Genetic Analysis Reveals Dual Roles of Speckle-Type

The Pennsylvania State University

The Graduate School

Eberly College of Science

The Huck Institutes of the Life Sciences

GENETIC ANALYSIS REVEALS DUAL ROLES OF

SPECKLE-TYPE POZ PROTEIN

IN HEDGEHOG SIGNALING AND MOUSE DEVELOPMENT

A Dissertation in

The Molecular, Cellular and Integrative Biosciences

Hongchen Cai

 2017 Hongchen Cai

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

May 2017

The dissertation of Hongchen Cai was reviewed and approved* by the following:

Aimin Liu

Associate Professor of Biology

Dissertation Advisor

Chair of Committee

Andrea M. Mastro

Professor of Microbiology and Cell Biology

Douglas Cavener

Professor of Biology

Verne M. Willaman Dean of Eberly College of Science

Yingwei Mao

Assistant Professor of Biology

Melissa Rolls

Associate Professor of Biochemistry and Molecular Biology

Chair of the Molecular, Cellular and Integrative Biosciences Graduate Program

* Signatures are on file in the Graduate School

iii

ABSTRACT

Hedgehog (Hh) signaling plays essential roles in animal development; and dysregulation of Hh pathway often results in diseases such as cancer. Glioma-associated oncogene homolog

(Gli) transcription factors interpret Hh signaling by serving as both activators and repressors.

Previous studies in Drosophila, Xenopus and mammalian cell culture implicated Speckle-type

POZ Protein (Spop), a E3 ubiquitin ligase component, as a potential regulator of Hh signaling by targeting Gli2 and Gli3 for ubiquitination; but the function of Spop remained largely unknown in mammals.

In this study, we present the genetics analysis of Spop loss-of-function mutants, and show that Spop plays an important role in bone development. By ubiquitination of Gli3 and its repressor form (Gli3R), Spop downregulates Gli3R level and promotes Indian Hedgehog (Ihh) signaling, which stimulates both chondrogenesis and osteogenesis. Since loss of Spop results in neonatal lethality, we conditionally deleted Spop with a limb-specific Cre, and found that these mutant mice exhibit lower bone mass, a symptom of osteoporosis and osteopenia. Therefore, loss of Spop may be associated with genetic predisposition to osteoporosis.

Furthermore, we showed that Spop inhibits Sonic Hedgehog (Shh) signaling and specification of ventral-most cell fates in the spinal cord by degradation of Gli3 activator. Since

Gli1, Gli2 and Gli3 activators play redundant roles in activating Shh target genes, the negative role of Spop is only revealed in Gli2 mutant or Sufu mutant background. Furthermore, Spop- induced degradation of Gli3 at least in part accounts for the destabilization of Gli proteins and decreased Hh target gene expression in ventral-most neural tube of Sufu mutant, suggesting a positive role of Sufu in Shh signaling by antagonizing Spop-mediated Gli3 degradation.

Therefore, our study indicates both positive and negative roles of Spop in Hh signaling by

iv degradation of Gli3 but not Gli2, which depicts a multifaceted role of Spop that was not posited by previous in vitro studies.

Finally, we present in vivo evidence that Spop-like (Spopl), a poorly-studied homolog of

Spop, does not play a redundant role with Spop in bone development or spinal cord patterning.

The function and substrate of Spopl remains an interesting question to be addressed in the future.

TABLE OF CONTENTS

LIST OF FIGURES ...... x

LIST OF TABLES ...... xiii

LIST OF ABBREVIATIONS ...... xiv

ACKNOWLEDGEMENTS ...... xv

Chapter 1 Introduction ...... 1

1.1 The Hh signaling cascade ...... 1

1.1.1 Hh signaling in Drosophila ...... 2

1.1.2 Primary cilium organizes Hh signaling molecules in mammals ...... 6

1.1.3 Hh signaling cascade in mammals...... 8

1.1.4 Gli repression and activation are mediated by a series of events ...... 13

1.2 Hh signaling plays critical roles in animal development ...... 19

1.2.1 Shh signaling regulates pattern formation of spinal neural tube ...... 19

1.2.2 Ihh signaling regulates skeletal development ...... 21

1.3 Spop targets Gli2 and Gli3 for ubiquitination and proteasomal degradation ...... 25

1.3.1 Spop is part of an E3 ubiquitin ligase complex ...... 25

1.3.2 Spop targets Gli2/3 proteins for degradation ...... 30

1.3.3 Spop-like is a potential regulator of Gli proteins ...... 31

1.4 Hypotheses and specific aims ...... 32

1.4.1 Specific aim 1 will test the hypothesis that Spop and Spopl regulate skeletal

development by inducing Gli2 and Gli3 ubiquitination and subsequent degradation...... 32

1.4.2 Specific aim 2 will test the hypothesis that Spop and Spopl regulate spinal cord

patterning by ubiquitination of Gli2 and Gli3...... 34

Chapter 2 Materials and methods ...... 36

2.1 Genetics Analysis ...... 36

2.1.1 Mouse strains and genotyping strategy ...... 36

2.1.2 β-galactosidase (lacZ) staining ...... 42

2.1.3 Alcian blue and Alizarin red staining ...... 44

2.1.4 RNA in situ hybridization (on cryosections) ...... 46

2.1.5 Quantitative Reverse Transcriptase Polymerase Chain Reaction (qRT-PCR) ...... 51

2.1.6 Hematoxylin and Eosin (H&E) stain ...... 53

2.1.7 Von Kossa stain ...... 53

2.1.8 Immunohistochemical staining ...... 54

2.1.9 Micro-Computed Tomography (µCT) ...... 57

2.2 Biochemistry, Molecular and Cell Biology ...... 58

2.2.1 Molecular Cloning ...... 58

2.2.2 site-directed mutagenesis ...... 60

2.2.3 Cell Culture and Transfection ...... 62

2.2.4 Immunoblot ...... 63

vii

2.2.5 Immunoprecipitation (IP) and Co-IP ...... 65

2.2.6 In vivo Ubiquitination Assay ...... 67

2.2.7 Immunocytochemistry ...... 68

Chapter 3 Requirement of Spop for the skeletal development and Ihh signaling ...... 70

3.1 Loss of Spop induces neonatal lethality ...... 72

3.1.1 Three Spop mutant alleles were generated ...... 72

3.1.2 Loss of Spop induced neonatal lethality ...... 74

3.2 Spop regulates skeletal development ...... 77

3.2.1 Spop is highly expressed in developing cartilage and bones...... 77

3.2.2 Ossification is defective in Spop null mutants ...... 78

3.2.3 Spop regulates hypertrophic differentiation of chondrocytes ...... 81

3.2.4 Spop promotes bone formation and osteoblast differentiation ...... 83

3.3 Spop promotes skeletal development by degrading Gli3 repressors ...... 85

3.3.1 Loss of Spop results in impaired Ihh signaling ...... 85

3.3.2 Spop targets Gli3R for ubiquitination and degradation ...... 87

3.3.3 The increase in Gli3 underlies the ossification defects in Spop mutants ...... 90

3.4 Loss of Spop induces brachydactyly and osteopenia ...... 92

3.4.1 Defective hypertrophic differentiation of chondrocyte results in brachydactyly ...... 92

3.4.2 Limb-specific Spop conditional mutants exhibit osteopenia ...... 96

3.4.3 Gli3R dosage reduction rescues brachydactyly and osteopenia in SpopcKO mutants ... 98

viii

3.5 Spop-like does not regulate skeletal development ...... 99

3.6 Conclusion ...... 102

Chapter 4 Roles of Spop in spinal cord neural patterning and Shh signaling ...... 104

4.1 Spop inhibits Shh signaling by downregulating Gli3 activator activity ...... 106

4.1.1 Spop is expressed at a low level in the neural tube ...... 106

4.1.2 Gli3 is stabilized in Spop mutant embryos ...... 107

4.1.3 The dorsal/ventral patterning of the Spop mutant neural tube was normal ...... 110

4.1.4 Loss of Spopl does not alter spinal cord patterning in wild type or Spop mutants .... 113

4.1.5 Loss of Spop restores normal patterning in Gli2 mutant spinal cord ...... 115

4.1.6 The ventral neural tube was properly patterned in Spop;Gli3 double mutants ...... 120

4.2 Sufu plays a positive role in spinal cord neural patterning by antagonizing Spop ...... 125

4.2.1 Loss of Spop exacerbates neural tube patterning defects in Sufu mutants ...... 126

4.2.2 Loss of Spop rescues the loss of ventral-most cell fates in the Gli1;Sufu double

mutants...... 129

4.3 Conclusion ...... 135

Chapter 5 Discussion ...... 140

5.1 The molecular mechanisms underlying Spop regulation of Gli proteins ...... 140

5.1.1 Conclusion ...... 140

5.1.2 Spop may exhibit a preference of substrate ...... 141

5.1.3 Spop and Btrc potentially induces ubiquitination on different lysine residues in Gli3

...... 144

5.2 Other potential functions of Spop ...... 146

5.2.1 Spop mutants unlikely die of respiratory or renal failure ...... 146

5.2.2 Spop mutants exhibit cardiovascular defect ...... 150

5.2.3 Spop potentially promotes carcinogenesis in bone tumors ...... 153

References ...... 157

LIST OF FIGURES

Figure 1-1. Hh signaling in Drosophila...... 3

Figure 1-2. The mammalian and Drosophila Hh signaling diverge at the primary cilium...... 7

Figure 1-3. The Hh signaling cascade in mammals...... 10

Figure 1-4. Modifications and binding motifs of Gli proteins...... 15

Figure 1-5. Shh patterns spinal cord by regulating the expression of various toolkit genes...... 20

Figure 1-6. Ihh signaling regulates endochondral ossification...... 23

Figure 1-7. Recurrent Spop mutations are identified in cancers...... 27

Figure 1-8. Spopl and Spop share high sequence identity...... 32

Figure 2-1. Schematics of mouse loss-of-function mutant and reporter strains...... 37

Figure 2-2. Synthesis of antisense RNA probes...... 48

Figure 3-1. Spop knockout strategy...... 72

Figure 3-2. Spop protein are not detected in SpoplacZKI and Spop∆Ex embryos...... 73

Figure 3-3. The growth of Spop mutants are attenuated after birth...... 77

Figure 3-4. Spop is highly expressed in the developing skeleton...... 78

Figure 3-5. Loss of Spop induces wide-spread ossification defect...... 79

Figure 3-6. Metatarsals are not calcified before death of pups...... 80

Figure 3-7. Loss of Spop disrupts hypertrophic differentiation of chondrocytes...... 82

Figure 3-8. Loss of Spop impairs bone formation and osteoblast differentiation...... 84

Figure 3-9. Loss of Spop induces Ihh signaling...... 86

Figure 3-10. Gli3 is stabilized in Spop mutant...... 87

Figure 3-11. Spop targets Gli31-700 for ubiquitination and degradation...... 89

Figure 3-12. Gli3 heterozygosity rescues the ossification defect in Spop mutants...... 90

Figure 3-13. Gli3, but not Gli2, is epistatic to Spop in skeletal development...... 92

Figure 3-14. Prx1-Cre induces recombination in multiple tissues including limb bud...... 93

Figure 3-15. Tissue-specific ablation of Spop leads to brachydactyly...... 94

Figure 3-16. Defective hypertrophic differentiation accounts for the shortness of digit bones. .. 95

Figure 3-17. SpopcKO exhibits osteopenia...... 96

Figure 3-18. Osteoblast, but not osteoclast, is affected in Spop mutants...... 98

Figure 3-19. The bone defects of limb-specific Spop mutants result from increased Gli3...... 99

Figure 3-20. Schematic illustration of Spopl mutant allele...... 100

Figure 3-21. Loss of Spopl does not lead to obvious skeletal defects...... 101

Figure 3-22. Model: Spop regulates skeletal development by degrading Gli3R...... 102

Figure 4-1. Spop-lacZ shows a low-level expression in the spinal neural tube at E9.5...... 107

Figure 4-2. Loss of Spop leads to stabilization of Gli3 but not Gli2...... 109

Figure 4-3. Removal of Spop does not alter the neural patterning...... 111

Figure 4-4. Ptch1 and Gli1 expression in mutant embryos...... 112

Figure 4-5. Spopl mutant and Spopl;Spop double mutant exhibit normal spinal cord patterning.

...... 115

Figure 4-6. Floor plate and pV3 are rescued in Spop;Gli2 double mutants...... 117

Figure 4-7. Floor plate and pV3 are rescued in Spop;Gli2 double mutants...... 119

Figure 4-8. Spop;Gli3 double mutant resembles Gli3 mutant morphologically and patterning- wise...... 123

Figure 4-9. Loss of Spop does not alter the neural patterning in Gli3 mutant embryos...... 124

Figure 4-10. Moderate increase in the level of Gli3, but not Gli2, in Spop;Sufu double mutant embryos...... 126

xii

Figure 4-11. Loss of Spop promotes the ventralization of Sufu mutant spinal cord...... 128

Figure 4-12. Loss of Spop rescues the floor plate and V3 interneuron progenitor fates in

Gli1;Sufu double mutant embryos...... 133

Figure 4-13. Spop overexpression alters the subcellular localization of Gli3...... 135

Figure 4-14. A model of the roles of Spop in spinal cord patterning...... 137

Figure 5-1. Spop plays negative and positive roles in Hh signaling by Gli3 degradation...... 140

Figure 5-2. Two models depicting Gli3 ubiquitination catalyzed by Spop and Btrc...... 145

Figure 5-3. Spop mutants exhibit exencephaly at a low penetrance...... 147

Figure 5-4. Loss of Spop does not cause cleft palate...... 148

Figure 5-5. Loss of Spop does not result in major lung defect...... 149

Figure 5-6. Loss of Spop does not cause major kidney defects...... 150

Figure 5-7. Loss of Spop affects heart development...... 151

xiii

LIST OF TABLES Table 1-1. Spop substrates and their functions affected by Spop-induced degradation...... 27

Table 2-1 Genotyping primers and condition for mouse strains...... 40

Table 2-2 RNA probes for in situ hybridization...... 50

Table 2-3. qRT-PCR primers...... 52

Table 2-4. Primary antibodies for immunohistochemistry...... 55

Table 2-5. Primers for subcloning into expression vectors...... 59

Table 2-6. Primers for site-directed mutagenesis of Gli2...... 60

Table 2-7. Primary antibodies for immunoblot...... 64

Table 3-1. Number of embryos or pups that survive to indicated stages from a crossing between heterozygotes (SpoplacZKI/+ × SpoplacZKI/+)...... 74

Table 3-2. Number of embryos or pups that survive to given stages from a crossing between heterozygotes (Spopflox/+ × Spopflox/+) ...... 75

Table 3-3. Number of embryos or pups that survive to given stages from a crossing between heterozygotes (SpopΔEx/+ × SpopΔEx/+) ...... 75

Table 5-1. Hh pathway mutations identified in Ewing tumors...... 154

Table 5-2. Missense mutations that potentially destabilize Gli3 by ubiquitination...... 155

xiv

LIST OF ABBREVIATIONS

Btrc β-transducin repeat containing protein ci cubitus interruptus

CK1 Casein kinase 1

Cos2 Costal 2

Cul1/3 Cullin 1 or 3

Gli1/2/3 Glioma-associated oncogene homolog 1, 2 or 3

Gli3A Gli3 activator

Gli3FL Full length Gli3 (may be inactive or active)

Gli3R Gli3 Repressor

GSK3β Glycogen synthase kinase 3β

Hh Hedgehog

Ift Intraflagellar transport

Ihh Indian Hedgehog

PBS Phosphate Buffered Saline

PFA paraformaldehyde

PKA Protein kinase A

Ptch1 Patched homolog 1

Shh Sonic Hedgehog

Smo Smoothened

Spop Speckle-type POZ Protein

Spopl Speckle-type POZ Protein-like

Sufu Suppressor of Fused

UTR Untranslated region

ACKNOWLEDGEMENTS

First, I would like to express my sincere gratitude to my advisor Dr. Aimin Liu, for his enormous help with my career development, and for his offers of the tremendous opportunities. From Dr. Liu, I received spectacular training on project design, scientific writing, critical thinking, and lab management.

What’s more important, Dr. Liu opened up an amazing door of developmental biology to me, revealing the elegance of genetics analysis and Hedgehog signaling, which encourages me to pursue a career in in vivo models.

I also want to express great appreciation to my committee, Dr. Andrea Mastro, Dr. Douglas

Cavener and Dr. Yingwei Mao, for their generous suggestions and technical supports. I thank our collaborators Dr. Neil Sharkey and Noriaki Okita (Department of Kinesiology) for kindly helping us with the micro-CT experiment.

Furthermore, I would like to give my special thanks to my lab colleagues Huiqing Zeng, Xuan

Ye, Jinling Liu and Xiaoxu Yang. Huiqing and Xuan actively participated in my experimental design and always shared ideas, and I won’t be able to accomplish the lab duties without everybody. In addition, I received technical supports from John Cantolina (Microscopy and Cytometry Facility), Walter Jackson

(Mastro lab), Jingjie Hu, Jianwen Wei, Siying Zhu (Cavener lab), Yijing Zhou, Fengping Dong (Mao lab) and Xin Tang (Gong Chen lab). Moreover, I learnt a lot in the Center for Molecular Investigation of

Neurological Disorders community. Thanks to all my friends for their support!

Finally, I thank my family for their understanding and support. I would like to dedicate this work to my wife Liting Guo, my daughter Claire, and my parents, for making my life complete.

Chapter 1

Introduction

Hedgehog (Hh) signaling pathway is one of the essential signaling pathways that regulate animal development 1. Disruption of Hh signaling contributes to various diseases. Previous in vitro studies identified Speckle-type POZ Protein (Spop) as a novel regulator of Hh pathway effector Ci/Gli proteins and many other substrates; and Spop mutation is frequently identified in cancers. However, the function of Spop in mammalian development remained poorly elucidated.

In this work, we identified the roles of Spop in mouse bone development and spinal cord neural patterning, and demonstrated that loss of Spop may cause diseases and anomaly such as osteopenia and brachydactyly.

1.1 The Hh signaling cascade

Hh signaling pathway is well demonstrated in Drosophila (Figure 1-1) and mouse models

(Figure 1-3). In Drosophila, Hh signal transmits from secreted Hh proteins, to receptor Patched

(Ptc), a seven-span transmembrane protein Smoothened (Smo), and finally is interpreted by transcription factor Cubitus Interruptus (ci). Similarly, the mammalian Hh signaling cascade consists of Hh, receptors Patched homolog (Ptch), Smo, and the transcription factors Glioma- associated oncogene homolog (Gli) (reviewed by Ye and Liu, 2011). Although Hh signaling components are mostly conserved from Drosophila to mammals, flies and mammals evolved some different mechanisms in relaying the signal 2.

1.1.1 Hh signaling in Drosophila

Hh was initially identified as one of the segment polarity genes in Drosophila 3. In the wing imaginal disc, Hh is secreted by posterior compartment and acts cell nonautonomously to regulate anterior cells 4,5. After the amino-terminal signal peptide is cleaved, an autoproteolytic cleavage between amino acid residues G257 and C258 further gives rise to a 18kDa amino- terminus (Hh-Np) that is associated to the cell, and a 25kDa carboxyl-terminal domain that is released 6,7. Hh-Np is the active species that triggers pathway activation, and the carboxyl- terminus is indispensable for the autoproteolytic processing 6,7. A palmitoylation at C85 residue of Hh, which is catalyzed by an acyltransferase Skinny Hedgehog, is required for Hh activity 8. In addition, the carboxyl-terminal domain tethers a cholesterol moiety to G257 of Hh-Np during the autoprocessing 9, which retains the Hh on the cell and confines Hh signaling activation in a smaller region 10. Dispatched, a sterol-sensing transmembrane protein, extracts the cholesterol modified

Hh-Np and releases it from signaling cells 10.

Patched (Ptc), initially identified as a segment polarity gene in Drosophila 11, encodes a twelve-span membrane receptor for Hh that is conserved from Drosophila to mammals 12. Unlike receptors in most pathways, Ptc acts as a negative regulator in Hh pathway; and Hh abolishes the repressive activity of Ptc (Figure 1-1) 13. Furthermore, Ptc induces the endocytosis of Hh to control the Hh gradient, although this endocytosis does not play a major role in signal transduction 14,15.

Hh induces Ptc internalization and degradation, and relieves Ptc inhibition of Smo 16. Interestingly,

Ptc is a Hh target gene, resulting in a negative feedback that resists subtle changes in Ptc level 16,17.

Smoothened (Smo), a protein that relays the Hh signal, consists of an extracellular amino- terminal Cysteine-Rich Domain, seven transmembrane domains, and a cytoplasmic carboxyl- terminal tail 18,19. Smo does not function as a coreceptor of Hh 20. In the absence of Hh, Ptc

induces Smo turnover 16. Multi-ubiquitination in the cytoplasmic carboxyl-terminal tail of Smo leads to its endocytosis and lysosomal degradation 21,22. Hh relieves the inhibition of Smo by Ptc, resulting in the stabilization and cell-surface accumulation of Smo 16. The inverse correlation of

Ptc and Smo localization indicates that Ptc likely represses Smo indirectly 16. Mutations in the sterol-sensing domain (second to sixth transmembrane segments) of Ptc attenuates its repression of Smo while retaining its interaction with Hh 23,24, suggesting the involvement of sterols in the function of Ptc (see Chapter 1.1.3 for more mechanistic studies in mammals).

Figure 1-1. Hh signaling in Drosophila.

(A) In the absence of Hh, Ptc inhibits Smo and induces its endocytosis and lysosomal degradation. Cos2 and Sufu sequester ci in cytoplasm, and a Hh signalosome complex that comprises a scaffold protein

Cos2 and multiple kinases trigger the phosphorylation of ci and subsequent ubiquitination by Cul1/slimb.

Ci is then proteolytically processed into the repressor form and represses Hh target gene expression.

(B) When Hh binds to Ptc, both proteins are endocytosed. Smo is activated by dimerization and phosphorylation in the cytoplasmic carboxyl-terminal tail. Subsequently, activated kinase Fused phosphorylates Cos2 and Sufu, leading to ci activation.

Drosophila Smo functions in the plasma membrane 25. Forced Smo localization to the plasma membrane activates Hh signaling, while trapping Smo in the endoplasmic reticulum blocks

Hh signaling 25. The kinesin-related protein Costal 2 (Cos2) causes plasma membrane accumulation of Smo in response to Hh reception 26.

The carboxyl-terminal cytoplasmic tail of Smo accounts for the activation of downstream signaling cascade 27. The arginine clusters in the cytoplasmic tail mediate intramolecular electrostatic interaction that arrests Smo in a closed inactive conformation 27. The carboxyl- terminal cytoplasmic tail of Smo is phosphorylated by the kinases PKA, CK1α, CK2α/β, GRK2, and Gilgamesh (Gish) (Figure 1-1B) 28–31. PKA and CK1 phosphorylation abolishes intramolecular electrostatic interaction and induces an open active conformation 27, and phosphorylation clusters also recruit Ubiquitin-specific protease 8 (USP8) to deubiquitinate Smo which prevents its endocytosis 21,22. Dephosphorylation by Protein Phosphatase 1 (PP1), PP2A and PP4 inhibits Smo activity 32,33. Hh gradient blocks phosphatase activity and triggers progressive phosphorylation in Smo by kinases, leading to conformational switch and dimerization of Smo cytoplasmic tails, hence translates the Hh gradient into graded responses 27,28,33. High level

Hh signaling also triggers oligomerization of Smo in the lipid rafts 34.

Cubitus interruptus (ci), a segment polarity gene, encodes a transcription factor that interprets Hh signal 17,35,36. In the anterior compartment of the wing imaginal discs where Hh is absent or only present at a minimum level, ci is ubiquitinated by an E3 ubiquitin ligase slimb and partially degraded in 26S proteasome to a 75kDa transcriptional repressor 36,37. PKA, CK1 and

GSK3 phosphorylate ci and primes ci for Slimb-mediated ci processing 38–40. Cos2 and Sufu both retain ci in the cytoplasm, hence attenuate the pathway activation 41,42.

Hh abolishes the production of ci repressor and activates ci through different mechanisms

43. Derepression of ci may result from activation G protein Gαi by Smo, which leads to lower cAMP level and decreased PKA activity 44. An increase in Smo abundance may also cause the switch of PKA substrate from ci to Smo 45, which compromises the interaction between PKA and ci 46. Furthermore, Hh dissociates CK1α and GSK3 from Cos2 and ci 46.

The activation of ci is triggered by a Hh signalosome complex containing Smo, Cos2,

Fused, PKA and CK1, which is assembled at Smo cytoplasmic tail by Cos2 26,47–49. The interaction of ci and Smo is abolished by Hh 49. PKA is constantly associated with Fused 45. Fused is activated by autophosphorylation and subsequent CK1 phosphorylation, and subsequently phosphorylates

Cos2 and Sufu, leading to ci activation 26,50. With or without Hh, Sufu does not interact with Smo or Cos2, suggesting that Sufu is not present in the Hh signalosome complex 26,51.

Apart from the partial degradation mediated by Cul1/slimb E3 ligase, the stability of ci is also regulated by a Cul3-based E3 ubiquitin ligase 52. Hh-induced MATH and BTB domain containing protein (also named roadkill) (hib/rdx), the substrate-recognition subunit of this Cul3- based E3 ligase, is expressed in the posterior compartment of wing disc and targets ci for ubiquitination 53,54. Unlike slimb-mediated ubiquitination which leads to partial degradation, hib/rdx triggers ci destruction 53,54. CK1 phosphorylates ci in the hib/rdx-binding sites, which blocks ci turnover 55. CK2 also stabilizes ci in a slimb-independent mechanism 31, possibly through antagonism of hib/rdx.

Finally, Hh signaling results in expression of Hh target genes, such as decapentaplegic, hence regulates the anterior/posterior patterning of wing pouch 56–58.

1.1.2 Primary cilium organizes Hh signaling molecules in mammals

The primary cilium is an antenna-like organelle that plays an essential role in mammalian signaling pathways (reviewed by Ye and Liu, 2011). During interphase, microtubules grow from the centriole and form the axoneme, which serves as the backbone for the primary cilium and railway for protein trafficking 59. A diffusion barrier containing Septin 2 is assembled at the base of the cilium and limits the entrance and exit of membrane proteins 60. Diffusion of lipid at the ciliary base is also limited, allowing exclusion of PIP2 from the cilium 61. Similar to the nuclear pore complex, a ciliary pore complex that contain nucleoporins gates the ciliary entry of cytoplasmic proteins 62, and Importin β2 and RanGTP regulate the ciliary import of a kinesin-2 motor KIF17 63. Transition fibers, which project from distal end of basal body to ciliary membrane, likely block the ciliary entry of vesicles 64. A component of transition fibers, FBF1, promotes ciliary entry of IFT components likely by providing docking sites 65.

Since the cilium is a relatively isolated compartment, signaling molecules are enriched in or excluded from the cilium. An example is the enrichment of Inpp5e, which result in accumulation of phosphatidylinositol-4-phosphate (PI(4)P) and elimination of phosphatidylinositol 4,5-bisphosphate (PIP2), thereby attenuating actin polymerization and stabilizing the axoneme 66,67. Furthermore, enrichment of phosphatidylinositol (3,4,5)- trisphosphate (PIP3) induces higher basal activity of adenylyl cyclase 5 and 6, which elevates basal cAMP level to five fold higher than in the whole cell, resulting in activation of PKA in the cilium

68.

Anterograde trafficking, i.e. the shuttle into the cilium and toward the tip, of Gli proteins, is possibly mediated by IFT-B 69,70 and kinesin motors 71, and Sufu is transported in a complex with Gli proteins 72,73, whereas the transport of Smo into the cilium requires both IFT-B and IFT-

A complex as well as dynein motor 74. The retrograde transport, which moves proteins towards the ciliary base and out of the cilium, is mediated by IFT-A complex and a dynein motor 75–77.

Loss of ciliary functions, such as by mutation of Ift proteins, Kif3a or Dnchc2, disrupts both Hh signaling activation and repression 70,75,77–79.

Figure 1-2. The mammalian and Drosophila Hh signaling diverge at the primary cilium.

(A) In mammals, the primary cilium plays essential roles in Hh signaling. Ptch is localized to the cilium in the absence of Hh, while Smo is localized to the cilium with Hh stimulation. Gli proteins and Sufu are enriched in the ciliary tip, and ciliary localization is required for Gli2 activation. PKA is localized in the cilium, and becomes active due to a high cAMP level in the cilium. Therefore, the cilium enriches the component of Hh pathway to enhance the efficiency of signaling. The ciliary compartment is segregated from cytoplasm by a membrane diffusion barrier containing Septin 2, ciliary pore complexes containing nucleoporins that filters soluble proteins, and transition fibers that likely blocks vesicle trafficking and serves as docking sites for IFT components. IFT-B complex, with kinesin motors, accounts for the anterograde trafficking, whereas IFT-A complex mediate retrograde trafficking through dynein motors.

(B) A comparison of Drosophila and mammalian pathway components. In Drosophila, Hh signals to Ptc.

In mammals, three Hh ligands and two Ptch receptors relay the signal. The majority of Drosophila cells are not ciliated but assemble Cos2-based Hh signalosome complexes to bring the pathway components together, whereas in mammals primary cilium enriches pathway components. Three homologs of

Drosophila ci play different roles in mammals.

The investigations of Hh signaling were mostly conducted in the imaginal eye, wing and leg discs of fruit flies, where cilium appears not to play a role (see Chapter 1.1.1). Plasma membrane localization of Smo is sufficient to trigger Hh signaling activation 25, and ci does not carry a ciliary localization signal 72. As a functional substitute, Costal 2 (Cos2) acts as a scaffold protein to bring together Smo, ci, Fused, CK1 and PKA 25,26,47,49. Nevertheless, recent studies revealed a ciliary targeting of Smo by Cos2 in olfactory sensory neurons 80,81. Since the only ciliated cells in Drosophila are the sensory neurons and sperm/spematocytes 82–84, the requirement of cilia for Hh signaling should be limited to a few cell lineages.

1.1.3 Hh signaling cascade in mammals

In mammals, three homologues of Drosophila Hh have been identified, including Sonic

Hedgehog (Shh), Indian Hedgehog (Ihh) and Desert Hedgehog (Dhh), which all contain an N- terminal signal peptide that is cleaved upon secretion of the protein 85. Similar to Drosophila homolog, mammalian Shh undergoes proteolytic cleavages that remove the signal peptide and glycosylated carboxyl-terminus, and eventually generate Shh-Np that carries all known signaling activity of the full-length protein 6,86. Ihh, and possibly Dhh, are proteolytically cleaved as well

7,87. Shh-Np is covalently conjugated to a cholesterol at the carboxyl terminus 9. Cholesterol modification is suggested to limit the free diffusion of Shh-Np in the limb bud 88. Subsequently,

Shh-Np is palmitoylated at C25 residue 89,90. The palmitate moiety is critical for Shh-Np to bind

and attenuate Ptch1 91, and it also enriches Shh in lipid rafts to enhance its activity 92. The lipid anchors allow Shh to be trafficked in lipid rafts by Dispatched 93,94. Palmitate and cholesterol modifications also assemble Shh to a potent soluble multimeric complex that likely contributes to long range signaling 95,96. Furthermore, vesicular trafficking of Shh in the specialized filopodia contributes to efficient long range signaling in the limb mesenchyme 97. Subsequently, Dispatched

1 extracts Shh and releases it from source cells 93,94. In vitro studies show that proteases ADAM17 and Scube2 also help releasing Shh-Np by truncating the carboxyl terminus which is tethered to cholesterol 98–100. Finally, palmitate and cholesterol modifications contributes to binding of Shh to target cells 101.

A synthetic palmitoylated amino-terminal Shh that contains residues 1-22 is sufficient to block Ptch1 activity without inducing Ptch1 internalization, whereas a truncated Shh containing residues 10-197 induces Ptch1 internalization without activating the pathway, suggesting that clearance of Ptch1 from the cilium is not sufficient or required for Smo activation 91.

Two homologs of Drosophila Ptc are present in mammals, Patched homolog 1 (Ptch1) and

Ptch2 102. Ptch1 encodes the primary receptor for Shh and Ihh 12,103–105. Ptch2 plays a redundant role as a Shh receptor whose function is only revealed in the absence of Ptch1 106. Ptch binds Hh through the two big extracellular loops 103, and it confers structural homology to bacterial proton- driven transmembrane transporters 107. By suppressing the pathway activity in the absence of Hh,

Ptch1 plays a permissive rather than instructive role in Hh signaling activation 108,109. Furthermore,

Ptch1 mediates the endocytosis of Shh via clathrin-coated pits, limiting the free diffusion of Shh, and this endocytosis does not play a major role in signal transduction 110,111. Shh binding leads to the internalization and degradation of Ptch1, and segregates Ptch1 from Smo 112, thereby

alleviating the Ptch1 repression of Smo. Since Ptch1 expression is induced by Hh signaling activation, it forms a negative feedback to limit the Hh gradient 105,108.

Figure 1-3. The Hh signaling cascade in mammals.

(A) In the cells that do not receive Hh, Ptch inhibits Smo by efflux of 7-dehydrocholesterol and depletion of cholesterol. Smo is ubiquitinated, internalized and degraded in lysosomes. PKA, CK1 and GSK3β phosphorylate Gli2 and Gli3, leading to the ubiquitination by Cul1/Btrc E3 ubiquitin ligase complex and subsequent proteolytic processing in 26S proteasome that produces Gli repressors. Consequentially, Gli repressors (and possibly Sufu) recruit transcriptional corepressors such as mSin3 and SAP18 to suppress

Hh target gene expression.

(B) Upon Hh reception, Ptch is inhibited and endocytosed. As cholesterol accumulates and binds to Smo,

Smo is laterally relocalized to the cilium, leading to the ciliary localization and activation of Gli proteins.

The Gli activators, though labile, enter the nucleus and recruit mediator and coactivators to Hh target gene for transcription.

Several membrane proteins regulate Hh signaling at the level of Ptch by interacting with

Hh. Three other membrane proteins HHIP, CDON and Ihog bind Hh and regulate Hh signaling at the level of Ptch in a tissue-specific manner 113. Glypican-3 inhibits Shh signaling by competing with Ptch1 to bind Shh 114.

Smo, a seven-span membrane protein that relays Hh signal, plays a similar role from

Drosophila to mammals possibly through a different molecular mechanism 2. Since Smo does not physically interact with Shh 104, it is generally not perceived as a coreceptor. Early studies showed a physical interaction between Smo and Ptch1 104,115 which was not observed in later reports 112.

Ptch1 induces multi-ubiquitination of Smo in its cytoplasmic carboxyl-terminal tail, which causes its endocytosis and lysosomal degradation 21. Shh relieves the inhibition of Smo by Ptc, resulting in the stabilization and ciliary accumulation of Smo 116,117. Since Ptch1 efficiently inhibits Smo at a 1:50 ratio and intercellularly, Ptch1 likely catalytically suppresses Smo instead of forming a stable complex with Smo and locking it in an inactive conformation 107, and this regulation depends

on a proton-driven antiporter activity of Ptch1 106. Clearance of ciliary Ptch is not required for

Smo activation 118. In line with this notion, Ptch1 pumps 7-dehydrochosterol, a cholesterol precursor, out of the cell, and 7-dehydrochosterol and its derivative vitamin D3 bind and inhibit

Smo activity 119. Ptch1 also causes cholesterol efflux, and Hh-bound Ptch1 fails to interact with cholesterol, resulting in accumulation of intracellular cholesterol that enriches Smo in the cilium

120. As Smo binds plenty of exogenous small molecules 121–125, it has been hypothesized that Hh activates Smo through an unknown secondary messenger. This hypothesis is recently validated as several studies show that cholesterol binding to the extracellular Cysteine-Rich Domain of Smo induces a conformational switch in transmembrane domains, which is both required and sufficient for activation of Smo by Hh 126–128.

Smo functions in the primary cilium 129. Cell surface Smo enter the cilium, not through vesicles, but instead through lateral transport to the ciliary membrane 130, by IFT-B/kinesin-2 and

IFT-A/dynein complex, and β-arrestin which acts as an adaptor to mediate Hh-induced Smo-

KIF3A interactions 74,131. The ciliary localization of Smo requires a hydrophobic and basic motif carboxyl-terminal to the seventh transmembrane domain, mutation of which abolishes Smo activity and response to Hh 129. The cilium-enriched phospholipid PI(4)P binds to Smo and promotes its phosphorylation 132. However, ciliary localization of Smo may not be sufficient to trigger Hh signaling activation, as a chemical inhibitor cyclopamine enriches Smo in the cilium while blocking Hh pathway activity 117,133,134.

Smo activates downstream signaling through the carboxyl-terminal cytoplasmic tail 112.

The carboxyl-terminal cytoplasmic tail of Smo is phosphorylated by the kinases CK1α, CK1γ and

GRK2, which triggers its ciliary localization and activation 135,136. Hh deubiquitinates Smo which prevents its endocytosis 21. Shh decreases cAMP level mainly by elevating Ca2+ level through a

Gd3+ sensitive channel 68, although Gα12 and Gαi2 also relay the signal from Smo in certain contexts 137,138.

Finally, Gli transcription factors function as the effectors to interpret Hh signaling 139. It remains unclear whether Gli proteins are present in a complex with Smo, and how Smo converts

Gli proteins into the activator form.

1.1.4 Gli repression and activation are mediated by a series of events

Gli proteins bind DNA through five consecutive zinc finger motifs 140. An activator domain at the carboxyl-terminus of Gli proteins binds the mediator 141 and coactivators such as

CREBP and TAF9 139,142–144. Gli2 and Gli3, but not Gli1, also harbor a repressor domain that represses Hh target gene expression through a Ski- and HDAC-independent 145 but SAP18- dependent mechanism 146. Therefore, Gli2 and Gli3 act as both transcriptional activators and repressors, but Gli1 acts as an obligatory activator 139. The crystal structure of zinc fingers is available 140, and NMR shows that the amino-terminus and repressor domain is intrinsically disordered 147,148, but the 3-dimensional structure of other domains in Gli proteins remains unknown. The nuclear import of Gli proteins is mediated by a PY-NLS which interacts with

Transportin 149, and a bipartite nuclear localization signal (NLS) 43 which likely interacts with

Importin β1 150. A nuclear export signal (NES) is identified in Gli1 and Gli2, mutation of which accumulates Gli1 and Gli2 in the nucleus 151,152, but the putative NES in Gli2 and Gli3 repressor domains has not been studied.

Figure 1-4. Modifications and binding motifs of Gli proteins.

The functional difference between Gli2 and Gli3 is coded by a protein processing domain in the central region (Gli2-585-780; Gli3-648-844), which accounts for the much higher

153 proteolytic processing efficiency in Gli3 than Gli2 . Substitution of 620VE621 in Gli2 with KR dramatically improves the efficiency of proteolytic processing, while substitution of 685KR686 in

Gli3 with VE shifts the cleavage site towards the carboxyl-terminus 153. Hh reception abolishes the proteolytic processing of Gli2 and Gli3 154,155. An E3 ubiquitin ligase complex composed of the ring finger scaffold protein Cullin 1 (Cul1) and β-transducin repeat containing protein (Btrc) binds to the center region of Gli2 and Gli3 (see Figure 1-4 for binding sites), resulting in the ubiquitination of Gli2 (sites unknown) and Gli3 (K773, K779, K784, K800) and subsequent

degradation in proteasomes 155,156. A series of phosphorylations in the Btrc-binding motifs of

Gli2 and Gli3 turn on this interaction 155,156. Gli1 is not proteolytically processed 157 although

Btrc also induces Gli1 destruction in vitro 158. In cerebellar granule progenitor cells, Gli1 is ubiquitinated by the HECT-type E3 ubiquitin ligase Itch and subsequently degraded, which arrests cell growth and promotes differentiation 159,160.

Repression of Gli proteins is triggered by phosphorylation and ubiquitination. PKA phosphorylates Gli1 (T374) in the zinc finger in juxtaposition to nuclear localization signal, inhibiting the nuclear import 161, and it is unknown whether this mechanism is shared by Gli2 and Gli3 although this region is highly conserved among paralogues (Figure 1-4). Similar to ci

39, Gli2 and Gli3 are phosphorylated by PKA (Gli2-S789, S805, S817, S848; Gli3-S849, S865,

S877, S907) , CK1 (Gli2-S792, S808, S820, S851; Gli3-S852, S868, S880, S910) and GSK3β

(Gli2-S801, S813, S844; Gli3-S861, S873, S903), which prime Gli2 and Gli3 for Btrc binding and subsequent proteolytic processing 155,156,162,163. In addition, more potential phosphorylation within these clusters (Gli2-S662, S795, S796, S831, S835, S840; Gli3-S855, S856, S864, S894,

S899) are required for Btrc binding, likely through converting the local environment to Btrc

155,156,164 binding consensus (DSGX2-4S) . Another kinase, DYRK, phosphorylates Gli2 (S385,

S1011) and Gli3 (sites unknown), promoting their degradation 165. Phosphorylation of two additional PKA sites (Gli2-S923, S956; Gli3-S980, S1006) represses Gli activities although they do not contribute to the partial degradation 162. PKA phosphorylation of Gli1 (S642), Gli2

(S956) and Gli3 (S1006) promote their binding to 14-3-3ε, thereby inhibiting their transcriptional activity 166. PKA phosphorylation also inhibit ciliary localization of Gli proteins 72. One of the mechanisms by which the above PKA phosphorylations suppressing transcriptional activation is through inhibiting Gli SUMOylation by Pias1 167.

The activation of Gli proteins also involves a series of events. Upon Hh activation, a cluster of serine/threonines in the repressor domain of Gli2 (S216, T220, S224, S230, S248) and Gli3

(S281, S285, S289, S295, S313) are phosphorylated by an unknown kinase (possibly PKA), which enhance transcriptional activities 162. SUMOylation of Gli1 (K183, K815), Gli2 (K630, K716) and Gli3 (K696, K779) by Pias1 enhances transcriptional activity 167. Interestingly, Gli3-K779 is also a ubiquitination site targeted by Btrc 156, and Gli1 (K183, K815) and Gli2 (K630)

SUMOylation sites are also putative ubiquitination sites (http://www.ubpred.org/). Mutation of

Gli2 SUMOylation sites into arginine, which prevents both SUMOylation and potential ubiquitination, increases transcriptional activity 168, raising the possibility that SUMOylation promotes Gli activation by blocking ubiquitination.

The activation and proteolytic processing of Gli proteins both require primary cilium 70,77,78.

Unlike their Drosophila counterpart, Gli proteins harbor a ciliary localization cue that directs them to the ciliary tip, and Gli proteins also receive signal from ciliary Smo 72,129, suggesting that the cilium regulates Gli proteins directly and indirectly. Mutations of IFT-B complex component such as Ift27, Ift52, Ift88 and Ift172, anterograde motor Kif3a, and retrograde motor Dnchc2, all compromise Gli activator activity 70,77–79,169. The kinesin-2 complex subunits KIF3A and KAP3A both bind and colocalize with Gli proteins, suggesting that Gli proteins may be actively transported by kinesin-2 in the cilium 71. The central region of Gli proteins (human Gli1-391-655; Gli2-570-

967 or 852-1183; human Gli3-633-1018 or 912-1223) is required for their ciliary localization 72,170.

Gli2 ciliary localization region mutant fails to activate Hh target genes in vivo although it retains intrinsic transcriptional activity, suggesting that Gli2 (and possibly Gli3) activation requires its translocation to the cilium 171. However, it remains unknown how this region mediates the ciliary localization of Gli proteins. Gli1 overexpression in ciliary mutants activates Hh target genes,

suggesting that Gli1 activation may not require its ciliary localization 70, but it is also possibly an overexpression artifact since Gli2, when overexpressed in ciliary mutant cells, also activates Hh target genes 172,173. The ciliary exit of Gli proteins depends on dynein motor Dnchc2 174 and is possibly also mediated by Exportin 1 170. Mutation of Dnchc2, though increases ciliary localization of Gli2, compromises the nuclear localization of Gli2 and Hh target gene activation

174. Furthermore, IFT, KIF3A and Dnchc2 loss-of-function mutants exhibit defects in proteolytic processing of Gli3, although overexpressed Gli3 repressor does not require cilium to be functional

70,77,78. Despite the recent progress, the molecular mechanisms underlying the ciliary entry and exit of Gli proteins, as well as the events that occur to Gli proteins in the cilium, remain largely unknown.

Sufu is well known as a negative regulator of Gli proteins by promoting their partial degradation, sequestering them in the cytoplasm and regulating their transcriptional activity

151,152,175,176. Sufu is required for the proteolytic processing of Gli2 and Gli3 172 since Sufu recruits

Btrc 177 and GSK3β 178. One of the two major Sufu-binding sites on Gli proteins is the SYGHLS motif in proximity to the PY-NLS 179,180. By masking the PY-NLS, Sufu blocks nuclear import of

Gli proteins 149. Moreover, Sufu exports Gli proteins in an Exportin 1-dependent manner 136,152.

It remains controversial whether Hh activation and repression requires dissociation of Sufu from

Gli proteins. Based on the physical interactions of Gli proteins and Sufu, a classic model depicts that Gli proteins are separated from Sufu upon activation by Hh signaling, and Gli3 repressor also enters the nucleus without the company of Sufu 73,176. However, this notion is challenged by the finding that Sufu is associated with Gli1 and Gli3-binding promoters in a Hh-dependent manner, which suggests that Sufu is likely a component of both Gli1 activator and Gli3 repressor complexes

136,151,181. Consistent with the latter model, the interaction of Sufu with Gli activator domain

impedes CREBP recruitment, thereby inhibiting Gli transcriptional activity 182. Furthermore, the interaction between Sufu and transcriptional coactivator pCIP and corepressors SAP18, mSIN3 and p66β indicates that Sufu possibly recruits coactivators and corepressors to Gli-loaded promoters 146,181,183. Finally, Hh signaling regulates Sufu activity through various proteins and modifications, such as GSK3β and PKA-induced phosphorylations which stabilize Sufu 184.

1.2 Hh signaling plays critical roles in animal development

1.2.1 Shh signaling regulates pattern formation of spinal neural tube

Despite its complexity, the spinal cord neurons are specified from a monolayer of neural epithelium by a morphogen that also induces a polarizing activity in the wing bud 185. This morphogen is Shh, which is initially expressed in the notochord and later also in the floor plate 85, and the Shh gradient patterns neural tube and somite (Figure 1-5) 186. Ectopic expression of Shh results in severe central nervous system defects including open neural tube 85. Consistently, loss of Ptch1 leads to ventralized neural tube and neural tube closure defect 108, while ectopic Ptch1 dorsalizes the neural tube 109. Cells expressing a Ptch1 mutant that fails to bind Shh fail to specify ventral neural fate 111, and Smo mutant cells in chimera embryos fail to adopt the designated cell fates 187, suggesting that Shh signal input is directly required by the target cells throughout the ventral-dorsal axis.

A series of transcription factors define the Hh signaling domains. FoxA2 is a transcription factor required for formation of intervertebral discs 188. Shh signaling induces FoxA2 expression in the ventral-most neural tube through a binding site in the FoxA2 enhancer 189. FoxA2 also transcriptionally activates Shh along with Arx, thereby turning on Shh expression in the floor plate

190. Nkx2.2 is required for specifying V3 interneuron progenitors 191,192. Olig2 is a basic helix-

loop-helix transcription factor required for motoneuron and oligodendrocyte lineage 193.

Interestingly, Nkx2.2 and Olig2 interacts with and cross-repress each other, which helps drawing the boundary between V3 and mononeuron progenitor cells 194. Nkx6.1 is expressed in all ventral cell types, including floor plate, V3, motor neuron, V2, V1 and V0 interneurons 191. Pax6 exhibits a low expression in the ventral while high expression in the dorsal neural tube, suggesting that its expression is inhibited by Hh signaling 195. The cis regulatory modules for ventral-most genes exhibit strong binding affinity to Gli proteins, and the cis regulatory modules for more dorsal genes confer weaker binding affinity towards Gli repressors 196. As a result, Gli activators play an instructive role in specifying the ventral-most cell fates (floor plate and V3 interneuron progenitor), and a permissive role in long-range Shh signaling that specifies motor neuron to V0 interneuron fates, in which circumstances Gli repressors play an instructive role 196.

Figure 1-5. Shh patterns spinal cord by regulating the expression of various toolkit genes.

Shh gradient specifies the progenitors of floor plate (pFP), V3 interneuron (p3), motor neuron (pMN), and

V2-V0 interneuron (p2-p0). The expression of FoxA2, Nkx2.2, Olig2, Dbx2, Nkx6.1 and Pax6 define these neural types and is frequently used as markers to evaluate Shh signaling output. In the ventral-most neural tube, Gli activators play an instructive role, whereas in long range Shh signaling that specifies pMN to p0 neuroprogenitors, Gli repressors determine the cell fates although Gli activators play a permissive role.

Gli2 and Gli3 are ubiquitously expressed in the neural tube, while Gli1 is only induced by

Hh signaling in the ventral neural tube and thus acts secondarily to Gli2 to amplify Hh signal output 189,197. Gli1 and Gli2 both act primarily as activators, and Gli1 can almost functionally substitute Gli2 198. Loss of Gli2 results in diminishment of floor plate and V3 interneuron progenitors, suggesting that Gli2 activator is required for specifying the ventral-most cell fates

199,200. In contrary, loss of Gli1 does not alter the ventral spinal cord (from floor plate to V2 interneuron) patterning unless on a Gli2 mutant background, suggesting a functional redundancy among Gli1 and Gli2A 197,201. Loss of Gli3 leads to a dorsal expansion of Nkx6.2, Dbx1 and Dbx2, which label V1 and more dorsal interneuron progenitors, but has no effect on the ventral cell fates

202, and substitution of Gli2 gene with Gli3 dampens Shh signaling 203, indicating a role of Gli3 repressor in spinal cord patterning. Removal of Gli3 from Gli2 mutant further increases Shh signaling and exacerbate the spinal cord patterning defect, which suggests that Gli3A also plays a redundant role in promoting Shh signaling 203,204. Importantly, Gli1 is not expressed in Gli2;Gli3 double mutant spinal cord, suggesting that Gli1 acts as a secondary activator to amplify Shh signaling output 203,204.

1.2.2 Ihh signaling regulates skeletal development

The skeleton provides essential mechanical support, mobility and mineral storage critical for health. Aging and mutations result in a lower bone density and microarchitectural deterioration that is known as osteoporosis or to a less extent, osteopenia 205,206. As a result of osteoporosis, millions of people suffer from hip fracture worldwide 205.

The vertebrate skeleton develops through two mechanisms: intramembranous ossification, which is the formation of bones directly from mesenchymal cells, and endochondral ossification, which is the replacement of a pre-formed cartilage mold by bone tissues 207.

The endochondral ossification begins with condensation of mesenchymal cells which differentiate into chondrocytes composing the growth plate cartilage (Mackie et al., 2011).

Chondrocytes proliferate, undergo hypertrophic differentiation and eventually die (Figure 1-6A).

Death of chondrocytes allows the invasion of cells constructing the bone, including osteoblasts, osteoclasts and cells composing the vasculature (Mackie et al., 2011).

The perichondrium and periosteum are layered fibrous membranes covering the surface of cartilage and bone, respectively, and they both contain osteoblast progenitors 207. Hypertrophic chondrocytes and osteoblasts deposit the extracellular matrix (ECM) by secreting alkaline phosphatase, osteopontin and osteonectin as well as inducing mineralization of ECM (Lefebvre et al., 1995). Mineralization then leads to the formation of cortical bone (initially the bone collar) beneath the perichondrium/periosteum and trabecular bone inside the long bones 208.

Osteoclasts, derived from hematopoietic lineage, resorb bone matrix for remodeling 209.

Vasculature not only paves the way for osteoblast and osteoclast invasion but also prompts osteoblast fate (Hilton et al., 2005; Long, 2012).

Ihh is produced by prehypertrophic chondrocytes and regulates endochondral ossification through multiple mechanisms 208,210. Firstly, Ihh acts as a long-range morphogen to antagonize

Gli3 in resting chondrocytes, thereby inducing the switch of resting chondrocytes to proliferating chondrocytes, a process termed columnar differentiation (Figure 1-6B) 211–213. Conditional activation and removal of Smo in cartilage both indicate the requirement of direct Ihh input for the proliferation of chondrocytes 214. This regulation is independent of Pthlh signaling 212.

Secondly, Ihh regulates chondrocyte hypertrophy by Pthlh-dependent and independent mechanisms. On the one hand, Ihh counteracts Gli3 repressors to induce Pthlh expression in periarticular perichondrium (Figure 1-6B), which signals to resting and proliferating chondrocytes to keep them in undifferentiated state (Figure 1-6D) 210,212,215. Pthlh regulates chondrocyte hypertrophic switch in part through enhancing PKA activity and Gli3R level in a Sufu dependent manner 216,217. On the other hand, Ihh signaling promotes hypertrophic differentiation independently of Pthlh (Figure 1-6D) 218.

Figure 1-6. Ihh signaling regulates endochondral ossification.

(A) A long bone comprises chondrocytes, perichondrium/periosteum, osteoblast, osteoclast, blood vessels and extracellular matrix. Mineralization beneath the perichondrium/periosteum flanking hypertrophic

chondrocytes forms the bone collar which is cortical bone, and mineralization of the cartilage and bone matrix (osteoid mineralization) contributes to the initial trabecular network.

(B-E) Ihh signaling to different populations of cells causes different effects through antagonism of Gli repressors (GliR) and/or stimulation of Gli activators (GliA). Ihh, Pthlh and BMP signaling orchestrate to regulate the processes.

Thirdly, Ihh also directly signals to perichondrium to induce osteoblasts (Figure 1-6E), as indicated by the loss of osteoblast fate in perichondrium-specific Smo conditional knockout, and failure of Smo mutant cells to contribute to the osteoblast lineage in chimeric embryos 208,219.

Activation of Gli2 activator and antagonism of Gli3R are both required for the osteoblast lineage

219–221.

Besides, Ihh also induces vascular invasion indirectly 220. This induction does not require direct Ihh signaling to vascular endothelia, as Smo mutant cells contribute to the vasculature in chimeric embryos with wild type cells 219.

Furthermore, Ihh coordinates with multiple signaling pathways to regulate endochondral ossification. Wnt/β-catenin signaling acts downstream of Ihh to stimulate osteoblast maturation 222.

Loss of Wnt5a significantly reduces Ihh expression and delays chondrocyte differentiation as well as osteoblast maturation 223. Ihh also induces BMP signaling which in turn activates Pthlh pathway

224. BMP2 induces Ihh expression, and the BMP2-induced osteoblast differentiation requires Ihh signaling 219,225. Fgf18 is expressed in joints and perichondrium, inhibiting Indian hedgehog signaling and chondrogenesis while stimulating osteogenesis 226–228.

Finally, loss of Ihh also disrupts intramembranous ossification 208.

1.3 Spop targets Gli2 and Gli3 for ubiquitination and proteasomal degradation

Speckle-type POZ Protein (Spop) is a substrate-recognition subunit of a Cul3-based E3 ubiquitin ligase that was shown to target Gli2 and Gli3 for degradation in vitro 53,173. Spop is homologous to Drosophila hib/rdx, which accounts for ci turnover in the posterior compartment in wing disc 53,54. The role of Spop in animal development remains largely unknown. A paralog of Spop, Speckle-type POZ Protein-like (Spopl) shares high sequence identity and also confers ubiquitin ligase activity 229, but the function of Spopl in Hh signaling and animal development is a mystery.

1.3.1 Spop is part of an E3 ubiquitin ligase complex

Ubiquitin is a 76-residue peptide that, when tagged to the lysine sidechain of a substrate, often labels it for degradation in 26S proteasome (for cytoplasmic proteins) or lysosome (for membrane proteins) 230. The ubiquitination is catalyzed by a three-step cascade that comprise E1 activase which activates ubiquitin, E2 conjugase which passes the ubiquitin to substrate, and E3 ligase which binds the substrate and brings E2 conjugase to a juxtaposition 230. Since E3 ligases directly interact with substrates, and the characterized E3 ubiquitin ligases (more than 600) outnumbers E2 (about 40) and E1, E3 ligases confer the substrate recognition specificity 231. The majority of E3 ligases carry a RING finger scaffold protein such as Cul1 and Cul3 that interacts with E2 conjugase, and a substrate-recognition subunit such as Btrc and Spop 53,164.

Spop was initially identified as a protein localized to the nuclear speckles along with the splicing factor snRNP B’/B 232. Coexpression with its substrates Pdx1 and Gli3 alters the sub- nuclear distribution of Spop 233,234. Under hypoxia, Spop is re-localized to the cytoplasm in clear cell renal cell carcinoma and Hela cells 235. Spop is also observed in the cytosol of MEFs 173.

Interestingly, Spop also induces its own ubiquitination 236,237.

Spop comprises a MATH domain that interacts with substrates, a BTB-domain that adheres to Cul3, a “3-box” that allows its homodimerization, and a BACK domain that mediates oligomerization into higher order clusters 229. The dimerization and oligomerization is critical for catalytic activity of Spop E3 ubiquitin ligase since it increases the binding affinity to the substrates carrying multiple Spop degrons 229,238.

Spop targets plenty of substrate proteins (Table 1-1) for ubiquitination and subsequent degradation, thereby regulating different developmental and oncogenic pathways. Recurrent

Spop mutations (Figure 1-7) and downregulations are identified in prostate cancers 239–248, glioma 249,250, hepatoblastoma 251, malignant mixed Mullerian tumors 252, follicular thyroid carcinoma 253, uterine endometrial carcinoma 254, non-small cell lung cancers 255,256, ovarian cancer 257, breast cancers 258,259, liver cancer 260, gastric cancer 261 and colorectal cancers 261–263, suggesting Spop as a tumor suppressor. Furthermore, Spop interacts with a protein kinase

Ataxia-Telangiectasia Mutated (ATM) and is recruited to double strand break sites, and Spop mutation is associated with genomic instability 264,265. Spop functions as a tumor suppressor by degradation of androgen receptor (AR) 266,267, NCOA3 259,268, ERG 269,270, EglN2/PHD1 271, DEK

237, Cdc20 272, DDIT3 273, estrogen receptor α (ERα) 274, progesterone receptors 258, SIRT2 256 and Gli2 263,275. However, Spop functions as an oncogene in renal cell carcinoma through ubiquitinating and degrading PTEN and DUSP7 235,276.

Figure 1-7. Recurrent Spop mutations are identified in cancers.

Data from 148 studies in cBioPortal cancer genomics database (http://www.cbioportal.org/). Green circle: missense mutations; black circle: nonsense mutation; brown circle: frame shift.

Table 1-1. Spop substrates and their functions affected by Spop-induced degradation.

Substrate Context Function

Pdx1 Mouse pancreas Pdx1 induces insulin transcription and promotes β-cell survival, and

Spop induces Pdx1 degradation 234,277,278.

DAXX HEK 293T and DAXX represses ETS1 and p53-dependent transcription and inhibits

Hela cells apoptosis, which are reversed by Spop-induced degradation 279,280.

PIPKIIβ HEK 293T and PIPKIIβ depletes PI(5)P by synthesizing PI(4,5)P2, thereby inhibiting

Hela cells the p38/MAPK pathway, and Spop induces degradation 236.

ERα HEK 293 cells Spop attenuates estrogen signaling by turnover of Estrogen Receptor α

(Erα) 274,281.

ATM MDA-MB-231 Spop is recruited to double strand break sites and interacts with ATM,

and U2OS cells promoting DNA repair, but it remains unknown whether this requires

its E3 ubiquitin ligase activity 264,265.

SENP7 HEK 293T and SENP7 is a deSUMOylase that suppresses HP1β, and Spop-induced

IMR90 cells SENP7 degradation promotes cellular senescence 282.

SETD2 HEK 293 cells SETD2 is a histone H3K36 trimethyltransferase that regulates

alternative splicing and acts as a tumor suppressor, and Spop induces

its degradation 283.

BMI1 HEK 293 and Polycomb group protein BMI1 and the histone MACROH2A are

MEF cells recruited to the inactivated X chromosome, and Spop-catalyzed

ubiquitination promotes this recruitment without affecting stability 284.

Puckered Drosophila eye Puckered is a Jun kinase phosphatase, and Spop-induced degradation

triggers TNF/Eiger-dependent apoptosis 285. xGli1/2, Xenopus Ectopic Spop expression in Xenopus embryos induces degradation of hGli3 embryo Xenopus Gli1, Gli2 and ectopic human Gli3, and tunes down Hh

signaling activity 286.

Gli2 Gastric cancer Spop inhibits tumor cell proliferation and migration while promoting

cell lines and apoptosis, and the roles of Spop in suppressing gastric cancer are at

patients least partially mediated by Gli2 turnover 275.

Gli2 Colorectal Spop induces Gli2 degradation, thereby inhibiting Bcl-2 and promoting

cancer cell lines apoptosis 263.

and patients

SIRT2 A549 cells SIRT2 is a NAD-dependent deacetylase and an oncogene, and Spop

suppresses lung cancer by induceing SIRT2 degradation 256.

PR HEK 293T and Spop induces degradation of progesterone receptors (PR) and

T47D cells attenuates progesterone signaling, S phase entry and Erk1/2 activation

in breast cancers 258.

AR Four prostate SPOP degrades androgen receptor (AR) and hence functions as a tumor

cancer cell lines suppressor in prostate cancers 266,267.

and patients

NCOA3 MCF7, MDA- NCOA3 (also known as SRC-3) is a p160 steroid receptor coactivator

MB-231, that promotes AR transcriptional activity, proliferation and migration,

LNCaP-Abl and thereby promoting prostate and breast cancers, and Spop induces SRC-

HepG2 cells 3 degradation 259,268.

EglN2 RV1, LNCaP, EglN2 acts as an oncogene hydroxylating FOXO3a which induces

/PHD1 C4-2 and PC3 cyclin D1 expression in breast cancers, and it also promotes prostate

cells cancer, while Spop induces its turnover 271.

DDIT3 LNCaP and DDIT3 is induced by ER stress and activates proapoptotic gene

HEK 293T cells expressions, and Spop inhibits apoptosis by inducing DDIT3

degradation 273.

Cdc20 PC3 and DU145 SPOP suppresses prostate cancer by destruction of Cdc20, a cell cycle

cells regulator 272.

ERG Prostate cancer ERG is often fused to TMPRSS2 in prostate cancers and becomes

cell lines and activated. SPOP induces ERG degradation 269,270.

patients

DEK Immortalized DEK promotes prostate epithelial cell invasion, suggesting DEK as an

prostate oncogene, and SPOP induces DEK degradation. This screening also

epithelial cells identified TRIM24, SCAF1, CAPRIN1, WIZ, G3BP1 and GLYR1 as

potential SPOP substrates 237.

DUSP7, Caki-2 and Hela SPOP induces DUSP7 and PTEN degradation, and downregulates

PTEN cells DUSP6, Gli2 and DAXX levels, suggesting that SPOP likely promotes

oncogenesis by degrading these proteins 235.

BRMS1 MCF7 and Breast cancer metastasis suppressor 1 (BRMS1) suppresses tumor

MDA-MB-435S metastasis, and its degradation appears to suggest SPOP as an oncogene

cells 287.

A Spop gene trap allele with the insertion in the first intron was previously generated, but studies were only conducted in the pancreas where Spop regulates β cell mass and function by catalyzing Pdx1 turnover 277. The function of Spop in Hedgehog signaling and animal development remains largely unknown.

1.3.2 Spop targets Gli2/3 proteins for degradation

A potential role of Spop in Hh signaling starts to be uncovered after the discovery that

HIB (also named roadkill), the Drosophila homolog of Spop, targets ci for proteasomal degradation, thereby negatively regulating Hh signaling 53,54.

Similar to ci activator, Gli activators are labile, as evidenced by the lower Gli3 level upon

Hh induction 176,288. Knockdown of Cul3, Spop and Btrc all stabilize Gli3 in the presence or absence of Hh, suggesting that Gli activators are sensitive to degradation by multiple E3 ubiquitin ligases including Cul3/Spop 288. Furthermore, Spop is enriched in the nucleus 232 where Gli activators and repressors function. Therefore, it is generally believed that Spop selectively targets

Gli activators for degradation 288,289. However, there has been no direct evidence that supports a preference of Gli activators over inactive forms as a Spop substrate. It is possible, though unlikely, that Spop downregulates GliA level by destabilizing full-length Gli proteins (see Chapter 5.1.2 for discussion). Furthermore, Spop interacts with a Gli3 repressor mimic Gli31-660 and catalyzes ubiquitination of Gli31-90 148,238 but does not drastically decrease the level of Gli31-700 when

coexpressed in C3H10T1/2 cells 289, raising a question on whether Spop also targets Gli repressors for ubiquitination and degradation.

Spop interacts with multiple serine and threonine-rich sequences, named Spop degrons, on Gli2 and Gli3 238. Hh induces phosphorylation of these serines and threonines by CK1, which abolishes Spop binding and stabilize ci and Gli proteins 55. An analysis of Gli31-90 reveals five ubiquitination sites on the amino-terminus: K15, K22, K32, K70 and K87 148, but it remains unknown whether the inactive Gli3 and Gli3 repressors are ubiquitinated on the same sites, and whether Btrc and Spop elicit distinct Gli responses by ubiquitinating different lysines.

We previously uncovered a positive role of Sufu in Hh signaling by the lack of maximal

Hh signaling activation in Gli1;Sufu double mutant neural tubes 290. This positive role was recently confirmed by another group by the lower Hh signaling activity in Sufu;Ptch1 double mutants in comparison to Ptch1 mutant 136. Since loss of Sufu destabilizes Gli2 and Gli3, Sufu likely plays this positive role by stabilizing Gli2 and Gli3 172,289. Based on the in vitro evidence that Sufu antagonizes Spop-directed Gli turnover 173,289, we hypothesized that Sufu positively regulate Hh signaling by competitive inhibition of Spop.

1.3.3 Spop-like is a potential regulator of Gli proteins

Spop-like (Spopl) is a paralog of Spop that shares 81% sequence identity, but the function of Spopl is poorly understood. Similar to Spop, Spopl forms a complex with Cul3 and targets Puc for degradation in 26S proteasome 229. Furthermore, Spopl targets EPS15, an endocytosis regulator and not a Spop substrate, for ubiquitination and proteasomal degradation

291. Compared with Spop, Spopl carries an 18-residue insertion in BACK domain that prevents it from higher order assembly and impairs its catalytic activity (Figure 1-8) 229. By heterodimerization with Spop, Spopl also shifts Spop into smaller complexes, hence impairs the

E3 ubiquitin ligase activity of Spop 229. Spopl mutation is associated with prostate cancer 248.

However, it remains unknown whether Spopl targets Gli for ubiquitination and degradation, and whether Spopl plays a role in Hh signaling and animal development.

Figure 1-8. Spopl and Spop share high sequence identity.

Similar to Spop, Spopl also comprise a MATH domain that binds substrate, a BTB domain that interacts with Cul3, and a BACK domain. However, the extra 18 amino acids in BACK domain prevents Spopl from higher-order oligomerization.

1.4 Hypotheses and specific aims

Previous in vitro studies suggested Gli2 and Gli3 as Spop substrates. Because Gli2 and

Gli3 play critical roles in skeletal development and spinal cord patterning, we hypothesize that

Spop regulates skeletal development and spinal cord patterning by triggering Gli2 and Gli3 degradation. Since Spopl shares high sequence identity with Spop, these two proteins potentially play redundant roles in animal development. In the current study, we will test these hypotheses in the following two specific aims.

1.4.1 Specific aim 1 will test the hypothesis that Spop and Spopl regulate skeletal development by inducing Gli2 and Gli3 ubiquitination and subsequent degradation.

Four major questions need to be addressed in this specific aim. First, does loss of Spop affect skeletal development? We hypothesize that Spop plays a role in skeletal development. If this hypothesis is true, which event in skeletal development is regulated by Spop? Bone formation involves multiple events including chondrogenesis and osteogenesis (see Chapter

1.2.2 ). It is possible that either chondrocyte or osteoblast lineage is regulated by Spop.

Furthermore, what is the molecular mechanism underlying the regulation of skeletal development by Spop? We hypothesize that Spop regulate skeletal development by degrading

Gli2 and/or Gli3. Finally, does Spopl regulate skeletal development? We hypothesize that Spop and Spopl play redundant role in skeletal development.

To test the hypothesis that Spop plays a role in skeletal development, we will generate loss-of-function Spop mutant mouse strains (Chapter 3.1.1 ). We will test whether Spop is expressed in the skeleton (Chapter 3.2.1 ). Subsequently, we will characterize the skeletal defects in Spop mutants. Since loss of Spop in the whole embryo induces perinatal lethality 277, we will conditionally delete Spop in the limbs to investigate whether loss of Spop affects the size, density and microarchitecture of adult bones (Chapter 3.4 ).

The proliferation and maturation of chondrocytes contribute the growth of long bones. If loss of Spop alters the size of bones, it is likely mediated by a deficiency in chodrogenesis. To test the hypothesis that Spop regulates chondrogenesis, we will investigate the influence of loss of Spop on the hypertrophic differentiation of chondrocytes (Chapter 3.2.3 ). Furthermore, we will examine whether the proliferation and apoptosis of chondrocyte are altered in Spop mutant bones (Chapter 3.4.1 ). Subsequently, we will test the hypothesis that Spop regulates osteogenesis by investigating the formation of bone collar and differentiation of osteoblasts in

Spop mutant bones (Chapter 3.2.4 ). Furthermore, we will determine whether a change in osteoblast or osteoclast activity accounts for altered bone density in Spop mutant bones (Chapter

3.4.2 ).

To test the hypothesis that Spop regulates skeletal development by degrading Gli2 and

Gli3, we will evaluate the Ihh signaling activity, and examine Gli2 and Gli3 protein levels, in

Spop mutant bones (Chapter 3.3 ). Three potential mechanisms may underlie Spop regulation of skeletal development. Firstly, Spop may inhibit Pthlh expression by degrading Gli2 and Gli3A in the periarticular perichondrium, thereby promoting chondrocyte maturation and osteoblast differentiation. Secondly, Spop may directly trigger Gli2 and Gli3A degradation in the cartilage growth plate and osteoblasts, thereby tuning down Ihh signaling activity and inhibiting chondrocyte and osteoblast differentiation. Lastly, Spop may downregulate Gli3R level in cartilage and osteoblasts, which triggers increased Ihh target gene expression, resulting in stimulated chondrocyte maturation and osteoblast differentiation.

Finally, by characterizing the skeletal phenotype of Spopl mutant and Spopl;Spop double mutant embryos, we will test the hypothesis that Spopl and Spop play a redundant role in skeletal development (Chapter 3.5 ). If loss of Spopl causes skeletal defect, or exacerbates the skeletal defect on Spop mutant background, we will further investigate the mechanisms by which Spopl regulates Ihh signaling and skeletal development.

1.4.2 Specific aim 2 will test the hypothesis that Spop and Spopl regulate spinal cord patterning by ubiquitination of Gli2 and Gli3.

Gli2 and Gli3 interpret Shh gradient to graded responses for spinal cord patterning. It is possible that Spop and Spopl, by inducing Gli2 and Gli3 ubiquitination. To reveal potential roles of Spop and Spopl in spinal cord patterning, we will address the following questions.

First, what role does Spop play in spinal cord patterning? We hypothesize that Spop inhibits Shh signaling activity, resulting in restriction of ventral-most cell fates, by inducing the ubiquitination and subsequent degradation of Gli2 and Gli3A. Alternatively, Spop may promote

Shh signaling and ventral-most cell fates by downregulating Gli3R. We will test these hypotheses by analyses of spinal cord patterning and Shh signaling activity in Spop mutant,

Spop;Gli2 and Spop;Gli3 double mutant embryos (Chapter 4.1 ).

We previously revealed a positive role of Sufu in promoting ventral-most cell fates by examining Gli1;Sufu double mutant and more mutant spinal cords 290. However, the mechanism underlying this positive role of Sufu remains a question. Since loss of Sufu results in a drastic decrease in Gli2 and Gli3 abundance 172,173,289, we hypothesize that Sufu promotes ventral-most cell fates by antagonism of Spop-induced Gli2 and Gli3 degradation. To test this hypothesis, we will investigate whether removal of Spop rescues Gli2 and Gli3 level and restores ventral-most cell fates in the context of Gli1;Sufu double mutant embryos (Chapter 4.2 ).

Finally, the role of Spopl in spinal cord patterning remains a question. It is possible that

Spopl also regulates spinal cord patterning by inducing Gli2 and Gli3 degradation. We will test the hypothesis that Spopl and Spop play redundant roles in spinal cord patterning by analyzing

Spopl mutant and Spopl;Spop double mutant embryos (Chapter 4.1.4 ).

Chapter 2

Materials and methods

2.1 Genetics Analysis

2.1.1 Mouse strains and genotyping strategy

The animal husbandry in this study was approved by the IACUC (#37372, #45676) at the

Pennsylvania State University. Mouse colonies were maintained on 129S2/SvPasCrl or

C3H/HeNCrl genetic background (Charles River Lab).

Spoptm1a(KOMP)Mbp (hereafter abbreviated as SpoplacZKI) and Spopltm1(KOMP)Vlcg (hereafter abbreviated as Spopl-) mouse strains were purchased from the Knockout Mouse Project

(https://www.komp.org/). The knockout strategies are described in Chapter 3.1.1 and Chapter

3.5, respectively.

Gli1, Gli2, Gli3 and Sufu loss-of-function mutant strains (Figure 2-1A-D) were generous gifts from Dr. Alex Joyner (Memorial Sloan Kettering Cancer Center) and Dr. Rune Toftgård

(Karolinska Institutet, Sweden). In Gli1tm2Alj (hereafter abbreviated as Gli1-) allele, part of exon

3 and the entire exon 4 to 11 that encode the N terminus and the five zinc fingers are replaced by a lacZ expression cassette, resulting in a loss of Gli1 function even if splicing occurs between the exons flanking lacZ insertion 197. In Gli2tm2.1Alj (hereafter abbreviated as Gli2-) allele, a promoterless lacZ knock-in abolishes Gli2 function 198. In Gli3Xt-J (hereafter abbreviated as

Gli3-) allele, a 51.5kb deletion immediately after exon 9 and a chromosomal rearrangement result in a truncated Gli3 fused to a short peptide that is not functional 292–294. However, the presence of this amino-terminally truncated protein in this allele limits the use of Gli3 mutant as a negative control in immunofluorescence, as the available Gli3 antibodies all target the amino-

terminus. In Sufutm1Rto (hereafter abbreviated as Sufu-) allele, exon 1 is replaced by a neomycin resistance cassette in an opposite orientation, and this abolishes the expression of Sufu 295.

Figure 2-1. Schematics of mouse loss-of-function mutant and reporter strains.

(A) In Gli1tm2Alj allele, a lacZ expression cassette with transcriptional stop sequence replaces part of exon

3 and the entire exon 4 to 11. As a result, Gli1 is either not translated, or missing the N terminus and the five zinc fingers if splicing occurs between exons 2 and 12.

(B) In Gli2tm2.1Alj allele, a promoterless lacZ is inserted into exon 2. The transcriptional stop sequence abolishes the expression of Gli2.

(C) In Gli3Xt-J allele, a 51.5kb deletion removes exons 9 to 15. Furthermore, an unprecedented chromosomal rearrangement brings together the Gli3 locus and an exon which is originally located 140kb downstream and encodes 10 residues. As a result, the Gli3 mutant is truncated after the first zinc finger and is not functional.

(D) In Sufutm1Rto allele, exon 1, which harbors both 5’-UTR and the start codon, is replaced by a neomycin resistance (neoR) cassette in an opposite orientation. As a result, Sufu is not expressed.

(E) In Ptch1tm1Mps, part of exon 1 and the entire exon 2 are replaced by a promoterless lacZ and neomycin resistance cassette. Although the mechanism is unclear, a byproduct Ptch1tm2Mps allele does not abolish the Ptch1 function but exhibits a similar lacZ expression profile as Ptch1tm1Mps.

(F) In Tg(ACTFLPe)9205Dym, a FLP3 transgene directed by human ACTB promoter is integrated into an unknown genomic locus.

(G) In Tg(Prrx1-cre)1Cjt, a Cre transgene directed by a 2kb-long Prx1 enhancer is integrated into an unknown genomic locus. The insulator prevents activation or repression by the local environment.

(H) In Tg(EIIa-Cre)C5379Lmgd, a Cre transgene directed by a adenovirus EIIa promoter is integrated into an unknown genomic locus.

(I) In Gt(ROSA)26Sortm1(EYFP)Cos, an Enhanced Yellow Fluorescent Protein (EYFP) gene is inserted in

Gt(ROSA)26Sor locus. Preceding the EYFP gene is a neoR cassette with transcriptional stop sequence, and this cassette is flanked by loxP sites. This allele hence serves as a popular reporter for spatiotemporal expression of Cre.

Ptch1tm2Mps (also known as Ptch1D11, hereafter abbreviated as Ptch1-lacZ) was kindly provided by Dr. Matthew Scott (Stanford University) (Figure 2-1E). This allele is a derivative from an aberrant recombination in the process of generating Ptch1tm1Mps (also known as

Ptch1KO1). In Ptch1KO1, part of exon 1 and the entire exon 2 which encode the first transmembrane domain are replaced by a promoterless lacZ and neomycin resistance cassette, resulting in lethality at E8-E10.5 stage 108. Although the mechanism is unclear, Ptch1-lacZ does not abolish the Ptch1 function, but exhibits a lacZ expression profile similar to the endogenous

Ptch1 expression, and thus serves as a popular lacZ reporter for Ptch1 expression 296. Since

Ptch1-lacZ homozygotes are sterile though viable 296, we bred this strain as heterozygotes. We also use Ptch1-lacZ reporter as heterozygotes to ensure fair comparison between Spop mutants and littermates.

Tg(ACTFLPe)9205Dym and Tg(Prrx1-cre)1Cjt were purchased from Jackson Lab.

Tg(ACTFLPe)9205Dym, a human ACTB promoter directs the expression of FLPe, a genetically engineered Saccharomyces cerevisiae FLP1 recombinase, in germline and other tissues 297. The locus of insertion is unknown. In Tg(Prrx1-cre)1Cjt (hereafter abbreviated as Prx1-Cre), a Cre transgene preceded by an insulator and an enhancer of Paired-related homeobox gene-1 (Prx1) is inserted into an unknown locus 298. The Prx1 enhancer directs the expression of Cre in forelimb mesenchyme as early as E9.5, and at E10.5 Cre induces efficient loxP recombination in the mesenchyme of both forelimb and hindlimb, while blood and the apical epidermal ridge remain unaffected. 298. Since Prx1-Cre expression is leaky in female germline, even though Prx1-Cre homozygotes are viable and fertile 298, female loxP;Prx1-Cre double carriers are not recommended as breeders. Therefore, we maintain this allele by crossing male Spopflox/flox;Prx1-

Cre carriers with female Spopflox/flox carriers.

Tg(EIIa-Cre)C5379Lmgd was a generous gift from Dr. Douglas Cavener (Pennsylvania

State University). In Tg(EIIa-Cre)C5379Lmgd (hereafter abbreviated as EIIa-Cre), a Cre transgene under the control of an adenovirus EIIa promoter is inserted into an unknown locus 299.

Since the EIIa promoter is active at the earliest stage of embryonic development, most likely the zygote, the Cre induces efficient loxP recombination in all cells in the embryo 299. Although the

EIIa-Cre homozygotes are viable, we maintain this strain by mating heterozygous males with wild type females to minimize the influence.

Gt(ROSA)26Sor encodes a long non-coding RNA constitutively expressed in whole embryo and in adult hematopoietic lineage 300,301. In Gt(ROSA)26Sortm1(EYFP)Cos (hereafter abbreviated as R26R-YFP), an Enhanced Yellow Fluorescent Protein (EYFP) gene preceded by loxP sites-flanked transcriptional stop sequence is inserted in Gt(ROSA)26Sor locus 300.

Excision of the transcriptional stop sequence with Cre activates Gt(ROSA)26Sor promoter driven EYFP expression in Cre-expressing cells and their descendants. Therefore, the R26R-

YFP allele is a popular reporter for Cre activity in embryos, but in postnatal development the application of this reporter is limited by the promoter activity.

Table 2-1 Genotyping primers and condition for mouse strains.

Strain Primers Remark

Spopflox PK50: 5’- TGCAGAGAAACTTGCCTTGA -3’ PK50/51 targets WT (268bp) and

PK51: 5’- GGAGCGTTCACATCCCTTAC -3’ flox (448bp)

SpoplacZKI PK50: 5’- TGCAGAGAAACTTGCCTTGA -3’ See above for PK50/51; PK54/51

PK51: 5’- GGAGCGTTCACATCCCTTAC -3’ targets SpoplacZKI (364bp)

PK54: 5’- ATCCGGGGGTACCGCGTCGAG -3’

SpopΔEx PK50: 5’- TGCAGAGAAACTTGCCTTGA -3’ See above for PK50/51;

PK51: 5’- GGAGCGTTCACATCCCTTAC -3’ PK50/PM12 targets SpopΔEx

PM12: 5’- GAAGGGAGGTGAACTGATGG -3’ (224bp).

Strain Primers Remark

Spopl- PN77: 5’- TCTCTTCCTGGTTTTCTCAAAGA -3’ PN77/78 targets WT (230bp);

PN78: 5’- GCATGTGACAAGCAGCAAAT -3’ PJ23/24 targets lacZ (327bp)

PJ23: 5’- CCGAACCATCCGCTGTGGTAC -3’

PJ24: 5’- CATCCACGCGCGCGTACATC -3’

Gli1- PK19: 5’- PK19/20 targets WT (493bp);

CCAGTTTCTGAGATGAGGGTTAGAGGC -3’ PK19/PL40 targets mutant

PK20: 5’- TTGAATGGGGAATACAGGGGCTTAC - (360bp)

3’

PL40: 5’- GTGCTGCAAGGCGATTAAGT -3’

Gli2- PK43: 5’- CCAGCGCACTCATTAAATCC -3’ PK43/45 targets WT (323bp);

PK44: 5’- GGGTTATTGAATATGATCGGAAT -3’ PK43/44 targets mutant (250bp)

PK45: 5’- TGTGGACCTGGGTTGGTATT -3’

Gli3- PI62: 5’- GGCCCAAACATCTACCAACACATAG - PI62/63 targets WT (193bp);

3’ PI64/65 targets mutant (590bp)

PI63: 5’- GTTGGCTGCTGCATGAAGACTGAC -3’

PI64: 5’- TACCCCAGCAGGAGACTCAGATTAG -

3’

PI65: 5’- AAACCCGTGGCTCAGGACAAG -3’

Sufu- PC5: 5’- CCCTTTTTGTCAATAGTTCC -3’ PC5/6 targets WT (274bp);

PC6: 5’- TGACAATAGACTCCGCCTCC -3’ PC5/7 targets mutant (194bp)

PC7: 5’- GCCTTCTATCGCCTTCTTGAC -3’

Strain Primers Remark

Ptch1lacZ PI66: 5’- TGTCTGTGTGTGCTCCTGAATCAC -3’ PI70/71 targets WT exon 2

PI67: 5’- TGGGGTGGGATTAGATAAATGCC -3’ (217bp); PI66/67 targets mutant

PI70: 5’- CTGCGGCAAGTTTTTGGTTG -3’ (neo) (501bp)

PI71: 5’- AGGGCTTCTCGTTGGCTACAAG -3’

ACT-FLPe PM23: 5’- CACTGATATTGTAAGTAGTTTGC -3’ PM23/24 targets the flippase

PM24: 5’- CTAGTGCGAAGTAGTGATCAGG -3’ transgene (500bp); PM25/26

PM25: 5’- CTAGGCCACAGAATTGAAAGATCT -3’ targets interleukin-2 (324bp).

PM26: 5’- GTAGGTGGAAATTCTAGCATCATCC - This primer set does not

3’ distinguish hemizygotes and

homozygotes.

Prx1-Cre PC2: 5’- CCTGGAAAATGCTTCTGTCCGTTTG -3’ PC2/1 targets Cre (650bp). This

PC1: 5’- GAGTTGATAGCTGGCTGGTGGCAGATG primer set does not distinguish

-3’ hemizygotes and homozygotes.

EIIa-Cre Same as Prx1-Cre genotyping

R26R-YFP PI29: 5’- AAAGTCGCTCTGAGTTGTTAT -3’ PI29/31 targets WT (603bp);

PI30: 5’- GCGAAGAGTTTGTCCTCAACC -3’ PI30/31 targets mutant (310bp)

PI31: 5’- GGAGCGGGAGAAATGGATATG -3’

2.1.2 β-galactosidase (lacZ) staining

Bacterial β-galactosidase, encoded by lacZ gene, is a popular reporter for gene expression. When a promoterless lacZ cassette is targeted to the genomic locus of gene of interest and subject to the regulation of endogenous promoters and enhancers, lacZ staining provides a faithful measure of the gene expression profile.

Solutions

Wash buffer

Mix 100µl 1M MgCl2, 50µl 10%(vol/vol) sodium deoxylcholate and 100µl

10%(vol/vol) NP-40 into 50ml PBS. Store at room temperature.

X-gal staining solution

Dissolve 0.106g potassium ferrocyanide, 0.082g potassium ferricyanide and 2ml

25mg/ml X-gal in 48ml wash buffer. Store at -20°C. Minimize freeze-and-thaw

cycles as precipitation results.

Procedure

(1) Fix embryos lightly in PBS/4% PFA at 4°C. Fix 20min for E9.5, 30min for E10.5, 1

hour for E11.5 to E12.5. The permeability for older embryos is limited, and thus it is

necessary to isolate the tissues of interest by microdissection or frozen section. Avoid

over-fixation as it kills β-galactosidase activity. Glutaraldehyde is a better fixative with

less impact on β-galactosidase activity 302.

(2) Remove fixatives by three washes of PBS at 4°C for 15min each on a nutator. Store

samples at 4°C while genotyping.

(3) Wash three times with wash buffer at RT for 15min each on a nutator.

(4) Stain in X-gal staining solution at 37°C in dark. Seal the chamber well to minimize

evaporation. Flip the sample in between to ensure even staining from top to bottom.

Stop reaction when the blue signal is strong enough, and before the negative control starts

to turn blue. The sample may turn brown before turning blue.

(5) Wash three times with wash buffer at RT for 5min each on a nutator.

(6) Postfix in PBS/4% PFA at 4°C overnight on a nutator.

(7) Wash three times with PBS/0.1% Tween20 at RT for 10min each on a nutator.

(8) Take photos. Process the samples for cryosection.

2.1.3 Alcian blue and Alizarin red staining

Alcian blue and Alizarin red staining are a special set of dyes for visualizing cartilage

(blue) and bones (red). At pH 1.0, Alcian blue binds sulphate-ester groups which are rich in cartilage, whereas at pH 2.5, Alcian blue also binds carboxyl groups which are rich in most acidic mucins 303. Fortunately, Alcian blue does not stain nuclear acid due to steric hindrance.

Alizarin red specifically binds calcium and forms a red precipitate 303.

Solutions

Alcian blue staining solution

Mix 40ml 95% ethanol and 10ml glacial acetic acid. Dissolve 7.5mg Alcian blue

8GX (Sigma A3157). This staining solution has a pH of about 2.1.

Alizarin red staining solution

Dissolve 2.5mg Alizarin red (Sigma A5533) and 1g potassium hydroxide in 50ml

water. This solution is stable for 6 weeks before Alizarin red is degraded through

a light catalyzed reaction.

Procedure

(1) Prepare samples. For embryos younger than E15.0, use whole embryo for staining, since

severing the embryo may result in compromised Alizarin red staining. For embryos older

than E16.0 and newborns, remove viscera, eyes and adipose tissue since the staining of

these tissues may affect imaging quality. Also remove the skin as it affects permeability

of dyes. To deskin the embryos, leave them in water for 1-24 hours to soften the skin,

and then soak in hot water (65-70°C) for up to 30 seconds. Subsequently, carefully peel

off the skin. For mice aged 1 week or older, also free the bones of muscles.

(2) Move samples to scintillation vials. Fix in 95% ethanol for 1-3 days on a nutator.

(3) Stain in Alcian blue for 24 hours on a nutator.

To keep the pH low and minimize mucin staining, use plenty of staining solutions for big

samples.

(4) Rinse twice in 95% ethanol. Wash in 95% ethanol overnight. Repeat once, nutation

allowed.

(5) (Skip this step for embryos and newborns.) For mice aged 1 week or older, rehydrate

through an ethanol series to water. Wash with 30% BORAX, pH 8.0 overnight. Change

the buffer and add 0.25% trypsin. Trypsinize for several days to remove stainings on

other tissues such as ligaments and tendons. Incubate in potassium hydroxide for 2

hours.

(6) Stain in Alizarin red for 1-3 hours. Change solution and stain overnight.

Caution: from this step on, the samples become very fragile! Handle the samples gently

without shaking or nutation. For embryos, make sure they are always submerged when

changing solution. Don’t remove all the liquid.

(7) (Skip this step for embryos younger than E15.0.) Clear samples in 1% KOH for a few

hours.

(8) Change to 20, 40, 60 and 80% Glycerol in 1% KOH. Leave the samples overnight in

each solution.

2.1.4 RNA in situ hybridization (on cryosections)

RNA in situ hybridization is a powerful method to reveal the RNA distribution and thus gene expression profile. Our lab employs a non-radioactive chemical substrate on cryosections to achieve a purple staining which works with nuclear fast red counterstaining.

Solutions

Acetylation solution

Dissolve 0.1M TEA-HCl in 40ml RNase-free water. Add 360µl 5M sodium

hydroxide to adjust the pH to 8.0. Add 100µl acetic anhydride right before use.

Hybridization solution

Mix 5ml deionized formamide, 2.5ml 20x SSC, 50µl 10mg/ml tRNA, 1ml 10%

SDS, 5µl 100mg/ml heparin into 1.445ml DEPC water. Aliquot and store at -

20°C.

20x Sodium chloride and sodium citrate (SSC)

Dissolve 87.6g sodium chloride and 44.1g sodium citrate in water. Adjust pH to

7.0 with HCl and bring final volume to 500ml. Autoclave to sterilize.

High stringency buffer

Mix 10ml 20x SSC, 50ml formamide with 40ml water.

RNase buffer

Mix 5ml 1M Tris-HCl, pH7.5, 1ml 0.5M EDTA, 50ml 5M sodium chloride with

444ml water.

NTMT

Mix 10ml 1M Tris, pH9.5, 5ml 1M MgCl2, 2ml 5M NaCl, 0.1ml Tween-20 into

82.9ml water. Add 2mM levamisole right before use.

Procedure

(1) Prepare DIG-labeled antisense RNA probes. Store at -80°C.

a. Linearize 10µg of DNA template with restriction endonucleases (New England

Biolab) (Figure 2-2). Run an agarose gel with 0.2µg to check the digestion

efficiency.

b. Bring the volume to 200µl with water. Add 20µl 3M sodium acetate, mix well.

Phenol/chloroform, chloroform extract. Add 600µl ethanol, mix well and

precipitate at -80°C for 30min. Re-dissolve in 10µl RNase-free water.

c. Set up in vitro transcription reactions. Transcribe at 37°C for 2h. Run 1µl

product in agarose gel and take pictures at 10min and 30min to check.

d. Add 180µl RNase-free water, 20µl 3M sodium acetate and mix well. Add 600µl

ethanol, mix well and precipitate at -80°C for 30min. Spin 16,000g at 4°C for

10min. Keep the precipitate. Redissolve in 50µl RNase-free water.

Figure 2-2. Synthesis of antisense RNA probes.

(A) For Ihh probe, linearize the vector with BamHI and transcribe with T7 RNA polymerase 208.

(B) For Gli1 probe, linearize the vector with NotI and transcribe with T3 RNA polymerase 304,305.

(NM_008957.3)

(D) For Col2a1 probe, linearize the vector with EcoRI and transcribe with T3 RNA polymerase. Probe targets 405bp of mouse Col2a1 (NM_031163.3) in the coding sequence and 3’-UTR.

(E) For Col10a1 probe, linearize the vector with NotI and transcribe with T3 RNA polymerase 306.

(F) For Pthlh probe, linearize the vector with SalI and transcribe with T7 RNA polymerase. Probe targets the 1512bp full-length mouse Pthlh transcript (should be 1455bp as inferred from NM_008970.4).

(G) For Pth1r probe, linearize the vector with EcoRI and transcribe with SP6 RNA polymerase.

(H) For Runx2 probe, linearize the vector with EcoRI and transcribe with T7 RNA polymerase 307. Probe targets 270bp of 5'UTR and 45bp of coding sequence.

(I) For Bmp2 probe, linearize the vector with XbaI and transcribe with T3 RNA polymerase 308. Probe targets mouse Bmp2 (NM_007553.3) with 164bp of 5’-UTR and 1026bp of coding sequence.

(J) For Bglap probe, linearize the vector with XbaI and transcribe with T3 RNA polymerase. Probe targets 37..304 of Bglap (NM_007541.3), CDS 52..339.

(2) Fix the samples overnight in PBS/4% PFA. Sink in two changes of PBS/30%(w/v)

sucrose at 4°C overnight. Embed in OCT (Sakura), cut longitudinal (for bones) or

transverse (for spinal cord) sections at 10μm thickness, dry sections on slides for a

minimum of 30min and store in -80°C.

Note:

a. Proper decalcification is required for sectioning E18.5 mouse bones. Prior to

sucrose immersion, decalcify in 3 changes of PBS/14%(w/v) EDTA, pH 7.4 at

4°C for 1 day each.

b. For E9.5 and E10.5 embryos, decapitate and cut the embryo in half at caudal level

(to a forelimb piece and a hindlimb piece) for a better sucrose penetration.

(3) Post fix slides in RNase-free PBS/4% PFA for 10min. Wash in two changes of RNase-

free PBS for 5min each. Drain the slides.

(4) Immerse slides in RNase-free PBS/20ug/ml proteinase K for 6min. Wash with RNase-

free PBS for 5min.

(5) Refix in RNase-free PBS/4% PFA. Wash with RNase-free PBS for 5min.

(6) Acetylate sections with 0.1M TEA-HCl/0.25%(v/v) acetic anhydride (pH8, adjusted with

NaOH, prepare fresh) for 10min. Wash with RNase-free PBS for 5min.

(7) Dehydrate in 70% ethanol for 5min, 95% ethanol for 2min, and air dry for 30min.

(8) Heat RNA probes (Table 2-2) in filtered hybridization solution at 80°C for 2min. Add

200µl to sections. Cover with a parafilm coverslip and avoid bubbles. Hybridize at 55°C

for 16-18 hours in a humid box.

Table 2-2 RNA probes for in situ hybridization.

Gene Probe length dilution BM purple incubation time

Ihh 1.2kb 1:250 2-3 days

Gli1 1.7kb 1:250 3 days

Ptch1 840bp 1:500 2 days

Col2a1 404bp 1:500 1 hour

Col10a1 1.3kb 1:500 2 hours

Pthlh 1.5kb 1:100 3 days

Pth1r 444bp 1:250 2-3 days

Runx2 315bp 1:250 3 days

BMP2 1.2kb 1:250 3 days osteocalcin 307bp 1:250 2 days

(9) Float off coverslips in 5x SSC.

(10) Wash in high stringency buffer 2x SSC/50% formamide at 65°C for 30min.

(11) Triple wash in RNase buffer at 37°C for 10min each. Digest with 20ug/ml RNase A in

RNase buffer at 37°C for 30min. Wash with RNase buffer at 37°C for 15min.

(12) Wash twice in high stringency buffer at 65°C for 20min.

(13) Wash with 2x SSC at 37°C for 15min, then 0.1x SSC at 37°C for 15min.

(14) Wash with PBST for 15min.

(15) Block with 300µl PBST/10% (v/v) goat serum for 1h in a humid box.

(16) Incubate in 300µl alkaline phosphatase-coupled anti-digooxigenin antibody 1/5000 in

PBST at 4°C overnight.

(17) Wash five times in PBST for 1 hour each.

(18) Wash twice in NTMT for 10min each.

(19) Incubate in 300µl BM purple/levemisol in humid box in dark. Stop reactions when an

obvious purple color is developed and background staining is still minor (Table 2-2).

(20) Lightly counterstain with Nuclear Fast Red for 1-5min. Rinse the excessive dye off with

water. Mount the slides.

a. For a quick mounting, mount with water-based mounting medium.

b. For a better differentiation of nuclei, dehydrate with ethanol series, rinse briefly

with xylene substitute and mount with Xylene Mounting Medium (Thermo).

2.1.5 Quantitative Reverse Transcriptase Polymerase Chain Reaction (qRT-PCR)

We employed qRT-PCR to quantitatively analyze the RNA level in bone samples.

(1) Isolate the samples. For postnatal bones, remove muscle and marrow and pulverize them

in liquid nitrogen.

(2) Isolate the RNA with a NucleoSpin RNA kit (Macherey-Nagel).

(3) Measure RNA concentration with Nanodrop.

(4) Set up reverse transcription with qScript cDNA SuperMix (Quanta Biosciences).

Use 1μg of RNA.

(5) Test the cDNA by running a regular PCR with AccuStart and GAPDH primers. Dilute

the cDNA 1:10 and store at -20°C.

(6) Performed the qRT-PCR in a StepOne Plus Real-time PCR system (Applied Biosystems)

with PerfeCTa SYBR Green SuperMix (Quanta Biosciences).

(7) Analyze the data and perform statistical test (t test).

Table 2-3. qRT-PCR primers.

Target Primers Product (bp)

Ptch1 Forward: 5’-CTCCAAGTGTCGTCCGGTTT-3’ 51

Reverse: 5’-ACCCATTGTTCGTGTGACCA-3’

Gli1 Forward: 5′-CGTTTGAAGGCTGTCGGAAG-3′ 51

Reverse: 5′-GCGTCTTGAGGTTTTCAAGGC-3′

Ihh Forward: 5’-CCCCAACTACAATCCCGACATC-3’ 256

Reverse: 5’-CGCCAGCAGTCCATACTTATTTCG-3’ 309

Pthlh Forward: 5′-TTCAGCAGTGGAGTGTCCTG-3′ 130

Reverse: 5′-TTGCCCTTGTCATGCAGTAG-3′ 310

Col1a1 Forward: 5'-CACCCTCAAGAGCCTGAGTC-3' 253

Reverse: 5'-GTTCGGGCTGATGTACCAGT-3' 311

SP7 Forward: 5'-CCAGCCTCTGGCTATGCAAA-3' 51

Reverse: 5'-AGGAAATGAGTGAGGGAAGGGT-3' 311

Acp5 Forward: 5'-CAGGAGACCTTTGAGGACGTG-3' 175

Reverse: 5'-GTGGAATTTTGAAGCGCAAAC-3' 312

Ctsk Forward: 5'-TGGCTCGGAATAAGAACAACG-3' 207

Reverse: 5'-GCACCAACGAGAGGAGAAATG-3' 312

Gapdh Forward: 5’-GTCGGTGTGAACGGATTTGG-3’ 278

Reverse: 5’-GACTCCACGACATACTCAGC-3’ 313

Spopl Forward: 5'-TCAACGTTTTCTTCAGGCCC-3' 237

Reverse: 5'-AAATCCCCAGTCCTTCCCCT-3'

2.1.6 Hematoxylin and Eosin (H&E) stain

Hematoxylin is a dye that stains the nuclei blue in the presence of aluminum ions 303.

Eosin is an anionic dye that binds cationic groups of proteins in almost all tissues except those rich in proteoglycans and glycoproteins such as the cartilage matrix 303. Therefore, H&E stain is a popular method to access tissue morphology.

Procedure

(1) Sacrifice the mice, dissect out the tissues for analysis, and fix in PBS/4% PFA at 4°C.

(2) (optional, for bones only) Decalcify the calcified tissue in PBS/14% EDTA (pH 7)

thoroughly.

(3) Embed the samples in paraffin. Store at 4°C until sectioning.

(4) Cut paraffin sections at 5um thickness. Float the sections on a warm water bath to unfold

the wrinkles for up to 1 hour, and then collect desired sections with glass slides (VWR,

cat #).

(5) Incubate slides in a slide warmer at 35°C overnight for adherence. Store the slides at

room temperature.

(6) Load the Gemini ES Automated Slide Stainer (Thermo) with SelecTech hematoxylin and

eosin staining system (Leica). All reagents can be reused for up to 10 runs.

(7) Run the program H&E with heat.

(8) Mount the slides with Xylene mounting medium (Thermo, cat #).

2.1.7 Von Kossa stain

Von Kossa method is a light-catalyzed reaction to stain the bone matrix. The silver ion reacts with phosphate groups, forming a light sensitive yellow precipitate which is subsequently converted to black precipitate 303. The sensitivity of von Kossa staining is lower than Alizarin

red, but it works better on sections as the high pH for Alizarin red tend to detach sections from glass slides.

(1) Samples should be processed without decalcification. Embed samples in paraffin blocks

and cut 5um sections.

(2) Deparaffinize sections.

(3) Rinse in water for three times.

(4) Immerse slides in 1%AgNO3. Place samples under a 150W lamp for 15min.

(5) Stop the reaction by rinsing briefly in water and then immersing in 5%NaS2O3.

(6) Rinse briefly and lightly counterstain with hematoxylin (tuloidine blue is not as dark as

hematoxylin and possibly provides a better contrast).

(7) Dehydrate with an ethanol series. Clear with xylene substitute, and mount the slides.

2.1.8 Immunohistochemical staining

Method 1

To measure the percentage of proliferating and apoptotic cells in bones, immunostaining of Ki67 and cleaved caspase 3 were performed on paraffin sections. Paraffin sections allow the best preservation of tissue morphology. Due to strong autofluorescence in the tissue, chemical chromogen (DAB) method was employed.

Solutions

Citrate buffer for antigen retrieval

Dissolve 10mM citrate and 0.05% Tween-20. Adjust pH to 6.0 with sodium

hydroxide.

Blocking buffer

Mix 0.1%Triton X-100 and 1% goat serum into PBS. Store at 4°C.

Procedure

(1) Deparaffinize sections.

(2) Antigen retrieval.

a. Preheat citrate buffer to 100°C.

b. Immerse slides in citrate buffer and keep 95-100°C for 10 min (Do not boil).

c. Allow slides to cool in buffer for 20min.

d. Rinse in dH2O three times, 2min each.

(3) Quench endogenous peroxidase activity.

a. 1%H2O2 in dH2O 10min.

b. Rinse in dH2O three times, 2min each.

Table 2-4. Primary antibodies for immunohistochemistry.

Antigen Donor species Type Reference Dilution

Ki67 Rabbit Polyclonal Abcam, ab15580 1:250

Cleaved caspase 3 Rabbit Polyclonal Cell Signaling, 9661 1:300

GFP Rabbit Polyclonal Life Technologies, A11122 1:500

FoxA2 Mouse Monoclonal DHSB, 4C7 1:40

Nkx2.2 Mouse Monoclonal DHSB, 74.5A5 1:40

Olig2 Rabbit Polyclonal Millipore, AB9610 1:1000

Nkx6.1 Mouse Monoclonal DHSB, F55A12 1:500

Pax6 Mouse Monoclonal DSHB, P3U1 1:500

Note: The monoclonal antibodies against Foxa2, Nkx2.2, Nkx6.1 and Pax6 developed by Drs. Jessell,

Madsen and Kawakami were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological

Sciences, Iowa City, IA 52242.

(1) Block and permeabilize sections in blocking buffer for 1h.

(2) Primary antibody (Table 2-4) incubation in blocking buffer at 4°C overnight.

(3) Wash three times in PBS for 10min each.

(4) Incubate slides in blocking buffer with horseradish peroxidase-conjugated secondary

antibody for 2 hours.

(5) Wash three times in PBS for 10min each.

(6) Develop with 0.5mg/mL DAB/ 0.05%NiCl2 for 3min. If cobalt is used in place of nickel,

staining will be a darker.

(7) Counterstain with Toluidine blue.

(8) Dehydrate with ethanol series and mount with Shandon™ Xylene Substitute Mountant

(Thermo Scientific, #1900231).

Method 2

To investigate the domains of spinal cord progenitors, immunofluorescence of a series of markers were analyzed. E9.5 and E10.5 embryos were lightly fixed (for 30min to 1 hour) in

PBS/4% PFA and cryosectioned at 10μm thickness (see RNA in situ hybridization protocol).

Sections were stored in -80°C.

Solution

Blocking buffer

Mix 0.1%Triton X-100 and 1% goat serum into PBS. Store at 4°C.

Procedure

(1) Thaw slides in the air for 1 hour.

(2) Block in blocking buffer for 1 hour.

(3) Incubate in 300μl blocking buffer with primary antibody (Table 2-4) at 4°C overnight.

(4) Wash three times with blocking buffer for 10min each.

(5) Incubate in 300μl blocking buffer with cy3-conjugated goat-anti-mouse or goat-anti-

rabbit secondary antibody (Jackson lab) for 2 hours.

(6) Wash three times with blocking buffer for 10min each.

(7) Mount with Dabco 33-LV (Sigma-Aldrich, 290734). Preserve slides in dark at 4°C.

(8) Take photos with a Nikon E600 microscope and a QImaging Micropublisher Digital

Camera.

2.1.9 Micro-Computed Tomography (µCT)

µCT provides a comprehensive measurement of the 3-dimensional structure of bones 314.

The distal femur and proximal humerus are the ideal bones for µCT analysis. We performed the

µCT scanning on 10-week-old femurs with the help of Noriaki Okita and Dr. Neil Sharkey

(Department of Kinesiology, Pennsylvania State University). The data extraction was also done by Noriaki Okita.

Procedure

(1) Sacrifice the mice at 10-weeks-old. Dissect the femurs out gently and remove muscles.

Fix the femurs in PBS/4% PFA at 4°C for 24 hours. Then preserve femurs in PBS/NaN3

at 4°C for up to 3 months.

(2) Dry the femurs since the scan medium is air, and move to 8-well cartridge with distal end

facing downward. Immobilize the bones with tissue paper.

(3) Scan the femurs in a Scanco μCT40 Desktop MicroCT Scanner (SCANCO Medical AG,

Zurich, Switzerland) at an isotropic voxel size of 15μm with the acquisition setting of

55kVp, 145μA and 200ms integration time. (By Noriaki Okita)

(4) View the images with company-provided software. Identify the proximal end of distal

growth plate. Count 50 slices (totaled 750μm) toward the mid-shaft to ensure the

analysis is performed in the metaphysis, and then select 100 slices for trabecular analysis.

(5) Contour.

(6) 3D modeling and analysis of parameters. (By Noriaki Okita)

2.2 Biochemistry, Molecular and Cell Biology

2.2.1 Molecular Cloning

In the present study, we subcloned Spop coding sequence into pEGFP-C1 and pFLAG-

CMV2 vectors, and Gli3 into pcDNA3.2-3xFLAG vector. pcDNA3.2-3xFLAG is a generous gift from Dr. Yingwei Mao (Pennsylvania State University).

Procedure

(1) Amplify the target gene sequence with PCR. Use Phusion Hot Start II as the polymerase.

Primers are shown in Table 2-5.

(2) Digest the PCR product and vector with NEB endonucleases for 4 hours at 37°C.

(3) Add 0.5µl Alkaline Phosphatase (NEB, M0290S) to the digestion reaction of vector.

Continue incubation for 1 hour at 37°C. Alkaline Phosphatase dephosphorylates the 5’

and 3’ ends of digestion product, preventing the self-annealing of linearized vectors in

subsequent steps.

(4) Purify the PCR product with QIAquick PCR purification kit (Qiagen, #28104). Elute in

30µl elution buffer.

Table 2-5. Primers for subcloning into expression vectors. gene Target vector Primer sequence

Spop (XhoI pFLAG-CMV2 Forward: 5’-gaattcCTCGAGTCAAGGGTTCCAAGTCCTCCA-3’

+ BamHI) (SalI + BamHI) Reverse: 5’-gaattcGGATCCTTAGGATTGCTTCAGGCGTTTG-3’

Spop pEGFP-C1 Same primer and restriction endonuclease as above.

Gli3 (NotI pcDNA3.2- Forward: 5’-gattaagcggccgcTGAGGCCCAGTCCCACAG-3’

+ XbaI) 3xFLAG (NotI Reverse: 5’-

+ XbaI) cattaatctagaCTATTGCATAACTGCAAGGAATTTGCTT-3’

Gli3-1-700 pcDNA3.2- Forward: 5’-gattaagcggccgcTGAGGCCCAGTCCCACAG-3’

(NotI + 3xFLAG (NotI Reverse: 5’- cattaatctagaTCTTCTCTGCCTTGACGGTTTTCA-3’

XbaI) + XbaI)

(5) Load the vector digestion product to a 0.8% agarose gel and run 90V to separate the

linearized vector from supercoils and nicked plasmids. Purify the linearized vector with

QIAquick Gel Extraction Kit (Qiagen, #28704). Elute in 30µl elution buffer.

(6) Set up a 15µl ligation reaction with T4 DNA Ligase (NEB, M0202S). The ratio of insert

and vector is 3:1. Hold 16°C overnight.

(7) Transform E. coli DH5α competent cells. Add 10µl to 100µl competent cells, incubate

on ice for 30min, heat shock 42°C 90s, return to ice for 2min, add 900µl SOC medium

and incubate 37°C for 30min, and plate all cells. Use ampicillin plates for pFLAG-

CMV2 and pcDNA3.2-3xFLAG, and kanamycin plates for pEGFP-C1 vector.

(8) Pick colonies and run crude minipreps to extract plasmids. Verify the insert with

restriction analysis and subsequent sequencing.

(9) Perform clean minipreps with NucleoSpin Plasmid (Macherey-Nagel, #740588) to

amplify the plasmids carrying desired inserts.

2.2.2 site-directed mutagenesis

In this study, we mutated the five Spop degrons on Gli2 sequentially through site-directed mutagenesis to generate Gli25m. This was achieved through whole plasmid PCR amplification with Phusion and mutagenesis primers.

Procedure

(1) Set up PCR reaction.

Final concentration Volume

5x HF buffer 1x 4µl

10mM dNTP 0.2mM 0.4 µl

Primers 0.5µM each 2ul each

Template 10ng

Phusion Hot Start II 0.02U/µl 0.4µl

Bring the total volume to 20 µl with water

Primers used for the mutagenesis were listed in Table 2-6.

Table 2-6. Primers for site-directed mutagenesis of Gli2.

Locus Primers Tm

Spop degron Forward: 5’-gcTgcCgcCAACTGTCTAAATGATGCCAACCAG-3’ Tm=67.56

1 Reverse: 5’-GgcGgcGAGTTGGGTAGGCATGGTGCT-3’ Tm=67.64

Spop degron Forward: 5’-gCCgcCCACACTGTGGAGGACTGCC-3’ Tm=68.67

2 Reverse: 5’-GgcAgcGGCCTCCACACTCTCCTCAG-3’ Tm=66.8

Spop degron Forward: 5’-gCTGGCgCCGTGGATGCC-3’ Tm=69.73

3 Reverse: 5’-AgcCACCgCATTCCACTGCACAGGCAT G-3’ Tm=68.19

Spop degron Forward: 5’-gcCgcCgCCGGAGGTCTAGACAGCACC-3’ Tm=66.45

4 Reverse: 5’-CTGCACGGcTTGTGGATTATATCCTG-3’ Tm=69.1

Spop degron Forward: 5’-GTGGcCgCCCAGCTCCTGGAGCCC-3’ Tm=68.96

5 Reverse: 5’-GGcGgcAGcCACCTGGTTGACTCCCGG-3’ Tm=67.24

Since the melting temperature (Tm) of all these primers fell in the range of 66-70°C, 67°C was used as the annealing temperature for PCR.

(2) Run PCR program:

Cycle

number

1x 98°C, 30s to 3min Initial denaturation

35x 98°C, 10s Denaturation

67°C, 30s Annealing

72°C, 5min (30s/kb) Extension

1x 72°C, 10min Final extension

1x 4°C, on hold

(3) Purify the PCR product with QIAquick PCR purification kit (Qiagen, #28104). Elute in

30µl nuclease-free water. This step removes the free nucleotides, enzymes and salt.

(4) Phosphorylate the 5’ end with 0.5µl T4 PNK (NEB, M0201S) at 37°C for 30min to

enable the self-ligation of the PCR product.

(5) Add 1µl T4 DNA ligase (NEB, M0202S) and hold at 16°C overnight to ligate the PCR

product.

(6) Add DpnI (NEB, R0176S) at 37°C for 2 hours to cleave the plasmid templates which are

methylated.

(7) Transform DH5α competent cells with 10µl of digestion product.

(8) Identify the clones with desired sequence through restriction analysis and sequencing.

2.2.3 Cell Culture and Transfection

HEK 293T cells and immortalized mouse embryonic fibroblasts (MEFs) were cultured in complete culture medium, i.e. DMEM (Cellgro) supplemented with 10% fetal bovine serum

(Thermo Fisher), glutamax, penicillin and streptomycin. For serum starvation to induce ciliation, cells were rinsed briefly with PBS and then cultured in starvation medium, i.e. DMEM supplemented with 0.25% FBS, glutamax, penicillin and streptomycin.

HEK 293T and MEFs were transfected with PEI (Polysciences), prepared by Huiqing

Zeng in our lab, transfection protocol as follows:

(1) Pass the cells one day before transfection at a density that cells grow to 60-80%

confluence at the time of transfection. Keep HEK 293T and MEF cells in complete

culture medium.

(2) Add DNA to opti-MEM. Mix briefly and spin down.

Culture vessel Culture Opti-MEM plasmid PEI

medium

24-well plate 500µl 50µl 0.5µg 3.5 µl

6-well plate or 35mm 2ml 250µl 2.5µg 17.5µl

dish

60mm dish 5ml 500µl 5µg 35µl

100mm dish 10ml 500µl 15µg 90µl

Test the cytotoxicity and transfection efficiency in advance to determine the optimal

condition.

(3) Add PEI to opti-MEM. Mix by vortex for 10sec. Spin briefly.

(4) Incubate for 10min. Add the mixture to cells.

(5) Change medium 6 to 12 hours later.

2.2.4 Immunoblot

Immunoblot is a classic approach to detect and semi-quantitate proteins by antibodies that specifically target epitopes on them. Proteins of different molecular weights are first separated through a sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to nitrocellulose or PVDF membranes for antibody detection. Since we used the

Odyssey CLx Infrared Imaging System (LI-COR) for immunoblot, we revised our protocol based on the manufacturer’s instruction.

Solutions

SDS-PAGE loading buffer (6x)

Dissolve 1g SDS, 0.93g DTT and 1.2mg Bromophenol blue in 7ml 0.5M Tris,

pH6.8. Add 3ml glycerol. Mix well, aliquot and store at -20°C.

SDS-PAGE running buffer

Dissolve 12.1g Tris, 57.6g glycine and 4g SDS in 4 liters of water.

Transfer buffer

For 10x stock, dissolve 3g Tris and 14.4g glycine in 1 litter of water. Before

transfer, mix 100ml 10x stock and 200ml methanol into 700ml water.

Procedure

(1) Prepare samples. Measure protein concentration with Pierce BCA Protein Assay Kit

(Thermo Fisher, 23225) to determine the loading amount, and then adjust the volume of

all samples to the same with the same lysis buffer used to lyse the samples. Mix with

SDS-PAGE loading buffer.

(2) Prepare SDS-PAGE. Load the samples along with Odyssey One-Color Protein

Molecular Weight Marker (LI-COR, P/N 928-40000).

(3) Run the SDS-PAGE at 10mA per gel until the Bromophenol blue is out of the gel.

(4) Set up the wet transfer sandwich. Use 100V for 3 hours with ice-water bath (low-

molecular weight proteins) or 25V for 16 hours in cold room (large proteins such as Gli)

to transfer the proteins to nitrocellulose membrane.

(5) Rinse the membrane twice with PBS. Then block with PBS/5% milk (for mouse and

rabbit origin primary antibodies) or PBS/5% chicken serum (for goat-derived primary

antibodies) for 1 hour with rocking.

(6) Wash twice with PBS for 5min each. This prevents milk contamination of primary

antibody so it can be reused.

(7) Incubate in primary antibody diluted in PBS/5% BSA/0.1% Tween-20/sodium azide at

4°C overnight on a shaker.

Table 2-7. Primary antibodies for immunoblot.

Antigen Donor species Type Reference Dilution

Spop Goat Polyclonal Santa Cruz, C-14 1:200

Gli2 Goat Polyclonal R&D Systems, AF3635 1:500

Gli3 Goat Polyclonal R&D Systems, AF3690 1:200

GFP Rabbit Polyclonal Life Technologies, A11122 1:500

FLAG Mouse Monoclonal Sigma-Aldrich, F1804 1:2500

HA Mouse Monoclonal Covance, MMS-101R 1:2000

Myc Mouse Monoclonal Santa Cruz, 9E10 1:2500

Cul3 Goat Polyclonal Santa Cruz, C-18 1:2500

Ubiquitin Mouse Monoclonal Santa Cruz, P4D1 1:200

β-tubulin Mouse Monoclonal Sigma-Aldrich, T5201 1:10,000

(8) Wash three times with PBS/5% BSA/0.1% Tween-20 for 5min each.

(9) Incubate in secondary antibody diluted in PBS/5% milk for 1 hour in dark. Secondary

antibodies are donkey-anti-mouse conjugated with IRDye-680RD (LI-COR, P/N 925-

68072), and donkey-anti-rabbit (LI-COR, P/N 925-32213) or donkey-anti-goat

conjugated with IRDye-800CW (LI-COR, P/N 925-32214).

(10) Wash three times with PBS/5% BSA/0.1% Tween-20 for 5min each. Then rinse twice

with PBS. Dry the membrane with clean filter paper.

(11) Scan with LI-COR Odyssey CLx (LI-COR) imager and quantify the relative protein

abundance with Image Studio (LI-COR).

2.2.5 Immunoprecipitation (IP) and Co-IP

Immunoprecipitation was performed with a FLAG-IP kit (Sigma-Aldrich, FLAGIPT1) following the manufacturer’s instruction.

Solutions

Cell lysis buffer (for extraction of nuclear proteins)

Mix 1ml 1% Triton X-100, 800µl 5M NaCl, 100µl 1M Hepes, pH 7.5, 2µl 0.5M

EGTA, 5µl 1M DTT, 2ml glycerol into 6.1ml water. Store at -20°C. Add

protease inhibitors aprotinin, leupeptin and pepstatin before use.

Wash buffer

Mix 2.5ml 1M Tris HCl, pH 7.5 and 1.5ml 5 M NaCl into 46ml water. Add

protease inhibitors before use.

Procedure

(1) Pass cells to 60mm dishes and transfect the cells the next day at 60-80% confluence with

5µg plasmids and 35 µl PEI.

(2) Rinse the cells twice with PBS and scrape the cells in PBS/0.1% EDTA. Spin 400g 4°C

5min to collect the cells.

(3) Lyse the cells with 400µl lysis buffer 25min on ice. Spin 16000g 4°C 10min and collect

the supernatant. Repeat the spin to get rid of debris completely.

(4) Aspirate beads, 20µl per sample. Beads are already bound with anti-FLAG antibody.

Wash the beads three times with wash buffer. Spin 5000g 30s to collect the beads each

time.

(5) Incubate lysate with beads 4°C overnight on an end-over-end rotator.

(6) Wash beads with 1ml lysis buffer at 4°C for 5min on an end-over-end rotator. Spin

5000g 4°C 30s and collect the beads. Use an insulin syringe and needle to remove the

supernatant thoroughly.

(7) Repeat Step (6) twice but with wash buffer.

(8) Add 30 to 45µl 1x SDS loading buffer to beads. Boil 5min to release the proteins. Cool

down and spin briefly. Store at -20°C.

2.2.6 In vivo Ubiquitination Assay

In vivo ubiquitination assay is employed here to determine whether the presence of the E3 ubiquitin ligase containing Spop promotes the ubiquitination of substrates. We referred to Zhang et al. (2006) when developing this protocol. HEK 293T cells were used for maximum overexpression efficiency.

Solutions

denaturation buffer

Mix 50µl 20% SDS into 1ml 10mM Tris-HCl, pH7.4. Add 1mM DTT right

before use.

1% Triton lysis buffer

Mix 5ml 1% Triton X-100, 15ml 5M NaCl, 10ml 1M Hepes, pH 7.5, 50ml

glycerol, 1ml 0.5M EDTA into 419ml water.

Procedure

(1) Pass cells in the previous day to 60mm dishes and transfect the cells at 60-80%

confluence with 5µg plasmids and 35 µl PEI.

(2) 12 hours later, change to starvation medium and starve the cells for 40 hours.

(3) Treat with 50μM MG132 (Sigma-Aldrich) in starvation medium for 8 hrs prior to

harvesting.

(4) Rinse cells twice with PBS. Then scrape the cells in 1ml PBS/0.1%EDTA with a cell

scraper (Greiner bio-one, #541070). Spin 400x g at 4°C for 10min to collect the cells.

(5) Lyse the cells with 90µl denaturation buffer. Sonicate at 30% output for 10 cycles. Add

810µl of 1% Triton lysis buffer and incubate on ice for 15min to ensure efficient lysis

and dilute SDS.

(6) Continue with immunoprecipitation using the FLAG-IP kit and subsequent immunoblot.

2.2.7 Immunocytochemistry

Solutions

Blocking buffer

Mix 10µl 10% Triton X-100 and 100µl goat serum into 890µl water. Make right

before use.

Procedure

(1) Prepare the coverlips. Sterilize the coverslips in 70% ethanol and air dry. Place

coverslips in a 24-well plate. Coat the coverslips with 0.1% gelatin for 1 hour. Remove

gelatin solution.

(2) Grow MEF cells on the coverslip for 1 day to about 50% confluence. Transfect cells

with PEI as in Chapter 2.2.3 and starve cells for 24 to 48 hours before cells are confluent.

(3) Rinse cells once gently with warm PBS.

(4) Fix in fresh PBS/4% PFA for 10min.

(5) Rinse and permeabilize twice in PBS/0.1% Triton X-100 for 5min each.

(6) Block in 35µl blocking buffer for 10min.

(7) Incubate with primary antibody rabbit-anti-GFP (Life Technologies, A11122) 1:500 in

35µl PBS/10% goat serum for 2 hours.

(8) Rinse three times in PBS/0.1% Triton X-100 for 5min each.

(9) Incubate with secondary antibody Cy3-conjugated goat-anti-rabbit (Jackson lab) 1:500 in

35µl PBS/10% goat serum for 2 hours.

(10) Rinse three times in PBS/0.1% Triton X-100 for 5min each.

(11) Rinse in water.

(12) Mount with VECTASHIELD Antifade Mounting Medium with DAPI (Vector

Laboratories, H-1200) and store in dark.

(13) Take photos with a Nikon E600 microscope and a QImaging Micropublisher

Digital Camera.

Chapter 3

Requirement of Spop for the skeletal development and Ihh signaling

This chapter is recently published in PNAS 315.

The human skeleton forms through endochondral ossification in which chondrocytes in the growth plate proliferate, undergo hypertrophic differentiation and secrete calcium-containing extracellular matrix 207. The osteoblasts in the perichondrium, a thin layer of tissue surrounding the cartilage, replace the dying chondrocytes and secret more bone matrix. On the other hand, osteoclasts, derived from white blood cells, invade and digest bone matrix. The balance between the osteoblast and osteoclast activities allows calcium homeostatic control and bone health 207. In osteopenia and osteoporosis, conditions afflicting more than 10 million Americans, an abnormal decrease in osteoblast activity or increase in osteoclast activity results in the loss of bone mass 316.

Unfortunately, our understanding of these bone diseases has been hindered by incomplete knowledge in the molecular mechanisms underlying endochondral bone development and remodeling.

Indian Hedgehog (Ihh), a member of the Hedgehog (Hh) family of signaling proteins, is essential for endochondral bone development 208. Ihh regulates gene expression through the Gli family of transcription factors, which act as both transcriptional activators and repressors 1. In the absence of Hh, efficient proteolytic processing turns Gli3 into a transcriptional repressor (Gli3R), while processing of Gli2 is rather inefficient 155,317,318. Hh inhibits Gli processing and converts the full-length Gli proteins into transcriptional activators 176. Both Gli2 and Gli3 play important roles in the regulation of bone formation downstream of Ihh 219–221,319.

Ihh maintains the expression of parathyroid hormone-like peptide (Pthlh) in periarticular perichondrium, which then stimulates the proliferation and delays the hypertrophic differentiation of chondrocytes, in part through Gli3R 216,217. Ihh also promotes chondrocyte proliferation and hypertrophic differentiation through Pthlh-independent mechanisms 209,214,320. In addition, Ihh signaling is required in the perichondrium for osteoblast differentiation and bone formation 208,219.

Speckle-type POZ Protein (Spop) is the substrate-recognition subunit of a Cullin-Ring E3 ubiquitin ligase that targets mammalian Gli2 and the full-length form of Gli3 (Gli3FL) for ubiquitination and degradation in vitro 173,238,288,289. The Spop homolog in Drosophila, known as hib/rdx, also mediates the degradation of Ci, the sole Gli family member in flies, and inhibits Hh signaling 53,54. Overexpression of Spop in Xenopus laevis similarly reduces Hh pathway activation

286. A Spop mutant mouse strain was analyzed previously, but only defects in endocrine pancreas were reported 277. Another Speckle-type POZ protein, Spop-like (Spopl), exhibits similar substrate specificity and similar, albeit somewhat weaker, ubiquitination activity as Spop 229. Its in vivo biological function has not been studied.

Here we report that loss of Spop, but not Spopl, disrupts chondrocyte hypertrophy and osteoblast differentiation in the mouse, suggesting requirement for Spop-mediated protein degradation in mouse skeletal development, Surprisingly, loss of Spop results in an increase in the level of Gli3R and a decrease in Ihh signaling. Consistent with this in vivo observation, we find that overexpressed Spop targets both Gli3FL and Gli3R for ubiquitination and degradation.

Supporting the role of increased Gli3R in Spop mutant phenotype, reducing the dosage of Gli3 restores normal Ihh signaling and endochondral ossification. Finally, we show that limb mesenchyme-specific loss of Spop results in shorter distal limb bones and lower bone density in the adults, which can be rescued by reducing the dosage of Gli3.

3.1 Loss of Spop induces neonatal lethality

3.1.1 Three Spop mutant alleles were generated

The previously characterized Spop gene-trap allele (Pcifgt) had two caveats. First, the investigations of Spop homozygous mutants were limited to embryonic stages since no viable mutants were recovered after postnatal day 1 277; and second, in rare cases gene trap gave rise to hypomorphic alleles rather than null alleles due to alternative splicing 321. To unravel the roles of

Spop in mouse development, we generated three new mouse strains, among which SpoplacZKI

(Spoptm1a(KOMP)Mbp) and Spop∆Ex were predicted to be null alleles, and Spopflox can be used for generating conditional mutants that survive to later postnatal stages (Figure 3-1).

Figure 3-1. Spop knockout strategy.

The SpoplacZKI allele contained a bacterial β-galactosidase (lacZ) reporter inserted into the

3rd intron of Spop (Figure 3-1). A splicing acceptor preceding the lacZ reporter was predicted to induce the splicing of the first two exons to the inserted sequences, which enabled the translation

of lacZ protein from the internal ribosome entry site (IRES). As a result, only the initial 26 residues of Spop protein could be translated, and the entire POZ and MATH domains would be missing.

By immunoblotting with a goat polyclonal antibody against endogenous Spop, we confirmed that

Spop protein was not produced in the embryos, and SpoplacZKI was a loss-of-function allele (Figure

3-2).

Figure 3-2. Spop protein are not detected in SpoplacZKI and Spop∆Ex embryos.

Immunoblots of E10.5 embryos with Spop and β-tubulin antibodies. Asterisks: non-specific bands.

Spopflox mouse strain was generated by breeding the SpoplacZKI strain to

Tg(ACTFLPe)9205Dym/J transgenic mice, which express flipase in the germline under the control of the human ACTB promoter 297. The flipase was predicted to induce the recombination of the

FRT sites flanking the lacZ expression cassette, resulting in the deletion of the sequence in between

(Figure 3-1).

Spop∆Ex strain was then generated by breeding Spopflox strain to Tg(EIIa-Cre)C5379Lmgd/J mice, which express Cre recombinase in the germline driven by adenovirus EIIa promoter 299. Cre can induce the recombination of the loxP sites flanking exons 4 and 5, causing the deletion of sequences in between (Figure 3-1). Exons 4 and 5 encode residues 27-117 which contribute to two-thirds of the MATH substrate-binding domain. The resulting Spop protein product was

predicted to be severely truncated at residue 26, and fused to a frameshift mutant if splicing occurs between exons 3 and 6. Immunoblotting confirmed that Spop∆Ex was a null allele (Figure 3-2).

3.1.2 Loss of Spop induced neonatal lethality

Similar to the previous gene-trap allele 277, the heterozygous carriers of SpoplacZKI were healthy and fertile, but the homozygous mutants only survived a few hours to days after birth

(Table 3-1). No SpoplacZKI/lacZKI mutants were recovered at P3.5 and at weaning, and the survival of mutants to P2.5 may require a mixed genetic background, as the SpoplacZKI/lacZKI mutants that survived to P1.5 and P2.5 all resulted from crosses between C3H/HeJ and 129S1/SvImJ bred parents.

Table 3-1. Number of embryos or pups that survive to indicated stages from a crossing between heterozygotes (SpoplacZKI/+ × SpoplacZKI/+).

Stage Total Spop+/+ SpoplacZKI/+ SpoplacZKI/lacZKI mutant%

E8.5-16.5 212 60 103 (1) 49 23.1%

E17.5-18.5 41 9 24 8 (1) 19.5%

P0 37 11 19 7 18.9%

P1.5 47 12 29 (1) 6 (5) 12.8%

P2.5 25 5 18 2 (3) 8.0%

P3.5 8 4 4 0 0.0%

At weaning(P21) 90 28 62 0 0.0%

The number in parentheses is the number of embryos/pups recovered at indicated stages that were already dead. This number was excluded from the subtotal. Mutant%, the percentage of homozygous mutants among all of the genotypes.

Removal of the lacZ expression cassette in the germline gave rise to Spopflox strain that was viable and fertile (Error! Not a valid bookmark self-reference.), suggesting that the neonatal lethality in SpoplacZKI strain was not caused by an off-target effect. Similar to the SpoplacZKI mice,

Spop∆Ex strain also exhibited neonatal lethal phenotype (Table 3-3).

Table 3-2. Number of embryos or pups that survive to given stages from a crossing between heterozygotes (Spopflox/+ × Spopflox/+)

Stage Total Spop+/+ Spopflox/+ Spopflox/flox flox/flox%

At weaning(P21) 61 17 26 18 29.5% flox/flox%, the percentage of Spopflox/flox pups among all of the genotypes

Table 3-3. Number of embryos or pups that survive to given stages from a crossing between heterozygotes (SpopΔEx/+ × SpopΔEx/+)

Stage Total Spop+/+ SpopΔEx/+ SpopΔEx/ΔEx mutant%

E9.5-16.5 412 91 (1) 235 (3) 86 (4) 20.9%

E17.5-18.5 97 35 (1) 47 15 (3) 15.5%

P0 40 7 28 5 12.5%

P1.5 22 7 13 2 (2) 9.1%

P2.5 38 15 23 0 0.0%

P3.5 10 5 5 0 0.0%

At weaning(P21) 50 21 29 0 0.0%

The percentage of Spop null mutant embryos (SpoplacZKI/lacZKI and SpopΔEx/ΔEx combined) was already significantly lower than the Mendelian rate at late gestation stage (E17.5-18.5)

(p<0.05 by χ2 test), indicating an embryonic lethality at a low penetrance; however, the majority of Spop mutants died after birth. To identify the cause of neonatal lethality, we tracked the newborns. At P0.5, we recovered Spop mutants alive or dead. The dead mutants (3/7

SpoplacZKI/lacZKI and 1/5 SpopΔEx/ΔEx) all had air in their lungs, suggesting that they survived the labor, and unlikely died of respiratory failure. We did not find any morphological difference between live Spop mutants and their littermates. Spop mutants exhibited similar body size and body weight as littermates at P0.5 (Figure 3-3), and the pups retrieved later than P0.5 had milk in their stomach. However, on P1.5 and P2.5, Spop mutants exhibit a significantly lower body weight and a much smaller body size (Figure 3-3 and data not shown), suggesting that the growth of Spop mutants were attenuated. The growth arrest was potentially caused by the reduced nutritional uptake as evidenced by the smaller milk spot in Spop mutants. The autopsy showed that the internal organs (thymus, lung, heart, liver, stomach, intestine, pancreas, spleen, kidney and adrenal glands) were present. We did not identify any defect except a reduction of heart size and thinned heart walls and septum that emerged from embryonic stages (see Chapter 5 for further analyses).

Autopsy also showed that Spop mutants did not die of cleft palate.

Figure 3-3. The growth of Spop mutants are attenuated after birth.

(A) Lateral views of Spop mutants and wild type littermates at P0 and P1.

(B) Spop mutant pups weigh significantly less than wild type or heterozygous controls at P1 (control: n=12; Spop mutants: n=8) but not P0 (control: n=24, Spop mutants: n=5). ** p<0.01; ns: p>0.05 (t test).

3.2 Spop regulates skeletal development

The requirement of Spop for the survival suggested that Spop is critical for the development. However, it remained unclear which developmental event was regulated by Spop.

Expression profiles may provide a clue.

3.2.1 Spop is highly expressed in developing cartilage and bones

To investigate the expression pattern of Spop, we took advantage of the lacZ reporter gene inserted into the Spop locus in the SpoplacZKI allele. We found that at embryonic stages starting from E12.5, high levels of the Spop-lacZ reporter expression were present in the developing skeleton including long bones, costal cartilage, ribs, vertebrae and the primordium of dermal bones

(Figure 3-4 and data not shown). A closer look revealed lacZ expression in the chondrocytes, perichondrium, as well as mesenchymal cells surrounding the long bones that are possibly muscle precursors.

Figure 3-4. Spop is highly expressed in the developing skeleton.

1: mesodermal precursors of frontal bone; 2: mesodermal precursors of parietal bone; 3: cartilage primordium of exoccipital bone; 4: Cartilage primordium of petrous part of the temporal bone; 5: ear cartilage; 6: cartilage of long bones; 7: umbilical cord; 8: retina; 9: radius; 10: ulna; 11: humerus; arrowheads: perichondrium; *: chondrocytes.

(A) A lacZ-stained E13.5 SpoplacZKI/+ embryo.

(B) A horizontal section of a lacZ-stained E14.5 SpoplacZKI/+ head.

(D) A longitudinal section of a lacZ-stained E14.5 SpoplacZKI/+ forelimb.

3.2.2 Ossification is defective in Spop null mutants

The high levels of Spop-lacZ expression in the developing skeleton indicate that Spop potentially plays a role in the bone development. To investigate whether loss of Spop affected the skeletal development, we employed Alcian blue to stain acidic polysaccharides including

glucosaminoglycans in the cartilages, and Alizarin red to stain calcium which labels the ossification centers a red color. The calcification of ribs was already evident in the wild type embryos but not the Spop mutant littermates at E14.5 (Figure 3-5A). Similarly, the calcification of vertebrae, sternum and long bones were affected (Figure 3-5B, C, D).

Figure 3-5. Loss of Spop induces wide-spread ossification defect.

(A) Ossification centers are already evident in wild type ribs (arrowhead) at E14.5, but still look faint in

Spop mutants.

(B) Ossification centers appear in wild type, but not Spop mutant, vertebrae (arrowhead) at E15.5.

(C) Ossification centers are much smaller in Spop mutant ulna, radius, humerus and scapula (arrowhead) than in wild type littermates at E14.5.

(D) Arrowheads indicate the calcification not detected in the lower part of Spop mutant sternum and

Xiphoid process.

(E) Ossification centers appear in wild type, but not Spop mutant, metacarpals 2-4 (arrowhead) at E16.5.

(F) Ossification centers appear in wild type, but not Spop mutant, metatarsals 2-5 (arrowhead) at E17.5.

(G,H) Top view of skulls at E17.5 and P0 shows fenestrated dermal bones and expansion of fontanelles in

Spop mutants. AF: anterior fontanelle; PF: posterior fontanelle. n≥3 for all experiments.

In particular, the calcification of the metacarpals and metatarsals was severely compromised in Spop mutants (Figure 3-5E, F). At later stages, the ossification defect in Spop mutant ribs, vertebrae and long bones in the proximal appendage appeared less noticeable by

Alizarin red staining, but the defect in Spop mutant metacarpals and metatarsals became more prominent (Figure 3-6).

Figure 3-6. Metatarsals are not calcified before death of pups.

Length measurement of calcified region in metacarpal 3 and metatarsal 3. *: p<0.05 (t test). Data represent mean ±SEM (n≥3).

The ossification defect was not limited to the endochondral skeleton. The anterior and posterior fontanelles were enlarged in the absence of Spop, and there were more fenestrations in

Spop mutant frontal, parietal, interparietal and occipital bones, suggesting that intramembranous ossification was also affected by the loss of Spop (Figure 3-5G, H).

3.2.3 Spop regulates hypertrophic differentiation of chondrocytes

The robust expression of Spop in the cartilage suggests that a disruption in the hypertrophic differentiation of the chondrocytes potentially underlies the mineralization defects in Spop mutants.

To determine whether the hypertrophic differentiation was affected by loss of Spop, we sectioned the long bones and stained them with hematoxylin and eosin (H&E). The compromised hypertrophic differentiation was evidenced by the shortened distance between the proliferating zones in Spop mutant humeri (Figure 3-7A, A’). More severe defects were observed in metatarsals.

Whereas chondrocytes underwent hypertrophic differentiation beginning at E16.5 in wild type, the chondrocytes remained undifferentiated in the Spop mutant metatarsals at E18.5 (Figure 3-7B, B’ and data not shown). To better evaluate the hypertrophic differentiation of chondrocytes in Spop mutants, we examined the expression of genes specific to chondrocytes at various stages of differentiation in E13.5 humerus. At the transition from proliferation to hypertrophic differentiation, chondrocytes turned off Col2a1 expression and turned on Col10a1 expression

(Figure 3-7C, D). Interestingly, the Col2a1-/Col10a1+ domain was greatly reduced in the Spop mutant humerus, suggesting that hypertrophic differentiation was defective (Figure 3-7C’, D’). In addition, the two narrow domains of Pth1r and Ihh expression, both labeling the prehypertrophic chondrocytes separated by the hypertrophic zones in wild type, were much closer to each other in

Spop mutants, confirming the hypertrophic differentiation defects (Figure 3-7E-F’).

Figure 3-7. Loss of Spop disrupts hypertrophic differentiation of chondrocytes.

(A-B’) H&E-stained longitudinal sections of humeri and metatarsal 3. P: proliferating zone; H: hypertrophic zone; OC: primary ossification center. Arrows: bone collar.

(C-G’) RNA in situ hybridization (purple) on longitudinal sections of E13.5 humeri.

(C-D’) Hypertrophic chondrocytes (bracket) express Col10a1, but not Col2a1.

(E-F’) Pth1r and Ihh are expressed in prehypertrophic chondrocytes (bracket).

(G,G’) Pthlh is expressed in periarticular perichondrium. Hu: humerus; Ul: ulna. n≥3 for all experiments.

3.2.4 Spop promotes bone formation and osteoblast differentiation

The expression of Spop in the perichondrium, where osteoblasts are derived (Figure 3-4), suggests that Spop may be important for bone formation and osteoblast differentiation. Indeed, in striking contrast to the wild type metatarsal 3 where bone collar was already formed at E18.5, the bone collar appeared to be absent in Spop mutant metatarsal 3 (Figure 3-7B, B’). To further investigate the bone formation defects in Spop mutants, we performed the von Kossa staining on

E15.5 humeri. The orthotropic bone collar was present around the prehypertrophic chondrocytes in wild type, but not in Spop mutants, suggesting that Spop was indeed required for normal bone development (Figure 3-8A, A’). Detailed histological analysis indicated that the inner layer of the perichondrium in Spop mutants failed to properly differentiate into cuboidal osteoblasts, and the perichondrium was thicker than that in Spop heterozygous littermates (Figure 3-8B, B’).

A set of markers were commonly used for evaluation of osteoblast differentiation. Runx2 is a transcription factor expressed in osteoblast precursors which programs osteoblast differentiation 208. BMP2 is produced by mature osteoblast and critical for osteogenesis 322. Bglap, also known as osteocalcin, is a bone matrix protein produced by terminally differentiated osteoblasts 208. By RNA in situ hybridization, we found the Runx2 expression was downregulated in the perichondrium of Spop mutant metacarpals at E13.5, and Bmp2 expression was also weaker in the perichondrium at E15.5 (Figure 3-8C-D’). Of note, chondrocytes in the metacarpals did not undergo hypertrophic differentiation until E15.5, suggesting that the lack of Runx2 and Bmp2 expressions was not secondary to the defects in chondrocyte hypertrophy, and Spop likely promoted osteoblast differentiation in a cell-autonomous manner. Similarly, the expression of

Bglap was reduced in the perichondrium and primary spongiosa of Spop mutant humeri at E18.5

(Figure 3-8E, E’), indicating the lack of terminally differentiated osteoblasts and osteocytes in both cortical and trabecular bones.

Figure 3-8. Loss of Spop impairs bone formation and osteoblast differentiation.

(A) von Kossa-stained longitudinal sections of E15.5 humerus. Insets: the perichondrium (brackets) by the prehypertrophic zone. The orthotropic bone collar (arrows) is missing in Spop mutant.

(B) H&E-stained sections showing the perichondrium (between the dash lines) by the prehypertrophic zone of E15.5 proximal humerus. Arrows: cuboidal osteoblasts. P: proliferating zone; H: hypertrophic zone.

(C-E’) RNA in situ hybridization of Runx2, Bmp2 and Bglap.

Finally, we employed quantitative Reverse Transcription Polymerase Chain Reaction

(qRT-PCR) to measure the expression levels of Col1a1, a major extracellular matrix protein produced by osteoblast, and Sp7/osterix, a transcription factor required for osteoblast maturation, and found that both genes were already significantly downregulated at E13.5 stage in the Spop mutant forelimbs (Figure 3-19A). These results suggest that Spop promotes osteoblast differentiation and bone formation.

3.3 Spop promotes skeletal development by degrading Gli3 repressors

Previous in vitro studies identified Gli2 and Gli3 as the substrates of Spop, and suggested that Gli activators but not repressors were major targets of Spop for ubiquitination 53,289. In line with this notion, the Drosophila counterpart of Spop was found to attenuate Hh signaling by destruction and cytoplasmic sequestration of ci 53,54,323, and a negative role of Spop in Hh signaling by inducing Gli2 turnover was also uncovered in cancers 235,275. These findings suggested that

Spop potentially acted as a negative regulator of Ihh signaling by degrading Gli activators, and a boosted Pthlh production in the periarticular perichondrium resulted in the skeletal defect in the absence of Spop. To test this hypothesis, we analyzed the Ihh signaling activity and the abundance of Gli2 and Gli3.

3.3.1 Loss of Spop results in impaired Ihh signaling

To investigate whether Ihh signaling was affected by the loss of Spop, we examined the expression of Ptch1, a direct transcriptional target of Ihh, in E13.5 forelimbs. In contrast to previous in vitro observation suggesting Spop as a negative regulator of Hh signaling, we found a significant reduction in Ptch1 expression in E13.5 Spop mutant forelimbs through qRT-PCR, indicating a positive role of Spop in Ihh signaling (Figure 3-9A). RNA in situ hybridization showed

that Ptch1 expression was reduced in both chondrocytes and the perichondrium (Figure 3-9B). A

Ptch1-lacZ reporter was similarly expressed at a lower level in both chondrocytes and perichondrium of Spop mutants (Figure 3-9C), suggesting that Spop regulates bone formation by promoting Ihh signaling. In contrast, Ptch1-lacZ expression in the posterior part of E12.5 Spop mutant limb buds was indistinguishable from that of the littermates (Figure 3-9D), suggesting that loss of Spop has no obvious effect on Shh signaling from the zone of polarizing activity.

Figure 3-9. Loss of Spop induces Ihh signaling.

(A) qRT-PCR analyses of Ptch1 and Pthlh in E13.5 forelimbs.

(B) RNA in situ hybridization of Ptch1 on E13.5 forelimb sections.

(D) Xgal-stained E12.5 limbs showing Ptch1-lacZ expression.

Consistent with decreased Ihh signaling, Pthlh, whose expression is dependent on Ihh signaling, was also expressed at a lower level (Figure 3-9A). In situ hybridization also showed that

Pthlh expression was not expanded in Spop mutants (Figure 3-7G, G’). These results suggest that

Spop actually regulates endochondral ossification by positively regulating Ihh signaling, and the skeletal defects in Spop mutants do not result from elevated Pthlh production.

3.3.2 Spop targets Gli3R for ubiquitination and degradation

To investigate the molecular mechanisms underlying this positive role of Spop in Ihh signaling, we examined the levels of Gli2 and Gli3 in E13.5 forelimbs by immunoblot analyses.

Surprisingly, the level of Gli2, the primary activator of the Ihh pathway, was not significantly changed in Spop mutants. In contrast, the levels of both Gli3FL and Gli3R in the Spop mutant forelimbs were more than twice as those in wild type (Figure 3-10). As Gli3R was known to antagonize Ihh signaling in endochondral ossification, we conclude that Spop promotes Ihh signaling in skeletal development by downregulating Gli3R.

Figure 3-10. Gli3 is stabilized in Spop mutant.

Immunoblots of E13.5 forelimbs with Gli2, Gli3 and β-tubulin antibodies (n=3 for both groups). Gli2FL,

Gli3FL and Gli3R bands are quantified and normalized to β-tubulin expression levels.* p<0.05; ns: p>0.05 (t test).

As the levels of both Gli3FL and Gli3R were higher in Spop mutants, it was possible that

Spop targeted exclusively Gli3FL for turnover, which resulted in a decline of proteolytic processing that gave rise to Gli3R. Alternatively, since Spop interacts with the amino-terminus of

Gli3 238,289 and promotes ubiquitination of Gli31-90 148, Spop may regulate the level of Gli3R by directly targeting it for degradation. To test the latter possibility, we investigated the ability of

Spop to degrade Gli3R in cultured cells. Interestingly, co-expressing Spop in HEK 293T cells reduced the protein abundance of both Gli3FL and Gli31-700, a truncated form that mimics Gli3R

(Figure 3-11A). We then tested whether Spop physically interacts with Gli31-700. FLAG-Spop was co-expressed with GFP-tagged Gli3, Gli31-700 and Gli2, respectively, and then immunoprecipitated with FLAG antibody-conjugated beads. We found that similar to Gli3 and Gli2, Gli31-700 co- precipitated with Spop efficiently (Figure 3-11B). As a negative control, we found that Gli25m, a previously reported Gli2 variant with mutated Serine/Threonin-rich degrons, did not co-precipitate with Spop 238 (Figure 3-11B). Reverse immunoprecipitation confirmed that GFP-Spop co- precipitated with FLAG-tagged Gli3 and Gli31-700 (Figure 3-11C).

Figure 3-11. Spop targets Gli31-700 for ubiquitination and degradation.

Immunoblots with the indicated antibody. All experiments were conducted in HEK 293T cells.

(A) Co-expression of GFP-Spop, FLAG-Gli3, FLAG-Gli31-700 and Myc-Cul3 with Myc-GFP as a transfection control. Relative abundance of FLAG-Gli3 and FLAG-Gli31-700 was quantified and normalized to Myc-GFP, and the mean and SEM from three experiments were indicated under the panel.

(B) Co-immunoprecipitation (IP) between FLAG-Spop and GFP-tagged Gli3, Gli31-700, Gli2 or Gli25m.

(D, E) In vivo ubiquitination assay on Gli3 and Gli31-700. Cells were treated with MG132 for 8 hours prior to IP and immunoblot analyses. n=3 for all experiments.

Finally, we performed a ubiquitination assay in HEK 293T cells and found that overexpressed Cul3 and Spop promoted the ubiquitination of both Gli3 and Gli31-700 (Figure

3-11D, E). These results indicate that Gli3R is a direct target of Spop, and the increase in Gli3R in Spop mutants resulted at least partially from lack of Spop-mediated degradation.

3.3.3 The increase in Gli3 underlies the ossification defects in Spop mutants

Figure 3-12. Gli3 heterozygosity rescues the ossification defect in Spop mutants.

(A) Immunoblots of Gli3 and β-tubulin on E17.5 metatarsals.

(B) Alcian blue and Alizarin red stained P0 autopods.

To determine whether the increase of Gli3 abundance was responsible for the skeletal defects in Spop mutants, we generated Spop∆Ex/∆Ex;Gli3+/- double mutants in which the levels of

Gli3FL and Gli3R were comparable to those in wild type (Figure 3-12A). Importantly, Alcian blue and Alizarin red staining showed that the calcification of metacarpal 2, 5 and metatarsals was restored in the double mutants (Figure 3-12B). Histological analyses indicated that the defects in chondrocyte hypertrophy in Spop mutants were also rescued in Spop∆Ex/∆Ex;Gli3+/- double mutants

(Figure 3-12C).

Consistent with our observation that the levels of Gli3, but not Gli2, were altered in Spop mutant limbs, Spop;Gli2 double homozygous mutants exhibited the same ossification defects as

Spop single mutants at E17.5 (Figure 3-13A). Histological analysis showed that similar to Spop mutants, Spop;Gli2 double homozygous mutants failed to deposit bone matrix in the digit bones

(Figure 3-13B), and the chondrocytes also failed to undergo hypertrophic differentiation at this stage (Figure 3-13C). In contrast, loss of Spop had no effect on either digit formation or the ossification of the metacarpals and metatarsals in the absence of Gli3 (Figure 3-13D). The cartilage and bone matrix mineralization in Spop;Gli3 double homozygous mutants and that in Gli3 mutants were comparable (Figure 3-13E), and the hypertrophic zones were of similar length (Figure 3-13F).

These results suggest that Spop regulates endochondral bone development mainly through Gli3, but not Gli2.

Figure 3-13. Gli3, but not Gli2, is epistatic to Spop in skeletal development.

(A, D) Alcian blue and Alizarin red stained E17.5 autopods. Arrows: the ossification centers.

(B,E) von Kossa-stained E17.5 metacarpal 3. Arrowheads: the orthotopic bone collar.

(C,F) H&E-stained E17.5 metatarsal 3. Brackets: the hypertrophic zone. n≥3 for all experiments.

3.4 Loss of Spop induces brachydactyly and osteopenia

3.4.1 Defective hypertrophic differentiation of chondrocyte results in brachydactyly

As the neonatal lethality of Spop null mutants prevented us from assessing skeletal defects in adults, we generated a limb-specific Spop mutant allele (SpopcKO) by removing Spop in Prx1- expressing limb mesenchymal cells. Consistent with previous report 298, Prx1-Cre induced efficient recombination in limb mesenchyme starting from E9.5 (Figure 3-14 and data not shown).

Recombination was also observed in a salt and pepper pattern in cranial mesenchyme, brain, spinal cord, body wall, heart and interlimb flank mesoderm (Figure 3-14).

Figure 3-14. Prx1-Cre induces recombination in multiple tissues including limb bud.

Shown are a series of sagittal sections of an E11.5 R26R-YFP;Prx1-Cre double heterozygous embryo.

Sections were immunostained with GFP antibody for EYFP expression and counterstained with DAPI.

Both during postnatal development and in adulthood, the metatarsals of SpopcKO mutants were significantly shorter than their littermates (Figure 3-15A, B). In addition, quantitative analyses showed that metacarpals and phalanges, but not long bones in more proximal parts of the limbs, were significantly shorter in SpopcKO mutants than in their littermate controls (Figure 3-15C).

Therefore, SpopcKO can serve as a model for brachydactyly, a common but poorly understood congenital anomaly.

Figure 3-15. Tissue-specific ablation of Spop leads to brachydactyly.

(A) Alcian blue and Alizarin red-stained 5-month-old hindlimb autopods. Arrows: metatarsal 3.

(B) The length of metatarsal 3 at various stages (n≥3 per group and stage).

To determine whether changes in cell proliferation and/or apoptosis contribute to the short digits in Spop mutants, we characterized metatarsal bones at E17.5, when the length difference between Spop mutants and wild type was already obvious (Figure 3-16A). We estimated the number of chondrocytes along the longitudinal axis by counting the total number of nuclei enclosed in a 100μm-wide rectangle spanning metatarsal 3. Interestingly, the cell number in Spop mutant metatarsals was similar to that in wild type littermates (Figure 3-16B), suggesting that the shortness of metatarsals was not caused by a reduction in cell number. In line with this notion, the numbers of proliferating cells (Ki67+) and apoptotic cells (Cleaved Caspase 3+) were also comparable between Spop mutant and wild type metatarsals (Figure 3-16D, E). On the other hand, although cells were densely packed in the resting zone of wild type metatarsals, the size of cells in the hypertrophic zone was greatly expanded (Figure 3-16B, C). In contrast, the average cell size in the Spop mutant metatarsals was similar to that in the resting zone of wild type. These results suggest that compromised hypertrophic differentiation of chondrocytes underlies the brachydactyly in Spop mutants.

Figure 3-16. Defective hypertrophic differentiation accounts for the shortness of digit bones.

(A) Measurement of metatarsal 3 length (n≥3 for each group). * p<0.05.

(B) H&E stained metatarsal 3 at E17.5. Total cell numbers in the rectangles (all 100μm wide) were counted and labeled in the image (n=3 samples for each genotype). R: resting zone; P: proliferating zone;

H: hypertrophic zone. (a-c): close-up views of the hypertrophic, proliferating and resting zones of wild type; (d): a close-up view of Spop mutant cartilage.

(D-E) Percentage of Ki67 and cleaved caspase 3 positive cells in E16.5 metatarsal 3. Sections were immunostained and cells confined in a rectangle across the resting, proliferating and hypertrophic zones were counted.

3.4.2 Limb-specific Spop conditional mutants exhibit osteopenia

Figure 3-17. SpopcKO exhibits osteopenia.

(A) H&E stained-sections of 5-week-old distal femur. 1°: primary ossification center; the areas inside the rectangles are shown in the close-up view. 2°: secondary ossification center.

(B-H) μCT analyses of the trabecular bone in 10-week-old distal metaphysis of femurs. Voxel size: 15μm. n=3 males and 3 females per genotype.

(B) 3D reconstruction.

(D) Specific bone surface.

(E) Trabecular number per unit length.

(F) Trabecular thickness.

(G) Trabecular separation.

(H) Connectivity density. * p<0.05; ** p<0.01; ns: p>0.05 (two-way ANOVA).

Although the length of stylopod and zeugopod bones was not altered in SpopcKO mutants, these bones nevertheless exhibit structural defects. By detailed histological analyses of the femurs,

we found an obvious reduction in the density of bone matrix in both the primary and secondary ossification centers, as well as thinner bone spicules, in adult SpopcKO mutants (Figure 3-17A). To further investigate the structural defects of the bones of SpopcKO mutants, we evaluated the femurs with X-ray Micro Computed Tomography (μCT). The 3D reconstruction of distal metaphysis suggested an obvious reduction in trabecular bone density in SpopcKO mutants (Figure 3-17B).

Quantitative analyses of the μCT data indicated a significant reduction in bone volume fraction, along with a significant increase in the specific bone surface, i.e. ratio of bone surface to bone volume, in SpopcKO mutants (Figure 3-17C, D). The trabecular number per unit length, trabecular thickness and connectivity density were also significantly reduced (Figure 3-17E, F, H), whereas trabecular separation, i.e. the distance between trabeculae, was significantly increased (Figure

3-17G).

Throughout development, as well as in adulthood, bones undergo constant remodeling with osteoblasts producing bone matrix and osteoclasts resorbing it 207. To determine the mechanism underlying the loss of bone mass, we performed qRT-PCR to analyze the expression of osteoblast and osteoclast markers. Similar to earlier stages, the expression of both Col1a1 and Sp7 was compromised in E17.5 Spop mutant femurs, suggesting continued defects in osteoblast differentiation (Figure 3-18A, B). In contrast, the expression of Tartrate Resistant Acid

Phosphatase 5 (Acp5) and Cathepsin K (Ctsk), both encoding enzymes secreted by osteoclasts to remodel the bone matrix, were not significantly changed in E17.5 and p10 Spop mutants (Figure

3-18B, C), suggesting that the loss of bone mass was not the result of too much osteoclast activity.

Therefore, osteopenia in SpopcKO mutants is likely the result of impaired osteoblast function. ‘

Figure 3-18. Osteoblast, but not osteoclast, is affected in Spop mutants. qRT-PCR analyses of osteoblast markers Col1a1 and Sp7, and osteoclast markers Acp5 and Ctsk in Spop null (A, B), SpopcKO (C) and control samples. * p<0.05; ns: p>0.05 (t test). n≥3 for all experiments.

3.4.3 Gli3R dosage reduction rescues brachydactyly and osteopenia in SpopcKO mutants

Because the increase in Gli3R underlies the defects in chondrocyte and osteoblast differentiation in Spop null mutants, we hypothesized that increased Gli3R activity also accounted for the osteopenia and brachydactyly in SpopcKO mutant mice. To test this hypothesis, we reduced the dosage of Gli3R by generating SpopcKO;Gli3+/- double mutants. We found that the length of metacarpals, phalanges and metatarsals in 2-week-old SpopcKO;Gli3+/- pups was comparable to that in Spopflox/flox littermates, and significantly longer than that in SpopcKO mutant littermates

(Figure 3-19A). In addition, the histology of distal femur showed an increase in the bone density and thickness of bone spicules in SpopcKO;Gli3+/- than in SpopcKO mutants (Figure 3-19B).

Therefore, we conclude that dysregulation of Gli3R underlies both brachydactyly and osteopenia in SpopcKO mutant mice.

Figure 3-19. The bone defects of limb-specific Spop mutants result from increased Gli3.

(A) Alcian blue and Alizarin red-stained P14 hindlimb autopods, and quantification of bone length in digit

3. n=4 per genotype. * p<0.05; ns: p>0.05 (t-test). Arrows: metatarsal 3.

(B) H&E-stained distal metaphysis of femurs. Arrows: the range of bone spicules.

3.5 Spop-like does not regulate skeletal development

A recent in vitro study showed that Spop-like (Spopl), which shares 85% sequence identity with Spop, exhibits similar, but weaker E3 ubiquitin ligase activity than Spop 229. To investigate whether Spop and Spopl play redundant roles in mouse development, we characterized Spopl mutants and Spop;Spopl double mutants (Figure 3-20). Spopl homozygous mutants were viable and fertile with no apparent morphological and skeletal defects (data not shown), and Spop;Spopl double mutants exhibit similar skeletal defects to Spop mutants (Figure

3-21). These results indicate that Spopl does not compensate for the loss of Spop in mouse development.

Figure 3-20. Schematic illustration of Spopl mutant allele.

(A) Schematics of Spopl knockout strategy. The whole Spopl coding sequence was replaced by a lacZ- neomycin expression-selection cassette. The resulting Spopltm1(KOMP)Vlcg allele is a null allele and hereby abbreviated as Spopl-.

(B) Quantitative real time PCR showing the absence of Spopl transcript in E9.5 Spopl mutant embryos

(mean ± SEM, n=3 wild type, 4 Spopl+/- and 5 Spopl-/- embryos).

100

Figure 3-21. Loss of Spopl does not lead to obvious skeletal defects.

(A) Alcian blue and Alizarin red staining of E18.5 limb autopods. Loss of Spopl did not cause obvious change in the size of ossification centers in the metacarpals or metatarsals.

(B) Measurement of the length of metacarpal 3.

(C) The ratio of ossified bones was derived from the length of ossification center on metacarpal 3 divided by its total length. ns: p>0.05 (t-test). There was no significant difference between Spopl mutants and wild type, or between Spop;Spopl double mutants and Spop mutants (n=3 for all genotypes).

101

3.6 Conclusion

In conclusion, we provide here evidence for essential roles of Spop-mediated ubiquitination in chondrocyte and osteoblast differentiation during skeletal development, and normal bone size and density in adult. Importantly, we revealed a surprising positive function of

Spop in Ihh signaling through specific downregulation of Gli3, particularly its repressor form

(Figure 3-22). Clinically, this knowledge allows us to better understand the pathology and potential intervention of skeletal disorders such as brachydactyly and osteoporosis from a new perspective.

Figure 3-22. Model: Spop regulates skeletal development by degrading Gli3R.

(A) Spop promotes Ihh signaling by downregulating Gli3R, thereby activating genes critical for chondrocyte and osteoblast differentiation.

(B) Loss of Spop in chondrocytes compromises hypertrophic differentiation, which results in shorter digits in Spop mutants. Loss of Spop affects osteoblast differentiation and production of bone matrix, leading to osteoporosis or osteopenia.

102

In addition to Gli proteins, Spop mediates the ubiquitination of at least 20 more proteins

(Table 1-1). It is theoretically possible that changes in the activities of other Spop substrates may also contribute to the skeletal and neural defects observed in Spop single or compound mutants. However, the near-complete rescue of bone and cartilage development in Spop-

/-;Gli3+/- and SpopcKO;Gli3+/- double mutants, as well as the similar bone and cartilage phenotype in Gli3-/- mutant and Spop-/-;Gli3-/- double mutant, suggests that Spop-mediated ubiquitination of other substrates plays minimal roles in skeletal development at best.

Acknowledgements

We thank Drs. Andrea Mastro, Philip Reno, and Yingwei Mao for critically reading this manuscript; Drs. Lee Niswander (University of Colorado), Matthew Hilton (Duke University),

Henry Kronenberg (Harvard Medical School), Patricia Ducy (Columbia University), Jin Jiang

(University of Texas Southwestern Medical Center), Matt Scott (Carnegie Institute), Alex Joyner

(Sloan-Kettering Institute), Pao-tien Chuang (University of California, San Francisco), Edward

Yeh (University of Texas MD Anderson Cancer Center), and Bernard Luscher and Yingwei Mao

(Pennsylvania State University) for sharing reagents; Dr. Neil Sharkey and Noriaki Okita for great help in acquiring and analyzing the μCT data; and the Microscopy and Cytometry Facility and Genomics Core Facility of Pennsylvania State University for technical support. This work was supported by NIH Grant HD083625, a Pennsylvania State University Start-up Fund (to

A.L.), and a J. Lloyd Huck Dissertation Research Grant (to H.C.).

103

Chapter 4

Roles of Spop in spinal cord neural patterning and Shh signaling

Sonic hedgehog (Shh) plays a central role in the stereotypical arrangement of neural progenitors along the dorsal/ventral (D/V) axis of the ventral spinal cord in vertebrates 324. The

Shh protein, produced in the notochord and subsequently in the floor plate, forms a ventral-to- dorsal gradient in the ventral spinal cord, and regulates the expression of a multitude of target genes, which subsequently interact with each other to form an intricate gene regulatory network that helps to sharpen the borders between them and define the locations of various groups of neural progenitors325.

The direct transcriptional responses to Shh are mediated by the effector Glioma-associated oncogene (Gli) family of transcriptional factors, Gli1, Gli2 and Gli3 85,187,199,200,202. In the absence of Hh ligands, Gli2 and Gli3 are ubiquitinated by Cul1/Btrc, an E3 ubiquitin ligase 155,317. The ubiquitinated Gli3 is partially degraded in the proteasome to produce a strong transcriptional repressor (Gli3R), whereas ubiquitinated Gli2 mostly undergoes complete degradation. Loss of

Gli3 repressor activity results in ventralization of the forebrain and severe polydactyly in the limbs, but very subtle defects in the D/V patterning of the spinal cord, suggesting that the requirements for Gli3R activity vary dependent on the developmental contexts (Hui and Joyner, 1993; Persson et al., 2002; Theil et al., 1999; Tole et al., 2000).

Shh inhibits the Cul1/Btrc-based proteolytic processing of Gli3 and degradation of Gli2, and turns them into transcriptional activators (Gli2A and Gli3A) (Humke et al., 2010; Pan et al.,

2006; Wang et al., 2000). Loss of Gli2 results in the loss of floor plate and great reduction of V3 interneurons, ventral-most spinal cord cell types dependent on high concentrations of Shh,

104

suggesting that Gli2 is the primary activator of the Shh pathway 199,200. Interestingly, loss of both

Gli2 and Gli3 results in more complete loss of all floor plate and V3 interneurons, as well as the mixing of other ventral cell types including the motor neurons, V1 and V2 interneurons, suggesting that Gli3A also contributes to Shh pathway activation in the spinal cord (Bai et al., 2004; Lei et al., 2004). The third member of the Gli family, Gli1, whose transcription is dependent on Hh signaling, contributes to, but is not essential for, ventral spinal cord patterning (Bai et al., 2004;

Park et al., 2000).

In addition to Cul1-based proteolytic processing, Gli2 and Gli3 are also subject to degradation mediated by a ubiquitin ligase containing Cul3 and Speckle-type POZ protein (Spop)

(Chen et al., 2009; Wang et al., 2010; Wen et al., 2010; Zhang et al., 2009; Zhang et al., 2006). It was reported that Spop specifically targeted activated Gli3 for degradation (Wen et al., 2010). This appears to be consistent with reports that suggested antagonistic roles of Sufu and Spop in Gli2 and Gli3 degradation (Chen et al., 2009; Wang et al., 2010). These in vitro studies raised the possibility that Spop may be responsible for the great reduction in the levels of Gli2 and Gli3 in

Sufu mutants. However, the in vivo roles of Spop in Shh signaling and spinal cord development, especially its role in Gli degradation in the absence of Sufu, have not been revealed.

Here we show that loss of Spop does not change the patterning of ventral spinal cord.

This lack of spinal cord patterning defects does not result from redundancy with another

Speckle-type POZ protein, Spop-like (Spopl), as the Spopl;Spop double mutant spinal cord is also properly patterned along its D/V axis. Interestingly, the level of Gli3 protein increases moderately in Spop mutants, and removing Spop rescues the loss of floor plate and V3 interneurons in Gli2 mutants, suggesting that Spop negatively regulates Gli3A. On the other hand, no change in the level of Gli2 protein was observed in the absence of Spop, and the D/V

105

patterning of Spop;Gli3 double mutant spinal cord resembles that of Gli3 mutants, suggesting that Gli2 is not the major target of Spop in the spinal cord. Furthermore, loss of Spop fails to restore the levels of Gli2 and Gli3 in Sufu mutants, but does exacerbate the ectopic activation of

Shh signaling and ventralization of the spinal cord, and restores the formation of the floor plate and V3 interneurons in the Gli1;Sufu double mutant spinal cord. In summary, our data suggest that Spop negatively regulates Gli3A activity and Shh signaling, but does not underlie the drastic decrease in the levels of Gli2 and Gli3 in Sufu mutants.

4.1 Spop inhibits Shh signaling by downregulating Gli3 activator activity

4.1.1 Spop is expressed at a low level in the neural tube

In a previous work it was reported that Spop RNA was present ubiquitously in E9.5 mouse embryos (Chen et al., 2009). We sought to confirm this expression pattern by examining the expression of the lacZ reporter in SpoplacZ heterozygotes. Consistent with the previous report, we observed low levels of lacZ expression throughout E9.5 and E10.5 embryo, including the neural tube (Figure 4-1). Interestingly, we found enriched lacZ expression in a thin layer of cells surrounding the neural tube, which are possibly perineural vascular plexus. The presence of Spop in the neural epithelium indicate that Spop may play a role in the specification of neural progenitor fates, which is tightly regulated by Shh signaling and Gli effectors 290.

106

Figure 4-1. Spop-lacZ shows a low-level expression in the spinal neural tube at E9.5.

(A) X gal-stained E10.5 whole embryos.

(B) X gal-stained E9.5 sections of spinal neural tube. 1: spinal cord; 2: a layer of cells with robust Spop- lacZ staining, possibly perineural vascular plexus; 3: forelimb bud; 4: heart.

4.1.2 Gli3 is stabilized in Spop mutant embryos

Previous in vitro studies suggested that Spop targeted overexpressed Gli2 and Gli3 for degradation 173,238,288,289. To investigate whether loss of Spop stabilizes Gli2 and Gli3 in spinal neural tube, we examined the levels of these proteins in E10.5 trunk through immunoblotting analyses (Figure 4-2A). Consistent with a role of Spop in Gli3 degradation, the levels of both full- length (Gli3FL) and repressor (Gli3R) forms of Gli3 were significantly increased in Spop mutant embryos (Figure 4-2C and D). In contrast, we did not observe any increase in the level of Gli2 in

Spop mutant embryos (Figure 4-2B and D).

107

Multiple kinases phosphorylate Gli3FL and prime it for Btrc-induced ubiquitination and subsequent ubiquitination (Figure 1-3A). On the other hand, phosphorylation on other groups of serine/threonine both promotes the transcriptional activity of Gli3A and stabilizes Gli3A (Figure

1-4) 55,162. To better separate phosphorylated form of Gli3FL (pGli3FL) from non-phosphorylated form (uGli3FL), we ran the samples in a 5% gel for long time. We confirmed the top band was phosphorylated form through a λ-protein phosphatase treatment, which greatly reduced the intensity of top band without affecting the bottom band (non-phosphorylated). Interestingly, pGli3FL was increased by 3 fold in Spop mutant spinal cord, but uGli3FL was only increased by

50% (Figure 4-2F). Therefore, loss of Spop had a more profound stabilization effect on pGli3FL, which possibly represented the majority of Gli3A.

108

Figure 4-2. Loss of Spop leads to stabilization of Gli3 but not Gli2.

(A) A schematic illustration of tissue (highlighted region) used for immunoblot in (B) and (C). The trunks of E10.5 embryos were freed of viscera and limbs to minimize the effect of Gli proteins in these tissues.

(B) Immunoblots with antibodies against Gli2 and β-tubulin.

(D) Quantification of (B) and (C) (mean ± SEM from n=6 embryos per group). t-test showed a significant increase in Gli3FL and Gli3R levels but not Gli2 level in Spop mutants.

(E) An immunoblot with antibody against Gli3. Samples were either not treated or treated with λ-protein phosphatase for indicated amount of time. A 5% gel was run for a long time to achieve a better separation of phosphorylated (pGli3FL) and unphosphorylated Gli3FL (uGli3FL).

(F) Quantification of pGli3FL and uGli3FL abundance (mean ± SEM from n=6 embryos per group).

109

4.1.3 The dorsal/ventral patterning of the Spop mutant neural tube was normal

A small number of Spop mutant embryos exhibited exencephaly (n= 11/157 homozygotes and 5/407 heterozygotes, combining SpopΔEx and SpoplacZKI alleles) and spina bifida (n= 4/155 homozygotes, combining both alleles), suggesting that Spop may play a role in the development of the nervous system. The majority of Spop mutants did not exhibit morphological defect (Figure

4-3A, A’, G, G’).

To investigate whether the increase in the level of Gli3 disrupted the dorsal/ventral patterning of the neural tube, we examined the expression of transcription factors specific for various progenitor cells in the ventral neural tube. At E9.5, Foxa2 and Nkx2.2 labeled the ventral-most floor plate and adjacent progenitors for V3 interneurons (p3) (Figure 4-3B, C). At

E10.5, Foxa2 expression was limited to floor plate (Figure 4-3H), and Nkx2.2 labeled V3 interneuron progenitors (Figure 4-3I). Olig2 labeled motor neuron progenitors (pMN) immediately dorsal to p3 (Figure 4-3D, J). Nearly all ventral progenitor cells (p1, 2, 3 and pMN) expressed Nkx6.1 (Figure 4-3E, K). Pax6 expression was found in dorsal and lateral regions of the neural tube (Figure 4-3F, L). These genes were all expressed in their normal patterns in the

Spop mutant spinal cords (Figure 4-3B’-F’, H’-L’), suggesting that the increase in the levels of

Gli3 protein was not sufficient to alter the early development of the neural tube. Consistent with this observation, Ptch1 and Gli1 expression levels were comparable between wild type and Spop mutant embryos, suggesting that Shh signaling activity was not affected by loss of Spop (Figure

4-4A).

110

Figure 4-3. Removal of Spop does not alter the neural patterning.

(A, A’) Lateral views of E9.5 control (A) and Spop (A’) mutant embryos.

111

(B-F’) Immunofluorescent images of transverse sections of E9.5 control (n=4 embryos) and Spop mutant

(n=4) spinal cords at the forelimb level with indicated antibodies.

(G, G’) Lateral views of E10.5 control (G) and Spop (G’) mutant embryos.

(H-L’) Immunofluorescent images of transverse sections of E10.5 control (n=4 embryos) and Spop mutant (n=4) spinal cords at the forelimb level. The spinal cords are outlined with dash lines. Brackets indicate the expression domains. The span of each domain along the D/V axis is quantified and shown on the right. t-test suggests no significant difference between Spop mutant and control groups.

Figure 4-4. Ptch1 and Gli1 expression in mutant embryos. qRT-PCR was performed to measure the levels of Ptch1 and Gli1 expression in E9.5 embryos with

Gapdh as the internal control.

(A) Ptch1 and Gli1 were expressed at a similar level in Spop homozygous mutants (n=4) as in heterozygous (n=4) and wild type (n=3) littermates.

(B) Ptch1 and Gli1 were expressed at a similar level in Spopl homozygous mutants (n=5) as in heterozygous (n=4) and wild type (n=3) littermates.

(C) Ptch1 expression was moderately reduced in Gli2 mutants (n=3) and rescued in Spop;Gli2 double mutants (n=3), although the difference did not reach statistical significance. Gli1 was expressed at a significantly lower level in Gli2 mutants, and was significantly increased in Spop;Gli2 double mutants. t- test was performed to compare the groups. *: p<0.05.

112

4.1.4 Loss of Spopl does not alter spinal cord patterning in wild type or Spop mutants

Spopl shares similar substrate specificity with Spop, albeit exhibiting weaker ubiquitin- ligase activity (Choo et al., 2010; Errington et al., 2012). To investigate whether Spopl functionally compensates for the loss of Spop in the D/V patterning of the spinal cord, we generated Spopl mutants and Spopl;Spop double mutants. Spopl homozygous mutants were healthy and fertile (>55 mutants weaned), and the E10.5 mutant spinal cords were patterned properly along the D/V axis, suggesting that Spopl did not play an essential role in mouse development and spinal cord D/V patterning (Figure 4-5A-A”). Importantly, Spopl;Spop double mutant mice exhibit postnatal lethality, similar to Spop single mutants. We also found that the spinal cords were properly patterned in E10.5 Spopl;Spop double mutants (Figure 4-5B-F”). Furthermore, loss of Spopl did not affect the expression levels of Ptch1 and Gli1 (Figure 4-4B). These data suggest that Spop and Spopl are dispensable for normal spinal cord patterning.

113

114

Figure 4-5. Spopl mutant and Spopl;Spop double mutant exhibit normal spinal cord patterning.

(A-A”) Lateral views of E10.5 control (A), Spop mutant (A’) and Spopl;Spop double mutant (A”) embryos.

(B-F”) Immunofluorescent images of transverse sections of E10.5 control (n=4 embryos), Spopl mutant

(n=4) and Spopl;Spop double mutant (n=4) spinal cords at the forelimb level with indicated antibodies are shown. The spinal cords are outlined with dash lines. Brackets indicate the expression domains. The morphology and spinal cord neural patterning of both Spopl mutant and Spopl;Spop double mutant resemble wild type.

4.1.5 Loss of Spop restores normal patterning in Gli2 mutant spinal cord

Neither Gli1 nor Gli3 was essential for the mammalian ventral spinal cord development

(Bai et al., 2002; Park et al., 2000; Persson et al., 2002); however, the importance of these Gli proteins in supporting maximal Shh signaling and ventral spinal cord development was revealed with simultaneous reduction or ablation of Gli2, suggesting that these three Gli proteins share redundant functions in spinal cord patterning (Bai et al., 2004; Lei et al., 2004; Park et al., 2000).

This functional redundancy among the three mouse Gli proteins suggests that a moderate change in the activities of any member of this family may have a more obvious impact in the spinal cord pattering in the absence of another Gli protein. We hence removed Gli2 in Spop mutants to see whether the moderate increase in the level of Gli3 rescues the spinal cord patterning defects resulted from the loss of Gli2. The morphology of Spop;Gli2 double mutant embryos resembled that of Gli2 mutants at E10.5 (Figure 4-6A-A”). As reported previously (Ding et al., 1998; Matise et al., 1998), the Foxa2-expressing FP was missing and Nkx2.2-expressing p3 cells were diminished in the Gli2 mutant spinal cord (Figure 4-6B-C’), whereas Olig2 expression was

115

expanded to the ventral midline (Figure 4-6D and D’). Interestingly, Foxa2 was expressed, and

Nkx2.2 expression was increased, in the Spop;Gli2 double mutant spinal cord (Figure 4-6B” and

C”). Meanwhile, Olig2 expression was excluded from the ventral-most part of the Spop;Gli2 double mutant spinal cord (Figure 4-6D”). The Nkx6.1 expression domain, which includes cell types that do not require high levels of Shh pathway activation, and Pax6 expression domain, which includes cell types that are inhibited by high levels of Shh activation, were similar in wild type, Gli2 mutants and Spop;Gli2 double mutants (Figure 4-6E-F”). We observed similar restoration of the spinal cord patterning in E9.5 Spop;Gli2 double mutants (Figure 4-7). The restored formation of FP and p3 in the Spop;Gli2 double mutant spinal cords suggested an increase in Shh pathway activation. Supporting this hypothesis, the expression levels of Ptch1 and Gli1 were decreased in Gli2 mutant embryos, and restored in Spop;Gli2 double mutants (Figure 4-4C), suggesting that loss of Spop enhanced Shh signaling activity, which underlies the rescue of spinal cord patterning defect caused by loss of Gli2. Theoretically, changes in the activities of both Gli1 and Gli3 could contribute to this increase; however, previous in vitro studies suggested that Spop did not target Gli1 for degradation (Chen et al., 2009; Zhang et al., 2009). Therefore, the increase in the activation of the Shh pathway in Spop;Gli2 double mutants likely results from an increase in the Gli3A activity, which is better revealed in the absence of Gli2.

116

Figure 4-6. Floor plate and pV3 are rescued in Spop;Gli2 double mutants.

117

(A–A”) Lateral views of E10.5 control (A), Gli2 mutant (A’) and Spop;Gli2 double mutant (A”) embryos.

(B–F”) Immunofluorescent images of transverse sections of E10.5 control (n=3 embryos), Gli2 mutant

(n=3) and Spop;Gli2 (n=3) double mutant spinal cords at the forelimb level with indicated antibodies. The spinal cords are outlined with dash lines. Brackets indicate the expression domains. The span of each domain along D/V axis is shown on the right. The width of Nkx2.2 domain is also quantified (hollow columns).

(B-B”) Foxa2 was present in the ventral-most region of the control and diminished in the Gli2 mutant, but restored in the Spop;Gli2 double mutant spinal cords. t-test was employed to compare the width of floor plate. *: p<0.05.

(C-C”) Nkx2.2 was present in juxtaposition to floor plate of the control but diminished in the Gli2 mutant, and was restored in the Spop;Gli2 double mutant spinal cords. t-test was employed to compare the size of

V3 interneuron progenitor domain. *:p<0.05.

(D-D”) Olig2 expression was excluded from the ventral-most region of control and Spop;Gli2 double mutant but expanded ventrally in Gli2 mutant spinal cords.

(E-F”) Nkx6.1 and Pax6 expression domains remained unchanged in the Gli2 mutant and Spop;Gli2 double mutant spinal cords.

118

Figure 4-7. Floor plate and pV3 are rescued in Spop;Gli2 double mutants.

119

(A–C) Lateral views of E9.5 control (A), Gli2 mutant (B) and Spop;Gli2 double mutant (C) embryos.

(D–O) Immunofluorescent images of transverse sections of E9.5 control (n=2 embryos), Gli2 mutant

(n=2) and Spop;Gli2 (n=2) double mutant spinal cords at the forelimb level with indicated antibodies. The spinal cords are outlined with dash lines. Brackets indicate the expression domains.

(D-I) Foxa2 and Nkx2.2 were present in the ventral-most region of the control and diminished in the Gli2 mutant but restored in the Spop;Gli2 double mutant spinal cords.

(J-L) Olig2 expression was excluded from the ventral-most region of control and Spop;Gli2 double mutant but expanded ventrally in Gli2 mutant spinal cords.

(M-O) Nkx6.1 expression domain remained unchanged in Gli2 mutant and Spop;Gli2 double mutant spinal cords.

(P-R) Pax6 expression domain remained unchanged in Gli2 mutant but appears to be ventrally expanded in Spop;Gli2 double mutant spinal cords.

4.1.6 The ventral neural tube was properly patterned in Spop;Gli3 double mutants

In contrast to the significant increase in the level of Gli3, the level of Gli2 was not changed in Spop mutants. However, we could not rule out the possibility that loss of Spop affected Gli2 activity through an alternative mechanism, especially as a recent study in

Drosophila suggested that HIB inhibited ci activity by attenuating its nuclear translocation when

Hh signaling was highly activated (Seong et al., 2010). As the presence of Gli3R in the lateral and dorsal neural tube may conceal the effect of any moderate increase in Gli2 activator activity, we removed Gli3 to sensitive the neural tube for the study of Gli2 activity. As previously reported, the loss of Gli3 led to frequent exencephaly (Figure 4-8A, A’; Figure 4-9A, A’), but had no impact on the expression of Foxa2, Nkx2.2, Olig2, Nkx6.1 and Pax6 (Figure 4-8B-F’;

Figure 4-9B-F’). Spop;Gli3 double mutant embryos resemble Gli3 mutants morphologically

120

(Figure 4-8A”, Figure 4-9”), and did not exhibit any ventral patterning defects in the neural tube

(Figure 4-8B”-F”; Figure 4-9B”-F”). This result suggests that the loss of Spop did not result in a noticeable change in Gli2 activity, consistent with the unchanged Gli2 level in Spop mutants.

121

122

Figure 4-8. Spop;Gli3 double mutant resembles Gli3 mutant morphologically and patterning-wise.

(A–A”) Lateral views of E10.5 control (A), Gli3 mutant (A’) and Spop;Gli3 double mutant (A”) embryos.

Arrows in A’ and A” indicate exencephaly.

(B–F”) Immunofluorescent images of transverse sections of E10.5 spinal cords in Gli3 mutant (n=4 embryos) and Spop;Gli3 double mutant (n=5) resembled those in the control (n=5) at the forelimb level with indicated antibodies. The spinal cords are outlined with dash lines. Brackets indicate the expression domains. The span of each domain is quantified and shown on the right.

123

Figure 4-9. Loss of Spop does not alter the neural patterning in Gli3 mutant embryos.

124

(A–C) Lateral views of E9.5 wild type (A), Gli3 mutant (B) and Spop;Gli3 double mutant (C) embryos.

Arrows in B and C indicate exencephaly.

(D–R) Immunohistochemical analysis of transverse sections of E9.5 spinal cords at the forelimb level.

The spinal cords are outlined with dash lines. Brackets indicate the expression domains. The expression domains of the five marker genes in Gli3 (n=2) mutant and Spop;Gli3 (n=2) double mutant resembles control (n=2) neural tube.

4.2 Sufu plays a positive role in spinal cord neural patterning by antagonizing Spop

Suppressor of Fused (Sufu) inhibits Shh signaling and ventral spinal cord development in mammals (Cooper et al., 2005; Svard et al., 2006). In vitro studies suggested that Sufu represses

Gli proteins both by sequestering them in the cytoplasm and by inhibiting Gli-mediated transcription in the nucleus (Barnfield et al., 2005; Cheng and Bishop, 2002; Ding et al., 1999;

Dunaeva et al., 2003; Kogerman et al., 1999; Lin et al., 2014; Merchant et al., 2004; Murone et al.,

2000; Paces-Fessy et al., 2004; Stone et al., 1999). Recent findings suggest that Hh signaling results in the primary cilium-dependent dissociation between Sufu and Gli proteins (Humke et al.,

2010; Lin et al., 2014; Tukachinsky et al., 2010). Interestingly, our group previously found that the floor plate and V3 interneurons failed to form in Gli1;Sufu double mutants, in striking contrast to normal spinal cord patterning in Gli1 mutants, suggesting that Sufu is required for the maximal activation of Shh signaling in the absence of Gli1 290. This positive function of Sufu correlates with the observation that the levels of Gli2 and Gli3 proteins are greatly reduced in the absence of

Sufu, suggesting that Sufu may promote Shh signaling by protecting the Gli proteins from degradation 172,173,289. Therefore, we hypothesize that Sufu promotes Hh signaling in the ventral- most neural tube by inhibiting Spop.

125

4.2.1 Loss of Spop exacerbates neural tube patterning defects in Sufu mutants

To test the hypothesis that Spop underlies the degradation of Gli2 and Gli3 in the absence of Sufu, we performed immunoblot analyses on E9.5 Spop;Sufu double mutant embryos. We found a moderate increase in the level of Gli3FL in Spop;Sufu double mutants than in Sufu single mutants, although the difference failed to reach statistical significance (Figure 4-10). In contrast, the levels of Gli2 and Gli3R were comparable between Spop;Sufu double mutants and Sufu single mutants.

Nevertheless, the levels of both Gli2 and Gli3 in Spop;Sufu double mutants were still significantly lower than those in wild type littermates, suggesting that Spop was not solely accountable for the drastic reduction of Gli2 and Gli3 in Sufu mutants (Figure 4-10).

Figure 4-10. Moderate increase in the level of Gli3, but not Gli2, in Spop;Sufu double mutant embryos.

(A) Immunoblots of E9.5 embryo lysates with indicated antibodies.

(B) Quantitative analyses of the levels of Gli2, Gli3FL and Gli3R (normalized to β-tubulin, mean ± SEM) based on the data of immunoblot analyses. Statistical significance was determined with student t test.

126

We next sought to investigate whether the moderate increase in the level of Gli3FL had an impact on the morphology and spinal cord patterning of Spop;Sufu double mutants. Consistent with previous report 290, Sufu mutant embryos exhibited twisted body axis and severe neural tube closure defects (compare Figure 4-11A to A’). The Sufu mutant spinal cord was severely ventralized, with strong Foxa2, Nkx2.2 and Nkx6.1 expression throughout the D/V axis (Figure

4-11B’-D’). Olig2 expression was strong in the dorsal spinal cord, but appeared weaker and scattered ventrally, consistent with its being activated by low levels of Shh signaling but repressed by high levels of Shh signaling (Figure 4-11E’). Consistent with the hyperactivation of Shh signaling, Pax6 expression was restricted to just a few cells in the dorsal spinal cord of Sufu mutants (Figure 4-11F’). Spop;Sufu double mutant embryos failed to turn, and also exhibited severe neural tube closure defects (Figure 4-11A”). Similar to the Sufu mutant, Spop;Sufu double mutant spinal cord exhibited widespread expression of Foxa2, Nkx2.2 and Nkx6.1 (Figure 4-11B”-

D”). However, Olig2 expression was only found in a few dorsal cells (Figure 4-11E”), and Pax6 expression was completely absent, in the Spop;Sufu double mutant spinal cord (Figure 4-11F”).

These results suggest that loss of Spop exacerbates the hyperactivation of the Shh pathway and ventralization of the spinal cord in Sufu mutants.

127

Figure 4-11. Loss of Spop promotes the ventralization of Sufu mutant spinal cord.

(A–C) Lateral views of E9.5 wild type (A), Sufu mutant (A’) and Spop;Sufu double mutant (A”) embryos.

Arrows in B and C indicate exencephaly; arrowheads indicate spina bifida.

(B–F”) Immunofluorescent images of transverse sections of E9.5 spinal cords at the thoracic level. The spinal cords are outlined with dash lines. Brackets indicate the expression domains. n=2 embryos for each genotype were analyzed.

(B-C”) Foxa2 and Nkx2.2 were expressed in the ventral-most region of wild type but expanded to the dorsal region in the Sufu mutant and Spop;Sufu double mutant spinal cords.

128

(D-D”) Nkx6.1 expression domain was expanded to the dorsal region in both Sufu mutant and Spop;Sufu double mutant spinal cords.

(E-E”) Olig2 expression was shifted to the dorsal region in Sufu mutant and even more dorsally shifted in

Spop;Sufu double mutant spinal cords.

(F-F”) Pax6 expression domain was shifted dorsally in Sufu mutant and was absent in Spop;Sufu double mutant spinal cord.

Although it is formally possible that Spop may regulate an unknown process that indirectly affects the D/V patterning of the spinal cord, the similarity between the morphological and patterning defects between Spop;Sufu double mutants and those of previously reported Ptch1 mutants 108 and Gli3;Sufu double mutants 290 suggests that the extreme ventralization of the spinal cord was likely the direct result of the hyperactivation of the Shh pathway due to loss of both Spop and Sufu.

4.2.2 Loss of Spop rescues the loss of ventral-most cell fates in the Gli1;Sufu double mutants

We previously found that the FP and p3, the ventral-most cell types in the spinal cord, were lost in Gli1;Sufu double mutants, but not in Gli1 single mutants, indicating a positive role of Sufu in Hh pathway activation 290. We hypothesized that Spop-mediated degradation of activated Gli2 and/or Gli3 in the absence of Sufu might underlie the loss of ventral cell types in Gli1;Sufu double mutants. Although the above immunoblot analyses (Figure 4-10) suggested that loss of Spop could not fully restore the levels of Gli2 and Gli3, we were encouraged by the exacerbation of the Sufu mutant spinal cord defects resulting from the additional loss of Spop (Figure 4-11), and tested the impact of loss of Spop on the spinal cord patterning defects of Gli1;Sufu double mutants. E9.5

Gli1;Sufu double mutant embryos exhibited less severe neural tube closure defects than Sufu single mutants, especially in the anterior part of the embryos (compare Figure 4-12A’ to Figure 4-11A’).

129

Consistent with our previous report 290, fewer Foxa2- and Nkx2.2-expressing cells were present in the E9.5 Gli1;Sufu double mutant spinal cord compared to Gli1 single mutants (Figure 4-12B-C’).

Meanwhile, the Olig2 and Pax6 expression domains were expanded toward the ventral midline

(Figure 4-12D, D’, F and F’), suggesting that Sufu was required for the full activation of Hh pathway in the absence of Gli1. The Olig2 and Nkx6.1 domains were also expanded more dorsally in Gli1;Sufu double mutants than in Gli1 single mutants, suggesting that lower levels of ectopic activation of Shh signaling persisted in the double mutants (Figure 4-12D-E’).

In contrast, Spop;Gli1;Sufu triple mutants exhibited widespread edema in addition to more severe neural tube closure defects than those of Gli1;Sufu double mutants (Figure 4-12A”).

Interestingly, the expression domains of Foxa2, Nkx2.2 and Nkx6.1 were all dorsally expanded in

Spop;Gli1;Sufu triple mutants, indicating higher levels of Shh pathway activation in the absence of Spop (Figure 4-12B”, C” and E”). Moreover, the expression of Olig2 and Pax6, which was inhibited by high levels of Shh signaling, was restricted to the dorsal regions of the spinal cord in

Spop;Gli1;Sufu triple mutants (Figure 4-12D” and F”). These results suggest that Spop plays a negative role in Shh pathway activation and ventral spinal cord development, and accounts for the lack of maximal activation of Shh pathway in Gli1;Sufu double mutant embryos.

To assess the Shh pathway activity more directly, we examined the expression of the lacZ reporter inserted into the Gli1 locus (Gli1-lacZ). As reported previously, Gli1-lacZ expression formed a ventral-to-dorsal gradient in the Gli1 mutant spinal cord, consistent with the existence of a ventral-to-dorsal Shh signaling gradient (Figure 4-12G)197. A notable exception was the floor plate, in which prolonged exposure to extremely high Shh activity resulted in a downregulation of Gli1-lacZ expression, as previously reported (Figure 4-12G) 326. Consistent with the widespread low and intermediate levels of Shh pathway activation but lack of maximal

130

Shh pathway activation, strong Gli1-lacZ expression is present throughout the entire Gli1;Sufu double mutant spinal cord (Figure 4-12G’). Interestingly, Gli1-lacZ is downregulated in the ventral region of the Spop;Gli1;Sufu triple mutant spinal cords, suggesting the restoration of maximal activation of Shh signaling (Figure 4-12G”). Combined with the above spinal cord D/V patterning analyses, we conclude that loss of Spop allows maximal activation of Shh pathway in the absence of Gli1 and Sufu.

131

132

Figure 4-12. Loss of Spop rescues the floor plate and V3 interneuron progenitor fates in

Gli1;Sufu double mutant embryos.

(A–A”) Lateral views of E9.5 Gli1 mutant (A), Gli1;Sufu double mutant (A’) and

Spop;Gli1;Sufu triple mutant (A”) embryos. Arrows in A’ and A” indicate exencephaly; white arrowheads indicate spina bifida; red arrowheads indicate edema.

(B–G”) Immunofluorescent (B-F”) or Xgal-stained (G-G”) images of transverse sections of the

E9.5 Gli1 mutant (n=3 embryos), Gli1;Sufu double mutant (n=3) and Spop;Gli1;Sufu triple mutant (n=3) spinal cords at the thoracic level. The spinal cords are outlined with dash lines.

Brackets indicate the expression domains. The span of each domain along the D/V axis is shown on the right. t-tests were performed to compare the sizes of expression domains of various genes.

*: p<0.05.

(B-C”) The expression of Foxa2 and Nkx2.2 in the ventral-most region of Gli1 mutant was diminished in Gli1;Sufu double mutant, but expanded dorsally in the Spop;Gli1;Sufu triple mutant spinal cords.

(D-D”) Olig2 expression was expanded both ventrally and dorsally in Gli1;Sufu double mutant, but was dorsally restricted in the Spop;Gli1;Sufu triple mutant spinal cords.

(E-E”) Nkx6.1 expression domain was expanded dorsally in the Gli1;Sufu double mutant and more dorsally in the Spop;Gli1;Sufu triple mutant spinal cords.

(F-F”) Pax6 expression domain was expanded ventrally in the Gli1;Sufu double mutant, and restricted to the more dorsal region in the Spop;Gli1;Sufu triple mutant spinal cord.

(G) Gli1-lacZ expression formed a ventral-to-dorsal gradient in the Gli1 mutant spinal cord, with the exception of the floor plate, in which Gli1-lacZ has been downregulated. (G’) Strong Gli1- lacZ expression is present throughout the entire spinal cord of Gli1;Sufu double mutant. (G”)

133

Gli1-lacZ is highly expressed in the dorsal, but not ventral region of the Spop;Gli1;Sufu triple mutant spinal cords.

Our observation that loss of Spop restored maximal activation of the Hh pathway and ventral-most cell types in the Gli1;Sufu double mutant spinal cord without a significant increase in the levels of Gli2 and Gli3 suggested that Spop may inhibit Gli activator activity through an additional, protein degradation-independent mechanism. Interestingly, a recent work showed that hib/rdx, the Drosophila homolog of Spop, appeared to inhibit the maximal activation of Hh signaling at the border between the anterior and posterior compartments of the wing disc by sequestering cubitus interruptus (ci), the sole Gli protein in Drosophila, in the cytoplasm (Seong et al., 2010). To investigate whether mouse Spop can also sequester Gli2 and/or Gli3 in the cytoplasm, we overexpress Gli2 and Gli3 together with Spop in Sufu mutant mouse embryonic fibroblasts (MEFs). We found that both Gli2 and Gli3 were predominantly in the nucleus when they were expressed alone (Figure 4-13A, D and G). Co-expression with Sufu shifted both Gli proteins to the cytoplasm (Figure 4-13B, E and G). Interestingly, Spop co-expression greatly decreased the nuclear localization of Gli3, but not Gli2 (Figure 4-13C, F and G), suggesting that cytoplasmic sequestration of Gli3 may be a potential mechanism for the negative function of Spop in Shh signaling and ventral spinal cord development. Our result is consistent with a previous overexpression experiment in wild type MEFs 173.

134

Figure 4-13. Spop overexpression alters the subcellular localization of Gli3.

(A-C) Sufu mutant MEFs co-transfected with GFP-Gli2 and empty vector (A), FLAG-Sufu (B) or FLAG-

Spop (C) are immunostained with GFP antibody.

(D-F) Sufu mutant MEFs co-transfected with GFP-Gli3 and empty vector (D), FLAG-Sufu (E) or FLAG-

Spop (F) are immunostained with GFP antibody.

(G) Cells are categorized by the localization of Gli2 or Gli3 into primarily nuclear (N>C), whole cell

(N=C) and primarily cytoplasmic (N

4.3 Conclusion

In our present study, we demonstrate that Spop downregulates the protein level of Gli3, but not Gli2, during spinal cord patterning (Figure 4-14A). Since Gli2A acts as the primary effector in specifying ventral-most cell fates, the increase of Gli3A in Spop mutants does not affect specification of floor plate to pV2 fates (Figure 4-14B). However, in Gli2 mutant spinal cord where ventral-most cell fates are diminished, the increase in Gli3A level by loss of Spop restores maximal Hh signaling activity and ventral-most cell fates (Figure 4-14B). Furthermore, we show a role of Spop in dampening Shh pathway over-activation in the absence of Sufu, especially in

135

preventing the maximal activation of Shh signaling and the ventral-most cell fates in the Gli1;Sufu double mutant spinal cord. The inhibition of Gli3 activity by Spop may be achieved by both cytoplasmic sequestration and destruction (Figure 4-14B). Our results provide insight into the roles of Spop in the regulation of Gli3 activity and spinal cord development in vivo, and reveal more complexity in the mechanisms of Spop function unappreciated in previous in vitro studies.

Our data also raise new questions such as what, if not Spop, is responsible for the drastic reduction of Gli2 and Gli3 in Sufu mutants, which will be discussed in Chapter 5.1.3.

136

Figure 4-14. A model of the roles of Spop in spinal cord patterning.

(A) In the lateral region of the spinal cord (long range) including p0-pMN, low to intermediate levels of

Shh reduce the production of Gli3R from Gli3FL by antagonizing the repressive function of Sufu, allowing

137

the expression of genes such as Olig2 and Nkx6.1. In pFP and p3, high levels of Shh promote the production of Gli2A and Gli3A, which then activate the expression of Foxa2 and Nkx2.2. Gli2A plays a more predominant role than Gli3A in this context. Spop targets Gli3 for degradation, preventing over activation of the Shh pathway.

(B) The levels of both Gli3A and Gli3R increase in the absence of Spop, but the effect of the increased

Gli3A on the formation of pFP and p3 is only revealed in Spop;Gli2 double mutants, in which the much more potent Gli2A is absent. In Gli1;Sufu double mutants, the reduced levels of Gli2A and Gli3A are insufficient to support pFP and p3 formation. Loss of Spop increases Gli3A and rescues the formation of pFP and p3 in Sufu;Gli1;Spop triple mutants. The moderate increase in Gli3R in the absence of Spop does not show apparent effect in ventral spinal cord patterning in various single and compound mutants.

Our result does not rule out the possibility that, Spop regulates spinal cord patterning mainly by targeting an unknown substrate other than Gli2 or Gli3 for degradation. It is possible that stabilization of this unknown substrate in Spop mutant spinal cord accounts for the increase of Shh signaling activity and rescue of ventral-most cell fates in Spop;Gli2 double mutant (Figure

4-4C) and Spop;Gli1;Sufu triple mutant embryos (Figure 4-12G”). This alternative model also provides a reasonable explanation to that the Gli3FL level only increased moderately in Spop;Sufu double mutant embryos as compared to Sufu mutants (Figure 4-10), but the ventral-most cell fates were greatly expanded in Spop;Gli1;Sufu triple mutants as compared to Gli1;Sufu mutants (Figure

4-12). However, based on this alternative hypothesis, the stabilization of this substrate in the context of Gli3 mutant spinal cord, which is sensitized to subtle elevation of Hh signaling activity, would likely cause an expansion of ventral-most cell fates, whereas our data showed that the ventral spinal cord patterning of Spop;Gli3 double mutants resemble Gli3 mutant and controls

(Figure 4-8). Therefore, this possibility appears to be minor. A direct method to rule out this possibility is to test whether Spop;Gli2;Gli3 triple mutant resembles Gli2;Gli3 double mutant in

138

spinal cord patterning. Unfortunately, we failed to generate any Spop;Gli2;Gli3 triple mutant embryo by mating triple heterozygous carriers due to low odds (1/64 if not lethal prior to E9.5).

Acknowledgement

We thank Drs Zhi-Chun Lai and Wendy Hanna-Rose for critically reading the manuscript.

We thank Drs Joyner and Toftgard for Gli1, Gli2, Gli3 and Sufu mutant mice. The monoclonal antibodies against Foxa2, Nkx2.2, Nkx6.1 and Pax6 developed by Drs. Jessell, Madsen and

Kawakami were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological

Sciences, Iowa City, IA 52242. A.L. was supported by the National Institutes of Health [grant number HD083625] and a Penn State Start-up Fund.

139

Chapter 5

Discussion

5.1 The molecular mechanisms underlying Spop regulation of Gli proteins

5.1.1 Conclusion

Figure 5-1. Spop plays negative and positive roles in Hh signaling by Gli3 degradation.

Spop targets Gli3FL or Gli3A for degradation, thereby inhibiting Shh target gene expression during specification of spinal neural progenitors. Spop also downregulates Gli3R level at least in part by directly targeting Gli3R for ubiquitination, thereby promoting Ihh target gene expression which is critical for chondrocyte and osteoblast differentiation. It remains unclear whether Gli3A or Gli3FL is a preferred

Spop substrate, and whether Spop promotes partial degradation of Gli3FL.

In the present study, we characterized two null alleles and a limb-specific conditional mutant allele of Spop, and discovered critical roles for Spop in skeletal and neural development.

Interestingly, by ubiquitinating and downregulating both Gli3FL and Gli3R, Spop plays a positive

140

role in Ihh signaling and skeletal development, but a negative role in Shh signaling and spinal neural patterning.

5.1.2 Spop may exhibit a preference of substrate

Previous studies using cultured mammalian cells or mRNA injection in Xenopus eggs suggested a negative role of Spop in Hh signaling 173,238,286,288,289. It is worth noting that a key difference between these studies and our in vivo data lies in the roles of Spop in the regulation of

Gli2, the primary activator of the mammalian Hh pathway. In these gain-of-function studies, overexpressed Spop interacts with Gli2 and targets it for degradation, an observation we confirmed in our own in vitro analysis (Figure 3-11). In contrast, we found no significant change in Gli2 protein level in Spop null mutants (Figure 3-10). It is not clear why the level of Gli2 is not altered by the loss of Spop in vivo. One possible explanation is that Spop specifically targets the activated, labile form of Gli2 that only exists in a small number of cells with the highest level of Hh pathway activation as suggested by previous in vitro studies 173,238,289. Alternatively, Spop may preferentially interact with other substrates, but not Gli2, under physiological condition. Since

Spop interacts with both the amino-terminus and carboxyl-terminus of Gli3 but only with the carboxyl-terminus of Gli2 (Wang et al., 2010; Zhang et al., 2009), potential modification and protein-protein interaction that occur to the carboxyl-terminus may abolish Gli2-Spop interaction without affecting Gli3-Spop interactions. It is also possible that Gli2 is associated with proteins that protects from Spop-induced ubiquitination and/or subsequent proteasomal degradation. These possibilities await testing in future studies.

The evidence showing opposite roles of Spop in different organs suggests that the function of Spop in Hh signaling depends on the context. In Drosophila, hib/rdx expression is directly induced by ci 327, and it is restricted to cells with high Hh signaling activity where

141

production of ci repressor was blocked 53,54. Therefore, hib/rdx consistently functions as a negative regulator of Hh signaling 53,54. In fact, misexpression of hib/rdx in areas of Drosophila wing disc with moderate Hh pathway activation resulted in the decrease in both full-length and processed ci 54. In contrast, we and another group 173 both showed that the expression of mammalian Spop is not limited to cells with active Hh signaling. Rather, the robust expression of Spop in the apical epidermal ridge during limb patterning (Figure 3-4) and in a layer of cells surrounding the neural tube (likely the perineural vascular plexus) (Figure 4-1) indicates a correlation between Spop expression and Wnt signaling 328–330. The discrepancy between the expression profiles of hib/rdx and Spop indicates an evolution of the cis-regulatory module that controls Spop transcription, which eventually results in a new function of Spop.

Our findings that Spop interacts with Gli3R and regulates its ubiquitination and degradation is consistent with previous reports of physical interaction between the N-terminus of

Gli3 and Spop 53,148,289. A recent study further showed that Spop induces ubiquitination of Gli31-

90 148. However, our data appear to be at odds with a study showing no effect of Spop overexpression on the stability of Gli31-700 in C3H10T1/2 cells 289. This could result from the use of different antibodies for immunoblots. A Gli3 antibody that recognizes both the endogenous Gli3R and overexpressed Gli31-700 was employed by Wang et al. (2010). Since the lipofectamine-mediated transfection efficiency was low in C3H10T1/2 cells 289, the majority of endogenous Gli3R in untransfected cells is not influenced by Spop overexpression, a moderate decrease in Gli31-700, which co-migrates with endogenous Gli3R, may be difficult to detect. It is also possible that the expression levels of Spop in C3H10T1/2 and HEK 293T cells underlie the different effects on Gli31-700 degradation. Unfortunately, we have failed to express Spop at sufficient levels in C3H10T1/2 cells with PEI (Polysciences) or JetPrime (Polyplus transfection,

142

#114-07) reagents, preventing us from reaching a solid conclusion. Nevertheless, C3H10T1/2 and HEK 293T cells may not faithfully recapitulates the in vivo condition, and thus better approaches are needed to resolve this controversy, such as by studying the genetic interactions of

Spop and Gli3R using a Gli31-700 knockin mouse strain 331.

The molecular mechanism underlying Spop inhibition of Gli3A activity in spinal neural patterning remains poorly understood, in part due to the technical difficulty to distinguish Gli3A and inactive Gli3FL in direct measurements. Although immunoblots reveal an accumulation of

Gli3FL in Spop mutants, the relative abundance of Gli3A and inactive Gli3FL remains unclear.

Since Spop promotes Hh target gene expression in the skeleton (Figure 3-9), the excessive

Gli3FL in Spop mutants are likely inactive, suggesting that Spop selectively targets inactive

Gli3FL for ubiquitination and subsequent degradation, and the increase in Gli3A activity in Spop mutant spinal cord may result from the stabilization of Gli3FL. This notion is supported by the finding that Hh stimulates CK1-catalyzed phosphorylation of Spop-binding sites in Gli3, which protects Gli3A from Spop-triggered ubiquitination 55. Nevertheless, there is a minor possibility that the increase in Gli3FL in Spop mutant skeleton (Figure 3-10) represents a substantial increase in Gli3A level, despite that the increase in Gli3A is counteracted by the accumulation of

Gli3R. Since Sufu binds inactive Gli3FL and competitively inhibits Spop-mediated Gli degradation 173,238,289, while Gli activation accompanies dissociation of Sufu 176, Spop possibly selectively targets Gli3A for degradation. Further investigation is needed to determine the key modification or processing that triggers Gli3 activation, which may help demonstrating whether

Spop preferentially targets Gli3A or inactive Gli3FL for degradation.

143

5.1.3 Spop and Btrc potentially induces ubiquitination on different lysine residues in Gli3

Two E3 ubiquitin ligases, Cul1/Btrc and Cul3/Spop, have been shown to induce Gli2 and

Gli3 degradation 332. However, how Btrc and Spop differentially regulate Gli protein degradation remains to be elucidated. In Drosophila, the partial degradation of ci into its repressor form is mediated by slimb, the Drosophila homolog of Btrc 37, and hib/rdx induces ci turnover 53,54. In vitro studies revealed a partially conserved mechanism in mammals, as Btrc is essential for proteolytic processing of Gli3 in HEK 293 cells 163 and induces destruction of Gli1 and Gli2 158,164. Similar to hib/rdx, Spop triggers the destruction of Gli2 and Gli3 in vitro

173,238,289. Therefore, it is possible that Spop and Btrc induce ubiquitination at different sites that determine the different fates of Gli3 (Figure 5-2A). Previous studies show that Spop induces ubiquitination of Gli31-90 at K15/22/32/70/87 148, and Btrc induces ubiquitination of Gli3 at

K773/779/784/800 156. Since four of the six Spop-binding sites are clustered in the amino- terminus of Gli3 (residues 1-167) 238, and Btrc-binding motifs are mostly located in the central domain (residues 849-899) 156, it is possible that the K15/22/32/70/87 ubiquitination is Spop- specific, and K773/779/784/800 sites are only ubiquitinated by Btrc. Spop and Btrc may then trigger polyubiquitination of Gli3 through K48-linked and K11-linked mechanisms, respectively

(Figure 5-2A) 333. Future studies are necessary to determine Spop-specific and Btrc-specific ubiquitination sites in Gli3, and test the hypothesis that different ubiquitination signatures trigger partial degradation and complete destruction of Gli3.

144

Figure 5-2. Two models depicting Gli3 ubiquitination catalyzed by Spop and Btrc.

(A) Spop induces poly-ubiquitination of K15/22/32/70/87 on Gli3. The ubiquitin moieties are linked to

K48 of the previous ubiquitin. The ubiquitinated Gli3 is subsequently destructed in 26S proteasome.

Preceded by phosphorylation by multiple kinases, Gli3 is ubiquitinated by Btrc at K773/779/784/800, and each ubiquitin is linked to K11 of the previous ubiquitin, labeling Gli3FL for subsequent proteolytic processing that gives rise to Gli3R.

(B) Phosphorylation of Gli3 allows ubiquitination at K773/779/784/800 by both Spop and Btrc in K11- linked manner, labeling Gli3FL for proteolytic processing that gives rise to Gli3R. Subsequently, Spop induces poly-ubiquitination of K15/22/32/70/87 on Gli3R in K48-linked manner. The ubiquitinated

Gli3R is then destructed in 26S proteasome.

Interestingly, a previous report shows that overexpression of Spop promotes proteolytic processing of Gli3 in HEK 293T cells 289. However, the context of Spop-induced proteolytic processing remains unclear. Our data showed that loss of Spop resulted in accumulation, but not decrease, of Gli3R in both E10.5 embryo and limb bud. Since Spop also targets Gli3R for ubiquitination and degradation (Figure 3-11), even though loss of Spop may slow down Gli3R

145

production, the stabilization of Gli3R in Spop mutants may result in a Gli3R accumulation. It is possible that a special context in Wang et al. (2010), such as high intrinsic PKA, CK1 and GSK3 kinase activity in the cultured cells, stabilized Gli3R when Spop is overexpressed, which allowed them to uncover a role of Spop in facilitating Gli3 proteolytic processing. Nevertheless, this finding prompts us to hypothesize that Spop induces Gli3 turnover through a two-step route

(Figure 5-2B). When Gli3 is phosphorylated by PKA, CK1 and GSK3, both Spop and Btrc potentially trigger the ubiquitination at K773/779/784/800 in a K11-linked mechanism that results in proteolytic processing of Gli3 into the repressor form. Subsequently, Spop induces

K15/22/32/70/87 ubiquitination that labels Gli3R for destruction. More studies are required to determine the Spop and Btrc-specific ubiquitination sites and test these two models.

Previous studies show that loss of Sufu destabilizes Gli2 and Gli3, and abolishes production of the repressor forms 172,173,289. Sufu binds Gli2 and Gli3, and protect them from

Spop-induced degradation, suggesting that Sufu stabilizes Gli2 and Gli3 by antagonizing Spop

173. However, loss of Spop in Sufu mutant background did not restore Gli2 and Gli3 to a level of wild type or Sufu heterozygous littermates (Figure 4-10), suggesting that during embryonic development, Sufu stabilizes Gli2 and Gli3 not only by antagonizing Spop but also by attenuating other degradation mechanisms. Therefore, it is possible that Btrc also induces ubiquitination and subsequent destruction of Gli2 and Gli3 in the absence of Sufu.

5.2 Other potential functions of Spop

5.2.1 Spop mutants unlikely die of respiratory or renal failure

Our data showed that the number of Spop mutants decreases progressively throughout the embryonic stages (Table 3-1), and a majority of Spop mutant pups die shortly after birth (Chapter

146

3.1.2). The exposure of internal organ by the low-penetrance defects exencephaly and spina bifida apparently accounts for the loss of a small proportion of Spop mutants (Figure 5-3), but the major cause of neonatal lethality remains undetermined. To identify the potential cause of lethality in Spop mutant pups, we performed autopsy.

Figure 5-3. Spop mutants exhibit exencephaly at a low penetrance.

Three Spop mutants and a control from the same litter at E18.5. Arrows indicate exencephaly.

A common cause of the neonatal lethality is cleft palate, which result from defective fusion of palate shelves 334. In wild type embryos, the palate shelves are already fused at E14.75

335. However, mutations of several genes such as transcription factor Msh homeobox 1 (Msx1), signaling ligand TGF-β3, etc. lead to cleft palate 335–337. These mutant pups typically exhibit cyanotic skin, gasping respirations and air-filled stomach, and die within 24 hours of birth, suggesting that cleft palate contribute to neonatal lethality by affecting respiration and suckling

335–337. Of note, Gli2 and Gli3 mutants both exhibit cleft palate at an incomplete penetrance and tooth defects (Figure 5-4B) 338. Therefore, we examined whether Spop mutants die of cleft palate or tooth defects, and found that Spop did not exhibit a defect in palate fusion or size of incisor (Figure 5-4).

147

Figure 5-4. Loss of Spop does not cause cleft palate.

(A) The palate shelves of a Spop mutant and a littermate are both intact.

(B) Alcian blue and Alizarin red-stained palate of a Gli2 mutant and a Gli3 mutant from different litters.

While their control littermates have already undergone fusion of palate shelves (not shown), fusion is incomplete in Gli2 and Gli3 mutants (highlighted by arrows).

(C) Alcian blue and Alizarin red-stained palate of a Spop mutant and a littermate at E18.5. Although loss of Spop compromises ossification, it does not affect the fusion of palate shelves.

(D) Alcian blue and Alizarin red-stained palate of a Spop mutant and a littermate at P1. Spop mutant skull exhibits a smaller size due to a global growth arrest in Spop mutant, but the ossification of palate shelf resembles control littermates.

Respiratory failure such as by defects in lung branching morphogenesis, lung maturation, gas exchange and rib cage size also accounts for a fraction of perinatal lethality 334. To determine whether Spop mutants exhibit lung defects, we analyzed the histology of Spop mutant

148

lungs (Figure 5-5), and no major defect was identified. Furthermore, Spop mutant pups breathed normally, and autopsy showed that the lungs of dead Spop mutants were fully inflated. The size of Spop mutant rib cages also resembles littermates. Therefore, lung defects unlikely accounts for lethality of Spop mutant pups.

Figure 5-5. Loss of Spop does not result in major lung defect.

H&E staining of 8µm-thick embryonic lung sections at E18.5 show normal histology in Spop mutants.

Kidney agenesis induces lethality within 24 hours of birth 334. To determine whether

Spop mutants exhibit morphological defects in the kidney, we analyzed the histology of Spop mutant kidneys (Figure 5-6), but no major defect was identified. Since renal failure may occur without apparent morphological defect, such as in the mutant of β-epithelial Na+ channel

(βENaC) 339, we further investigate whether kidney-specific deletion of Spop results in lethality.

We took advantage of Hoxb7-Cre, which induces efficient loxP recombination in the mesonephric duct and its derivatives including the ureter, collective duct and ureteric bud 340.

Since Spopflox/flox;HoxB7-Cre conditional mutant mice survived and reproduced, renal failure appeared unlikely to be the major cause of neonatal lethality in Spop mutant pups.

149

Figure 5-6. Loss of Spop does not cause major kidney defects.

H&E staining of 8µm-thick embryonic kidney sections at E18.5 show normal histology in Spop mutants.

5.2.2 Spop mutants exhibit cardiovascular defect

Cardiovascular defects often results in lethality ranging from embryonic to postnatal stages 341,342. We investigated whether cardiovascular defects may account for the neonatal lethality in Spop mutants. Spop mutant hearts exhibited thinner ventricle wall and septum

(Figure 5-7). We did not observe ventricular septal defects in Spop mutants, though. The substrate of Spop in heart development remains unknown. One of the Spop substrates NcoA3

(Table 1-1) is potentially regulated by Spop in heart development, since loss of NcoA3 resulted

150

in thinned septum and ventricle walls 343, although it remains unknown whether stabilization of

NcoA3 would cause the same defect. More than 60% of NcoA3 loss-of-function mutants survive to weaning 344.

Figure 5-7. Loss of Spop affects heart development.

(A) Frontal and back views of a Spop mutant heart and littermate control at E18.5.

(B) H&E stained 5µm-thick sections of Spop mutant hearts and littermate control at the same level. la: left atrium; ra: right atrium; lv: left ventricle; rv: right ventricle; s: septum; w: wall of ventricles; m: mitral valve; t: tricuspid valve. The size of hearts in (B) is smaller than (A) due to the shrinkage of hearts through the process of dehydration and paraffin embedding.

Neonatal lethality, which was observed in Spop mutants, indicates a potential patent ductus arteriosus defect. The ductus arteriosus, a blood vessel that directs the pulmonary-to- aorta blood flow in fetal circulation, closes within 3 hours of birth to adapt to the respiration 345.

The closure of ductus arteriosus is regulated by PGE2 through its receptor EP4 345,346. EP4 mutant mice exhibit open ductus arteriosus, and 95% of pups die within 24 to 72 hours of birth, while 5% of homozygous mutants of mixed genetic background (129/SvEv and 129/Olac

151

substrains) survive to adulthood 345,346. Loss of prostaglandin dehydrogenase, the gene encoding an enzyme involved in PGE2 metabolism, also results in lethality within 12 to 48 hours of birth

342. Therefore, whether Spop mutants exhibit open ductus arteriosus remains an interesting question to be addressed.

Importantly, loss of Spop caused an expansion of pancreatic β cell population at E18.5 due to stabilization of Pdx1 277, which potentially disrupt glucose homeostasis in Spop mutants.

Previous studies showed that hypoglycemia may also cause neonatal lethality, such as in a eIF2α mutant 347. Therefore, it is worth testing the blood glucose level of Spop mutant pups, and investigating whether injection of glucose may prolong the life span of Spop mutants pups.

It is noteworthy that the milk spot of Spop mutant pups appear much smaller than control littermates at P1 and P2, suggesting that Spop mutants have feeding problems that likely result in their growth arrest and possibly contribute to the lethality (Figure 3-3). Whereas we cannot exclude the possibility that abnormal feeding results from neuromuscular defects, the cardiovascular defects in Spop mutants may influence the activity of pups. Whether the abnormal feeding is caused by the cardiovascular defect or other problems remains an interesting question to be addressed.

Although Gli3 acts downstream of Spop in endochondral ossification and spinal neural patterning, we failed to recover Spop∆Ex/∆Ex;Gli3+/- double mutants at weaning, suggesting that the increase in the level of Gli3 does not account for the lethality of Spop null mutants. Since

Spop induces ubiquitination of at least 20 substrate proteins (Table 1-1), the mechanism underlying the requirement of Spop for the survival of pups remains a question. Based on our hypothesis that cardiovascular defect in Spop mutants accounts for the lethality, a proteomic screen of major Spop substrates in the developing heart will help addressing this question.

152

5.2.3 Spop potentially promotes carcinogenesis in bone tumors

SPOP has been defined as a tumor suppressor in most reported cases, but it acts as an oncogene in kidney cancer (Table 1-1). Whether Spop promotes or inhibits cancer depends on its major substrate in the specific cancer. Primary bone tumors are the cancers that originate in the bone 348. In 2016, 3300 new bone cancers are reported in United States, which accounts for

0.2% of new cancer cases 349. The most prevalent subtypes are osteosarcoma that arises from osteoid, chondrosarcoma that arises from cartilage, and Ewing sarcoma that is initiated by a fusion of EWSR1 and other members of ETS family transcription factors 348,350. An increase in

IHH, PTCH1 and GLI1 expression is found in osteosarcoma patients and negatively correlates with the systemic disease-free survival 351, and Hh signaling inhibitors effectively suppress osteosarcoma growth in mouse models and in vitro studies 352. Similar tumorigenic role of Hh signaling was observed in chondrosarcoma 353. Furthermore, a Hh target gene NKX2.2 is highly expressed in Ewing sarcoma and promotes oncogenesis 354. Therefore, Hh signaling stimulates progression of primary bone tumors. Although the role of Spop in primary bone tumors has not been defined, based on its positive role in Ihh signaling (Figure 3-22), I hypothesize that Spop potentially acts as an oncogene in bone tumors by inducing Gli3 degradation.

Based on this hypothesis, two types of gene mutations are expected in bone tumors. The first type is a gain of SPOP function, likely through an amplification of copy number.

Unfortunately, only two studies of bone tumors (217 samples sequenced) are available in the cancer genomics database (http://www.cbioportal.org/), both of which conducted in Ewing sarcoma, and no Spop mutation has been identified in these two studies (Table 5-1) 350,355. The second type is a GLI3 mutation that sensitizes GLI3 to SPOP-induced ubiquitination and

153

degradation, thereby downregulating GLI3 repressors. This mechanism potentially underlies a

GLI3-E453K mutation in an Ewing tumor patient (Table 5-1). This mutated site is predicted to be a preferred ubiquitination site (scored 0.93) (http://www.ubpred.org) 356. Alternatively, the alteration of local charges switches the conformation of E453K mutant protein.

Table 5-1. Hh pathway mutations identified in Ewing tumors.

Gene Mutation Potential mechanism

IHH N/A

PTCH1 V1131M Conformation switch that compromises the inhibition of SMO by PTCH1.

SMO I408T Conformation switch that activates SMO.

GLI1 N/A

GLI2 E1577K Conformation switch that activates GLI2.

GLI3 E453K (1) Conformation switch that activates GLI3 activator or inactivates GLI3

repressor. (2) New ubiquitination site that decreases GLI3 repressor level.

GLI3 X158_splice Premature truncation at residue 158 that attenuates Gli3 function.

SUFU I200I Synonymous substitution.

SPOP N/A

Mutation sites were collected from two Ewing sarcoma studies 350,355.

I further analyzed GLI3 mutations in other cancer types that give rise to new lysine residues (Table 5-2). Two of the mutations, R175K and R1206K, do not change the charge or conformation of Gli3, and thus likely compromises Gli3 function by serving as new ubiquitination sites. Several other mutations that substitute asparagine or glutamine to lysine also generate potential ubiquitination sites though the longer and positively charged lysine side

154

chain may have an effect on the GLI3 conformation and function. I did not analyze other substitutions by lysine since those mutations may have a major effect on GLI3 function by either converting negative charge to positive or increasing steric hindrance.

Table 5-2. Missense mutations that potentially destabilize Gli3 by ubiquitination.

Cancer type (Sequencing Mutation Allelic Likelihood of new ubiquitination site

center) Frequency Score Confidence

Stomach (TCGA) R175K N/A 0.663319 High

NSCLC (TCGA 2016) R1206K 0.46 0.473003 Medium

NSCLC (TCGA 2016) N725K 0.15 0.561067 High

Cholangiocarcinoma N1203K N/A 0.403814 Medium

(NCCS)

Colorectal (DFCI 2016) N1221K N/A 0.42539 Medium

Colorectal (Genentech) N1530K N/A No

Lung adeno (MSKCC) Q472K 0.06 0.346258 Medium

Lung adeno (Broad) Q676K N/A 0.456283 Medium

NSCLC (TCGA 2016) Q676K 0.2 0.456283 Medium

Head & neck (TCGA) Q1047K 0.42 No

pRCC (TCGA) Q1315K 0.06 No

Mutation information was collected from http://www.cbioportal.org/ 357,358. N/A: data not available.

Mutations that likely disrupt conformation by altering charge or steric hindrance (E to K mutations,

A353K and M789K) were not considered in this table. Subsequently, the potential of the new lysine as new ubiquitination sites was evaluated with http://csb.cse.yzu.edu.tw/UbiSite/ 359.

155

In summary, SPOP potentially acts as an oncogene in bone tumors and in specific context that sensitizes tumor suppressors to SPOP-induced degradation. Therefore, more investigations on Spop expression and Gli3 mutations in bone tumors are necessary. Of note, RNA silencing of

SPOP effectively suppresses tumor cell survival and migration in renal cell carcinoma, suggesting SPOP as a drug target in kidney cancer 276. It is also important to find out whether targeting Spop effects in therapy of bone tumors and other cancers that are driven by destabilization of Spop substrates.

156

References

1. Ye, X. & Liu, A. Hedgehog signaling: mechanisms and evolution. Front. Biol. (Beijing). 6, 504– 521 (2011).

2. Huangfu, D. & Anderson, K. V. Signaling from Smo to Ci/Gli: conservation and divergence of Hedgehog pathways from Drosophila to vertebrates. Development 133, 3–14 (2006).

3. Nüsslein-Volhard, C. & Wieschaus, E. Mutations affecting segment number and polarity in Drosophila. Nature 287, 795–801 (1980).

4. Mohler, J. Requirements for hedgehod, a segmental polarity gene, in patterning larval and adult cuticle of Drosophila. Genetics 120, 1061–1072 (1988).

5. Lee, J. J., Von Kessler, D. P., Parks, S. & Beachy, P. a. Secretion and localized transcription suggest a role in positional signaling for products of the segmentation gene hedgehog. Cell 71, 33– 50 (1992).

6. Lee, J. J. et al. Autoproteolysis in hedgehog protein biogenesis. Science (80-. ). 266, 1528–1537 (1994).

7. Porter, J. A. et al. The product of hedgehog autoproteolytic cleavage active in local and long-range signalling. Nature 374, 363–366 (1995).

8. Chamoun, Z. et al. Skinny Hedgehog, an Acyltransferase Required for Palmitoylation and Activity of the Hedgehog Signal. Science (80-. ). 293, 2080–2084 (2001).

9. Porter, J. A., Young, K. E. & Beachy, P. A. Cholesterol Modification of Hedgehog Signaling Proteins in Animal Development. Science (80-. ). 274, 255–259 (1996).

10. Burke, R. et al. Dispatched, a novel sterol-sensing domain protein dedicated to the release of cholesterol-modified hedgehog from signaling cells. Cell 99, 803–815 (1999).

11. Nakano, Y. et al. A protein with several possible membrane-spanning domains encoded by the Drosophila segment polarity gene patched. Nature 341, 508–13 (1989).

12. Goodrich, L. V., Johnson, R. L., Milenkovic, L., McMahon, J. A. & Scott, M. P. Conservation of the hedgehog/patched signaling pathway from flies to mice: Induction of a mouse patched gene by Hedgehog. Genes Dev. 10, 301–312 (1996).

13. Ingham, P. W., Taylor, A. M. & Nakano, Y. Role of the Drosophila patched gene in positional signalling. Nature 353, 184–7 (1991).

14. Chen, Y. & Struhl, G. Dual roles for patched in sequestering and transducing Hedgehog. Cell 87, 553–563 (1996).

15. Torroja, C., Gorfinkiel, N. & Guerrero, I. Patched controls the Hedgehog gradient by endocytosis in a dynamin-dependent manner, but this internalization does not play a major role in signal transduction. Development 131, 2395–2408 (2004).

16. Denef, N., Neubüser, D., Perez, L. & Cohen, S. M. Hedgehog induces opposite changes in

157

turnover and subcellular localization of patched and smoothened. Cell 102, 521–31 (2000).

17. Forbes, A. J., Nakano, Y., Taylor, A. M. & Ingham, P. W. Genetic analysis of hedgehog signalling in the Drosophila embryo. Dev Suppl 115–124 (1993).

18. van den Heuvel, M. & Ingham, P. W. Smoothened Encodes a Receptor-Like Serpentine Protein Required for Hedgehog Signalling. Nature 382, 547–551 (1996).

19. Alcedo, J., Ayzenzon, M., Von Ohlen, T., Noll, M. & Hooper, J. E. The Drosophila smoothened gene encodes a seven-pass membrane protein, a putative receptor for the hedgehog signal. Cell 86, 221–232 (1996).

20. Chen, Y. & Struhl, G. In vivo evidence that Patched and Smoothened constitute distinct binding and transducing components of a Hedgehog receptor complex. Development 125, 4943–8 (1998).

21. Xia, R., Jia, H., Fan, J., Liu, Y. & Jia, J. USP8 promotes smoothened signaling by preventing its ubiquitination and changing its subcellular localization. PLoS Biol. 10, (2012).

22. Li, S. et al. Hedgehog-regulated ubiquitination controls smoothened trafficking and cell surface expression in Drosophila. PLoS Biol. 10, (2012).

23. Strutt, H. et al. Mutations in the sterol-sensing domain of patched suggest a role for vesicular trafficking in smoothened regulation. Curr. Biol. 11, 608–613 (2001).

24. Martín, V., Carrillo, G., Torroja, C. & Guerrero, I. The sterol-sensing domain of patched protein seems to control smoothened activity through patched vesicular trafficking. Curr. Biol. 11, 601– 607 (2001).

25. Zhu, A. J., Zheng, L., Suyama, K. & Scott, M. P. Altered localization of Drosophila smoothened protein activates hedgehog signal transduction. Genes Dev. 17, 1240–1252 (2003).

26. Lum, L. et al. Hedgehog signal transduction via Smoothened association with a cytoplasmic complex scaffolded by the atypical kinesin, Costal-2. Mol. Cell 12, 1261–1274 (2003).

27. Zhao, Y., Tong, C. & Jiang, J. Hedgehog regulates smoothened activity by inducing a conformational switch. Nature 450, 252–258 (2007).

28. Jia, J., Tong, C., Wang, B., Luo, L. & Jiang, J. Hedgehog signalling activity of Smoothened requires phosphorylation by protein kinase A and casein kinase I. Nature 432, 1045–1050 (2004).

29. Zhang, C., Williams, E. H., Guo, Y., Lum, L. & Beachy, P. a. Extensive phosphorylation of Smoothened in Hedgehog pathway activation. Proc. Natl. Acad. Sci. U. S. A. 101, 17900–17907 (2004).

30. Li, S. et al. Regulation of Smoothened Phosphorylation and High-Level Hedgehog Signaling Activity by a Plasma Membrane Associated Kinase. PLoS Biol. 14, 1–24 (2016).

31. Jia, H. et al. Casein kinase 2 promotes Hedgehog signaling by regulating both smoothened and Cubitus interruptus. J. Biol. Chem. 285, 37218–37226 (2010).

32. Jia, H., Liu, Y., Yan, W. & Jia, J. PP4 and PP2A regulate Hedgehog signaling by controlling Smo and Ci phosphorylation. Development 136, 307–316 (2009).

158

33. Su, Y. et al. Sequential phosphorylation of smoothened transduces graded hedgehog signaling. Sci. Signal. 4, ra43 (2011).

34. Shi, D. et al. Smoothened oligomerization/higher order clustering in lipid rafts is essential for high hedgehog activity transduction. J. Biol. Chem. 288, 12605–12614 (2013).

35. Orenic, T. V., Slusarski, D. C., Kroll, K. L. & Holmgren, R. A. Cloning and characterization of the segment polarity gene cubitus interruptus Dominant of Drosophila. Genes Dev. 4, 1053–1067 (1990).

36. Aza-Blanc, P., Ramírez-Weber, F. A., Laget, M. P. P., Schwartz, C. & Kornberg, T. B. B. Proteolysis that is inhibited by hedgehog targets Cubitus interruptus protein to the nucleus and converts it to a repressor. Cell 89, 1043–53 (1997).

37. Jiang, J. & Struhl, G. Regulation of the Hedgehog and Wingless signalling pathways by the F- box/WD40-repeat protein Slimb. Nature 391, 493–6 (1998).

38. Wang, G., Wang, B. & Jiang, J. Protein kinase A antagonizes Hedgehog signaling by regulating both the activator and repressor forms of Cubitus interruptus. Genes Dev. 13, 2828–2837 (1999).

39. Smelkinson, M. G., Zhou, Q. & Kalderon, D. Regulation of Ci-SCFSlimb Binding, Ci Proteolysis, and Hedgehog Pathway Activity by Ci Phosphorylation. Dev. Cell 13, 481–495 (2007).

40. Smelkinson, M. G. & Kalderon, D. Processing of the Drosophila Hedgehog signaling effector Ci- 155 to the repressor Ci-75 is mediated by direct binding to the SCF component slimb. Curr. Biol. 16, 110–116 (2006).

41. Methot, N. & Basler, K. Suppressor of fused opposes hedgehog signal transduction by impeding nuclear accumulation of the activator form of Cubitus interruptus. Development 127, 4001–4010 (2000).

42. Wang, G., Amanai, K., Wang, B. & Jiang, J. Interactions with Costal2 and Suppressor of fused regulate nuclear translocation and activity of Cubitus interruptus. Genes Dev. 14, 2893–2905 (2000).

43. Wang, Q. T. & Holmgren, R. A. The subcellular localization and activity of Drosophila cubitus interruptus are regulated at multiple levels. Development 126, 5097–5106 (1999).

44. Ogden, S. K. et al. G protein Gαi functions immediately downstream of Smoothened in Hedgehog signalling. Nature 456, 967–970 (2008).

45. Ranieri, N., Thérond, P. P. & Ruel, L. Switch of PKA substrates from Cubitus interruptus to Smoothened in the Hedgehog signalosome complex. Nat. Commun. 5, 5034 (2014).

46. Zhang, W. et al. Hedgehog-regulated Costal2-kinase complexes control phosphorylation and proteolytic processing of cubitus interruptus. Dev. Cell 8, 267–278 (2005).

47. Ogden, S. K. et al. Identification of a Functional Interaction between the Transmembrane Protein Smoothened and the Kinesin-Related Protein Costal2. Curr. Biol. 13, 1998–2003 (2003).

48. Jia, J., Tong, C. & Jiang, J. Smoothened transduces Hedgehog signal by physically interacting

159

with Costal2/Fused complex through its C-terminal tail. Genes Dev. 17, 2709–2720 (2003).

49. Ruel, L., Rodriguez, R., Gallet, A., Lavenant-Staccini, L. & Thérond, P. P. Stability and association of Smoothened, Costal2 and Fused with Cubitus interruptus are regulated by Hedgehog. Nat. Cell Biol. 5, 907–13 (2003).

50. Zhou, Q. & Kalderon, D. Hedgehog Activates Fused through Phosphorylation to Elicit a Full Spectrum of Pathway Responses. Dev. Cell 20, 802–814 (2011).

51. Ogden, S. K. et al. Smoothened regulates activator and repressor functions of Hedgehog signaling via two distinct mechanisms. J. Biol. Chem. 281, 7237–7243 (2006).

52. Ou, C. Y., Lin, Y. F., Chen, Y. J. & Chien, C. T. Distinct protein degradation mechanisms mediatead by Cul1 and Cul3 controlling Ci stability in Drosophila eye development. Genes Dev. 16, 2403–2414 (2002).

53. Zhang, Q. et al. A hedgehog-induced BTB protein modulates hedgehog signaling by degrading Ci/Gli transcription factor. Dev. Cell 10, 719–29 (2006).

54. Kent, D., Bush, E. W. & Hooper, J. E. Roadkill attenuates Hedgehog responses through degradation of Cubitus interruptus. Development 133, 2001–10 (2006).

55. Shi, Q. et al. Hedgehog-induced phosphorylation by CK1 sustains the activity of Ci/Gli activator. Proc. Natl. Acad. Sci. U. S. A. 111, E5651-60 (2014).

56. Capdevila, J. & Guerrero, I. Targeted expression of the signaling molecule decapentaplegic induces pattern duplications and growth alterations in Drosophila wings. EMBO J. 13, 4459–68 (1994).

57. Ingham, P. W. & Fietz, M. J. Quantitative effects of hedgehog and decapentaplegic activity on the patterning of the Drosophila wing. Curr. Biol. 5, 432–440 (1995).

58. Johnson, R. L., Grenier, J. K. & Scott, M. P. Patched Overexpression Alters Wing Disc Size and Pattern: Transcriptional and Post-Transcriptional Effects on Hedgehog Targets. Development 121, 4161–4170 (1995).

59. Kim, S. & Dynlacht, B. D. Assembling a primary cilium. Curr. Opin. Cell Biol. 25, 506–511 (2013).

60. Hu, Q. et al. A septin diffusion barrier at the base of the primary cilium maintains ciliary membrane protein distribution. Science 329, 436–9 (2010).

61. Jensen, V. L. et al. Formation of the transition zone by Mks5/Rpgrip1L establishes a ciliary zone of exclusion (CIZE) that compartmentalises ciliary signalling proteins and controls PIP2 ciliary abundance. EMBO J. e201488044 (2015). doi:10.15252/embj.201488044

62. Takao, D. et al. An assay for clogging the ciliary pore complex distinguishes mechanisms of cytosolic and membrane protein entry. Curr. Biol. 24, 2288–2294 (2014).

63. Dishinger, J. F. et al. Ciliary entry of the kinesin-2 motor KIF17 is regulated by importin-beta2 and RanGTP. Nat. Cell Biol. 12, 703–710 (2010).

160

64. Reiter, J. F., Blacque, O. E. & Leroux, M. R. The base of the cilium: roles for transition fibres and the transition zone in ciliary formation, maintenance and compartmentalization. EMBO Rep. 13, 608–618 (2012).

65. Wei, Q. et al. Transition fibre protein FBF1 is required for the ciliary entry of assembled intraflagellar transport complexes. Nat. Commun. 4, 2750 (2013).

66. Phua, S. C. et al. Dynamic Remodeling of Membrane Composition Drives Cell Cycle through Primary Cilia Excision Article Dynamic Remodeling of Membrane Composition Drives Cell Cycle through Primary Cilia Excision. Cell 168, 264–279. (2017).

67. Chávez, M. et al. Modulation of Ciliary Phosphoinositide Content Regulates Trafficking and Sonic Hedgehog Signaling Output. Dev. Cell 34, 338–350 (2015).

68. Moore, B. S. et al. Cilia have high cAMP levels that are inhibited by Sonic Hedgehog-regulated calcium dynamics. Proc. Natl. Acad. Sci. U. S. A. 113, 13069–13074 (2016).

69. Friedland-Little, J. M. et al. A novel murine allele of intraflagellar transport protein 172 causes a syndrome including VACTERL-like features with hydrocephalus. Hum. Mol. Genet. 20, 3725– 3737 (2011).

70. Haycraft, C. J. et al. Gli2 and Gli3 localize to cilia and require the intraflagellar transport protein polaris for processing and function. PLOS Genet. 1, e53 (2005).

71. Carpenter, B. S., Barry, R. L., Verhey, K. J. & Allen, B. L. The heterotrimeric kinesin-2 complex interacts with and regulates GLI protein function. J. Cell Sci. 128, 1034–1050 (2015).

72. Zeng, H., Jia, J. & Liu, A. Coordinated translocation of mammalian Gli proteins and suppressor of fused to the primary cilium. PLoS One 5, (2010).

73. Tukachinsky, H., Lopez, L. V. & Salic, A. A mechanism for vertebrate Hedgehog signaling: Recruitment to cilia and dissociation of SuFu-Gli protein complexes. J. Cell Biol. 191, 415–428 (2010).

74. Liem, K. F., He, M., Ocbina, P. J. R. & Anderson, K. V. Mouse Kif7/Costal2 is a cilia-associated protein that regulates Sonic hedgehog signaling. Proc. Natl. Acad. Sci. U. S. A. 106, 13377–13382 (2009).

75. May, S. R. et al. Loss of the retrograde motor for IFT disrupts localization of Smo to cilia and prevents the expression of both activator and repressor functions of Gli. Dev. Biol. 287, 378–389 (2005).

76. Liem, K. F. et al. The IFT-A complex regulates Shh signaling through cilia structure and membrane protein trafficking. J. Cell Biol. 197, 789–800 (2012).

77. Huangfu, D. & Anderson, K. V. Cilia and Hedgehog responsiveness in the mouse. Proc. Natl. Acad. Sci. U. S. A. 102, 11325–11330 (2005).

78. Liu, A., Wang, B. & Niswander, L. A. Mouse intraflagellar transport proteins regulate both the activator and repressor functions of Gli transcription factors. Development 132, 3103–3111 (2005).

161

79. Huangfu, D. et al. Hedgehog signalling in the mouse requires intraflagellar transport proteins. Nature 426, 83–87 (2003).

80. Sanchez, G. M. et al. Hedgehog Signaling Regulates the Ciliary Transport of Odorant Receptors in Drosophila. Cell Rep. 14, 464–470 (2016).

81. Kuzhandaivel, A., Schultz, S. W., Alkhori, L. & Alenius, M. Cilia-Mediated Hedgehog Signaling in Drosophila. Cell Rep. 7, 672–680 (2014).

82. Sarpal, R. et al. Drosophila KAP Interacts with the Kinesin II Motor Subunit KLP64D to Assemble Chordotonal Sensory Cilia, but Not Sperm Tails. Curr. Biol. 13, 1687–1696 (2003).

83. Han, Y. G., Kwok, B. H. & Kernan, M. J. Intraflagellar Transport Is Required in Drosophila to Differentiate Sensory Cilia but Not Sperm. Curr. Biol. 13, 1679–1686 (2003).

84. Riparbelli, M. G., Callaini, G. & Megraw, T. L. Assembly and Persistence of Primary Cilia in Dividing Drosophila Spermatocytes. Dev. Cell 23, 425–432 (2012).

85. Echelard, Y. et al. Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 75, 1417–1430 (1993).

86. Bumcrot, D. A., Takada, R. & McMahon, A. P. Proteolytic processing yields two secreted forms of sonic hedgehog. Mol. Cell. Biol. 15, 2294–303 (1995).

87. Valentini, R. P. et al. Post-translational processing and renal expression of mouse Indian hedgehog. J. Biol. Chem. 272, 8466–8473 (1997).

88. Li, Y., Zhang, H., Litingtung, Y. & Chiang, C. Cholesterol modification restricts the spread of Shh gradient in the limb bud. Proc. Natl. Acad. Sci. U. S. A. 103, 6548–53 (2006).

89. Pepinsky, R. B. et al. Identification of a Palmitic Acid-modified Form of Human Sonic hedgehog Identification of a Palmitic Acid-modified Form of Human Sonic hedgehog *. J. Biol. Chem. 273, 14037–14045 (1998).

90. Kohtz, J. D. et al. N-terminal fatty-acylation of sonic hedgehog enhances the induction of rodent ventral forebrain neurons. Development 128, 2351–63 (2001).

91. Tukachinsky, H., Petrov, K., Watanabe, M. & Salic, A. Mechanism of inhibition of the tumor suppressor Patched by Sonic Hedgehog. Proc. Natl. Acad. Sci. 201606719 (2016). doi:10.1073/pnas.1606719113

92. Long, J. et al. Identification of a family of fatty-acid-speciated sonic hedgehog proteins, whose members display differential biological properties. Cell Rep. 10, 1280–1287 (2015).

93. Etheridge, L. A., Crawford, T. Q., Zhang, S. & Roelink, H. Evidence for a role of vertebrate Disp1 in long-range Shh signaling. Development 137, 133–140 (2010).

94. Ma, Y. et al. Hedgehog-mediated patterning of the mammalian embryo requires transporter-like function of dispatched. Cell 111, 63–75 (2002).

95. Zeng, X. et al. A freely diffusible form of Sonic hedgehog mediates long-range signalling. Nature 411, 716–20 (2001).

162

96. Chen, M. H., Li, Y. J., Kawakami, T., Xu, S. M. & Chuang, P. T. Palmitoylation is required for the production of a soluble multimeric Hedgehog protein complex and long-range signaling in vertebrates. Genes Dev. 18, 641–659 (2004).

97. Sanders, T. a, Llagostera, E. & Barna, M. Specialized filopodia direct long-range transport of SHH during vertebrate tissue patterning. Nature 497, 628–32 (2013).

98. Dierker, T., Dreler, R., Petersen, A., Bordych, C. & Grobe, K. Heparan sulfate-modulated, metalloprotease-mediated sonic hedgehog release from producing cells. J. Biol. Chem. 284, 8013– 8022 (2009).

99. Jakobs, P. et al. Scube2 enhances proteolytic Shh processing from the surface of Shh-producing cells. J. Cell Sci. 127, 1726–37 (2014).

100. Jakobs, P. et al. Bridging the gap: heparan sulfate and Scube2 assemble Sonic hedgehog release complexes at the surface of producing cells. Sci. Rep. 6, 26435 (2016).

101. Grover, V. K., Valadez, J. G., Bowman, A. B. & Cooper, M. K. Lipid modifications of Sonic hedgehog ligand dictate cellular reception and signal response. PLoS One 6, 1–12 (2011).

102. Carpenter, D. et al. Characterization of two patched receptors for the vertebrate hedgehog protein family. Proc. Natl. Acad. Sci. U. S. A. 95, 13630–4 (1998).

103. Marigo, V., Davey, R. A., Zuo, Y., Cunningham, J. M. & Tabin, C. J. Biochemical evidence that Patched is the Hedgehog receptor. Nature 384, 176–9 (1996).

104. Stone, D. M. et al. The tumour-suppressor gene patched encodes a candidate receptor for Sonic hedgehog. Nature 384, 129–34 (1996).

105. Marigo, V. & Tabin, C. J. Regulation of patched by sonic hedgehog in the developing neural tube. Proc. Natl. Acad. Sci. U. S. A. 93, 9346–51 (1996).

106. Alfaro, A. C., Roberts, B., Kwong, L., Bijlsma, M. F. & Roelink, H. Ptch2 mediates the Shh response in Ptch1-/- cells. Development 141, 3331–9 (2014).

107. Taipale, J., Cooper, M. K., Maiti, T. & Beachy, P. A. Patched acts catalytically to suppress the activity of Smoothened. Nature 418, 892–7 (2002).

108. Goodrich, L. V, Milenković, L., Higgins, K. M. & Scott, M. P. Altered neural cell fates and medulloblastoma in mouse patched mutants. Science (80-. ). 277, 1109–13 (1997).

109. Goodrich, L. V., Jung, D., Higgins, K. M. & Scott, M. P. Overexpression of ptc1 inhibits induction of Shh target genes and prevents normal patterning in the neural tube. Dev. Biol. 211, 323–334 (1999).

110. Incardona, J. P. et al. Receptor-mediated endocytosis of soluble and membrane-tethered Sonic hedgehog by Patched-1. Proc. Natl. Acad. Sci. U. S. A. 97, 12044–12049 (2000).

111. Briscoe, J., Chen, Y., Jessell, T. M. & Struhl, G. A hedgehog-insensitive form of Patched provides evidence for direct long-range morphogen activity of Sonic hedgehog in the neural tube. Mol. Cell 7, 1279–1291 (2001).

163

112. Incardona, J. P., Gruenberg, J. & Roelink, H. Sonic hedgehog induces the segregation of patched and smoothened in endosomes. Curr. Biol. 12, 983–995 (2002).

113. Bosanac, I. et al. The structure of SHH in complex with HHIP reveals a recognition role for the Shh pseudo active site in signaling. Nat. Struct. Mol. Biol. 16, 691–697 (2009).

114. Capurro, M. I. et al. Glypican-3 Inhibits Hedgehog Signaling during Development by Competing with Patched for Hedgehog Binding. Dev. Cell 14, 700–711 (2008).

115. Murone, M., Rosenthal, A. & de Sauvage, F. J. Sonic hedgehog signaling by the patched- smoothened receptor complex. Curr. Biol. 9, 76–84 (1999).

116. Rohatgi, R., Milenkovic, L. & Scott, M. P. Patched1 Regulates Hedgehog Signaling at the Primary Cilium. Science (80-. ). 317, 372–376 (2007).

117. Wang, Y., Zhou, Z., Walsh, C. T. & McMahon, A. P. Selective translocation of intracellular Smoothened to the primary cilium in response to Hedgehog pathway modulation. Proc. Natl. Acad. Sci. U. S. A. 106, 2623–2628 (2009).

118. Kim, J. et al. The role of ciliary trafficking in Hedgehog receptor signaling. Sci. Signal. 8, ra55 (2015).

119. Bijlsma, M. F. et al. Repression of smoothened by patched-dependent (pro-)vitamin D3 secretion. PLoS Biol. 4, 1397–1410 (2006).

120. Bidet, M. et al. The hedgehog receptor patched is involved in cholesterol transport. PLoS One 6, 1–11 (2011).

121. Chen, J. K., Taipale, J., Cooper, M. K. & Beachy, P. A. Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened. Genes Dev. 16, 2743–2748 (2002).

122. Chen, J. K., Taipale, J., Young, K. E., Maiti, T. & Beachy, P. A. Small molecule modulation of Smoothened activity. Proc. Natl. Acad. Sci. U. S. A. 99, 14071–14076 (2002).

123. Yauch, R. L. et al. Smoothened mutation confers resistance to a Hedgehog pathway inhibitor in medulloblastoma. Science 326, 572–4 (2009).

124. Frank-Kamenetsky, M. et al. Small-molecule modulators of Hedgehog signaling: identification and characterization of Smoothened agonists and antagonists. J. Biol. 1, 10 (2002).

125. Sinha, S. & Chen, J. K. Purmorphamine activates the Hedgehog pathway by targeting Smoothened. Nat. Chem. Biol. 2, 29–30 (2006).

126. Luchetti, G. et al. Cholesterol activates the G-protein coupled receptor Smoothened to promote morphogenetic signaling. Elife 5, 1–22 (2016).

127. Huang, P. et al. Cellular Cholesterol Directly Activates Smoothened in Hedgehog Signaling. Cell 166, 1176–1187.e14 (2016).

128. Byrne, E. F. X. et al. Structural basis of Smoothened regulation by its extracellular domains. Nature 535, 1–21 (2016).

164

129. Corbit, K. C. et al. Vertebrate Smoothened functions at the primary cilium. Nature 437, 1018– 1021 (2005).

130. Milenkovic, L., Scott, M. P. & Rohatgi, R. Lateral transport of Smoothened from the plasma membrane to the membrane of the cilium. J. Cell Biol. 187, 365–374 (2009).

131. Kovacs, J. J. et al. Β-Arrestin-Mediated Localization of Smoothened To the Primary Cilium. Science (80-. ). 320, 1777–1781 (2008).

132. Jiang, K. et al. PI(4)P Promotes Phosphorylation and Conformational Change of Smoothened through Interaction with Its C-terminal Tail. PLoS Biol. 14, 1–26 (2016).

133. Rohatgi, R., Milenkovic, L., Corcoran, R. B. & Scott, M. P. Hedgehog signal transduction by Smoothened: pharmacologic evidence for a 2-step activation process. Proc. Natl. Acad. Sci. U. S. A. 106, 3196–201 (2009).

134. Wilson, C. W., Chen, M. H. & Chuang, P. T. Smoothened adopts multiple active and inactive conformations capable of trafficking to the primary cilium. PLoS One 4, (2009).

135. Chen, Y. et al. Sonic Hedgehog dependent Phosphorylation by CK1alpha and GRK2 is required for Ciliary Accumulation and Activation of Smoothened. PLoS Biol. 9, (2011).

136. Zhang, Z. et al. Suppressor of Fused chaperones Gli proteins to generate transcriptional responses to Sonic Hedgehog signaling. Mol. Cell. Biol. 37, e00421-16 (2017).

137. Kasai, K. et al. The G12 family of heterotrimeric G proteins and Rho GTPase mediate Sonic hedgehog signalling. Genes to Cells 9, 49–58 (2004).

138. Villanueva, H. et al. An essential role for Gai2 in Smoothened-stimulated epithelial cell proliferation in the mammary gland. Sci. Signal. 8, 1–11 (2015).

139. Sasaki, H., Nishizaki, Y., Hui, C., Nakafuku, M. & Kondoh, H. Regulation of Gli2 and Gli3 activities by an amino-terminal repression domain: implication of Gli2 and Gli3 as primary mediators of Shh signaling. Development 126, 3915–3924 (1999).

140. Pavletich, N. P. & Pabo, C. O. Crystal structure of a five-finger GLI-DNA complex: new perspectives on zinc fingers. Science 261, 1701–7 (1993).

141. Zhou, H., Kim, S., Ishii, S. & Boyer, T. G. Mediator Modulates Gli3-Dependent Sonic Hedgehog Signaling. Mol. Cell. Biol. 26, 8667–8682 (2006).

142. Kalff-Suske, M. et al. Point mutations throughout the GLI3 gene cause Greig cephalopolysyndactyly syndrome. Hum. Mol. Genet. 8, 1769–1777 (1999).

143. Dai, P. et al. Sonic hedgehog-induced activation of the Gli1 promoter is mediated by GLI3. J. Biol. Chem. 274, 8143–8152 (1999).

144. Yoon, J. W. et al. P53 modulates the activity of the GLI1 oncogene through interactions with the shared coactivator TAF9. DNA Repair (Amst). 34, 9–17 (2015).

145. Tsanev, R. et al. Identification of the gene transcription repressor domain of Gli3. FEBS Lett. 583, 224–228 (2009).

165

146. Cheng, S. Y. & Bishop, J. M. Suppressor of Fused represses Gli-mediated transcription by recruiting the SAP18-mSin3 corepressor complex. Proc. Natl. Acad. Sci. U. S. A. 99, 5442–7 (2002).

147. Tsanev, R. et al. The Transcriptional Repressor Domain of Gli3 Is Intrinsically Disordered. PLoS One 8, 1–9 (2013).

148. Pierce, W. K. et al. Multiple Weak Linear Motifs Enhance Recruitment and Processivity in SPOP- Mediated Substrate Ubiquitination. J. Mol. Biol. 428, 1256–1271 (2016).

149. Shi, Q., Han, Y. & Jiang, J. Suppressor of fused impedes Ci/Gli nuclear import by opposing Trn/Kapβ2 in Hedgehog signaling. J. Cell Sci. 127, 1092–103 (2014).

150. Szczepny, A. et al. Overlapping binding sites for importin β1 and suppressor of fused (SuFu) on glioma-associated oncogene homologue 1 (Gli1) regulate its nuclear localization. Biochem. J. 461, 469–76 (2014).

151. Kogerman, P. et al. Mammalian Suppressor-of-Fused modulates nuclear – cytoplasmic shuttling of GLI-1. Nat. Cell Biol. 1, 312–319 (1999).

152. Barnfield, P. C., Zhang, X., Thanabalasingham, V., Yoshida, M. & Hui, C. C. Negative regulation of Gli1 and Gli2 activator function by Suppressor of fused through multiple mechanisms. Differentiation 73, 397–405 (2005).

153. Pan, Y. & Wang, B. A novel protein-processing domain in Gli2 and Gli3 differentially blocks complete protein degradation by the proteasome. J. Biol. Chem. 282, 10846–52 (2007).

154. Aza-Blanc, P., Lin, H. Y., Ruiz i Altaba, a & Kornberg, T. B. Expression of the vertebrate Gli proteins in Drosophila reveals a distribution of activator and repressor activities. Development 127, 4293–301 (2000).

155. Pan, Y., Bai, C. B., Joyner, A. L. & Wang, B. Sonic hedgehog Signaling Regulates Gli2 Transcriptional Activity by Suppressing Its Processing and Degradation Sonic hedgehog Signaling Regulates Gli2 Transcriptional Activity by Suppressing Its Processing and Degradation †. Mol. Cell. Biol. 26, 3365–3377 (2006).

156. Tempé, D., Casas, M., Karaz, S., Blanchet-Tournier, M.-F. & Concordet, J.-P. Multisite protein kinase A and glycogen synthase kinase 3beta phosphorylation leads to Gli3 ubiquitination by SCFbetaTrCP. Mol. Cell. Biol. 26, 4316–26 (2006).

157. Schrader, E. K., Harstad, K. G., Holmgren, R. A. & Matouschek, A. A Three-part signal governs differential processing of Gli1 and Gli3 proteins by the proteasome. J. Biol. Chem. 286, 39051– 39058 (2011).

158. Huntzicker, E. G. et al. Dual degradation signals control Gli protein stability and tumor formation. Genes Dev. 20, 276–281 (2006).

159. Di Marcotullio, L. et al. Numb is a suppressor of Hedgehog signalling and targets Gli1 for Itch- dependent ubiquitination. Nat. Cell Biol. 8, 1415–23 (2006).

160. Di Marcotullio, L. et al. Numb activates the E3 ligase Itch to control Gli1 function through a novel

166

degradation signal. Oncogene 30, 65–76 (2011).

161. Sheng, T., Chi, S., Zhang, X. & Xie, J. Regulation of Gli1 localization by the cAMP/protein kinase A signaling axis through a site near the nuclear localization signal. J. Biol. Chem. 281, 9–12 (2006).

162. Niewiadomski, P. et al. Gli protein activity is controlled by multisite phosphorylation in vertebrate hedgehog signaling. Cell Rep. 6, 168–181 (2014).

163. Wang, B. & Li, Y. Evidence for the direct involvement of {beta}TrCP in Gli3 protein processing. Proc. Natl. Acad. Sci. U. S. A. 103, 33–8 (2006).

164. Bhatia, N. et al. Gli2 is targeted for ubiquitination and degradation by beta-TrCP ubiquitin ligase. J. Biol. Chem. 281, 19320–6 (2006).

165. Varjosalo, M. et al. Application of Active and Kinase-Deficient Kinome Collection for Identification of Kinases Regulating Hedgehog Signaling. Cell 133, 537–548 (2008).

166. Asaoka, Y. et al. Identification of a suppressive mechanism for Hedgehog signaling through a novel interaction of gli with 14-3-3. J. Biol. Chem. 285, 4185–4194 (2010).

167. Cox, B., Briscoe, J. & Ulloa, F. SUMOylation by pias1 regulates the activity of the hedgehog dependent gli transcription factors. PLoS One 5, (2010).

168. Han, L., Pan, Y. & Wang, B. Small ubiquitin-like modifier (SUMO) modification inhibits GLI2 protein transcriptional activity in Vitro and in Vivo. J. Biol. Chem. 287, 20483–20489 (2012).

169. Eguether, T. et al. IFT27 links the bbsome to ift for maintenance of the ciliary signaling compartment. Dev. Cell 31, 279–290 (2014).

170. Santos, N. & Reiter, J. F. A central region of Gli2 regulates its localization to the primary cilium and transcriptional activity. J. Cell Sci. 127, 1500–10 (2014).

171. Liu, J., Zeng, H. & Liu, A. The loss of Hh responsiveness by a non-ciliary Gli2 variant. Development 142, 1651–60 (2015).

172. Jia, J. et al. Suppressor of Fused inhibits mammalian Hedgehog signaling in the absence of cilia. Dev. Biol. 330, 452–60 (2009).

173. Chen, M.-H. et al. Cilium-independent regulation of Gli protein function by Sufu in Hedgehog signaling is evolutionarily conserved. Genes Dev. 23, 1910–28 (2009).

174. Kim, J., Kato, M. & Beachy, P. a. Gli2 trafficking links Hedgehog-dependent activation of Smoothened in the primary cilium to transcriptional activation in the nucleus. Proc. Natl. Acad. Sci. U. S. A. 106, 21666–21671 (2009).

175. Merchant, M. et al. Suppressor of fused regulates Gli activity through a dual binding mechanism. Mol. Cell. Biol. 24, 8627–41 (2004).

176. Humke, E. W., Dorn, K. V, Milenkovic, L., Scott, M. P. & Rohatgi, R. The output of Hedgehog signaling is controlled by the dynamic association between Suppressor of Fused and the Gli proteins. Genes Dev. 24, 670–82 (2010).

167

177. Stone, D. M. et al. Characterization of the human suppressor of fused, a negative regulator of the zinc-finger transcription factor Gli. J. Cell Sci. 112 ( Pt 2, 4437–48 (1999).

178. Kise, Y., Morinaka, A., Teglund, S. & Miki, H. Sufu recruits GSK3β for efficient processing of Gli3. Biochem. Biophys. Res. Commun. 387, 569–574 (2009).

179. Dunaeva, M., Michelson, P., Kogerman, P. & Toftgard, R. Characterization of the physical interaction of Gli proteins with SUFU proteins. J. Biol. Chem. 278, 5116–5122 (2003).

180. Zhang, Y. et al. Structural insight into the mutual recognition and regulation between Suppressor of Fused and Gli/Ci. Nat. Commun. 4, 2608 (2013).

181. Lin, C. et al. Regulation of Sufu activity by p66b and Mycbp provides new insight into vertebrate Hedgehog signaling. Genes Dev. 28, 2547–2563 (2014).

182. Han, Y., Shi, Q. & Jiang, J. Multisite interaction with Sufu regulates Ci/Gli activity through distinct mechanisms in Hh signal transduction. Proc. Natl. Acad. Sci. U. S. A. 112, 6383–8 (2015).

183. Paces-Fessy, M., Boucher, D., Petit, E., Paute-Briand, S. & Blanchet-Tournier, M.-F. The negative regulator of Gli, Suppressor of fused (Sufu), interacts with SAP18, Galectin3 and other nuclear proteins. Biochem. J. 378, 353–62 (2004).

184. Chen, Y. et al. Dual phosphorylation of suppressor of fused (Sufu) by PKA and GSK3?? regulates its stability and localization in the primary cilium. J. Biol. Chem. 286, 13502–13511 (2011).

185. Wagner, M., Thaller, C., Jessell, T. & Eichele, G. Polarizing activity and retinoid synthesis in the floor plate of the neural tube. Nature 345, 819–22 (1990).

186. Johnson, R. L., Laufer, E., Riddle, R. D. & Tabin, C. Ectopic expression of Sonic hedgehog alters dorsal-ventral patterning of somites. Cell 79, 1165–1173 (1994).

187. Wijgerde, M., McMahon, J. A., Rule, M. & McMahon, A. P. A direct requirement for hedgehog signaling for normal specification of all ventral progenitor domains in the presumptive mammalian spinal cord. Genes Dev. 16, 2849–2864 (2002).

188. Maier, J. A., Lo, Y. T. & Harfe, B. D. Foxa1 and Foxa2 Are Required for Formation of the Intervertebral Discs. PLoS One 8, (2013).

189. Sasaki, H., Hui, C., Nakafuku, M. & Kondoh, H. A binding site for Gli proteins is essential for HNF-3beta floor plate enhancer activity in transgenics and can respond to Shh in vitro. Development 124, 1313–22 (1997).

190. Cho, G., Lim, Y., Cho, I. T., Simonet, J. C. & Golden, J. A. Arx together with FoxA2, regulates Shh floor plate expression. Dev. Biol. 393, 137–148 (2014).

191. Briscoe, J., Pierani, A., Jessell, T. M. & Ericson, J. A homeodomain protein code specifies progenitor cell identity and neuronal fate in the ventral neural tube. Cell 101, 435–445 (2000).

192. Briscoe, J. et al. Homeobox gene Nkx2.2 and specification of neuronal identity by graded Sonic hedgehog signalling. Nature 398, 622–627 (1999).

193. Takebayashi, H. et al. The basic helix-loop-helix factor Olig2 is essential for the development of

168

motoneuron and oligodendrocyte lineages. Curr. Biol. 12, 1157–1163 (2002).

194. Sun, T. et al. Cross-repressive interaction of the Olig2 and Nkx2.2 transcription factors in developing neural tube associated with formation of a specific physical complex. J. Neurosci. 23, 9547–9556 (2003).

195. Ericson, J. et al. Pax6 controls progenitor cell identity and neuronal fate in response to graded Shh signaling. Cell 90, 169–180 (1997).

196. Oosterveen, T. et al. Mechanistic Differences in the Transcriptional Interpretation of Local and Long-Range Shh Morphogen Signaling. Dev. Cell 23, 1006–1019 (2012).

197. Bai, C. B., Auerbach, W., Lee, J. S., Stephen, D. & Joyner, A. L. Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129, 4753–4761 (2002).

198. Bai, C. B. & Joyner, a L. Gli1 can rescue the in vivo function of Gli2. Development 128, 5161– 5172 (2001).

199. Ding, Q. et al. Diminished Sonic hedgehog signaling and lack of floor plate differentiation in Gli2 mutant mice. Development 125, 2533–43 (1998).

200. Matise, M. P., Epstein, D. J., Park, H. L., Platt, K. A. & Joyner, A. L. Gli2 is required for induction of floor plate and adjacent cells, but not most ventral neurons in the mouse central nervous system. Development 125, 2759–2770 (1998).

201. Park, H. L. et al. Mouse Gli1 mutants are viable but have defects in SHH signaling in combination with a Gli2 mutation. Development 127, 1593–605 (2000).

202. Persson, M. et al. Dorsal-ventral patterning of the spinal cord requires Gli3 transcriptional repressor activity. Genes Dev. 16, 2865–2878 (2002).

203. Bai, C. B., Stephen, D. & Joyner, A. L. All mouse ventral spinal cord patterning by Hedgehog is Gli dependent and involves an activator function of Gli3. Dev. Cell 6, 103–115 (2004).

204. Lei, Q., Zelman, A. K., Kuang, E., Li, S. & Matise, M. P. Transduction of graded Hedgehog signaling by a combination of Gli2 and Gli3 activator functions in the developing spinal cord. Development 131, 3593–3604 (2004).

205. Sambrook, P. & Cooper, C. Osteoporosis. Lancet 367, 2010–8 (2006).

206. Khosla, S. & Melton III, L. J. Osteopenia. N. Engl. J. Med. 356, 2293–2300 (2007).

207. Long, F. Building strong bones: molecular regulation of the osteoblast lineage. Nat. Rev. Mol. Cell Biol. 13, 27–38 (2012).

208. St-Jacques, B. Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev. 13, 2072–2086 (1999).

209. Mak, K. K. et al. Hedgehog signaling in mature osteoblasts regulates bone formation and resorption by controlling PTHrP and RANKL expression. Dev. Cell 14, 674–88 (2008).

169

210. Vortkamp, A., Lee, K. & Lanske, B. Regulation of rate of cartilage differentiation by Indian hedgehog and PTH-related protein. Science (80-. ). 273, 613–22 (1996).

211. Koziel, L., Kunath, M., Kelly, O. & Vortkamp, A. Ext1-dependent heparan sulfate regulates the range of Ihh signaling during endochondral ossification. Dev. Cell 6, 801–813 (2004).

212. Koziel, L., Wuelling, M., Schneider, S. & Vortkamp, A. Gli3 acts as a repressor downstream of Ihh in regulating two distinct steps of chondrocyte differentiation. Development 1, 5249–5260 (2005).

213. Kobayashi, T. & Soegiarto, D. Indian hedgehog stimulates periarticular chondrocyte differentiation to regulate growth plate length independently of PTHrP. J. Clin. … 115, (2005).

214. Long, F., Zhang, X. M., Karp, S., Yang, Y. & McMahon, a P. Genetic manipulation of hedgehog signaling in the endochondral skeleton reveals a direct role in the regulation of chondrocyte proliferation. Development 128, 5099–108 (2001).

215. Kobayashi, T. et al. PTHrP and Indian hedgehog control differentiation of growth plate chondrocytes at multiple steps. Development 129, 2977–86 (2002).

216. Mau, E. et al. PTHrP regulates growth plate chondrocyte differentiation and proliferation in a Gli3 dependent manner utilizing hedgehog ligand dependent and independent mechanisms. Dev. Biol. 305, 28–39 (2007).

217. Hsu, S.-H. C. et al. Suppressor of fused (Sufu) mediates the effect of parathyroid hormone-like hormone (Pthlh) on chondrocyte differentiation in the growth plate. J. Biol. Chem. 287, 36222–8 (2012).

218. Mak, K. K., Kronenberg, H. M., Chuang, P.-T., Mackem, S. & Yang, Y. Indian hedgehog signals independently of PTHrP to promote chondrocyte hypertrophy. Development 135, 1947–1956 (2008).

219. Long, F. et al. Ihh signaling is directly required for the osteoblast lineage in the endochondral skeleton. Development 131, 1309–18 (2004).

220. Joeng, K. S. & Long, F. The Gli2 transcriptional activator is a crucial effector for Ihh signaling in osteoblast development and cartilage vascularization. Development 136, 4177–85 (2009).

221. Hilton, M. J., Tu, X., Cook, J., Hu, H. & Long, F. Ihh controls cartilage development by antagonizing Gli3, but requires additional effectors to regulate osteoblast and vascular development. Development 132, 4339–51 (2005).

222. Day, T. F. & Yang, Y. Wnt and hedgehog signaling pathways in bone development. J. Bone Joint Surg. Am. 90 Suppl 1, 19–24 (2008).

223. Yang, Y. Wnt5a and Wnt5b exhibit distinct activities in coordinating chondrocyte proliferation and differentiation. Development 130, 1003–1015 (2003).

224. Zou, H., Wieser, R., Massague, J. & Niswander, L. Distinct roles of type I bone morphogenetic protein receptors in the formation and differentiation of cartilage. Genes Dev. 11, 2191–2203 (1997).

170

225. Hojo, H. et al. Hedgehog-Gli activators direct osteo-chondrogenic function of bone morphogenetic protein toward osteogenesis in the perichondrium. J. Biol. Chem. 288, 9924–32 (2013).

226. Liu, Z., Xu, J., Colvin, J. S. & Ornitz, D. M. Coordination of chondrogenesis and osteogenesis by fibroblast growth factor 18. Genes Dev. 16, 859–69 (2002).

227. Ohbayashi, N. et al. FGF18 is required for normal cell proliferation and differentiation during osteogenesis and chondrogenesis. Genes Dev. 16, 870–9 (2002).

228. Minina, E., Kreschel, C., Naski, M. C., Ornitz, D. M. & Vortkamp, A. Interaction of FGF, Ihh/Pthlh, and BMP signaling integrates chondrocyte proliferation and hypertrophic differentiation. Dev. Cell 3, 439–49 (2002).

229. Errington, W. J. et al. Adaptor protein self-assembly drives the control of a cullin-RING ubiquitin ligase. Structure 20, 1141–53 (2012).

230. Ciechanover, A. Proteolysis: From the lysosome to ubiquitin and the proteasome. Nat. Rev. Mol. Cell Biol. 6, 79–86 (2005).

231. Metzger, M. B., Hristova, V. A. & Weissman, A. M. HECT and RING finger families of E3 ubiquitin ligases at a glance. J. Cell Sci. 125, 531–537 (2012).

232. Nagai, Y. et al. Identification of a novel nuclear speckle-type protein, SPOP. FEBS Lett. 418, 23– 26 (1997).

233. Marzahn, M. R. et al. Higher-order oligomerization promotes localization of SPOP to liquid nuclear speckles. EMBO J. 35, 1–22 (2016).

234. Liu, A., Desai, B. M. & Stoffers, D. A. Identification of PCIF1, a POZ Domain Protein That Inhibits PDX-1 (MODY4) Transcriptional Activity. Mol. Cell. Biol. 24, 4372–4383 (2004).

235. Li, G. et al. SPOP promotes tumorigenesis by acting as a key regulatory hub in kidney cancer. Cancer Cell 25, 455–468 (2014).

236. Bunce, M. W., Boronenkov, I. V. & Anderson, R. A. Coordinated activation of the nuclear ubiquitin ligase Cul3-SPOP by the generation of phosphatidylinositol 5-phosphate. J. Biol. Chem. 283, 8678–8686 (2008).

237. Theurillat, J.-P. P. et al. Ubiquitylome analysis identifies dysregulation of effector substrates in SPOP-mutant prostate cancer. Science (80-. ). 346, 85–89 (2014).

238. Zhang, Q. et al. Multiple Ser/Thr-rich degrons mediate the degradation of Ci/Gli by the Cul3- HIB/SPOP E3 ubiquitin ligase. Proc. Natl. Acad. Sci. U. S. A. 106, 21191–6 (2009).

239. García-Flores, M. et al. Clinico-pathological significance of the molecular alterations of the SPOP gene in prostate cancer. Eur. J. Cancer 50, 2994–3002 (2014).

240. Khani, F. et al. Evidence for molecular differences in prostate cancer between african American and Caucasian Men. Clin. Cancer Res. 20, 4925–4934 (2014).

241. Barbieri, C. E. et al. Exome sequencing identifies recurrent SPOP, FOXA1 and MED12 mutations in prostate cancer. Nat. Genet. 44, 685–9 (2012).

171

242. Lee, S. H. et al. Genetic Progression of High Grade Prostatic Intraepithelial Neoplasia to Prostate Cancer. Eur. Urol. 69, 823–830 (2016).

243. Spans, L. et al. Genomic and epigenomic analysis of high-risk prostate cancer reveals changes in hydroxymethylation and TET1. Oncotarget 7, (2016).

244. Zuhlke, K. A. et al. Identification of a novel germline SPOP mutation in a family with hereditary prostate cancer. Prostate 74, 983–990 (2014).

245. Buckles, E. et al. Identification of speckle-type POZ protein somatic mutations in African American prostate cancer. Asian J. Androl. 16, 829–32 (2014).

246. Manson-Bahr, D. et al. Mutation detection in formalin-fixed prostate cancer biopsies taken at the time of diagnosis using next-generation DNA sequencing. J Clin Pathol 68, 212–217 (2015).

247. Blattner, M. et al. SPOP mutations in prostate cancer across demographically diverse patient cohorts. Neoplasia 16, 14–20 (2014).

248. Abeshouse, A. et al. The Molecular Taxonomy of Primary Prostate Cancer. Cell 163, 1011–1025 (2015).

249. Ding, D., Song, T., Jun, W., Tan, Z. & Fang, J. Decreased expression of the SPOP gene is associated with poor prognosis in glioma. Int. J. Oncol. 46, 333–341 (2015).

250. Liu, Y. et al. Insight in glioma susceptibility through an analysis of 6p22.3, 12p13.33-12.1, 17q22- 23.2 and 18q23 SNP genotypes in familial and non-familial glioma. Hum. Genet. 131, 1507–1517 (2012).

251. Jia, D. et al. Exome sequencing of hepatoblastoma reveals novel mutations and cancer genes in the Wnt pathway and ubiquitin ligase complex. Hepatology 1–112 (2014). doi:10.1002/hep.27243

252. Jones, S. et al. Genomic analyses of gynaecologic carcinosarcomas reveal frequent mutations in chromatin remodelling genes. Nat. Commun. 5, 5006 (2014).

253. Jung, S. et al. Mutational burdens and evolutionary ages of thyroid follicular adenoma are comparable to those of follicular carcinoma. Oncotarget Epub ahead, (2016).

254. Le Gallo, M. et al. Exome sequencing of serous endometrial tumors identifies recurrent somatic mutations in chromatin-remodeling and ubiquitin ligase complex genes. Nat. Genet. 44, 1310– 1315 (2012).

255. Yoo, S.-K. et al. Comprehensive Analysis of the Transcriptional and Mutational Landscape of Follicular and Papillary Thyroid Cancers. PLOS Genet. 12, e1006239 (2016).

256. Luo, J. et al. Biochemical and Biophysical Research Communications SPOP promotes SIRT2 degradation and suppresses non-small cell lung cancer cell growth. Biochem. Biophys. Res. Commun. 483, 880–884 (2017).

257. Hu, X. et al. Speckle-type POZ (pox virus and zinc finger protein) protein gene deletion in ovarian cancer: Fluorescence in situ hybridization analysis of a tissue microarray. Oncol Lett 12, 658–662 (2016).

172

258. Gao, K. et al. Tumor suppressor SPOP mediates the proteasomal degradation of progesterone receptors (PRs) in breast cancer cells. Am. J. Cancer Res. 5, 3210–20 (2015).

259. Li, C. et al. Tumor-suppressor role for the SPOP ubiquitin ligase in signal-dependent proteolysis of the oncogenic co-activator SRC-3/AIB1. Oncogene 30, 4350–64 (2011).

260. Huang, Y. et al. Speckle-type POZ protein is negatively associated with malignancies and inhibits cell proliferation and migration in liver cancer. Tumor Biol. 36, 9753–9761 (2015).

261. Kim, M. S., Je, E. M., Oh, J. E., Yoo, N. J. & Lee, S. H. Mutational and Expressional Analyses Of SPOP, A Candidate Tumor Suppressor Gene, In Prostate, Gastric and Colorectal Cancers. Apmis 121, 626–633 (2013).

262. Xu, J. et al. Properties and Clinical Relevance of Speckle-Type POZ Protein in Human Colorectal Cancer. J. Gastrointest. Surg. 19, 1484–96 (2015).

263. Zhi, X. et al. Silencing speckle-type POZ protein by promoter hypermethylation decreases cell apoptosis through upregulating Hedgehog signaling pathway in colorectal cancer. Cell Death Dis. 7, e2569 (2016).

264. Zhang, D. et al. Speckle-type POZ protein, SPOP, is involved in the DNA damage response. Carcinogenesis 35, 1691–1697 (2014).

265. Boysen, G. et al. SPOP mutation leads to genomic instability in prostate cancer. Elife 4, 1–18 (2015).

266. Geng, C. et al. Androgen receptor is the key transcriptional mediator of the tumor suppressor SPOP in prostate cancer. Cancer Res. 74, 5631–43 (2014).

267. An, J., Wang, C., Deng, Y., Yu, L. & Huang, H. Destruction of Full-Length Androgen Receptor by Wild-Type SPOP, but Not Prostate-Cancer-Associated Mutants. Cell Rep. 6, 657–669 (2014).

268. Geng, C. et al. Prostate cancer-associated mutations in speckle-type POZ protein (SPOP) regulate steroid receptor coactivator 3 protein turnover. Proc. Natl. Acad. Sci. U. S. A. 110, 6997–7002 (2013).

269. Gan, W. et al. SPOP Promotes Ubiquitination and Degradation of the ERG Oncoprotein to Suppress Prostate Cancer Progression. Mol. Cell 59, 917–930 (2015).

270. An, J. et al. Truncated ERG Oncoproteins from TMPRSS2-ERG Fusions Are Resistant to SPOP- Mediated Proteasome Degradation. Mol. Cell 59, 904–916 (2015).

271. Zhang, L. et al. Tumor suppressor SPOP ubiquitinates and degrades EglN2 to compromise growth of prostate cancer cells. Cancer Lett. 390, 11–20 (2017).

272. Wu, F. et al. Prostate cancer-associated mutation in SPOP impairs its ability to target Cdc20 for poly-ubiquitination and degradation. Cancer Lett. 385, 207–214 (2017).

273. Zhang, P. et al. Destruction of DDIT3/CHOP protein by wild-type SPOP but not prostate cancer- associated mutants. Hum. Mutat. 35, 1142–1151 (2014).

274. Zhang, P. et al. Endometrial cancer-associated mutants of SPOP are defective in regulating

173

estrogen receptor-α protein turnover. Cell Death Dis. 6, e1687 (2015).

275. Zeng, C. et al. SPOP suppresses tumorigenesis by regulating Hedgehog/Gli2 signaling pathway in gastric cancer. J. Exp. Clin. Cancer Res. 33, 75 (2014).

276. Liu, X., Sun, G. & Sun, X. RNA interference-mediated silencing of speckle-type POZ protein promotes apoptosis of renal cell cancer cells. Onco. Targets. Ther. 9, 2393–2402 (2016).

277. Claiborn, K. C. et al. Pcif1 modulates Pdx1 protein stability and pancreatic β cell function and survival in mice. The Journal of clinical investigation 120, 3713–21 (2010).

278. Liu, A., Oliver-Krasinski, J. & Stoffers, D. A. Two conserved domains in PCIF1 mediate interaction with pancreatic transcription factor PDX-1. FEBS Lett. 580, 6701–6706 (2006).

279. Kwon, J. E. et al. BTB domain-containing speckle-type POZ protein (SPOP) serves as an adaptor of Daxx for ubiquitination by Cul3-based ubiquitin ligase. J. Biol. Chem. 281, 12664–72 (2006).

280. La, M. et al. Daxx-mediated transcriptional repression of MMP1 gene is reversed by SPOP. Biochem. Biophys. Res. Commun. 320, 760–765 (2004).

281. Byun, B. & Jung, Y. Repression of Transcriptional Activity of Estrogen Receptor α by a Cullin3/SPOP Ubiquitin E3 Ligase Complex. Mol. Cells 25, 289–293 (2007).

282. Zhu, H. et al. SPOP E3 Ubiquitin Ligase Adaptor Promotes Cellular Senescence by Degrading the SENP7 deSUMOylase. Cell Rep. 13, 1183–1193 (2015).

283. Zhu, K. et al. SPOP-containing complex regulates SETD2 stability and H3K36me3-coupled alternative splicing. Nucleic Acids Res. 45, 1–14 (2017).

284. Hernández-Muñoz, I. et al. Stable X chromosome inactivation involves the PRC1 Polycomb complex and requires histone MACROH2A1 and the CULLIN3/SPOP ubiquitin E3 ligase. Proc. Natl. Acad. Sci. U. S. A. 102, 7635–40 (2005).

285. Liu, J. et al. Analysis of Drosophila segmentation network identifies a JNK pathway factor overexpressed in kidney cancer. Science 323, 1218–22 (2009).

286. Schwend, T. et al. Stabilization of speckle-type POZ protein (Spop) by daz interacting protein 1 (Dzip1) is essential for gli turnover and the proper output of hedgehog signaling. J. Biol. Chem. 288, 32809–32820 (2013).

287. Kim, B. et al. Breast cancer metastasis suppressor 1 (BRMS1) is destabilized by the Cul3-SPOP E3 ubiquitin ligase complex. Biochem. Biophys. Res. Commun. 415, 720–726 (2011).

288. Wen, X. et al. Kinetics of hedgehog-dependent full-length Gli3 accumulation in primary cilia and subsequent degradation. Mol. Cell. Biol. 30, 1910–22 (2010).

289. Wang, C., Pan, Y. & Wang, B. Suppressor of fused and Spop regulate the stability, processing and function of Gli2 and Gli3 full-length activators but not their repressors. Development 137, 2001–9 (2010).

290. Liu, J., Heydeck, W., Zeng, H. & Liu, A. Dual function of suppressor of fused in Hh pathway activation and mouse spinal cord patterning. Dev. Biol. 362, 141–53 (2012).

174

291. Gschweitl, M. et al. A SPOPL/cullin-3 ubiquitin ligase complex regulates endocytic trafficking by targeting EPS15 at endosomes. Elife 5, 1–26 (2016).

292. Büscher, D., Grotewold, L. & Rüther, U. The Xt(J) allele generates a Gli3 fusion transcript. Mamm. Genome 9, 676–678 (1998).

293. Hui, C. C. & Joyner, a L. A mouse model of greig cephalopolysyndactyly syndrome: the extra- toesJ mutation contains an intragenic deletion of the Gli3 gene. Nat. Genet. 3, 241–246 (1993).

294. Maynard, T. M., Jain, M. D., Balmer, C. W. & LaMantia, A. S. High-resolution mapping of the Gli3 mutation extra-toesJ reveals a 51.5-kb deletion. Mamm. Genome 13, 58–61 (2002).

295. Svärd, J. et al. Genetic elimination of suppressor of fused reveals an essential repressor function in the mammalian hedgehog signaling pathway. Dev. Cell 10, 187–197 (2006).

296. Oro, A. E. & Higgins, K. Hair cycle regulation of Hedgehog signal reception. Dev. Biol. 255, 238– 248 (2003).

297. Rodríguez, C. I. et al. High-efficiency deleter mice show that FLPe is an alternative to Cre-loxP. Nat. Genet. 25, 139–140 (2000).

298. Logan, M. et al. Expression of Cre Recombinase in the developing mouse limb bud driven by a Prxl enhancer. Genesis 33, 77–80 (2002).

299. Lakso, M. et al. Efficient in vivo manipulation of mouse genomic sequences at the zygote stage. Proc. Natl. Acad. Sci. U. S. A. 93, 5860–5 (1996).

300. Srinivas, S. et al. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol 1, 4 (2001).

301. Zambrowicz, B. P. et al. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc. Natl. Acad. Sci. U. S. A. 94, 3789–94 (1997).

302. Ma, W., Rogers, K., Zbar, B. & Schmidt, L. Effects of different fixatives on beta-galactosidase activity. J. Histochem. Cytochem. 50, 1421–4 (2002).

303. Kiernan, J. A. Histological and Histochemical Methods - Theory and practice. (Cold Spring Harbor Laboratory Press, 2008).

304. Hui, C.-C., Slusarski, D., Platt, K. A., Holmgren, R. & Joyner, A. L. Expression of three mouse homologs of the Drosophila segment polarity gene cubitus interruptus, Gli, Gli-2, and Gli-3, in ectoderm- and mesoderm-derived tissues suggests multiple roles during postimplantation development. Dev. Biol. 162, 402–413 (1994).

305. Walterhouse, D. et al. Gli, a Zinc Finger Transcription Factor and Oncogene, Is Expressed During Normal Mouse Development. Dev. Dyn. 196, 91–102 (1993).

306. Elima, K. et al. The mouse collagen X gene : complete nucleotide and expression pattern structure. Biochem. J. 289, 247–253 (1993).

307. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A. L. & Karsenty, G. Osf2/Cbfa1: A Transcriptional

175

Activator of Osteoblast Differentiation. Cell 89, 747–754 (1997).

308. Dickinson, M. E. et al. Chromosomal localization of seven members of the murine TGF-beta superfamily suggests close linkage to several morphogenetic mutant loci. Genomics 6, 505–520 (1990).

309. Seki, K. & Hata, A. Indian Hedgehog Gene Is a Target of the Bone Morphogenetic Protein Signaling Pathway. J. Biol. Chem. 279, 18544–18549 (2004).

310. Mamillapalli, R., VanHouten, J., Zawalich, W. & Wysolmerski, J. Switching of G-protein usage by the calcium-sensing receptor reverses its effect on parathyroid hormone-related protein secretion in normal versus malignant breast cells. J. Biol. Chem. 283, 24435–24447 (2008).

311. Wei, J., Sheng, X., Feng, D., McGrath, B. & Cavener, D. R. PERK is essential for neonatal skeletal development to regulate osteoblast proliferation and differentiation. J. Cell. Physiol. 217, 693–707 (2008).

312. Hikata, T. et al. PIAS3 negatively regulates RANKL-mediated osteoclastogenesis directly in osteoclast precursors and indirectly via osteoblasts. Blood 113, 2202–2212 (2009).

313. Burcelin, R. et al. Impaired Glucose Homeostasis in Mice Lacking the α 1b-Adrenergic Receptor Subtype. J. Biol. Chem. 279, 1108–1115 (2004).

314. Bouxsein, M. L. et al. Guidelines for assessment of bone microstructure in rodents using micro- computed tomography. J. Bone Miner. Res. 25, 1468–86 (2010).

315. Cai, H. & Liu, A. Spop promotes skeletal development and homeostasis by positively regulating Ihh signaling. Proc. Natl. Acad. Sci. (2016). doi:10.1073/pnas.1612520114

316. (US), O. of the S. G. Bone Health and Osteoporosis: A Report of the Surgeon General. Reports Surg. Gen. (2004).

317. Wang, B., Fallon, J. F. & Beachy, P. A. Hedgehog-Regulated Processing of Gli3 Produces an Anterior/Posterior Repressor Gradient in the Developing Vertebrate Limb. Cell 100, 423–434 (2000).

318. Pan, Y. & Wang, B. A novel protein-processing domain in Gli2 and Gli3 differentially blocks complete protein degradation by the proteasome. J. Biol. Chem. 282, 10846–10852 (2007).

319. Miao, D. et al. Impaired endochondral bone development and osteopenia in Gli2-deficient mice. Exp. Cell Res. 294, 210–22 (2004).

320. Karp, S. J. et al. Indian hedgehog coordinates endochondral bone growth and morphogenesis via parathyroid hormone related protein dependent and independent pathways. Development 127, 543–548 (2000).

321. Stanford, W. L., Cohn, J. B. & Cordes, S. P. Gene-trap mutagenesis: past, present and beyond. Nat. Rev. Genet. 2, 756–68 (2001).

322. Bandyopadhyay, A. et al. Genetic analysis of the roles of BMP2, BMP4, and BMP7 in limb patterning and skeletogenesis. PLoS Genet. 2, 2116–2130 (2006).

176

323. Seong, K.-H. et al. Inhibition of the nuclear import of cubitus interruptus by roadkill in the presence of strong hedgehog signal. PLoS One 5, e15365 (2010).

324. Briscoe, J. & Thérond, P. P. The mechanisms of Hedgehog signalling and its roles in development and disease. Nat. Rev. Mol. Cell Biol. 14, 416–429 (2013).

325. Cohen, M., Briscoe, J. & Blassberg, R. Morphogen Interpretation: The Transcriptional Logic Of Neural Tube Patterning. Curr. Opin. Genet. Dev. 23, 423–428 (2013).

326. Ribes, V. et al. Distinct Sonic Hedgehog signaling dynamics specify floor plate and ventral neuronal progenitors in the vertebrate neural tube. Genes Dev. 24, 1186–1200 (2010).

327. Gurdziel, K. et al. Identification and validation of novel hedgehog-responsive enhancers predicted by computational analysis of Ci/Gli binding site density. PLoS One 10, 1–24 (2015).

328. Stenman, J. M. et al. Canonical Wnt Signaling Regulates Organ-Specific Assembly and Differentiation of CNS Vasculature. Science (80-. ). 322, 1247–1250 (2008).

329. Kawakami, Y. et al. WNT signals control FGF-dependent limb initiation and AER induction in the chick embryo. Cell 104, 891–900 (2001).

330. Cygan, J., Johnson, R. & McMahon, A. Novel regulatory interactions revealed by studies of murine limb pattern in Wnt-7a and En-1 mutants. Development 124, 5021–32 (1997).

331. Cao, T., Wang, C., Yang, M., Wu, C. & Wang, B. Mouse limbs expressing only the Gli3 repressor resemble those of Sonic hedgehog mutants. Dev. Biol. 379, 221–228 (2013).

332. Jiang, J. Regulation of Hh/Gli signaling by dual ubiquitin pathways. Cell Cycle 5, 390–393 (2006).

333. Zhang, Z. et al. Ter94 ATPase Complex Targets K11-Linked Ubiquitinated Ci to Proteasomes for Partial Degradation. Dev. Cell 155, 1–9 (2013).

334. Turgeon, B. & Meloche, S. Interpreting neonatal lethal phenotypes in mouse mutants: insights into gene function and human diseases. Physiol. Rev. 89, 1–26 (2009).

335. Satokata, I. & Maas, R. Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nat. Genet. 6, 348–356 (1994).

336. Proetzel, G. et al. Transforming growth factor-beta 3 is required for secondary palate fusion. Nat. Genet. 11, 409–14 (1995).

337. Kaartinen, V. et al. Abnormal lung development and cleft palate in mice lacking TGF-beta 3 indicates defects of epithelial-mesenchymal interaction. Nat. Genet. 11, 415–421 (1995).

338. Mo, R. et al. Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development. Development 124, 113–123 (1997).

339. McDonald, F. J. et al. Disruption of the beta subunit of the epithelial Na+ channel in mice: hyperkalemia and neonatal death associated with a pseudohypoaldosteronism phenotype. Proc. Natl. Acad. Sci. U. S. A. 96, 1727–31 (1999).

340. Yu, J., Carroll, T. J. & McMahon, A. P. Sonic hedgehog regulates proliferation and differentiation

177

of mesenchymal cells in the mouse metanephric kidney. Development 129, 5301–5312 (2002).

341. Hoover, A. N. et al. C2cd3 is required for cilia formation and Hedgehog signaling in mouse. Development 135, 4049–4058 (2008).

342. Coggins, K. G. et al. Metabolism of PGE2 by prostaglandin dehydrogenase is essential for remodeling the ductus arteriosus. Nat. Med. 8, 91–92 (2002).

343. Chen, X. et al. Knockout of SRC-1 and SRC-3 in mice decreases cardiomyocyte proliferation and causes a noncompaction cardiomyopathy phenotype. Int. J. Biol. Sci. 11, 1056–1072 (2015).

344. Xu, J. et al. The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc. Natl. Acad. Sci. U. S. A. 97, 6379–84 (2000).

345. Segi, E. et al. Patent ductus arteriosus and neonatal death in prostaglandin receptor EP4-deficient mice. Biochem. Biophys. Res. Commun. 246, 7–12 (1998).

346. Nguyen, M. et al. The prostaglandin receptor EP4 triggers remodelling of the cardiovascular system at birth. Nature 390, 78–81 (1997).

347. Scheuner, D. et al. Translational control is required for the unfolded protein response and in vivo glucose homeostasis. Mol. Cell 7, 1165–76 (2001).

348. Dai, X., Ma, W., He, X. & Jha, R. K. Review of therapeutic strategies for osteosarcoma, chondrosarcoma, and Ewing’s sarcoma. Med. Sci. Monit. 17, RA177-190 (2011).

349. American Cancer Society. Cancer Facts & Figures 2016. Cancer Facts Fig. 2016 1–9 (2016). doi:10.1097/01.NNR.0000289503.22414.79

350. Tirode, F. et al. Genomic landscape of ewing sarcoma defines an aggressive subtype with co- association of STAG2 and TP53 mutations. Cancer Discov. 4, 1342–1353 (2014).

351. Lo, W. W., Pinnaduwage, D., Gokgoz, N., Wunder, J. S. & Andrulis, I. L. Aberrant hedgehog signaling and clinical outcome in osteosarcoma. Sarcoma 2014, 261804 (2014).

352. Kumar, R. M. R. & Fuchs, B. Hedgehog signaling inhibitors as anti-cancer agents in osteosarcoma. Cancers (Basel). 7, 784–794 (2015).

353. Tiet, T. D. et al. Constitutive hedgehog signaling in chondrosarcoma up-regulates tumor cell proliferation. Am. J. Pathol. 168, 321–330 (2006).

354. Smith, R. et al. Expression profiling of EWS/FLI identifies NKX2.2 as a critical target gene in Ewing’s sarcoma. Cancer Cell 9, 405–416 (2006).

355. Crompton, B. D. et al. The genomic landscape of pediatric Ewing sarcoma. Cancer Discov. 4, 1326–1341 (2014).

356. Radivojac, P. et al. Identification, analysis, and prediction of protein ubiquitination sites. Proteins 78, 365–80 (2010).

357. Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the

178

cBioPortal. Sci. Signal. 6, pl1 (2013).

358. Cerami, E. et al. The cBio Cancer Genomics Portal: An open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2, 401–404 (2012).

359. Huang, C.-H. et al. UbiSite: incorporating two-layered machine learning method with substrate motifs to predict ubiquitin-conjugation site on lysines. BMC Syst. Biol. 10, 6 (2016).

179

VITA

Hongchen Cai

Education: Aug 2011 — present The Pennsylvania State University, Pennsylvania, USA Ph.D. candidate, Molecular, Cellular and Integrative Biosciences. Advisor: Dr. Aimin Liu. Sep 2008 — Jul 2011 Tsinghua University, Beijing, China Master of Science, Biology. Advisor: Dr. Yong-Bin Yan. Aug 2004 — Jul 2008 Tsinghua University, Beijing, China Bachelor of Science, Biological Science.

Teaching experience: Spring 2013 — 2016 Biol 240W: Function and Development of Organisms

Grant and Award: Aug 2014 Huck Graduate Dissertation Research Award (primary applicant, mini-grant, $5000) May 2015 Center for Molecular Investigation of Neurological Disorders research presentation prize

Formal Meeting Presentations: Aug 2016 Society for Developmental Biology 75th Annual Meeting Poster: Spop promotes skeletal development by degrading Gli3 repressor

Publications: [1] Hongchen Cai and Aimin Liu. Spop specifically regulates Gli3 activity and Shh signaling in dorsoventral patterning of the mouse neural tube. In review. [2] Hongchen Cai and Aimin Liu (2016) Spop promotes skeletal development and homeostasis by positively regulating Ihh signaling. PNAS 113(51):14751–14756. [3] Xiao-Qiao Li *, Hong-Chen Cai *, Shi-Yi Zhou *, Ju-Hua Yang *, Yi-Bo Xi *, Xiao-Bo Gao, Wei-Jie Zhao, Peng Li, Guang-Yu Zhao, Yi Tong, Fan-Chen Bao, Yan Ma, Sha Wang, Yong-Bin Yan, Cai-Ling Lu and Xu Ma (2012) A novel mutation impairing the tertiary structure and stability of γC-crystallin (CRYGC) leads to cataract formation in humans and zebrafish lens. Human Mutation 33(2):391-401. (* These authors contributed equally to this work.) [4] Wang Zhang *, Hong-Chen Cai *, Fei-Feng Li *, Yi-Bo Xi, Xu Ma and Yong-Bin Yan (2011) The congenital cataract-linked G61C mutation destabilizes γD-crystallin and promotes non-native aggregation. PLoS ONE 6(5):e20564. [5] Binbin Wang, Changhong Yu，Yi-Bo Xi, Hong-Chen Cai, Jing Wang, Sirui Zhou, Shiyi Zhou, Yi Wu, Yong- Bin Yan, Xu M and Lixin Xie (2010) A novel CRYGD mutation (p. Trp43Arg) causing autosomal dominant congenital cataract in a Chinese family. Human Mutation 32(1):E1939-47. [6] Yanping Ding, Yujie Huang, Nan Song, Xiaobin Gao, Shaopeng Yuan, Xiaofeng Wang, Hongchen Cai, Yan Fu, and Yongzhang Luo (2010) NFAT1 Mediates Placental Growth Factor-Induced Myelomonocytic Cell Recruitment via the Induction of TNF-alpha. Journal of Immunology 184(5):2593-601. [7] Yujie Huang, Nan Song, Yanping Ding, Shaopeng Yuan, Xuhui Li, Hongchen Cai, Hubing Shi and Yongzhang Luo (2009) Pulmonary Vascular Destabilization in the Premetastatic Phase Facilitates Lung Metastasis. Cancer Research 69(19):7529-37.