Assembly and Disassemly of the Wnt Beta-Catenin Destruction Complex

ASSEMBLY AND DISASSEMLY OF THE WNT BETA-CATENIN DESTRUCTION COMPLEX

Kristina Schaefer

A dissertation submitted to the faculty at the University of North Carolina at Chapel Hill in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Curriculum of Genetics and Molecular biology in the School of Medicine.

Chapel Hill 2018

Approved by:

Mark Peifer

Robert Duronio

Amy Shaub Maddox

Kevin Slep

Benjamin Major

ABSTRACT

Kristina N. Schaefer: Assembling and Disassembling the Wnt Beta-catenin destruction complex (Under the direction of Mark Peifer)

Wnt signaling provides a key example for cell-cell signaling pathways that regulate embryonic development and stem cell homeostasis and then are frequently inappropriately activated in cancers. Wnt signaling acts by regulating levels of β-catenin (βcat), an essential transcriptional coactivator of Wnt target genes. The tumor suppressors APC and Axin, along with the kinases GSK3 and CK1, form the core of the multiprotein destruction complex (DC), which targets βcat for phosphorylation, ubiquitination and destruction. In the presence of Wnt ligands, DC function is down-regulated, allowing levels of βcat levels to rise, eventually entering the nucleus to activate transcription of Wnt target genes. Based on earlier work, we hypothesize that the DC is a supramolecular entity that self-assembles by Axin and APC polymerization, and that regulation of assembly and stability of the DC regulates its function. We tested this hypothesis in Drosophila embryos by combining biochemistry, genetic tools to manipulate Axin and APC2 levels, advanced imaging, and molecule counting. By expressing Axin:GFP at near endogenous levels we found that in the absence of Wnt signals, Axin and APC2 co-assemble into large cytoplasmic complexes containing tens to hundreds of Axin proteins. Wnt signals trigger recruitment of these puncta to the membrane, a decrease in the number of Axin molecules, while cytoplasmic Axin levels increase, suggesting altered assembly. GSK3 regulates DC recruitment to the membrane and the release of Armadillo/βcat from the DC.

Manipulating Axin or APC2 levels had no effect on DC activity when Wnt signals were absent, but, surprisingly, had opposite effects on the DC when Wnt signals were present. Elevating Axin made the complex more resistant to inactivation, while elevating APC2 levels enhanced

iii inactivation. We also found that endogenous Axin and APC2 proteins and their antagonist

Dishevelled accumulate at roughly similar levels. Hetero-polymerization between Dishevelled and Axin via their DIX domains is essential for Wnt down regulation of the DC, yet the mechanism of regulation is unknown. Our data suggest both absolute levels and the ratio of these three core components affect destruction complex function, supporting models in which competition among Axin partners determines complex activity.

PREFACE

“One, remember to look up at the stars and not down at your feet. Two, never give up work.

Work gives you meaning and purpose and life is empty without it. Three, if you are lucky enough

to find love, remember it is there and don't throw it away.”

Stephan Hawking

We were invited by Developmental Cell to write a review on the current view of Wnt signaling. Therefore, instead of a typical introduction chapter, we wrote a review. This review centered around our hypothesis that the destruction complex is really a biomolecular condensate. We discuss recent research in the field and describe how several recent reports support our hypothesis. This chapter was a collaboration between Mark and I. We have now submitted the manuscript.

Chapter 2 describes the majority of my research. This chapter was published in PLoS

Genetics in 2018 (PLoS Genet 14(4): e1007339). Here we investigated the mechanistic role of

Axin, APC2, GSK3, and Dsh in regulating destruction complex both in the presence and absence of Wnt signaling. We utilized Drosophila genetics, biochemistry, and high-resolution microscopy. This work was completed under the supervision of Mark Peifer. Teresa Bonello analyzed protein levels of Axin, APC, and Dsh and critical reviewed the manuscript. Shiping

Zhang aided in visualization of Axin:GFP and analyzing Arm levels, and statistical analyses.

Clara Williams analyzed Engrailed expression, and helped with determining embryonic lethality, terminal cuticle phenotypes, and prepping embryos for imaging analysis. Dave Roberts imaged and analyzed Arm levels in APC2 mutants. Dan McKay analyzed RNAseq data. I helped organize the team of authors on this paper and performed all other experiments. The manuscript was written by me and Mark Peifer with input from the other authors.

Chapter 3 describes my work on how beta-catenin is transferred from the destruction complex to the E3 ligase. The groundwork of this project was started when I was a rotation student, and many tools utilized in this project were created by an undergraduate honors research student Lauren Bauer, who I supervised. At the time of this preface, we still finishing up a few experiments and plan to include this data as part of a manuscript we hope to submit later this year. Mira Pronobis conducted and analyzed the FRAP data. I performed the biochemistry experiments, acquired the confocal and super resolution images.

ACKNOWLEDGEMENTS

Before I even entered graduated school, I was given a series of article about how difficult it would be, and then I was questioned whether I really wanted to purse my Ph.D. Throughout my graduate school career I have regretted this decision and have been elated I went back to school. I want to thank those who helped me back on the path of science and discovery.

1) To Dr. Xin Zhang, without him I would have never found my love for research or

decided to attend graduate school in the first place.

2) To Mark Peifer who pushed me to be a greater thinker, to make a better argument,

and to be healthily skeptical about my own data.

3) To my other lab members and peers - thanks for being my sounding board, my late-

night confidant, for helping me stay the course, and for getting me out the lab and doing

something fun every once and a while.

4) Thanks to my family who have always supported me in any they could. I would not

have made it through without you guys.

5) Lastly to Nathan, thank you for being my practice audience, for believing in me when I

was unable, and for supporting my crazy career choice knowing what we would have to

sacrifice to accomplish it.

vii

TABLE OF CONTENTS LIST OF FIGURES ...... xii LIST OF TABLES...... xiv CHAPTER 1: EMERGING IDEAS IN REGULATING WNT SIGNALING: REGULATION BY POLYMERIZATION AND PARALLEL BEHAVIOR WITH BIOMOLECUAR CONDENSATES .. 1 OVERVIEW: ...... 1 The textbook model of Wnt signaling ...... 6 The Wnt-regulatory destruction complex—is it a biomolecular condensate? ...... 8 The destruction complex is an internally ordered structure that assembles by polymerization ...... 10 A functional destruction complex contains many more than four proteins ...... 13

destruction ...... 15 ...... 17 Other conserved sequences in APC’s intrinsically disordered region also play key functions...... 18 APC may play additional positive and negative roles in Wnt signaling ...... 19 Regulating a biomolecular condensate: Wnt signaling changes destruction complex localization and assembly ...... 20 Wnt signaling causes a switch in the destruction complex mix, destabilizing it ...... 21 One consequence of supermolecular assembly: the kinase GSK3 plays both positive and negative roles in the destruction complex via its access to many targets ...... 23 Axin post-translational regulation plays complex roles ...... 24 APC mutations in colorectal cancer target specific aspects of destruction complex function ...... 27 The destruction complex is a multifunctional machine with other targets including the cytoskeleton ...... 28 REFERENCES ...... 33 CHAPTER 2: SUPRAMOLECULAR ASSEMBLY OF THE BETA-CATENIN DESTRUCTION COMPLEX AND THE EFFECT OF WNT SIGNALING ON ITS LOCALIZATION, MOLECULAR SIZE, AND ACTIVITY IN VIVO ...... 46

viii

OVERVIEW: ...... 46 Author Summary ...... 46 INTRODUCTION ...... 47 RESULTS ...... 52 axin and APC1/APC2 are transcribed at similar levels ...... 52 Axin and APC2 proteins accumulate at similar levels during early-mid embryogenesis ...... 54 -fold, it inhibits Wg-regulated cell fate choice during embryogenesis ...... 59 Elevating Axin levels has no effect on Arm levels in cells not receiving Wg signals, but does render the destruction complex more resistant to inactivation by physiological levels of Wg signaling ...... 61 Levels of APC2 can be substantially elevated without significantly affecting viability or Wg-regulated cell fates ...... 68 Elevating levels of APC2 strongly promotes downregulation of the destruction complex in response to physiological levels of Wg signaling ...... 68 Simultaneously elevating levels of both APC2 and Axin inhibits Wg signaling more than elevating levels of Axin alone ...... 71 The relative ratio of APC2:Axin levels determines the effectiveness of Arm destruction ...... 73 Axin assembles into cytoplasmic multiprotein destruction complexes, and Wnt/Wg signaling leads to their membrane-recruitment and elevates levels of cytoplasmic Axin ...... 77 Wg signaling and GSK3 activity are each required for membrane recruitment of Axin puncta ...... 86 Simultaneously elevating Axin and APC2 makes destruction complex puncta more resistant to disassembly by Wg signaling ...... 89 Each destruction complex punctum includes tens to hundreds of APC2 or Axin proteins ...... 90 Dsh accumulates at levels similar to those of APC2 and Axin and localizes to Axin puncta in cells that receive Wg signals ...... 94 DISCUSSION ...... 98 In vivo levels of APC2 and Axin are similar rather than orders of magnitude different ...... 98 In the absence of Wg signaling, Axin assembles into large cytoplasmic multiprotein complexes that each contain tens to hundreds of Axin proteins ...... 99

Wg signaling triggers membrane recruitment of Axin and may destabilize destruction complex assembly ...... 100 Elevating Axin levels renders the destruction complex less sensitive to inactivation by Wg signaling ...... 102 APC2 is not rate-limiting for destruction complex activity but elevating its levels facilitates destruction complex inactivation ...... 103 Effects of altering the Axin:APC2 ratio suggest APC2 can play both positive and negative roles in Wnt regulation ...... 103 A proposed model of how Wnt signaling regulates destruction complex assembly and function ...... 106 MATERIALS AND METHODS ...... 107 Fly stocks, embryonic lethality, and cuticles ...... 107 Immunostaining and antibodies ...... 108 Assessing effects on Engrailed expression ...... 109 Quantitative analysis of Arm accumulation ...... 109 Statistics ...... 111 Immunoblotting ...... 111 RNA-Seq ...... 112 Cell culture and transfections ...... 112 Yeast fluorescence comparison ...... 112 REFERENCES: ...... 122 CHAPTER 3: THE DANCE BETWEEN THE DESTRUCTION COMPLEX AND THE E3 LIGASE ...... 128 OVERVIEW: ...... 128 INTRODUCTION: ...... 128 1. Recruitment ...... 129 2.Phosphorylation ...... 130 3. Ubiquitination ...... 130 4.Degradation ...... 131 RESULTS ...... 133 A system to examine whether the destruction complex and the E3 ligase co-localize ...... 133 Axin and not APC2 can recruit Slimb into the destruction complex...... 134 The RGS domain of Axin is required for efficient Slimb recruitment ...... 135 Slimb is a dynamic component of the destruction complex ...... 139

Slimb localizes along Axin cables ...... 141 DISCUSSION ...... 141 Defining the DC and E3 ligase interaction...... 141 The APC2:Axin complex recruits the F-box E3 adaptor Slimb ...... 143 APC2 and SCFSlimb don’t mix, but Axin does ...... 145 The RGS domain of Axin is necessary to efficiently recruit Slimb ...... 145 occurs in a supra-molecular factory ...... 147 METHODS: ...... 149 Cell Culture and transfection ...... 149 Immunofluorescence and Microscopy ...... 149 Immunoprecipitation and Western blotting ...... 149 REFERENCES ...... 151 CHAPTER 4: DISCUSSION ...... 155 And the debate goes on: Wnt regulation of the destruction complex ...... 155 The basics: Wnt signaling pathway simplified ...... 155 Size matters: Why is polymerization necessary for complex function? ...... 157 Condensates: the more the merrier ...... 157 Condensates: A way to bring multiple complexes together? ...... 160 Multiple mechanisms of turing down of the destruction complex ...... 161 Axin degradation or disassembly: control the scaffold, control the complex? ...... 162 Does a change in scenery result in a change in function? ...... 163 The destruction complex loses a friend...... 165 Dishevelled: The regulator of Axin puncta ...... 167 Revised Model: It’s all about who your friends are at the time ...... 171 Putting all the pieces together ...... 171 Shifting friends ...... 172 REFERENCES ...... 174

LIST OF FIGURES

Figure 1.1 - Properties of a biomolecular condensate….....……………………………………...…4

Figure 1.2 - The Wnt signaling pathway and APC, Axin and Dvl proteins contain properties of proteins found in condensates………...... ……………..………...... 7

Figure 1.3 - Axin and Axin plus APC accumulate in non-membrane bound puncta…...... …….…...... ….11

Figure 1.4 - In vivo recruitment of APC2 into Axin:GFP puncta…………………….………....…..22

Figure 1.5 - A revised model of the destruction complex...... 30

Figure 2.1 - Endogenous APC2 and Axin proteins accumulate at similar levels…….…...... ……...... 53

Figure 2.2 - Crosses used to achieve different level and timing of Axin elevation...... 56

Figure 2.3 - Developing tools to differentially elevate levels of Axin:GFP………...... …...... …....58

Figure 2.4 - Elevating Axin produces dose-sensitive inhibition of Wg signaling, while increasing APC2 levels does not……………….………………………….60

Figure 2.5 - Increasing Axin levels reduces the ability of endogenous Wg signaling to turn down the destruction complex but has little or no effect in Wg-Off cells…………………...... ……………...... …63

Figure 2.6 - Assessing gradation of Arm levels across the segment and absolute levels of Arm in Wg stripes and interstripes...... 66

Figure 2.7 - The opposite effects of Axin versus APC2 overexpression on Arm levels in Wg-ON cells are observed in both the cytoplasmic and membrane-associated pools...... 67

Figure 2.8 - Elevating APC2 levels increases the ability of endogenous Wg signaling to turn down the destruction complex, thus increasing Arm levels in cells receiving Wg…………...... ……...... …….70

Figure 2.9 - The relative ratios of APC2 to Axin levels determine effects on embryonic viability and Wg-regulated cell fates………………………………………...... 72

Figure 2.10 - Illustration of how embryos were sorted as to inferred genotype...... 75

Figure 2.11 - The relative ratios of APC2 to Axin levels determine effects on Arm destruction……...... ……………………………………………………...... ……….…..76

Figure 2.12 - Axin:GFP can largely restore normal Wnt signaling after

Axin RNAi...... 78

xii

Figure 2.13: Flag-tagged Axin assembles into puncta indistinguishable from those assembled by Axin:GFP...... 80

Figure 2.14 - Axin assembles into cytoplasmic multiprotein destruction complexes together with APC2, and Wg signaling leads to their membrane-recruitment and elevates levels of cytoplasmic Axin………..….…………..…82

Figure 2.15: When Axin is localized using an antibody to the GFP epitope-tag, it emphasizes the elevation in cytoplasmic Axin in Wg-ON cells and de-emphasizes Axin puncta in Wg-OFF cells...... 85

Figure 2.16 - Wg signal and GSK3/Zw3 activity are important for destruction complex membrane recruitment and GSK3/Zw3 regulates release of Arm from the destruction complex...... ……...... ……...87

Figure 2.17 - Ubiquitous expression of Wg increases embryonic lethality and induces a loss of denticle belts, whereas Dsh overexpression has little effect on viability and cuticle phenotype ...... 88

Figure 2.18 - The destruction complex contains thousands of APC2 or Axin molecules after over-expression in SW480 cells, and 10-100s of Axin molecules in vivo in embryos...... 92

Figure 2.19 - Dsh accumulates at similar levels to Axin and APC2, and co-localizes with Axin puncta in Wg-ON but not Wg-OFF cells…………………...... 95

Figure 3.1 - Axin recruits Slimb into cytoplasmic puncta…………...... ….136

Figure 3.2 - Axin is unable to recruit Cul1 or SkpA into the complex………...... ……138

Figure 3.3 - Slimb is not recruited by mito-APC2……………...... …..140

Figure 3.4 - The RGS domain of Axin is necessary for Slimb recruitment into Axin puncta…...... …142

Figure 3.5 - Slimb turnover in the destruction complex is unaffected by co-localization with Axin or Axin and APC2………………...... 144

Figure 3.6 - SIM imaging reveals that Slimb is recruited along Axin cables……………...... …...... ….148

xiii

LIST OF TABLES

Table 2.1 - Normalized densitometry values...... 114

Table 2.2 - Embryonic viability...... 115

Table 2.3 - Embryonic and first instar larva cuticle phenotype...... 116

Table 2.4 - Rows of En-expressing cells per segment...... 117

Table 2.5 - Effects on Arm levels of elevating Axin and/or APC2 levels...... 118

Table 2.6 - Quantification of the differences in Arm levels in Wg-stripe versus interstripe cells...... 119

Table 2.7 - Quantification of the different pools of Arm levels in Wg-stripe versus interstripe-cells...... 120

Table 2.8 - Fluorescence comparison values...... 121

xiv

CHAPTER 1: EMERGING IDEAS IN REGULATING WNT SIGNALING: REGULATION BY POLYMERIZATION AND PARALLEL BEHAVIOR WITH BIOMOLECUAR CONDENSATES

OVERVIEW:

Wnt/ βCatenin signaling plays key roles in cell fate decisions in embryonic and post embryonic development across the animal kingdom, and also helps maintain homeostasis in many tissues. As a result, loss- and gain-of-function mutations in the pathway are found in both developmental disorders and in many human cancers. In the absence of Wnt ligands, signaling is kept off by the multiprotein destruction complex, while pathway activation requires the destruction complex to be downregulated. Here we describe recent advances in the field that have provided new insights into the activity of the destruction complex and the mechanisms of its downregulation and point out parallels to other cell biological processes carried out by biomolecular condensates that form by phase separation.

The transformation of a fertilized egg into the body of an animal is among the most remarkable events in biology. Individual cells must choose fates based on their position, and then maintain those fates for a lifetime through tissue homeostasis. Cell-cell communication is critical for this, and a handful of cell-cell signaling pathways play especially important roles.

Among these is the Wnt pathway (Nusse and Clevers, 2017), which directs cell fates from the initial establishment of the vertebrate body axes to the detailed architecture of the kidney or nervous system. The key developmental roles of these pathways mean that mutations in pathway components lead to congenital diseases, which in the case of the Wnt pathway include

1 bone density and growth disorders (e.g. Robinow Disease) and progressive vision loss (Familial exudative vitreoretinopathy).

The same signaling pathways play critical roles in tissue homeostasis, maintaining proper

cell numbers by regulating tissue stem cell proliferation. To ensure signaling occurs only at the

right time and place, dedicated negative regulatory machinery has evolved to keep signaling

completely off in the absence of ligand. In the Wnt pathway this is accomplished by the

destruction complex, a multiprotein machine that targets the key Wnt-effector beta-catenin

(βCat) for phosphorylation, and ultimate ubiquitination and destruction. Mutations in destruction

complex proteins like Adenomatous polyposis coli (APC) occur in a wide variety of cancers and

play the initiating role in virtually all colorectal tumors. As a result, mechanisms by which Wnt

signaling is regulated are the subject of intensive research, potentially providing new cancer

therapies. Here we summarize current knowledge about Wnt signaling, framing it in the context

of the emerging idea that many key cellular processes are carried out in large non-membrane

bound cellular compartments, an idea we think provides new insights into destruction complex

function and regulation. Due to space limitations, we focus on canonical Wnt signaling, not its

variants, and on its core conserved components; other proteins with tissue- or animal phyla-

specific roles will be neglected, though they are important for a full picture of Wnt signaling and

its regulation (e.g. (Adler and Wallingford, 2017; Green et al., 2014; Malinauskas and Jones,

2014).

Centralized cellular boutiques: biomolecular condensates create cellular signaling and regulatory compartments The cell is a complex place. Like a city, within its boundaries hundreds of different activities

occur simultaneously at different places, from transcription to translation to metabolism to

transport to cell signaling. To organize this complexity, cells dedicate particular locations to

particular tasks. Some of this sequestration of activities is accomplished via membrane-bound

2 compartments, ranging from the ER or Golgi to the smallest exocytic vesicle. Within them contents are segregated from the bulk cytoplasm and interchange occurs via specialized transport systems. However, relying on specialized transport is insufficient to organize the vast volume of cytoplasm and nucleoplasm not encompassed within a membrane-bound organelle.

To solve this problem, cells evolved an additional mechanism of organizing cellular compartments that does not require membrane enclosure. Some of these structures were large enough to merit recognition by cell biology’s pioneers (Gall, 2000)—e.g., nucleoli or Cajal

bodies, locations of ribosome or spliceosome assembly within nuclei, or the germplasm of

animal eggs, containing determinants specifying germ cell fate.

In the past decade scientists recognized that these entities are examples of a much broader

group of non-membrane bound cellular compartments that organize specific proteins and RNAs.

They are key to diverse cellular processes including transcription, the DNA damage response,

and cellular signaling (Banani et al., 2017; Holehouse and Pappu, 2018). Pioneering work on

the C. elegans germline P granules and on signaling centers organized by SH3 domain proteins

led to the idea that these structures assemble by “liquid-liquid phase separation” (Brangwynne

et al., 2009; Li et al., 2012a). Multivalent interactions among their protein and/or RNA

constituents lead to self-assembly, creating compartments separated from the bulk cytoplasm

where the concentration of key players is exceptionally high, significantly speeding intricate

reactions and/or processes (reviewed in Banani et al., 2017). The field emerged from concepts

from soft-matter physics and polymer chemistry, which provide a biophysical basis and

theoretical framework for this behavior. Critically, molecules can freely diffuse within, into and

out of these structures, as they are not enclosed in a lipid bilayer and are often liquid-like in

nature. This allows them to serve as centralized functional hubs for particular cellular processes,

in which substrate molecules can enter, assemble, disassemble, or be modified, and products

Figure 1.1

Figure 1.1: Steps for building a biomolecular condensate A) Two separate proteins or RNAs (Green backbone and blue backbone) each contain multiple protein:protein or protein:RNA interaction sites sites (e.g. Orange, Purple, and Magenta circles and ovals) B) Homo and heteromeric interactions between molecules form multivalent linakges. C) These interactions induce condensate formation. D) Example of a biomolecular condensate in vitro. The droplets are formed from 8 uM Whi3 protein and 5 nM BNI1 RNA in 150 mM KCl after 4 hours incubation—picture provided Erin Langdon and Amy Gladfelter. E) Key properties of biomolecular condensates.

4 leave. They also serve as storage depots for key players to be deployed at later times.

Structures like these recently were given the broad name “biomolecular condensates”, reflecting the broad range of cellular and molecular processes that occur within them.

Condensates have a number of defining properties (Banani et al., 2017; Fig. 1.1), though precise definitions are still being established. Each is a non-membrane bounded structure ranging up to micron scale that concentrates proteins and/or RNAs at a particular cellular site. They assemble by multivalent interactions mediated by multidomain proteins and/or

RNAs with multiple protein or RNA interaction sites (Fig. 1.1). Many of the proteins involved

contain “intrinsically disordered regions (IDRs)” -- these lack tertiary structure, are often not

highly conserved in sequence, and self-interact or include within them interaction sites for other

proteins (Fig. 1.1A-B). IDRs are often tethered to folded domains (Mittal et al., 2018). Even after

phase separation, protein components freely diffuse into and out of the structures. Some

condensates can transition to a more gel-like state (Wang et al., 2018), with reduced exchange

with the bulk cytosol, a process that can contribute both to function and to pathogenesis. One

key to understanding assembly of condensates is the ability to reconstitute phase separation

behavior in vitro, with minimal components (Fig. 1.1D). Both in vitro and in vivo, liquid

condensates can fuse and relax to minimize surface tension. The rapidly expanding universe

of biological processes and structures encompassed under the biomolecular condensate

umbrella and the challenge of defining the rules governing their assembly, disassembly, and

function have made this one of the fastest growing areas of cell biology. As we’ll see below,

structures that regulate and transduce Wnt signals share many features with biomolecular

condensates.

The textbook model of Wnt signaling

Like most key signaling pathways regulating development, the primary output of the

canonical Wnt pathway is a change in the cell’s transcriptional program. This occurs by

regulating levels of βCat, a co-activator of transcription (reviewed in Gammons and Bienz, 2017;

Nusse and Clevers, 2017; Stamos and Weis, 2013). In the absence of Wnt signaling βCat levels

are kept low by the βCat destruction complex (Fig. 1.2A). At the core of this complex are the

tumor suppressor APC, the scaffold Axin, and two kinases, GSK3 and CK1. This complex

recruits βCat, where it is sequentially phosphorylated by CK1 and then GSK3. Once βCat is

phosphorylated, it is transferred to the Cullin-based E3 Ligase SCFβTrCP, polyubiqitinated, and then recognized by the proteasome and degraded. As a result, Wnt-regulated transcription is

OFF.

Wnt ligands bind both the 7 transmembrane Frizzled (Fz) and the single-pass transmembrane LRP5/6 receptors (Fig. 1.2B; reviewed in DeBruine et al., 2017; Nusse and

Clevers, 2017). The Wnt/Fz/LRP complex recruits the cytoplasmic protein Disheveled (Dvl in mammals, fly Dsh). GSK3 phosphorylates LRP5/6 (fly Arrow), creating a binding site for Axin and recruiting it to the membrane. This downregulates destruction complex activity. The primary mechanism of downregulation is not yet clear, as data support diverse mechanisms ranging from disassembly of the complex, inhibition of GSK3 kinase activity, Axin degradation, sequestration of destruction complex core proteins, and loss of E3-ligase interaction.

Destruction complex inhibition allows βCat levels to rise and it enters the nucleus, binding to T-

Cell Factor (TCF)/Lymphoid Enhancer factor (LEF) family DNA binding proteins. TCF/LEF proteins bind Wnt regulated genes, initiating different multiprotein complexes in cells where Wnt signaling is ON or OFF (reviewed by Gammons and Bienz, 2017). Thus the ultimate output of

Wnt signaling occurs when the TCF/LEF:βCat complex activates transcription of Wnt target genes.

Figure 1.2

Figure 1.2: Textbook model of Wnt signaling. A) In the absence of a Wnt ligand, the destruction complex (APC, Axin, CK1, GSK3) recruits βcat for phosphorylation. Once phosphorylated, βcat can be recognized by the E3 ligase, SCFβTrCP, ubiquitinated, and then passed to the proteasome for ultimate protein degradation. B) Wnt signaling induces the formation of the Wnt receptor complex of Wnt/Frizzled/LRP5/6/Dvl. This complex recruits Axin and induces down-regulation of the destruction complex. Levels of cytoplasmic βcat rise, allowing βcat to enter the nucleus, bind to TCF/LEF family of transcription factors, and co- activate transcription of Wnt target genes.

Although the field agrees on the main components of the destruction complex, key questions about its function and regulation remain unanswered. For example, what is the role of

APC within the destruction complex? The fact that APC is mutated in >80% of colon cancers emphasizes that it is essential for destruction complex function (Kandoth et al., 2013), yet what

APC does within the complex remained unclear. Second, Axin has a self-polymerizing domain,

and loss of this domain reduces destruction complex function. This suggests that the destruction

complex is larger than a simple hetero-tetramer of Axin, APC and the two kinases, but how does

the sum of these parts create and affect complex function? Another key question involves the

primary mechanism used to down regulate destruction complex function. A few prominent

theories were mentioned above, but which is the primary mechanism? New research is

shedding light on these and other areas, providing insights into the mechanisms of destruction

complex function and regulation.

The Wnt-regulatory destruction complex—is it a biomolecular condensate?

Looking back with hindsight at our unfolding understanding of the components, regulation

and function of the destruction complex reveals striking parallels between its properties and

many of those of biomolecular condensates. Two key non-enzymatic components, APC and

Axin, are complex multidomain scaffolding proteins containing folded domains that bind

other proteins along with long intrinsically disordered regions that contain binding sites for

other destruction complex proteins, including βCat (Fig. 1.3A, reviewed in Stamos and Weis,

2013). For example, human APC is predicted to have 50% disordered content (Piovesan et al.,

2018). Axin has an N-terminal Regulator of G-protein signaling (RGS) domain that binds APC, a

C-terminal DIX domain that mediates head-to-tail polymerization, and an intervening intrinsically

disordered region containing binding sites for βCat and the two key kinases, as well as for the

phosphatase PP2A (Behrens et al., 1998; Fagotto et al., 1999; Hart et al., 1998; Ikeda et al.,

1998; Kishida et al., 1998; Sakanaka et al., 1999; Sakanaka et al., 1998; Sakanaka and

Williams, 1999; Zeng et al., 1997; Fig.1.3A). Axin’s multiple binding sites allow it to bring βCat into proximity to the kinases GSK3 and CK1. APC has an conserved N-terminal region that mediates oligomerization (Kunttas-Tatli et al., 2014) and includes an Armadillo (Arm) repeat domain that can bind diverse partners. This is followed by a long intrinsically disordered region, embedded within which are multiple copies of two distinct types of binding sites for βCat (Fig.

3A), multiple binding sites for Axin, and other conserved sites for which the binding partners remain undetermined (reviewed in Stamos and Weis, 2013). The multivalent nature of APC and Axin and their intrinsically disordered regions are shared features with known components of biomolecular condensates, suggesting that they may also form a condensate.

The localization of key destruction complex players is also striking when considered through

the lens of phase separation. When Axin is expressed in many different cell types, both in vitro

and in vivo, it forms large protein “puncta”, and recruits into them APC and other

destruction complex proteins, thus increasing their effective local concentrations (Figs.

1.3 and 1.4, e.g. Cliffe et al., 2003; Fagotto et al., 1999; Faux et al., 2008; Thorvaldsen et al.,

2015). Recent analysis by correlative fluorescence and electron microcopy confirmed these puncta are not enclosed in a membrane (Thorvaldsen et al., 2015). When Axin is expressed at endogenous levels, puncta are also seen (Faux et al., 2008). As we discuss in detail below,

APC is required for puncta assembly in vivo, and puncta localization is regulated by Wnt signaling, consistent with puncta as active players in Wnt regulation (Cliffe et al., 2003;

Mendoza-Topaz et al., 2011; Schaefer et al., 2018). Puncta formation and destruction complex function depend at least in part on the ability of Axin’s C-terminal DIX domain to oligomerize

(Sakanaka and Williams, 1999). The DIX/DAX (Dishevelled/Axin; referred to as DIX below) domain was initially defined because it is conserved with Dvl, a positive effector of Wnt signaling. Dvl also forms puncta, both when expressed in cells (Axelrod et al., 1998; Yang-

Snyder et al., 1996) and in its endogenous state (Miller et al., 1999). Like Axin, Dvl puncta

9 formation also depends on its DIX domain (Schwarz-Romond et al., 2005). Strikingly, Dvl and

Axin physically interact and co-localize in puncta (Fig. 3A) (Fagotto et al., 1999; Julius et al.,

2000; Kishida et al., 1999). Dvl puncta are recruited to the membrane by the Wnt receptor

(Axelrod et al., 1998; Miller et al., 1999; Yang-Snyder et al., 1996). Early work suggested that

Dvl’s DIX domain associated with vesicles (Capelluto et al., 2002). However, subsequent work

failed to reveal any co-localization of Dvl with vesicular markers. Instead, live imaging revealed

that Dvl puncta are protein oligomers that can grow by fusion (Schwarz-Romond et al.,

2005). Puncta containing Axin and APC can also fuse (Kunttas-Tatli et al., 2014). FRAP

analysis further revealed that Dvl, Axin, and APC all freely diffuse into and out of puncta

(Pronobis et al., 2015; Schwarz-Romond et al., 2007b). Together, these data suggest that Dvl

and Axin puncta meet most of the criteria for biomolecular condensates (Fig. 3A) and

demonstrate that puncta assembly is key for destruction complex function and regulation.

The destruction complex is an internally ordered structure that assembles by polymerization

Some biomolecular condensates form via a network of multivalent interactions without a

strong underlying structural scaffold, while others can assemble into a more gel-like polymerized

state (Banani et al., 2017). Early studies of the destruction complex suggested it is more

structured, supporting a model of ‘signaling by reversible polymerization’ (Schwarz-Romond et

al., 2007a). The DIX domains of Dvl and Axin polymerize by head to tail interactions, forming

filaments that can be visualized by EM or X-ray crystallography (Schwarz-Romond et al.,

2007a), similar to tubulin, actin, and septins. Critically, mutations in their respective DIX

domains that block polymerization reduce Dvl’s ability to promote Wnt signaling and attenuate

Axin’s ability to inhibit Wnt signaling (Fiedler et al., 2011; Schwarz-Romond et al., 2007a).

These data suggested that destruction complex puncta have an internal structure conferred by

DIX domain polymerization. Evidence for this model was subsequently obtained using high

Figure 1.3

Figure 1.3: APC, Axin, and Dvl are multidomain proteins with intrinsically disordered regions that accumulate in structured non-membrane bound puncta. A) Cartoon of the structures of APC, Axin and Dvl highlighting domains mediating self-interaction as well as interaction sites with other proteins. Proteins found in condensates often have intrinsically disordered regions (black lines) and can polymerize and/or oligomerize other components in the condensates. Domains and motifs are as labelled. Solid green lines indicate direct interactions. Dotted orange line indicates identified interaction regions. Circled arrows = regions of self- interaction. B-D. Drosophila APC2 and Axin constructs expressed in SW480 cells. A) Close up of APC and Axin puncta visualized using standard confocal microscopy shows colocalization, but no underlining structure (Image originally published in Pronobis et al., 2015: DOI: 10.7554/eLife.08022) B-D) SIM Imaging. B) When Axin is expressed alone, many small puncta are formed. C) Close-up of a puncta from B, revealing that Axin polymers form knots. D) Co-

11 transfection of APC2 and Axin. They accumulate together in puncta. Note that puncta are larger and fewer than in B. E) SIM images of a punctum, revealing that APC2 and Axin each form intertwined cables with multiple potential interaction sites, thus revealing the internal structure of the destruction complex. (Image originally published in Pronobis et al., 2015.: DOI: 10.7554/eLife.08022)

resolution microscopy. Structured illumination super-resolution microscopy (SIM) of Drosophila

Axin and APC2 expressed in SW480 cells resolved the puncta into structured entities (Fig.

1.3B-F) (Pronobis et al., 2015). When expressed alone, Axin formed puncta with an internal structure that resembled toroids or knots, potentially representing DIX domain filaments (Fig.

1.3C-D). Co-expressing Axin and APC2 led to their co-recruitment in puncta and puncta assembly was enhanced, with the largest puncta on the order of micron size. Strikingly, intertwined homo-filaments of Axin and APC2 were resolved, further supporting the idea that regulated polymerization underlies destruction complex assembly (Fig. 1.3E-F). In parallel,

scientists examined endogenous Axin puncta stabilized by inhibiting the enzyme Tankyrase, a

known regulator of Axin levels, using both SIM and correlative fluorescence and electron

microscopy (Thorvaldsen et al., 2015). These data also revealed micron scale puncta in which

Axin, βCat, and Tankyrase formed an intermeshed network of filaments, which electron

microscopy verified are not membrane-bounded. Together, these data suggest that the

destruction complex has a structured scaffold, built around intertwined polymers of Axin and

APC (Fig. 1.5).

A functional destruction complex contains many more than four proteins

Early work in Xenopus oocytes suggested that destruction complex assembly is limited by

Axin’s very low protein abundance compared to all other destruction complex proteins—up to

5000-fold lower (Lee et al., 2003; Salic et al., 2000). This suggested that Axin was exquisitely

rate-limiting, a factor built into many mathematical models of signaling (e.g.Lee et al., 2003).

Interestingly, recent work in both flies and mammals suggest Axin levels are in fact quite similar

to those of APC (Kitazawa et al., 2017; Schaefer et al., 2018; Tan et al., 2012). These new data

will power updated models and perhaps new insights of Wnt signaling and its regulation.

For decades textbook models of the destruction complex represented it as a simple 4-

protein complex of APC:Axin:GSK:CK1 (Fig. 1A), despite the fact that we knew for years that

Axin polymerization is necessary for efficient βCat regulation. Defining the number of molecules in a functional destruction complex has been a challenge. Some destruction complex proteins localize to other locations where they have distinct roles --e.g., GSK3 regulates Wnt, Hedgehog,

Insulin, PI3K, and Erk signaling (Cormier and Woodgett, 2017) and APC regulates the

cytoskeleton (Nelson and Nathke, 2013). Thus not every molecule of GSK3 or APC in the cell

localizes to the destruction complex. Second, effective antibodies to key players were not

available. This was particularly true for Axin. To overcome the lack of antibodies to endogenous

Axin, investigators over-expressed Axin and/or APC, hoping the larger complexes formed would

serve as expanded representations of endogenous complexes. This provided important insights.

However, new reagents recently allowed scientists to look at Axin expressed at endogenous or

near endogenous levels, using either a new antibody against fly Axin (Wang et al., 2016a) or

epitope-tagged Axin expressed at near-endogenous levels (Schaefer et al., 2018; Wang et al.,

2016a; Yang et al., 2016). This revealed that in Wnt-OFF cells Axin assembles into puncta

similar to those that assemble after overexpression in cultured cells, and these puncta recruit

APC (Schaefer et al., 2018). This validates early work examining endogenous Axin in MDCK

cells (Faux et al., 2008) and supports the idea that the destruction complex is a supermolecular

machine containing tens to hundreds of molecules rather than a simple 4-protein complex. But

how large is this complex? The ability to express GFP-tagged Axin at near endogenous levels

provided insight. Fluorescence intensity measurements compared to GFP-labeled complexes of

known molecular composition revealed that in Drosophila embryos, active destruction complex

puncta contain on average ~260 Axin molecules (range ~60-930; Schaefer et al., 2018). Mass

spectroscopy provides an alternate mechanism of putting numbers on destruction complex

proteins—recent analyses suggest HEK293 cells each contain ~13,000 Axin proteins (Kitazawa

et al., 2017)—however, with current technology this assessment requires many simplifying

assumptions. Together, these data emphasize that the destruction complex is a supermolecular

14 machine, consistent with it being a form of biomolecular condensate.

Like many biomolecular condensates, the destruction complex assembles via many multivalent interactions. This makes it surprisingly robust to removal of some but not all protein interaction motifs. For example, individually deleting most of APC’s βCat binding sites, Axin’s

βCat binding site, or Axin’s RGS domain have only modest effects in vivo (Kremer et al., 2010;

Kunttas-Tatli et al., 2012; Peterson-Nedry et al., 2008; Roberts et al., 2011; Yamulla et al.,

2014). However, some binding sites are essential individually (Axin’s GSK3 binding site) or when deleted in concert (Axin∆RGS∆Arm; Peterson-Nedry et al., 2008). This powered synthetic biology approaches to design a “minimal βCat destruction machine” which retains function in colorectal cancer cells (Pronobis et al., 2017).

Stabilizing destruction complex supermolecular assembly is a key factor in βCat destruction

Axin’s DIX domain is necessary for its self-polymerization but how is polymerization regulated? Newly synthesized Axin molecules can either nucleate a new Axin filament or add to an existing polymer. Several studies focused on APC’s role in forming or stabilizing Axin filaments. APC is required for assembly of Axin puncta and therefore active destruction complexes (Mendoza-Topaz et al., 2011). Co-expression revealed that APC2 stabilizes Axin assembly, as measured by destruction complex volume, increased complexity of Axin filaments, and decreased Axin turnover (Pronobis et al., 2015). Strikingly, as APC co-expression increased the size of Axin puncta, it simultaneously decreased the number of puncta, consistent with the idea that APC2 promotes Axin addition to existing polymers over nucleation of new polymers (Pronobis et al., 2015).

When the destruction complex was visualized, both Axin and APC2 appeared as intertwined filaments, suggesting that APC may also polymerize (Fig. 3E-F). Human APC1 has an N- terminal coiled-coil oligomerization domain (Joslyn et al., 1993), but this is not conserved in

Drosophila family members. However, another N-terminal region of APC, the APC self-

association domain (ASAD), is conserved in Drosophila, Xenopus, and humans, mediates

Drosophila APC2 self-association (Kunttas-Tatli et al., 2014), and together with APC’s Arm

repeats can mediate puncta formation (Pronobis et al., 2015). The ASAD and adjacent Arm

repeats are required for destruction complex function (Kunttas-Tatli et al., 2014; McCartney et

al., 2006; Roberts et al., 2012a). Interestingly, loss of APC’s oligomerization domain eliminated

APC’s ability to stabilize Axin within the destruction complex (Kunttas-Tatli et al., 2014; Pronobis

et al., 2015). These data suggest that APC polymerization is required to initiate and stabilize

formation of functional destruction complexes.

APC’s stabilization of the destruction complex requires two different Axin:APC interactions

mediated by different domains (Fig.3A; Pronobis et al., 2015). The first is via the well-

established interaction of the Axin-RGS domain with APC’s SAMPS (Spink et al., 2000).

Retention of at least one SAMP is essential for APC function in both mice (Smits et al., 1999)

and flies (Roberts et al., 2011). The second APC:Axin interaction is between APC’s Arm repeats

and a less well-defined region in Axin’s central IDR? intrinsically disordered region (Pronobis et

al., 2015). Recent data suggest that not all SAMP motifs are functionally similar. Fly APC2 has 2

SAMPs, and data suggest one recruits Axin while the other aids in efficient βCat destruction by

an unknown mechanism (Kunttas-Tatli et al., 2015). Intriguingly, in colorectal cancer cells the

SAMP interaction is dispensable if APC and Axin are fused into a single minimal polypeptide

(Pronobis et al., 2017). The ability of APC to stabilize Axin in the destruction complex is further

enhanced by a bridging interaction involving βCat’s ability to bind both APC and Axin, an

interaction disrupted by the vertebrate-specific βCat binding protein ICAT (Ji et al., 2018).

Together, these data support the idea that one key role of APC in the destruction complex is to

stabilize Axin, increasing destruction complex size and thus its effectiveness.

The destruction complex serves a second role as a sink for cytoplasmic βCat

Another mystery with regard to APC is the role of its multiple βCat binding motifs (Fig. 3A).

Most APC proteins have multiple copies of two distinct types of binding sites for βCat embedded in the central intrinsically disordered region, the 15- and 20-amino acid repeats (15R and

20R;Eklof Spink et al., 2001; Ha et al., 2004; Liu et al., 2006; Xing et al., 2004). Each 20R has a different affinity for βCat, with an affinity range of 100-fold (Liu et al., 2006). Phosphorylation of

20Rs by GSK3 dramatically increases their affinity for βCat (Ha et al., 2004; Liu et al., 2006).

This led to the hypothesis that high affinity binding sites are needed when βCat levels are low and the others come into use when βCat levels are high, helping sequester βCat in the cytoplasm (Ha et al., 2004; Krieghoff et al., 2006). This model was tested in colorectal cancer cells and in Drosophila, by systematically deleting 15R and 20R βCat binding sites. Strikingly, the highest affinity βCat binding sites are dispensable in targeting βCat for destruction—instead the binding sites collaborate to fine-tune Wnt signals in an additive fashion by cytoplasmic retention of βCat, supporting the sequestration hypothesis (Kunttas-Tatli et al., 2012; Roberts et al., 2011; Yamulla et al., 2014). Interestingly, a fly APC2 mutant lacking all the 15R and 20Rs retains ability to restore APC function in colorectal cancer cells, although it is not is not fully functional in destruction in vivo in Drosophila (Yamulla et al., 2014). Axin also plays a role in cytoplasmic retention of βCat in Drosophila (Tolwinski and Wieschaus, 2001).The single βCat binding site in Axin’s intrinsically disordered region (Xing et al., 2003) may serve a redundant

role, as a designed synthetic minimal destruction complex containing essential regions of APC

and Axin that restored βCat regulation in colorectal cancer cells solely utilized Axin’s βCat

binding site (Pronobis et al., 2017). In vivo analysis of Drosophila APC2 mutants lacking βCat binding sites revealed an aspect of in vivo regulation that remains to be understood—rather than restoring the graded levels of βCat seen in wildtype, they led to a sharp ON/OFF transition

(Yamulla et al., 2014), reminiscent of the behavior of certain mutations in Drosophila βCat

(Orsulic and Peifer, 1996). This is consistent with some sort of threshold feedback response.

Other conserved sequences in APC’s intrinsically disordered region also play key functions

In addition to the βCat and Axin binding sites, APC’s intrinsically disordered region also

contains another highly conserved motif which did not have a known binding partner, variously

called conserved sequence B (B) or the catenin inhibitory domain (CID, Fig. 3A). Strikingly the

B/CID motif is essential for APC function in Wnt regulation in both flies and mammals (Kohler et

al., 2009; Roberts et al., 2011) and may be the sequence targeted for removal in the protein

truncations found in colorectal tumors (Kohler et al., 2009). Intriguingly, an immediately adjacent

motif, 20R2, which resembles other 20Rs but lacks key residues that mediate binding to βCat

(Kohler et al., 2008; Liu et al., 2006), is also essential for Wnt regulation. Together, B and 20R2

may form a binding site for an unidentified partner. Further examination revealed that 20R2/B

regulate one of the two APC:Axin binding interactions, that between APC’s Arm repeats and

Axin’s mid-region. The function of 20R2/B requires phosphorylation by GSK3 (Pronobis et al.,

2015). Together, these data led to a model in which phosphorylation of the B and 20R2 motifs triggers a conformational change in APC, releasing one of the two APC;Axin interactions and allowing transfer of phosphorylated βCat to the E3 ubiquitin ligase, as part of a catalytic cycle.

This model is consistent with other data, revealing that loss of GSK3 in fly embryos leads to

βCat accumulation in the destruction complex (Schaefer et al., 2018), and suggesting that inhibiting βCat release to the E3 ligase is a key step by which Wnt signaling inactivates the destruction complex (Li et al., 2012b). This model also helps explain a paradox in the field— colorectal cancer cells are defective in βCat destruction but not in βCat phosphorylation (Yang et al., 2006).

These data left open the identity of the interacting partner of conserved region B and/or

20R2. The same screen that identified βCat as an APC binding partner also identified α-catenin

(αcat; Rubinfeld et al., 1993), another component of cadherin-based cell-cell junctions, but the function of the APC:αcat interaction remained a mystery. In 2013 evidence emerged that αcat binds to the B/CID region, and assays in cultured cells supported the idea that αcat facilitates

βCat ubiquitination and proteolysis (Choi et al., 2013). These data further suggested that αcat binds APC via its VH1 domain and that αcat/βCat interaction is also critical for βCat destruction.

These data are intriguing, but the physiological role of αcat in Wnt regulation remains in question. In Drosophila neither zygotic αcat mutants (Desai et al., 2013; Sarpal et al., 2012) nor zygotic βCat mutants deleting the αcat binding site (Orsulic and Peifer, 1996) have defects in

Wnt signaling or its regulation. Mutations in αcat in C. elegans also do not cause obvious

defects in Wnt signaling (Costa et al., 1998). A role for αcat in transcriptional regulation of Wnt target genes has also been proposed, supported by mass spectrometry suggesting αcat forms a complex with TCF/LEF family members (Choi et al., 2013). Similar nuclear roles have been suggested for APC (e.g. Sierra et al., 2006), but sequestering APC at a variety of cytoplasmic locations does not disrupt regulation of βCat destruction in flies or mammalian cells (Roberts et al., 2012b) suggesting a nuclear role of APC is not essential. Continued work is needed to further clarify the relevant binding partner of B/20R2 and its function in destruction complex function.

APC may play additional positive and negative roles in Wnt signaling

Current data support roles for APC in stabilizing the destruction complex, promoting transfer of βCat to the E3 ligase, and sequestering βCat in the cytoplasm. However, these may not encompass its full range of functions. Recent work suggested that APC also acts at the level of the Wnt receptor, inhibiting baseline activity in the absence of Wnt ligands by promoting clathrin- dependent receptor endocytosis (Saito-Diaz et al., 2018)—however, this role is only exhibited by certain APC family members. Another intriguing hypothesis is that APC both inhibits and promotes Wnt signaling (Tacchelly-Benites et al., 2018; Takacs et al., 2008). The challenge in

19 defining a positive regulatory role of APC is in designing experiments that allow it to be distinguished from its essential negative regulatory role. Genetic studies in which the levels of both Drosophila APC proteins, APC1 and APC2, were manipulated in parallel provided the right system. For example, reducing APC2 function attenuated activated Wnt signaling in fly eyes induced by loss of APC1 (Takacs et al., 2008). The mechanism by which this occurs was not fully defined, though effects on Axin stability and phosphorylation were suggested (Tacchelly-

Benites et al., 2018; Takacs et al., 2008; Wang et al., 2016a). Interestingly, overexpressing

APC2 in Drosophila embryos enhanced accumulation of βCat only in Wnt-ON cells, supporting the idea that APC2 can aid in turning down of the destruction complex activity in response to

Wnt signals (Schaefer et al., 2018). These data further illustrate the intricate nature of Wnt regulation.

Regulating a biomolecular condensate: Wnt signaling changes destruction complex localization and assembly

The data above summarize our knowledge of active destruction complexes. Another challenge is to define how its function is down-regulated by Wnt signaling. Many of us initially spoke of turning the destruction complex “OFF”, but this is inaccurate. Wnt signaling does not fully inactivate the complex—it retains, at least initially, the ability to phosphorylate βCat (Kim et al., 2013; Li et al., 2012b), such that rate of βCat turnover is reduced but not halted (Hernandez et al., 2012). In retrospect, this was apparent in early work in Drosophila, as mutational inactivation of GSK3 or APC function led to much higher levels of βCat accumulation than those seen in cells receiving Wnt signals (e.g. Ahmed et al., 2002; Akong et al., 2002). How is destruction complex activity repressed? After Wnt ligands bind their receptors, Axin is recruited into a second protein complex with many properties of a biomolecular condensate, the

“signalasome” (reviewed in Gammons and Bienz, 2017). Axin recruitment to the Fz:LRP coreceptor is mediated both by direct interactions with the phosphorylated LRP5/6 tail (Davidson

20 et al., 2005; Mao et al., 2001; Tamai et al., 2004; Zeng et al., 2005) and by a less well defined

role of Dvl (Bilic et al., 2007; Cliffe et al., 2003). Consistent with this, Drosophila Axin (at

endogenous or near endogenous levels) is found in cytoplasmic puncta in the absence of Wnt

signaling, while in cells receiving Wnt signals Axin puncta are recruited to the plasma

membrane (Schaefer et al., 2018). Interestingly, in Wnt receiving cells, the number of Axin

molecules in puncta is reduced (Schaefer et al., 2018) while the cytoplasmic Axin pool is

increased (Schaefer et al., 2018; Wang et al., 2016a; Yang et al., 2016)) suggesting that after

Axin is recruited to the membrane some change occurs that either inhibits Axin self-

polymerization or inhibits its stability within condensates. But what is the nature of this change?

Wnt signaling causes a switch in the destruction complex mix, destabilizing it

One potential change involves a switch in binding partners. Dvl, one of the first proteins

implicated in Wnt signaling, inhibits destruction complex function in response to Wnt signaling

(reviewed in Mlodzik, 2016). As noted above, Dvl and Axin both have a DIX domain. This

shared domain mediates both self-polymerization and hetero-polymerization (Fiedler et al.,

2011; Kishida et al., 1999; Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b;

Smalley et al., 1999). Dvl is essential for Wnt receptor phosphorylation, Axin recruitment to the

membrane and signalasome endocytosis, thus turning down destruction complex function (Bilic

et al., 2007; Cliffe et al., 2003). Dvl binding to Fz, via Dvl’s DEP and/or PDZ domains (Axelrod et

al., 1998; Wong et al., 2003), is followed by a conformational change that crosslinks Dsh

polymers, increasing local concentration of Dvl at the receptor as it is endocytosed (Gammons

et al., 2016). DEP-mediated Dvl cross-linking may drive Axin recruitment by increased avidity,

driving Dvl:Axin hetero-polymerization. FRAP analysis revealed that Dvl:Axin co-assembly

enhances Axin turnover in puncta (Schwarz-Romond et al., 2007b). Dvl destabilization of Axin

puncta thus provides one mechanism by which Dvl could inactivate the destruction complex. In

contrast, APC inclusion in Axin puncta stabilizes them, increasing destruction complex efficiency

Figure 1.4

Figure 1.4: In vivo recruitment of APC2 into Axin:GFP puncta. A) Model illustrating Wnt signaling in a Drosophila embryo. One row of cells per body segment produce and secrete the Wnt Wingless (Wg). It forms a graded distribution and stabilizes the fly βCat (Arm), leading to graded Arm levels across the body segment. B) Stage 9 Drosophila embryos expressing Axin:GFP at near endogenous levels. Anterior to the left. Axin:GFP localization is dependent on Wnt/Wg expression. In the absence of Wg, Axin:GFP is found in cytoplasmic puncta containing 10-100s of Axin molecules. In the presence of Wg, Axin puncta are located along the membrane and there is an increase in cytoplasmic GFP. Staining for endogenous APC2 shows that APC2 is recruited to both cytoplasmic Axin puncta and membrane localized Axin puncta. (Image originally published in Schaefer et al., 2018.: DOI: 10.1371/journal.pgen.1007339).

(Pronobis et al., 2015). Intriguingly, co-expressing Axin, APC, and Dvl2 in cultured cells revealed that APC:Axin:Dvl2 complexes are rare while Axin:APC or Axin:Dvl2 complexes are more frequent, consistent with a competition between APC and Dvl for interaction with Axin

(Mendoza-Topaz et al., 2011). Competition for Axin binding is also consistent with the fact that protein levels of APC:Axin:Dsh in Drosophila embryos are all in the same order of magnitude

(Schaefer et al., 2018). The possibility that DIX:DIX interactions between Dvl and Axin inhibit

Axin is also consistent with fact that Drosophila Axin∆DIX is constitutively active in βCat destruction (Peterson-Nedry et al., 2008). One remaining question is how Axin decides between binding its different partners? Condensates form via multiple multivalent interactions. Perhaps changes in Axin ADP-ribosylation or phosphorylation in response to Wnt signaling, altering the

charge of the intrinsically disordered region, reduce interaction between Axin:APC or promote

Axin:Dvl interaction. Future research into the rules regulating the competition between

assembly/disassembly of the destruction complex and that of the signalasome will provide

essential insights.

One consequence of supermolecular assembly: the kinase GSK3 plays both positive and

negative roles in the destruction complex via its access to many targets

GSK3, first discovered as a kinase regulating glycogen metabolism, plays pleiotropic roles in

the cell, regulating multiple signaling pathways (Cormier and Woodgett, 2017). GSK3 was one

of the first proteins with a known biochemical role to be placed in the Wnt pathway and the first

negative regulator, a role defined via genetic analysis in Drosophila (Peifer et al., 1994;

Siegfried et al., 1992; Siegfried et al., 1990; Siegfried et al., 1994). This raises the question of

how is pathway specificity maintained? Recruitment into different supermolecular complexes

provides a mechanism. Both CK1 and GSK3 are recruited by Axin into the destruction complex,

where they sequentially phosphorylate βCat, priming it for destruction. However, subsequent

work revealed that GSK3 plays many roles in the pathway, both positive and negative. Perhaps

23 it is not surprising that recruiting an active kinase into a multiprotein complex allows it to phosphorylate many proteins within. Within the active destruction complex, GSK3 phosphorylates Axin to keep Axin “open” for βCat interaction (Kim et al., 2013), phosphorylates

APCs 20Rs to increase affinity for βCat (Ha et al., 2004; Liu et al., 2006; Xing et al., 2004), and phosphorylates R2/B to facilitate βCat release to the E3 ligase (Pronobis et al., 2015). However,

GSK3 is also recruited into the Wnt signalsome, where it plays important roles. In response to

Wnt signaling, CK1 and GSK3 phosphorylate the tail of LRP5/6, creating a binding site that facilitates Axin recruitment to the receptor complex for inactivation (Tamai et al., 2004; Zeng et al., 2005), a process that is visualized in Drosophila as GSK3-dependent recruitment of Axin puncta to the membrane (Cliffe et al., 2003; Schaefer et al., 2018). Intriguingly, the LRP5/6 phosphorylated tail then can act as a GSK3 inhibitor (Piao et al., 2008; Stamos et al., 2014; Wu et al., 2009), providing another mechanism by which Wnt activation turns down the destruction complex. GSK3 phosphorylation of Axin also allows it to be “open” for binding LRP5/6 (Kim et al., 2013). In these two roles GSK3 is a positive effector of Wnt signaling. Vertebrates have two

GSK isoforms, and questions remain about the roles they play (or do not) in regulating this pathway. Genetic analysis demonstrated that the two isoforms are largely redundant for Wnt regulation, with single mutants having tissue-specific defects (Doble et al., 2007; Hoeflich et al.,

2000), though recent studies with isoform specific inhibitors suggest possible differences in isoform function in Wnt regulation (Chen et al., 2017).

Axin post-translational regulation plays complex roles

Axin features in most studies of destruction complex downregulation, with models ranging from enhanced Axin degradation, to its dissociation from GSK3, or changes in its assembly state (reviewed in MacDonald and He, 2012; Nusse and Clevers, 2017). These diverse data may reflect differences in different animals and/or tissues or, as we think more likely, may reflect initial versus more long-term mechanisms for elevating βCat levels and thus Wnt signaling.

Many models of destruction complex inhibition involve Axin regulation via post-translational modifications, including phosphorylation/dephosphorylation, ADP-ribosylation, or ubiquitination.

As noted above, Axin phosphorylation plays important but diverse roles. Interestingly a recent report suggested that all of these Axin phosphorylation events are dependent on APC

(Tacchelly-Benites et al., 2018).

One common consequence of long-term Wnt signaling is down-regulation of Axin protein

levels (Kofron et al., 2007; Liu et al., 2005; Mao et al., 2001; Tolwinski et al., 2003; Yang et al.,

2016). GSK3 phosphorylation of Axin can stabilize the protein, while Axin de-phosphorylation by

PP2A leads to degradation (Willert et al., 1999; Yamamoto et al., 1999). Recent work highlights

the role of kinases as “dissolvases” of other biomolecular condensates (Rai et al., 2018). Two

other post-translational modifications of Axin, ADP-ribosylation and ubiquitination, are also

proposed to regulate Axin stability and thus Wnt signaling. Tankyrase is a poly(ADP-ribose)

polymerase, adding ADP-ribose moieties to target proteins, which are then often ubiquitinated

and destroyed (Hsiao and Smith, 2008; Mariotti et al., 2017). Tankyrase binds Axin and ADP-

ribosylates it (Huang et al., 2009), targeting it for ubiquitination by RNF146, a poly(ADP-ribose)-

directed E3 ligase (Callow et al., 2011; Zhang et al., 2011; Zhou et al., 2011). Wnt signaling

induces Axin ADP-ribosylation within 30 min (Yang et al., 2016). Thus it seemed plausible that

Tankyrase caused the Axin degradation seen after longer exposure to Wnt signaling. Consistent

with this, Tankyrase inhibition stabilizes Axin puncta formation in both mammalian and

Drosophila cells (de la Roche et al., 2014; Thorvaldsen et al., 2015; Waaler et al., 2011).

However, things are not that simple. ADP-ribosylation of Axin also enhances Axin’s ability to

immunoprecipitate with active LRP5/6 (Yang et al., 2016). Further, Axin is still degraded after

long Wnt signaling exposure in fly Tankyrase mutants, or when Axin lacks its Tankyrase binding

domain (AxinΔTBD) (Wang et al., 2016b). This suggests Tankyrase is not solely responsible for

Axin degradation. In fact, in embryos expressing AxinΔTBD the initial accumulation of Axin

25 shortly after Wnt exposure is now seen in all cells, not just those receiving Wg, suggesting other potential effects of ADP-ribosylation (Yang et al., 2016). Intriguingly, Tankyrase can also bind

APC (Croy et al., 2016). It is important to keep in mind that Drosophila lacking Tankyrase are viable and fertile (Feng et al., 2014), and mice lacking both Tankyrase proteins, while embryonic lethal, survive to E10 without obvious defects in Wnt signaling. Thus while Tankyrase can finely regulate destruction complex activity, as it does in the Drosophila intestine (Wang et al., 2016c), it is not an essential regulator of the pathway.

If Tankyrase is not essential for Axin degradation, then how are Axin levels decreased after

Wnt signaling? This may be triggered by the RING domain E3 ligase Seven in absentia (Sina) homolog SIAH 1/2. SIAH’s potential role in Wnt signaling was first identified in a yeast 2-hybrid screen for novel APC binding partners. Further exploration of this interaction in human cells and

Xenopus embryos indicated that SIAH can ubiquitinate βCat, labeling it for proteasomal degradation, independent of βTrCP (Liu et al., 2001; Matsuzawa and Reed, 2001). However, recently SIAH 1 and SIAH 2 were identified as novel regulators of Axin levels in HEK293T cells.

Co-crystallization revealed that SIAH directly binds Axin near its GSK3 binding site.

Interestingly, while SIAH and GSK3 can simultaneously interact with Axin, there is competition between the two, since the binding of one can allosterically inhibit binding of the other (Ji et al.,

2017). Based on these data, one could hypothesize that after Wnt signaling inhibits GSK3 interaction with Axin, SIAH binds Axin and labels it for proteasomal degradation. Once again, however, it is important to note that the single Drosophila Sina family member, is adult viable without obvious defects in Wnt signaling (Carthew and Rubin, 1990).

While long-term Wnt exposure decreases Axin levels, recent studies in vivo in Drosophila suggest that Axin levels initially increase after Wnt signaling is activated (Wang et al., 2016a;

Yang et al., 2016). Axin levels in the cytoplasm initially increase in cells receiving Wnt signals, and only decrease several hours later (Yang et al., 2016). Total embryonic Axin levels increase

26 over the same time frame, but it is hard to rule out that this is not simply an effect of activating zygotic Axin expression in all cells. This degree of Axin elevation is at the threshold at which

Axin can inhibit Wnt signaling (3-9x; Peterson-Nedry et al., 2008; Schaefer et al., 2018; Wang et al., 2016b). Why should Wnt signaling elevate Axin levels when this should enhance βCat destruction and thus inhibit Wnt signaling? One possibility is that the increased pool of cytoplasmic Axin is largely “inactivated” Axin molecules that cannot form stable puncta. In fact, membrane-localized Axin puncta in Wnt-ON cells harbor only half the number of Axin molecules

as cytoplasmic puncta in Wnt-OFF cells, while cytoplasmic Axin levels rise (Schaefer et al.,

2018). Together, these data reveal many levels at which Axin is regulated by post-translational

modification and open up questions for future research, defining which changes are the initial

response to Wnt signals and which are adjustments allowing longer term modulation of

signaling.

APC mutations in colorectal cancer target specific aspects of destruction complex function

Activating mutations in the Wnt pathway play roles in many cancers, including endometrial

and liver cancer, but are most prominent in colorectal tumors, where they initiate oncogenesis

(reviewed in Zhang and Shay, 2017). ~10% of colorectal tumors have gain-of-function βCat

mutations disrupting phosphorylation and thus destruction, a few have loss-of-function Axin

mutations, but >80% are APC mutant (Kandoth et al., 2013). These mutations have a very

striking feature—unlike most tumor suppressors where selection favors homozygous null mutations, all or virtually all colorectal tumors have at least one APC allele expressing a truncated protein. Intriguingly, the truncations occur in a small region of the protein, the mutation cluster region (MCR; Kohler et al., 2008), leading researchers to explore what properties are lost or retained to favor selection. Most now accept the “just right” hypothesis (Albuquerque et al., 2002), which proposes that complete loss of APC function leads to such high levels of Wnt activity that oncogenic stress sensors trigger apoptosis. This suggests the truncated APC

protein retains some function. What functions are retained and which lost? One critical thing lost

in the truncated proteins are the SAMP motifs, the high affinity Axin binding sites. A mouse

mutant with one allele truncated to lack all SAMPs is tumor prone while mice carrying an allele

with a slightly longer truncation retaining one SAMP are not tumor prone (and in fact are

homozygous viable!; Smits et al., 1999). However, closer examination of the MCR suggested

more is going on, focusing attention on the B/CID motif, which is just N-terminal to the last

SAMP (Fig. 3A) and thus also disrupted in most or all tumor truncations (Kohler et al., 2008).

Thus selection may favor loss of both the SAMPS and B/CID. What function is retained by

truncated APC to prevent selection for null mutations? Truncated APC proteins like those in

tumors cannot promote βCat destruction but can still bind βCat (Roberts et al., 2011) and

mediate its phosphorylation (Yang et al., 2006). Retaining βCat in the cytoplasm may thus be

how destruction complexes containing truncated APC dampen but do not eliminate Wnt

signaling. It will be intriguing to further explore assembly and function of destruction complex

condensates carrying truncated APC, as deleting the SAMP motifs reduces but does not

eliminate APC incorporation into puncta (Pronobis et al., 2015; Roberts et al., 2011).

The destruction complex is a multifunctional machine with other targets including the cytoskeleton

While best known for roles in Wnt regulation, destruction complex proteins also have other

functions. The first evidence emerged even before the connection to Wnt signaling was made,

when scientists found that human APC co-localizes with and binds microtubules and the microtubule plus end protein EB1(Munemitsu et al., 1994; Smith et al., 1994; Su et al., 1995).

Subsequent work on APC led to suggested roles in spindle orientation, chromosome

segregation, and polarity of neurons and migrating cells, but many of these studies rely on

overexpressing full-length or truncated APC (reviewed in Nelson and Nathke, 2013; Rusan and

Peifer, 2008). Use of Drosophila allowed analysis of complete loss of APC function, revealing

roles in genome stability via regulating centrosome migration and function of the formin

Diaphanous (McCartney et al., 2001; Poulton et al., 2013; Webb et al., 2009), mitotic spindle

orientation (Yamashita et al., 2003) and microtubule dynamics in neuronal dendrites (Mattie et

al., 2010; Weiner et al., 2016), but casting doubt on suggested roles in axon or dendrite polarity

(Rusan et al., 2008). C. elegans APR-1 helps orient mitotic spindles by attenuating cortical

spindle pulling forces (Sugioka et al., 2018; Sugioka et al., 2011). At least some APC proteins

also regulate actin dynamics, in part through a “rocket launching” mechanism in which they work

with Diaphanous to stimulate filament nucleation and extension (Breitsprecher et al., 2012;

Jaiswal et al., 2013), and thus regulate focal adhesion turnover (Juanes et al., 2017). When

considering cytoskeletal roles for APC, however, it is critical to rule out places where effects on

Wnt signaling lead to downstream cytoskeletal alterations (e.g., Elbaz et al., 2016; Eom et al.,

2014; Hayden et al., 2007; Nakagawa et al., 2017; Yokota et al., 2009).

Other destruction complex or signalasome proteins also may have cytoskeletal roles. For

example, Axin has suggested roles in mouse oocyte meiosis (He et al., 2016) and in axon and

dendrite morphogenesis and intermediate neuronal progenitor differentiation in the cerebral

cortex (Chen et al., 2015; Fang et al., 2013; Fang et al., 2011). Once again, however, one must

be cautious in differentiating Wnt-independent from Wnt dependent roles. The Wnt regulator Dvl

and receptor Fz have well known roles in planar cell polarity and in cilia genesis/orientation, but

these appear to be largely independent of the destruction complex and canonical βCat signaling

(reviewed in Adler and Wallingford, 2017). There is one common cytoskeletal thread among

different destruction complex proteins and regulators—APC, Axin, Dvl and βCat are all reported

to localize to centrosomes in at least some cell types (reviewed in Bryja et al., 2017; Mbom et

al., 2013). While some data suggests roles for centrosomes and cilia in canonical Wnt signaling,

both flies and mice lacking centrosomes develop without strong defects in Wnt signaling (Basto

et al., 2006; Bazzi and Anderson, 2014). Likewise, basic spindle assembly functions of

Figure 1.5

Figure 1.5: A revised model of the destruction complex. Polymers of Axin and of APC, mediated by polymerization of their respective DIX and ASAD/Arm repeat domains, intertwine. Polymers interact via the RGS:SAMP and Arm repeat-Axin motif interactions. Polymers concentrate ßcat, GSK3, and CK1 (not shown), accelerating phosphorylation.

30 centrosomes do not absolutely depend on APC, Axin, Dvl or βCat. Defining the full suite of

Wnt-independent cytoskeletal roles remains a challenge for the field, as is determining whether these roles are independent roles for individual destruction complex proteins or whether the complex acts as an entity. In a few cases, APC and Axin share roles (e.g., Mattie et al., 2010;

Poulton et al., 2013; Weiner et al., 2016), but whether this involves the destruction complex itself, and whether the complex regulates phosphorylation of targets in addition to βCat (e.g.,

Kim et al., 2009) remain to be determined.

Looking toward the future

The sections above outline current questions raised by new insights into the regulation of

Wnt signaling. Looking more globally, it’s an exciting time for the Wnt field and for the broader

fields of signaling pathways and biomolecular condensates. Virtually every week brings new

examples of cellular structures assembled by phase separation (e.g. Du and Chen, 2018; Shan

et al., 2018) or defining the mechanisms by which they assemble (e.g. (Harmon et al., 2017; Rai

et al., 2018; Zeng et al., 2018). Current data strongly support the idea that Axin and APC

multimerization and APC-dependent destruction complex stabilization, mediated by its

multivalent interactions with Axin, increase the efficiency of the destruction complex (Fig. 5). We

predict this is elicited via effects of both elevated protein concentration and increased avidity

within puncta. Considering the destruction complex and signalosome as potential biomolecular

condensates, many broad questions remain, which can be addressed using ever more powerful

tools ranging from in vitro reconstitution to in vivo genetic analysis to molecular modeling. What

proteins nucleate formation of each supermolecular complex, and how do the core proteins

choose between nucleating a new assembly or adding to an old one? How do post-translational

modifications govern condensate assembly or disassembly?

How do different supermolecular complexes, like the destruction complex and the SCF-E3 ligase, communicate with one another to control flow of molecules and information? How do proteins like APC, βCat, and GSK3, which play multiple roles within each cell, choose between different supermolecular assemblies? These and other questions will keep many of us happily employed for many years to come.

REFERENCES

Adler, P.N., and Wallingford, J.B. (2017). From Planar Cell Polarity to Ciliogenesis and Back: The Curious Tale of the PPE and CPLANE proteins. Trends Cell Biol 27, 379-390.

Ahmed, Y., Nouri, A., and Wieschaus, E. (2002). Drosophila Apc1 and Apc2 regulate Wingless transduction throughout development. Development 129, 1751-1762.

Akong, K., Grevengoed, E., Price, M., McCartney, B., Hayden, M., DeNofrio, J., and Peifer, M. (2002). Drosophila APC2 and APC1 Play Overlapping Roles in Wingless Signaling in the Embryo and Imaginal Discs. Dev Biol 250, 91-100.

Albuquerque, C., Breukel, C., van der Luijt, R., Fidalgo, P., Lage, P., Slors, F.J., Leitao, C.N., Fodde, R., and Smits, R. (2002). The 'just-right' signaling model: APC somatic mutations are selected based on a specific level of activation of the beta-catenin signaling cascade. Hum Mol Genet 11, 1549-1560.

Axelrod, J.D., Miller, J.R., Shulman, J.M., Moon, R.T., and Perrimon, N. (1998). Differential recruitment of Dishevelled provides signaling specificity in the planar cell polarity and Wingless signaling pathways. Genes Dev 12, 2610-2622.

Banani, S.F., Lee, H.O., Hyman, A.A., and Rosen, M.K. (2017). Biomolecular condensates: organizers of cellular biochemistry. Nat Rev Mol Cell Biol 18, 285-298.

Basto, R., Lau, J., Vinogradova, T., Gardiol, A., Woods, C.G., Khodjakov, A., and Raff, J.W. (2006). Flies without centrioles. Cell 125, 1375-1386.

Bazzi, H., and Anderson, K.V. (2014). Acentriolar mitosis activates a p53-dependent apoptosis pathway in the mouse embryo. Proc Natl Acad Sci U S A 111, E1491-1500.

Behrens, J., Jerchow, B.-A., Würtele, M., Grimm, J., Asbrand, C., Wirtz, R., Kühl, M., Wedlich, D., and Birchmeier, W. (1998). Functional interaction of an axin homolog conductin, with ß- catenin, APC, and GSK3ß. Science 280, 596-599.

Bilic, J., Huang, Y.L., Davidson, G., Zimmermann, T., Cruciat, C.M., Bienz, M., and Niehrs, C. (2007). Wnt induces LRP6 signalosomes and promotes dishevelled-dependent LRP6 phosphorylation. Science 316, 1619-1622.

Brangwynne, C.P., Eckmann, C.R., Courson, D.S., Rybarska, A., Hoege, C., Gharakhani, J., Julicher, F., and Hyman, A.A. (2009). Germline P granules are liquid droplets that localize by controlled dissolution/condensation. Science 324, 1729-1732.

Breitsprecher, D., Jaiswal, R., Bombardier, J.P., Gould, C.J., Gelles, J., and Goode, B.L. (2012). Rocket launcher mechanism of collaborative actin assembly defined by single-molecule imaging. Science 336, 1164-1168.

Bryja, V., Cervenka, I., and Cajanek, L. (2017). The connections of Wnt pathway components with cell cycle and centrosome: side effects or a hidden logic? Crit Rev Biochem Mol Biol 52, 614-637.

Callow, M.G., Tran, H., Phu, L., Lau, T., Lee, J., Sandoval, W.N., Liu, P.S., Bheddah, S., Tao, J., Lill, J.R., et al. (2011). Ubiquitin ligase RNF146 regulates tankyrase and Axin to promote Wnt signaling. PLoS ONE 6, e22595.

Capelluto, D.G., Kutateladze, T.G., Habas, R., Finkielstein, C.V., He, X., and Overduin, M. (2002). The DIX domain targets dishevelled to actin stress fibres and vesicular membranes. Nature 419, 726-729.

Chen, X., Wang, R., Liu, X., Wu, Y., Zhou, T., Yang, Y., Perez, A., Chen, Y.C., Hu, L., Chadarevian, J.P., et al. (2017). A Chemical-Genetic Approach Reveals the Distinct Roles of GSK3alpha and GSK3beta in Regulating Embryonic Stem Cell Fate. Dev Cell 43, 563-576 e564.

Chen, Y., Liang, Z., Fei, E., Chen, Y., Zhou, X., Fang, W., Fu, W.Y., Fu, A.K., and Ip, N.Y. (2015). Axin Regulates Dendritic Spine Morphogenesis through Cdc42-Dependent Signaling. PLoS One 10, e0133115.

Choi, S.H., Estaras, C., Moresco, J.J., Yates, J.R., 3rd, and Jones, K.A. (2013). alpha-Catenin interacts with APC to regulate beta-catenin proteolysis and transcriptional repression of Wnt target genes. Genes Dev 27, 2473-2488.

Cliffe, A., Hamada, F., and Bienz, M. (2003). A role of Dishevelled in relocating Axin to the plasma membrane during wingless signaling. Curr Biol 13, 960-966.

Cormier, K.W., and Woodgett, J.R. (2017). Recent advances in understanding the cellular roles of GSK-3. F1000Res 6.

Costa, M., Raich, W., Agbunag, C., Leung, B., Hardin, J., and Priess, J.R. (1998). A putative catenin-cadherin system mediates morphogenesis of the Caenorhabditis elegans embryo. Journal of Cell Biology 141, 297-308.

Croy, H.E., Fuller, C.N., Giannotti, J., Robinson, P., Foley, A.V., Yamulla, R.J., Cosgriff, S., Greaves, B.D., von Kleeck, R.A., An, H.H., et al. (2016). The Poly(ADP-ribose) Polymerase Enzyme Tankyrase Antagonizes Activity of the beta-Catenin Destruction Complex through ADP-ribosylation of Axin and APC2. J Biol Chem 291, 12747-12760.

Davidson, G., Wu, W., Shen, J., Bilic, J., Fenger, U., Stannek, P., Glinka, A., and Niehrs, C. (2005). Casein kinase 1 gamma couples Wnt receptor activation to cytoplasmic signal transduction. Nature 438, 867-872. de la Roche, M., Ibrahim, A.E., Mieszczanek, J., and Bienz, M. (2014). LEF1 and B9L shield beta-catenin from inactivation by Axin, desensitizing colorectal cancer cells to tankyrase inhibitors. Cancer Res 74, 1495-1505.

DeBruine, Z.J., Xu, H.E., and Melcher, K. (2017). Assembly and architecture of the Wnt/beta- catenin signalosome at the membrane. Br J Pharmacol 174, 4564-4574.

Desai, R., Sarpal, R., Ishiyama, N., Pellikka, M., Ikura, M., and Tepass, U. (2013). Monomeric alpha-catenin links cadherin to the actin cytoskeleton. Nat Cell Biol 15, 261-273.

Doble, B.W., Patel, S., Wood, G.A., Kockeritz, L.K., and Woodgett, J.R. (2007). Functional redundancy of GSK-3alpha and GSK-3beta in Wnt/beta-catenin signaling shown by using an allelic series of embryonic stem cell lines. Dev Cell 12, 957-971.

Du, M., and Chen, Z.J. (2018). DNA-induced liquid phase condensation of cGAS activates innate immune signaling. Science 361, 704-709.

Eklof Spink, K., Fridman, S.G., and Weis, W.I. (2001). Molecular mechanisms of beta-catenin recognition by adenomatous polyposis coli revealed by the structure of an APC-beta-catenin complex. EMBO Journal 20, 6203-6212.

Elbaz, B., Traka, M., Kunjamma, R.B., Dukala, D., Brosius Lutz, A., Anton, E.S., Barres, B.A., Soliven, B., and Popko, B. (2016). Adenomatous polyposis coli regulates radial axonal sorting and myelination in the PNS. Development 143, 2356-2366.

Eom, T.Y., Stanco, A., Guo, J., Wilkins, G., Deslauriers, D., Yan, J., Monckton, C., Blair, J., Oon, E., Perez, A., et al. (2014). Differential regulation of microtubule severing by APC underlies distinct patterns of projection neuron and interneuron migration. Dev Cell 31, 677- 689.

Fagotto, F., Jho, E.H., Zeng, L., Kurth, T., Joos, T., Kaufmann, C., and Costantini, F. (1999). Domains of Axin involved in protein-protein interactions, Wnt pathway inhibition, and intracellular localization. Journal of Cell Biology 145, 741-756.

Fang, W.Q., Chen, W.W., Fu, A.K., and Ip, N.Y. (2013). Axin directs the amplification and differentiation of intermediate progenitors in the developing cerebral cortex. Neuron 79, 665- 679.

Fang, W.Q., Ip, J.P., Li, R., Ng, Y.P., Lin, S.C., Chen, Y., Fu, A.K., and Ip, N.Y. (2011). Cdk5- mediated phosphorylation of Axin directs axon formation during cerebral cortex development. J Neurosci 31, 13613-13624.

Faux, M.C., Coates, J.L., Catimel, B., Cody, S., Clayton, A.H., Layton, M.J., and Burgess, A.W. (2008). Recruitment of adenomatous polyposis coli and beta-catenin to axin-puncta. Oncogene 27, 5808-5820.

Feng, Y., Li, X., Ray, L., Song, H., Qu, J., Lin, S., and Lin, X. (2014). The Drosophila tankyrase regulates Wg signaling depending on the concentration of Daxin. Cell Signal 26, 1717-1724.

Fiedler, M., Mendoza-Topaz, C., Rutherford, T.J., Mieszczanek, J., and Bienz, M. (2011). Dishevelled interacts with the DIX domain polymerization interface of Axin to interfere with its function in down-regulating beta-catenin. Proceedings of the National Academy of Sciences of the United States of America 108, 1937-1942.

Gall, J.G. (2000). Cajal bodies: the first 100 years. Annu Rev Cell Dev Biol 16, 273-300. Gammons, M., and Bienz, M. (2017). Multiprotein complexes governing Wnt signal transduction. Curr Opin Cell Biol 51, 42-49.

Gammons, M.V., Renko, M., Johnson, C.M., Rutherford, T.J., and Bienz, M. (2016). Wnt Signalosome Assembly by DEP Domain Swapping of Dishevelled. Mol Cell 64, 92-104.

Green, J., Nusse, R., and van Amerongen, R. (2014). The role of Ryk and Ror receptor tyrosine kinases in Wnt signal transduction. Cold Spring Harb Perspect Biol 6.

Ha, N.C., Tonozuka, T., Stamos, J.L., Choi, H.J., and Weis, W.I. (2004). Mechanism of phosphorylation-dependent binding of APC to beta-catenin and its role in beta-catenin degradation. Mol Cell 15, 511-521.

Harmon, T.S., Holehouse, A.S., Rosen, M.K., and Pappu, R.V. (2017). Intrinsically disordered linkers determine the interplay between phase separation and gelation in multivalent proteins. eLife 6.

Hart, M.J., de los Santos, R., Albert, I.N., Rubinfeld, B., and Polakis, P. (1998). Downregulation of beta-catenin by human Axin and its association with the APC tumor suppressor, ß-catenin and GSK3ß. Current Biology 8, 573-581.

Hayden, M.A., Akong, K., and Peifer, M. (2007). Novel roles for APC family members and Wingless/Wnt signaling during Drosophila brain development. Dev Biol 305, 358-376.

He, X.Q., Song, Y.Q., Liu, R., Liu, Y., Zhang, F., Zhang, Z., Shen, Y.T., Xu, L., Chen, M.H., Wang, Y.L., et al. (2016). Axin-1 Regulates Meiotic Spindle Organization in Mouse Oocytes. PLoS One 11, e0157197.

Hernandez, A.R., Klein, A.M., and Kirschner, M.W. (2012). Kinetic responses of beta-catenin specify the sites of Wnt control. Science 338, 1337-1340.

Hoeflich, K.P., Luo, J., Rubie, E.A., Tsao, M.S., Jin, O., and Woodgett, J.R. (2000). Requirement for glycogen synthase kinase-3beta in cell survival and NF-kappaB activation. Nature 406, 86-90.

Holehouse, A.S., and Pappu, R.V. (2018). Functional Implications of Intracellular Phase Transitions. Biochemistry 57, 2415-2423.

Hsiao, S.J., and Smith, S. (2008). Tankyrase function at telomeres, spindle poles, and beyond. Biochimie 90, 83-92.

Huang, S.-M.A., Mishina, Y.M., Liu, S., Cheung, A., Stegmeier, F., Michaud, G.A., Charlat, O., Wiellette, E., Zhang, Y., Wiessner, S., et al. (2009). Tankyrase inhibition stabilizes axin and antagonizes Wnt signalling. Nature 461, 614.

Ikeda, S., Kishida, S., Yamamoto, H., Murai, H., Koyama, S., and Kikuchi, A. (1998). Axin, a negative regulator of the Wnt signaling pathway, forms a complex with GSK-3ß and ß- catenin and promotes GSK-3ß-dependent phosphorylation of ß-catenin. EMBO J 17, 1371- 1384.

Jaiswal, R., Stepanik, V., Rankova, A., Molinar, O., Goode, B.L., and McCartney, B.M. (2013). Drosophila homologues of adenomatous polyposis coli (APC) and the formin diaphanous collaborate by a conserved mechanism to stimulate actin filament assembly. J Biol Chem 288, 13897-13905.

Ji, L., Jiang, B., Jiang, X., Charlat, O., Chen, A., Mickanin, C., Bauer, A., Xu, W., Yan, X., and Cong, F. (2017). The SIAH E3 ubiquitin ligases promote Wnt/beta-catenin signaling through mediating Wnt-induced Axin degradation. Genes Dev 31, 904-915.

Ji, L., Lu, B., Wang, Z., Yang, Z., Reece-Hoyes, J., Russ, C., Xu, W., and Cong, F. (2018). Identification of ICAT as an APC Inhibitor, Revealing Wnt-Dependent Inhibition of APC-Axin Interaction. Mol Cell.

Joslyn, G., Richardson, D.S., White, R., and Alber, T. (1993). Dimer formation by an N-terminal coiled coil in the APC protein. Proc Natl Acad Sci U S A 90, 11109-11113.

Juanes, M.A., Bouguenina, H., Eskin, J.A., Jaiswal, R., Badache, A., and Goode, B.L. (2017). Adenomatous polyposis coli nucleates actin assembly to drive cell migration and microtubule-induced focal adhesion turnover. J Cell Biol 216, 2859-2875.

Julius, M.A., Schelbert, B., Hsu, W., Fitzpatrick, E., Jho, E., Fagotto, F., Costantini, F., and Kitajewski, J. (2000). Domains of axin and disheveled required for interaction and function in wnt signaling. Biochem Biophys Res Commun 276, 1162-1169.

Kandoth, C., McLellan, M.D., Vandin, F., Ye, K., Niu, B., Lu, C., Xie, M., Zhang, Q., McMichael, J.F., Wyczalkowski, M.A., et al. (2013). Mutational landscape and significance across 12 major cancer types. Nature 502, 333-339.

Kim, N.G., Xu, C., and Gumbiner, B.M. (2009). Identification of targets of the Wnt pathway destruction complex in addition to beta-catenin. Proc Natl Acad Sci U S A 106, 5165-5170.

Kim, S.E., Huang, H., Zhao, M., Zhang, X., Zhang, A., Semonov, M.V., MacDonald, B.T., Zhang, X., Garcia Abreu, J., Peng, L., et al. (2013). Wnt stabilization of beta-catenin reveals principles for morphogen receptor-scaffold assemblies. Science 340, 867-870.

Kishida, S., Yamamoto, H., Hino, S.I., Ikeda, S., Kishida, M., and Kikuchi, A. (1999). DIX domains of Dvl and axin are necessary for protein interactions and their ability to regulate beta-catenin stability. Molecular and Cellular Biology 19, 4414-4422.

Kishida, S., Yamamoto, H., Ikeda, S., Kishida, M., Sakamoto, I., Koyama, S., and Kikuchi, A. (1998). Axin, a negative regulator of the wnt signaling pathway, directly interacts with adenomatous polyposis coli and regulates the stabilization of beta-catenin. J Biol Chem 273, 10823-10826.

Kitazawa, M., Hatta, T., Ogawa, K., Fukuda, E., Goshima, N., and Natsume, T. (2017). Determination of Rate-Limiting Factor for Formation of Beta-Catenin Destruction Complexes Using Absolute Protein Quantification. Journal of proteome research 16, 3576-3584.

Kofron, M., Birsoy, B., Houston, D., Tao, Q., Wylie, C., and Heasman, J. (2007). Wnt11/beta- catenin signaling in both oocytes and early embryos acts through LRP6-mediated regulation of axin. Development 134, 503-513.

Kohler, E.M., Chandra, S.H., Behrens, J., and Schneikert, J. (2009). Beta-catenin degradation mediated by the CID domain of APC provides a model for the selection of APC mutations in colorectal, desmoid and duodenal tumours. Hum Mol Genet 18, 213-226.

Kohler, E.M., Derungs, A., Daum, G., Behrens, J., and Schneikert, J. (2008). Functional definition of the mutation cluster region of adenomatous polyposis coli in colorectal tumours. Hum Mol Genet 17, 1978-1987.

Kremer, S.A., Erdeniz, N., Peterson-Nedry, W., Swanson, E.A., and Wehrli, M. (2010). In vivo analysis in Drosophila reveals differential requirements of contact residues in Axin for interactions with GSK3beta or beta-catenin. Dev Biol 337, 110-123.

Krieghoff, E., Behrens, J., and Mayr, B. (2006). Nucleo-cytoplasmic distribution of beta-catenin is regulated by retention. J Cell Sci 119, 1453-1463.

Kunttas-Tatli, E., Roberts, D.M., and McCartney, B.M. (2014). Self-association of the APC tumor suppressor is required for the assembly, stability, and activity of the Wnt signaling destruction complex. Mol Biol Cell 25, 3424-3436.

Kunttas-Tatli, E., Von Kleeck, R.A., Greaves, B.D., Vinson, D., Roberts, D.M., and McCartney, B.M. (2015). The two SAMP repeats and their phosphorylation state in Drosophila Adenomatous polyposis coli-2 play mechanistically distinct roles in negatively regulating Wnt signaling. Mol Biol Cell 26, 4503-4518.

Kunttas-Tatli, E., Zhou, M.N., Zimmerman, S., Molinar, O., Zhouzheng, F., Carter, K., Kapur, M., Cheatle, A., Decal, R., and McCartney, B.M. (2012). Destruction complex function in the Wnt signaling pathway of Drosophila requires multiple interactions between Adenomatous polyposis coli 2 and Armadillo. Genetics 190, 1059-1075.

Lee, E., Salic, A., Kruger, R., Heinrich, R., and Kirschner, M.W. (2003). The roles of APC and Axin derived from experimental and theoretical analysis of the Wnt pathway. PLoS Biol 1, E10.

Li, P., Banjade, S., Cheng, H.C., Kim, S., Chen, B., Guo, L., Llaguno, M., Hollingsworth, J.V., King, D.S., Banani, S.F., et al. (2012a). Phase transitions in the assembly of multivalent signalling proteins. Nature 483, 336-340.

Li, V.S., Ng, S.S., Boersema, P.J., Low, T.Y., Karthaus, W.R., Gerlach, J.P., Mohammed, S., Heck, A.J., Maurice, M.M., Mahmoudi, T., et al. (2012b). Wnt Signaling through Inhibition of beta-Catenin Degradation in an Intact Axin1 Complex. Cell 149, 1245-1256.

Liu, J., Stevens, J., Rote, C.A., Yost, H.J., Hu, Y., Neufeld, K.L., White, R.L., and Matsunami, N. (2001). Siah-1 mediates a novel beta-catenin degradation pathway linking p53 to the adenomatous polyposis coli protein. Mol Cell 7, 927-936.

Liu, J., Xing, Y., Hinds, T.R., Zheng, J., and Xu, W. (2006). The third 20 amino acid repeat is the tightest binding site of APC for beta-catenin. J Mol Biol 360, 133-144.

Liu, X., Rubin, J.S., and Kimmel, A.R. (2005). Rapid, Wnt-induced changes in GSK3beta associations that regulate beta-catenin stabilization are mediated by Galpha proteins. Curr Biol 15, 1989-1997.

MacDonald, B.T., and He, X. (2012). Frizzled and LRP5/6 receptors for Wnt/beta-catenin signaling. Cold Spring Harb Perspect Biol 4.

Malinauskas, T., and Jones, E.Y. (2014). Extracellular modulators of Wnt signalling. Curr Opin Struct Biol 29, 77-84.

Mao, J., Wang, J., Liu, B., Pan, W., Farr, G.H., 3rd, Flynn, C., Yuan, H., Takada, S., Kimelman, D., Li, L., et al. (2001). Low-density lipoprotein receptor-related protein-5 binds to Axin and regulates the canonical Wnt signaling pathway. Mol Cell 7, 801-809.

Mariotti, L., Pollock, K., and Guettler, S. (2017). Regulation of Wnt/beta-catenin signalling by tankyrase-dependent poly(ADP-ribosyl)ation and scaffolding. Br J Pharmacol 174, 4611- 4636.

Matsuzawa, S.I., and Reed, J.C. (2001). Siah-1, SIP, and Ebi collaborate in a novel pathway for beta-catenin degradation linked to p53 responses. Molecular Cell 7, 915-926.

Mattie, F.J., Stackpole, M.M., Stone, M.C., Clippard, J.R., Rudnick, D.A., Qiu, Y., Tao, J., Allender, D.L., Parmar, M., and Rolls, M.M. (2010). Directed microtubule growth, +TIPs, and kinesin-2 are required for uniform microtubule polarity in dendrites. Curr Biol 20, 2169-2177.

Mbom, B.C., Nelson, W.J., and Barth, A. (2013). beta-catenin at the centrosome: discrete pools of beta-catenin communicate during mitosis and may co-ordinate centrosome functions and cell cycle progression. Bioessays 35, 804-809.

McCartney, B.M., McEwen, D.G., Grevengoed, E., Maddox, P., Bejsovec, A., and Peifer, M. (2001). Drosophila APC2 and Armadillo participate in tethering mitotic spindles to cortical actin. Nature Cell Biology 3, 933-938.

McCartney, B.M., Price, M.H., Webb, R.L., Hayden, M.A., Holot, L.M., Zhou, M., Bejsovec, A., and Peifer, M. (2006). Testing hypotheses for the functions of APC family proteins using null and truncation alleles in Drosophila. Development 133, 2407-2418.

Mendoza-Topaz, C., Mieszczanek, J., and Bienz, M. (2011). The APC tumour suppressor is essential for Axin complex assembly and function, and opposes Axin’s interaction with Dishevelled. Biology Open 1.

Miller, J.R., Rowning, B.A., Larabell, C.A., Yang-Snyder, J.A., Bates, R.L., and Moon, R.T. (1999). Establishment of the Dorsal/Ventral Axis in Xenopus Embryos Coincides with the Dorsal Enrichment of Dishevelled That Is Dependent on Cortical Rotation. Journal of Cell Biology 146, 427-438.

Mittal, A., Holehouse, A.S., Cohan, M.C., and Pappu, R.V. (2018). Sequence-to-Conformation Relationships of Disordered Regions Tethered to Folded Domains of Proteins. J Mol Biol 430, 2403-2421.

Mlodzik, M. (2016). The Dishevelled Protein Family: Still Rather a Mystery After Over 20 Years of Molecular Studies. Curr Top Dev Biol 117, 75-91.

Munemitsu, S., Souza, B., Müller, O., Albert, I., Rubinfeld, B., and Polakis, P. (1994). The APC gene product associates with microtubules in vivo and promotes their assembly in vitro. Cancer Res 54, 3676-3681.

Nakagawa, N., Li, J., Yabuno-Nakagawa, K., Eom, T.Y., Cowles, M., Mapp, T., Taylor, R., and Anton, E.S. (2017). APC sets the Wnt tone necessary for cerebral cortical progenitor development. Genes Dev 31, 1679-1692.

Nelson, S., and Nathke, I.S. (2013). Interactions and functions of the adenomatous polyposis coli (APC) protein at a glance. J Cell Sci 126, 873-877.

Nusse, R., and Clevers, H. (2017). Wnt/beta-Catenin Signaling, Disease, and Emerging Therapeutic Modalities. Cell 169, 985-999.

Orsulic, S., and Peifer, M. (1996). An in vivo structure-function analysis of armadillo, the ß- catenin homologue, reveals both separate and overlapping regions of the protein required for cell adhesion and wingless signaling. Journal of Cell Biology 134, 1283-1301.

Peifer, M., Sweeton, D., Casey, M., and Wieschaus, E. (1994). wingless signal and Zeste-white 3 kinase trigger opposing changes in the intracellular distribution of Armadillo. Development 120, 369-380.

Peterson-Nedry, W., Erdeniz, N., Kremer, S., Yu, J., Baig-Lewis, S., and Wehrli, M. (2008). Unexpectedly robust assembly of the Axin destruction complex regulates Wnt/Wg signaling in Drosophila as revealed by analysis in vivo. Dev Biol 320, 226-241.

Piao, S., Lee, S.H., Kim, H., Yum, S., Stamos, J.L., Xu, Y., Lee, S.J., Lee, J., Oh, S., Han, J.K., et al. (2008). Direct inhibition of GSK3beta by the phosphorylated cytoplasmic domain of LRP6 in Wnt/beta-catenin signaling. PLoS One 3, e4046.

Piovesan, D., Tabaro, F., Paladin, L., Necci, M., Micetic, I., Camilloni, C., Davey, N., Dosztanyi, Z., Meszaros, B., Monzon, A.M., et al. (2018). MobiDB 3.0: more annotations for intrinsic disorder, conformational diversity and interactions in proteins. Nucleic Acids Res 46, D471- D476.

Poulton, J.S., Mu, F.W., Roberts, D.M., and Peifer, M. (2013). APC2 and Axin promote mitotic fidelity by facilitating centrosome separation and cytoskeletal regulation. Development 140, 4226-4236.

Pronobis, M.I., Deuitch, N., Posham, V., Mimori-Kiyosue, Y., and Peifer, M. (2017). Reconstituting regulation of the canonical Wnt pathway by engineering a minimal beta- catenin destruction machine. Mol Biol Cell 28, 41-53.

Pronobis, M.I., Rusan, N.M., and Peifer, M. (2015). A novel GSK3-regulated APC:Axin interaction regulates Wnt signaling by driving a catalytic cycle of efficient betacatenin destruction. eLife 4, e08022.

Rai, A.K., Chen, J.X., Selbach, M., and Pelkmans, L. (2018). Kinase-controlled phase transition of membraneless organelles in mitosis. Nature 559, 211-216.

Roberts, D.M., Pronobis Mi Fau - Alexandre, K.M., Alexandre Km Fau - Rogers, G.C., Rogers Gc Fau - Poulton, J.S., Poulton Js Fau - Schneider, D.E., Schneider De Fau - Jung, K.-C., Jung Kc Fau - McKay, D.J., McKay Dj Fau - Peifer, M., and Peifer, M. (2012a). Defining components of the beta-catenin destruction complex and exploring its regulation and

mechanisms of action during development. PLoS ONE 7, e31284.

Roberts, D.M., Pronobis, M.I., Poulton, J.S., Kane, E.G., and Peifer, M. (2012b). Regulation of Wnt signaling by the tumor suppressor APC does not require ability to enter the nucleus nor a particular cytoplasmic localization. Mol Biol Cell 23, 2041-2056.

Roberts, D.M., Pronobis, M.I., Poulton, J.S., Waldmann, J.D., Stephenson, E.M., Hanna, S., and Peifer, M. (2011). Deconstructing the beta-catenin destruction complex: mechanistic roles for the tumor suppressor APC in regulating Wnt signaling. Mol Biol Cell 22, 1845-1863.

Rubinfeld, B., Souza, B., Albert, I., Muller, O., Chamberlain, S.H., Masiarz, F.R., Munemitsu, S., and Polakis, P. (1993). The APC gene product associates with ß-catenin. Science 262, 1731-1734.

Rusan, N.M., Akong, K., and Peifer, M. (2008). Putting the model to the test: are APC proteins essential for neuronal polarity, axon outgrowth, and axon targeting? J Cell Biol 183, 203- 212.

Rusan, N.M., and Peifer, M. (2008). Original CIN: reviewing roles for APC in chromosome instability. J Cell Biol 181, 719-726.

Saito-Diaz, K., Benchabane, H., Tiwari, A., Tian, A., Li, B., Thompson, J.J., Hyde, A.S., Sawyer, L.M., Jodoin, J.N., Santos, E., et al. (2018). APC Inhibits Ligand-Independent Wnt Signaling by the Clathrin Endocytic Pathway. Dev Cell 44, 566-581 e568.

Sakanaka, C., Leong, P., Xu, L., Harrison, S.D., and Williams, L.T. (1999). Casein kinase I- epsilon in the Wnt pathway: Regulation of beta-catenin function. Proc Nat Acad Sci USA 96, 12548-12552.

Sakanaka, C., Weiss, J.B., and Williams, L.T. (1998). Bridging of ß-catenin and glycogen synthase kinase-3ß by axin and inhibition of ß-catenin-mediated transcription. Proc Nat Acad Sci USA 95, 3020-3023.

Sakanaka, C., and Williams, L.T. (1999). Functional domains of axin. Importance of the C terminus as an oligomerization domain. J Biol Chem 274, 14090-14093.

Salic, A., Lee, E., Mayer, L., and Kirschner, M.W. (2000). Control of beta-catenin stability: reconstitution of the cytoplasmic steps of the wnt pathway in Xenopus egg extracts. Mol Cell 5, 523-532.

Sarpal, R., Pellikka, M., Patel, R.R., Hui, F.Y., Godt, D., and Tepass, U. (2012). Mutational analysis supports a core role for Drosophila alpha-catenin in adherens junction function. J Cell Sci 125, 233-245.

Schaefer, K.N., Bonello, T.T., Zhang, S., Williams, C.E., Roberts, D.M., McKay, D.J., and Peifer, M. (2018). Supramolecular assembly of the beta-catenin destruction complex and the effect of Wnt signaling on its localization, molecular size, and activity in vivo. PLoS Genet 14, e1007339.

Schwarz-Romond, T., Fiedler, M., Shibata, N., Butler, P.J., Kikuchi, A., Higuchi, Y., and Bienz, M. (2007a). The DIX domain of Dishevelled confers Wnt signaling by dynamic polymerization. Nat Struct Mol Biol 14, 484-492.

Schwarz-Romond, T., Merrifield, C., Nichols, B.J., and Bienz, M. (2005). The Wnt signalling effector Dishevelled forms dynamic protein assemblies rather than stable associations with cytoplasmic vesicles. J Cell Sci 118, 5269-5277.

Schwarz-Romond, T., Metcalfe, C., and Bienz, M. (2007b). Dynamic recruitment of axin by Dishevelled protein assemblies. J Cell Sci 120, 2402-2412.

Shan, Z., Tu, Y., Yang, Y., Liu, Z., Zeng, M., Xu, H., Long, J., Zhang, M., Cai, Y., and Wen, W. (2018). Basal condensation of Numb and Pon complex via phase transition during Drosophila neuroblast asymmetric division. Nat Commun 9, 737.

Siegfried, E., Chou, T.-B., and Perrimon, N. (1992). wingless signaling acts through zeste-white 3, the Drosophila homolog of glycogen synthase kinase-3, to regulate engrailed and establish cell fate. Cell 71, 1167-1179.

Siegfried, E., Perkins, L., Capaci, T., and Perrimon, N. (1990). Putative protein kinase product of the Drosophila segment polarity gene zeste-white 3. Nature 345, 825-829.

Siegfried, E., Wilder, E., and Perrimon, N. (1994). Components of wingless signaling in Drosophila. Nature 367, 76-80.

Sierra, J., Yoshida, T., Joazeiro, C.A., and Jones, K.A. (2006). The APC tumor suppressor counteracts beta-catenin activation and H3K4 methylation at Wnt target genes. Genes Dev 20, 586-600.

Smalley, M.J., Sara, E., Paterson, H., Naylor, S., Cook, D., Jayatilake, H., Fryer, L.G., Hutchinson, L., Fry, M.J., and Dale, T.C. (1999). Interaction of Axin and Dvl-2 proteins regulates Dvl-2-stimulated TCF-dependent transcription. EMBO Journal 18, 2823-2835.

Smith, K.J., Levy, D.B., Maupin, P., Pollard, T.D., Vogelstein, B., and Kinzler, K.W. (1994). Wild- type but not mutant APC associates with the microtubule cytoskeleton. Cancer Res 54, 3672-3675.

Smits, R., Kielman, M.F., Breukel, C., Zurcher, C., Neufeld, K., Jagmohan-Changur, S., Hofland, N., van Dijk, J., White, R., Edelmann, W., et al. (1999). Apc1638T: a mouse model delineating critical domains of the adenomatous polyposis coli protein involved in tumorigenesis and development. Genes Dev 13, 1309-1321.

Spink, K.E., Polakis, P., and Weis, W.I. (2000). Structural basis of the Axin-adenomatous polyposis coli interaction. Embo J 19, 2270-2279.

Stamos, J.L., Chu, M.L., Enos, M.D., Shah, N., and Weis, W.I. (2014). Structural basis of GSK-3 inhibition by N-terminal phosphorylation and by the Wnt receptor LRP6. eLife 3, e01998.

Stamos, J.L., and Weis, W.I. (2013). The beta-catenin destruction complex. Cold Spring Harb Perspect Biol 5, a007898.

Su, L.K., Burrell, M., Hill, D.E., Gyuris, J., Brent, R., Wiltshire, R., Trent, J., Vogelstein, B., and Kinzler, K.W. (1995). APC binds to the novel protein EB1. Cancer Res 55, 2972-2977.

Sugioka, K., Fielmich, L.E., Mizumoto, K., Bowerman, B., van den Heuvel, S., Kimura, A., and Sawa, H. (2018). Tumor suppressor APC is an attenuator of spindle-pulling forces during C. elegans asymmetric cell division. Proc Natl Acad Sci U S A 115, E954-E963.

Sugioka, K., Mizumoto, K., and Sawa, H. (2011). Wnt regulates spindle asymmetry to generate asymmetric nuclear beta-catenin in C. elegans. Cell 146, 942-954.

Tacchelly-Benites, O., Wang, Z., Yang, E., Benchabane, H., Tian, A., Randall, M.P., and Ahmed, Y. (2018). Axin phosphorylation in both Wnt-off and Wnt-on states requires the tumor suppressor APC. PLoS Genet 14, e1007178.

Takacs, C.M., Baird, J.R., Hughes, E.G., Kent, S.S., Benchabane, H., Paik, R., and Ahmed, Y. (2008). Dual positive and negative regulation of wingless signaling by adenomatous polyposis coli. Science 319, 333-336.

Tamai, K., Zeng, X., Liu, C., Zhang, X., Harada, Y., Chang, Z., and He, X. (2004). A mechanism for Wnt coreceptor activation. Mol Cell 13, 149-156.

Tan, C.W., Gardiner, B.S., Hirokawa, Y., Layton, M.J., Smith, D.W., and Burgess, A.W. (2012). Wnt signalling pathway parameters for mammalian cells. PLoS One 7, e31882.

Thorvaldsen, T.E., Pedersen, N.M., Wenzel, E.M., Schultz, S.W., Brech, A., Liestol, K., Waaler, J., Krauss, S., and Stenmark, H. (2015). Structure, Dynamics, and Functionality of Tankyrase Inhibitor-Induced Degradasomes. Mol Cancer Res 13, 1487-1501.

Tolwinski, N.S., Wehrli, M., Rives, A., Erdeniz, N., DiNardo, S., and Wieschaus, E. (2003). Wg/Wnt signal can be transmitted through arrow/LRP5,6 and Axin independently of Zw3/Gsk3beta activity. Dev Cell 4, 407-418.

Tolwinski, N.S., and Wieschaus, E. (2001). Armadillo nuclear import is regulated by cytoplasmic anchor Axin and nuclear anchor dTCF/Pan. Development 128, 2107-2117.

Waaler, J., Machon, O., von Kries, J.P., Wilson, S.R., Lundenes, E., Wedlich, D., Gradl, D., Paulsen, J.E., Machonova, O., Dembinski, J.L., et al. (2011). Novel synthetic antagonists of canonical Wnt signaling inhibit colorectal cancer cell growth. Cancer Res 71, 197-205. Wang, J., Choi, J.M., Holehouse, A.S., Lee, H.O., Zhang, X., Jahnel, M., Maharana, S., Lemaitre, R., Pozniakovsky, A., Drechsel, D., et al. (2018). A Molecular Grammar Governing the Driving Forces for Phase Separation of Prion-like RNA Binding Proteins. Cell 174, 688- 699 e616. Wang, Z., Tacchelly-Benites, O., Yang, E., and Ahmed, Y. (2016a). Dual Roles for Membrane Association of Drosophila Axin in Wnt Signaling. PLoS Genet 12, e1006494.

Wang, Z., Tacchelly-Benites, O., Yang, E., Thorne, C.A., Nojima, H., Lee, E., and Ahmed, Y. (2016b). Wnt/Wingless Pathway Activation Is Promoted by a Critical Threshold of Axin Maintained by the Tumor Suppressor APC and the ADP-Ribose Polymerase Tankyrase. Genetics 203, 269-281.

Wang, Z., Tian, A., Benchabane, H., Tacchelly-Benites, O., Yang, E., Nojima, H., and Ahmed, Y. (2016c). The ADP-ribose polymerase Tankyrase regulates adult intestinal stem cell proliferation during homeostasis in Drosophila. Development 143, 1710-1720.

Webb, R.L., Zhou, M.N., and McCartney, B.M. (2009). A novel role for an APC2-Diaphanous complex in regulating actin organization in Drosophila. Development 136, 1283-1293.

Weiner, A.T., Lanz, M.C., Goetschius, D.J., Hancock, W.O., and Rolls, M.M. (2016). Kinesin-2 and Apc function at dendrite branch points to resolve microtubule collisions. Cytoskeleton (Hoboken) 73, 35-44.

Willert, K., Shibamoto, S., and Nusse, R. (1999). Wnt-induced dephosphorylation of axin releases beta-catenin from the axin complex. Genes Dev 13, 1768-1773.

Wong, H.C., Bourdelas, A., Krauss, A., Lee, H.J., Shao, Y., Wu, D., Mlodzik, M., Shi, D.L., and Zheng, J. (2003). Direct binding of the PDZ domain of Dishevelled to a conserved internal sequence in the C-terminal region of Frizzled. Mol Cell 12, 1251-1260.

Wu, G., Huang, H., Garcia Abreu, J., and He, X. (2009). Inhibition of GSK3 phosphorylation of beta-catenin via phosphorylated PPPSPXS motifs of Wnt coreceptor LRP6. PLoS One 4, e4926.

Xing, Y., Clements, W.K., Kimelman, D., and Xu, W. (2003). Crystal structure of a beta- catenin/axin complex suggests a mechanism for the beta-catenin destruction complex. Genes Dev 17, 2753-2764.

Xing, Y., Clements, W.K., Le Trong, I., Hinds, T.R., Stenkamp, R., Kimelman, D., and Xu, W. (2004). Crystal structure of a beta-catenin/APC complex reveals a critical role for APC phosphorylation in APC function. Mol Cell 15, 523-533.

Yamamoto, H., Kishida, S., Kishida, M., Ikeda, S., Takada, S., and Kikuchi, A. (1999). Phosphorylation of axin, a Wnt signal negative regulator, by glycogen synthase kinase- 3beta regulates its stability. J Biol Chem 274, 10681-10684.

Yamashita, Y.M., Jones, D.L., and Fuller, M.T. (2003). Orientation of Asymmetric Stem Cell Division by the APC Tumor Suppressor and Centrosome. Science 301, 1547-1550.

Yamulla, R.J., Kane, E.G., Moody, A.E., Politi, K.A., Lock, N.E., Foley, A.V., and Roberts, D.M. (2014). Testing models of the APC tumor suppressor/beta-catenin interaction reshapes our view of the destruction complex in Wnt signaling. Genetics 197, 1285-1302.

Yang, E., Tacchelly-Benites, O., Wang, Z., Randall, M.P., Tian, A., Benchabane, H., Freemantle, S., Pikielny, C., Tolwinski, N.S., Lee, E., et al. (2016). Wnt pathway activation by ADP-ribosylation. Nat Commun 7, 11430.

Yang, J., Zhang, W., Evans, P.M., Chen, X., He, X., and Liu, C. (2006). Adenomatous polyposis coli (APC) differentially regulates beta-catenin phosphorylation and ubiquitination in colon cancer cells. J Biol Chem 281, 17751-17757.

Yang-Snyder, J., Miller, J.R., Brown, J.D., Lai, C.-J., and Moon, R.T. (1996). A frizzled homolog functions in a vertebrate Wnt signaling pathway. Current Biology 6, 1302-1306.

Yokota, Y., Kim, W.Y., Chen, Y., Wang, X., Stanco, A., Komuro, Y., Snider, W., and Anton, E.S. (2009). The adenomatous polyposis coli protein is an essential regulator of radial glial polarity and construction of the cerebral cortex. Neuron 61, 42-56.

Zeng, L., Fagotto, F., Zhang, T., Hsu, W., Vasicek, T.J., Perry, W.L., 3rd, Lee, J.J., Tilghman, S.M., Gumbiner, B.M., and Costantini, F. (1997). The mouse Fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation. Cell 90, 181- 192.

Zeng, M., Chen, X., Guan, D., Xu, J., Wu, H., Tong, P., and Zhang, M. (2018). Reconstituted Postsynaptic Density as a Molecular Platform for Understanding Synapse Formation and Plasticity. Cell 174, 1172-1187.e1116.

Zeng, X., Tamai, K., Doble, B., Li, S., Huang, H., Habas, R., Okamura, H., Woodgett, J., and He, X. (2005). A dual-kinase mechanism for Wnt co-receptor phosphorylation and activation. Nature 438, 873-877.

Zhang, L., and Shay, J.W. (2017). Multiple Roles of APC and its Therapeutic Implications in Colorectal Cancer. J Natl Cancer Inst 109.

Zhang, Y., Liu, S., Mickanin, C., Feng, Y., Charlat, O., Michaud, G.A., Schirle, M., Shi, X., Hild, M., Bauer, A., et al. (2011). RNF146 is a poly(ADP-ribose)-directed E3 ligase that regulates axin degradation and Wnt signalling. Nat Cell Biol 13, 623-629.

Zhou, Z.D., Chan, C.H., Xiao, Z.C., and Tan, E.K. (2011). Ring finger protein 146/Iduna is a poly(ADP-ribose) polymer binding and PARsylation dependent E3 ubiquitin ligase. Cell adhesion & migration 5, 463-471.

CHAPTER 2: SUPRAMOLECULAR ASSEMBLY OF THE BETA-CATENIN DESTRUCTION COMPLEX AND THE EFFECT OF WNT SIGNALING ON ITS LOCALIZATION, MOLECULAR SIZE, AND ACTIVITY IN VIVO1

OVERVIEW:

Wnt signaling provides a paradigm for cell-cell signals that regulate embryonic development and stem cell homeostasis and are inappropriately activated in cancers. The tumor suppressors APC and Axin form the core of the multiprotein destruction complex, which targets the Wnt-effector beta-catenin for phosphorylation, ubiquitination and destruction. Based on earlier work, we hypothesize that the destruction complex is a supermolecular entity that self- assembles by Axin and APC polymerization, and that regulating assembly and stability of the destruction complex underlie its function. We tested this hypothesis in Drosophila embryos, a premier model of Wnt signaling. Combining biochemistry, genetic tools to manipulate Axin and

APC2 levels, advanced imaging and molecule counting, we defined destruction complex assembly, stoichiometry, and localization in vivo, and its downregulation in response to Wnt signaling. Our findings challenge and revise current models of destruction complex function.

Endogenous Axin and APC2 proteins and their antagonist Dishevelled accumulate at roughly similar levels, suggesting competition for binding may be critical. By expressing Axin:GFP

1This chapter previously appeared as an article in the PLoS Genetics. The original citation is as follows: Kristina N. Schaefer, Teresa T. Bonello , Shiping Zhang , Clara E. Williams, David M.

Roberts, Daniel J. McKay, Mark Peifer. “Supramolecular assembly of the beta-catenin destruction complex and the effect of Wnt signaling on its localization, molecular size, and activity in vivo, PLoS Genetics 14, e1007339

46 at near endogenous levels we found that in the absence of Wnt signals, Axin and APC2 co- assemble into large cytoplasmic complexes containing tens to hundreds of Axin proteins. Wnt signals trigger recruitment of these to the membrane, while cytoplasmic Axin levels increase, suggesting altered assembly/disassembly. Glycogen synthase kinase3 regulates destruction complex recruitment to the membrane and release of Armadillo/beta-catenin from the destruction complex. Manipulating Axin or APC2 levels had no effect on destruction complex activity when Wnt signals were absent, but, surprisingly, had opposite effects on the destruction complex when Wnt signals were present. Elevating Axin made the complex more resistant to inactivation, while elevating APC2 levels enhanced inactivation. Our data suggest both absolute levels and the ratio of these two core components affect destruction complex function, supporting models in which competition among Axin partners determines destruction complex activity.

Author Summary

Cell-cell communication is critical for cells to choose fates during embryonic development and often goes wrong in diseases like cancer. The Wnt cell signaling pathway provides a superb example. Loss of negative regulatory proteins like APC and Axin takes the brakes off cell proliferation and thus contributes to colon cancer. We study how APC, Axin and their partners keep cell signaling off, and how cell-to-cell Wnt signals reverse this. We use the fruit fly embryo, combining biochemical and genetic tools with advanced microscopy. We found that the destruction complex proteins APC2, Axin, and their antagonist Dishevelled are present at similar levels, allowing them to effectively compete with one another. We further find that the ability of Wnt signaling to turn off the negative regulatory destruction complex machine is influenced both by the levels of Axin and APC2 and by the ratio of their levels. We visualize the active destruction complex in the animal, and count the number of Axin proteins in this complex.

Finally, we find that Wnt signals have two effects on the destruction complex—recruiting it to the

46 plasma membrane and altering its assembly/disassembly. We then propose a new model for how this important signaling pathway is regulated.

INTRODUCTION

Cell-cell signaling is critical for cell fate decisions during embryonic development and cell fate maintenance during adult homeostasis. Altered signaling by these same pathways underlies most solid tumors. The Wnt signaling pathway provides a paradigm—it regulates cell fate choice in tissues throughout the body, maintains stem cell identity in many adult tissues, and is inappropriately activated in colorectal and other cancers (Clevers and Nusse, 2012).

Thus, understanding the mechanisms by which signaling occurs and is regulated are key issues for cell, developmental, and cancer biology.

Work in both animal models and cultured mammalian cells provided a broad outline of

Wnt signaling and its regulation (Gammons and Bienz, 2018; Nusse and Clevers, 2017). The key effector is the transcriptional co-activator β-catenin (βcat; Drosophila Armadillo; Arm). In the absence of signaling, βcat is captured by a multiprotein complex called the destruction complex.

The scaffold proteins Adenomatous polyposis coli (APC) and Axin bind βcat and present it to the kinases glycogen synthase kinase-3 (GSK3) and casein kinase 1 (CK1). They phosphorylate βcat, creating a binding site for an E3 ubiquitin ligase, thus targeting βcat for proteasomal destruction. When Wnt ligands bind to receptors, the destruction complex is downregulated, allowing βcat to accumulate, enter the nucleus and act together with the DNA binding proteins in the TCF/LEF family to transcriptionally activate Wnt-regulated genes.

Work in cultured mammalian cells has added important aspects to this model (Nusse and Clevers, 2017)—here we focus on the action of the destruction complex and its regulation by Wnt signaling. Several different mechanisms have emerged by which Wnt signaling downregulate βcat destruction and thus activate downstream signaling. Wnt binding to the

Frizzled:LRP5/6 receptors triggers assembly of the receptors along with the Wnt effector

Dishevelled (Dvl; Drosophila Dsh) into a higher order signalosome (Bilic et al., 2007; Gammons 47 et al., 2016a; Gammons et al., 2016b; Metcalfe et al., 2010). LRP5/6 becomes phosphorylated and recruits the destruction complex to the plasma membrane, at least in part by interactions between the phosphorylated tail of LRP5/6 and Axin (Zeng et al., 2005). The phosphorylated

LRP5/6 tail can directly inhibit GSK3 (Stamos et al., 2014). Alternate mechanisms for destruction complex inhibition also exist. Dsh can co-polymerize with Axin via their shared DIX domains, antagonizing its function (Fiedler et al., 2011). Careful kinetic analysis revealed that

Wnt stimulation reduces the rate of ßcat phosphorylation by both CK1 and GSK3, reducing but not eliminating destruction complex activity (Hernandez et al., 2012). Wnt signaling can trigger

Axin dephosphorylation, reducing its interaction with both ßcat and LRP5/6, thus reducing ßcat destruction (Kim et al., 2013). Finally, another study suggested that after Wnt signaling the destruction complex remains intact and capable of phosphorylating βcat, but its transfer to the

E3 ligase is prevented (Li et al., 2012). These studies provide important insights into key regulatory mechanisms by which Wnt signaling can inactivate the destruction complex, but leave as an open question which mechanism(s) is most prominent during signaling in vivo.

The Wnt pathway is part of an emerging theme in cell signaling, in which self-assembly of multiprotein supermolecular signaling hubs creates non-membrane bound cellular compartments (Toretsky and Wright, 2014). Three key steps in Wnt signaling are catalyzed by distinct supermolecular machines—the signalasome, involved in Wnt reception and destruction complex downregulation, the destruction complex itself, and the enhancesome, which mediates

Wnt-regulated gene expression (Gammons and Bienz, 2018). Key questions remain about the mechanism by which the active destruction complex targets ßcat for destruction in the absence of Wnt signaling. APC was originally viewed as the scaffold around which the destruction complex assembled, but subsequent work revealed that Axin fulfills this function, leaving APC’s molecular role a mystery. Further, while the destruction complex is typically represented in models as a simple four-protein complex, considerable evidence supports the idea that it is a large supermolecular protein assembly, built by self-polymerization of Axin and APC

(e.g.,(Kishida et al., 1999; Kunttas-Tatli et al., 2014; Pronobis et al., 2015; Schwarz-Romond et al., 2007a)).

Recent work provided new mechanistic insights into the molecular mechanisms by which

APC functions, helping begin to transform the static, low-resolution textbook model of Wnt signaling into a more dynamic, high resolution view. Super-resolution microscopy of Axin and

APC complexes assembled after overexpression in colorectal cancer cells provided the first look inside the active destruction complex. Axin and APC containing “puncta” were resolved into intertwined strands of each protein, presumably assembled by polymerization (Pronobis et al.,

2015). Combining this with assessment of APC and Axin dynamics and genetic and biochemical dissection of the two proteins provided novel mechanistic insights and a new model. First, they suggest APC promotes/stabilizes Axin multimerization, thus increasing destruction complex efficiency (Pronobis et al., 2015). Second, they revealed a key role for two peptide motifs in

APC, 20 amino acid repeat 2 and sequence B/the CID, both essential for destruction complex function (Kohler et al., 2009; Roberts et al., 2011). These motifs appear to play two roles. They are binding sites for alpha-catenin, stabilizing ßcat association with APC and preventing its dephosphorylation (Choi et al., 2013). After Axin-mediated βcat phosphorylation, these APC motifs are also phosphorylated, triggering a regulated conformational change that transfers βcat out of the destruction complex to the E3 ligase, to restart the catalytic cycle (Pronobis et al.,

2015). These data fit with other studies suggesting that Wnt signaling does not totally turn off the destruction complex, but reduces the rate of destruction. Instead, the destruction complex remains intact and capable of phosphorylating βcat, but βCat transfer to the E3 ligase is inhibited (Hernandez et al., 2012; Kim et al., 2013; Li et al., 2012).

However, this work was largely done in cultured cells, which provide a simple place to explore pathway circuitry but do not provide a physiologically relevant situation with all regulatory mechanisms intact. We thus took these insights back into the Drosophila embryonic epidermis, arguably the system where our understanding of the roles and regulation of the Wnt

pathway is strongest. Stripes of cells in each body segment produce a fly Wnt, Wingless (Wg),

creating a field of cells experiencing high, moderate and low levels of Wg signaling. Taking

advantage of new genetic approaches and high-resolution microscopy, we addressed several

key issues in the field, exploring the structure, assembly and stoichiometry of the destruction

complex in vivo during normal development and how it is downregulated by Wnt signaling.

To understand a complex multiprotein machine, one key issue involves the relative levels of its

component parts. Most current models of Wnt regulation suggest Axin accumulates at levels

dramatically lower than those of other proteins in the destruction complex. This hypothesis

derives from influential early work in Xenopus oocyte extracts. By adding in known amounts of recombinant Axin and measuring the resulting destruction complex activity, they estimated Axin concentrations were as much as 5000-fold lower than those of APC and other destruction

complex proteins. Their mathematical model of Wnt signaling and many subsequent ones are

based on these estimates (Lee et al., 2001; Lee et al., 2003; Mirams et al., 2010). In contrast,

recent work in cultured mammalian cells suggests Axin and APC levels are more similar (Tan et

al., 2012). Thus, defining the relative levels of Axin and APC in tissues undergoing Wnt

signaling in vivo is a key issue, and the Drosophila embryo provided a superb place to

accomplish this end.

With relative protein levels defined, different models for the function and regulation of the

destruction complex can be tested by varying absolute levels of Axin or APC and their relative

ratios to one another. Substantially elevating Axin levels in Drosophila embryos strongly inhibits

Wnt signaling (Cliffe et al., 2003). Further analyses suggested there is a threshold below which

elevating Axin does not substantially alter signaling, since more subtle elevation of Axin levels

(2–5 fold) had little effect in Drosophila embryos or imaginal discs (Peterson-Nedry et al., 2008;

Wang et al., 2016b; Yang et al., 2016) and mutating tankyrase, which elevates Axin levels 2–3 fold, does not substantially perturb Wnt signaling (Feng et al., 2014; Yang et al., 2016). In contrast, a 9-fold increase in Axin levels inhibited Wnt signaling in imaginal discs (Wang et al.,

2016b). However, these studies used multiple tissues or systems in parallel, and left the mechanisms underlying the dose-sensitive response unclear. The Drosophila embryo provided a place to assess how altering Axin levels affects cell fate choice, Wnt-target gene expression and ßcat levels in parallel, and to directly compare effects on cells receiving and not receiving

Wnt signals. It also offered the opportunity to manipulate APC levels, the other key scaffolding protein in the destruction complex. Whether APC levels are rate-limiting remains an open question, because APC has been viewed as present in substantial excess. The Drosophila embryo also allowed us to test effects of varying the Axin:APC ratio, another key parameter of any molecular model.

Finally, to effectively understand destruction complex assembly and function, we need to visualize it directly. Our recent super-resolution imaging of Axin:APC puncta in cultured cells provided the first insights into the internal structure and dynamics of this multiprotein machine, but these experiments involved significant over-expression. The Drosophila embryo provided a place to assess whether similar complexes assemble at near endogenous levels. Recent advances in molecular counting technology also offered the possibility of directly assessing the number of Axin proteins assembled in a complex.

Visualizing the destruction complex in the embryo would also allow us to address how

Wnt signaling inactivates it. Work in cultured cells led to a model in which Wnt binding the

Frizzled:LRP5/6 receptor complex triggers LRP5/6 phosphorylation, and Axin and Dvl/Dsh membrane recruitment (MacDonald and He, 2012). What happens next is disputed, with many events suggested to play a part. For example, some data suggest the destruction complex is disassembled because Dsh competes for Axin (Fiedler et al., 2011) or Wnt signaling destabilizes Axin (Mao et al., 2001). Interestingly, examining effects of Wg signaling on the destruction complex in Drosophila embryos led to starkly divergent conclusions. One group reported that Wg signaling strongly reduced Axin levels, as assessed both by immunofluorescence and immunoblotting (Tolwinski et al., 2003). A second, visualizing GFP-

tagged Axin, found little or no effect of Wg on Axin levels—instead their data suggested that Wg

signaling causes a Dsh-dependent relocalization of Axin from cytoplasmic puncta to the plasma

membrane (Cliffe et al., 2003). Finally, a third group reported that Wg signaling initially stabilizes Axin, as assessed by immunofluorescence, increasing both membrane bound and

cytoplasmic pools (Wang et al., 2016a; Yang et al., 2016). Thus, the effects of Wg signaling on

Axin, a key part of the mechanism underlying βcat stabilization, also remain an open question.

Our system, allowing direct detection of fluorescently-tagged Axin expressed at near

endogenous levels, allowed us to address this issue.

RESULTS

axin and APC1/APC2 are transcribed at similar levels

Most current models of Wnt regulation suggest Axin accumulates at levels dramatically

lower than those of other destruction complex proteins, potentially making destruction complex

activity sensitive to very small increases in its levels. However, the literature contains indications

that this is not universally true (e.g. (Tan et al., 2012)). To better understand how APC and Axin

levels affect Wnt signaling in vivo we directly compared levels of APC family members and Axin

in Drosophila embryos.

We first compared mRNA levels of Drosophila axin with those encoding the two fly APC

family proteins, APC1 and APC2, using RNAseq data from staged embryos. In embryos, APC2

plays the predominant role in Wnt regulation during early to mid-embryogenesis ((Ahmed et al.,

2002; Akong et al., 2002a; McCartney et al., 1999); 2–4 or 6–8 hours after egg laying,

respectively), while APC1 is expressed at low levels early but becomes prominent later in the

central nervous system (Ahmed et al., 1998; Akong et al., 2002b). Consistent with

this, APC2 mRNA levels are ~19x higher than APC1 during early embryogenesis, and ~7x

higher during mid-embryogenesis (484 versus 26 Fragments Per Kilobase of transcript per

Million mapped reads (FPKM), and 201 versus 27 FPKM, respectively). However, in late

Figure 2.1

Figure 2.1: Endogenous APC2 and Axin proteins accumulate at similar levels. (A) mRNA levels (RNAseq) of APC1 (light blue), APC2 (blue), and Axin (yellow) during Drosophila embryogenesis. Levels are Fragments Per Kilobase of transcript per Million mapped reads (FPKM). (B-D) Immunoblots, 4-8hr old Drosophila embryos. Tubulin is loading control. n = # of blots quantified (S1 Table). (B) Anti-Axin antibody. Endogenous Axin levels versus those in Axin RNAi or Zyg Axin:GFP embryos. Endogenous Axin runs as doublet ~75kDa (red arrowheads) while Axin:GFP runs at ~105kDa (yellow arrowhead). * = background band. (C) Anti-APC2 antibody. Endogenous APC2 levels versus those of a GFP:APC2 transgene expressed under its endogenous promoter in an APC2 null (APC2g10) background. (D) Anti-GFP antibody. Relative levels of GFP:APC2 expressed under its endogenous promoter versus Zyg Axin:GFP.

53 embryogenesis, as the nervous system is assembled, APC1 mRNA levels are ~5x more abundant than APC2 (120 vs. 23 FPKM). Since APC2 and APC1 can act redundantly in regulating Wnt signaling (Ahmed et al., 2002; Akong et al., 2002a), we compared axin mRNA levels with combined mRNA abundance of APC1 plus APC2. Surprisingly, RNAseq reads for axin were roughly comparable to those of APC1 plus APC2 at three different stages of embryonic development (Fig2.1A), indicating that there are not dramatic differences between

APC family members versus Axin at the mRNA level.

Axin and APC2 proteins accumulate at similar levels during early-mid embryogenesis

These data did not rule out differences in protein translation or stability. To determine if similar mRNA levels led to similar protein levels, we compared Axin and APC2 protein levels in early to mid-embryogenesis (4–8 hrs), when APC2 is the predominant family member expressed. Since antibodies to APC2 and Axin may have different affinities, one cannot simply compare antibody-labeled endogenous proteins. To overcome this, we utilized GFP-tagged proteins expressed at near-endogenous levels. This allowed us to compare endogenous versus

GFP-tagged Axin, or endogenous versus GFP-tagged APC2 proteins, using antibodies against the endogenous proteins, followed by comparing GFP-tagged Axin and GFP-tagged APC2 proteins, using anti-GFP antibodies. We used the GAL4-UAS system (Brand and Perrimon,

1993; Rorth, 1998) to express Axin:GFP, using the driver that gave the lowest level of Axin:GFP expression (act5c-GAL4 provided by male parents). Axin:GFP was expressed at 1.0±0.5 fold that of endogenous Axin, as assessed by immunoblotting with anti-Axin antibodies (Fig

2.1B, Table 2.1). We next used transgenic flies expressing GFP:APC2 under control of the endogenous APC2 promotor, in an APC2 null mutant background (Roberts et al., 2011). Using anti-APC2 antibodies, we re-confirmed that APC2-driven GFP:APC2 was expressed at the same level as endogenous APC2 (0.9±0.4 fold endogenous APC2; Fig 2.1C, Table1.1). To complete the comparison, we then compared APC2-driven GFP:APC2 to Axin:GFP driven by

zygotic act5c-GAL4. Immunoblotting with anti-GFP antibodies revealed that GFP:APC2 is

expressed ~4-fold the levels of Axin:GFP (Fig 2.1D; 4.3±1.4; Table 2.1). These three

comparisons—endogenous Axin to act5c-GAL4 driven Axin:GFP, act5c-GAL4 x Axin:GFP

to APC2-driven GFP:APC2, and APC2-driven GFP:APC2 to endogenous APC2—provided a

reasonable estimate of the relative levels of endogenous APC2 to Axin: APC2 accumulates at a

~5-fold higher level than Axin (4.7±1.4). This is in contrast to the 5000-fold difference in

accumulation observed in Xenopus extracts that forms the basis of some current models, but is

consistent with the similar levels of mRNAs revealed by RNAseq.

Developing methods to vary Axin levels during embryogenesis

Axin is the key scaffold on which the destruction complex is built, and thus most models

of Wnt signaling suggest Axin is rate limiting for destruction complex function. Previous

experiments in fly embryos and imaginal discs strongly support this, as over-expressing Axin

can shut down Wnt signaling. Our knowledge of the relative levels of APC2 versus Axin in

the Drosophila embryonic epidermis allowed us to confirm and extend the analysis. We first

developed ways to vary Axin levels systematically, exploring how increasing Axin levels to

different degrees altered viability, cell fate and expression of a Wg target gene. We next

explored the underlying mechanism, by examining how different Axin levels affected destruction

complex activity and ßcat levels, both in cells receiving and not receiving Wg signals. We then

brought APC2 into this picture, examining effects of elevating APC2 levels, and of altering the

ratios of Axin to APC2.

To manipulate Axin levels systematically, we used the GAL4-UAS system. Four crosses

using two different GAL4 drivers provided different levels and timing of Axin over-expression (

Fig 2.2; Methods; Table 2.1). act5c-GAL4 is expressed during oogenesis and relatively

ubiquitously during embryonic development. 1. By crossing UAS-Axin:GFP females to act5c-

GAL4/+ males, we achieved lower-level and later elevation of Axin:GFP levels, which was driven by zygotically-expressed GAL4 (hereafter Zyg Axin). 2. By crossing act5c-GAL4/+

Figure 2.2

Figure 2.2: Crosses used to achieve different level and timing of Axin elevation. Crosses and expected progeny are diagrammed, along with features of expression characteristic of the GAL4 driver used.

females to UAS-Axin:GFP males (hereafter Mat/Zyg Axin), we achieved relatively high-level

overexpression, which began early due to maternally-contributed GAL4 and continued

zygotically. The second GAL4 driver stock was MatGAL4, which includes two GAL4 lines

expressed during oogenesis; they are not expressed zygotically but maternally expressed GAL4

protein perdures in the embryo. 3. For maternal and zygotic over-expression, we assessed

progeny of females trans-heterozygous for MatGAL4 and UAS-Axin:GFP (hereafter Mat Axin).

4. To achieve levels of Axin elevation intermediate between that produced by Zyg Axin and Mat

Axin, we used MatGAL4 to co-express UAS-Axin:GFP with a second UAS-driven transgene

encoding RFP (hereafter Mat RFP&Axin). When two different UAS-driven transgenes are

present, this reduces expression of both transgenes. We directly measured protein levels by

immunoblotting with antibodies to either GFP or to endogenous Axin.

These four schemes produced an excellent range of Axin expression levels in stage 9

embryos, when Wnt signaling is at its peak. Zyg Axin effectively tripled normal Axin levels in

embryos in which it was expressed (Fig 2.3A–2C, Table 2.1; taking into account endogenous

Axin and the fact that only 50% of embryos inherit the GAL4 driver). Mat RFP&Axin led to an

~4-fold increase, while both Mat/Zyg Axin and Mat Axin led to 8–9 fold elevation in total Axin levels (Fig 2.3A–2C, Table 2.1). When we examined the pattern of Axin:GFP accumulation, we

noted Mat/Zyg Axin led to substantially more variable expression from cell-cell than MatGAL4-

driven Axin:GFP. Thus, in most subsequent functional assays we used Mat Axin for high-level

overexpression. In addition to differences in expression levels, these lines also differed in timing

of Axin:GFP expression (Fig 2.3F). Zyg Axin levels started very low (as expected with no

maternal GAL4 expression) and continued to rise throughout development. Mat Axin levels started somewhat higher (driven by maternal GAL4), increased during stages 9–11 (4–9 hrs)

and then slowly decayed. Mat/Zyg Axin exhibited initially modest Axin:GFP levels, which

continued to rise throughout development. Mat RFP&Axin accumulation followed a similar

expression pattern as Mat Axin, but at decreased levels due to the presence of two UAS-driven

Figure 2.3

Figure 2.3 Developing tools to differentially elevate levels of Axin:GFP. (A-E) Immunoblots and quantification, 4-8hr old Drosophila embryos. Tubulin is loading control. (A) Anti-Axin antibody—samples were run on same blot with intervening lanes removed. Levels of Axin:GFP when expressed with different GAL4 drivers. * = background band. (B) Quantification of Axin:GFP, normalized to levels of endogenous Axin. # of blots quantified is in Table 2.1. (C) Relative levels of total Axin (thus including both endogenous Axin + Axin:GFP) accumulation—see Table 2.1 for standard deviation and # of blots assessed. (D,E) Anti-APC2 antibody. Endogenous APC2, GFP:APC2 expressed via its endogenous promoter, or GFP:APC2 expressed using MatGAL4. (F,G) Immunoblots of Drosophila embryos of the indicated ages, anti-GFP Antibody. Time courses of Axin:GFP (F) or GFP:APC2 (G) accumulation when expressed with different GAL4 drivers. (H) Immunoblot of 4-8hr old Drosophila embryos, with anti-Axin antibody, comparing endogenous Axin levels to Axin:GFP levels in lines expressing both Axin:GFP and a second transgene (RFP or GFP:APC2). From same gel with intervening lanes removed. (I) Same samples stained with an anti-GFP antibody, thus comparing levels of Axin:GFP and GFP:APC2. (J) Quantification of Axin:GFP levels normalized to wildtype. N = 9 blots.

transgenes (Fig 2.3F, right). These tools allowed us to vary Axin levels systematically, and we

thus used them to assess how altering Axin levels and timing of accumulation affect Wg

signaling and its regulation, by assessing effects on embryonic viability, cell fate choice, Wg

target gene expression, and Arm (fly βcat) levels.

When Axin expression is elevated by ≥4-fold, it inhibits Wg-regulated cell fate choice during embryogenesis

We first assessed effects of elevating Axin levels on embryonic viability and cell fate choice— these assays integrate effects on Wnt signaling across embryonic development, and thus have to be interpreted in light of effects on Axin levels both at stage 9 and later, as Wg signaling affects cell fate choice through stage 11 (9 hours; (Bejsovec and Martinez Arias,

1991)). The relatively subtle (3-fold) Axin elevation produced by Zyg Axin at stages 9–11 did not result in embryonic lethality (Fig 2.4A, Table 2.2; 5% lethality vs. 3% lethality of wildtype controls

(controls carried UAS-Axin without a GAL4 driver)). We then examined larval cuticles to look for

more subtle effects on Wg signaling. Reducing Wg signaling affects cell fate, causing loss of

naked cuticle fates and merger of denticle belts—Fig 2.4C illustrates the graded series of

defects with successively reduced Wg signaling. Most Zyg Axin embryonic cuticles (3-fold

increase) were near wildtype (Fig 2.4B and 3C, Table 2.3), though the occasional defects seen

suggest subtle reduction of Wg signaling in some embryos. Consistent with this possibility, no

hatching Zyg Axin larvae survived to adulthood—this may reflect the fact that as a consequence

of zygotic GAL4 expression Axin levels continued to rise throughout development (Fig 2.1F).

The slightly higher level expression of Axin:GFP in Mat RFP&Axin embryos (4-fold increase)

and the earlier onset of expression led to some embryonic lethality (32% lethal; Fig 2.4A; Table

2.1), and a larger fraction of embryos had moderate inhibition of Wg signaling, as assessed by

cell fate choices (Fig 2.4B and 3C, Table 2.3). In contrast, higher-level, earlier overexpression of

Axin (8–9 fold) led to substantial embryonic lethality—90% lethality for Mat/Zyg Axin and 78%

lethality for Mat Axin (Fig 2.4A; Table2.2). In both crosses, there were two genotypes of

Figure 2.4

Figure 2.4 Elevating Axin produces dose-sensitive inhibition of Wg signaling, while increasing APC2 levels does not. (A) Embryonic viability of indicated genotypes. (B,C) Assessing the effect of elevating Axin or APC2 levels on Wg-regulated cell fates. (B) Range of cuticle phenotypes of embryos/larvae of each genotype—since not all genotypes are lethal, phenotypes include those of hatched larvae. (C) Representative images of cuticle phenotypes used in B. Anterior to the top. 1: Wildtype. 2: 1–2 merged denticle belts (brackets). 3: 3–4 merged denticle belts. 4: Most denticle belts merged, mouth parts still present. 5: wg null phenotype–denticle lawn and no head (arrow). (D) Quantification of number of rows of En-expressing cells per segment. Embryos analyzed: WT- 21, AxinRNAi-5, APC2g10- 5, wgIG22–14, Zyg Axin- 9, Mat RFP&Axin- 11, Mat Axin- 18, Mat APC2- 12. * = p<0.05 using a one-way ANOVA test. (E-J) Representative images, En expression, as quantified in D. Anterior to the left.

embryonic progeny; for Mat/Zyg Axin these differed by whether or not they had a zygotic copy

of act5c-GAL4 and for Mat Axin by whether they had one or two copies of the UAS-Axin:GFP

transgene zygotically (Fig 2.2). Cuticle analysis of cell fates revealed that Wg-signaling was

strongly reduced in many Mat/Zyg Axin and Mat Axin progeny (Fig 2.4B and 3C, Table 2.3), but

there were variations in the strength of this effect that likely reflect the two different zygotic

genotypes in each cross. Thus, when levels of Axin exceed ~4–5 fold endogenous during

stages 9–11, this led to embryonic lethality and strong inhibition of Wg-regulated cell fates.

To assess effects of Axin levels on a Wg-regulated target gene, examined engrailed (en) expression, using antibodies to its protein product. En usually accumulates in the two most posterior cell rows in each body segment (Fig 2.4D and 3E, Table 2.4), and maintenance of En

expression requires Wg signaling—thus in wg mutants En stripes are narrowed ((Bejsovec and

Martinez Arias, 1991); Fig 2.4D, Table 4.4). In contrast, in APC2g10 null mutants or after Axin

RNAi, En expression expands to additional cell rows (Fig 2.4D, 4F and 4G, Table 4.4). The 3-

fold elevation of Axin levels via ZygGAL4 did not affect En expression (Fig 2.4D and 3H, Table

4.4). In contrast, the 9-fold increase of Axin via MatGAL4 led to partial loss of En expression

(Fig 2.4D and 4I, Table 2.4), though on average this was not as severe as that seen

in wg mutants. Thus, mildly elevating Axin levels during the critical period (stage 9–11) has little

effect on embryonic viability, Wg regulated cell fates or target genes, but when Axin levels are

elevated ≥ 8-fold, Wg signaling is strongly inhibited, consistent with previous data suggesting

that Axin is rate-limiting.

Elevating Axin levels has no effect on Arm levels in cells not receiving Wg signals, but does render the destruction complex more resistant to inactivation by physiological levels of Wg signaling

The primary role of the Axin/APC2 based destruction complex is to regulate levels of

Arm/βcat. We thus measured effects of different Axin levels on Arm accumulation. Arm has two

roles: as part of the cadherin-based cell adhesion complex and as a transcriptional co-activator in the Wnt pathway. Thus, all cells have a pool of Arm at the cortex in adherens junctions. In 61

wildtype, Wg is expressed by one row of cells in each segment, and moves to neighboring cells,

resulting in a gradient of Wg signaling across the segment. In cells not receiving Wg, the

destruction complex binds to newly synthesized Arm, which targets it for destruction (Fig 2.5A

and 5B). Thus, levels of cytoplasmic Arm are low. However, they are not zero; instead Arm that

is not immediately destroyed is retained in the cytoplasm by binding to the multiple Arm binding

sites on APC2 (Roberts et al., 2011). Together cytoplasmic retention and destruction mean little

or no Arm can translocate to the nucleus and co-activate Wnt target genes (Fig 2.5B, arrows). In

cells receiving Wg, the destruction complex is turned down, and Arm accumulates in both the

cytoplasm and nucleus, leading to activation of Wnt target genes (Peifer et al., 1994). Together,

these inputs create a gradient of Arm accumulation across the segment, with the highest level of

cytoplasmic/nuclear Arm accumulation in Wg-expressing cells and their immediate neighbors,

and gradually decreasing levels of cytoplasmic/nuclear accumulation in cells more distant from

the Wg source (Fig 2.5A and 5B; diagrammed in 5B’). In wg mutants the destruction complex

downregulates Arm in all cells, eliminating the stripes of Arm accumulation (Peifer et al., 1994).

Because different GAL4 drivers changed both the level and the timing of Axin expression, we

focused our attention on stage 9, when Wnt signaling is maximal, to alleviate the complication of

differences in timing.

We developed methods to quantify the effects of elevating Axin levels on two different aspects of Arm stabilization. To quantify the graded effects of Wg signal across the segments

(Fig 2.5B’), we used a digital image mask (Fig 2.6A’) to remove the cortical Arm in cell-cell

adherens junctions ( Fig 2.6A vs. Fig 2.6A” ), and then measured fluorescence levels of

cytoplasmic/nuclear Arm pixel by pixel across two to three body segments (Fig 2.6A” box; two

wildtype examples are in Fig 2.5G left). In wildtype embryos, both our images and quantitative

analysis revealed a smooth gradation of Arm accumulation, from peaks centered on Wg stripes

to troughs in the interstripes (Fig 2.5A, 5B and 5G). As a control, we examined wg null mutants,

in which Arm levels were not elevated in any cells (Fig 2.5G center; each mutant was analyzed

Figure 2.5

Figure 2.5 Increasing Axin levels reduces the ability of endogenous Wg signaling to turn down the destruction complex but has little or no effect in Wg-Off cells. (A-F) Fixed stage 9 embryos. Anterior to the top except B where anterior is to the right. (A,B) Representative images of Arm accumulation in wildtype embryos. (B) Close-up. “Wg” shows location of Wg-expressing cells, which accumulate elevated levels of Arm in the nucleus and cytoplasm. Arrows illustrate that in Wg-OFF cells, there are still detectable levels of Arm in the cytoplasm and also illustrate cytoplasmic retention. (B’) Diagrammatic illustration of Arm levels across two embryonic segments, illustrating the graded nature of Arm accumulation, and the parameters we assessed: absolute levels in Wg stripes or in interstripes (black arrows), and the difference between these levels (red arrow). (C,D) Representative images, Mat Axin:GFP embryos with higher or lower levels of Axin:GFP expression (depending on zygotic copy number of UAS- construct), taken from the same slide under the same microscope settings. Arm accumulation in Wg–stripes is reduced. (E,F) Representative images, Mat RFP&Axin embryos with higher or lower levels of Axin:GFP expression, taken from the same slide under the same microscope settings. Arm accumulation in Wg–stripes is reduced at higher levels of Axin:GFP expression, but less affected at lower levels. (G,H) When Axin levels exceed 4xendogenous,

63 this flattens the usual graded level of Arm accumulation across each embryonic segment. Plots of Arm accumulation over 2 segments for each genotype indicated. Dot plots = raw data from 3 separate embryos. Line graphs underneath = averages of these data. (G) Left. In wildtype embryos Arm accumulation varies smoothly over the segment. Middle. Loss of Wg flattens the Arm stripes. Right. Expressing Axin:GFP using the MatGAL4 driver blunts or eliminates Arm stripes. (H) The slightly lower levels of Axin expression in Mat RFP&Axin embryos have more variable effects on Arm stripes. (I-L) Elevating Axin levels reduces Arm accumulation in Wg stripes but does not significantly affect Arm levels in interstripes. (I,K) Box and whisker plot comparing Arm accumulation levels in Wg-expressing stripes versus Arm levels in the interstripes for the indicated genotypes. n = 10 pairs. Boxes extend from 25th to 75th percentiles, and whiskers indicate minimum to maximum values. Median = middle line of the box and mean = +. (Full values are in S5 Table). (J,L) Scatter plots showing difference in Arm accumulation between the Wg stripes and interstripes within individual embryos. Each point = a single embryo. Error bars = mean+S.D. Statistical analysis: A paired t-test was used to determine the significance between intragroup values in I and K. To assess the significance between intergroup values, an unpaired t-test was used in I and J, and an ordinary one-way ANOVA followed by Dunnett's multiple comparisons test was applied in K and L. ns, not significant i.e. p≥ 0.05. * = p<0.05. ** = p<0.01. *** = p<0.001. **** = p<0.0001. Scale bars = 30μm.

in parallel with the wildtype shown to its left). 9-fold elevation of Axin (Mat Axin) led to either

complete loss of this graded stabilization of Arm in cells receiving Wg signal, or a reduction in

the height of the peaks, relative to wildtype (Fig 2.5C and 5D, quantified in G). The changes in

Arm peak heights were dependent on the level of Axin:GFP expression; this was best visualized

in Mat RFP&Axin embryos where the lower level Axin expression only partially flattened the Arm

distribution (Fig 2.5E and 5F, quantified in 5H).

To measure absolute levels of Arm stabilization by Wg signaling, we assessed Arm

fluorescence in two groups of cells: 1–2 cell rows centered on cells expressing Wg (the Wg

stripes; Fig 2.6B and B’, yellow boxes) and 1–2 cell rows farthest from the Wg-expressing cells

(the interstripes; Fig 2.6B, white boxes). Wildtype embryos were included on the same slides as

a control. We quantified absolute Arm levels in both Wg stripes and interstripes (Fig 2.5B’, black

arrows, Fig 2.5I, Table 2.5) and also the difference in levels between these two cell types (Fig

2.5B’ red arrow, Fig 2.5J, Table 2.6). 9-fold overexpression of Axin (Mat Axin) substantially

reduced Arm accumulation in Wg stripes, to levels similar to those normally seen in interstripes

(Fig 2.5C and 5D vs. 5A; quantified in 5I and 5J, Tables 2.5 and 2.6). However, strikingly, Arm

accumulation in interstripes was unaffected. The 4-fold Axin overexpression in Mat RFP&Axin

embryos also reduced Wg-stabilization of Arm, but when we sorted embryos by level of

Axin:GFP expression, this was less pronounced in embryos with lower levels of Axin:GFP (Fig

2.5E and 5F vs. 5A, quantified in 5K and 5L, Tables 2.5 and 2.6). To complete this analysis, we

examined whether elevating Axin levels affected only the signaling pool of Arm (cytoplasmic

plus nuclear) or also affected the pool at cell junctions. Using a membrane-mask, we separately

assessed these two pools. Elevating Axin levels 9-fold (Mat Axin) reduced Arm accumulation

in both the junctional and cytoplasmic/nuclear pool in Wg-ON cells, without significantly affecting

either pool in Wg-OFF cells, relative to wildtype embryos (Fig 2.7A-C, Table 2.7).

Together, these data suggest that when Axin levels are elevated ≥4–5 fold, the destruction complex cannot be effectively inactivated by physiological levels of Wg signaling,

Figure 2.6

Figure 2.6 Assessing gradation of Arm levels across the segment and absolute levels of Arm in Wg stripes and interstripes. (A-A”) Representative example illustrating assessment of graded cytoplasmic and nuclear Arm levels across 2–3 embryonic segments (A) Original image of Arm (A’) Plasma membrane mask created using pTyr staining of the same embryo in A. (A”) Resulting Arm image after subtracting A’ from A. The green box illustrates a region of interest (ROI) selected. A profile of the ROI was plotted along the anterior-posterior axis. ROI profiles were adjusted for embryo length and to bring valleys to zero. (B-B”) Representative example illustrating calculations of absolute levels of Arm in Wg-expressing cells (Wg-stripes) and Wg-OFF cells (interstripes). Yellow boxes indicate regions sampled for Wg stripes. White boxes represent the regions sampled for the interstripe regions. The blue box represents the region sampled for background. In all cases, wildtype and mutant embryos were imaged together using constant imaging conditions

Figure 2.7

Figure 2.7: The opposite effects of Axin versus APC2 overexpression on Arm levels in Wg-ON cells are observed in both the cytoplasmic and membrane-associated pools. We separately assessed how elevating levels of Axin or APC2 affected total Arm levels (A), levels of Arm in the cytoplasmic/nuclear pool (B), or levels of Arm in the junctional (membrane) pool (C), by using a membrane marker to create an image mask (see Methods). Elevating Axin levels 9-fold (Mat Axin) reduced Arm levels in each of these pools in Wg-ON cells, without affecting levels of Arm in any of the pools in Wg-OFF cells, relative to wildtype. Conversely, elevating APC2 levels 11-fold (MatAPC2) increased Arm levels in each of these pools in Wg-ON cells, without affecting levels of Arm in any of the pools in Wg-OFF cells, relative to wildtype. (D) Embryo expressing a mutant APC2 protein deleting all of the ßcat binding sites (APC2Δ15Δ20R1,R3-R5 (expressed in the APC null background = APC2g10 APC1Q8; referred to here as APCΔßcat). Arm accumulation is enhanced in Wg-ON cells without affecting accumulation in Wg-OFF cells. (E) Quantification of Arm levels. (F) Difference in Arm levels between Wg stripes and interstripes. Statistical analysis: a paired t-test was used to determine the significance between intragroup values in E, and an unpaired t-test was used to determine the significance between intergroup values in E and F. ns, not significant i.e. p≥ 0.05. ** = p<0.01. *** = p<0.001. **** = p<0.0001.

confirming previous observations that Axin is rate-limiting in this regard. However, it was also

striking that elevating Axin levels did not further increase Arm destruction in cells not receiving

Wg signal (Fig 2.5I, Table 2.5, Fig 2.7A-C. Table 2.7), suggesting that Axin is not rate-limiting

for destruction complex activity in those cells.

Levels of APC2 can be substantially elevated without significantly affecting viability or Wg-regulated cell fates

We next investigated whether Wg signaling was similarly affected by altered APC2 levels—since it is the other key component of the destruction complex and our data revealed that its levels are not substantially different from those of Axin, we suspected it might also be rate-limiting and thus over-expression would inhibit Wg signaling. We used a similar approach to elevate GFP:APC2 levels. Using the MatGAL4 driver, we achieved an ~12-fold increase in

APC2 levels (Fig 2.3D and 3E; Table 2.1; hereafter Mat APC2). As we observed with Mat Axin,

in Mat APC2 progeny GFP:APC2 levels started high and slowly decreased (Fig 2.3G).

Strikingly, elevating APC2 levels 12-fold had no effect on embryonic viability (94% viable; Fig

2.4A, Table 2.2); in fact, these embryos could develop to adulthood and produce viable

offspring. We next examined whether elevating APC2 levels affected Wg-regulated cell fate

choices, as assessed by cuticle phenotype. Little or no effect on embryonic patterning was seen

(Fig 2.4B and 4C, Table 2.3), and the few denticle belt fusions observed were in hatched larvae.

Finally, we examined effects on expression of the Wg target gene en. This was also unaffected

by overexpression of APC2 (Fig 2.4D and 4J, Table 2.4). Thus, in stark contrast to Axin,

embryonic viability, cell fate choice and Wg target gene expression are not sensitive to

substantially elevated levels of APC2.

Elevating levels of APC2 strongly promotes downregulation of the destruction complex in response to physiological levels of Wg signaling

As a final exploration of the effects of elevating APC2 levels, we examined Wg-

regulation of Arm stability, using the same assays we employed for analyzing effects of altering

Axin levels (Fig 2.6). We were surprised to find APC2 overexpression led to a striking change in

Arm levels, suggesting reduced activity of the destruction complex. Levels of Arm in Wg-

expressing cells and their immediately adjacent neighbors were strongly elevated (Fig 2.8B–

8D vs. 8A), leading to a more defined pattern of stipes in Arm accumulation across each

segment. Quantification confirmed that while interstripe Arm levels were unchanged, Arm levels

in Wg stripes were significantly higher (Fig 2.8E and 8F, Tables 2.5 and 2.6). The sharpened

stripes and elevated Arm levels in Wg-ON cells were also apparent in our analysis of Arm levels

across each segment (Fig 2.8G). Finally, the same differences were also apparent when we

used a membrane mask to examine only cytoplasmic/nuclear Arm or only the membrane pool of

Arm (Fig 2.7A-C). These data were quite surprising, as they were the exact opposite of the

effects of elevating Axin levels. We examined whether these effects result from reducing Axin

levels, a function previously suggested for APC2 (Wang et al., 2016b), but immunoblotting

suggested this was not the case (Fig 2.8H, Table 2.2). Instead, these data suggest that when

APC2 levels are elevated in a way that accentuates the endogenous APC2:Axin ratio,

stabilization of Arm by Wg signaling is enhanced. This could occur by direct effects on the ability

of Wg signaling to downregulate the destruction complex, or via the ability of APC2 to bind and

sequester Arm (Roberts et al., 2011)—we consider these possibilities more completely in the

Discussion. However, this further elevation of Arm levels in cells already receiving Wg signals

had little effect on Wnt-target gene expression or cell fate (Fig 2.4A, 4B and 4D, Tables 2.2 -

2.4). Finally, elevating APC2 levels did not alter destruction complex activity in cell not receiving

Wg signals, similar to what we observed with Axin (Fig 2.8E, Fig 2.7A).

These effects on Arm levels—sharpened and enhanced Arm stripes—were reminiscent of effects previously seen when analyzing APC2 mutants in which the motifs that act as binding sites for Arm (the 15- and 20-amino acid repeats) were reduced in number or eliminated

((Yamulla et al., 2014); Fig 2.7D). These APC2 mutants were expressed at endogenous levels

in a null APC2g10background, rather than overexpressed. We analyzed the most extreme of

Figure 2.8

Figure 2.8 Elevating APC2 levels increases the ability of endogenous Wg signaling to turn down the destruction complex, thus increasing Arm levels in cells receiving Wg. (A-D) Fixed Stage 9 embryos. Anterior to the left. (A-C) Representative images, wildtype (A) or Mat GFP:APC2 embryos with higher (B) or lower (C) levels of GFP:APC2 expression. Elevating APC2 levels increases levels of Arm specifically in cells receiving Wg signal. (D) Close-up, embryo expressing elevated levels of GFP:APC2. The boundary of cells with elevated levels of Arm is quite sharp, and does not expand much farther than the cells adjacent to those expressing Wg. (E) Elevating APC2 levels increases Arm accumulation in Wg stripes but does not affect Arm levels in interstripes. Box and whisker plot (as in Fig 4I and 4K), comparing Arm accumulation levels in Wg-expressing stripes versus Arm in the interstripes in wildtype or Mat GFP:APC2 embryos imaged on the same slide. (F) Difference in Arm accumulation between the Wg stripes and interstripes within individual embryos. Each point = a single embryo. (G) Plots of Arm accumulation pattern over 2 segments. Dot plots = raw data from 3 separate embryos. Line graphs underneath = averages of these data. Elevating APC2 levels exaggerates and sharpens the Arm stripes. (H) Immunoblotting with anti-Axin antibodies reveals that embryos overexpressing APC2 have no change in Axin levels. Statistical analysis: a paired t-test was used to assess the significance between intragroup values in E, and an unpaired t-test was used to determine the significance between intergroup values in E and F. A one-way t-test was used to assess the significance of difference in Axin levels in H. ns, not significant i.e. p≥ 0.05. ** = p<0.01. *** = p<0.001. **** = p<0.0001.

these—APC2Δ15Δ20R1,R3-R5, which deletes all of the ßcat binding sites, and expressed it in the APC null background = APC2g10 APC1Q8. This allele has a paradoxical phenotype: it strongly reduces APC2 function in Wnt regulation, as assessed by cell fates, but still promotes destruction of Arm in Wg-OFF cells (Yamulla et al., 2014).

To determine if the quantitative effects on Arm levels paralleled those we saw after elevating levels of wildtype APC2, we applied our quantitative toolkit to measure Arm levels on

Wg-ON and Wg-OFF cells in that mutant. Intriguingly, interstripe Arm levels were unchanged,

while Arm levels in Wg stripes were significantly higher (Fig 2.7E and 7F, Tables 2.5 and 2.6).

This phenotype mimics what we observed after elevating levels of wildtype APC2. This may

suggest that Wg-ON cells are more sensitive to any perturbation that reduces the function of the

destruction complex. We consider the interpretation of this similarity further in the Discussion.

Simultaneously elevating levels of both APC2 and Axin inhibits Wg signaling more than elevating levels of Axin alone

These data reveal that elevating Axin levels or elevating APC2 levels had opposite

effects on the ability of Wg signaling to regulate destruction complex function. To explore this

further, we varied the levels of both proteins simultaneously, and also varied the ratios of their

expression levels. We began by expressing both Axin:GFP and GFP:APC2 simultaneously

(progeny of GFP:APC2/MatGal4; Axin:GFP/Mat Gal4 females crossed to GFP:APC2; Axin:GFP

males; hereafter, Mat APC2&Axin). The progeny of this cross differ in their zygotic genotypes

and thus in the relative levels of Axin:GFP and GFP:APC2 (Fig 2.9A). We first examined the

average overexpression levels in embryos including all four zygotic genotypes combined.

Immunoblotting revealed that, on average, they accumulate Axin:GFP at levels 4-fold above

endogenous Axin (Fig 2.3H–2J, Table 2.1) similar to Mat RFP&Axin (which also contains two

UAS transgenes), and accumulate GFP:APC2 at ~20x endogenous levels (Table 2.1). However

embryonic lethality of embryos overexpressing both Axin:GFP and GFP:APC2 was substantially

higher than that of Mat RFP&Axin embryos (63% versus 32% lethal; Fig 2.9C vs.

Figure 2.9

Figure 2.9 The relative ratios of APC2 to Axin levels determine effects on embryonic viability and Wg-regulated cell fates. (A) Cross used to generate embryos expressing different ratios of APC2 and Axin, with the four categories of progeny, their relative levels of Axin and APC2 overexpression, and the criteria used to identify them. (B) Control crosses used to assess how different ratios of Axin and APC2 overexpression differentially affect embryonic lethality and Wg-regulated cell fate choice. (C) Embryonic viability of different genotypes with differentially altered APC2:Axin ratios. (D) Quantification of the effects of elevating APC2 and Axin on cell fate, as assessed by cuticle pattern. Representative cuticles are in Fig 2.4C.

Fig 2.4A; Table 2.2), despite similar average levels of Axin:GFP accumulation (Fig 2.3H–

2J, Table 2.1). In parallel, cell fates were shifted more towards the wg null phenotype (Fig

2.9D, Table 2.3) than was seen in Mat RFP&Axin embryos (Fig 2.4B, Table 2.3). Therefore, co-

expressing APC2 and Axin inhibits Wg signaling to a greater extent than expressing either Axin

or APC2 alone, despite similar average levels of Axin:GFP and GFP:APC2 accumulation (Fig

2.3H–2J, Table 2.1). These data are consistent with the hypothesis that co-expressing Axin and

APC2 enhances the resistance of the destruction complex to inactivation by Wg signal.

We suspected that these averages hid differences in outcome among the four different genotypes present among the progeny (Fig 2.9A), which would express different ratios of APC2

and Axin. To determine which genotypes exhibited elevated embryonic lethality and defects in

Wg-regulated cell fates, we set up two additional crosses, in which the relative zygotic

expression of Axin and APC2 differed (Fig 2.9B): 1) APC2>>Axin = average zygotic dose of

GFP:APC2 is higher than that of Axin:GFP, and 2) Axin>>APC2 = average zygotic dose of

GFP:APC2 is lower than that of Axin:GFP. These two crosses had strikingly different results.

APC2>>Axin progeny had only 20% embryonic lethality and Wg-regulated cell fates were only

mildly affected (Fig 2.9C and 9D, Tables 2.2 and 2.3), while Axin>>APC2 progeny had 70%

embryonic lethality and had very strong effects on Wg-regulated fates, with 39% having a wgnull phenotype (Fig 2.9C and 9D, Tables 2.2 and 2.3). Thus, while APC2 overexpression alone does

not affect cell fates, elevating levels of both APC2 and Axin levels inhibits Wg signaling to a

greater degree than elevating levels of Axin alone, suggesting both total levels and the relative

ratios of Axin and APC2 are important.

The relative ratio of APC2:Axin levels determines the effectiveness of Arm destruction

These data made strong predictions about how different relative levels of Axin and APC2 would affect Arm destruction. While we could not directly determine genotypes of fixed and stained embryos, we developed a method to infer genotypes from levels and localization of

GFP-tagged proteins. Since total protein levels of GFP:APC2 were, on average, higher than

those of Axin (Fig 2.3I, Fig 2.10A vs. B), we first separated embryos into two categories, directly

quantifying total GFP expression by immunofluorescence and using low versus high GFP levels

as a surrogate for zygotically UAS-GFP:APC2/+ versus zygotically UAS-GFP:APC2/UAS-

GFP:APC2 embryos (e.g., Fig 2.11A’” and B’” vs. 11C’” and D’”).

To further subdivide the embryos, we made use of the assembly of Axin:GFP into cytoplasmic puncta in interstripes (Cliffe et al., 2003). If we could easily visualize cytoplasmic puncta (Fig 2.11B’” and D’” insets), we categorized embryos as zygotically UAS-Axin:GFP/UAS-

Axin:GFP rather than zygotically UAS-Axin:GFP/+. This produced four presumptive genotypes

with different degrees of overexpression of Axin and APC2 (Fig 6A):

1) APC2+Axin+. Presumptive zygotic genotype = UAS-GFP:APC2/+; UAS-Axin:GFP/+,

2) APC2+Axin++. Presumptive zygotic genotype = UAS-GFP:APC2/ +; UAS-Axin:GFP/UAS-

Axin:GFP.

3) APC2++Axin+. Presumptive zygotic genotype = UAS-GFP:APC2/UAS-GFP:APC2; UAS-

Axin:GFP/+

4) APC2++ Axin++. Presumptive zygotic genotype = UAS-GFP:APC2/UAS-GFP:APC2; UAS-

Axin:GFP/UAS-Axin:GFP.

We then analyzed Arm accumulation in these four embryo categories, using the quantitative tools described above to assess absolute Arm levels in Wg stripes and interstripes relative to wildtype controls. To our surprise, despite the four presumptive genotypes, the embryos divided into two phenotypic categories with regard to Arm accumulation. In embryos of the two genotypes that overexpressed Axin at the highest levels (APC2+Axin++ (Fig 2.11B);

and APC2++Axin++ (Fig 2.11D)), Arm levels were strongly reduced in the Wg stripes (Fig 2.11B

and D: quantified in Fig 2.11E and F; Tables 2.5 and 2.6). Thus, they resembled embryos

overexpressing only Axin (Fig 2.5C, D, I and J). In contrast, the two genotypes that

overexpressed APC2 but had lower levels of Axin elevation (APC2+Axin+ (Fig 2.11A) and

Figure 2.10

Figure 2.10 Illustration of how embryos were sorted as to inferred genotype. (A,B) Fixed stage 9 embryos imaged under the same microscope conditions, illustrating that when expressed using the MatGAl4 drivers, GFP:APC2 is substantially brighter under our conditions than Axin:GFP, allowing us define GFP:APC2 genotypes by overall GFP fluorescence. (A) GFP brightness of Mat Axin embryo. (B) Example of a MatAPC2 & Axin embryo we scored in the GFP high category. This embryo would also have been scored as Axin high because of the bright puncta.

Figure 2. 11

Figure 2.11 The relative ratios of APC2 to Axin levels determine effects on Arm destruction. (A-D) Fixed stage 9 embryos. Anterior to the left. Representative images of the four different categories of the Mat APC2 & Axin phenotypes from Fig 2.9. Images were taken under the same microscope settings. Insets are close-ups. See Fig 2.9A for key to identifying presumptive genotype. (A,C) Both genotypes in which GFP:APC2 elevation exceeds that of Axin:GFP have elevated Arm accumulation in Wg stripes. (B,D) Both genotypes with the highest levels of Axin:GFP have reduced Arm accumulation in Wg stripes. (E) Effects on Arm accumulation in Wg stripes or interstripes in embryos with different ratios of Axin and APC2 accumulation-criteria used to distinguish embryos are in Fig 2.9. Box and whisker plot (as in Fig 2.5I and K), Arm accumulation in Wg-expressing stripes versus Arm in the interstripes. Wildtype and different presumptive genotypes of Mat APC2&Axin embryos imaged on the same slide. (F) Difference in Arm accumulation between Wg stripes and interstripes within individual embryos. Statistical analysis, a paired t-test was used to determine the significance between intragroup values in E, and an ordinary one-way ANOVA followed by Dunnett's multiple comparisons test was applied between intergroup values in E and F. ns, not significant i.e. p≥ 0.05. ** = p<0.01. *** = p<0.001. **** = p<0.0001.

APC2++Axin+ (Fig 2.11C)), Arm levels were strongly elevated in the Wg stripes (Fig 2.11A and

2.11C: quantified in Fig 2.11E and F, Tables 2.5 and 2.6). Thus, they resembled embryos

overexpressing APC2 alone (Fig 2.8B and C).

Combined with the phenotypic data above, these data suggest that the ratio of APC2 to

Axin plays a very important role in determining sensitivity of the destruction complex to being inactivated by Wg signaling, such that when Axin is expressed at or over the levels of APC2, the destruction complex is resistant to inactivation, but when levels of APC2 are substantially higher than those of Axin, the destruction complex is more easily inhibited.

Axin assembles into cytoplasmic multiprotein destruction complexes, and Wnt/Wg signaling leads to their membrane-recruitment and elevates levels of cytoplasmic Axin

One major question still debated in the Wnt field is what happens to the destruction

complex after Wnt stimulation. Wnt signaling leads to Axin recruitment to the transmembrane

receptor LRP5/6 (MacDonald and He, 2012). Work in both cultured human cells

and Drosophila embryos suggest that both core components of the destruction complex, APC

and Axin, can be recruited to the membrane after Wnt stimulation (Cliffe et al., 2003; Li et al.,

2012). However, three studies of the resulting effects of Wg signaling on Axin levels and

localization in the Drosophila embryonic epidermis yielded to three distinct conclusions: 1) Wg

signaling destabilizes Axin (Tolwinski et al., 2003), 2) Wg signaling initially stabilizes Axin (Yang

et al., 2016), or 3) Wg signaling leads to Axin membrane recruitment (Cliffe et al., 2003).

We thus revisited the question, taking advantage of our ability to express Axin:GFP at

defined levels below those at which it significantly inhibits Wg signaling. We first verified that

GFP-tagging does not alter physiological roles of Axin: the ability of Axin:GFP to downregulate

Arm levels, or its ability to be inactivated in cells that receive Wg. To do so, we expressed

Axin:GFP in embryos in which endogenous Axin was knocked down by RNAi (we co-expressed

UAS-RFP to account for effects of different copy numbers of UAS–driven transgenes). Axin

RNAi led to highly penetrant embryonic lethality (Fig 2.12A, Table 2.2), transformation of cell

Figure 2.12

Figure 2.12 Axin:GFP can largely restore normal Wnt signaling after Axin RNAi. We compared the effects of Axin RNAi with or without expressing Axin:GFP. Crosses: matGAL4/+; matGAL4/Axin shRNA females to either UAS-Axin:GFP males or UAS-RFP males as a control. We also crossed matGAL4/UAS:RFP; matGAL4/+ females to UAS-Axin:GFP males to control for effects of Axin:GFP expression. (A) Assessment of embryonic viability. Axin RNAi leads to highly penetrant embryonic lethality which is largely rescued by expression of Axin:GFP. (B) Assessment of effect on Wg regulated cell fates via cuticle analysis. Categories are illustrated in Fig 2.4 (reduced Wg signaling) or S8E Fig (increased Wg signaling). Axin-RNAi expands the Wg-promoted naked cuticle fates. This is largely rescued to wildtype by expression of Axin:GFP, though a few embryos lose naked cuticle, as is seen in the control expressing only Axin:GFP. (C-F) Stage 9 embryos, visualize Wg, Arm and Axin:GFP. (C) Wildtype. (D) Axin- RNAi. Note elevated Arm levels and expansion of Wg stripes. (E) Axin-RNAi combined with expression of Axin:GFP. The normal segmental stripes of Arm and the single-cell wide stripes of Wg expression are restored. (F) Expression of Axin:GFP. At this level of expression most embryos have near normal Arm stripes.

fates toward Wg-ON fates (as assessed by cuticle analysis; Fig 2.12, Table 2.3), strong

elevation of Arm levels and altered Wg expression (Fig 2.12C vs 5D ). Axin:GFP substantially

restored embryonic viability and Wg-regulated cell fate choices (Fig 2.12A and B, Tables 2.2

and 2.3), downregulated Arm and restored normal Wg expression (Fig 2.12E). Most embryos

had a wildtype cuticle, though a small fraction had Wg-signaling inhibited ( Fig 2.12B Fig, Table

2.3). Together, these data suggest GFP-tagging does not substantially affect Axin function or its

ability to be downregulated by Wg signals.

To further verify that the GFP tag on Axin does not alter its function, we analyzed Axin self-assembly in cultured colorectal cancer cells, where Axin self-assembles into multiprotein

“puncta” and recruits APC into these structures (Faux et al., 2008). We hypothesize these puncta are larger versions of the normal multiprotein destruction complex (Pronobis et al., 2015;

Roberts et al., 2011). Because GFP can dimerize under some conditions, we verified that similar puncta form and recruit APC2 when Axin is tagged with a Flag-epitope rather than with

GFP or one of its derivatives (Fig 2.13A–F). We also created a version of Axin tagged with a

monomeric mutant of GFP (Zacharias et al., 2002), and observed no difference in Axin-self-

assembly into puncta or recruitment of APC2 (Fig 2.13A G and H). Similar puncta were

previously observed in Drosophila embryos when Axin:GFP was significantly overexpressed, at levels that inhibit Wg signaling (Cliffe et al., 2003). Membrane-associated endogenous Axin

puncta were also seen in imaginal discs, and Axin tagged with the V5 epitope also accumulated

in membrane-associated puncta and in the cytoplasm of cells in embryos that received Wg

signal (Wang et al., 2016a).

Our system allowed us to directly visualize Axin:GFP localization in embryos expressing it at levels near those of endogenous Axin (Mat RFP&Axin = 4-fold elevated); ~70% of these embryos are viable and >60% have no disruption of Wg-regulated cell fates (Fig 2.4A and

B, Tables 2.2 and 2.3). Visualizing Axin:GFP directly avoided issues with antibody accessibility

to Axin assembled into large multiprotein complexes versus protein diffuse in the cytoplasm,

Figure 2.13

Figure 2.13: Flag-tagged Axin assembles into puncta indistinguishable from those assembled by Axin:GFP. Constructs expressing the indicated proteins were transfected into SW480 cells and visualized either using the fluorescent tag or using an anti-Flag epitope antibody. (A,C,E,G) Axin:RFP (A), Flag:Axin (C), Axin:GFP (E), and Axin:monomeric GFP (mGFP) (G) all assemble into numerous puncta-no differences were seen in this regard (B) RFP:APC2 is diffusely cytoplasmic. (D,F,H) Flag:Axin (D), Axin:GFP (F) and Axin:mGFP (H) can all recruit RFP:APC2 into puncta. Insets = closeups of puncta, illustrating co-localization.

an issue we observed in cultured colorectal cancer cells ((Pronobis et al., 2015); Fig

2.13D’ insets). We examined Axin:GFP localization throughout the stages at which Wg signals

regulate cell fate. wg mRNA expression initiates at the blastoderm stage. As germband

extension starts, Wg protein is just beginning to accumulate in stripes (Fig 2.14A and B; [42]). At

this stage, most cells had small puncta of Axin:GFP, both membrane-proximal and cytoplasmic,

along with a cytoplasmic pool. In some cells near to those initiating Wg expression, Axin:GFP

containing puncta were beginning to be enriched at the cortex (Fig 2.14B arrows). In contrast, at

stage 9, when Wg signaling begins to regulate Arm levels and shape cell fate, we observed a

prominent difference in Axin:GFP localization in cells receiving or not receiving Wg signal (Fig

2.14C and D). In cells far from the source of Wg, much of the Axin:GFP was assembled into

bright cytoplasmic puncta, with relatively low levels in the cytoplasm (Fig 2.14D and E yellow

arrows). In contrast, in cells receiving Wg signal, Axin:GFP assembled into less bright

membrane-associated puncta, and elevated levels of Axin:GFP were seen in the cytoplasm (Fig

2.14D and E magenta arrows). A similar pattern was observed using GAL4 drivers that led to

higher levels of Axin:GFP (act5c-GAL4 = Mat/Zyg Axin or MatGAL4 without RFP = Mat Axin).

This resembled the pattern previously observed by Cliffe et al. (2003) using a strong GAL4

driver (Cliffe et al., 2003). During stage 10, when the Wg stripes become interrupted, with separate midline and lateral stripes (Fig 2.14G, brackets), the pattern of Axin:GFP localization

became more complex in parallel. Differences in intracellular localization remained between

cells near those expressing Wg (Fig 2.14G, magenta arrows) and those farther away (yellow

arrows).

To quantitatively assess levels of Axin:GFP in different subcellular structures, we

thresholded our images to different degrees, assessing which structures were brightest and thus

likely contained the highest density of Axin:GFP proteins. The results were quite striking. The

brightest 0.1% of pixels and most of the brightest 0.3% of pixels, which represent the highest

Figure 2.14

Figure 2.14 Axin assembles into cytoplasmic multiprotein destruction complexes together with APC2, and Wg signaling leads to their membrane-recruitment and elevates levels of cytoplasmic Axin.

(A,B) Fixed stage 8 Mat RFP&Axin embryo. Axin:GFP accumulates in puncta in all cells. Anterior to the left. Inset = close-up of B. Red arrows = initiation of Wg stripes. (C-E) Fixed stage 9 Mat RFP&Axin embryos. Anterior to the left. Red arrows = Wg expressing cells. (D,E) Close-ups of embryo in C showing Axin:GFP localization change in response to Wg. Yellow arrows = cytoplasmic puncta. Magenta arrows = membrane-associated Axin:GFP puncta. (F- F”‘) Image thresholding to determine the relative brightness of different pools of Axin:GFP. (F,F’) The brightest Axin:GFP pixels are in the cytoplasmic puncta in the interstripe cells (brackets). (F”) The next brightest pixels are in membrane–associated puncta in the Wg stripe cells (arrows). (F’”) Diffuse cytoplasmic staining is higher in Wg-stripe cells (arrows) than in interstripes (brackets). (G) Fixed stage 10 embryo. As the Wg stripe separates into medial and lateral domains (brackets), Axin:GFP continues to exhibit differential localization near or distant from Wg-expressing cells. Anterior to the left. Yellow arrows = cytoplasmic puncta. Magenta arrows = membrane-associated Axin:GFP puncta. (H) Localization of endogenous APC2 in a wildtype embryo. (I,J) Expression of Axin:GFP leads to recruitment of endogenous APC2 into both membrane-associated puncta in Wg-ON cells (magenta arrows) and into cytoplasmic puncta in Wg-OFF cells (yellow arrows). (K,L) Mat APC2&Axin. L = closeup. Presumptive APC2++ Axin++ embryo. Simultaneously highly elevating levels of bothAPC2 and Axin enhances resistance of the destruction complex to be turned off by Wg signaling. Yellow arrow = very bright cytoplasmic puncta. Cyan arrows = bright puncta found near Wg-positive cells. Magenta = membrane-associated puncta in Wg expressing cells. Scale bars = 15μm.

levels of Axin:GFP accumulation, were located in the cytoplasmic puncta in Wg-OFF cells (Fig

2.14F and F’). When we lowered the threshold intensity to visualize the brightest 1% of pixels,

the next structures to appear were the membrane-associated puncta in Wg-ON cells (Fig

2.14F”). It was only when we visualized the brightest 15% of the pixels that the relatively high

levels of diffuse cytoplasmic Axin:GFP in the Wg-ON cells were revealed (Fig 2.14F”‘). This

contrasted with the lower cytoplasmic levels of Axin:GFP in Wg-OFF cells.

We next sought to reconcile our observations with recent publications, whose data suggested that the primary effect of Wg signaling was to stabilize Axin in both the cytoplasm and at the membrane (Wang et al., 2016a; Yang et al., 2016). These studies used an antibody to an epitope to visualize epitope-tagged Axin. We therefore used a GFP-antibody to visualize

Axin:GFP expression (Fig 2.15A–E). Intriguingly, the bright Axin cytoplasmic puncta in the

interstripe regions were less apparent (e.g., Fig 2.14C’ vs. Fig 2.15A’ or C’)—thus use of an

antibody emphasized the stronger cytoplasmic signal in Wg-ON cells, reproducing the earlier

observations. This suggested that directly visualizing Axin:GFP provides a more complete

picture of the effects of Wg signaling on Axin localization and levels.

Earlier work suggested that when Axin is significantly over-expressed, Axin puncta also contain APC2 (Cliffe et al., 2003; Mendoza-Topaz et al., 2011). We revisited this issue, using our ability to visualize Axin puncta at near endogenous levels (4x-elevated; Mat RFP&Axin) in embryos where Wnt signaling is not substantially inhibited. In wildtype embryos, APC2 is cortically enriched, with a strong cytoplasmic pool (Fig 2.14H;(McCartney et al., 1999)).

Expressing Axin:GFP at 4x endogenous levels significantly altered APC2 localization (Fig 2.14I

and J). APC2 was now recruited into both the large cytoplasmic puncta in Wg-OFF cells (Fig

2.14I and J, yellow arrows) and to the smaller, membrane-bound puncta in Wg-ON cells (Fig

2.14I and J, magenta arrows). Intriguingly, recruitment of APC2 into Axin puncta seemed more

robust in Wg-ON than in Wg-OFF cells (Fig 2.14J and J’-yellow vs. magenta arrows).

Figure 2.15

Figure 2.15: When Axin is localized using an antibody to the GFP epitope-tag, it emphasizes the elevation in cytoplasmic Axin in Wg-ON cells and de-emphasizes Axin puncta in Wg-OFF cells. (A-D). Stage 9 embryos, anterior to the left. (E) Late stage 9/stage 10 embryo. All are expressing Axin:GFP using the matGAL4 driver (Mat Axin) and all stained with antibodies to GFP and Wg, along with Neurotactin (Nrt) to visualize the plasma membrane. B and D are close-ups of A and C, respectively. (A,C) Antibody staining clearly reveals elevated cytoplasmic Axin:GFP in cells receiving Wg signal (arrows). (B) In optimally stained embryos, close-ups also reveal both cytoplasmic puncta in Wg-OFF cells (yellow arrows) and membrane- associated puncta in Wg-ON cells (magenta arrows). (D) In many embryos cytoplasmic puncta in Wg-OFF cells are either not visible or less apparent (yellow arrows), while membrane- associated puncta in Wg-ON cells remain visible (magenta arrows) and elevation of cytoplasmic Axin in Wg-ON cells becomes prominent. (E). By late stage 9-early stage 10, we can visualize the decrease in cytoplasmic Axin in cells expressing Wg (arrows), as was previously reported (Wang et al., 2016b).

Together, these data suggest that in the absence of Wg signals, Axin self-assembles into large cytoplasmic multiprotein destruction complexes and diffuse cytoplasmic levels of Axin are reduced. Axin recruits APC2 into these puncta and thus they are likely to represent active destruction complexes. In contrast, in cells receiving Wg signal, Axin:APC2 puncta are recruited to the plasma membrane, these puncta diminish in intensity, and the cytoplasmic pool of Axin is correspondingly increased—these changes occur in parallel with and may cause the reduction in destruction complex activity.

Wg signaling and GSK3 activity are each required for membrane recruitment of Axin puncta

These data suggest that Wg signaling leads to destruction complex recruitment to the plasma membrane, as was observed in cultured cells. To confirm that Wg was required for this response, we visualized Axin:GFP localization in embryos zygotically mutant for the genetically null allele wgIG22 (these mutants produce reduced levels of a non-functional protein, allowing us

to identify mutants by reduced Wg accumulation and loss of Arm destruction). Consistent with

the hypothesis that reception of Wg triggers membrane recruitment of Axin:GFP puncta,

Axin:GFP localized to cytoplasmic puncta in all cells in wgIG22 mutants (Fig 2.16A vs. B), while

levels of diffuse cytoplasmic Axin:GFP were relatively low in all cells. These data are consistent

with what Cliffe et al. (2003) observed when expressing Axin:GFP at higher levels. We also

carried out the converse experiment, using the matGAL4 driver to ubiquitously express UAS-

Wg:HA (Hays et al., 1997), and examined effects on localization of Axin:GFP. Ubiquitous Wg

expression led to highly penetrant embryonic lethality and strong expansion of the Wg-regulated

naked cuticle fates (Fig 2.17A, B and E, Tables 2.2 and 2.3). Ubiquitous Wg expression led all

cells to accumulate Axin:GFP in membrane puncta, with elevated levels of Axin:GFP in the

cytoplasm (Fig 2.16C). These data confirm that the alterations of Axin:GFP localization are

driven by reception of Wg signal.

Figure 2.16

Figure 2.16 Wg signal and GSK3/Zw3 activity are important for destruction complex membrane recruitment and GSK3/Zw3 regulates release of Arm from the destruction complex. (A,B) Localization of Axin:GFP in stage 9 sibling control embryo (A) and wgIG22 mutant (B). Neurotactin serves as a membrane marker. Both the patterned recruitment of Axin:GFP puncta to the membrane and elevation of cytoplasmic pool of Axin:GFP in Wg-ON cells (double arrows) are lost in wgIG22 mutants. (C) Stage 9 embryo ubiquitously expressing Wg, using the MatGAL4 driver driving both UAS-Wg:HA and UAS-Axin:GFP. Now all cells accumulate Axin:GFP in membrane puncta and also accumulate elevated levels of Axin:GFP in the cytoplasm. (D-F) Stage 9 zw3 maternal/zygotic RNAi embryos expressing UAS-Axin:GFP, both driven by matGAL4 drivers (= zw3 RNAi x Axin in Methods). (D) Arm levels are highly elevated in all cells. (E,F) Membrane recruitment of Axin:GFP puncta in Wg-ON cells is lost, and Arm accumulates in Axin puncta (F, arrowheads). (G) Immunoblot with anti-Axin antibodies and quantification. Axin levels remain unchanged after zw3 RNAi (note: UAS:Axin:GFP was not present in this cross = zw3 RNAi in Methods). * = 100 kDa band is non-specific cross-reacting band, as is indicated by the Axin RNAi control. Tubulin was a loading control. A one-way t-test was used to assess the significance of difference in Axin levels. Scale bars = 15μm.

Figure 2.17

Figure 2.17: Ubiquitous expression of Wg increases embryonic lethality and induces a loss of denticle belts, whereas Dsh overexpression has little effect on viability and cuticle phenotype. (A) Ubiquitous expression of UAS-Wg:HA using the MatGAL4 driver (Mat Wg) reduces embryonic viability to 1.2%. (B) Ubiquitous Wg expression promotes Wg-regulated cell fates and thus reduces the number of denticle belts formed. (C) Overexpressing UAS- Dsh:Myc using the MatGAL4 driver (Mat Dsh) slightly reduces embryonic viability (83.3 +/- 7.3%). (D) >60% of embryos and larva have a wildtype cuticle phenotype while 29% have a partial loss of denticle belts. (E) Representative images of the phenotypic categories used in B and D. Category 0 = wildtype; Category -1 = At most one denticle belt, or part of 2 belts are absent (arrow); Category -2 = more than half of total denticle belts are present (arrows represent lost denticle belts); Category -3 = less than half of total denticle belts remain, strong head defects or head hole (arrow); Category -4 = A small patch of denticle to none, head hole (arrow).

The kinase GSK3, encoded in Drosophila by the zw3 gene, plays multiple roles in Wnt signaling (MacDonald and He, 2012; Nusse and Clevers, 2017). In addition to its essential role

in regulating Arm/ßcat levels (Bejsovec and Martinez Arias, 1991; Siegfried et al., 1992) by

phosphorylating its degron (Pai et al., 1997; Yost et al., 1996), GSK3 also phosphorylates the

LRP5/Arrow co-receptor, creating Axin binding sites (Zeng et al., 2005). GSK3 also

phosphorylates Axin to regulate its stability and association with ßcat (Willert et al., 1999b;

Yamamoto et al., 1999), and phosphorylates APC on distinct sites to increase its affinity for ßcat

(Ha et al., 2004; Rubinfeld et al., 1996; Xing et al., 2004) or to promote ßcat release to the E3

ligase (Pronobis et al., 2015). The most upstream of these roles is in membrane recruitment of

the destruction complex via receptor phosphorylation—we thus explored whether reducing

GSK3 activity would alter this. We knocked down maternal/zygotic zw3 by RNAi, and examined

Axin:GFP localization. Strikingly, the membrane recruitment of Axin:GFP observed in Wg-ON

cells was lost—instead Axin:GFP formed cytoplasmic puncta in all cells (Fig 2.16D–F). As

expected, Arm levels were strongly elevated (Fig 2.16D). Moreover, we also observed notable

Arm enrichment in the Axin:GFP puncta (Fig 2.16F). This is intriguing; it is consistent with the

role of GSK3 in phosphorylating Arm/ßcat to create an E3 ligase binding site, and with the

proposed role of GSK3 in phosphorylating APC2’s R2/B motifs to stimulate transfer of Arm/ßcat

from the destruction complex to the E3 ligase (Pronobis et al., 2015). Finally, we saw no

significant changes in Axin levels (Fig 2.16G), suggesting any effects on Axin stability were not

substantial.

Simultaneously elevating Axin and APC2 makes destruction complex puncta more resistant to disassembly by Wg signaling

Our phenotypic data above suggest that simultaneously elevating levels of both Axin and

APC2 leads to synergistic inhibition of Wnt signaling. We thus examined how elevating levels of

both Axin and APC2 altered destruction complex assembly and localization. GFP:APC2

expressed alone was primarily cortical (Fig 2.8B and C), as observed for endogenous APC2

(McCartney et al., 1999). In embryos expressing both GFP:APC2 and Axin:GFP at strongly

elevated levels (APC2++Axin++ embryos; Fig 2.14K and L), we observed two notable

differences from what we observed when each was expressed alone. First, the cytoplasmic

puncta in Wg-OFF cells were brighter (Fig 2.14K and L yellow arrows), likely due at least in part

to accumulation of two different GFP-tagged proteins into the puncta. Second and more interesting, the region occupied by bright cytoplasmic puncta became much broader, expanding right up to the Wg-expressing cells (Fig 2.14L, blue arrows), and the region with membrane-

associated puncta became narrower, now largely restricted to the single row of Wg-expressing

cells (Fig 2.14L, magenta arrows). Together with the phenotypic data above (Figs 2.9 and 2.11),

these data suggest that if Axin levels are limiting relative to those of APC2 (i.e., APC2>>Axin),

the destruction complex is more susceptible to being turned down by Wg signaling. In contrast,

if Axin levels are not limiting relative to those of APC2 (Axin≈APC2), then elevating APC2 levels

makes the destruction complex less susceptible to being turned down by Wg signaling. This

state correlates with accumulation in large cytoplasmic puncta, consistent with the idea that this

occurs by stabilizing destruction complex assembly to the effects of Wg signaling.

Each destruction complex punctum includes tens to hundreds of APC2 or Axin proteins

Data from both cultured cells and Drosophila suggest the ability of Axin and APC to polymerize into a large multimeric complex is critical for targeting βcat for destruction.

Polymerization is driven by DIX-domain-mediated head-to-tail Axin polymerization (previously visualized by crystallography and SEM (Schwarz-Romond et al., 2007a)) and by APC’s ability to oligomerize via its N-terminal region and Arm repeats (Kunttas-Tatli et al., 2014).

Overexpressing Drosophila Axin in colorectal cancer cells leads to assembly into large “puncta”,

which we hypothesize are enlarged versions of the normal destruction complex. APC2 is

recruited into these. We used super resolution microscopy to begin to look inside these puncta,

revealing that APC2 and Axin form intertwining filaments (Pronobis et al., 2015). To fully

understand destruction complex assembly and function, one key parameter is to estimate the

number of proteins assembled into active destruction complexes. This has not been possible, either with respect to the large puncta observed after overexpression in colorectal cancer cells, or the presumably smaller complexes produced when Axin and APC2 are expressed at endogenous levels.

To estimate the number of APC2 or Axin molecules within an active destruction complex, we adapted a fluorescence comparison technique developed to quantify numbers of

GFP-tagged proteins in multimeric complexes (Lawrimore et al., 2011; Verdaasdonk et al.,

2014). This technique utilized macromolecular structures containing a known number of GFP

molecules as standards (e.g., purified eGFP = 2 molecules and a virus-like particle = 120

molecules), and from these developed methods to define the number of proteins in yeast

multiprotein complexes where molecule number had not been previously defined. We used 2

yeast strains from this study as standards (Fig 2.18A): one expressing Ndc80:GFP (calculated

to have 306 molecules) and the other expressing Mif2:GFP (calculated to have 58 molecules)

(Lawrimore et al., 2011). Since we thought it likely that destruction complexes did not have a

fixed size, our goal was to get an order of magnitude estimate of the number of proteins in each

destruction complex punctum.

We first examined GFP-tagged Drosophila Axin over-expressed in SW480 cells. Axin uses its DIX domain to polymerize, forming cytoplasmic puncta in a large range of sizes and brightnesses (Pronobis et al., 2015). We compared living yeast and Axin-expressing SW480 cells in parallel (Fig 2.18A), using identical imaging conditions (see Methods for details). Puncta

size in these cells varies over several orders of magnitude (Pronobis et al., 2015), and thus the

brightest puncta in each cell exceeded the linear range of our yeast standards and could not be

analyzed. We determined brightness of individual puncta and used the two yeast standards to

estimate relative brightness and thus relative molecule number. This allowed us to obtain order-

of magnitude estimates of the number of Axin molecules per punctum. In the set we analyzed,

Figure 2.18

Figure 2.18 The destruction complex contains thousands of APC2 or Axin molecules after over-expression in SW480 cells, and 10-100s of Axin molecules in vivo in embryos. (A) Representative images of live samples used for fluorescence comparisons to calculate GFP molecule numbers. Each panel is scaled to the same size and brightness. Ndc80:GFP assembles into a structure containing ~306 GFP molecules while Mif2:GFP assembles into a structure containing ~58 GFP molecules. (B) Pattern of Axin:GFP accumulation and localization in a live embryo. Comparison to our fixed samples allowed identification of regions receiving Wg signal (dimmer puncta) or not receiving Wg signal (brighter puncta). (C-E) Estimated number of GFP molecules per punctum. Each dot = an individual punctum analyzed. Means and standard deviation are in Table 2.8. (C) GFP Molecule counts from SW480 colorectal cancer cells expressing Axin:GFP alone, Axin:GFP plus RFP:APC2, or GFP:APC2 in addition to Axin:RFP. (D-E) GFP molecule counts in vivo from stage 9 embryos expressing RFP and Axin:GFP under the control of MatGAL4 (Mat RFP&Axin). (E) Quantification of puncta GFP molecule counts from D, after being separated into those in presumptive regions receiving or not receiving Wg signals (as in B). Statistical analysis via an unpaired t-test.

the number of Axin:GFP molecules per punctum ranged from 163–1327 (mean ~700; Fig

2.18C; Table 2.8). When APC2 is expressed along with Axin in SW480 cells, it is recruited into

the Axin puncta (Pronobis et al., 2015). We thus also examined SW480 cells coexpressing both

to get order of magnitude comparisons of the number of Axin or APC2 molecules in puncta. In

cells co-transfected for Axin:GFP and RFP:APC2, the number of Axin:GFP molecules ranged

from 104–2041 (Fig 2.18C; Table 2.8), while in cells transfected with a GFP:APC2 and

Axin:RFP, the number of GFP:APC2 molecules per punctum ranged from 162–3297 (Fig

2.18C; Table 2.8), suggesting puncta contain roughly comparable numbers of both proteins.

Because the brightest puncta were outside the dynamic range of our camera, and thus were not

quantifiable using our yeast standards, these data provide a lower bound for molecule number

in the largest puncta. These data suggest that when over-expressed in SW480 cells, APC2 and

Axin can assemble into destruction complexes containing at least 100s to 1000s of each

protein, and within the complex are likely to be present at the same order of magnitude in

molecule number.

While this offered insights into the assembly ability of Axin and APC2, it involved very significant overexpression in an APC mutant colorectal cancer cell line. To assess molecule numbers in an active destruction complex in a natural context and at more normal expression levels, we turned to live Drosophila embryos from the Mat RFP&Axin line. They express

Axin:GFP at 4x-endogenous levels and >60% of these embryos are viable with no or subtle defects in Wg-regulated cell fates (Figs 2.3A, B, Fig 2.4A and B, Tables 2.1 - 2.3). We imaged

Mat RFP&Axin embryos live, in parallel with yeast expressing each of our two protein number

standards (Fig 2.18A). In embryos even the brightest puncta were within the dynamic range of

the camera, and thus could be accurately compared to our yeast standards. Fluorescence

comparison revealed that the Axin:GFP puncta range from 46–931 Axin molecules per punctum

(at stage 9; average ~200; Fig 2.18D, Table 2.8). As noted above, subcellular localization and

apparent brightness of Axin:GFP puncta changed in response to Wg signaling, with the

brightest puncta in the cytoplasm of Wg-OFF cells and dimmer, membrane-bound puncta in

Wg-ON cells. This difference across the segment was apparent in our live Mat RFP&Axin flies

(Fig 2.18B). We used these criteria to separate the puncta into those in Wg-ON versus Wg-OFF

cells. There was a significant difference between the numbers of Axin:GFP molecules in puncta

found in Wg-OFF (average ~260 molecules) versus Wg-ON regions (average ~130 molecules; Fig 2.18E; Table 2.8), although the distributions overlapped. These data provide the

first insight into the scale of macromolecular assembly in an endogenous destruction complex,

suggesting each contains 10s to 100s of Axin molecules. They also support the idea that the

number of Axin molecules per destruction complex decreases in response to Wg signaling.

Dsh accumulates at levels similar to those of APC2 and Axin and localizes to Axin puncta in cells that receive Wg signals

Dsh is a key positive effector of Wnt signaling, acting downstream of the receptors to

downregulate the destruction complex. Dsh can co-polymerize with Axin, competing with Axin

self-polymerization (Fiedler et al., 2011). Data in vivo suggest Dsh and APC can compete for

Axin interaction (Mendoza-Topaz et al., 2011). This suggested the possibility that these two

forms of competition might be part of the mechanism by which Dsh downregulates destruction

complex activity. One key factor in evaluating this possibility are the relative levels of the three

proteins. Our analysis above revealed that Axin and APC2 accumulate at levels within a few-

fold of one another. We adopted a similar strategy to assess the relative levels of Dsh. We

obtained a set of Dsh:GFP transgenes driven by the endogenous promotor (Axelrod, 2001), and

used immunoblotting with an anti-Dsh antibody (Shimada et al., 2001) to explore their levels of

expression (Fig 2.19A). We chose the line that accumulated Dsh:GFP at levels closest to

endogenous Dsh (line Dsh:GFP 2) and compared accumulation of Dsh:GFP to that of

GFP:APC2 and Zyg Axin:GFP (Fig 2.19B, Table 2.1). These data revealed that Dsh

accumulates at levels 2.4±0.5 times that of Axin, suggesting that all three proteins are

Figure 2.19

Figure 2.19 Dsh accumulates at similar levels to Axin and APC2, and co-localizes with Axin puncta in Wg-ON but not Wg-OFF cells. (A) Immunoblot with anti-Dsh antibody, wildtype embryos and three different lines expressing Dsh:GFP driven by its endogenous promotor. Tubulin was the loading control. (B) Left. Immunoblot with anti-GFP antibody, embryos expressing GFP:APC2 driven by its endogenous promotor, Zyg Axin:GFP, or expressing Dsh:GFP driven by its endogenous promotor. Right. Quantification of levels of Dsh:GFP normalized to those of Zyg Axin:GFP. (C-G) Stage 9 embryos, anterior to the left. (C) Dsh localization in a wildtype embryo. There is subtle

enrichment of Dsh at the membrane in cells receiving Wg signal (double arrows). (D) In embryos expressing Axin:GFP at 9x endogenous levels (Mat Axin), Dsh co-localizes with membrane associated Axin puncta in Wg-ON cells (white arrows) but not with cytoplasmic Axin puncta in Wg-OFF cells (blue arrowheads). (E) A similar pattern of Dsh recruitment is seen when Axin:GFP is expressed at 4x endogenous levels (MatAxin&RFP). Insets show higher magnification views of region in box. (F,G) Stage 9 wildtype embryo (F) versus embryo overexpressing Dsh (G), using MatGAL4 to drive UAS-Dsh:Myc (Mat Dsh). The elevation of Dsh levels is apparent, but there is little or no effect on Arm regulation. (H) Immunoblot with anti-Dsh antibody. Wildtype embryos versus embryos overexpressing Dsh (Mat Dsh). Tubulin is a loading control. (I) Immunoblot with anti-Axin antibodies and quantification. Axin levels remain unchanged after Dsh overexpression. A one-way t-test was used to assess the significance of difference in Axin levels.

within a few-fold of one another in abundance and thus competition for oligomerization or

binding among them are plausible.

We next examined whether endogenous Dsh is recruited into the Axin puncta, either in

Wg-ON cells, where it acts to antagonize destruction complex activity, or in Wg-OFF cells where

it does not have a known role. We used an antibody to Dsh (Shimada et al., 2001). In wildtype embryos Dsh was largely cytoplasmic, with weak membrane recruitment in Wg-ON cells (Fig

2.19C double arrows). However, when we elevated Axin expression (Mat Axin; 9x

overexpressed), we saw a striking enhancement of Dsh membrane recruitment in Wg-ON cells.

In those cells, Dsh was recruited into membrane-associated puncta that largely co-localized with

Axin:GFP (Fig 2.19D white arrows). In contrast, in Wg-OFF cells Dsh localization remained

diffusely cytoplasmic, with little or no enrichment in the cytoplasmic Axin puncta (Fig 11D, blue

arrowheads). We saw similar preferential co-localization of Dsh with Axin:GFP in Wg-ON cells

when we expressed Axin at near endogenous levels (4x-elevated) such that Wnt signaling is not

substantially inhibited (Fig 2.19E). These data are consistent with co-recruitment of Dsh and the

destruction complex to the Wnt receptors in response to Wg signal.

Earlier work revealed that substantially overexpressing Dsh can lead to activation of Wg signaling (e.g. (Cliffe et al., 2003; Sokol et al., 1995; Yanagawa et al., 1995)). Given our new

knowledge revealing that relative Dsh and Axin are within a few-fold of one another in

abundance, we examined the effect of elevating Dsh levels, using the MatGAL4 driver to

express UAS-myc-tagged Dsh ((Penton et al., 2002); here after Mat Dsh). This elevated Dsh

levels roughly 7-fold (Fig 2.19F vs G and H), but had only a modest effect on embryonic

viability, reducing it to 83% (Fig 2.17C, Tables 2.1 and 2.2). Cell fates in most embryos were

wildtype, with 29% of the embryos showing mild expansion of naked cuticle (Fig 2.17D and

E, Table 2.3). At this level of overexpression there was little or no effect on the regulation of Arm

levels, as assessed by immunofluorescence (Fig 2.19F’ vs G’), and no apparent effect on levels

of Axin, as assessed by immunoblotting (Fig 2.19I, Table 2.1). These data suggest that at

endogenous levels of Axin and APC2, Dsh is not strongly rate-limiting for Wnt signaling.

DISCUSSION

Wnt signaling plays key roles in cell fate choice and stem cell homeostasis in normal development, and mutational activation underlies colorectal and other cancers. The key regulated step in signaling is regulation of the stability of the Wnt effector ßcat by a multiprotein supermolecular machine, the destruction complex. Despite the intense interest in this pathway, the mechanisms by which Wnt signaling regulates destruction complex activity remains a key question in the Wnt field. Our data provide new insights into this in several ways.

In vivo levels of APC2 and Axin are similar rather than orders of magnitude different

Pioneering work in Xenopus egg extracts defined key parameters underlying the biochemical action of the destruction complex, by assembling and measuring destruction complex activity. These studies lacked reagents to directly measure protein levels of all components, and thus used addition of purified Axin to estimate its relative levels. These data suggested Axin is present at levels much lower than the other components of the destruction complex, with an APC:Axin ratio ~5000:1 (Lee et al., 2001; Lee et al., 2003). Many mathematical and other models in the field use these data as an underlying premise, and thus they have been influential in thinking about Wnt signaling.

However, work in cultured mammalian cells cast doubt on the universality of this ratio— in some cell lines APC levels are much more similar to those of Axin (<2-fold higher) while in others Axin was actually present at higher levels than APC (Tan et al., 2012). We thus used a well-characterized model where the consequences of Wnt signaling are well known: the Drosophila ectoderm during mid-embryogenesis, when cell fate is tightly regulated by Wg

signaling. Using both RNAseq and direct comparisons of protein levels, we found that, in

contrast to Xenopus oocyte extracts, APC and Axin levels are quite similar: our protein data

suggest the APC2:Axin ratio is 5:1. In fact, this may overestimate available levels of APC2, the

primary APC family member at this time. APC proteins have distinct cytoskeletal roles (Aoki and

Taketo, 2007) at times including those just prior to those we examined (Poulton et al., 2013), and thus the pool of APC2 available for Wg signaling may be even lower. While it is possible that the difference in results involve the species used (Xenopus vs. Drosophila), our data and

the mammalian cultured cell data suggest that the difference may be in comparing tissues

where Wnt signaling is active, versus those, like Xenopus egg extracts, in which Wnt signaling

is not yet active. Thus, future mathematical modeling of Wnt signaling should include states in

which APC and Axin are present at similar levels.

In the absence of Wg signaling, Axin assembles into large cytoplasmic multiprotein complexes that each contain tens to hundreds of Axin proteins

Previous work provided conflicting results on Drosophila Axin localization in the absence

and presence of Wg signaling. In interstripe cells, which receive little or no Wg signal, the

destruction complex is in an active state. Cliffe et al. (2003) suggested that Axin and APC2 co-

localize in cytoplasmic puncta in these cells (Cliffe et al., 2003), while others did not see any

notable subcellular localization of Axin in Wg-OFF cells (Tolwinski et al., 2003; Yang et al.,

2016). To address this, we examined Axin:GFP localization in embryos expressing Axin below

the level that substantially alters embryonic viability or cell fates (Mat RFP&Axin;

4xendogenous). Our data confirm and extend the work of Cliffe et al.(Cliffe et al., 2003). In Wg-

OFF interstripe cells, Axin assembled into large cytoplasmic puncta, presumably driven by DIX-

domain mediated head-to-tail Axin polymerization, as previously visualized by crystallography

and SEM (Schwarz-Romond et al., 2007a). In these cells, levels of cytoplasmic Axin were

relatively low, suggesting that much of the Axin self-assembles into puncta. Earlier data and our

work reveal that these puncta also contain APC2 ((Cliffe et al., 2003); Fig 8), and thus APC2’s

ability to multimerize may also be relevant (Kunttas-Tatli et al., 2014; Pronobis et al., 2015). Our

molecular counting experiments also provided the first assessment of the number of Axin

molecules in the multiprotein destruction complex. These data suggest active destruction

complexes contain tens to low hundreds of Axin proteins, thus helping explain the critical role of the Axin DIX domain (Kishida et al., 1999; Peterson-Nedry et al., 2008), which mediates Axin

polymerization (Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b). Our recent

work to engineer the minimal Wnt regulatory machine confirmed that both Axin’s DIX domain

and APC2’s Arm repeats, implicated in polymerization, are among the domains most critical for

destruction complex function (Pronobis et al., 2016).

Wg signaling triggers membrane recruitment of Axin and may destabilize destruction complex assembly

One key and controversial question in the field involves the mechanism(s) by which Wg

signaling turns down the destruction complex. Different studies in cultured mammalian cells

and Drosophila (see Introduction) led to quite different conclusions, ranging from total

destruction complex disassembly to inactivation of an intact complex to Axin stabilization. Our

new tools allowed us to examine Axin localization directly using a GFP-tagged protein

expressed at near endogenous levels, in a tissue where we can examine cells before the onset

of Wg signaling, as well as in side-by-side cells experiencing high or low levels of signaling. Our data suggest that in this tissue, Wg signaling leads to membrane recruitment of the destruction complex and are consistent with the idea that it destabilizes assembly, thus increasing the cytoplasmic Axin pool. Before Wg signaling is initiated, Axin:GFP was in cytoplasmic puncta in all cells. However once Wg signaling initiated, Axin:GFP localization differed between cells. In

cells not receiving Wg, much of the Axin assembled into large cytoplasmic puncta, leaving

relatively low levels diffuse in the cytoplasm. However, in cells receiving Wg signal, the Axin

puncta were recruited to the membrane. Our molecular counting and image thresholding

experiments suggest these dimmer membrane-proximal puncta contain fewer Axin molecules.

Our image thresholding experiments further suggest that in Wg-receiving cells, diffuse

cytoplasmic levels of Axin are elevated in Wg-ON relative to Wg-OFF cells.

100

These data support and extend the earlier work of Cliffe et al. (2003) (Cliffe et al., 2003), who expressed GFP-tagged Axin at more elevated levels. Our observation of elevated cytoplasmic levels of Axin in Wg-ON cells is also consistent with earlier work (Yang et al., 2016).

However, we did not see clear evidence that this results from Axin protein stabilization, as none of our manipulations of Wnt signaling components significantly altered total Axin levels. Our data further suggest that previous use of antibody staining of an epitope-tagged protein rather than direct visualization emphasized the diffuse cytoplasmic pools of Axin in Wg-ON cells while simultaneously de-emphasizing the larger cytoplasmic puncta in Wg-OFF cells, due to differential antibody accessibility. Thus, the stabilization of Axin proposed previously (Tan et al.,

2012) may largely involve a change in protein localization, rather than a change in total Axin

levels. This is consistent with the immunoblotting experiments of Cliffe et al (2003), who did not

detect altered Axin:GFP levels upon ubiquitous expression of Wg. While Yang et al. 2016

observed an increase in Axin levels by immunoblotting after the onset of Wg signaling in

wildtype embryos (Yang et al., 2016), this simply reflects activation of the zygotic genome.

Our data and earlier data also help define how different components of the Wnt pathway

regulate destruction complex localization and assembly. APC2 and Axin co-assemble into large

cytoplasmic puncta in Wg-OFF cells. In response to Wg signaling these puncta are recruited to

the membrane and reduced in Axin protein number—this recruitment does not occur

in wgmutants whereas Wg overexpression triggers membrane recruitment in all cells ((Cliffe et

al., 2003); Fig 9). Mendoza-Topaz et al. (2011) found that APC2 is critical for assembly of Axin

puncta in both Wg-ON and Wg-OFF cells (Mendoza-Topaz et al., 2011). GSK3/Zw3 is important

for membrane-recruitment of Axin puncta, labeling Arm to be recognized by the E3 ligase, and

for release of Arm from these puncta (Fig 9). Dsh is specifically recruited into Axin puncta at the

membrane of Wg-ON cells (Fig 11), and its overexpression accentuates membrane recruitment

of both APC2 and Axin (Cliffe et al., 2003). Integrating our data with this earlier work, we

hypothesize that in the presence of Wg signaling, Axin puncta are recruited to the membrane,

101

presumably by binding to the activated Wg-receptor. We further hypothesize that Wg signaling, acting at least in part via Dsh, either destabilizes puncta or inhibits puncta assembly, increasing the relative amount of Axin in the cytoplasmic pool.

Elevating Axin levels renders the destruction complex less sensitive to inactivation by Wg signaling

Previous work revealed that sufficiently elevating Axin levels could inactivate Wnt signaling either in cultured mammalian cells (Nakamura et al., 1998) or in Drosophila embryos

(Cliffe et al., 2003; Willert et al., 1999a), demonstrating that Axin is rate-limiting. More recent work revealed that this only occurred when Axin levels were elevated over a certain level

(Peterson-Nedry et al., 2008; Wang et al., 2016b). Our knowledge of absolute levels of APC2 and Axin allowed us to vary levels of each individually or together, thus varying both levels and ratios of the two proteins in the Drosophila embryo where effects of Wg signaling are well characterized. By assessing the effect on embryonic viability, expression of the target gene en, and cell fates choices, we defined the effects of different Axin levels and different timing of Axin accumulation. When Axin:GFP was expressed at ≤4x endogenous Axin, we observed little or no effect on any of these parameters, while at >8x endogenous Axin there was a dramatic increase in embryonic lethality, reduced En expression, and a shift towards a more wg-null like

phenotype. These data confirm that Axin can be rate-limiting in vivo. Our data also provided

insight into the underlying mechanism: increasing Axin ≥4x rendered the destruction complex

less sensitive to inactivation by Wg signaling, and thus decreased Arm levels specifically in Wg-

ON cells. Further mechanistic insights remain to be determined, but a key parameter may be

the levels of “active Dsh” protein, which is activated by Wg signaling and then can

heteropolymerize with Axin and compete with APC (Cliffe et al., 2003; Mendoza-Topaz et al.,

2011; Schwarz-Romond et al., 2007a). Our data reveal that Dsh levels are in the same order of magnitude as those of Axin, making competition plausible. If down-regulation involves a competition between Axin homo-multimerization and Dsh hetero-multimerization, sufficient

102

elevation of Axin levels may saturate the available Dsh molecules and therefore inhibit its ability

to inactivate the destruction complex, thus rendering a subset of destruction complexes immune

to downregulation. In contrast, our data suggest Axin is not rate-limiting in Wg-OFF cells—Arm

levels there were not further decreased by elevating Axin levels.

APC2 is not rate-limiting for destruction complex activity but elevating its levels facilitates destruction complex inactivation

We next asked whether APC2, the second core component of the destruction complex, is also rate-limiting for destruction complex activity. Expressing GFP:APC2 at >10x endogenous levels had little to no effect on embryonic lethality, En expression, or Wg-regulated cell fate choices. This might be because Axin is rate-limiting—thus additional APC2 would not trigger assembly of additional destruction complexes once it exceeded the available pool of Axin.

Intriguingly, however, we observed an unexpected effect of elevated levels of APC2. In wildtype the normal gradient of Wg creates a gradient of Arm accumulation. In contrast, in embryos with high APC2 expression there is an essentially binary change in Arm accumulation in response to

Wg signaling. Wg–expressing cells and their immediately adjacent neighbors accumulate Arm at levels ~1.5x higher than the same cells in wildtype. However, in cells more distant from the

Wg-expressing cells, Arm levels are unchanged from wildtype. These data suggest a potential positive role for APC2 in turning the destruction complex down in the presence of Wg signaling.

While this paper was under review, another paper was published, suggesting a different role of

APC in inhibiting the destruction complex in response to Wg signal, by modulating Axin phosphorylation by GSK3 (Tacchelly-Benites et al., 2018).

Effects of altering the Axin:APC2 ratio suggest APC2 can play both positive and negative roles in Wnt regulation

These paradoxically opposite effects of elevating levels of Axin or APC2 on the ability of

Wg signaling to inactivate the destruction complex were surprising. Our dual over-expression of

APC2 and Axin provided potential insight into the underlying mechanisms, emphasizing the

103

importance of the relative ratios of different destruction complex proteins. In embryos with

elevated expression of both APC2 and Axin, the destruction complex was even less effectively

turned down by Wg signals than after overexpression of Axin alone, and large cytoplasmic

Axin/APC2 puncta were found in cells immediately adjacent to those expressing Wg, rather than

being confined to cells with lower levels of Wg signaling. As noted above, this may occur in

situations where Axin levels exceed those of active Dsh. These data are consistent with the

known role of APC proteins in promoting Arm/ßcat destruction, and fit well with our work in

cultured mammalian cells, which demonstrated that APC can stabilize Axin multimerization, thus

increasing destruction complex size and its effective activity (Pronobis et al., 2015).

However, if Axin was limiting, effects of APC2 elevation were quite different. Now elevated levels of APC2 allowed Wg signaling to more effectively turn down the destruction

complex, thus elevating Arm levels. It is possible that when APC2 levels exceed those of Axin, it

forms incomplete subcomplexes with other destruction complex proteins, thus titrating their

levels. Since Axin directly interacts with GSK3 and CK1 while APC2 does not, we do not think

the effects of elevating APC2 occurred solely by sequestering those proteins in partially

assembled and inactive complexes. One potential explanation is that if APC2 levels exceed

those of Axin, it binds and sequesters Arm in binary APC2:Arm complexes, protecting Arm from

destruction by the remaining functional destruction complexes. Previous work revealed that the

ability of APC2 to retain Arm in the cytoplasm fine-tunes Wg signaling independent of its role in

destruction (Roberts et al., 2011; Yamulla et al., 2014). Alternately, these data are consistent

with a model in which APC2 has dual positive and negative roles in Wnt regulation, as was

previously suggested to occur in the eye and wing imaginal discs (Takacs et al., 2008).

However, the mechanism by which this occurs remains unclear. Our immunolocalization and

immunoblotting data do not support the hypothesis that APC2 overexpression promotes Axin

turnover, as was previously suggested based on loss-of-function analysis (Takacs et al., 2008;

Wang et al., 2016b). Perhaps the membrane-associated pool of APC2 brings both Axin and the

104

destruction complex in proximity to the Wg receptors and Dsh, allowing more rapid and effective

turndown of destruction complex function and assembly.

Our data are also consistent with the hypothesis that relative levels of Axin and active

Dsh are key, as we found Axin and Dsh accumulate at quite similar total levels. It would be interesting to examine the effect of varying Dsh levels on destruction complex assembly, localization and function. If Dsh levels are a rate-limiting factor in turning off the destruction complex, then increasing Dsh may balance the effects of elevating Axin. We only saw modest effects of elevating Dsh levels 7-fold, suggesting that elevating Dsh levels alone may not be sufficient, if the ability of the Wg receptor to “activate” the pool of Dsh is limiting. Our data on

Dsh localization suggest that its ability to co-localize with Axin requires “activation” by Wg signaling. Moving forward, it would be interesting to count the number of Dsh molecules in the membrane-associated Axin puncta to determine how Dsh molecule numbers within puncta relate to those of Axin or APC2. We also need to explore the ratio of APC2:Axin within the

destruction complex in vivo, to parallel the work after overexpression in cultured mammalian

cells reported here. Wg signaling may alter these ratios and thus regulate destruction complex

function. Intriguingly, after Axin over-expression, recruitment of APC2 into Axin puncta seemed

more robust in Wg-ON than in Wg-OFF cells (Fig 8J and 8J’-yellow vs. magenta arrows).

One striking aspect of our manipulations was that the effects of elevating levels of APC2 or Axin appeared largely confined to the cells receiving Wg signal. These data are consistent with the idea that normal levels of Axin and APC2 are not rate-limiting in Wg-OFF cells, and

thus the alterations in APC2:Axin ratio we created do not alter the efficiency of Arm turnover

there. In contrast, the cells receiving Wg signal seem much more sensitive to any manipulation

that reduced destruction complex function. These include the elevated APC2:Axin ratio created

by elevating APC2 levels, or the reduced ability of APC2Δ15ΔR1,ΔR3-5 to bind and retain Arm

in the cytoplasm (Yamulla et al., 2014). One speculative explanation for this is that in those

cells, the system is finely balanced, due to competitive interactions between Dsh and Axin, Axin

105 and Axin, and Axin and APC2. In such a scenario, relatively small changes in the levels of any one of these proteins could change the balance of active versus inactive destruction complexes.

A proposed model of how Wnt signaling regulates destruction complex assembly and function

Our data suggest that the relative levels of different Wnt signaling regulators are a critical determining factor, and that modulating relative levels of different components may make cells more or less sensitive to Wnt signaling. Integrating our data with data from many labs using Drosophila and cultured mammalian cells, we propose the following model. During embryogenesis, cells begin with relatively similar levels of the two core scaffolds of the destruction complex (4-5x more APC2 than Axin), and levels of total Dsh protein are in the same range. When Wg signaling is off, Axin self-assembles into cytoplasmic complexes of tens to hundreds of molecules, which we believe represent the functional destruction complex. In this state, much of the Axin in the cell is assembled into puncta, with less free in the cytoplasm. In these cells, we propose that there are 2 pools of APC2, a pool localized to the cortex that mediates APC2’s cytoskeletal functions, and another that is associated with and stabilizes the assembly of the Axin puncta. This would represent a high activity state of the destruction complex, and it would rapidly bind, sequester and turnover all newly synthesized Arm that is not assembled into adherens junctions. In these cells, which lack activation of the Wg receptors,

Dsh is not competent to integrate into Axin complexes at levels sufficient to antagonize destruction complex function. In the presence of Wg signaling, LRP5/6 is recruited to the Wg receptor Frizzled, and phosphorylation of LRP5/6 by GSK3 and other kinases recruits Axin and

Dsh to the membrane, in the process activating Dsh so it can be incorporated into Axin complexes. We hypothesize that Axin membrane-recruitment involves largely intact destruction complexes. Our observations are consistent with recent data suggesting that the destruction complex is not fully disassembled in response to Wg signaling nor is Arm phosphorylation by the destruction complex completely inhibited (Hernandez et al., 2012; Kim et al., 2013; Li et al.,

106

2012; Pronobis et al., 2015). Instead the ability of the destruction complex to target Arm for

destruction is reduced, perhaps in part by blocking the ability to transfer Arm to the E3 ligase.

Dsh contains a DIX domain, like Axin, which allows Dsh to hetero-dimerize with Axin. We

hypothesize that this Dsh:Axin interaction aids in puncta re-localization and stimulates the

decrease in destruction complex size and function. Dsh may actively reduce the size of

destruction complexes by competition with Axin:Axin multimerization, or it may compete with

APC2 for access to Axin. Other longer-term effects may then reinforce this initial event,

including ubiquitination and destruction of Axin or inhibition of GSK3 kinase activity.

MATERIALS AND METHODS

Fly stocks, embryonic lethality, and cuticles

All crosses were performed at 25°C. Wildtype was either y w or act5c-Gal4/CyO. The following stocks were obtained from the Bloomington Stock Center: act5c-

GAL4 (4414), Maternal alpha tubulin GAL4 (referred to as MatGAL4; a stock carrying both of

the GAL4 lines in 7062 and 7063), UAS-Axin:GFP (7225), UAS-Dsh:Myc (9453), UAS-RFP

(30556), wgIG22 (5351), UAS-Axin-RNAi (62434), UAS-zw3-RNAi (35364), and UAS-Wg:HA

(5918). We also used UAS-GFP:APC2 (Roberts et al., 2011) and an APC2 transgene which

expresses APC2 under its endogenous promoter (Roberts et al., 2011). Dsh:GFP 2, a Dsh:GFP

transgene expressed under its endogenous promoter is from (Axelrod, 2001). Dsh:GFP 2.33

and Dsh:GFP 2.35, which are derivatives of Dsh:GFP after local gene hopping, were both kind

gifts from J. Axelrod (Stanford).

Cross Abbreviations (Female x Male):

GFP:APC2 = APC2 promoter-GFP:APC2; APC2g10 x APC2 promoter-GFP:APC2; APC2g10

Zyg Axin = UAS-Axin:GFP x act5c-GAL4/+

Mat RFP&Axin = UAS-RFP/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-RFP; UAS-Axin:GFP

Mat Axin = +/MatGAL4; UAS-Axin:GFP /MatGAL4 x UAS-Axin:GFP

107

Mat/Zyg Axin = act5c-GAL4/+ x UAS-Axin:GFP.

Mat APC2 = UAS-GFP:APC2/MatGAL4; +/MatGAL4 x UAS-GFP:APC2

Mat APC2& Axin = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-GFP:APC2;

UAS-Axin:GFP

APC2 >> Axin = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-GFP:APC2

Axin >> APC2 = UAS-GFP:APC2/MatGAL4; UAS-Axin:GFP/MatGAL4 x UAS-Axin:GFP

Axin RNAi = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin-RNAi/+

Mat Axin-RNAi x Axin:GFP = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin:GFP

Mat Axin-RNAi x RFP = UAS-Axin-RNAi/MatGAL4; +/MatGAL4 x UAS-RFP

Mat RFP x Axin:GFP = UAS-RFP/MatGAL4; +/MatGAL4 x UAS-Axin:GFP

wgIG22 mutant = wgIG22/MatGAL4; +/MatGAL4 x wgIG22/+; UAS-Axin:GFP/+ zw3 RNAi = UAS-zw3-RNAi/MatGAL4; +/MatGAL4 x UAS-zw3-RNAi

zw3 RNAi x Axin = UAS-zw3-RNAi/MatGAL4; +/MatGAL4 x UAS-Axin:GFP

Mat x Wg = UAS-Wg:HA/MatGAL4; +/MatGAL4 x UAS-Wg:HA/UAS-Axin:GFP

Mat Dsh = UAS-Dsh:Myc/MatGAL4; +/MatGAL4 x UAS-Dsh:Myc

Embryonic lethality assays and cuticle preparations were as in (Riggleman et al., 1989).

Inhibition of Wg signaling was assessed by analyzing embryonic and first instar larvae cuticles

with the scoring criteria found in Fig 3C and S8 Fig.

Immunostaining and antibodies

Embryos were prepared as in (Fox and Peifer, 2007). Briefly, flies were allowed to lay eggs on apple juice/agar plates with yeast paste for up to 7 hours. Embryos were collected in

0.1% Triton-X in water using a paintbrush, then dechorionated for 5 minutes in 50% bleach.

Embryos were fixed for 20 minutes in 1:1 heptane to 9% formaldehyde, with 8mM EGTA added to preserve GFP expression. Embryos were then devitillenized by vortexing in 1:1 heptane to methanol. Embryos were then washed in methanol followed by 0.1% Triton-X in PBS, then

108

incubated in blocking buffer (1:1000 normal goat serum diluted in 0.1% Triton-X in PBS) for 30 minutes. Embryos were incubated in primary overnight at 4°C, washed in 0.1% Triton-X in PBS, then incubated in secondary antibody for 1 hr at room temperature. Embryos were mounted in

Aqua polymount (Polyscience). Primary antibodies were: Wingless (Wg, Developmental Studies

Hybridoma Bank (DSHB):4D4, 1:1000), Arm (DSHB:N27 A1, 1:75), phospho-tyrosine (pTyr,

Millipore:4G10, 1:1000), En (DSHB:4D9, 1:50), GFP (Abcam:ab13970, 1:10,000), Neurotactin

(Nrt, DHSB:BP 106, 1:100), APC2 (McCartney et al., 1999), 1:1000), and Dsh ((Shimada et al.,

2001); 1:4000).

Assessing effects on Engrailed expression

Stage 9 embryos were stained with antibody to Engrailed and imaged on a Zeiss LSM

710 or 880 scanning confocal microscope. Images were processed using FIJI (Fiji Is Just

ImageJ) as follows: maximum intensity projections 8μm thick were created and thresholded to highlight cells expressing Engrailed and eliminate background noise. Three lines parallel to the midline were drawn to intersect with bands 2 through 5 of Engrailed expressing cells relative to the head, two on either side of the embryo and one just to the left of the midline. The cells in each Engrailed band which were intersected by each line were included in our measurements.

The number of cells per Engrailed stripe was then determined by averaging these three values.

Embryos were scored blind. Significance was assessed using a one-way ANOVA test.

Quantitative analysis of Arm accumulation

Graded accumulation of Arm.

To quantify effects of our manipulations on the graded accumulation of Arm across the segment, control and experimental embryos were stained in parallel for phosphotyrosine, which marks the adherens junctions, and Arm. Since 70% of Arm protein accumulates at the adherens junction (Peifer, 1993), and we wanted to focus on the cytoplasmic and nuclear accumulation of

Arm, thus we removed the adherens junction pool of Arm by creating a mask. First, all embryos

109

were rotated to have the anterior on the left and the midline at 180°. Next sum intensity

projections were created that went 8μm deep into the embryo. We next used FIJI’s trainable

WEKA tool with anti-phosphotyrosine staining to create a membrane mask. This mask was overlaid and subtracted from the Arm image. Next a rectangular region of interest (ROI) was drawn (446 W x 60 H pixels, spanning approximately 3 Wg stripes and 4 cells wide) starting at the first interstripe in the thorax. A profile of the ROI was plotted. ROI profiles were adjusted for embryo length and to bring valleys to zero. See Fig 2.6A for a visualization this process.

Wg stripes versus interstripes.

To calculate the absolute levels of Arm accumulation in cells receiving or not receiving

Wg signals, stage 9 embryos were collected and stained as previously described. For each genotype, we added act5c-GAL4/Cyo embryos to the same tube as a wildtype control, allowing

immunostaining and microscopy imaging on the same slide. Control and experimental embryos

were distinguished by the presence or absence of GFP-fluorescence. To calculate the level of

Arm accumulation, we choose a specified boxed region (100 pixels wide x 30 pixels high)

spanning the width of the Wg-expressing cells, and measured the mean gray value of Arm by

FIJI (Fig 2.6B). Three Wg stripe regions from parasegments 2 to 4 were measured, and the

average Arm value minus the background value from a region outside the embryo was defined

as the Wg stripe Arm value. In the adjacent interstripe regions we used the same box size to

measure and calculate Interstripe Arm values. We also measured the relative difference in Arm accumulation between the Wg Stripes and Interstripes. We also measured the GFP fluorescence of the same boxed regions, allowing us to determine the GFP expression level in different embryos—at times this was used to infer possible genotypes.

To separate and measure the membrane and cytoplasmic pools of Arm, images were rotated so anterior was at the top. A ROI was created (160 W x 20 H pixels, spanning approximately 10–15 cells wide and 2 cells high) which was confined to cells in Wnt OFF or Wnt

ON regions. First, a total fluorescence intensity of the ROI was measured and adjusted for

110

background, and then divided by the area of the ROI. Next a membrane mask was created as

above and used to measure the intensity and area of the membrane pool of Arm. Next, the

membrane mask was used to subtract all of the membrane intensity. Lastly, the remaining

intensity in the ROI was measured and recorded as the cytoplasmic/nuclear pool of Arm. To determine the area of the cytoplasmic/nuclear pool of Arm, the area of the membrane was subtracted from the area of the ROI.

Statistics

Wg-Stripe and Interstripe Arm level values were generally normally distributed, as tested by the D’Agostino-Pearson omnibus normality test as well as the Shapiro-Wilk normality test, and thus parametric tests were employed in statistical analysis. The Paired t-test was used to determine the significance between intragroup values, and an unpaired t-test was used to determine the significance between intergroup values. For multiple comparisons, ordinary one- way ANOVA followed by Dunnett's multiple comparisons test were applied.

Immunoblotting

4-8hr old embryos were collected in 0.1% Trition-X100, dechorionated in 50% bleach, and then homogenized with a pestle in RIPA buffer (1% NP-40, 0.5% Na deoxycholate, 0.1%

SDS, 50mM Tris pH 8, 300 mM NaCl; 1x Halt Protease and Phosphatase Inhibitor (Thermo

Scientific)). Protein concentrations were calculated using Protein Assay Dye (BioRad) following the manufacturer’s recommendations. Samples were mixed with SDS-PAGE sample buffer, boiled for 5 minutes and then run on an 8% SDS-PAGE gel and transferred to a nitrocellulose membrane. Westerns were visualized using a CLX Licor machine which allowed blots to be imaged over a 4-log range. Band densitometry was calculated using LICOR Image Studio and significance was assessed using a one-sample t test using GraphPad. When band densitometry

values differed by more than 5-fold, serial dilutions of samples were used to verify values.

Values acquired from these dilutions were reported. Primary Antibodies: anti-GFP (JL-8

111

Clontech, mouse monoclonal, 1:1000), anti-Axin (a kind gift from Y. Ahmed, guinea pig

polyclonal, 1:1000), anti-γ-tubulin (Sigma-Aldrich, mouse monoclonal, 1:2000), anti-APC2

(Rabbit polyclonal, a kind gift of M. Bienz (Yu et al., 1999), 1:1000), and anti-Dsh ((Shimada et

al., 2001), 1:1000). Secondary Antibodies: IRDye680RD anti-Rabbit (Licor, 1:10,000),

IRDye680RD anti-Guinea pig (Licor, 1:10,000), and IRDye800CW anti-Mouse (Licor 1:10,000).

RNA-Seq

mRNA collection and RNAseq analysis are described in (McKay and Lieb, 2013) (GEO

accession number GSE38727).

Cell culture and transfections

SW480 cells were cultured at 37° C at normal atmospheric levels of CO2 in L15-media

(Cellgro) supplemented with 10% FBS and 1x penicillin–streptomycin. Drosophila APC2 or Axin

protein constructs were transfected into SW480s using Lipofectamine 2000 (Invitrogen) as

recommended by the manufacturer. Cells were imaged 24 hours later. Full

length DrosophilaAPC2 and Axin were cloned with either a GFP, RFP, or Flag tag as in

(Pronobis et al., 2015).

To verify that Axin:GFP polymerization is not simply a result of di- or oligomerization of

the GFP protein, we created a monomeric GFP (mGFP) by changing Alanine 206 to Leucine

(Zacharias et al., 2002). We created the A206K amino acid change in our base plasmid (pCMV-

Axin:GFP;(Pronobis et al., 2015)) using QuikChange (Agilent Technologies) following

manufacturer’s recommendations. The full plasmid was sequenced to verify the amino acid

change in GFP and to ensure no other detrimental mutations were induced in the plasmid.

Yeast fluorescence comparison

Yeast Fluorescence comparison analysis was performed as described in (Lawrimore et

al., 2011; Verdaasdonk et al., 2014). Briefly, yeast were grown at 24°C in YPD media until they

reached an OD between 400–600. Yeast cells were then pelleted and resuspended in YC

112 complete media for live imaging. To adjust for possible background caused by media, both

SW480 cells and embryos were also suspended in YC complete medium for live imaging.

Images were taken using the same settings on the same day for each experiment. Each slide was imaged for no longer than 20 minutes at room temperature (~25°C). For analysis, a 15 x

15pixel ROI was created around each punctum, then a 21 x 21 pixel ROI was made around the smaller ROI for background subtraction. Puncta distance from the coverslip and depth of field

(number of Z slices containing a single punctum) were both taken into account when calculating the molecule numbers (as in (Verdaasdonk et al., 2014)). To verify molecular counting, each punctum was compared to 2 different yeast strains: Ndc80:GFP (~306 molecules of GFP) and

Mif2:GFP (~58 molecules of GFP; both kind gifts from K. Bloom (Lawrimore et al., 2011)). Data sets were only used when molecular numbers were consistent (+/- 15 molecules) when calculated with both the Ndc80 and Mif2 standards. Due to the dynamic range of the camera, we were limited in the brightness of puncta that we were able to image in SW480 cells.

Therefore, the brightest (presumed largest) puncta were omitted from analysis, suggesting that our molecular counts in SW480 cells are an underestimate, as noted in the Results. However, all puncta in embryos were within the camera’s dynamic range. Images were taken on a Zeiss

LSM 710 confocal microscope using a Licor LSM-T-PMT camera, with a 488nm diode for stable illumination on a 100x/1.4 NA objective lens. Images were analyzed using FIJI. For a more detailed description of how to calculate molecular number see (Verdaasdonk et al., 2014).

113

Table 2.1

Table S1: Normalized densitometry values Significant Cross (Female x Male) Abbreviation Antibody Mean Std. Dev. n= from Wt+ act5cGAL4/Cyo WT - 1.0 - - -

UAS-Axin:GFP x act5cGAL4/Cyo Zyg Axin Axin 1.0 0.5 No 10 UAS-RFP/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat RFP & Axin Axin 3.1 1.0 ** 5 UAS-RFP; UAS-Axin:GFP actc5cGAL4/Cyo x UAS-Axin:GFP Mat/Zyg Axin Axin 7.4 3.0 * 4 +/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat Axin Axin 7.9 3.0 * 4 UAS-Axin:GFP APC2>GFP:APC2; APC2 g10 GFP:APC2 APC2 0.9 0.4 No 4

APC2>GFP:APC2; APC2 g10 GFP:APC2 GFP# 4.3 1.4 *** 8 UAS-GFP:APC2/MatGAL4; +/MatGAL4 x Mat APC2 GFP## 10.9 0.4 *** 4 UAS-GFP:APC2 UAS-GFP:APC2/MatGAL4; UAS- Mat APC2 & Axin Axin 2.9 1.4 ** 9 Axin:GFP/MatGAL4 x UAS-GFP:APC2; UAS- UAS-GFP:APC2/MatGAL4; UAS- Mat APC2 & Axin GFP## 20.8 1.7 ** 3 Axin:GFP/MatGAL4 x UAS-GFP:APC2; UAS- + = significance calculated using one-sample student t test; * = p value < 0.05, ** = p value < 0.005, *** = p value < 0.0005; # = Normalized to Zyg Axin; ## = Normalized to GFP:APC2

Table 2.1: Normalized densitometry values. Quantification of protein levels of endogenous and GFP-tagged proteins via immunoblotting followed by band quantification using the CLX Licor, which allows samples to be quantified in a 4-log range. When differences were ≥ 5-fold, dilutions were used to verify levels. Crosses and their abbreviation are labelled. Means, standard deviation, and number of blots quantified are indicated. Significance was calculated using a one-sample t-test.

114

Table 2.2

Table S2: Embryonic viability Cross (Female x Male) Abbreviation Viability StdDev n=

UAS-Axin:GFP x UAS-Axin:GFP UAS-Axin:GFP Only 96.7 1.5 243

UAS-Axin:GFP x act5cGAL4/Cyo Zyg Axin 95.4 3.5 409

UAS-RFP/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat RFP & Axin 68.1 6.3 568 UAS-RFP; UAS-Axin:GFP

+/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat Axin 22.2 2.3 900 UAS-Axin:GFP actc5cGAL4/Cyo x UAS-Axin:GFP Mat/Zyg Axin:GFP 9.5 8.0 275

UAS-GFP:APC2/MatGAL4; +/MatGAL4 x Mat APC2 94.5 4.6 451 UAS-GFP:APC2 UAS-GFP:APC2/MatGAL4; UAS- Axin:GFP/MatGAL4 x UAS-GFP:APC2; UAS- Mat APC2 & Axin 36.9 3.9 515 Axin:GFP UAS-GFP:APC2/MatGAL4; UAS- APC2>>Axin 79.5 7.3 667 Axin:GFP/MatGAL4 x UAS-GFP:APC2

UAS-GFP:APC2/MatGAL4; UAS- Axin>>APC2 29.3 6.1 791 Axin:GFP/MatGAL4 x UAS-Axin:GFP

Table 2.2 Embryonic viability Quantification of embryonic viability after altering levels of GFP:APC2 and/or Axin:GFP, Wg:HA, Dsh:Myc, or Zw3. Crosses, embryonic viability, standard deviation and numbers of embryos assayed are indicated.

115

Table 2.3

Table S3: Embryonic and first instar larva cuticle phenotypes % in % in % in % in % in Cross (Female x Male) Abbreviation n= Category 1 Category 2 Category 3 Category 4 Category 5

UAS-Axin:GFP x act5cGAL4/Cyo Zyg Axin 75.1 18.3 6.2 0.4 0.0 149

UAS-RFP/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat RFP & Axin 62.5 18.7 6.2 11.7 1.0 368 UAS-RFP; UAS-Axin:GFP actc5cGAL4/Cyo x UAS-Axin:GFP Mat/Zyg Axin 10.0 10.5 13.0 19.1 47.5 323

+/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat Axin 26.9 9.5 12.6 17.3 33.7 524 UAS-Axin:GFP UAS-GFP:APC2/MatGAL4; +/MatGAL4 x Mat APC2 82.7 15.4 1.2 0.2 0.4 122 UAS-GFP:APC2 UAS-GFP:APC2/MatGAL4; UAS- Axin:GFP/MatGAL4 x UAS-GFP:APC2; UAS- Mat APC2 & Axin 29.0 14.8 4.7 24.9 26.7 329 Axin:GFP UAS-GFP:APC2/MatGAL4; UAS- APC2>>Axin 51.6 35.9 5.5 2.5 4.5 128 Axin:GFP/MatGAL4 x UAS-GFP:APC2 UAS-GFP:APC2/MatGAL4; UAS- Axin>>APC2 16.2 21.3 12.7 10.7 39.1 635 Axin:GFP/MatGAL4 x UAS-Axin:GFP

Table 2.3: Embryonic and first instar larva cuticle phenotype. (A)Effects of GFP:APC2 and/or Axin:GFP, Wg:HA, Dsh:Myc, or Zw3 manipulations on embryonic and first instar larva cuticle phenotypes. Fig 2.7C show the cuticle categories. (B) Effects of Axin shRNA, overexpression of Wg or Dsh on embryonic and first instar cuticle phenotypes. Fig 2.7C and Fig 2.17E show the cuticle categories. n = number of embryos/larva scored.

116

Table 2.4 Table S4: Rows of En-expressing cells per segment Avg. # of cell rows Significant Cross (Female x Male) Abbreviation Std. Dev. n= with En from wt+ act5cGAL4/Cyo WT 1.8 0.2 - 21

UAS-Axin-RNAi/MatGAl4; MatGAl4 x Axin RNAi 2.7 0.3 * 5 UAS-Axin-RNAi/MatGAl4; MatGAl4

APC2 g10 (APC2 null) APC2g10 2.9 0.6 **** 3 wg IG22 /+ (wg null) wgIG2 2 0.5 0.5 **** 14

UAS-Axin:GFP x act5cGAL4/Cyo Zyg Axin 1.9 0.3 n.s. 6

UAS-RFP/MatGAl4; UAS-Axin:GFP/MatGAL4 x Mat RFP & Axin 1.5 0.3 n.s. 12 UAS-RFP; UAS-Axin:GFP +/MatGAL4; UAS-Axin:GFP/MatGAL4 x Mat Axin 1.3 0.4 * 22 UAS-Axin:GFP UAS-GFP:APC2/MatGAL4; +/MatGAL4 x Mat APC2 2.2 0.2 n.s. 10 UAS-GFP:APC2

+ Significance was calculated using one-way ANOVA; n.s. = non-significant; * = p<.05; **** = p<.0001

Table 2.4: Rows of En-expressing cells per segment Quantification of the number of rows of En expressing cells per segment, in embryos in which APC2 or Axin levels are decreased or elevated. n = number of embryos scored. Significance was determined using one-way ANOVA with GraphPad.

117

Table 2.5

Table S5: Effects on Arm levels of elevating Axin and/or APC2 levels th th Abbrev. Genotype Category Mean Std. Dev. Minimum 25 Median 75 Maximum n= Percentile Percentile Wildtype Wg Stripe 189502 57641 115308 144436 165122 241961 278825 10 (act5cGAL4/Cyo) Interstripe 121351 37219 71619 96433 108979 151905 189887 10 Mat Axin Wg Stripe 128196 40878 79708 94402 122544 146699 228161 11 Interstripe 110603 23538 78025 90643 103552 131697 143606 11

Wildtype Wg Stripe 64909 12683 46595 56127 63731 72185 95562 21 (act5cGAL4/Cyo) Interstripe 48737 11239 34593 38754 47431 53488 80570 21 Mat RFP & Axin Wg Stripe 59219 15569 31098 46050 64029 72824 80381 13 Lower GFP Interstripe 45967 12967 26661 36118 46419 59746 64117 13 Mat RFP & Axin Wg Stripe 39758 7491 30852 33473 38051 46225 54182 10 Higher GFP Interstripe 36796 5180 27238 33647 36800 42773 43351 10

Wildtype Wg Stripe 239631 57740 138431 209218 229983 280509 333401 11 (act5cGAL4/Cyo) Interstripe 159690 49143 91128 126440 150837 180870 247271 11 Mat APC2 Wg Stripe 318561 58230 253471 265841 309908 372713 435928 11 Interstripe 187333 33560 145382 159956 181812 201144 265900 11

Wildtype Wg Stripe 204637 45223 79176 173770 207570 234588 288438 36 (act5cGAL4/Cyo) Interstripe 135451 32366 58278 112824 129429 152718 220457 36 Mat APC2 & Axin Wg Stripe 314265 47598 235343 278999 310783 346017 390044 11 APC2+; Axin+ Interstripe 163378 24203 119670 140780 167185 180692 198515 11 Mat APC2 & Axin Wg Stripe 124596 12761 105843 112659 128190 133803 143265 8 APC2+; Axin++ Interstripe 98061 14441 76147 87797 97210 111791 118670 8 Mat APC2 & Axin Wg Stripe 280121 104100 159805 175731 304371 366095 443029 9 APC2++; Axin+ Interstripe 154319 55917 76548 93721 152046 203967 226113 9 Mat APC2 & Axin Wg Stripe 143706 32493 93466 114848 155752 164689 197860 12 APC2++; Axin++ Interstripe 121121 25701 81880 94435 125967 143678 154060 12

Table 2.5 Effects on Arm levels of elevating Axin and/or APC2 levels. Arm accumulation levels in Wg-expressing stripes versus Arm levels in the interstripes for the indicated genotypes as in Fig 2.6B. These are the raw data used to create the box-and-whisker plots in the Figures. n = number of embryos examined.

118

Table 2.6

Table S6: Quantification of the difference in Arm levels in Wg-stripe versus interstripe-cells 25th 75th Abbrev. Genotype Mean Std. Dev. Minimum Median Maximum n= Percentile Percentile Wildtype 68151 22105 43689 48003 62445 89313 103911 10 (act5cGAL4/Cyo)

Mat Axin 17593 23556 -780.8 1683 12259 20734 84555 11

Wildtype 16172 3778 8073 12384 16355 19396 22201 21 (act5cGAL4/Cyo) Mat RFP & Axin 13252 5664 4103 9550 13012 18619 21253 13 Lower GFP Mat RFP & Axin 2962 3791 -2263 -117.5 3331 4674 10831 10 Higher GFP

Wildtype 79941 15078 47303 71183 82541 94366 99639 11 (act5cGAL4/Cyo)

Mat APC2 131228 34523 71658 112477 121318 170029 179568 11

Wildtype 69186 21282 20899 54668 68992 82661 111353 36 (act5cGAL4/Cyo) Mat APC2 & Axin 150886 40165 63303 125452 156161 187565 204232 11 APC2+; Axin+ Mat APC2 & Axin 26535 7269 16321 19646 26388 32595 37605 8 APC2+; Axin++ Mat APC2 & Axin 125801 55865 60347 81289 113078 176501 216917 9 APC2++; Axin+ Mat APC2 & Axin 22585 10484 9084 16869 20052 28007 49430 12 APC2++; Axin++

Table 2.6: Quantification of the differences in Arm levels in Wg-stripe versus interstripe cells. Quantification of difference in Arm accumulation between the Wg stripes and interstripes within individual embryos. These are the raw data behind the scatter plots in the Figures. n = number of embryos examined.

119

Table 2.7

S7 Table: Quantification of the different pools of Arm levels in Wg-stripe versus interstripe-cells Mean # of segments Abbrev. Genotype Arm Pool Std. Dev. Minimum Maximum (Intensity/Area) analyzed Total Arm 1216 143 1005 1367 9 Wildtype Cytoplasmic/Nuclear 1088 129 894 1225 9 Wg Stripe Membrane Arm 1420 141 1221 1575 9 Total Arm 760 81 635 871 9 Wildtype Cytoplasmic/Nuclear 653 57 572 752 9 Interstripe Membrane Arm 947 93 795 1046 9

Total Arm 732 74 589 827 9 Mat Axin Cytoplasmic/Nuclear 613 73 465 700 9 Wg Stripe Membrane Arm 894 68 759 978 9 Total Arm 667 64 551 765 9 Mat Axin Cytoplasmic/Nuclear 565 52 465 654 9 Interstripe Membrane Arm 811 77 720 944 9

Total Arm 1493 125 1283 1677 9 Mat APC2 Cytoplasmic/Nuclear 1375 128 1234 1568 9 Wg Stripe Membrane Arm 1774 160 1552 2044 9 Total Arm 831 124 675 1031 9 Mat APC2 Cytoplasmic/Nuclear 732 100 611 900 9 Interstripe Membrane Arm 1041 166 807 1256 9

Table 2.7 Quantification of the different pools of Arm levels in Wg-stripe versus interstripe-cells. Quantification of difference in Arm accumulation between the Wg stripes and interstripes within individual embryos. These are the raw data behind the scatter plots in the Fig 2.4. n = number of segments analyzed.

120

Table 2.8 Table S7: Fluorescence comparison values Mean # Name Organism Range Std. Dev. n= GFP molc. Human Colon Cancer cells Axin:GFP Only 728 163-1327 333 38 (SW480) Human Colon Cancer cells Axin:GFP + RFP:APC2 801 104-2041 457 39 (SW480) Human Colon Cancer cells GFP:APC2 + Axin:RFP 917 162-3297 608 55 (SW480)

Mat RFP & Axin:GFP Drosophila Embryo 195 46-931 152 70

Mat RFP & Axin:GFP Drosophila Embryo 128 46-305 61 35 Wg ON Mat RFP & Axin:GFP Drosophila Embryo 262 56-931 185 35 Wg OFF

Table 2.8 Fluorescence comparison values. Detailed results from calculating the number of GFP-tagged Axin or APC2 proteins in Axin puncta in live human SW480 colon cancer cells or in Drosophila embryo. These data form the basis of Fig 2.16. n = number of puncta examined.

121

REFERENCES

Ahmed, Y., Hayashi, S., Levine, A., and Wieschaus, E. (1998). Regulation of armadillo by a Drosophila APC inhibits neuronal apoptosis during retinal development. Cell 93, 1171- 1182.

Ahmed, Y., Nouri, A., and Wieschaus, E. (2002). Drosophila Apc1 and Apc2 regulate Wingless transduction throughout development. Development (Cambridge, England) 129, 1751- 1762.

Akong, K., Grevengoed, E.E., Price, M.H., McCartney, B.M., Hayden, M.A., DeNofrio, J.C., and Peifer, M. (2002a). Drosophila APC2 and APC1 play overlapping roles in wingless signaling in the embryo and imaginal discs. Dev Biol 250, 91-100.

Akong, K., McCartney, B.M., and Peifer, M. (2002b). Drosophila APC2 and APC1 have overlapping roles in the larval brain despite their distinct intracellular localizations. Dev Biol 250, 71-90.

Aoki, K., and Taketo, M.M. (2007). Adenomatous polyposis coli (APC): a multi-functional tumor suppressor gene. J Cell Sci 120, 3327-3335.

Axelrod, J.D. (2001). Unipolar membrane association of Dishevelled mediates Frizzled planar cell polarity signaling. Genes & development 15, 1182-1187.

Bejsovec, A., and Martinez Arias, A. (1991). Roles of wingless in patterning the larval epidermis of Drosophila. Development (Cambridge, England) 113, 471-485.

Brand, A.H., and Perrimon, N. (1993). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development (Cambridge, England) 118, 401-415.

Clevers, H., and Nusse, R. (2012). Wnt/beta-catenin signaling and disease. Cell 149, 1192- 1205.

Cliffe, A., Hamada, F., and Bienz, M. (2003). A role of Dishevelled in relocating Axin to the plasma membrane during wingless signaling. Current biology : CB 13, 960-966.

122

Feng, Y., Li, X., Ray, L., Song, H., Qu, J., Lin, S., and Lin, X. (2014). The Drosophila tankyrase regulates Wg signaling depending on the concentration of Daxin. Cellular signalling 26, 1717-1724.

Fiedler, M., Mendoza-Topaz, C., Rutherford, T.J., Mieszczanek, J., and Bienz, M. (2011). Dishevelled interacts with the DIX domain polymerization interface of Axin to interfere with its function in down-regulating β-catenin. Proceedings of the National Academy of Sciences 108, 1937-1942.

Fox, D.T., and Peifer, M. (2007). Abelson kinase (Abl) and RhoGEF2 regulate actin organization during cell constriction in Drosophila. Development (Cambridge, England) 134, 567-578.

Gammons, M., and Bienz, M. (2018). Multiprotein complexes governing Wnt signal transduction. Current opinion in cell biology 51, 42-49.

Gammons, M.V., Renko, M., Johnson, C.M., Rutherford, T.J., and Bienz, M. (2016a). Wnt Signalosome Assembly by DEP Domain Swapping of Dishevelled. Mol Cell 64, 92-104.

Gammons, M.V., Rutherford, T.J., Steinhart, Z., Angers, S., and Bienz, M. (2016b). Essential role of the Dishevelled DEP domain in a Wnt-dependent human-cell-based complementation assay. J Cell Sci 129, 3892-3902.

Hays, R., Gibori, G.B., and Bejsovec, A. (1997). Wingless signaling generates pattern through two distinct mechanisms. Development (Cambridge, England) 124, 3727-3736.

Hernandez, A.R., Klein, A.M., and Kirschner, M.W. (2012). Kinetic responses of beta-catenin specify the sites of Wnt control. Science (New York, NY) 338, 1337-1340.

123

Lawrimore, J., Bloom, K.S., and Salmon, E.D. (2011). Point centromeres contain more than a single centromere-specific Cse4 (CENP-A) nucleosome. The Journal of cell biology 195, 573-582.

Lee, E., Salic, A., and Kirschner, M.W. (2001). Physiological regulation of [beta]-catenin stability by Tcf3 and CK1epsilon. The Journal of cell biology 154, 983-993.

Lee, E., Salic, A., Kruger, R., Heinrich, R., and Kirschner, M.W. (2003). The roles of APC and Axin derived from experimental and theoretical analysis of the Wnt pathway. PLoS biology 1, E10.

Li, V.S., Ng, S.S., Boersema, P.J., Low, T.Y., Karthaus, W.R., Gerlach, J.P., Mohammed, S., Heck, A.J., Maurice, M.M., Mahmoudi, T., et al. (2012). Wnt signaling through inhibition of beta-catenin degradation in an intact Axin1 complex. Cell 149, 1245-1256.

MacDonald, B.T., and He, X. (2012). Frizzled and LRP5/6 receptors for Wnt/beta-catenin signaling. Cold Spring Harbor perspectives in biology 4.

Mao, J., Wang, J., Liu, B., Pan, W., Farr, G.H., Flynn, C., Yuan, H., Takada, S., Kimelman, D., Li, L., et al. (2001). Low-Density Lipoprotein Receptor-Related Protein-5 Binds to Axin and Regulates the Canonical Wnt Signaling Pathway. Molecular Cell 7, 801-809.

McCartney, B.M., Dierick, H.A., Kirkpatrick, C., Moline, M.M., Baas, A., Peifer, M., and Bejsovec, A. (1999). Drosophila APC2 is a cytoskeletally-associated protein that regulates wingless signaling in the embryonic epidermis. The Journal of cell biology 146, 1303-1318.

McKay, D.J., and Lieb, J.D. (2013). A common set of DNA regulatory elements shapes Drosophila appendages. Developmental cell 27, 306-318.

Mendoza-Topaz, C., Mieszczanek, J., and Bienz, M. (2011). The Adenomatous polyposis coli tumour suppressor is essential for Axin complex assembly and function and opposes Axin's interaction with Dishevelled. Open biology 1, 110013.

Metcalfe, C., Mendoza-Topaz, C., Mieszczanek, J., and Bienz, M. (2010). Stability elements in the LRP6 cytoplasmic tail confer efficient signalling upon DIX-dependent polymerization. J Cell Sci 123, 1588-1599.

Mirams, G.R., Byrne, H.M., and King, J.R. (2010). A multiple timescale analysis of a mathematical model of the Wnt/beta-catenin signalling pathway. Journal of mathematical biology 60, 131-160.

Nakamura, T., Hamada, F., Ishidate, T., Anai, K., Kawahara, K., Toyoshima, K., and Akiyama, T. (1998). Axin, an inhibitor of the Wnt signalling pathway, interacts with beta-catenin, GSK-3beta and APC and reduces the beta-catenin level. Genes to cells : devoted to molecular & cellular mechanisms 3, 395-403.

Nusse, R., and Clevers, H. (2017). Wnt/beta-Catenin Signaling, Disease, and Emerging Therapeutic Modalities. Cell 169, 985-999.

124

Pai, L.M., Orsulic, S., Bejsovec, A., and Peifer, M. (1997). Negative regulation of Armadillo, a Wingless effector in Drosophila. Development (Cambridge, England) 124, 2255-2266.

Peifer, M. (1993). The product of the Drosophila segment polarity gene armadillo is part of a multi-protein complex resembling the vertebrate adherens junction. J Cell Sci 105 ( Pt 4), 993-1000.

Penton, A., Wodarz, A., and Nusse, R. (2002). A mutational analysis of dishevelled in Drosophila defines novel domains in the dishevelled protein as well as novel suppressing alleles of axin. Genetics 161, 747-762.

Poulton, J.S., Mu, F.W., Roberts, D.M., and Peifer, M. (2013). APC2 and Axin promote mitotic fidelity by facilitating centrosome separation and cytoskeletal regulation. Development (Cambridge, England) 140, 4226-4236.

Pronobis, M.I., Deuitch, N., Posham, V., Mimori-Kiyosue, Y., Peifer, M., and Montell, D. (2016). Reconstituting regulation of the canonical Wnt pathway by engineering a minimal β- catenin destruction machine. Molecular biology of the cell 28, 41-53.

Riggleman, B., Wieschaus, E., and Schedl, P. (1989). Molecular analysis of the armadillo locus: uniformly distributed transcripts and a protein with novel internal repeats are associated with a Drosophila segment polarity gene. Genes & development 3, 96-113.

Roberts, D.M., Pronobis, M.I., Poulton, J.S., Waldmann, J.D., Stephenson, E.M., Hanna, S., and Peifer, M. (2011). Deconstructing the sscatenin destruction complex: mechanistic roles for the tumor suppressor APC in regulating Wnt signaling. Molecular biology of the cell 22, 1845-1863.

Rorth, P. (1998). Gal4 in the Drosophila female germline. Mechanisms of development 78, 113- 118.

Rubinfeld, B., Albert, I., Porfiri, E., Fiol, C., Munemitsu, S., and Polakis, P. (1996). Binding of GSK3beta to the APC-beta-catenin complex and regulation of complex assembly. Science (New York, NY) 272, 1023-1026.

125

Schwarz-Romond, T., Metcalfe, C., and Bienz, M. (2007b). Dynamic recruitment of axin by Dishevelled protein assemblies. Journal of Cell Science 120, 2402-2412.

Shimada, Y., Usui, T., Yanagawa, S., Takeichi, M., and Uemura, T. (2001). Asymmetric colocalization of Flamingo, a seven-pass transmembrane cadherin, and Dishevelled in planar cell polarization. Current biology : CB 11, 859-863.

Siegfried, E., Chou, T.B., and Perrimon, N. (1992). wingless signaling acts through zeste-white 3, the Drosophila homolog of glycogen synthase kinase-3, to regulate engrailed and establish cell fate. Cell 71, 1167-1179.

Sokol, S.Y., Klingensmith, J., Perrimon, N., and Itoh, K. (1995). Dorsalizing and neuralizing properties of Xdsh, a maternally expressed Xenopus homolog of dishevelled. Development (Cambridge, England) 121, 1637-1647.

Stamos, J.L., Chu, M.L., Enos, M.D., Shah, N., and Weis, W.I. (2014). Structural basis of GSK-3 inhibition by N-terminal phosphorylation and by the Wnt receptor LRP6. eLife 3, e01998.

Tan, C.W., Gardiner, B.S., Hirokawa, Y., Layton, M.J., Smith, D.W., and Burgess, A.W. (2012). Wnt signalling pathway parameters for mammalian cells. PloS one 7, e31882.

Toretsky, J.A., and Wright, P.E. (2014). Assemblages: Functional units formed by cellular phase separation. The Journal of cell biology 206, 579-588.

Verdaasdonk, J.S., Lawrimore, J., and Bloom, K. (2014). Determining absolute protein numbers by quantitative fluorescence microscopy. Methods in cell biology 123, 347-365.

Wang, Z., Tacchelly-Benites, O., Yang, E., and Ahmed, Y. (2016a). Dual Roles for Membrane Association of Drosophila Axin in Wnt Signaling. PLoS genetics 12, e1006494.

Willert, K., Logan, C.Y., Arora, A., Fish, M., and Nusse, R. (1999a). A Drosophila Axin homolog, Daxin, inhibits Wnt signaling. Development (Cambridge, England) 126, 4165-4173.

126

Willert, K., Shibamoto, S., and Nusse, R. (1999b). Wnt-induced dephosphorylation of axin releases beta-catenin from the axin complex. Genes & development 13, 1768-1773.

Xing, Y., Clements, W.K., Le Trong, I., Hinds, T.R., Stenkamp, R., Kimelman, D., and Xu, W. (2004). Crystal Structure of a β-Catenin/APC Complex Reveals a Critical Role for APC Phosphorylation in APC Function. Molecular Cell 15, 523-533.

Yamulla, R.J., Kane, E.G., Moody, A.E., Politi, K.A., Lock, N.E., Foley, A.V.A., and Roberts, D.M. (2014). Testing Models of the APC Tumor Suppressor/β-Catenin Interaction Reshapes Our View of the Destruction Complex in Wnt Signaling. Genetics 197, 1285- 1302.

Yanagawa, S., van Leeuwen, F., Wodarz, A., Klingensmith, J., and Nusse, R. (1995). The dishevelled protein is modified by wingless signaling in Drosophila. Genes & development 9, 1087-1097.

Yost, C., Torres, M., Miller, J.R., Huang, E., Kimelman, D., and Moon, R.T. (1996). The axis- inducing activity, stability, and subcellular distribution of beta-catenin is regulated in Xenopus embryos by glycogen synthase kinase 3. Genes & development 10, 1443- 1454.

Yu, X., Waltzer, L., and Bienz, M. (1999). A new Drosophila APC homologue associated with adhesive zones of epithelial cells. Nature cell biology 1, 144-151.

Zacharias, D.A., Violin, J.D., Newton, A.C., and Tsien, R.Y. (2002). Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells. Science (New York, NY) 296, 913-916.

127

CHAPTER 3: THE DANCE BETWEEN THE DESTRUCTION COMPLEX AND THE E3 LIGASE

OVERVIEW:

Wnt signaling is an essential cell-cell signaling pathway which regulates development and is often perturbed in cancer. Wnt signaling regulated the levels of beta-catenin, a co- activator of transcription. In the absence of Wnt signaling, the destruction complex comprised of the tumor suppressor APC, the scaffold protein Axin, and the kinases CK1 and GSK3 recruit beta-catenin and phosphorylate it and thus labeling beta-catenin for proteasomal degradation.

Once beta-catenin is phosphorylated, it can now be recognized by the Skp-Cullin-F-box containing E3 Ligase. Even though it has been known for over 20 years that beta-TrCP is the F- box protein that recognizes phosphorylated beta-catenin, the mechanism of beta-catenin transfer from the destruction complex to the E3 ligase is unknown. We utilized cell culture to visualize that Slimb, the fly homolog of beta-TrCP, is recruited to the destruction complex by

Axin. Co-immunoprecipitation and co-localization assays in which we sequentially deleted different domains of Axin reveal that the RGS domain of Axin is necessary for Axin and Slimb to interact. Our data suggests that Slimb shuttles phosphorylated beta-catenin from the destruction complex to the E3 ligase to be ubiquitinated and then passed to the proteasome for degradation.

INTRODUCTION:

During development, Wnt signaling is essential for setting up the body plan and determining cell fate in all animals. Either loss of or mis-regulation of this pathway leads to embryonic lethality. Wnt signaling is also essential for tissue homeostasis. For example, Wnt

128

signaling maintains the stem cells in the crypts of the colon, which divide to replace cells lost in

the villi. Over 90% of colon cancers contain a mutation inappropriately activating Wnt signaling.

The protein most often altered in colon cancer is the tumor suppressor Adenomatous polyposis

coli (APC). In fact, more than 80% of all colon cancers express only a truncated form of APC.

APC is a core component of the destruction complex (DC), a key negative regulator of the Wnt pathway. Βeta-catenin (βcat), main effector of Wnt signaling, forms a complex with

TCF/LEF family transcription factors to co-activate of transcription of Wnt target genes. The core components of the DC are APC, the scaffold protein Axin, and the kinases CK1 and GSK3.

These proteins form large supermolecular-complexes containing 10s-100s of Axin molecules

(Schaefer et al., 2018), in which APC and Axin work together to recruit βcat and target it for

destruction. To regulate levels of βcat it must be post-translationally modified, via addition of

ubiquitin. Once βcat has been poly-ubiquitinated it can now be recognized by the proteasome

for protein degradation. The process of βcat destruction occurs in four steps: 1-Recruitment, 2-

Phosphorylation, 3-E3-ligase ubiquitination, and 4-Degradation by the proteasome.

1. Recruitment

The βcat:DC interaction is mediated through APC and/or Axin. Both contain βcat binding

motifs that bind to the Armadillo (Arm) repeats of βcat (Xu and Kimelman, 2007). In Drosophila,

APC2 is the main APC family member that is expressed during embryonic development. APC2 contains 8 βcat binding motifs: three 15 amino acid repeats (15R), and five 20 amino acid repeats (20R). Despite having multiple βcat binding domains, a single APC2 molecule’s ability to recruit βcat is quite low (Spink et al., 2001; Xing et al., 2004), however phosphorylation of the

20Rs increases APC2´s affinity for βcat (Xing et al., 2004). Interestingly studies in cancer cell lines and in developing Drosophila embryo suggest many of APC2’s βcat binding motifs are dispensable for destruction complex function (Pronobis et al., 2016; Yamulla et al., 2014).

Axin contains a single βcat binding motif which is essential for DC function (Pronobis et al., 2016). However Axin’s affinity for βcat is also low (Salic et al., 2000). So how does the DC 129

efficiently recruit βcat? As noted above, the DC assembles into a large supra-molecular complex containing 10s to 100s of molecules of Axin (Schaefer et al., 2018). So perhaps by concentrating APC2 and Axin molecules into spots or puncta, the DC’s avidity for βcat is increased. Axin, can self-polymerize via is C-terminal DIX domain, which can polymerize in a

head to tail fashion and forming the base scaffold of the DC (Fiedler et al., 2011; Schwarz-

Romond et al., 2007a; Schwarz-Romond et al., 2007b). If Axin loses this ability to self- polymerize, then the DC is no longer to target βcat for destruction (Fiedler et al., 2011). Thus, in order for βcat to be recruited to the DC, Axin and APC2 need to form higher-order oligomers to increase their avidity for βcat.

2.Phosphorylation

The βcat binding pocket in the E3 ligase can only recognize a phosphorylated βcat (Wu et al., 2003). Phosphorylation of βcat occurs in 2 steps, via 2 different kinases, CK1 and GSK3.

These kinases are recruited to the DC via interaction with Axin (Dajani et al., 2003; Liu et al.,

2002). The first phosphorylation step is carried out by CK1α. The N-terminal end of βcat, before

the start of its Arm repeat domains, carries a series of conserved serine and threonine residues

spaced four amino acids apart. CK1 phosphorylates βcat at serine 45 (Liu et al., 2002). This phosphorylation primes βcat to now be bound and then phosphorylated by GSK3β at serine 33,

37, and threonine 41 (Liu et al., 2002). This series of sequential phosphorylation events are necessary for βcat to now be recognized by the E3 ligase.

3. Ubiquitination

Many substrates are post-translationally modified by attachment of the small protein

ubiquitin (Ub). Ubiquitination regulates several cell processes, including degradation by the

proteasome, cell cycle progression, regulation of transcription, DNA repair, and signal

transduction (Berndsen and Wolberger, 2014). Ubiquitination of βcat labels it for recognition and

destruction by the proteasome, thus preventing βcat from activating transcription of downstream

130

Wnt target genes. Ubiquitination requires a coordinated dance between 3 different enzymes,

E1, E2, and E3. The E1 catalyzes the activation of Ub in an ATP-dependent reaction (Berndsen and Wolberger, 2014). Ub is then transferred to an E2 enzyme. The E3 ligase then interacts with both the substrate and the E2-ubiquitin to promote attachment of Ub to the substrate

(Berndsen and Wolberger, 2014). Humans only have a few E1s, approximately 40 E2s, and more than 500 E3 ligases (Wenzel et al., 2011). The E3 responsible for βcat ubiquitination is

SCFβTrCP. This complex is comprised of: Cullin1 (Cul1), Skp1, the F-box protein βTrCP, and

Ring box (RBX) subunits, which work together to bind to phosphorylated βcat and attach multiple Ub subunits. Cul1 is the scaffold of the complex, at one end bringing together the

Rbx/E2-Ub proteins and at the other end binds Skp1. The small protein Skp1(SkpA in

Drosophila) is the linker between Cul1 and βTrCP. The F-box protein βTrCP (Slimb in

Drosophila) contains the substrate recognition domain of the E3 ligase. The βcat recognition site spans WD40 repeats on the c-terminal end of βTrCP (Wu et al., 2003). This domain forms a propeller structure with a pocket that binds only to phosphorylated proteins. βTrCP can to bind to several phospho-proteins and thus regulate diverse cell signaling pathways (ex: NFκB and

Hedgehog (Fuchs et al., 2004; Jiang and Struhl, 1998)) however we will focus on its role in regulating βcat levels. After βTrCP-βcat binding βcat is poly-ubiquitinated and can now be recognized by the proteasome.

4.Degradation

Protein degradation in Eukaryotes cells can occurs via 2 major ways: lysosomal or autosomal proteolysis or the ubiquitin proteasome system. The ubiquitin proteasome system is the major protein degradation pathway utilized in Eukaryotes (Kisselev et al., 2000). The proteasome is a large multi-protein complex composed of 2 subcomplexes: the 20S catalytic core and 1-2 19S regulatory particle(s). These subcomplexes get their name from the size of the proteins after run on a separation column. The 20S core complex is barrel shaped and is inactive until bound by the 19S complex at one or both ends (Tanaka, 2009). The 19S cap(s) 131 recognizes poly-ubiquitinated proteins and is believed to aid in transferring the substrate into the catalytic core. Within the barrel shaped 20S protein hydrolysis occurs to create 3-15 aa long peptides. This process is ATP-dependent (Tanaka, 2009). Thus, in the case of Wnt signaling, the 19S subcomplex of the proteasome recognizes a poly-ubiquitinated βcat and passes it to the 20S subcomplex to degrade βcat, keeping cytoplasmic levels of βcat low.

While regulation of βcat levels via protein degradation is a key function of Wnt signaling our understanding of how βcat is transferred from the DC to the E3 is minimal. Based on recent research from our lab we propose that the DC is a phase separated machine (see Ch1 for more detail). The E3 ligase is a well-studied complex, yet the number of molecules that form an active complex is unknown. If the E3 is also a supra-molecular machine, similar to the DC, we hypothesize 2 different mechanisms for βcat transfer to the E3. 1) Since the E3 is able to bind and Ub diverse phospho-proteins, it might make sense for the E3 to form a complex separate from the DC. In this scenario, the DC and the E3 form two separate complexes, requiring βcat to be shuttled between the complexes. 2) The E3 is a part of the phase separated DC. In this model once βcat is phosphorylated it is directly transferred to the E3. This would prevent dephosphorylation of βcat by cellular phosphatases during transit. Immuno-precipitation (IP) experiments in animals and in cell culture suggest that at least βTrCP interacts with Axin (Hart et al., 1999; Kitagawa et al., 1999; Li et al., 2012; Liu et al., 1999). However, none of these studies looked colocalization between Axin and βTrCP, or other components of the E3, leaving both models an option, especially if βTrCP was the shuttling protein.

We thus set out to examine these two models, using immuno-fluorescence, IPs, and super-resolution microscopy. Our data support the first mechanism. We found that the

Drosophila βTrcP homolog Slimb is recruited into Axin puncta, independent of Axin’s ability to interact with βcat or APC2. However, we do not detect other components of the SCF complex.

Our data also suggest that Slimb either directly or indirectly interacts with Axin’s RGS domain, a

132 known APC2 direct interaction domain, perhaps suggesting competitive interaction between

APC2-Axin-Slimb.

RESULTS

A system to examine whether the destruction complex and the E3 ligase co-localize

A key step in Wnt signaling is the transfer of phosphorylated βcat from the DC to the E3 ligase complex to begin βcat degradation. One key question in the field involves the mechanism by which βcat is transferred from one complex to another. Do these complexes form separate structures within the cytoplasm of the cells, thus requiring some form of protein shuttle to move

βcat? Or is the destruction complex more like a factory for βcat destruction, containing the machinery to first phosphorylate βcat, and then directly pass it down the assembly line to the E3 ligase? Previous work is more consistent with the latter model, as Axin can co- immunoprecipitate (coIP) with mammalian βTrcP (Hart et al., 1999; Li et al., 2012) and one role of APC is to protect βcat from dephosphorylation before it is ubiquitinated (Su et al., 2008). To further address this issue, we utilized a colon cancer cell line (SW480) to transfect components of the destruction complex and the E3 ligase to visualize whether they co-localize.

To visualize the destruction complex, we tagged both Drosophila Axin and APC2 with

GFP, RFP, or Flg epitope tags (Pronobis et al., 2015; Roberts et al., 2011). For our studies we utilized the Drosophila proteins, which can rescue βcat destruction in this colorectal cell line

(Pronobis et al., 2015; Roberts et al., 2011). Drosophila APC2 is also half the size of human

APC1 and therefore easier to transfect and express in cells (Kreiss et al., 1999). When GFP- tagged APC2 (GFP:APC2) is transfected in cells alone, APC2 is found throughout the cytoplasm (Fig 3.1A). When Axin with an RFP tag is transfected alone (Axin:RFP) it forms cytoplasmic puncta, due to Axin’s ability to self-polymerize via its DIX domain (Fig 3.1B)

(Kishida et al., 1999). When GFP:APC2 is expressed along with Axin:RFP, GFP:APC2 is recruited into Axin puncta (Fig 3.1C). Previous studies have shown that this APC2-Axin

133 interaction leads to larger stabilized destruction complexes (Kunttas-Tatli et al., 2014; Pronobis et al., 2015).

We next examined whether the E3 ligase had any specific localization pattern on its own. To accomplish this, we tagged Drosophila Cul1, SkpA, and Slimb with either GFP, RFP or

Flg tags. When each E3 component was expressed alone or when they were expressed in combination, each was diffusely localized in both the cytoplasm and nucleus, without obvious enrichment in any subcellular structure (Fig 3.1D, Fig 3.2A-B and data not shown). In a few cells, there was slight enrichment of proteins in puncta near the nucleus, which may be due to the E3’s known role in regulating centrosome duplication (Wojcik et al., 2000).

Axin and not APC2 can recruit Slimb into the destruction complex

Previous studies have shown that Axin and βTrcP co-IP with one another (Hart et al.,

1999; Kitagawa et al., 1999; Li et al., 2012; Liu et al., 1999). To test Axin and/or APC were responsible for this co-recruitment we expressed co-expressed Flg:Axin and GFP:APC2 with a

RFP-tagged Slimb. We found that the βTrcP homolog Slimb was robustly recruited to Axin/APC puncta (Fig 3.3A). However, this did not isolate whether Axin or APC recruited Slimb.

When APC was first discovered, it was believed to be the scaffold of the DC, as it can bind βcat and coIP with the kinase GSK3 (Polakis, 1997). However, subsequent work revealed that Axin is the actual scaffold of the DC, mediating complex assembly by directly binding all of the core destruction complex components: APC, GSK3, CK1, and βcat (Dajani et al., 2003; Liu et al.,

2002; Spink et al., 2000). To define which whether APC or Axin and recruit Slimb, we expressed either co expressed Axin and Slimb or APC2 and Slimb. The ability of Axin to form puncta made examining Slimb recruitment straightforward. Co-expression of an RFP-tagged Axin (Axin:RFP) with a GFP-tagged Slimb (Slimb:GFP) revealed that Axin can recruit Slimb:GFP into cytoplasmic puncta (Fig 3.1E-F). Interestingly, we observed that robust Slimb:GFP recruitment into Axin:RFP puncta was depentdent on the amount of Axin:RFP protein in the cell Fig 3.1E-F).

134

Since APC2 has no specific localization pattern when expressed on its own, it is difficult

to tell whether APC2 is able to recruit other proteins. We therefore utilized an APC2 construct

containing a mitochondrial localization signal (mito:APC2,(Roberts et al., 2012)). As shown

previously, even though mito:APC2 is recruited to the mitochondria, it remains functional, as

evidenced by reduction of βcat levels inSW480 cells and the ability to rescue Drosophila APC2

mutants (Fig 3.3B (Roberts et al., 2012)). Mito:APC2 can still recruit Axin (Fig 3.3C, (Roberts et

al., 2012)). We therefore expressed mito:APC2 with Slimb:GFP to test whether APC2 was

capable of E3 recruitment. GFP-tagged mito:APC2 was expressed with RFP- tagged Slimb and

colocalization between these proteins was observed. Mito:APC2 was unable to recruit any

Slimb (Fig 3.3D), suggesting APC2 is does not directly interact with the E3.

We next wanted to begin to test they hypothesis of whether the whole E3 is recruited by Axin, or

just Slimb. To test this we co-expressed Axin:RFP with either GPP-tagged Cul1 (GFP:Cul1) or

GFP-tagged SkpA (GFP:SkpA). When expressed alone, both SkpA and Cul1 were found

throughout the cytoplasm and nucleus (Fig 3.2A-B). We were surprised to find that while Slimb

was robustly recruited to the Axin puncta, Cul1 and SkpA were not (Fig 3.2C-D).

Consistent with previous work with the mammalian homologs, we could coIP Axin and Slimb,

but did not detect co-IP of with either Cul1 or SkpA (Fig 3.2E-F). These data suggest that Axin is

able to recruit Slimb to the DC and not the other components of the E3.

The RGS domain of Axin is required for efficient Slimb recruitment

Both Axin and Slimb directly bind to βcat, but at different locations on βcat (Xu and

Kimelman, 2007). Therefore previous studies have suggested that the Axin:Slimb interaction might not be direct, but instead a result of bridging by βcat (Liu et al., 1999). APC2 is also able to directly bind to βcat, in a similar location as Axin (Xu and Kimelman, 2007). If Axin is only recruiting Slimb via a βcat linker, then APC2 should also be able to recruit Slimb, something not supported by our data (Fig 3.2F, Fig 3.3D). To further test the hypothesis that Axin recruits

135

Figure 3.1

Figure 3.1: Axin recruits Slimb into cytoplasmic puncta. A-F) SW480 cells transfected with the labelled constructs and stained for βcat (except C). A) RFP-tagged wild type Drosophila Axin self-polymerizes into cytoplasmic puncta and reduces βcat levels. B) When expressed in cells, GFP:APC2 is diffuse throughout the cell and dramatically reduces βcat levels. C) When co-expressed, GFP:APC2 is recruited into Axin:RFP puncta. D) GFP tagged Slimb is expressed through out the cytoplasm and nucleus. Slimb expressed alone is unable to significantly reduce βcat levels. E-F) Co-expression of Axin:RFP and Slimb:GFP. Axin can recruit Slimb:GFP into cytoplasmic puncta and robust recruitment appears to be dependent on Axin concentration. Scale bar = 10 μm.

136

Slimb via a βcat linker, we removed the βcat binding site from Axin (AxinΔβcat:RFP) (Fig 3.4A)

and co-expressed it with Slimb. If βcat was acting as a bridge between Slimb and Axin, then we

would expect to no longer see Slimb recruitment into Axin puncta. In contrast, if Axin and Slimb

interact by another means, then Slimb should still be recruited into the puncta. When

AxinΔβcat:RFP was expressed alone, it was still able to form cytoplasmic puncta since it still

contains its self-polymerization domain (Fig 3.4A & C). When co-expressed with Slimb:GFP,

AxinΔβcat:RFP was still able to recruit Slimb:GFP into puncta (Fig 3.4F), suggesting that the

Slimb-Axin interaction is not solely a result of both proteins binding to βcat. It is interesting to note that cells expressing AxinΔβcat:RFP were unable to decrease βcat levels (Fig 3.4F’’’), however we think this is due to the essential role of Axin’s βcat binding motif in DC function

(Pronobis et al., 2016).

To further investigate which domain of Axin was required for Slimb recruitment, we generated mutants of Axin deleting different domains or regions (Fig 3.4A). Each RFP-tagged Axin mutant

was co-expressed with Flg-tagged Slimb. We then pulled down for RFP (Axin mutant) and

assessed if Slimb was co-IPed. We created four different deletion mutants: 1) AxinΔRGS:RFP-

removed the RGS domain - which allows Axin and APC to directly interact (Spink et al., 2000).

2) AxinΔβcat:RFP, which deleted Axin’s βcat binding motif. 3) AxinΔDIX:RFP - this removed

Axin’s polymerization domain and last 4) StAftRGS:RFP - (Start After RGS) this constructe

deleted the first third of Axin (Fig 3.4A) We were surprised to see that the two Axin mutants

lacking the RGS domain were unable to pull down Flg:Slimb (Fig 3.4B). These data suggest that

the RGS domain is necessary for Axin-Slimb interaction.

To verify that AxinΔRGS:RFP and SrtAftRGS:RFP were unable to recruit Slimb into

puncta, we co-expressed these mutants with GFP:Slimb in our colon cancer cells. Both of these

constructs formed normal looking puncta (Fig 3.4D-E). Neither Axin mutant was able to robustly

137

Figure 3.2

Figure 3.2: Axin is unable to recruit Cul1 or SkpA into the complex. A-D) SW480 cells transfected with the indicated Drosophila proteins. A-B) GFP tagged SkpA and Cul 1 are expressed throughout the cytoplasm and nucleus and has little effect on βcat destruction. C-D) Co-transfection of Axin:RFP with either GFP:SkpA or Cul 1 (as marked). Axin is unable to robustly recruit SkpA or Cul1. E) Co-immunoprecipitation of Flg-tagged Slimb for for GFP tagged APC2 or Axin. Slimb was able to pull down Axin and not APC2. F) Co- immunoprecipitation of Flg tagged Axin for each component of the E3 ligase. Flg:Axin was only able to recruit Slimb and not SkpA or Cul1. All blots are from the same gel, with excess lanes removed. Scale bar = 10 μm.

138 recruit GFP:Slimb (Fig 3.4G-H). These data suggest the Axin and Slimb interact through Axin’s

RGS domain.

Slimb is a dynamic component of the destruction complex

The destruction complex has many of the properties of a biomolecular condensate. One of these is the ability of individual components to rapidly exchange with the cytoplasmic pool. This property can be measured using fluorescence recovery after photobleaching (FRAP), in which fluorescently-tagged protein components of a protein complex are photobleached, and exchange with the unbleached cytoplasmic pool is assessed. FRAP analysis provides the amount of protein turnover in the form of fluorescence recovery percentage plateau (recovery plateau). For example, if there were 100 GFP tagged proteins in a punctum, and we observed

30% recovery of the total GFP fluorescence, this would indicate that 70 proteins remained in the complex and 30 new proteins entered. It also provides an assessment of the Time ½ (T1/2), the amount of time necessary to replace half of the total recovered fluorescence. This measure provides turnover rate. Previous analysis revealed that when expressed alone, Axin:RFP has a relatively quick turnover and is able to replace almost all Axin molecules after FRAP (Pronobis et al., 2015). However, when Axin is co- expressed with APC2, half as much Axin moves out of the complex and FRAP recovery takes twice as long (Pronobis et al., 2015)). When we repeated these experiments, we had similar results (Fig 3.5A-B). These data suggest that Axin is stabilized by APC2.

To gain an understanding of Slimb dynamics in the DC, we co-expressed Axin:RFP and

GFP:Slimb. We hypothesized that if Slimb was stable in the complex, then its T1/2 would be relative slow and its recovery plateau low. However, if Slimb shuttles βcat between complexes, then the turnover rate would be faster and recovery higher. When GFP:Slimb was co-expressed with Axin:RFP, Slimb’s recovery plateau was ~50% and it had a T1/2 of 100 seconds (Fig 3.5C-

D). This rate of Slimb turnover was very similar to Axin’s. This could suggest that both Axin and

Slimb are leaving the complex together. To test this hypothesis, we measured Slimb’s dynamics 139

Figure 3.3

Figure 3.3: Slimb is not recruited by mito-APC2. A-D) SW480 cells transfected with the marked constructs. A) Triple co-transfection of GFP:APC2, Flg:Axin, and RFP:Slimb. APC2 and Axin are able to robustly recruit Slimb into the destruction complex. B) Expression of a GFP- tagged APC2 with a mitochondrial localization signal. Even though APC2 is stuck at the membrane, it is still able to degrade βcat. C) To verify function mito:APC2, Axin:RFP was co- expressed. Mito:APC2 was still able to recruit Axin:RFP and degrade βcat. D) Mito-APC2 was expressed with RFP tagged Slimb. RFP:Slimb was not recruited by APC2. Scale bar = 10 μm.

140

when both Axin and APC2 were expressed. If Slimb and Axin were leaving the complex

together, then expression of APC2 would slow Slimb’s recover and T1/2 to that of Axin in

APC2:Axin complexes. FRAP experiments of Slimb in complex with APC2 and Axin showed no

significant changes in Slimb dynamics compared to Slimb in Axin:Slimb complexes. These data

are consistent with the idea that Slimb leaves the DC independently of Axin.

Slimb localizes along Axin cables

While standard confocal microscopy only resolves the destruction complex as co-localized dots of Axin and APC, structured illumination microscopy (SIM) allowed us to begin to discern its internal structure, revealing intertwined cables of Axin and APC2 (Pronobis et al., 2015). To further investigate the Axin:Slimb interaction, we thus sought to visualize this by SIM. When

Axin:RFP was expressed alone, puncta appeared to form tight cables (Fig 3.6A-C), and co- expression of GFP:APC2 with RFP:Axin caused an expansion in average puncta size, with

APC2 and Axin forming intertwined cables (Fig 3.6D-F, (Pronobis et al., 2015)). We next co- expressed both Axin:RFP and Slimb:GFP. The addition of Slimb appeared to have little effect on average puncta size. Interestingly, Axin and Slimb do not form intertwined cables. Instead

Slimb appears to coat segments of Axin cables (Fig 3.6G-I) These preliminary data suggest that

Axin forms a scaffold upon which Slimb binds.

DISCUSSION

Defining the DC and E3 ligase interaction

Wnt signaling is defined by the regulation of βcat levels in the cytoplasm: when levels are high, βcat enters the nucleus and works with TCF/LEF proteins to activate the transcription of Wnt target genes, inducing diverse cellular processes such as cell fate choice and proliferation (Clevers and Nusse, 2012). When βcat levels are kept low, Wnt target genes

141

Figure 3.4

Figure 3.4: The RGS domain of Axin is necessary for Slimb recruitment into Axin puncta. A) Diagram of different Axin mutant constructs created. B) Co-immunoprecipitation of RFP- tagged Axin mutants for Flg tagged Slimb. Axin constructs missing the RGS domain are unable to pull down Slimb. All blots are from the same gel, with excess lanes removed. C-E) Expression of different Axin mutants as labelled. F-H) Co-expression of Slimb:GFP with the labelled Axin mutant. Only AxinΔβcat:RFP is able to robustly recruit Slimb into puncta. Scale bar = 10 μm

142 remain inactive. In the colon, for example, low levels of βcat decrease cell proliferation. It has been known for decades that βcat sequentially interacts with the DC, E3, and proteasome

(Easwaran et al., 1999; Hart et al., 1999; Jiang and Struhl, 1998; Kitagawa et al., 1999; Liu et al., 1999), but how βcat is transferred from one complex to the next is poorly understood. One model for βcat movement through these complexes is a simple release and capture model, meaning that each complex remains in separate locations, and βcat physically moves from one complex to the next as each post translational modification alters it recognition by other proteins. This model leaves βcat open to dephosphorylation, and thus short-circuiting its destruction (Su et al., 2008). An alternative model is that that these complexes interact with one another, suggesting βcat is directly passed from one complex to the next.

The APC2:Axin complex recruits the F-box E3 adaptor Slimb

While previous studies have shown that Slimb/βTrCP can coIP with Axin, they have not examined whether they co-localize in Axin puncta. To begin to differentiate between these two proposed models, we first examined the localization of the DC and Slimb. When Slimb was expressed alone, it was found throughout the cell without obvious enrichment in a particular structure, similar to what we observe with APC2 when expressed alone. Axin expressed alone accumulates in cytoplasmic puncta, and co-expression of APC2 and Axin led to APC2 recruitment into puncta. When all three proteins were co-expressed, Slimb was recruited into the APC2:Axin complex (Fig 3.3A). These data suggest that APC2, Axin, or both can recruit this key E3 ligase component, consistent with the possibility that βcat may not be released from the from the DC to travel to the E3, but instead may be passed from one complex to the next.

143

Figure 3.5

Figure 3.5: Slimb turnover in the destruction complex is unaffected by co-localization with Axin or Axin and APC2. A-B) Slimb expression has no effect on Axin stability in puncta, even when expressed with APC2. C-D) Slimb stabilization in puncta is unchanged when expressed with Axin or APC2 and Axin.

144

APC2 and SCFSlimb don’t mix, but Axin does

We next asked whether Slimb is recruited to the destruction complex by interaction with

Axin, APC or both. Understanding whether APC2 or Axin recruits the E3 to the DC could be key to understanding what goes awry in cancer. In over 80% of colon cancers APC is mutated, often leading to a truncated protein (CGAN, 2012). If APC was responsible for recruiting the E3, it could provide insights into how an APC truncation, which might cause loss of APC:E3 interaction and thus induce elevated levels of βcat, leads to activated Wnt signaling. This was a challenge to assess, as APC2 alone does not have a particular localization. However, we took advantage of a fully functional form of APC2 tagged so that it localized to mitochondria (Roberts et al., 2012). This was unable to recruit Slimb, suggesting that APC2 is not responsible for recruiting Slimb into the complex, and thus making it unlikely that truncations of APC in colon cancers affect E3 localization.

Although Axin is mutated in colon cancers at a much lower frequency than APC (CGAN,

2012), it is the scaffold of the DC, which can directly bind to almost all other components of the

DC (Dajani et al., 2003; Liu et al., 2002; Spink et al., 2000). It thus is perhaps not surprising that

Axin also is sufficient to recruit Slimb to Axin puncta (Fig 3.1E-F). However, we were surprised to find that Axin only robustly colocalizes with Slimb and does not recruit two other E3 proteins, in contrast to the expectation if the entire SCFSlimb complex was at the DC. These data may

suggest that βcat is shuttled between the DC and the E3 via Slimb.

The RGS domain of Axin is necessary to efficiently recruit Slimb

Early studies in mice have proposed an Axin:Slimb interaction, but suggested that this

interaction not direct but instead was a result of a βcat “bridge” (Liu et al., 1999). To see if this

was the same case, we created an Axin mutant protein that lacked the βcat binding site. This

mutant was still able to recruit Slimb, suggesting that the Slimb:Axin interaction is not solely a

result of both proteins binding to βcat. One possible explanation for the difference in our results

could be due to endogenous Axin in our cell lines still being able to recruit βcat, however the 145

mice experiments also contained endogenous Axin. In the end this may be a moot point since

our IP experiments suggest that it is Axin’s RGS domain that is able to interact with Slimb.

The RGS domain of Axin was original thought to function in G-protein signaling, due to it similar structure and sequence to other Regulators of G-protein Signaling (RGS) proteins like

RGS-4 (Spink et al., 2000). However, the G-protein that interacts with Axin has yet to be discovered. Instead explanations of the function of the RGS domain have focused on its interaction with APC’s SAMPs (Spink et al., 2000). It has remained controversial whether the

RGS domain is absolutely essential for Axin and DC function. Studies in Xenopus revealed that loss of Axin’s RGS domain (AxinΔRGS:RFP) causes Axin to behave as a dominant-negative –

meaning embryos behave as though Wnt signaling is never turned off (Hedgepeth et al., 1999;

Itoh et al., 1998). However, in Drosophia, embryos in which the only Axin is AxinΔRGS are have

an almost wildtype cuticle phenotype and a slight elevation of βcat throughout the embryo,

suggesting only a weak loss-of-function (Peterson-Nedry et al., 2008). Our data suggest that the

RGS domain is important for recruiting Slimb into the DC. It is possible to imagine that loss of

the Slimb:Axin interaction would prevent the efficient transfer of βcat from the DC to the E3 and

thus result in higher βcat levels.

The known function of the RGS domain is to directly bind with the SAMPs of APC2. The

RGS domain is comprised of 9 alpha-helixes, and the SAMPs directly bind to regions of helixes

2, 3, 4, and 5 (Spink et al., 2000). On the 9th helix, there is a conserved residue (Y247 in human

Axin, Y171 in Drosophila Axin) which forms a pi-helix loop which has been shown to be

unnecessary for Axin-RGS:APC-SAMP (Spink et al., 2000). I hypothesize that this highly

conserved residue/pi-loop may be important for the Axin:Slimb interaction. This would be

interesting to test moving forward. The first 5 helices of the RGS domain fold onto the 9th (Spink et al., 2000). It is thus plausible that when APC is bound to the RGS it could block Slimb’s ability to interact with RGS. However, co-expression of APC2, Axin and Slimb revealed that all three proteins can co-localize. Thus it is possible that APC2 and Slimb could both bind to Axin at the

146

same time, or that different Axin proteins in the multiprotein DC are interacting with APC and

with Slimb. A competition experiment or a co-crystallization of Axin and Slimb could help resolve

this question.

βcat destruction occurs in a supra-molecular factory

βcat has 2 major functions within the cell: to maintain cell-cell adhesive contacts and the

other as a transcription factor. Because of βcat’s roll in cell junctions, βcat transcription and

translation is continual (Peifer et al., 1994; Riggleman et al., 1990). On average, 70% of βcat in

located at the cell membrane (Peifer, 1993). After βcat protein is made, it is first recruited to the

cell junctions until they are saturated, then allowing any additional βcat produced free to roam

the cytoplasm. In the absence of Wnt signaling this “free floating” βcat is sequestered by the

βcat destruction complex, composed of APC, Axin, CK1, and GSK3. Each DC contains 10-100s

of Axin molecules, and presumably 10-100s of the other core components since Axin can either

directly or indirectly bind all other core components of the DC (Dajani et al., 2003; Liu et al.,

2002; Spink et al., 2000). After binding with Axin or APC2, βcat is sequentially phosphorylated

by CK1 and then GSK3 (Liu et al., 2002). Our data suggest that Slimb, the substrate

recognizing protein of the E3 ligase, is also a part of the DC, via its interaction with Axin’s RGS

domain. This suggests that once phosphorylated, βcat is passed/released to Slimb, perhaps by

APC (Pronobis et al., 2015). Slimb:βcat is now free to interact with the other E3 ligase proteins

(Cul1 and SkpA) that are either bound toSlimb or nearby. SCFSlimb adds Ub to βcat, βcat is then either directly transferred to the proteasome (Thorvaldsen), shuttled by a helper protein to the proteasome, or recruited via the proteasome’s affinity for ub, thus resulting in breakdown of

βcat. Our data further support the idea that the destruction complex is more like a phase-phase transition complex, with multiple players kept in close proximity to perform a particular function in a quick and organized fashion.

147

Figure 3.6

Figure 3.6: SIM imaging reveals that Slimb is recruited along Axin cables. To further evaluate Axin:Slimb interaction we utilized structural illumination microscopy (SIM). A, D, and G) SIM images of whole cells with marked constructs expressed. B-C) Close-up images of Axin puncta. E-F) Close-up images of destruction complexes with intertwined cables of APC2 and Axin. H-I) Magnification of puncta in G. Slimb is closely localizes along Axin cables.

148

METHODS:

Cell Culture and transfection

The colorectal-cancer cell line, SW480, was used for all cell culture experiments. Cells were maintained in L-15 media (Corning) supplemented with 10% heat inactivated FBS and 1x

Pen/Strep (Gibco) at 37° C with ambient CO2 levels. Lipofectamine 2000 (Life Technologies)

was used for transient transfection of constructs following manufacturer’s instructions. All

constructs contained the pCMV-backbone and Drosophila genes were inserted using the

pCR8/Gateway protocol (Invitrogen) and tagged with GFP, RFP, or Flag as in (Pronobis et al.,

2015).

Immunofluorescence and Microscopy

Cells grown on coverslips were collected for immunofluorescence (IF) 24 hours after

transfection. Briefly, cells were washed in PBS and then fixed in 4% formaldehyde for 5

minutes. Cells were then permeabilized with 0.1% TritonX-100 in PBS for 5 minutes. After 30

minutes in block buffer (0.01% NGS in PBS), cells were incubated in primary antibody for 1-2

hours, washed with PBS, and then incubated in secondary for 1-2 hours. Cells were mounted

on microscope slides using poly-aquamount. Primary antibodies used: anti- βCatenin (BD

Transduction, 1:800) and anti-M2-Flag (Sigma, 1:1000).

Immunostained cells were imaged on the LSM Pascal microscope (Zeiss), LSM 710

(Zeiss) or LSM 880 (Zeiss). All images were processed using FIJI (Fiji Is Just ImageJ) or

IMARIS to create max intensity projections, and Photoshop CS6 (Adobe, San Jose, CA) was

used to adjust input levels so that the signal spanned the entire output grayscale and to adjust

brightness and contrast.

Immunoprecipitation and Western blotting

Cells were collected in lysis buffer (150 mM NaCl, 30 mM Tris pH 7.5, 1 mM EDTA, 1%

Triton-X-100, 10% glycerol, 0.5 mM DTT, 0.1 mM PMSF plus proteinase/phosphatase inhibitors

149

(EDTA-free, Pierce)) 24 hours after transfection. Antibodies were conjugated to magnetic beads and incubated will lysed cells overnight. After washing in lysis buffer, immuno-precipitated proteins were removed from the beads with 2xSDS buffer run on an 8 or 10% SDS PAGE gel and transferred to a nitrocellulose membrane. Westerns were visualized using Flim or the

Typhoon Imager. Primary Antibodies: anti-GFP (JL-8 Clontech, 1:1000), anti-Flag (Sigma-

Aldrich, 1:2000), anti-γ-tubulin (Sigma-Aldrich, 1:2000). Secondary Antibodies: IRDye680RD anti-Rabbit (Licor, 1:10,000), and IRDye800CW anti-Mouse (Licor 1:10,000).

150

REFERENCES

Berndsen, C.E., and Wolberger, C. (2014). New insights into ubiquitin E3 ligase mechanism. Nature structural & molecular biology 21, 301-307.

CGAN, C.G.A.N. (2012). Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330-337.

Clevers, H., and Nusse, R. (2012). Wnt/beta-catenin signaling and disease. Cell 149, 1192- 1205.

Dajani, R., Fraser, E., Roe, S.M., Yeo, M., Good, V.M., Thompson, V., Dale, T.C., and Pearl, L.H. (2003). Structural basis for recruitment of glycogen synthase kinase 3β to the axin—APC scaffold complex. The EMBO journal 22, 494-501.

Easwaran, V., Song, V., Polakis, P., and Byers, S. (1999). The ubiquitin-proteasome pathway and serine kinase activity modulate adenomatous polyposis coli protein-mediated regulation of beta-catenin-lymphocyte enhancer-binding factor signaling. The Journal of biological chemistry 274, 16641-16645.

Fiedler, M., Mendoza-Topaz, C., Rutherford, T.J., Mieszczanek, J., and Bienz, M. (2011). Dishevelled interacts with the DIX domain polymerization interface of Axin to interfere with its function in down-regulating β-catenin. Proceedings of the National Academy of Sciences 108, 1937-1942.

Fuchs, S.Y., Spiegelman, V.S., and Suresh Kumar, K.G. (2004). The many faces of β-TrCP E3 ubiquitin ligases: reflections in the magic mirror of cancer. Oncogene 23, 2028.

Hart, M., Concordet, J.P., Lassot, I., Albert, I., del los Santos, R., Durand, H., Perret, C., Rubinfeld, B., Margottin, F., Benarous, R., et al. (1999). The F-box protein beta-TrCP associates with phosphorylated beta-catenin and regulates its activity in the cell. Current biology : CB 9, 207-210.

Hedgepeth, C.M., Deardorff, M.A., Rankin, K., and Klein, P.S. (1999). Regulation of Glycogen Synthase Kinase 3β and Downstream Wnt Signaling by Axin. Molecular and cellular biology 19, 7147-7157.

Itoh, K., Krupnik, V.E., and Sokol, S.Y. (1998). Axis determination in Xenopus involves biochemical interactions of axin, glycogen synthase kinase 3 and β-catenin. Current Biology 8, 591-594.

Jiang, J., and Struhl, G. (1998). Regulation of the Hedgehog and Wingless signalling pathways by the F-box/WD40-repeat protein Slimb. Nature 391, 493-496.

Kishida, S., Yamamoto, H., Hino, S., Ikeda, S., Kishida, M., and Kikuchi, A. (1999). DIX domains of Dvl and axin are necessary for protein interactions and their ability to regulate beta- catenin stability. Molecular and cellular biology 19, 4414-4422. Kisselev, A.F., Songyang, Z., and Goldberg, A.L. (2000). Why does threonine, and not serine, function as the active site nucleophile in proteasomes? The Journal of biological chemistry 275, 14831-14837.

151

Kitagawa, M., Hatakeyama, S., Shirane, M., Matsumoto, M., Ishida, N., Hattori, K., Nakamichi, I., Kikuchi, A., Nakayama, K., and Nakayama, K. (1999). An F-box protein, FWD1, mediates ubiquitin-dependent proteolysis of beta-catenin. The EMBO journal 18, 2401- 2410.

Kreiss, P., Cameron, B., Rangara, R., Mailhe, P., Aguerre-Charriol, O., Airiau, M., Scherman, D., Crouzet, J., and Pitard, B. (1999). Plasmid DNA size does not affect the physicochemical properties of lipoplexes but modulates gene transfer efficiency. Nucleic acids research 27, 3792-3798.

Liu, C., Kato, Y., Zhang, Z., Do, V.M., Yankner, B.A., and He, X. (1999). beta-Trcp couples beta-catenin phosphorylation-degradation and regulates Xenopus axis formation. Proceedings of the National Academy of Sciences of the United States of America 96, 6273-6278.

Liu, C., Li, Y., Semenov, M., Han, C., Baeg, G.H., Tan, Y., Zhang, Z., Lin, X., and He, X. (2002). Control of beta-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell 108, 837-847.

Peifer, M. (1993). The product of the Drosophila segment polarity gene armadillo is part of a multi-protein complex resembling the vertebrate adherens junction. J Cell Sci 105 ( Pt 4), 993-1000.

Polakis, P. (1997). The adenomatous polyposis coli (APC) tumor suppressor. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer 1332, F127-F147. Pronobis, M.I., Deuitch, N., Posham, V., Mimori-Kiyosue, Y., Peifer, M., and Montell, D. (2016). Reconstituting regulation of the canonical Wnt pathway by engineering a minimal β- catenin destruction machine. Molecular biology of the cell 28, 41-53.

152

Riggleman, B., Schedl, P., and Wieschaus, E. (1990). Spatial expression of the Drosophila segment polarity gene armadillo is posttranscriptionally regulated by wingless. Cell 63, 549-560.

Roberts, D.M., Pronobis, M.I., Poulton, J.S., Kane, E.G., and Peifer, M. (2012). Regulation of Wnt signaling by the tumor suppressor adenomatous polyposis coli does not require the ability to enter the nucleus or a particular cytoplasmic localization. Molecular biology of the cell 23, 2041-2056.

Salic, A., Lee, E., Mayer, L., and Kirschner, M.W. (2000). Control of β-Catenin Stability: Reconstitution of the Cytoplasmic Steps of the Wnt Pathway in Xenopus Egg Extracts. Molecular Cell 5, 523-532.

Schwarz-Romond, T., Metcalfe, C., and Bienz, M. (2007b). Dynamic recruitment of axin by Dishevelled protein assemblies. Journal of Cell Science 120, 2402-2412.

Spink, K., Fridman, S.G., and Weis, W.I. (2001). Molecular mechanisms of beta-catenin recognition by adenomatous polyposis coli revealed by the structure of an APC-beta- catenin complex. The EMBO journal 20, 6203-6212.

Spink, K.E., Polakis, P., and Weis, W.I. (2000). Structural basis of the Axin–adenomatous polyposis coli interaction. The EMBO journal 19, 2270-2279. Su, Y., Fu, C., Ishikawa, S., Stella, A., Kojima, M., Shitoh, K., Schreiber, E.M., Day, B.W., and Liu, B. (2008). APC is essential for targeting phosphorylated beta-catenin to the SCFbeta-TrCP ubiquitin ligase. Mol Cell 32, 652-661.

Tanaka, K. (2009). The proteasome: overview of structure and functions. Proceedings of the Japan Academy Series B, Physical and biological sciences 85, 12-36.

Wenzel, D.M., Stoll, K.E., and Klevit, R.E. (2011). E2s: structurally economical and functionally replete. The Biochemical journal 433, 31-42.

Wojcik, E.J., Glover, D.M., and Hays, T.S. (2000). The SCF ubiquitin ligase protein Slimb regulates centrosome duplication in Drosophila. Current Biology 10, 1131-1134.

153

Wu, G., Xu, G., Schulman, B.A., Jeffrey, P.D., Harper, J.W., and Pavletich, N.P. (2003). Structure of a β-TrCP1-Skp1-β-Catenin Complex: Destruction Motif Binding and Lysine Specificity of the SCFβ-TrCP1 Ubiquitin Ligase. Molecular Cell 11, 1445-1456.

Xu, W., and Kimelman, D. (2007). Mechanistic insights from structural studies of β-catenin and its binding partners. Journal of Cell Science 120, 3337-3344.

154

CHAPTER 4: DISCUSSION

And the debate goes on: Wnt regulation of the destruction complex

Development is an amazing process in which a single cell divides and multiplies to create a multicellular organism. During this process cells must reorganize spatially and temporally to determine their final cell fate. This reorganization requires cells to communicate with one another. There are a handful of essential cell signaling pathways that aid in cell fate and organization (Basson, 2012). One of these essential signaling pathways is Wnt signaling.

The importance of this pathway in development is emphasized by its conservation throughout animals (Loh et al., 2016). Not only is Wnt signaling necessary for development, but it is also essential for tissue homeostasis. For example, Wnt signaling is responsible for maintaining the colon’s stem cell population. In fact, 90% of all colorectal cancers contain an activating mutation in this pathway (CGAN, 2012). Even though this pathway is ubiquitous and has been studied for over thirty years, the field is still debating the mechanisms of pathway is regulation.

The basics: Wnt signaling pathway simplified

The main function of Wnt signaling is to regulate the levels of β-catenin (βcat), a co- activator of transcription. In the absence of Wnt signaling, the destruction complex comprised of the tumor suppressor APC, the scaffold Axin, and the kinases CK1 and GSK3 keeps βcat levels low. The destruction complex recruits βcat to be sequentially phosphorylated by CK1 and GSK3

(Liu et al., 2002). This modification of βcat allows it to be recognized by a Skp-Cullin-Fbox E3 ligase (SCF βTrCP) which ubiquitinates βcat, labeling it for proteasomal degradation

155

(Aberle et al., 1997; Kitagawa et al., 1999; Liu et al., 2002). Thus, in the absence of Wnt

signaling βcat levels are kept low, preventing Wnt induced transcription.

During Wnt signaling, a Wnt ligand interacts with the seven transmembrane receptor

Frizzled and its co-receptor LDL-receptor-related protein (LRP) 5/6 (Arrow in Drosophila),

creating a 3-protein receptor-ligand complex which promotes and stabilizes Dishvelled interaction with this complex (Angers and Moon, 2009; DeBruine et al., 2017; Tamai et al.,

2000). Phosphorylation of LRP5/6 enhances recruitment of Axin to the membrane and destruction complex function is turned down (DeBruine et al., 2017). Throughout the years, many models have described how destruction complex function is decreased, ranging from complex disassembly, inhibition of GSK3, to degradation of Axin. These will be briefly described in the next section. However, it is agreed that Wnt signaling increases cytoplasmic levels of

βCat, and thus allowing it to enter the nucleus and works with the family of TCF and LEF transcription factors to behave as a co-activator of transcription to express different Wnt target genes.

There is agreement as to the main positive and negative regulators of the Wnt signaling pathway (ex: Frizzled, Wnt, LRP5/6, Dsh, APC, Axin, CK1, GSK3, and βCat), yet many key

questions of destruction complex regulation remain. For example: Axin polymerization is

necessary for destruction complex function and the destruction complex contains 100s of Axin

molecules (Faux et al., 2008; Fiedler et al., 2011; Kishida et al., 1999; Schaefer et al., 2018) -

does the size of the complex matter? Also, what do APC and Axin physically do to regulate

destruction complex function both in the presence and absence of Wnt signaling? Another gap

in the field is mechanism of βcat transfer from the destruction complex to the E3 ligase. Does

the destruction complex and the E3 ligase physically interact with one another? The F-box

protein βTrCP (Slimb in Drosophila), which can directly bind to a phosphorylated βcat (Wu et al.,

2003) can co-immunoprecipitate (co-IP) with Axin (Hart et al., 1999; Li et al., 2012; Liu et al.,

1999b), suggesting a possible role of Axin to recruit the E3 to the destruction complex. Lastly,

156 what is the mechanism of destruction complex turn down? Dishvelled is necessary to recruit

Axin to the Wnt receptor complex to turn down destruction complex function (Cliffe et al., 2003;

Fiedler et al., 2011; Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b), yet it is not fully understood what Dishevelled is doing to the complex. By understanding how these proteins interact with one another on a molecular level we can better understand how the system goes awry in cancer I will discuss each of these questions in more detail below.

Size matters: Why is polymerization necessary for complex function?

Most models of destruction complex function represent the complex as a simple tetramer of APC:Axin:CK1:GSK3 with a 1:1:1:1 ratio. Interestingly, Axin and APC self-association have been shown to be necessary for complex function (Fiedler et al., 2011; Kishida et al., 1999;

Kunttas-Tatli et al., 2014; Mendoza-Topaz et al., 2011). Axin contains a DIX domain which allows it to self-polymerize and form puncta (Faux et al., 2008), and loss of polymerization decreases destruction complex function (Fiedler et al., 2011; Kishida et al., 1999; Pronobis et al., 2016). APC also contains a conserved self-association domain (ASAD). When this domain is deleted in cancer cells, Axin was less able to form robust puncta (Kunttas-Tatli et al., 2014), and function in Wnt signaling is lost (Kunttas-Tatli et al., 2014; Kunttas-Tatli et al., 2015; Roberts et al., 2012). In concert with this, in Drosophila embryos completely lacking APC2, (the main

APC family member during Drosophila embryogenesis), no Axin puncta were seen (Mendoza-

Topaz et al., 2011). These data suggest that puncta formation is necessary for complex function, but why would polymerization matter?

Condensates: the more the merrier

A re-emerging field in science is that of “phase separation” biochemistry, allowing the creation of a large non-membrane compartment, called a condensate (review in (Banani et al.,

2017)). Within a cell there are 100s to 1000s of different chemical reactions occurring ranging from transcription to protein synthesis, and protein degradation. But how do reactants find one

157

another? By concentrating reactants in a condensate, different chemical processes can be

enhanced, while their separation can inhibit the same reactions. One example of this is the

nucleolus - a non-membrane-bound compartment which contains many of the components necessary for ribosome biogenesis (Boisvert et al., 2007; Hyman et al., 2014).

In the Introduction chapter, we proposed that the destruction complex is another

example of a cellular condensate. I propose that the destruction complex forms condensates to

increase its avidity for βCat and the E3 ligase. Interestingly, Axin’s binding affinity for βCat has

been suggested to be quite low, so like adding more people to a tug of war, concentrating Axin

molecules into puncta could increase Axin’s avidity/pull for βCat (Salic et al., 2000). In addition,

APC’s ability to recruit βCat is quite variable and dependent on its phosphorylation (Easwaran et

al., 1999; Rubinfeld et al., 1996; Xing et al., 2004). In the absence of phosphorylation, none of

APC’s 10 βCat binding sites (3x 15 amino acid repeats; 7x 20 amino acid repeats) are able to

sequester βCat and prevent TCF binding (Spink et al., 2001; Xing et al., 2004). Once APC is

phosphorylated, its affinity for βCat is enhance by 3-500 fold, and can now block TCF binding to

βCat (Xing et al., 2004). Taken together, these results suggest that Axin and APC may need to

form condensates simply to gain the “strength” to recruit βCat. Even though APC can robustly

recruit βCat, it must first be phosphorylated, yet when absent from the destruction complex APC

has not been shown to contain a binding site for any kinases. However, Axin can directly bind

both APC and the kinase GSK3, and associates with CK1(Dajani et al., 2003; Liu et al., 2002;

Spink et al., 2000), therefore Axin can bring a kinase within close proximity to APC, which can

then be phosphorylated to more robustly recruit βCat. In fact, earlier studies have shown that

APC phosphorylation is dependent on Axin (Rubinfeld et al., 2001).

APC has also been shown to stabilize Axin assembly into puncta and increase puncta volume compared to when Axin is expressed alone (Pronobis et al., 2015). Since Axin is the scaffold on which the complex is assembled, these data suggest that APC:Axin interaction could also stabilize other components of the destruction complex that bind Axin. My proposed model

158

would thus suggest that regulation of destruction complex size regulates its ability to recruit βcat

and other essential components of the destruction complex. Regulation of destruction complex

size could also be a means of destruction complex downregulation. In support of this, when a

GFP-tagged Axin is expressed at near endogenous levels, the number of Axin molecules per

punctum is decreased by half in cells receiving Wnt signaling (Schaefer et al., 2018). Therefore,

in the Wnt-ON cells, cells in which the destruction complex is turned down, there is a decrease in the Axin numbers within the complex. This further suggests that at this lower number of the scaffold protein, the destruction complex is less efficient at recruiting βCat to label it for

proteasomal degradation. To further test whether Wnt regulates complex size, it would be worth

counting the number of Axin molecules in Drosophila embryos that either over- or under-

express Wnt signaling. If Wingless (Wg, Drosophila Wnt) regulates complex size, then the

average number of Axin molecules per punctum in the Wg over-expression embryos would be

half that of wg mutant embryos. It would also be interesting to explore how many molecules of

APC are also in the complex, and whether this number also change in response to Wnt

signaling.

It is interesting to note that in esophageal cancers, low levels of Axin correlate with poor

patient prognosis (Li et al., 2009). This suggest that there may be a minimal level of total Axin

protein to form enough complexes to regulate βCat. Interestingly, early reports on protein levels

within Xenopus egg extracts have suggested core complex proteins were expressed at

dramatically different levels (ex: APC:Axin = 5000:1) (Lee et al., 2001; Lee et al., 2003). Based

on these data, it was suggested that Axin is the limiting component of destruction complex

function. However more recent studies have suggested that Axin and APC levels in humans and

Drosophila are within the same order or magnitude (Kitazawa et al., 2017; Schaefer et al., 2018;

Tan et al., 2012). I found that in Drosophila embryos in vivo, raising levels of APC or Axin via

protein over-expression does not enhance baseline destruction complex function, suggesting

that neither APC or Axin are rate limiting in βCat destruction in this cell type. It may make more

159

sense that one of the proteins that post-translationally alter βCat, such as CK1, GSK3, or the E3

would be rate limiting, causing βCat destruction to slow down. Interestingly, addition of GSK3 to

Xenopus egg extracts enhances βCat destruction (Salic et al., 2000).

Condensates: A way to bring multiple complexes together?

The main function of the destruction complex is to phosphorylate βCat so it can be

recognized by the Skp-Cullin-F-box (SCF) E3 ligase, which labels βCat for proteasomal

degradation. Even though βTrCP (Slimb in Drosophila) has been known to be the F-box protein

that recognizes βCat for 20 years (Jiang and Struhl, 1998), how βCat is transferred to the E3 has remained mysterious. I propose 2 models: 1) The destruction complex and the E3 ligase are two separate complexes and therefore βcat needs to be actively transported from one complex to another or 2) The destruction complex and the E3 are a part of the same condensate, thus allowing phosphorylated βCat to be passed from one protein to the next.

Several IP experiments have suggested the latter model (Hart et al., 1999; Kitagawa et al.,

1999; Li et al., 2012; Liu et al., 1999a), with βTrCP co-IPing with APC or Axin, but colocalization of these complexes has never been observed in vivo and the mechanism of this interaction has never been worked out.

Our preliminary work in SW480 cells (Chapter 3) has provided new insights in destruction complex interactions. First our data support previous reports that Axin, and not APC recruits Slimb into the complex. Interestingly, we rarely saw any other of the components of the

E3 (Cul1 or SkpA) associated with Axin puncta. These data suggest a hypothesis that Slimb shuttles phosphorylated βCat from the destruction complex to the rest of the E3 for ubiquitination. In my studies that examined the interaction between Axin and Slimb, robust recruitment of Slimb into Axin puncta occurred approximately 50% of the time, usually when

Axin levels were high, suggesting this interaction may be weak. These data suggest that the more Axin present, the more readily Slimb is recruited to the destruction complex.

160

As mentioned previously, APC promotes Axin is stabilization in puncta and puncta size,

and there are more Axin molecules per punctum in Wnt-OFF cells than in Wnt-ON cells

(Pronobis et al., 2015; Schaefer et al., 2018). Based on these data, we could hypothesize that in

Wnt-OFF cells, when complexes are at their largest size, the destruction complex is more

readily able to recruit βTrCP. Therefore, a decrease in complex size in Wnt-ON would therefore

cause a decrease in Slimb recruitment and slow the passing of βCat to the E3 thus resulting in

increased βCat levels. These data are supported by IP data that shows in response to Wnt

signaling, Axin is no longer able to recruit βTrCP (Li et al., 2012). A simple binding assay could

test whether Axin concentration affects its interaction with βTrCP. To test whether βTrCP is

shuttled between the destruction complex and the rest of the E3 ligase, it would also be

interesting to restrict βTrCP movement and observe whether this has an effect on βcat

destruction. If βcat remains stuck in the destruction complex, it could suggest that βTrCP is

necessary to shuttle βcat between complexes. However, if βcat is not stuck in the complex then

perhaps another protein that also interacts with βTrCP could be responsible for transferring βcat

to the E3. In vertebrates, WTX could be a candidate gene as it has been shown to enhance

βcat ubiquitination and can directly interact with Axin and βTrCP1/2 (Major et al., 2007).

However, there is no known WTX gene in Drosophila. Interestingly, preliminary results

comparing co-localization between fly Axin and Slimb and human Axin and βTrCP, βTrCP was

more robustly recruited to Axin puncta.

Multiple mechanisms of turing down of the destruction complex

Since most colorectal cancers contain an activating mutation in the Wnt pathway,

understanding how the destruction complex is inhibited could provide insights in cancer drug

treatments. Could a drug be created to promote destruction complex function? As mentioned

above, the exact mechanism of destruction complex inhibition is still a mystery. There are

several different proposed models for destruction complex regulation. Here I focus on three:

Axin degradation, destruction complex re-localization, and loss of interactants. 161

Axin is the scaffold of the destruction complex, as it contains known or presumptive binding sites for itself, APC, CK1, GSK3, βCat, Dvl, and LRP5/6 and can recruit core complex proteins into cytoplasmic puncta (Dajani et al., 2003; Faux et al., 2008; Fiedler et al., 2011;

Kishida et al., 1999; Liu et al., 2002; Spink et al., 2000). Early work in Xenopus egg extracts suggested that Axin was exceptionally rate-limiting, with levels as much as 5000-fold below those of other destruction complex proteins (Lee et al., 2003; Salic et al., 2000) Recent research

revealed that Axin levels are not universally much lower than any other protein with in the

destruction complex-in fact Axin levels in both Drosophila and in cultured normal and colorectal

cancer cells are in the same order of magnitude as those of APC (Kitazawa et al., 2017;

Schaefer et al., 2018; Tan et al., 2012). Interestingly, previous work in Drosophila examining the

role of Wg on destruction complex function suggest very different modes of complex regulation.

One group found that Axin levels were strongly reduced in response to Wg (Tolwinski et al.,

2003). Another suggested that in response to Wg signaling, Axin was initially stabilized, causing

an increase in both the membrane and cytoplasmic pools of Axin (Wang et al., 2016a; Wang et

al., 2016b; Yang et al., 2016). Finally, a third group suggested that Wg had no effect in Axin

levels, but instead Axin was recruited to the membrane in a Dsh dependent manner (Cliffe et al.,

2003). In the following sections, I’ll explore these different models, and consider how the

destruction complex may be regulated by changes in localization, phosphorylation state, and

protein levels.

Axin degradation or disassembly: control the scaffold, control the complex?

Several labs have suggested that activation of Wnt signaling initiates the degradation of

Axin, though they differ in whether this is a rapid, primary response or a longer adaptation (Lee

et al., 2003; Mao et al., 2001; Tolwinski et al., 2003; Wang et al., 2016a; Wang et al., 2016b;

Yang et al., 2016). But how is Axin lost? One possible mechanism for Axin loss is via

modification by the poly-ADP-ribosylating (PARP) enzyme Tankyrase (Tnks) (Huang et al.,

2009). PARsylation of Axin can lead to ubiquitination and then proteasomal degradation of Axin 162

(Huang et al., 2009). In Drosophila and human cultured cells, after a 15-minute incubation with

Wnt, Axin PARsylation is increased, and then 1-2 hours later Axin protein levels are decreased

(Li et al., 2012; Wang et al., 2016b). Regulation of the destruction complex scaffold makes intuitive sense, since degradation of Axin would result in loss of complex formation. However, loss of Tnks has no effect on embryonic viability in Drosophila; in fact adult flies are grossly normal (Feng et al., 2014). These data suggest that TNKS may be a fine-tuning regulator of Wnt

signaling or that there is another mechanism to label Axin for degradation. Interestingly, in Tnks

mutant flies, Axin levels are still decreased 1-2 hour after Wnt signaling (Yang et al., 2016) If

TNKS is unnecessary for Axin degradation, then how are Axin levels decreased after Wnt

signaling? Are there two phases of Axin degradation? Axin degradation after long Wnt exposure

could be mediated by the RING domain E3 ligase SIAH 1/2. Recently SIAH 1 and SIAH 2 have

been identified as novel inhibitor of Axin levels in HEK293T cells (Ji et al., 2017). Once again,

however, it is important to note that the single Drosophila Sina family member is adult viable

without obvious defects in Wnt signaling (Carthew and Rubin, 1990).

Does a change in scenery result in a change in function?

Early work exploring how Wnt signaling effected Axin often utilized Axin-over-expression

studies (Cliffe et al., 2003; Mendoza-Topaz et al., 2011; Tolwinski et al., 2003). However, over-

expression of Axin is known enhance degradation of βcat/Arm in several model systems

(Hamada et al., 1999; Nakamura et al., 1998). Therefore, is the observed regulation of Axin

seen after over-expression similar to how endogenous Axin would behave? Recent papers

looking at near endogenous levels of Axin have come to somewhat different conclusions about

Axin localization in Wnt-OFF and Wnt ON cells (Schaefer et al., 2018; Wang et al., 2016a;

Wang et al., 2016b; Yang et al., 2016).

One group expressed Axin at near endogenous levels during embryogenesis and

visualized Axin by immuno-staining with antibodies to an epitope tag. Based on their Axin

staining, in Wnt-OFF cells Axin was found throughout the cytoplasm at low levels, and no 163 cytoplasmic puncta were apparent. In Wnt-ON cells Axin levels were initially elevated both in the cytoplasm and at the membrane (visualized as an increase in staining), but dropped Axin levels several hours later (visualized as a decrease in staining). Interestingly, when we expressed a

GFP-tagged Axin (Axin:GFP) at near endogenous levels we observed a different pattern. By using Axin:GFP, we were able to visualize native destruction complex localization without the use of an antibody. In Wnt-OFF cells, Axin:GFP was robustly recruited into cytoplasmic puncta.

In Wnt-ON cells, Axin puncta were recruited to the membrane and the cytoplasmic pool of

Axin:GFP was increased (Chapter 2; (Schaefer et al., 2018)). The membrane bound puncta contained reduced numbers of Axin proteins, providing a possible mechanism by which Wnt signaling turns down the destruction complex.

Why was there a difference in Axin localization pattern between these different labs? I hypothesize that it is the difference at looking at native protein expression versus using an antibody. When we have immune-stained Axin after it is over-expressed in colon cancer cells, and noticed that antibodies have difficulty penetrating puncta, thus only labeling the outer layer of Axin proteins and none of the internal protein. This caused puncta fluorescence brightness to decrease, such that it stood out less from the surrounding free Axin. We thus believe directly visualizing Axin using a GFP tag provides a more realistic view—our localization patterns are also consistent with previous studies (Cliffe et al., 2003; Schwarz-Romond et al., 2007b).

Another proposed mechanism for downregulation of the destruction complex is that complex localization to the membrane induces disassembly of the destruction complex.

However, when visualized Axin:GFP and immuno-stained for APC2, we noticed that almost all

Axin:GFP puncta contained APC2, both the cytoplasmic puncta in Wnt-OFF cells and the membrane-associated puncta in Wnt-ON cells, suggesting that APC2 remains within the complex in response to Wnt signaling (Schaefer et al., 2018). This supported by co-IP data in a human cancer cell line. Axin was able to co-IP APC before and after treatment with Wnt (Li et al., 2012). To further verify that the destruction complex remains intact, immuno-staining for

164 both GSK3/Zw3 and CK1 during Drosophila embryogenesis would be worth visualizing. While we did not see changes in Axin or APC2 localization in the destruction complex, we did observe that Dsh was recruited into Axin:GFP puncta only in cells receiving Wg (Schaefer et al., 2018). I will discuss Dishevelled more below.

The destruction complex loses a friend

In contrast to proposed mechanisms regulating destruction complex function by Axin degradation, another model of complex regulation suggests that βCat levels increase in response to Wnt because the complex is unable to interact with the E3 ligase (Li et al., 2012).

Specifically, this model suggests that Wnt signaling inhibits the interaction between Axin and

βTrcP while Axin’s interaction with GSK3 is maintained so that βCat is still phosphorylated (Li et al., 2012). Hence βCat levels rise because they are no longer ubiquitinated by the E3.

Interestingly these authors accept the idea that Axin protein levels decrease after long exposure to Wnt signaling (1-2 hours) but argue that the initial response of the destruction complex to Wnt signaling is more important (Li et al., 2012). How might the Axin: βTrcP interaction be inhibited?

One possibility is that in response to Wnt signaling, the complex decreases in size and therefore the complex avidity for βTrcp is decreased as well (see above for more detail). The model suggesting E3 ligase interaction loss was based on data from co-IP experiments and therefore changes in destruction complex size could not be observed.

Another proposed model for destruction complex downregulation is through inhibition of

GSK3. The mechanisms of GSK3 inhibition range from blockage of the GSK3:Axin interaction, to suppression of kinase activity, to sequestration of GSK3 into multivesicular bodies (Cselenyi et al., 2008; Piao et al., 2008; Taelman et al., 2010; van Amerongen et al., 2005). In Xenopus,

Frat1 can directly bind to GSK3 to block GSK3:Axin interaction (Ferkey and Kimelman, 2002; Li et al., 1999). However, in mice Frat1 activity is not important for regulation of Wnt signaling (van

Amerongen et al., 2005), nor is there Frat1 homolog in Drosophila. These data suggest that while Frat1 may be important for Wnt regulation in Xenopus, it is not a universal mechanism. 165

Another mechanism for GSK3 inhibition is by inactivation of its kinase activity by Wnt signaling

(Cselenyi et al., 2008). In the presence of Wnt ligands, Wnts bind to Frizzled (Fzd) and/or

LRP5/6 to stabilize the Wnt/Fzd/LRP5/6/Dvl complex. (Angers and Moon, 2009; DeBruine et al.,

2017; Tamai et al., 2000). Axin is recruited to the membrane in a Dvl dependent manner and

LRP5/6 is phosphorylation is enhanced (Tamai et al., 2000). Phosphorylation of LRP5/6 at its

PPPSP motifs is required to inhibit GSK3 phosphorylation of βCat in an Axin independent manner (Cselenyi et al., 2008; Stamos et al., 2014). Using co-crystallization assays, phosphorylation of LRP5/6 creates and inhibitory substrate in which GSK3 binds, thus inactivating kinase activity (Stamos et al., 2014). Since Frat1 is only important in Wnt signaling regulation in Xenopus and not in mammals, it is worth knowing whether GSK3 kinase inhibition is universal.

The last model of GSK3 inhibition is though GSK3 sequestration (Taelman et al., 2010).

In this model, after activation of the Wnt receptor complex (also referred to as the signalsome), the receptor complex is taken up by endocytosis, creating an endosome. Multiple endosomes then fuse together, sequestering GSK3 and the Wnt receptor complexes inside multivesicular bodies (MVB) (Taelman et al., 2010). One caveat of this model is that the formation of MVBs occurred hours after initiation of Wnt signaling. What, if anything, happens to GSK3 activity immediately after Wnt signaling?

Whatever the exact mechanism is of GSK3 downregulation, pause in GSK3 activity could allow for accumulation of βCat and thus activation of transcription of Wnt target genes. It is interesting to note that in Drosophila embryos, loss of Zeste White 3 (zw3, the Drosophila homolog of GSK3) activity prevented the destruction complex from being localized to the membrane in Wnt-ON cells and also led to accumulation of βcat in the destruction complex, and

(Schaefer et al., 2018). GSK3 is also necessary to phosphorylate LRP5/6 to enhance Axin recruitment to the membrane (Zeng et al., 2008). GSK3 is necessary for Axin recruitment to the membrane/Wnt receptor complex, and thus immediate and complete inactivation GSK3 would

166

more than likely prevent destruction complex downregulation. Taking this into account, I

hypothesize that either GSK3 inhibition is delayed (not occurring until after it can phosphorylate

LRP5/6) or inhibition of its kinase activity is limited to phosphorylation of βCat. Since GSK3 is

essential for several signaling pathways (PI3K, ERK, etc), global inhibition of GSK3 in response to Wnt signaling, even if it was delayed, would be detrimental to the entire cell, and thus is unlikely. However, this provides an example of when a cellular condensate is useful. In this situation, only GSK3 in the complex would be inhibited in response to Wnt signaling.

Understanding this mechanism of GSK3 inhibition would be of excellent value to the cancer community.

Dishevelled: The regulator of Axin puncta

Dishevelled (Dvl in vertebrates, Dsh in Drosophila) was first discovered in Drosophila

and named for the misoriented hairs found on adult bodies and wings in mutant flies (Fahmy

and Fahmy, 1959), as part of what is now known as in the planar cell polarity pathway. Several

decades after its discovery, Dsh was found to have essential role as a signal transduction

protein for Wnt signaling (Perrimon and Mahowald, 1987). Cells lacking Dsh are unable to

transduce signals from Wnts through the Wnt/Fzd/LRP5/6 receptor complex and pass this

signal to the destruction complex (Klingensmith et al., 1994). Dsh’s essential roles in both Wnt

and planar cell polarity pathways suggests that Dsh is the fork in the road between these two

pathways (reviewed in (Sharma et al., 2018)). There is a single Dsh gene in Drosophila and 3

Dvls in Xenopus, mice, and humans. All Dvls contain 3 highly conserved domains: DIX, PDZ,

and DEP. The DIX domain of Dvl shares similarity to Axin’s DIX domain, and as in Axin, allows

for self-polymerization (Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b). The

PDZ domain allows for direct binding between Dvl and Frizzled after Dvl recruitment to the

membrane (MacDonald and He, 2012). Both domains are necessary for Wnt signaling. The

planar cell polarity pathway utilizes the DEP domain as well as the PDZ (Sharma et al., 2018;

Wallingford and Habas, 2005). Here, I will focus on Dvl/Dsh role in Wnt signaling. Even though 167

Dvl/Dsh has been studied for over 50 years, the exact mechanism of destruction complex down regulation by Dvl is still not clear.

Dvl’s DIX domain allows it to both self-polymerize and hetero-polymerize with Axin through its DIX domain. In HEK293 cells, in the absence of Wnt signaling, Dvl is found in cytoplasmic puncta. The addition of Wnt signaling moves Dvl localization from the cytoplasm to the cell membrane, similar to Axin (Schwarz-Romond et al., 2007b). Co-transfection of Axin and

Dvl in cells results in co-localization of Axin and Dvl in puncta (Schwarz-Romond et al., 2007b).

When we looked at Drosophila embryos, we were somewhat surprised to learn that while Dsh was expressed in all cells, Dsh predominately colocalized with the Axin puncta in cells that are receiving Wnt signaling, and not in those where Wnt signaling was off. This was even true in embryos expressing Axin at almost 10-times the endogenous level (Chapter 2, (Schaefer et al.,

2018)). If Axin and Dsh can hetero-polymerize, why were Axin:Dsh puncta not seen in all cells?

These data suggest that the ability of Axin and Dsh to interact is Wg/Wnt dependent, but what mechanism regulates Axin:Dsh interaction? In addition, Wnt signaling induced destruction complex localization to the membrane causes a decrease in Axin molecules within the complex.

Does the interaction between Axin and Dsh cause this decrease in Axin molecules?

In the Introduction chapter we introduced the idea that the destruction complex is a cellular condensate, a group of proteins that are gathered together via low-affinity binding to catalyze cellular reactions. Some unique properties of these complexes are the ability of the proteins within to mix. Sometimes within the same mixture two proteins separate, suggesting differential affinities (Shin and Brangwynne, 2017), and sometimes separation within the condensate reduces condensate function. Interestingly, my preliminary results from co- expression of Axin and Dsh in colon cancer cells suggest that when Axin levels are higher than

Dsh, Axin and Dsh more directly colocalize/mix. As Dsh levels increase co-localization decreases and Axin begins to form smaller puncta, with Dsh surrounding them but not co- localizing with Axin. These data could suggest that changes in the ratio of Axin and Dsh can

168 control Axin:Dsh interaction. This is an interesting observation since in vivo, Axin and Dsh protein levels are within the same order of magnitude. It would be intriguing to look at the effects of Dsh overexpression in vivo in Drosophila on co-localization with Axin. It would also be interesting to look at Dsh levels in embryos either over-expressing Wg or in wg null embryos. If

Wnt signaling controls Dsh levels (e.g., Dsh levels are lower in Wg over-expressing embryos, thus preventing Axin and Dsh mixing) then perhaps regulation of Dsh levels and not Axin levels regulates destruction complex function.

If Dsh levels remain the same, then perhaps changes in post-translational modification such as phosphorylation alter Dsh “activity” and thus formation of Dsh:Axin complexes.

Condensates form due to low affinity interactions between proteins, so changes in phosphorylation could change the charge of Dvl to encourage either increased affinity or separation between Dvl and Axin. Such modification of Dvl could create an “active” pool. Dvl is a known phospho-protein containing 100s of possible phosphorylation sites, suggesting that Dvl may be activated by phosphorylation. Since 1995, it has been known that hyper-phosphorylation of Dsh occurs in response to Wnt signaling (Yanagawa et al., 1995). A few studies have tried to map out the function of some of conserved phosphorylation sites, but none have been found to be individually essential for Wnt signaling, although several kinases that phosphorylate Dvl have been found (reviewed in (Mlodzik, 2016; Sharma et al., 2018)). Perhaps changes in multiple phosphorylation sites may be necessary to affect Dvl’s function in Wnt signaling. It is interesting to note that between the DIX and PDZ domain lies a conserved basic region containing many serines and threonines (Sharma et al., 2018). Is this necessary for the Axin:Dvl interaction? The proximity of the basic region to the DIX domain could suggest that phosphorylation of the basic region could regulate Axin and Dvl interaction. Can Axin and Dvl still interact when the basic region cannot be phosphorylated? To test this, it would be interesting to create both a phospho- mimetic version and one that cannot be phosphorylated. If prevention of dephosphorylation or phosphorylation inhibited Dvl:Axin complexes, this could provide insights into the importance of

169

this domain in Dvl:Axin interaction. It is also possible that changes in phosphorylation are a red- herring, or that changes in phosphorylation are only necessary for Dsh function in the planar cell polarity pathway - either to inhibit or promote planar cell polarity.

It is interesting to note that even though the DIX domains of Axin and Dvl are quite

similar, they are not identical (Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b).

Surprisingly, preliminary results from super resolution microscopy suggest that when expressed

alone, Axin puncta and Dvl puncta have different architectures. Axin tends to form doughnuts or

pretzel structures whereas Dvl forms honeycomb like structures. Is this difference in puncta a

result of differences in DIX domain sequence or in its localization within the protein? Axin’s DIX

domain is found at its C-terminal end, whereas Dvl’s DIX domain is at the N-terminus. Perhaps

there is something about the position of the DIX domain that affects interaction between the

proteins? Surprisingly, my preliminary data suggest that when Axin’s DIX domain is expressed

without the rest of Axin, it does not interact with Dsh. Is there another piece of Axin that is

necessary for Axin-DIX:Dvl-DIX interaction? Previous studies have suggested that the affinity

between Axin-DIX:Dvl-DIX is very low (Kishida et al., 1999), leading to questions about what

other part of these proteins or other proteins aid in this interaction Consistent with this, Axin’s

DIX domain is not absolutely essential for Axin:Dvl interaction, but Dsh’s DIX domain is

(Schwarz-Romond et al., 2007a; Schwarz-Romond et al., 2007b). These data suggest that Dvl’s

ability to regulate Axin’s assembly into puncta is more complex than a simple DIX:DIX

interaction.

Another protein that could regulate Axin:Dvl complexes is APC. When APC, Axin, and

Dvl2 are co-expressed in HEK293 cells, APC:Axin:Dvl2 complexes are rare while Axin:APC or

Axin:Dvl complexes are more frequent, suggesting a competition between APC and Dvl for

interaction with Axin (Mendoza-Topaz et al., 2011). In Drosophila embryos APC, Axin, and Dsh

protein levels are all with in the same order of magnitude (Schaefer et al., 2018), consistent with

the idea that Dsh and APC may compete for interaction with Axin. To test this idea, it would be

170 interesting to perform a competition assay. Each of these proteins have been notoriously difficult to purify in full length versions, therefore to test for competition between APC and Dsh we could use inducible-promotors to gradually adjust the amounts of each protein, while holding the others constant. If there is more APC than Dsh, will Axin prefer APC interactions over Dsh? If

APC and Dsh are expressed at the same level, is Axin more often found with APC or Axin? If

Dsh levels are higher than Axin and the two proteins form separate puncta - where is APC found?

It also would be interesting to further explore the effect of each of these proteins on the others dynamics. APC can stabilize Axin within the destruction complex and increase volume of the complex (Pronobis et al., 2015). In vivo, we only see Dsh and Axin interacting in Wnt-ON cells, when the number of Axin molecules within the destruction complex decrease and cytoplasmic levels of Axin increase. A previous study in cells using FRAP analysis has suggest that Dsh increases Axin turnover in puncta (Schwarz-Romond et al., 2007b). To test Dsh’s effect on Axin puncta, we could measure the number and volume of Axin puncta as we altered

Dsh levels (as was done in (Pronobis et al., 2015)). It would also be interesting to count the number of Axin molecules in embryos with Dsh over-expression compared to dsh mutants.

Does the number of Axin molecules in destruction complexes change in response to Wnt in the absence of Dsh? These data would provide insight into the role of APC and Dsh on Axin puncta size and assembly state.

Revised Model: It’s all about who your friends are at the time

No good dissertation about a signaling pathway ends without providing a revised model on how the pathway works. Here goes:

Putting all the pieces together

In the absence of Wnt signaling, in order for the DC to form, it must first be seeded.

APC’s ASAD promotes APC self-association, which in turn signals recruitment of Axin and

171 initiates Axin self-assembly. Driven by multiple interactions sites between APC and Axin, a protein complex containing 10s-100s of Axin molecules is formed. Binding sites on Axin for CK1 and GSK3β recruit them into the complex. Although Axin can form puncta on its own, the addition of APC stabilizes the complex in size, increasing the destruction complex’s avidity for

βcat and the kinases thus, increasing their local concentration and promoting GSK3β phosphorylation of Axin and βcat. This increase in Axin number may also increase the destruction complex’s avidity for βTrCP. While APC’s ability to bind βCat is not essential for

βCat phosphorylation, APC’s association with Axin induces phosphorylation of some APC’s

βCat binding sites to increase APC recruitment of free βCat. After the initial APC:Axin interaction, APC may utilize one of its SAMP motifs to recruit Axin and then use another SAMP motif to move Axin into the right position to be phosphorylated by CK1/GSK3 or so CK1/GSK3 can more readily phosphorylate βCat of APC. Phosphorylation of Axin prevents its degradation and phosphorylation of βCat primes it for recognition by βTrCP. GSK3β also phosphorylates region B and R2 of APC. This induces a conformational change, in which one end of APC is released from Axin, thus exposing phosphorylated βCat to be removed from the complex.

βTrCP then binds to βCat, and a βTrCP: βCat complex then leaves the destruction complex to meet up with the rest of the E3 ligase (Cul1 and SkpA) for βCat ubiquitination and eventual proteasomal degradation. Phosphatases (such as PP1 or PP2A) then dephosphorylate R2/B and perhaps Axin to reset the system, allowing the complex to recruit and phosphorylate another βCat.

Shifting friends

During Wnt signaling, Wnt ligands bind to Frizzled receptors inducing clustering of

LRP5/6 and creating a Wnt:Frizzled/LRP5/6 complex. Wnt ligand binding to Frizzled causes a conformational change to its intracellular domain, thus recruiting Dvl for direct binding. Perhaps by utilizing Dvl DEP domain for receptor oligomerization with DIX domain self-polymerization,

10-100s of Wnt/Frizzled/LRP5/6/Dvl complexes group together to form the signalsome 172 condensate. This condensate formation may “activate” Dvl to enhance recruitment the destruction complex to the membrane—Dvl phosphorylation is one possible mechanism. During this time the cytoplasmic tail of LRP5/6 is highly phosphorylated, enhancing destruction complex recruitment. Once Axin and Dvl are within close proximity, they are able to interact. I think this interaction blocks some of APC’s multiple Axin binding sites, thus destabilizing APC:Axin interaction. This destabilization of APC:Axin may also expose Axin’s TNKS binding domain, thus allowing TNKS to PARsylate Axin. During this time, Axin is also being de-phosphorylated by activated phosphatases. The changes in Axin stability within the complex and perhaps its change in charge due to post-translational modifications therefore cause Axin to “leave” the destruction complex condensate and thus increase the cytoplasmic pool of Axin. As the number of Axin molecules within the complex decrease, Axin’s avidity for βTrCP also decreases, thus preventing transfer of a phosphorylated βCat to the E3 ligase. Cytoplasmic levels of βCat rise and it moves into the nucleus. βCat then binds to the transcription factor families of TCF/LEF to induce transcription of Wnt target genes. A few hours after the initiation of Wnt signaling, TNKS

PARsylation and SIAH1 ubiquitination label Axin for proteasomal degradation.

173

REFERENCES

Aberle, H., Bauer, A., Stappert, J., Kispert, A., and Kemler, R. (1997). beta-catenin is a target for the ubiquitin-proteasome pathway. The EMBO journal 16, 3797-3804.

Angers, S., and Moon, R.T. (2009). Proximal events in Wnt signal transduction. Nature reviews Molecular cell biology 10, 468-477.

Banani, S.F., Lee, H.O., Hyman, A.A., and Rosen, M.K. (2017). Biomolecular condensates: organizers of cellular biochemistry. Nature Reviews Molecular Cell Biology 18, 285.

Basson, M.A. (2012). Signaling in cell differentiation and morphogenesis. Cold Spring Harbor perspectives in biology 4.

Boisvert, F.M., van Koningsbruggen, S., Navascues, J., and Lamond, A.I. (2007). The multifunctional nucleolus. Nature reviews Molecular cell biology 8, 574-585.

Carthew, R.W., and Rubin, G.M. (1990). seven in absentia, a gene required for specification of R7 cell fate in the Drosophila eye. Cell 63, 561-577.

CGAN, C.G.A.N. (2012). Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330-337.

Cliffe, A., Hamada, F., and Bienz, M. (2003). A role of Dishevelled in relocating Axin to the plasma membrane during wingless signaling. Current biology : CB 13, 960-966.

Cselenyi, C.S., Jernigan, K.K., Tahinci, E., Thorne, C.A., Lee, L.A., and Lee, E. (2008). LRP6 transduces a canonical Wnt signal independently of Axin degradation by inhibiting GSK3's phosphorylation of β-catenin. Proceedings of the National Academy of Sciences 105, 8032-8037.

Dajani, R., Fraser, E., Roe, S.M., Yeo, M., Good, V.M., Thompson, V., Dale, T.C., and Pearl, L.H. (2003). Structural basis for recruitment of glycogen synthase kinase 3β to the axin— APC scaffold complex. The EMBO journal 22, 494-501.

DeBruine, Z.J., Xu, H.E., and Melcher, K. (2017). Assembly and architecture of the Wnt/beta- catenin signalosome at the membrane. British journal of pharmacology 174, 4564-4574.

Fahmy, O.G., and Fahmy, M.J. (1959). Complementation Among the Sub-Genic Mutants in the r-Locus of Drosophila Melanogaster. Nature 184, 1927.

174

Feng, Y., Li, X., Ray, L., Song, H., Qu, J., Lin, S., and Lin, X. (2014). The Drosophila tankyrase regulates Wg signaling depending on the concentration of Daxin. Cellular signalling 26, 1717-1724.

Ferkey, D.M., and Kimelman, D. (2002). Glycogen synthase kinase-3 beta mutagenesis identifies a common binding domain for GBP and Axin. The Journal of biological chemistry 277, 16147-16152.

Fiedler, M., Mendoza-Topaz, C., Rutherford, T.J., Mieszczanek, J., and Bienz, M. (2011). Dishevelled interacts with the DIX domain polymerization interface of Axin to interfere with its function in down-regulating β-catenin. Proceedings of the National Academy of Sciences 108, 1937-1942.

Hamada, F., Tomoyasu, Y., Takatsu, Y., Nakamura, M., Nagai, S.-i., Suzuki, A., Fujita, F., Shibuya, H., Toyoshima, K., Ueno, N., et al. (1999). Negative Regulation of Wingless Signaling by D-Axin, a Drosophila Homolog of Axin. Science (New York, NY) 283, 1739-1742.

Hyman, A.A., Weber, C.A., and Jülicher, F. (2014). Liquid-Liquid Phase Separation in Biology. Annual Review of Cell and Developmental Biology 30, 39-58.

Jiang, J., and Struhl, G. (1998). Regulation of the Hedgehog and Wingless signalling pathways by the F-box/WD40-repeat protein Slimb. Nature 391, 493-496.

175

Klingensmith, J., Nusse, R., and Perrimon, N. (1994). The Drosophila segment polarity gene dishevelled encodes a novel protein required for response to the wingless signal. Genes & development 8, 118-130.

Lee, E., Salic, A., and Kirschner, M.W. (2001). Physiological regulation of [beta]-catenin stability by Tcf3 and CK1epsilon. The Journal of cell biology 154, 983-993.

Lee, E., Salic, A., Kruger, R., Heinrich, R., and Kirschner, M.W. (2003). The roles of APC and Axin derived from experimental and theoretical analysis of the Wnt pathway. PLoS biology 1, E10.

Li, A.F., Hsu, P.K., Tzao, C., Wang, Y.C., Hung, I.C., Huang, M.H., and Hsu, H.S. (2009). Reduced axin protein expression is associated with a poor prognosis in patients with squamous cell carcinoma of esophagus. Annals of surgical oncology 16, 2486-2493.

Li, L., Yuan, H., Weaver, C.D., Mao, J., Farr, G.H., Sussman, D.J., Jonkers, J., Kimelman, D., and Wu, D. (1999). Axin and Frat1 interact with Dvl and GSK, bridging Dvl to GSK in Wnt‐mediated regulation of LEF‐1. The EMBO journal 18, 4233-4240.

Liu, C., Kato, Y., Zhang, Z., Do, V.M., Yankner, B.A., and He, X. (1999a). beta-Trcp couples beta-catenin phosphorylation-degradation and regulates Xenopus axis formation. Proceedings of the National Academy of Sciences of the United States of America 96, 6273-6278.

176

Liu, C., Kato, Y., Zhang, Z., Do, V.M., Yankner, B.A., and He, X. (1999b). β-Trcp couples β- catenin phosphorylation-degradation and regulates Xenopus axis formation. Proceedings of the National Academy of Sciences 96, 6273-6278.

Liu, C., Li, Y., Semenov, M., Han, C., Baeg, G.H., Tan, Y., Zhang, Z., Lin, X., and He, X. (2002). Control of beta-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell 108, 837-847.

Loh, K.M., van Amerongen, R., and Nusse, R. (2016). Generating Cellular Diversity and Spatial Form: Wnt Signaling and the Evolution of Multicellular Animals. Developmental cell 38, 643-655.

MacDonald, B.T., and He, X. (2012). Frizzled and LRP5/6 receptors for Wnt/beta-catenin signaling. Cold Spring Harbor perspectives in biology 4.

Major, M.B., Camp, N.D., Berndt, J.D., Yi, X., Goldenberg, S.J., Hubbert, C., Biechele, T.L., Gingras, A.-C., Zheng, N., MacCoss, M.J., et al. (2007). Wilms Tumor Suppressor WTX Negatively Regulates WNT/ß-Catenin Signaling. Science (New York, NY) 316, 1043- 1046.

Mlodzik, M. (2016). Chapter Five - The Dishevelled Protein Family: Still Rather a Mystery After Over 20 Years of Molecular Studies. In Current Topics in Developmental Biology, P.M. Wassarman, ed. (Academic Press), pp. 75-91.

Perrimon, N., and Mahowald, A.P. (1987). Multiple functions of segment polarity genes in Drosophila. Developmental Biology 119, 587-600.

Piao, S., Lee, S.-H., Kim, H., Yum, S., Stamos, J.L., Xu, Y., Lee, S.-J., Lee, J., Oh, S., Han, J.- K., et al. (2008). Direct Inhibition of GSK3β by the Phosphorylated Cytoplasmic Domain of LRP6 in Wnt/β-Catenin Signaling. PloS one 3, e4046.

177

Pronobis, M.I., Deuitch, N., Posham, V., Mimori-Kiyosue, Y., Peifer, M., and Montell, D. (2016). Reconstituting regulation of the canonical Wnt pathway by engineering a minimal β-catenin destruction machine. Molecular biology of the cell 28, 41-53.

Roberts, D.M., Pronobis, M.I., Alexandre, K.M., Rogers, G.C., Poulton, J.S., Schneider, D.E., Jung, K.C., McKay, D.J., and Peifer, M. (2012). Defining components of the ss-catenin destruction complex and exploring its regulation and mechanisms of action during development. PloS one 7, e31284.

Rubinfeld, B., Tice, D.A., and Polakis, P. (2001). Axin-dependent phosphorylation of the adenomatous polyposis coli protein mediated by casein kinase 1epsilon. The Journal of biological chemistry 276, 39037-39045.

Schwarz-Romond, T., Metcalfe, C., and Bienz, M. (2007b). Dynamic recruitment of axin by Dishevelled protein assemblies. Journal of Cell Science 120, 2402-2412.

Sharma, M., Castro-Piedras, I., Simmons, G.E., and Pruitt, K. (2018). Dishevelled: A masterful conductor of complex Wnt signals. Cellular signalling 47, 52-64.

Shin, Y., and Brangwynne, C.P. (2017). Liquid phase condensation in cell physiology and disease. Science (New York, NY) 357.

178

Spink, K.E., Polakis, P., and Weis, W.I. (2000). Structural basis of the Axin–adenomatous polyposis coli interaction. The EMBO journal 19, 2270-2279.

Stamos, J.L., Chu, M.L., Enos, M.D., Shah, N., and Weis, W.I. (2014). Structural basis of GSK-3 inhibition by N-terminal phosphorylation and by the Wnt receptor LRP6. eLife 3, e01998.

Taelman, V.F., Dobrowolski, R., Plouhinec, J.-L., Fuentealba, L.C., Vorwald, P.P., Gumper, I., Sabatini, D.D., and De Robertis, E.M. (2010). Wnt Signaling Requires Sequestration of Glycogen Synthase Kinase 3 inside Multivesicular Endosomes. Cell 143, 1136-1148.

Tamai, K., Semenov, M., Kato, Y., Spokony, R., Liu, C., Katsuyama, Y., Hess, F., Saint-Jeannet, J.P., and He, X. (2000). LDL-receptor-related proteins in Wnt signal transduction. Nature 407, 530-535.

Tan, C.W., Gardiner, B.S., Hirokawa, Y., Layton, M.J., Smith, D.W., and Burgess, A.W. (2012). Wnt signalling pathway parameters for mammalian cells. PloS one 7, e31882.

Tolwinski, N.S., Wehrli, M., Rives, A., Erdeniz, N., DiNardo, S., and Wieschaus, E. (2003). Wg/Wnt Signal Can Be Transmitted through Arrow/LRP5,6 and Axin Independently of Zw3/Gsk3β Activity. Developmental cell 4, 407-418. van Amerongen, R., Nawijn, M., Franca-Koh, J., Zevenhoven, J., van der Gulden, H., Jonkers, J., and Berns, A. (2005). Frat is dispensable for canonical Wnt signaling in mammals. Genes & development 19, 425-430.

Wallingford, J.B., and Habas, R. (2005). The developmental biology of Dishevelled: an enigmatic protein governing cell fate and cell polarity. Development (Cambridge, England) 132, 4421-4436.

Wang, Z., Tacchelly-Benites, O., Yang, E., and Ahmed, Y. (2016a). Dual Roles for Membrane Association of Drosophila Axin in Wnt Signaling. PLoS genetics 12, e1006494.

179

Yanagawa, S., van Leeuwen, F., Wodarz, A., Klingensmith, J., and Nusse, R. (1995). The dishevelled protein is modified by wingless signaling in Drosophila. Genes & development 9, 1087-1097.

Zeng, X., Huang, H., Tamai, K., Zhang, X., Harada, Y., Yokota, C., Almeida, K., Wang, J., Doble, B., Woodgett, J., et al. (2008). Initiation of Wnt signaling: control of Wnt coreceptor Lrp6 phosphorylation/activation via frizzled, dishevelled and axin functions. Development (Cambridge, England) 135, 367-375.

180