Structure, Catalytic Mechanism, and Membrane Interaction of the Mtor Activator Rheb

Structure, catalytic mechanism, and membrane interaction of the mTOR activator Rheb

Mohammad Taghi Mazhab Jafari

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Medical Biophysics University of Toronto

Structure, catalytic mechanism, and membrane interaction of the mTOR activator Rheb

Mohammad Taghi Mazhab Jafari

Doctor of Philosophy

Department of Medical Biophysics University of Toronto

2014 Abstract

The activator of mammalian target of rapamycin complex 1 (mTORC1), Ras homolog enriched in brain

(Rheb), is a membrane-associated protein belonging to the Ras subfamily of small GTPases. Rheb’s slow GTPase activity is stimulated by the GTPase activating protein (GAP) domain of tuberous sclerosis

complex 1 and 2 (TSC1/2). Rheb hyperactivation, through its overexpression or loss of TSC1/2 GAP

function, results in hyperactivated mTORC1 signaling culminating in tumourigenesis. The molecular

details of Rheb GTP hydrolysis and the effect of membrane association on Rheb structure, dynamics and

its GTPase function are currently not fully understood. The studies presented in this thesis focus on two

key determinants of Rheb function i) the mechanism of Rheb GTP hydrolysis and ii) the structural and

functional consequence of Rheb-bilayer membrane interaction. Through studies of fluorescent

nucleotides, we revealed that the conserved G2-box residue Tyr35 auto-inhibits GTP hydrolysis in

Rheb. We demonstrated that a non-canonical catalytic residue, Asp65, in the switch II region of Rheb,

contributed more to the GTP hydrolysis rate than Gln64, which corresponds to the canonical Ras Gln61.

This non-canonical auto-inhibited mechanism of GTP hydrolysis was required for optimal mTORC1

regulation. These structural insights were then used to guide the design of novel gain- and loss-of

function mutants by substitutions of the ultra-conserved G3-box Gly63 of Rheb. Finally, using solution

NMR spectroscopy, we monitored the Rheb GTPase cycle and characterized its nucleotide-dependent membrane orientations on nanodisc-based phospholipid bilayers. Rheb was shown to sample two orientations in which its C-terminal helix was semi-perpendicular or semi-parallel with respect to the

bilayer plane. The semi-parallel orientation, where switch II residues critical for mTORC1

communication are accessible, was favored in the GTP bound conformation, suggesting that membrane-

tethering modulates Rheb function. These structural insights into the catalytic machinery of Rheb and its

membrane interface suggest new approaches to modulate these key determinants of Rheb function

through small molecules towards development of therapeutic avenues for Rheb-mediated pathogenesis

such as tuberous sclerosis or cancer.

iii

Acknowledgments

First and foremost, I would like to thank my supervisor Dr. Mitsu Ikura for his guidance, support, and understanding. I could have not accomplished any portion of this research without his supervision. His passion for science was a continuous source of inspiration for me over the years. I learned key aspects of successful research from him, including critical thinking.

I am grateful to my supervisory committee members; Dr. Lewis Kay and Dr. Vuk Stambolic. Their expert inputs were critical to the success of this thesis. Thank you for your advise, supervision and encouragement, which kept me on the right path.

I am also thankful to my collaborators, Vanessa De Palma and Jason Ho from Dr. Stambolic lab for the cell biology experiments.

I had the honor of working with and learning from many excellent current and former members of Ikura’s lab including; Dr. Christopher Marshall, Dr. Peter Stathopulos, Dr. Genevieve Seabrook, Dr. Noboru Ishiyama, Le Zheng, Dr. Feng Wang, Carol Liu, and Dr. Fernando Amador. I would like to specially thank Chris, who was involved in all aspects of my projects and gave valuable inputs and suggestions.

Finally, I would like to thank my family; my parents Mohammad Mazhab-Jafari and Ziba Fadavi and my brother Hamed Mazhab-Jafari for all their sacrifice that provided me with the opportunity to carry out this research. Words cannot express my grateful feelings for you.

Table of Contents

Acknowledgements……………………………………………………………………………………….iv

Table of Contents ………………………………………………………………………………………...v

List of Tables …………………………………………………………………………………………….x

List of Figures …………………………………………………………………………………………...xi

List of Appendices ……………………………………………………………………………………...xiv

List of Abbreviations …………………………………………………………………………………...xv

Chapter 1: Introduction and Thesis Overview ………………………………………………………...1

1.1 General introduction to small GTPase proteins……………………………………………………….2

1.2 Introduction to Rheb………………………………………………………………...... 3

1.3 Biophysical and biochemical properties of Rheb…………………………………...... 6

1.4 Pathogenesis of Rheb………………………………………………………………………………….9

1.5 Thesis Overview and Rationale……………………………………………………………………...11

1.6 Attributions…………………………………………………………………………………………..12

1.7 References……………………………………………………………………………………………12

Chapter 2: Fluorescent-tagged Nucleotides Alter the Native GTPase Cycle …………………….…17

2.1 Abstract………………………………………………………………………………………………18

2.2 Introduction………………………………………………………………………...... 19

2.3 Results………………………………………………………………………………………………..20

2.3.1 The effect of mant on the intrinsic rate of nucleotide hydrolysis………………………….20

2.3.2 The effect of mant on the rate of GAP-catalyzed nucleotide hydrolysis…………………..23

2.3.3 The effect of mant on the GEF-mediated nucleotide exchange…………………………....25

2.4 Discussion……………………………………………………………………………………………28

2.4.1 Intrinsic hydrolysis of GTP and mantGTP………………………………………………...28

2.4.2 GAP-catalyzed GTP hydrolysis by HRas and Rheb…………………………………….…30

2.4.3 GEF-accelerated nucleotide exchange of HRas and RhoA………………………………..31

2.5 Experimental Procedures…………………………………………………………………………….32

2.5.1 Protein preparation…………………………………………………………………………32

2.5.2 NMR-based GTPase, GAP and GEF assays……………………………………………….33

2.5.3 Fluorescence-based GTPase assay…………………………………………………………34

2.6 References……………………………………………………………………………………………34

Chapter 3: Mechanism of GTP hydrolysis by Rheb ……………………………………………..…...36

3.1 Abstract…………………………………………………………………………………………..…..37

3.2 Introduction………………………………………………………………………………………..…38

3.3 Results………………………………………………………………………………………………..39

3.3.1 Rheb Tyr35 inhibits intrinsic GTPase activity……………………………………………..39

3.3.2 Structural basis for the Tyr35 auto-inhibitory function……………………………………40

3.3.3 Identification of a catalytic residue for GTP hydrolysis…………………………………...42

3.3.4 Involvement of Rheb’s Asp65 and Tyr35 in TSC2GAP-mediated GTP hydrolysis…...... 47

3.3.5 Thermodynamic basis for the Tyr35 auto-inhibitory function…………………………….49

3.3.6 Regulation of mTORC1 by growth factors involves the non-canonical catalytic and autoinhibitory mechanisms…………………………………………………………….………...54

3.4 Discussion……………………………………………………………………………………….…...57

3.5 Experimental Procedures……………………………………………………………………….……60

3.5.1 Protein preparation………………………………………………………………….……...60

3.5.2 Crystallization and data collection………………………………………………….……...61

3.5.3 Structure determination and refinement…………………………………………….……...61

3.5.4 NMR-based GTPase assays…………………………………………………………….…62

3.5.5 Thermodynamic measurements……………………………………………………….…..63

3.5.6 Cell-based phosphorylation assay……………………………………………………..…...63

3.5.7 Nucleotide binding in vivo…………………………………………………………………64

3.6 References……………………………………………………………………………………………65

Chapter 4: Structure-guided design of novel active and inactive Rheb mutants using single site modifications ……………………………………………………………………………………………68

4.1 Abstract……………………………………………………………………………………………....69

4.2 Introduction…………………………………………………………………………………………..70

4.3 Results and Discussion……………………………………………………………………………....72

4.4 Experimental Procedures…………………………………………………………………………….78

4.4.1 Protein Preparation…………………………………………………………………………78

4.4.2 NMR-based Real-time GTPase assay……………………………………………………...78

4.4.3 Crystallization and Data Collection………………………………………………………..79

vii

4.4.4 Structure Determination and Refinement………………………………………………….79

4.4.5 Cell-based Phosphorylation Assays………………………………………………………..80

4.5 References……………………………………………………………………………………………82

Chapter 5: Structural and Functional consequence of Rheb-membrane interaction ……………...84

5.1 Abstract………………………………………………………………………………………………85

5.2 Introduction…………………………………………………………………………………………..86

5.3 Results and Discussion………………………………………………………………………………88

5.4 Experimental Procedures……………………………………………………………………….……99

5.4.1 Protein preparations………………………………………………………………….…….99

5.4.2 Preparation of Rheb-nanodisc complex…………………………………………………..100

5.4.3 NMR- measurements……………………………………………………………………..101

5.4.4 Real-time NMR-based GTPase assay………………………………………………….…103

5.4.5 Molecular docking simulation……………………………………………………………103

5.5 References……………………………………………………………………………………….….106

Chapter 6: Conclusions and Future directions ……………...………………………………………108

6.1 Conclusions…………………………………………………………………………………………109

6.2 Future Directions…………………………………………………………………………………...110

6.2.1 Engineering GTPase probes by structure-guided mutations of G3-box glycine………....110

6.2.2 Drugging Rheb……………………………………………………………………………111

6.2.3 Probing membrane-dependent regulation of KRas-effector interaction………………….111

viii

6.3 Closing Remarks……………………………………………………………………………………112

6.4 References…………………………………………………………………………………………..114

Appendix A……………………………………………………………………………………………..116

Appendix B……………………………………………………………………………………………..118

List of Tables

Table 2.1 The rates of intrinsic and GAP mediated nucleotide hydrolysis for three small GTPase proteins…………………………………………………………………………………………………..23

Table 2.2: The rates of GEF mediated nucleotide exchange for HRas and RhoA proteins……………..27

Table 3.1 Summary of thermodynamic activation parameters…………………………………………..54

Table 3.2 Data collection and refinement statistics……………………………………………………...62

Table 4.1 Data collection and refinement statistics……………………………………………………...81

Table 5.1 Resonance assignment of Rheb HVR………………………………………………….…....101

List of Figures

Figure 1.1 Sequence alignment of Rheb with other small GTPase proteins investigated in this thesis...... 3

Figure 1.2 Schematic illustration of Rheb in the mTORC1 signaling pathway…………………………..5

Figure 1.3 Overall structure of active Rheb and comparison with that of HRas……………………….....8

Figure 2.1 Effect of mant-substituted GTP on the intrinsic nucleotide hydrolysis of Ras, Rheb and RhoA……………………………………………………………………………………………………..19

Figure 2.2 Comparison of the real-time NMR and fluorescence-substituted nucleotide-based GTPase assays…………………………………………………………………………………………………….21

Figure 2.3 Perturbation of the native structure of GTPase proteins by mant-tagged nucleotides……….22

Figure 2.4 Effects of mant-substituted GTP on GAP-mediated nucleotide hydrolysis by HRas and Rheb……………………………………………………………………………………………………...24

Figure 2.5 Effects of mant-substituted nucleotides on GEF-mediated nucleotide exchange for HRas and RhoA……………………………………………………………………………………………………..26

Figure 2.6 Complementary assays of the effects of mant-substituted nucleotides on GEF-mediated nucleotide exchange for HRas and RhoA………………………………………………………………..27

Figure 2.7 mantGTP-induced chemical shift perturbations in Ras, Rheb and RhoA……………………29

Figure 2.8 Rheb Gln64 is not involved in the rapid hydrolysis of mantGTP……………………………30

Figure 3.1 Rapid hydrolysis of mantGTP by Rheb is related to autoinhibitory role of Tyr35…………..40

Figure 3.2 Structure and dynamics of Rheb Y35A……………………………………………………....41

Figure 3.3 Mutagenic analysis of potential catalytic residues in Rheb Y35A…………………………...43

Figure 3.4 Role of Asp65 in intrinsic and GAP-mediated GTP hydrolysis by Rheb……………………45

Figure 3.5 Effect on Rheb GTPase activity of mutations of acidic and electronegative residues in switch II………………………………………………………………………………………………………….46

Figure 3.6. Structure-function analysis of the position of putative catalytic residues and the conserved switch I Tyrosine in small GTPases……………………………………………………………………...48

Figure 3.7 Perturbation of Rheb HSQC spectrum by Tyr35 mutation…………………………………..50

Figure 3.8 Minimal chemical shift perturbation associated with mutation of Rheb Asp65…………….51

Figure 3.9 Tyr35 hydroxyl is required for TSC2GAP-assisted GTP hydrolysis………………………...52

Figure 3.10 Effect of Tyr35 and Asp65 mutations on the thermodynamic activation parameters for GTP hydrolysis by Rheb………………………………………………………………………………………53

Figure 3.11 Mutations of Rheb catalytic and autoinhibitory residues impact Rheb’s activation level and mTORC1 phosphorylation of p70 S6K…………………………………………………………………55

Figure 3.12 Effects of Y35A and D65A mutations in Rheb on mTORC1 activity……………………...56

Figure 3.13 Schematic model of intrinsic and TSC2GAP-stimulated GTP hydrolysis by Rheb………..59

Figure 4.1) Manipulation of the GTPase cycle of Rheb and mTOR signaling through substitutions of Gly63…………………………………………………………………………………………………….71

Figure 4.2 Conservation of the G3-box glycine and structural basis for the functional properties of Rheb Gly63Ala/Val mutants…………………………………………………………………………………..73

Figure 4.3 Electron densities of the catalytic sites of wild-type Rheb and its Gly63 mutations………..74

Figure 4.4 Intrinsic nucleotide hydrolysis and exchange of Rap1A GTPase and its Gly60 mutations…77

Figure 5.1 Tethering Rheb to a bilayer membrane inhibits nucleotide exchange and activation……….87

Figure 5.2 Preparation of Rheb-nanodisc complex……………………………………………………...89

xii

Figure 5.3 Backbone 15N relaxation data for Rheb-GDP 1-181 in free (black) and nanodisc-bound (red) states……………………………………………………………………………………………………...91

Figure 5.4 Backbone 15N relaxation data for Rheb-GMPPNP 1-181 in free (black) and nanodisc-bound (red) states………………………………………………………………………………………………..92

Figure 5.5 Identification of the Rheb-membrane interface and its modulation by the bound nucleotide.93

Figure 5.6 Residues affected by PRE localized on the HADDOCK models…………………………….94

Figure 5.7 Cluster analysis of final HADDOCK solutions………………………………………………95

Figure 5.8 Subtle changes in surface electrostatics upon nucleotide exchange………………………….97

Figure 5.9 Formation of a Rheb-PDE δ complex is compatible with Rheb-nanodisc model 2, but not model 1…………………………………………………………………………………………………..98

Figure 6.1 Backbone and side chain assignments of free KRas4B in complex with GMPPNP…….…113

xiii

List of Appendices

Appendix A……………………………………………………………………………………………..116

Appendix B…………………………………………………………………………………………….118

xiv

List of Abbreviations

4E-BP (eIF)4E-binding proteins

Arf ADP ribosylation factor

BMRB Biological Magnetic Resonance Bank

CNS Crystallography & NMR System

Coot Crystallographic Object-Oriented Toolkit

DH-PH Dbl homology-Pleckstrin-homology

DOPC 1,2-dioleoyl-sn -glycero-3-phosphocholine

DOPS 1,2-dioleoyl-sn-glycero-3-phospho-L-serine

DTT Dithiothreitol

ER Endoplasmic reticulum

EDTA Ethylenediaminetetraacetic acid

FKBP FK506-binding protein

FRET Fluorescence resonance energy transfer

FTI Farnesyltransferase Inhibitors

GAP GTPase activating protein

GDI Guanine Dissociation Inhibitor

GEF Guanine Exchange Factor

GMPPNP 5'-Guanylyl imidodiphosphate

cat H2O Catalytic water

HADDOCK High Ambiguity Driven biomolecular DOCKing

HVR Hyper Variable Region

IPTG Isopropyl β-D-1-thiogalactopyranoside

LAM lymphangioleiomyomatosis

LC3 Light Chain 3

Mant 2'(3')-O-(N-Methylanthraniloyl)

MCF-7 Michigan Cancer Foundation-7

MD Molecular Dynamics

MSP Membrane Scaffold Protein

mTORC1 mammalina Target of Rapamycin Complex 1

NBS Nucleotide Binding Site

NMR Nuclear Magnetic Resonance

NOE Nuclear Overhauser effect

PAM Associated with Myc

PDB Protein Data Bank

PDE phosphodiesterase

PE-DTPA(Ga 3+ ) 1,2-distearoyl-sn -glycero-3-phosphoethanolamine-N- diethylenetriaminepentaacetic acid (gadolinium salt)

PE-MCC 1,2-dioleoyl-sn -glycero-3-phosphoethanolamine-N-[4-(p- maleimidomethyl)cyclohexane-carboxamide]

PI3K Phosphatidylinositol 3-kinases

xvi

POPC 1-palmitoyl-2-oleoyl-sn -glycero-3-phosphocholine

PRAK p38-regulated/activated kinase

PRE Paramagnetic Relaxation Enhancement

Ran Ras-related nuclear

Ras Rat sarcoma

RBD Ras Binding Domain

Rheb Ras homolog enriched in brain

RhebL1 Ras homolog enriched in brain-like 1

Rho Ras homolog

RMSD Root Mean Square Deviation

S6K S6 Kinase

PAGE polyacrylamide gel electrophoresis

SDS sodium dodecyl sulfate

SOS Son of Sevenless

SRP signal recognition particle

TCEP tris(2-carboxyethyl)phosphine

TRIS tris(hydroxymethyl)aminomethane

UV Ultraviolet

WT Wild Type

xvii 1

CHAPTER 1

Introduction and Thesis Overview

1.1 General introduction to small GTPase proteins: The Rat sarcoma (Ras) superfamily of small GTPases are ~21 kDa globular protein-nucleotide complexes that control diverse signaling networks affecting cell growth, proliferation and differentiation, intracellular trafficking, cellular morphology, mobility, chemotaxis and apoptosis (Wittinghofer and Vetter 2011; Cherfils and Zeghouf 2013). Deregulation of GTPase signaling has immense implications on human health and can result in life-threatening diseases such as cancer. The superfamily was originally divided into five subfamilies, based on cellular function and sequence similarities (Wennerberg et al. 2005). These include Ras, Ras homolog (Rho), Ras related in brain (Rab), ADP ribosylation factor (Arf), and Ras-related nuclear (Ran) subfamilies. A more recent phylogenetic analysis has identified additional small GTPases that are yet to be classified into a specific subfamily (Rojas et al. 2012). The superfamily members contain five conserved sequence motifs, identified as the G1 to G5 boxes, which are required for interactions with nucleotides, effectors, and regulators (Figure 1.1). The G1, G2 and G3 boxes reside within functionally important and dynamic regions of GTPase proteins. The G1 box forms a loop/helix structural motif termed the P-loop, which is responsible for phosphate binding, whereas G2 and G3 are found within switch I and II, respectively, which are responsible for interaction with effectors and regulators of GTPases (Wittinghofer and Vetter 2011). Additionally, most members of the superfamily are targeted to intracellular membrane compartments through post-translational modifications whereby hydrophobic moieties are covalently attached to either the N- or C-terminus (Ahearn et al. 2012). A universal feature of the small GTPases is their ability to bind guanine nucleotides, specifically, guanosine di-(GDP) and tri-(GTP) phosphates, and cycle between GDP- and GTP-bound states. When bound to GTP they adopt a conformation that allows recognition and activation of downstream effector proteins. Hydrolysis of the γ-phosphate of GTP converts the GTPases to a GDP-bound inactive state, where they can no longer functionally interact with their effectors. The GTPase can then be reactivated by exchange of GDP for a new molecule of GTP, driven in part by the higher concentrations of GTP in eukaryotic cells (Traut 1994). The rate of this cycle depends in part on intrinsic properties of each GTPase, including their relative affinities for GDP and GTP and the efficiency of their nucleotide hydrolyzing machinery (Li and Zhang 2004). However the rate of the GTPase cycle is tightly linked to internal and external stimuli of the cell that are transmitted to the small GTPase by the action of upstream regulators, such as guanine exchange factors (GEF), which accelerate the intrinsic nucleotide exchange, and GTPase activating proteins (GAP), which accelerate the intrinsic GTP hydrolysis.

Figure 1.1 Sequence alignment of Rheb with other small GTPase proteins investigated in this thesis. Alignment of Rheb is shown with its close homologues K- and H-Ras isoforms and Rap1A, as well as the more distant homologue RhoA, from the Rho subfamily. Conserved amino acids are highlighted in green, those conserved in more than half of the alignment are shown in yellow, and semi- conserved amino acids in cyan. G1 to G5 boxes are indicated with solid bars and the P-loop, switch I and switch II regions are indicated in red. The hyper variable region (HVR) is indicated with a green bar, and the CaaX (C is Cys, a is aliphatic aa, X is any aa) box is highlighted with a rectangle.

The Ras superfamily is an excellent example of achieving functional diversity within a similar structural framework. Therefore, detailed understanding of the structural and dynamic properties of these switch proteins is critical to the understanding of their function in cells and can aid in therapeutic intervention for patients suffering from diseases associated with aberrant GTPase function.

1.2 Introduction to Rheb:

The main focus of this thesis is an important disease-related member of the Ras subfamily called Ras homolog enriched in brain (Rheb). Rheb was originally discovered in a screen for genes induced in hippocampal granule cells of Rat by seizures and by N-methyl-aspartate (NMDA)-dependent synaptic activity (Yamagata et al. 1994). Sequence analysis identified Rheb as a member of the Ras superfamily of small GTP binding proteins with the highest similarity to Rap1A and H-Ras GTPases. Subsequently, the gene encoding human Rheb was identified on chromosome 7 (Gromov et al. 1995; Mizuki et al. 1996), and a second Rheb-like 1 (RhebL1) gene was later identified on chromosome 12 (Patel et al. 2003). Rheb is expressed ubiquitously with higher expression levels in skeletal and cardiac muscle whereas RhebL1 is primarily expressed in the brain. Rheb homologs have been found in yeast, slime mold, fungi, fruit fly and zebra fish (Reuther and Der 2000). Rheb was shown to bind to and hydrolyze GTP with a rate slower than that of H-Ras (Marshall et al. 2009). Like its close homolog H-Ras, Rheb was shown to be farnesylated in cells (Clark et al. 1997) suggesting membrane association. Rheb was originally described to interact with Raf-1 kinase and antagonize Ras signaling and transformation (Clark et al. 1997; Yee and Worley 1997). Later it was shown that Rheb affinity for Raf-1 is 1000-fold lower than that of Ras (Karassek et al.), thus Raf might not be a physiological effector. Loss of the Rheb ortholog Rhb1 in fission yeast mimics the nitrogen-starvation induced phenotype and growth arrest (Mach et al. 2000), and Drosophila Rheb is required for cell growth and cell cycle progression (Patel et al. 2003). In 2003, several independent groups showed that Rheb is involved in the target of rapamycin complex 1 (TORC1) signaling pathway (Figure 1.2) in Drosophila melanogaster and mammalian cells (Garami et al. 2003; Saucedo et al. 2003; Stocker et al. 2003). Rheb was found to promote cellular growth in a mTOR- (mammalian TOR) dependent manner in response to insulin and nutrient availability. This activity required phosphorylation and activation of the TOR substrate p70 S6 Kinase, driving protein translation and cell growth (Ruvinsky and Meyuhas 2006). Another major Rheb- regulated target of mTORC1 was identified as eukaryotic initiation factor (eIF)4E-binding proteins (4E- BPs), involved in cap-dependent translation (Beretta et al. 1996; Bommareddy et al. 2009). The molecular mechanism by which Rheb activates mTORC1 is not yet fully understood and is still a matter of debate (Bai et al. 2007; Sun et al. 2008; Wang et al. 2008; Sato et al. 2009). It is however clear that Rheb activates mTORC1 kinase activity in a GTP-dependent manner (Sancak et al. 2007), but paradoxically both GDP- and GTP-bound states of Rheb were reported to interact with mTORC1 (Long et al. 2005). Biochemical and genetic studies placed Rheb downstream of the tumor suppressors tuberous sclerosis complex 1 and 2 (TSC 1/2) in both Drosophila and mammalian cells (Garami et al.

2003; Inoki et al. 2003; Tee et al. 2003; Zhang et al. 2003). Rheb was shown to be a direct target of TSC2 GTPase activating domain (GAP) activity both in vitro and in vivo (Inoki et al. 2003). The isolated TSC2 GAP domain was later shown to accelerate the slow intrinsic GTP hydrolysis rate of

Figure 1.2 Schematic illustration of Rheb in the mTORC1 signaling pathway. Rheb-GTP stimulates the kinase activity of mammalian target of rapamycin complex 1 (mTORC1) that phosphorylates its substrates including 4E-BP1 and p70S6K1 promoting protein biosynthesis and cell cycle progression. Rheb intrinsic GTP hydrolysis rate is accelerated via the GAP activity of TSC1/2, which is regulated in response to the availability of growth factors, and energy stress.

Rheb in vitro (Scrima et al. 2008; Marshall et al. 2009). The GAP function of TSC 1/2 is tightly regulated by upstream signaling via phosphorylation events in response to internal and external

6 nutritional and energy status of the cell (Huang and Manning 2008). Hence, Rheb acts as a switch to regulate cellular growth and cell cycle progression in response to nutrient availability, energy status and hypoxia, via bridging TSC 1/2 to mTORC1 activity. Interestingly, it was recently recognized that under certain stress conditions such as UV irradiation and cellular intoxication, Rheb can switch from a pro- growth signaling molecule to a pro-apoptotic one using both canonical (mTORC1-dependent) and non- canonical (mTORC1-independent) signaling networks (Ehrkamp et al. 2013). Some of the interesting non-canonical and membrane-dependent interactions reported for Rheb include; interaction with mitochondrial autophagic receptor Nix and the autophagosomal protein LC3-II in promoting mitophagy (Melser et al. 2013); regulation of FKBP38 interaction with anti-apoptotic proteins Bcl-2 and Bcl-XL (Ma et al. 2010); an inhibitory interaction between Rheb and Bnip3 in hypoxia-mediated inhibition of mTORC1 signaling (Li et al. 2007); and sequestration of farnesylated Rheb by cAMP hydrolyzing enzyme phosphodiesterase 4D (PDE4D) (Kim et al. 2010), and the proposed solubilizing factor for farnesylated Ras-subfamily proteins, PDE δ (Ismail et al. 2011).

Currently, there is no consensus on the identity of the protein, if any, that plays the role of a Rheb GEF. There is debate whether the translationally controlled tumor protein (TCTP) is the physiological GEF for this small GTPase (Hsu et al. 2007; Rehmann et al. 2008; Wang et al. 2008; Dong et al. 2009). Interestingly, recent studies have proposed other potential GEFs for Rheb including; the E3 ubiquitin ligase Protein Associated with Myc (PAM) (Maeurer et al. 2009) and deacetylated soluble αβ -tubulin (Lee et al. 2013). In both cases, it remains to be determined whether the newly identified candidates possess direct GEF-function toward Rheb in vitro using purified recombinant proteins.

1.3 Biophysical and biochemical properties of Rheb: Rheb is a 184-amino acid GTPase, containing 6 conserved sequence elements that are hallmarks of the Ras subfamily. These include G1 to G5 boxes, as descried above plus a C-terminal CaaX box (where C is cysteine, a is aliphatic amino acid, and X is any amino acid). Rheb is post-translationally modified via farnesylation, cleavage of the aaX tripeptide and carboxymethylation of the C-terminal cysteine by the concerted actions of farnesyltransferase, Ras converting enzyme 1 (Rce1) and isoprenylcysteine carboxyl methyltransferase (Icmt ) enzymes, respectively (Seabra 1998). These modifications target Rheb to intracellular membrane compartments such as the ER, Golgi apparatus, and lysosome

(Takahashi et al. 2005; Buerger et al. 2006). Inhibition of Rheb farnesylation, via mutation of the CaaX box Cys residue or treatment with farnesyltransferase inhibitors (FTIs), drastically reduces Rheb’s ability to stimulate mTORC1 activity in cells (Buerger et al. 2006; Finlay et al. 2007). Although Rheb shares 34% sequence identity with H-Ras and harbors many conserved sequence elements through G1- G5 boxes, two key sequence variations were observed upon its discovery. In Ras, codons 12 and 13 encode two conserved Glycine residues, mutations of which result in oncogenic Ras molecules that are constitutively GTP bound (Figure 1.3) (Prior et al. 2012). The structural basis for the Gly12 and Gly13 mutation-mediated oncogenic transformation in Ras is well characterized (Krengel et al. 1990). For example, H-Ras G12V, a high frequency mutation in various tumor types, sterically blocks the canonical catalytic residue, Gln61 (H-Ras numbering), from the nucleotide binding site, impairing GTP hydrolysis. Gln61 is positioned at the N-terminus of switch II and it is shown to stimulate the GTP hydrolysis reaction by interacting with a water molecule positioned in-line with the γ-phosphate of GTP cat (aka , the catalytic water, H2O ). In Rheb the corresponding codons (15 and 16) encode Arginine and Serine residues, respectively. However, the slow GTP hydrolysis by Rheb is not due to the presence of these residues at codons 15 and 16. In cells over-expressing Rheb, R15G and S16G mutations did not affect its activation state (Im et al. 2002), suggesting that Arg15 and Ser16 did not affect the Rheb-GTP levels, where the function of TSC2GAP is limiting.

Compared to Ras, Rheb exists in a high activation state in cells (Im et al. 2002). Initial characterization of Rheb’s intrinsic GTP hydrolysis rate suggested Rheb possesses a slower GTP hydrolysis rate than Ras (Aspuria and Tamanoi 2004). This was confirmed by the use of the novel real-time NMR-based GTPase assay that accurately measured the nucleotide hydrolysis rate of Rheb to be 10 times slower than that of H-Ras (Marshall et al. 2009). A potential structural explanation for this low GTPase activity was found eleven years after Rheb was first discovered. Crystal structures of Rheb were solved in both an inactive GDP-bound state (PDB: 1XTQ) and active states bound to GTP (PDB: 1XTS) and GMPPNP-bound, a non-hydrolyzable analog of GTP (PDB: 1XTR) (Yu et al. 2005). The overall fold is typical of Ras subfamily members, namely a six-stranded β-sheet and five α-helices forming a globular 20 kDa protein (Figure 1.3). However, two important structural differences were observed in the switch I and II regions of the GMPPNP-bound state compared to that observed in H-Ras. In Rheb, switch I forms a lid that covers the phosphate groups of the nucleotide and simultaneously creates a pore to the

cat bulk solvent, which is occupied by the H 2O . This lid is formed by the H-bonding interaction between the hydroxyl group of the conserved Tyr35 in switch I of Rheb and the γ-phosphate (Figure 1.3).

Figure 1.3 Overall structure of active Rheb and comparison with that of H-Ras. Crystal structure of Rheb bound to GMPPNP (PDB: 1XTR) is shown in green. The nucleotide is shown with stick. The inset

9 shows the positions of key residues within the nucleotide binding site of Rheb and H-Ras (cyan) (PDB: cat 5P21). The H 2O is presented with spheres. The backbones of Tyr35, Gln61 and Gln64 are not shown.

In H-Ras, the equivalent tyrosine (Tyr32) is in an open conformation that does not interact with the γ- phosphate, thus exposing the nucleotide phosphate groups to the solvent (Figure 1.3). This conformational divergence could aid in differential recognition of interactors by Ras versus Rheb. For example, the open conformation of Ras Tyr32 allows p120 RasGAP to insert an Arginine ( aka the Arginine finger) into the nucleotide binding site to stimulate GTP hydrolysis (Scheffzek et al. 1997). The second difference involves the conformation of switch II. Whereas the H-Ras switch II harbors a 10-residue curved alpha helix and is relatively detached from the globular G-domain, in Rheb switch II is mainly unstructured and forms tight contacts with the G-domain (Figure 1.3). The structural consequence of this conformation is a net removal of the switch II residues from the nucleotide binding site of Rheb compared to that of H-Ras. The point of divergence begin after the conserved DxxG motif 2+ cat at the N-terminus of switch II, a conserved curved loop structure responsible for Mg and H 2O coordination. This unique switch II configuration flips the Gln64 (corresponding to H-Ras Gln61) side chain away from the nucleotide binding pocket and traps it within a hydrophobic pocket underneath switch II formed by Leu12, Phe70, Pro71, Tyr74, and Ile99. A salt bridge between Glu66 in switch II and Lys91 of the α3 helix stabilizes switch II away from the nucleotide binding site. Further stabilization is achieved by polar interaction between the Lys102 of α3 helix and the backbone carbonyl of Glu66. This switch II conformation of Rheb provides additional space that accommodates the Arg15 side chain that is not available in the H-Ras structure. Hence, the structural analysis indicated that although Rheb possesses a glutamine residue corresponding to H-Ras Gln61, this residue does not participate in GTP hydrolysis. This was later confirmed via NMR-based GTPase assay, demonstrating that the Q64L mutation in Rheb has minor effects on both intrinsic and TSC2GAP-mediated GTP hydrolysis (Marshall et al. 2009).

1.4 Pathogenesis of Rheb: Rheb is involved in the pathogenesis of the tuberous sclerosis disease, which is an autosomal dominant genetic disorder with a prevalence of ~1 in 8,000 (Napolioni and Curatolo 2008). Mutations in tsc 1 and tsc 2 genes that disrupt TSC1/2 GAP function in tuberous sclerosis and lymphangioleiomyomatosis (LAM) diseases result in high levels of Rheb-GTP that constitutively activate mTORC1. Tuberous sclerosis disease is characterized by multiple benign tumor formation in multiple organ systems. The

symptoms include autism, developmental delay, hematuria (blood in urine), cardiac arrhythmia and skin disfigurement.

Additionally, numerous studies have suggested Rheb to be directly involved in tumorgenesis. Constitutively activated Rheb mutants have been shown to induce oncogenic transformation in cell culture (Jiang and Vogt 2008) and it has been well established that transcriptional- and translational- mediated elevation of Rheb expression contributes to its hyperactivation (Norsted Gregory et al. 2010; Cao et al. 2013). Elevated levels of Rheb result in high activation levels due to limiting GAP function of TSC1/2. The highly GTP-loaded Rheb constitutively signals to mTORC1 to promote cellular growth and cell cycle progression. The elevated expression of Rheb has been found in fibroadenoma (Eom et al. 2008), shown to be critical and sufficient for skin epithelial carcinogenesis (Lu et al. 2010), and to cause prostate cancer (Nardella et al. 2008; Kobayashi et al. 2010) and lymphomagenesis (Mavrakis et al. 2008). An increase in Rheb activation state has also been demonstrated to drive 17-beta estradiol (E(2))- dependent proliferation of the MCF-7 breast cell line (Yu and Henske 2006).

Where mutations of Gly12, Gly13 and Gln61 result in oncogenic transformation of Ras, Rheb has not been associated with any particular oncogenic mutation. However, recent high-thoughput cancer genome sequencing analyses of tumor samples from patients with various tumour types including lung, breast, urinary tract, endometrium, colon, kidney and stomach has revealed multiple missense mutations throughout the Rheb gene (COSMIC database). The mutations map to the important structural regions in Rheb including the P-loop, switch I and II. Residues with the highest frequency mutations are Tyr35 (to Asn) (Lawrence et al. 2014) and Glu139 (to Lys, Glu, and Asp). Tyr35 is in the middle of switch I (G2 box) and, as noted above, covers the nucleotide-binding pocket from the bulk solvent. Glu139 is at the C-terminus of α5 near a site of post-translational modification (Ser130 at the N-terminus of α5). Rheb nucleotide loading is inhibited by phosphorylation of Ser130 by p38-regulated/activated kinase (PRAK) (Zheng et al. 2011). The functional properties of these Rheb mutants are yet to be characterized to determine weather they drive tumorgenesis.

To target deregulated mTORC1 signaling, several therapeutic strategies are being investigated. First is the direct inhibition of mTORC1 via small molecules. Rapamycin and its synthetic analogs (Rapalogs), also known as the first-generation mTOR inhibitors, inhibit phosphorylation of a subset of mTORC1

substrates (Kang et al. 2013). Rapmycin directly blocks substrate recruitment by forming a complex with FK506-binding protein 12 (FKBP12) and binding to the FKBP12–rapamycin-binding (FRB) domain of mTORC1, close to the kinase catalytic site (Choi et al. 1996; Yang et al. 2013). Clinical trials have shown that rapalogs are cytostatic and tend to be stabilizers of diseases rather than causing disease regression (Wander et al. 2011). In addition, rapalog-mediated inhibition of mTORC1 fails to suppress a negative feedback loop that results in phosphorylation-mediated activation of AKT (Wan et al. 2007). More recently, ATP-binding site inhibitors, also known as second-generation mTOR inhibitors, were developed, which directly compete with ATP binding and inhibit the kinase activity of both mTORC1 and mTORC2 for all substrates (Zaytseva et al. 2012). However, they display cytotoxic effects and are only partially effective in K-Ras driven tumors (Zaytseva et al. 2012). An alternate strategy involves prevention of Rheb association with bilayer membranes through the use of FTIs, which block farnesylation of Rheb and its ability to signal to mTORC1 in the cells, but suffer from non-specificity (Mavrakis et al. 2008). Interestingly, a cytotoxic biphenyl compound, 4,4 ′-biphenol was shown to interact with Rheb switch II region, block mTORC1 signaling and induce cell death (Schopel et al. 2013). However the affinity of the compound was low (Kd in millimolar range), indicating the requirement for structure-guided optimization of the compound.

1.5 Thesis Overview and Rationale: From the above discussion it is clear that two key physiological properties of Rheb are directly linked to human diseases: (a) its GTP hydrolysis rate and (b) its membrane association. The studies presented in this thesis are basic research aimed at gaining an atomic-resolution understanding of Rheb’s catalytic mechanism and its interaction with phospholipid bilayers. The information gained, could aid in the i) mechanistic understanding of disease-associated mutations in Rheb, ii) development of small molecules that specifically target the disease-associated Rheb mutant states, iii) development of novel probes to study the biology of the mTORC1 pathway, and iv) characterization of the Rheb-membrane interface and its implication for Rheb signaling.

The catalytic mechanism of Rheb GTP hydrolysis is described in Chapters 2, 3 and 4, representing three published manuscripts. When I commenced my Ph. D. studies, Dr. Marshall published a novel method for quantitatively measuring the rate of the GTPase cycle using native nucleotides via real-time NMR- based GTPase assay (Marshall et al. 2009). Hence, in my first study (chapter 2), I examined the effect of

widely used fluorescent nucleotide analogs, 2'(3')-O-(N-Methylanthraniloyl)-(mant)-GDP and mant- GTP in modulating the kinetics of the GTPase cycle. Chapter 3 describes a structure-function study of the non-canonical GTP catalytic mechanism of Rheb that was initially revealed by studies of Rheb hydrolysis of the mant-nucleotide. Chapter 4 describes the structure-guided development of novel gain- of-function and loss-of-function mutants of Rheb using the knowledge gained from its nucleotide catalytic mechanism.

The Rheb-membrane interaction is described in Chapter 5, representing our published work on a novel NMR-based method for studying GTPase-membrane interaction at atomic resolution using nanodisc- based lipid bilayers. These methods demonstrated effects of membrane binding on Rheb’s GTPase cycle, as well as determining the orientation of Rheb with respect to the lipid bilayer in both GDP- and GTP- bound states.

Chapter 6 includes concluding remarks and future directions, including (i) potential applications of the gain- and loss-of-function Rheb mutants generated via structure-guided design, and (ii) the feasibility of applying nanodisc-GTPase methodology to other disease-related GTPases, such as K-Ras, and its complexes with different Ras Binding Domains (RBDs).

1.6 Attributions: The cell-based assays of nucleotide binding (Chapter 3) and mTORC1 phosphorylation (Chapter 3 and 4) were performed by our collaborators, Vanessa Di Palma and Jason Ho, from Dr. Vuk Stambolic’s Lab (Department of Medical Biophysics, Campbell Family Cancer Research Institute, Ontario Cancer Institute, Princess Margaret Cancer Centre, University Health Network, University of Toronto).

1.7 References:

Ahearn IM, Haigis K, Bar-Sagi D, Philips MR. 2012. Regulating the regulator: post-translational modification of RAS. Nat Rev Mol Cell Biol 13 : 39-51. Aspuria PJ, Tamanoi F. 2004. The Rheb family of GTP-binding proteins. Cell Signal 16 : 1105-1112. Bai X, Ma D, Liu A, Shen X, Wang QJ, Liu Y, Jiang Y. 2007. Rheb activates mTOR by antagonizing its endogenous inhibitor, FKBP38. Science 318 : 977-980. Beretta L, Gingras AC, Svitkin YV, Hall MN, Sonenberg N. 1996. Rapamycin blocks the phosphorylation of 4E-BP1 and inhibits cap-dependent initiation of translation. Embo J 15 : 658- 664.

Bommareddy A, Hahm ER, Xiao D, Powolny AA, Fisher AL, Jiang Y, Singh SV. 2009. Atg5 regulates phenethyl isothiocyanate-induced autophagic and apoptotic cell death in human prostate cancer cells. Cancer Res 69 : 3704-3712. Buerger C, DeVries B, Stambolic V. 2006. Localization of Rheb to the endomembrane is critical for its signaling function. Biochem Biophys Res Commun 344 : 869-880. Cao M, Tan X, Jin W, Zheng H, Xu W, Rui Y, Li L, Cao J, Wu X, Cui G et al. 2013. Upregulation of Ras homolog enriched in the brain (Rheb) in lipopolysaccharide-induced neuroinflammation. Neurochem Int 62 : 406-417. Cherfils J, Zeghouf M. 2013. Regulation of small GTPases by GEFs, GAPs, and GDIs. Physiol Rev 93 : 269-309. Choi J, Chen J, Schreiber SL, Clardy J. 1996. Structure of the FKBP12-rapamycin complex interacting with the binding domain of human FRAP. Science 273 : 239-242. Clark GJ, Kinch MS, Rogers-Graham K, Sebti SM, Hamilton AD, Der CJ. 1997. The Ras-related protein Rheb is farnesylated and antagonizes Ras signaling and transformation. J Biol Chem 272 : 10608-10615. Dong X, Yang B, Li Y, Zhong C, Ding J. 2009. Molecular basis of the acceleration of the GDP-GTP exchange of human ras homolog enriched in brain by human translationally controlled tumor protein. J Biol Chem 284 : 23754-23764. Ehrkamp A, Herrmann C, Stoll R, Heumann R. 2013. Ras and rheb signaling in survival and cell death. Cancers (Basel) 5: 639-661. Eom M, Han A, Yi SY, Shin JJ, Cui Y, Park KH. 2008. RHEB expression in fibroadenomas of the breast. Pathol Int 58 : 226-232. Finlay GA, Malhowski AJ, Liu Y, Fanburg BL, Kwiatkowski DJ, Toksoz D. 2007. Selective inhibition of growth of tuberous sclerosis complex 2 null cells by atorvastatin is associated with impaired Rheb and Rho GTPase function and reduced mTOR/S6 kinase activity. Cancer Res 67 : 9878- 9886. Garami A, Zwartkruis FJ, Nobukuni T, Joaquin M, Roccio M, Stocker H, Kozma SC, Hafen E, Bos JL, Thomas G. 2003. Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2. Mol Cell 11 : 1457-1466. Gromov PS, Madsen P, Tomerup N, Celis JE. 1995. A novel approach for expression cloning of small GTPases: identification, tissue distribution and chromosome mapping of the human homolog of rheb. FEBS Lett 377 : 221-226. Hsu YC, Chern JJ, Cai Y, Liu M, Choi KW. 2007. Drosophila TCTP is essential for growth and proliferation through regulation of dRheb GTPase. Nature 445 : 785-788. Huang J, Manning BD. 2008. The TSC1-TSC2 complex: a molecular switchboard controlling cell growth. Biochem J 412 : 179-190. Im E, von Lintig FC, Chen J, Zhuang S, Qui W, Chowdhury S, Worley PF, Boss GR, Pilz RB. 2002. Rheb is in a high activation state and inhibits B-Raf kinase in mammalian cells. Oncogene 21 : 6356-6365. Inoki K, Li Y, Xu T, Guan KL. 2003. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev 17 : 1829-1834. Ismail SA, Chen YX, Rusinova A, Chandra A, Bierbaum M, Gremer L, Triola G, Waldmann H, Bastiaens PI, Wittinghofer A. 2011. Arl2-GTP and Arl3-GTP regulate a GDI-like transport system for farnesylated cargo. Nat Chem Biol 7: 942-949. Jiang H, Vogt PK. 2008. Constitutively active Rheb induces oncogenic transformation. Oncogene 27 : 5729-5740.

Kang SA, Pacold ME, Cervantes CL, Lim D, Lou HJ, Ottina K, Gray NS, Turk BE, Yaffe MB, Sabatini DM. 2013. mTORC1 phosphorylation sites encode their sensitivity to starvation and rapamycin. Science 341 : 1236566. Karassek S, Berghaus C, Schwarten M, Goemans CG, Ohse N, Kock G, Jockers K, Neumann S, Gottfried S, Herrmann C et al. Ras homolog enriched in brain (Rheb) enhances apoptotic signaling. J Biol Chem 285 : 33979-33991. Kim HW, Ha SH, Lee MN, Huston E, Kim DH, Jang SK, Suh PG, Houslay MD, Ryu SH. 2010. Cyclic AMP controls mTOR through regulation of the dynamic interaction between Rheb and phosphodiesterase 4D. Mol Cell Biol 30 : 5406-5420. Kobayashi T, Shimizu Y, Terada N, Yamasaki T, Nakamura E, Toda Y, Nishiyama H, Kamoto T, Ogawa O, Inoue T. 2010. Regulation of androgen receptor transactivity and mTOR-S6 kinase pathway by Rheb in prostate cancer cell proliferation. Prostate 70 : 866-874. Krengel U, Schlichting I, Scherer A, Schumann R, Frech M, John J, Kabsch W, Pai EF, Wittinghofer A. 1990. Three-dimensional structures of H-ras p21 mutants: molecular basis for their inability to function as signal switch molecules. Cell 62 : 539-548. Lawrence MS, Stojanov P, Mermel CH, Robinson JT, Garraway LA, Golub TR, Meyerson M, Gabriel SB, Lander ES, Getz G. 2014. Discovery and saturation analysis of cancer genes across 21 tumour types. Nature . Lee MN, Koh A, Park D, Jang JH, Kwak D, Jeon H, Kim J, Choi EJ, Jeong H, Suh PG et al. 2013. Deacetylated alphabeta-tubulin acts as a positive regulator of Rheb GTPase through increasing its GTP-loading. Cell Signal 25 : 539-551. Li G, Zhang XC. 2004. GTP hydrolysis mechanism of Ras-like GTPases. J Mol Biol 340 : 921-932. Li Y, Wang Y, Kim E, Beemiller P, Wang CY, Swanson J, You M, Guan KL. 2007. Bnip3 mediates the hypoxia-induced inhibition on mammalian target of rapamycin by interacting with Rheb. J Biol Chem 282 : 35803-35813. Long X, Lin Y, Ortiz-Vega S, Yonezawa K, Avruch J. 2005. Rheb binds and regulates the mTOR kinase. Curr Biol 15 : 702-713. Lu ZH, Shvartsman MB, Lee AY, Shao JM, Murray MM, Kladney RD, Fan D, Krajewski S, Chiang GG, Mills GB et al. 2010. Mammalian target of rapamycin activator RHEB is frequently overexpressed in human carcinomas and is critical and sufficient for skin epithelial carcinogenesis. Cancer Res 70 : 3287-3298. Ma D, Bai X, Zou H, Lai Y, Jiang Y. 2010. Rheb GTPase controls apoptosis by regulating interaction of FKBP38 with Bcl-2 and Bcl-XL. J Biol Chem 285 : 8621-8627. Mach KE, Furge KA, Albright CF. 2000. Loss of Rhb1, a Rheb-related GTPase in fission yeast, causes growth arrest with a terminal phenotype similar to that caused by nitrogen starvation. Genetics 155 : 611-622. Maeurer C, Holland S, Pierre S, Potstada W, Scholich K. 2009. Sphingosine-1-phosphate induced mTOR-activation is mediated by the E3-ubiquitin ligase PAM. Cell Signal 21 : 293-300. Marshall CB, Ho J, Buerger C, Plevin MJ, Li GY, Li Z, Ikura M, Stambolic V. 2009. Characterization of the intrinsic and TSC2-GAP-regulated GTPase activity of Rheb by real-time NMR. Sci Signal 2: ra3. Mavrakis KJ, Zhu H, Silva RL, Mills JR, Teruya-Feldstein J, Lowe SW, Tam W, Pelletier J, Wendel HG. 2008. Tumorigenic activity and therapeutic inhibition of Rheb GTPase. Genes Dev 22 : 2178-2188.

Melser S, Chatelain EH, Lavie J, Mahfouf W, Jose C, Obre E, Goorden S, Priault M, Elgersma Y, Rezvani HR et al. 2013. Rheb regulates mitophagy induced by mitochondrial energetic status. Cell Metab 17 : 719-730. Mizuki N, Kimura M, Ohno S, Miyata S, Sato M, Ando H, Ishihara M, Goto K, Watanabe S, Yamazaki M et al. 1996. Isolation of cDNA and genomic clones of a human Ras-related GTP-binding protein gene and its chromosomal localization to the long arm of chromosome 7, 7q36. Genomics 34 : 114-118. Napolioni V, Curatolo P. 2008. Genetics and molecular biology of tuberous sclerosis complex. Curr Genomics 9: 475-487. Nardella C, Chen Z, Salmena L, Carracedo A, Alimonti A, Egia A, Carver B, Gerald W, Cordon-Cardo C, Pandolfi PP. 2008. Aberrant Rheb-mediated mTORC1 activation and Pten haploinsufficiency are cooperative oncogenic events. Genes Dev 22 : 2172-2177. Norsted Gregory E, Codeluppi S, Gregory JA, Steinauer J, Svensson CI. 2010. Mammalian target of rapamycin in spinal cord neurons mediates hypersensitivity induced by peripheral inflammation. Neuroscience 169 : 1392-1402. Patel PH, Thapar N, Guo L, Martinez M, Maris J, Gau CL, Lengyel JA, Tamanoi F. 2003. Drosophila Rheb GTPase is required for cell cycle progression and cell growth. J Cell Sci 116 : 3601-3610. Prior IA, Lewis PD, Mattos C. 2012. A comprehensive survey of Ras mutations in cancer. Cancer Res 72 : 2457-2467. Rehmann H, Bruning M, Berghaus C, Schwarten M, Kohler K, Stocker H, Stoll R, Zwartkruis FJ, Wittinghofer A. 2008. Biochemical characterisation of TCTP questions its function as a guanine nucleotide exchange factor for Rheb. FEBS Lett 582 : 3005-3010. Reuther GW, Der CJ. 2000. The Ras branch of small GTPases: Ras family members don't fall far from the tree. Curr Opin Cell Biol 12 : 157-165. Rojas AM, Fuentes G, Rausell A, Valencia A. 2012. The Ras protein superfamily: evolutionary tree and role of conserved amino acids. J Cell Biol 196 : 189-201. Ruvinsky I, Meyuhas O. 2006. Ribosomal protein S6 phosphorylation: from protein synthesis to cell size. Trends Biochem Sci 31 : 342-348. Sancak Y, Thoreen CC, Peterson TR, Lindquist RA, Kang SA, Spooner E, Carr SA, Sabatini DM. 2007. PRAS40 is an insulin-regulated inhibitor of the mTORC1 protein kinase. Mol Cell 25 : 903-915. Sato T, Nakashima A, Guo L, Tamanoi F. 2009. Specific activation of mTORC1 by Rheb G-protein in vitro involves enhanced recruitment of its substrate protein. J Biol Chem 284 : 12783-12791. Saucedo LJ, Gao X, Chiarelli DA, Li L, Pan D, Edgar BA. 2003. Rheb promotes cell growth as a component of the insulin/TOR signalling network. Nat Cell Biol 5: 566-571. Scheffzek K, Ahmadian MR, Kabsch W, Wiesmuller L, Lautwein A, Schmitz F, Wittinghofer A. 1997. The Ras-RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science 277 : 333-338. Schopel M, Jockers KF, Duppe PM, Autzen J, Potheraveedu VN, Ince S, Yip KT, Heumann R, Herrmann C, Scherkenbeck J et al. 2013. Bisphenol A Binds to Ras Proteins and Competes with Guanine Nucleotide Exchange: Implications for GTPase-Selective Antagonists. J Med Chem 56 : 9664-9672. Scrima A, Thomas C, Deaconescu D, Wittinghofer A. 2008. The Rap-RapGAP complex: GTP hydrolysis without catalytic glutamine and arginine residues. Embo J 27 : 1145-1153. Seabra MC. 1998. Membrane association and targeting of prenylated Ras-like GTPases. Cell Signal 10 : 167-172.

Stocker H, Radimerski T, Schindelholz B, Wittwer F, Belawat P, Daram P, Breuer S, Thomas G, Hafen E. 2003. Rheb is an essential regulator of S6K in controlling cell growth in Drosophila. Nat Cell Biol 5: 559-565. Sun Y, Fang Y, Yoon MS, Zhang C, Roccio M, Zwartkruis FJ, Armstrong M, Brown HA, Chen J. 2008. Phospholipase D1 is an effector of Rheb in the mTOR pathway. Proc Natl Acad Sci U S A 105 : 8286-8291. Takahashi K, Nakagawa M, Young SG, Yamanaka S. 2005. Differential membrane localization of ERas and Rheb, two Ras-related proteins involved in the phosphatidylinositol 3-kinase/mTOR pathway. J Biol Chem 280 : 32768-32774. Tee AR, Manning BD, Roux PP, Cantley LC, Blenis J. 2003. Tuberous sclerosis complex gene products, Tuberin and Hamartin, control mTOR signaling by acting as a GTPase-activating protein complex toward Rheb. Curr Biol 13 : 1259-1268. Traut TW. 1994. Physiological concentrations of purines and pyrimidines. Mol Cell Biochem 140 : 1-22. Wan X, Harkavy B, Shen N, Grohar P, Helman LJ. 2007. Rapamycin induces feedback activation of Akt signaling through an IGF-1R-dependent mechanism. Oncogene 26 : 1932-1940. Wander SA, Hennessy BT, Slingerland JM. 2011. Next-generation mTOR inhibitors in clinical oncology: how pathway complexity informs therapeutic strategy. J Clin Invest 121 : 1231-1241. Wang X, Fonseca BD, Tang H, Liu R, Elia A, Clemens MJ, Bommer UA, Proud CG. 2008. Re- evaluating the roles of proposed modulators of mammalian target of rapamycin complex 1 (mTORC1) signaling. J Biol Chem 283 : 30482-30492. Wennerberg K, Rossman KL, Der CJ. 2005. The Ras superfamily at a glance. J Cell Sci 118 : 843-846. Wittinghofer A, Vetter IR. 2011. Structure-function relationships of the G domain, a canonical switch motif. Annu Rev Biochem 80 : 943-971. Yamagata K, Sanders LK, Kaufmann WE, Yee W, Barnes CA, Nathans D, Worley PF. 1994. rheb, a growth factor- and synaptic activity-regulated gene, encodes a novel Ras-related protein. J Biol Chem 269 : 16333-16339. Yang H, Rudge DG, Koos JD, Vaidialingam B, Yang HJ, Pavletich NP. 2013. mTOR kinase structure, mechanism and regulation. Nature 497 : 217-223. Yee WM, Worley PF. 1997. Rheb interacts with Raf-1 kinase and may function to integrate growth factor- and protein kinase A-dependent signals. Mol Cell Biol 17 : 921-933. Yu J, Henske EP. 2006. Estrogen-induced activation of mammalian target of rapamycin is mediated via tuberin and the small GTPase Ras homologue enriched in brain. Cancer Res 66 : 9461-9466. Yu Y, Li S, Xu X, Li Y, Guan K, Arnold E, Ding J. 2005. Structural basis for the unique biological function of small GTPase RHEB. J Biol Chem 280 : 17093-17100. Zaytseva YY, Valentino JD, Gulhati P, Evers BM. 2012. mTOR inhibitors in cancer therapy. Cancer Lett 319 : 1-7. Zhang Y, Gao X, Saucedo LJ, Ru B, Edgar BA, Pan D. 2003. Rheb is a direct target of the tuberous sclerosis tumour suppressor proteins. Nat Cell Biol 5: 578-581. Zheng M, Wang YH, Wu XN, Wu SQ, Lu BJ, Dong MQ, Zhang H, Sun P, Lin SC, Guan KL et al. 2011. Inactivation of Rheb by PRAK-mediated phosphorylation is essential for energy- depletion-induced suppression of mTORC1. Nat Cell Biol 13 : 263-272.

CHAPTER 2

Real-time NMR assays on Rheb, Ras, and RhoA GTPase cycles

This chapter has been reformatted from the original publication: Mazhab-Jafari MT, Marshall CB, Smith M, Gasmi-Seabrook GM, Stambolic V, Rottapel R, Neel BG, Ikura M. Real-time NMR study of three small GTPases reveals that fluorescent 2'(3')-O-(N-methylanthraniloyl)-tagged nucleotides alter hydrolysis and exchange kinetics. J Biol Chem. 2010 Feb 19;285(8):5132-6.

A link to the published paper can be found at: http://www.jbc.org/content/285/8/5132.long

2.1 Abstract The Ras family of small GTPases control diverse signaling pathways through a conserved ‘switch’ mechanism, which is turned on by binding of GTP and turned off by GTP hydrolysis to GDP. Full understanding of GTPase ‘switch’ functions requires reliable, quantitative assays for nucleotide binding and hydrolysis. Fluorescently labeled guanine nucleotides, such as 2'(3')-O-(N-Methylanthraniloyl)- (mant)-substituted GTP and GDP analogs, have been widely used to investigate the molecular properties of small GTPases, including Ras and Rho. Using a recently-developed NMR method, we show that the kinetics of nucleotide hydrolysis and exchange by three small GTPases, alone and in the presence of their cognate GTPase activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), are affected by the presence of the fluorescent mant moiety. Intrinsic hydrolysis of mantGTP by Rheb is ~10-times faster than that of GTP, whereas it is 3.4-times slower with RhoA. On the other hand, the mant tag inhibits TSC2GAP-catalyzed GTP hydrolysis by Rheb but promotes p120 RasGAP-catalyzed GTP hydrolysis by HRas. GEF catalyzed nucleotide exchange for both HRas and RhoA were inhibited by mant-substituted nucleotides, and the degree of inhibition depends highly on the GTPase and whether the assay measures association of mantGTP with, or dissociation of mantGDP from the GTPase. These results indicate that the mant moiety has significant and unpredictable effects on GTPase reaction kinetics and underscore the importance of validating its use in each assay.

2.2 Introduction: The Ras superfamily of small GTPases plays vital roles in the integrated network of cellular signaling. They are “turned on” by binding to GTP and adopting a conformation that allows modulation of their downstream effectors. The proteins are then “turned off” by hydrolysis of the γ-phosphate of GTP and conversion to GDP (Figure 2.1a). The relative amounts of activated GTP-bound and inactive GDP- bound forms of GTPases are tightly regulated by GAPs, which catalyze nucleotide hydrolysis, and GEFs, which promote nucleotide exchange. Mutation or unregulated expression of the small GTPase proteins or their respective GAPs and GEFs can deregulate the GTPase cycle and lead to diseases such as cancer, neurodegeneration and mental disabilities (Ahmadian et al. 1996). A B

Figure 2.1 Effect of mant-substituted GTP on the intrinsic nucleotide hydrolysis of Ras, Rheb and RhoA. a) Nucleotide dependent conformational changes in GTPases. Nucleotide hydrolysis and exchange are correlated with conformational changes that are readily probed in two dimensional 1H-15 N HSQC spectra (PDB: 1XTS green, 1XTQ cyan). b) The chemical structure of mantGTP (ChemSketch (Spessard 1998)) is shown with the mant moiety at the 3’ position of the ribose ring. In solution, this moiety exhibits slow chemical exchange between the 2’ and 3’ positions of the ribose ring, producing an equilibrium ratio of 4:6 (Neal et al. 1990). c) The time course of intrinsic GTP (black) and mantGTP (green) hydrolysis by HRas, Rheb and RhoA probed via real time solution NMR spectroscopy. All the proteins were fully loaded with GTP or mantGTP before the start of the assay, as assessed by 1H-15 N HSQC. The rates are displayed with a histogram for each curve in inserted panel c. See appendix A for description of equations.

Although several methods for monitoring GTP hydrolysis and nucleotide exchange of small GTPase proteins have been developed (Ahmadian et al. 1996; Zhang and Zheng 1998; Albert and Gallwitz 2000; Futai et al. 2004), the assay most widely used to monitor kinetics employs the fluorescently labeled guanosine nucleotide analogs 2'(3')-O-(N-Methylanthraniloyl) - GTP/GDP (mantGTP / mantGDP), which are sensitive to the hydrophobic environment of proteins (Figure 2.1b). Because of high sensitivity and selectivity, mantGTP and mantGDP have been widely used in the field (Sondermann et al. 2004; Mou et al. 2005; Gureasko et al. 2008), however, the use of these nucleotide analogs is justified only if they report reaction kinetics and thermodynamics that are consistent with the natural ligands GTP and GDP. Previously, some inconsistencies between native GTP and mantGTP were observed in nucleotide hydrolysis assays of Ras and RhoA with their cognate GAPs (Moore et al. 1993; Eberth et al. 2005). However, these fluorescent probes have never been fully assessed due to the lack of appropriate methodology.

Recently we developed an NMR-based real-time assay to monitor the rate of GTP hydrolysis of Ras homolog enriched in brain (Rheb), enabling us to monitor GTPase reactions using native GTP and GDP (Figure 2.2b) (Marshall et al. 2009). We have demonstrated that the real-time NMR methodology can be successfully used to assay nucleotide exchange in RhoA (Gasmi-Seabrook et al. 2010). This methodology requires no chemical modification of the protein or the nucleotide, which can perturb the native structure of the protein (Figure 2.3), and has the ability to sense subtle changes in the rate of catalysis in real-time fashion. In this study we employed the NMR methodology to examine how the fluorescent adduct on mantGTP and mantGDP affects the kinetics of nucleotide hydrolysis and exchange of three small GTPases (HRas, Rheb and RhoA) alone and in the presence of their GAPs or GEFs.

2.3 Results: 2.3.1 The effect of mant on the intrinsic rate of nucleotide hydrolysis :Mant-substituted nucleotides have been used extensively to investigate many aspects of G-protein signaling, including the kinetics and mechanisms of RasGAP-mediated GTP hydrolysis (Moore et al. 1993; Ahmadian et al. 1997; Ahmadian et al. 2002; Phillips et al. 2003) and Cdc25 SOS -mediated nucleotide exchange (Sondermann et al. 2004; Gureasko et al. 2008). We first used NMR methodology to examine whether the addition of the mant

moiety alters the intrinsic hydrolysis rate of the GTPase domain of HRas (1-171) (Figure 2.1c) and found that it had little effect (Table 2.1), consistent with previous reports (Remmers et al. 1994). A

Figure 2.2 Comparison of the real-time NMR and fluorescence-substituted nucleotide-based GTPase assays. a) Fluorescence assay. Left panel: The emission spectra of Rheb-mantGTP (black) and Rheb-mantGDP (red) illustrating the decline in the fluorescence intensity accompanied by a red shift upon nucleotide hydrolysis. Right panel: The time-dependent intrinsic hydrolysis of mantGTP by Rheb monitored by fluorescence and fitted to a one-phase exponential decay. b) GTP hydrolysis by Rheb observed using NMR. Snapshots of 1H-15 N HSQC spectra collected over the time course of hydrolysis. The green box indicates the resonance specific to Gly13 of Rheb-GTP and the red box corresponds to that of Rheb-GDP.

Figure 2.3 Perturbation of the native structure of GTPase proteins by mant-tagged nucleotides. Selected regions of 1H-15 N HSQC spectra from HRas, Rheb, and RhoA bound to GTP (green) or mantGTP (red) are shown. Dashed boxes indicate resonances that exhibit chemical shift changes between GTP and mantGTP.

Table 2.1 The rates of intrinsic and GAP mediated nucleotide hydrolysis for three small GTPase proteins. All rates are reported in min -1.

1 2 1

1: A molar ratio of 1/2,500 HRasGAP to HRas was used. 2: A molar ratio of 1/2.2 TSC2GAP to Rheb was used.

Next, we analyzed GTP hydrolysis by Ras homolog enriched in brain (Rheb), which shares ~33% sequence identity with HRas. Surprisingly, the hydrolysis rate of mantGTP by Rheb (3.2×10 -3 min -1) was >10 times faster than that of native GTP when both reactions were monitored by NMR (Figure 2.1c). The rate constant obtained with mantGTP by NMR was identical to the value obtained using fluorescence spectroscopy (Figure 2.1c and 2.2a), demonstrating that the results are independent of the detection method. Using mantGTP to compare the activity of different GTPases, the intrinsic GTPase activity of Rheb would appear ~2 times lower than that of HRas, whereas it is actually ~32 times lower with native GTP, demonstrating that reliance on mantGTP would overlook this biologically important difference.

As a third case, the effect of the mant substitution on the GTPase activity of RhoA, which shares ~31% sequence identity with HRas and Rheb, was investigated and found to be opposite to that observed with Rheb (Figure 2.1c). The half-life of mantGTP bound to RhoA was ~155 min, 3.5 fold longer than that of native GTP (~45 min) (Table 2.1). Taken together, the results show that mant-substituted nucleotides can substantially alter the kinetics of nucleotide hydrolysis by small GTPases in a manner that could not have been predicted a priori .

2.3.2 The effect of mant on the rate of GAP-catalyzed nucleotide hydrolysis :Having established that mant can have a substantial effect on intrinsic GTPase reaction rates, we investigated the effect of

mantGTP on GAP-accelerated GTPase reactions. First, we used the GAP domain of the human p120- RasGAP. GTP was hydrolyzed with a rate of 2.6×10 -2 min -1 by HRas in the presence of a 1/2,500 molar ratio of RasGAP, whereas mantGTP was turned over ~5 times faster (Figure 2.4a and Table 2.1). Indeed, Moore et al. (Moore et al. 1993) previously noted that in the presence of RasGAP, p21 N-Ras hydrolyzes mantGTP more rapidly than native GTP, although the cleavage of the two nucleotides were not monitored by the same method. A

Figure 2.4 Effects of mant-substituted GTP on GAP-mediated nucleotide hydrolysis by HRas and Rheb. a) Time course of RasGAP-mediated nucleotide hydrolysis of HRas with GTP (black) and mantGTP (green). A molar ratio of 1:2,500 RasGAP to HRas was used. b) Time course of TSC2GAP- mediated hydrolysis of GTP (black) and mantGTP (green) bound to Rheb. A molar ratio of 1:2.2 TSC2GAP to Rheb was used for both assays. Rates extracted from each curve are displayed with an inserted histogram. Each experiment was performed in duplicate, and each curve represents a single representative experiment.

We then examined how mantGTP might affect GAP-mediated GTP hydrolysis of a second, unrelated GAP with a distinct mechanism of action, by studying Rheb and its well-characterized GAP, TSC2GAP

(Marshall et al. 2009). At a TSC2GAP to Rheb ratio of 1/2.2, GTP was hydrolyzed at a rate of 1.9×10 -2 min -1 whereas mantGTP was hydrolyzed ~2.5 fold more slowly (Figure 2.4b and Table 2.1). Comparing Rheb’s intrinsic rate of mantGTP hydrolysis to the GAP-catalyzed rate, TSC2GAP only produced a two- fold enhancement (Table 2.1). This contrasts sharply with the ~50 fold stimulation of Rheb GTPase activity by TSC2GAP when native GTP is used in the assay. Hence, exclusive use of mantGTP would lead to a gross underestimation of the GAP activity of TSC2 towards Rheb, underscoring the utility of the NMR-based methodology.

2.3.3 The effect of mant on the GEF-mediated nucleotide exchange :We examined how mant-substituted nucleotides affect the DH-PH PRG -mediated nucleotide exchange of RhoA (Figure 2.5a) using a procedure described elsewhere (Gasmi-Seabrook et al. 2010). In the NMR GEF assay with hydrolysable nucleotides, the readout (i.e, GDP- and GTP-specific protein cross-peaks) is determined by both exchange and intrinsic nucleotide hydrolysis. This is evident in the case of Ras and RhoA (Figure 2.5) which do not become 100% saturated with GTP or mantGTP. Hence the observed data was fit to an equation that considers exchange and hydrolysis (see appendix A), to derive the true exchange rate. We have shown that this derived rate agrees well with the rate determined for the non-hydrolysable GTP γS (Gasmi-Seabrook et al. 2010). In an assay of nucleotide association with RhoA, mantGTP exhibits a 30% lower exchange rate (4.9×10 -2 min -1) compared to native GTP (7.0×10 -2 min -1) (Table 2.2). In the literature, both the incorporation (Hutchinson and Eccleston 2000; Derewenda et al. 2004; Oleksy et al. 2004) and dissociation (Tan et al. 2002; Hemsath and Ahmadian 2005) of the mant-substituted nucleotide have been employed to study the function of RhoA and its interaction with GEFs. Hence, we also performed the ‘dissociation’ assay, initially loading RhoA with GDP or mantGDP, and monitoring exchange to GTP. In this reaction the rate of DH-PHPRG mediated nucleotide dissociation was approximately six times slower for mantGDP than for GDP (Figure 2.6a and 6c).

Previous studies have used mantGDP extensively to probe the Son of Sevenless (Cdc25 SOS ) catalyzed nucleotide exchange of Ras GTPase. Generally, Ras is preloaded with mantGDP and the decay in fluorescent intensity is monitored as this fluorescent nucleotide is displaced by unlabeled GTP (Margarit et al. 2003; Sondermann et al. 2004; Ford et al. 2005; Ford et al. 2006; Freedman et al. 2006; Gureasko et al. 2008). Here, using an NMR-based protocol similar to that described for RhoA, we show (Figure 2.5b and Table 2.2) that the rate of Cdc25 SOS -catalyzed mantGDP dissociation (7.2×10 -3 min -1) is

approximately 30% slower compared with that of native GDP. GEF assays using association of mant- tagged nucleotide with Ras have also been reported (Sacco et al. 2006), thus we used NMR to compare the Cdc25 SOS -catalyzed association of mantGTP and GTP. We found that the rate of association of mantGTP with HRas is ~3-fold slower than that of GTP (Figure 2.6b and 6d). A

Figure 2.5 Effects of mant-substituted nucleotides on GEF-mediated nucleotide exchange for HRas and RhoA. a) Time course of DH-PH PRG mediated nucleotide exchange of RhoA-GDP to GTP (black) and mantGTP (green). A molar ratio of 1/30,000 RhoA to DH-PH PRG was used for both assays. b) Time course of CDC25 SOS mediated nucleotide exchange of HRas-GDP (black) and HRas-mantGDP (green) to GTP with 1/30,000 HRas to CDC25 SOS molar ratio. Observed data (continuous lines) were fitted to an equation that considers both exchange and hydrolysis to extract the true exchange rate k ex in each experiment (see appendix A). Using this rate, exponential decay curves (dashed lines) were generated to approximate nucleotide exchange in the absence of hydrolysis. The corrected rates are displayed with an inserted histogram for each curve. Each experiment was performed in duplicate, and each curve represents a single representative experiment.

Table 2.2 The rates of GEF mediated nucleotide exchange for HRas and RhoA proteins. These exchange rates are reported in min -1 and determined by fitting data to an equation that considers both exchange and hydrolysis.

1: A ratio of 1/30,000 GEF to GTPase was used with 10 molar excess of nucleotide relative to GTPase.

A B

C D

Figure 2.6 Complementary assays of the effects of mant-substituted nucleotides on GEF-mediated nucleotide exchange for HRas and RhoA. a) The DH-PH PRG mediated nucleotide exchange of RhoA- GDP to GTP (black) and mantGDP to GTP (green). A molar ratio of 1/30,000 RhoA to DH-PH PRG was used for both assays. b) The CDC25 SOS -mediated nucleotide exchange of HRas-GDP to GTP (black) or mantGTP (green). A molar ratio of 1/30,000 HRas to CDC25 SOS was used for both assays. Observed data (circles and triangles) were fitted to an equation that considers both exchange and hydrolysis to extract the true exchange rate, which was used to plot exponential decay curves (gray and green lines) (see Appendix A). These rates are displayed with a histogram for each curve in panels c and d.

2.4 Discussion: Fluorescently-labeled guanosine nucleotides have been used extensively to study nucleotide hydrolysis and exchange of GTPases. However, covalent modification of the nucleotide with a bulky fluorophore raises concerns about how this reporter moiety may perturb enzymatic activity. The NMR methodology recently developed by our group (Marshall et al. 2009) does not require any chemical modification of GTP or GDP because it makes direct observation of protein resonances that depend on nucleotide- induced changes in chemical environment of protein (Figure 2.2b). In this study, we employed this NMR method to compare native versus mant-labeled nucleotides in the kinetics of intrinsic GTPase reactions, GAP-mediated nucleotide hydrolysis, and GEF-mediated nucleotide exchange reactions of three small GTPases, HRas, Rheb, and RhoA. Our results clearly demonstrate that mant-labeled nucleotides had substantial effects on the kinetics of these reactions and that these effects were remarkably different and unpredictable with each GTPase, GAP, and GEF.

2.4.1 Intrinsic hydrolysis of GTP and mantGTP :HRas exhibited a small decrease in the intrinsic hydrolysis rate of mantGTP versus native GTP, however Rheb and RhoA were affected more drastically by this fluorescent tag. Remarkably, mant had opposite effects on Rheb and RhoA; Rheb hydrolyzed mantGTP ~10 fold faster than native GTP, whereas RhoA hydrolyzed this analog 3 fold more slowly than GTP. Thus, the effects of mant are specific to the structure and catalytic mechanism of each GTPase rather than the inherent lability of the nucleotide. With all three GTPases, the mant-adduct perturbed proximal residues in the P-loop, switch I and G-5 box, but also induced long range perturbations in several regions including switch II (Figure 2.7). The catalytic Gln of RhoA is found in switch II, thus distortion of the structure of this loop could inhibit the hydrolysis of mantGTP. Interestingly, the analogous Gln in Rheb is in an orientation that does not contribute to catalysis (Yu et al. 2005; Marshall et al. 2009). To understand the rapid hydrolysis of mantGTP by Rheb, we asked

Figure 2.7 mantGTP-induced chemical shift perturbations in Ras, Rheb and RhoA. Residues exhibiting chemical shift perturbations induced by binding of mantGTP (versus GTP) are mapped onto

the structures of a) Ras (PDB:5P21) , b) Rheb (PDB:1XTS) and c) RhoA (PDB:1KMQ) . Magenta colored residues correspond to small chemical shift changes (i.e., peaks partially overlap) whereas residues exhibiting large chemical shift changes or severe line broadening are colored red. Unassigned residues including prolines are indicated with dark gray. The nucleotide is shown in spheres with carbon, oxygen, nitrogen, and phosphate shown in green, red, blue, and orange respectively. The magnesium ion is shown with cyan sphere. The predicted position of the mant-moiety, which exists in equilibrium between the 2’ and 3’ hydroxyl groups of the ribose ring, is indicated in each structure by a red circle. Switch I residues of all three GTPases are broadened beyond detection in both the GTP- and mantGTP-bound states, however chemical shift perturbations would be expected based on the structure of Ras in complex with mant-dGppNHp (PDB:1GNP) , in which the mant moiety faces the switch I region of Ras.

whether the mant-induced perturbation of switch II might favor a catalytically competent conformation of Gln64. However the Q64L mutation had no effect on hydrolysis of mantGTP, indicating that this reaction occurs through more complex mechanism (Figure 2.8).

Figure 2.8 Rheb Gln64 is not involved in the rapid hydrolysis of mantGTP. The time course of mantGTP hydrolysis by wild type Rheb (black) and RhebQ64L mutant (green) probed via real time solution NMR spectroscopy. The proteins were fully loaded with mantGTP before the start of the assay, as assessed by 1H-15 N HSQC.

2.4.2 GAP-catalyzed GTP hydrolysis by HRas and Rheb:p120-RasGAP-catalyzed hydrolysis of mantGTP by HRas was ~5 fold faster than that of native GTP whereas TSC2GAP-mediated hydrolysis of mantGTP by Rheb was slower by a factor of ~2.5 relative to unmodified GTP. Considering the rapid intrinsic GTPase activity of Rheb towards mantGTP, it is apparent that TSC2GAP activity is severely inhibited by mant. These results demonstrate the unpredictable effects of mant-tagged nucleotides on intrinsic and GAP-mediated GTPase activities and highlight the utility of the NMR approach.

Previous studies using a phosphate release assay showed that the K m of p120 RasGAP-mediated hydrolysis of p21 N-Ras-mantGTP is lower than that of the p21 N-Ras-GTP complex (Moore et al. 1993). This result suggests that mant may increase the affinity of the HRas-nucleotide complex for RasGAP, thus increasing the rate of nucleotide hydrolysis. Conversely, we propose that for Rheb the mant moiety probably hinders docking of the TSC2GAP domain to Rheb. Note that TSC2 and p120 RasGAP are not homologous, have different folds and function through distinct catalytic mechanisms. Further structural insights are required to address the mechanistic basis for the ‘mant effect’ on GAP- mediated nucleotide hydrolysis for both HRas and Rheb.

2.4.3 GEF-accelerated nucleotide exchange of HRas and RhoA :Using the aforementioned NMR-based approach (Gasmi-Seabrook et al. 2010), we performed GEF assays to measure both association and dissociation of mant-tagged nucleotides, as fluorescence-based experiments have been reported both ways in the literature (Hutchinson and Eccleston 2000; Tan et al. 2002; Margarit et al. 2003; Derewenda et al. 2004; Oleksy et al. 2004; Sondermann et al. 2004; Ford et al. 2005; Hemsath and Ahmadian 2005; Ford et al. 2006; Freedman et al. 2006; Sacco et al. 2006; Gureasko et al. 2008). Comparing dissociation, Cdc25 SOS -mediated nucleotide exchange was 30% slower when starting with mantGDP- bound HRas (mantGDP to GTP) than with GDP-bound HRas (GDP to GTP). Measuring association under the same conditions, the GDP to mantGTP exchange was ~3 times slower than GDP to GTP exchange. Similarly, DH-PH PRG -mediated nucleotide exchange was 30% slower for the GDP to mantGTP than the GDP to GTP exchange in the association assay and ~6 fold slower for the mantGDP to GTP than GDP to GTP exchange in the dissociation assay. In the crystal structure of HRas-mant dGppNHp (Scheidig et al. 1995), mant is near residue Tyr32 in switch I, which would introduce steric clashes at the primary contact point between the GTPase and Cdc25 SOS and could interfere with nucleotide exchange. Assuming the mant moiety of the nucleotide is similarly positioned on the switch I region of RhoA, it would also hinder DH-PH PRG binding to RhoA (Derewenda et al. 2004). Furthermore, mant-induced structural perturbations of the GEF binding sites in the GTPase switch regions may inhibit interactions with GEFs. The exchange kinetics reported by mant are more reliable when the dissociation assay is used for the Ras-CDC25 SOS system and the association assay is used for the RhoA-DH-PH PRG system. Note that these are the conventional fluorescence methods for each protein, nevertheless the less accurate alternative approaches are still in use. Finally, the reduced sensitivity of mantGDP-bound GTPases to the action of GEFs suggests an avenue for design of GTPase inhibitors.

In conclusion, we have demonstrated that mant-labeled nucleotides can alter the intrinsic GTPase activity of Rheb, RhoA, the GAP-catalyzed GTP hydrolysis of HRas-RasGAP and Rheb-TSC2GAP, as well as the DH-PH PRG and CDC25 SOS -mediated nucleotide exchange of RhoA and HRas, respectively. These results reveal that the fluorescent probes could yield biochemically inaccurate data and potentially lead to misleading conclusions. The significant and unpredictable effects of the mant tag clearly indicate that mant-tagged nucleotides should be used with caution and should be validated for each GTPase system studied. At the same time, these findings provide clues as to how one could inhibit or activate specific signaling pathways using small organic molecules, which may mimic the effects of the mant-tag on small GTPases. This study also extends the utility and value of the NMR-based assays for both GTPase and GEF reactions for three small GTPases, suggesting that it will be broadly applicable to the GTPase superfamily.

2.5 Experimental procedures: 2.5.1 Protein preparation : Murine Rheb (residues 1-169), human HRas (residues 1-171) and murine RhoA (residues 1-181) were prepared according to previous protocols (Scheidig et al. 1995; Derewenda et al. 2004; Marshall et al. 2009). In brief, the three proteins were expressed in Escherichia coli (BL21), 15 o grown in minimal media supplemented with NH 4Cl at 15 C with 0.25 mM IPTG. Rheb, HRas and RhoA were expressed using pGEX2T, pET15b and pET28 vectors, respectively. All purified proteins were exchanged into NMR buffer (25 mM sodium phosphate pH 7.0, 100 mM NaCl, 5 mM MgCl 2 and 1mM DTT) in a PD MidiTrap TM G-25 column (GE healthcare). Small GTPase-proteins expressed in E. coli co-purified primarily as complexes with GDP. Tuberous sclerosis 2 (TSC2) GAP domain (residues 1525-1742: hereafter termed TSC2GAP) and the DH-PH fragment of PDZ-RhoGEF (residues 713- 1081: hereafter termed DH-PH PRG ) were prepared using pGEX2T and pGEX4T1 vectors, respectively (Derewenda et al. 2004; Marshall et al. 2009). Catalytic domain constructs of the Son of Sevenless (SOS, Cdc25 residues 566-1049: hereafter referred to as Cdc25 SOS ) and the GTPase activating domain of human GTPase activating protein p120GAP (residues 715-1047: hereafter referred to as RasGAP) were prepared as His-tagged proteins from the pET15b vector. All proteins were cleaved from their tags via thrombin.

2.5.2 NMR-based GTPase, GAP and GEF assays : HRas and Rheb were loaded with GTP or mantGTP by incubation with a 10 fold excess of nucleotide in the presence of 10mM EDTA. RhoA was loaded with GTP or mantGTP in the presence 0.5 M urea and 10 mM EDTA. A 1H-15 N HSQC (Heteronuclear Single Quantum Coherence) spectrum was collected to confirm full nucleotide loading, and the mixture was then passed through a de-salting column (PD MidiTrap TM G-25 (GE healthcare)) equilibrated with NMR buffer, to produce a 1:1 complex of GTPase and the nucleotide.

All NMR experiments were run on a Bruker AVANCE II 800 MHz spectrometer equipped with a 5mm TCI CryoProbe. Sensitivity enhanced 1H-15 N HSQCs with 2 scans (5 min) were run in succession to monitor the intrinsic GTP hydrolysis activity of GTPases (0.1-0.3 mM) at 20 oC. The spectra were processed with NMRPipe (Delaglio et al. 1995) and the peak heights were analyzed with Sparky (Goddard and Kneller) via Gaussian line fitting. Residues from switch I&II, P-loop, β3 & β4 and the α3 helix that exhibit distinct well-resolved peaks in each nucleotide-bound state were used as reporters of the reaction rates for each of the three GTPases, as described previously for Rheb (Marshall et al. 2009).

For the intrinsic nucleotide hydrolysis, the fraction of GTPase protein in the GDP-bound state was calculated for each reporter residue using the equation 1 for Ras and Rheb relying on both GDP and GTP peaks (see appendix A). In the case of RhoA the active state exhibited broadened and split peaks that complicate peak integration, hence we monitored (equation 2; appendix A) the appearance of GDP peaks. Data fitting was done using PRISM (GraphPad software).

To assay GAP-mediated nucleotide hydrolysis, RasGAP or TSC2GAP was added to GTP-loaded HRas or Rheb at a GAP to GTPase molar ratio of 1/2,500 or 1/2.2, respectively. The hydrolysis rate was determined by fitting the data to equation 1 in appendix A.

For GEF assays, DH-PH PRG or Cdc25 SOS were added at a molar ratio of 1/30,000 GEF to RhoA or Ras, in the presence of a 10 fold molar excess of GTP or mantGTP. All the experiments were performed in duplicate with 0.1 mM GTPase. The observed rates of nucleotide exchange assays performed with hydrolysable nucleotides are affected by intrinsic hydrolysis. Thus the observed data were fitted to an equation that considers both exchange and hydrolysis (see appendix A) to extract the true exchange rate.

Using this rate, an exponential decay curve was generated to approximate nucleotide exchange in the absence of hydrolysis.

2.5.3 Fluorescence-based GTPase assay :Emission spectra of 5 M Rheb-mantGTP and Rheb-mantGDP (380-525 nm) were collected at 20 oC on a Shimadzu RF-5301PC spectrofluorophotometer with excitation at 370 nm. Fluorescence emission intensity at 436 nm was monitored during the hydrolysis reaction with excitation at 370 nm. Ten measurements were collected per minute and the average and standard deviation were reported.

2.6 References:

Ahmadian MR, Hoffmann U, Goody RS, Wittinghofer A. 1997. Individual rate constants for the interaction of Ras proteins with GTPase-activating proteins determined by fluorescence spectroscopy. Biochemistry 36 : 4535-4541. Ahmadian MR, Wiesmuller L, Lautwein A, Bischoff FR, Wittinghofer A. 1996. Structural differences in the minimal catalytic domains of the GTPase-activating proteins p120GAP and neurofibromin. J Biol Chem 271 : 16409-16415. Ahmadian MR, Wittinghofer A, Herrmann C. 2002. Fluorescence methods in the study of small GTP- binding proteins. Methods Mol Biol 189 : 45-63. Albert S, Gallwitz D. 2000. Msb4p, a protein involved in Cdc42p-dependent organization of the actin cytoskeleton, is a Ypt/Rab-specific GAP. Biol Chem 381 : 453-456. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. 1995. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR 6: 277-293. Derewenda U, Oleksy A, Stevenson AS, Korczynska J, Dauter Z, Somlyo AP, Otlewski J, Somlyo AV, Derewenda ZS. 2004. The crystal structure of RhoA in complex with the DH/PH fragment of PDZRhoGEF, an activator of the Ca(2+) sensitization pathway in smooth muscle. Structure 12 : 1955-1965. Eberth A, Dvorsky R, Becker CF, Beste A, Goody RS, Ahmadian MR. 2005. Monitoring the real-time kinetics of the hydrolysis reaction of guanine nucleotide-binding proteins. Biol Chem 386 : 1105- 1114. Ford B, Hornak V, Kleinman H, Nassar N. 2006. Structure of a transient intermediate for GTP hydrolysis by ras. Structure 14 : 427-436. Ford B, Skowronek K, Boykevisch S, Bar-Sagi D, Nassar N. 2005. Structure of the G60A mutant of Ras: implications for the dominant negative effect. J Biol Chem 280 : 25697-25705. Freedman TS, Sondermann H, Friedland GD, Kortemme T, Bar-Sagi D, Marqusee S, Kuriyan J. 2006. A Ras-induced conformational switch in the Ras activator Son of sevenless. Proc Natl Acad Sci U S A 103 : 16692-16697. Futai E, Hamamoto S, Orci L, Schekman R. 2004. GTP/GDP exchange by Sec12p enables COPII vesicle bud formation on synthetic liposomes. Embo J 23 : 4146-4155. Gasmi-Seabrook GM, Marshall CB, Cheung M, Kim B, Wang F, Jang YJ, Mak TW, Stambolic V, Ikura M. 2010. Real-time NMR study of guanine nucleotide exchange and activation of RhoA by PDZ-RhoGEF. J Biol Chem 285 : 5137-5145.

Goddard TD, Kneller D. SPARKY 3. in University of California, San Francisco . Gureasko J, Galush WJ, Boykevisch S, Sondermann H, Bar-Sagi D, Groves JT, Kuriyan J. 2008. Membrane-dependent signal integration by the Ras activator Son of sevenless. Nat Struct Mol Biol 15 : 452-461. Hemsath L, Ahmadian MR. 2005. Fluorescence approaches for monitoring interactions of Rho GTPases with nucleotides, regulators, and effectors. Methods 37 : 173-182. Hutchinson JP, Eccleston JF. 2000. Mechanism of nucleotide release from Rho by the GDP dissociation stimulator protein. Biochemistry 39 : 11348-11359. Margarit SM, Sondermann H, Hall BE, Nagar B, Hoelz A, Pirruccello M, Bar-Sagi D, Kuriyan J. 2003. Structural evidence for feedback activation by Ras.GTP of the Ras-specific nucleotide exchange factor SOS. Cell 112 : 685-695. Marshall CB, Ho J, Buerger C, Plevin MJ, Li GY, Li Z, Ikura M, Stambolic V. 2009. Characterization of the intrinsic and TSC2-GAP-regulated GTPase activity of Rheb by real-time NMR. Sci Signal 2: ra3. Moore KJ, Webb MR, Eccleston JF. 1993. Mechanism of GTP hydrolysis by p21N-ras catalyzed by GAP: studies with a fluorescent GTP analogue. Biochemistry 32 : 7451-7459. Mou TC, Gille A, Fancy DA, Seifert R, Sprang SR. 2005. Structural basis for the inhibition of mammalian membrane adenylyl cyclase by 2 '(3')-O-(N-Methylanthraniloyl)-guanosine 5 '- triphosphate. J Biol Chem 280 : 7253-7261. Neal SE, Eccleston JF, Webb MR. 1990. Hydrolysis of GTP by p21NRAS, the NRAS protooncogene product, is accompanied by a conformational change in the wild-type protein: use of a single fluorescent probe at the catalytic site. Proc Natl Acad Sci U S A 87 : 3562-3565. Oleksy A, Barton H, Devedjiev Y, Purdy M, Derewenda U, Otlewski J, Derewenda ZS. 2004. Preliminary crystallographic analysis of the complex of the human GTPase RhoA with the DH/PH tandem of PDZ-RhoGEF. Acta Crystallogr D Biol Crystallogr 60 : 740-742. Phillips RA, Hunter JL, Eccleston JF, Webb MR. 2003. The mechanism of Ras GTPase activation by neurofibromin. Biochemistry 42 : 3956-3965. Remmers AE, Posner R, Neubig RR. 1994. Fluorescent guanine nucleotide analogs and G protein activation. J Biol Chem 269 : 13771-13778. Sacco E, Metalli D, Busti S, Fantinato S, D'Urzo A, Mapelli V, Alberghina L, Vanoni M. 2006. Catalytic competence of the Ras-GEF domain of hSos1 requires intra-REM domain interactions mediated by phenylalanine 577. FEBS Lett 580 : 6322-6328. Scheidig AJ, Franken SM, Corrie JE, Reid GP, Wittinghofer A, Pai EF, Goody RS. 1995. X-ray crystal structure analysis of the catalytic domain of the oncogene product p21H-ras complexed with caged GTP and mant dGppNHp. J Mol Biol 253 : 132-150. Sondermann H, Soisson SM, Boykevisch S, Yang SS, Bar-Sagi D, Kuriyan J. 2004. Structural analysis of autoinhibition in the Ras activator Son of sevenless. Cell 119 : 393-405. Spessard GO. 1998. ACD Labs/LogP dB 3.5 and ChemSketch 3.5. Journal of Chemical Information and Computer Sciences 38 : 1250-1253. Tan YC, Wu H, Wang WN, Zheng Y, Wang ZX. 2002. Characterization of the interactions between the small GTPase RhoA and its guanine nucleotide exchange factors. Anal Biochem 310 : 156-162. Yu Y, Li S, Xu X, Li Y, Guan K, Arnold E, Ding J. 2005. Structural basis for the unique biological function of small GTPase RHEB. J Biol Chem 280 : 17093-17100. Zhang B, Zheng Y. 1998. Regulation of RhoA GTP hydrolysis by the GTPase-activating proteins p190, p50RhoGAP, Bcr, and 3BP-1. Biochemistry 37 : 5249-5257.

CHAPTER 3

Identification of critical amino acids in Rheb GTP hydrolysis

This chapter has been reformatted from the original publication: Mazhab-Jafari MT, Marshall CB, Ishiyama N, Ho J, Di Palma V, Stambolic V, Ikura M. An autoinhibited noncanonical mechanism of GTP hydrolysis by Rheb maintains mTORC1 homeostasis. Structure. 2012 Sep 5;20(9):1528-39.

A link to the published paper can be found at: http://www.sciencedirect.com/science/article/pii/S096921261200247X

3.1 Abstract: Rheb, an activator of mammalian target of rapamycin (mTOR), displays low intrinsic GTPase activity favoring the biologically activated, GTP-bound state. We identified a Rheb mutation (Y35A) that increases its intrinsic nucleotide hydrolysis activity ~10-fold, and solved structures of both its active and inactive forms, revealing an unexpected mechanism of GTP hydrolysis involving Asp65 in switch II and Thr38 in switch I. In the wild-type protein, this non-canonical mechanism is markedly inhibited by Tyr35, which constrains the active site conformation, restricting the access of the catalytic Asp65 to the nucleotide-binding pocket. Rheb-Y35A mimics the enthalpic and entropic changes associated with GTP hydrolysis elicited by the GTPase activating protein (GAP) TSC2, and is insensitive to further TSC2 stimulation. Overexpression of Rheb-Y35A impaired the regulation of mTORC1 signaling by growth factor availability. We demonstrate that the opposing functions of Tyr35 in the intrinsic and GAP- stimulated GTP catalysis are critical for optimal mTORC1 regulation.

3.2 Introduction: Small GTPases act as molecular switches to regulate diverse cellular functions. When bound to guanosine trisphosphate (GTP), they adopt an ‘on’ conformation that elicits a biological response. GTP hydrolysis is accompanied by a conformational change into a GDP-bound ‘off’ conformation. Cycling between the active and inactive states of each GTPase is a result of the intrinsic nucleotide hydrolysis and exchange rates, and regulatory proteins that catalyze these processes. GTPase activating proteins (GAPs) stimulate GTP hydrolysis, whereas guanine nucleotide exchange factors (GEFs) mediate the displacement of GDP, allowing a new GTP molecule to bind (Bos et al. 2007). GTPase proteins possess either complete or partial catalytic machinery for hydrolysis of GTP. In most cases, an electronegative group is used for stabilization/polarization of the hydrolytic water for in-line nucleophilic attack of the γ-phosphate (Maegley et al. 1996; Li and Zhang 2004). In most Ras and Rho subfamily GTPases, this is achieved by the carboxamide oxygen of a conserved Gln in a dynamic loop called switch II. Ras and Rho GAPs work by stabilizing this Gln in a catalytic conformation, while an Arg residue referred to as an “Arginine finger” neutralizes the developing negative charge on the α- and β-phosphates of GTP (Scheffzek et al. 1997). In other systems, such as Rap-RapGAP, a catalytic asparagine is provided in trans by the GAP (Scrima et al. 2008).

Ras homolog enriched in brain (Rheb) is a key regulator of the mTOR complex 1 (mTORC1) signaling pathway (Inoki et al. 2003; Dunlop et al. 2009). Rheb-GTP promotes phosphorylation of mTORC1 targets, resulting in enhanced protein translation and cellular growth (Garami et al. 2003). Rheb has an unusually slow intrinsic GTPase activity, which is regulated by the GAP activity of tuberous sclerosis complex 2 (TSC2), a tumor suppressor frequently inactivated in human patients with the tumor predisposition syndrome tuberous sclerosis (Garami et al. 2003; Tee et al. 2003). Rheb overexpression has been observed in certain cancer cell lines (Im et al. 2002; Eom et al. 2008; Nardella et al. 2008) and constitutively activated Rheb mutants can induce oncogenic transformation in cell culture (Jiang and Vogt 2008). The low intrinsic GTPase activity of Rheb has been attributed to the catalytically incompetent conformation of Gln64 (Yu et al. 2005), which is homologous to Ras Gln61, but does not participate in GTP hydrolysis (Li et al. 2004; Marshall et al. 2009). TSC2GAP is thought to utilize Asn1643 to promote GTP hydrolysis by substituting for Gln64 in an “Asn thumb”-type mechanism (Inoki et al. 2003; Marshall et al. 2009) similar to that of RapGAP (Scrima et al. 2008).

Here, utilizing site-directed mutagenesis, crystallography and real-time NMR-based GTPase assays, we discovered that Rheb Tyr35, a residue that is highly conserved across the small GTPase superfamily (Wennerberg et al. 2005), maintains the high activation state of Rheb by inhibiting intrinsic GTP hydrolysis. Mutation of this residue substantially accelerated intrinsic nucleotide hydrolysis through a catalytic mechanism that did not require Gln64, but also conferred resistance to the activity of TSC2. Crystal structures of RhebY35A led us to identify the backbone carbonyl of Thr38 and side chain of Asp65 as candidate residues that contribute to the intrinsic GTPase activity. Mutagenesis studies confirm that Asp65 contributes significantly to the intrinsic GTPase activity of both wild-type Rheb and the Y35A mutant. Further, Asp65 was absolutely essential for the sensitivity of Rheb to the GAP activity of TSC2, whereas Gln64 was dispensable. Consistent with the in vitro data, expression of Rheb Y35A and D65A mutants in mammalian cells affected transduction of growth factor signals to mTORC1. Taken together, our observations reveal an efficient non-canonical mechanism of GTP hydrolysis by Rheb, and suggest that autoinhibition of catalysis maintains Rheb in its highly activated state upon growth factor stimulation, which is necessary for the proper signal transduction to mTORC1.

3.3 Results: 3.3.1 Rheb Tyr35 inhibits intrinsic GTPase activity: We previously showed that fluorescent-tagged nucleotides can alter the hydrolysis and exchange rates governing the GTPase cycle (Mazhab-Jafari et al. 2010). The most striking example we observed was that 2'-/3'-O-(N'- Methylanthraniloyl) (mant)GTP was hydrolyzed by Rheb ~10-fold faster than GTP. This is not an intrinsic property of the modified nucleotide as the mant moiety inhibited GTP hydrolysis by RhoA and did not affect hydrolysis by Ras. The rate of mantGTP hydrolysis by Rheb is similar to that of Ras (Figure 3.1a), indicating that Rheb has a latent capacity for efficient catalysis. Interestingly however, the rapid hydrolysis of mantGTP was independent of Rheb Gln64 (Mazhab-Jafari et al. 2010). The position of the fluorophore in a structure of Ras bound to a non-hydrolyzable analog of mantGTP (Scheidig et al. 1995) suggested it may interact with the phenol ring of Tyr35 in switch I of Rheb. Remarkably, mutation of Tyr35 to Ala recapitulated the mant effect, increasing the rate of GTP hydrolysis by an order of magnitude (Figure 3.1b). Furthermore, the mant tag had no further effect on the catalytic activity of Rheb Y35A, suggesting that the mutation and the fluorophore stimulate hydrolysis through the same mechanism (Figure 3.1b). These observations indicate that Tyr35 auto-inhibits the intrinsic GTPase activity of Rheb.

Figure 3.1 Rapid hydrolysis of mantGTP by Rheb is related to autoinhibitory role of Tyr35. a) Hydrolysis of GTP or mantGTP by Rheb (black and green, respectively) and Ras (red and blue, respectively). Reaction rates derived by curve fitting are presented in the insets. b) Hydrolysis of GTP by wild-type Rheb (black), and GTP and mantGTP by Rheb Y35A (green and red respectively).

3.3.2 Structural basis for the Tyr35 auto-inhibitory function: We crystallized GDP-bound Rheb Y35A in the presence of excess GMPPNP (a non-hydrolyzable analog of GTP) and to our surprise, the asymmetric unit contained two molecules of Rheb, one bound to GDP and one to GMPPNP (Figure 3.2a-c). The overall protein fold is very similar to WT Rheb (Yu et al. 2005) (backbone RMSD of 0.44 Å) with a few key differences. The nucleotide-binding pocket is completely solvent exposed in the GMPPNP-bound structure of Rheb Y35A, whereas in the WT protein the triphosphate group of the

Figure 3.2 Structure and dynamics of Rheb Y35A. (a) Asymmetric unit and electron density of nucleotide substrates. The asymmetric unit of the RhebY35A crystal containing one GDP-bound (cyan) and one GMPPNP-bound (green) Rheb molecule is shown in the center. 2F o-Fc electron density maps at 1.5 σ of the nucleotide binding site with GMPPNP (left) fitted into one Rheb Y35A molecule and GDP (right) in the second molecule of the crystal asymmetric unit. (b) Ribbon model of GMPPNP-bound Rheb Y35A. (c) Ribbon model of GDP-bound Rheb Y35A in the same orientation as b. Panels d, e, f and g show overlays of GMPPNP- or GDP-bound Rheb Y35A (colored as above) and WT Rheb in

complex with GMPPNP (1XTR) (magenta) or GDP (1XTQ) (gray), as indicated. d) Mutation of Tyr35 affects the position of the catalytic water (spheres) and γ-phosphate with respect to the carbonyl of Thr38. e) Minor conformational change of Ala35 in the Rheb mutant upon GTP hydrolysis. The Ala35 C α and C β translocation distances from the GMPPNP-bound form to the GDP-bound form are shown. f) Major conformational rearrangement of WT Rheb Tyr35 upon GTP hydrolysis, with translocations indicated as in e. g) Position of switch II residues relative to the nucleotide binding site in the activated form of WT Rheb versus that of the Y35A mutant. The backbone of N-terminal switch II residues Gly63-Ser68 and side chains of Asp65 and Glu66 are shown. Two conformations were observed for the Asp65 side chain. h) 1H-15 N HSQC spectra illustrating cross-peaks from switch II residues in WT Rheb (black) and Rheb Y35A (red) in complex with GTP. The panel showing Ser68 is illustrated at a higher contour level for clarity. The reduction in height of the peak from the mutant relative to the wild-type peak is indicated as a percentage at the bottom of each panel in which it is measurable. The full spectra are shown in figure 3.7a.

nucleotide is shielded from the solvent by the phenol ring of Tyr35, which forms a hydrogen bond with the γ-phosphate. In addition, the γ-phosphate is 0.5 Å closer to Thr38 in the absence of Tyr35 (Figure 3.2d), which in the WT structure “pulls” the γ-phosphate toward the middle of switch I. Interestingly, the hydrolytic water is closer to the backbone carbonyl of Thr38 in the mutant (2.7 Å versus 3.8 Å in the WT protein) (Figure 3.2d), placing it in a more electron-rich environment that may enhance its polarization for an in-line nucleophilic attack to the γ-phosphate. It has been proposed that the corresponding backbone carbonyl of Ras (Thr35) contributes to the stabilization/activation of the catalytic water during intrinsic GTP hydrolysis (Frech et al. 1994; Buhrman et al. 2010). Comparison of our structure with that of wild-type Rheb indicates that Tyr35 pulls the γ-phosphate and catalytic water away from the Thr38 carbonyl, thus reducing its catalytic contribution.

Switch I of Rheb Y35A does not undergo any substantial conformational change upon nucleotide hydrolysis, whereas this region of the WT protein exhibits a large structural change mediated by an interaction between Tyr35 and the γ-phosphate (Yu et al. 2005) (Figure 3.2e&f). It was hypothesized that a similar nucleotide-dependent rearrangement of Rap Tyr32 would be energetically unfavourable to the GTPase reaction (Cherfils et al. 1997), consistent with our observation that nucleotide hydrolysis is accelerated by a mutation that disrupts this conformational change.

3.3.3 Identification of a catalytic residue for GTP hydrolysis: Previous work has shown that Gln64, corresponding to the catalytic Gln61 of Ras, is not involved in GTP hydrolysis by WT Rheb (Inoki et al. 2003; Li et al. 2004; Marshall et al. 2009). Likewise, Gln64 remains in a non-catalytic conformation in the structure of Rheb Y35A (Figure 3.3a) and is not required for the accelerated hydrolysis of GTP by

Figure 3.3 Mutagenic analysis of potential catalytic residues in Rheb Y35A. a) Position in the structure of GMPPNP-bound Rheb Y35A of potential catalytic residues selected for mutagenesis,. b-e) Intrinsic GTPase assays for hydrolysis of GTP or mantGTP by WT Rheb and mutants as indicated: Hydrolysis of (b) GTP by Rheb Y35A and the double mutant Y35A, S16A, (c) mantGTP by WT Rheb and Rheb D36A, (d) GTP by Rheb Y35A and the double mutant Y35A, Q64L, (e) GTP by Rheb Y35A and the double mutant Y35A, R15G. Hydrolysis of mantGTP by Rheb Y35A is shown in Figure 3.1b, and hydrolysis of GTP by additional Rheb mutants is shown in Figures 3.4, 3.5 and 3.9. f) Multiple conformations of Arg15, Asp65, and Ser68 in the structure of Rheb Y35A in complex with GMPPNP.

The 2F o-Fc electron density map at 1.0 σ is shown for the three aforementioned residues with double- headed arrows indicating the movement of the side chains.

Rheb Y35A (Figure 3.3d and 3.5a). Since the catalytic residues of other small GTPase superfamily members are found in the N-terminus of switch II, we examined this region for residues with electronegative side chains that may contribute to the hydrolytic reaction. Immediately downstream of Gln64 are two residues with acidic side chains, Asp65 and Glu66 (Figure 3.2g). The crystal structure of Rheb Y35A shows that the backbone of the N-terminal loop of switch II of this mutant is displaced by an average of 1 Å toward the nucleotide binding pocket relative to the wild type, which brings the side- WT chain carboxylate of Asp65 closer to the nucleotide by 1 Å (average Asp65O δ1,2 - average Y35A Asp65O δ1,2 ) (Figure 3.2g). Mutation of Asp65 to Ala reduced the intrinsic hydrolysis of Rheb Y35A by more than 60% and that of wild type by 30% (Figure 3.4a), as did the conservative substitution of Asp65 by Asn (Figure 3.5b). On the other hand, mutations of Glu66 had no effect on intrinsic GTPase activity (Figure 3.5c), consistent with its perpendicular orientation away from the nucleotide (Figure 3.2g). We also tested all other residues found within 10 Å of the hydrolytic water in the Rheb Y35A structure that could potentially provide (i) a negative charge to activate this water molecule, or (ii) a positive charge to stabilize the β- and γ-phosphates in the transition state for hydrolysis (Figure 3.3). There was no change in the rate of intrinsic nucleotide hydrolysis associated with R15G, S16A, or D36A mutations (Figure 3.3b, c, e). The only other charged residues within 10 Å of the hydrolytic water are Lys19 and Asp60 of the highly conserved G1 and G3 box motifs, respectively. The Rheb K19A mutant failed to express, presumably due to impaired nucleotide binding, and D60A was highly unstable and could not be loaded with GTP, consistent with the role of this residue in Mg ++ coordination (Yu et al. 2005). These data strongly suggest that Asp65 is the sole candidate for a catalytic residue in Rheb. Notably, carboxylates are more potent nucleophiles than carboxamides, and consistently, the Q61E substitution increased the GTPase activity of Ras (Frech et al. 1994).

In the structure of wild-type Rheb, the carboxylate of Asp65 is 12 Å (average Asp65O δ1,2 ) from the γ- phosphate in a single conformation, whereas the electron density of Rheb Y35A indicates that Asp65 exists in two conformations, 11.0 and 12.0 Å from the γ-phosphate, respectively (Figure 3.3f). By comparison, the catalytic carboxamide of Ras (Gln61O ε) has been found at distances varying from 4.7 to 12.2 Å from the γ-phosphate (median distance of 8.1Å) (Figure 3.6a) in available crystallographic snapshots, consistent with the dynamic nature of switch II determined by NMR studies (Ito et al. 1997).

Figure 3.4 Role of Asp65 in intrinsic and GAP-mediated GTP hydrolysis by Rheb. a) Hydrolysis of GTP by wild-type Rheb (black), the mutants D65A and Y35A (blue and green, respectively), as well as the double mutant Y35A-D65A (red). Reaction rates derived by curve fitting are presented in the insets. b) Sensitivity of WT Rheb and Asp65 mutants to TSC2GAP-stimulated GTP hydrolysis. WT, WT+GAP, D65A+GAP, D65E+GAP, and D65N+GAP are shown with black, blue, green, yellow, and red respectively.

Thus, despite its established role as a catalytic residue (Frech et al. 1994), Gln61 is rarely found in a catalytically competent conformation in Ras crystal structures, presumably because this state is transient and energetically unfavorable (Grant et al. 2009; Fraser et al. 2011). Similarly, our Y35A structure and

the previous wild-type Rheb structure (Yu et al. 2005) both appear to be energetically stable states, with the conformations of Asp65 stabilized primarily by ionic and polar interactions with the Arg15 and Ser68 side chains, which are also found in two alternate conformations in our structure (Figure 3.3f). Interestingly, comparison of WT and Y35A 1H-15 N heteronuclear single quantum coherence (HSQC) spectra revealed increased line broadening for residues in the P-loop and the N-terminus of switch II of GTP-bound Rheb Y35A (Figure 3.2h & 3.7a), suggesting elevated dynamics in µs-ms time scale. This

Figure 3.5 Effect on Rheb GTPase activity of mutations of acidic and electronegative residues in switch II. a) Intrinsic GTPase assays for hydrolysis of mantGTP by WT Rheb (black), Rheb D65A (green), and Rheb Q64L (red). b) Intrinsic GTPase assays for hydrolysis of GTP by WT Rheb (black) and Rheb D65N (green). c) Mutation of Glu66 impairs TSC2 GAP-catalyzed but not intrinsic GTPase activity of Rheb. Intrinsic GTP hydrolysis by WT and E66A Rheb are shown in black and green, respectively. TSC2GAP-mediated GTP hydrolysis by Rheb WT and E66A are shown in blue and red, respectively. Note that the GTPase activity of Rheb E66A is modestly stimulated by TSC2GAP, but this mutant exhibits reduced sensitivity to GAP activity.

could allow the N-terminus of switch II to sample alternate conformations closer to the nucleotide and the catalytic water. The elevated dynamics of the N-terminal region of switch II and its proximity to the nucleotide binding site in Rheb Y35A is consistent with the greater impact on catalysis of Asp65 mutations in the Y35A mutant than in wild-type Rheb (Figure 3.4a). Hence, in addition to affecting the orientation of the nucleotide and hydrolytic water, Tyr35 may reduce the intrinsic GTPase activity of Rheb by restricting the dynamics of switch II and displacing it from the nucleotide binding site. Relative to Ras, the catalytic Gln residues of Rho subfamily GTPases were found closer to the γ-phosphate (median distance of 5.5Å) (Figure 3.6a and c), which may contribute to their faster intrinsic nucleotide hydrolysis rate (Mazhab-Jafari et al. 2010). On the other hand Gln63, which was recently proposed to be a non-canonical catalytic residue of Rap in GAP1 IP4BP -mediated GTP hydrolysis (Sot et al. 2010), is found with a median distance of 11.8Å from the γ-phosphate in structures of free Rap (Figure 3.6a), consistent with the slow nucleotide hydrolysis of this GTPase.

In Ras, Gly12, Gly13 and Gln61 are the major sites of oncogenic mutations. Mutation of Ras Gly12 to any other residue hinders GTP hydrolysis by sterically occluding access of the catalytic residue Gln61 to the hydrolytic water and nucleotide (Krengel et al. 1990). However, Rheb has an Arg in this position and its mutation to Gly (R15G) does not increase the catalytic activity of Rheb Y35A (Figure 3.3e) or WT (Yamagata et al. 1994; Im et al. 2002; Li et al. 2004; Marshall et al. 2009). The distinctive impact of P-loop residues on the activities of Ras and Rheb lends further support to the different molecular mechanisms of action of these two closely related GTPase homologs.

3.3.4 Involvement of Rheb’s Asp65 and Tyr35 in TSC2GAP-mediated GTP hydrolysis: Mutation of the solvent exposed residue Asp65 to Ala (D65A) did not perturb the structure of Rheb, on the basis of minimal and localized chemical shift perturbations in the 1H 15 N HSQC spectra that were mainly confined to switch II (Figure 3.8), but totally abolished the susceptibility of Rheb to the GAP activity of

Figure 3.6 Structure-function analysis of the position of putative catalytic residues and the conserved switch I Tyrosine in small GTPases. a) Crystallographic ensemble of interatomic distances between the γ-phosphate of GTP (or analogue) and the proposed catalytic residues of Ras, Rheb, Rho, and Rap GTPases. Histogram of distances measured from the γ-phosphate (P γ) to Ras Gln61O ε (blue), Rheb average Asp65O δ1,2 (red), Rho Gln63O ε (green), and Rap Gln63O ε (purple) from crystal structures of free protein in complex with GTP and non-hydrolyzable GTP analogs. All available PDB-archived structures of Ras with wild-type GTPase activity where Q61 is resolved were included, [PDB code] (distance in Å): [5P21] (6.1), [6Q21] (6.3, 7.0, 6.8, 7.6), [1QRA] (7.8, 6.7), [1CTQ] (6.7), [1GNP] (9.7),

[3K8Y] (5.9), [3K9N] (11.7), [1P2S] (6.3), [1P2V] (6.7), [1PLK] (4.7), [2CL0] (12.2), [2CL7] (10.4), [2CLC] (10.3), [121P] (8.1), [3PIR] (9.0), [3PIT] (9.8), [3I3S] (7.2). Some structures contain more than one Ras molecule in the asymmetric unit, in which case the distance was measured for each molecule separately. 1GNP and 3K9N are structures of WT-Ras in complex with mantGMPPNP and Ras Y32F in complex with GMPPNP, respectively, both of which have the same intrinsic nucleotide hydrolysis rate as WT Ras (Yamasaki et al. 1994; Mazhab-Jafari et al. 2010). 2CL0, 2CLC and 2CL7 are structures of Ras bearing a fluorescent tag on switch I (with Y32C mutation) in complex with GMPPNP or GTP (Klink et al. 2006). 3PIR, 3PIT, and 3I3S are structures of Ras D41E and T50I, mutations that are solvent exposed and far away from switch II. For Rheb, distances were measured for WT Rheb- GMPPNP [1XTR] (12.0), WT Rheb-GTP [1XTS] (12.1) and two separate measurements were made for Rheb Y35A-GMPPNP, in which Asp65 exists in two conformations (11.0, 12.0). The distances for Rho subfamily GTPases (Rac1, TC10/RhoQ, RhoC, Rac3, Cdc42) were measured using the following structures: [1MH1] (5.7), [2ATX] (6.8, 7.4), [2GCO] (5.7, 5.7), [2GCP] (5.6), [2IC5] (4.8), [2QRZ] (4.4, 3.8). 1MH1 is structure of Rac1 M1P-F78S double mutant, both of which are far away from nucleotide binding site. Finally, the distances for Rap were from structures of Rap2A: [2RAP] (12.7), [3RAP] (10.4, 12.2). b) Curves showing the intrinsic GTP hydrolysis rate of WT RhoA (black) and the Y34A mutant (green). c) The positions of the Tyr34, Gln63 and the non-hydrolyzable nucleotide GTP γS in RhoA G14V mutant (PDB code: 1A2B). d) Curves showing the intrinsic GTP hydrolysis rate of WT Ras (black) and the Y32A mutant (green). e) The positions of the Tyr32, Gln61 and the non- hydrolyzable nucleotide GMPPNP in Ras (PBD code: 5P21).

TSC2 (Figure 3.4b). Furthermore, even conservative mutations of Asp65 (D65E/N) rendered Rheb totally insensitive to the activity of TSC2 GAP. The strict requirement for the geometry and charge of this side chain suggest that it might be a critical catalytic residue for the GAP-mediated hydrolysis reaction. We also tested the sensitivity of the GTPase activity of Rheb Y35A to the action of TSC2GAP and found that the GTPase activity of this mutant was not further stimulated by the addition of the GAP domain of TSC2 (Figure 3.9a). An analagous mutation (Y32A), impaired the sensitivity of Rap GTPase to the function of RapGAP (Brinkmann et al. 2002; Scrima et al. 2008), however a conservative mutation (Y32F) was tolerated. Interestingly, the Y35F mutation was sufficient to render Rheb insensitive to the function of the TSC2GAP (Figure 3.9b), highlighting differences in the details of molecular recognition in these two homologous systems.

3.3.5 Thermodynamic basis for the Tyr35 auto-inhibitory function: To better understand the energetic basis of Tyr35 auto-inhibition, we analyzed the thermodynamics of the GTP hydrolysis reaction using an Arrhenius plot (Figure 3.10). This powerful technique allows one to extract energetic parameters, such as enthalpy, entropy and free energy, from the highly unstable and low populated transition state of an enzymatic reaction. The increased catalytic activity of Rheb Y35A was associated with a large reduction in the activation enthalpy for GTP hydrolysis (Figure 3.10, Table 3.1). However, the

Figure 3.7 Perturbation of Rheb HSQC spectrum by Tyr35 mutation. a) HSQC spectrum of Rheb Y35A-GTP (red) overlaid with WT Rheb-GTP (black). Resonances that are present in the wild-type

spectrum but exhibit large chemical shift changes or severe broadening in the spectrum of the mutant are circled. Resonances showing minor chemical shift changes due to the mutation are indicated by double headed arrows. Note that in the GTP-bound form there are no extra peaks in the Y35A mutant, indicating that most circled peaks correspond to resonances broadened beyond detection in the mutant. b) HSQC spectrum of Y35A-GDP (red) overlaid with WT Rheb-GDP (black). Resonances that are unique to the wild-type spectrum are circled and those unique to the mutant spectrum are highlighted by squares. Note that in the GDP-bound form, the total number of observable resonances remains constant between WT Rheb and the Y35A mutant (circles ≈ squares), indicating that most of these resonances undergo long range chemical shift changes.

activation entropy was also reduced (unfavorable contribution), resulting in a modest decrease in the overall activation free energy of the nucleotide hydrolysis reaction in the mutant. Because there is a build-up of negative charge on the β-γ bridging oxygen during GTP hydrolysis (Cepus et al. 1998; Du et al. 2000; Allin et al. 2001), the proximity of the electron rich phenol ring of Rheb Tyr35 could destabilize the transition state, which is consistent with the reduction in activation enthalpy associated with mutation of this residue. Interestingly, the Arg fingers of Ras- and Rho-GAPs accelerate nucleotide hydrolysis of their cognate GTPases by providing positive charge in a position equivalent to that of Rheb Tyr35. Another contribution to the enthalpic term may come from the strengthened hydrogen bond between the Thr38 carbonyl in the mutant and the repositioned catalytic water, which may be more reactive toward the γ-phosphate (Frech et al. 1994).

Figure 3.8 Minimal chemical shift perturbation associated with mutation of Rheb Asp65. HSQC spectrum of Rheb D65A-GTP (red) overlaid with WT Rheb-GTP (black). Resonances that are strongly perturbed by the mutation (severe broadening or large chemical shift change) are circled and those showing minor chemical shift changes are indicated by double headed arrows.

The larger negative value of S‡ for Rheb Y35A indicates that formation of the transition state requires the mutant to undergo a larger increase in “order” than the wild-type protein. In the crystal structure of WT Rheb-GMPPNP, a hydrogen bond between the hydroxyl of Tyr35 and the γ-phosphate of GMPPNP stabilizes switch I, contributing to the order of the ground state. Disruption of this contact by mutation of Tyr35 increases the disorder in switch I, as illustrated by partial spectral broadening of peaks from residues 27-30 of GTP-bound Rheb Y35A (Figure 3.7). Thus assembly of the ordered transition state from the more flexible Rheb Y35A ground state would be more entropically unfavourable. The conservative mutation Y35F increased intrinsic hydrolysis almost as much as Y35A with similar thermodynamic effects (Figure 3.10, Table 3.1), suggesting that the H-bond between the hydroxyl of Tyr35 and the γ-phosphate is critical for the auto-inhibition of Rheb’s GTPase activity. It is very interesting to note that the thermodynamic landscape of intrinsic GTP hydrolysis in the Y35A mutant

Figure 3.9 Tyr35 hydroxyl is required for TSC2GAP-assisted GTP hydrolysis. a) Intrinsic and GAP-catalyzed GTP hydrolysis by Rheb Y35A. b) Intrinsic and GAP-catalyzed GTP hydrolysis by Rheb Y35F. In both graphs the GTPase activity of mutant Rheb with and without TSC2GAP are shown in black and green, respectively, and the TSC2GAP-catalyzed GTP hydrolysis by WT Rheb is shown in blue.

(reduced activation enthalpy with an entropic penalty) is similar to that reported for TSC2GAP-mediated GTP hydrolysis in WT Rheb (Marshall et al. 2009), suggesting that TSC2GAP may promote hydrolysis in part by disrupting the electrostatic contact between Tyr35 and the γ-phosphate. Consistent with this hypothesis, the increased rate of GTP hydrolysis by the Rheb Y35A/F mutants is not further accelerated by the addition of the GAP (Figure 3.9). The larger reduction of the activation enthalpy by TSC2GAP- mediated catalysis compared to Y35A mutation suggests the GAP provides additional stimulatory electrostatic contributions to GTP hydrolysis, perhaps via complementation of the intrinsic catalytic machinery by the Asn thumb. On the other hand the larger unfavorable reduction in entropy of the GAP- mediated reaction could be due to complex formation between Rheb and the GAP domain of TSC2.

Figure 3.10 Effect of Tyr35 and Asp65 mutations on the thermodynamic activation parameters for GTP hydrolysis by Rheb. Arrhenius plots for intrinsic GTP hydrolysis by Rheb WT (black), Y35A (green), Y35F (brown), Y35A-D65A (blue) and D65A (red). Each data point represents a rate

determined from a representative GTPase assay consisting of 25 time points performed at a given temperature. Error bars represent standard error associated with derivation of the rate by curve fitting.

Mutation of Asp65 substantially impairs the stimulatory effect of the Y35A mutation on intrinsic hydrolysis, hence we measured the thermodynamic parameters of the transition state for the Rheb double mutant Y35A-D65A (Table 3.1). Mutation of Asp65 increased the activation enthalpy (H‡ ) of Rheb Y35A, indicating that the negatively charged carboxylic acid side chain of Asp65 stabilizes the transition state since the enthalpic term originates primarily from electrostatic interactions (Kötting and Gerwert 2004). In Ras, the enthalpic contribution to hydrolysis was attributed to the charge shift from the γ- toward the β-phosphate (Kötting and Gerwert 2004). We propose that electrostatic interactions between Rheb Asp65 and the nucleotide similarly shift charge in the transition state to promote hydrolysis. Interestingly, Tyr35 reduces the enthalpic contribution of Asp65 to GTP hydrolysis, H‡(Y35A-Y35A,D65A) > H‡(WT-D65A) (Table 3.1), which is consistent with our kinetic data (Figure 3.4a).

Table 3.1 Summary of thermodynamic activation parameters. [free energy of activation ( G‡), activation enthalpy ( H‡), and activation entropy (T S‡) in kJ/mol] calculated for GTP hydrolysis by Rheb WT and mutants. T is set to 298 K in the equation; G‡ = H‡ - T S‡.

Protein WT Y35A Y35F Y35A- D65A TSC2GAP* D65A H‡ 86.0 52.3 52.7 55.6 87.0 41.3

TS‡ -13.5 -43 -43.3 -41.6 -13.3 -51.8

G‡ 99.5 95.3 96.0 97.2 100.3 93.1

*: Values from (Marshall et al. 2009).

3.3.6 Regulation of mTORC1 by growth factors involves the non-canonical catalytic and autoinhibitory mechanisms: To investigate the role of Rheb’s newly discovered catalytic machinery, and its autoinhibition, in the activation of mTORC1, HeLa cells were transfected with wild-type Rheb or the Y35A or D65A mutants. Phosphorylation of p70 S6K Thr389, a measure of mTORC1 signaling throughput (Figure 3.11a), was monitored by immunoblotting, upon growth factor starvation, as well as 15 min and 6 h after growth factor stimulation (Figure 3.11b & Figure 3.12). Following growth factor stimulation, when the TSC1/2 complex is down-regulated (Benvenuto et al. 2000; Chong-Kopera et al. 2006), mutation of Rheb Tyr35 to Ala reduced the activation of mTORC1, as evidenced by slightly lower p70

S6K phosphorylation despite higher Rheb Y35A expression compared to Rheb WT, an effect made more apparent with longer stimulation (Figure 3.11b & 3.12). Remarkably, in the absence of stimuli upon serum starvation, Rheb Y35A led to a significant increase in mTORC1 activity (Figure 3.11b), as did the expression of the GAP-resistant mutant Rheb D65A (Figure 3.11b). Mutation of Asp65 (but not Tyr35) perturbed the chemical shift of the neighboring switch II residue Tyr67 (Figure 3.7a & 3.8), previously shown to be critical for activation of mTORC1 (Long et al. 2007), which may affect the ability of Rheb D65A to fully activate mTORC1. This may explain the fact that under conditions of growth factor stimulation, Rheb D65A induced less S6K phosphorylation than the wild type protein, despite its reduced intrinsic GTPase activity (Figure 3.11b).

Figure 3.11 Mutations of Rheb catalytic and autoinhibitory residues impact Rheb’s activation level and mTORC1 phosphorylation of p70 S6K. a) Signaling pathway by which growth factors

stimulate mTORC1 phosphorylation of p70 S6K. b) Normalized values of p70 S6K phosphorylation for cells starved (-) or stimulated (+) with serum and insulin for 15 min (left) or 6 h (right) are shown as bar graphs. More details in experimental procedures. Representative Western blots are shown in Figure 3.12. c) Rheb nucleotide loading monitored in HEK293 cells with 32 P-labeling experiment as described in experimental procedures.

Guanine nucleotide loading of WT Rheb and the mutants in vivo was determined by their immunoprecipitation from HEK293 cells metabolically labeled with 32 P-orthophosphate and nucleotide resolution by thin layer chromatography (Figure 3.11c). Consistent with its increased GTPase activity, Rheb Y35A displayed reduced loading of GTP, whereas Rheb D65A showed modestly increased GTP loading. Reflected in the elevated loading of WT Rheb with GTP (Figure 3.11c), the overabundance of transfected Rheb proteins likely negates the effects of endogenous TSC2 on the overexpressed Rheb

Figure 3.12 Effects of Y35A and D65A mutations in Rheb on mTORC1 activity. Western blots detecting p70 S6K phosphorylated at Thr389, total p70 S6K, and Myc-tagged Rheb in HeLa cells transfected with wild-type Rheb, or D65A or Y35A mutants, or empty vector, in the absence (-) or presence (+) of serum and insulin for a) 15min and b) 6 h. More details in experimental procedures.

mutants in this cell line. While the lower GTP loading of Rheb Y35A indicates higher intrinsic hydrolysis rate than that of the wt protein, modestly increased GTP loading of D65A signifies the autoinhibitory effects of Tyr35 on Asp65’s contribution to intrinsic catalysis, consistent with their in vitro behavior (Figure 3.4a).

3.4 Discussion: GTPases are versatile molecular switches that utilize surprisingly diverse mechanisms to mediate the interconversion between the active and inactive states. The study presented here illustrates how Rheb evolved a GTP hydrolysis mechanism drastically different from its close homolog H-Ras. Rheb employs an autoinhibitory mechanism to maintain a high activation state in cells essential for the proper maintenance of mTORC1 signaling and cellular growth. The novel auto-inhibitory mechanism functions via an interaction between Rheb Tyr35-OH and the γ- phosphate of GTP, which hinders GTP hydrolysis. Interestingly, our investigation of this inhibitory mechanism led to the elucidation of an unusual Rheb catalytic mechanism involving Asp65, which is one position downstream of the canonical catalytic Gln, equivalent to Ras Gln61 and Rho Gln63. In the canonical mechanism, the catalytic water is activated/stabilized by interaction with an electronegative group, the carboxamide oxygen of a glutamine, provided either in cis or trans (Bos et al. 2007). Although an equivalent residue (Gln64) is present in the sequence of Rheb switch II, our work demonstrates that the catalytic function is carried out by the adjacent Asp65.

A Tyr residue in switch I is highly conserved amongst the GTPase superfamily, however the functional role of this residue varies. For example, in RhoA, Tyr34 stimulates intrinsic hydrolysis, presumably by stabilizing the catalytic conformation of Gln63 in switch II, whereas Ras Tyr32 is solvent exposed and does not impact the intrinsic hydrolysis (Figure 3.6). In Rheb, Tyr35 counters the contribution of Asp65 to catalysis by restricting the dynamics of switch II and reducing its access to the catalytic site (Figure 3.13). Severe peak broadening was observed for the amides of Gly63, Gln64, Asp65 and Glu66 in the 1H-15 N HSQC spectrum of the Y35A mutant (Figure 3.2h), presumably the result of disrupting the H- bond network from Tyr35 to the amide of Gly63 through the catalytic water and the γ-phosphate (Figure

3.2d). Moreover, Tyr35 restricts the position of the hydrolytic water and reduces its polarization by the carbonyl of Thr38, which is thought to promote hydrolysis (Frech et al. 1994; Buhrman et al. 2010). The residual catalytic activity found in Rheb D65A may reflect the stimulatory effect of the Thr38 backbone carbonyl in polarization and/or stabilization of the hydrolytic water. Upon mutation of Tyr35, Asp65 adopts two conformations, one of which is more proximal (relative to the WT) to the nucleotide binding site and appears better positioned for hydrolysis (Figure 3.2g), consistent with our finding that mutation of Asp65 has a greater impact on the catalytic activity of Rheb Y35A than that of WT (Figure 3.4a). On the basis of the multiple crystallographic conformations of Asp65, and the NMR peak broadening, we propose that in both wild-type and Y35A Rheb, switch II exists in an ensemble of conformations, some of which allow Asp65 to adopt a catalytic conformation closer to the nucleotide.

Structural evidence implicating Asp as a catalytic residue exists for at least one other member of the GTPase superfamily, the signal recognition particle (SRP) Ffh/FtsY (Focia et al. 2004), which is a large dimeric prokaryotic GTPase that is highly divergent from the Ras family. Nevertheless, to our knowledge, a catalytic Asp has not been previously reported within the Ras subfamily. On the other hand, the inhibitory role for a switch I Tyr we discovered in Rheb may be relevant in inhibiting the intrinsic hydrolytic machinery in certain other GTPases, such as Ran (Brucker et al. 2010).

Our structural analysis of Rheb also shed further light on the mechanism of TSC2GAP-mediated hydrolysis of GTP. We propose that while providing an Asn thumb as a means of accelerating catalysis, TSC2GAP may also stimulate the GTPase activity of Rheb by relieving autoinhibition and aligning Rheb’s newly identified catalytic machinery (Figure 3.13c). Interaction of TSC2 with Rheb switch I may disrupt the electrostatic contact between Tyr35 and the γ-phosphate, reducing the auto-inhibitory effect of this residue on GTP hydrolysis, explaining the functional and thermodynamic similarities between WT Rheb in the presence of the TSC2GAP and the Rheb Y35A mutant alone. A common theme in GAP-stimulated GTP hydrolysis is repositioning of the N-terminus of switch II relative to the nucleotide binding site to allow for efficient catalysis. For example, the RanGAP Asn thumb serves to properly orient the catalytic Gln69 of Ran (Seewald et al. 2002; Bos et al. 2007). Similarly, GAP IP4BP has been proposed to promote Rap GTP hydrolysis by repositioning a non-canonical catalytic glutamine residue, Gln63, located two positions C-terminal to the position corresponding to Ras Gln61 (Sot et al. 2010). In the case of Rheb, we propose that Rheb-TSC2GAP interaction stabilizes Asp65 closer to the γ-

59 phosphate to catalyze GTP hydrolysis in synergy with the Asn thumb (Asn1643) provided by TSC2GAP. The delineation of the precise details of TSC2-mediated catalysis will require a more detailed structural analysis of the Rheb:TSC2GAP complex that is hindered by the transient nature of their interaction observed by us and others (Scrima et al. 2008; Marshall et al. 2009), preventing us from distinguishing impaired binding from impaired catalysis.

Figure 3.13 Schematic model of intrinsic and TSC2GAP-stimulated GTP hydrolysis by Rheb. a) In wild-type Rheb, Tyr35 displaces the hydrolytic water away from the backbone carbonyl of Thr38, reducing the polarization and reactivity of this water molecule. Further, Tyr35 reduces the electrostatic contribution of Asp65 to the GTP hydrolysis reaction by displacing this residue away from the nucleotide and the catalytic water. b) Mutation of Tyr35 allows Asp65 and Thr38 to assemble a more efficient catalytic site. c) Interaction of the GAP domain of TSC2 with Rheb may reduce the inhibitory effects of Tyr35, allowing Asp65 to promote catalysis in synergy with the Asn thumb (Asn1643) provided in trans by TSC2GAP.

The Tyr35-mediated auto-inhibition of Rheb’s GTPase reaction is necessary to maintain the appropriate level of activation of this small GTPase, and thus mTORC1 signaling, in response to growth factors. When growth factors are available, TSC2GAP activity becomes limiting, and Tyr35 inhibits GTP hydrolysis, resulting in mTORC1 up-regulation. In the absence of stimulation, Rheb Tyr35 is required for productive TSC2GAP-mediated acceleration of GTP hydrolysis to shut down mTORC1 signaling. Indeed, overexpression of Rheb Y35A substantially uncouples mTOR signaling from growth factors. In cells overexpressing wild-type Rheb, mTOR signaling was strongly responsive to the availability of serum and insulin, whereas this response was significantly dampened in cells overexpressing the Y35A mutant.

To our knowledge this is the first example of a distinct mechanism of intrinsic nucleotide hydrolysis within the Ras subfamily, which may be relevant to some other Ras superfamily GTPases, particularly those that lack the canonical catalytic Gln in switch II (e.g., Rap GTPases). This study provides a view into an unusual mechanism of GTP hydrolysis by Rheb and an intriguing autoinhibitory interaction that blocks this GTPase reaction.

3.5 Experimental Procedures: 3.5.1 Protein preparation : Mus musculus Rheb (residues 1-169) WT and mutants and Tuberous sclerosis 2 (TSC2) GAP domain (residues 1525-1742) were prepared according to previous protocols (Mazhab-Jafari et al. 2010). In brief, the proteins were expressed in Escherichia coli (BL21) using 14 15 o pGEX2T vector, grown in minimal media supplemented with either NH 4Cl or NH 4Cl at 15 C with 0.25 mM IPTG. Mutagenesis was performed with Quikchange site-directed mutagenesis. Small GTPase- proteins expressed in E. coli co-purified mainly as complexes with GDP nucleotide. Rheb and TSC2 proteins were initially purified using glutathione Sepharose, cleaved from the GST tag by thrombin, and further purified via Superdex 75 size exclusion chromatography. Human H-Ras (residues 1-171) and

murine RhoA (residues 1-181) were expressed using pET15b and pET28, respectively, and purified by Ni-NTA followed by Superdex 75 after removal of the His tag with thrombin.

3.5.2 Crystallization and data collection: Crystals of Rheb (residues 1-169) Y35A mutant were grown at room temperature with seeding using the hanging drop vapor diffusion method. The protein

solution contained 20mM Tris hydrochloride pH 8.0, 100 mM NaCl, 5 mM MgCl 2 and 0.02% NaN 3 w/v. The protein was concentrated to 870 µM, and GMPPNP was added to a final concentration of 3.8 mM. The solution was allowed to sit at room temperature for 4h for nucleotide exchange and then placed at 4 oC overnight. Crystallization experiments were set at room temperature by mixing equal

volumes of the protein solution with the well solution (100 mM Tris hydrochloride pH 8.5, 200 mM Sodium acetate trihydrate and 30% w/v Polyethylene glycol 4,000). Seed crystals (grown at 4 oC from a

360 µM protein solution containing 1.56 mM GMPPNP and a well solution of 100 mM Tris hydrochloride pH 8.5, 200 mM Sodium acetate trihydrate and 30% w/v Polyethylene glycol 4,000) were added in serial dilution to hanging drops to promote crystal growth. Resulting crystals were soaked in a cryoprotective solution containing 25% PEG 400 and flash frozen in liquid Nitrogen. The diffraction data were collected using an R-Axis IV++, Rigaku RUH3R rotating anode generator equipped with Osmic optics and an X-stream cryosystem for data collection at a temperature of 100K with a wavelength of 1.54 Å and processed using HKL2000 (Otwinowski and Minor 1997).

3.5.3 Structure determination and refinement: The phase problem was solved by molecular replacement method using the structure of WT Rheb (PDB code: 1XTR) as a search model. Successive rounds of refinements and manual model building were performed to construct the final model. Phenix was used for both phase determination and structure refinement (Adams et al. 2010), while manual model building was performed in Coot (Emsley and Cowtan 2004). During the course of refinement, density for the nucleotides could be clearly seen in difference electron density maps, which allowed us to manually position GMPPNP bound to one Rheb molecule and GDP bound to the other Rheb molecule in the asymmetric unit. The refinement statistic can be found in Table 3.2. The structure is released with the PDB ID code 3SEA.

Table 3.2 Data collection and refinement statistics.

Rheb-Y35A Data collection Space group P 2 2 1 21 Cell dimensions  a, b, c (Å) 57.2, 69.9, 79.2 α, β, γ  ( °) 90, 90, 90 Resolution (Å) 46.4-2.0 (2.07-2.0) * Rsym 9.3 (41.2) I / σI 20.7 (4.6) Completeness (%) 99.7 (100) Redundancy 7.0 (6.6)

Refinement Resolution (Å) 26.9-2.0 No. reflections 21742 Rwork / Rfree 16.2 / 21.4 No. atoms Protein 2766 Ligand / Mg 2+ ion 62 / 2 Water 252 B-factors Protein 26.3 Ligand / Mg 2+ ion 21.6 / 24.2 Water 30.3 R.m.s. deviations Bond lengths (Å) 0.007 Bond angles ( °) 1.14 Ramachandran statistics Most favorable regions (%) 96.7 Allowed regions (%) 3.3 Disallowed regions (%) 0 *Data set was collected from one crystal. *Values in parentheses are for highest-resolution shell.

3.5.4 NMR-based GTPase assays : Rheb WT and mutants were loaded with GTP or mantGTP by incubation with a 10-20-fold molar excess of nucleotide in the presence of 10mM EDTA. Full nucleotide loading was confirmed by collecting a sensitivity-enhanced 1H-15 N HSQC (Heteronuclear Single Quantum Coherence) spectrum, and the mixture was then passed through a de-salting column (PD MidiTrap TM G-25 (GE healthcare)) equilibrated with NMR buffer (20 mM sodium phosphate pH

7.0, 100 mM NaCl, 5 mM MgCl 2 1mM DTT and 5% D 2O), to produce a 1:1 complex of GTPase and the nucleotide for the kinetic measurements. Nucleotide loading of Ras was achieved by the same method,

however in the case of RhoA, 0.5 M urea was included to facilitate the EDTA-mediated nucleotide exchange and was removed by the de-salting column (Mazhab-Jafari et al. 2010).

NMR experiments were run on a Bruker AVANCE II 800 MHz spectrometer equipped with a 5mm TCI CryoProbe TM or a 600 MHz spectrometer equipped with TCI 1.7 mm MicroCryoProbe TM . Sensitivity enhanced 1H-15 N HSQCs with 4 scans (10 min) were run in succession for monitoring the GTP hydrolysis reactions at a temperature of 20 oC using GTPase concentrations of 0.1-0.3 mM. The spectra were then processed with NMRPipe (Delaglio et al. 1995) and GDP- and GTP-specific peak heights were analyzed via Gaussian line fitting using Sparky (Goddard and Kneller). Residues from switch I&II, P-loop, β3 & β4 and the α3 helix that exhibit distinct well-resolved resonances in each nucleotide-bound form were used as reporters of the reaction rates, as described previously (Marshall et al. 2009; Mazhab- Jafari et al. 2010). The fraction of GTPase protein in the GDP-bound state was calculated for each

reporter residue using the following equation: I GDP /(I GDP +I GTP ), where I represents intensity, and plotted

against time. In the case of RhoA, the decay of IGDP peaks were used in the rate calculation, as described previously (Gasmi-Seabrook et al. 2010; Mazhab-Jafari et al. 2010). Data fitting was done using PRISM (GraphPad software). To assay GAP-mediated nucleotide hydrolysis, TSC2GAP was added to GTP- loaded Rheb at a GAP to GTPase molar ratio of 1:2.

3.5.5 Thermodynamic measurements: All assays for thermodynamic measurements of WT, Y35A, Y35F, Y35A-D65A, and D65A Rheb were run on 600 MHz spectrometer at a protein

concentration of 0.5 mM in 20mM HEPES pH 8.0, 100mM NaCl, 5mM MgCl 2, 2mM DTT and 10%

D2O with 4 scans. The GTPase assays were run at four temperatures (287, 292, 296, and 301.5 K) and Arrhenius plots were constructed by plotting ln(k) as a function of 1/T (K), where k is the rate of GTP hydrolysis in sec -1. Activation energy (Ea) and activation entropy ( S‡) values were calculated from the slope and Y-intercept, respectively, as described previously. The activation enthalpy ( H‡ = Ea – RT) and the free energy of activation ( G‡ = H‡ - T S‡) were then calculated with T set to 298 K. The G‡ values are reported in kJ/mol.

3.5.6 Cell-based phosphorylation assay: All chemicals were purchased from Sigma unless stated otherwise. Antibodies against p70 S6 Kinase and p70 S6 Kinase phosphorylated at Thr389 were from Cell Signaling Technology and Anti-Myc 9E10 is described previously (Buerger et al. 2006). Murine

Rheb cDNA was obtained from Open Biosystems and subcloned into pcDNA3.1 myc-His. The Rheb mutations were generated by site-directed mutagenesis (Agilent).

HeLaBT cells were cultured in Dulbecco's modified Eagle's medium containing 10% foetal calf serum (DMEM/10%FCS) and transfected using the calcium phosphate precipitation method. 24h after transfection, the media was replaced with serum-free DMEM. 48h after transfection, cells were either further starved, or stimulated for 15 minutes or 6 h with DMEM/10% FCS and 10 µg/mL insulin. Cells were then lysed in CHAPS lysis buffer (40mM Hepes, pH 7.5, 0.3% CHAPS, 120mM NaCl, imM

EDTA, 10mM pyrophosphate, 50mM NaF, 1mM Na 2VO 3, 20mM β-glycerophosphate, and protease inhibitors). Total protein in lysates was quantified using Bradford reagent (Bio-Rad), and equalized. Samples were then boiled in SDS-PAGE sample buffer, separated by SDS-PAGE, and transferred to PVDF membranes. After blocking in 5% bovine serum albumin in TBS-T (50mM Tris-HCl, pH 7.5, 150mM NaCl and 0.1% Tween 20), membranes were probed with the indicated antibodies and visualized with HRP-conjugated secondary antibodies using the ECL system (Amersham).

The band intensities of p70 S6K P-Thr389, total S6K and Myc-Rheb for each lysate were measured using ImageJ software (NIH). Normalized p70 S6K P-Thr389 values were determined using the formula [(p70 S6K P-Thr389) / (total S6K)] / (Myc-Rheb) and the value for Rheb WT under growth factor stimulated condition was set to 100 for each time point after starvation/stimulation. The background intensity was used as estimation for the error in the intensity measurements and propagated appropriately throughout division.

3.5.7 Nucleotide binding in vivo: Labeling was done according to previous protocols (Wolthuis et al. 1997) with the following modification. HEK293 cells were seeded in a 6-well plate and transfected with 2.5ug of DNA using the calcium-phosphate method. The following morning, cells were washed and maintained in DMEM containing 10% FBS and 6 hours later, their media replaced with phosphate- and serum-free DMEM and the cells incubated for 1hr. Cells were labeled with 250uCi/ml of [ 32 P] orthophosphoric acid (Perkin Elmer) overnight. Cells were washed with cold PBS and lysed in 750ul of

50mM Tris pH 7.4, 140mM NaCl, 1mM KCl, 2mM MgCl 2, 1% Triton X-100, protease inhibitor cocktail. Lysates were precleared with 20ul of protein A sepharose for 20 minutes. Myc-Rheb was immunoprecipitated with 1ug of anti-myc antibody (Cell Signaling) for 2hrs and 20ul of protein A

sepharose for 30 minutes. Immunoprecipitates were washed 4 times in lysis buffer containing 500mM NaCl and twice in lysis buffer with 0.1% Triton X-100. Nucleotides were eluted in 20ul of 2mM EDTA, 2mM DTT, 0.2% SDS, 1mM GDP, 1mM GTP at 68 oC for 15 minutes. Eluted nucleotides were

separated by spotting 10ul on PEI cellulose TLC plates and resolved in 0.75M KH 2PO 4 pH 3.4. Air dried TLC plates were exposed using a phosphorimager and GTP/GDP ratios calculated using ImageQuant 5.2 software.

3.6 References:

Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW et al. 2010. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66 : 213-221. Allin C, Ahmadian MR, Wittinghofer A, Gerwert K. 2001. Monitoring the GAP catalyzed H-Ras GTPase reaction at atomic resolution in real time. Proc Natl Acad Sci U S A 98 : 7754-7759. Benvenuto G, Li S, Brown SJ, Braverman R, Vass WC, Cheadle JP, Halley DJ, Sampson JR, Wienecke R, DeClue JE. 2000. The tuberous sclerosis-1 (TSC1) gene product hamartin suppresses cell growth and augments the expression of the TSC2 product tuberin by inhibiting its ubiquitination. Oncogene 19 : 6306-6316. Bos JL, Rehmann H, Wittinghofer A. 2007. GEFs and GAPs: critical elements in the control of small G proteins. Cell 129 : 865-877. Brinkmann T, Daumke O, Herbrand U, Kuhlmann D, Stege P, Ahmadian MR, Wittinghofer A. 2002. Rap-specific GTPase activating protein follows an alternative mechanism. J Biol Chem 277 : 12525-12531. Brucker S, Gerwert K, Kotting C. 2010. Tyr39 of ran preserves the Ran.GTP gradient by inhibiting GTP hydrolysis. J Mol Biol 401 : 1-6. Buerger C, DeVries B, Stambolic V. 2006. Localization of Rheb to the endomembrane is critical for its signaling function. Biochem Biophys Res Commun 344 : 869-880. Buhrman G, Holzapfel G, Fetics S, Mattos C. 2010. Allosteric modulation of Ras positions Q61 for a direct role in catalysis. Proc Natl Acad Sci U S A 107 : 4931-4936. Cepus V, Scheidig AJ, Goody RS, Gerwert K. 1998. Time-resolved FTIR studies of the GTPase reaction of H-ras p21 reveal a key role for the beta-phosphate. Biochemistry 37 : 10263-10271. Cherfils J, Menetrey J, Le Bras G, Janoueix-Lerosey I, de Gunzburg J, Garel JR, Auzat I. 1997. Crystal structures of the small G protein Rap2A in complex with its substrate GTP, with GDP and with GTPgammaS. Embo J 16 : 5582-5591. Chong-Kopera H, Inoki K, Li Y, Zhu T, Garcia-Gonzalo FR, Rosa JL, Guan KL. 2006. TSC1 stabilizes TSC2 by inhibiting the interaction between TSC2 and the HERC1 ubiquitin ligase. J Biol Chem 281 : 8313-8316. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. 1995. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR 6: 277-293. Du X, Frei H, Kim SH. 2000. The mechanism of GTP hydrolysis by Ras probed by Fourier transform infrared spectroscopy. J Biol Chem 275 : 8492-8500.

Dunlop EA, Dodd KM, Seymour LA, Tee AR. 2009. Mammalian target of rapamycin complex 1- mediated phosphorylation of eukaryotic initiation factor 4E-binding protein 1 requires multiple protein-protein interactions for substrate recognition. Cell Signal 21 : 1073-1084. Emsley P, Cowtan K. 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60 : 2126-2132. Eom M, Han A, Yi SY, Shin JJ, Cui Y, Park KH. 2008. RHEB expression in fibroadenomas of the breast. Pathol Int 58 : 226-232. Focia PJ, Shepotinovskaya IV, Seidler JA, Freymann DM. 2004. Heterodimeric GTPase core of the SRP targeting complex. Science 303 : 373-377. Fraser JS, van den Bedem H, Samelson AJ, Lang PT, Holton JM, Echols N, Alber T. 2011. Accessing protein conformational ensembles using room-temperature X-ray crystallography. Proc Natl Acad Sci U S A 108 : 16247-16252. Frech M, Darden TA, Pedersen LG, Foley CK, Charifson PS, Anderson MW, Wittinghofer A. 1994. Role of glutamine-61 in the hydrolysis of GTP by p21H-ras: an experimental and theoretical study. Biochemistry 33 : 3237-3244. Garami A, Zwartkruis FJ, Nobukuni T, Joaquin M, Roccio M, Stocker H, Kozma SC, Hafen E, Bos JL, Thomas G. 2003. Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2. Mol Cell 11 : 1457-1466. Gasmi-Seabrook GM, Marshall CB, Cheung M, Kim B, Wang F, Jang YJ, Mak TW, Stambolic V, Ikura M. 2010. Real-time NMR study of guanine nucleotide exchange and activation of RhoA by PDZ-RhoGEF. J Biol Chem 285 : 5137-5145. Goddard TD, Kneller D. SPARKY 3. in University of California, San Francisco . Grant BJ, Gorfe AA, McCammon JA. 2009. Ras conformational switching: simulating nucleotide- dependent conformational transitions with accelerated molecular dynamics. PLoS Comput Biol 5: e1000325. Im E, von Lintig FC, Chen J, Zhuang S, Qui W, Chowdhury S, Worley PF, Boss GR, Pilz RB. 2002. Rheb is in a high activation state and inhibits B-Raf kinase in mammalian cells. Oncogene 21 : 6356-6365. Inoki K, Li Y, Xu T, Guan KL. 2003. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev 17 : 1829-1834. Ito Y, Yamasaki K, Iwahara J, Terada T, Kamiya A, Shirouzu M, Muto Y, Kawai G, Yokoyama S, Laue ED et al. 1997. Regional polysterism in the GTP-bound form of the human c-Ha-Ras protein. Biochemistry 36 : 9109-9119. Jiang H, Vogt PK. 2008. Constitutively active Rheb induces oncogenic transformation. Oncogene 27 : 5729-5740. Klink BU, Goody RS, Scheidig AJ. 2006. A newly designed microspectrofluorometer for kinetic studies on protein crystals in combination with x-ray diffraction. Biophys J 91 : 981-992. Kötting C, Gerwert K. 2004. Time-resolved FTIR studies provide activation free energy, activation enthalpy and activation entropy for GTPase reactions. Chemical Physics 307 : 227-232. Krengel U, Schlichting I, Scherer A, Schumann R, Frech M, John J, Kabsch W, Pai EF, Wittinghofer A. 1990. Three-dimensional structures of H-ras p21 mutants: molecular basis for their inability to function as signal switch molecules. Cell 62 : 539-548. Li G, Zhang XC. 2004. GTP hydrolysis mechanism of Ras-like GTPases. J Mol Biol 340 : 921-932. Li Y, Inoki K, Guan KL. 2004. Biochemical and functional characterizations of small GTPase Rheb and TSC2 GAP activity. Mol Cell Biol 24 : 7965-7975.

Long X, Lin Y, Ortiz-Vega S, Busch S, Avruch J. 2007. The Rheb switch 2 segment is critical for signaling to target of rapamycin complex 1. J Biol Chem 282 : 18542-18551. Maegley KA, Admiraal SJ, Herschlag D. 1996. Ras-catalyzed hydrolysis of GTP: a new perspective from model studies. Proc Natl Acad Sci U S A 93 : 8160-8166. Marshall CB, Ho J, Buerger C, Plevin MJ, Li GY, Li Z, Ikura M, Stambolic V. 2009. Characterization of the intrinsic and TSC2-GAP-regulated GTPase activity of Rheb by real-time NMR. Sci Signal 2: ra3. Mazhab-Jafari MT, Marshall CB, Smith M, Gasmi-Seabrook GM, Stambolic V, Rottapel R, Neel BG, Ikura M. 2010. Real-time NMR study of three small GTPases reveals that fluorescent 2'(3')-O- (N-methylanthraniloyl)-tagged nucleotides alter hydrolysis and exchange kinetics. J Biol Chem 285 : 5132-5136. Nardella C, Chen Z, Salmena L, Carracedo A, Alimonti A, Egia A, Carver B, Gerald W, Cordon-Cardo C, Pandolfi PP. 2008. Aberrant Rheb-mediated mTORC1 activation and Pten haploinsufficiency are cooperative oncogenic events. Genes Dev 22 : 2172-2177. Otwinowski Z, Minor W. 1997. Processing of X-ray Diffraction Data Collected in Oscillation Mode. Methods in Enzymology 276 : 307-326. Scheffzek K, Ahmadian MR, Kabsch W, Wiesmuller L, Lautwein A, Schmitz F, Wittinghofer A. 1997. The Ras-RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science 277 : 333-338. Scheidig AJ, Franken SM, Corrie JE, Reid GP, Wittinghofer A, Pai EF, Goody RS. 1995. X-ray crystal structure analysis of the catalytic domain of the oncogene product p21H-ras complexed with caged GTP and mant dGppNHp. J Mol Biol 253 : 132-150. Scrima A, Thomas C, Deaconescu D, Wittinghofer A. 2008. The Rap-RapGAP complex: GTP hydrolysis without catalytic glutamine and arginine residues. Embo J 27 : 1145-1153. Seewald MJ, Korner C, Wittinghofer A, Vetter IR. 2002. RanGAP mediates GTP hydrolysis without an arginine finger. Nature 415 : 662-666. Sot B, Kotting C, Deaconescu D, Suveyzdis Y, Gerwert K, Wittinghofer A. 2010. Unravelling the mechanism of dual-specificity GAPs. Embo J 29 : 1205-1214. Tee AR, Manning BD, Roux PP, Cantley LC, Blenis J. 2003. Tuberous sclerosis complex gene products, Tuberin and Hamartin, control mTOR signaling by acting as a GTPase-activating protein complex toward Rheb. Curr Biol 13 : 1259-1268. Wennerberg K, Rossman KL, Der CJ. 2005. The Ras superfamily at a glance. J Cell Sci 118 : 843-846. Wolthuis RM, de Ruiter ND, Cool RH, Bos JL. 1997. Stimulation of gene induction and cell growth by the Ras effector Rlf. Embo J 16 : 6748-6761. Yamagata K, Sanders LK, Kaufmann WE, Yee W, Barnes CA, Nathans D, Worley PF. 1994. rheb, a growth factor- and synaptic activity-regulated gene, encodes a novel Ras-related protein. J Biol Chem 269 : 16333-16339. Yamasaki K, Shirouzu M, Muto Y, Fujita-Yoshigaki J, Koide H, Ito Y, Kawai G, Hattori S, Yokoyama S, Nishimura S et al. 1994. Site-directed mutagenesis, fluorescence, and two-dimensional NMR studies on microenvironments of effector region aromatic residues of human c-Ha-Ras protein. Biochemistry 33 : 65-73. Yu Y, Li S, Xu X, Li Y, Guan K, Arnold E, Ding J. 2005. Structural basis for the unique biological function of small GTPase RHEB. J Biol Chem 280 : 17093-17100.

CHAPTER 4

Structure-guided design of a constitutively active mutant of Rheb

This chapter has been reformatted from the original publication: Mohammad T. Mazhab-Jafari, Christopher B. Marshall, Jason Ho, Noboru Ishiyama, Vuk Stambolic, and Mitsuhiko Ikura. Structure- guided mutation of the conserved G3-box glycine in Rheb generates a constitutively activated regulator of mTOR. J Biol Chem . In Press. doi: 10.1074/jbc.C113.543736

A link to the published paper can be found at: http://www.jbc.org/content/289/18/12195.long

4.1 Abstract: Constitutively activated variants of small GTPases, which provide valuable functional probes of their role in cellular signaling pathways, can often be generated by mutating the canonical catalytic residue (e.g., RasQ61L) to impair GTP hydrolysis. However, this general approach is ineffective for a substantial fraction of the small GTPase family in which this residue is not conserved (e.g., Rap) or not catalytic (e.g., Rheb). Using a novel engineering approach, we have manipulated nucleotide binding through structure-guided substitutions of an ultra-conserved glycine residue in the G3-box motif (DxxG). Substitution of Rheb Gly63 with alanine impaired both intrinsic and TSC2 GTPase activating protein (GAP)-mediated GTP hydrolysis by displacing the hydrolytic water molecule, while introduction of a bulkier valine sidechain selectively blocked GTP binding by steric occlusion of the γ-phosphate. Rheb G63A stimulated phosphorylation of the mTORC1 substrate p70S6 kinase more strongly than wild-type, thus offering a new tool for mTOR signaling.

4.2 Introduction: The Ras-superfamily comprises 167 members, which share a common fold and function as ‘switches’ to regulate diverse cellular signaling pathways (Wittinghofer and Vetter 2011). Small GTPases cycle between an activated guanosine triphosphate (GTP)-bound state that interacts with and activates downstream effector proteins to drive signaling, and an inactive guanosine diphosphate (GDP)-bound state. Thus GTPases are inactivated by hydrolysis of the γ-phosphate of GTP, and can be reactivated by exchange of GDP for a new molecule of GTP, processes that occur slowly, but are catalyzed by the upstream regulators GTPase activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), respectively. Mutations that impair GTP hydrolysis perturb this cycle and generate hyperactivated variants (e.g., Ras G12V and Q61L) that are often associated with disease processes. Nevertheless, constitutively activated and inactive GTPase mutants are indispensible research tools for probing the function of GTPases, and dissecting the signaling pathways they regulate. The strategy most widely used to generate activated GTPases employs point mutations of a solvent-exposed glutamine residue in the G3-box motif (Der et al. 1986) in switch II (five G-boxes comprise conserved sequence elements for nucleotide binding and effector recognition). This glutamine is conserved in 73% of human GTPases (Wennerberg et al. 2005) and has been demonstrated to be a key catalytic residue in several cases. Structural and enzymatic studies of Ras have shown that the carboxamide oxygen of this glutamine sidechain (Gln61) catalyzes GTP hydrolysis by increasing the nucleophilicity of a hydrolytic cat water molecule (H 2O ) positioned in-line with the γ-phosphate (Frech et al. 1994). Although this mechanism is conserved amongst many GTPases, this residue is substituted in one quarter of small GTPases, and in some Ras superfamily members, this glutamine is present but non-catalytic. These GTPases have lower catalytic activity, which may proceed through alternate mechanisms. In Ras homolog enriched in brain (Rheb), the G3-box glutamine residue (Gln64) is found in a non-catalytic conformation, and its mutation has no impact on intrinsic hydrolysis and only modestly reduces GAP- mediated hydrolysis (Li et al. 2004; Yu et al. 2005; Marshall et al. 2009). Mutations of Ras Gly12 impair GTP hydrolysis because the presence of a sidechain restricts the proper alignment of the catalytic residue, however Rheb already has a bulky residue (Arg15) in this position and its mutation has little effect on hydrolysis (Marshall et al. 2009). Thus, despite the modest impact of the Q64L mutation, it remains well used for lack of a more robust activated Rheb variant (Li et al. 2004; Jiang and Vogt 2008; Mavrakis et al. 2008; Zhou et al. 2009). In mutagenic analyses of residues potentially mediating Rheb’s non-canonical catalytic mechanism, mutation of Asp65 had the greatest impact on catalytic activity

(Mazhab-Jafari et al. 2012), however the reduction in intrinsic GTPase activity was limited to a modest ~30% (D65A). Tuberous sclerosis complex 2 (TSC2) possesses a GAP activity that accelerates the intrinsic GTP hydrolysis of wild-type Rheb (Inoki et al. 2003; Marshall et al. 2009). Mutation of Rheb Asp65 also impaired sensitivity to TSC2 GAP, but Rheb D65A did not fully activate mTORC1, possibly due to weakened interaction (Mazhab-Jafari et al. 2012). Therefore we sought to develop novel strategies to control the GTPase cycle of Rheb and other GTPase proteins that lack the canonical catalytic machinery.

Figure 4.1 Manipulation of the GTPase cycle of Rheb and mTOR signaling through substitutions of Gly63. a), Schematic illustration of the mTORC1 signaling pathway. b), Hydrolysis of GTP by wild- type Rheb and G63A in the presence and absence of the TSC2 GAP domain, monitored by real-time NMR. c), Intrinsic nucleotide exchange (GDP to GTP) of wild-type Rheb, G63A and G63V. Reaction rates are in min -1. (d-e) Rheb G63A mutation increases activation of mTORC1 signaling under starvation. (d) HeLa and (e) TSC2-deficient MEF cells were transfected with wild-type Rheb, G63A or empty vector, and starved or stimulated with insulin prior to lysis. Each experiment was performed in duplicate with one representative blot shown.

4.3 Results and Discussion: cat The hydrolytic H 2O has been observed in almost all high-resolution GTPase structures. It is coordinated by hydrogen bonds with the γ-phosphate of GTP and a common set of highly conserved residues including the backbone carbonyl of a threonine in the G2 box, and the backbone amide of a glycine in the G3 box. The threonine (Thr35 in Ras) is in the C-terminal part of switch I and is 90% identical amongst human GTPases (Wennerberg et al. 2005), whereas the glycine (Gly60 in Ras, 93% identity) (Wennerberg et al. 2005) is in the N-terminal part switch II. The curved-loop configuration of the G3-box DxxG motif (where x stands for any amino acid) allows the aspartic acid to coordinate a Mg 2+ ion required for nucleotide binding (John et al. 1993), and positions the C α of the glycine next to cat H2O and proximal to the γ-phosphate (Wittinghofer and Vetter 2011). On the basis of this structure, we hypothesized that by substituting various residues for glycine, we could introduce side-chain steric cat cat clashes to selectively perturb binding of the H 2O and/or the γ-phosphate. Displacing the H 2O would impair GTP hydrolysis, yielding an activated variant, whereas blocking γ-phosphate binding would yield an inactive mutant that can only accommodate GDP. We first applied this novel engineering approach to the mTOR activator Rheb (Figure 4.1a), examining how mutations of Gly63 affect its non-canonical GTPase cycle.

Real-time NMR-based GTPase assays (Marshall et al. 2009) demonstrated that mutation of Gly63 to alanine substantially reduced Rheb’s intrinsic GTP hydrolysis rate by ~4.5 fold, (Figure 4.1b), cat consistent with our notion that the introduction of a methyl side chain could destabilize the H 2O . Most remarkably, no TSC2 GAP activity could be detected for the Rheb-G63A mutant (Figure 4.1b), which cat was predicted since GAP-catalyzed hydrolysis utilizes the same H 2O , stimulating the reaction by stabilizing and complementing the intrinsic catalytic machinery. Finally, to investigate whether the mutation affects intrinsic nucleotide exchange of GDP for GTP, we carried out real-time NMR exchange

73 assays. Rheb-G63A exchanged >10 times faster than the wild-type protein (Figure 4.1c), and although this effect was not anticipated a priori , it would further drive this mutant towards the activated state.

Figure 4.2 Conservation of the G3-box glycine and structural basis for the functional properties of Rheb Gly63Ala/Val mutants. a) Alignment of the G3-boxes of selected small GTPase protein sequences illustrating conservation of the glycine residue. GTPases lacking the catalytic glutamine residue were selected. b-c) Overall fold of b) wild-type Rheb-GTP, c) Rheb G63A-GTP, and d) Rheb G63V-GDP, with each nucleotide binding site enlarged (right). The nucleotide is represented with

spheres (left) or balls-and-sticks (right) and the glycine, alanine, and valine are shown in yellow. For clarity Tyr35 is not shown.

Conversely, mutation of Gly63 to valine resulted in a Rheb variant that failed to bind GTP (Figure 4.1c), consistent with steric interference between the bulkier side chain and the γ-phosphate. As expected, Rheb G63V remained fully GDP-bound and inactive. Therefore, we have identified a single, highly conserved residue whose mutation can result in two diametrically opposed effects on activation.

Figure 4.3 Electron densities of the catalytic sites of wild-type Rheb and its Gly63 mutations. 2Fo- Fc electron density maps of the nucleotide and residue 63 at 1.0 σ are shown for a) WT Rheb-GTP b), G63A Rheb-GTP and c), G63V Rheb-GDP. The 2Fo-Fc electron density map of the hydrolytic water is shown at 1.0 σ in a, but is absent in b and c. The predicted position of the γ-phosphate, based on global alignment of WT Rheb-GTP with G63V Rheb-GDP, is indicated in C (semi-transparent). The insets in each panel are chemical schematics illustrating the relative positions of the γ-phosphate, the G3-box glycine (the mutated alanine and valine), the hydrolytic water, and the nucleophilic GTPase reaction.

Rheb activates the mammalian target of rapamycin complex 1 (mTORC1), which regulates protein biosynthesis, cell growth, proliferation and autophagy (Saucedo et al. 2003; Stocker et al. 2003; Kim et al. 2011). Consistent with our biochemical data (Figure 4.1b), overexpression of Rheb G63A in HeLa cells stimulated phosphorylation of the mTORC1 kinase substrate p70 S6K more strongly than wild- type Rheb in the absence of serum, when TSC2 is active (Figure 4.1d). Upon serum stimulation, which inactivates TSC2, p70 S6K is highly phosphorylated and its phosphorylation cannot be further enhanced by expression of wild-type or G63A Rheb, suggesting that under this condition, p70 S6K phosphorylation is limited by mTORC1 rather than the availability of endogenous Rheb-GTP (Figure 4.1d). Similarly, high level of p70 S6K phosphorylation in TSC2-null MEFs cannot be further enhanced by expression of wild-type or G63A Rheb (Figure 4.1e). Although Rheb G63V expressed well in E. coli, in mammalian cells its expression was ~90% lower than the wild type, thus we could not reliably probe the effects of this mutation on mTORC1 signaling. Reduced expression of nucleotide-free Rheb mutants has been reported previously (Li et al. 2004). The inability to bind GTP, which is much more abundant than GDP in mammalian cells (Traut 1994), might explain the poor expression of G63V, despite the stability of its GDP-bound form.

To validate that these mutations of the ultra-conserved G3-box glycine (Figure 4.2a) exert their effects through the anticipated mechanisms, we solved the crystal structures of wild-type Rheb and the G63A mutant bound to GTP, and the Rheb G63V mutant bound to GDP (Figure 4.2b-d and 4.3a-c). The overall fold was not substantially affected by the mutations (all atoms RMSD WT-GTP vs . G63A-GTP = 0.23 Å and WT-GDP(Yu et al. 2005) vs . G63V-GDP = 0.74 Å), however, the electron density of the cat nucleotide-binding pocket confirmed the positioning of the Ala63 side chain in the H 2O -γ-phosphate cat binding cavity of the mutant (Figure 4.2c and 4.3b). Whereas the electron density of the H 2O is clearly visible in the WT structure, no corresponding density is observed in the G63A mutant structure (Figure cat 4.3a and b). Hence we hypothesize that the reduced residency of the hydrolytic H 2O decreases the

ability of the G63A mutant to hydrolyze GTP, in full agreement with the biochemical observations. In the structure of Rheb G63V in complex with GDP, the bulkier valine side chain occupies additional space within the nucleotide-binding pocket. Val63 Cγ was positioned less than 2 Å from the γ-phosphate location determined in the GTP-bound wild-type structure (Figure 4.2d and 4.3c). Thus enthalpic costs and steric clashes prevent the γ-phosphate from approaching the methyl protons, explaining the lack of GTP binding and nucleotide exchange for Rheb G63V.

Previously, an analogous substitution (G60A) in Ras was biochemically and structurally characterized. This G60A mutation reduced the intrinsic GTP hydrolysis rate of Ras as well, however this was mediated by a distinct mechanism involving substantial distortions of both switch regions that displace the catalytic glutamine away from the nucleotide binding site by ~5-8 Ǻ (Ford et al. 2005). This conformation resembled that of nucleotide-free Ras in complex with its exchange factor Son of Sevenless (SOS) (Sung et al. 1996; Boriack-Sjodin et al. 1998; Ford et al. 2005), and consequently Ras G60A was shown to exert a dominant negative effect in cells by forming a stable non-productive complex with SOS that sequesters this GEF (Sung et al. 1996; Ford et al. 2005). In contrast, the switch I and II regions of Rheb retain the wild-type conformation in the G63A mutant structure (all atoms RMSD of 0.22 and 0.35 Ǻ for switch I and II, respectively). Thus the in vivo phenotypes of G3-box glycine mutants are GTPase specific, and must be characterized. Nevertheless, the ability to generate gain-of- function mutants for some non-canonical GTPases is especially valuable for developing probes. We therefore investigated the effect of substituting Gly60 in Rap1A (37% sequence identity with Rheb), a small GTPase for which no mutation impairing GTP hydrolysis is available because the canonical catalytic glutamine is replaced by a threonine. The G60A mutation reduced the GTP hydrolysis rate of Rap1A by 3 fold (Figure 4.4a), which is remarkable considering that no other mutation has been reported to inactivate Rap GTPase activity. In the case of Rap1A, however, the G60V mutation did not perturb nucleotide exchange activity (Figure 4.4b), suggesting that the valine side chain adopts a different conformation from that of Rheb G63V. Hence the structural and functional properties of G3- box Glycine mutations should be investigated for each GTPase. Intriguingly, a few examples of G3-box Gly residue mutations have been listed in the catalogue of somatic mutations in cancer (COSMIC) database (K-Ras G60A/V/D, H-Ras G60S, and Rheb G63W). While these mutations were identified in colon, thyroid and hematopoietic cancers, it remains to be determined whether they are oncogenic ‘driver’ mutations.

Herein this report, we have successfully demonstrated a structure-guided approach whereby mutating a single residue in the conserved G3 box can control nucleotide binding thereby locking Rheb in an activated (G63A-GTP) or an inactive (G63V-GDP) conformation. In serum-starved HeLa cells, the G63A mutant strongly activated the downstream effector mTORC1, thus enhancing p70 S6K phosphorylation, whereas the inability to bind GTP rendered G63V unstable in these cells. The crystal structures of the two Rheb mutants provided the structural basis for the unique features of each mutant. To our knowledge this is the first report of a mutation that severely impairs GTP hydrolysis of GTPases with non-canonical catalytic mechanisms. This diverse group of GTPases includes several members from four of the five GTPase subfamilies: Ras, Rab, Rho, and Arf (Figure 4.2a), thus this strategy should be a valuable biological tool for probing the functions of these unique proteins.

Figure 4.4 Intrinsic nucleotide hydrolysis and exchange of Rap1A GTPase and its Gly60 mutations. a) GTP hydrolysis by wild-type Rap1A (black) and the G60A mutant (green). b) GDP to

GTP exchange monitored for wild-type Rap1A (black), and the G60A (green) and G60V (blue) mutants. Reaction rates derived by curve fitting are shown with the same color codes (in min -1).

4.4 Experimental Procedures: 4.4.1 Protein Preparation: Mus musculus Rheb (residues 1–169) WT and G63A and G63V mutants and TSC2 GAP domain (residues 1,525–1,742) were expressed as glutathione-S-transferase (GST) fusions from the pGEX2T in Escherichia coli (BL21 strain) vector as described previously (Marshall et al. 14 15 2012). Briefly, bacteria were grown in minimal M9 media supplemented with either NH 4Cl or NH 4Cl o o at 37 C to OD 600 = 0.6, and protein expression was initiated with 0.25 mM IPTG at 15 C. Mutations were introduced using QuikChange site-directed mutagenesis (Stratagene). The proteins were affinity purified using glutathione Sepharose, cleaved from the GST tag by thrombin, and further purified using Superdex 75 (GE Healthcare) size exclusion chromatography. Human Rap1A (residues 1–167) was expressed using pET28 vector and purified using Ni-NTA resin, followed by removal of the His tag with thrombin and further purification by Superdex 75 size exclusion chromatography. Small GTPase proteins expressed in E. coli co-purified with guanine nucleotide.

4.4.2 NMR-based Real-time GTPase assay: For GTP hydrolysis assays, the small GTPase proteins were loaded with GTP in the presence of EDTA and 10-fold excess GTP, and following addition of MgCl 2 excess reagents were removed by passage through a desalting column (PD MidiTrap G-25; GE

Healthcare) equilibrated with NMR buffer (20 mM Tris pH 7.4, 100 mM NaCl, 5 mM MgCl 2, and 1 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP)). The final protein concentration used in NMR experiments was 0.3 mM. The proteins were fully GTP loaded at the start of each assay, as judged by 1H-15 N heteronuclear single-quantum coherence (HSQC) spectrum. Prior to nucleotide exchange assays, the GTPase proteins were incubated at room temperature for a minimum of a week and monitored by 1H-15 N HSQC spectra to ensure that E. coli-derived GTP had hydrolyzed, and then exchange reactions were initiated by addition of 1.5 mM GTP (5-fold molar excess). For GAP assays, TSC2 GAP domain was added to the GTP-loaded Rheb (WT or mutant) at a molar ratio of 1:2. All NMR spectra were acquired on Bruker AVANCE II 800 MHz spectrometer equipped with a 5 mm TCI CryoProbe or a 600 MHz spectrometer with a TCI 1.7 mm MicroCryoProbe. 1H-15 N HSQC spectra with sensitivity enhancement were collected in succession with 4 scans (10 min acquisition time) to monitor the GTPase reaction at 293.2 K. Spectral processing was carried out with NMRPipe(Delaglio et

al. 1995) and the nucleotide sensitive peaks were integrated via Gaussian line fitting using SPARKY (Goddard and Kneller). Residues exhibiting distinct and well-resolved peaks for each nucleotide-bound state were monitored as reporters of the reaction rates, as described previously (Marshall et al. 2009). For each pair of reporter resonances, the fraction of protein in the GDP-bound state was calculated as

IGDP /(I GDP +I GTP ), where I represents intensity, plotted against time and fit to one-phase exponential decay (instead of equation 4 of appendix A, since Kex >> K hyd ) or association functions for nucleotide exchange and hydrolysis, respectively. The error was estimated as described previously (Marshall et al. 2009). Because resonance assignments are not available for Rap1A GTPase the fraction of protein in the avg avg avg avg avg GDP-bound state was calculated as I GDP /(I GDP +I GTP ), where I GDP and I GTP represent the average intensities of GDP- and GTP-specific peaks. The error was then estimated from spectral noise and propagated accordingly. Data fitting was performed with PRISM (GraphPad software).

4.4.3 Crystallization and Data Collection: Crystals of Rheb (residues 1–169) WT, G63A, and G63V mutants were grown with seeding using the hanging drop vapor diffusion method at room temperature. The protein solution contained 0.8 mM protein, 20 mM Tris hydrochloride (Tris-HCl) [pH 8.0], 200 mM NaCl, 5 mM MgCl2, and 0.02% NaN3 w/v. The protein solution was mixed with an equal volume of the well solution (100 mM Tris-HCl [pH 8.5], 200 mM sodium acetate trihydrate, and 30% w/v Polyethylene Glycol 4000). Crystals of Rheb WT and G63A mutant were soaked in a solution containing 20 mM GTP and 25% Polyethylene Glycol 400 for 2 hours at room temperature to allow nucleotide exchange whereas G63V crystals were soaked in 25% Polyethylene Glycol 400 alone. Crystals were then flash frozen in liquid nitrogen. Diffraction data (WT and G63A) were collected at 100K on a Rigaku FR-E super-bright rotating anode generator equipped with Rigaku Saturn A200 CCD detector, Osmic VariMAx HF optics and an Oxford cryosystem with a wavelength of 1.54 Ǻ at Structural Genomics Consortium, Toronto, and processed with HKL2000 (Otwinowski and Minor 1997). Rheb WT-GTP and G63A-GTP diffracted to a resolution of 2.2 and 2.40 Ǻ, respectively. The diffraction data for Rheb G63V-GDP was collected at 100K with Cu K α radiation on a Bruker Microstar X8 Proteum SMART CCD system with a wavelength of 1.54 Ǻ to a resolution of 2.25 Ǻ, and processed with PROTEUM suite of programs.

4.4.4 Structure Determination and Refinement: The structures of GTP-bound murine Rheb-WT and – G63A were solved by molecular replacement using the structure of human Rheb-WT in GTP-bound

conformation [PDB: 1XTS] as the initial search model, whereas the structure of murine Rheb-G63V in complex with GDP was solved using the structure of human Rheb in GDP-bound conformation [PDB: 1XTQ]. The final models were generated after successive rounds of refinements using PHENIX (Adams et al. 2010) accompanied by manual model building with Coot (Emsley and Cowtan 2004). During the course of refinement, densities for the nucleotides and the hydrolytic water molecules (in the case of Rheb-WT) were clearly visible in difference electron density maps, which enabled us to manually position the molecules. WT and mutant proteins were crystallized with two Rheb molecules in the asymmetric unit, with one of the molecule’s switch I and II regions being involved in the crystal packing. All the structures discussed in the manuscript are from the Rheb molecules where the switch I and II are not involved in crystal packing, except for the GDP-bound Rheb-G63V, where the electron density of switch II was weak in the absence of crystal contact but well resolved upon crystal contact formation. It worth to note that switch II in GDP-bound Rheb-G63V adopted similar conformation as to the GDP-bound Rheb-WT (PDB: 1XTQ). The data collection and refinement statistics can be found in table 4.1. The structures has been deposited in the protein data bank with the following PDB codes: 4O25, 4O2L, and 4O2R for Rheb WT, G63A, and G63V, respectively.

4.4.5 Cell-based Phosphorylation Assays: Antibodies against p70 S6 Kinase and its Thr389- phosphorylated form were from Cell Signaling Technology and the M2-FLAG antibody was from Sigma. Murine Rheb cDNA (Open Biosystems) was subcloned into FLAG-tagged pcDNA5. The mutant Rheb construct was generated via site-directed mutagenesis (Agilent). Myc-p70 S6K was cloned in to pcDNA3.1. The p70 S6K phosphorylation assays were performed according to a previously described protocol (Mazhab-Jafari et al. 2012) with the following modifications. In brief, 3 x 10 5 HeLa H1 and 2.7 x 10 6 TSC2 -/- MEF cells, cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (DMEM/10% FCS), were co-transfected with FLAG-Rheb-WT or –G63A and myc-S6K at a ratio of 20:1 using Effectene (Qiagen) and 5:1 using GenJet reagent (SignaGen) for HeLa H1 and MEF cells, respectively. Twenty-four hours post-transfection, the media were exchanged with serum-free DMEM. Forty-eight hours post-transfection, cells were either further starved or stimulated with 100 nM insulin for 15 minutes prior to harvesting. FLAG-Rheb and myc-S6K were immunoprecipitated from normalized lysates with anti-Flag and anti-myc antibodies and Protein A/G Mix PureProteome magnetic beads (Millipore) for western blot analysis.

Table 4.1 Data collection and refinement statistics. Statistic information regarding the crystal structures of Rheb WT and G63A mutant in GTP bound form and G63V mutant in GDP bound form is presented.

Rheb-WT-GTP Rheb-G63A-GTP Rheb-G63V-GDP Data collection Space group P 2 2 1 21 P 2 2 1 21 P 21 21 2 Cell dimensions  a, b, c (Å) 57.7, 70.5, 79.8 57.6, 70.2, 79.7 70.3, 79.2, 57.3 α, β, γ  ( °) 90, 90, 90 90, 90, 90 90, 90, 90 Resolution (Å) 50.0-2.20 (2.24-2.20)* 50.0-2.40 (2.44-2.40)* 50-2.25 (2.35-2.25)* Rsym 0.127 (0.549) 0.155 (0.574) 0.073 (0.344) I / σI 15 (4.1) 14.9 (3.8) 12.3 (3.4) Completeness (%) 99.0 (98.6) 100 (100) 99.9 (100) Redundancy 7.0 (7.1) 7.7 (7.8) 13 (8.0)

Refinement Resolution (Å) 39.9-2.2 38.9-2.4 28.7-2.25 No. reflections 15674 12371 14672 Rwork / Rfree 20.3 / 25.6 19.5 / 24.3 20.6 / 25.9 No. atoms Protein 2738 2736 2742 Ligand / Mg 2+ ion 62 / 2 62 / 2 54 / 2 Water 130 119 125 B-factors Protein 25.2 30.7 21.9 Ligand/ion 26.4 33.3 17.8 Water 28.1 34.3 23.1 R.m.s. deviations Bond lengths (Å) 1.04 1.11 1.13 Bond angles ( °) 0.007 0.008 0.008 Ramachandran statistics Most Favorable region (%) 94.8 95.1 94.6 Allowed region (%) 5.2 4.9 5.4 Disallowed region (%) 0 0 0 *Data set was collected from one crystal. *Values in parentheses are for highest-resolution shell.

4.5 References:

Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW et al. 2010. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66 : 213-221. Boriack-Sjodin PA, Margarit SM, Bar-Sagi D, Kuriyan J. 1998. The structural basis of the activation of Ras by Sos. Nature 394 : 337-343. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. 1995. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR 6: 277-293. Der CJ, Finkel T, Cooper GM. 1986. Biological and biochemical properties of human rasH genes mutated at codon 61. Cell 44 : 167-176. Emsley P, Cowtan K. 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60 : 2126-2132. Ford B, Skowronek K, Boykevisch S, Bar-Sagi D, Nassar N. 2005. Structure of the G60A mutant of Ras: implications for the dominant negative effect. J Biol Chem 280 : 25697-25705. Frech M, Darden TA, Pedersen LG, Foley CK, Charifson PS, Anderson MW, Wittinghofer A. 1994. Role of glutamine-61 in the hydrolysis of GTP by p21H-ras: an experimental and theoretical study. Biochemistry 33 : 3237-3244. Goddard TD, Kneller D. SPARKY 3. in University of California, San Francisco . Inoki K, Li Y, Xu T, Guan KL. 2003. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev 17 : 1829-1834. Jiang H, Vogt PK. 2008. Constitutively active Rheb induces oncogenic transformation. Oncogene 27 : 5729-5740. John J, Rensland H, Schlichting I, Vetter I, Borasio GD, Goody RS, Wittinghofer A. 1993. Kinetic and structural analysis of the Mg(2+)-binding site of the guanine nucleotide-binding protein p21H- ras. J Biol Chem 268 : 923-929. Kim J, Kundu M, Viollet B, Guan KL. 2011. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nat Cell Biol 13 : 132-141. Li Y, Inoki K, Guan KL. 2004. Biochemical and functional characterizations of small GTPase Rheb and TSC2 GAP activity. Mol Cell Biol 24 : 7965-7975. Marshall CB, Ho J, Buerger C, Plevin MJ, Li GY, Li Z, Ikura M, Stambolic V. 2009. Characterization of the intrinsic and TSC2-GAP-regulated GTPase activity of Rheb by real-time NMR. Sci Signal 2: ra3. Marshall CB, Meiri D, Smith MJ, Mazhab-Jafari MT, Gasmi-Seabrook GM, Rottapel R, Stambolic V, Ikura M. 2012. Probing the GTPase cycle with real-time NMR: GAP and GEF activities in cell extracts. Methods 57 : 473-485. Mavrakis KJ, Zhu H, Silva RL, Mills JR, Teruya-Feldstein J, Lowe SW, Tam W, Pelletier J, Wendel HG. 2008. Tumorigenic activity and therapeutic inhibition of Rheb GTPase. Genes Dev 22 : 2178-2188. Mazhab-Jafari MT, Marshall CB, Ishiyama N, Ho J, Di Palma V, Stambolic V, Ikura M. 2012. An autoinhibited noncanonical mechanism of GTP hydrolysis by Rheb maintains mTORC1 homeostasis. Structure 20 : 1528-1539. Otwinowski Z, Minor W. 1997. Processing of X-ray Diffraction Data Collected in Oscillation Mode. Methods in Enzymology 276 : 307-326. Saucedo LJ, Gao X, Chiarelli DA, Li L, Pan D, Edgar BA. 2003. Rheb promotes cell growth as a component of the insulin/TOR signalling network. Nat Cell Biol 5: 566-571.

Stocker H, Radimerski T, Schindelholz B, Wittwer F, Belawat P, Daram P, Breuer S, Thomas G, Hafen E. 2003. Rheb is an essential regulator of S6K in controlling cell growth in Drosophila. Nat Cell Biol 5: 559-565. Sung YJ, Hwang MC, Hwang YW. 1996. The dominant negative effects of H-Ras harboring a Gly to Ala mutation at position 60. J Biol Chem 271 : 30537-30543. Traut TW. 1994. Physiological concentrations of purines and pyrimidines. Mol Cell Biochem 140 : 1-22. Wennerberg K, Rossman KL, Der CJ. 2005. The Ras superfamily at a glance. J Cell Sci 118 : 843-846. Wittinghofer A, Vetter IR. 2011. Structure-function relationships of the G domain, a canonical switch motif. Annu Rev Biochem 80 : 943-971. Yu Y, Li S, Xu X, Li Y, Guan K, Arnold E, Ding J. 2005. Structural basis for the unique biological function of small GTPase RHEB. J Biol Chem 280 : 17093-17100. Zhou X, Ikenoue T, Chen X, Li L, Inoki K, Guan KL. 2009. Rheb controls misfolded protein metabolism by inhibiting aggresome formation and autophagy. Proc Natl Acad Sci U S A 106 : 8923-8928.

CHAPTER 5

Structural and Functional consequences of Rheb-membrane interaction

This chapter has been reformatted from the original publication: Mazhab-Jafari MT, Marshall CB, Stathopulos PB, Kobashigawa Y, Stambolic V, Kay LE, Inagaki F, Ikura M. Membrane-dependent modulation of the mTOR activator Rheb: NMR observations of a GTPase tethered to a lipid-bilayer nanodisc. J Am Chem Soc. 2013 Mar 6;135(9):3367-70.

A link to the published paper can be found at: http://pubs.acs.org/doi/abs/10.1021/ja312508w

5.1 Abstract: Like most Ras superfamily proteins, the GTPase domain of Ras homolog enriched in brain (Rheb) is tethered to cellular membranes through a prenylated cysteine in a flexible C-terminal region, however little is known about how Rheb or other GTPases interact with the membrane or how this environment may affect their GTPase functions. We used NMR methods to characterize Rheb tethered to nanodiscs, monodisperse protein-encapsulated lipid bilayers with a diameter of 10 nm. Membrane conjugation markedly reduced the rate of intrinsic nucleotide exchange, while GTP hydrolysis was unchanged. NMR measurements revealed that the GTPase domain interacts transiently with the surface of the bilayer in two distinct preferred orientations, which are determined by the bound nucleotide. We propose models of membrane-dependent signal regulation by Rheb that shed light on previously unexplained in vivo properties of this GTPase. The study presented provides a general approach for direct experimental investigation of membrane-dependent properties of other Ras-superfamily GTPases.

5.2 Introduction: Most Ras superfamily small GTPase proteins are targeted by prenylation to cellular membranes, where they function as switches in a variety of signaling networks. Following prenylation of a CaaX-box Cysteine residue (C; a is an aliphatic residue; X is the C-terminal amino acid) in the C-terminal hypervariable region (HVR) (Brunsveld et al. 2009), the aaX peptide is cleaved by C-terminal prenylprotein peptidases, and Cys is carboxymethylated. Blockage of these post-translational modifications prevents GTPase-mediated signal transduction, highlighting the critical role of membrane localization. Hence, farnesyltransferase inhibitors have been developed to inhibit the biological activity of oncogenic Ras mutants (Berndt et al. 2011). While fluorescence-studies and molecular dynamic (MD) simulations have provided insight into small GTPase interactions with bilayer membranes (Gorfe et al. 2007a; Gorfe et al. 2007b; Abankwa et al. 2008b; Gureasko et al. 2008), atomic scale pictures of small GTPases on membranes are lacking due to the inherent challenges imposed by bilayer-membrane systems to high-resolution structural biology techniques.

Recent advances in the assembly of stable nano-scale bilayer membranes have made it possible to study structure, dynamics, and functions of peripheral and membrane-integrated proteins at the atomic level. Here, we used nanodiscs (Ritchie TK 2009), which are 5 by 10 nm discoidal lipoprotein complexes comprised of a bilayer of 120-160 lipid molecules bounded and stabilized by two copies of an engineered variant of Apo-lipoprotein A(Denisov et al. 2004), to enable nuclear magnetic resonance (NMR)-based characterization of the structure, dynamics, and function of membrane-tethered Rheb. Rheb is targeted to endomembrane compartments including Golgi, endoplasmic reticulum and lysosomes via farnesylation of its sole Cys residue (Cys181), processing that is required for its activation of mTORC1 (Buerger et al. 2006). To mimic processed Rheb, a truncated construct (residues 1-181) was covalently linked through Cys181 to a maleimide-conjugated lipid (Gureasko et al. 2008) incorporated into assembled nanodiscs (Figure 5.1a&b) or vesicles, a strategy previously described in studies of H- Ras on unilamellar vesicles (Gureasko et al. 2008). Relative to vesicles, for which diameters of 20-500 nm are typical, the smaller size of nanodiscs results in rapid tumbling, enabling high-resolution NMR studies of membrane-anchored Rheb and NMR-based real-time assays (Marshall et al. 2009) of the GTPase cycle of Rheb on the bilayer membrane.

Figure 5.1 Tethering Rheb to a bilayer membrane inhibits nucleotide exchange and activation. a) Lipids used for nanodisc assembly (full names in text). Reaction between Rheb Cys181 and the maleimide-moiety of PE-MCC was used to tether Rheb to the nanodisc bilayer. Drawn with ChemSketch (Advanced Chemistry Development, Inc.) b) Schematic diagram of the Rheb-nanodisc complex. c,d) Nucleotide hydrolysis and exchange reactions were monitored using NMR-based real- time GTPase assays of Rheb free in solution or tethered to nanodisc membrane bilayers . c) Intrinsic GTP hydrolysis by free and nanodisc-bound Rheb are shown in black and green, respectively. GTP hydrolysis by free and nanodisc-bound Rheb in the presence of extract of HEK-293 cells overexpressing TSC1/TSC2 are shown in red and blue, respectively. d) Intrinsic nucleotide exchange for free and

nanodisc-bound Rheb are shown in black and green, respectively. I(GDP) and I(GTP) are intensities of several peaks in HSQC spectra that are specific to Rheb-GDP and -GTP, respectively. The higher plateau of the green curve reflects the greater impact of GTP hydrolysis when nucleotide exchange is slow.

5.3 Results and Discussion: Tethering Rheb to the nanodisc did not alter the rate of intrinsic GTP hydrolysis or the GTPase activating protein (GAP) TSC2-catalyzed GTP hydrolysis rates by Rheb (Figure 5.1c); however, the intrinsic nucleotide exchange rate was significantly reduced upon membrane association (Figure 5.1d), indicating that membrane conjugation directly impacts the GTPase cycle of Rheb. To investigate the structural and dynamical properties of the interactions between the membrane and the GTPase domain of tethered Rheb, we used chemical shift perturbation (CSP) analysis, 15 N R1, 15 N R2, steady state 1H- 15 N NOE backbone relaxation measurements, and paramagnetic relaxation enhancement (PRE) experiments.

Conjugation of Rheb to the membrane did not detectably perturb the chemical shifts of resonances in the 1H-15 N HSQC spectrum, with the exception of Cys181 (Figure 5.2), suggesting that the GTPase domain (1-169) and the HVR (170-180) of Rheb do not tightly associate with the bilayer. The overall rotational 15 15 1 15 correlation time ( τM) was calculated from N R1, N R2, and steady state H- N NOE (Farrow NA 1995) for GDP- and GMPPNP- (non-hydrolyzable analog of GTP) bound Rheb conjugated to the

nanodisc and free in solution. Overall tumbling-times ( τM) for Rheb-GDP were calculated to be 17.0±0.2 ns in solution and 35.4±1.3 ns when conjugated to the nanodiscs (20 °C). A protein rigidly associated

with the >100 kDa Rheb-nanodisc complex would have an expected τM >60 ns (based on Stoke’s law

assuming a spherical protein-nanodisc complex), thus the intermediate τM value indicates that the GTPase domain of Rheb-GDP exhibits a high degree of freedom (dynamics) despite being tethered to

the membrane. The τM values for Rheb-GMPPNP in solution and conjugated to nanodiscs were 17.0±0.5 ns and 32.6±1.4 ns, respectively, similar to what was observed for Rheb-GDP. These results further establish the transient nature of Rheb’s interactions with the membrane.

The most significant local changes in dynamics upon membrane binding were at the C-terminal HVR, where the backbone ps-ns motions were partially quenched, independent of the bound nucleotide (Figure 5.3 and 5.4). Within the GTPase domain, the only significant change in local dynamics upon membrane conjugation was an increase in the s-ms time-scale motions in the α3 helix (R2 and R1×R2 of residues

Figure 5.2 Preparation of Rheb-nanodisc complex. a) Superdex S200 chromatograms of nanodiscs before and after conjugation with Rheb. SDS-PAGE analysis of the fraction containing Rheb-tethered nanodiscs (*) indicates a ~1:1 ratio of Rheb to MSP. b) 15 N1H-HSQC spectra of 0.3 mM free (black) and 0.6 mM nanodisc-conjugated (red) Rheb-GDP. The only peak (Cys181) exhibiting appreciable chemical shift change upon membrane conjugation is shown with a dashed rectangle. Resonances with low S/N in free Rheb that are broadened beyond detection upon membrane conjugation are shown with dashed circles. c) Linewidths of free (black) versus nanodisc-conjugated (red) Rheb. The region illustrated is indicated by a solid rectangle in b and is shown at the same contour level in c.

90-95) of Rheb-GDP, which was not observed for the GMPPNP-bound form (Figure 5.3&5.4). The α3 helix forms ionic contacts with switch II in Rheb, and in the homolog H-Ras, α3 is allosterically coupled to the nucleotide binding site (NBS) (Fraser et al. 2011), thus the effect of membrane conjugation on the rate of nucleotide exchange may involve modulation of the NBS by perturbation of α3 dynamics. Another factor that might impair nucleotide exchange is occlusion of the NBS by the membrane surface, thus we explored this possibility more fully by performing PRE experiments as detailed below.

The PRE effect can be powerful for identifying transient interactions, such as those between Rheb and the bilayer plane, and for determining “preferred conformations” of a protein such as membrane- associated Rheb. Rheb-tethered nanodiscs were prepared with the incorporation of lipids (5%) in which the head group was conjugated to the paramagnetic ion gadolinium (Gd 3+ ) (see Methods). Intensities of Rheb resonances were compared in the presence and absence of Gd 3+ to identify specific residues that interact with the membrane (Figure 5.5a). For Rheb-GDP the residues broadened by the paramagnetic ion (defined as >50% decrease in peak intensity) were distributed throughout the GTPase domain (30 peaks), including the N-terminus, β2-β3 loop, C-terminus of switch II, C-terminus of α3 and, as expected, the HVR (11 peaks). Mapping these regions to the Rheb-GDP structure illuminates a patch surrounding residue 171, which links the GTPase domain to the HVR, which is prenylated in cells (Figure 5.5b) (Clark et al. 1997). The unusually slow rate of GTP hydrolysis by Rheb (Marshall et al. 2009) allowed us to extend this PRE approach to Rheb bound to native GTP. The perturbed residues (24 in the GTPase-domain and 11 in the HVR) form a pattern similar to that seen for Rheb-GDP, however the C-terminus of switch II was less broadened in the GTP-bound form, as was the C-terminus of α3, albeit with smaller differences here (Figure 5.5b). To generate atomic-resolution models of the Rheb- membrane complex, we performed simulations with High Ambiguity Driven biomolecular DOCKing (HADDOCK) (Dominguez et al. 2003) using our experimentally derived PRE-restraints (as described in experimental procedures).

We generated 3000 models of Rheb docked to the membrane surface and the 300 models with lowest HADDOCK scores (a weighted sum of energy terms related to the interface and the individual components comprising the complex) were selected for cluster analysis. These models were grouped into clusters based on the orientation of Rheb relative to the membrane surface (i.e., RMSD following translation in the plane of the nanodisc and rotation about the normal axis, see experimental procedures).

15 Figure 5.3 Backbone N relaxation data for Rheb-GDP 1-181 in free (black) and nanodisc-bound (red) states. From top to bottom, transverse ( 15 N R2) relaxations, longitudinal ( 15 N R1) relaxations, 1 15 R1×R2 products, and the { H} N steady state NOE (I saturated /I unsaturated ), each plotted against the residue number. The α3-helix, which exhibits increased R2 and R1R2, is highlighted in red.

15 Figure 5.4 Backbone N relaxation data for Rheb-GMPPNP 1-181 in free (black) and nanodisc- bound (red) states. From top to bottom, transverse relaxations (R2), longitudinal ( 15 N R1) relaxations, R1×R2 products, and the {1H} 15 N steady state NOE, each plotted against the residue number.

Notably, mapping the PRE-affected residues on the structure of Rheb (Figure 5.5b and 5.6) illustrates that a single Rheb orientation cannot simultaneously satisfy all of the PRE-derived restraints,

and indeed, two major clusters of solutions were identified for the GDP-bound form (Figure 5.6a and 5.7a). Cluster 1 (backbone RMSD cutoff = 8 Å) was oriented such that the α6 helix (residues 153-171) was semi-perpendicular (58˚ relative to the bilayer plane for the cluster center, i.e., model with the lowest RMSD to all other models within the cluster), while cluster 2 (backbone RMSD cutoff = 8 Å) was semi-parallel (6˚ relative to the bilayer plane). Cluster 1 represented ~52% of the final solutions for Rheb·GDP docked to nanodiscs, while cluster 2 represented ~17% of the solutions. The remainder of the solutions formed small clusters, each of which comprised < 10% of total solutions, or did not cluster. In both cluster 1 and 2 models, the NBS, comprising the P-loop, switch I, G4 (residues 118-123), and G5 (residues 149-151) boxes were facing away from the membrane bilayer and were > 10 Å away (Figure 5.6a), consistent with the lack of broadening of these resonances, while the N-terminus, β2-β3 loop, and HVR were each within 10 Å of the bilayer plane. The major difference between the models in clusters 1

Figure 5.5 Identification of the Rheb-membrane interface and its modulation by the bound nucleotide. a) Paramagnetic relaxation enhancement (PRE) of Rheb H( 15 N) peak intensities induced by Gd 3+ -conjugated lipids incorporated into nanodiscs. PRE effects detected for each residue of Rheb-GDP (black) and Rheb-GTP (red) presented as a ratio of H( 15 N) resonance intensities of Rheb conjugated to nanodiscs in the presence (I*) or absence (Io) of Gd 3+ . Error bars based on spectral noise. b) PRE- affected residues mapped on Rheb G-domain . Residues that exhibited > 50% and > 80% peak broadening in the presence of Gd 3+ are shown as orange and red surfaces, respectively, for Rheb-GDP (left) and -GTP (right), and two interfaces are identified. c) PRE-driven HADDOCK models of Rheb- nanodisc complexes. Cluster center models of GDP- (left) and GTP- (right) bound Rheb in the semi- perpendicular and semi-parallel orientations. The non-transparent models reflect the favored conformation in each nucleotide bound state. The axis of Rheb α6 helix is shown with a brown arrow.

PDB 1XTQ and 1XTS (Yu et al. 2005) were used in the calculations, and results were deposited as 2M4A and 2M4B for GDP- and GTP-bound states, respectively.

versus 2 was the re-orientation of the G-domain with respect to the bilayer due to rotation about residue 171. This re-orientation causes the C-terminal regions of switch II and α3 to sample positions proximal to the bilayer plane (<5 Å) in cluster 1, and distal from the bilayer plane (>10 Å) in cluster 2. The re- orientation also brought α5 and β6 (residues 132-148) to within 10 Å of the bilayer in cluster 2. Likewise, two major clusters of models were identified for docking of ‘activated’ Rheb-GTP to the nanodisc (Figure 5.6b and 5.7b), with orientations similar to those determined for Rheb-GDP (i.e., cluster 1 exhibited a semi-perpendicular orientation with an α6 angle of 54˚, and cluster 2 was semi- parallel with an α6 angle of 3˚). Interestingly, relative to its GDP-counterpart, clusters 1 and 2 were more equally populated for Rheb-GTP, where 39 and 35 % of total models were found in clusters 1 and 2, respectively. The relative cluster sizes were independent of the overall number of models calculated and were consistently reproducible. It is intriguing that on the basis of FRET analysis and MD simulations, remarkably similar semi-perpendicular and semi-parallel orientations have been proposed for H-Ras-GDP and -GTP interactions with the membrane, respectively (Gorfe et al. 2007b; Abankwa et al. 2008b).

Figure 5.6 Residues affected by PRE localized on the HADDOCK models. The cluster center models of Rheb-nanodisc complexes are shown as follows; a) Rheb-GDP in orientation 1 (left) and 2 (right), b) Rheb-GTP in orientation 1 (left) and 2 (right). Residues exhibiting I*/Io < 0.5 and < 0.2 as described in Figure 2a are colored orange and red, respectively for Rheb-GDP and -GTP. The orange text indicates that the highlighted region is at the back of the page.

Our calculations show that the measured PRE restraints are consistent with an equilibrium between two distinguishable Rheb configurations, and that both of these conformers are populated by each of the GDP- and GTP-bound states, as illustrated in Figure 5.5c. The C-terminus of switch II of Rheb-GTP (residues 74-79) is significantly less perturbed by the PRE-tagged lipid than that of Rheb-GDP (Figure 5.5a), whereas the magnitude and distribution of the PRE effects are otherwise similar for both states; thus, the population of each clustered orientation is largely dictated by the nucleotide-specific PRE differences in switch II.

Figure 5.7 Cluster analysis of final HADDOCK solutions. The final 300 complex structures, represented with empty black circles are displayed in a plot of HADDOCK score versus rmsd to the global mean structure (defined as the structure with lowest rmsd to all other 299 structures) for a) Rheb- GDP-nanodisc and b) Rheb-GTP-nanodisc simulations. Complex structures belonging to cluster 1 (semi-perpendicular orientation, as described in the main text) and cluster 2 (semi-parallel orientation) are highlighted with green and magenta, respectively. The average HADDOCK score and rmsd (± standard deviation) of the 30 complex structures with lowest HADDOCK scores in each cluster is plotted as a filled black rectangle.

The low τM value we determined for membrane-tethered Rheb indicates considerable mobility on the membrane, most likely reflecting the fact that both semi-perpendicular and semi-parallel orientations are sampled during transient interactions between Rheb and the membrane surface. However, the semi- perpendicular configuration 1 appears to be less favoured in the GTP-bound state, due to subtle redistribution of surface electrostatics. The surface of the switch II-α3 region that forms the membrane interface in orientation 1 becomes less positively charged in the GTP-bound structure (Figure 5.8), which may diminish interactions with exposed negative charges in the phosphocholine bilayer(Semchyschyn and Macdonald 2004). Concomitantly, the α5-β6 of interface 2 becomes more positive in the GTP-bound structure, stabilizing the semi-parallel orientation (Figure 5.8). Consistent with this notion, a GDP/GTP-dependent shift in the equilibrium between perpendicular and parallel orientations has been described for H-Ras-membrane interactions by MD simulations (Gorfe et al. 2007b; Abankwa et al. 2008b). The NBS is not sterically occluded by the membrane surface in either perpendicular or parallel orientations; thus, we suggest that the mechanism by which membrane association impairs nucleotide exchange likely involves allosteric effects mediated via the membrane interaction, propagated from helix α3. The lower kinetic rate indicates a higher activation energy for the rate-limiting step in the intrinsic exchange reaction, suggesting that the proposed allosteric effect might stabilize the NBS in the GDP-bound state.

Avruch and colleagues previously performed a comprehensive Ala-mutation scan of surface accessible residues in Rheb and analyzed their effects on cellular mTORC1 signaling (Long et al. 2007). Ectopic expression of Rheb with mutations in the C-terminal region of switch II that did not impair GTP binding, nevertheless abolished the rescue of mTORC1 signaling in nutrient starved cells, and relative to wild-type Rheb, significantly reduced mTORC1 output in replenished cells with endogenous Rheb knocked down. This phenotype is similar to that observed with the nucleotide-deficient Rheb mutant D60I, but appears to be mediated by a different mechanism (Tabancay et al. 2003). Our data show that

Figure 5.8 Subtle changes in surface electrostatics upon nucleotide exchange. The lowest energy annealed structures of Rheb (residues 1-181) before docking simulations. Rheb-GDP and -GTP are shown on the left and right of each panel, respectively. Panels a and b represent the electrostatic surfaces of the switch II-α3 (interface 1) and α5-β6 (interface 2) motifs, respectively. In the electrostatic surface, red, blue and gray represent negative, positive and neutral (hydrophobic) surfaces. The surface electrostatics were generated using the macromolecular electrostatics calculation program Adaptive Poisson-Boltzmann Solver (APBS) (Baker et al. 2001), and visualized via PyMOL.

the C-terminal residues of switch II form part of the Rheb-membrane interface in cluster 1 of the GDP and GTP-bound states, but become solvent exposed in cluster 2. In light of these observations, we propose that the membrane acts as a regulatory platform in Rheb-mTORC1 communication in cells, in synergy with the GTPase cycle of Rheb. In our proposed model, the nucleotide-dependent re-orientation of Rheb regulates communication with mTORC1 in cells by GTP-mediated exposure of switch II. A similar mode of GTPase-effector regulation has been suggested for H-Ras interactions with the Ras binding domain (RBD) of C-Raf and Phosphatidylinositol 3-kinases γ (PI3K γ) (Abankwa et al. 2008a). However, as the Rheb-binding domain in mTOR has not been defined, the precise molecular details of mTORC1 regulation by the C-terminal switch II segment of Rheb in vivo remain to be fully elucidated.

Figure 5.9 Formation of a Rheb-PDE δ complex is compatible with Rheb-nanodisc model 2, but not model 1. Cluster center structures for models 1 and 2 from Rheb-GDP docking simulations are shown in green in a and b, respectively. The Rheb and PDE δ components of the Rheb-PDE δ complex (PDB: 3T5G) are shown in orange and blue, respectively, and were aligned with the G-domain of Rheb (residues 4-171). The nucleotide, Mg 2+ ion, and all atoms of POPC are represented as spheres.

Interestingly, a recent structure of Rheb in complex with PDE δ, a nucleotide-independent guanine dissociation inhibitor (GDI)-like protein (Ismail et al. 2011), reveals that PDE δ captures the farnesyl tail of Rheb in a deep pocket, and makes additional contacts (~20% of the total buried surface area) with the N-terminal β1 strand, and to a lesser extent, the C-terminal loop of switch II. Thus, overlaying the Rheb- PDE δ structure with Rheb-nanodisc models 1 and 2, it is evident that steric clashes between PDE δ and the membrane bilayer would preclude the Rheb-PDE δ complex from adopting a model 1 orientation, and likewise block PDE δ from binding this state of Rheb (Figure 5.9a). However, because of the transient nature of the Rheb-membrane interaction, other orientations sampled by the G-domain of Rheb including model 2 would allow the β1 strand and the C-terminal loop of switch II to interact with PDE δ (Figure 5.9b). Formation of this complex would then restrict G-domain interactions with the membrane and may initiate solubilization of Rheb, involving extraction of the farnesyl moiety from the bilayer into the hydrophobic pocket of PDE δ.

To our knowledge, the present study is the first experimental work to directly probe membrane- dependent regulation of the structure and function of a GTPase at atomic resolution, and provides a general approach for direct experimental studies on membrane-dependent regulation of other GTPases.

5.4 Experimental procedures: 5.4.1 Protein preparations: Mus musculus Rheb (residues 1-181) was prepared according to previous protocols (Marshall et al. 2009). In brief, the protein was expressed in Escherichia coli (BL21) using a 15 o pGEX2T vector and grown in minimal media supplemented with NH 4Cl and induced at 15 C with 0.25 mM IPTG. The protein was initially purified using glutathione Sepharose, cleaved from the GST tag by thrombin, with subsequent purification using Superdex 75 size exclusion chromatography in

buffer A (20mM Tris pH 7.4, 100mM NaCl, 5mM MgCl 2, 1mM TCEP). Rheb typically yields ~20 mg/L of culture, and like other members of the small GTPase superfamily, it co-purified with endogenous guanine nucleotide. Membrane scaffold protein 1D1 (MSP1D1) (Ritchie TK 2009) was prepared as described previously (Kobashigawa et al. 2011), with the following modifications. The

100

protein was expressed in Escherichia coli (BL21) with pGBHPS–MSP in 2× TY media using a LEX TM bioreactor system at 37 ˚C with 1 mM IPTG for 1 h followed by further 2.5h incubation at 28 ˚C. MSP was purified using His-tagged affinity purification followed by HRV3C protease-mediated His-tag cleavage and subsequent size exclusion chromatography using Superdex 75, with a typical yield of 40 mg/L of culture.

5.4.2 Preparation of Rheb-nanodisc complex: All lipids were purchased from Avanti Polar Lipids, Inc. Nanodiscs were prepared according to previous protocols (Kobashigawa et al. 2011) with the following modifications. 1,2-dioleoyl-sn -glycero-3-phosphocholine (DOPC) and the thiol-reactive lipid 1,2-dioleoyl-sn -glycero-3-phosphoethanolamine-N-[4-(p-maleimidomethyl)cyclohexane-carboxamide] (PE-MCC), were mixed with a molar ratio of 20:1 in a chloroform-ethyl alcohol solution. For PRE experiments, 1,2-distearoyl-sn -glycero-3-phosphoethanolamine-N-diethylenetriaminepentaacetic acid (gadolinium salt) (PE-DTPA (Gd 3+ )) was added to the lipids at a molar ratio of 20 DOPC : 1 PE-MCC : 1 PE-DTPA (Gd 3+ ). The organic solvents were then removed using gentle nitrogen flow followed by vacuum, and the dried lipid film was solubilized in aqueous buffer containing detergent (20mM Tris pH 7.4, 100mM NaCl, and 100mM sodium cholate). This mixture was subjected to three freeze/thaw cycles, vortexed until clarified, then MSP1D1 was added at a 1:40 molar ratio relative to the lipids (Kobashigawa et al. 2011). Following 1 h incubation with mild rotation at 20 °C, sodium cholate was removed from the MSP-lipid mixture with Bio-Beads SM-2 Adsorbents (Bio-RAD) using a batch method with 2h incubation at room temperature with mild rotation. The nanodisc particles were then purified via size exclusion chromatography using a 26-60 Superdex 200 column equilibrated with buffer A without TCEP. The particle size were analyzed with dynamic light scattering (DLS) and corresponded with a 10nm diameter particle. Rheb 1-181 was passed through a 10-30 Superdex S75 column equilibrated with buffer A without TCEP at 4 oC to remove reducing agent, and was immediately added to the nanodisc preparation at a 2:1 molar ratio (one Rheb molecule for each face of the nanodisc). The conjugation reaction was allowed to proceed for 16h at room temperature, then the mixture was passed through a Superdex 200 column equilibrated with buffer A to separate free Rheb from nanodisc-bound Rheb. The concentration of the nanodisc-Rheb complex was estimated by SDS-PAGE and size exclusion chromatography analysis.

101

5.4.3 NMR- measurements: NMR measurements were carried out on either a Bruker AVANCE II 800 MHz spectrometer equipped with a 5 mm TCI CryoProbeTM or a 600 MHz spectrometer equipped with TCI 1.7 mm MicroCryoProbe TM . The spectra were acquired at a temperature of 20 o C using samples containing 0.3mM free Rheb or 0.6mM nanodisc-conjugated Rheb. Backbone assignment of Rheb G- domain (residues 1-169) has been previously carried out and the HVR (residues 171-181) were assigned using HNCACB and CbCa(co)NH experiments (Table 5.1). Table 5.1) Resonance assignment of Rheb HVR.

Residue Residue Atom chemical ID # type shift Asp 171 N 121.793 Asp 171 H 8.334 Gly 172 N 108.277 Gly 172 H 8.199 Ala 173 N 123.686 Ala 173 H 7.975 Ala 174 N 122.079 Ala 174 H 8.157 Ser 175 N 114.197 Ser 175 H 8.172 Gln 176 N 121.812 Gln 176 H 8.272 Gly 177 N 109.788 Gly 177 H 8.44 1 Lys 178 N 120.904 Lys 178 H 8.2 Ser 179 N 117.489 Ser 179 H 8.484 Ser 180 N 118.296 Ser 180 H 8.475 Cys 181 N 124.733 Cys 181 H 8.194

102

To monitor GTPase reactions in real-time, sensitivity enhanced 1H-15 N heteronuclear single quantum coherence (HSQC) spectra with 8 scans (18 min) each were collected in succession as previously described(Marshall et al. 2009). For analysis of chemical shift perturbation and PRE experiments, 32 scans (72min) were used. No GDP resonances were observable above the spectral noise in the 32-scan 15 N1H-HSQC of a GTP-loaded Rheb sample indicating that any hydrolysis occurring during data collection was not appreciable. For PRE measurements, the resonance intensities of 15 N Rheb on PE- DTPA (Gd 3+ )-containing nanodiscs were compared to those of a control sample prepared without PE- DTPA (Gd 3+ ). No PE-DTPA-chelated diamagnetic ion is available to serve as a control, however the absence of any detectable chemical shift differences in the Rheb spectrum in the absence or presence of 5% (mole fraction to total lipid) PE-DTPA suggests that PE-DTPA-free nanodiscs provide an appropriate control to monitor the PRE-effect. The cross-peak intensities were measured using Sparky(Goddard and Kneller) and Gaussian line fitting. The difference in concentration between the PRE and the control samples was less than 10%, as judged by the size exclusion chromatogram and SDS-PAGE. This difference in concentration was corrected by normalization of the calculated intensity ratios against the highest observed I*/Io (where I* is the resonance intensities of Rheb conjugated to nanodiscs incorporating 5% Gd3+-conjugated PE-DTPA, and Io is that in the paramagnetic ion-free nanodiscs) for each GDP and GTP plot. Residues in the β1-α1 and β4-α3 loops, which are in close proximity to each other and close to the nucleotide binding pocket, exhibited the least PRE and were therefore used as an internal standard for the normalization, for GDP and GTP, respectively. For comparing GDP and GTP plots, this normalization process was evaluated using the PRE-effect on Cys181, which is covalently linked to the surface of the membrane, and thus would be expected to exhibit similar PRE, independent of the bound nucleotide. Indeed, Cys181 was similarly broadened in GDP- and GTP-bound samples (96.0% and 95.5 %, respectively). For relaxation experiments, the 15 N 1 15 longitudinal and transverse relaxation rates R 1 and R 2, as well as the { H} N steady-state NOEs were measured using 0.3mM samples of free 15 N Rheb-GDP and -GMPPNP (182-184) and samples of 0.6mM 15 N Rheb-GDP and -GMPPNP (182-184) conjugated to 0.3 mM nanodiscs. Rheb was preloaded with GMPPNP in the presence of EDTA- and alkaline phosphatase, as described previously(Marshall et al. 2009), prior to nanodisc conjugation. Backbone relaxation measurements were performed at 800MHz. Recycle delays of 3.0, 2.5, and 4.0 sec were used for 15 N R1, 15 N R2, and {1H} 15 N steady state NOE acquisitions, respectively. The relaxation delays ( 15 N R1 experiments:10, 180, 900, 265, 700, 1400, 350, 1100, 265, and 10 ms, and R2 experiments: 15.84, 63.36, 31.68, 95.04,

103

47.52, 84.20, 110.9, 63.36, and 15.84 ms) were varied in non-sequential order (as listed) to avoid systematic stability-related fluctuations in peak intensities, and duplicate delay times were used to ensure sample stability and estimate errors in fitting curves. Sample stability was monitored by intermediate HSQCs between T1, T2, and { 1H} 15 N steady state NOE measurements. 15 N R1 and 15 N R2 experiments were run with 16 scans, and 64 scans were acquired for { 1H} 15 N steady state NOEs. The overall rotational correlation time ( τM) was calculated as previously detailed (Farrow NA 1995). In brief,

the values of the spectral density functions J(0.870 ωH), J( ωN), and J(0) were determined from the mean of R1, R2 and {1H} 15 N steady state NOEs using equations 12-14 from Farrow et al (Farrow NA 1995), in which no prior assumption is made on the structure of the molecule. Of the values of 15 N R1, 15 N R2, and { 1H} 15 N steady state NOEs determined for each peak, the top and bottom 10% were excluded from the calculation. The value of the spectral density function at zero frequency was then used to derive the

(τM) using the form of the spectral density function derived from model-free formalism (equation 4 in Farrow et al (Farrow NA 1995)) assuming limited (large order parameter, S 2) and fast (<10ps) internal motions. The variations in the trimmed means of R1, R2, and { 1H} 15 N steady state NOEs were used in error calculations.

5.4.4 Real-time NMR-based GTPase assay: Assays of Rheb GTPase function were carried out as described previously (Marshall et al. 2009). In brief, nucleotide exchange reactions were initiated by adding a 5-fold molar excess of GTP to a GDP-loaded 15 N Rheb sample, and monitored via successive 8-scan 15 N 1H HSQC spectra (18min). TSC2GAP-catalyzed reactions were initiated by adding extracts of HEK-293 cells over-expressing full length TSC1 and TSC2 as described previously (Marshall et al. 2012). The reactions were initiated with cell extracts prepared by mechanical lysis. Expression of TSC2 in these lysates was confirmed by Western blots. The hydrolysis data were fitted to a one phase association function. Since the intrinsic hydrolysis was unchanged upon membrane association, a simple one phase exponential decay function was used to fit the intrinsic exchange data.

5.4.5 Molecular docking simulation: All docking simulations were performed with gpc supercomputer at the SciNet (Chris Loken 2010) HPC Consortium using High Ambiguity Driven biomolecular DOCKing (HADDOCK) sofware version 2.1 (Dominguez et al. 2003; de Vries et al. 2007). The structures of GDP- and GTP-bound of Mus musculus Rheb (residues 1-181) were annealed in CNS using the RECOORD scripts (Nederveen et al. 2005) from starting extended polypeptide chains using

104

distance, hydrogen bond, and torsion angle restraints derived from the crystal structures of human Rheb GDP and GTP (1XTQ and 1XTS, respectively)(Yu et al. 2005), which share 99% sequence identity. No restraints were given for 1-MP-2 and 172-DGAASQGKSSC-181 residues during the annealing protocol so that these Rheb termini sampled a large conformational variability in the final ensemble of structures consistent with our { 1H} - 15 N steady state NOEs. The G-domains (residues 4-171) of the 20 lowest energy annealed structures were all within 0.1 Å backbone rmsd of the template crystal structures. Each of the 20 lowest-energy structures for Rheb-GDP and Rheb-GTP were used as starting models in HADDOCK. The coordinate file for a nanodisc model (Segrest et al. 1999) contained 80 1-palmitoyl-2- oleoyl-sn -glycero-3-phosphocholine (POPC) lipids per leaflet (i.e. 160 total), encompassed by 2 MSP polypeptide chains. The acyl chain of POPC is shorter than that of DOPC by two C atoms, but they share the same headgroup, thus their bilayer surfaces would be virtually identical.

HADDOCK ambiguous restraints were generated using the PRE-measurements on nanodisc-tethered Rheb. Because the lateral diffusion rate of DOPC at room temperature (~8.2×103 nm 2/ms) (Andrey Filippov 2003) is high relative to T2 relaxation times of Rheb on nanodiscs (~30 ms), an assumption was made that the PRE tag uniformly sampled all positions on the nanodisc surface (78 nm 2) during the time course of the measurement. Thus, ambiguous restraints were generated between PRE-affected residues on Rheb (e.g. Gly51) and “any” lipid headgroup atom on the nanodisc surface ( e.g. G51 to lipid-1[N1 or C13 or C14 or C15 or P1 or O11 …] or lipid-2[N1 or C13 or C14 or C15 or P1 or O11 …] or …). Residues in Rheb with peaks displaying >80% broadening were defined in HADDOCK as ‘active’ residues and those broadened between 50% and 80% were defined as passive residues. For the Rheb-GDP-nanodisc complex the active residues were 1, 3, 6, 51, 77, 110, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, and the passive residues were 5, 7, 47, 48, 49, 53, 55, 74, 75, 76, 78, 106, 107, 108, 132, 114, 141, 143, 144, 145, 154, 162, 164, 169, 171. For the Rheb-GTP-nanodisc the active residues were 1, 3, 6, 48, 51, 110, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, and the passive residues were 5, 7, 49, 53, 54, 108, 109, 114, 115, 132, 140, 141, 143, 144, 145, 154, 162, 166, 171. The ambiguous restraints were assigned a distance range of 2-10 Å from any atom of any lipid headgroup. Although Gd 3+ is known to affect nuclear spin magnetization within a 20 Å radius (Otting 2010), the upper limit of 10 Å was selected because i) each lipid head group on the membrane surface is only partially occupied by Gd 3+ - conjugated lipid which comprises 5% of the total lipids, and ii) active residues were stringently defined (i.e. >80% broadening) in HADDOCK. An upper limit of 2 Å was selected for Cys181 of Rheb, since it

105

was covalently linked to the membrane surface. The CNS topology and the parameter files for the nucleotides and POPC lipids were initially generated using the HIC-UP server (Kleywegt et al. 2003) and modified according to the data listed for these small molecules in the Automated Topology Builder (ATB) and Repository (Malde AK 2011). The docking protocol consisted of a 3000 rigid-body docking stage, where the top 300 ranked structures based on the HADDOCK score were refined using semi- flexible simulated annealing. The docking protocol was executed using the default HADDOCK script parameters except an additional Powell energy minimization step was performed on the lipid headgroups prior to the semi-flexible refinement stage.

Pairwise backbone rmsd values were tabulated for the Rheb G-domain (residues 4-171) after alignment of the nanodiscs on the same plane and a translational and 360 o rotational (in 5 ˚ increments) rmsd minimization search confined only to movements within the 2D plane of the membrane surface in order to structurally align the Rheb molecules relative to the large lateral dimension of the nanodisc. The Rheb-nanodisc position in each solution was kept constant during pairwise rmsd calculation. Cluster analysis was then performed on the pairwise rmsd values using a previously described algorithm (Xavier Daura 1999), setting an rmsd cutoff of 8 Å and cluster size cutoff of 30 structures. The HADDOCK scores of all 300 structures were plotted against rmsd relative to the global mean structure, defined as the solution with the lowest average pairwise rmsd to all other 299 solutions (Figure 5.7). To estimate the orientation angles, a vector was defined along the α6-helix of Rheb’s G-domain, and the minimum angle between the long axis of α6 and the nanodisc membrane surface was measured for each complex. All structural manipulations and measurements in three-dimensional space were performed using Crystallography & NMR System (CNS) (Brunger et al. 1998). The top 20% HADDOCK-scored complex structures of each cluster, as well as the cluster center models for each nucleotide-bound state, have been deposited in PDB/BMRB (PDB codes 2M4A (GDP) and 2M4B (GTP); BMRB accession numbers 18996 (GDP) and 18997 (GTP)).

106

5.5 References:

Abankwa D, Gorfe AA, Hancock JF. 2008a. Mechanisms of Ras membrane organization and signalling: Ras on a rocker. Cell Cycle 7: 2667-2673. Abankwa D, Hanzal-Bayer M, Ariotti N, Plowman SJ, Gorfe AA, Parton RG, McCammon JA, Hancock JF. 2008b. A novel switch region regulates H-ras membrane orientation and signal output. Embo J 27 : 727-735. Andrey Filippov GO, and Göran Lindblom. 2003. Influence of Cholesterol and Water Content on Phospholipid Lateral Diffusion in Bilayers. Langmuir 19 : 6397–6400. Baker NA, Sept D, Joseph S, Holst MJ, McCammon JA. 2001. Electrostatics of nanosystems: application to microtubules and the ribosome. Proc Natl Acad Sci U S A 98: 10037-10041. Berndt N, Hamilton AD, Sebti SM. 2011. Targeting protein prenylation for cancer therapy. Nat Rev Cancer 11 : 775-791. Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS et al. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54 : 905-921. Brunsveld L, Waldmann H, Huster D. 2009. Membrane binding of lipidated Ras peptides and proteins-- the structural point of view. Biochim Biophys Acta 1788 : 273-288. Buerger C, DeVries B, Stambolic V. 2006. Localization of Rheb to the endomembrane is critical for its signaling function. Biochem Biophys Res Commun 344 : 869-880. Chris Loken DG, Leslie Groer, Richard Peltier, Neil Bunn, Michael Craig, Teresa Henriques, Jillian Dempsey, Ching-Hsing Yu, Joseph Chen, L Jonathan Dursi, Jason Chong, Scott Northrup, Jaime Pinto, Neil Knecht and Ramses Van Zon. 2010. SciNet: Lessons Learned from Building a Power-efficient Top-20 System and Data Centre. J Phys: Conf Ser 256 : 012026. Clark GJ, Kinch MS, Rogers-Graham K, Sebti SM, Hamilton AD, Der CJ. 1997. The Ras-related protein Rheb is farnesylated and antagonizes Ras signaling and transformation. J Biol Chem 272 : 10608-10615. de Vries SJ, van Dijk AD, Krzeminski M, van Dijk M, Thureau A, Hsu V, Wassenaar T, Bonvin AM. 2007. HADDOCK versus HADDOCK: new features and performance of HADDOCK2.0 on the CAPRI targets. Proteins 69 : 726-733. Denisov IG, Grinkova YV, Lazarides AA, Sligar SG. 2004. Directed self-assembly of monodisperse phospholipid bilayer Nanodiscs with controlled size. J Am Chem Soc 126 : 3477-3487. Dominguez C, Boelens R, Bonvin AM. 2003. HADDOCK: a protein-protein docking approach based on biochemical or biophysical information. J Am Chem Soc 125 : 1731-1737. Farrow NA ZO, Szabo A, Torchia DA, Kay LE. 1995. Spectral density function mapping using 15N relaxation data exclusively. J Biomol NMR 6: 153-162. Fraser JS, van den Bedem H, Samelson AJ, Lang PT, Holton JM, Echols N, Alber T. 2011. Accessing protein conformational ensembles using room-temperature X-ray crystallography. Proc Natl Acad Sci U S A 108 : 16247-16252. Goddard TD, Kneller D. SPARKY 3. in University of California, San Francisco . Gorfe AA, Babakhani A, McCammon JA. 2007a. H-ras protein in a bilayer: interaction and structure perturbation. J Am Chem Soc 129 : 12280-12286. Gorfe AA, Hanzal-Bayer M, Abankwa D, Hancock JF, McCammon JA. 2007b. Structure and dynamics of the full-length lipid-modified H-Ras protein in a 1,2-dimyristoylglycero-3-phosphocholine bilayer. J Med Chem 50 : 674-684.

107

Gureasko J, Galush WJ, Boykevisch S, Sondermann H, Bar-Sagi D, Groves JT, Kuriyan J. 2008. Membrane-dependent signal integration by the Ras activator Son of sevenless. Nat Struct Mol Biol 15 : 452-461. Ismail SA, Chen YX, Rusinova A, Chandra A, Bierbaum M, Gremer L, Triola G, Waldmann H, Bastiaens PI, Wittinghofer A. 2011. Arl2-GTP and Arl3-GTP regulate a GDI-like transport system for farnesylated cargo. Nat Chem Biol 7: 942-949. Kleywegt GJ, Henrick K, Dodson EJ, van Aalten DM. 2003. Pound-wise but penny-foolish: How well do micromolecules fare in macromolecular refinement? Structure 11 : 1051-1059. Kobashigawa Y, Harada K, Yoshida N, Ogura K, Inagaki F. 2011. Phosphoinositide-incorporated lipid- protein nanodiscs: A tool for studying protein-lipid interactions. Anal Biochem 410 : 77-83. Long X, Lin Y, Ortiz-Vega S, Busch S, Avruch J. 2007. The Rheb switch 2 segment is critical for signaling to target of rapamycin complex 1. J Biol Chem 282 : 18542-18551. Malde AK ZL, Breeze M, Stroet M, Poger D, Nair PC, Oostenbrink C, Mark AE. 2011. An Automated force field Topology Builder (ATB) and repository: version 1.0. Journal of Chemical Theory and Computation 7: 4026-4037. Marshall CB, Ho J, Buerger C, Plevin MJ, Li GY, Li Z, Ikura M, Stambolic V. 2009. Characterization of the intrinsic and TSC2-GAP-regulated GTPase activity of Rheb by real-time NMR. Sci Signal 2: ra3. Marshall CB, Meiri D, Smith MJ, Mazhab-Jafari MT, Gasmi-Seabrook GM, Rottapel R, Stambolic V, Ikura M. 2012. Probing the GTPase cycle with real-time NMR: GAP and GEF activities in cell extracts. Methods 57 : 473-485. Nederveen AJ, Doreleijers JF, Vranken W, Miller Z, Spronk CA, Nabuurs SB, Guntert P, Livny M, Markley JL, Nilges M et al. 2005. RECOORD: a recalculated coordinate database of 500+ proteins from the PDB using restraints from the BioMagResBank. Proteins 59 : 662-672. Otting G. 2010. Protein NMR Using Paramagnetic Ions. Annu Rev Biophys 39 : 387–405. Ritchie TK GY, Bayburt TH, Denisov IG, Zolnerciks JK, Atkins WM, Sligar SG. 2009. Reconstitution of Membrane Proteins in Phospholipid Bilayer Nanodiscs. Methods Enzymol 464 : 211-231. Segrest JP, Jones MK, Klon AE, Sheldahl CJ, Hellinger M, De Loof H, Harvey SC. 1999. A detailed molecular belt model for apolipoprotein A-I in discoidal high density lipoprotein. J Biol Chem 274 : 31755-31758. Semchyschyn DJ, Macdonald PM. 2004. Conformational response of the phosphatidylcholine headgroup to bilayer surface charge: torsion angle constraints from dipolar and quadrupolar couplings in bicelles. Magn Reson Chem 42 : 89-104. Tabancay AP, Jr., Gau CL, Machado IM, Uhlmann EJ, Gutmann DH, Guo L, Tamanoi F. 2003. Identification of dominant negative mutants of Rheb GTPase and their use to implicate the involvement of human Rheb in the activation of p70S6K. J Biol Chem 278 : 39921-39930. Xavier Daura KG, Bernhard Jaun, Dieter Seebach, Wilfred F. van Gunsteren and Alan E. Mark. 1999. Peptide Folding: When Simulation Meets Experiment. Angewandte Chemie International Edition 38 : 236-240. Yu Y, Li S, Xu X, Li Y, Guan K, Arnold E, Ding J. 2005. Structural basis for the unique biological function of small GTPase RHEB. J Biol Chem 280 : 17093-17100.

108

Chapter 6

Conclusions and Future directions

109

6.1 Conclusions:

In our studies of the widely used mant-conjugated nucleotides, we demonstrated how perturbation of exchange and hydrolysis reaction rates due to the fluorescent tag can lead to potential misinterpretations of enzymatic functions of three small GTPases, H-Ras, Rheb, and RhoA. The effects of the mant moiety on the reaction kinetics were unpredictable for each GTPase tested, demonstrating the importance of the NMR-based GTPase assay, which is ‘label-free’ and enables determination of kinetic parameters of the GTPase reactions with native nucleotide. This work was published in J. Biol. Chem (Mazhab-Jafari et al. 2010).

The fluorescent mant-GTP analog unexpectedly enhanced Rheb’s intrinsic nucleotide hydrolysis rate by ~10 fold, through a mechanism that is independent of Gln64 (which corresponds to the Ras catalytic residue Gln61). Through our studies to understand the mechanism of rapid hydrolysis of mant-GTP by Rheb, we identified and characterized a non-canonical nucleotide hydrolysis mechanism involving Tyr35 in switch I and Asp65 in switch II. Tyr35 played a dual role in Rheb nucleotide catalysis. On one hand it auto-inhibited the intrinsic nucleotide hydrolysis by stabilizing a switch II conformation in which Asp65 was further away form the nucleotide through an H-bonding network from the side-chain hydroxyl of Tyr35 to the backbone amide of Gly63. Mutation of Tyr35 to Ala increased GTP hydrolysis 10-fold. On the other hand, Tyr35 was necessary to facilitate productive TSC2GAP-mediated nucleotide hydrolysis by Rheb. Indeed, we showed that Tyr35 was necessary for mTORC1 homeostasis in cells, as its mutation decoupled mTORC1 signaling from regulation by serum. We also showed that Asp65 contributed significantly to GTP hydrolysis in Rheb, implicating this residue as a potential catalytic residue in this GTPase. This work was published in Structure (Mazhab-Jafari et al. 2012).

cat Our structural analyses of the Rheb Y35A mutant showed that the catalytic water molecule H 2O is coordinated by the amide of Gly63 in the conserved G3 box at the N-terminus of switch II. We utilized this observation in a novel protein engineering strategy to generate GDP- and GTP-locked Rheb variants by mutations of the single ultra-conversed residue Gly63. Replacing the H α with a larger methyl group cat (G63A mutation) displaced the H 2O , resulting in severely reduced intrinsic GTP hydrolysis and resistance to the GAP activity of TSC2. However, introducing a bulkier side chain (G63V mutation) sterically occluded the space occupied by the γ-phosphate of GTP in the WT protein, resulting in an inactive Rheb mutant that could only accommodate GDP. We demonstrated that Rheb G63A over-

110

expression results in elevated mTORC1 signaling in cells. This work was published in J. Biol. Chem (In Press, doi: 10.1074/jbc.C113.543736).

In the final study of this thesis, we sought to characterize Rheb’s interactions with membranes, and investigate how this affects its GTPase functions. This is an important study, since Rheb requires farnesylation and membrane association to carry out its physiological functions in cells, however very little is known about the details of Rheb-membrane interactions. These challenging NMR studies were enabled by the use of nanodiscs, 5 by 10 nm discoidal lipoprotein complexes, as a model membrane system. Rheb was covalently conjugated through its sole C-terminal cysteine to a maleimide-conjugated lipid incorporated into nanodiscs. The GTPase domain of Rheb loosely associates with the surface of the bilayer and we defined two preferred specific orientations, which are dependent on the bound nucleotide. Using real-time NMR-based GTPase assays, we showed that membrane conjugation of Rheb directly impacts the GTPase cycle. This work, published in J. Am. Chem. Soc. (Mazhab-Jafari et al. 2013), revealed previously unknown biological properties of Rheb, and sets the stage for future studies of other membrane-associated GTPases of the Ras super-family.

6.2 Future Directions: 6.2.1 Engineering GTPase probes by structure-guided mutations of the G3-box glycine: Our protein engineering study showed that manipulation of the conserved G3-box glycine in Rheb by mutations can be used to control the GTPase cycle, and led to the development of the gain-of-function mutant Rheb G63A, which is constitutively GTP-bound. This mutant, which is resistant to intrinsic and TSC2GAP- mediated nucleotide hydrolysis reactions, should provide a valuable probe to identify potential novel Rheb protein-protein interactions in cells, which appear to be transient and difficult to detect. In the last few years, there have been a number of controversial reports of Rheb protein:protein interactions that could not be reproduced by other groups such as FKPB38 (Wang et al. 2008; Uhlenbrock et al. 2009) and TCTP (Rehmann et al. 2008; Wang et al. 2008). Paradoxically, mTOR was reported to interact more tightly with inactive Rheb-GDP than –GTP (Long et al. 2005), but was only activated by Rheb-GTP (Sancak et al. 2007). More reliable activated and inactivated Rheb mutants may contribute to a better understanding of its structure and function. Furthermore, the G63V mutant, although it expressed at lower concentrations in eukaryotic cells, may be useful to identify potential GEFs for Rheb, since it cannot bind GTP, and thus may engage in longer-lived interactions with GEFs. Indeed, we have

111

promising preliminary data in a recent collaboration with Dr. Anne-Claude Gingras to identify proteins that are associated with Rheb WT, G63A, and G63V mutants stably expressed in cells using the proximity-dependent biotin identification (BioID) method (Roux et al. 2012). In addition, the high degree of conservation of this G3-box residue across the Ras superfamily of GTPases, both at the level of the sequence and tertiary structure, suggests that analogous mutations may serve as probes for other GTPases. This would be particularly valuable for those GTPases in which the canonical catalytic Gln (i.e. Ras Q61) is not conserved, and thus catalytically impaired hyperactivated mutants are not available. However, it will be important to characterize whether mutations of the G3-box glycine confer the expected biochemical properties and phenotype for each GTPase. For example, in Ras, G60A mutation reduces the GTP hydrolysis rate, however this mutant acts as a dominant negative mutant in cells by sequestering SOS (Ford et al. 2005).

6.2.2 Towards small molecule inhibitors of Rheb: Traditionally small GTPases were thought to be undruggable because GTP binds with very high affinity (Kd ~ nM) and is abundant in the cell, and these proteins generally lack other defined binding site on the surface of the protein. However, recent structure-guided studies have produced promising results in targeting H- and K-Ras with small molecule compounds (Maurer et al. 2012; Sun et al. 2012; Ostrem et al. 2013; Hunter et al. 2014; Shima et al. 2013; Lim et al. 2014). Our studies suggest at least two strategies for targeting Rheb with small molecules that stabilize its inactive form. (i) The NBS is in an open conformation in the GDP-bound state due to the lack of a Try35-γ-phosphate interaction. Since we showed that positioning of bulky valine side chain in close proximity to Gly63 can inhibit GTP binding, a small molecule that specifically binds Rheb-GDP and occupies this space close to Gly63 may block GTP binding to WT Rheb by occluding the γ-phosphate binding site. (ii) Since we found that Rheb-GDP adopts a specific orientation with respect to the membrane surface that masks switch II from the solvent, one can analyze the protein- membrane interface to develop a small compound with a hydrophobic side chain that inserts into the bilayer and a head group that specifically binds Rheb-GDP at the membrane interface. This type of molecule may stabilize a Rheb orientation in which switch II is masked from the solvent, and diminish Rheb’s ability to signal to mTORC1.

6.2.3 Probing membrane-dependent regulation of K-Ras-effector interaction: The methods we developed to analyze Rheb GTPase interactions with membranes at high resolution using nanodisc-

112

based lipid bilayers can be extended to other members of the Ras superfamily that are targeted to biological membrane compartments in cells to carry out their physiological functions. Hence, we have begun preliminary analysis of K-Ras4B (Figure 6.1) conjugated to nanodiscs, using a protocol described in detail in appendix B. The K isoform of Ras displays the highest frequency of somatic mutations in cancer (Prior et al. 2012). Thus understanding its physiological and oncogenic properties as well as interactions with its effectors on membrane bilayers is crucial to rational design of small molecular inhibitors for therapeutic interventions.

These nanodisc methodologies have enabled preliminary investigations of biophysical properties of K- Ras on bilayer membranes including; i) orientation with respect to the bilayer plane and its modulation by the bound nucleotide (GDP versus GMPPNP), ii) structural and dynamical effects of membrane binding on the K-Ras G-domain, iii) examination of the interactions between K-Ras and Ras binding domains (RBDs) on the membrane, iv) effects of oncogenic mutations on Ras-membrane orientation and interaction with RBDs, and v) effect of membrane association on the K-Ras GTPase cycle.

6.3 Closing Remarks: Since the discovery of Rheb two decades ago, intense studies of its physiological and pathological roles have been carried out and an understanding of the role of this small GTPase in different cellular signaling pathways is emerging, however many questions remain. Far fewer structure-function studies of Rheb have been carried out, relative to those of Ras. The structural and functional works presented in this thesis, revealed the structural basis of autoinhibition of Rheb GTPase activity that contributes to its high activation state in cells, which can become pathogenic upon limiting TSC2 GAP activity. Using this knowledge, we have also developed novel probes for dissecting the signaling pathways that Rheb regulates. These probes can also be applied to studies of other GTPases with non-canonical GTP hydrolysis mechanisms (i.e., no functional catalytic Gln corresponding to Ras Gln61). Furthermore, our structural and functional work on the bilayer membranes have revealed potential mechanisms of membrane-dependent regulation of Rheb function in cells, and this high-resolution methodology can be extended to the studies of other membrane associated GTPases. The structural and functional knowledge of Rheb catalytic function and membrane association, which are two pillars of Rheb function, can aid in development of novel therapeutic strategies that combat Rheb-mediated pathogenesis.

113

114

Figure 6.1) Backbone and side chain assignments of free K-Ras4B in complex with GMPPNP. a) Assignment of K-Ras4B residues 1-185. 60% of the G-domain and all of the HVR were assigned. b) Chemical shift assignment of K-Ras4B Ile δ-CH 3 (aided by assignment transfer from of H-Ras-GDP, courtesy of Dr. Sharon Campbell, see appendix B for more details).

6.4 References:

Ford B, Skowronek K, Boykevisch S, Bar-Sagi D, Nassar N. 2005. Structure of the G60A mutant of Ras: implications for the dominant negative effect. J Biol Chem 280 : 25697-25705. Hunter JC, Gurbani D, Ficarro SB, Carrasco MA, Lim SM, Choi HG, Xie T, Marto JA, Chen Z, Gray NS et al. 2014. In situ selectivity profiling and crystal structure of SML-8-73-1, an active site inhibitor of oncogenic K-Ras G12C. Proc Natl Acad Sci U S A 111 : 8895-8900. Lim SM, Westover KD, Ficarro SB, Harrison RA, Choi HG, Pacold ME, Carrasco M, Hunter J, Kim ND, Xie T et al. 2014. Therapeutic targeting of oncogenic K-Ras by a covalent catalytic site inhibitor. Angew Chem Int Ed Engl 53 : 199-204. Long X, Lin Y, Ortiz-Vega S, Yonezawa K, Avruch J. 2005. Rheb binds and regulates the mTOR kinase. Curr Biol 15 : 702-713. Maurer T, Garrenton LS, Oh A, Pitts K, Anderson DJ, Skelton NJ, Fauber BP, Pan B, Malek S, Stokoe D et al. 2012. Small-molecule ligands bind to a distinct pocket in Ras and inhibit SOS-mediated nucleotide exchange activity. Proc Natl Acad Sci U S A 109 : 5299-5304. Mazhab-Jafari MT, Marshall CB, Ishiyama N, Ho J, Di Palma V, Stambolic V, Ikura M. 2012. An autoinhibited noncanonical mechanism of GTP hydrolysis by Rheb maintains mTORC1 homeostasis. Structure 20 : 1528-1539. Mazhab-Jafari MT, Marshall CB, Smith M, Gasmi-Seabrook GM, Stambolic V, Rottapel R, Neel BG, Ikura M. 2010. Real-time NMR study of three small GTPases reveals that fluorescent 2'(3')-O- (N-methylanthraniloyl)-tagged nucleotides alter hydrolysis and exchange kinetics. J Biol Chem 285 : 5132-5136. Mazhab-Jafari MT, Marshall CB, Stathopulos PB, Kobashigawa Y, Stambolic V, Kay LE, Inagaki F, Ikura M. 2013. Membrane-dependent modulation of the mTOR activator Rheb: NMR observations of a GTPase tethered to a lipid-bilayer nanodisc. J Am Chem Soc 135 : 3367-3370. Ostrem JM, Peters U, Sos ML, Wells JA, Shokat KM. 2013. K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature 503 : 548-551. Prior IA, Lewis PD, Mattos C. 2012. A comprehensive survey of Ras mutations in cancer. Cancer Res 72 : 2457-2467. Rehmann H, Bruning M, Berghaus C, Schwarten M, Kohler K, Stocker H, Stoll R, Zwartkruis FJ, Wittinghofer A. 2008. Biochemical characterisation of TCTP questions its function as a guanine nucleotide exchange factor for Rheb. FEBS Lett 582 : 3005-3010. Roux KJ, Kim DI, Raida M, Burke B. 2012. A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells. J Cell Biol 196 : 801-810. Sancak Y, Thoreen CC, Peterson TR, Lindquist RA, Kang SA, Spooner E, Carr SA, Sabatini DM. 2007. PRAS40 is an insulin-regulated inhibitor of the mTORC1 protein kinase. Mol Cell 25 : 903-915.

115

Shima F, Yoshikawa Y, Ye M, Araki M, Matsumoto S, Liao J, Hu L, Sugimoto T, Ijiri Y, Takeda A et al. 2013. In silico discovery of small-molecule Ras inhibitors that display antitumor activity by blocking the Ras-effector interaction. Proc Natl Acad Sci U S A 110 : 8182-8187. Sun Q, Burke JP, Phan J, Burns MC, Olejniczak ET, Waterson AG, Lee T, Rossanese OW, Fesik SW. 2012. Discovery of small molecules that bind to K-Ras and inhibit Sos-mediated activation. Angew Chem Int Ed Engl 51 : 6140-6143. Uhlenbrock K, Weiwad M, Wetzker R, Fischer G, Wittinghofer A, Rubio I. 2009. Reassessment of the role of FKBP38 in the Rheb/mTORC1 pathway. FEBS Lett 583 : 965-970. Wang X, Fonseca BD, Tang H, Liu R, Elia A, Clemens MJ, Bommer UA, Proud CG. 2008. Re- evaluating the roles of proposed modulators of mammalian target of rapamycin complex 1 (mTORC1) signaling. J Biol Chem 283 : 30482-30492.

116

APPENDIX A

Formula used for data fitting of intrinsic- and GAP-mediated nucleotide hydrolysis: Intrinsic- and GAP- mediated nucleotide hydrolysis data were fitted to one phase exponential association to determine the rates according to:

* * * [I GDP /(I GDP +I GTP )] (t) = 1 - exp (-khy × t) equation 1

* I GDP = Peak heights of Ras or Rheb bound to GDP or mantGDP

* I GTP = Peak heights of Ras or Rheb bound to GTP or mantGTP

khy = nucleotide hydrolysis rate

In the case of RhoA, due to line broadening of the activated (GTP-bound) state, we used the following one phase exponential association formula using the increasing intensities of the resonances related to the inactive (GDP-bound) conformation of the GTPase:

* *f [I GDP /I GDP ] (t) = 1 - exp (-khy × t) equation 2

*f I GDP = Peak heights of the GDP or mantGDP after complete GTP or mantGTP hydrolysis.

Determination of the true nucleotide exchange rate k ex in the presence of intrinsic hydrolysis. When nucleotide exchange assays are performed with hydrolysable nucleotides (e.g., GTP, mantGTP), the observed exchange rate (plotted by continuous lines in figures 2.5 and 2.6) is determined by a combination of the GEF-mediated exchange rate and the intrinsic hydrolysis of the nucleotide bound to the small GTPase (Ras or RhoA). If the rate of intrinsic nucleotide hydrolysis is known, the observed data can be fitted to equation 4 to obtain the true exchange rate k ex . This model makes the assumption that all of the GTPase protein is bound to either GTP or GDP in solution, due to nM affinities for the nucleotide. The true exchange rate can be used to plot a one phase exponential decay curve approximating exchange in the absence of hydrolysis (dashed lines in figures 2.5 and 2.6).

* * * * * * * * * d [I GDP /(I GDP +I GTP )] / d t = - kex [I GDP /(I GDP +I GTP )] + k hy [1 – (I GDP /(I GDP +I GTP ))] equation 3

* * * I GDP /(I GDP +I GTP ) (t=0) = 1

* * * [I GDP /(I GDP +I GTP )](t)= 1 – [k ex /(k ex +k hy ) - kex /(k ex + k hy ) exp –(k ex + k hy ) t] equation 4

117

kex = nucleotide exchange rate

As mentioned above, GTP-bound RhoA exhibited line broadening and split peaks that complicated peak integration hence we used the resonances of the GDP-bound state to calculate the exchange rate.

* *0 * *0 * *0 d [I GDP /I GDP ] / d t = - kex [I GDP /I GDP ] + k hy [1 – (I GDP /I GDP )] equation 5

* *0 I GDP /I GDP (t=0) = 1

* *0 [I GDP /I GDP ](t)= 1 – [k ex /(k ex +k hy ) - kex /(k ex + k hy ) exp –(k ex + k hy ) t] equation 6

*0 I GDP = Peak heights of the GDP or mantGDP before the addition of GEF and excess GTP or mantGTP nucleotide.

118

APPENDIX B

Preparation of K-Ras4B-nanodisc complex:

A construct encoding K-ras4b (amino acids 1-185, C51S, C80S, C118S) was synthesized (Genscript) and ligated into pET-28 vector. The buried Ser51 and Ser80 were back mutated to native Cys residues via site directed mutagenesis to improve expression. The protein expression was induced in E. Coli o 15 BL21 (DE3) at an OD 600 of 0.6 and a temperature of 15 C in M9 media supplemented with NH 3. 70 mg/L of 2-Ketobutyric acid-4-13 C was added to the media one hour prior to IPTG-meditated induction of expression to label the C δ position of the isoleucine residues with 13 C isotope (Tugarinov et al. 2006). K-Ras4B was then purified using standard His-tag affinity purification methods, as described for H-Ras in chapter 2.

The nanodisc assembly was carried out as described in detail in chapter 5, with the exception of a different lipid composition (75% DOPC, 20% DOPS, 5% PE-MCC, mol ratio) to mimic the mammalian plasma membrane. For PRE experiments, 2.5% (final mol ratio) of 1,2-distearoyl-sn -glycero-3- phosphoethanolamine-N-diethylenetriaminepentaacetic acid (gadolinium salt) was used in the assembly of nanodiscs. All lipids were from Avanti Polar Lipids. The GTPase-nanodisc conjugation reaction and size exclusion chromatography were performed as described for Rheb-nanodisc conjugation, in chapter 5, except that K-Ras4B constructs were incubated with 100 units of bovine thrombin (per 20 mg of KRas at a concentration of 1mM) for 2h at room temperatures to ensure complete cleavage of the His- tag prior to the conjugation reaction. The cleavage reaction was confirmed via mass spectroscopy. Nucleotide exchange was carried out prior to nanodisc conjugation in the presence of GMPPNP, EDTA and calf intestinal alkaline phosphatase, (Cip) followed by passage through S75 size exclusion chromatography. Unlabeled RBDs [A-Raf, RalGDS, and PLC ε1-2, prepared as described previously (Smith and Ikura 2014)] were added to the uniformly 15N- Ile13C δ-labeled K-Ras-nanodisc conjugate at an excess of 1:1.2 KRas:RBD to ensure saturation. To analyze interactions of the RBDs with the bilayer, similar experiments were also performed with uniformly 15N- Ile13C δ-labeled RBDs and unlabeled K-Ras-nanodisc conjugate. Assignment of K-Ras Ile side chains were transferred from the H- Ras-GMPPNP T35S mutant (BMRB entry: 17610) and H-Ras-GDP (courtesy of Dr. Sharon Campbell). The assignment was then completed via site directed mutagenesis of residues with low confidence

119

assignment transfer. Five mutants were prepares including; Ile21Leu, Ile36Leu, Ile100Leu, Ile139Leu, and Ile163Leu.

1H-15 N HSQC, 1H-15 N TROSY-HSQC, and 1H-13 C HMQC spectra were collected at 298.2 K o on a Bruker AVANCE II 800 MHz spectrometer equipped with a 5 mm TCI CryoProbe and a 600 MHz spectrometer with a TCI 1.7 mm MicroCryoProbe, for nanodisc-bound and free samples, respectively. Spectra were processed and analyzed as described in chapter 5.

References:

Smith, M. J. and M. Ikura (2014). "Integrated RAS signaling defined by parallel NMR detection of effectors and regulators." Nat Chem Biol . Tugarinov, V., V. Kanelis and L. E. Kay (2006). "Isotope labeling strategies for the study of high- molecular-weight proteins by solution NMR spectroscopy." Nat Protoc 1(2): 749-54.