(GDF8) Latency, Activation, and Antagonism

Molecular mechanisms of growth differentiation factor 8 (GDF8) latency, activation, and antagonism

A dissertation submitted to the University of Cincinnati in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

In the Department of Molecular Genetics, Biochemistry, and Microbiology of the College of medicine by

Jason C. McCoy B.S. Miami University August 2020

Committee Chair: Thomas Thompson, Ph.D.

Abstract

Growth differentiation factor 8 (GDF8), a.k.a. myostatin, is a member of the activin subclass within the larger TGFβ superfamily of signaling ligands that contains over 30 distinct members. Discovered in 1997,

GDF8 quickly became characterized as a negative regulator of muscle mass because a highly active GDF8 caused muscle wasting or atrophy. In contrast, when GDF8 is rendered inactive through extracellular inhibitors or mutations, massive muscle gain was observed. This discovery generated massive interest within the pharmaceutical industry to manufacture inhibitors for GDF8 to combat a variety of muscle wasting diseases. A better understanding of how GDF8 is regulated in vivo is imperative to develop efficacious and novel inhibitors against GDF8. The activity of GDF8 is tightly regulated in vivo by several different processes. TGFβ ligands are synthesized as large precursor proteins with a N-terminal signal sequence and prodomain followed by the mature signaling domain. The prodomain is cleaved from the mature domain, but unlike most other TGFβ family members, GDF8 forms a high affinity interaction with its prodomain rendering it inactive or latent. Prior to our work, the molecular mechanisms dictating latency were not well understood but were hypothesized to be like the latent TGFβ1 procomplex. Using the TGFβ1 procomplex as a model, we identified residues critical for a stable latent GDF8 procomplex.

Furthermore, the structure of the latent GDF8 procomplex was solved by another laboratory and our mutants could be further characterized to determine the molecular mechanisms of GDF8 latency. In order to signal, latent GDF8 needs to be activated by a member of the tolloid family of metalloproteases which proteolytically cleaves the prodomain. While the site of tolloid cleavage on the GDF8 prodomain has been identified, how tolloid recognized the prodomain as a substrate was unknown. Unlike other proteases, tolloid has no concrete consensus sequence required for cleavage. We sought to characterize what molecular features of the GDF8 cut site were required for tolloid processing. To this end, we identified several residues near the cut site that when mutated significantly reduced latent GDF8 activation by tolloid. Although GDF8 is tightly regulated through the formation of a latent procomplex, there are also

ii extracellular antagonists that will bind to GDF8 and prevent it from signaling. Of the antagonists that target GDF8, the WFIKKN family is by far the most specific. However, the driver of this specificity was not well defined. Here, we characterized the follistatin domain (FSD) of WFIKKN2. Previously, the FSD had been identified as a primary driver of the high affinity interaction between GDF8 and WFIKKN. We solved the crystal structure of the WFIKKN2 FSD and identified key residues for antagonism and characterized how the FSD of WFIKKN contributes to antagonism. Together, our in vitro data provide valuable insight into the molecular mechanisms dictating GDF8 regulation in vivo and has the potential to be leveraged by the pharmaceutical industry to develop novel approaches for GDF8 inhibition to combat muscle wasting.

iii

Dedication

This work was made possible due to many people:

To my parents, John and Judy McCoy. They have always pushed and supported me in every aspect of my life. Without them behind me I would not be where I am today. I cannot thank them enough for their love and compassion in every aspect of my life. I owe them more than I could ever repay.

To my sister Jennifer. Although at times we have not seen eye to eye we both know that no matter what we will be there for each other. I could not ask for a better role model and sister to look up to.

To the many friends I have made along the way. Graduate school is both mentally and emotionally challenging. I have made too many friends to list them all here. They kept me sane and lifted me up when

I was down. Thank you all. A special thanks to Erich Goebel and his wife Jessica Goebel. I was the best man at their wedding, and they will be the best man/maid of honor in mine. I have known them since day 1 of graduate school or longer and I could not ask for a better couple to spend these years and the future with.

To my fiancée, Victoria Jensen. Anyone who knows me knows that I sometimes float a little too much,

Victoria is my rock. She keeps me grounded, focused, and gives me the motivation to keep moving forward. Without her support this journey would have been infinitely harder. I cannot wait for our future together and to overcome every challenge in our way.

Acknowledgements

First and foremost, I would like to thank my mentor, Tom Thompson, for giving me an amazing opportunity within his laboratory. Having known many of his previous and now current graduate students one thing remains consistent, Tom will push you. Not over an edge, but to greater heights. He motivates his students to not only be better scientist, but writers through involvement in grants, communicators through lab meetings and conferences, and overall, more well-rounded members of the scientific community. Without his guidance and constant support, I would not be writing this today. What is more, while pushing me to be a better scientist, he gave me room. Room to grow, make mistakes, and learn from those many, many mistakes. He provided the tools, the guidance, and the motivation to be better and for that I cannot thank him enough.

I would also like to thank my committee members: Sean Davidson, Jeff Molkentin, and David

Wieczorek for their valuable insight, feedback, constructive criticism, and overall support during my academic pursuits. Two other professors deserve recognition alone: Rhett Kovall and Bill Miller. While

Tom was my primary mentor Rhett and Bill always supported me. Offering any expertise, reagents, or advice at any time. I cannot thank them enough for being another source of excellent mentorship during this journey.

The Thompson Lab is full of great scientists, friends, and peers: Erich Goebel, Greg Gipson, Kaitlin Hart,

Emily Kappes, Chandra Kattamuri, and Magda Czepnik. Every member of the current lab and two past members, Ryan Walker and Kristoff Nolan, have been integral to my success. Whether it is the exchange of ideas or the release of stress I owe them all a debt of gratitude. I would like to give a special thanks to

Magda, her and I have worked very closely on many experiments and projects. Her dedication and work ethic are unparalleled. I cannot thank her enough for helping me finish a multitude of experiments and projects.

Lastly, I would like to thank my soon-to-be wife Victoria. More than anyone else she has pushed me to where I am. We both went through graduate school together, supporting one another and succeeding together. Her strength is unrivaled, and I cannot wait for our future together.

vii

Table of Contents Abstract……………………………………………………………………………………………………………………………………………………ii Dedication…………………………………………………………………………………………………………………..………………………….iii Acknowledgements. …………………………………………………………………………………………………………………….………….v Table of Contents. ………………………………………………………………………………………………………………………………..viii List of Figures and Tables. ……………………………………………………………………………………………………………………...xi

Chapter I: Introduction………………………………………………………………………………………………………………………..1 Introduction………………………………………………………………………………………………………………………………..2 Ligand domain architecture and signaling…………………………………………………………………………………...3 TGFβ superfamily ligand prodomains facilitate proper complex formation and regulation………….6 Structure of the TGFβ1 procomplex and mechanisms of activation……………………………………………..7 GDF8/11 procomplexes and mechanisms of activation……………………………………………………………...10 Extracellular antagonism of the TGFβ superfamily……………………………………………………………………..11 Aims of this dissertation..14 Chapter II: Molecular characterization of latent GDF8……………………………………………………………………… 16 Abstract…………………………………………………………………………………………………………………………………… 17 Significance……………………………………………………………………………………………………………………………… 17 Introduction…………………………………………………………………………………………………………………………….. 18 Results…………………………………………………………………………………………………………………………………….. 19 -Prodomain-GDF8 can exist in a latent and active complex…………………………………………. 19 -SAXS analysis reveals conformational differences between GDF8 and other TGFβ prodomain-ligand complexes…………………………………………………………………………………….… 22 -Specific mutation within the prodomain enhance GDF8 activity………………………………… 25 -GDF8 prodomain mutations exhibit reduced antagonism………………………………………….. 31 -Reformed complexes using the GDF8 prodomain mutants are more active and exhibit decreased thermal stability…………………………………………………………………………………………. 32 -GDF8 mutants enhance muscle atrophy compared with WT GDF8……………………………. 34 Discussion……………………………………………………………………………………………………………………………….. 36 Materials and methods…………………………………………………………………………………………………..………. 42 Acknowledgements………………………………………………………………………………………………………………… 49

viii

Chapter III: Residues adjacent to the tolloid cut site of GDF8 and important for Tolloid mediated Activation………………………………………………………………………………………………………………………………………… 50 Abstract…………………………………………………………………………………………………………………………………… 51 Introduction…………………………………………………………………………………………………………………………….. 51 Results…………………………………………………………………………………………………………………………………….. 54 -Lucferase assays screening GDF8 alanine mutations………………………………………………….. 54 -Is the proteolytic astacin domain sufficient for latent GDF8 activation………………………. 57 -Direct comparison of astacin-domain mediated processing of GDF8 prodomain mutations……………………………………………………………………………………………………………………. 59 -Fully latent GDF8 mutatn processing by tolloid………………………………………………………….. 61 -GDF8 mutants defective in tolloid activation can suppress WT GDF8…………………………. 63 Discussion………………………………………………………………………………………………………………………………… 64 Materials and Methods……………………………………………………………………………………………………………. 68 Chapter IV: Crystal Structure of the WFIKKN2 follistatin domain reveals insight into how it inhibits GDF8……………………………………………………………………………………………………………………………………………….. 74 Abstract…………………………………………………………………………………………………………………………………… 75 Introduction…………………………………………………………………………………………………………………………….. 75 Results…………………………………………………………………………………………………………………………………….. 77 -Production of WFIKKN2 follistatin domain…………………………………………………………………. 77 -Comparison of WFIKKN2 FSD and WFIKKN2 full length protein………………………………….. 79 -WFIKKN2 binds GDF8 at the Type II receptor epitope………………………………………………… 81 -Crystallization and Structural determination of the WFIKKN2 FSD…………………………….. 83 -WFIKKN2 FSD mutagenesis and inhibition…………………………………………………………………. 84 -Comparison of the WFIKKN2 FSD to the FSDs of Follistatin……………………………………….. 85 Discussion……………………………………………………………………………………………………………………………….. 87 Materials and Methods……………………………………………………………………………………………………………. 92 Chapter-V: Summary and Future Directions……………………………………………………………………………………… 96

Dissertation summary……………………………………………………………………………………………………………… 97

Summary of prodomain mediated latency of GDF8…………………………………………………………………. 97

Future directions of prodomain mediated latency…………………………………………………………………… 97

Summary of the GDF8 tolloid cut site and their contribution to activation…………………………….. 101

Future directions to further understand GDF8 actviation………………………………………………………. 102

Summary of the WFIKKN2 follistatin domain and GDF8 antagonism……………………………………… 103

Future directions to determine molecular mechanisms of WFIKKN antagonism……………………. 104

Future directions for a translational impact: Moving beyond the biochemistry…………………….. 105

Using an in vitro assay to predict the location of GDF8 expression and subsequent activation in

vivo……………………………………………………………………………………………………………………………………….. 105

Final Conclusions…………………………………………………………………………………………………………………… 109

Bibliography………………………………………………………………………………………………………………………………….. 110

Appendix A: Abstracts from publications authored and co-authored…………………………………………….. 130

Appendix B: Curriculum Vitae………………………………………………………………………………………………………… 137

Figures and Table List

Chapter I

Figure 1. Schematic of GDF8 synthesis and processing

Figure 2. GDF8 structure and schematic of signaling

Figure 3. TGFβ procomplex structures

Figure 4. Follistatin antagonisms of Activin and GDF8

Figure 5. Chapter Overview

Chapter II

Figure 1. Activity and analysis of the latent GDF8 prodomain complex

Figure 2. SAXS analysis of latent, acid activated, and reformed GDF8 prodomain complexes

Table 1. Experimentally determined parameters from SAXS analysis of GDF8 prodomain

complexes

Figure 3. GDF8 activation by tolloid

Figure 4. Mutations within the GDF8 prodomain and activation by tolloid

Figure 5. Purification of bacterially produced GDF8 prodomains and mammalian-produced GDF8

prodomain complexes

Table 2. Calculated IC50 values for various mutant GDF8 prodomain constructs

Table 3. Calculated EC50 values for various mutant GDF8 prodomain constructs

Figure 6. Characterization of bacterially produced GDF8 prodomains and purified

prodomain:ligand complexes

Figure 7. Activation mutations in GDF8 increase in vivo activity

Figure 8. Relative protein expression following transient transfection of various GDF8 prodomain

mutant constructs

Figure 9. Crystal structure of the GDF8 prodomain complex

Figure 10. Comparison of the experimental SAXS scattering profile of various GDF8 prodomain

complexes to the GDF8 prodomain crystal structure

Chapter III

Figure 1. Latent GDF8 and Tll1 astacin domain architecture and structure

Table 1. Clustal O(1.2.4) multiple sequence alignment GDF8 residues adjacent to tolloid cut site

Figure 2. Transfection based luciferase assay of GDF8 alanine screen

Figure 3. Expression and cleavage test of transfected GDF8

Figure 4. Analysis of Tll1 astacin domain activity vs GDF8 prodomain states

Figure 5. Western blot analysis of GDF8 prodomain mutant processing by the astacin domain and

BMP1

Figure 6. Analysis of fully latent GDF8 mutants

Figure 7. Western blot analysis of latent GDF8 procomplex processing

Figure8. Transfection based luciferase assay of dominant negative GDF8

Chapter IV

xii

Figure 1. WFIKKN2 domain architecture and FSD purification

Figure 2. Binding and antagonism of WFIKKN2 FSD to GDF8

Table 1. Surface plasmon resonance of WFIKKN2 interaction with GDF8

Figure 3. Competitive biding between WFIKKN2 FSD and ActRIIB to GDF11

Figure 4. WFIKKN2 FSD structure and surface hydrophobicity

Table 2. X-ray crystallographic statistics for structural determination

Figure 5. Full length WFIKKN2 mutant selection and inhibition

Figure 6. Comparison of WFIKKN2 FSD and Fs288 FSD1-3

Table 3. Clustal O(1.2.4) multiple sequence alignment WFIKKN2

Figure 7. Brinding interface of WFIKKN2 FSD is distinct from the FSDs of Fs288

Chapter V

Figure 1. Transfection-based luciferase assay to determine dominant negative effects of GDF8 and

GDF11 mutants.

Figure 2. Cloning ring experimental design and preliminary data

xiii

Chapter I

Introduction*

*The following text is, in part, within the publication: Erich J. Goebel, Kaitlin N Hart, Jason C McCoy, and

Thomas B. Thompson. Structural biology of the TGFβ family. Mini Review. Experimental Biology and

Medicine December 2019. DOI: 10.1177/1535370219880894 PMID: 31594405 PMCID: PMC6920667, In addition to: Gregory R. Gipson, Erich J. Goebel, Kaitlin N. Hart, Emily C. Kappes, Chandramohan Kattamuri,

Jason C. McCoy, and Thomas Thompson. “Structural Perspective of BMP Ligands and Signaling.” Accepted to Bone, July 3, 2020. Ref. No.: BONE-D-20-00699R1

Introduction

Signaling within the body is a highly complex system that remains largely unknown. Broadly speaking, there are three major components involved in signaling. For the purpose of this dissertation, the (1) ligand is a protein that binds to and forms a complex with its (2) receptor protein. These interactions can be facilitated by (3) agonists or hindered by antagonists. Of all of the signaling proteins within the body, the Transforming Growth Factor β (TGFβ) superfamily represents one of the largest families and includes over 30 different members divided into 3 subclasses: the TGFβs, Activins, and bone morphogenic proteins (BMPs). The TGFβ superfamily is involved in many processes throughout life from embryogenesis, tissue homeostasis and even tumorigenesis. This dissertation focuses on two members of the activin subclass, growth differentiation factor 8 (GDF8) and growth differentiation factor 11

(GDF11), which have recently become recognized as important regulators of adult tissue homeostasis.

GDF8 (also known as myostatin) has been well characterized as the primary negative regulator of muscle mass1–5. GDF8 was discovered in 1997 after studying Belgian blue cattle. These cattle have significantly more muscle mass compared to other cattle breeds, a phenotype that is caused by the insertion of 11 nucleotides within the GDF8 gene1. This insertion causes a frameshift that prevents the majority of the mature signaling ligand to be produced, leading to their hyper muscular phenotype.

Further studies demonstrated that GDF8 knockout mice also exhibited a ~7-fold increase in skeletal muscle mass5, cementing GDF8’s role as a negative regulator of skeletal muscle. Along these same lines,

GDF8 has also been implicated as a negative regulator of adipogenesis or fat formation6,7.

Although GDF8 and GDF11 are 90% identical within their mature signaling domains, knockout mice exhibit diverging phenotypes. Rather than resulting in a hypermuscular phenotype, GDF11 knockout mice die less than 24hours after birth8. In addition, GDF11 knockout mice have additional vertebrae and ribs, suggesting GDF11 plays a role in axial/ventral patterning during embryogenesis8. Recently, GDF11

2 has also been implicated in improved neurogenesis following stroke and other regenerative processes

(but this remains highly debated within the field9–11). Interestingly, when mice are treated with inhibitors or antagonists that prevent both GDF8 and GDF11 signaling, increased muscle mass was observed12–17.

As such, there is significant interest in the development of anti-GDF8/11 inhibitors to treat a variety of muscle wasting diseases, such as cancer cachexia and muscular dystrophies15,18–22.

Ligand domain architecture and signaling

While the TGFβ superfamily contains over 30 members, there is a striking conservation in domain architecture and ligand structure. TGFβ superfamily ligands are all synthesized as pro-peptides containing an N- terminal signal sequence, prodomain, and a C- terminal mature signaling domain (Figure 1A).

The N-terminal prodomain is essential for the proper folding of the ligand and is cleaved from the mature domain by proprotein convertases such as furin23–28. TGFβ ligands are characterized by the presence of a cystine knot within the mature domain and an inter- disulfide bond with another mature domain forming a covalently linked dimer29,30. Figure1. Schematic of GDF8 synthesis and processing. A) Domain architecture of GDF8 including relative size of each domain (below). Arrows represent respective proteolytic processing sites. B) Currently, every subclass, the Activins, BMPs, Unprocessed GDF8 dimer with mature domains covalently linked by a disulfide bond shown as ‘S-S’. C) Schematic of the latent GDF8 and TGFβs have at least one mature ligand procomplex with tolloid cut sites denoted by red dotted line. D) Prodomain release from active mature GDF8 dimer following tolloid dimer structure solved using x-ray cleavage.

3 crystallography29,31–39. Structurally, ligand monomers are typically described as a hand, with an α-helical wrist, and two sets of anti-parallel β-strands resembling fingers (Figure 2). The cystine knot mentioned above occurs between finger 2 and 4 forming a ring where an additional disulfide bond between fingers

1 and 3 is formed. Ligand dimers are formed by an additional disulfide bond near the wrist helix that packs the wrist into the palm of the other ligand. The resulting ligand dimer resembles a propeller or butterfly shape with a concave and convex surface important for receptor binding (Figure 2a). While there are similarities in the overall shape of the mature dimer, there are important differences that are translated to differences in receptor recruitment and downstream signaling outcomes. For instance, the activin members contain an additional cysteine within an N-terminal segment that forms an additional disulfide bond within finger 1 of the ligand37. In contrast, BMP ligands lack the additional cysteine and subsequent disulfide bond33. Overall, different signaling outcomes by different ligands is extremely complex and involve a number of different factors ranging from overall fold to minute residues differences40.

Interestingly, despite containing over 30 different ligands, the TGFβ superfamily signals through a limited number of receptors consisting of only five type I and 7 type II serine threonine kinase receptors30,41,42. To signal, ligands engage two type I and two type II receptors through extensive hydrophobic interactions forming a heterotrimeric complex (Figure 2b)42–44. When brought into a complex the constitutively active type II receptors will phosphorylate and activate the type I receptors. Once the type I receptors are activated, they will subsequently phosphorylate and activate SMAD 1/5/7 or SMAD2/3 transcription factors (Figure 2b)43. Activated SMADs will form a complex with a co-SMAD (SMAD4) and be translocated to the nucleus for downstream gene regulation (Figure 2b)45. The SMADs activated and subsequent signaling outcomes are dependent on the receptor complex formed. As such, receptor utilization and sequence identity of the TGFβ superfamily ligands is used to divide them into the 3 subclasses: the TGFβs, BMPs and Activins30.

Figure 2. GDF8 structure and schematic of signaling. A) Representative ligand structure of GDF8. One monomer is shown in pale cyan and the other in blue. Disulfide bonds are represented by yellow sticks. The front of the ligand (left) denotes ligand wrist helix and fingers labeled 1-4. The dimer is rotated 90 degrees to show the top-down view reavealing the propeller shape with the convex type II and concave type I binding sites. B) Schematic of GDF8 signaling. First the ligand dimer coordinates type I and type II receptors. Upon recruitment the activated receptor complex activates the canonical SMAD pathway and non-canonical JNK, ERK and p38MAPK cascades. Following activation SMAD2/3 forms a complex with the co-SMAD, SMAD4, and is translocated to the nucleus for gene regulation.

As mentioned, GDF8 and GDF11 are members of the Activin subclass and utilize two type II receptors, ActRIIa and ActRIIb (Activin receptor II a and b), and three different type I receptors (Alk4, Alk5, and Alk73,39,46). Formation of the GDF8/11 receptor complex, like other activins, results in SMAD 2/3 transcription factor activation (Figure 2b)47,48. In addition, there are other non-canonical signaling pathways activated by GDF8 and GDF11, such as the JNK, P38MAPK, and ERK signaling cascades (Figure

2b)11,49–53, highlighting the importance of a tightly regulated signaling output by not only GDF8 and 11, but in addition to the other ligands within the TGFβ super family. As such, there are many mechanisms that have evolved for the spatial and temporal control of ligands.

TGFβ superfamily ligand prodomains facilitate proper complex formation and regulation

Due to the importance of the TGFβ superfamily in a wide range of biological activity, they are tightly regulated. As mentioned, TGFβ ligands are synthesized as larger precursor proteins that include an

N-terminal prodomain and a C-terminal mature signaling growth factor domain54. The prodomain facilitates proper folding of the ligands and is subsequently cleaved from the mature domain by furin proteases27,54,55. In most cases, the prodomain remains noncovalently bound to the growth factor in what is typically referred to as a procomplex27,55–60. The functional significance of the procomplex appears to be highly variable for different family members. Commonly, the procomplex can still signal where the prodomain is believed to be readily displaced by higher-affinity interactions with the receptors56,59–61.

Whether the prodomain is fully displaced or remains partially bound during signaling remains to be determined for individual family members.

In specific cases, the prodomain can render a ligand inactive, keeping it trapped in a latent procomplex. This prodomain-mediated latency is observed within the TGFβ class and the activin class members GDF8 and GDF1124,58,60,62–64. For these ligands to signal, different mechanisms are used to

6 liberate the ligand from the latent procomplex. For TGFβ1 and TGFβ3, the prodomain interacts with αvβ6 integrin and either latent TGFβ binding proteins (LTBPs) or glycoprotein-A repetitions predominant protein (GARP)27,64–68. Interactions with LTBP and GARP involve a direct disulfide bond, which anchors the procomplex to the extracellular matrix or the cell surface, respectively64,68–70. These larger latent complexes can interact with αvβ6 integrin, which can initiate a tensile force across the complex, releasing the ligand for signaling27,64,66,71. In contrast, GDF8 and GDF11 are activated by an additional proteolytic event mediated by the tolloid family of metalloproteases58,63,72,73. Processing of the prodomain by tolloid fragments the inhibitory prodomain, weakening its interactions with the growth factor57,72,74. Recently, structures of the TGFβ1 and GDF8 procomplexes, including higher order complexes of the TGFβ1 procomplex bound to integrin and GARP, have provided a better understanding of prodomain-mediated latency and activation55,57,67,69.

Structure of the TGFβ1 procomplex and mechanisms of activation

The first structure of latent TGFβ1 (prodomain bound to the growth factor) was solved in 2011 and revealed key features of the prodomain that contributed to latency, including the α1-helix, latency lasso, and fastener region55. These elements inhibit TGFβ signaling by blocking both the type I and type II receptor binding sites. The α1-helix of the prodomain lies perpendicular to the fingers of the ligand within the concave palm of the type I receptor interface (Figure 3). While only a handful of TGFβ ligands contain inhibitory prodomains, the majority of prodomains are predicted to contain an α1-helix, suggesting that this element plays a major role in procomplex formation across the family30. From the C-terminus of the

α1-helix, the prodomain wraps around the fingertips of the ligand to block TGFβ type II receptor binding, which is referred to as the latency lasso. Following the latency lasso, the prodomain traverses over the knuckles and back toward the C-terminus of the α1-helix. Here, the prodomain loops back to interact with

α1-helix, forming the “fastener”, which includes key hydrogen bonds between distant prodomain segments that appear to pinch the base of the fingers. Interestingly, mutations of fastener residues have

7 been shown to remove prodomain mediated latency39,55,57,75. Collectively, these inhibitory elements make up the “straitjacket” of the prodomain. In addition, the two prodomain chains continue away from the ligand in the same direction and join together through an inter-prodomain disulfide bond. The top of this structure is termed the “bowtie”, with the prodomains forming either a compact or closed conformation

(Figure 3a). The structure of the procomplex revealed that the consensus integrin binding motif – RGD - is surface exposed at the base of the bowtie distal from the inhibitory straitjacket components, indicating that long distance interactions are important for triggering TGFβ activation.

In 2017, the structure of the TGFβ1 procomplex bound to the αvβ6 integrin head was solved. In the lattice, only one chain of the TGFβ1 prodomain was bound to αvβ6 integrin, leaving one RGD motif

69 unoccupied . In the structure, the integrin head is positioned such that the β6 chain is proximal to the growth factor while the αv chain is positioned more distal, yet both the αv and β6 chains interact with the

RGD motif. Comparison to the unbound structure showed that significant structural differences occurred in the conformation of the RGD motif and surrounding area. Interestingly, conformational differences were also observed in the unbound prodomain, implying that integrin binding might have long-range effects that prime the procomplex for force-dependent release of the TGFβ1 ligand. Molecular dynamic simulations showed that a force applied through the β6 subunit of integrin, while artificially anchoring the

N-terminus of the prodomains, can disrupt the ligand:straightjacket interactions69. This study concluded that since the prodomains are covalently linked through a disulfide bond, a force applied through one prodomain monomer is sufficient to fully activate TGFβ1.

In order to gain insight on how TGFβ is anchored at the N-terminus of the prodomain, a recent

67 structure of TGFβ1 was solved in complex with GARP with the aid of a stabilizing antibody MHG-8FAB . In the immune system, the immunosuppressive activity of Tregs is enhanced by TGFβ1 and has become a recent therapeutic target for cancer immunotherapy74,76,77. GARP is a transmembrane protein expressed on the surface of Tregs and forms a disulfide bond with each TGFβ1 prodomain chain, tethering latent

TGFβ1 to the cell surface70,78,79. GARP contains leucine rich repeats that adopts a horseshoe-like structure similar to toll-like receptors. Within GARP, two unpaired cysteines, C211 and C350, are involved in linking both prodomain chains of latent TGFβ1 procomplex through the N-terminus of the α1-helix. Together, these studies highlight a TGFβ activation mechanism that requires both chains of the prodomain to be tethered to GARP, while a force is generated by a single molecule of integrin64,67,69. While GARP is bound to cell surface for TGFβ activation, TGFβ is also found tethered to the ECM through LTBPs64,65. Since LTBPs

Figure 3. TGFb procomplex structures. (a) TGFb1 procomplex in a closed conformation mediated by the bowtie. Ligand monomers are pale-cyan and blue, and prodomain monomers are orange and light orange with inhibitory elements labeled. (b) The TGFb1 procomplex bound to avb6 integrin with the RGD motif of the prodomain circled. (c) TGFb1 procomplex bound to GARP rotated 90? to reveal both disulfide bonds anchoring the two proteins. (d) The structure of GDF8 bound to its prodomain revealing the same inhibitory elements as TGFb1 but adopting an open conformation. (A color version of this figure is available in the online journal.

9 are structurally distinct from GARP, containing multiple domains, future studies are needed to determine if LTBPs have a similar mechanism of activation for the latent TGFβ procomplexes.

GDF8/11 procomplexes and mechanisms of activation

While the molecular mechanisms of TGFβ-1 activation have been extensively characterized, much remains unknown regarding latent GDF8 and GDF11 activation; however, it is clear that an additional cleavage event is required. It has been established that latent GDF8/11 procomplex requires an additional cleavage event within the prodomain by a member of the tolloid metalloprotease family (Figure

1c)58,63,73,80,81. Tolloids are members of the astacin super family, meaning they contain an N-terminal, zinc and calcium dependent protease72,82,83. All tolloid family members from Drosophila to humans exhibit a conserved domain architecture: an N-terminal prodomain, the proteolytic astacin domain, two tandem

CUB (Compliment/Uegf/BMP1) domains (CUB1/2), an EGF (epidermal growth factor) domain (EGF1), a third CUB domain (CUB3), a second EGF domain (EGF2), and two more tandem CUB domains (CUB4/5).

There are four family members: (1) mammalian tolloid (mTLD) and (2) its splice form lacking domains from

EGF2 to CUB5, (3) tolloid like 1 (Tll1) and (4) tolloid like 2 (Tll2). Importantly, all of these tolloid family members are able to activate latent GDF858,73,81. In addition to GDF8 and 11 activation, tolloids are also essential for proper extracellular matrix (ECM) formation by processing pro-collagen to collagen72,84–86.

Notably, tolloid is also needed during embryogenesis where it activates BMP bound to the antagonist chordin for proper axial patterning72,87–90.

Recently, the structure of the latent GDF8 procomplex was solved and provided valuable insight into latentcy and activation57. The GDF8 procomplex contains the same inhibitory elements described from TGFβ155,57,58 (see Chapter II). Similarly, mutations of the fastener residues resulted in decreased prodomain-mediated latency58. Interestingly, mutation of hydrophobic residues of GDF8 within the α1-

10 helix (specifically, L56 and L53) also removed latency, but were shown to be active without the need for tolloid proteolysis. These results will be discussed further in Chapter II58. However, despite sharing similar inhibitory elements, the TGFβ1 and GDF8 procomplexes form distinct conformations. Unlike TGFβ, the prodomain of GDF8 lacks an inter-disulfide bond, which allows for a more open V-like conformation as

Figure 4. Follistatin antagonism of Activin and GDF8. A) Top down view of an activin dimer, surface representation with monomers in pale-cyan and blue, bound to two monomers of follistatin (cartoon). Domains of one follistatin monomer are labeled, ND: N-terminal domain (gray), FSD1–3: follistatin domains 1–3, light blue, green and light orange. B) Top down view of a GDF8 dimer bound to the Fst spliceform, Fs288 with the same coloration and labeling as in (A). C) Top down view of a GDF8 dimer bound to Fstl3 with the same coloration and labeling as (A) and (B) compared to the closed conformation of TGFβ1 (Figure 3d). Interestingly, this open conformation exposes the tolloid cleavage site required for latent GDF8 activation. Further, studies using hydrogen-deuterium exchange revealed that processing by tolloid primes the prodomain for dissociation by weakening its interaction with the ligand73. These proteolytic fragments are easily displaced by the receptors allowing for GDF8 to signal. Currently, the mechanisms of latent GDF8 activation by tolloid, including recognition of the procomplex by tolloid and the spatial and temporal regulation of activation in-vivo remains under investigation and will be address in Chapter III.

Extracellular antagonism of the TGFβ superfamily

In addition to the prodomain, TGFβ family ligands are also regulated by a wide array of structurally diverse extracellular protein antagonists. Protein antagonists consist of varying domain architectures and inhibit ligands through different binding mechanisms23. The majority function by blocking both the two

11 type I and two type II receptor binding sites to prevent receptor assembly and signaling. Often antagonists accomplish this by binding in a 2:1 stoichiometric ratio (two antagonists, one ligand dimer), where one antagonist molecule will block a single type I and a single type II site. For example, the Follistatin family of antagonists inhibits numerous TGFβ superfamily ligands by wrapping around the ligand using 2 follistatin monomers, each blocking one type I and one type II receptor binding site (Figure 4). However, antagonists such as Chordin and WFIKKN2 (named for its domain architecture described below) are unique in that they each bind in an asymmetric 1:1 ratio91–93. GDF8 and GDF11 are primarily antagonized by two antagonist families: the Follistatin family and the WFIKKN family.

The first structure of a Follistatin family member in complex with a ligand was published in 2005 between Follistatin (Fst) and ActA (Figure 4)94. Fst is a multi-domain protein containing an N-terminal

Domain (ND) and 3 tandem follistatin domains (FSDs). Two molecules of Fst bind the ActA dimer in a ring- like structure. The ND of each Fst occupies the type I receptor binding site with FSD1 and FSD2 covering the type II receptor site (Figure 4). The two Fst molecules interact in a head-to-tail mechanism where the

ND of one Fst interacts with the FSD3 of the neighboring Fst. Several structures of Fst in complex with various ligands have highlighted a conformational selection model where both the antagonist and ligand adopt a preferred conformation for binding94–98. In particular, the ND of Fst can adopt different conformations to accommodate the wrist region of different ligands. This mechanism provides the basis for Fst’s ability to be promiscuous and target multiple ligands, including activin (ActA/B, GDF8/11) and

BMP (BMP2/4/6/7) class ligands. The structure of Fst bound to Activin A and GDF8 bound to the Fst spliceform, Fs288, are shown in Figure 4. Similarly, Follistatin-Like 3 (Fstl3), which lacks the third FSD, was solved in complex with both ActA and GDF8 (Figure 4)95,99. Here, the ND ligand interactions are less flexible, partially explaining why Fstl3 antagonizes fewer family members. Collectively, these structures have been important for understanding how an antagonist binds different ligands and how a particular ligand is neutralized by various antagonists.

WFIKKN is the other family responsible for GDF8 and 11 antagonism. WFIKKN contains two family members, WFIKKN1 and WFIKKN2, and is named for its conserved domain architecture: a whey acidic protein, follistatin domain, immunoglobulin domain, 2 tandem Kunitz domains and a netrin domain100–

10296-98. Unlike the Follistatin family, WFIKKN is extremely specific for only GDF8 and GDF11. Limited structural information is available for how WFIKKN1/2 bind and antagonize ligands. Currently, the Kunitz

2 domain of WFIKKN1 and the follistatin domain of WFIKKN2 (Chapter IV) have been solved by X-ray crystallography in the absence of ligand103,104. The Kunitz 2 domain resembles a typical Kunitz protease inhibitor, containing two anti-parallel β-strands, an adjacent α-helix and a flexible loop that is predicted to insert into target proteases, namely trypsin103. The structure of WFIKKN2 FSD was solved by our lab

(Chapter IV) and revealed a similar two domain architecture with an epidermal growth factor (EGF) and

Kazal subdomain that is conserved in other FSDs. However, the orientation of the WFIKKN2 FSD subdomains is unique from Fst FSDs 1-3, including FSD1 of Fst unbound to ligand and will be discussed further in Chapter IV104. Currently, the molecular mechanisms utilized by the multiple domains of WFIKKN to inhibit GDF8 and GDF11 is unknown, as no structure of the domains or full-length WFIKKN has been solved.

In general, TGFβ ligands, including GDF8/11, are highly regulated by both the prodomains they are synthesized with and a variety of extracellular protein antagonists. Over the years, structural work has highlighted major differences in the mechanism of binding. However, many questions remain, including how higher-order complexes with the extracellular matrix impact ligand activity. Further, only four procomplexes of family members are available, each describing vastly different structures. Additional structural data is needed to define these mechanisms for individual ligands and to further to understand how large multi-domain extracellular antagonists (such as WFIKKN) neutralize ligands. These studies will help support the growing need to develop therapeutics that modulate TGFβ ligands.

Figure 5. Chapter Overview. Each schematic depicts a gap, red ‘?’, targeted within the thesis regarding GDF8 regulation, including latency (Chapter II), activation (Chapter II) and antagonism (Chapter IV). The GDF8 ligand is represented by the blue dimer with one monomer in pale blue and the other in dark blue. The prodomain monomers are depicted as light brown and brown. Tolloid and WFIKKN are represented by an orange hexagon and red octagon, respectively.

Aims of this dissertation

GDF8 and GDF11 remain important therapeutic targets for combating muscle wasting. Given the relatively recent discovery, many facets of GDF8/11 regulation remain to be elucidated. The following aims describe the work in this dissertation that provide valuable insight into GDF8/11 regulation.

Aims Chapter II: Determining the molecular mechanisms of ligand prodomain mediated latency. As mentioned earlier, GDF8 and GDF11 form a high affinity interaction with their respective prodomains, leading to an inactive or latent procomplex. While it was hypothesized that GDF8/11 shared the same inhibitory elements as TGFβ1, there was no structure of the GDF8 or GDF11 procomplex. As such, further study needed to be conducted to determine what features of the prodomain led to latency. Therefore, our goal was to conduct a rigorous analysis of the latent GDF8 procomplex to determine which residues on the prodomain are important for latency. To this end, we constructed a number of prodomain mutants and assessed the ability of the prodomain to trap the GDF8 ligand in a latent state. Through this analysis we were able to identify mutations within the prodomain that completely removed the need for tolloid

14 activation. Effectively making a GDF8 prodomain that was not able to form a latent complex with GDF8 in vitro and in vivo. Together with the GDF8 procomplex structure, solved by the Marko Hyvonen lab57, we were able to identify the molecular mechanisms leading to prodomain mediated latency of GDF8.

Aims Chapter III: Determining if residues adjacent to the GDF8 tolloid cut site are important for activation. The latent GDF8 and GDF11 procomplexes require tolloid cleavage for activation. However, there is not consensus sequence for tolloid recognition and processing. As such, we sought to determine what residues, if any, adjacent to the tolloid cut site of GDF8 are important for activation. To this end, we generated sequential alanine mutations near the tolloid cut site of GDF8 and tested the ability of tolloid to cleave and activate the latent complex. From this, we were able to identify a cluster of N-terminal residues important for the ability of tolloid to activated latent GDF8. Together, these data provide valuable insight into tolloid recognition and activation of latent GDF8.

Aims Chapter IV: Determining how the WFIKKN2 follistatin domain contributes to GDF8 and 11 antagonism. The WFIKKN family is exquisitely specific to only GDF8 and GDF11. Previous research has shown the follistatin domain (FSD) of WFIKKN serves a pivotal role in antagonism, but the molecular mechanism that dictates this specificity is unknown. As such, we sought to structurally characterize the

WFIKKN2 FSD as it relates to GDF8 and GDF11 antagonism. We were able to identify a number of residues on the FSD that were important for the full length WFIKKN2 antagonist to inhibit GDF8. In addition, we solved the x-ray crystal structure of the WFIKKN2 FSD and found it did not resemble other well characterized FSDs involved in GDF8 antagonism. Together, these data displays that the FSD of WFIKKN2 utilizes unique molecular mechanism for GDF8 inhibition, offering insight into the exceptional specificity seen by WFIKKN members.

Chapter II

Molecular characterization of latent GDF8 *

*The following text is part of the publication: Ryan G. Walker, Jason C. McCoy#, Magdalena Czepnik,

Melanie J. Mills, Adam Hagg, Kelly L. Walton, Thomas R. Cotton, Marko Hyvönen, Richard T. Lee, Paul

Gregorevic, Craig A. Harrison, and Thomas B. Thompson. “Molecular Characterization of Latent GDF8

Reveals Mechanisms of Activation”. PNAS December 7, 2017. Doi: 10.1073/pnas.1714622115 PMID:

29348202 PMCID: PMC5798348. #Co-First Author

Abstract

Growth/differentiation factor 8 (GDF8) or myostatin negatively regulates muscle mass. GDF8 is held in a latent state through interactions with its N-terminal prodomain, much like TGF-β. Using a combination of small angle X-ray scattering and mutagenesis, we characterized the interactions of GDF8 with its prodomain. Our results show that the prodomain:GDF8 complex can exist in a fully latent state and an activated or ‘triggered’ state where the prodomain remains in complex with the mature domain.

However, these states are not reversible, indicating the latent GDF8 is ‘spring-loaded’. Structural analysis shows that the prodomain:GDF8 complex adopts an ‘open’ configuration, distinct from the latency state of TGF-β and more similar to the ‘open’ state of Activin A and BMP9 (non-latent complexes). We determined that GDF8 maintains similar features for latency, including the alpha-1 helix and fastener elements, and identified a series of mutations in the prodomain of GDF8 that alleviate latency, including

I56E, which does not require activation by the protease Tolloid. In vivo, active GDF8 variants were potent negative regulators of muscle mass, compared to wild-type GDF8. Collectively, these results help characterize the latency and activation mechanisms of GDF8.

Significance

GDF8 is a signaling protein that inhibits muscle mass. Inhibitors of GDF8 are highly sought as therapeutics for the treatment of muscle-wasting diseases. During synthesis, GDF8 is made as a precursor where the signaling segment is cleaved from the N-terminal prodomain, which remains associated and inhibits signaling. Activation involves an additional cleavage of the prodomain. We demonstrate GDF8 signaling could be gained through a conformational change where the prodomain remains associated with the signaling segment. Alteration of the prodomain can weaken the interactions causing GDF8 to signal, thus alleviating inhibition by the prodomain. This study illuminates how GDF8 transitions from an

17 inhibited state to an active state - information that will help to understand the mechanism of GDF8 signaling.

Introduction

One of the most thoroughly described negative regulators of skeletal muscle mass is the TGF-

β superfamily ligand growth/differentiation factor 8 (GDF8), also known as myostatin1,5. Genetic disruption of Gdf8 results in substantial skeletal muscle growth1,5. Further, a significant increase in muscle fiber size is also observed when adult animals are treated with agents that bioneutralize GDF8 (reviewed in105). As such, targeted inhibition of GDF8 is currently being pursued for the treatment of skeletal muscle- related disorders and associated symptoms106,107.

GDF8, like numerous TGF-β family members, is a disulfide-linked dimer that is synthesized as a precursor protein which requires cleavage by a furin-like protease to yield a N-terminal prodomain and a C-terminal mature, signaling domain3. Interestingly, for a number of TGF-β ligands the role of the prodomain extends beyond ligand maturation and folding support26,27, remaining non-covalently associated with the mature ligand following secretion in either a low-affinity, non-inhibitory or high- affinity, inhibitory fashion (reviewed in30). For example, the prodomains of TGF-β1, TGF-β2, TGF-β3,

GDF11, and GDF8 hold the mature ligand in a latent or inactive state mediated by a non-covalent, yet high affinity, ligand-specific interaction3,108–111 whereas mature Activin A and BMP9 remain associated with, but are not inhibited by, their prodomain56,112. Activation of TGF-β1 and TGF-β3 requires covalent interactions with the extracellular matrix and cellular contractile forces to release the mature ligand55,65,113. In fact, resolution of the latent TGF-β1 crystal structure provided a molecular explanation for how latency is exerted by the prodomain via a coordinated interaction between the N-terminal alpha helix (alpha-1), latency lasso, and fastener of the prodomain with type I and type II receptor epitopes of the mature domain55. On the other hand, GDF8 activation requires a second cleavage event within

18 the prodomain via proteases from the BMP1/Tolloid (TLD) family of metalloproteases111. However, the molecular and structural details of the GDF8 latent state have yet to be determined.

Based on sequence conservation and prior biochemical data describing the N-terminal portion of the GDF8 prodomain108, it is plausible that the molecular interactions and overall structure of the GDF8 latent complex may be similar to that of TGF-β1. However, the prodomains of a number of TGF-β family members share similar sequence conservation, yet they do not regulate the mature ligand in the same fashion and also exhibit significant structural diversity56,112. Therefore, while one might expect that GDF8 and TGF-β1 would share certain elements for how the prodomain binds and confers latency, it is possible that significant structural and molecular differences in these interactions occur as they exhibit profoundly different mechanisms of activation. However, this comparison is hindered by a lack of understanding of the GDF8 latent complex at the molecular level.

In this study, we utilized small angle X-ray scattering (SAXS) and mutagenesis to characterize the

GDF8 latent complex. Interestingly, SAXS analysis reveals that the GDF8 latent complex adopts a more

‘open’ conformation, similar to the overall structure of the BMP9 and Activin A prodomain complexes, which are not latent. The ‘open’ conformation of the GDF8 latent complex is in stark contrast to the

‘closed’ conformation adopted by the TGF-β1 latent complex. Furthermore, we identify key residues in the GDF8 prodomain that are responsible for promoting latency indicating that GDF8 and TGF-β1 share similar features for latency including a latency lasso. We further show that certain mutations in the prodomain of GDF8 can reduce latency, producing a more active ligand both in vitro and in vivo. Overall, our data provide insight toward the molecular mechanisms of GDF8 latency and activation.

Results

Prodomain-GDF8 can exist in a latent and active complex.

Initial characterization in adult mice showed that GDF8 is secreted into the systemic circulation as a latent protein complex that requires activation to trigger GDF8 signaling109. While the biological mechanism for activation remained unknown, it was shown that a GDF8-specific signal derived from the serum of a wild-type mouse, but not a Gdf8-/- mouse, could be detected following exposure to acidic conditions, referred to here as “acid-activation”109. The premise for acid-activation stemmed from a similar observation that was made during the characterization of TGF-β, which is similarly regulated by its prodomain114,115, and provided the initial basis that latent GDF8 and latent TGF-β are likely to be very similar in terms of activation and prodomain release.

While a molecular basis to describe how acid-activation alleviates ligand latency remains unknown, it is thought that the acidic-conditions simply dissociate the prodomain from the mature domain, thereby freeing the ligand from inhibition109. However, our initial attempts to purify the mature domain from the prodomain after acid-activation using an affinity column to the high-affinity antagonist, follistatin, failed, even though the complex exhibited significant activity. This observation suggested that perhaps the prodomain remained bound to the mature domain, but was not in a fully inhibitory state. To extend these initial observations, we isolated the mammalian-derived latent proGDF8 complex (GDF8L) and compared its signaling activity to both the acid-activated state (GDF8AA) and to the

apo mature, unbound GDF8 (GDF8 ) using a SMAD3-responsive (CAGA)12 luciferase reporter HEK293 cell line39,91,95,97,116. As expected and consistent with our previous report, GDF8apo readily signaled with a

39 L calculated half-maximal effective concentration (EC50) of 0.72 nM , whereas media containing GDF8 did not readily signal and required nearly 10,000-fold more protein to achieve a similar response as compared to GDF8apo (Figure 1a). In contrast, acid-activation of media containing GDF8L at pH 2 to generate

GDF8AA resulted in a significant gain in activity compared to non-acid activated latent GDF8 (Figure 1a).

AA Interestingly, the calculated EC50 for GDF8 (5.7 nM)

apo still did not reach the EC50 of GDF8 , suggesting that under these conditions we were unable to observe the full signaling potential of mature GDF8. Since

GDF8apo is stable and stored in 10 mM HCl, we do not expect this difference in activity to be caused by subjecting GDF8L to extreme conditions. We next evaluated the activation of GDF8L as a function of pH by subjecting the complex to various pH ranges (pH 2-

10) for 1 hour, followed by neutralization and

(CAGA)12 activation (Figure 1b). We determined that at the concentration tested (40 nM), the level of activation increases with a decrease in pH (Figure 1b), however substantial activation was observed Figure 1. Activity and analysis of the latent GDF8 prodomain complex. (a) HEK293 (CAGA)12 cells treated with latent throughout the pH range examined. The shape of the GDF8 prodomain complex (GDF8L; black), acid-activated (GDF8AA; orange), and free mature (GDF8apo; green) titration experiment suggests that multiple ionizable ligand. Experiments were performed at least twice with each data point measured in triplicate. Shown is a groups could be involved in the latency mechanism or representative experiment. Data were fit by nonlinear regression to a variable slope to determine the EC50.(b) Activity measurement following 1-h incubation of purified that shifts in the pH cause disruption in the structure GDF8L (40 nM) at the indicated pH, followed by neutralization before exogenous administration to HEK293 of the prodomain that effects its ability to inhibit (CAGA)12 cells. Data are shown as the mean ± SEM. (c) SEC analysis of GDF8L, GDF8AA, and reformed (GDF8R) GDF8. prodomain:ligand complexes. (Inset) The protein composition of the peak visualized by SDS/PAGE under nonreducing conditions. Given that titration of the GDF8 prodomain against mature GDF8 results in potent ligand inhibition108, we hypothesized that we were unable to recover the full signal from acid activated latent

GDF8 due to the possibility that the prodomain may still be able to provide some level of antagonism, through a non-covalent interaction, despite being acid activated. To test this hypothesis, we subjected

GDF8L and GDF8AA to size exclusion chromatography (SEC) followed by SDS-PAGE/Coomassie staining. In addition, we combined isolated GDF8 prodomain with GDF8apo, both components derived from mammalian cell expression, and applied the mixture to SEC. We determined that all three variations of

GDF8 complexes had similar retention volumes with both components (co-elution of the prodomain and

GDF8apo), indicative of complex formation (Figure 1c). This result supports the idea that during the acid activation, the prodomain can re-associate with the mature domain and partially inhibit signaling.

However, since the GDF8AA has significant activity, it also suggests that the latent interaction between the prodomain and mature domain is not completely reversible. Nevertheless, our finding that following exposure to acidic conditions, mature GDF8 remains associated with the prodomain, but in an active state, provided the opportunity for further comparison to latent GDF8.

SAXS analysis reveals conformational differences between GDF8 and other TGF-β prodomain-ligand complexes.

While the aforementioned data suggests that the latent and acid-activated forms of GDF8 could adopt different molecular states, limited structural information is available for the prodomain:GDF8 complex. Given that high-resolution structural information for prodomain:ligand complexes of latent TGF-

β155 and non-latent ligands, BMP9112 and Activin A56 have demonstrated a series of configurations that range in compactness, we next wanted to determine how the prodomain:GDF8 complex compared to these other prodomain:ligand complexes. Therefore, we used the solution-based technique small angle

Figure 2. SAXS analysis of latent, acid-activated, and reformed GDF8 prodomain complex. (a) SAXS scattering profile showing the intensity distribution and (b) the pairwise distribution function for the various GDF8 prodomain complexes. (d) The crystal structures of various prodomain:ligand complexes used to generate theoretical scattering profiles for comparison [TGF-β1, Protein Data Bank (PDB) ID code 3RJR (18); BMP9, PDB ID code 4YCG (14); ActA (Activin A), PDB ID code 5HLY (15)]. The chi (χ) value was determined using the FoXS webserver (26). Residuals for each comparison are shown below the scattering profiles. Note that the latent TGF-β1 structure exemplifies a closed conformation unlike the non- latent, but prodomain:ligand-associated BMP9 and Activin A (ActA) structures are in an open conformation. (d) Ab initio SAXS envelope (DAMFILT model) of the GDF8L (black), GDF8AA (orange), and GDF8R (gray) complexes. Note that the reconstruction of the GDF8AA complex appears more elongated compared with the other GDF8 prodomain complexes. The recently resolved GDF8 prodomain complex crystal structure [PDB ID code 5NTU (37)], shown in teal, is superimposed on the various ab initio molecular envelopes. X-ray scattering (SAXS) to analyze the purified prodomain:GDF8 complexes, including GDF8L, recombined and purified prodomain:GDF8 (GDF8R), and GDF8AA (Figure 2 and Table 1). Samples were well behaved

23 in solution and did not show evidence of interparticle repulsion or aggregation over multiple protein concentrations (Figure 2a and Table 1). From the Gunier analysis, we determined that GDF8L has a lower

Rg than GDF8AA, 41.1 ± 0.85 versus 46.8 ± 0.86 Å (Table 1), respectively, which suggests that acid activation of GDF8L altered the overall conformation of the complex. This is further supported by the appearance of a more ‘featured’ pairwise distribution plot (P(r)) for the GDF8AA complex compared to

GDF8L complex (Figure 2b). Additionally, we determined that the GDF8R complex had a similar scattering profile, pairwise distribution curve, and associated SAXS derived values as the GDF8L complex (Figure

2a, b and Table 1).

Table 1: Experimentally determined parameters from SAXS analysis of GDF8 prodomain complexes. Rg (Å) Sample Concentration I(0) (cm-1) Gunier Real Space Dmax (Å) Volume (Å3) Mass (kDa) (mg/mL) Native (GDF8L) 3 4,400 41.7 39.3 136 310,000 90.0 2 2,400 40.5 38.7 133 300,000 76.0

Acid Activated (GDF8AA) 1.3 1,600 47.6 43.0 148 340,000 77.0 1.15 1,700 47.0 41.6 147 350,000 70.0 1 1,500 45.9 40.0 134 340,000 59.0

Reformed (GDF8R) 1.6 190 40.0 39.0 131 270,000 89.0

Theoreticala proTGF-β1 NA NA 28.5 NA 92 NA 82.4 proBMP9 NA NA 37.8 NA 137 NA 90.2 proActivin A NA NA 31.6 NA 106 NA 90.1 proGDF8 NA NA 34.9 NA 120 NA 80.1 aValues derived from deposited PDB coordinates for each prodomain:ligand complex To determine if the GDF8L complex adopted a similar conformation to the other known prodomain:ligand structures (Figure 2c), we compared our experimental scattering profile to the theoretical scattering profile using FoXS (Figure 2d117). We first compared the experimental profile of

GDF8L to the theoretical profiles based on the prodomain:ligand structures of TGF-β1, BMP9, and Activin

A. This analysis showed that the overall structure of the GDF8L complex did not show substantial similarity to any structure as indicated by the calculated chi values (Figure 2d), which is also consistent with a larger Rg value than the other prodomain:ligand structures. Nevertheless, the most similarity was found with Activin A (c=4.65), which has an ‘open’ conformation, whereas the least similarity was found with

TGF-β1 (c=8.74; Figure 2d). Interestingly, both the GDF8AA and GDF8R complexes were more similar to

Activin A (c=1.96 and c=1.43, respectively) while still a poor fit with BMP9 (c=3.13 and c=3.35, respectively) and TGF-β1 (c=3.48 and c=3.04, respectively), suggesting that GDF8AA and GDF8R are also likely in an ‘open’ conformation and that there are likely additional differences in these complexes compared to the GDF8L complex. To extend these observations, we calculated the SAXS-derived ab initio molecular envelopes for each state. The overall shape of the envelopes for each state further supported our initial observation that structural differences likely exist between the activity states (Figure

2d). However, following superposition of the prodomain:GDF8 complex crystal structure, which was resolved during the preparation of this manuscript (see discussion), we observed that there are poorly defined regions within the envelopes, which may be the result of structural flexibility inherent to the

GDF8 prodomain complexes.

Specific mutations within the prodomain enhance GDF8 activity.

Our SAXS analysis revealed that the GDF8L complex likely adopts a different overall conformation compared to TGF-β1. Despite this, GDF8 and TGF-β1 share high sequence conservation in the N-terminal alpha-1 helix, latency lasso, alpha-2 helix, and fastener regions (Figure 4a). Thus, we hypothesized that these regions could interact with the mature GDF8 ligand and are important for forming the non-covalent interactions required for latency, such that removing these interactions might generate a more active

GDF8 ligand (i.e. remove latency). One might also expect that disruption of these interactions might disrupt folding, as observed for TGF-β1118. Therefore, to test our hypothesis, we utilized the TGF-β1 structure as a guide to systematically mutate specific residues in regions of the GDF8 prodomain and compared their activity to wild-type GDF8L. For this evaluation we developed a robust cell-based

(CAGA)12 luciferase-reporter assay where we could assess the variants through transient

Figure 3. GDF8 activation by Tolloid. (CAGA)12 HEK293 cells were co- transfected with varying amounts of GDF8 DNA within the pRK5 expression vector, 50 ng of Furin (human) and 50 ng of Tll2, Tll1, or BMP1 (human). Media was exchanged for serum free media 6 hours post transfection and luminescence was read 24 hours later. The resulting luminescence was then divided by the signal from cells co- transfected with 50ng of Furin, 50ng of Tll2, Tll1, or BMP1 and 25, 50, or 100ng of pRK5 empty vector to report Fold Activation. b) HEK293 (CAGA)12 were co-transfected with 25ng of the indicated GDF8 mutant in addition to 50ng Furin, 50ng of Tll2, Tll1, or BMP1 DNA, and 25ng of empty psF-IRES vector as a transfection control (Renilla). Duration of transfection and incubation prior to activity measurements was the same as in (Figure 3b). Luciferase signal was read first and then quenched followed by detection of Renilla. The luciferase signal for each well was normalized to the corresponding Renilla signal Following normalization, the background signal from empty vector transfected wells, similar to (Figure 3b), was used to divide all samples to report Fold Activation c) GDF8 mutants were cloned into the psF-IRES vector and HEK293 (CAGA)12 were co- transfected with 25ng or 50ng of DNA in addition to 50ng of Furin and 5 or 25ng of Tll2. As in (Figure 3b) and (c) media was exchanged after 6 hours and the plate read 24hours later. Signal was read and normalized as it was in (b). Error is shown as mean ± SEM. Bar graphs were compared using two-way ANOVA with Bonferroni correction against wild type (*P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001).

transfection. Our first goal was to determine which TLD family protease member (e.g.

BMP1/mTLD, tolloid-like 1 (Tll1) or tolloid-like 2 (Tll2)) yielded the most optimal activation of wild-type

GDF8L 81,111. Using a similar assay format as previously described39, we compared the activity of wild-type

GDF8 following transient co-transfection of wild-type GDF8, furin, and either BMP1, Tll1 or Tll2 using

HEK293 (CAGA)12 luciferase cells (Figure 3a). As predicted, we observed little to no signal when the TLDs

26 were not included in the assay, indicating that little to no basal TLD is present and incapable of activating

GDF8L (Figure 3a). However, when cells were co-transfected with DNA from one of the three TLDs, we observed a dose-dependent increase in signal with increasing concentrations of wild-type GDF8 DNA

(Figure 3a). As predicted, we observed differences in the fold activation of wild-type GDF8 when co- transfected with the various TLDs, where the highest activation resulted from Tll2

(Tll2>Tll1>BMP1; Figure 3a). Although this result is consistent with previous reports81,111, indicating that differences in the magnitude of activation by TLDs, we cannot rule out the possibility this increase in activity is due to differences in TLD protein expression levels or differential regulation of TLD maturation needed for activation119–121. Regardless, since Tll2 was the most effective activator of wild-type (WT) GDF8 with increases ranging from 20 to 60-fold activation, it was used in the remaining assays unless otherwise noted.

The panel of mutations is shown in Figure 4b and is categorized based on the anticipated location in the prodomain. Within these regions, we primarily focused our attention on mutation of hydrophobic residues, since hydrophobic interactions commonly drive known inhibitory interactions within the TGF-β family (reviewed in30). For example, GDF8 maintains a number of hydrophobic resides that are predicted to align to one side of the alpha-1 helix, similar to the register of TGF-β1. Of particular interest, we identified two hydrophobic residues in the alpha-1 helix, I53 and I56, which showed more than 2-fold higher activity compared to wild-type GDF8 when mutated to either an alanine (I53A, I56A) or glutamate

(I53E, I56E; Figure 4b). Additionally, mutation of residues outside of the alpha-1 helix, I77A within the latency lasso, and H112A within the fastener showed an increase in activity when compared to wild-type, whereas the mutants generated in the alpha-2 helix did not show any significant gain in activity compared to wild-type (Figure 4b). In contrast, Y94A resulted in little to no activity. As a control, we tested the activity of D99A, which has previously been shown to eliminate activation by TLD111. As expected, introduction of D99A abolished activity, supporting that the assay is specific to the plasmid carrying the

Figure 4. Mutations within the GDF8 prodomain and activation by Tolloid. a) Latent TGF-β1 structure highlighting inhibitory elements of the prodomain: alpha-1 helix (blue), latency lasso (cyan), and the fastener (magenta). Sequence alignment between TGF-β1, GDF8, and GDF11 are shown with residues investigated marked with an asterisk. Labels correspond to TGF-β1 with human GDF8 in parentheses. b) Transfection assay to determine GDF8 mutant activity. HEK293 (CAGA)12 cells were co- transfected with 25 ng of each GDF8 mutant, 50 ng of furin (human), and 25 ng of Tll2 (human) DNA. Fold activation was determined by dividing the signal from no GDF8 plasmid (25 ng of empty vector, 50 ng of furin, and 25 ng of Tll2). c) HEK293 (CAGA)12 cells were co-transfected with 25 ng of GDF8 mutant DNA, 50 ng of furin, 25 ng of empty psF-IRES, and 0, 5, or 25 ng of Tll2. The luciferase signal was normalized to Renilla. Fold activation was calculated similar to (b). d) Transfection assay to determine GDF11 mutant activity. HEK293 (CAGA)12 were co-transfected with 3 or 10 ng of the GDF11 mutant DNA and 50 ng of furin, with or without 50 ng of Tll2. Transfection, normalization, and fold activation were calculated as they were in c). All mentioned experiments were performed at least twice where individual points were measured in triplicate. Error is shown as mean ± SEM. Bar graphs were compared using one-way (b) or two-way (c and d) ANOVA with Bonferroni correction against WT (*P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001). 28

GDF8 gene. In addition, we tested the K153R mutant, which was previously shown to enhance furin processing but not influence activation by TLD. K153R had similar activity as wild-type, indicating that TLD processing is optimal122.

To validate these observations and perform a more rigorous cross-comparison between WT GDF8 and these mutants, we inserted the ligand DNA into the pSF-CMV-FMDV-Rluc vector, which allowed us to normalize our data for transfection efficiency. We focused on the I53A/E and I56A/E mutants within the alpha-1, as well as the Y111A and H112A mutants within the fastener region due to their apparent importance when examining the structure of TGF-β1. This approach was used because previous efforts to detect the secreted ligand in the conditioned medium in this assay format were unsuccessful, likely due to protein levels below the limit of detection. We determined that all mutants retain significantly higher activity than WT GDF8 in a dose-dependent fashion with respect to titration of ligand DNA and Tll2 DNA

(Figure 3b). Given that activation of WT GDF8 is differentially regulated by the various TLDs, we tested whether or not our mutants retained higher activity when activated by the other TLDs, Tll1 and

BMP1. Overall, our results indicated that our mutants were more active than WT GDF8, though there were a few differences in the activation across the various TLDs (Figure 3c). Except for the I53A and I56A variants of the Ile mutations co-transfected with Tll2, all mutants showed enhanced activity in the presence of either Tll2 or Tll1, while only the Y111A and H112A mutants showed enhanced activity when co-transfected with BMP1 (Figure 3c). These results were unexpected and likely suggest that the enhanced activity of our mutants may occur because of multiple mechanisms, such as whether or not TLD is still required for activation.

To determine if the enhanced ligand activity was dependent on TLD activity (i.e. TLD-dependent), we compared the activity of these mutants transfected with and without Tll2 (Figure 4c). Interestingly, of the mutants tested, the I56E mutant showed significant activity compared to WT GDF8 in the absence of Tll2 (Figure 4c), whereas the other mutants (I53A/E, Y111A, and H112A) required the presence of TLD

29 for enhanced activity (Figure 4c). To confirm that GDF8 with the I56E mutation is not dependent on TLD, we generated the double mutant, I56E/D99A, which would eliminate the potential for Tll2 activation. Similar to I56E, transfection of the I56E/D99A mutant showed enhanced activity thus demonstrating that the I56E mutation results in non-latent and active GDF8 ligand (Figure 4c). However, we did observe that co-transfection of Tll2 further enhanced the activity of the I56E mutant suggesting that more activity from this mutant can still be gained, but not in the presence of D99A. Thus, I56E has activity without the requirement of TLD, but TLD can further potentiate I56E’s activity. The I56A mutant did not show the same Tll2 independence as I56E (Figure 4c) suggesting that introduction of the charged residue may destabilize the interaction between the prodomain and the mature ligand, perhaps by disrupting a hydrophobic pocket or core.

GDF11 is a closely related ligand to GDF8 and is regulated in a similar fashion as GDF8 in terms of latency and the requirement of TLD processing to alleviate latency63. Therefore, we tested if mutation of similar residues in GDF11 as GDF8 would also enhance ligand activity (Figure 4a, d). To test this hypothesis, we assessed ligand activity following co-transfection of the HEK293 (CAGA)12 cells with and without Tll2 (Figure 4d). Similar to GDF8, WT GDF11 activity was increased when Tll2 was present. Furthermore, mutation of similar residues in the alpha-1 (L76E and I79E) and fastener (Y135A and H136A) regions in GDF11 significantly enhanced ligand activity. Specifically, GDF11 L76E showed significantly enhanced ligand activity compared to WT that was independent of Tll2 (Figure 4d).

Interestingly, unlike in GDF8 (Y111A), GDF11 Y135A also showed enhanced ligand activity compared to

WT GDF11 in the absence of Tll2 (Figure 4d) suggesting that there may be unique molecular contacts in this region, which may account for these differences. Overall, our results indicate that mutation of specific residues in the GDF11 prodomain can affect ligand latency and activity.

GDF8 prodomain mutations exhibit reduced antagonism.

As mentioned earlier, GDF8 mature ligand signaling can be antagonized by titrating increasing amounts of purified GDF8 prodomain. Therefore, we next wanted to determine if the prodomains with activating mutations had an altered capacity to inhibit the mature GDF8 ligand. To accomplish this, we produced and purified the

GDF8 prodomain mutants in bacteria (Figure 5a) and determined their half-maximal inhibitory potential

(IC50) against a constant concentration of mammalian derived, mature GDF8 (Figure 6a). To improve the production and solubility of the bacteria- derived GDF8 prodomain mutants, we mutated all four cysteines in the prodomain to serine (GDF84xCtoS; see Figure 5. Purification of bacterial-produced GDF8 prodomains and mammalian-produced GDF8 prodomain complexes. a) SDS gel of bacterially expressed and purified Materials and Methods). Using the SMAD3-responsive GDF84xCtoS prodomain mutants. b) GDF8 prodomain mutant complexes were purified from expi293F (CAGA)12 luciferase reporter HEK293 cell line described conditioned media using liquid chromatography. The coomassie stained SDS-PAGE gel depicts GDF8L, I56ENC, above, we determined the IC50 for several of the and H112ANC following purification using a Lentil Lectin Sepharose B column followed by SEC in non-reducing and activating prodomain mutations (Figure 6a; Table 2). reducing conditions. The prodomain is shown by the blue arrow and the mature domain by the green arrow. Protein was normalized to mature dimer under non-reducing Results show that mutations in the fastener region, conditions using to GDF8L. Note that all proteins were run on the same gel that is shown here. For clarity purposes, Y111A and H112A, had a similar IC50 to the space in between some lanes is due to removal of gel lanes between samples pertinent to this study. GDF84xCtoS whereas, mutations in the alpha-1 helix (I53A, I56A and I56E) were 3 to 4-fold less potent.

However, the most dramatic effect was observed with I56E, which was ~16-fold less potent than

GDF84xCtoS.

Table 2: Calculated IC50 values for various mutant GDF8 prodomain constructs.

Construct LogIC50 ± SEM (M) IC50 (nM) Log 95% CI (M) Fold over 4xCtoS 4xCtoS -8.73 ± 0.02 1.85 -8.69 to -8.77 1.00

GDF84xCtoS I53A -8.42 ± 0.01 3.77 -8.39 to -8.46 2.04

GDF84xCtoS I53E -7.55 ± 0.01 28.27 -7.42 to -7.68 15.28

GDF84xCtoS I56A -8.56 ± 0.01 2.78 -8.52 to -8.59 1.50

GDF84xCtoS I56E -8.47 ± 0.01 3.35 -8.30 to -8.64 1.81

GDF84xCtoS Y111A -8.52 ± 0.02 3.01 -8.47 to -8.58 1.63

GDF84xCtoS H112A -8.58 ± 0.06 2.03 -8.53 to -8.60 1.10

Table 3: Calculated EC50 values for various mutant GDF8 prodomain constructs.

apo Construct LogEC50 ± SEM (M) EC50 (nM) Log 95% CI (M) Fold over GDF8 GDF8apo -8.82 ± 0.03 1.53 -8.75 to -8.88 1.00

GDF8L NC NC NC NC

GDF8Native I56E -8.55 ± 0.06 2.79 -8.37 to -8.73 1.82

GDF8Native H112A -7.86 ± 0.06 13.67 -7.70 to -8.03 8.93

GDF84xCtoS -7.93 ± 0.02 10.62 -7.92 to -8.03 6.94

GDF84xCtoS I53A -9.23 ± 0.03 0.59 -8.90 to -9.57 0.39

GDF84xCtoS I56A -9.53 ± 0.06 0.30 -8.74 to -10.32 0.20

GDF84xCtoS Y111A -8.98 ± 0.09 1.05 -7.78 to -10.18 0.69

aNC=not calculable Reformed complexes using the GDF8 prodomain mutants are more active and exhibit decreased thermal stability.

We next wanted to determine if we could reform the prodomain:ligand complex using the various mutant prodomain constructs and subsequently assess their signaling activity. Therefore, we combined the mutated prodomains with the mature ligand and isolated the complex by SEC. We calculated the

Figure 6. Characterization of bacterially produced GDF8 prodomains and purified prodomain:ligand complexes. a) Representative IC50 curve of serially diluted bacterially expressed prodomains mixed with exogenous, mammalian- derived GDF8apo (0.62 nM) and added to HEK293 (CAGA)12 cells. Fraction activation was calculated using the signal of GDF8apo treated with the prodomain divided by GDF8apo alone or maximum signal. Data were fit to a nonlinear regression with variable slope to determine the IC50. b)EC50 curves of reformed GDF8 (mammalian-derived) prodomain (bacterial derived) complexes denoted with Rbac superscript. Data were fit by nonlinear regression to a variable slope to determine the EC50. c) Representative derivative plot of melt curves from 24 to 100 °C generated by thermal shift and reported as fluorescent units. d) Representative Tm (Co) for each mutant shown in c. All experiments were performed at least twice where individual points were measured in triplicate for (a) and (b) and duplicate for (c) and (d). All data are shown as mean ± SEM.

EC50 of the complexes using the HEK293 (CAGA)12 luciferase-reporter cells and compared these results to

GDF8L complex, GDF84xCtoS complex (referred to as GDF8Rbac), and GDF8apo (Figure 6b; Table 3). We were unable to isolate a stable complex using the prodomain mutations of I53E and I56E, presumably due to a loss in affinity for the mature GDF8 (see Figure 6a and Table 2). Interestingly, all reformed mutant complexes showed significant activity with EC50 values similar to the mature GDF8 indicating that the prodomain:ligand inhibitory complex was less stable during the assay and could not function to inhibit

GDF8 signaling. This is in contrast to the GDF8Rbac complex, which had significantly less activity, but still had more activity than GDF8L complex (Figure 6b; Table 3).

To further determine if the enhanced activity shown by the mutants, specifically the alpha-1 mutants (I53, I56), may be explained in part by destabilization of the prodomain:mature ligand complex, we performed a thermal shift assay. In this assay the binding of the hydrophobic dye Rox dye was measured as a function of temperature (Figure 6c, d). We determined that the mammalian-derived

L 4xCtoS Rbac GDF8 complex had the highest Tm whereas the reformed GDF8 complex (GDF8 ) and GDF8 I53A and I56A mutant complexes showed a lower Tm suggestive of diminished stability or differences in the binding mode compared to the GDF8L complex (Figure 6c, d). Both the GDF8L and mutant complexes showed increased stability when compared to GDF8apo and the unbound

4xCtoS apo GDF8 prodomain indicating that the difference in Tm is not due to excess GDF8 ligand within the sample or as a result of dissociated, unbound prodomain (Figure 6c, d). In addition to the higher

L Tm maxima for the GDF8 complex, we detected a second maxima at a lower temperature, not observed in the reformed complexes, suggesting that a complex destabilization event occurred for the

GDF8L complex (Figure 6c). Taken together, these data suggest that mutation of the residues within the alpha-1 helix alleviates GDF8 latency through disruption of the interaction between the prodomain and mature domain.

GDF8 mutants enhance muscle atrophy compared to wild-type GDF8

Having demonstrated that mutation of specific residues within the GDF8 prodomain result in a more active or less latent ligand in vitro, we next wanted to determine if the enhanced activity would be recapitulated in vivo in a model of skeletal muscle atrophy. To test this, we generated AAV vectors encoding either WT GDF8 or the activating GDF8 mutants, I56E and H112A, and locally injected them

Figure 7. Activating mutations in GDF8 increase in vivo activity. The right TA muscles of 6- to 8-wk-old male C57BL/6 mice were injected with AAV6 vectors encoding for GDF8, GDF8 (I56E), or GDF8 (H112A) (left TA muscles were injected with equivalent doses of an AAV6 vector lacking a transgene). a) Eight weeks after AAV6 injection, the TA muscles were harvested and weighed (n = 4–6, paired Student’s t test, data groups with different letters achieved significance, P < 0.05; *significantly different from WT GDF8, P < 0.05). b) Hematoxylin and eosin staining of TA muscles was performed on cryosections (scale bar, 100 μm) and c) muscle fiber diameter quantified (n = 3, paired Student’s t test, data groups with different letters achieved significance of P < 0.05, at least 150 myofibers were counted per TA muscle). d) Mammalian- derived (NC, native complex) and purified GDF8 prodomain complexes were serially titrated and exogenously added to HEK293 (CAGA)12 cells and reported as fraction activation compared with GDF8apo. Data shown are representative of one of three independent experiments that were performed with duplicate wells for each data point. Data were fit by nonlinear regression to a variable slope to determine the EC50. All data shown as mean ± SEM. in into the tibialis anterior (TA) muscles of 6-8-week-old male

C57Bl/6 mice. Eight weeks post-AAV injections, WT GDF8, which is secreted in a latent form, induced a modest (~7%) decrease in TA mass

(from 61.2 mg to 56.7 mg) (Figure 7a). In contrast, the decrease in muscle mass induced by GDF8 I56E

(from 53.2 mg to 40 mg) and GDF8 H112A (from 50.4 mg to 37.3 mg) was much greater (~25%) (Figure

7a), which is consistent with the in vitro finding that these mutations enhance the ligand activity

(see Figure 3; Figure 4). Histological analysis, using hematoxylin and eosin staining of GDF8-treated TA muscles, revealed that the decreased muscle mass was a product of muscle fiber atrophy (Figure 7b), as indicated by decreased fiber diameter (Figure 7c). GDF8 I56E and GDF8 H112A also provoked a significant endomysial cellular infiltration, which was not evident in WT GDF8-treated muscles (Figure 7b).

As we have shown previously with Activin A, these cells are likely collagen-secreting myofibroblasts123 and

35 their presence is indicative of enhanced GDF8 activity. Collectively, these data indicate that activating mutations in GDF8 markedly increase in vivo activity of this TGFβ superfamily ligand.

Given the apparent activity differences between GDF8L, GDF8R, and GDF8Rbac prodomain complexes, it is plausible that the natively produced mutant prodomain:GDF8 complexes, such as in the AAV experiments above, may show differences in their TLD dependence and signaling potential. Therefore, we next wanted to determine if the activity of mammalian produced mutant prodomain:GDF8 complexes were more active than the WT prodomain:GDF8 complex. Following expression and purification of the prodomain:ligand complexes, we determined the EC50 of the complexes using the HEK293

L (CAGA)12 luciferase-reporter cells and compared these results to GDF8 and

GDF8apo (Figure 7d; Figure 5b; Table 3). While not as active as GDF8apo, titration of the I56E complex showed significantly enhanced activity compared to both the H112A and WT GDF8L complexes and did not require the presence of TLD for enhanced activity (Figure 7d). The H112A complex showed minimal elevation in activity compared to WT GDF8L (Figure 7d). Together, these results support the notion that the muscle atrophy observed in our in vivo studies are likely the result of enhanced ligand activity.

Discussion

The goal of this study was to elucidate mechanisms of latent GDF8 activation and identify the residues within the prodomain that contribute to latency. Although many ligands have been shown to loosely associate with their prodomains, only the prodomains of TGF-β ligands, GDF8, and the highly- related GDF11, have been shown to potently inhibit their respective ligands (reviewed in30). Through sequence alignment and structural modeling, we hypothesized that, despite the different modes of activation, the GDF8 prodomain confers latency through a similar binding mechanism as observed in the latent TGF-β1 crystal structure55. Using the low-resolution solution-based technique, SAXS, we

36 demonstrated that the GDF8L complex exhibits an ‘open’ conformation, unlike the ‘closed’ conformation adopted by the latent TGF-β1 structure. This difference is not unexpected given the mechanistic differences required for their respective activation. It is possible that an ‘open’ conformation is required for TLD activation of GDF8L in order to improve accessibility of the TLD-cleavage site or TLD-recognition motif whereas a ‘closed’ conformation may impede access. However, through site directed mutagenesis of the GDF8 prodomain, based on sequence alignment to TGF-β1, we identified important residues within either the alpha-1 helix or fastener region, which when mutated, significantly enhance ligand signaling activity in vitro and in vivo. Together, our data supports the conclusion that the GDF8 and TGF-

β1 prodomains both utilize similar residues to confer latency yet, we have identified that significant overall structural differences exist between the two complexes.

Apart from biological mechanisms of activation, it has been shown that exposure of latent TGF-

β114,115 and GDF8L 39,109 to acidic conditions results in activation of the latent complexes. A molecular explanation for this mode of activation has yet to be determined but it has been postulated that ‘acid activation’ causes the prodomain and mature domain to dissociate, thus explaining the gain in ligand activity109. Interestingly, our biophysical data strongly support that acid activation of the GDF8L does not dissociate the complex but rather the pro- and mature domains remain associated, yet in a different molecular state, referred to as a ‘triggered’ state. Interestingly, the triggered state is not as active as

GDF8apo suggesting that the prodomain needs to be dissociated for full activity. This might be through partially interfering with receptor binding and is consistent with exogenous addition of prodomain to inhibit GDF8apo. Moreover, we determined that reconstitution of the GDF8 prodomain:ligand complex

(GDF8R) from individual components did not result in a fully latent complex as the GDF8R complex shows significant activity compared to the GDF8L complex, suggesting that the latent state and ‘triggered’ state are not fully reversible. The notion that the GDF8 prodomain:ligand complex may exist in multiple activity states may explain, in part, why bacterially-derived and refolded GDF8 prodomain:ligand complex has

37 been shown to have significant ligand activity124 and, therefore may better represent the acid-activated or ‘triggered’ state. Nonetheless, our findings raise the possibility that mature GDF8 may be held in a locked or ‘spring-loaded’ state by its prodomain following biosynthesis, which can be ‘triggered’ when exposed to changes in pH.

In order to extend our understanding of the molecular interactions that drive GDF8 latency, we performed a targeted mutagenesis on the

GDF8 prodomain, based on the latent TGF-β1 structure55 and corresponding sequence alignment.

Consistent with our hypothesis, we identified specific Figure 8. Relative protein expression following transient transfection of various GDF8 prodomain mutant constructs. Conditioned media from (CAGA) HEK293 cells was probed residues in the alpha-1 helix and the fastener region 12 for mature GDF8 under reducing conditions to confirm expression within our luciferase assay. that when mutated, resulted in a more active ligand compared to WT whereas mutation of hydrophobic residues in the latency lasso region did not increase activity. Importantly, our data suggests that the increase in activity was not due to increased protein expression (Figure 8). In fact, our most active mutant, I56E, showed the least detectable expression, perhaps due to rapid turnover of the mature ligand following receptor binding. Nonetheless, this observation is consistent with other groups that observed a reduction in ligand detection when corresponding residues were mutated in other TGF-β growth factors, though the effect of these mutations on TGF-β latency was not tested118,125 .

Due to the overall complexity of GDF8 biosynthesis, latency, and activation, we are unable to define the molecular mechanisms to describe or explain why these mutations enhance GDF8 activity.

Surprisingly, all activating GDF8 mutants required the presence of TLD except the I56E mutant, which remained significantly active despite the incorporation of the TLD cleavage-resistant mutation, D99A

(GDF8 I56E/D99A111). It is possible that incorporation of the I56E mutation disrupts the interaction

38 between the alpha-1 helix and the mature ligand, which allows competition with GDF8 receptors. On the other hand, mutation of these regions may prevent GDF8 from fully entering the latent or ‘spring-loaded’ state during biosynthesis. Instead, this variant may be secreted in a form similar to the ‘triggered’ state that we have identified. This idea is supported by our data showing that the recombined mutant

GDF8 prodomain:ligand complexes had similar signaling activity as GDF8apo. It is clear that further characterization of the mutant GDF8 prodomain:ligand complexes is necessary to pinpoint the molecular mechanism responsible for enhanced activity. However, we have identified specific residues within the GDF8 prodomain that can be modified to alleviate ligand latency without disrupting the function of the prodomain in folding and biosynthesis of the mature ligand26,27.

We extended our analysis of the GDF8 prodomain activating mutants in vivo using a model of skeletal muscle atrophy to determine if these mutants recapitulated our in vitro experiments. We focused our efforts on I56E, which showed the greatest activity independent of TLD and H112A where increased activity is completely dependent on TLD. In both cases, AAV delivery of I56E or H112A decreased the size of the muscle fibers relative to control mice and mice that received AAV encoding WT GDF8. Similar to our in vitro experiments, we were unable to detect evidence of mature GDF8 in the muscle of mice that received the AAV encoding the I56E mutant whereas we could reliably detect the mature ligand in the muscle from mice that received either the AAV encoding the H112A or WT proteins. As mentioned above, we speculate that loss of latency independent of TLD may enhance ligand turnover rate, thus making it

39 challenging to detect the mature ligand. Together, these results are consistent with our previous observation that mutation of these residues results in a more active ligand.

Figure 9. Crystal structure of the GDF8 prodomain complex. The GDF8 prodomain complex containing the mature dimer (gray and pale green), alpha-1 helix (blue), latency lasso (cyan), and fastener (magenta). (Middle Inset) Depiction of the alpha-1 helix and fastener regions following a 75° rotation about the y axis (Middle) Note the location of residues I53, I56, I60, and I64 within the alpha-1 helix and Y111 and H112 within the fastener regions. See also Fig. 10 While this manuscript was in preparation, it became apparent that an X-ray crystal structure had been determined in the laboratory of Marko Hyvonen (Figure 9, Figure 1057). Therefore, we wanted to compare how our low-resolution SAXS data compared to the overall shape of the

GDF8 prodomain:ligand complex. Consistent with our initial SAXS-based observations, the crystal structure of the GDF8 prodomain:ligand adopts a more open conformation that is drastically different from that of TGF-β1 and more similar to that of Activin A or BMP9. In agreement with our hypothesis that

GDF8 prodomain contains similar inhibitory elements comparable to TGF-β1, the alpha-1 helix, latency lasso, and fastener features are all present in the GDF8 prodomain:ligand complex. However, the conformation of the GDF8 prodomains in relation to their mature domain with which monomer they interact is significantly different than that of TGF-β1. For instance, the prodomain of one TGF-β1 monomer sits atop the other monomer of the homodimer, with all inhibitory elements imposed by one prodomain. However, the prodomain of one GDF8 monomer crosses over to interact with both mature domains of the dimer (Figure 9). Notably, the alpha-1 helix and latency lasso inhibit the GDF8 monomer from the same chain while the fastener interacts with the adjacent monomer. The significance of this binding strategy on inhibition is unknown. However, this ‘fastener-swap’ may play a role to ensure

Figure 10. Comparison of the experimental SAXS scattering profile of various GDF8 prodomain complexes to the GDF8 prodomain crystal structure. The panels show the theoretical scattering profile derived from the GDF8 prodomain complex crystal structure (cyan) compared to our experimental SAXS data on the various GDF8 prodomain complexes (GDF8L: top; GDF8AA: middle; GDF8R: bottom). The chi (Χ) value, determined using FoXS webserver, for each comparison is shown adjacent to each scattering profile. The residuals for each comparison are shown below scattering profiles. homodimer formation and/or aid in exposure of the TLD protease site. Nevertheless,

I53 and I56 in the alpha-1 helix are shown to interact directly with the GDF8 ligand. It is possible that mutation of I53 or I56 would destabilize the alpha-1 helix and disrupt binding of the prodomain to GDF8. One would also expect that mutation of I60 would show a similar, if not more, exaggerated phenotype as the I53 or I56 mutants. However, mutation of I60 did not result in enhanced activity, but rather even lower activity than WT GDF8. It is possible that I60 may be important for protein folding and loss of this residue is detrimental to this process. Furthermore, the GDF8 prodomain:ligand crystal structure supports our finding that mutation of the fastener residues, Y111 and H112, would destabilize the fastener- interaction with the alpha-1 helix. This is similar to TGF-β1 where mutation of the fastener residues created a more active

TGF-β1 ligand55. Taken together, our mutational analysis of the

GDF8 prodomain is highly consistent with the structure of the prodomain:GDF8 complex and also consistent with previous truncation analysis75,124,126,127. Our results are also consistent with results

41 from the laboratory of Tim Springer who performed a rigorous hydrogen-deuterium exchange followed by MS to map the interactions of the prodomain with the mature in solution73.

In summary, we determined that the latent GDF8 prodomain:ligand complex adopts a more

‘open’ structural conformation unlike that of the TGF-β1 latent complex (reviewed in30). Interestingly, both ligands share commonality with respect to the alpha-1 and latency lasso inhibitory elements, but show significant divergence with respect to the coordination of their respective fastener regions to confer latency. While it is unknown how this binding mode impacts or confers latency to GDF8 compared to TGF- β1, our data strongly supports the notion that the GDF8 prodomain:ligand complex can exist in multiple conformational states which ultimately dictate ligand activity and that the interactions between the prodomain and mature domain can be modified to generate a less latent and more active signaling ligand. It is plausible that GDF8 circulates within serum110 in these various conformational ‘activity’ states, thus making it tempting to speculate that GDF8 biological regulation may include shifts in the balance of these ‘activity’ states depending on the physiological context.

Materials and Methods

HEK293-(CAGA)12 luciferase-reporter assay

Luciferase reporter assays for activation and inhibition were performed as previously described39,91,95,97,116. Briefly, HEK293 (CAGA)12 cells (from RRID: CVCL_0045) stably transfected with plasmid containing Firefly luciferase reporter gene under the control of SMAD3-responsive promoter were seeded in growth media at 20,000 cells per well in a 96-well poly-D-lysine coated flat-bottom plate

(Cat. No. 655940 Greiner Bio-One GmbH, Germany) and incubated at 37°C / 5% CO2 until 75-85% confluent. For transient expression experiments, 200 ng total DNA in a final volume of 25 μL (25-75 ng ligand DNA, 50 ng full length human furin in pcDNA4, 5-50 ng of appropriate TLD DNA in pRK5 or pcDNA3, filled to 200 ng with empty vector) per well was added directly to the growth media, incubated for 6 h

42 and exchanged into serum-free media. OPTI-MEM reduced serum media (31985-070, Gibco, Life

Technologies, USA) and TransIT-LT1 Reagent (MIR 2300, Mirus Bio LLC, USA) were utilized for transfection according to manufacturer instructions. Cells were lysed 30 h post-transfection using 20 μL per well 1x

Passive Lysis Buffer (E1941, Promega, USA), on a plate shaker (800 rpm, 20 min, 20°C). The lysates were transferred to opaque black and white 96 well plates, 40 μL of LAR (E1501 and E1960, Promega, USA) was added, Firefly luminescence was recorded on Synergy H1 Hybrid Plate Reader (BioTek). When necessary, subsequent addition of 40 μL of Stop&Glo substrate (E1960, Promega, USA) was added and

Renilla luminescence was recorded. To determine EC50 and IC50 values, the growth media was removed and the appropriate dilutions of either ligand alone or with antagonist, respectively, were serially titrated, and added to the cells in a 100 L total volume of serum-free media. Luminescence was recorded as mentioned 18-24 h post ligand or antagonist addition. Experiments were independently performed at least 2 times and all data points were performed in triplicate. The EC50 and IC50 values were derived from non-linear regression with variable slope using GraphPad Prism 5 software. The EC50 and IC50 mean and standard error was calculated for each experiment and the mean weighted to the standard error was calculated using the following formulas, where ‘a’ is the standard error of the EC50 or IC50 determination, etc128:

푨 푩 ( )+( )+⋯ 풂ퟐ 풃ퟐ Weighted mean= ퟏ ퟏ (1a) ( )+( )+⋯ 풂ퟐ 풃ퟐ and

ퟏ ퟏ ( )+( )+⋯ 풂 풃 Weighted error= ퟏ ퟏ (1b) ( )+( )+⋯ 풂ퟐ 풃ퟐ

Production and purification of GDF8 prodomain from E. coli.

The prodomain of human GDF8 (residues 24-262) was cloned into a modified pET28a expression vector that contains an N-terminal 6x histidine tag, maltose binding protein (MBP) containing the mutations

D82A/K83A/E172A/N173A/K239A for surface entropy reduction (Moon AG, Mueller GA, a synergistic approach), and a HRV-3C protease cleavage site (6xHis-MBP-HRV3C cleavage site-GDF8 (residues 24-

262)). The cysteine residues in the human GDF8 prodomain (C39/C41/C137/C138) were mutated to serine to improve expression and solubility and were shown to form a stable complex with mature GDF8 similar to mammalian derived GDF8 prodomain. E.coli Rosetta (DE3) strain carrying the appropriate prodomain construct was grown at 37°C, 220 rpm until an optical density (OD) of 0.8 at 600 nm was achieved, followed by cold induction with 0.5 mM IPTG, addition of 2% ethanol, and incubation at 20°C overnight.

Cells were lysed and soluble 6xHis-MBP-GDF8 prodomain was applied to nickel affinity column (GE

Lifesciences) equilibrated in 20 mM Tris pH 7.4, 500 mM NaCl followed by elution with a linear gradient using 20 mM Tris pH 7.4, 500 mM NaCl, 500 mM imidazole over 5 column volumes. The eluted protein was then dialyzed into 20 mM Tris pH 7.4, 500 mM NaCl and HRV-3C protease was added and incubated for 24 h to remove the 6xHis-MBP-fusion protein. Following cleavage, the protein was dialyzed into 10 mM HCl and applied to a C4 reverse phase column (Sepax) equilibrated in 0.1 % TFA, 5 % acetonitrile and eluted with a linear gradient to 0.1 % TFA, 95 % acetonitrile over 30 column volumes. The fractions containing GDF8 prodomain protein were pooled and buffer exchanged into 10 mM HCl for storage at -

80° C for future use.

Mammalian derived latent GDF8 complex (GDF8L) and mutant complexes.

Chinese hamster ovary (CHO) cells stably producing GDF8 were used as previously described39,91,112,129,130. Conditioned media containing GDF8 was concentrated ~10-fold and applied to a Lentil Lectin Sepharose 4B (Amersham Biosciences) column. The protein was then applied to an S200 size exclusion column (Pharmacia Biotech). Molarity of the GDF8 latent complex was determined

44 as previously described, using SDS-PAGE/Coomassie staining and the quantified GDF8 mature as a standard39.

For the expression of mutant prodomain:GDF8 complexes, expi293 cells (Life Technologies) were transiently co-transfected with the mutant DNA and furin DNA. Conditioned medium was applied to a

Lentil Lectin Sepharose 4B (Amersham Biosciences) column and the protein was applied to an SRT-

SEC300 size exclusion column (Sepax). Molarity of the mutant complexes was determined as previously described, using SDS-PAGE/Coomassie staining and normalization to the mature dimer under non- reducing conditions39.

Acid Activation

Acid activation of GDF8 complex was performed as previously described39,109. In short, GDF8 complex was acidified to pH 2 - 7 using 1M HCl and incubating for 1 h followed by neutralization with 1M

NaOH back to pH 8. Conversely, when a pH>8 was required 1M NaOH was used which was neutralized accordingly with 1M HCl. This material was then used in luciferase, and SAXS analysis.

Small Angel X-ray Scattering (SAXS)

SAXS data was collected using SIBYLS mail-in SAXS service (Berkley, CA). GDF8 latent complex was purified as described above with the exception that the protein was reapplied to a Phenomenex HPLC

S2000 size exclusion column. Generation of the reformed GDF8 complex required separation of mature

GDF8 from the prodomain using previously described methods39,91 (24, 25). Briefly, following purification of the GDF8L complex, the complex was adjusted to 4 M guanidinium hydrochloride, 0.1 % TFA and applied to a C4 reverse phase column (Sepax) to yield the individuals components. The two proteins were then

45 mixed together with an excess molar ratio of prodomain to mature ligand dimer (2.25 prodomain:1 ligand dimer) and then applied to a Phenomenex HPLC S2000 size exclusion column as described above. SAXS data were collected on purified at least 2 concentrations of GDF8L, GDF8AA, and GDF8R in 20 mM HEPES pH 7.4, 500 mM NaCl, 1 mM EDTA, 2 % glycerol at 10°C. Exposures exhibiting radiation damage were discarded. Buffer matched controls were used for buffer subtraction. ScÅtter (SIBYLS) and the ATSAS program suite (EMBL) were used for data analysis. Comparison of the experimental scattering profiles to known crystal structures was performed using the FoXS webserver117. Ab initio molecular envelopes were calculated using DAMMIN (ATSAS, EMBL131), averaged using DAMAVER (ATSAS, EMBL132), and filtered using DAMFILT (ATSAS, EMBL). SUPCOMB (ATSAS, EMBL133) was used to superimpose the crystal structure of the latent GDF8 protein complex57.

Western analysis

To test protein expression following transfection, 500,000 HEK293 (CAGA)12 cells mentioned above (from RRID: CVCL_0045) were plated in a 6-well plate coated with poly-D lysine and incubated at

37oC until 75-85% confluency. A mixture of 625ng of GDF8 DNA, 1.25 μg of Furin, and 3.125 μg of pRK5

EV was used totaling 5 μg of DNA, ~25x the DNA used in a 96-well in order to closely mimic conditions within our luciferase assay. OPTI-MEM reduced serum media (31985-070, Gibco, Life Technologies, USA) and TransIT-LT1 Reagent (MIR 2300, Mirus Bio LLC, USA) were utilized for transfection according to manufacturer instructions. 12 h post-transfection media was removed and replaced with serum-free media. 30 h post transfection media was removed and concentrated 25x and run under reducing conditions on an SDS-PAGE gel. Standard western protocols were utilized and the anti-GDF8 antibody from RnD Biosystems (AF788) was used as described by the manufacturer. Western blot was developed

46 using the SuperSignal West Pico detection reagent (ThermoFisher) per manufacture instructions and detected using the C-DiGit blot scanner (LI-COR).

Protein Thermal Shift

Protein thermal shift assays were conducted using an OneStep real-time PCR system (Applied

Biosystems), run by the StepOne Software v2.3, as described by the manufacturer. In short, 1 g of protein was placed in 20 μL of 20 mM HEPES pH7.4, 500 mM NaCl in the presence of 1x ROX reagent from the

TM Protein Thermal Shift Dye Kit (Applied Biosystems). The melting temperature and Tm of each protein was conducted on a 1% gradient from 25°C-100°C taking approximately 40 min. Data was analyzed using

Protein thermal shift software v1.3, and curves were plotted from triplicate measurements.

Production of AAV vectors

The cDNA constructs encoding for WT GDF8, GDF8 I56E and GDF8 H112A were cloned into an AAV expression plasmid consisting of a CMV promoter/enhancer and SV40 poly-A region flanked by AAV2 terminal repeats. These AAV plasmids were co-transfected with pDGM6 packaging plasmid into HEK293 cells to generate type-6 pseudotyped viral vectors. Briefly, HEK293 cells were seeded onto culture and were transfected with a vector-genome-containing plasmid and the helper plasmid pDGM6 by calcium phosphate precipitation. After 72 h, the media and cells were collected and subjected to three cycles of freeze-thaw followed by 0.22 µm clarification (Millipore). Vectors were purified from the clarified lysate by affinity chromatography using heparin columns (HiTrapTM, GE Healthcare), the eluent was ultra- centrifuged overnight, and the vector-enriched pellet was re-suspended in sterile physiological Ringer's solution and quantified with a customized sequence-specific quantitative PCR-based reaction (Life

Technologies).

Administration of AAV6 vectors to mice

All experiments were conducted in accordance with the relevant code of practice for the care and use of animals for scientific purposes (National Health & Medical Council of Australia, 2016). Vectors carrying transgenes of GDF8 mutants were injected into the right tibialis anterior (TA) muscle of 6-8-week- old male C57Bl/6 mice under isoflurane anesthesia at 1010 vector genomes (vg). As controls, the left TA muscles were injected with AAVs carrying an empty vector at equivalent doses. At the experimental endpoint, mice were humanely euthanized via cervical dislocation, and TA muscles were excised rapidly and weighed before subsequent processing.

Histological analysis

Harvested muscles were placed in OCT cryoprotectant and frozen in liquid nitrogen-cooled isopentane. The frozen samples were cryosectioned through the middle of the muscle at 10 μm thickness and stained with hematoxylin and eosin. Tissue sections were imaged using a U-TV1X-2 camera mounted to an IX71 microscope, and an Olympus PlanC 10X/0.25 objective lens. DP2-BSW acquisition software

(Olympus) was used to acquire images. Images were separated into 8 fields covering the whole of the TA muscle (designated A1-A4 and B1-B4). The minimum Feret’s diameter of muscle fibers in fields A2, B2 and

B3 were determined using ImageJ software (US National Institutes of Health, Bethesda, MD, USA) by measuring at least 300 fibers per mouse muscle. The same fields were compared for each TA muscle examined.

Acknowledgements

We thank Monash Micro Imaging staff, AMREP campus for technical guidance and AMREP

Precinct Animal Centre staff for animal husbandry. We also thank Georgia Goodchild for help with the histological analysis. This work was supported, in part, by National Institutes of Health, the National

Health and Medical Research Council Grant, the Muscular Dystrophy Association, the University of

Cincinnati Graduate Dean Fellowship, and the American Heart Association (R01AG047131,

R01AG040019, and R03AG049657 to RTL; 1078907 to CAH; App 1117835 for Gregorevic Senior Research

Fellowship from NHMRC (Aust.); Graduate Dean Fellowship and 12PRE11790027 to RGW; R01GM114640 and Muscular Dystrophy Association Grant 240087 to TBT). This work was also supported by the

Integrated Diffraction Analysis Technologies Program of the Department of Energy Office of Basic Energy

Sciences awarded to the Advanced Light Source at Lawrence Berkeley National Laboratory. The Baker

Heart and Diabetes Institute is supported in part by the Operational Infrastructure Support Program of the Victorian Government. The Authors thank Dr. Hongwei Qian (Baker Heart and Diabetes Institute) for assistance with production of recombinant AAV vectors.

Chapter III

Residues near the GDF8 tolloid cut site are important for

cleavage and subsequent activation of latent GDF8*

*The following text is being submitted to the Journal of Biochemistry for publication: Jason C. McCoy and

Thomas B. Thompson. “Residues near the GDF8 tolloid cut site are important for cleavage and subsequent activation of latent GDF8”

Abstract

Growth differentiation factor 8 (GDF8), a.k.a. myostatin, is a member of the activin subclass within the larger TGFβ superfamily of signaling ligands. GDF8 has been well characterized as a negative regulator of muscle mass. As such, it is tightly regulated within the body. One such mechanism is the formation of a latent procomplex between GDF8 and its prodomain. Activation of latent GDF8 requires proteolytic cleavage of the prodomain by a member of the tolloid family of metalloproteases. Unlike other characterized proteases tolloid has no concrete consensus sequence and substrates need to be evaluated on a case by case basis. Here we investigate the tolloid cleavage site of the GDF8 prodomain to determine what residues contribute to tolloid recognition and subsequent proteolysis. Using sequential alanine mutations, we identified several residues that when mutated abolish tolloid mediated activation of latent

GDF8. Using western blot analysis with purified prodomain mutants and the proteolytic astacin domain of tolloid we determined that mutants were resistant to proteolysis. A reduction in processing was also observed when using the full length tolloid (BMP1) and purified latent complexes. Taken together our data demonstrates that while there is no consensus sequence for tolloid recognition tolloid utilizes the primary sequence of their substrate to dictate proteolysis.

Introduction

Growth differentiation factor 8 (GDF8), commonly known as myostatin, is a member of the larger

TGFβ superfamily of signaling ligands that functions as a potent negative regulator of muscle mass1,3,5,105,134. Genetic deletion of GDF8 or use of GDF8 inhibitors results in a drastic increase in muscle mass1,5,13,14,21,135,136. In contrast, overexpression of GDF8 results in muscle atrophy109,137–139. Thus, over the last two decades significant effort has been put forth toward the development of a therapy that can boost muscle mass by regulating GDF815,18,21,22,140–143.

GDF8, like all TGFβ ligands, is synthesized as a precursor protein containing an N-terminal signal sequence and prodomain followed by a C-terminal growth factor or mature signaling domain (Figure 1a).

The prodomain is essential for proper folding, localization, and plays an important role in the regulation of ligands23,24,27,28,41. The prodomain component is cleaved from the mature domain by proprotein convertases (PCs) such as furin24,27, ultimately producing the signaling ligand, which consists of two mature domain chains linked through a disulfide bond. To signal, ligands coordinate a complex between two type

I and two type II serine-threonine kinase receptors that when brought into proximity by the ligand will activate downstream SMAD transcription factors.

In most cases, after cleavage by furin, the prodomain remains non-covalently bound to the mature ligand dimer forming a procomplex, and for the majority of the TGFβ superfamily this procomplex is non-inhibitory.

However, a handful of ligands including GDF8,

GDF11, and TGFβ subclass ligands form interactions with the prodomain that render the ligand inactive or latent complex23,55,57.

These latent procomplexes require an Figure 1: Latent GDF8 and Tll1 astacin domain architecture and additional activation event to liberate the structure. a) Domain architecture of GDF8, signal sequence (SS), prodomain and mature domain. Tolloid and Furin cut sites denoted ligand to allow signaling. For example, the with size in kDa of each fragment after furin and tolloid processing shown. b) Structure of the GDF8 procomplex. Mature domain procomplex of TGFβ1 is activation by monomers in pale cyan and blue, prodomain monomers in light brown and brown. Dotted lines indicated resides not in density. mechanical force where the protein is pulled Predicted tolloid cut sites labeled. c) GDF8 procomplex shown in (a) rotated 90 degrees about the vertical axis. d) Domain architecture of the tolloid family including the prodomain, astacin domain, CUB and upon through interactions with αVβ6 integrins EGF domains. e) Structure of the Tll1 astacin domain, active site cleft shown. f) GDF8 prodomain structure shown in (a) super imposed and TGFβ binding proteins (LTBPs) or with the astacin domain (e) rotated 90 degrees about the horizontal axis. 52

GARP66,68,69,78. For GDF8 and GDF11, procomplex activation requires an additional proteolytic cleavage event mediated by the tolloid family of metalloproteases58,63,72,73,111.

Tolloids are metalloproteases characterized by having an astacin domain responsible for the proteolytic activity. The tolloid family shares a conserved domain architecture from Drosophila to humans, with an N-terminal prodomain, the proteolytic astacin domain, 2 tandem CUB (Compliment/Uegf/BMP1) domains (CUB1/2), an EGF (epidermal growth factor) domain (EGF1), a third CUB domain (CUB3), a second

EGF domain (EGF2) and 2 more tandem CUB domains (CUB4/5) (Figure 1d). There are 4 members of the tolloid family, mammalian tolloid (mTLD) and its splice form BMP1 (truncated at CUB3), tolloid like 1 (Tll1) and tolloid like 2 (Tll2). Tolloids are essential for the proper processing of ECM components such as procollagen and the release of TGFβ ligands from antagonism71–73,80,84,119,144–147. Notably tolloid is essential during embryogenesis by cleaving the BMP antagonists, Chordin, to ensure proper axial patterning72,87,90,148–151. The requirement of tolloid during embryogenesis is reflected in the embryonic lethal homozygous Tll1/BMP1 null mice that have improper heart maturation and abnormal collagen maturation85,146,152,153.

Despite several known substrates of tolloid processing, very little is known about how tolloid family members selectively cleave different substrates. In part, this is because tolloid protease sites are highly variable with limited consensus71,149. When comparing sequences of known substrates, the most common residue is an aspartic acid in the P1’ site, directly C-terminal of the scissile bond71. Analysis of tolloid cleavage of the GDF8 prodomain identified D99 as point of cleavage111. In fact, mutation of D99 to alanine abolished tolloid activation of GDF858,111. Beyond a preference for aspartate, little is known about the molecular mechanism that determine tolloid substrate recognition and processing. Additional studies have shown that the non-catalytic domains have a role in substrate recognition and/or activity148,151,154,155.

Structural and functional studies also showed that the C-terminal domains are involved in homodimer formation, restricting the activity of Tll1 and mTLD as compared to the astacin domain alone or BMP1

53 which lacks the domains involved in dimerization148,151,155. In contrast, when the CUB4 and CUB5 domains are removed from Tll2, it loses the ability to proteolyze chordin154. Thus, substrate specificity appears to be complex and can involve both inter and intra molecular interactions with the CUB domains.

For GDF8, preliminary evidence suggests that all tolloids can remove latency through processing the prodomain. However, it is not known what the role of the additional domains of the tolloid proteins play in the recognition of the specific cut site of GDF8 or what prodomain sequences are necessary for tolloid-mediated activation recognition. To investigate this, we characterized the residues adjacent to the scissile bond of GDF8 to determine their impact on tolloid-mediated activation. Our results indicate that residues toward the N-terminus of the scissile bond are important for tolloid-mediated activation.

Furthermore, we demonstrate that the astacin domain alone is a potent activator of GDF8 latency, indicating that the additional domains of tolloids are not required for substrate recognition.

Results

Recent structural studies have revealed different conformations of the GDF8 procomplex relative to that of TGFβ1 giving insight into tolloid-mediated processing. Unlike latent TGFβ1, GDF8 adopts a more elongated “open” conformation resembling a ‘V’ (Figure 1b)55,57. This places the Tolloid cut site of the prodomain at the center of the complex, in a crevice adjacent to the ligand (Figure 1b). Using the crystal structure of the astacin domain, it is clear that this open conformation is necessary for Tolloid to gain access to the scissile bone (Figure 2e, f). However, the GDF8 procomplex structure does not contain residues V96-L106 indicating that this region is highly flexible. Sequence alignments show that the residues adjacent to the D99 (P1’) cut site are highly conserved (Table 1) . Thus, we wanted to investigate whether these residues were important for tolloid-mediated activation of GDF8.

Luciferase assays screening GDF8 alanine mutations

In previous studies, we established a cell-based assay where plasmids containing GDF8, furin and

58 tolloid could be transfected and shown to stimulate luciferase under control of the (CAGA)12 promoter .

Using a similar assay format, we tested whether residues adjacent to the D99A P1’ are important to GDF8 activity. For this, we made individual alanine mutations from D92 and D110 (Figure 2a). The affect alanine mutations had on GDF8 activation by tolloid was measured using a GDF8-responsive, transfection-based luciferase assay. In short, GDF8, furin, and one of the three Tolloid family members, Tll1, Tll2, or mTLD

DNA was transfected into HEK293T (CAGA)12 cells. Results of the point mutations are plotted as fold activation compared to wild type (WT) GDF8 in Figure 1b-d. All luciferase assays within this study include a legend in the top left denoting what proteins are transfected (Trx) and what proteins are added

Table 1: CLUSTAL O(1.2.4)195 multiple sequence alignment GDF8 residues adjacent to the tolloid cut site (D99). GDF8 sequences were identified using uniprot2 and aligned using Clustal sequence alignment web software. Across 21 species the identity was 87%. Sequences are in FASTA format with the uniprot accession number followed by the species abbreviation in bold. D99 required for tolloid processing is red, bolded, and underlined. Asterisks at the bottom of the sequences identify identical amino acid residues, colons and periods represent highly similar residues.

55 exogenously (Exo) for clarity. As expected, mutating the P1’ site, D99, to an alanine completely abolishes GDF8 activity58,111.

Interestingly, residues N-terminal to D99, including D92A, Y94A, D95A, V96A, and Q97A, were identified that significantly reduced

GDF8 activity when co-transfected with Tll1,

Tll2, or mTLD (Figure 1b-d). On the other hand, residues C-terminal to the cut-site had less of an impact. While four C-terminal mutants

E107A, D108A, D109A and D110A also resulted in a significant reduction in GDF8 activity,

D100A, S101A, S102A and S105A had a little impact with a modest increase when co- transfected with mTLD. Curiously, two mutants R98A and D103A significantly Figure 2: Transfection based luciferase assay of GDF8 alanine screen. a) Sequence of the GDF8 tolloid cut site with P3-P3’ residues denoted and the essential D99 residue shown in red. b) 100ng of increased Tll1, Tll2, and mTLD mediated GDF8 WT or mutant DNA transfected with 50ng of Furin and Tll1 DNA. c) Similar to (b) but with Tll2 DNA transfected. d) Similar to (b- activation. To ensure the impact of signaling c) with mTLD DNA transfected. The DNA transfected (Trx) and exogenous (Exo) protein added (if applicable) is denoted in the top was not a result of differences in expression, left each graph. Data plotted as mean±SD and were conducted at least three times with experimental triplicate. Bar graphs were WT, D92A, Y94A, D95A, V96A, Q97A and D99A compared using one-way anova with Bonferroni correction against WT GDF8 (*P<0.05, **P<0.01, ***P<0.001). DNA were transfected into HEK293T cells with furin and analyzed by an anti-GDF8 prodomain western blot (Figure 3). These results show that loss of GDF8 signaling is not from a loss a protein expression. We also transfected each GDF8 mutations with Tll2 DNA to determine if prodomain processing was altered

(Figure 3). Tll2 was previously shown to yield the highest activation of GDF8 compared to other tolloids

56 when transfected58. This showed that a reduction in Tll2 mediated prodomain processing occurred for several of the point mutations, including D92A, Y94A, Q97A and D99A when compared to WT GDF8 (Figure

Figure 3: Expression and cleavage test of transfected GDF8 mutants. (a-d). GDF8 DNA transfected is denoted at the top of each lane and if Tll2 DNA was added (+) or not (-). The band corresponding to full length prodomain and the C-terminal cleavage product are labeled. 3).

Is the proteolytic astacin domain sufficient for latent GDF8 activation?

Given the consistent results across the different tolloids for the GDF8 mutations tested, we next wanted to determine if the astacin domain alone was sufficient for processing the prodomain and activating GDF8. Using previously established refolding protocols, we generated active TLL1 astacin domain which was capable of processing the commercially available Mca-YVADAPK(Dnp)-OH fluorogenic substrate with a specific activity of 35.3±6.96 pmol/min/µg156. Using this assay format, we first wanted to determine if the prodomain alone could serve as a competitive inhibitor. Prodomain (ProW) was produced in bacteria as previously published and titrated against a constant concentration of the fluorogenic peptide and the astacin domain, 2.5µM and 250nM, respectively (Figure 4a). Addition of ProWT

57 shows a clear reduction in fluorescence indicating that the astacin domain activity for the fluorogenic peptide is being inhibited.

Next, we wanted to determine if the astacin domain could activate latent GDF8. Previously, we were able to produce and purify latent GDF8, which has minimal activity when administered to (CAGA)12- luc cells39,58,91,95,129. To test this, we titrated a constant amount (0.62nM) of latent GDF8 (GDF8L) with increasing amounts of the astacin domain (Figure 4b). Titration of the astacin domain provides robust activation of GDF8L with an EC50 29.55nM for the astacin domain, indicating that the astacin domain alone is fully capable of activating latent GDF8.

Figure 4: Analysis of Tll1 astacin domain activity vs GDF8 prodomain states. a) Fluorogenic peptide incubated with the astacin domain and treated with a titration GDF8 prodomain, plotted as fold specific activity. b) Latent GDF8 treated with a titration of the Tll1 astacin domain. Data were fit by non-linear regression with a variable slope to generate the EC50 curve. c) Luciferase assay of latent GDF8 (GDF8L), acid activated GDF8 procomplex (GDF8AA), and a mixture of GDF8 prodomain:mature domain in a 3:1 molar ratio, respectively, treated with or without the Tll1 astacin domain. The DNA transfected (Trx) and exogenous (Exo) protein added (if applicable) is denoted in the top left of each luciferase assay. d) Anti-GDF8 prodomain western blot of GDF8L, GDF8AA, or GDF8 prodomain alone (ProWT) treated with the Tll1 astacin domain for 30min or 1hr. Experiments a-c were conducted at least three times with experimental triplicate.

Previous studies by our lab and others have shown that the procomplex of GDF8 can exists in a partially active state58. For instance, treating latent procomplex with acidic conditions can activate or trigger the procomplex without full dissociation of the prodomains, referred to as acid-activated (GDF8AA).

As shown in Figure 4c, acid activation results in a gain of signaling activity. However, treatment of GDF8AA with the astacin greatly enhances activation, indicating that the semi-active state of GDF8AA can be fully activated (Figure 4c). In fact, the combination of acid-activation plus astacin domain treatment is greater than astacin domain alone. Similar results to GDF8L were observed when combining excess ProWT with apo-GDF8 at a 3:1 molar ratio (Figure 4c). Cleavage analysis by Western Blot showed that both GDF8L and

GDF8AA were processed by the astacin domain, albeit less efficiently than the prodomain alone (Figure

4d). This shows that a reduction in prodomain proteolysis occurs when the ligand is bound.

Direct comparison of astacin-domain mediated processing of GDF8 prodomain mutations.

We next wanted to determine if residues adjacent to the scissile bond, when mutated, had a direct impact on astacin-mediated proteolysis. First, the GDF8 prodomain and select mutants, namely D92A, Y94A,

D95A, V96A, and D99A were expressed and purified from E. coli. The extent of prodomain processing, without the GDF8 ligand present, was monitored by Western blot analysis (Figure 5a). As expected, the uncut GDF8 prodomain (ProWT) migrates at 34kD, while incubation with the astacin domain results in fragmentation as noted by the appearance of a band around 20kD. Figure 4a shows that all mutants showed slight to complete reduction in processing when compared to ProWT. D92A, Y94A, D95A and V96A were processed to a much lesser extent than WT prodomain, while D99A showed no processing. We next analyzed processing of the prodomain in a reconstituted procomplex with ligand. WT and various prodomain mutants were combined with the highly similar GDF8 relative, GDF11 (90% identical) and purified through size-exclusion chromatography. The reconstituted procomplexes showed reduced processing by the astacian-domain, indicating that the procomplex is more difficult to proteolyze (Figure

5b). Prodomain mutants had similar negative effects on astacin-domain mediated processing as compared

59 to WT of the reconstituted procomplexes (Figure 5b). Similar analysis was performed with full-length tolloid (BMP1) under the same conditions (Figure 5c & 5d). While the overall trends of the prodomain mutants was conserved using both the astacin domain and BMP1, there was a significant reduction in processing by BMP1. Thus, residues upstream to the scissile bond directly impact astacin-domain mediated processing of the prodomain, both alone and in the presence of the ligand.

Figure 5: Western blot analysis of GDF8 prodomain mutant processing by astacin domain and BMP1. All westerns are anti-GDF8 prodomain blots a) Bacterially produced GDF8 prodomain treated with astacin domain in a 1.2:1 molar ratio (astacin:prodomain) for 1hr. b) Reformed complexes, of bacterially produced GDF8 prodomain and GDF11 treated with astacin domain in a 2.4:1 molar ratio (astacin:prodomain) for 1hr c) Bacterially produced GDF8 prodomain treated with BMP1 in a 1.2:1 molar ratio (BMP1:prodomain) for 1hr. d) Reformed complexes of bacterially produced GDF8 prodomain and GDF11 treated with BMP1 in a 2.4:1 molar ratio (BMP1:prodomain) for 1hr.

Fully latent GDF8 mutant processing by tolloid.

Previously we have shown that purified procomplex from CHO cells is fully latent whereas, reconstitution of the procomplex from purified components retains significant activity58. Therefore, we wanted to investigate if the prodomain mutations had an impact on fully latent GDF8 procomplex. Here we titrated exogenous astacin domain to cells transfected with GDF8 prodomain mutants, along with furin in the (CAGA)12 reporter cell line. EC50 curves for the addition of astacin domains was generated for each of the following mutants: WT, D92A, Y94A, D95A, V96A, Q97A, and D99A (Figure 6a). While WT GDF8 DNA

Figure 6: Analysis of fully latent GDF8 mutants. a) 100ng of GDF8 DNA and 50ng Furin DNA were transfected into HEK293T

(CAGA)12 cells and treated with a 1:2 serial dilution of the Tll1 astacin domain. b) 0.62nM of purified GDF8 procomplexes treated with a serial dilution of the Tll1 astacin domain, similar to (a) and added to HEK293T (CAGA)12 cells. Data from (a) and (b) were fit by non-linear regression with a variable slope to generate EC50 curves. c) Transfection based luciferase assay were full length

Tll1, Tll2 or mTLD were transfected into (CAGA)12 cells. Cells were then treated with 0.62nM of purified GDF8 procomplex. Bar graphs were compared using one-way anova with Bonferroni correction against WT GDF8 (*P<0.05, **P<0.01, ***P<0.001). d) 20nM GDF8 procomplex was acid activated by adding HCl and neutralized by NaOH addition. The acid activated (AA) procomplex was then added to (CAGA)12 cells and plotted as fold over serum free background. The DNA transfected (Trx) and exogenous (Exo) protein added (if applicable) is denoted in the top left each graph. Experiments were conducted at least three times with experimental triplicate and plotted as mean±SD.

61 was effectively activated with an EC50 concentration of 137nM, the rest of the prodomain mutations were significantly impaired, despite similar levels of procomplex formation (Figure 3). Consistent with the

Figure 2 data where the tolloids were transfected, exogenous addition of the astacin domain also failed to activate Y94A, Q97A and D99A, with significant reductions observed for D92A, D95A and V96A with

EC50s of, 648nM, 1.97µM, and 462nM, respectively.

To further the analysis, we expressed and purified latent complexes from Expi293F and analyzed their cleavage and activation by the astacin domain. The purified latent complexes had little to no activity in the (CAGA)12 reporter cell line.

Titration of the astacin domain was used to activate a constant concentration (0.62nM) the purified latent WT and GDF8 mutants D92A, Y94A and D99A

(Figure 6b). The WT procomplex was readily activated by the astacin domain with an EC50 of Figure7: Western blot analysis of latent GDF8 procomplex 66.5nM. Both Y94A and D99A were not activated by processing. Both westerns blots are anti-GDF8 prodomain run under non-reducing conditions a) Latent GD8 the astacin domain, whereas D92A had an EC50 of procomplexes incubated with the Tll1 astacin domain overnight at 37oC in a 2.4:1 molar ratio (astacin:prodomain). b) Latent GD8 procomplexes incubated with BMP1 overnight 201nM. Furthermore, the maximum signal for D92A at 37oC in a 2.4:1 molar ratio (astacin:prodomain). was approximately half that of WT. Western blot analysis also showed that each mutation was not processed by the astacin domain under conditions where the WT was readily processed (24hrs at 37C in a 2.4:1 molar ratio of astacin:Prodomain) (Figure 7). We next tested if the exogenous latent complexes could be activated by full-length tolloids. Here, we transfected Tll1, Tll2 and mTLD and added purified latent GDF8 procomplexes (Figure 6c). WT latent complex was activated by all three tolloids, with Tll2 giving the strongest activation. Y94A and D99A remained latent and were not activated by the different

62 tolloids whereas D92A was only activated by Tll2 (Figure 6c). Again, of the fully latent complexes only WT showed significant processing via western blot when using BMP1 (Tll1, Tll2, or mTLD are not available commercially) (Figure 7). To rule out the possibility that the GDF8 prodomain mutants were improperly folded, we acid-activated the purified latent complexes and tested their activity in (CAGA)12 cells (Figure

6d). As expected acid-activation of WT latent procomplex resulted in robust signals. Furthermore, acid- activation of D92A, Y94A and D99A similarly resulted in robust activation, indicating that the GDF8 procomplex can be activated by a process other than tolloid-mediated proteolysis.

GDF8 mutants defective in Tolloid activation can suppress WT GDF8.

Previous research has demonstrated that plasmid, or AAV, delivery of the D99A prodomain mutant alone (without the mature domain) is able to promote muscle growth157,158. In addition, a heterozygous mutation in humans within GDF11, R298Q, that prevents furin cleavage and activation, results in orofacial clefting159. We showed that Figure 8: Transfection-based luciferase assay of dominant negative GDF8. 50ng of WT GDF8, Furin, Tll2, and 50ng of empty vector or R298Q can suppress WT GDF11 when both mutant GDF8 DNA (denoted under the X-axis) were transfected into HEK293T (CAGA)12 cells. Data plotted as mean±SD and were conducted at least two times with experimental triplicate. Bar graphs proteins are expressed in HEK293T cells, were compared using one-way anova with Bonferroni correction against WT GDF8 (*P<0.05, **P<0.01, ***P<0.001). The DNA indicating the mutant GDF11 was dominant transfected (Trx) and exogenous (Exo) protein added (if applicable) is denoted in the top left each graph. negative to WT GDF11 signaling 159. While the prodomain mutant, D99A, has been shown to reduce WT

GDF8 signaling by function as an inhibitor, we wanted to determine if full-length GDF8 that could not be cleaved by Tolloid could serve as a dominant negative regulator and suppress WT GDF8 signaling. For this experiment, we focused on the established tolloid site, D99A, and Y94A identified in this study to hinder tolloid cleavage. In addition, we tested mutants D108A, D109A, and D110A which had a modest decrease

63 in activity in Figure 1b-d but not to the extent of Y94A which completely abolished signaling. To test this, we conducted transfection-based luciferase where WT GDF8 and mutant GDF8 DNA were transfected into

HEK293T (CAGA)12 cells in a 1:1 ratio with furin and Tll2. As expected, WT GDF8 without mutant DNA was able to signal while both Y94A and D99A did not and D108A, D109A, and D110A has approximately half the signalign of WT (Figure 8). However, when WT GDF8 was transfected with Y94A or D99A a ~50% reduction in signaling was observed (Figure 6). In contrast, when WT GDF8 was transfected with D108A and D110A there was no significant change in signaling while transfection with D109A resulted in a significant increase. Thus, mutations in the prodomain that are resistant to tolloid activation can suppress

WT GDF8 signaling.

Discussion

Tolloids are essential for the proper maintenance of ECM components and activation of TGFβ superfamily ligands. Despite the prevalence of tolloid activity in vivo, little is known about the molecular mechanisms that dictate substrate selection. Previous studies have shown that the non-catalytic domains of Tll1 and mTLD serve a negative regulatory role while in contrast, Tll2 non-catalytic domains were shown to enhance binding to chordin 145,151,155. Examining different substrate cut sites reveals no concrete consensus sequence, outside of the regular occurrence of an aspartic acid in the P1’ site 71,149. Despite no consensus sequence, tolloid cleavage is highly specific. Thus, individual substrates require a more focused analysis to ascertain the molecular requirements for specific tolloid-mediated cleavage.

Given its role in suppressing muscle mass, significant effort has been undertaken to understand the function and regulation of GDF8. GDF8 activity is tightly regulated through the formation of a latent procomplex, where its N-terminal prodomain is cleaved by Tolloid at aspartate 99 (D99). Mass spectrometry analysis showed this is a highly specific cleavage event despite an adjacent aspartic acid at

D100. In this study we investigated the molecular requirements of tolloid-mediated cleavage of GDF8 prodomain, and the subsequent activation from latency.

Unlike other substrates, latent GDF8 can be activated by all four tolloid family members, Tll1, Tll2, mTLD and BMP1. Through sequence alignment of 21 different species residues near the tolloid cut site

(aa89-aa112) are 87% identical (Table 1). Interestingly, recent structural analysis of the GDF8 procomplex shows that this region is disordered. Given the high sequence conservation of the flexible loop and proximity to the tolloid cut site implies a functional role. Thus, we hypothesized that this conservation is important for tolloid recognition and activation of latent GDF8. High-throughput transfection-based luciferase assays revealed that mutants made N-terminal of the tolloid cut site abolished activation by all three family members tested, Tll1, Tll2 and mTLD (Figure 2b-d). Due to the proximity of the cut site, the relative size of the astacin domain to the GDF8 prodomain, and that single point mutations were able to disrupt activation by all family members suggests that mutants impaired the astacin domain, specifically.

While difficult to distinguish, these mutations could disrupt direct interactions with the astacin domain or possibly alter the conformation of the loop to one that is less favorable for proteolysis. Interestingly, mutation of R98A, and to a lesser extent D103A, increased tolloid-mediated activation of GDF8. This suggests that residues near the cut site have both a positive and negative impact on tolloid activation.

Given that all tolloids had similar responses to the panel of mutations, suggests that the astacin domain, which is 88% identical across the different tolloids, is the primary factor driving proteolysis. This is supported by the recent structure of the GDF8 procomplex, which shows a conformation amenable to bind the astacin domain57. Thus, we asked whether the astacin domain alone was sufficient for processing the prodomain and activating latent GDF8. Using purified astacin domain from Tll1 we demonstrated that indeed the astacin domain is sufficient for not only GDF8 prodomain cleavage but latent GDF8 activation

(Figure 4b-d). While the non-catalytic domains might fine-tune cleavage, they are not required for latent

GDF8 recognition and activation. We also found that acid-activated procomplex was more readily

65 processed by tolloid than the fully latent GDF8 procomplex, suggesting that acid-activation might make the tolloid recognized loop more accessible. In addition, an increase in total activity was observed for the acid activated versus fully latent procomplex, indicating that the prodomain fragments might be more readily displaced post-processing when the complex is acid-activated (Figure 4c). However, despite more activity, western blot analysis revealed that the latent and acid activated GDF8 procomplex had similar levels of proteolysis (Figure 4d). Interestingly, the prodomain alone was cleaved more efficiently than either the latent procomplex or the acid-activated form implying that the ligand or the conformation of the prodomain in complex with the ligand dampens tolloid proteolysis. The result that the astacin domain can process the prodomain alone suggests that a specific conformation of the prodomain in the procomplex is not required for tolloid activation, and the recognition might be more dependent on the primary sequence. However, it is possible the prodomain retains necessary structural elements for astacin recognition and cleavage even in the absence of ligand.

Using recombinant prodomain expressed and purified from E. coli. we showed that the mutants identified to have reduced GDF8 activity also displayed a decrease in sensitivity to astacin domain proteolysis. As expected, complete loss of processing occurred when D99 (P1’ site) was altered to an alanine. Other mutations, D92A, Y94A, D95A, and V96A all showed a significant decrease in astacin domain processing when not in complex with the GDF8 ligand (Figure 5). Again, indicating that these residues are directly impacting the ability of the astacin domain to cleave the prodomain even in the absence of ligand. Similarly, we observed a significant reduction in proteolysis when the prodomain mutants were combined with the GDF8 ligand. Comparison of the astacin domain to the Full length BMP1 showed even a further reduction in processing, indicating that the non-catalytic domains attenuate processing. This is similar to other studies that show BMP1 is less efficient at processing the collagen prodomain when compared to the astacin domain alone145. Nevertheless, prodomain mutants D92A,

Y94A, D95A, and V96A still resulted in less proteolysis by BMP1 than WT prodomain alone, or in complex

66 with ligand. Taken together, these results show that the prodomain is more readily processed than the procomplex, and that the astacin domain is more potent than BMP1.

To further investigate activation of the GDF8 procomplex we determined that exogenous delivery of the astacin domain could readily activate WT prodomain complex that was generated via transfection in HEK293T (CAGA)12 cells (Figure 7a). Consistent with previous results, transfected prodomain mutants had displayed reduced activation. In fact, Y94A, Q97A, and D99A could not be activated by the astacin domain. Other mutants, D95A and V96A could be activated by the astacin domain but with a higher EC50 than WT. Focusing on selected prodomain mutants, D92A, Y94A, and D99A we purified recombinant versions of the fully latent complex and assessed their ability to be directly activated by the astacin domain. At low concentrations of astacin domain all procomplexes showed little to no activity. Titration of recombinant precomplex with astacin domain resulted in GDF8 signaling activity with an EC50 similar to transfection of GDF8 WT. However, titration of the astacin domain could not activate Y94A or D99A, and was significantly less effective in activating D92A. All mutant versions also displayed a reduced sensitivity to proteolysis by the astacin domain, further support that the mutations in the prodomain hindered the activation of the latent complex by suppressing tolloid proteolysis. Similar results were observed when the different tolloids were transfected in HEK293T (CAGA)12 cells and treated with the different procomplexes. Consistent with previous results, Tll2 was the most effective tolloid in activating

GDF8. Critically, we show that the purified mutant procomplexes can be acid-activated resulting in potent

GDF8 signaling. This supports that the mutant procomplexes are resistant to tolloid activation. While previous work has described the importance of D99, this is the first time that additional residues in the prodomain have been identified to render the procomplex tolloid resistant.

A current strategy for therapeutic strategy for boosting muscle mass has been through the inhibition of tolloid-mediate activation of GDF8. This has been supported by studies that show administering D99A alone, without the mature domain, results in muscle growth. In addition, an antibody

67 has also been developed that binds the GDF8 procomplex and prevents prodomain cleavage by tolloid.

Interestingly, the antibody does not bind directly to the tolloid cut site, but on the top of the prodomain proximal to the ligand. The method of inhibition is thought to occur through exclusion of tolloid and/or by restricting the procomplex in a conformation that is unable to be proteolyzed. However, most methods rely on either inhibiting an already activated GDF8 or inhibiting a GDF8 procomplex post-translationally.

Recent studies have implied that non-functional variants of GDF11 can have a dominant negative effect on WT GDF11. Since GDF11 is a dimer, similar to GDF8, this is likely due to the non-functional form combining with the WT version, forming a less functional procomplex. In this study, we wanted to test if, by combining a tolloid-resistant GDF8 with WT GDF8, could we suppress GDF8 activity. Indeed, our results show that co-transfection of the GDF8 prodomain mutations that are tolloid resistant with WT GDF8 suppressed GDF8 signaling activity in the (CAGA)12 reporter assay. Thus, another option for suppressing

GDF8 in vivo might be to co-express tolloid-resistant versions which could target endogenously produced

GDF8 before secretion.

Prior to this study only D99 of the GDF8 prodomain was known to be required for tolloid mediated activation. Due to the lack of a tolloid consensus sequence we investigated the residues near the scissile bond of prodomain of GDF8 and identified additional residues important for proteolysis and activation.

We also show that the astacin domain alone is sufficient for processing the prodomain and activating

GDF8 procomplex. Further evidence suggests that the prodomain processing is attenuated when bound to the ligand and, further, that the astacin domain is more effective in activating GDF8 procomplex than the full-length tolloids. Collectively, this study provides insight into the specific mechanism of GDF8 activation by the tolloid family or proteases.

Materials and Methods

HEK293-(CAGA)12 Luciferase-Reporter Assay

Luciferase assays were conducted, in large part, as previously described by our lab39,58,91,95,104,116. In short,

HEK293-T cells stably transfected with a firefly luciferase reporter gene under the control of the SMAD3- responsive (CAGA)12 promoter were seeded in growth media at 20,000 cells per well in a 96 well, poly-D lysine coated plate. Cells were then treated 24hours after seeding. Transfection based luciferase assays in

Figure 2 were conducted by transfecting 50ng of Furin, 50ng of the appropriate Tolloid DNA, and 100ng of GDF8 DNA. After 24hours growth media was swapped for serum free media. Data was collected

24hours after the media swap and plotted using GraphPad Prism 5 software.

EC50 curve in Figure 4b and 6b were generated using exogenous protein. 24hours after seeding, serum free media containing a constant concentration of latent GDF8 (0.62nM) was treated with a titration of the Tll1 astacin domain. Data was collected 24hours later and fit to a nonlinear regression with variable slope via GraphPad Prism 5 software to generate the EC50 curves. Data in Figure 3c was generated by treating cells with 0.62nM of Latent GDF8 (GDF8L), Acid Activated GDF8, or Apo-GDF8 mixed with 1.86nM of bacterially produced GDF8 prodomain with or without 175nM of the Tll1 astacin domain.

The EC50 curve in Figure 6a was generated by transfecting 100ng of GDF8 DNA and 50ng of furin. After

24hrs growth media was swapped with serum free media containing a titration of the Tll1 astacin domain.

The resulting EC50 curve was fit using GraphPad Prism 5 software as before. Figure 5c was generated by transfecting 50ng of tolloid DNA and 50ng of furin DNA. After 24hrs media was replaced with serum free containing 0.62nM of latent recombinant GDF8 and plotted using GraphPad Prism 5 software. Figure 6 was generated by transfecting in 50ng furin and Tll2 DNA, with 50ng of GDF8 DNA with either 50ng of empty vector or 50ng of the WT GDF8. Media was swapped 24hrs later for serum free and read 24hours after media swap.

HEK293T expression test of GDF8 prodomain mutants

1.5ug of GDF8 DNA, and 0.75ug Furin DNA was transfected into HEK293T cells in a 6-well format with

0.75ug of empty vector or Tll2 DNA to replicate 96well conditions. Media was swapped 24hours later and conditioned media (CM) was harvested 24hours after the media swap. CM was then concentrated and analyzed via anti-GDF8 prodomain western blots under non-reducing conditions. An anti-GDF8 prodomain western was conducted using standard western-blot procedures.

Tll1 Astacin domain production and refolding from Escherichia coli.

The astacin domain of Tll1 was produced and refolded as previously described156. In short, the astacin domain of Tll1, residues 148-347, was cloned into the pET28a(+) expression vector without N- or C- terminal tags and used to transform RosettaTM (DE3) competent cells. Cells were cultured until an OD of

0.8 at 600 nm then subjected to a cold drunk induction where cells are placed in an ice bath and EtOH is added to a final concentration of 2% (v/v). After approximately 10minutes cells are induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for culture at 20oC overnight. Cells were harvested by centrifugation and suspended in 5mM EDTA, 5mM benzamidine, and 50mM Tris-HCl pH 8.0 (lysis buffer) before sonication for cell lysis. The inclusion body containing the astacin domain was then isolated by centrifugation and washed three time with lysis buffer with a final wash with 4M urea, 1% (v/v) Triton-

X100, 5mM EDTA, 50mM Tris-HCl pH 8.0. Inclusion bodies were solubilized to a final concentration of approximately 5mg/ml in 8M guanidinium chloride, 100mM DTT, and 50mM Tris-HCL pH 8.0 as previously described156. The solubilized astacin domain was then refolded by rapid dilution in 50mM Tris-HCl pH 8.5,

0.8M l-arginine, 125mM NaCl, 1mM CaCl2, 10µM ZnCl2, 1mM reduced glutathione, 0.25mM oxidized glutathione to a final concentration of 50µg/ml and stirred at 4oC for 14days. The refolded astacin domain

o was then thoroughly dialyzed into 20mM Hepes pH 8.0, 125mM NaCl, 1mM CaCl2, 10µM ZnCl2 at 4 C.

Dialyzed material was then concentrated filtered to remove misfolded material and tested for activity using the RnD systems fluorogenic peptide (Cat#: ES007).

Astacin domain processing of Fluorogenic peptide (MCA-YVADAPK(DNP)-OH)

The fluorogenic peptide used is commercially available from RnD systems (Cat#:ES007) and experiments were conducted via manufacturer protocol. In short, 500nM of Tll1 astacin domain was mixed with 2.5µM of the fluorogenic peptide with or without increasing amounts of bacterially produced GDF8 prodomain in assay buffer, 25mM Hepes, 0.1% Brij-35 (w/v), pH 7.5. Specific activity was calculated using the below equation.

푅퐹푈 ∗ 푝푚표푙 푉푚푎푥( )푥퐶표푛푣푒푟푠푖표푛 퐹푎푐푡표푟 ( ) Specific Activity= 푚푖푛 푅퐹푈 Amount of Enzym (µg)

*Calibration standard MCA-P-L-OH (Bachem Catalog#: M-1975)

Production and purification of GDF8 prodomain from Escherichia coli

The production and purification of the GDF8 prodomain from E. coli was conducted as previously published58. In short, GDF8 prodomain mutants were cloned into a modified pET28a expression vector containing an N-terminal 6x-His tag, a maltose binding protein (MBP), followed by an HRV-3C protease cleavage site. The four cysteine residues within the GDF8 prodomain (C39/C41/C137/C138) were mutated to serine residues to improve solubility. E. coli RosettaTM (DE3) cells transformed with the proper GDF8 construct were grown to an OD of 0.8 at 600nm and induced with 0.5mM IPTG at 20oC overnight. Cells were lysed via sonication and soluble 6xHis-MBP-GDF8 prodomain was applied to a nickel affinity column

(GE Lifesciences) followed by elution with 20mM Tris pH 7.5, 500mM NaCl, 500mM Immidazole. Elutions were dialyzed into 20mM Tris pH 7.5, 500mM NaCl, 1mM DTT, 4% ethylene glycol (v/v) and HRV-3C protease was added. Following cleavage, the protein was dialyzed into 10mM HCl and applied to a C4 reverse phase column (Sepax) and eluted with a linear gradient to 0.1% TFA, 95% acetonitrile over 30 column volumes and prodomain containing fractions were pooled.

Production and purification of recombinant latent WT GDF8 and mutants.

Latent GDF8 in Figure 2 was purified from a stably transfected CHO cell line. Latent GDF8 and GDF8 mutants were transiently cotransfected with furin into expi293 cells (Life Technologies) and purified from conditioned media. Purification from CHO conditioned media (CM) is as follows. CM was concentrated

10-fold and dialyzed into 50mM Tris, pH 7.4, and 500mM NaCl and applied to Lentil Lectin Sepharose 4B

(Amershame Biosciences). Elution was conducted using the same buffer with 500mM Methyl Mannose.

Elutions were then applied to S200 size exclusion chromatography equilibrated with 20mM Hepes, pH 7.4,

500mM NaCl (Pharmacia Biotech). Purification of GDF8 from expi293 cells was conducted using the same method as CHO CM without the initial concentration. Quantification was conducted as previously described using western blot analysis and quantified using a GDF8 mature standard under non-reduced conditions 39,58,95,129,130.

Western blot processing of the GDF8 prodomains

Western blot analysis of astacin domain processing of the GDF8 prodomain was conducted by mixing 50ng of GDF8 prodomain, alone or in complex, with 50ng of the astacin domain for apo-prodomain experiments or 100ng for prodomain complexes, 1.2:1 or 2.4:1 molar ratio (protease:prodomain), respectively in

25mM HEPES pH 7.5, 0.1% Brij-35. The only exception being Figure 1d where 100ng of prodomain was used, still with a 2.4:1 molar ratio of astacin:prodomain. Processing via BMP1 was conducted using a 2.4:1 molar ratio for both apo- and prodomain reformed or latent complexes. The mixture was then placed at

37oC for 1hour. The only exception being fully latent complexes which required overnight incubation with

BMP1 for detectable processing. SDS-PAGE gels were run under non-reducing conditions and transferred using standard western blot protocols. Western blots were anti-GDF8 prodomain using the anti-pro antibody, AF1539 (RnD), per manufacturer’s protocol. Western blot signal was captured using the C-DiGit blot scanner (Li-COR).

Acid Activation

Acid activation of latent complexes was conducted as previously described58. In short, latent GDF8 was rapidly acidified by 100% HCl for 10minutes and neutralized by addition of 10M NaOH. This material was then used for western blot analysis of astacin processing and luciferase assays.

Chapter IV

Crystal structure of the WFIKKN2 follistatin domain reveals

insight into how it inhibits growth differentiation factor 8

(GDF8) and GDF11*

*The following text is part of the publication: Jason C. McCoy, Ryan G. Walker, Nathan H. Murray, and

Thomas B. Thompson. “Crystal structure of the WFIKKN2 follistatin domain reveals insight into how it inhibits growth differentiation factor 8 (GDF8) and GDF11”. JBC February 27, 2019. Doi:

10.1074/jbc.RA118.005831 PMID: 30814254 PMCID: PMC6484119

Abstract

Growth Differentiation Factor 8 (GDF8), a.k.a. Myostatin and GDF11 are closely related members of the TGF-β family. While GDF8 is a strong negative regulator of muscle, GDF11 has been implicated in various age-related pathologies, such as cardiac hypertrophy. GDF8 and GDF11 signaling is controlled by the extracellular protein antagonists follistatin, FSTL3 and WFIKKN. In common, each contains a follistatin domain (FSD) that is important for ligand binding and antagonism. Here, we investigate the structure and function of the follistatin domain (FSD) from WFIKKN2 and draw comparisons to the FSDs of follistatin and FSTL3. We determined that WFIKKN2 FSD could interact with both GDF8 and GDF11 and block interactions with the type II receptor, ActRIIB. Further, we solved the crystal structure of the WFIKKN2 FSD to 1.39Å and identified surface exposed residues that when mutated to alanine, reduced antagonism of GDF8 in full-length WFIKKN2. Comparison of the WFIKKN2

FSD with follistatin and FSTL3 highlights differences in both the FSD structure, and position of residues within the domain that are important for ligand antagonism. Taken together, our results indicate that WFIKKN and follistatin utilize the FSD to block the type II receptor, albeit through different binding interactions.

Introduction

Growth Differentiation Factor 8 (GDF8), also known as Myostatin, is a member of the TGF-β superfamily of ligands30. GDF8 is a potent negative regulator of skeletal muscle growth where genetic deletion results in a hyper-muscular phenotype1,5. In the adult animal, several studies have established that downregulation, or inhibition, of GDF8 induces muscle hypertrophy1,2,5,13,51,134. In contrast, transgenic overexpression of GDF8 causes muscle wasting, consistent with the idea that certain tumors increase GDF8 levels and contribute to cancer cachexia109,139,160,161. GDF11, which shares 90% identity at the amino acid level, has distinct functions and is important for proper anterior-posterior

75 patterning8,162. Recent evidence supports a beneficial role of GDF11 in neurogenesis, however excessive treatment of mice with GDF11 results in muscle wasting, similar to GDF89,163,164. Given that GDF8 and

GDF11 have prominent roles in tissue homeostasis, there has been a strong interest in developing therapeutics that modulate their signaling, especially ones that aim to boost muscle by targeting GDF8 in muscle-related pathologies15,18,140,141,165,166.

To signal, TGF-β ligands binds to, and coordinate the assembly of two type II and two type I serine-threonine kinase receptors30. GDF8 and GDF11 signal by using a select combination of receptors; ActRIIA or ActRIIB (type II) coupled with Alk4 or Alk5 (type I)39,129,167. Ligand activity is selectively controlled by extracellular protein antagonists that bind to ligands and interfere with receptor binding. Although TGF-β ligands are structurally similar, antagonists have variable sizes and domain architectures129,130,168–171. This structural diversity allows different antagonists to selectively inhibit subsets of TGF-β family members. Over the years, structural studies have provided insight into how different antagonists can adopt different mechanisms to bind and neutralize ligands. However, it appears that in most cases antagonists block both the type I and type II receptor binding interfaces95,129,170–172.

A number of antagonists have been shown to bind and block GDF8 signaling, including follistatin splice variants follistatin288 (Fs288) and follistatin315 (Fs315),

FSTL3 (follistatin-like 3), decorin, and WFIKKN101,129,130,168,169,173. While follistatin and FSTL3 antagonize multiple ligands including activin A and activin B, WFIKKN is exceptionally specific for GDF8 and

GDF11101,102,130,174.

WFIKKN is named from the conserved multi-domain architecture - Whey Acidic Protein

(WAP), Follistatin domain (FSD), Immunoglobulin domain (Ig), two tandem Kunitz (K1, K2) domains and a

Netrin (N) domain (Figure 1a). Most animals have two related versions of WFIKKN, WFIKKN1 and

WFIKKN2, which share 56% identity91,100. Differences in potency and binding stoichiometry have previously been reported where WFIKKN2 has an IC50 of 0.26nM for GDF8 antagonism and is nearly 100-

76 fold more potent than WFIKKN191. In addition, WFIKKN2 forms a 1:1 complex with GDF8/11 (1 WFIKKN2 and 1 ligand dimer), whereas WFIKKN1 forms a 2:1 complex. Interestingly, removing the Kunitz 2 and

Netrin domains both reduces the potency of WFIKKN2 (IC50=7.2nM) and shifts the binding stoichiometry to that of full length WFIKKN191. While multiple domains are thought to interact with GDF8 and GDF11, previous studies have determined that the FSD plays a significant role in antagonism91,101.

Follistatin domains are also functionally important for the antagonists follistatin and FSTL3. FSTL3 and follistatin contain two and three FSDs, respectively, which have been shown through multiple X-ray crystal structures to directly contact the ligand94,95,129. Interestingly, each of the FSDs (FSD1-

3) within follistatin/FSTL3 adopt different molecular conformations and functions differently at the ligand interface, indicating that the FSDs are not functionally redundant. In fact, biochemical experiments showed that changing the order of the FSDs can severely alter ligand binding and specificity97,168.

While WFIKKN selectively inhibits GDF8 and GDF11, the molecular basis for ligand selectivity has not been established. Given that WFIKKN and follistatin/FSTL3 each have FSDs important for binding, whether they exhibit unique or common ligand mechanisms has not been determined. To investigate this, we characterized the WFIKKN FSD and contrasted the binding features of the follistatin FSDs with the WFIKKN FSD. We found that the FSD of WFIKKN blocks the type II receptor binding interface similar to the FSD2 of follistatin and FSTL3. However, WFIKKN2 FSD displays different structural features, and residues that interact with the ligand that map to a different location of the FSD, indicating a different binding mode when compared to the FSDs of follistatin and FSTL3.

Results

Production of WFIKKN2 Follistatin Domain.

Recombinant murine WFIKKN2

FSD, containing residues 104 to

172, was produced in bacteria. WFIKKN2

FSD protein formed inclusion bodies which were solubilized and subjected to oxidative refolding to induce disulfide bond formation.

Properly folded material was purified using reverse phase chromatography as depicted in Figure 1b. First, Native-PAGE was used to analyze the refolded

WFIKKN2 FSD and to evaluate binding interactions with GDF8. FSD alone migrated as a single band indicating the isolation of a single refolded species. Similar to previous published results, GDF8 alone does not enter the Native-PAGE and is not visible39. Upon mixing WFIKKN2 Figure 1: WFIKKN2 domain architecture and FSD purification. a) Domain architecture of the WFIKKN FSD. Conserved cysteines are represented as yellow spheres. b) Chromatograph depicting the C18 column purification of FSD with GDF8 an additional WFIKKN2 FSD. Inlet shows respective peak fractions run on an SDS-PAGE gel under non-reducing conditions. Solid line: WFIKKN2 FSD, dotted line: band appeared indicating complex Misfolded FSD and contaminants c-d) Native-PAGE of WFIKKN2 FSD and GDF8/11 complex formation. Molar ratio of WFIKKN2:Ligand are formation (Figure 1c). Titration of GDF8 annotated. Inlet shows the band (dotted box) that was excised and analyzed using SDS-PAGE under reducing conditions. with increasing concentrations of WFIKKN2 FSD showed an increase in the complex band intensity. Analysis of the newly formed band by SDS-PAGE shows that both WFIKKN2 FSD and GDF8 are

78 present, supporting complex formation (Figure 1c). Similar results were observed with GDF11 where the complex band was significantly sharper (Figure 1d). These results indicate that WFIKKN2 FSD is properly folded and can bind both GDF8 and GDF11, as previously implicated101.

Comparison of WFIKKN2 FSD and WFIKKN2 Full length protein.

We next wanted to compare the WFIKKN2 FSD to full-length (FL) WFIKKN2 for their ability to bind and antagonize GDF8 and GDF11. Surface plasmon resonance (SPR) was used to measure binding affinity where GDF8 and GDF11 were coupled to a CM5 SPR chip. Binding analysis was performed by injecting increasing concentrations of WFIKKN2 FL or WFIKKN2 FSD (Figure 2a-d). Similar to previous reports WFIKKN2 FL bound GDF8 and GDF11 with a high-affinity as demonstrated by the slow dissociation rate101,174. Consistent with the Native-PAGE analysis, WFIKKN2 FSD exhibited a significant interaction with

GDF8 and GDF11, albeit with a much faster dissociation rate as compared to WFIKKN2 FL. The association and dissociation rate constants and equilibrium constant, KD are reported in Table 1. Overall, WFIKKN2 FL exhibited a high-affinity interaction with GDF8 (KD=0.74nM) and GDF11 (0.24nM) while the WFIKKN2 FSD was approximately 1000-fold weaker (KD=0.66µM for GDF8 and 0.12µM for GDF11). WFIKKN2 FL and

WFIKKN2 FSD were able to bind TGFβ1, albeit weakly, but have no affinity for activin A, consistent with previous studies174. We next sought to determine if the affinity of WFIKKN2 FSD for GDF8 and

GDF11 was sufficient to inhibit signaling. Using a luciferase reporter assay responsive to GDF8 and GDF11, we titrated increasing concentrations of WFIKKN2 FL and WFIKKN2 FSD. While WFIKKN2 FL inhibited

GDF8 and GDF11 signaling with an IC50 value of 0.34nM and 0.13nM, respectively, WFIKKN2

FSD exhibited a significantly reduced inhibition of GDF8 and GDF11 with an IC50 value of 0.85µM and

0.28µM, respectively (Figure 2d), similar to the 1000x difference in binding affinity determined by SPR.

This demonstrates that while the WFIKKN2 FSD binding affinity is severely reduced compared to full length WFIKKN2, it is still able to antagonize GDF8 and GDF11, albeit with weaker potency.

Figure 2: Binding and antagonism of WFIKKN2 FSD to GDF8. a-d) SPR of WFIKKN2 Full length (FL) and WFIKKN2 FSD over immobilized GDF8, GDF11. Red: experimental binding trace Black: data fit using a 1:1 binding model a) WFIKKN2 FL applied to immobilized GDF8. b) WFIKKN2 FSD applied to immobilized GDF8. c) WFIKKN2 FL applied to immobilized GDF11. d) WFIKKN2 FSD applied to immobilized GDF11. e) WFIKKN2 FL (red) or WFIKKN2 FSD (black) applied to immobilized TGF-β1. WFIKKN2 FL (grey) or WFIKKN2 FSD (orange) applied to immobilized Activin A. f) Luciferase assay using (CAGA)12 HEK293 cells treated with GDF8/GDF11 alone or titrated with increasing concentrations of WFIKKN2 FL or WFIKKN2 FSD. Data were fit using nonlinear regression to determine an IC50 and plotted as mean ± SD conducted at least twice with each point measured in triplicate.

Table 1: Surface Plasmon Resonance of WFIKKN2 interactions with GDF8. Values shown represent the association constant (ka), dissociation constant (kd), equilibrium dissociation constant (KD)) of WFIKKN2 Full length or the FSD injected over primary amine coupled GDF8 or GDF11. WFIKKN2 GDF8 ka GDF8 kd GDF8 KD GDF11 ka GDF11 kd GDF11 KD (1/Ms) (1/s) (1/Ms) (1/s)

WFIKKN2 Full 3.37x105 2.48x10-4 0.74nM 6.00x105 1.34x10-4 0.24nM Length WFIKKN2 FSD 6.9x103 4.6x10-3 0.66µM 3.52x104 4.25x10-3 0.12µM

Does the WFIKKN2 FSD bind GDF8 at the type II receptor epitope?

To signal, GDF8 binds the extracellular domain of the type II receptor, ActRIIB with high affinity

175,176 (KD ~1nM) . Two ActRIIB receptors are expected to bind GDF8, or GDF11, at each knuckle region of the dimeric ligand, similar to the observed binary crystal structures of ActRIIB in complex with activin A and BMP7177,178. Previous studies have demonstrated that full-length WFIKKN2 interferes with type II receptor binding130. To determine if WFIKKN2 FSD could interfere with type II receptor binding we performed a competition analysis using Native-

PAGE. GDF11 mixed with WFIKKN2

FSD was titrated with increasing Figure 3: Competitive binding between WFIKKN2 FSD and ActRIIB to GDF11. a) Native-PAGE analysis using pre-formed WFIKKN2 FSD:GDF11 complex mixed with increasing amounts of ActRIIB concentrations of the extracellular domain (ARIIB). b) Reduced SDS gel of WFIKKN2 FSD, ARIIB, GDF11 and excised bands represented in (a) by a black box. c) Co-injection SPR of ActRIIB (ActRIIB-ECD) and analyzed binding experiment. First, ActRIIB-Fc was captured onto a protein A sensor chip and baseline was normalized. Next, GDF11 was injected by Native-PAGE (Figure 3a). GDF11 was used forto analysisform the binarysince ActRIIB:GDF11the complex complex. bands Subsequently,bound to WFIKKN2 WFIKKN2

FSD and ActRIIB-ECD are more distinct than those with GDF8. Our results show that ActRIIB-ECD easily displaces WFIKKN2 FSD leading to the formation of the ActRIIB-ECD:GDF11 complex (Figure 3a). A higher order complex that contained all three components was not visible indicating that the WFIKKN2 FSD and ActRIIB-ECD were mutually exclusive and likely bind to the same position on the ligand. Bands

81 corresponding to WFIKKN2 FSD:GDF11 and ActRIIB-ECD:GDF11 were excised and verified by SDS-PAGE gel under reducing conditions (Figure 3b). These results indicate the WFIKKN2 FSD and ActRIIB compete for the same binding interface of GDF11.

To further test the idea that WFIKKN2 FSD and ActRIIB bind GDF8 and GDF11 at similar locations competition experiments were performed using SPR. Experiments were conducted using the extracellular domain of ActRIIB fused to the Fc portion of an antibody (ActRIIB-Fc). ActRIIB-

Fc was captured onto a protein A chip followed by an injection of GDF11 to form the binary

ActRIIB-Fc:GDF11 complex. We next injected

WFIKKN2 FSD at 500nM and observed no increase in binding, indicating that the

WFIKKN2 FSD could not bind GDF11 in the presence of ActRIIB (Figure 3c). Interestingly, when similar experiments were performed with WFIKKN2 FL a noticeable mass increase was observed indicating that WFIKKN2 FL was able to bind GDF11 that was already in complex with ActRIIB (Figure 3c).

Crystallization and Structural determination Figure 4: WFIKKN2 FSD structure and surface hydrophobicity. a) of WFIKKN2 FSD. Ribbon diagram depicting WFIKKN2 FSD with structural components (α-helix and β-sheets) in red, flexible loops in grey, and disulfide bonds shown as yellow sticks rotated about the Y-axis 180o. b) Since the WFIKKN2 FSD can bind and Hydrophobicity of the GASP1 Fs domain surface, red being the most hydrophobic and white being the least hydrophobic using Color_H antagonize GDF8 and GDF11, we next wanted pymol script(46). Structures are in the same orientation as shown in (a).

Table 2: X-ray crystallographic statistic for structural determination. Overall statistics of native and heavy atom datasets for WFIKKN2 FSD structural determination. Parenthesis represent the highest resolution shell. Data Collection Native K2PtCl4 Resolution (Å) 41.80-1.39 (1.42-1.39) 41.85-1.54 (1.56-1.54) Wavelength 0.98 1.07 # of Observations 277,020 259,934 # of Unique Reflections 21,677 29,887 # of Heavy Atom Derivatives (Pt) 9 Space Group P43 21 2 P43 21 2 Unit Cell a,b,c (Å) 46.5, 46.5, 95.7 46.6, 46.6, 95.1 α,β,γ (o) 90, 90, 90 90, 90, 90 Completeness (%) 99.7 (99.8) 99.9 Redundancy 12.8 24.1 Anomalous Completeness (%) 99.9 Anomalous Redundancy 13.4 Rmeas 0.063 (0.184) 0.239 (3.508) Rpim 0.024 (0.080) 0.05 (0.943) Mean ((I)/σ(I)) 25.5 (8.7) 13.2 (1.1) BAYES-CC 55.5 ± 15.0 FOM initial (after DM) 0.357(0.61) Refinement Resolution (Å) 41.80-1.39 (1.42-1.39) Reflections (total/free) 21,708 Cutoff for Refinement F>0σ Rwork/Rfree, % 16.45/18.01 Atoms total/protein 770/598 Root mean square deviations Bonds (Å) 0.004 Angles (°) 0.75 Average B Factors (Å2) 17 Amino Acids 13.97 Ligands 25.6 Water 32.26 Wilson B Factors (Å2) 11.5 Ramachandran plot Favored (%) 98.68 Allowed (%) 1.32 Outliers (%) 0.00 Clashscore 2.57

Footnote: Rmeas, overall measure of error between multiple measurements of a reflection within I+/I-, independent of redundancy. Rpim, standard error of the mean for intensity measurements within I+/I-. FOM, Figure of Merit, DM, Density Modification

풏 풏 ퟏ 풏 หۄ 푰ۃ− ห σ ට σ ห푰ۄ푰풉풌풍ۃ−σ풉풌풍 ට σ ห푰풉풌풍풋 풏−ퟏ 풋=ퟏ 풉풌풍 풏−ퟏ 풋=ퟏ 풉풌풍풋 풉풌풍 푹풎풆풂풔 = 푹풑풊풎 = σ풉풌풍 σ풋 푰풉풌풍풋 σ풉풌풍 σ풋 푰풉풌풍풋 to determine the molecular structure and draw comparison to the FSD of follistatin and FSTL3. The X-ray

crystal structure of WFIKKN2 FSD was solved by single isomorphous replacement

with anomalous scattering to 1.39 Å resolution. Data collection and refinement statistics are presented

in Table 2. Figure 4a depicts the overall structure of the WFIKKN2 FSD. Similar to other FSDs, the

structure contains two subdomains; an N-terminal EGF-like portion followed by a Kazal-like protease

inhibitor subdomain. The EGF-like portion contains anti-parallel β-strands (β1, β2), whereas, the Kazal-

83 like protease inhibitor subdomain contains a central helix (1) and another set of anti-parallel β-strands

(β3, β4) that caps the Kazal subdomain. Similar to the FSDs from follistatin and FSTL3, the WFIKKN2 FSD contains 5 conserved disulfide bonds. Two are positioned in the EGF subdomain that connect the β1 and

β2 strands (Figure 4a). The other three are located in the Kazal subdomain - two link the 1 to the segment connecting the EGF and Kazal subdomains and one links β3 to the C-terminus.

WFIKKN2 FSD mutagenesis and inhibition.

Utilizing the structure, we wanted to identify residues that are important for the interaction with

GDF8 and GDF11. TGF-β Ligands, including BMP2, BMP4, BMP7, activin A, activin B, GDF8 and

GDF11, predominantly interact with receptors and extracellular antagonists using hydrophobic interactions129,179. Therefore, we hypothesized that surface exposed hydrophobic residues of WFIKKN2

FSD might be important for ligand binding and antagonism. Several surface exposed hydrophobic residues in both the EGF and Kazal subdomains were readily apparent (Figure 4b)180. Residues were selected for alanine mutagenesis, including two that are positioned in the EGF domain (F109,

W121) and three positioned within the Kazal subdomain (F139, F153, and I163) (Figure 5a). In order to determine if these residues were important for

WFIKKN2 antagonism, we generated single point mutations Figure 5: Full length WFIKKN2 mutant selection and inhibition a) Ribbon diagram of WFIKKN2 FSD, of the selected labeled sticks represent residues mutated to alanine and colored based on the inhibition curve. b) IC50 curves generated via luciferase assay with increasing amount of WFIKKN2 mutants titrated residues in the full against a constant concentration of GDF8. c) IC50 values of mutants tested in (b) and depicted on the WFIKKN2 FSD structure in (a). length WFIKKN2. WT and mutant versions of WFIKKN2 FL were expressed in human embryonic kidney

84 cells (HEK-293F), purified, and tested for GDF8 inhibition in a cell-based luciferase reporter assay.

HEK293T-CAGA cells were treated with 0.62nM GDF8 and titrated with increasing concentrations of either purified wildtype or mutant WFIKKN2 FL (Figure 5b). Data were fit to a dose-response curve to determine an IC50 for WFIKKN2 FL and mutations. Mutations within the EGF domain (F109A and

W121A) resulted in little to no change in the antagonism of GDF8. However, WFIKKN2 mutants tested within the Kazal subdomain (F139A, F153A, and I163A) all displayed weaker GDF8 inhibition. The most striking effects were seen with the mutations F139A and F153A which were 10- and 18-fold weaker than WT WFIKKN2 (Figure 5). Thus, the three residues that had a negative impact on GDF8 antagonism are located on the highly hydrophobic face of the WFIKKN2 FSD Kazal subdomain as shown in Figure 4.

Comparison of the WFIKKN2 FSD to the FSDs of Follistatin.

Using the structure of WFIKKN2 FSD, we can draw structural comparisons to the different FSDs in

FS288 and FSTL3. Since the FSD1 and FSD2 in both FS288 and FSTL3 share similar structures, the comparison focuses on the FSDs within FS288. The WFIKKN2 FSD and the FSDs from FS288 and FSTL3 contain a similar domain architecture with EGF and Kazal protease inhibitor subdomains, including the spacing and alignment of the 5 disulfide bonds. In addition, all of the FSDs contain two highly conserved residues, T146 and Y147 highlighted in Figure 6a, which are conserved in the broader Kazal domain of protease inhibitors (e.g. ovomucoids and Serine protease inhibitor Kazal-type

1)181–183. Despite these conservations, there are differences in the relative position of the subdomains making it challenging to perform an overall alignment of the FSDs.

However, the Kazal subdomains exhibit similar structures where alignment results in a root mean square of 3.1, 2.1 and 2.9 Å2 for FSD1, FSD2 and FSD3 of follistatin, respectively (40  positions), when aligned to

WFIKKN2 FSD. The moderate differences of the C positions between the Kazal subdomains localize in the loops connecting the 1-helix and the position of β-strands 3 and 4. Despite their similar

85 structure, the Kazal subdomain of WFIKKN2 FSD only contains 2 of the 3 β-strands found in other FSDs

(Figure 6b).

Figure 6: Comparison of WFIKKN2 FSD and Fs288 FSD1-3 a) Sequence alignment of WFIKKN2 FSD to Fs288 FSDs with the EGF and Kazal subdomains annotated, rectangles represent beta strands, cylinders represent alpha helices, black bars are conserved disulfide bonds, asterisks are conserved residues. Residues determined to be important for GDF8 binding are highlighted. Underlined residues are involved in the EGF- Kazal interaction b) WFIKKN2 FSD (Red) to FSD1 (Blue), FSD2 (Green), and FSD3 (Yellow) of Fs288 were aligned using the Kazal subdomain. (top left) Schematic representation depicting the relative orientation of the Kazal and EGF subdomains. The inlet shows the 2Fo-Fc (2α) map of the EGF-Kazal interaction of WFIKKN2 FSD. Dotted lines represent missing segments in the crystal structures. Residues involved in the EGF-Kazal interaction are indicated in ball-and- stick. (far right) Alignment of WFIKKN2 and Fs288 FSD1-3 only depicting the alpha helix and β1-2 of the EGF. The relative difference in orientation between the tip of the EGF domain and alpha-helix is shown by a black line. While the individual subdomains are structurally similar, there are significant conformational differences in the relative position of the subdomains. In retrospect, this explains why using the available

FSD structures (FSD1-3) as search models failed to provide a molecular replacement solution for the

WFIKKN2 FSD X-ray diffraction data. Thus, we aligned the structures using only the Kazal subdomain (Figure 6b). Comparison of the structures shows that different FSDs can adopt either an “open” or “closed” conformation. An open conformation is observed in FSD1 and WFIKKN2 FSD where the EGF β-strands are extended away from the Kazal domain, which creates a more linear appearance of the FSD domain. In contrast, a closed, more compact conformation is observed in FSD2 and FSD3 where the EGF domain, specifically β-strand 1, interacts with the Kazal domain. In

FSD2, T178, V180, and V181 within β-strand 1 of the EGF domain form hydrophobic interactions with the Kazal subdomain (Figure 6b). Similarly, in FSD3, K256, L258, and F261 of β-strand 1 in the EGF domain interact with the Kazal domain. While the FSD of WFIKKN2 is in the open conformation, it should be noted that the N-terminus contains tandem glutamine residues (Q113 and Q114) that wrap back into the Kazal subdomain to interact with the 1-helix. Residues at the interfaces of all the FSD are highly conserved across species, including Q113 and Q114, suggesting that they might serve a role to stabilize the different conformations of the FSDs.

Discussion

WFIKKN proteins are evolutionarily conserved in all vertebrates and can be found in other organisms such as sea urchins and worms100. When compared to other extracellular protein antagonists of the TGF-β family, WFIKKN proteins are remarkably specific for the ligands GDF8 and GDF11130,174. While the basis for this specificity remains unknown, it is clear that the FSD plays an important role in binding and antagonism91,101. For example, a construct containing only the WF domains of WFIKKN2 is sufficient for antagonism of GDF8, and the FSD of WFIKKN1 has a higher affinity for GDF8 than the Kunitz2 and

Netrin domain91,101. Interestingly, comparing the different domains of WFIKKN2, the FSD is the most conserved domain across mammalian species, experiencing the lowest rate of substitution100.

In other ligand antagonists, such as follistatin and FSTL3, FSDs play an important role in ligand binding and antagonism95,97,136. However, whether the FSD from WFIKKN serves a similar role remains to be determined. To address this, we produced recombinant WFIKKN2 FSD and characterized its interaction with GDF8 and GDF11.

A complex between WFIKKN2 FSD and GDF8 or GDF11 was observed using Native-PAGE analysis, which SPR experiments revealed to involve sub µM affinity interactions. Using Native-PAGE we also demonstrated that the addition of the type II receptor, ActRIIB, was able to dissociate a WFIKKN2

FSD:GDF11 complex, suggesting a direct competition between WFIKKN2 FSD and ActRIIB. Consistent with

87 these results, SPR experiments showed that GDF11 in complex with ActRIIB-Fc was unable to bind

WFIKKN2 FSD. Therefore, similar to their role in follistatin and FSTL3, the FSD of WFIKKN2 appears to block or at least compete for type II receptor binding on the ligand. However, full length WFIKKN2 retained the ability to bind the GDF11:ActRIIB-Fc complex. This interaction is likely due to other domains of WFIKKN2 associating with regions of GDF8 and GDF11 apart from the type II binding site. Thus, while the other domains of WFIKKN2 are important for potent antagonism, the WFIKKN2 FSD is sufficient to antagonize GDF8 and GDF11 by blocking type II receptor binding. Whether WFIKKN2 FSD or full-length WFIKKN2 interferes with type I receptor binding has not been resolved and is complicated by the low-affinity interaction of the type I receptor with GDF8 and GDF1139.

To help draw comparisons to the FSD of follistatin, we solved the crystal structure of WFIKKN2

FSD. The structure reveals a similar domain architecture with an N-terminal EGF subdomain and a C- terminal Kazal subdomain as previously observed for other FSDs95,97,129,184. Using the structure, we were able to identify hydrophobic residues at the surface that contribute to GDF8 antagonism. Single point mutations within the FSD of WFIKKN2 FL reduced the capacity to inhibit GDF8, presumably by reducing the ability of WFIKKN2 FL to interact with the type II receptor binding interface. These results are consistent with antagonists across the family which use hydrophobic interactions to engage the ligand surfaces important for binding type II receptors. However, it is possible that the residues identified could be important for interacting with other domains within WFIKKN2 FL, functioning to stabilize the WFIKKN2-ligand complex.

Comparison of WFIKKN2 FSD across 54 species reveals a high sequence identity (68%) (Table 3).

Interestingly, the residues identified (F139, F153, I163) within the FSD to be are important for antagonism of GDF8 are highly conserved. Further, these residues are not conserved in other FSDs (Figure 6), indicating they are not necessary to the structural integrity of the Kazal subdomain. Thus, the decrease in

Table3: CLUSTAL O(1.2.4)63 multiple sequence alignment WFIKKN2. WFIKKN2 sequences were identified using uniprot64 and Supplement1: CLUSTAL O(1.2.4)1 multiple sequence alignment WFIKKN2. WFIKKN2 sequences were identified using uniprot2 and aligned used alignedClustal sequence used Clustal alignment sequence web software. alignment Across web all 54 software. species the Across amino acid all 54identity specie wass 68.1%the amino. Sequences acid areidentity in FASTA was format 68.1%. with Sequences the uniprot are inaccession FASTA numberformat followed with the by uniprot the species accession abbreviation number in bold. followed Asterisks byat th thee bottom species of the abbreviation sequences identify in bold. identical Asterisks amino at acid the residues. bottom of the sequencesUnderlined residues identify are identical suspected amino to be structural acid residues. among Kazal Underlined domains asresidues a whole ,are Q113 susp andected Q114 suspectedto be structural to stabilize among the EGF Kazal-Kazal domains interaction as a bolded. Disulfide cysteines are connected by black lines. whole, Q113 and Q114 suspected to stabilize the EGF-Kazal interaction bolded. Disulfide cysteines are connected by black lines.

>Q8TEU8_HUMAN ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >Q7TQN3_MOUSE ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACSKGITLSVVTCRY >D3Z9K5_RAT ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGVTLSVVTCRY >F1MRR8_BOVIN ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >F7CRY5_HORSE ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >E1C6W9_CHICK ATCDRFMCIQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNKCYMDAEACIKGITLNVVTCRY >F7GFJ7_MACMU ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >E2RSS4_CANLF ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >D2HTT7_AILME ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >G3WTE3_SARHA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLTVVTCRY >G3RT83_GORGO ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >M3YVT3_MUSPF ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >G1P586_MYOLU ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >F1RTB4_PIG ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >I3MTQ8_ICTTR ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSRGITLAVVTCRY >U3JFA3_FICAL ATCDHFMCIQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNKCYMDAEACIKGITLSVVTCRY >H2NTM3_PONAB ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A0D9QZL6_CHLSB ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >F6VGY3_MONDO ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A286XEP6_CAVPO ATCDHFMCPQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2I3LXY3_PAPAN ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2K5VVF3_MACFA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >G3UN82_LOXAF ATCDHFMCLQQGSECDIWDGQPVCKCKERCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >G1N923_MELGA ATCDRFMCIQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNKCYMDAEACIKGITLNVVTCRY >A0A2K6QNA4_RHIRO ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGISLAVVTCRY >A0A2K6UIM8_SAIBB ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACSKGITLAVVTCRY >A0A2K5QS22_CEBCA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACSKGITLAVVTCRY >A0A2K5YEH8_MANLE ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2K5CH66_AOTNA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACSKGITLAVVTCRY >F7ISD7_CALJA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACSKGITLAVVTCRY >A0A2K5MJ03_CERAT ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2K5IBZ3_COLAP ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2K6GQ51_PROCO ATCDHFMCPQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2K6D706_MACNE ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A1U7QW10_MESAU ARCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGISLAVVTCRY >G1K9Q5_ANOCA ATCDRFMCIQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNKCYMDAEACTKGITLNVVTCRY >H0WZE7_OTOGA ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A1L8EN30_XENLA ATCDLFMCTQQGSECDIWDGQPICKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACTKGITLTVVTCKY >H2TAP0_TAKRU ATCDKFMCTQQGSECDIWDGQPVCKCRDRCEREPHFTCASDGMTYYNKCYMDAEACSKGISISEVTCRY >E7FCX1_DANRE ASCDQFMCTQQGSECDIRDGQPVCKCRDRCEREPQFTCASDGMTYYNKCYMDAEACSKGISLSVVSCRF >A0A1W5AU64_9TELE ATCDKFMCTQQGSECDIWDGQPVCKCRDRCEREPNFTCASDGMTYYNRCYMDAEACSKGISLSVVTCRY >A0A1S3GZ71_DIPOR ATCDHFMCVQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A1S2ZWC7_ERIEU ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >F7CCB5_XENTR ATCDLFMCTQQGSECDIWDGQPICKCKDRCEKEPSFTCASDGLTYYNRCFMDAEACTKGITLTVVTCKY >H3A6B1_LATCH ATCDRFMCTQQGSECDIWDGQPVCKCKDRCEKEPNFTCASDGLTYYNKCYMDAEACSKGITLNVVTCRY >A0A2I4C8R5_9TELE ATCDKFMCAQQGSECDIWDGQPVCKCRDRCEREPHFTCASDGMTYYNKCYMDAEACSKGISISEVTCRY >A0A1U7RT46_ALLSI ATCDRFMCIQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNKCYMDAEACIKGITLNVVTCRY >A0A2U3YAL5_LEPWE ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2U3VK32_ODORO ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2U3UZE4_TURTR ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2Y9EKV3_PHYCD ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2Y9I8S8_NEOSC ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2Y9LWC5_DELLE ATCDHFMCLQQGSECDIWDGQPVCKCRDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY >A0A2I3T853_PANTR ATCDHFMCLQQGSECDIWDGQPVCKCKDRCEKEPSFTCASDGLTYYNRCYMDAEACSKGITLAVVTCRY * ** *** ******** **** *** *** ** ******* **** * ****** * * * antagonism is not likely due to a misfolded FSD within WFIKKN2. Taken together, these results support a functional rather than structural role of F139, F153 and I163 in the binding and antagonism of GDF8 and

GDF11. Interestingly, when compared to follistatin and FSTL3, WFIKKN2 interacts with the ligands

89 through a different surface of the FSD. The WFIKKN2 FSD residues, F139, F153 and I163 are located within the Kazal subdomain, centered on the hydrophobic face of the 1 opposite of the N-terminal loop (Figure 4, Figure 7). In FSD1 and FSD2, residues that interact with the ligands are located in a completely different location of the domain, especially for FSD2 where the interaction surface is on the opposite side as compared to WFIKKN2 FSD (Figure 7).

While it is possible that the conformational differences observed between the EGF and Kazal subdomains of FSDs are related to crystallization, the linker region is well ordered in the crystal lattice and exhibits similar temperature factors throughout the structure suggesting the conformation of the domain is stable. Consistent with the FSD having a stable structure, the open conformation of the follistatin FSD1 was observed in both the bound and unbound state, indicating that the domain structure is not affected by ligand binding and or alternative crystallization conditions.

In addition to WFIKKN, follistatin and FSTL3, FSDs are found in several proteins with various roles in TGF-β signaling. For instance, FSDs are also found in agrin, tomoregulin and FSTL1, all of which are implicated in regulating TGF-β ligand activity185–187. This raises the question as to the general function of the FSD and whether the domain serves a common role in ligand interactions. While the FSD in both

WFIKKN and follistatin/FSTL3 both interact at the type II receptor interface, differences are readily apparent in both their structures and the placement of residues utilized for this interaction. Thus, it does not appear that the FSD represents a common binding domain, but rather a stable scaffold that allows the presentation of surface exposed hydrophobic residues that can be important for protein-protein interactions.

Given the differences in domain architecture and the differences in the position of surface residues important for ligand binding, we anticipate that the FSD of WFIKKN2 does not resemble the ligand binding scheme of any of the three FSDs of follistatin. Certainly, resolution of WFIKKN2 FSD or FL in

complex with a ligand will ultimately resolve these differences and help to understand how WFIKKN

specifically engages GDF8 and GDF11 over other TGF-β ligands.

Figure 7: Binding interface of WFIKKN2 FSD is distinct from the FSDs of FS288. a) Structure of Fs288 bound to GDF8 (PDB 3HH2(23)) The central GDF8 dimer is depicted as ribbon and colored orange and wheat. One Fs288 molecule (white) is shown as a surface representation. The second FS288 molecule is shown as a surface and colored based on the schematic, FS288 FSD1- FSD2 surface that interacts with GDF8 is colored grey. b) WFIKKN2 FSD (red), FS288 FSD1 (blue), and FS288 FSD2 (green) are shown in similar orientation based on alignment of the Kazal subdomain. Residues within each FSD important for GDF8 antagonism are displayed. c) Rotation of (b) along the vertical axis by 90o.

Experimental procedures

WFIKKN2 FSD production and purification.

Mouse WFIKKN2 FSD (residues 104-172, 97% identical to human) with a cleavable N-terminal 6x-

His tag was cloned into pET28a(+) and expressed in BL21 Rosetta cells. Cells were spun down and resuspended in PBS before sonication for cell lysis. Pellets were washed two times with PBS, 0.1%

TritonX100 followed by a final wash with PBS. Inclusion bodies containing the WFIKKN2 FSD were solubilized using 10mM sodium tetrathionate, 100mM sodium sulfite, 100mM Tris pH 8.5, 8M urea, and

100mM DTT. Solubilized inclusion bodies were dialyzed into 4M Urea, 50mM sodium acetate pH4.5,

100mM NaCl, 50mM Tris, 15mM imidazole and rapidly diluted into refolded buffer containing 100mM Tris pH 8.5, 150mM NaCl, 1mM EDTA, 0.1mM oxidize glutathione, 0.5mM reduced glutathione, and 0.5M arginine to a final concentration of 0.1mg/ml. After three days protein was diluted 1:5 in Ni-NTA running buffer: 50mM Tris pH 8, 500mM NaCl, and 15mM imidazole and applied to a HisTrapTM Excel column (GE

Healthcare). WFIKKN2 FSD was eluted with 500mM imidazole and dialyzed into 10mM HCl prior to separation on a C18 reverse phase column (phenomenex).

ActRIIB-ECD production and purification.

ActRIIB-ECD was produced within SF+ insect cells using pFastBac expression plasmid. Purification was conducted as previously described178. In short, ActRIIB-ECD containing a 6xHis tag is purified from

SF+ condition media using a Ni-NTA affinity column. Bound protein was eluted with 20mM Tris pH 7.4,

500mM NaCl containing 500mM imidazole and subsequently applied to an S75 size exclusion column.

Fractions containing the ActRIIB-ECD were pooled and used for Native-Page analysis.

Native-PAGE gel and Western blot.

3ug of GDF11/8 was mixed with WFIKKN2 FSD at different molar ratios, starting with 4:1

(WFIKKN2 FSD:GDF11/8) and WFIKKN2 FSD was serially diluted 1:2 5x for a final ratio of 0.13:1. Native-

PAGE gels (12%) were run at 20oC for 140min at 110V and stained using colloidal coomassie. For staining,

92 gels were fixed using 40% EtOH and 10% Acetic Acid for at least 1 h before washing 3x using dH2O. A working dye solution was used containing 80% colloidal coomassie and 20% Methanol and stained overnight.

Native-PAGE to SDS-PAGE transfer.

Native-PAGE gels were run as described above. However, rather than a titration, one condition was repeated in 7 lanes. One lane was cut, stained, and realigned to the unstained gel. The desired stained band was used as a guide to excise the other 6 unstained bands. Excised gel was then placed in a dialysis bag with 1mL of SDS running buffer and electro-eluted with 180V for 1h. The buffer containing the eluted band was then concentrated and subjected to SDS-PAGE under reducing conditions followed by colloidal staining as described above.

Luciferase assay.

Luciferase assays were conducted in HEK293 (CAGA)12 cells as previously described by our laboratory39,58,91. For the assay, 20,000 cells were seeded in growth media in a 96 well format on poly-D- lysine coated plates (CAT. No. 655940 Greiner Bio-One GmbH, Germany). Cells were grown at 37°C with 5% CO2 until reaching 75-85% confluency. Media was then removed and treated with 100uL of serum free media containing GDF8 in the presence or absence of antagonists. GDF8 was kept at a constant concentration (0.62nM) while antagonists (WFIKKN2 FSD and Full Length) were titrated in using 1:2 dilutions. After 1day cells were lysed using 1x passive lysis buffer (E1941, Promega, USA) on a plate shaker

(800rpm, 20min, 20°C). Lysates were transferred to black and white 96 well plates, 40uL of LAR (E1501 and E1960, Promega, USA) was added. Firefly luminescence was measured using the Synergy H1 Hybrid

Plate Reader (BioTek). All experiments were conducted independently at least twice with all data points being done in triplicate. The concentration of antagonists at which 50% of GDF8 activity is lost or IC50, was calculated using non-linear regression with variable slope using GraphPad Prism 5 software.

Surface Plasmon Resonance.

SPR experiments were conducted on the Biacore T200 microfluidic system. Ligands were primary amine coupled to the GE series S CM5 sensor chip (Ca. No. BR-1005-30) using manufacture’s protocols. GDF8, GDF11, activin A, and TGFβ1 at 1ug/ml was sufficient to achieve ~500 response units bound to the chip. SPR experiments were conducted in HBS-EP buffer (10mM HEPES, pH 7.4, 150mM

NaCl, 3mM EDTA, 0.005% P-20 surfactant (Biacore AB)). WFIKKN2 full length or WFIKKN2 FSD was diluted in HBS-EP buffer to a concentration of 500nM then serially diluted 2-fold 10 times and applied to the chip at 15ul/min. Association was measured for 180seconds followed by 120 seconds of buffer to measure dissociation. The dissociation equilibrium constant, KD, was calculated using Biacore T200 software using a 1:1 binding model. For the receptor binding experiments, ActRIIB-Fc at a concentration of

312.5ng/ml (~4.2nM) was immobilized on a Protein A chip. GDF11 was then applied at a concentration of

250nM for 180s. Subsequently, binding of 125nM and 500nM WFIKKN2 full-length or WFIKKN2 FSD to the receptor-ligand complex was measured.

Structural Determination and Experimental Phasing.

Crystals of WFIKKN2 FSD were grown in 2.5M AmNO3 and 0.1M NaCitrate pH 4.6 via hanging drop vapor diffusion at 4mg/ml. Rectangular crystals grew to a size of 100um x 200uM. Diffraction data were collected at the Advance photon source (APS) beamline GM/CA 23-ID-D at Argonne National Laboratory.

Molecular replacement techniques for phasing were unsuccessful requiring experimental phasing via single isomorphous replacement with anomalous scattering (SIRAS). In short, crystals were soaked in 5mM KCl4Pt for 96h before being transferred to a cryogenic solution containing mother liquor and 35% PEG 550 and flash frozen. Heavy atom positions were identified using phenix AutoSol188,189.

Phasing and model refinement were carried out using autosol and phenix.refine190. Model validation was performed using the program molprobity191,192. Coordinates are available at the Protein Data Bank,

94 accession code 6MAA. Hydrophobicity depicted in Figure 4 was determined using Color_H pymol script180.

WFIKKN2 FSD and Fs288 Alignment and structural representation.

Sequence alignments between WFIKKN2 FSD and other FSDs were conducted using T-

Coffee193. Structural alignments were conducted using ce-alignment within Pymol194. Alignments were conducted between the Kazal subdomain of each FSD starting with the conserved cysteine (residue 132 in WFIKKN2) to the end of the FSD (residues 172 in WFIKKN2). Residues used for each Fs288 are as follows,

FSD1: 119-166, FSD2: 194-242, FSD3: 271-318. Supplemental sequence alignments were conducted using

CLUSTAL O (1.2.4) through uniprot195,196.

Mutagenesis and purification full length WFIKKN2.

Full Length WFIKKN2 was purified as previously described91. In short, conditioned media from CHO cells stably expressing WFIKKN2 was collected and subjected to a butyl-sepharose and heparin columns. Elutions containing WFIKKN2 were dialyzed into 50mM Tris pH 7.4, 20mM NaCl, 1mM EDTA and applied to a MonoQ 10/100 GL column and eluted with a linear gradient of NaCl. Following Mono

Q, elutions were subjected to S2000 size exclusion chromatography to obtain pure WFIKKN2. Full length

WFIKKN2 was cloned into pcDNA4 mammalian expression vector and subjected to site-directed mutagenesis. WFIKKN2 mutants were transfected and produced via HEK 293F cells. Following expression, WFIKKN2 mutants were purified similar to full length WFIKKN2 produced in CHO cells.

Chapter V

Summary, conclusions, and future directions

Dissertation Summary

The studies described in Chapters II-IV describe our work that determined the molecular mechanisms surrounding GDF8 latency, activation, and antagonism. Specifically, we showed (1) how specific residues within the GDF8 prodomain are critical for maintaining a latent procomplex,(2) that residues adjacent to the GDF8 tolloid cut site are important for tolloid recognition and proteolysis, (3) the crystal structure of the follistatin domain of WFIKKN2 and how it contributes to GDF8 antagonism by binding the type II receptor binding site. While these studies have provided invaluable insight into the mechanisms of GDF8 signaling that may be therapeutically targeted to restore altered muscle mass homeostasis, some aspects of GDF8 signaling remain unknown and are discussed below.

Summary of prodomain mediated latency of GDF8

First, we sought to determine what molecular mechanisms contribute to GDF8 latency. Only a handful of ligands within the TGFβ superfamily form a high affinity, inactive complex with their prodomain: the TGFβ subclass and two activin subclass members GDF8 and GDF1141,55,57,197. Prior to our work, it was hypothesized that GDF8 utilized similar molecular mechanisms for latency as TGFβ1, which have been extensively characterized55,58,66,79,113. Through low resolution SAXS analysis, we determined that unlike the compact or “closed” TGFβ1 procomplex, the GDF8 procomplex was elongated or “open”, presumably to allow access by tolloid metalloproteases for activation.

Despite differences in overall conformation between TGFβ1 and GDF8 procomplexes, sequence homology and secondary structure prediction revealed that they likely shared similar inhibitory elements including the α-1 helix, latency lasso, and fastener region. Using site directed mutagenesis, we were able to identity a number of residues that contributed to latent GDF8 activation and latency. Interestingly, mutants within the α-helix were able to remove latency. Specifically, I56E was able to overcome the need

97 for tolloid mediated activation. We hypothesized that removal of GDF8 latency was due to a weakened prodomain-ligand interaction, which would lead to prodomain dissociation upon receptor binding. This hypothesis was supported by inhibition and thermal shift assays that revealed prodomain mutants formed a less stable complex with the mature signaling domain, subsequently removing the ability of the prodomain to form a latent complex.

Mutations within other inhibitory elements (such as the fastener region) still required tolloid for activation, including Y111A, H112A, and I56A. Interestingly, these GDF8 mutants were significantly more active than WT GDF8 following tolloid cleavage. This is likely due to a weakened triggered state described by the Springer Laboratory73. In short, the prodomain does not immediately dissociate following tolloid cleavage, but instead remains bound in a “triggered” state. The prodomain is then more readily displaced when it is in the triggered state and dissociates upon receptor binding. This triggered state can also be achieved with acid-activation that is described in Chapter II. When either I56E or H112A were administered to mice by AAV injection (performed by the Craig Harrison laboratory), there was a significant decrease in the tibialis anterior mass/body mass ratio and a reduction in myofiber size. Overall, these results indicate that the prodomain mutants resulted in a more active GDF8 that led to muscle atrophy and corroborate the in vitro data describing a more active GDF8 procomplex.

While we were conducting our mutagenesis study, the structure of the GDF8 procomplex was solved by the Marko Hyvonen laboratory57. Using this structure, we identified the residues we mutated and determined their contribution to latency. For example, I56 is buried within the palm of the GDF8 mature domain, and the fastener residues (Y111 and H112) form a direct interaction with the α-1 helix.

Thus, mutating I56 would destabilize the prodomain-ligand interaction and remove latency while mutating the Y111 or H112 fastener residues would destabilize the fastener:α-1 helix interaction, supporting our experimental results described in Chapter II. Together, these data help describe the

98 molecular mechanisms behind GDF8 prodomain mediated latency and how mutations within the prodomain are able to destabilize a the prodomain:ligand complex.

Future directions of prodomain mediated latency

The goal of understanding GDF8 latency is to determine how GDF8 is regulated in vivo with the hope of unveiling new therapeutic avenues for GDF8 inhibition to treat muscle wasting. Experimental results in Chapter II identified key residues that are critical to prodomain mediated latency within the inhibitory elements of the GDF8 prodomain. However, more research needs to be conducted to elucidate the role of other prodomain elements. This study focuses on the inhibitory elements between residues

24-112, but the role of the “shoulders” (residues 113-266) need to be elucidated further. Specifically, it is currently unknown if and how the shoulders dictate tolloid recognition for activation, dimer formation, and localization. Currently, attempts have been made by other laboratories to determine the minimal prodomain sequence required for latency in the hope to develop peptides capable of GDF8 inhibition127.

However, it has been shown that a reformed complex with GDF8 and its prodomain does not yield a fully latent complex but instead retains a large amount of signaling capabilities58,124. As such, an excess of prodomain would be required to achieve complete antagonism when using it as an inhibitor.

A novel, promising approach being pursued by our lab to achieve GDF8 inhibition involves the development of a dominant negative GDF8 complex during endogenous GDF8 production in vivo. In short, our goal is to produce GDF8 mutants that cannot be activated to reduce or abolish native GDF8 signaling.

We hypothesize that if an inactive GDF8 mutant is expressed with WT GDF8, the two GDF8 monomers will dimerize resulting in a complex that cannot be activated. In fact, Dr. Timothy Cox identified a heterozygous

GDF11 mutation, R298Q, which prevents furin processing159. In humans, this GDF11 mutation causes orofacial clefting. In collaboration with Timothy Cox, we demonstrated that transfecting GDF11 R298Q with WT GDF11 completely abolished GDF11 signaling within our luciferase assays. Further

99 characterization using western blot analysis revealed that GDF11 R298Q was able to dimerize with GDF11, thus forming a dominant negative complex that could not be activated by tolloid. The findings were published in Human Mutations in 2019159.

Using the same methodology, we made a similar mutant within GDF8 R266A. Like GDF11, when

GDF8 R266A was transfected with WT GDF8, we saw a significant reduction in signaling (Figure 1). Despite

GDF8 and GDF11 being 90% identical within their mature domain, they are only ~53% identical within their prodomains. Due to the high homology between the ligands, we hypothesized that GDF11 R298Q would also be dominant negative to GDF8 and R266A would be dominant negative to GDF11. However, our initial experiment showed that while GDF11 R298Q can fully inhibit GDF8 and GDF11 signaling, GDF8

R266A is only able to inhibit GDF8 signaling and to a lesser extent than GDF11 R298Q (Figure 1). In addition, the GDF8 mutant D99A (discussed at length in Chapter III with Y94A), is also able to reduce GDF8 signaling but not GDF11 signaling. Interestingly, when using a GDF8 prodomain:GDF11 ligand chimera,

GDF8 mutants R266A and D99A are able to inhibit signaling (Figure 1). Together, these data reveal the

Figure1: Transfection-based luciferase assay to determine dominant negative effects of GDF8 and GDF11 mutatns. 50ng of GDF8, GDF11, or the GDF8pro:GDF11 mature domain chimera, with 50ng of Furin and Tll2 is transfected with either 50ng of empty vector (Black), 50ng of GDF8 D99A (Red), 50ng of GDF8 R266A, or 50ng of GDF11 R298Q. After transfection, the media is swapped for serum free 24hrs later and luciferase signal is read 24hrs after the media swap. Data presented as mean±SD and was conducted at least twice with experimental triplicate.

100 importance of the prodomain and its high specificity despite the mature domains of GDF8 and GDF11 being 90% identical.

Together, these preliminary data reveal a layer of complexity of prodomain mediated regulation of GDF8 and GDF11 signaling. Currently, the molecular mechanisms surrounding why GDF11 R298Q inhibits both ligands while GDF8 R266A and D99A appear to be specific to GDF8 (or at least a ligand with the GDF8 prodomain) remains to be elucidated. However, due to GDF11s role in regeneration, a highly specific therapeutic that only antagonizes GDF8 signaling can be highly beneficial. Future studies using

AAV delivery of GDF8 R266A and/or GDF11 R298Q into murine skeletal muscle is being discussed as a possible approach to promote muscle growth.

Summary of the GDF8 tolloid cut site and their contribution to activation.

In order for a latent GDF8 procomplex to become activated, it requires an additional proteolytic event mediated by the tolloid family of metalloproteases58,63,73,111. Despite multiple substrates of tolloid being identified, there is no consensus sequence for tolloid cleavage and very little is understood regarding the molecular interactions required for tolloid proteolysis71,149. Prior to this study, only the aspartic acid residue, D99, on GDF8 was known to be required for tolloid cleavage111. Tolloid cleaves the

GDF8 prodomain directly at the N-terminal of D99. When D99 is mutated to an alanine, tolloid is no longer able to activate latent GDF858,111. However, due to the apparent specificity of cleavage, we hypothesized other residues adjacent to the cleavage site aid in recognition and cleavage by tolloid. To address this, we made a series of sequential alanine mutants near the tolloid cut site and were able to identify a number of residues that, when mutated, removed tolloid mediated activation. Further characterization revealed that mutations were directly impacting the ability of the proteolytic astacin domain of tolloid to proteolyze the GDF8 prodomain. In addition, similar to D99A, one mutant, Y94A, was able to have a dominant negative affect on WT GDF8 when both constructs were transfected in a luciferase assay. Together, these

101 data demonstrated that while the aspartic acid in the P1’ position is required for tolloid processing, other molecular mechanisms are utilized for tolloid mediated activation of latent GDF8.

Future directions to further understand GDF8 activation.

Currently, our lab has been able to purify and refold the proteolytic astacin domain of Tll1.

However, tolloid consists of other non-catalytic domains (namely CUB and EGF domains). These domain architectures are constructed as follows, starting at the N-terminus: prodomain, astacin domain, CUB1,

CUB2, EGF1, CUB3, EGF2, CUB4, and CUB5 (Figure 1 Chapter III pg. 53). Non-catalytic domains have been shown to play important (yet varied) roles in tolloid’s ability to process substrates with a high specificity87,145,151,154,155. For instance, when the CUB4 and CUB5 domain are removed from Tll2, it loses its ability to process another substrate, Chordin (a BMP antagonist)154. In contrast, the mTLD spliceform,

BMP1, which also lacks CUB4 and CUB5, is more active than the full length mTLD variant155. In addition, increased cleavage events occur when only the astacin domain of mTLD is used to process collagen, suggesting that there is a loss in specificity of proteolysis145. Moreover, these data highlight the importance of non-catalytic domains in activity, substrate recognition, and specificity. Recently, we have been able to produce full length Tll2 within HEK293F Expi cells. However, we have not purified Tll2 from condition media. Moving forward, we aim to characterize the effect that non-catalytic domains have on latent GDF8 recognition and activation by generating sequential C to N-terminal truncations. Currently, we have DNA constructs of Tll2 with these truncations. However, more work needs to be done to express and purify full length and Tll2 truncations from conditioned media. In addition, while other substrates seem to prefer one tolloid family member, latent GDF8 is able to be activated by all four: Tll1, Tll2, mTLD, and BMP1. However, it is unknown which tolloid is being utilized in vivo and needs to be elucidated for translational medicine. Within our transfection-based luciferase assays, Tll2 appears to be the most active.

Unfortunately, we cannot rule out that this is not due to expression difference and needs to be controlled

102 for in future experiments. However, it is promising that all Tll2 family members are highly expressed within skeletal muscle144, leading us to hypothesize that Tll2 is likely the primary driver of GDF8 activation in vivo.

Summary of the WFIKKN2 follistatin domain and GDF8 antagonism

Understanding how GDF8 is regulated within the body after activation is essential to the development of inhibitors that can treat muscle wasting. While there are a large number of extracellular antagonists that inhibit GDF8, the WFIKKN family is remarkably specific to only GDF8 and GDF11101,130. As such, it has been a focus of study to determine the molecular mechanisms that contribute to this specificity. Work by our laboratory and others have demonstrated that the follistatin domain (FSD) of

WFIKKN plays a significant role in binding and specificity to GDF8 and GDF11, even though the molecular mechanisms of antagonism were not well understood91,101. Results described in Chapter IV demonstrate that we were able to solve the X-ray crystal structure of the WFIKKN2 FSD. Using this structure, we were able to identify a number of highly hydrophobic amino acids that contribute to GDF8 and GDF11 antagonism. Due to the importance of hydrophobic amino acids on the WFIKKN2 FSD, we hypothesized that it was binding a receptor binding site on GDF8 and GDF11. Using surface plasmon resonance and native-PAGE analysis, we further discovered that the WFIKKN2 FSD was able to antagonize GDF8 and

GDF11 by binding the type II receptor site.

We next compared the WFIKKN2 FSD to other well characterized FSDs within the follistatin family for which the domain is named95,97,129,198. Interestingly, the WFIKKN2 FSD adopts a unique conformation compared to other FSDs (described at length in Chapter IV). In short, the WFIKKN2 FSD adopts a “semi- open” conformation and utilizes a different binding interface than other FSDs. However, it remains unknown how this conformation is able to contribute to specificity or binding. More research needs to be conducted to determine exactly where the WFIKKN2 FSD binds the ligand and in what orientation. Efforts have been made to determine the structure of WFIKKN2 FSD bound to GDF8. Preliminary crystals of a

103

WFIKKN2 FSD:GDF8 complex have been made, but do not diffract to suitable resolution for structure determination.

Future Directions to determine molecular mechanisms of WFIKKN antagonism

We successfully characterized the role that the follistatin domain of WFIKKN2 plays in GDF8 antagonism, but there are still five other domains within WFIKKN whose roles remain unknown, including whey acid protein (W), immunoglobulin (I), two tandem Kunitz domains (KK), and a netrin domain (N)

(Figure 1 Chapter IV pg. 79). Thus, additional experiments are required to fully characterize how WFIKKN2 antagonizes GDF8. Furthermore, WFIKKN1 and WFIKKN2 exhibit different characteristics of GDF8 and

GDF11 antagonism. Notably, WFIKKN2 binds in a 1:1 stoichiometric ratio (1 WFIKKN2:1 GDF8/11 dimer) and WFIKKN1 binds in a 2:1 ratio91. Work by our lab has shown that when the Kunitz 2 and netrin domain are removed from WFIKKN2, it reverts to a 2:1 biding ratio similar to WFIKKN191. While we know the domains that contribute to the differences in binding ratio, we do not know the molecular mechanisms that dictate these binding differences. In addition, WFIKKN1 is able to bind the prodomain of GDF8, whereas WFIKKN2 cannot124. Work done by the Patthy laboratory also demonstrated that WFIKKN1 may potentiate tolloid activation in the presence of heparin199. Currently, there is a sizeable gap in knowledge surrounding the molecular mechanism of the WFIKKN family and its role in GDF8 and GDF11 regulation.

Specifically, it is unknown which domains are responsible for the exquisite specificity, why WFIKKN1 is able to bind the GDF8 prodomain and WFIKKN2, and what biological impact these large multi-protein complex have in vivo.

One promising approach to determine how the WFIKKN family is able to antagonize GDF8 and

GDF11 involves structurally characterizing additional domains. Currently, only the WFIKKN2 FSD and

WFIKKN1 Kunitz 2103 domain have been solved, but neither have been solved in complex with a ligand.

Attempts at a WFIKKN2 FSD:GDF8 crystal structure have thus far been unsuccessful. This could be due to

104 potentially weak interactions between the WFIKKN2 FSD and GDF8 that prevent crystallization. However,

SPR studies have demonstrated that additional domains contribute to the binding affinity of WFIKKN to the ligand. As such, attempting crystallography with constructs containing multiple WFIKKN domains or the full length WFIKKN protein would be advantageous. Currently, we have constructs within the lab of various WFIKKN domain combinations that can be explored further for biochemical and structural studies.

In addition, the University of Cincinnati is in the process of acquiring Cryo-EM facilities and could serve as a valuable tool for determining the structure of WFIKKN1/2 in complex with GDF8/11.

Future Directions for a Translational Impact: Moving Beyond the Biochemistry

Using an in vitro assay to predict the location of GDF8 expression and subsequent activation in vivo.

The majority of the work presented in this dissertation utilizes protein crystallography and biochemical assays to determine the signaling mechanisms of GDF8. While this work has successfully provided invaluable insight, it is still poorly understood where GDF8 activation occurs in vivo. GDF8 can be found in blood serum, but the ratio of latent:antagonists-bound:fully active GDF8 has not been well characterized. As such, it is unknown whether GDF8 in serum is biochemically active or if it is bound to extracellular antagonists. In addition, tolloid is primarily found within the extracellular matrix (ECM) or within tissues, but it is unknown if tolloid in serum is able to activate serum GDF8 or if serum GDF8 becomes localized to the ECM for activation71,72,144. In short, the contribution of endocrine, paracrine, and autocrine GDF8 signaling is not known. A better understanding of where GDF8 activation and signaling takes place has a drastic impact on the efficacy of therapeutics. For example, if serum GDF8 only accounts for a small portion of GDF8 signaling, then inhibiting serum GDF8 would be less successful than inhibiting

GDF8 within specific tissues or during expression. Thus, we ask this fundamental question: where does

GDF8 become activated and how does that affect signaling?

105

I have developed a novel luciferase assay format to begin addressing this question with an in vitro assay. For activity assays, we commonly use HEK283T cells stably transfected with a luciferase gene under the control of a SMAD2/3 (CAGA)12 promoter that becomes activated upon GDF8 signaling. However, these proteins are expressed randomly when transfected in this assay. Some cells will be transfected with tolloid DNA but not GDF8 whereas others will express both proteins simultaneously. As such, we cannot determine if GDF8 expressed in one cell can be activated by tolloid expressed in a distal cell, just as we cannot determine if ECM bound tolloid can activate serum GDF8 and vice versa. To overcome this, I have developed a method of inserting borosilicate cloning rings (Sigma) within a 24 well plate format. In short, a cloning ring is placed within a 12 well plate separating two cell populations (Figure 2a-b). The ring is held in place by waterproof silicon glue to prevent mixing of media within and outside of the ring. GDF8 or tolloid can then be transfected either inside or outside the ring. Following transfection, the cells can then be washed, and media can be added to the plate. Enough media is added so that the two wells can “mix” by making the top of the media higher than the cloning ring. By extension, tolloid expressing cells and

GDF8 expressing cells are kept separate, but free GDF8 or tolloid can travel into the other compartment.

Signaling can then be measured independently to determine whether GDF8 or tolloid can be released from the cell surface and signal or activate distally from the site of production. To date, I have demonstrated that signaling can successfully occur (Figure 2d). If both elements (GDF8 and Tolloid) are transfected in the “outer” or “inner” ring that section has significantly higher signaling than the other ring

(Figure 2d). Excitingly, initial data suggests that the ring transfected with Tll2 has a higher signal than the ring transfected with GDF8 (Figure 2e). This indicates that GDF8, not Tll2, is able to release from the cell surface and migrate to the other ring supporting in vivo observations of GDF8 within serum and Tll2 localization to the ECM144,200. While this preliminary data is exciting, rings that have only GDF8 and not

Tll2 transfected are still able to signal (Figure 2e). This is potentially due to our luciferase assays relying on over expression of transfected components, thus allowing more tolloid or ligand to escape from the

106 cell surface when they may otherwise remain tethered to the cell surface under physiological conditions.

Moving forward, a more sensitive signaling output may be required. For example, immunohistochemistry can be used to detect GDF8 and tolloid localization at much lower levels and without the need for massive overexpression. Additionally, the plate set up can be conducted with a variety of different cell types to assess cell specific effects.

107

Figure 2: Cloning ring experimental design and preliminary data. a) Image of the cloning ring placed within a well in a 24well plate. b) Demonstrating that silicon based glue is sufficient for a waterproof seal and does not result in media acidification. c) Analysis of the cell density inside (left) and outside (right) of the cloning ring, and demonstrating that no cells enter the cloning ring. d) Transfection based luciferase assay using cloning ring set up with both GDF8 and Tll2 transfected within the same ring.. Location of transfected components are shown in the ring schematic and written under the ring schematic. Signal from the inner ring is shown in black and the outer ring shown in orange. e) Transfection based luciferase assay conducted the same way as (d) but components are separated into different rings denoted beneath each pair. Experiments plotted as mean±SD conducted once with experimental triplicate.

108

Final Conclusions

In conclusion, through extensive biochemical analysis we have identified molecular mechanisms of GDF8 latency, activation, and antagonism that successfully expanded the field’s understanding of the complex regulation of GDF8 signaling. A better understanding of how GDF8 is regulated in vivo is critical to achieve efficacious GDF8 inhibition to develop better therapies to prevent muscle wasting. However, there are still important questions that arise regarding GDF8 signaling and regulation in vivo. While we have explored the inhibitory elements of the GDF8 extensively, further work is required to investigate the other elements of the GDF8 prodomain and their role in localization, dimer formation, and activation. In addition, while we have characterized residues near the tolloid cut site critical for activation, there are other features of tolloid itself and the prodomain that mediate activation. After activation, a number of antagonists are able to inhibit GDF8 but none are as specific as the WFIKKN family. Our characterization of the WFIKKN2 FSD is only one piece of the large picture of GDF8 inhibition in vivo. Together, our findings have provided invaluable insight into GDF8 regulation, whereas the remaining gaps in our knowledge provide exciting directions for future research that may be leveraged into new therapeutic approaches for treating muscle wasting diseases.

109

Bibliography

1. McPherron AC, Lee SJ. Double muscling in cattle due to mutations in the myostatin gene. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 1997 Nov 11 [cited 2017 Oct 26];94(23):12457–61. Available from: http://www.ncbi.nlm.nih.gov/pubmed/9356471 PMID: 9356471 2. Lee S-J. Regulation of muscle mass by myostatin. Annu Rev Cell Dev Biol. 2004;20:61–86. PMID: 15473835 3. Lee SJ, McPherron AC. Regulation of myostatin activity and muscle growth. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 2001 Jul 31 [cited 2017 Jan 17];98(16):9306–11. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11459935 PMID: 11459935 4. Rodriguez J, Vernus B, Chelh I, Cassar-Malek I, Gabillard JC, Hadj Sassi A, Seiliez I, Picard B, Bonnieu A. Myostatin and the skeletal muscle atrophy and hypertrophy signaling pathways. Cellular and molecular life sciences : CMLS. 2014. p. 4361–4371. PMID: 25080109 5. McPherron AC, Lawler AM, Lee SJ. Regulation of skeletal muscle mass in mice by a new TGF-β superfamily member. Nature [Internet]. 1997 May 1 [cited 2017 Jan 9];387(6628):83–90. Available from: http://www.ncbi.nlm.nih.gov/pubmed/9139826 PMID: 9139826 6. Shan T, Liang X, Bi P, Kuang S. Myostatin knockout drives browning of white adipose tissue through activating the AMPK-PGC1α-Fndc5 pathway in muscle. FASEB J [Internet]. The Federation of American Societies for Experimental Biology; 2013 May [cited 2017 Oct 26];27(5):1981–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23362117 PMID: 23362117 7. Deng B, Zhang F, Wen J, Ye S, Wang L, Yang Y, Gong P, Jiang S. The function of myostatin in the regulation of fat mass in mammals [Internet]. Nutrition and Metabolism. BioMed Central Ltd.; 2017 [cited 2020 Jul 12]. Available from: /pmc/articles/PMC5360019/?report=abstract 8. McPherron AC, Lawler AM, Lee SJ. Regulation of anterior/posterior patterning of the axial skeleton by growth/differentiation factor 11. Nat Genet [Internet]. Nature Publishing Group; 1999 Jul 1 [cited 2017 Aug 24];22(3):260–264. Available from: http://www.nature.com/doifinder/10.1038/10320 PMID: 10391213 9. Lu L, Bai X, Cao Y, Luo H, Yang X, Kang L, Shi M-J, Fan W, Zhao B-Q. Growth Differentiation Factor 11 Promotes Neurovascular Recovery After Stroke in Mice. Front Cell Neurosci [Internet]. Frontiers Media SA; 2018 [cited 2018 Aug 23];12:205. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30061815 PMID: 30061815 10. Loffredo FS, Steinhauser ML, Jay SM, Gannon J, Pancoast JR, Yalamanchi P, Sinha M, Dall’Osso C, Khong D, Shadrach JL, Miller CM, Singer BS, Stewart A, Psychogios N, Gerszten RE, Hartigan AJ, Kim MJ, Serwold T, Wagers AJ, Lee RT. Growth differentiation factor 11 is a circulating factor that reverses age-related cardiac hypertrophy. Cell. 2013;153(4):828–839. PMID: 23663781 11. Zhang Y, Li Q, Liu D, Huang Q, Cai G, Cui S, Sun X, Chen X. GDF11 improves tubular regeneration after acute kidney injury in elderly mice. Sci Rep [Internet]. 2016 Oct 5 [cited 2016 Oct 14];6:34624. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27703192 PMID: 27703192 12. McPherron AC, Huynh T V, Lee S-J. Redundancy of myostatin and growth/differentiation factor 11 function. BMC Dev Biol [Internet]. BioMed Central; 2009 Mar 19 [cited 2017 Aug 24];9:24. Available

110

from: http://www.ncbi.nlm.nih.gov/pubmed/19298661 PMID: 19298661 13. Monestier O, Brun C, Heu K, Passet B, Malhouroux M, Magnol L, Vilotte JL, Blanquet V. Ubiquitous Gasp1 overexpression in mice leads mainly to a hypermuscular phenotype. BMC Genomics [Internet]. BioMed Central; 2012 Oct 10 [cited 2016 Dec 31];13(1):541. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23046573 PMID: 23046573 14. Nakatani M, Takehara Y, Sugino H, Matsumoto M, Hashimoto O, Hasegawa Y, Murakami T, Uezumi A, Takeda S, Noji S, Sunada Y, Tsuchida K. Transgenic expression of a myostatin inhibitor derived from follistatin increases skeletal muscle mass and ameliorates dystrophic pathology in mdx mice. FASEB J. 2008; PMID: 17893249 15. Mendell JR, Sahenk Z, Malik V, Gomez AM, Flanigan KM, Lowes LP, Alfano LN, Berry K, Meadows E, Lewis S, Braun L, Shontz K, Rouhana M, Clark KR, Rosales XQ, Al-Zaidy S, Govoni A, Rodino-Klapac LR, Hogan MJ, Kaspar BK. A phase 1/2a follistatin gene therapy trial for becker muscular dystrophy. Mol Ther [Internet]. 2015;23(1):192–201. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4426808&tool=pmcentrez&rendert ype=abstract PMID: 25322757 16. Haidet AM, Rizo L, Handy C, Umapathi P, Eagle A, Shilling C, Boue D, Martin PT, Sahenk Z, Mendell JR, Kaspar BK. Long-term enhancement of skeletal muscle mass and strength by single gene administration of myostatin inhibitors. Proc Natl Acad Sci [Internet]. National Academy of Sciences; 2008 Mar 18 [cited 2018 Oct 3];105(11):4318–4322. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18334646 PMID: 18334646 17. St. Andre M, Johnson M, Bansal PN, Wellen J, Robertson A, Opsahl A, Burch PM, Bialek P, Morris C, Owens J. A mouse anti-myostatin antibody increases muscle mass and improves muscle strength and contractility in the mdx mouse model of Duchenne muscular dystrophy and its humanized equivalent, domagrozumab (PF-06252616), increases muscle volume in cynomolgus monkeys. Skelet Muscle [Internet]. BioMed Central Ltd.; 2017 Nov 9 [cited 2020 Jul 12];7(1). Available from: /pmc/articles/PMC5679155/?report=abstract PMID: 29121992 18. Singh P, Rong H, Gordi T, Bosley J, Bhattacharya I. Translational Pharmacokinetic/Pharmacodynamic Analysis of MYO-029 Antibody for Muscular Dystrophy. Clin Transl Sci [Internet]. 2016 Dec [cited 2018 Mar 2];9(6):302–310. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27700008 PMID: 27700008 19. Pirruccello-Straub M, Jackson J, Wawersik S, Webster MT, Salta L, Long K, McConaughy W, Capili A, Boston C, Carven GJ, Mahanthappa NK, Turner KJ, Donovan A. Blocking extracellular activation of myostatin as a strategy for treating muscle wasting. Sci Rep [Internet]. Nature Publishing Group; 2018 Feb 2 [cited 2018 Mar 2];8(1):2292. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29396542 PMID: 29396542 20. Walker RG, Thompson TB. Fibronectin-based scaffold domain proteins that bind myostatin: a patent evaluation of WO2014043344. Expert Opin Ther Pat. 2015;25(5):619–624. PMID: 25632990 21. Puolakkainen T, Ma H, Kainulainen H, Pasternack A, Rantalainen T, Ritvos O, Heikinheimo K, Hulmi JJ, Kiviranta R. Treatment with soluble activin type IIB-receptor improves bone mass and strength in a mouse model of Duchenne muscular dystrophy. BMC Musculoskelet Disord [Internet]. 2017 Dec 19 [cited 2017 Jan 30];18(1):20. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28103859 PMID: 28103859

111

22. Hatakeyama S, Summermatter S, Jourdain M, Melly S, Minetti GC, Lach-Trifilieff E. ActRII blockade protects mice from cancer cachexia and prolongs survival in the presence of anti-cancer treatments. Skelet Muscle [Internet]. 2016 Dec 26 [cited 2017 Jan 29];6(1):26. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27462398 PMID: 27462398 23. Hinck AP, Mueller TD, Springer TA. Structural biology and evolution of the TGF-β family. Cold Spring Harb Perspect Biol. Cold Spring Harbor Laboratory Press; 2016;8(12). 24. Constam DB. Regulation of TGFβ and related signals by precursor processing. Seminars in Cell and Developmental Biology. 2014. p. 85–97. 25. Jin HJ, Dunn MA, Borthakur D, Kim YS. Refolding and purification of unprocessed porcine myostatin expressed in Escherichia coli. Protein Expr Purif. 2004;35(1):1–10. PMID: 15039059 26. Gray AM, Mason AJ. Requirement for activin a and transforming growth factor-β1 pro-regions in homodimer assembly. Science (80- ) [Internet]. American Association for the Advancement of Science; 1990 Mar 16 [cited 2020 Jul 12];247(4948):1328–1330. Available from: https://science.sciencemag.org/content/247/4948/1328 PMID: 2315700 27. Harrison C a, Al-Musawi SL, Walton KL. Prodomains regulate the synthesis, extracellular localisation and activity of TGF-β superfamily ligands. Growth Factors [Internet]. 2011;29(5):174– 86. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21864080 PMID: 21864080 28. Sengle G, Ono RN, Sasaki T, Sakai LY. Prodomains of transforming growth factor β (TGFβ) Superfamily members specify different functions: Extracellular matrix interactions and growth factor bioavailability. J Biol Chem. American Society for Biochemistry and Molecular Biology; 2011 Feb 18;286(7):5087–5099. 29. Mittl PRE, Priestle JP, Cox DA, Mcmaster G, Cerletti N, Grütter MG. The crystal structure of TGF-β3 and comparison to TGF-β2: Implications for receptor binding. Protein Sci [Internet]. Blackwell Publishing Ltd; 1996 [cited 2020 Jul 12];5(7):1261–1271. Available from: https://pubmed.ncbi.nlm.nih.gov/8819159/ PMID: 8819159 30. Hinck AP, Mueller TD, Springer TA. Structural biology and evolution of the TGF-β family. Cold Spring Harb Perspect Biol [Internet]. 2016 Dec 1 [cited 2017 Feb 19];8(12):a022103. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27638177 PMID: 27638177 31. Hinck AP, Archer SJ, Qian SW, Roberts AB, Sporn MB, Weatherbee JA, Tsang MLS, Lucas R, Zhang BL, Wenker J, Torchia DA. Transforming growth factor β1: Three-dimensional structure in solution and comparison with the X-ray structure of transforming growth factor β2. Biochemistry [Internet]. American Chemical Society ; 1996 [cited 2020 Jul 12];35(26):8517–8534. Available from: https://pubs.acs.org/doi/abs/10.1021/bi9604946 32. Daopin S, Piez KA, Ogawa Y, Davies DR. Crystal structure of transforming growth factor-β2: An unusual fold for the superfamily. Science (80- ) [Internet]. Science; 1992 [cited 2020 Jul 12];257(5068):369–373. Available from: https://pubmed.ncbi.nlm.nih.gov/1631557/ PMID: 1631557 33. Scheufler C, Sebald W, Hülsmeyer M. Crystal structure of human bone morphogenetic protein-2 at 2.7 A resolution. J Mol Biol [Internet]. 1999;287(1):103–15. Available from: http://www.ncbi.nlm.nih.gov/pubmed/10074410 PMID: 10074410 34. Allendorph GP, Isaacs MJ, Kawakami Y, Belmonte JCI, Choe S. BMP-3 and BMP-6 structures

112

illuminate the nature of binding specificity with receptors. Biochemistry [Internet]. Biochemistry; 2007 Oct 30 [cited 2020 Jul 12];46(43):12238–12247. Available from: https://pubmed.ncbi.nlm.nih.gov/17924656/ 35. Griffith DL, Keck PC, Sampath TK, Rueger DC, Carlson WD. Three-dimensional structure of recombinant human osteogenic protein 1: Structural paradigm for the transforming growth factor β superfamily. Proc Natl Acad Sci U S A [Internet]. Proc Natl Acad Sci U S A; 1996 Jan 23 [cited 2020 Jul 12];93(2):878–883. Available from: https://pubmed.ncbi.nlm.nih.gov/8570652/ PMID: 8570652 36. Brown MA, Zhao Q, Baker KA, Naik C, Chen C, Pukac L, Singh M, Tsareva T, Parice Y, Mahoney A, Roschke V, Sanya I, Choe S. Crystal structure of BMP-9 and functional interactions with pro-region and receptors. J Biol Chem [Internet]. J Biol Chem; 2005 Jul 1 [cited 2020 Jul 12];280(26):25111– 25118. Available from: https://pubmed.ncbi.nlm.nih.gov/15851468/ 37. Harrington AE, Morris-Triggs SA, Ruotolo BT, Robinson C V., Ohnuma SI, Hyvönen M. Structural basis for the inhibition of activin signalling by follistatin. EMBO J [Internet]. EMBO Press; 2006 Mar 8 [cited 2018 Oct 15];25(5):1035–1045. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16482217 PMID: 16482217 38. Nickel J, Kotzsch A, Sebald W, Mueller TD. A single residue of GDF-5 defines binding specificity to BMP receptor IB. J Mol Biol [Internet]. Academic Press; 2005 Jun 24 [cited 2020 Jul 12];349(5):933– 947. Available from: https://pubmed.ncbi.nlm.nih.gov/15890363/ 39. Walker RG, Czepnik M, Goebel EJ, McCoy JC, Vujic A, Cho M, Oh J, Aykul S, Walton KL, Schang G, Bernard DJ, Hinck AP, Harrison CA, Martinez-Hackert E, Wagers AJ, Lee RT, Thompson TB. Structural basis for potency differences between GDF8 and GDF11. BMC Biol [Internet]. BMC Biology; 2017 Dec 3 [cited 2017 May 31];15(1):19. Available from: http://bmcbiol.biomedcentral.com/articles/10.1186/s12915-017-0350-1 40. Goebel EJ, Corpina RA, Hinck CS, Czepnik M, Castonguay R, Grenha R, Boisvert A, Miklossy G, Fullerton PT, Matzuk MM, Idone VJ, Economides AN, Kumar R, Hinck AP, Thompson TB. Structural characterization of an activin class ternary receptor complex reveals a third paradigm for receptor specificity. Proc Natl Acad Sci U S A. National Academy of Sciences; 2019 Jul 30;116(31):15505– 15513. 41. Goebel EJ, Hart KN, Mccoy JC, Thompson TB. Structural biology of the TGFb family. Exp Biol Med. 2019;244:1530–1546. 42. Hinck AP. Structural studies of the TGF-βs and their receptors - Insights into evolution of the TGF- β superfamily [Internet]. FEBS Letters. Elsevier B.V.; 2012 [cited 2020 Jul 12]. p. 1860–1870. Available from: https://pubmed.ncbi.nlm.nih.gov/22651914/ 43. Weiss A, Attisano L. The TGFbeta superfamily signaling pathway. Wiley Interdisciplinary Reviews: Developmental Biology. 2013. p. 47–63. PMID: 23799630 44. Wrana JL, Attisano L, Wieser R, Ventura F, Massagué J. Mechanism of activation of the TGF-β receptor. Nature. Nature Publishing Group; 1994;370(6488):341–347. 45. Massagué J, Seoane J, Wotton D. Smad transcription factors [Internet]. Genes and Development. Genes Dev; 2005 [cited 2020 Jul 12]. p. 2783–2810. Available from: https://pubmed.ncbi.nlm.nih.gov/16322555/ PMID: 16322555

113

46. Andersson O, Reissmann E, Ibáñez CF. Growth differentiation factor 11 signals through the transforming growth factor-β receptor ALK5 to regionalize the anterior-posterior axis. EMBO Rep [Internet]. EMBO Rep; 2006 Aug [cited 2020 Jul 12];7(8):831–837. Available from: https://pubmed.ncbi.nlm.nih.gov/16845371/ 47. Myostatin signals through a transforming growth factor β-like signaling pathway to block adipogenesis. Mol Cell Biol [Internet]. 2003;23(20):7230–7242. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=230332&tool=pmcentrez&renderty pe=abstract PMID: 14517293 48. Duran J, Troncoso MF, Lagos D, Ramos S, Marin G, Estrada M. GDF11 modulates ca2+-dependent Smad2/3 signaling to prevent cardiomyocyte hypertrophy. Int J Mol Sci [Internet]. MDPI AG; 2018 May 18 [cited 2020 Jul 12];19(5). Available from: /pmc/articles/PMC5983757/?report=abstract 49. Philip B, Lu Z, Gao Y. Regulation of GDF-8 signaling by the p38 MAPK. Cell Signal [Internet]. Cell Signal; 2005 Mar [cited 2020 Jul 12];17(3):365–375. Available from: https://pubmed.ncbi.nlm.nih.gov/15567067/ 50. Yang W, Chen Y, Zhang Y, Wang X, Yang N, Zhu D. Extracellular Signal–Regulated Kinase 1/2 Mitogen-Activated Protein Kinase Pathway Is Involved in Myostatin-Regulated Differentiation Repression. Cancer Res [Internet]. American Association for Cancer Research (AACR); 2006 Feb 1 [cited 2020 Jul 12];66(3):1320–1326. Available from: https://pubmed.ncbi.nlm.nih.gov/16452185/ 51. Brun C, Périé L, Baraige F, Vernus B, Bonnieu A, Blanquet V. Absence of hyperplasia in Gasp-1 overexpressing mice is dependent on myostatin up-regulation. Cell Physiol Biochem [Internet]. Karger Publishers; 2014 [cited 2017 Feb 3];34(4):1241–59. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25277978 PMID: 25277978 52. Watanabe H, de Caestecker MP, Yamada Y. Transcriptional cross-talk between Smad, ERK1/2, and p38 mitogen-activated protein kinase pathways regulates transforming growth factor-beta- induced aggrecan gene expression in chondrogenic ATDC5 cells. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology Inc.; 2001 Apr 27 [cited 2020 Jul 12];276(17):14466–73. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11278290 PMID: 11278290 53. Biesemann N, Mendler L, Kostin S, Wietelmann A, Borchardt T, Braun T. Myostatin induces interstitial fibrosis in the heart via TAK1 and p38. Cell Tissue Res [Internet]. Springer Verlag; 2015 Sep 28 [cited 2020 Jul 12];361(3):779–787. Available from: https://pubmed.ncbi.nlm.nih.gov/25725788/ 54. Gray AM, Mason AJ. Requirement for activin a and transforming growth factor-β1 pro-regions in homodimer assembly. Science (80- ). 1990;247(4948):1328–1330. 55. Shi M, Zhu J, Wang R, Chen X, Mi L, Walz T, Springer T a. Latent TGF-β structure and activation. Nature [Internet]. 2011;474(7351):343–349. Available from: http://dx.doi.org/10.1038/nature10152 PMID: 21677751 56. Wang X, Fischer G, Hyvönen M. Structure and activation of pro-activin A. Nat Commun [Internet]. 2016 [cited 2019 Jul 12];7. Available from: www.nature.com/naturecommunications 57. Cotton TR, Fischer G, Wang X, McCoy JC, Czepnik M, Thompson TB, Hyvönen M. Structure of the human myostatin precursor and determinants of growth factor latency. EMBO J [Internet]. EMBO

114

Press; 2018 Jan 12 [cited 2018 Jan 31];e97883. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29330193 PMID: 29330193 58. Walker RG, McCoy JC, Czepnik M, Mills MJ, Hagg A, Walton KL, Cotton TR, Hyvönen M, Lee RT, Gregorevic P, Harrison CA, Thompson TB. Molecular characterization of latent GDF8 reveals mechanisms of activation. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 2018 Jan 30 [cited 2018 Apr 2];115(5):E866–E875. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29348202 PMID: 29348202 59. Salmon RM, Guo J, Wood JH, Tong Z, Beech JS, Lawera A, Yu M, Grainger DJ, Reckless J, Morrell NW, Li W. Molecular basis of ALK1-mediated signalling by BMP9/BMP10 and their prodomain- bound forms. Nat Commun. Nature Research; 2020 Dec 1;11(1). PMID: 32238803 60. Mi L, Brown CT, Gao Y, Tian Y, Le VQ, Walz T, Springer TA. Structure of bone morphogenetic protein 9 procomplex. Proc Natl Acad Sci U S A. 2015;112(12):3710–3715. PMID: 25751889 61. Jiang H, Salmon RM, Upton PD, Wei Z, Lawera A, Davenport AP, Morrell NW, Li W. The prodomain- bound form of bone morphogenetic protein 10 is biologically active on endothelial cells. J Biol Chem. 2016;291(6):2954–2966. 62. Ge G, Greenspan DS. BMP1 controls TGFbeta1 activation via cleavage of latent TGFbeta-binding protein. J Cell Biol [Internet]. The Rockefeller University Press; 2006 Oct 9 [cited 2018 Jan 3];175(1):111–20. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17015622 PMID: 17015622 63. Ge G, Hopkins DR, Ho W-B, Greenspan DS. GDF11 forms a bone morphogenetic protein 1-activated latent complex that can modulate nerve growth factor-induced differentiation of PC12 cells. Mol Cell Biol [Internet]. 2005;25(14):5846–58. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1168807&tool=pmcentrez&rendert ype=abstract PMID: 15988002 64. Miyazono K, Olofsson A, Colosetti P, Heldin C-H. A role of the latent TGF-3 1-binding protein in the assembly and secretion of TGF-B1 [Internet]. The EMBOJournal. 1991. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC452762/pdf/emboj00103-0066.pdf 65. Annes JP, Chen Y, Munger JS, Rifkin DB. Integrin αvβ6-mediated activation of latent TGF-β requires the latent TGF-β binding protein-1. J Cell Biol [Internet]. The Rockefeller University Press; 2004 Jun 7 [cited 2019 Jul 16];165(5):723–734. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15184403 PMID: 15184403 66. Sarrazy V, Koehler A, Chow ML, Zimina E, Li CX, Kato H, Caldarone CA, Hinz B. Integrins αvβ5 and αvβ3 promote latent TGF-β1 activation by human cardiac fibroblast contraction. Cardiovasc Res [Internet]. Oxford University Press; 2014 Jun 1 [cited 2018 Apr 23];102(3):407–17. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24639195 PMID: 24639195 67. Liénart S, Merceron R, Vanderaa C, Lambert F, Colau D, Stockis J, van der Woning B, De Haard H, Saunders M, Coulie PG, Savvides SN, Lucas S. Structural basis of latent TGF-β1 presentation and activation by GARP on human regulatory T cells. Science (80- ) [Internet]. American Association for the Advancement of Science; 2018 Nov 23 [cited 2019 Apr 22];362(6417):952–956. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30361387 PMID: 30361387 68. Stockis J, Dedobbeleer O, Lucas S. Role of GARP in the activation of latent TGF-β1. Molecular

115

BioSystems. 2017. p. 1925–1935. 69. Dong X, Zhao B, Iacob RE, Zhu J, Koksal AC, Lu C, Engen JR, Springer TA. Force interacts with macromolecular structure in activation of TGF-β. Nature [Internet]. NIH Public Access; 2017 [cited 2018 Apr 23];542(7639):55–59. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28117447 PMID: 28117447 70. Tran DQ, Andersson J, Wang R, Ramsey H, Unutmaz D, Shevach EM. GARP (LRRC32) is essential for the surface expression of latent TGF- on platelets and activated FOXP3+ regulatory T cells [Internet]. Proceedings of the National Academy of Sciences. 2009. Available from: www.pnas.orgcgidoi10.1073pnas.0901944106 71. Hopkins DR, Keles S, Greenspan DS. The bone morphogenetic protein 1/Tolloid-like metalloproteinases. Matrix Biol [Internet]. NIH Public Access; 2007 Sep [cited 2018 Jan 4];26(7):508–23. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17560775 PMID: 17560775 72. Ge G, Greenspan DS. Developmental roles of the BMP1/TLD metalloproteinases. Birth Defects Res Part C - Embryo Today Rev [Internet]. John Wiley & Sons, Ltd; 2006 Mar 1 [cited 2019 May 2];78(1):47–68. Available from: http://doi.wiley.com/10.1002/bdrc.20060 73. Le VQ, Iacob RE, Tian Y, McConaughy W, Jackson J, Su Y, Zhao B, Engen JR, Pirruccello‐Straub M, Springer TA. Tolloid cleavage activates latent GDF8 by priming the pro‐complex for dissociation. EMBO J [Internet]. 2018;37(3):384–397. Available from: https://doi. 74. Stockis J, Liénart S, Colau D, Collignon A, Nishimura SL, Sheppard D, Coulie PG, Lucas S. Blocking immunosuppression by human Tregs in vivo with antibodies targeting integrin αVβ8. Proc Natl Acad Sci [Internet]. 2017;114(47):E10161–E10168. Available from: www.pnas.org/cgi/doi/10.1073/pnas.1710680114 75. Jiang MS, Liang LF, Wang S, Ratovitski T, Holmstrom J, Barker C, Stotish R. Characterization and identification of the inhibitory domain of GDF-8 propeptide. Biochem Biophys Res Commun. 2004;315(3):525–531. PMID: 14975732 76. Cuende J, Liénart S, Dedobbeleer O, Van Der Woning B, De Boeck G, Stockis J, Huygens C, Colau D, Somja J, Delvenne P, Hannon M, Baron F, Dumoutier L, Renauld JC, De Haard H, Saunders M, Coulie PG, Lucas S. Monoclonal antibodies against GARP/TGF-β1 complexes inhibit the immunosuppressive activity of human regulatory T cells in vivo. Sci Transl Med. 2015;7(284):284ec65. 77. Manlove LS, Berquam-Vrieze KE, Pauken KE, Williams RT, Jenkins MK, Farrar MA. Adaptive Immunity to Leukemia Is Inhibited by Cross-Reactive Induced Regulatory T Cells. J Immunol. 2015;195(8):4028–4037. 78. Wang R, Zhu J, Dong X, Shi M, Lu C, Springer TA. GARP regulates the bioavailability and activation of TGFβ. Mol Biol Cell. 2012;23(6):1129–1139. 79. Stockis J, Colau D, Coulie PG, Lucas S. Membrane protein GARP is a receptor for latent TGF-β on the surface of activated human Treg. Eur J Immunol [Internet]. 2009;39(12):3315–3322. Available from: www.eji-journal.eu 80. Troilo H, Bayley CP, Barrett AL, Lockhart-Cairns MP, Jowitt TA, Baldock C. Mammalian tolloid proteinases: role in growth factor signalling. FEBS Lett [Internet]. Wiley-Blackwell; 2016 [cited 2018

116

Jan 3];590(15):2398–407. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27391803 PMID: 27391803 81. Lee SJ. Genetic analysis of the role of proteolysis in the activation of latent myostatin. PLoS One [Internet]. Public Library of Science; 2008 Feb 20 [cited 2019 Feb 28];3(2):e1628. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18286185 PMID: 18286185 82. Takahara K, Lyons GE, Greenspan DS. Bone morphogenetic protein-1 and a mammalian tolloid homologue (mTld) are encoded by alternatively spliced transcripts which are differentially expressed in some tissues. J Biol Chem [Internet]. 1994 Dec 23 [cited 2018 Mar 2];269(51):32572– 8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/7798260 PMID: 7798260 83. Hopkins DR, Keles S, Greenspan DS. The bone morphogenetic protein 1/Tolloid-like metalloproteinases. Matrix Biol [Internet]. 2007 [cited 2019 Feb 19];26(7):508–523. Available from: www.sciencedirect.com 84. Muir AM, Massoudi D, Nguyen N, Keene DR, Lee SJ, Birk DE, Davidson JM, Marinkovich MP, Greenspan DS. BMP1-like proteinases are essential to the structure and wound healing of skin. Matrix Biol [Internet]. NIH Public Access; 2016 [cited 2019 May 2];56:114–131. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27363389 PMID: 27363389 85. Muir AM, Ren Y, Butz DH, Davis NA, Blank RD, Birk DE, Lee SJ, Rowe D, Feng JQ, Greenspan DS. Induced ablation of Bmp1 and Tll1 produces osteogenesis imperfecta in mice. Hum Mol Genet [Internet]. Oxford University Press; 2014 Jun 15 [cited 2019 Feb 28];23(12):3085–3101. Available from: https://academic.oup.com/hmg/article-lookup/doi/10.1093/hmg/ddu013 86. Vadon-Le Goff S, Hulmes DJS, Moali C. BMP-1/tolloid-like proteinases synchronize matrix assembly with growth factor activation to promote morphogenesis and tissue remodeling. Matrix Biol [Internet]. Elsevier; 2015 May 1 [cited 2019 Feb 26];44–46:14–23. Available from: https://www.sciencedirect.com/science/article/pii/S0945053X15000438 87. Lee HX, Mendes FA, Plouhinec JL, De Robertis EM. Enzymatic regulation of pattern: BMP4 binds CUB domains of Tolloids and inhibits proteinase activity. Genes Dev [Internet]. 2009;23(21):2551– 2562. Available from: http://www.genesdev.org. 88. Plouhinec JL, Zakin L, De Robertis EM. Systems control of BMP morphogen flow in vertebrate embryos [Internet]. Current Opinion in Genetics and Development. NIH Public Access; 2011 [cited 2020 Jul 12]. p. 696–703. Available from: /pmc/articles/PMC3224208/?report=abstract 89. Plouhinec JL, Zakin L, Moriyama Y, De Robertis EM. Chordin forms a self-organizing morphogen gradient in the extracellular space between ectoderm and mesoderm in the Xenopus embryo. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 2013 Dec 17 [cited 2020 Jul 12];110(51):20372–20379. Available from: www.pnas.org/cgi/doi/10.1073/pnas.1319745110 90. Ploper D, Lee HX, De robertis EM. Dorsal-Ventral patterning: Crescent is a dorsally secreted Frizzled-related protein that competitively inhibits Tolloid proteases. Dev Biol [Internet]. Academic Press Inc.; 2011 [cited 2020 Jul 12];352(2):317–328. Available from: /pmc/articles/PMC3088358/?report=abstract 91. Walker RG, Angerman EB, Kattamuri C, Lee Y-S, Lee S-J, Thompson TB. Alternative Binding Modes Identified for Growth and Differentiation Factor-associated Serum Protein (GASP) Family Antagonism of Myostatin. J Biol Chem [Internet]. 2015 Mar 20 [cited 2016 Jun 7];290(12):7506–

117

7516. Available from: http://www.jbc.org/lookup/doi/10.1074/jbc.M114.624130 92. Troilo H, Zuk A V., Tunnicliffe RB, Wohl AP, Berry R, Collins RF, Jowitt TA, Sengle G, Baldock C. Nanoscale structure of the BMP antagonist chordin supports cooperative BMP binding. Proc Natl Acad Sci [Internet]. 2014;111(36):13063–13068. Available from: www.pnas.org/cgi/doi/10.1073/pnas.1404166111 93. Zhang JL, Huang Y, Qiu LY, Nickel J, Sebald W. Von Willebrand factor type C domain-containing proteins regulate bone morphogenetic protein signaling through different recognition mechanisms. J Biol Chem [Internet]. 2007;282(27):20002–20014. Available from: http://www.jbc.org/ 94. Thompson TB, Lerch TF, Cook RW, Woodruff TK, Jardetzky TS. The structure of the follistatin: Activin complex reveals antagonism of both type I and type II receptor binding. Dev Cell [Internet]. Cell Press; 2005 Oct 1 [cited 2017 Nov 28];9(4):535–543. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16198295 PMID: 16198295 95. Cash JN, Angerman EB, Kattamuri C, Nolan K, Zhao H, Sidis Y, Keutmann HT, Thompson TB. Structure of myostatin·follistatin-like 3: N-terminal domains of follistatin-type molecules exhibit alternate modes of binding. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2012 Jan 6 [cited 2017 Jan 9];287(2):1043–53. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22052913 PMID: 22052913 96. Cash JN, Rejon CA, McPherron AC, Bernard DJ, Thompson TB. The structure of myostatin:follistatin 288: Insights into receptor utilization and heparin binding. EMBO J [Internet]. European Molecular Biology Organization; 2009 Sep 2 [cited 2017 Jan 9];28(17):2662–2676. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19644449 PMID: 19644449 97. Cash JN, Angerman EB, Keutmann HT, Thompson TB. Characterization of follistatin-type domains and their contribution to myostatin and activin A antagonism. Mol Endocrinol [Internet]. 2012;26(7):1167–78. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3385792&tool=pmcentrez&rendert ype=abstract PMID: 22593183 98. Innis CA, Hyvönen M. Crystal structures of the heparan sulfate-binding domain of follistatin. Insights into ligand binding. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2003 Oct 10 [cited 2018 Oct 12];278(41):39969–39977. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12867435 PMID: 12867435 99. Stamler R, Keutmann HT, Sidis Y, Kattamuri C, Schneyer A, Thompson TB. The structure of FSTL3??activin A complex: Differential binding of N-terminal domains influences follistatin-type antagonist specificity. J Biol Chem. 2008;283(47):32831–32838. PMID: 18768470 100. Monestier O, Brun C, Cocquempot O, Petit D, Blanquet V. GASP/WFIKKN proteins: Evolutionary aspects of their functions. Schubert M, editor. PLoS One [Internet]. Public Library of Science; 2012 Aug 24 [cited 2017 Jan 8];7(8):e43710. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22937083 PMID: 22937083 101. Kondás K, Szláma G, Trexler M, Patthy L. Both WFIKKN1 and WFIKKN2 have high affinity for growth and differentiation factors 8 and 11. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2008 Jun 4 [cited 2016 Jun 8];283(35):23677–23684. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18596030 PMID: 18596030

118

102. Hill JJ, Qiu Y, Hewick RM, Wolfman NM. Regulation of Myostatin in Vivo by Growth and Differentiation Factor-Associated Serum Protein-1: A Novel Protein with Protease Inhibitor and Follistatin Domains. Mol Endocrinol [Internet]. Oxford University Press; 2003 Jun 1 [cited 2018 Apr 3];17(6):1144–1154. Available from: https://academic.oup.com/mend/article- lookup/doi/10.1210/me.2002-0366 PMID: 12595574 103. Liepinsh E, Nagy A, Trexler M, Otting G. Second Kunitz-type protease inhibitor domain of the human WFIKKN1 protein. [cited 2018 Jan 3]; Available from: https://link.springer.com/content/pdf/10.1007%2Fs10858-006-9013-1.pdf 104. McCoy JC, Walker RG, Murray NH, Thompson TB. Crystal structure of the WFIKKN2 follistatin domain reveals insight into how it inhibits growth differentiation factor 8 (GDF8) and GDF11. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2019 Feb 27 [cited 2019 Mar 28];294(16):6333–6343. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30814254 PMID: 30814254 105. Lee S-J. Extracellular Regulation of Myostatin: A Molecular Rheostat for Muscle Mass. Immunol Endocr Metab Agents Med Chem [Internet]. NIH Public Access; 2010 [cited 2018 May 23];10:183– 194. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21423813 PMID: 21423813 106. Suragani RNVS, Cadena SM, Cawley SM, Sako D, Mitchell D, Li R, Davies M V, Alexander MJ, Devine M, Loveday KS, Underwood KW, Grinberg A V, Quisel JD, Chopra R, Pearsall RS, Seehra J, Kumar R. Transforming growth factor-β superfamily ligand trap ACE-536 corrects anemia by promoting late- stage erythropoiesis. Nat Med [Internet]. 2014 Apr 23 [cited 2018 Mar 2];20(4):408–14. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24658078 PMID: 24658078 107. Lach-Trifilieff E, Minetti GC, Sheppard K, Ibebunjo C, Feige JN, Hartmann S, Brachat S, Rivet H, Koelbing C, Morvan F, Hatakeyama S, Glass DJ. An Antibody Blocking Activin Type II Receptors Induces Strong Skeletal Muscle Hypertrophy and Protects from Atrophy. Mol Cell Biol [Internet]. American Society for Microbiology; 2014 Feb 15 [cited 2020 Jul 16];34(4):606–618. Available from: /pmc/articles/PMC3911487/?report=abstract 108. Thies RS, Chen T, Davies M V, Tomkinson KN, Pearson AA, Shakey QA, Wolfman NM. GDF-8 propeptide binds to GDF-8 and antagonizes biological activity by inhibiting GDF-8 receptor binding. Growth Factors. 2001;18(4):251–259. PMID: 11519824 109. Zimmers T a, Davies M V, Koniaris LG, Haynes P, Esquela AF, Tomkinson KN, McPherron AC, Wolfman NM, Lee S-J. Induction of cachexia in mice by systemically administered myostatin. Science. 2002;296(5572):1486–1488. PMID: 12029139 110. Hill JJ, Davies M V., Pearson AA, Wang JH, Hewick RM, Wolfman NM, Qiu Y. The myostatin propeptide and the follistatin-related gene are inhibitory binding proteins of myostatin in normal serum. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2002 Oct 25 [cited 2017 Jan 19];277(43):40735–40741. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12194980 PMID: 12194980 111. Wolfman NM, McPherron AC, Pappano WN, Davies M V, Song K, Tomkinson KN, Wright JF, Zhao L, Sebald SM, Greenspan DS, Lee S-J. Activation of latent myostatin by the BMP-1/tolloid family of metalloproteinases. Proc Natl Acad Sci U S A [Internet]. 2003;100(26):15842–6. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=307655&tool=pmcentrez&renderty pe=abstract PMID: 14671324

119

112. Mi L-Z, Brown CT, Gao Y, Tian Y, Le VQ, Walz T, Springer TA. Structure of bone morphogenetic protein 9 procomplex. Proc Natl Acad Sci [Internet]. 2015 [cited 2019 Jul 12];201501303. Available from: www.pnas.org/cgi/doi/10.1073/pnas.1501303112 113. Wipff PJ, Hinz B. Integrins and the activation of latent transforming growth factor β1 - An intimate relationship [Internet]. European Journal of Cell Biology. Eur J Cell Biol; 2008 [cited 2020 Jul 16]. p. 601–615. Available from: https://pubmed.ncbi.nlm.nih.gov/18342983/ 114. Lawrence DA, Pircher R, Jullien P. Conversion of a high molecular weight latent β-TGF from chicken embryo fibroblasts into a low molecular weight active β-TGF under acidic conditions. Biochem Biophys Res Commun [Internet]. Biochem Biophys Res Commun; 1985 Dec 31 [cited 2020 Jul 16];133(3):1026–1034. Available from: https://pubmed.ncbi.nlm.nih.gov/3866579/ 115. Schultz-Cherry S, Murphy-Ullrich JE. Thrombospondin causes activation of latent transforming growth factor-β secreted by endothelial cells by a novel mechanism. J Cell Biol [Internet]. Rockefeller University Press; 1993 [cited 2020 Jul 16];122(4):923–932. Available from: https://pubmed.ncbi.nlm.nih.gov/8349738/ PMID: 8349738 116. Cash JN, Angerman EB, Kirby RJ, Merck L, Seibel WL, Wortman MD, Papoian R, Nelson S, Thompson TB. Development of a Small-Molecule Screening Method for Inhibitors of Cellular Response to Myostatin and Activin A. J Biomol Screen [Internet]. 2013 Aug 1 [cited 2016 Jun 7];18(7):837–844. Available from: http://jbx.sagepub.com/cgi/doi/10.1177/1087057113482585 117. Schneidman-Duhovny D, Hammel M, Sali A. FoXS: a web server for rapid computation and fitting of SAXS profiles. [cited 2020 Jul 16]; Available from: http://salilab.org/foxs 118. Walton KL, Makanji Y, Wilce MC, Chan KL, Robertson DM, Harrison CA. A common biosynthetic pathway governs the dimerization and secretion of inhibin and related transforming growth factor β(TGFβ) ligands. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2009 Apr 3 [cited 2020 Jul 16];284(14):9311–9320. Available from: /pmc/articles/PMC2666583/?report=abstract PMID: 19193648 119. Lee S, Solow-Cordero DE, Kessler E, Takahara K, Greenspan DS. Transforming growth factor-β regulation of bone morphogenetic protein- 1/procollagen C-proteinase and related proteins in fibrogenic cells and keratinocytes. J Biol Chem [Internet]. J Biol Chem; 1997 [cited 2020 Jul 13];272(30):19059–19066. Available from: https://pubmed.ncbi.nlm.nih.gov/9228090/ 120. Leighton M, Kadler KE. Paired basic/furin-like proprotein convertase cleavage of pro-BMP-1 in the trans-golgi network. J Biol Chem. American Society for Biochemistry and Molecular Biology; 2003 May 16;278(20):18478–18484. 121. Jasuja R, Ge G, Voss NG, Lyman-Gingerich J, Branam AM, Pelegri FJ, Greenspan DS. Bone morphogenetic protein 1 prodomain specifically binds and regulates signaling by bone morphogenetic proteins 2 and 4. J Biol Chem [Internet]. 2007;282(12):9053–9062. Available from: http://www.jbc.org/ 122. Szláma G, Trexler M, Buday L, Patthy L. K153R polymorphism in myostatin gene increases the rate of promyostatin activation by furin. FEBS Lett [Internet]. Elsevier B.V.; 2015 Jan 30 [cited 2020 Jul 16];589(3):295–301. Available from: https://pubmed.ncbi.nlm.nih.gov/25543063/ PMID: 25543063 123. Chen JL, Walton KL, Winbanks CE, Murphy KT, Thomson RE, Makanji Y, Qian H, Lynch GS, Harrison

120

CA, Gregorevic P. Elevated expression of activins promotes muscle wasting and cachexia. FASEB J. 2014;28(4). PMID: 24378873 124. Szláma G, Trexler M, Patthy L. Latent myostatin has significant activity and this activity is controlled more efficiently by WFIKKN1 than by WFIKKN2. FEBS J [Internet]. 2013 Aug [cited 2017 Jan 9];280(16):3822–39. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23829672 PMID: 23829672 125. Young GD, Murphy-Ullrich JE. Molecular interactions that confer latency to transforming growth factor-β. J Biol Chem [Internet]. J Biol Chem; 2004 Sep 3 [cited 2020 Jul 16];279(36):38032–38039. Available from: https://pubmed.ncbi.nlm.nih.gov/15208302/ 126. Ohsawa Y, Takayama K, Nishimatsu SI, Okada T, Fujino M, Fukai Y, Murakami T, Hagiwara H, Itoh F, Tsuchida K, Hayashi Y, Sunada Y. The inhibitory core of the myostatin prodomain: Its interaction with both type I and II membrane receptors, and potential to treat muscle atrophy. PLoS One [Internet]. Public Library of Science; 2015 [cited 2017 Jul 28];10(7):e0133713. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26226340 PMID: 26226340 127. Takayama K, Noguchi Y, Aoki S, Takayama S, Yoshida M, Asari T, Yakushiji F, Nishimatsu SI, Ohsawa Y, Itoh F, Negishi Y, Sunada Y, Hayashi Y. Identification of the minimum peptide from mouse myostatin prodomain for human myostatin inhibition. J Med Chem. 2015;58(3):1544–1549. PMID: 25569186 128. Jones DC, Hallyburton I, Stojanovski L, Read KD, Frearson JA, Fairlamb AH. Identification of a κ- opioid agonist as a potent and selective lead for drug development against human African trypanosomiasis. Biochem Pharmacol [Internet]. Elsevier; 2010 Nov 15 [cited 2020 Jul 16];80(10):1478–1486. Available from: /pmc/articles/PMC3025325/?report=abstract 129. Cash JN, Rejon CA, McPherron AC, Bernard DJ, Thompson TB. The structure of myostatin:follistatin 288: Insights into receptor utilization and heparin binding. EMBO J. 2009 Sep 2;28(17):2662–2676. 130. Lee Y-S, Lee S-J. Regulation of GDF-11 and myostatin activity by GASP-1 and GASP-2. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 2013 Sep 24 [cited 2017 Feb 15];110(39):E3713- 22. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3785741&tool=pmcentrez&rendert ype=abstract PMID: 24019467 131. Svergun DI. Restoring low resolution structure of biological macromolecules from solution scattering using simulated annealing. Biophys J [Internet]. Biophysical Society; 1999 [cited 2020 Jul 16];76(6):2879–2886. Available from: /pmc/articles/PMC1300260/?report=abstract 132. Volkov V V., Svergun DI. Uniqueness of ab initio shape determination in small-angle scattering. J Appl Crystallogr. 2003. p. 860–864. 133. Kozin MB, Svergun DI. Automated matching of high- and low-resolution structural models. J Appl Crystallogr [Internet]. International Union of Crystallography; 2001 Feb 1 [cited 2020 Jul 16];34(1):33–41. Available from: http://scripts.iucr.org/cgi-bin/paper?pe0064 134. Thomas M, Langley B, Berry C, Sharma M, Kirk S, Bass J, Kambadur R. Myostatin, a negative regulator of muscle growth, functions by inhibiting myoblast proliferation. J Biol Chem [Internet]. 2000 Dec 15 [cited 2017 Feb 3];275(51):40235–40243. Available from: http://www.ncbi.nlm.nih.gov/pubmed/10976104 PMID: 10976104

121

135. Parenté A, Boukredine A, Baraige F, Duprat N, Gondran‐Tellier V, Magnol L, Blanquet V. GASP‐2 overexpressing mice exhibit a hypermuscular phenotype with contrasting molecular effects compared to GASP‐1 transgenics. FASEB J [Internet]. 2020 Jan 20 [cited 2020 Jan 26];fj.201901220R. Available from: https://onlinelibrary.wiley.com/doi/abs/10.1096/fj.201901220R 136. Lee S-J, Lee Y-S, Zimmers T a, Soleimani A, Matzuk MM, Tsuchida K, Cohn RD, Barton ER. Regulation of muscle mass by follistatin and activins. Mol Endocrinol. 2010; PMID: 20810712 137. Funkenstein B, Rebhan Y. Expression, purification, renaturation and activation of fish myostatin expressed in Escherichia coli: Facilitation of refolding and activity inhibition by myostatin prodomain. Protein Expr Purif. 2007;54(1):54–65. PMID: 17383894 138. Sakuma K, Aoi W, Yamaguchi A. Molecular mechanism of sarcopenia and cachexia: recent research advances. Pflügers Arch - Eur J Physiol [Internet]. 2017 Jan 19 [cited 2017 Jan 29]; Available from: http://www.ncbi.nlm.nih.gov/pubmed/28101649 PMID: 28101649 139. Costelli P, Muscaritoli M, Bonetto A, Penna F, Reffo P, Bossola M, Bonelli G, Doglietto GB, Baccino FM, Fanelli FR. Muscle myostatin signalling is enhanced in experimental cancer cachexia. Eur J Clin Invest [Internet]. 2008 Jul [cited 2017 Jan 29];38(7):531–538. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18578694 PMID: 18578694 140. Tsuchida K. Targeting myostatin for therapies against muscle-wasting disorders. Curr Opin Drug Discov Devel [Internet]. 2008;11(4):487–94. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18600566 PMID: 18600566 141. Smith RC, Lin BK. Myostatin inhibitors as therapies for muscle wasting associated with cancer and other disorders. Curr Opin Support Palliat Care [Internet]. 2013;7(4):352–60. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3819341&tool=pmcentrez&rendert ype=abstract PMID: 24157714 142. Chen JL, Walton KL, Al-Musawi SL, Kelly EK, Qian H, La M, Lu L, Lovrecz G, Ziemann M, Lazarus R, El-Osta A, Gregorevic P, Harrison CA. Development of novel activin-targeted therapeutics. Mol Ther. 2015; PMID: 25399825 143. Pirruccello-Straub M, Jackson J, Wawersik S, Webster MT, Salta L, Long K, McConaughy W, Capili A, Boston C, Carven GJ, Mahanthappa NK, Turner KJ, Donovan A. Blocking extracellular activation of myostatin as a strategy for treating muscle wasting. Sci Rep [Internet]. 2018 [cited 2019 Jul 12];8(1). Available from: www.nature.com/scientificreports/ 144. Scott IC, Blitz IL, Pappano WN, Imamura Y, Clark TG, Steiglitz BM, Thomas CL, Maas S a, Takahara K, Cho KW, Greenspan DS. Mammalian BMP-1/Tolloid-related metalloproteinases, including novel family member mammalian Tolloid-like 2, have differential enzymatic activities and distributions of expression relevant to patterning and skeletogenesis. Dev Biol. 1999;213:283–300. PMID: 10479448 145. Wermter C, Hö Wel M, Hintze V, Bombosch B, Aufenvenne K, Yiallouros I, Stö W. The protease domain of procollagen C-proteinase (BMP1) lacks substrate selectivity, which is conferred by non- proteolytic domains. Biol Chem [Internet]. Walter de Gruyter; 2007 [cited 2018 Jan 3];388:513– 521. Available from: https://www.degruyter.com/downloadpdf/j/bchm.2007.388.issue- 5/bc.2007.054/bc.2007.054.pdf

122

146. Pappano WN, Steiglitz BM, Scott IC, Keene DR, Greenspan DS. Use of Bmp1/Tll1 Doubly Homozygous Null Mice and Proteomics To Identify and Validate In Vivo Substrates of Bone Morphogenetic Protein 1/Tolloid-Like Metalloproteinases. Mol Cell Biol [Internet]. American Society for Microbiology Journals; 2003 Jul 1 [cited 2019 Feb 28];23(13):4428–4438. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12808086 PMID: 12808086 147. Li SW, Sieron AL, Fertala A, Hojima Y, Arnold W V., Prockop DJ. The C-proteinase that processes procollagens to fibrillar collagens is identical to the protein previously identified as bone morphogenic protein-1. Proc Natl Acad Sci U S A [Internet]. National Academy of Sciences; 1996 May 14 [cited 2020 Jul 13];93(10):5127–5130. Available from: /pmc/articles/PMC39418/?report=abstract 148. Petropoulou V, Garrigue-Antar L, Kadler KE. Identification of the minimal domain structure of bone morphogenetic protein-1 (BMP-1) for chordinase activity: chordinase activity is not enhanced by procollagen C-proteinase enhancer-1 (PCPE-1). J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2005 Jun 17 [cited 2018 Jan 4];280(24):22616–23. Available from: http://www.ncbi.nlm.nih.gov/pubmed/15817489 PMID: 15817489 149. Yang RT, Lim GL, Yee CT, Fuller RS, Ritchie HH. Site specificity of DSP-PP cleavage by BMP1. Connect Tissue Res [Internet]. NIH Public Access; 2014 Aug [cited 2018 Sep 27];55(SUPPL. 1):142–145. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25158199 PMID: 25158199 150. Canty EG, Garrigue-Antar L, Kadler KE. A complete domain structure of Drosophila tolloid is required for cleavage of short gastrulation. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2006 May 12 [cited 2018 Sep 19];281(19):13258–13267. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16488892 PMID: 16488892 151. Berry R, Jowitt TA, Garrigue-Antar L, Kadler KE, Baldock C. Structural and functional evidence for a substrate exclusion mechanism in mammalian tolloid like-1 (TLL-1) proteinase. FEBS Lett [Internet]. Elsevier; 2010 Feb 19 [cited 2018 Jan 3];584(4):657–61. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20043912 PMID: 20043912 152. Timothy C, Conway S, Scott I, Labosky P, Winnier G, Bundy J, Hogan B, Greenspan D. The mammalian Tolloid-like 1 gene, Tll1, is necessary for normal septation and positioning of the heart - PubMed [Internet]. Development. 1999 [cited 2020 Jul 13]. p. 2631–2642. Available from: https://pubmed.ncbi.nlm.nih.gov/10331975/ 153. Suzuki N, Labosky P, Furuta Y, Hargett L, Dunn R, Fogo A, Takahara K, Peters D, Greenspan D, Hogan B. Failure of ventral body wall closure in mouse embryos lacking a procollagen C-proteinase encoded by Bmp1, a mammalian gene related to Drosophila tolloid [Internet]. Development. 1996 [cited 2020 Jul 13]. p. 3587–3595. Available from: https://dev.biologists.org/content/develop/122/11/3587.full.pdf PMID: 8951074 154. Bayley CP, Ruiz Nivia HD, Dajani R, Jowitt TA, Collins RF, Rada H, Bird LE, Baldock C. Diversity between mammalian tolloid proteinases: Oligomerisation and non-catalytic domains influence activity and specificity. Sci Rep [Internet]. Nature Publishing Group; 2016 Feb 23 [cited 2018 Jan 3];6:21456. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26902455 PMID: 26902455 155. Berry R, Jowitt TA, Ferrand J, Roessle M, Grossmann JG, Canty-Laird EG, Kammerer RA, Kadler KE, Baldock C. Role of dimerization and substrate exclusion in the regulation of bone morphogenetic protein-1 and mammalian tolloid. Proc Natl Acad Sci U S A [Internet]. National Academy of

123

Sciences; 2009 May 26 [cited 2018 Jan 3];106(21):8561–6. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19429706 PMID: 19429706 156. Sweeney⁎ A Mac, Gil-Parrado S, Vinzenz D, Bernardi A, Hein A, Bodendorf U, Erbel P, Logel C, Gerhartz B. Structural Basis for the Substrate Specificity of Bone Morphogenetic Protein 1/Tolloid- like Metalloproteases. J Mol Biol [Internet]. 2008 [cited 2018 Jan 4];384:228–239. Available from: https://ac.els-cdn.com/S0022283608011595/1-s2.0-S0022283608011595- main.pdf?_tid=4adfec74-f17b-11e7-b1ec- 00000aab0f02&acdnat=1515089930_9faccb93929a5220438d79e3fdc985ae 157. Hu S, Chen C, Sheng J, Sun Y, Cao X, Qiao J. Enhanced Muscle Growth by Plasmid-Mediated Delivery of Myostatin Propeptide. J Biomed Biotechnol. Hindawi Publishing Corporation; 2010;2010. 158. Morine KJ, Bish LT, Pendrak K, Sleeper MM, Barton ER, Sweeney HL. Systemic Myostatin Inhibition via Liver-Targeted Gene Transfer in Normal and Dystrophic Mice. Lewin A, editor. PLoS One [Internet]. Public Library of Science; 2010 Feb 11 [cited 2020 Jul 26];5(2):e9176. Available from: https://dx.plos.org/10.1371/journal.pone.0009176 159. Cox TC, Lidral AC, McCoy JC, Liu H, Cox LL, Zhu Y, Anderson RD, Moreno Uribe LM, Anand D, Deng M, Richter CT, Nidey NL, Standley JM, Blue EE, Chong JX, Smith JD, Kirk EP, Venselaar H, Krahn KN, Bokhoven H, Zhou H, Cornell RA, Glass IA, Bamshad MJ, Nickerson DA, Murray JC, Lachke SA, Thompson TB, Buckley MF, Roscioli T. Mutations in GDF11 and the extracellular antagonist, Follistatin, as a likely cause of Mendelian forms of orofacial clefting in humans. Hum Mutat [Internet]. John Wiley and Sons Inc.; 2019 Oct 18 [cited 2020 Jun 12];40(10):1813–1825. Available from: https://onlinelibrary.wiley.com/doi/abs/10.1002/humu.23793 160. Loumaye A, De Barsy M, Nachit M, Lause P, Frateur L, Van Maanen A, Trefois P, Gruson D, Thissen JP. Role of activin A and myostatin in human cancer cachexia. J Clin Endocrinol Metab [Internet]. Endocrine Society Chevy Chase, MD ; 2015 May [cited 2017 Jan 17];100(5):2030–2038. Available from: http://press.endocrine.org/doi/10.1210/jc.2014-4318 161. Lokireddy S, Wijesoma IW, Bonala S, Wei M, Sze SK, McFarlane C, Kambadur R, Sharma M. Retraction: Myostatin is a novel tumoral factor that induces cancer cachexia. Biochem J [Internet]. Portland Press Ltd; 2016 Aug 15 [cited 2017 Jan 17];473(8):1111–1111. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22621320 PMID: 22621320 162. Paul Oh S, Yeo CY, Lee Y, Schrewe H, Whitman M, Li E. Activin type IIA and IIB receptors mediate Gdf11 signaling in axial vertebral patterning. Genes Dev. 2002;16(21):2749–2754. PMID: 12414726 163. Ozek C, Krolewski RC, Buchanan SM, Rubin LL. Growth Differentiation Factor 11 treatment leads to neuronal and vascular improvements in the hippocampus of aged mice. Sci Rep [Internet]. Nature Publishing Group; 2018 Nov 23 [cited 2019 Feb 8];8(1):17293. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30470794 PMID: 30470794 164. Zimmers TA, Jiang Y, Wang M, Liang TW, Rupert JE, Au ED, Marino FE, Couch ME, Koniaris LG. Exogenous GDF11 induces cardiac and skeletal muscle dysfunction and wasting. Basic Res Cardiol [Internet]. Springer Berlin Heidelberg; 2017 Jul 24 [cited 2017 Jun 26];112(4):48. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28647906 PMID: 28647906 165. Rinaldi F, Zhang Y, Mondragon-Gonzalez R, Harvey J, Perlingeiro RCR, Wallace G, McNally E, Hoffman E, Brown R, Kunkel L, Campbell K, Emery A, Sinha M, Jang Y, Oh J, Khong D, Wu E, Manohar R, Nakashima M, Toyono T, Akamine A, Joyner A, McPherron A, Lawler A, Lee S, Patel K, Amthor H,

124

Kirk S, Oldham J, Kambadur R, Sharma M, Dobbie P, Bass J, Wagner K, McPherron A, Winik N, Lee S, Wagner K, Liu X, Chang X, Allen R, Loffredo F, Steinhauser M, Jay S, Gannon J, Pancoast J, Yalamanchi P, Katsimpardi L, Litterman N, Schein P, Miller C, Loffredo F, Wojtkiewicz G, Thavasu P, Longhurst S, Joel S, Slevin M, Balkwill F, Egerman M, Cadena S, Gilbert J, Meyer A, Nelson H, Swalley S, Rodgers B, Eldridge J, Gamer L, Cox K, Small C, Rosen V, Souza T, Chen X, Guo Y, Sava P, Zhang J, Hill J. Treatment with rGDF11 does not improve the dystrophic muscle pathology of mdx mice. Skelet Muscle [Internet]. BioMed Central; 2016 Dec 14 [cited 2016 Dec 5];6(1):21. Available from: http://skeletalmusclejournal.biomedcentral.com/articles/10.1186/s13395-016-0092-8 166. Wagner KR, Fleckenstein JL, Amato AA, Barohn RJ, Bushby K, Escolar DM, Flanigan KM, Pestronk A, Tawil R, Wolfe GI, Eagle M, Florence JM, King WM, Pandya S, Straub V, Juneau P, Meyers K, Csimma C, Araujo T, Allen R, Parsons SA, Wozney JM, LaVallie ER, Mendell JR. A phase I/IItrial of MYO-029 in adult subjects with muscular dystrophy. Ann Neurol [Internet]. Wiley-Blackwell; 2008 May 1 [cited 2018 Sep 12];63(5):561–571. Available from: http://doi.wiley.com/10.1002/ana.21338 167. Rebbapragada A, Benchabane H, Wrana JL, Celeste AJ, Attisano L. Myostatin Signals through a Transforming Growth Factor β -Like Signaling Pathway To Block Adipogenesis Myostatin Signals through a Transforming Growth Factor ␤ -Like Signaling Pathway To Block Adipogenesis. Mol Cell Bioligy. 2003;23(20):7230–42. 168. Sidis Y, Mukherjee A, Keutmann H, Delbaere A, Sadatsuki M, Schneyer A. Biological activity of follistatin isoforms and follistatin-like-3 is dependent on differential cell surface binding and specificity for activin, myostatin, and bone morphogenetic proteins. Endocrinology. 2006;147(7):3586–3597. PMID: 16627583 169. Miura T, Kishioka Y, Wakamatsu JI, Hattori A, Hennebry A, Berry CJ, Sharma M, Kambadur R, Nishimura T. Decorin binds myostatin and modulates its activity to muscle cells. Biochem Biophys Res Commun [Internet]. Academic Press; 2006 Feb 10 [cited 2018 May 23];340(2):675–680. Available from: https://www.sciencedirect.com/science/article/pii/S0006291X0502807X?via%3Dihub PMID: 16380093 170. Nolan K, Kattamuri C, Rankin SA, Read RJ, Zorn AM, Thompson TB. Structure of Gremlin-2 in Complex with GDF5 Gives Insight into DAN-Family-Mediated BMP Antagonism. Cell Rep [Internet]. NIH Public Access; 2016 Aug 23 [cited 2018 Mar 29];16(8):2077–2086. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27524626 PMID: 27524626 171. Nolan K, Thompson TB. The DAN family: Modulators of TGF-β signaling and beyond. Protein Sci [Internet]. Wiley-Blackwell; 2014 Aug [cited 2019 Jul 16];23(8):999–1012. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24810382 PMID: 24810382 172. Kattamuri C, Luedeke DM, Nolan K, Rankin SA, Greis KD, Zorn AM, Thompson TB. Members of the DAN family are BMP antagonists that form highly stable noncovalent dimers. J Mol Biol [Internet]. NIH Public Access; 2012 Dec 14 [cited 2018 Sep 12];424(5):313–27. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23063586 PMID: 23063586 173. Amthor H, Nicholas G, McKinnell I, Kemp CF, Sharma M, Kambadur R, Patel K. Follistatin complexes Myostatin and antagonises Myostatin-mediated inhibition of myogenesis. Dev Biol. 2004;270(1):19–30. PMID: 15136138 174. Szláma G, Kondás K, Trexler M, Patthy L. WFIKKN1 and WFIKKN2 bind growth factors TGFβ1, BMP2

125

and BMP4 but do not inhibit their signalling activity. FEBS J [Internet]. 2010 Dec [cited 2016 Dec 30];277(24):5040–5050. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21054789 PMID: 21054789 175. Sako D, Grinberg A V., Liu J, Davies M V., Castonguay R, Maniatis S, Andreucci AJ, Pobre EG, Tomkinson KN, Monnell TE, Ucran JA, Martinez-Hackert E, Pearsall RS, Underwood KW, Seehra J, Kumar R. Characterization of the ligand binding functionality of the extracellular domain of activin receptor type IIB. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2010 Jul 2 [cited 2016 Dec 31];285(27):21037–21048. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20385559 PMID: 20385559 176. Heinecke K, Seher A, Schmitz W, Mueller TD, Sebald W, Nickel J. Receptor oligomerization and beyond: a case study in bone morphogenetic proteins. BMC Biol [Internet]. BioMed Central; 2009 Sep 7 [cited 2019 Feb 8];7:59. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19735544 PMID: 19735544 177. Greenwald J, Groppe J, Gray P, Wiater E, Kwiatkowski W, Vale W, Choe S. The BMP7/ActRII extracellular domain complex provides new insights into the cooperative nature of receptor assembly. Mol Cell [Internet]. Cell Press; 2003 Mar 1 [cited 2018 Sep 18];11(3):605–617. Available from: https://www.sciencedirect.com/science/article/pii/S1097276503000947?via%3Dihub 178. Thompson TB, Woodruff TK, Jardetzky TS. Structures of an ActRIIB:activin a complex reveal a novel binding mode for TGF-β ligand:receptor interactions. EMBO J [Internet]. European Molecular Biology Organization; 2003 Apr 1 [cited 2018 Sep 18];22(7):1555–1566. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12660162 PMID: 12660162 179. Walker RG, Poggioli T, Katsimpardi L, Buchanan SM, Oh J, Wattrus S, Heidecker B, Fong YW, Rubin LL, Ganz P, Thompson TB, Wagers AJ, Lee RT. Biochemistry and Biology of GDF11 and Myostatin: Similarities, Differences, and Questions for Future Investigation. Circulation Research. 2016. PMID: 27034275 180. Eisenberg D, Schwarz E, Komaromy M, Wall R. Analysis of membrane and surface protein sequences with the hydrophobic moment plot. J Mol Biol [Internet]. 1984 Oct 15 [cited 2019 Jan 28];179(1):125–42. Available from: http://www.ncbi.nlm.nih.gov/pubmed/6502707 PMID: 6502707 181. Ibrahim BS, Pattabhi V. Crystal structure of trypsin–turkey egg white inhibitor complex. Biochem Biophys Res Commun [Internet]. Academic Press; 2004 Jan 2 [cited 2018 Apr 4];313(1):8–16. Available from: https://www.sciencedirect.com/science/article/pii/S0006291X03024574?via%3Dihub 182. JOHANSSON MW, KEYSER P, SODERHALL K. Purification and cDNA cloning of a four-domain Kazal proteinase inhibitor from crayfish blood cells. Eur J Biochem [Internet]. Wiley/Blackwell (10.1111); 1994 Jul 1 [cited 2018 Apr 5];223(2):389–394. Available from: http://doi.wiley.com/10.1111/j.1432-1033.1994.tb19005.x 183. Pariani S, Contreras M, Rossi FR, Sander V, Corigliano MG, Simón F, Busi M V., Gomez-Casati DF, Pieckenstain FL, Duschak VG, Clemente M. Characterization of a novel Kazal-type serine proteinase inhibitor of Arabidopsis thaliana. Biochimie [Internet]. Elsevier; 2016 Apr 1 [cited 2018 May 17];123:85–94. Available from: https://www.sciencedirect.com/science/article/pii/S0300908416000456?via%3Dihub

126

184. Keutmann HT, Schneyer AL, Sidis Y. The Role of Follistatin Domains in Follistatin Biological Action. Mol Endocrinol [Internet]. Oxford University Press; 2004 Jan 1 [cited 2018 Apr 4];18(1):228–240. Available from: https://academic.oup.com/mend/article-lookup/doi/10.1210/me.2003-0112 185. Geng Y, Dong Y, Yu M, Zhang L, Yan X, Sun J, Qiao L, Geng H, Nakajima M, Furuichi T, Ikegawa S, Gao X, Chen Y-G, Jiang D, Ning W. Follistatin-like 1 (Fstl1) is a bone morphogenetic protein (BMP) 4 signaling antagonist in controlling mouse lung development. Proc Natl Acad Sci [Internet]. National Academy of Sciences; 2011 Apr 26 [cited 2018 Sep 18];108(17):7058–7063. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21482757 PMID: 21482757 186. Patthy L, Nikolics K. Functions of agrin and agrin-related proteins. Trends Neurosci [Internet]. Elsevier Current Trends; 1993 Feb 1 [cited 2018 Sep 18];16(2):76–81. Available from: https://www.sciencedirect.com/science/article/pii/016622369390021D?via%3Dihub 187. Chang C, Eggen BJL, Weinstein DC, Brivanlou AH. Regulation of nodal and BMP signaling by tomoregulin-1 (X7365) through novel mechanisms. Dev Biol [Internet]. Academic Press; 2003 Mar 1 [cited 2018 Sep 18];255(1):1–11. Available from: https://www.sciencedirect.com/science/article/pii/S0012160602000751?via%3Dihub 188. Adams PD, Afonine P V., Bunkóczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Terwilliger TC, Zwart PH. PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr Sect D Biol Crystallogr [Internet]. International Union of Crystallography; 2010 Feb 1 [cited 2018 Sep 16];66(2):213–221. Available from: http://scripts.iucr.org/cgi-bin/paper?S0907444909052925 189. Terwilliger TC, Adams PD, Read RJ, McCoy AJ, Moriarty NW, Grosse-Kunstleve RW, Afonine P V., Zwart PH, Hung LW. Decision-making in structure solution using Bayesian estimates of map quality: The PHENIX AutoSol wizard. Acta Crystallogr Sect D Biol Crystallogr [Internet]. International Union of Crystallography; 2009 Jun 1 [cited 2018 Sep 16];65(6):582–601. Available from: http://scripts.iucr.org/cgi-bin/paper?S0907444909012098 190. Afonine P V., Grosse-Kunstleve RW, Echols N, Headd JJ, Moriarty NW, Mustyakimov M, Terwilliger TC, Urzhumtsev A, Zwart PH, Adams PD. Towards automated crystallographic structure refinement with phenix.refine. Acta Crystallogr Sect D Biol Crystallogr [Internet]. International Union of Crystallography; 2012 Apr 1 [cited 2018 Sep 16];68(4):352–367. Available from: http://scripts.iucr.org/cgi-bin/paper?S0907444912001308 191. Williams CJ, Headd JJ, Moriarty NW, Prisant MG, Videau LL, Deis LN, Verma V, Keedy DA, Hintze BJ, Chen VB, Jain S, Lewis SM, Arendall WB, Snoeyink J, Adams PD, Lovell SC, Richardson JS, Richardson DC. MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci [Internet]. 2018 Jan [cited 2018 Sep 21];27(1):293–315. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29067766 PMID: 29067766 192. Chen VB, Arendall WB, Headd JJ, Keedy DA, Immormino RM, Kapral GJ, Murray LW, Richardson JS, Richardson DC. MolProbity: All-atom structure validation for macromolecular crystallography. Acta Crystallogr Sect D Biol Crystallogr [Internet]. International Union of Crystallography; 2010 Jan [cited 2018 Sep 21];66(1):12–21. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20057044 PMID: 20057044 193. Di Tommaso P, Moretti S, Xenarios I, Orobitg M, Montanyola A, Chang JM, Taly JF, Notredame C.

127

T-Coffee: A web server for the multiple sequence alignment of protein and RNA sequences using structural information and homology extension. Nucleic Acids Res [Internet]. Oxford University Press; 2011 Jul 1 [cited 2018 Sep 16];39(SUPPL. 2):W13–W17. Available from: https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkr245 194. L DW. Pymol: An open-source molecular graphics tool [Internet]. {CCP4} Newsletter On Protein Crystallography. 2002. Available from: https://www.ccp4.ac.uk/newsletters/newsletter40/11_pymol.pdf 195. Larkin MA, Blackshields G, Brown NP, Chenna R, Mcgettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG. Clustal W and Clustal X version 2.0. Bioinformatics [Internet]. Oxford University Press; 2007 Nov 1 [cited 2018 Sep 16];23(21):2947– 2948. Available from: https://academic.oup.com/bioinformatics/article- lookup/doi/10.1093/bioinformatics/btm404 196. Bateman A, Martin MJ, O’Donovan C, Magrane M, Alpi E, Antunes R, Bely B, Bingley M, Bonilla C, Britto R, Bursteinas B, Bye-AJee H, Cowley A, Da Silva A, De Giorgi M, Dogan T, Fazzini F, Castro LG, Figueira L, Garmiri P, Georghiou G, Gonzalez D, Hatton-Ellis E, Li W, Liu W, Lopez R, Luo J, Lussi Y, MacDougall A, Nightingale A, Palka B, Pichler K, Poggioli D, Pundir S, Pureza L, Qi G, Rosanoff S, Saidi R, Sawford T, Shypitsyna A, Speretta E, Turner E, Tyagi N, Volynkin V, Wardell T, Warner K, Watkins X, Zaru R, Zellner H, Xenarios I, Bougueleret L, Bridge A, Poux S, Redaschi N, Aimo L, ArgoudPuy G, Auchincloss A, Axelsen K, Bansal P, Baratin D, Blatter MC, Boeckmann B, Bolleman J, Boutet E, Breuza L, Casal-Casas C, De Castro E, Coudert E, Cuche B, Doche M, Dornevil D, Duvaud S, Estreicher A, Famiglietti L, Feuermann M, Gasteiger E, Gehant S, Gerritsen V, Gos A, Gruaz- Gumowski N, Hinz U, Hulo C, Jungo F, Keller G, Lara V, Lemercier P, Lieberherr D, Lombardot T, Martin X, Masson P, Morgat A, Neto T, Nouspikel N, Paesano S, Pedruzzi I, Pilbout S, Pozzato M, Pruess M, Rivoire C, Roechert B, Schneider M, Sigrist C, Sonesson K, Staehli S, Stutz A, Sundaram S, Tognolli M, Verbregue L, Veuthey AL, Wu CH, Arighi CN, Arminski L, Chen C, Chen Y, Garavelli JS, Huang H, Laiho K, McGarvey P, Natale DA, Ross K, Vinayaka CR, Wang Q, Wang Y, Yeh LS, Zhang J. UniProt: The universal protein knowledgebase. Nucleic Acids Res [Internet]. Oxford University Press; 2017 Jan 4 [cited 2018 Sep 17];45(D1):D158–D169. Available from: https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkw1099 197. Sengle G, Ono RN, Sasaki T, Sakai LY. Prodomains of transforming growth factor β (TGFβ) Superfamily members specify different functions: Extracellular matrix interactions and growth factor bioavailability. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2011 Feb 18 [cited 2020 Jul 12];286(7):5087–5099. Available from: /pmc/articles/PMC3037620/?report=abstract 198. Innis CA, Hyvönen M. Crystal structures of the heparan sulfate-binding domain of follistatin. Insights into ligand binding. J Biol Chem [Internet]. American Society for Biochemistry and Molecular Biology; 2003 Oct 10 [cited 2018 Apr 5];278(41):39969–77. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12867435 PMID: 12867435 199. Szláma G, Vásárhelyi V, Trexler M, Patthy L. Influence of WFIKKN1 on BMP1-mediated activation of latent myostatin. FEBS J [Internet]. 2016 Dec [cited 2017 Oct 26];283(24):4515–4527. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27782377 PMID: 27782377 200. Poggioli T, Vujic A, Yang P, MacIas-Trevino C, Uygur A, Loffredo FS, Pancoast JR, Cho M, Goldstein J, Tandias RM, Gonzalez E, Walker RG, Thompson TB, Wagers AJ, Fong YW, Lee RT. Circulating Growth Differentiation Factor 11/8 Levels Decline with Age. Circ Res. 2016;118(1):29–37. PMID:

128

26489925

129

Appendix A

Abstracts from publications authored and co-authored*

*Title, authors, doi, PMID, and PMCID are shown when applicable. Publications are in chronological order.

130

1) Structural Basis for Potency Differences between GDF8 and GDF11. BMC biology

Walker R.G., Czepnik M., Goebel E.J., McCoy J.C., Vujic A., Cho M., Oh J., Aykul S., Walton K.L., Schang G.,

Bernard D.J., Hinck A.P., Harrison C.A., Martinez-Hackert E., Wagers A.J., Lee R.T., Thompson T.B. (2017).

Mar 3;15(1):19 PMCID:PMC5336696.

Background

Growth/differentiation factor 8 (GDF8) and GDF11 are two highly similar members of the transforming growth factor β (TGFβ) family. While GDF8 has been recognized as a negative regulator of muscle growth and differentiation, there are conflicting studies on the function of GDF11 and whether GDF11 has beneficial effects on age-related dysfunction. To address whether GDF8 and GDF11 are functionally identical, we compared their signaling and structural properties.

Results

Here we show that, despite their high similarity, GDF11 is a more potent activator of SMAD2/3 and signals more effectively through the type I activin-like receptor kinase receptors ALK4/5/7 than GDF8. Resolution of the GDF11:FS288 complex, apo-GDF8, and apo-GDF11 crystal structures reveals unique properties of both ligands, specifically in the type I receptor binding site. Lastly, substitution of GDF11 residues into

GDF8 confers enhanced activity to GDF8.

Conclusions

These studies identify distinctive structural features of GDF11 that enhance its potency, relative to GDF8; however, the biological consequences of these differences remain to be determined.

2) Molecular characterization of latent GDF8 reveals mechanisms of activation.

Walker RG, McCoy JC (Co-First Author), Czepnik M, Mills MJ, Hagg A, Walton KL, Cotton TR, Hyvönen M,

Lee RT, Gregorevic P, Harrison CA, Thompson TB. (2018) PNAS 115(5):E866-E875, PMCID: PMC5798348

Abstract

131

Growth/differentiation factor 8 (GDF8), or myostatin, negatively regulates muscle mass. GDF8 is held in a latent state through interactions with its N-terminal prodomain, much like TGF-β. Using a combination of small-angle X-ray scattering and mutagenesis, we characterized the interactions of GDF8 with its prodomain. Our results show that the prodomain:GDF8 complex can exist in a fully latent state and an activated or “triggered” state where the prodomain remains in complex with the mature domain.

However, these states are not reversible, indicating the latent GDF8 is “spring-loaded.” Structural analysis shows that the prodomain:GDF8 complex adopts an “open” configuration, distinct from the latency state of TGF-β and more similar to the open state of Activin A and BMP9 (nonlatent complexes). We determined that GDF8 maintains similar features for latency, including the alpha-1 helix and fastener elements, and identified a series of mutations in the prodomain of GDF8 that alleviate latency, including I56E, which does not require activation by the protease Tolloid. In vivo, active GDF8 variants were potent negative regulators of muscle mass, compared with WT GDF8. Collectively, these results help characterize the latency and activation mechanisms of GDF8.

3) Structure of the human myostatin precursor and determinants of growth factor latency.

Cotton TR, Fischer G, Wang X, McCoy JC, Czepnik M, Thompson TB, Hyvönen M. (2018) EMBO J 37(3):367-

383. doi: 10.15252/embj.201797883, PMID: 29330193, PMCID: PMC5793801

Abstract

Myostatin, a key regulator of muscle mass in vertebrates, is biosynthesised as a latent precursor in muscle and is activated by sequential proteolysis of the pro-domain. To investigate the molecular mechanism by which pro-myostatin remains latent, we have determined the structure of unprocessed pro-myostatin and analysed the properties of the protein in its different forms. Crystal structures and SAXS analyses show that pro-myostatin adopts an open, V-shaped structure with a domain-swapped arrangement. The pro-mature complex, after cleavage of the furin site, has significantly reduced activity compared with the mature growth factor and persists as a stable complex that is resistant to the natural antagonist follistatin.

132

The latency appears to be conferred by a number of distinct features that collectively stabilise the interaction of the pro-domains with the mature growth factor, enabling a regulated stepwise activation process, distinct from the prototypical pro-TGF-β1. These results provide a basis for understanding the effect of missense mutations in pro-myostatin and pave the way for the design of novel myostatin inhibitors

4) Crystal structure of the WFIKKN2 follistatin domain reveals insight into how it inhibits growth differentiation factor 8 (GDF8) and GDF11.

McCoy JC, Walker RG, Murray NH, Thompson TB. (2019) Journal of Biochemistry. RA118.005831. doi:

10.1074/jbc.RA118.005831.

Abstract

Growth differentiation factor 8 (GDF8; also known as myostatin) and GDF11 are closely related members of the transforming growth factor β (TGF-β) family. GDF8 strongly and negatively regulates skeletal muscle growth, and GDF11 has been implicated in various age-related pathologies such as cardiac hypertrophy.

GDF8 and GDF11 signaling activities are controlled by the extracellular protein antagonists follistatin; follistatin-like 3 (FSTL3); and WAP, follistatin/kazal, immunoglobulin, Kunitz, and netrin domain- containing (WFIKKN). All of these proteins contain a follistatin domain (FSD) important for ligand binding and antagonism. Here, we investigated the structure and function of the FSD from murine WFIKKN2 and compared it with the FSDs of follistatin and FSTL3. Using native gel shift and surface plasmon resonance analyses, we determined that the WFIKKN2 FSD can interact with both GDF8 and GDF11 and block their interactions with the type II receptor activin A receptor type 2B (ActRIIB). Further, we solved the crystal structure of the WFIKKN2 FSD to 1.39 Å resolution and identified surface-exposed residues that, when substituted with alanine, reduce antagonism of GDF8 in full-length WFIKKN2. Comparison of the WFIKKN2

FSD with those of follistatin and FSTL3 revealed differences in both the FSD structure and position of residues within the domain that are important for ligand antagonism. Taken together, our results indicate

133 that both WFIKKN and follistatin utilize their FSDs to block the type II receptor but do so via different binding interactions.

5) Mutations in GDF11 and the extracellular antagonist, Follistatin, as a likely cause of Mendelian forms of orofacial clefting in humans.

Timothy C. Cox, Andrew C. Lidral, Jason C. McCoy, et. al (2019) Human Mutation. Human Mutation

40(10):1813-1825. doi: 10.1002/humu.23793

Abstract

Cleft lip with or without cleft palate (CL/P) is generally viewed as a complex trait with multiple genetic and environmental contributions. In 70% of cases, CL/P presents as an isolated feature and/or deemed nonsyndromic. In the remaining 30%, CL/P is associated with multisystem phenotypes or clinically recognizable syndromes, many with a monogenic basis. Here we report the identification, via exome sequencing, of likely pathogenic variants in two genes that encode interacting proteins previously only linked to orofacial clefting in mouse models. A variant in GDF11 (encoding growth differentiation factor

11), predicting a p.(Arg298Gln) substitution at the Furin protease cleavage site, was identified in one family that segregated with CL/P and both rib and vertebral hypersegmentation, mirroring that seen in

Gdf11 knockout mice. In the second family in which CL/P was the only phenotype, a mutation in FST

(encoding the GDF11 antagonist, Follistatin) was identified that is predicted to result in a p.(Cys56Tyr) substitution in the region that binds GDF11. Functional assays demonstrated a significant impact of the specific mutated amino acids on FST and GDF11 function and, together with embryonic expression data, provide strong evidence for the importance of GDF11 and Follistatin in the regulation of human orofacial development.

6) Structural biology of the TGFβ family. Experimental Biology and Medicine.

134

Goebel EJ, Hart KN, McCoy JC, Thompson TB (2019) 0:1-17. doi: 10.1177/1535370219880894, PMID:

31594405. PMCID: PMC6920667

Abstract

The transforming growth factor beta (TGFβ) signaling pathway orchestrates a wide breadth of biological processes, ranging from bone development to reproduction. Given this, there has been a surge of interest from the drug development industry to modulate the pathway – at several points. This review discusses and provides additional context for several layers of the TGFβ signaling pathway from a structural biology viewpoint. The combination of structural techniques coupled with biophysical studies has provided a foundational knowledge of the molecular mechanisms governing this high impact, ubiquitous pathway, underlying many of the current therapeutic pursuits. This work seeks to consolidate TGFβ-related structural knowledge and educate other researchers of the apparent gaps that still prove elusive. We aim to highlight the importance of these structures and provide the contextual information to understand the contribution to the field, with the hope of advancing the discussion and exploration of the TGFβ signaling pathway.

7) Structural Perspective of BMP Ligands and Signaling

Gregory R. Gipson, Erich J. Goebel, Kaitlin N. Hart, Emily C. Kappes, Chandramohan Kattamuri, Jason C.

McCoy, Dr. Thomas Thompson (2020). Bone. doi: 10.1016/j.bone.2020.115549

Abstract

The Bone Morphogenetic Proteins (BMPs) are the largest class signaling molecules within the greater

Transforming Growth Factor Beta (TGFβ) family, and are responsible for a wide array of biological functions, including dorsal-ventral patterning, skeletal development and maintenance, as well as cell homeostasis. As such, dysregulation of BMPs results in a number of diseases, including fibrodysplasia ossificans progressiva (FOP) and pulmonary arterial hypertension (PAH). Therefore, understanding BMP signaling and regulation at the molecular level is essential for targeted therapeutic intervention. This

135 review discusses the recent advances in the structural and biochemical characterization of BMPs, from canonical ligand-receptor interactions to co-receptors and antagonists. This work aims to highlight how

BMPs differ from other members of the TGFβ family, and how that information can be used to further advance the field. Lastly, this review discusses several gaps in the current understanding of BMP structures, with the aim that discussion of these gaps will lead to advancements in the field.

136

Appendix B

Curriculum Vitae

137

Jason C. McCoy Curriculum Vitae ______

720 Clifton Colony Dr. Apt2 ● Cincinnati, Ohio 45220 ● (740) 497-1616 ● [email protected]

Education 08/11-05/14 Miami University, Oxford, OH Bachelor of Science, Microbiology Minor: Molecular Genetics Thematic Sequence, Biochemistry Cumulative GPA: 3.61 Science GPA: 3.81

08/15-present University of Cincinnati, Cincinnati, OH PhD Graduate Student, Department of Molecular Genetics, Biochemistry and Microbiology. Cumulative GPA: 3.59 Dr. Tom Thompson Laboratory

Honors and Awards

2013-2015 Dean’s List, last 5 semesters of my undergraduate education.

2015 The Orton K. Stark Award. Miami University, Department of Microbiology.

2014 Howard Hughes Summer Scholar, 9 paid week internship to work within a research laboratory full time. 2014 Undergraduate Research Award, supplied funding for my undergraduate research 2017-2018 T32 Training Grant: Environmental Carcinogenesis and Mutagenesis

2018-2020 American Heart Association Pre-doctoral Fellowship. 2 years of outside funding for my ongoing research. 18PRE33990312 2018-2020 Albert J. Ryan Foundation Fellowship. 2 years of additional stipend funding.

2018 International BMP Conference Travel Award Recipient. Tokyo, Japan

2019 Young investigator award. Society for Experimental Biology and Medicine. Experimental biology meeting. Orlando, Florida 2019 ASBMB 2019 Graduate/Postdoctoral Travel Award Recipient. Experimental biology meeting. Orlando, Florida 2020 Scientific Research Award. University of Cincinnati department of molecular genetics, biochemistry, and microbiology. Cincinnati, Ohio Publications

1. Garretson TA, McCoy JC, Cheng WX. (2016) Baculovirus FP25K Localization: Role of the Coiled-Coil Domain. Journal of virology 90(21):9582-9597

138

2. Walker R.G., Czepnik M., Goebel E.J., McCoy J.C., Vujic A., Cho M., Oh J., Aykul S., Walton K.L., Schang G., Bernard D.J., Hinck A.P., Harrison C.A., Martinez-Hackert E., Wagers A.J., Lee R.T., Thompson T.B. (2017) Structural Basis for Potency Differences between GDF8 and GDF11. BMC biology. Mar 3;15(1):19 PMCID:PMC5336696.

3. Walker RG, McCoy JC (Co-First Author), Czepnik M, Mills MJ, Hagg A, Walton KL, Cotton TR, Hyvönen M, Lee RT, Gregorevic P, Harrison CA, Thompson TB. (2018) Molecular characterization of latent GDF8 reveals mechanisms of activation. PNAS 115(5):E866-E875, PMCID: PMC5798348

4. Cotton TR, Fischer G, Wang X, McCoy JC, Czepnik M, Thompson TB, Hyvönen M. (2018) Structure of the human myostatin precursor and determinants of growth factor latency. EMBO J 37(3):367-383

5. McCoy JC, Walker RG, Murray NH, Thompson TB. (2019) Crystal structure of the WFIKKN2 follistatin domain reveals insight into how it inhibits growth differentiation factor 8 (GDF8) and GDF11. JBC. RA118.005831. doi: 10.1074/jbc.RA118.005831.

6. Timothy C. Cox, Andrew C. Lidral, Jason C. McCoy, et. al (2019). Mutations in GDF11 and the extracellular antagonist, Follistatin, as a likely cause of Mendelian forms of orofacial clefting in humans. Human Mutation. Human Mutation 40(10):1813-1825. doi: 10.1002/humu.23793

7. Goebel EJ, Hart KN, McCoy JC, Thompson TB (2019). Structural biology of the TGFβ family. Experimental Biology and Medicine. 0:1-17. doi: 10.1177/1535370219880894

8. Gregory R. Gipson, Erich J. Goebel, Kaitlin N. Hart, Emily C. Kappes, Chandramohan Kattamuri, Jason C. McCoy, Dr. Thomas Thompson (2020). Structural Perspective of BMP Ligands and Signaling. Bone. Under Review.

Research and Laboratory Experience

2012-2015 Miami University Department of Chemistry. Laboratory assistant under Lijie Yang. 2014-2015 Miami University Department of Microbiology. Dr. Xiao-Wen Cheng, undergraduate researcher. Led to a publication within the Journal of Virology 2015-present University of Cincinnati Department of Molecular Genetics, Biochemistry and Microbiology. PhD candidate within the Dr. Thomas Thompson lab with two publications to date. 2018 CCP4 summer workshop. One-week intensive X-ray crystallographic workshop learning data collection and processing to solve protein structure.

Teaching/Mentorship

2016-2018 Mentored the Undergraduate Austin Songer in mutagenesis and protein purification. 2020 Mentored the Graduate rotation student Kierra Ware in bacterial protein production and purification. Presentations and Conferences

139

2017 Hitchhikers guide to the biomolecular galaxy. Purdue university, poster presentation. 2017 Gibbs society of biological thermodynamics. Poster presentation.

2017 Graduate student research forum. University of Cincinnati poster presentation.

2018 International BMP Conference. Oral presentation in Tokyo, Japan.

2019 Graduate Student Expo. Poster Presentation University of Cincinnati.

2019 Experimental Biology Conference. Oral and Poster presentation in Orlando, Florida 2019 Ryan Symposium. Oral Presentation.

2019 The TGF-β Superfamily Conference. Poster presentation in West Palm Beach, Florida.

Other Experience and Extracurricular

2012-2016 Miami Microbiology Club.

2012-2016 Phi Delta Epsilon Premedical Fraternity. Graduate Committee Chair 2016

2017-2018 Health Sciences Graduate Association (GSGA). Philanthropy Chair

2018-present American Society for Biochemistry and Molecular Biology (ASBMB) Member

2018-Present Society for Experimental Biology and Medicine (SEBM) Member

140