pattern in the PSMs of Lfng∆FCE1/∆FCE1 and Lfng-/- embryos, distinct from the dynamic banding pattern seen in wild-type embryos (Figure

2.8 A). Thus, during primary body formation, the loss of Lfng prevents the cyclic transcription of Hes7. In sharp contrast, we find that during secondary body formation

42 Hes7 mRNA transcription oscillates similar to wild-type expression patterns in both

Lfng∆FCE1/∆FCE1 and Lfng-/- embryos (Figure 2.8 B). Similar results were seen utilizing a

Hes7 mRNA probe, indicating that post-transcriptional regulation of Hes7 mRNA levels

is also normal in these embryos (Figure 2.8 C). Hes7 cyclic expression was confirmed by

half tail culture experiments. PSMs were bisected, with one half fixed immediately and

the other half cultured before fixation. After one hour of culture the Hes7 expression

pattern in the cultured half is different from the uncultured half regardless of genotype

(Figure 2.8 D). These results suggest that Hes7 transcription may be differentially controlled during different stages of somitogenesis, requiring Notch oscillations during

primary body formation, but not during secondary body formation.

We confirmed and extended these observations by analyzing the expression of

other oscillatory genes in Lfng∆FCE1/∆FCE1 embryos. Similar to our results with Hes7, we

find that Nrarp expression is differentially affected during primary and secondary body

formation. Before tailbud formation distinctive banding patterns are observed in

wildtype, but not in Lfng∆FCE1/∆FCE1 embryos (Figure 2.9 A). After tailbud formation,

oscillatory Nrarp expression recovers in Lfng∆FCE1/∆FCE1 embryos, although the expression

patterns in Lfng∆FCE1/∆FCE1 PSMs are less distinct than those in wildtype embryos. Other

Notch pathway genes also oscillate during secondary body formation in Lfng∆FCE1/∆FCE1 embryos, including Dll1 (Figure 2.9 B). However, the expression of some Notch targets, including Hes5 (Barrantes et al., 1999) and Hesr1 (Leimeister et al., 1999) is perturbed at this stage (Figure 2.10). Thus, although multiple genes that may be involved in the segmentation clock mechanism exhibit oscillatory expression in the absence of cyclic

Notch activation during secondary body formation in Lfng∆FCE1/∆FCE1 embryos, cyclic 43 Notch1 activity is required for proper expression of some genes in the region at this stage.

2.3.7 Wnt targets continue to cycle, and Brachyury expression is not disrupted in

Lfng∆FCE1/∆FCE1 embryos

Not only do members of the Notch pathway cycle in the mouse PSM, but

members of the Wnt pathway such as Axin2 also cycle. We examined the expression pattern of Axin2 in wild-type and Lfng∆FCE1/∆FCE1 embryos and found no difference

between the expression patterns (Figure 2.11), confirming that cyclic Notch activity is not

necessary for Axin2 expression during secondary body formation (Aulehla et al., 2003)

similar to Hes7. To determine if the mesoderm being produced in the tailbud was affected by the ∆FCE1 allele or by the complete loss of Lfng in the PSM, we examined the expression of Brachyury (Wilson et al., 1993) in wild-type, Lfng∆FCE1/∆FCE1, and Lfng-

/- embryos at 12.5 dpc. There are no differences in Brachyury expression at this stage,

suggesting the tailbud is able to produce mesoderm in the absence of Lfng.

2.4 Discussion

2.4.1 The FCE1 enhancer is necessary for cyclic expression of Lfng in Region I of

the PSM

Ascertaining the functions of Notch signaling in segmentation is complicated by

the fact that the pathway plays multiple roles during somitogenesis. It is unclear what

aspects of the Lfng-/- phenotype can be ascribed to its role in R/C patterning as opposed to

any role in clock function, as both aspects of expression are perturbed throughout

development in the mouse knockout. Indeed, in zebrafish, Lfng is expressed solely in the

anterior-most region of the PSM indicating that in this organism Lfng plays no role in the

44 segmentation clock (Prince et al., 2001). We therefore specifically disrupted the

oscillatory expression of Lfng in Region I of the PSM to examine the role it plays in the

segmentation clock during mouse somitogenesis. Targeted deletion of FCE1 sequences

eliminates expression of Lfng in the posterior region of the PSM, indicating that the

enhancer is required for cyclic Lfng transcription in Region I of the PSM. However, the

anterior band of Lfng expression in Lfng∆FCE1/∆FCE1 embryos is weaker than that seen in

wild-type embryos, perhaps supporting an additional role for FCE1 in enhancing expression of Lfng in the anterior PSM. It is difficult to eliminate the possibility that

there are low levels of residual expression of Lfng in the posterior PSM, however these

levels would not be sufficient to support overt oscillation of Notch1 activation.

2.4.2 Oscillatory Lfng expression and Notch signaling are critical for the proper

segmentation during primary, but not secondary, body formation

As predicted, the loss of oscillatory Lfng expression in Region I of the PSM affects segmentation in Lfng∆FCE1/∆FCE1 embryos, but distinct effects are seen during primary and secondary body formation. In the thoracic and lumbar skeleton, malformed vertebral condensations and rib abnormalities were seen in Lfng∆FCE1/∆FCE1 skeletons

(Figure 2.4). The appearance of the vertebrae resembles the phenotypes seen in some

cases of autosomal recessive spondylocostal dysostosis caused by mutations in

Deltalike3, Lunatic fringe, or HES7; both the thoracic and lumbar spine are affected and

vertebral bodies are irregularly shaped and fitted together (Bulman et al., 2000; Sparrow

et al., 2006; Sparrow et al., 2008). To our surprise, we found that the caudal skeletal

regions (sacral and tail vertebrae) were invariably less severely affected in Lfng∆FCE1/∆FCE1 animals than in Lfng-/- animals. Especially striking is the fact that in the sacral region of 45 the spine, Lfng∆FCE1/∆FCE1 animals exhibit essentially normal vertebral formation, while

irregularities are still seen at this level in Lfng-/- skeletons. The point of phenotype recovery at the lumbo-sacral junction indicates that secondary body formation occurs relatively normally in LfngΔ FCE1/Δ FCE1 embryos.

This differential severity is reflected in the process of somitogenesis throughout

development. During primary body formation, somites are frequently abnormal in

LfngΔ FCE1/Δ FCE1 embryos. The production of irregularly sized somites suggests that the loss of oscillatory Lfng expression in Region I of the PSM interferes with segmentation

clock function during primary body formation. In contrast, during secondary body

formation, somites formed in Lfng∆FCE1/∆FCE1 embryos are evenly spaced and of regular

size, and the phenotypes in the sacral and caudal skeleton are correspondingly milder

(Figure 2.4). Thus, our data suggest that segmentation of the embryo during primary

body formation (contributing to the thoracic and lumbar skeleton) is more sensitive to the

loss of cyclic Lfng expression than is segmentation during secondary body formation.

This sheds new light on one of the classical questions of developmental biology: the

extent to which primary and secondary body formation represent distinct mechanisms of

development.

Interestingly, although Dll3 null embryos exhibit similar Lfng expression patterns as those observed in LfngΔ FCE1/Δ FCE1 mice, at least after 9.5 dpc (Dunwoodie et al., 2002;

Kusumi et al., 2004), they exhibit disordered somitogenesis along the length of the

vertebral column, suggesting that Lfng expression in Region II of the PSM is not, in and

of itself, sufficient to rescue secondary body formation. This may reflect a requirement

for Dll3 expression in the anterior PSM during secondary body formation. Alternatively, 46 it was recently shown that the loss of Dll3 in the PSM leads to a loss or reduction in

NICD levels in Region I of the PSM, in contrast to the ubiquitous Notch1 activation

observed in Lfng∆FCE1/∆FCE1 embryos (Geffers et al., 2007). This raises the possibility that

while oscillatory Notch1 activation in the posterior PSM is not required during secondary

body formation, some level of Notch1 activation is still necessary during this process and

may be especially interestingly in light of the observation that constitutive overexpression

of Lfng in the mouse PSM, which might be predicted to repress Notch1 activation, also

perturbs somitogenesis along the entire axial skeleton (Serth et al., 2003).

2.4.3 Lfng plays separable roles in the segmentation clock and R/C patterning

Lfng-/- embryos have severe defects in R/C somite patterning (Evrard et al., 1998;

Zhang and Gridley, 1998). This could arise due to downstream effects of the loss of Lfng

in the segmentation clock, or more directly due to the loss of Lfng expression in Region II

of the PSM. Analysis of R/C patterning in LfngΔ FCE1/Δ FCE1 embryos directly assesses

whether the retention of Lfng expression in the presumptive anterior compartment of the

forming somite can rescue R/C somite patterning in the absence of oscillatory Notch

activity in the clock. During secondary body formation, Lfng∆FCE1/∆FCE1 embryos produce

regular pairs of somites, and these somites are properly patterned. More surprisingly, the

irregular somites produced during primary body formation in Lfng∆FCE1/∆FCE1 embryos are

also patterned into clear rostral and caudal compartments (Figure 2.5), although this

patterning may be somewhat delayed. We propose that in LfngΔ FCE1/Δ FCE1 embryos,

Mesp2 expression in the anterior compartment of the developing somite is able to

stabilize the pattern of Notch activation in somites S-0 and S-1, at least in part via its

specific activation of Lfng transcription. This allows the Notch pathway to function in 47 R/C patterning in LfngΔ FCE1/Δ FCE1 embryos despite the loss of cyclic Lfng expression in

Region I of the PSM. It is not clear at this time whether the delay in robust R/C patterning of thoracic somites is due to some underlying disorganization of somites S-1

and S0 as a result of perturbed clock function, or if it may be a result of the reduced Lfng

levels seen in the anterior PSM in LfngΔ FCE1/Δ FCE1 embryos. However, the successful patterning of irregularly sized somites during primary body formation in LfngΔ FCE1/Δ FCE1

embryos suggests that the Notch-based processes involved in the segmentation clock can

be largely divorced from its roles in R/C somite patterning, and that the processes

regulated by oscillatory Notch signaling in Region I of the PSM are not prerequisites for

the patterning of the pre-somites in Region II.

2.4.4 Differential segmentation clock regulation at distinct levels of the axial

skeleton?

The loss of cyclic Notch1 activation has distinct effects during primary and

secondary body formation. During early stages of somitogenesis, the loss of oscillatory

Lfng expression interferes with oscillatory Notch activation (Figure 2.7) and causes

phenotypes (irregular somite size and positioning, alteration of oscillatory gene

expression) that suggest defects in segmentation clock function (Figure 2.8 A). In

contrast, during later stages of segmentation, despite the continued absence of oscillatory

Lfng and the presence of ubiquitous Notch1 activation, somitogenesis proceeds relatively

normally in LfngΔ FCE1/Δ FCE1 embryos, and the oscillatory expression of several clock

genes largely recovers at these stages (Figure 2.8 B-D, 2.9). Although expression of

some Notch target genes is slightly affected during secondary body formation in

LfngΔ FCE1/Δ FCE1 embryos (Figure 2.10), the mild phenotypes observed in the caudal axial 48 skeleton suggest that these perturbations are relatively unimportant. Thus it appears that segmentation clock function is more sensitive to the loss of oscillatory Lfng expression during primary body formation than during secondary body formation.

Differential regulation of somitogenesis at different axial levels of the embryo is not unprecedented. The first 5-6 somites are frequently spared in mutations that affect the Notch signaling pathway, although in zebrafish these segments can be affected by the simultaneous downregulation of several clock components (Oates et al., 2005), perhaps indicating multiple, parallel mechanisms of regulation. More recently, it was shown that in zebrafish the anlagen of the anterior trunk, posterior trunk and tail are specified before somitogenesis begins, raising the possibility that different genetic pathways may affect these regions in distinct ways (Szeto and Kimelman, 2006).

Our data expand on these observations, suggesting that in the mouse somitogenesis during secondary body formation is controlled largely by pathways that do not require oscillatory Notch1 activation. The robust nature of somitogenesis may reflect the existence of multiple, overlapping, and interacting feedback loops controlling the oscillation of numerous genes in the Notch, Wnt and FGF pathways (Dequeant et al.,

2006). For instance, FGF signaling is thought to be required for oscillatory function of both the Notch and Wnt pathways in the PSM (Wahl et al., 2007), though most observations were confined to primary body formation. Additionally, recent findings have shown that Wnt signaling can rescue cyclic Lfng in the absence of FGF signaling, suggesting that the various signaling pathways synergize with one another to allow oscillations (Aulehla et al., 2008). Our data support the additional hypothesis that oscillation of any individual pathway or component may be more or less important during 49 different stages of somitogenesis. Our finding that Hes7 oscillations recover during

secondary body formation is especially interesting in light of recent findings that Hes7

oscillations are regulated in part by the FGF pathway, and that oscillatory HES7 protein

regulates the expression of FGF pathway components (Niwa et al., 2007). It is clear that

regulated crosstalk among these pathways is important, however, our results suggest that specific interactions may be differentially regulated during primary and secondary body formation.

Wnt activity may play especially important roles in the regulation of posterior somitogenesis. Reductions in Wnt signaling levels can preferentially affect segmentation of the posterior embryo: the Wnt3avt hypomorphic allele develops segmentation defects

in the lumbar, sacral and tail regions, and mutations in Lrp6, encoding a Wnt coreceptor, affect the caudal axial skeleton more severely than anterior skeletal regions (Kokubu et al., 2004; Pinson et al., 2000). These data may indicate that the caudal skeleton is more

sensitive to perturbations in Wnt pathway activity. Conversely, based on our data, Notch

oscillations may play a more important role during the development of the thoracic and

lumbar skeleton. Additionally, we find that Axin2 is not disrupted in LfngΔ FCE1/Δ FCE1

embryos during secondary body formation (Figure 2.11). It will clearly be important to

carefully dissect the interactions among these three pathways to fully clarify the

possibility that the segmentation clock mechanism is differentially regulated during

primary and secondary body formation.

50 2.4.5 R/C patterning of anterior somites may affect ongoing segmentation during

secondary body formation

We find that many aspects of clock function recover in both Lfng-/- and

LfngΔ FCE1/Δ FCE1 embryos after tailbud formation, however, the posterior skeletal

phenotypes of Lfng-/- animals are much more severe than those seen in LfngΔ FCE1/Δ FCE1

embryos. Therefore, we propose that the truncation of posterior skeletal structures in

Lfng-/- animals is caused by the perturbation of R/C patterning in these embryos rather

than the loss of oscillatory Notch activity in the clock. Several lines of evidence suggest

that R/C somite patterning contributes to the continued segmentation of the posterior

embryo. Targeted deletion of Mesp2, which is exclusively expressed in Region II of the

PSM, causes disrupted R/C somite patterning and truncation of the posterior skeleton

(Saga et al., 1997). Further, it appears that the total dosage of MESP activity (comprising

the additive effects of MESP1 and MESP2) in Region II of the PSM is important during

R/C somite patterning. Manipulating the dosage of MESP proteins can both partially

rescue the R/C patterning of the somites and mitigate the caudal truncation of the axial

skeleton (Morimoto et al., 2006). More recent results suggest that in newly developed

Mesp2 knockout alleles, Mesp1 expression is elevated leading to partial rescue of

somitogenesis during secondary body formation (Takahashi et al., 2007). Interestingly, a mutation in MESP2 seen in human spondylocostal dysostosis, also has more severe effects on the thoracic vertebrae than the more caudal skeleton (Whittock et al., 2004b).

One possible explanation for these results is that continued R/C somite patterning is necessary for segmentation to proceed normally during secondary body formation.

51 This would suggest that information transfer in the PSM can occur from the anterior to the tailbud, and that the segmentation of the most caudal embryonic structures may be reliant on proper patterning of more anterior structures. We propose that the expression of Lfng in the anterior PSM and the subsequent amelioration of R/C patterning defects in

LfngΔ FCE1/Δ FCE1 embryos, permits posterior segmentation to proceed relatively normally, preventing the tail truncation seen in Lfng-/- animals, underscoring the potential for the

transfer of information between the anterior and posterior regions of the PSM, at least

during secondary body formation. The expression of Brachyury is not altered in Lfng-/- embryos (Figure 2.11), therefore it remains to be seen what other components of the tailbud may be affected by the loss of R/C patterning.

Thus, the work reported in this chapter uncovers new levels of complexity linking differential regulation of clock function and R/C somite patterning to the long- known but little-understood processes of primary and secondary body formation.

Figure 2.1: Deletion of the FCE1 enhancer. A. The Lfng endogenous locus is shown (boxes signify coding exons and FCE1). The targeting event replaces the 110 bp FCE1 sequence with an EcoRV site. External probes (small boxes) are shown below. Primers used for PCR genotyping of mice are numbered arrows 1-4. B. A representative colony containing the Lfng∆FCE1 allele and a representative mouse genotyping PCR reaction are shown. Arrows: endogenous band. Arrowheads: targeted bands.

+ /∆FCE1 ∆FCE1/ ∆FCE1 Lfng Lfng

a b c d

e f g h

i j k l

Figure 2.2: Deletion of FCE1 from the endogenous locus alters Lfng expression in the posterior PSM. RNA in situ analysis demonstrates cyclic Lfng expression that can be divided into three phases in wild-type embryos at (A) 8.5 dpc (a, n=7; b, n=4; c, n=8), (B) 9.5 dpc (e, n=3; f, n=4; g, n=4), and (C) 10.5 dpc (i, n=4; j, n=5; k, n=5). In LfngΔ FCE1 homozygous mutant embryos, expression is seen only in the anterior PSM (pink bar) at all stages (d, n=4; h, n=4; l, n=11).

+ /∆FCE1 ∆FCE1/ ∆FCE1 Lfng Lfng

a b c d e

Figure 2.3: Lfng expression is localized to the proper somite compartment in Lfng∆FCE1/∆FCE1 embryos. Double RNA in situ hybridizations of Lfng and Mesp2, a marker for the caudal half of somitomere S-1 in Region II of the PSM, are shown (Lfng is purple, Mesp2 is orange). In 10.5 dpc wild-type embryos, there is overlapping expression of Lfng and Mesp2 in Region II of the PSM at all phases of the segmentation clock (a, n=3; b, n=2; c, n=2). In Lfng∆FCE1/∆FCE1 embryos, Lfng and Mesp2 again display overlapping expression patterns (d, n=5), although occasional expression of Lfng is observed rostral to Mesp2 (arrow; e, n=3), possibly reflecting the disorganization of somitomere S-1 as it enters Region II to be patterned.

Courtesy of John Franklin

Figure 2.4: The Lfng ∆FCE1 allele interferes with normal skeletal development during primary body formation. A. Representative phenotypes of Lfng+/-, Lfng∆FCE1/∆FCE1 and Lfng-/- mice. The Lfng∆FCE1/∆FCE1 mouse has a shortened body and kinked tail. B. Skeletal preparations of wild-type (a,b,g), Lfng∆FCE1/∆FCE1 (c,d,h) and Lfng-/- (e,f,i) mice. Ventral (a,c,e) and dorsal (b,d,f) views of the ribs and dorsal views of the lumbar and sacral spine (g-i) are shown. The thoracic regions of Lfng∆FCE1/∆FCE1 (c,d) and Lfng-/- (e,f) mice exhibit rib fusions (arrows) and disorganized vertebrae. In Lfng∆FCE1/∆FCE1 skeletons, vertebral disorganization extends through the lumbar region (bar, h), but normal vertebral condensations are seen in the sacral spine (*). By contrast, vertebral disorganization extends throughout the lumbar (bar) and sacral (*) regions of Lfng-/- skeletons (i), and the tail appears severely truncated.

Figure 2.5: R/C patterning in Lfng∆FCE1/∆FCE1 embryos. A. Whole-mount in situ hybridization for Mesp2, defining the presumptive rostral compartment of the pre-somite. In wild-type (a,c) and Lfng∆FCE1/∆FCE1 (b,d) embryos, a single clear band of Mesp2 expression is seen at both 9.0 (a,b) and 10.5 (c,d). The anterior border of this band is sometimes less defined in Lfng∆FCE1/∆FCE1 embryos (arrows). B. Whole-mount in situ hybridization with a probe against Uncx4.1, which demarcates the caudal half of the somites. At 9.5 and 10.5 dpc, wild-type somites have clear rostral and caudal compartments (a-d). During primary body formation, Lfng∆FCE1/∆FCE1 embryos exhibit some compartmentalization with stronger staining in the caudal region of the somite (e,f, arrowheads), although compartments are frequently irregular (e, bracket). At this stage, little compartmentalization in seen in Lfng-/- embryos (i,j). The mature derivatives of these somites are patterned; clear rostral compartments are observed in Lfng∆FCE1/∆FCE1 embryos in the sclerotome of mature somites in the thoracic region, but compartments may be misshapen or irregularly spaced (arrow, g). Somites in the thoracic region of Lfng-/- embryos exhibit no compartmentalization at this stage (k). During secondary body formation, somites in Lfng∆FCE1/∆FCE1 embryos are of regular size and are correctly patterned (h), while in Lfng-/- embryos at this stage, little to no compartmentalization is observed (bar, l). Anterior is towards the left in all panels.

Figure 2.5

+ / +

Lfng

a b c

FCE1 ∆

Lfng

d e f

a b

Figure 2.6: Somites differentiate properly in Lfng∆FCE1/∆FCE1 embryos. A. Whole- mount in situ analysis of Mox1 mRNA demonstrates a regular pattern of mature somitic derivatives in the thoracic region of wild-type embryos at 10.5 dpc (a) and 9.5 dpc (b). In Lfng∆FCE1/∆FCE1 embryos, somitic derivatives in this region are distinct but irregularly spaced (d, e; arrow, bar). During secondary body formation, Mox1 expression in 10.5 dpc tails is indistinguishable between wild-type (c) and Lfng∆FCE1/∆FCE1 somites (f). B. Whole-mount in situ analysis of Tbx18 mRNA demonstrates that the anterior, lateral sclerotome is specified properly in wild-type (a) and Lfng∆FCE1/∆FCE1 (b) 10.5 dpc embryos. C. Whole-mount in situ analysis of Pax9 mRNA demonstrates that the posterior, ventrolateral sclerotome is specified properly in wild-type (a) and Lfng∆FCE1/∆FCE1 (b) 9.5 dpc embryos. Anterior is to the left.

Courtesy of Kellie VanHorn

Figure 2.7: Notch1 signaling is altered in Lfng∆FCE1/∆FCE1 embryos. Whole-mount immunohistochemistry using an antibody specific for activated Notch1 was performed. A. At 8.5 dpc, dynamic domains of Notch activation are seen in wild-type embryos, with anterior bands and a posterior band of varying width (a,b, n=29). In both Lfng∆FCE1/∆FCE1 (c, n=12) and Lfng-/- (d, n=7) embryos, Notch1 activation is seen ubiquitously throughout the PSM. B. At 10.5 dpc, dynamic Notch1 activation is observed in wild- type embryos, with four distinct phases observed [a-d n=9/38 Phase 1, 8/38 Phase 2, 11/38 Phase 3 and 10/38 Phase 4, as defined in Morimoto et al. (Morimoto et al., 2005)]. By contrast, Lfng∆FCE1/∆FCE1 (e, n=17) and Lfng-/- (f, n=8) embryos exhibit a gradient of Notch1 activation throughout the PSM. Yellow bars indicate the extent of the stained regions.

Figure 2.8: Hes7 transcription is affected in Lfng∆FCE1/∆FCE1 embryos. A. Hes7 expression in 8.5 dpc embryos, using a probe specific for intronic sequences. In wild-type embryos, several distinct patterns of expression are seen with a Hes7 intron probe (a-c, n=20), reflecting cyclic Hes7 transcription. In Lfng∆FCE1/∆FCE1 (d, n=6) or Lfng-/- (e, n=4) embryos, Hes7 mRNA is transcribed ubiquitously throughout the PSM, suggesting that, at this stage, Lfng activity is required for Hes7 oscillation. B. Hes7 expression, as detected with a probe specific for intronic sequences in 10.5 dpc embryos. At 10.5 dpc, Hes7 mRNA expression levels and transcription oscillate in wild-type (a-c, n=13/51 phase 1, 18/51 phase 2, 20/51 phase 3), Lfng∆FCE1/∆FCE1 (d-f, n=13/32 phase 1, 8/32 phase 2, 11/32 phase 3) and Lfng-/- (g-i, n=10/22 phase 1, 5/22 phase 2, 7/22 phase 3) embryos. C. Hes7 RNA expression was examined using a cDNA probe that reveals the steady- state levels of mature Hes7 mRNA. In wild-type 8.5 dpc embryos, several distinct patterns of expression are seen (a-c, n=9), while in Lfng∆FCE1/∆FCE1 (d, n=7) embryos, Hes7 mRNA is found ubiquitously throughout the PSM. At 10.5 dpc, oscillatory expression is seen in both wild-type (e, n=11; f, n=9) and Lfng∆FCE1/∆FCE1 (g, n=8, h, n=9) embryos. D. 10.5 dpc embryos were bisected along the neural tube, and one half was fixed (a,c,e), while the other half was cultured for 1 hour prior to fixation (b,d,f). The Hes7 expression pattern is altered between the fixed and cultured halves of wild-type (a,b), Lfng∆FCE1/∆FCE1 (c,d) and Lfng-/- (e,f) embryos, confirming that Hes7 RNA levels can oscillate in the absence of LFNG activity.

Figure 2.8

Figure 2.9: Oscillatory gene expression is differentially perturbed in Lfng∆FCE1/∆FCE1 embryos. A. Nrarp expression oscillates in wild-type embryos at 8.5 dpc (a,b, n=10), but is stably expressed in Lfng∆FCE1/∆FCE1 embryos (c,d, n=6). At 10.5 dpc, wild-type embryos exhibit two distinct patterns of expression (e, n=7; f, n=8). Both these phases are seen in Lfng∆FCE1/∆FCE1 embryos, but the pattern is more diffuse than that observed in wild-type embryos (g, n=5; h, n=2). B. Cyclic expression of the Dll1 intron probe is observed in both wild-type (a, n=6; b, n=7) and Lfng∆FCE1/∆FCE1 (c, n=6; d, n=4) embryos at 10.5 dpc.

Hes5 Hesr1

a b c d e

a b c d

Figure 2.10: Hes expression is disrupted in Lfng∆FCE1/∆FCE1 embryos. A. In wild- type embryos, Hes5 exhibits cyclic expression (a, n=1; b, n=3; c, n=3) and two phases of Hesr1 expression are seen, with a narrow band in the anterior PSM and either a broad (d, n=4) or narrow (e, n=2) band in the posterior PSM. B. In Lfng∆FCE1/∆FCE1 embryos, Hes5 cycles, but the bands are diffuse and disorganized compared to wild-type embryos (a, n=2; b, n=3; c, n=3). Hesr1 expression is also perturbed in Lfng∆FCE1/∆FCE1 embryos, with a diffuse band of staining extending from the posterior into the anterior PSM that overlies a narrow band of expression in the anterior PSM (d, n=6). Black bars indicate the extent of the stained regions.

A +/+ ΔFCE1 Lfng Lfng

a b c d

+/+ ΔFCE1 −/− B Lfng Lfng Lfng

a b c

Figure 2.11: The Lfng∆FCE1 allele does not prevent cyclic Wnt in the PSM or mesoderm induction in the tail bud. A. Whole-mount in situ hybridization of Axin2 shows varying expression in wild-type and Lfng∆FCE1/∆FCE1 10.5 dpc embryos, with a narrow band in the anterior PSM and either expression around the tail tip (a, n=7; c, n=3) or expression throughout most of the posterior PSM (b, n=8; d, n=6). Pink bars indicate the extent of the stained regions. B. Whole-mount in situ hybridization of Brachyury shows similar expression in the tail-bud of wildtype (a, n=9), Lfng∆FCE1/∆FCE1 (b, n=4), and Lfng-/- (c, n=3) 12.5 dpc embryos.

CHAPTER 3

THE LUNATIC FRINGE 3’UTR CONTRIBUTES TO SHORT RNA HALF-LIFE IN THE PRESOMITIC MESODERM

3.1 Introduction

Several genes exhibit cyclic RNA expression patterns in the PSM, but are not apparently necessary for somitogenesis, for instance, several Hes genes appear to simply

cycle downstream of Notch, but no phenotypes are seen in the knock-outs (Bessho et al.,

2001b; Gessler et al., 2002; Ishibashi et al., 1995; Jouve et al., 2000; Ohtsuka et al.,

1999). This does not appear to be the case with Lfng. As we show in Chapter 2, Lfng expression in the segmentation clock is functionally important for proper somitogenesis.

Not only does Lfng need to be expressed in Region I of the PSM, but its expression must be oscillatory for proper segmentation clock function. For instance, overexpression of

Lfng in mouse embryos leads to abnormal Hes7 and endogenous Lfng expression patterns

(Serth et al., 2003). Similarly in chick embryos where Lfng is overexpressed, it disrupts the cyclic expression patterns of other oscillatory genes, suggesting that the cyclic expression of Lfng RNA in the segmentation clock is important in multiple species (Dale et al., 2003).

66 We became interested in the mechanism by which Lfng RNA is rapidly turned over in the PSM. Researchers have demonstrated that transcription of Lfng oscillates in mice with a periodicity that matches the clock (Cole et al., 2002; Aulehla and Johnson,

1999; Forsberg et al., 1998; Jiang et al., 2000; McGrew et al., 1998). In order to observe cyclic RNA levels, the periodic transcription of Lfng must be coupled with quick RNA turn-over in PSM. Indeed, mathematical models of the segmentation clock highlight the importance of RNA half-life in shaping oscillations. For instance, incorporating Hes1

RNA half-life along with a number of other variables influencing RNA and protein levels allowed researchers to model the 2 hour Hes1 oscillations observed in cells (Momiji and

Monk, 2008). The rapid half-life of Lfng RNA could be achieved either through the regulated, periodic destruction of RNA or through signals that confer a short half-life on

RNA molecules. In either case, sequences that control Lfng RNA stability would be encoded in the RNA.

Our lab has examined the ability of specific Lfng sequences to confer an Lfng-like

RNA half-life on exogenous RNAs by expressing LacZ and GFP transgenes in mice

(Figure 3.1). When lacZ is expressed under control of the Lfng promoter (including the

Lfng 5’UTR), cyclic lacZ RNA is observed in the PSM. Thus, the Lfng promoter and upstream sequences contain the regulatory elements necessary for cyclic gene expression

(Cole et al., 2002). However, the cyclic lacZ RNA observed in the PSM of transgenic embryos could be due to either sequences in the 5’UTR that confer instability on the

RNA or due to lacZ RNA coincidentally having a short half-life, similar to that of Lfng

RNA. To distinguish between these possibilities, a similar transgene was constructed expressing destabilized GFP (dGFP, Clontech) under the control of the Lfng promoter

67 and 5’UTR. In transgenic embryos carrying the dGFP construct, GFP RNA is detected

ubiquitously throughout the PSM and perduring into mature somites (Figure 3.1). This

suggests that the cyclic lacZ RNA observed was due to cyclic transcription coupled with

lacZ RNA having a similar half-life to Lfng, rather than to sequences in the 5’UTR, and that the dGFP RNA half-life is long enough that mRNA accumulation masks the transcriptional oscillation. Often the stability of an mRNA is controlled by sequences

found in the 3’UTR such as AREs (Wilusz and Wilusz, 2004). Therefore, another dGFP

transgene was constructed with the Lfng 3’UTR included. This transgene exhibits cyclic

GFP RNA expression in the PSM (Figure 3.1). This supports the hypothesis that the

3’UTR of Lfng contains sequences sufficient to confer a short RNA half-life on an exogenous mRNA. Other researchers have also found that the 3’UTR of Lfng is

important for conferring a short RNA half-life on exogenous RNAs in tissue culture

(Chen et al., 2005) as well as in chick embryos (Hilgers et al., 2005).

Knowing that the Lfng 3’UTR is important for conferring a short RNA half-life,

we identified sequences within the 3’UTR that might be responsible for this activity.

Identifying these sequences would allow us to examine possible binding factors that

might regulate not only Lfng RNA, but possibly other cyclic clock genes as well. We

hypothesized that these sequences would be evolutionarily conserved among species that

utilize Lfng in their segmentation clocks. We find that there are high levels of sequence

conservation between mouse, human, and chicken in regions of the Lfng 3’UTR (Figure

3.2). These regions may be functionally important sequences for regulating Lfng RNA

half-life. Therefore, we examined the conserved regions of the Lfng 3’UTR for their

ability to promote RNA turn-over.

68 To preliminarily identify regions of the Lfng 3’UTR that may act in rapid RNA

turn-over, we used a luciferase reporter assay to screen for sequences that can confer a

short RNA half-life in cell culture. Deletion analysis of the Lfng 3’UTR allowed us to

localize sequences that may contribute to mRNA stability. Based on these experiments,

we examined the embryonic expression of 3 transgenes containing distinct 3’UTR

sequences, and expressing dGFP under control of the Lfng promoter. These analyses

examined which regions are important for RNA stability in vivo. By examining the

expression patterns of these transgenes and comparing them to endogenous Lfng mRNA

expression patterns, we were able to identify sequences that may be important for

conferring an RNA half-life consistent with the quick Lfng oscillations observed during

somitogenesis.

3.2 Materials and Methods

3.2.1 Transgene Construction

Transgenes encoding destabilized GFP expressed under the control of the Lfng

promoter were constructed. Fragment E, Fragment B, or Fragment B+D were included in

the 3’UTR. Transgenes were injected by the OSUCCC Transgenic/ES Core Facility

following standard protocols. Mice and embryos were identified as carrying the

transgene with PCR as in Chapter 2 using the following primers: FNG325 (5’-

AGAGTTCCTGAAGCGAGAG-3’), FNG237 (5’-CGACATTTTGCAGCACAG-3’),

FNG239 (5’-TTCACCGATGGAGACGAC-3’), and SC157 (5’-

GAACTTCAGGGTCAGCTTGC-3’), which amplify a 500bp control band and a 230bp

GFP band.

69 3.2.2 Luciferase Reporter Assay

Fragments of Lfng 3’UTR were cloned into the pGL3 luciferase vector (Promega) between the stop codon and polyadenylation site. Steady-state levels of luciferase were measured from transfected NIH3T3 tissue culture cells as in Chapter 3. Luciferase values for each construct were normalized to a pSVβgal control and the pGL3 control value was set at 1. Experiments were performed in triplicate.

3.2.3 Whole Mount in situ Analysis

Single and double whole mount in situ hybridization was performed as described in Chapter 2. The dGFP probe contains full length EGFP sequences. For the half-tail in situ hybridization, the tails were cut after fixation, but before incubation with the prehybridization buffer. The half-tails were then transferred to a 96 well plate for the remainder of the protocol.

3.3 Results

3.3.1 Localizing Lfng 3’UTR signals affecting RNA stability

Our lab constructed a number of expression vectors containing different conserved regions of the Lfng 3’UTR inserted into a luciferase reporter and expressed them in NIH3T3 tissue culture cells. We predicted that destabilization of the luciferase

RNA would lead to decreased steady state levels of both luciferase RNA and luciferase activity. Indeed, the full length Lfng 3’UTR and two overlapping Fragments (A and C) significantly reduced steady state luciferase levels 2.5-3 fold compared to the pGL3 control vector (Figure 3.3). The highly conserved Fragment E, contained in the overlap between Fragments A and C, had a smaller, but significant effect by itself, reducing luciferase levels by ~30%. We also examined the 3’UTR without Fragment E by

70 examining the activity of Fragments B and D. Interestingly, Fragments B or D have no

significant effect on luciferase levels by themselves, but are able to significantly reduce

levels 2.5-3 fold when included together as Fragment B+D, suggesting a cooperative

effect. This screen identified Fragment E as a partial modifier and Fragment B+D as

cooperative modifiers of luciferase levels suggesting the Lfng 3’UTR contains modules of sequences regulating RNA turn-over (Figure 3.3 B).

3.3.2 Sections of the Lfng 3’UTR have differential effects on RNA stability in the mouse PSM

We selected Fragment E, Fragment B, and Fragment B+D to make transgenic mice in order to test their effects on RNA stability in vivo. First, these fragments would

allow us to determine the effects of the conserved Fragment E and cooperative Fragment

B+D which our model suggests may have independent effects on RNA stability. In contrast, Fragment B, which lacks sequences contained in Fragment E and the 3’ end of the Lfng 3’UTR did not affect steady state luciferase levels in our assays, but has been suggested to destabilize Lfng RNA in cell culture (Chen et al., 2005). As mentioned earlier, when dGFP alone is expressed from the Lfng promoter, we observe ubiquitous dGFP RNA expression in the PSM, but when the Lfng 3’UTR is included in the transgene, dGFP RNA is destabilized, revealing the underlying cyclic transcription

(Figure 3.1). We predicted that transgenes including various sub-sections of the 3’UTR might exhibit one of three different expression patterns. If the transgene is expressed ubiquitously, this would indicate that the included Lfng 3’UTR sequences are not sufficient to destabilize an exogenous RNA. On the other hand, if the transgene is able to completely recapitulate the endogenous, cyclic Lfng expression pattern, then the

71 incorporated sequences are sufficient to destabilize an exogenous RNA. However, given

the modular nature we predict for Lfng 3’UTR regulation, we might expect some

sequences (for instance, Fragment E) to have partial effects on RNA stability. In this case, transgenes incorporating these sequences might exhibit oscillatory expression patterns due to the partially decreased RNA half-life, however, we would expect that endogenous Lfng expression pattern would not be completely recapitulated. Specifically, the dGFP transgene might be expressed as broad or fuzzy bands compared to tight bands

of cyclic Lfng RNA in the PSM.

3.3.2.1 Fragment E has a partial effect on RNA stability

To determine if Fragment E can destabilize dGFP mRNA in the PSM, whole

mount in situ hybridization was performed probing for dGFP RNA in embryos

expressing an dGFP/Fragment E transgene (Figure 3.4 B). Cyclic expression of GFP is observed in the PSM of transgenic embryos (Phase I, 15%; Phase II, 62%; Phase III,

23%), however a second band of GFP is often observed in the anterior PSM during Phase

III suggesting that the transgene is not exactly recapitulating Lfng RNA expression

patterns. To further examine this question we directly compared dGFP expression driven

by the transgene with endogenous Lfng expression. Double-label in situ hybridization

was performed on transgenic embryos, directly comparing GFP RNA patterns with Lfng

expression (Figure 3.5). We find that dGFP/Fragment E transcripts persist not only in the

anterior PSM during Phase III, but also in the posterior PSM, in groups of cells where

Lfng RNA is no longer detected. We are currently analyzing a second dGFP/Fragment E

transgenic line and preliminary data suggests that dGFP transcripts persist both areas of

the PSM similar to the line shown here. These data suggest that Fragment E has a partial

72 effect on RNA stability, but is not fully sufficient to confer an Lfng-like half-life on

exogenous RNA molecules.

3.3.2.2 Sequences in Fragment B+D confer an Lfng-like RNA half-life on GFP

dGFP transgenes containing Fragment B+D exhibit cyclic expression patterns of the dGFP/FragmentB+D transcript in the PSM. We observe that 23% of the embryos are in Phase I, 47% in Phase II, and 30% in Phase III (Figure 3.4 B). These ratios are very similar to the distribution we observe for Lfng expression in wild-type embryos with 37% in Phase I, 40% in Phase II, and 23% in Phase III (Figure 3.4 A). Similar to the Fragment

E transgenic line, occasionally an additional band of GFP is observed in the anterior PSM of Fragment B+D embryos. To determine if this additional band corresponds to Lfng expression, we performed double label whole mount in situ hybridization comparing dGFP/FragmentB+D transcript and endogenous Lfng (Figure 3.5). Transgenes containing Fragment B+D exhibit an expression pattern closely matching endogenous

Lfng. GFP is occasionally observed persisting in the PSM as the extra band in the anterior PSM in Phase III or as a longer streak in the posterior PSM during Phase II. To further compare the RNA expression patterns in Fragment B+D transgenic embryos, half- tail in situ hybridization experiments were performed where one half of the tail was probed for GFP while the other half was probed for endogenous Lfng. Most tails exhibited matching expression patterns, however an extra GFP band was occasionally observed in the anterior PSM that did not correspond to endogenous Lfng expression

(Figure 3.6). The persistence of GFP RNA in Fragment B+D transgenic embryos does not occur as often as in Fragment E embryos, where GFP is observed persisting in multiple areas of the PSM during all phases. This in vivo data matches the in vitro

73 findings that Fragment E by itself is not as effective as Fragment B+D at promoting RNA

instability.

3.3.2.3 Fragment B appears to have an effect on RNA stability in the PSM

A transgene containing Fragment B displays cyclic GFP expression in the PSM

(Figure 3.4 B), which matches endogenous Lfng expression patterns fairly well (Figure

3.5). There do appear to be fuzzy bands of GFP RNA in the anterior PSM compared to

tight bands of endogenous Lfng suggesting Fragment B may not be totally sufficient for

conferring an Lfng-like RNA half-life. However, more embryos and a more strongly expressing transgenic line will be needed for a complete analysis of the effects of

Fragment B. Interestingly, based on the luciferase data, we did not predict that Fragment

B would be sufficient to drive RNA turn-over in the PSM, however Fragment B does

appear to have an effect on GFP RNA in vivo allowing for visualization of cyclic

expression patterns.

3.4 Discussion

Having confirmed that Lfng plays an important role in segmentation clock

function during primary body formation, we hypothesized that the cyclic activity of Lfng

in the PSM is achieved through multiple mechanisms including transcriptional and post-

transcriptional control. The importance of post-transcriptional levels of control is

highlighted by Momiji et al. findings that the transcriptional time delay and subsequent stability of an mRNA involved in a feedback loop are important variables to consider

when modeling the segmentation clock. By including these variables, they were able to model large amplitudes in their oscillations similar to the amplitudes observed with experimental data (Momiji and Monk, 2008). To examine the control of Lfng mRNA

74 stability, we identified several conserved regions of the Lfng 3’UTR that may play important roles in regulating RNA stability. Using a luciferase reporter assay, we suggested that the Lfng 3’UTR may contain modular sequences that coordinate to regulate RNA turn-over in vitro (Figure 3.2). However, as this assay quantifies luciferase protein activity, it does not distinguish between effects that 3’UTR sequences might have on RNA stability or on the efficiency of protein translation. Further, mRNA stability in cell culture may not accurately reflect findings in vivo. To better address this question, we next examined the function of the Lfng 3’UTR in vivo.

We examined the functions of conserved regions of the Lfng 3’UTR using transgenic mice. We find that Fragment E partially contributes to mRNA turnover both in vitro and in vivo. When incorporated in a dGFP transgene, Fragment E destabilizes dGFP mRNA sufficiently to reveal cyclic RNA transcription in the PSM. However, all three oscillatory phases exhibit GFP RNA persisting in the PSM compared to endogenous

Lfng (Figure 3.5) suggesting Fragment E is not sufficient to confer an Lfng-like half-life to dGFP mRNA. Transgenes incorporating Fragment B+D exhibit cyclic expression patterns highly similar to endogenous Lfng, with only occasional instances of persistent

GFP expression. This suggests that these sequences may coordinate to confer a short mRNA half-life sufficient for Lfng to quickly cycle in the PSM. Preliminary data suggests that the transgene incorporating Fragment B appears to confer an Lfng-like RNA half-life in vivo, however analyzing more embryos will help clarify these results.

Interestingly, Fragment B was not able to promote RNA turn-over in vitro. This may suggest that cell culture systems do not accurately recapitulate the situation in the PSM.

Overall, there appears to be redundant, modular control of Lfng mRNA stability, which

75 may be important for the tight regulation of Lfng mRNA levels during somitogenesis.

A number of caveats should be considered when moving forward with this

analysis. First, the expression levels driven by different transgenic lines varies. The

transgenic line incorporating Fragment B expresses at fairly low levels in the PSM. In

contrast, the transgenic line shown incorporating Fragment B+D expresses at high levels,

sometimes appearing in regions of the PSM at much higher levels than endogenous Lfng.

It is therefore important to examine multiple lines for each distinct transgene in order to

control for over/under-expression issues as well as to verify the results presented here.

We have begun analysis with a second transgenic line incorporating Fragment B+D that

expresses at lower levels than the line examined here, which will be useful in comparing

the expression pattern of dGFP/FragmentB+D transcripts to endogenous Lfng transcripts.

Additionally, to examine whether any of the transgenes’ dosage is influencing the dGFP expression patterns observed, we would like to generate homozygote transgenic mice.

These mice would express the transgenes at higher levels and allow us to determine, for instance if the tight expression patterns of dGFP/FragmentB+D are maintained or become streaky with increased expression.

In the future, it will be interesting to examine the mechanisms by which the conserved sequences of the 3’UTR are used to confer a short mRNA half-life, for instance identifying what factors bind to these regions and how they are cyclically activated themselves. One possibility is that microRNAs, which can bind to the 3’UTR of mRNAs and either promote their degradation or prevent their translation (Carthew and

Sontheimer, 2009), may bind to conserved regions in the Lfng 3’UTR. Overall, the Lfng

3’UTR appears to have complex, modular activity in its regulation of Lfng RNA, a

76 phenomenon which may be conserved in other clock genes requiring rapid RNA turn- over.

A Endogenous Lfng locus

5’UTR LacZ

LacZ Lfng

a b

5’UTR GFP

dGFP

GFP 5’UTR 3’UTR Lfng dGFP

d e

Courtesy of Susan Cole

Figure 3.1: Lfng 3’UTR is necessary for RNA turn-over in the PSM. A. Schematic of the Lfng locus showing the endogenous promoter (arrow), 5’UTR and 3’UTR (hatched boxes), and coding exons (blue boxes). B. Transgenes were constructed as diagramed and single or double RNA in situ analysis was performed on transgenic mice as specified on the right. In the first transgene with LacZ under the control of Lfng promoter, LacZ RNA is observed to cycle in the PSM (compare a and b) similar to endogenous Lfng. However, in the second transgene with dGFP under the control of Lfng promoter, GFP RNA is detected throughout the PSM (c). When the 3’UTR of Lfng is included in the third transgene, GFP RNA now cycles in the PSM with endogenous Lfng (compare d and e).

Mouse

vs. Human

Mouse vs. Chicken

Human vs. Chicken

Conserved in human Conserved in chicken

Figure 3.2: Conservation of Lfng 3’UTR. An mVista comparison shows several conserved regions in the Lfng 3’ UTR among mouse, human, and chicken (Frazer et al., 2004). Diagramed underneath are the identified sections conserved among species (hatched boxes, conserved from mouse to human; green boxes, conserved from mouse to chicken).

79 0 0.2 0.4 0.6 0.8 1

A pGL3 control pGL

** Full Full ** ** A A B B

C C

D D * E E B+D B+D

B Conserved in human Conserved in chicken

Cooperative Partial Cooperative

effect effect effect

Courtesy of Kellie VanHorn

Figure 3.3: Conserved regions of Lfng 3’UTR contribute to RNA turn-over in vitro. A. A deletion analysis was designed to determine which conserved regions of the 3’ UTR contribute to RNA instability. Blue striped boxes represent sequences conserved from mice to human, green boxes represent sequences conserved from mice to chicken. Dotted line indicates regions that were deleted from the transgene. Each fragment was cloned into a luciferase reporter vector and steady-state levels of luciferase were measured in tissue culture cells. The graph represents normalized luciferase values for each construct. Experiments were performed in triplicate and statistically significant results are marked with asterisks. The full length 3’ UTR and several overlapping fragments (A, C, B+D) produced a 2.5-3 fold reduction in luciferase levels (**p<0.005). The highly conserved Fragment E (base pairs 550-811) reduced levels by about one third (*p<0.03). B. A model representing a possible mechanism for Lfng RNA turn-over, in which different elements of the 3’ UTR interact for a cooperative effect on RNA half-life: Fragment E has a partial effect on its own, and Fragments B+D do not have an effect on their own, but are able to cooperate together or with Fragment E to drive RNA turn-over.

Figure 3.4: Sections of Lfng 3’UTR are sufficient to reveal cyclic RNA expression in the PSM. A. Whole mount in situ hybridization for endogenous Lfng was performed on wild-type embryos and reveals cyclic Lfng expression (Phase I, n=22; Phase II, n=24; Phase III, n=14). B. Whole mount in situ hybridization for GFP RNA was performed on transgenic embryos. All three phases of cyclic gene expression were detected for Fragment E (Phase I, n=2; Phase II, n=8; Phase III, n=3), Fragment B (Phase I, n=1; Phase II, n=1; Phase III, n=4), and Fragment B+D (Phase I, n=7; Phase II, n=14; Phase III, n=9). Occasionally in Phase III, an extra band of GFP was detected in the anterior PSM (red arrows) suggesting the possible perdurance of GFP RNA in this region (d, n=2; k, n=4).

Phase I Phase II Phase III Phase III A with extra anterior band Wild-type

Full length N/A 3’UTR a b c

Fragment E

dGFP

a b c d

Fragment B

dGFP

e f g

Fragment B+D

dGFP

h i j k

Figure 3.4

Phase I Phase II Phase III

Fragment E

Lfng GFP

a b c

{ Fragment B { GFP Lfng

d e f

Fragment B+D

Lfng GFP

g h i j

Figure 3.5: Different sections of the Lfng 3’UTR are differentially able to promote an Lfng-like RNA expression pattern in the PSM. Double whole-mount in situ was performed for GFP and Lfng. For Fragment E and Fragment B+D, Lfng is purple, GFP is orange, while for Fragment B, GFP is purple and Lfng is orange. Areas of GFP expression not overlapping with Lfng expression are marked with red arrows. Fragment E reveals cyclic expression similar to Lfng (Phase I, n=3; Phase II, n=2; Phase III, n=2), however GFP RNA is often observed persisting in regions of the PSM where Lfng is no longer detected. Fragment B reveals cyclic expression similar to Lfng (Phase I, n=3; Phase II, n=3; Phase III, n=3), although GFP bands appear fuzzy compared to Lfng (brackets). Fragment B+D reveals cyclic expression that closely matches Lfng (Phase I, n=4; Phase II, n=5; Phase III, n=6), however GFP can occasionally be observed persisting longer than endogenous Lfng. Additionally, GFP is observed reappearing in the tip of the PSM before Lfng can be detected there (j, arrowhead) in approximately half of the tails in Phase III, most likely due to the high expression level of the B+D transgene.

GFP Lfng

Phase I

a b

Phase II

c d

Phase III

e f

Phase III

g h

Figure 3.6: Fragment B+D reveals a cyclic expression pattern similar to Lfng. Half- tail in situ hybridization for GFP (a,c,e,g) and Lfng (b,d,f,h) was performed on embryos with the Fragment B+D transgene. Examples of all three phases of cyclic gene expression were observed (Phase I, n=5; Phase II, n=3; Phase III, n=7). Occasionally in Phase III, a band of GFP was observed in the anterior PSM (red arrow) that did not correspond to Lfng expression (compare g and h, n=3).

CHAPTER 43

LUNATIC FRINGE PROTEIN PROCESSING BY PROPROTEIN CONVERTASES CONTRIBUTES TO THE SHORT PROTEIN HALF-LIFE IN THE SEGMENTATION CLOCK

4.1 Introduction

As discussed in Chapter 1, Notch signaling can be fine-tuned through

modification of Notch receptors by fringe family proteins, including Lunatic, Manic, and

Radical fringe (Lfng, Mfng and Rfng) (Cohen et al., 1997; Johnston et al., 1997). Fringe

genes encode glycosyltransferases that modulate Notch signaling by modification of the

Notch extracellular domain (Bruckner et al., 2000; Moloney et al., 2000). FRINGE

proteins transfer N-acetylglucosamine to fucose on extracellular EGF repeats of NOTCH

receptors, but it is not well understood how sugar addition alters the interactions between

Notch and its ligands (Stanley, 2007). However, it is known that the different fringe

proteins have distinct effects on Notch signaling, allowing for context-dependent fine-

tuning of Notch signaling (for example Hicks et al., 2000; Shimizu et al., 2001; Yang et

al., 2005).

3 Portions of this chapter were originally published in Biochimica et Biophysica Acta: Molecular Cell Research, Volume 1783, Pages 2384-2390. 85 Somitogenesis requires the cyclic modulation of NOTCH1 by LFNG during

primary body formation (Rida et al., 2004; Shifley et al., 2008; Weinmaster and Kintner,

2003). For Lfng to play a role in the segmentation clock, not only must its mRNA levels

cycle, but its protein activity levels must also oscillate with a short period (two hours in

the mouse). During chick segmentation, LFNG protein levels, as well as Lfng transcript

levels, oscillate with a period that matches somite formation, linking LFNG protein

activity to the clock (Dale et al., 2003). As discussed in Chapter 3, cyclic Lfng

expression is regulated transcriptionally (Cole et al., 2002; Morales et al., 2002) and the

Lfng 3’UTR contains sequences important for mRNA turn-over. However, little is

known about the post-translational mechanisms that contribute to the rapid periodicity of

LFNG function in the segmentation clock.

The functions of FRINGE proteins within the Notch pathway are cell autonomous

(Hicks et al., 2000; Irvine and Wieschaus, 1994; Moloney et al., 2000; Munro and

Freeman, 2000; Panin et al., 1997), but interestingly, both Drosophila FRINGE and

LFNG protein are cleaved following a conserved dibasic site, and are rapidly secreted into the media when expressed in tissue culture cells (Johnston et al., 1997) (Figure 4.1).

This suggests the possibility that LFNG secretion could provide a mechanism to terminate LFNG function in the Notch pathway, facilitating the rapid oscillations of

LFNG activity in the segmentation clock.

The sequence of the identified LFNG processing site (RARR) suggests that the

protein may be cleaved by members of the subtilisin-like proprotein convertase family

(SPCs) (Johnston et al., 1997). This family has nine known members that play diverse

roles in the processing and maturation of many substrates including proteases, hormones 86 and growth factors (Taylor et al., 2003). Seven of these proteins (furin/Pcsk3/SPC1,

SPC2/Pcsk2/PC2, SPC3/Pcsk1/PC1/3, SPC4/Pcsk6/PACE4, SPC5/Pcsk4/PC4,

SPC6/Pcsk5/PC6, and SPC7/Pcsk7) process their substrates at multibasic sites with the

motif (K/R)XX(K/R), and in many cases this protein processing is required for activation

of the substrate.

To better understand the post-translational regulation of LFNG activity, we

examined the roles of LFNG processing by SPC convertases. We find that LFNG

processing is promoted by specific pro-protein convertases including furin and SPC6.

Mutations that alter LFNG processing increase its intracellular half-life without

preventing its secretion. These mutations do not affect the specificity of LFNG function

in the Notch pathway, thus regulation of protein half-life affects the duration of LFNG

activity without altering its function. Targeted expression of a Golgi-resident LFNG

protein in mice results in severe somitogenesis defects suggesting that the cleavage and

clearance of the protein is necessary for proper segmentation clock function.

4.2 Materials and Methods

4.2.1 LFNG mutants

Vectors encoding AP-tagged mouse Lfng, Mfng or Rfng have been described

(Johnston et al., 1997). The vectors utilized here were created by removing sequences

encoding the AP-tagged fringe protein from the described vectors with HindIII and XbaI

and ligating them into the HindIII and XbaI sites of pcDNA3 (Invitrogen). To create HA

tagged vectors the fringe coding sequences were excised from AP tag vectors using NotI

and XbaI, and transferred into a pcDNA3 based vector containing a C-terminal HA tag.

Rfng or Mfng N-terminal sequences were amplified and the EagI to AatII fragment, 87 RFNG aa 1-59 or MFNG aa 1-54, replaced LFNG aa 1-112 in R/LFNG and M/LFNG

respectively. LFNGm1 and LFNGm2 were created by 2 step PCR based mutagenesis

(Primer sequences in Table 4.1).

4.2.2 Alkaline phosphatase assays

4x104 NIH3T3 cells (grown in DMEM supplemented with 10% FBS, 50mM

glutamine) were plated in 24-well plates and co-transfected 24 hours later with 800ng of

AP-fringe plasmid and 200ng of pSVβgal (Promega) using Lipofectamine 2000

(Invitrogen). After 24 hours, media was collected and cells were lysed with 100µl

Passive Lysis Buffer (Promega). 50µl of the cellular extracts or 50µl of heat inactivated

culture were mixed with 50µl of AP Assay Reagent A (GenHunter) to determine AP

activity following manufacturers instructions. AP activity was calculated as

(OD405*54)/(Reaction time*Sample Volume) minus the background AP activity of

pCDNA3 control and was normalized to β-gal activity levels as a control for transfection

efficiency. 30μl of the cell extracts were subjected to a β-gal assay as described

(Sambrook, 2001) using the substrate ONPG. β-gal values were measured at 420nm. For

each experiment the percent of AP activity in the media and in the cellular fraction were

calculated.

4.2.3 Immunoflourescence

Cells were plated on glass cover slips, transfected as described above and fixed in

8%PFA. Coverslips were incubated with anti-AP antibody (Fitzgerald Industries, 1:100)

and anti-GM130 antibody (BD Biosciences, 1:200). Secondary Alexaflour antibodies

(594 anti-rabbit and 488 anti-mouse, Invitrogen) were diluted 1:1000. Cells were

counterstained with Hoechst dye. Coverslips were mounted with Citifluor and examined 88 with an Olympus 1X81 microscope.

4.2.4 Notch1 signaling assay

An established Notch1 signaling assay was utilized to assess the effects of fringe

proteins on Jagged1 induced signaling (Hicks et al., 2000). NIH3T3 cells were plated as

described above and transfected with 100ng of pBOSrNotch1 (Nofziger et al., 1999),

100ng of AP-tagged fringe expression vector or empty APtag4 expression vector, 200ng

of a CBF1-luciferase reporter construct (Hsieh et al., 1997), and 200ng pSVβGal for

normalization of transfection efficiency. After 16 hours, the cells were co-cultured for 24

hours with 1.24x106 control L-cells or L-cells stably expressing Jagged1 (Lindsell et al.,

1995). 20µl of cell lysates were analyzed by luciferase assay (Promega). Luciferase values were normalized to βgal expression measured at 420nm. Notch-induced activation

of CBF1 is expressed as a ratio of normalized luciferase values induced by the Jagged-

expressing cells compared to that obtained with parental L-cells.

4.2.5 Western blot analysis

NIH3T3 cells were plated and transfected as above with expression vectors

encoding HA-tagged fringe proteins, and 6 h after transfection the media was changed to

DMEM+2% FBS. Expression vectors encoding α1PDX protease inhibitor (Jean et al.,

1998), Spc1/furin, Spc4, or Spc7 (Constam and Robertson, 1999), Spc6a or Spc6b (from

D. Constam) were co-transfected as indicated. For Figure 4.2A, plasmid amounts were:

350 ng LFNG, α1PDX as indicated, and pcDNA3 to bring up total DNA to 900 ng. For

Figure 4.2B plasmid amounts were: 300 ng LFNG, 100 ng α1PDX, 500 ng SPC vector as

indicated and pcDNA3 to bring up total DNA to 900 ng. For Figure 4.3C a total of 1000

ng of DNA was transfected, either fringe expression vector alone or equal amounts of 89 fringe expression vector, α1PDX and/or SPC6A expression vector. After incubating cells

in DMEM+2% FBS for 24 h, 500 μl of cell media was concentrated to ~100 μl with

Microcon columns (Millipore). 11.5 μl of concentrated media were mixed with 2×

Laemmli loading buffer, run on a 12% polyacrylamide gel and transferred to an

Immobilon membrane (Millipore). The membrane was incubated with an anti-HA

antibody (HA-7, 1:1000, Sigma-Aldrich) and analyzed using the ECL System following

the manufacturer's protocols (GE Healthcare).

4.2.6 Cycloheximide treatment

2.6x105 NIH3T3 cells were plated in a 6 well plate and transfected with 300ng

(LFNG, LFNGm1) or 100ng (LFNGm1/2, R/LFNG) HA-tagged expression vectors using

Lipofectamine 2000 (Invitrogen) for low expression levels. After 6 hours, transfected

cells were split and 1.25x105 cells were plated into five wells of a 24 well plate, ensuring

that all samples in the timecourse were transfected with equivalent efficiency. After 16 hours, cells were incubated in media with 20µg/ml cycloheximide (Sigma). Cell extracts were collected at 20 minute intervals lysed directly in 100µl loading buffer, 23µl of each time point was analyzed by Western blot utilizing Alexafluor anti-mouse 680 secondary antibody (1:20,000, Invitrogen). Blots were imaged and quantitated on a Li-Cor Odyssey.

Blots were re-probed using a monoclonal anti-tubulin antibody (1:1000, Sigma).

Exponential trend lines were fitted to data points to calculate protein half-lives.

Experiments were repeated at least six times and outliers >2xIQR from the median were excluded.

90 4.2.7 Whole Mount in situ hybridization

Wild-type embryos were collected from timed pregnancies with noon of the day

of plug identification designated at 0.5 d.p.c. Double RNA in situ hybridization using

digoxigen and fluorescein-labeled probes was performed as described (Cole et al., 2002).

Probes: Mesp2 (Saga et al., 1997), Spc6 (Constam et al., 1996), which recognizes both

Spc6A and Spc6b spliceforms, and Uncx4.1 (Mansouri et al., 1997).

4.2.8 Targeted replacement of LFNG pre/pro region

Rfng pre sequence was amplified with primers (Table 4.1) cut with EagI and

AatII, and inserted into the Lfng locus from the NotI to AatII site in exon 1. The final

allele replaces the pre/pro region and 28aa of the mature region of endogenous Lfng with

the signal sequence and 21aa of Rfng fused to exon1 of Lfng. The fusion is at aa 113 of

LFNG, the first conserved amino acid in the fringe family. The targeting vector 5’ arm

extended from the 5’XhoI site in Lfng 5’ flank to the HindIII site in intron 1 and included a floxed cassette containing a neo /testis cre cassette (Bunting et al., 1999). 3’ flanking sequences extended from the HindIII site to the XhoI site in intron 1. Linearized vector

was electroporated into TC1 cells (Deng et al., 1996) and G418 resistant colonies were

screened by Southern Blot analysis using 3’ flanking probes. Two independent ES cell

lines were injected by the Nationwide Children’s Hospital Transgenic and ES Core

Facility, one of which transmitted through the germline. Mice were maintained on a

mixed 129/SvxC57BL/6J background. Mice were maintained in an SPF facility under the

care of the Ohio State University ILACUC.

91 4.2.9 Genotyping

Genomic DNA was prepared from tail clips via proteinase K saltout or from yolk sac fragments via the HOTSHOT procedure (Truett et al., 2000). Animals were genotyped by PCR. LfngRL primers SC286 (5’-TTGGGTCTATCTGGGAAACG-3’) and

SC287 (5’- GCGACTCATCCAGACACAGA-3’) produce a 149 bp wildtype band and a

250 bp mutant band.

4.2.10 Skeletal preparations

Skeletal preparations of neonates or 18.5 d.p.c embryos were performed

essentially as described (Kessel and Gruss, 1991).

4.3 Results

4.3.1 SPC1/furin, SPC6A and SPC6B promote LFNG processing

LFNG cleavage occurs after a conserved RARR sequence, a consensus

recognition site for the SPC family of proprotein convertases (Figure 4.1) (Johnston et al.,

1997). To examine whether LFNG processing is reliant on specific SPC proteins, we

assessed whether the SPC inhibitor α1 Antitrypsin Portland (α1PDX) (Jean et al., 1998)

could inhibit this processing. HA-tagged mouse LFNG (LFNG) was co-expressed with

increasing amounts of α1PDX (Figure 4.2 A). In the absence of α1PDX, a 35 kDa LFNG

fragment is detected in the media, consistent with the predicted size of the fully

processed, mature fragment (34.1 kDa). The mature fragment is reduced as α1PDX

levels increase, and we observe a corresponding increase in the release of a 43.6 kDa

fragment, which is the predicted size for full-length, unprocessed LFNG protein (43

kDa). These data indicate that inhibition of SPC proteins interferes with LFNG processing at the conserved dibasic cleavage site, but suggest that proteolytic processing 92 of LFNG is not required for its secretion.

α1PDX is reported as a specific inhibitor of furin and SPC6 (Jean et al., 1998),

suggesting that only a subset of SPC family members may efficiently process LFNG.

LFNG was expressed along with intermediate levels of α1PDX expression vector to

inhibit endogenous cleavage. Co-transfection of expression vectors encoding different

SPC family members assessed which convertases could efficiently process LFNG

protein. Expression of SPC1/furin, SPC6A, or SPC6B results in the recovery of the processed LFNG fragment. In contrast, expression of either SPC4 or SPC7 results in limited LFNG processing (Figure 4.2 B). These results confirm that furin, SPC6A and

SPC6B are able to efficiently process LFNG protein, and indicate specificity among SPC family members in their recognition of LFNG as a substrate.

4.3.2 N-terminal sequences regulate the secretory behavior of fringe family proteins

Golgi retention of glycosyltransferases remains poorly understood, but in many glycosyltransferases, sequences at the N-terminus are important in protein localization and/or secretion (Young, 2004). The mammalian fringe proteins exhibit distinct Golgi retention and secretion behaviors when expressed in tissue culture cells, and the regulation of these different behaviors is unknown. LFNG is rapidly secreted as a processed, mature fragment. MFNG is secreted more slowly, while RFNG is a Golgi resident protein (Johnston et al., 1997). We assessed the secretory behavior of AP-tagged mouse fringe proteins by tracking the steady-state levels of AP activity in the cellular extract and media fractions of tissue culture cells (Figure 4.3). LFNG protein is observed primarily in the media fraction, reflecting its rapid secretion. In contrast, RFNG remains confined to the cellular fraction, while MFNG protein is found in both fractions (Figure 93 4.3 B). To localize the protein sequences regulating the secretory behavior of the fringe family proteins, chimeric fringe proteins were engineered, replacing the N-terminus of

LFNG (including the dibasic cleavage site) with the N-terminal domain of either RFNG

(R/LFNG) or MFNG (M/LFNG) (Figure 4.3 A). In both these chimeric proteins, we find that secretory behavior is controlled by the protein sequences found at the N-terminus.

Like RFNG, R/LFNG is detected mostly in cell extracts, thus the RFNG N-terminus is sufficient to confer Golgi retention on the LFNG mature fragment. The M/LFNG protein also mirrors the steady-state levels of MFNG, being detected in both the cell extracts and media at similar levels to MFNG (Figure 4.3 B). Thus, the rapid secretion of LFNG protein relies on sequences found at the N-terminus of the protein, and the N-terminal sequences of other FRINGE proteins are sufficient to properly regulate Golgi retention and or secretion.

4.3.3 SPC family proteases recognize two dibasic cleavage sites in LFNG, but protein processing is not required for secretion

To examine the impacts of the cleavage site on mouse LFNG protein secretion, the conserved RARR cleavage site was mutated to AAAA (LFNGm1), and protein secretion and processing were examined (Figure 4.4). Like LFNG, LFNGm1 is found largely in the media fraction, indicating that protein processing at the primary cleavage site is not required for protein secretion (Figure 4.4 B). Expression of HA-tagged LFNG protein results in the secretion of the 35 kDa, fully processed fragment. Longer exposure reveals small amounts of a 43.6 kDa band, presumably corresponding to the full length

LFNG protein, and an intermediate 40.4 kDa band. Co-expression of α1PDX reduces the secretion of the fully processed 35 kDa band, and increases the amounts of the 40.4 and

94 43.6 kDa bands, while overexpression of SPC6 results in the secretion of only the fully

processed LFNG mature fragment (Figure 4.4 C). Mutation of the major processing site

(LFNGm1) leads to the loss of the 35kDa band, confirming that the RARR sequence is

required for processing at this site. LFNGm1 protein is secreted as a predominant 40.4

kDa band and co-expression of α1PDX causes a reduction in this band and an increase of

the full length 43.6 kDa fragment. Overexpression of SPC6 results in the secretion of

only the 40.4 kDa band, indicating that this fragment arises from the processing of full

length LFNG by an SPC proconvertase at a site N-terminal to the previously described

site (Figure 4.4 C). We hypothesized that this fragment might result from cleavage of

LFNG after the dibasic RGRR site found at amino acid 40 and mutated that sequence to

AAAA either by itself (LFNGm2) or in combination with the RARR to AAAA mutation

described above (LFNGm1/2) (Figure 4.4 A). Both LFNGm2 and LFNGm1/2 are found

predominantly in the media fraction of transfected cells (Figure 4.4 B). As predicted, the

RGRR mutation (LFNGm2) causes the loss of the intermediate 40.4 kDa fragment,

confirming that this band results from LFNG processing after the RGRR sequence.

LFNGm1/2 is found in the media only as a full length 43.6 kDa band, confirming that

protein processing by SPC proprotein convertases is not required for the release of LFNG protein from tissue culture cells (Figure 4.4 C). Together these data indicate that LFNG is cleaved by SPC proprotein convertases at two sites in the protein. The originally described RARR site appears to constitute the primary cleavage site, but the RGRR site can be utilized. However, neither of these processing events is necessary for the secretion of LFNG from tissue culture cells. Regardless of their effects on protein secretion, all LFNG variant proteins localize to the Golgi, as previously described (Figure

95 4.5) (Hicks et al., 2000).

4.3.4 Mutation of SPC processing sites increases the intracellular half-life of the

LFNG protein

We find that LFNG processing is not a prerequisite for secretion, however, mutations that interfere with LFNG processing would affect the protein's intracellular

half-life if the mature, processed peptide is secreted more efficiently than the full-length

protein. To test this idea, we measured the intracellular half-life of mouse LFNG and of

LFNG mutant proteins. Cells expressing LFNG, LFNGm1, LFNGm1/2, or R/LFNG were

treated with cycloheximide to inhibit protein translation, and the amount of protein in the

cellular fraction was quantified over time (Figure 4.6). Under these conditions, we

calculate an intracellular half-life for LFNG of 70 minutes, consistent with rapid turnover

in the segmentation clock. Mutation of the primary or both the primary and secondary

SPC processing sites causes an increase in intracellular half-life, with both LFNGm1 and

LFNGm1/2 having calculated half-lives of 97 minutes. Tethering of the LFNG protein in

the Golgi further increases the intracellular half-life with a calculated half-life of 126

minutes for the R/LFNG chimeric protein. These findings suggest that cleavage of

LFNG by SPC family convertases may influence the duration of its activity by

modulating the protein's intracellular half-life.

4.3.5 Alterations of LFNG intracellular half-life do not affect the specificity of its

function in the Notch signaling pathway

Different fringe family proteins exhibit distinct effects on Notch signaling

depending on the fringe and ligand involved (Yang et al., 2005). It is likely that the

specificity of fringe activity maps to the catalytic domain of the protein, however, it is

96 possible that mutations that alter LFNG processing and secretion could alter its function in the Notch pathway by changing the secretory behavior of the protein in question (i.e.,

R/LFNG could exhibit the enzymatic activity of RFNG rather than LFNG). To address

this question, we assessed whether mutations that affect the secretory behavior of mouse

LFNG protein affect the specificity of its activity in the Notch pathway using an

established Notch1 signaling assay. As previously reported, we find that expression of

LFNG in the signal receiving cell inhibits the ability of JAGGED1 to activate signaling

through NOTCH1, while expression of RFNG enhances JAGGED1-induced signaling

(Figure 4.6 and reference Yang et al., 2005). We further find that LFNGm1, LFNGm2,

LFNGm1/2, and R/LFNG all inhibit JAGGED1-induced signaling to a similar extent as

observed for LFNG, thus these mutations alter LFNG processing and half-life without

appearing to change its function in the Notch pathway (Figure 4.7). These results predict

that mutations affecting protein processing will affect the duration, but not the nature of

LFNG activity.

4.3.6 The expression pattern of Spc6 suggests a role in clearance of LFNG from maturing somites

Lfng expression during embryonic segmentation is complex. In the posterior

PSM, Lfng expression is cyclic, and linked to the segmentation clock. In the anterior

PSM, Lfng expression is restricted to the anterior half of presomite S-1 where it plays important roles in the rostral/caudal patterning of presomites as they mature (Morimoto et al., 2005; Shifley et al., 2008). While Spc1/furin expression is reported to be ubiquitous during rat embryogenesis (Zheng et al., 1994), specific expression of Spc6 has been reported in the PSM of developing mouse embryos (Constam et al., 1996; Rancourt and

97 Rancourt, 1997). To assess the potential functional overlap between SPC6 and LFNG

during embryonic segmentation, Spc6 expression was localized in the developing mouse embryo by comparison to that of Mesp2, a robust marker of the anterior presomite

compartment. During embryonic segmentation, Spc6 expression localizes immediately anterior to Mesp2 in S0, the somite that will next bud from the PSM (Figure 4.8). The expression domains of Spc6 and Mesp2 do not overlap, instead a clear border is maintained between the anterior compartment of somite S-1, where Mesp2 and Lfng are co-expressed, and somite S0 where Spc6 expression is initiated. Spc6 is also expressed in the most recently formed somite (S1), initially throughout the somite and resolving into a graded expression pattern with higher expression in the anterior somite (Figure 4.8 B).

Thus Spc6 is expressed exclusively in cells that previously expressed Lfng, but which have recently downregulated that expression. This supports a possible role for SPC6 in the processing and rapid secretion of residual LFNG protein in somite S0, ensuring that

LFNG activity is rapidly inactivated once Lfng transcription ceases.

4.3.7 LfngRL/+ mice show significant segmentation defects

To determine if the processing and secretion of LFNG is important for

segmentation clock function, we decided to examine the in vivo effects of the R/LFNG

mutation. We replaced the pre/pro signal sequence of endogenous Lfng with the pre

sequence of Rfng creating the allele LfngRL (Figure 4.9). Our targeting vector contained a

floxed neomycin resistant marker (Neo) and testis cre cassette; the presence of this

cassette in the targeted locus should prevent expression in chimeras, however the allele

would be activated when the cassette is floxed out upon passage through the male germline. The LfngRL allele maintains the endogenous transcription state site and splicing

98 sites. We hypothesized that the R/LFNG mutation would create a hyperactive LFNG.

The N-terminus of RFNG would tether the protein to the Golgi and increase its intracellular half-life as we observed occurring in tissue culture (Figure 4.3, 4.6). By increasing the amount of time LFNG spends in the Golgi modifying the Notch receptor, we expected the feedback loops regulating the segmentation clock would be disrupted.

We therefore predicted that the R/LFNG allele would have a dominant effect in mice.

Indeed, LfngRL/+ mice show segmentation defects with shortened bodies and

severely kinked tails, similar to Lfng-/- mice (Figure 4.10). To analyze the segmentation

phenotype further, we performed Alcian blue/Alizarin red staining on an LfngRL/+ mouse.

The skeletal derivates of the somites, the vertebrae and rib cage are highly disorganized in LfngRL/+ mice at all levels of the A/P axis. The vertebrae are misshapen and frequently

fused and the ribs show multiple origins and fusions as well. This phenotype is quite

similar to the skeletal defects observed in Lfng-/- mice which display a characteristic

cobblestone backbone, as well as disorganized rib cages with multiple origins and

occasional fusions (Figure 4.11 A). When we compare the number and shape of the

cervical vertebrae, again there are similarities between LfngRL/+ and Lfng-/- mice. Unlike

Lfngwt mice with seven, normal cervical vertebrae, LfngRL/+ and Lfng-/- mice only form

approximately 5 cervical vertebrae. Additionally, the vertebrae are fused beginning at the second cervical vertebrae in the LfngRL/+ mouse and at the third cervical vertebrae in the

Lfng-/- mouse. Fused and misshapen vertebrae continue throughout the thoracic region of

LfngRL/+ and Lfng-/- mice (Figure 4.11 B). The similar phenotypes observed in LfngRL/+

and Lfng-/- mice implies that having too much LFNG activity in the PSM is as detrimental

to the segmentation clock as not having any LFNG at all.

99 4.3.8 Somitogenesis is highly disrupted in LfngRL/+ embryos

The severe phenotype of the LfngRL/+ mice suggests that the segmentation clock is

not functioning properly. In fact, somitogenesis is visibly disrupted in LfngRL/+ embryos at 10.5 dpc where the tails are shortened compared to wild-type littermates (Figure 4.12).

Additionally, while the formation of epithelial somites out of the PSM is visible in 10.5 dpc wild-type embryos, it is not occurring in LfngRL/+ embryos (Figure 4.12), similar to

Lfng-/- embryos. To examine the maturation and patterning of somites, we examined the expression of a marker for the posterior sclerotome, Uncx4.1. We find that the Uncx4.1

expression pattern is highly disorganized in LfngRL/+ embryos. Again, the characteristic

Uncx4.1 “salt and pepper” expression pattern observed in Lfng-/- embryos is also occurring in LfngRL/+ embryos. Although sclerotome is forming, the somites are not properly patterned at any level of the body (Figure 4.13). Only at the level of the forelimb does there appear to be some condensation of sclerotome with occasional bands of Uncx4.1 expression observed in LfngRL/+ embryos (Figure 4.13 panel f). If we measure

the distance between these bands, however, they are not equally spaced, which suggests

either faulty patterning or sizing of the trunk somites in LfngRL/+ embryos. As we

discussed in Chapter 2, Lfng plays roles in both the segmentation clock and in R/C

patterning, and the severe phenotype of Lfng-/- embryos is due to a combined disruption

of both of these functions. It appears that the LfngRL allele is causing similar disruptions

through its dominant effect in LfngRL/+ embryos.

4.4 Discussion

For LFNG to function in the segmentation clock or during R/C somite patterning,

there must exist post-translational mechanisms that confer a short half-life to its activity

100 in the Notch pathway. We propose that protein processing by SPCs promotes the

secretion and subsequent inactivation of the mature LFNG protein and that the rapid release of the protein into the extracellular space acts in combination with the cyclic transcription of Lfng to facilitate its activity in the segmentation clock and to refine and terminate the functions of LFNG during R/C somite patterning. This would represent a novel mechanism for the regulation of Notch activity allowing tight temporal and spatial modulation of Notch signaling during somitogenesis.

4.4.1 LFNG processing by furin/SPC proteases is not necessary for its secretion or modification of Notch receptors

We localized N-terminal sequences that regulate the secretory behavior of the fringe-family proteins, and find that these sequences are sufficient to direct protein processing and/or secretion in a fringe specific manner (Figure 4.3). This indicates that, like many glycosyltransferases, sequences at the N-terminus, including putative type II transmembrane domains, are important in fringe localization and/or secretion (Young,

2004). Examining the regulation of LFNG processing and secretion, we found that furin and SPC6 can efficiently process LFNG, and identify two distinct sites in LFNG that can

be processed by SPC proconvertases (Figure 4.2, 4.4C).

Processing by SPC family proteases is conserved between Drosophila FRINGE

and LFNG, but the functional roles of LFNG processing have been undetermined. The

original descriptions of the Drosophila and vertebrate fringes suggested that protein

processing might be required to create active, mature protein (Cohen et al., 1997; Irvine

and Wieschaus, 1994; Johnston et al., 1997; Wu et al., 1996). However, we find that

mutations that block LFNG processing (LFNGm1/2), or which tether it in the Golgi 101 (R/LFNG) do not alter activity of the protein in the Notch receiving cell (Figure 4.7).

Thus, uncleaved protein is active, and has the same specificity in the Notch pathway as

wild-type LFNG, and the function of SPC processing does not relate to protein activation.

We propose instead that the secretion of LFNG from the cell acts as a mechanism

to terminate LFNG glycosyltransferase activity in the Notch signal receiving cell. This

would represent a novel mechanism for the regulation of Notch activity allowing tight temporal and spatial modulation of Notch signaling during somitogenesis. Cleavage and

secretion of some glycosyltransferases is proposed to be a general mechanism of turnover

(Cho and Cummings, 1997; Colley, 1997). For example, the secretion of the ST6Gal I

isoform is suggested to limit the sialylation activity of the enzyme (Ma et al., 1997). The

function of LFNG in the segmentation clock offers an example of a glycosyltransferase which is cleaved and secreted and whose activity must be modulated temporally,

providing a situation where the regulated turnover of a glycosyltransferase may indeed be

functional in vivo.

4.4.2 LFNG intracellular half-life is critical for proper somitogenesis

In the segmentation clock, Lfng has been proposed to cyclically inhibit Notch signaling through its modifications of the Notch1 receptor in the Golgi (Shifley and Cole,

2007). Lfng transcription is periodically inhibited by another Notch target Hes7 (Bessho et al., 2003) and we have shown in Chapter 3 that the 3’UTR of Lfng contains sequences that contribute to mRNA turn-over. The cyclic regulation of Lfng mRNA in combination with its protein processing by SPC proconvertases and rapid secretion may function to regulate LFNG protein levels, facilitating the oscillations of LFNG activity within the segmentation clock. One intriguing implication of this data would be the idea that 102 processing by SPC proteins plays important roles in rapidly clearing LFNG from the cell,

influencing the duration of its activity post-translationally. Supporting this idea, we find

that mutating the SPC processing sites in LFNG extends its intracellular half-life from 70

to 97 minutes, presumably by decreasing the secretion rate of the protein (Figure 4.6).

Unlike LFNG, RFNG protein is not cleaved and is a Golgi resident protein

(Johnston et al., 1997), therefore, as expected, the R/LFNG mutation tethers the protein in

the cell (Figure 4.3 B) and extends the intracellular half-life from 70 to 126 minutes

(Figure 4.6). Based on these results, we expected the R/LFNG mutation to exhibit

dominant phenotypic effects by creating a hyperactive LFNG protein. Indeed, LfngRL/+

embryos exhibit severe segmentation defects and do not undergo proper somitogenesis

(Figures 4.10, 4.12). The skeletal derivates of somites are highly disorganized (Figure

4.11), similar to the phenotypes observed in knock-outs of Notch signaling modifiers

Lfng, Hes7, and Dll3 (Bessho et al., 2001b; Dunwoodie et al., 2002; Evrard et al., 1998;

Kusumi et al., 1998; Zhang and Gridley, 1998), suggesting that the increase in LFNG half-life has significantly disrupted segmentation clock function. Similarly, researchers have made knock-in mice where the half-life of HES7 was increased by approximately 8 minutes and were able to disrupt somitogenesis (Hirata et al., 2004). These results suggest that the careful regulation of protein activity is critical for proper segmentation clock function. However, unlike our proposal for LFNG regulation, HES7 turn-over has been found to be regulated by the proteosome (Bessho et al., 2003).

Thus far, neither the male or female LfngRL/+ mice generated have successfully

mated, so all analyses were performed on F1 offspring of chimeric males. The cause of

this infertility is not clear. Lfng -/- females are reported to be infertile due to improper

103 folliculogenesis, although this appears to be strain and background dependent (Hahn et al., 2005; Xu et al., 2006). Lfng -/- males have cysts in their rete testis (Hahn et al., 2009).

Therefore, it is possible that hyperactive LFNG activity from the Lfng RL allele may be adversely affecting fertility. Alternatively, skeletal defects may interfere with mating.

However, transgenic mice with constituative Lfng expression have been successfully bred (Serth et al., 2003), and only two LfngRL/+ mice have survived to adulthood due to

low transmission frequency from our chimeras. In the future, creating a conditional

LfngRL allele will circumvent fertility issues and allow for further in depth analysis of

LfngRL/+ and LfngRL/RL embryos. We predict that a number of segmentation clock and

R/C patterning genes will be disrupted in LfngRL embryos. Additionally, we predict that

unlike Lfng -/- embryos where NICD is constitutively activated in Region I of the PSM

(Morimoto et al., 2005), LfngRL embryos will have constitutive NICD repression in the

PSM, which would provide further evidence that Lfng normally inhibits Notch1 activation in the PSM.

4.4.3 SPC/furin cleavage of LFNG may be important for multiple aspects of somitogenesis

In the anterior PSM, we propose that LFNG activity is required in somite S-1 to modulate Notch1 signaling during R/C somite patterning, and that LFNG protein is then secreted from cells to terminate that function. The expression pattern we define for Spc6 supports a possible role for that protein specifically in regulating LFNG activity during

R/C somite patterning (Figure 4.8). We hypothesize that the activity of SPC6 in somites

S0 and S1 would promote the rapid cleavage and secretion of any residual LFNG protein in those regions, preventing unwanted LFNG activity. SPC6 has been deleted in mice,

104 and the best described allele results in embryonic death before gastrulation (Essalmani et

al., 2006). More recently, this gene has been conditionally inactivated in the mouse

epiblast, resulting in altered anterior/ posterior patterning, extra thoracic and lumbar

vertebrae, uneven rib attachments, and loss of tail structures, among other phenotypes.

The authors also find that GDF11 is an in vivo substrate of SPC6 (Essalmani et al., 2008).

Additional analysis will be required to determine whether LFNG is an in vivo substrate of

SPC6, and whether loss of LFNG protein processing contributes to the Spc6 knockout phenotype. However, LfngRL/+ embryos do not properly pattern their somites with

Uncx4.1 displaying a highly disorganized expression pattern (Figure 4.13), suggesting

that the increase in LFNG intracellular half-life may not only be disrupting the

segmentation clock, but also R/C patterning.

LFNG cleavage in the posterior PSM, during clock function must be regulated by

a distinct SPC family member, as we detect no SPC6 expression in this region. Furin,

which we show can efficiently process LFNG (Figure 4.2), is relatively ubiquitously

expressed during embryogenesis. Further, targeted deletion of this protein causes

irregular somitogenesis (Roebroek et al., 1998). Interestingly, SPC6B and furin have

been shown to target the same substrates like BMP4 and are both specifically inhibited by

α1PDX (Cui et al., 1998; Jean et al., 1998). However, SPC6B does not compliment the

defects seen in SPC1/furin null mice and SPC1/furin and SPC6 have been shown to localize to different compartments of the Golgi network suggesting similar, but distinct

activities (Roebroek et al., 1998; Xiang et al., 2000). This raises the intriguing possibility

that different SPC proteases may cleave LFNG with greater or lesser efficiency,

functionally creating subtle changes in half-life that may play important roles in the post-

105 translational modulation of LFNG function.

4.4.4 An extracellular role for LFNG?

One unaddressed question in fringe biology is the possibility that LFNG may in fact play some active role in the extracellular space. Extracellular roles have been suggested for fringe proteins, although the mechanisms behind these findings have not been elucidated (Mikami et al., 2001; Mustonen et al., 2002; Wu et al., 1996). Indeed, extracellular functions have been found for other glycosyltransferases. For example, secreted glycosyltransferases have been hypothesized to have lectin or adhesion functions in the intercellular space (Colley, 1997). In another case a secreted glycosyltransferase

(GnT-V) has been found to possess an independent, extracellular signaling function that does not require its transferase activity (Saito et al., 2002). It is unclear at this time whether LFNG plays a role in the extracellular space and what this role might be.

To date it has been difficult to address the functional significance of LFNG processing and secretion in vivo. The results reported in this Chapter define a panel of

LFNG mutations that differentially affect protein processing, intracellular half-life, and secretion. The LfngRL/+ embryos reveal the importance of LFNG turnover for somitogenesis. The segmentation defects observed in LfngRL/+ embryos closely mimic the phenotypes observed in Lfng-/- embryos, suggesting that the loss of rapid LFNG turn-over is as detrimental to segmentation clock function as a complete loss of Lfng expression.

106

TABLE 4.1: PRIMER SEQUENCES FOR MUTAGENESIS OF LFNG b

Mutation FORWARD PRIMER REVERSE PRIMER R/LFNG a 5'-atgcggccggcggccaccatgagccgtgcgcggcgg 5'-ggttcttccgagtggtcttg M/LFNG a 5'-atgcggccggcggccaccatgcactgccgacttt 5'-atgaagacgtcgcctagctg LFNGm1 5'-accgccgcggcagcagacgcggatccacc 5'-cgatgaagacgtcgcgag (5' fragment) LFNGm1 5'- agagttcctgaagcgagag 5'-tctgctgccgcggcggtgagtagactgaag (3' fragment) LFNGm2 5'-gccgcagcggccgcgctgcgtagcctg 5'-cgatgaagacgtcgcgag (5' fragment) LFNGm2 5'- agagttcctgaagcgagag 5'-ggccgctgcggcctcagcgggcatcgg (3' fragment) a. EagI sites were introduced in primers (underlined). EagI to AatII fragments were excised and ligated at the AatII site in LFNG b. Mutated nucleotides are underlined. After a second round of PCR with flanking primers, the mutated fragments were ligated into the AatII site of LFNG

107

Figure 4.1: Alignment of pre/pro region of fringe genes. BOXSHADE (Subramaniam, 1998) alignment of Drosophila fringe, mouse Lunatic, human LUNATIC, mouse Radical, and mouse Manic pre/pro sequences. Consensus dibasic cleavage sites are boxed and the described cleavage site is indicated (arrow) (Johnston et al., 1997). Proposed signal sequences are underlined (Nielsen et al., 1997). Amino acids conserved in all sequences are shaded black; amino acids identical in at least 50% of the sequences are shaded grey. 1st conserved amino acid indicated (*).

108

Figure 4.2: LFNG is processed by SPC proconvertases. A. Media fractions from NIH3T3 cells transfected with expression vectors encoding wild-type mouse LFNG wt (LFNG ) and α1PDX were analyzed by Western blot (ng of α1PDX expression vector indicated). Protein species are diagrammed, (black box = mature LFNG, gray oval = LFNG pro region). α1PDX expression reduces the amount of secreted mature LFNG (35 kDa) and increases the amount of full-length LFNG (43.6 kDa) suggesting a general inhibition of LFNG processing. B. NIH3T3 cells were transfected with LFNG expression vector, α1PDX expression vector, and SPC protease expression vectors as indicated, and media fractions were analyzed by Western blot. SPC1, SPC6A, and SPC6B efficiently compete with the inhibitor and cleave LFNG, while SPC4 and SPC7 are less efficient.

109

B LFNG

MFNG

Protein in Media M/LFNG Protein in Cells

RFNG

R/LFNG

25 50 75 100 125

Figure 4.3: The N-terminus of fringe proteins controls their secretory behavior. A. For chimeric fringe proteins the first 112 aa of wild-type LFNG (grey circles) were replaced with either the first 59 aa of RFNG (white circle) or the first 54 aa of MFNG (black circle). Fusion is at LFNG D113 the first conserved amino acid in fringe proteins. B. Steady-state AP activity was measured in the cell extracts and media of cells transfected with vectors encoding AP-tagged fringe proteins. Results are shown as the fraction of total AP activity in the cellular or media fraction and are the mean ± SD of at least three independent experiments. LFNG is found almost entirely in the media, while RFNG and R/LFNG are almost entirely retained in the cell. MFNG and M/LFNG are detected at similar levels in both cellular extracts and media.

110

Figure 4.4: LFNG protein is cleaved at two sites, but neither cleavage is necessary for secretion. A. Dibasic sites in mouse LFNG were mutated as indicated (low asterisk) in constructs LFNGm1, LFNGm2, LFNGm1/2. B. Steady-state AP activity was measured in the cell extracts and media of cells transfected with vectors encoding AP-tagged fringe proteins. Results are shown as the fraction of total AP activity in the cellular or media fraction and are the mean ± SD of at least three independent experiments. LFNG is found almost entirely in the media, as are the mutated cleavage constructs. C. Cells were transfected with HA-tagged LFNG and LFNGm1 (I), LFNGm2 (II), or LFNGm1/2 (III). α1PDX and/or SPC6A vectors were co-transfected as indicated. HA-tagged proteins in the media were analyzed by Western blot. LFNG is found in the media predominantly as the 35 kDa mature form. Bands corresponding to the predicted full-length protein (43.6 kDa) and a band of intermediate size (40.4 kDa) are also detected (lane 1). α1PDX blocks the formation of the mature protein and the amount of the full-length protein is increased (lane 2). SPC6A drives production of mature LFNG (lane 3). LFNGm1 is detected as either the intermediate or full-length protein when expressed. α1PDX blocks production of the intermediate LFNG band, which is recovered upon expression of SPC6A (I, lanes 4–6). LFNGm2 blocks production of the intermediate band and is secreted as the mature 35 kDa fragment. α1PDX expression blocks the formation of the mature protein, which is recovered with expression of SPC6A (II, lanes 4–6). LFNGm1/2 is found in the media exclusively as the full-length protein (III, lanes 4–6). Secreted protein species are diagrammed.

111

B LFNG

LFNGm1 Protein in Media Protein in Cells LFNGm2

LFNGm1/2

0 25 50 75 100 125

Figure 4.4

112

Figure 4.5: LFNG variant proteins localize to the Golgi. NIH3T3 cells expressing AP-tagged FRINGE expression constructs were analyzed by immunohistochemistry. The cis-Golgi marker GM130 (green, panels a-f) and the AP-tag of the fringe proteins listed (red, panels g-l), overlap, localizing the fringe variant proteins in the Golgi (merge, panels m-r, arrows). Nuclei were counter-stained with Hoechst dye (blue). Similar to previous descriptions LFNGwt localization overlaps with that of GM-130, a marker of the cis-Golgi. As expected, we find that all LFNG protein variants, regardless of their effects on LFNG processing or secretion exhibit similar localization, thus these mutations do not perturb the Golgi localization of LFNG.

113

Golgi AP-tag Merge

LFNG

m1 LFNG

m2 LFNG

LFNGm1/2

R/LFNG

M/LFNG

Figure 4.5

114

halflivesforthesisgraph

1.1

0.9 LFNG 0.8 LFNG m1 0.7 LFNG LFNG m1/2 0.6 R/LFNG 0.5

Relative amount of LFNG 0.4

0.3 0 20406080 Time (minutes)

Column 1

Figure 4.6: Mutations affecting LFNG processing result in an increased intracellular protein half-life. A. NIH3T3 cells expressing HA-tagged mouse LFNG proteins were treated with cycloheximide for an 80-minute time course. Cellular extracts were analyzed by Western blot. A representative LFNG time course is shown indicating that LFNG protein levels decrease over time (indicated in minutes). Tubulin is used as a loading control. B. Protein concentration in cellular extracts was quantified and normalized to tubulin concentration. Relative protein concentration is shown over time (initial timepoint set to 1) and standard deviation from at least 6 independent experiments is represented by error bars. Exponential trend lines were fitted to data points to calculate the intracellular protein half-lives. LFNG has a calculated half-life of 70 min, while LFNGm1 and LFNGm1/2 have intracellular half-lives of 97 min. The R/LFNG half-life is 126 min.

115

Figure 4.7: LFNG variant proteins modify Notch signaling with an LFNG-like activity. NIH3T3 cells were co-transfected with Notch1 expression vector and expression vectors encoding either secreted AP (SeAP) or an AP-tagged mouse fringe protein, along with a CBF1-luciferase reporter and pSVβgal. Transfected cells were co- cultured with cells expressing JAGGED1 or parental LTK- cells. Luciferase values were normalized to βgal values to control for transfection efficiency and are expressed as fold activation reflecting relative luciferase units (RLU) induced by JAGGED1 expressing cells over the RLUs obtained with parental L-cells. (JAGGED1-induction of Notch1 in the absence of fringe was set to 1). Expression of LFNG, LFNGm1, LFNGm2, LFNGm1/2, and R/LFNG significantly reduce JAGGED1-induced activation of Notch signaling (*p < 0.02, **p < 0.005) while RFNG potentiates JAGGED1-induced Notch1 activation. Error bars represent the standard deviation of the mean from at least three independent experiments performed in triplicate.

116

Figure 4.8: SPC6 expression patterns suggest a role in clearing LFNG from maturing somites. A. Whole mount RNA in situ hybridization of 10.5 dpc mouse embryos comparing the expression patterns of Mesp2 and Spc6. (a, b) Spc6 signal is purple, and Mesp2 signal is orange. (c, d) Spc6 signal is orange, and Mesp2 signal is purple. The expression of Mesp2 does not overlap with that of Spc6, forming a clear border at the boundary between presomites S0 and S-1. Spc6 is also expressed in the newly formed, mature somite S1 either throughout the somite (a, c) or in a graded manner with highest expression in the rostral half of the somite (b, d). Somite borders are indicated by short lines. B. Schematic of the PSM. In the posterior PSM, LFNG expression is cyclic, and the protein may be cleaved by SPC proconvertases, contributing to its short intracellular half-life. In the anterior PSM, where presomites are patterned, the expression domains of Lfng and Spc6 do not overlap, suggesting that SPC6 may serve to help cleave and clear any remaining LFNG from the maturing somites, especially from the rostral compartment of S0.

117

A R/Lfng targeting vector

HindIII EcoRV Neo Cre

Lfng endogenous locus

EcoRV HindIII EcoRV

5’ Target in ES cells 3’

EcoRV HindIII EcoRV EcoRV

Neo Cre

Targeted locus after Cre Excision EcoRV HindIII EcoRV

+/RL +/RL EcoRV Bands: Endogenous= 13.5 kb 5’ after target= 10.4kb 3’ after target=8.0 kb 3’ probe

Figure 4.9: Knock-in of the R/LFNG allele. A. Targeting of the Lfng endogenous locus is shown (boxes signify coding exons). The targeting event replaces 28aa of the mature region of exon 1 of wild type LFNG (hatched circles) with the pre signal sequence and 21aa of Radical Fringe (RFNG) (checkered circles). External probes (labeled boxes) are shown below. B. Southern blot showing targeted colonies containing the LfngRL/+ allele. Arrow: endogenous band, Arrowhead: targeted band.

118

Figure 4.10: LfngRL/+ mice have shortened tails and bodies. A. An LfngRL/+ mouse (arrow) has a shortened body and stubby tail compared to the wild-type mouse. B. 18.5 dpc embryos also show a segmentation phenotype; the LfngRL/+ pup has a shortened tail (arrow) compared to its wild-type littermate.

119

Figure 4.11: Skeletal derivates of somites are highly disorganized in LfngRL/+ mice. A. Lateral (a,b,c) and dorsal (d,e,f) views of skeletal preps of mice with the indicated genotypes are shown. Compared to wild-type mice, the rib cages of LfngRL/+ and Lfng-/- mice are highly disorganized with shared origins (arrowheads) and multiple rib fusions (arrows). Unfortunately, the LfngRL/+ skeletal preparation was damaged during the protocol causing the broken connections between the ribs and sternum and the loss of lumbar and sacral vertebrae. B. Close-ups of the vertebral column of wild-type (a,b), LfngRL/+ (c,d), and Lfng-/- (e,f) mice. In wild-type mice, there are seven cervical vertebrae, while in LfngRL/+ and Lfng-/- mice, there are, at the most, five cervical vertebrae with fusions starting at c2 and c3, respectively. The disorganization of the LfngRL/+ vertebrae extends throughout the thoracic region (d) with misshapen and fused vertebrae unlike the wild-type mice (b), but similar to Lfng-/- mice (f).

120

A + /+ Lfng RL/+ Lfng -/- Lfng

a b c

d e f

B c1 c2

+ c7

/ +

Lfng a b c1 c2 c5

RL/+

Lfng c d

c1 c2 c5

-/-

Lfng e f

Figure 4.11

121

+ / +

Lfng

a b

RL/+

Lfng

c d e

Figure 4.12: LfngRL/+ embryos are not forming epithelial somites properly. The shortened tail of LfngRL/+ embryos is evident in 10.5 dpc embryos (compare a and c, white arrows). Additionally, while the PSM of wild-type embryos is forming clear, epithelial somites (b, lines mark somite boundaries), LfngRL/+ embryos are not forming visible epithelial somites (d) or are showing only indentations rather than clear, somite boundaries (e, pink arrows).

122

+ /

Lfng

a b c

RL/+

{

Lfng

d e f

-/-

Lfng {

g h i

Figure 4.13: Somite patterning is disrupted in LfngRL/+ embryos. Whole-mount in situ hybridization with a probe against Uncx4.1, which demarcates the caudal half of the somites. At 9.5 dpc, wild-type somites have clear rostral and caudal compartments (a-c) while LfngRL/+ (d-f) and Lfng-/- embryos (g-i) have little compartmentalization. Uncx4.1 is expressed throughout the somites in a disorganized, spotted pattern in tails of LfngRL/+ and Lfng-/- embryos compared to the tails of wild-type embryos (anterior is up in panels b, e, and h), suggesting that the somites are not being properly patterned. Additionally, the spacing between somites in the trunk appears to be irregular in LfngRL/+ and Lfng-/- embryos (bracket) possibly suggesting the size of the somites is irregular (anterior is to the right in panels c, f, and i).

123

CHAPTER 5

CONCLUSION

5.1 Overview

The aim of this project was to examine the regulation of Lunatic fringe during the developmental process of somitogenesis. Somitogenesis is the morphological hallmark of vertebrate segmentation. Somites give rise to the most obvious segmental structures of the adult body plan including the vertebrae and ribs. The regulation of somitogenesis is complex: the segmentation clock organizes cohorts of cells into presomites, the wavefront specifies their positional differentiation, and somites undergo rostral/caudal patterning in Region II of the PSM. The Notch signaling pathway plays multiple roles during somitogenesis; it is an important component of the segmentation clock and also appears to be involved in the R/C patterning of somites. Targeted deletions of Notch pathway members in mouse models cause a number of segmentation defects reflecting the fact that multiple aspects of somitogenesis have been disrupted. Importantly, the role of the Notch signaling pathway in somitogenesis appears to be conserved in humans, as many familial cases of Spondylocostal Dysostosis are caused by mutations in Notch pathway members.

124 This study has focused on the role of the Notch signaling modifier Lfng during somitogenesis. We find that Lfng plays an important role in the segmentation clock, specifically during primary body formation, and also contributes to R/C patterning of somites. As part of the segmentation clock, Lfng mRNA is periodically transcribed and

LFNG protein levels cycle in the PSM. Lfng mRNA and LFNG protein molecules must therefore have rapid turnover rates in the PSM and we explore possible mechanisms for their turnover. We find that the short intracellular half-life of LFNG protein is essential for proper segmentation clock function. Overall, Lfng acts to spatially and temporally regulate Notch signaling and its tightly regulated activity plays an important role in vertebrate development.

5.2 Lfng is important for segmentation clock function during primary, but not secondary body formation

In Chapter 2, we show that the specific loss of Lfng expression in the segmentation clock caused by the Lfng∆FCE1 allele perturbs normal anterior skeletal

development. In contrast, posterior development is relatively normal in these mice. This

is due to a specific disruption of the segmentation clock during primary body formation.

Interestingly, the expression of oscillatory genes is differentially affected during primary and secondary body formation in Lfng∆FCE1/∆FCE1 embryos. For instance, Lfng expression

and cyclic Notch activity are necessary for maintaining Hes7 oscillations during primary

body formation, but dispensable during secondary body formation. The Lfng∆FCE1 allele

has also revealed that Lfng not only plays a role in the segmentation clock, but also in

R/C patterning of somites. By maintaining Lfng expression in Region II, R/C somite

patterning of maturing somites occurs properly unlike Lfng-/- embryos. Interestingly, it

125 appears that the R/C patterning of developing somites may affect ongoing segmentation

during secondary body formation, as Lfng∆FCE1/∆FCE1 embryos are able to complete

somitogenesis unlike Lfng-/- embryos.

Thus, the Lfng∆FCE1 allele has uncovered several new findings about the regulation

of somitogenesis. These data support the idea that the segmentation clock requires different components during primary versus secondary body formation. It will be interesting in the future to determine how the segmentation clock is controlled during secondary body formation. Two intriguing possibilities include the Wnt or FGF pathways which, as discussed, also exhibit cyclic expression patterns in the PSM. Our analysis of Lfng∆FCE1/∆FCE1 embryos also demonstrates that the functions of Lfng and the

Notch pathway in the segmentation clock can largely be separated from their roles in R/C

patterning; this has been difficult to determine with certainty from traditional knock-outs.

Finally, the R/C pattering region of the PSM appears to be able to communicate with the

posterior PSM and dictate whether or not somitogenesis should continue. Identifying the

signals used in this situation is another interesting future aim.

5.3 The Lfng 3’UTR contributes to short RNA half-life in the PSM

The segmentation clock requires the rapid turn-over of Lfng RNA molecules in

order to function properly. In Chapter 3, we identify sequences in the Lfng 3’UTR that

might help regulate Lfng RNA stability. The Lfng 3’UTR appears to utilize complex,

modular activity in its regulation of Lfng RNA. Conserved sub-fragments of the Lfng

3’UTR are sufficient to reveal cyclic expression of exogenous transgenes in the PSM

similar to endogenous Lfng expression. Continued analysis of these transgenes will help

identify which sequences are sufficient and, in the future, it will be interesting to identify

126 what regulators bind to these sequences, how they are able to cyclically regulate Lfng

RNA, and what other genes they might target.

5.4 LFNG protein processing by proprotein convertases contributes to the short protein half-life in the segmentation clock

In Chapter 4, we examine the cleavage and secretion of LFNG protein. We find that N-terminal sequences regulate the secretory behavior of fringe family proteins and that SPC1/furin, SPC6A and SPC6B promote LFNG processing. SPC family proteases can recognize two dibasic cleavage sites in LFNG, however this processing is not required for LFNG secretion. Prevent of LFNG processing by SPC proteases increases the intracellular half-life of the LFNG protein without affecting the specificity of its function in the Notch signaling pathway. We hypothesized that the short intracellular half-life of LFNG was important for its function in the segmentation clock and used the

R/LFNG mutation to test this hypothesis. LfngRL/+ mice have significant segmentation defects in their skeletons and somitogenesis is highly disrupted. This phenotype closely mirrors Lfng-/- mice, suggesting that having a hyperactive LFNG is just as detrimental as completely lacking LFNG protein activity in the segmentation clock. Perhaps the

LfngRL/RL homozygotes will have phenotype similar to LfngRL/+ embryos if the allele is truly dominant. We expect that both functions of Lfng are disrupted in the LfngRL/+ mice; the segmentation clock is not functioning properly because of the hyperactive LFNG activity, and the R/C patterning of somites may also be disrupted as SPC6 (which we hypothesize is expressed in maturing somites to cleave and clear LFNG) would not be able to cleave and clear R/LFNG protein from maturing somites. In the future, it will be interesting to determine how the LfngRL allele affects the expression patterns of other

127 clock genes and R/C patterning makers during both primary and secondary body formation. Its effects on Notch activity, as well as Wnt and FGF pathway members

might also shed some light on the possible interactions these signaling pathways use to

coordinate the segmentation clock.

Overall, we have found that Lfng acts to spatially and temporally regulate Notch

signaling during somitogenesis and is itself tightly regulated at multiple levels. The

studies in this dissertation have demonstrated the complex genetic controls that used to coordinate the development of somites, an important aspect of the vertebrate body plan.

128

LIST OF REFERENCES

Aulehla, A., and Herrmann, B. G., 2004, Segmentation in vertebrates: clock and gradient finally joined, Genes Dev 18(17):2060-7.

Aulehla, A., and Johnson, R. L., 1999, Dynamic expression of lunatic fringe suggests a link between notch signaling and an autonomous cellular oscillator driving somite segmentation, Dev Biol 207(1):49-61.

Aulehla, A., Wehrle, C., Brand-Saberi, B., Kemler, R., Gossler, A., Kanzler, B., and Herrmann, B. G., 2003, Wnt3a plays a major role in the segmentation clock controlling somitogenesis, Dev Cell 4(3):395-406.

Aulehla, A., Wiegraebe, W., Baubet, V., Wahl, M. B., Deng, C., Taketo, M., Lewandoski, M., and Pourquie, O., 2008, A beta-catenin gradient links the clock and wavefront systems in mouse embryo segmentation, Nat Cell Biol 10(2):186- 93.

Barrantes, I. B., Elia, A. J., Wunsch, K., Hrabe de Angelis, M. H., Mak, T. W., Rossant, J., Conlon, R. A., Gossler, A., and de la Pompa, J. L., 1999, Interaction between Notch signalling and Lunatic fringe during somite boundary formation in the mouse, Curr Biol 9(9):470-80.

Bessho, Y., Hirata, H., Masamizu, Y., and Kageyama, R., 2003, Periodic repression by the bHLH factor Hes7 is an essential mechanism for the somite segmentation clock, Genes Dev 17(12):1451-6.

Bessho, Y., Miyoshi, G., Sakata, R., and Kageyama, R., 2001a, Hes7: a bHLH-type repressor gene regulated by Notch and expressed in the presomitic mesoderm, Genes Cells 6(2):175-85.

Bessho, Y., Sakata, R., Komatsu, S., Shiota, K., Yamada, S., and Kageyama, R., 2001b, Dynamic expression and essential functions of Hes7 in somite segmentation, Genes Dev 15(20):2642-7.

129 Bonafe, L., Giunta, C., Gassner, M., Steinmann, B., and Superti-Furga, A., 2003, A cluster of autosomal recessive spondylocostal dysostosis caused by three newly identified DLL3 mutations segregating in a small village, Clin Genet 64(1):28-35.

Bruckner, K., Perez, L., Clausen, H., and Cohen, S., 2000, Glycosyltransferase activity of Fringe modulates Notch-Delta interactions, Nature 406(6794):411-5.

Bulman, M. P., Kusumi, K., Frayling, T. M., McKeown, C., Garrett, C., Lander, E. S., Krumlauf, R., Hattersley, A. T., Ellard, S., and Turnpenny, P. D., 2000, Mutations in the human delta homologue, DLL3, cause axial skeletal defects in spondylocostal dysostosis, Nat Genet 24(4):438-41.

Bunting, M., Bernstein, K. E., Greer, J. M., Capecchi, M. R., and Thomas, K. R., 1999, Targeting genes for self-excision in the germ line, Genes Dev 13(12):1524-8.

Bussen, M., Petry, M., Schuster-Gossler, K., Leitges, M., Gossler, A., and Kispert, A., 2004, The T-box transcription factor Tbx18 maintains the separation of anterior and posterior somite compartments, Genes Dev 18(10):1209-21.

Candia, A. F., Hu, J., Crosby, J., Lalley, P. A., Noden, D., Nadeau, J. H., and Wright, C. V., 1992, Mox-1 and Mox-2 define a novel homeobox gene subfamily and are differentially expressed during early mesodermal patterning in mouse embryos, Development 116(4):1123-36.

Carthew, R. W., and Sontheimer, E. J., 2009, Origins and Mechanisms of miRNAs and siRNAs, Cell 136(4):642-55.

Chen, J., Kang, L., and Zhang, N., 2005, Negative feedback loop formed by Lunatic fringe and Hes7 controls their oscillatory expression during somitogenesis, Genesis 43(4):196-204.

Cho, S. K., and Cummings, R. D., 1997, A soluble form of alpha1,3-galactosyltransferase functions within cells to galactosylate glycoproteins, J Biol Chem 272(21):13622- 8.

Christ, B., Jacob, H. J., and Jacob, M., 1974, [Somitogenesis in the chick embryo. Determination of the segmentation direction], Verh Anat Ges 68:573-9.

Christ, B., Schmidt, C., Huang, R., Wilting, J., and Brand-Saberi, B., 1998, Segmentation of the vertebrate body, Anat Embryol (Berl) 197(1):1-8.

Cohen, B., Bashirullah, A., Dagnino, L., Campbell, C., Fisher, W. W., Leow, C. C., Whiting, E., Ryan, D., Zinyk, D., Boulianne, G., Hui, C. C., Gallie, B., Phillips, R. A., Lipshitz, H. D., and Egan, S. E., 1997, Fringe boundaries coincide with

130 Notch-dependent patterning centres in mammals and alter Notch-dependent development in Drosophila, Nat Genet 16(3):283-8.

Cole, S. E., Levorse, J. M., Tilghman, S. M., and Vogt, T. F., 2002, Clock regulatory elements control cyclic expression of Lunatic fringe during somitogenesis, Dev Cell 3(1):75-84.

Colley, K. J., 1997, Golgi localization of glycosyltransferases: more questions than answers, Glycobiology 7(1):1-13.

Conlon, R. A., Reaume, A. G., and Rossant, J., 1995, Notch1 is required for the coordinate segmentation of somites, Development 121(5):1533-45.

Constam, D. B., Calfon, M., and Robertson, E. J., 1996, SPC4, SPC6, and the novel protease SPC7 are coexpressed with bone morphogenetic proteins at distinct sites during embryogenesis, J Cell Biol 134(1):181-91.

Constam, D. B., and Robertson, E. J., 1999, Regulation of bone morphogenetic protein activity by pro domains and proprotein convertases, J Cell Biol 144(1):139-49.

Cooke, J., 1975, Control of somite number during morphogenesis of a vertebrate, Xenopus laevis, Nature 254(5497):196-9.

Cooke, J., and Zeeman, E. C., 1976, A clock and wavefront model for control of the number of repeated structures during animal morphogenesis, J Theor Biol 58(2):455-76.

Cui, Y., Jean, F., Thomas, G., and Christian, J. L., 1998, BMP-4 is proteolytically activated by furin and/or PC6 during vertebrate embryonic development, Embo J 17(16):4735-43.

Dale, J. K., Maroto, M., Dequeant, M. L., Malapert, P., McGrew, M., and Pourquie, O., 2003, Periodic notch inhibition by lunatic fringe underlies the chick segmentation clock, Nature 421(6920):275-8.

Deng, C., Wynshaw-Boris, A., Zhou, F., Kuo, A., and Leder, P., 1996, Fibroblast growth factor receptor 3 is a negative regulator of bone growth, Cell 84(6):911-21.

Dequeant, M. L., Glynn, E., Gaudenz, K., Wahl, M., Chen, J., Mushegian, A., and Pourquie, O., 2006, A complex oscillating network of signaling genes underlies the mouse segmentation clock, Science 314(5805):1595-8.

Diez del Corral, R., Olivera-Martinez, I., Goriely, A., Gale, E., Maden, M., and Storey, K., 2003, Opposing FGF and retinoid pathways control ventral neural pattern,

131 neuronal differentiation, and segmentation during body axis extension, Neuron 40(1):65-79.

Dubrulle, J., McGrew, M. J., and Pourquie, O., 2001, FGF signaling controls somite boundary position and regulates segmentation clock control of spatiotemporal Hox gene activation, Cell 106(2):219-32.

Dubrulle, J., and Pourquie, O., 2004, Coupling segmentation to axis formation, Development 131(23):5783-93.

Dunty, W. C., Jr., Biris, K. K., Chalamalasetty, R. B., Taketo, M. M., Lewandoski, M., and Yamaguchi, T. P., 2008, Wnt3a/beta-catenin signaling controls posterior body development by coordinating mesoderm formation and segmentation, Development 135(1):85-94.

Dunwoodie, S. L., 2009, Mutation of the fucose-specific beta1,3 N- acetylglucosaminyltransferase LFNG results in abnormal formation of the spine, Biochim Biophys Acta 1792(2):100-11.

Dunwoodie, S. L., Clements, M., Sparrow, D. B., Sa, X., Conlon, R. A., and Beddington, R. S., 2002, Axial skeletal defects caused by mutation in the spondylocostal dysplasia/pudgy gene Dll3 are associated with disruption of the segmentation clock within the presomitic mesoderm, Development 129(7):1795-806.

Elmasri, H., Liedtke, D., Lucking, G., Volff, J. N., Gessler, M., and Winkler, C., 2004, her7 and hey1, but not lunatic fringe show dynamic expression during somitogenesis in medaka (Oryzias latipes), Gene Expr Patterns 4(5):553-9.

Essalmani, R., Hamelin, J., Marcinkiewicz, J., Chamberland, A., Mbikay, M., Chretien, M., Seidah, N. G., and Prat, A., 2006, Deletion of the gene encoding proprotein convertase 5/6 causes early embryonic lethality in the mouse, Mol Cell Biol 26(1):354-61.

Essalmani, R., Zaid, A., Marcinkiewicz, J., Chamberland, A., Pasquato, A., Seidah, N. G., and Prat, A., 2008, In vivo functions of the proprotein convertase PC5/6 during mouse development: Gdf11 is a likely substrate, Proc Natl Acad Sci U S A 105(15):5750-5.

Evrard, Y. A., Lun, Y., Aulehla, A., Gan, L., and Johnson, R. L., 1998, lunatic fringe is an essential mediator of somite segmentation and patterning, Nature 394(6691):377-81.

Flint, O. P., Ede, D. A., Wilby, O. K., and Proctor, J., 1978, Control of somite number in normal and amputated mutant mouse embryos: an experimental and a theoretical analysis, J Embryol Exp Morphol 45:189-202.

132

Forsberg, H., Crozet, F., and Brown, N. A., 1998, Waves of mouse Lunatic fringe expression, in four-hour cycles at two-hour intervals, precede somite boundary formation, Curr Biol 8(18):1027-30.

Frazer, K. A., Pachter, L., Poliakov, A., Rubin, E. M., and Dubchak, I., 2004, VISTA: computational tools for comparative genomics, Nucleic Acids Res 32(Web Server issue):W273-9.

Geffers, I., Serth, K., Chapman, G., Jaekel, R., Schuster-Gossler, K., Cordes, R., Sparrow, D. B., Kremmer, E., Dunwoodie, S. L., Klein, T., and Gossler, A., 2007, Divergent functions and distinct localization of the Notch ligands DLL1 and DLL3 in vivo, J Cell Biol 178(3):465-76.

Gessler, M., Knobeloch, K. P., Helisch, A., Amann, K., Schumacher, N., Rohde, E., Fischer, A., and Leimeister, C., 2002, Mouse gridlock: no aortic coarctation or deficiency, but fatal cardiac defects in Hey2 -/- mice, Curr Biol 12(18):1601-4.

Goldbeter, A., Gonze, D., and Pourquie, O., 2007, Sharp developmental thresholds defined through bistability by antagonistic gradients of retinoic acid and FGF signaling, Dev Dyn 236(6):1495-508.

Goldbeter, A., and Pourquie, O., 2008, Modeling the segmentation clock as a network of coupled oscillations in the Notch, Wnt and FGF signaling pathways, J Theor Biol 252(3):574-85.

Gomez, C., Ozbudak, E. M., Wunderlich, J., Baumann, D., Lewis, J., and Pourquie, O., 2008, Control of segment number in vertebrate embryos, Nature 454(7202):335- 9.

Gossler, A., and Hrabe de Angelis, M., 1998, Somitogenesis, Curr Top Dev Biol 38:225- 87.

Gossler, A. a. T., P.P., 2002, Somitogenesis: segmentation of the paraxial mesoderm and delineation of tissue compartments., in: Mouse Development (J. R. a. P. P. Tam, ed.), Academic Press, San Diego, pp. 122-149.

Hahn, K. L., Beres, B., Rowton, M. J., Skinner, M. K., Chang, Y., Rawls, A., and Wilson-Rawls, J., 2009, A deficiency of lunatic fringe is associated with cystic dilation of the rete testis, Reproduction 137(1):79-93.

Hahn, K. L., Johnson, J., Beres, B. J., Howard, S., and Wilson-Rawls, J., 2005, Lunatic fringe null female mice are infertile due to defects in meiotic maturation, Development 132(4):817-28.

133 Handrigan, G. R., 2003, Concordia discors: duality in the origin of the vertebrate tail, J Anat 202(Pt 3):255-67.

Hicks, C., Johnston, S. H., diSibio, G., Collazo, A., Vogt, T. F., and Weinmaster, G., 2000, Fringe differentially modulates Jagged1 and Delta1 signalling through Notch1 and Notch2, Nat Cell Biol 2(8):515-20.

Hilgers, V., Pourquie, O., and Dubrulle, J., 2005, In vivo analysis of mRNA stability using the Tet-Off system in the chicken embryo, Dev Biol 284(2):292-300.

Hirata, H., Bessho, Y., Kokubu, H., Masamizu, Y., Yamada, S., Lewis, J., and Kageyama, R., 2004, Instability of Hes7 protein is crucial for the somite segmentation clock, Nat Genet 36(7):750-4.

Hirata, H., Yoshiura, S., Ohtsuka, T., Bessho, Y., Harada, T., Yoshikawa, K., and Kageyama, R., 2002, Oscillatory expression of the bHLH factor Hes1 regulated by a negative feedback loop, Science 298(5594):840-3.

Holley, S. A., Geisler, R., and Nusslein-Volhard, C., 2000, Control of her1 expression during zebrafish somitogenesis by a delta-dependent oscillator and an independent wave-front activity, Genes Dev 14(13):1678-90.

Holmdahl, D. E., 1925, Experimentelle Untersuchungen uber die Lage der Grenze zwischen primarer und sekundarer Korperentwicklung beim Huhn, Anatomsicher Anzeiger 59:393-396.

Hrabe de Angelis, M., McIntyre, J., 2nd, and Gossler, A., 1997, Maintenance of somite borders in mice requires the Delta homologue DII1, Nature 386(6626):717-21.

Hsieh, J. J., Nofziger, D. E., Weinmaster, G., and Hayward, S. D., 1997, Epstein-Barr virus immortalization: Notch2 interacts with CBF1 and blocks differentiation, J Virol 71(3):1938-45.

Huang, M. T., and Gorman, C. M., 1990, The simian virus 40 small-t intron, present in many common expression vectors, leads to aberrant splicing, Mol Cell Biol 10(4):1805-10.

Huang, R., Zhi, Q., Neubuser, A., Muller, T. S., Brand-Saberi, B., Christ, B., and Wilting, J., 1996, Function of somite and somitocoele cells in the formation of the vertebral motion segment in avian embryos, Acta Anat (Basel) 155(4):231-41.

Huang, R., Zhi, Q., Patel, K., Wilting, J., and Christ, B., 2000, Contribution of single somites to the skeleton and muscles of the occipital and cervical regions in avian embryos, Anat Embryol (Berl) 202(5):375-83.

134 Huppert, S. S., Ilagan, M. X., De Strooper, B., and Kopan, R., 2005, Analysis of Notch Function in Presomitic Mesoderm Suggests a gamma-Secretase-Independent Role for Presenilins in Somite Differentiation, Dev Cell 8(5):677-88.

Irvine, K. D., and Wieschaus, E., 1994, fringe, a Boundary-specific signaling molecule, mediates interactions between dorsal and ventral cells during Drosophila wing development, Cell 79(4):595-606.

Ishibashi, M., Ang, S. L., Shiota, K., Nakanishi, S., Kageyama, R., and Guillemot, F., 1995, Targeted disruption of mammalian hairy and Enhancer of split homolog-1 (HES-1) leads to up-regulation of neural helix-loop-helix factors, premature neurogenesis, and severe neural tube defects, Genes Dev 9(24):3136-48.

Ishikawa, A., Kitajima, S., Takahashi, Y., Kokubo, H., Kanno, J., Inoue, T., and Saga, Y., 2004, Mouse Nkd1, a Wnt antagonist, exhibits oscillatory gene expression in the PSM under the control of Notch signaling, Mech Dev 121(12):1443-53.

Jean, F., Stella, K., Thomas, L., Liu, G., Xiang, Y., Reason, A. J., and Thomas, G., 1998, alpha1-Antitrypsin Portland, a bioengineered serpin highly selective for furin: application as an antipathogenic agent, Proc Natl Acad Sci U S A 95(13):7293-8.

Jen, W. C., Gawantka, V., Pollet, N., Niehrs, C., and Kintner, C., 1999, Periodic repression of Notch pathway genes governs the segmentation of Xenopus embryos, Genes Dev 13(11):1486-99.

Jiang, Y. J., Aerne, B. L., Smithers, L., Haddon, C., Ish-Horowicz, D., and Lewis, J., 2000, Notch signalling and the synchronization of the somite segmentation clock, Nature 408(6811):475-9.

Johnston, S. H., Rauskolb, C., Wilson, R., Prabhakaran, B., Irvine, K. D., and Vogt, T. F., 1997, A family of mammalian Fringe genes implicated in boundary determination and the Notch pathway, Development 124(11):2245-54.

Jouve, C., Palmeirim, I., Henrique, D., Beckers, J., Gossler, A., Ish-Horowicz, D., and Pourquie, O., 2000, Notch signalling is required for cyclic expression of the hairy- like gene HES1 in the presomitic mesoderm, Development 127(7):1421-9.

Kerszberg, M., and Wolpert, L., 2000, A clock and trail model for somite formation, specialization and polarization, J Theor Biol 205(3):505-10.

Kessel, M., and Gruss, P., 1991, Homeotic transformations of murine vertebrae and concomitant alteration of Hox codes induced by retinoic acid, Cell 67(1):89-104.

Keynes, R. J., and Stern, C. D., 1984, Segmentation in the vertebrate nervous system, Nature 310(5980):786-9.

135

Kokubu, C., Heinzmann, U., Kokubu, T., Sakai, N., Kubota, T., Kawai, M., Wahl, M. B., Galceran, J., Grosschedl, R., Ozono, K., and Imai, K., 2004, Skeletal defects in ringelschwanz mutant mice reveal that Lrp6 is required for proper somitogenesis and osteogenesis, Development 131(21):5469-80.

Kraus, F., Haenig, B., and Kispert, A., 2001, Cloning and expression analysis of the mouse T-box gene Tbx18, Mech Dev 100(1):83-6.

Kusumi, K., Mimoto, M. S., Covello, K. L., Beddington, R. S., Krumlauf, R., and Dunwoodie, S. L., 2004, Dll3 pudgy mutation differentially disrupts dynamic expression of somite genes, Genesis 39(2):115-21.

Kusumi, K., Sun, E. S., Kerrebrock, A. W., Bronson, R. T., Chi, D. C., Bulotsky, M. S., Spencer, J. B., Birren, B. W., Frankel, W. N., and Lander, E. S., 1998, The mouse pudgy mutation disrupts Delta homologue Dll3 and initiation of early somite boundaries, Nat Genet 19(3):274-8.

Leimeister, C., Dale, K., Fischer, A., Klamt, B., Hrabe de Angelis, M., Radtke, F., McGrew, M. J., Pourquie, O., and Gessler, M., 2000, Oscillating expression of c- Hey2 in the presomitic mesoderm suggests that the segmentation clock may use combinatorial signaling through multiple interacting bHLH factors, Dev Biol 227(1):91-103.

Leimeister, C., Externbrink, A., Klamt, B., and Gessler, M., 1999, Hey genes: a novel subfamily of hairy- and Enhancer of split related genes specifically expressed during mouse embryogenesis, Mech Dev 85(1-2):173-7.

Li, Y., Fenger, U., Niehrs, C., and Pollet, N., 2003, Cyclic expression of esr9 gene in Xenopus presomitic mesoderm, Differentiation 71(1):83-9.

Lindsell, C. E., Shawber, C. J., Boulter, J., and Weinmaster, G., 1995, Jagged: a mammalian ligand that activates Notch1, Cell 80(6):909-17.

Ma, J., Qian, R., Rausa, F. M., 3rd, and Colley, K. J., 1997, Two naturally occurring alpha2,6-sialyltransferase forms with a single amino acid change in the catalytic domain differ in their catalytic activity and proteolytic processing, J Biol Chem 272(1):672-9.

Mansouri, A., Yokota, Y., Wehr, R., Copeland, N. G., Jenkins, N. A., and Gruss, P., 1997, Paired-related murine homeobox gene expressed in the developing sclerotome, kidney, and nervous system, Dev Dyn 210(1):53-65.

Mara, A., and Holley, S. A., 2007, Oscillators and the emergence of tissue organization during zebrafish somitogenesis, Trends Cell Biol 17(12):593-9.

136

McGrew, M. J., Dale, J. K., Fraboulet, S., and Pourquie, O., 1998, The lunatic fringe gene is a target of the molecular clock linked to somite segmentation in avian embryos, Curr Biol 8(17):979-82.

Meinhardt, H., 1986, Models of Segmentation, in: Somites in Developing Embryos (R. Bellairs, Ede, D.A., Lash, J.W., ed.), Plenum, New York, pp. 179-189.

Menkes, B. a. S., S, 1977, Somitogenesis: regulation potencies, sequence determination and primordial interactions., in: Vertebrate Limb and Somite Morphogenesis (J. R. H. a. M. B. D.A. Ede, ed.), Cambridge University Press, Cambridge, pp. 405-419.

Mikami, T., Ohnaka, Y., Nakamura, A., Kurosaka, A., and Itoh, N., 2001, Radical fringe negatively modulates Notch signaling in postmitotic neurons of the rat brain, Brain Res Mol Brain Res 86(1-2):138-44.

Moloney, D. J., Panin, V. M., Johnston, S. H., Chen, J., Shao, L., Wilson, R., Wang, Y., Stanley, P., Irvine, K. D., Haltiwanger, R. S., and Vogt, T. F., 2000, Fringe is a glycosyltransferase that modifies Notch, Nature 406(6794):369-75.

Momiji, H., and Monk, N. A., 2008, Dissecting the dynamics of the Hes1 genetic oscillator, J Theor Biol 254(4):784-98.

Morales, A. V., Yasuda, Y., and Ish-Horowicz, D., 2002, Periodic Lunatic fringe expression is controlled during segmentation by a cyclic transcriptional enhancer responsive to notch signaling, Dev Cell 3(1):63-74.

Moreno, T. A., and Kintner, C., 2004, Regulation of segmental patterning by retinoic acid signaling during Xenopus somitogenesis, Dev Cell 6(2):205-18.

Morimoto, M., Kiso, M., Sasaki, N., and Saga, Y., 2006, Cooperative Mesp activity is required for normal somitogenesis along the anterior-posterior axis, Dev Biol 300(2):687-98.

Morimoto, M., Takahashi, Y., Endo, M., and Saga, Y., 2005, The Mesp2 transcription factor establishes segmental borders by suppressing Notch activity, Nature 435(7040):354-9.

Mortier, G. R., Lachman, R. S., Bocian, M., and Rimoin, D. L., 1996, Multiple vertebral segmentation defects: analysis of 26 new patients and review of the literature, Am J Med Genet 61(4):310-9.

Munro, S., and Freeman, M., 2000, The notch signalling regulator fringe acts in the Golgi apparatus and requires the glycosyltransferase signature motif DXD, Curr Biol 10(14):813-20.

137

Mustonen, T., Tummers, M., Mikami, T., Itoh, N., Zhang, N., Gridley, T., and Thesleff, I., 2002, Lunatic fringe, FGF, and BMP regulate the Notch pathway during epithelial morphogenesis of teeth, Dev Biol 248(2):281-93.

Neubuser, A., Koseki, H., and Balling, R., 1995, Characterization and developmental expression of Pax9, a paired-box-containing gene related to Pax1, Dev Biol 170(2):701-16.

Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijne, G., 1997, A neural network method for identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites, Int J Neural Syst 8(5-6):581-99.

Niwa, Y., Masamizu, Y., Liu, T., Nakayama, R., Deng, C. X., and Kageyama, R., 2007, The initiation and propagation of Hes7 oscillation are cooperatively regulated by Fgf and notch signaling in the somite segmentation clock, Dev Cell 13(2):298- 304.

Nofziger, D., Miyamoto, A., Lyons, K. M., and Weinmaster, G., 1999, Notch signaling imposes two distinct blocks in the differentiation of C2C12 myoblasts, Development 126(8):1689-702.

Oates, A. C., and Ho, R. K., 2002, Hairy/E(spl)-related (Her) genes are central components of the segmentation oscillator and display redundancy with the Delta/Notch signaling pathway in the formation of anterior segmental boundaries in the zebrafish, Development 129(12):2929-46.

Oates, A. C., Mueller, C., and Ho, R. K., 2005, Cooperative function of deltaC and her7 in anterior segment formation, Dev Biol 280(1):133-49.

Ohtsuka, T., Ishibashi, M., Gradwohl, G., Nakanishi, S., Guillemot, F., and Kageyama, R., 1999, Hes1 and Hes5 as notch effectors in mammalian neuronal differentiation, Embo J 18(8):2196-207.

Oka, C., Nakano, T., Wakeham, A., de la Pompa, J. L., Mori, C., Sakai, T., Okazaki, S., Kawaichi, M., Shiota, K., Mak, T. W., and Honjo, T., 1995, Disruption of the mouse RBP-J kappa gene results in early embryonic death, Development 121(10):3291-301.

Ozbudak, E. M., and Lewis, J., 2008, Notch signalling synchronizes the zebrafish segmentation clock but is not needed to create somite boundaries, PLoS Genet 4(2):e15.

Packard, D. S., Jr., 1976, The influence of axial structures on chick somite formation, Dev Biol 53(1):36-48.

138

Palmeirim, I., Henrique, D., Ish-Horowicz, D., and Pourquie, O., 1997, Avian hairy gene expression identifies a molecular clock linked to vertebrate segmentation and somitogenesis, Cell 91(5):639-48.

Panin, V. M., Papayannopoulos, V., Wilson, R., and Irvine, K. D., 1997, Fringe modulates Notch-ligand interactions, Nature 387(6636):908-12.

Panin, V. M., Shao, L., Lei, L., Moloney, D. J., Irvine, K. D., and Haltiwanger, R. S., 2002, Notch ligands are substrates for protein O-fucosyltransferase-1 and Fringe, J Biol Chem 277(33):29945-52.

Pinson, K. I., Brennan, J., Monkley, S., Avery, B. J., and Skarnes, W. C., 2000, An LDL- receptor-related protein mediates Wnt signalling in mice, Nature 407(6803):535- 8.

Prince, V. E., Holley, S. A., Bally-Cuif, L., Prabhakaran, B., Oates, A. C., Ho, R. K., and Vogt, T. F., 2001, Zebrafish lunatic fringe demarcates segmental boundaries, Mech Dev 105(1-2):175-80.

Rancourt, S. L., and Rancourt, D. E., 1997, Murine subtilisin-like proteinase SPC6 is expressed during embryonic implantation, somitogenesis, and skeletal formation, Dev Genet 21(1):75-81.

Rickmann, M., Fawcett, J. W., and Keynes, R. J., 1985, The migration of neural crest cells and the growth of motor axons through the rostral half of the chick somite, J Embryol Exp Morphol 90:437-55.

Rida, P. C., Le Minh, N., and Jiang, Y. J., 2004, A Notch feeling of somite segmentation and beyond, Dev Biol 265(1):2-22.

Riddle, R. D., Johnson, R. L., Laufer, E., and Tabin, C., 1993, Sonic hedgehog mediates the polarizing activity of the ZPA, Cell 75(7):1401-16.

Rifes, P., Carvalho, L., Lopes, C., Andrade, R. P., Rodrigues, G., Palmeirim, I., and Thorsteinsdottir, S., 2007, Redefining the role of ectoderm in somitogenesis: a player in the formation of the fibronectin matrix of presomitic mesoderm, Development 134(17):3155-65.

Roberts, A. P., Conner, A. N., Tolmie, J. L., and Connor, J. M., 1988, Spondylothoracic and spondylocostal dysostosis. Hereditary forms of spinal deformity, J Bone Joint Surg Br 70(1):123-6.

139 Roebroek, A. J., Umans, L., Pauli, I. G., Robertson, E. J., van Leuven, F., Van de Ven, W. J., and Constam, D. B., 1998, Failure of ventral closure and axial rotation in embryos lacking the proprotein convertase Furin, Development 125(24):4863-76.

Saga, Y., Hata, N., Koseki, H., and Taketo, M. M., 1997, Mesp2: a novel mouse gene expressed in the presegmented mesoderm and essential for segmentation initiation, Genes Dev 11(14):1827-39.

Saga, Y., and Takeda, H., 2001, The making of the somite: molecular events in vertebrate segmentation, Nat Rev Genet 2(11):835-45.

Saito, T., Miyoshi, E., Sasai, K., Nakano, N., Eguchi, H., Honke, K., and Taniguchi, N., 2002, A secreted type of beta 1,6-N-acetylglucosaminyltransferase V (GnT-V) induces tumor angiogenesis without mediation of glycosylation: a novel function of GnT-V distinct from the original glycosyltransferase activity, J Biol Chem 277(19):17002-8.

Sambrook, J. a. R., DW, 2001, Molecular Cloning: A Laboratory Manual.

Sawada, A., Fritz, A., Jiang, Y. J., Yamamoto, A., Yamasu, K., Kuroiwa, A., Saga, Y., and Takeda, H., 2000, Zebrafish Mesp family genes, mesp-a and mesp-b are segmentally expressed in the presomitic mesoderm, and Mesp-b confers the anterior identity to the developing somites, Development 127(8):1691-702.

Sawada, A., Shinya, M., Jiang, Y. J., Kawakami, A., Kuroiwa, A., and Takeda, H., 2001, Fgf/MAPK signalling is a crucial positional cue in somite boundary formation, Development 128(23):4873-80.

Schnell, S., and Maini, P. K., 2000, Clock and induction model for somitogenesis, Dev Dyn 217(4):415-20.

Serth, K., Schuster-Gossler, K., Cordes, R., and Gossler, A., 2003, Transcriptional oscillation of lunatic fringe is essential for somitogenesis, Genes Dev 17(7):912- 25.

Shifley, E. T., and Cole, S. E., 2007, The vertebrate segmentation clock and its role in skeletal birth defects, Birth Defects Res C Embryo Today 81(2):121-33.

Shifley, E. T., Vanhorn, K. M., Perez-Balaguer, A., Franklin, J. D., Weinstein, M., and Cole, S. E., 2008, Oscillatory lunatic fringe activity is crucial for segmentation of the anterior but not posterior skeleton, Development 135(5):899-908.

Shimizu, K., Chiba, S., Saito, T., Kumano, K., Takahashi, T., and Hirai, H., 2001, Manic fringe and lunatic fringe modify different sites of the Notch2 extracellular region, resulting in different signaling modulation, J Biol Chem 276(28):25753-8.

140

Sparrow, D. B., Chapman, G., Wouters, M. A., Whittock, N. V., Ellard, S., Fatkin, D., Turnpenny, P. D., Kusumi, K., Sillence, D., and Dunwoodie, S. L., 2006, Mutation of the LUNATIC FRINGE gene in humans causes spondylocostal dysostosis with a severe vertebral phenotype, Am J Hum Genet 78(1):28-37.

Sparrow, D. B., Guillen-Navarro, E., Fatkin, D., and Dunwoodie, S. L., 2008, Mutation of Hairy-and-Enhancer-of-Split-7 in humans causes spondylocostal dysostosis, Hum Mol Genet 17(23):3761-6.

Stanley, P., 2007, Regulation of Notch signaling by glycosylation, Curr Opin Struct Biol 17(5):530-5.

Subramaniam, S., 1998, The Biology Workbench--a seamless database and analysis environment for the biologist, Proteins 32(1):1-2.

Suriben, R., Fisher, D. A., and Cheyette, B. N., 2006, Dact1 presomitic mesoderm expression oscillates in phase with Axin2 in the somitogenesis clock of mice, Dev Dyn 235(11):3177-83.

Szeto, D. P., and Kimelman, D., 2006, The regulation of mesodermal progenitor cell commitment to somitogenesis subdivides the zebrafish body musculature into distinct domains, Genes Dev 20(14):1923-32.

Takahashi, Y., Koizumi, K., Takagi, A., Kitajima, S., Inoue, T., Koseki, H., and Saga, Y., 2000, Mesp2 initiates somite segmentation through the Notch signalling pathway, Nat Genet 25(4):390-6.

Takahashi, Y., Yasuhiko, Y., Kitajima, S., Kanno, J., and Saga, Y., 2007, Appropriate suppression of Notch signaling by Mesp factors is essential for stripe pattern formation leading to segment boundary formation, Dev Biol 304(2):593-603.

Takikawa, K., Haga, N., Maruyama, T., Nakatomi, A., Kondoh, T., Makita, Y., Hata, A., Kawabata, H., and Ikegawa, S., 2006, Spine and rib abnormalities and stature in spondylocostal dysostosis, Spine 31(7):E192-7.

Tam, P. P., 1981, The control of somitogenesis in mouse embryos, J Embryol Exp Morphol 65 Suppl:103-28.

Tam, P. P., 1986, A study of the pattern of prospective somites in the presomitic mesoderm of mouse embryos, J Embryol Exp Morphol 92:269-85.

Taylor, N. A., Van De Ven, W. J., and Creemers, J. W., 2003, Curbing activation: proprotein convertases in homeostasis and pathology, Faseb J 17(10):1215-27.

141 Truett, G. E., Heeger, P., Mynatt, R. L., Truett, A. A., Walker, J. A., and Warman, M. L., 2000, Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide and tris (HotSHOT), Biotechniques 29(1):52, 54.

Turnpenny, P. D., Whittock, N., Duncan, J., Dunwoodie, S., Kusumi, K., and Ellard, S., 2003, Novel mutations in DLL3, a somitogenesis gene encoding a ligand for the Notch signalling pathway, cause a consistent pattern of abnormal vertebral segmentation in spondylocostal dysostosis, J Med Genet 40(5):333-9.

Vermot, J., and Pourquie, O., 2005, Retinoic acid coordinates somitogenesis and left- right patterning in vertebrate embryos, Nature 435(7039):215-20.

Wahl, M. B., Deng, C., Lewandoski, M., and Pourquie, O., 2007, FGF signaling acts upstream of the NOTCH and WNT signaling pathways to control segmentation clock oscillations in mouse somitogenesis, Development 134(22):4033-41.

Weinmaster, G., and Kintner, C., 2003, Modulation of notch signaling during somitogenesis, Annu Rev Cell Dev Biol 19:367-95.

Whittock, N. V., Ellard, S., Duncan, J., de Die-Smulders, C. E., Vles, J. S., and Turnpenny, P. D., 2004a, Pseudodominant inheritance of spondylocostal dysostosis type 1 caused by two familial delta-like 3 mutations, Clin Genet 66(1):67-72.

Whittock, N. V., Sparrow, D. B., Wouters, M. A., Sillence, D., Ellard, S., Dunwoodie, S. L., and Turnpenny, P. D., 2004b, Mutated MESP2 causes spondylocostal dysostosis in humans, Am J Hum Genet 74(6):1249-54.

Wilson, V., Rashbass, P., and Beddington, R. S., 1993, Chimeric analysis of T (Brachyury) gene function, Development 117(4):1321-31.

Wilusz, C. J., and Wilusz, J., 2004, Bringing the role of mRNA decay in the control of gene expression into focus, Trends Genet 20(10):491-7.

Wu, J. Y., Wen, L., Zhang, W. J., and Rao, Y., 1996, The secreted product of Xenopus gene lunatic Fringe, a vertebrate signaling molecule, Science 273(5273):355-8.

Xiang, Y., Molloy, S. S., Thomas, L., and Thomas, G., 2000, The PC6B cytoplasmic domain contains two acidic clusters that direct sorting to distinct trans-Golgi network/endosomal compartments, Mol Biol Cell 11(4):1257-73.

Xu, J., Norton, C. R., and Gridley, T., 2006, Not all lunatic fringe null female mice are infertile, Development 133(4):579; author reply 579-80.

142 Yang, L. T., Nichols, J. T., Yao, C., Manilay, J. O., Robey, E. A., and Weinmaster, G., 2005, Fringe glycosyltransferases differentially modulate Notch1 proteolysis induced by Delta1 and Jagged1, Mol Biol Cell 16(2):927-42.

Young, W. W., Jr., 2004, Organization of Golgi glycosyltransferases in membranes: complexity via complexes, J Membr Biol 198(1):1-13.

Zhang, N., and Gridley, T., 1998, Defects in somite formation in lunatic fringe-deficient mice, Nature 394(6691):374-7.

Zheng, M., Streck, R. D., Scott, R. E., Seidah, N. G., and Pintar, J. E., 1994, The developmental expression in rat of proteases furin, PC1, PC2, and carboxypeptidase E: implications for early maturation of proteolytic processing capacity, J Neurosci 14(8):4656-73.

143