Using Light to Observe and Control Cellular Function: Improving Bioluminescence Imaging and Photocontrol of Rho Gtpase Activation States: a Dissertation

University of Massachusetts Medical School eScholarship@UMMS

GSBS Dissertations and Theses Graduate School of Biomedical Sciences

2011-09-30

Using Light to Observe and Control Cellular Function: Improving Bioluminescence Imaging and Photocontrol of Rho GTPase Activation States: A Dissertation

Katryn R. Harwood University of Massachusetts Medical School

Let us know how access to this document benefits ou.y Follow this and additional works at: https://escholarship.umassmed.edu/gsbs_diss

Part of the Amino Acids, Peptides, and Proteins Commons, Cells Commons, and the Cellular and Molecular Physiology Commons

Repository Citation Harwood KR. (2011). Using Light to Observe and Control Cellular Function: Improving Bioluminescence Imaging and Photocontrol of Rho GTPase Activation States: A Dissertation. GSBS Dissertations and Theses. https://doi.org/10.13028/k5sh-t587. Retrieved from https://escholarship.umassmed.edu/ gsbs_diss/569

This material is brought to you by eScholarship@UMMS. It has been accepted for inclusion in GSBS Dissertations and Theses by an authorized administrator of eScholarship@UMMS. For more information, please contact [email protected]. USING LIGHT TO OBSERVE AND CONTROL CELLULAR FUNCTION:

Improving Bioluminescence Imaging and Photocontrol of Rho GTPase Activation

States

A Dissertation Presented

KATRYN RENÉE HARWOOD

Submitted to the Faculty of the

University of Massachusetts Graduate School of Biomedical Sciences, Worcester

In partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

September 30, 2011

Biochemistry and Molecular Pharmacology

USING LIGHT TO OBSERVE AND CONTROL CELLULAR FUNCTION: Improving Bioluminescence Imaging and Photocontrol of Rho GTPase Activation States A Dissertation Presented By KATRYN RENÉE HARWOOD The signatures of the Dissertation Defense Committee signify completion and approval as to style and content of the Dissertation

______Stephen Miller, Ph.D., Thesis Advisor

______William Kobertz, Ph.D., Member of Committee

______Kirsten Hagstrom, Ph.D., Member of Committee

______David Lambright, Ph.D., Member of Committee

______Bruce Branchini, Ph.D., Member of Committee

The signature of the Chair of the Committee signifies that the written dissertation meets the requirements of the Dissertation Committee

______Kendall Knight, Ph.D., Chair of Committee

The signature of the Dean of the Graduate School of Biomedical Sciences signifies that the student has met all graduation requirements of the school

______Anthony Carruthers, Ph.D., Dean of the Graduate School of Biomedical Sciences

Program in Biochemistry and Molecular Pharmacology September 30, 2011 iii

DEDICATION

I dedicate this thesis to my wonderful family--my mother and father, Janice and

Jim Harwood, my brother, Brad Harwood, and my fiancé Mina Seedhom. Your love and support have been immense.

ACKNOWLEDGEMENTS

First, I would like to acknowledge my thesis advisor, Stephen Miller, without whom this thesis would not be possible. He gave me a second chance with my first project: I loved photocaging Rho GTPases and learned so much. I also have to thank him for giving me the opportunity to work on luciferase, which was just so much pure fun experimentally. Your guidance over the years has shaped me as a scientist.

I am very appreciative of my committee members, Kendall Knight, Bill Kobertz,

Kirsten Hagstrom, and David Lambright, who have been very supportive during my time as a graduate student. I am very grateful to Bruce Branchini for serving as my outside committee member. I really appreciate him taking the time to be part of this process.

Also, a huge thanks to Natalie Wayne and her P.I. Dan Bolon for all of their help with protein purification and efforts to generate full length Rho GTPases. Natalie was a great teacher and colleague throughout graduate school. In addition, I must thank Jinal

Patel and Karin Green for all of their help with mass spectrometry over the years.

Thank you as well to my lab mates and to my “step”-lab mates in the Kobertz lab for your invaluable help over the years. Steven Pauff and Karen Mruk, a profound thank you for lending your shoulders, and homes, to me when I needed them most. I also have to thank my incredible family and the great friends I met here at UMass for always being there for me. I love you all. Finally, I thank my amazing fiancé; your support during our time in graduate school has meant the world to me (and a nickel!).$

$ Dee v

ABSTRACT

The dynamic processes that occur at specific times and locations in cells and/or whole organisms during cellular division, migration, morphogenesis and development are critical. When these molecular events are not properly regulated, disease states can develop. Tools that can allow us to better understand the specific events that, when misregulated, result in disease development can also allow us to determine better ways to combat such misregulation. Specifically, tools that could allow us to better visualize cellular processes or those that allow us to control cellular functioning in a spatiotemporal manner could present great insight into the detailed inner workings of cells and/or whole organisms. Where chemistry and biology intersect presents a powerful starting point for the development of such tools.

The first half of this thesis addresses tools to allow the better visualization of cellular events, in particular the intriguing process of bioluminescence and the work that has been done to better understand and optimize its utilization, particularly in living organisms. The novel work presented here details a parallel approach to improve our ability to observe cellular functioning specifically by improving bioluminescence imaging through the generation and characterization of mutant luciferase proteins that can better utilize novel small molecule luciferin substrates.

The second half of this thesis discusses methods that have been developed to better control cellular events through the control of protein activity, specifically a family of proteins called the Rho GTPases. This family’s activation at specific times and vi

locations is essential to proper cellular function and exemplifies the need for spatiotemporal control. Described are methods to control the activation states of the Rho

GTPases to probe their cellular roles in a temporal and spatial manner using photosensitive small molecules. Taken together, the findings described herein demonstrate the application of chemistry to allow for the better observation and control of cellular processes, toward the ultimate goal of improving our understanding of the regulatory processes involved in the control of key factors leading to disease states.

vii

TABLE OF CONTENTS Dedication ...... iii Acknowledgements ...... iv Abstract ...... v Table of Contents ...... vii List of Tables ...... ix List of Schemes ...... ix List of Figures ...... xi List of Symbols and Abbreviations...... xiv Preface...... xviiii

CHAPTER I: Improving Bioluminescence Imaging Background ...... 1 Introduction ...... 1 Naturally occuring luciferases that emit red-shifted light ...... 5 Studies of red emititng native, chimeric and mutant luciferases ...... 7 Theories of bioluminescent color determination ...... 13 BRET as a method to generate red-emitting luciferases ...... 23 Novel luciferin substrates ...... 24

CHAPTER II: Identification of mutant firefly luciferases that efficiently utilize aminoluciferins...... 29 Summary ...... 29 Introduction ...... 29 Materials and Methods ...... 31 Results ...... 37 Discussion ...... 58

CHAPTER III: Improving Bioluminescence Imaging Discussion ...... 64

CHAPTER IV: Photocaging Background ...... 71 Methods to control protein activity ...... 71 Photocaging introduction ...... 74 Direct chemical modification ...... 76 Expressed protein ligation...... 78 Unnatural amino acid mutagenesis ...... 79 Phytochrome fusion proteins ...... 80

viii

CHAPTER V: Rho GTPase Background ...... 82 Introduction ...... 82 Factors that control Rho GTPse activation sates ...... 83 Active and Dominant Negative Rho GTPases ...... 86 Alternate methods to control Rho GTPase activation states ...... 88 Techniques to observe Rho GTPase activity ...... 91

CHAPTER VI: Leveraging a small-molecule modification to enable the photoactivation of Rho GTPases ...... 97 Introduction ...... 97 Materials and Methods ...... 100 Results and Discussion ...... 107

CHAPTER VII: Efforts to introduce caged Rho GTPases into cells ...... 118 Introduction ...... 118 Materials and Methods ...... 120 Results ...... 125 Discussion ...... 130

CHAPTER VIII: Using photocaging to inhibit Rho GTPases ...... 134 Introdction ...... 134 Materials and Methods ...... 137 Results and Discussion ...... 142

CHAPTER IX: Photocontrol of Rho GTPase's Activation States Disucssion ...... 151

Bibliography ...... 153

LIST OF TABLES

Table 1-1 Red-emitting luciferase mutants

Table 1-2 max with novel luciferins and WT luciferase

Table 2-1 Mutants identified in saturating mutagenesis screen

Table 2-2 Apparent Km and Vmax values

Table 2-3 Emission maxima x

LIST OF SCHEMES

Scheme 6-1 Photocages used in CHAPTER VI

LIST OF FIGURES

Figure 1-1 Luciferase mechanism

Figure 1-2 Keto and enol forms of oxyluciferin

Figure 1-3 Resonance delocalization of excited state oxyluciferin

Figure 1-4 Twisted intramolecular charge transfer (TICT) excited state

Figure 1-5 Structures of the emitter

Figure 1-6 Seventh proposed structure of the oxyluciferin emitter

Figure 1-7 Novel alkylamino luciferin structures

Figure 2-1 Preliminary characterization of phenylalanine 247 mutants

Figure 2-2 Dose-response profiles from purified luciferases and luciferases transfected into CHO-K1 cell lysate and in live cells

Figure 2-3 Creation of mutant luciferases

Figure 2-4 Characterization of luciferase mutants

Figure 2-5 Several S347 mutants show selectivity for CycLuc1 over D-Luciferin

Figure 2-6 Burst kinetics and emission profiles

Figure 2-7 Burst kinetic profiles with WT and R218K luciferases

Figure 2-8 Dose-response profiles from luciferin substrates in CHO-K1 cell lysate

Figure 2-9 Dose-response profiles from luciferases in expressed in CHO-K1 cell lysates and in live cells

Figure 2-10 Relative activity of mutant luciferases and novel substrates in vitro, in CHO cell lysate, and in intact CHO cells at 1 µM substrate concentration

Figure 2-11 Relative activity of mutant luciferases and novel substrates in vitro, in CHO cell lysate, and in intact CHO cells at 125 µM substrate concentration xii

Figure 3-1 Additional novel alkylamino luciferin substrates

Figure 3-2 Emission maximum with R218K and novel substrates

Figure 3-3 Combining favorable mutations can increase Vmax with our novel synthetic substrates, but at a cost to Km values

Figure 4-1 Ortho nitrobenzyl uncaging

Figure 5-1 Rho GTPase cycle

Figure 5-2 Biosensors to detect Rho GTPase activation

Figure 6-1 Using photolabile small molecules to control Rho GTPase activity

Figure 6-2 Effector binding domain assay

Figure 6-3 Photocaging of Cdc42

Figure 6-4 Fluorescence polarization assay

Figure 6-5 Uncaging Cdc42-4-HT

Figure 6-6 Specificity of modification and HaloTag recruitment

Figure 6-7 Photocaging of Rac1

Figure 7-1 Westerns following Chariot and BioPORTER treatment

Figure 7-2 Loading of polyarginine-tagged HT protein into HeLa cells

Figure 7-3 Loading of polyarginine-tagged SNAP protein into CHO cells

Figure 7-4 Differential cysteine modification scheme

Figure 8-1 GEF:GTPase binding interface

Figure 8-2 Alternate scheme to use light to inhibit a GTPase

Figure 8-3 Photocaging DN RhoA:GEF nucleotide exchange assays

Figure 8-4 Mass spectrometry of DN RhoA modified with 4 xiii

Figure 8-5 SDS-Page of fractions collected during FPLC purification of Ack-4-HT

Figure 8-6 Mass spectrometry of ACK42 purification products

Figure 8-7 ACK42 and Ack-4-HT are prone to degradation

xiv

LIST OF SYMBOLS AND ABBREVIATIONS

 Alpha ATP Adenosine triphosphate  Beta BCA bicinchoninic acid protein assay BSA bovine serum albumin BRET bioluminescence resonance energy transfer Ca Calcium CPP Cell-penetrating peptide CA Constitutively active CALI Chromophore assisted light inactivation cAMP adenosine 3’,5’-cyclic monophosphate CFP Cyan fluorescent protein CHO-K1 Chinese hamster ovary-K1 CID Chemical inducer(s) of dimerization cLogP The logarithm of the partition coefficient between n-octanol and water log(coctanol/cwater) for a compound, and is a measure of the compound's hydrophobicity cps counts per second DIC Differential interference contrast microscopy DLSA 5’-O- [N-(dehydroluciferyl)-sulfamoyl]adenosine DMEM Dulbecco's Modified Eagle Medium DN Dominant negative DNA Deoxyribonucleic acid DsRed Discosoma red fluorescent protein DTT Dithiothreitol EPL Expressed protein ligation FC Fast-cycling FKBP12 rapamycin-induced FK506 binding protein 12 FPLC fast protein liquid chromatography FRAP FKBP12-rapamycin-associated protein FRB FKBP12 rapamycin binding domain FRET Fluorescence resonance energy transfer GAP GTPase-activating proteins GEF Guanine-nucleotide exchange factors GDP Guanosine diphosphate GFP Green fluorescent protein GTP Guanosine triphosphate GTPS Guanosine gamma thio-phosphate: a non-hydrolyzable GTP analog HBSS Hank’s Buffered Saline solution HT the modified haloalkane dehalogenase HaloTag IPTG Isopropyl β-D-1-thiogalactopyranoside xv

K Potassium kDa kilodalton Km The Michaelis-Menten constant: the substrate concentration at which the reaction rate is half of Vmax max Wavelength maxima L-AMP dehydroluciferyl-AMP LC/ESI-MS liquid chromatography electrospray ionization mass spectrometry LH2 D-luciferin LOV light oxygen voltage LT lethal toxin 6’-MeNHLH2 monomethyl aminoluciferin 6’-Me2NLH2 dimethyl aminoluciferin MLCK myosin light chain kinase MMTS methyl methanethiosulfonate MTSET [2-(trimethylammonium) ethyl] methanethiosulfonate Na Sodium 6’-NH2LH2 aminoluciferin nIR near infrared PAK p21 activated kinase PBD p21-binding domain PBS Phosphate buffered saline PCR polymerase chain reaction PDK-1 3-phosphoinositide-dependent kinase 1 PhyB Phytochrome B Pif3 Phytochrome interacting factor 3 PKA cAMP-dependent protein kinase Raichu Ras and interacting chimeric unit RFP Red fluorescent protein RhoGDI Rho guanine nucleotide dissociation inhibitors RNA Ribonucleic acid RNAi RNA interference ROS Reactive oxygen species SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis siRNA Small interfering RNA SNAP O6-methylguanine-DNA methyltransferase derivative TCEP tris(2-carboxyethyl)phosphine TICT twisted intramolecular charge transfer TMR tetramethylrhodamine TRIP Trio Inhibitory Peptide alpha UV Ultraviolet Vmax The maximum velocity or rate of a reaction catalyzed by an enzyme when all enzyme active sites are saturated with substrate. WASP Wiskott-Aldrich Syndrome Protein WT Wild type xvi

YFP Yellow fluorescent protein xvii

PREFACE

Publications derived from work contained within this thesis:

In CHAPTER II: Harwood, K.R., Mofford, D.M., Reddy, G.R., and Miller S.C. “Identification of mutant firefly luciferases that efficiently utilize aminoluciferins.” Submitted 6/1/2011 to Chemistry and Biology: Accepted 9/21/2011.

In CHAPTER VI: Harwood, K. R. and S. C. Miller (2009). "Leveraging a small-molecule modification to enable the photoactivation of rho GTPases." Chembiochem 10(18): 2855- 2857.

I thank Dave Mofford for conducting the site directed mutagenesis and screens of Y340 and R218 presented in Table 2-1 and for generating a portion of the emission data presented in Figure 3-1, and G.R. Reddy for synthesizing the alkyl aminoluciferin substrates utilized in the studies reported in Chapter II.

CHAPTER 1: Improving Bioluminescence Imaging Background

Introduction

Bioluminescence is a fascinating process by which an organism converts the energy of a chemical reaction into light. While many different species possess bioluminescent properties, including bacteria, fungi, and fish (Greer and Szalay, 2002;

Johnson and Shimomura, 1972; Shimomura, 2006), the most commonly studied have been bioluminescent beetles. As many as 2000 different species of bioluminescent beetle have been observed (Hastings, 1995; Lloyd, 1978a), with the North American firefly

Photinus pyralis the best characterized (Deluca, 1976). These fireflies use the flashing light derived from the bioluminescent reaction to signal to potential mates, with different species adopting distinct patterns of flashes (Lloyd, 1978b).

The bioluminescence of the firefly relies on the chemical reaction between the luciferin substrate, ATP and oxygen in the context of the binding pocket of the luciferase enzyme to generate light (Figure 1-1). Specifically, in the presence of ATP-Mg2+, luciferase converts luciferin to luciferyl-AMP, with the displacement of inorganic pyrophosphate. Removal of a proton from the C-4 carbon of the adenylate and addition of molecular oxygen transforms luciferyl-AMP into the very reactive dioxetanone intermediate, which possesses a strained four-membered cyclic ring that spontaneously breaks down to generate the excited state oxyluciferin and carbon dioxide (Deluca, 1976;

White et al., 1971; White et al., 1980). Light is emitted upon relaxation of the excited state oxyluciferin molecule back to the ground state (Suzuki and Goto, 1972; Suzuki et 2

al., 1969). In addition, a dark reaction pathway in which luciferyl-AMP reacts with oxygen to generate the oxidized product dehydroluciferyl-AMP can also occur (Rhodes and McElroy, 1958).

Figure 1-1. Luciferase mechanism. Firefly luciferase catalyzes the formation of an activated AMP ester of its native substrate, D-luciferin. Subsequent oxidation within the luciferase binding pocket generates an excited state oxyluciferin molecule that is responsible for light emission.

The quantum yield of luciferase catalyzed light emission is very high and was originally reported to be 0.88 (Seliger and Mc, 1960), meaning that the ratio of reacted luciferin to photons of light emitted was initially thought to be almost 1:1. Using a newly 3

developed total-photon-flux spectrometer, which allowed for a more accurate determination of the quantum yield, Ando and colleagues reported the quantum yield from the American firefly Photinus pyralis to be a more modest but still quite high ~0.41

(Ando et al., 2008; Kazuki et al., 2007). This is especially impressive when compared to the quantum yields possible for most chemiluminescent reactions, which are typically

~0.01 (Roda et al., 2009).

Firefly luciferase is a member of the acyl-adenylate/thioester-forming superfamily of enzymes that also includes many acyl-CoA ligases (Babbitt et al., 1992; Chang et al.,

1997; Suzuki et al., 1990). Crystal structures of the luciferase proteins and other acyl- adenylate forming enzymes reveal a common structural motif made up of two domains, a larger N-terminal domain and a smaller C-terminal domain, linked by a flexible linker peptide (Conti et al., 1997; Franks et al., 1998; Gulick et al., 2003; May et al., 2002;

Nakatsu et al., 2006). Luciferase functions as a monomer of about 61 kDa (de Wet et al.,

1985), and does not possess any functionally significant posttranslational modifications or disulfide bonds (Alter and DeLuca, 1986; DeLuca et al., 1964). Further, the full length enzyme is fully active directly following translation from its mRNA (Wood et al., 1984), and protein expressed from E. coli behaves in a manner that is indistinguishable from the native luciferase (Hill et al., 1986).

Luciferase-based assays have broad applications in the study of mammalian biology. Due to the role ATP plays in the luminescent reaction, firefly luciferase has been widely used as a tool to measure ATP concentrations (Ludin, 1981). Expression of luciferase in E. coli (de Wet et al., 1985) and cloning of the full-length gene (Cohn et al., 4

1983; de Wet et al., 1987) expanded the application of luciferase to the study of protein metabolism and gene expression in mammalian cells. Furthermore, unlike fluorescence- based methods, bioluminescence does not require exogenous illumination. This means that in imaging studies using luciferase, background is greatly reduced, improving sensitivity.

The many applications of firefly luciferase assays have included the detection of bacteria and environmental toxins (Alexander et al., 2000; Campbell and Sala-Newby,

1993; Kricka, 2000; Roda et al., 2004; Squirrel et al., 2002), gene expression and regulation studies (de Ruijter et al., 2003; Gelmini et al., 2000; Gould and Subramani,

1988), monitoring of biomolecular interactions using bioluminescence resonance energy transfer (BRET) (Arai et al., 2002; Prinz et al., 2006; Xia and Rao, 2009; Yamakawa et al., 2002), and many others. Firefly luciferase has also been used for genetic regulation studies and the monitoring of tumor growth and metastasis in whole organisms (Contag and Bachmann, 2002; Contag et al., 1997; Rehemtulla et al., 2000; Yu et al., 2003).

The high relative quantum yield of light emission of the luciferase:luciferin reaction and its inherently low background make luciferase bioluminescence a particularly powerful tool for cellular imaging. However, at physiological pH, firefly luciferase emits yellow-green light (max 560 nm). Light of this wavelength exhibits poor tissue penetrance in whole organisms, and further can be absorbed by hemoglobin, melanin and other pigmented cellular components. Light above 600 nm in wavelength, however, is absorbed less strongly than shorter wavelength light and can travel a greater distance through tissue (Cheong et al., 1990; Villalobos et al., 2007; Zhao et al., 2005). 5

The emission of longer wavelengths of light from luciferase would greatly improve the utilization of bioluminescence imaging in intact tissues and whole organisms.

Naturally occurring luciferases that emit red-shifted light

In nature, beetles have been observed to emit light of many different colors ranging from green (~540 nm) to red (~635 nm) (Hastings, 1996; McElroy and Seliger,

1966). It has been suggested that the evolution of luciferases that emit in this wide range was to optimize detection under certain photic conditions or for distinct biological functions (Lall et al., 1980). The differences in bioluminescence color seen in the different beetle species are caused by the natural species variations in luciferase structure, as all use the same luciferin substrate, D-luciferin (Hastings, 1996; McElroy and Seliger,

1966; Wood, 1995; Wood et al., 1989b). The different beetle luciferases that have been cloned include those from fireflies (Lampyridae), click beetles (Elateridae), railroad worms (Phengodidae), and Japanese luminous beetles (Rhagophthalmidae) (Ugarova and

Brovko, 2002; Welsh and Noguchi, 2010). Of these, fireflies emit green-yellow light

(Biggley et al., 1967; Seliger et al., 1964), click beetles emit green-orange light (Biggley et al., 1967; Colepicolo et al., 1986), and railroad worms emit green-red light (Viviani and Bechara, 1993, 1997).

The Jamaican click beetle Pyrophorus plagiophthalamus is particularly interesting in that it has two sets of light producing organs, a ventral organ and a pair of organs on the dorsal surface of the prothorax. While the dorsal pair emits light ranging from green to yellow green, the ventral organ emits that ranging from green to orange, 6

with overall emission ranging from 547-594 nm (Biggley et al., 1967). Another fascinating species are the Phrixothrix railroad worms. They emit yellow-green light

(max 542-574) from dorsal lateral lanterns along the body, and red light (max 609-635) from cephalic and post cephalic organs (Viviani and Bechara, 1993, 1997).

Other notable non-beetle luciferases that have been cloned are those from several marine species, including the sea pansy Renilla reniformis (max 480) (Lorenz et al.,

1991) and the copepod Gaussia princeps (max 480) (Tannous et al., 2005). These differ from the beetle luciferases in that they are smaller, ATP-independent, and their substrate is not D-luciferin but coelenterazine. From a reagent perspective, relative to D-luciferin coelenterazine is more toxic and expensive, and less stable and soluble. It also is more autoluminescent than D-luciferin, increasing the background observed and reducing sensitivity (Welsh and Noguchi, 2010).

Interestingly, under certain conditions, the light emitted from firefly luciferases is shifted to the red as well. Some of these conditions include a shift in pH from neutral to acidic, temperature increase, or the presence of ATP analogs or certain divalent metal ions such as Hg2+ and Zn2+ (Deluca, 1976; McElroy et al., 1965; Rosendahl et al., 1982;

Seliger and McElroy, 1964; Viviani, 2002; Viviani et al., 2008; Wood et al., 1989a;

Wood et al., 1989c). Unlike firefly luciferases, railroad worm and click beetle luciferases are not sensitive to pH dependent changes with respect to the color of emitted light

(Viviani and Bechara, 1995). To begin to understand the factors that lead to the different colors emitted from different species’ luciferases, as well as the effect of pH on the color change seen with firefly luciferase under acidic conditions, much work has been done 7

with native species of luciferase, as well as genetically engineered chimeras and mutants proteins.

Studies of red emitting native, chimeric and mutant luciferases

While it was first observed in 1964 that the click beetles of Jamaica (Pyrophorus plagiophthalamus) displayed different profiles of light emission than North American fireflies (Seliger and McElroy, 1964), due to the limited availability of the beetles themselves, it was not until several years later that Wood and colleagues isolated cDNA from them and determined that they coded for four different luciferase homologs. These homologs possess between 94-99% amino acid identity and each displays a different color of light emission ranging from green to orange (max 546, 560, 578, 593 nm) (Wood et al., 1989c). Chimeras, or “rearrangement hybrids”, of these proteins were generated by exchanging restriction fragments from the different clones and single mutants were made by site directed mutagenesis to yield 31 recombinant luciferases in an attempt to determine which specific amino acid changes could account for the color changes observed (Wood, 1990; Wood et al., 1989b). From this analysis, two sets of amino acid changes were found to cause a red shift in the color of light emitted, the pair

R223E/L238V and the set S247G/D352V/S358T. Similarly, chimeric mutants of the luciferases from the fireflies Pyrocoelia miyako (max 550 nm) and Hotaria parvula (max

568 nm) were made to assess the fragments responsible for the difference in color of light emitted, and studies with these chimeras indicated that residues in the fragment between

Val209 and Ala318 were responsible for the difference in the spectra between the two 8

species’ luciferases (Ohmiya et al., 1996). Chimeric luciferases generated by Viviani and colleagues made up of residues 1-344 from the red light-emitting luciferase of railroad worm Phrixothrix hirtus (max 623 nm) and 345-545 from the green light-emitting luciferase of railroad worm Phrixothrix viviani (max 548 nm) showed that for these luciferases, only residues 1-344 were responsible for the color of light emitted (Viviani and Ohmiya, 2000).

In other studies, it was shown that random and site-directed mutagenesis of a single amino acid in different species’ luciferases could generate proteins that emitted red-shifted light at pH 7.8. Mutagenesis of Luciola cruciatai luciferase (max 562 nm) generated the red-shifted point mutants P452S (max 595 nm), S286N (max 607 nm),

G326S (max 609 nm), and H433Y (max 612 nm) (Kajiyama and Nakano, 1991). Based on the results Kajiyama and colleagues obtained from their random mutagenesis experiments, Mamaev and colleagues mutated serine 286 (Ser284 in Photinus pyralis) in

Luciola mingrelica firefly luciferase (max 582) to a number of residues and found that all of these mutations red-shifted light emission: S286K max 608 nm, S286Q max 609 nm,

S286Y max 613 nm, and S286L max 619 nm (Arslan et al., 1997; Mamaev and Laikhter,

1996). Random mutagenesis of Hotaria parvula luciferase also generated the point mutation H433Y, with max 610 nm (Ueda et al., 1996). Furthermore, the mutation

R215S in the green emitting railroad worm luciferase of Phrixothrix viviani was shown to shift its emission from 548 nm to 588 nm, but to have no effect on the light emitted from the red emitting luciferase of Phrixothrix hirtus (Viviani and Ohmiya, 2000). 9

In addition to random mutagenesis studies, work was also done to mutate residues that were conserved between the different beetle species to determine the roles played by different residues in color determination. Viviani and colleagues identified residues that differed between the pH sensitive Photinus pyralis luciferase and the pH insensitive click beetle and railroad worm luciferases and performed mutagenesis studies to attempt to explain why certain luciferases emit longer wavelength light at acid pH and others emission does not change. From this work, they found that substitution of many residues in the loop between residues 223-235, as well as at Ala243 (pH insensitive luciferase numbering) had dramatic effects on the emission properties from the various species’ luciferases studied (Viviani et al., 2001; Viviani et al., 2008; Viviani et al., 2002).

It was also observed that the naturally red emitting luciferase from Phrixothrix hirtus possesses an arginine, R353, which is deleted in other luciferases. Tafreshi and colleagues demonstrated that insertion of that corresponding residue into the luciferase from Lampyris turkestanicus (R356), a green emitting luciferase (max 555), generated a luciferase with a bimodal spectrum with an emission maximum in the red (max 615) and a small shoulder in the green (max 560) region (Tafreshi et al., 2007). Additionally, this insertion only minimally impacted the relative activity of the luciferase. Based on this work, Moradi and colleagues introduced four residues at the same position in Photinus pyralis luciferase that were similar in size to arginine, but which had different charges

(Arg, Lys, Glu, and Gln) (Moradi et al., 2009). They demonstrated that while the insertion of a residue with a positive side chain caused a red shift in the light emitted

(max 608), inserting a negative or neutral residue had little effect on emission. 10

Shapiro and colleagues reported on the generation of red emitting luciferases using error prone PCR in Photinus pyralis. Interestingly, the resulting red shifting mutations were all found to be located in one region of the N-terminus of luciferase located away from the proposed active site (Shapiro et al., 2005). Of the mutants generated, the brightest and most red emitting at neutral pH were A850G/S284G (max

603) and A848G/Q283R (max 609).

While many studies were done to pinpoint specific residue(s) that determine the color emitted by luciferase, the fact that mutations at so many locations were able to red shift the light emitted suggested that the solution might not be so simple. With the publication of the structures of firefly Photinus pyralis luciferase (Conti et al., 1996) and another acyl-adenylate forming enzyme, phenylalanine-activating enzyme (Conti et al.,

1997), modeling of the luciferase active site made rational site-directed mutagenesis to assess the structure-function basis of color determination possible (Branchini et al., 1998;

Sandalova and Ugarova, 1999). Based on the working model they developed, Branchini and colleagues performed several mutagenesis studies in Photinus pyralis luciferase of residues in the proposed binding site, including highly conserved residues (Branchini et al., 1999; Branchini et al., 1998; Branchini et al., 2001; Branchini et al., 2003). Mutation of many of these residues was found to red shift the bioluminescence maximum at neutral pH (Table 1-1). Further studies to better understand the adenylation and oxidation partial reactions in the overall production of light from firefly luciferase showed that the mutations K529A and K433A/K529A generated luciferase with broad red-shifted emission spectra (Branchini et al., 2005b). 11

Table 1-1. Red-emitting luciferase mutants. Mutants discovered in several mutagenesis screens by Branchini and colleagues (Branchini et al., 1999; Branchini et al., 1998; Branchini et al., 2001; Branchini et al., 2003; Branchini et al., 2005b) that red shift light emission from firefly luciferase. Firefly luciferase point mutants are shown in the left column and wavelength maxima (nm) for each mutant are shown in the right column.

Subsequent to the identification of many mutations that could red shift light emission from luciferase, efforts moved towards the generation of red shifted luciferases with improved properties, as often the mutations which red shift light emission also result in luciferases with decreased bioluminescent activity. Based on sequence information from a yellow-green emitting luciferase from Pyrophorus plagiophthalamus, Almond and colleagues at Promega performed further mutagenesis to yield a bright thermostable red emitting variant (max 613) optimized for used in dual reporter assays in mammalian cells

(Almond et al., 2003).

Branchini and colleagues performed site directed and random mutagenesis in

Photinus pyralis luciferase looking for a mutant that would emit red light but also maintain its activity, and identified the triple mutant (V241I/S284T/I351A) (Branchini et al., 2004a). Further analysis showed that the S284T mutation was sufficient for the red light emission, narrow bandwidth and favorable kinetic properties of this mutant, and that it could be used in dual reporter studies with the blue-shifted luciferase mutant

V241I/G246A/F250S (Branchini et al., 2005a). To improve the S284T mutant further for imaging and reporter gene application, the mutations

T214A/A215L/I232A/F295L/E354K were added, generating a thermostable pH insensitive red-emitting variant (Ppy RE-TS max 610 nm) with a narrow emission bandwidth (Branchini et al., 2007a). Further mutagenesis of Ppy RE-TS red shifted the emission of this mutant even further, generating the mutants Ppy RE8 and Ppy RE9 (max

617 nm) (Branchini et al., 2010a). Ppy RE8 has been used successfully in in vivo bioluminescence imaging studies (Mezzanotte et al., 2011). 13

Theories of bioluminescence color determination

Several theories have been proposed to account for the wide range of colors of light emitted by different luciferases. It was originally suggested that keto-enol tautomerization of the oxyluciferin molecule was responsible for the different colors of light emitted, where the ketonic species of oxyluciferin is responsible for red light emission and the enolic species is responsible for yellow-green light emission (Figure 1-

2), possibly with the involvement of a basic residue in the luciferin-binding pocket

(Viviani and Bechara, 1995; White and Branchini, 1975; White et al., 1969; White et al.,

1971; White and Roswell, 1991; White et al., 1980). This hypothesis was supported by the observation that adenyl-luciferin that had undergone spontaneous oxidation in aqueous buffer (White et al., 1971) and an analogue of luciferin that cannot be enolized,

5-dimethyl luciferin (White et al., 1969; White et al., 1971), both emitted red light.

However, more recently Branchini and colleagues showed that in the presence of a synthetic luciferin analog (5,5-dimethyloxyluciferin) that is constrained to the keto form, green light is emitted by the firefly luciferase from Photinus pyralis, while red light is emitted by the luciferase from the click beetle Pyrophorus plagiophthalamus, suggesting that keto-enol tautomerization is not responsible for the observed color modulation

(Branchini et al., 2002; Branchini et al., 2004b).

Figure 1-2. Keto and enol forms of oxyluciferin. One theory of color determination suggests that the keto form of the excited oxyluciferin emitter generates red light with the luciferase protein and that the enol form generates yellow-green light.

In addition, Branchini and colleagues also suggested that changes in resonance- based charge delocalization of the excited state oxyluciferin plays a role in modulating the color of light emitted (Branchini et al., 2002; Branchini et al., 2004b). Specifically, key conserved amino acids and interactions with AMP in the luciferase-binding pocket play a role in controlling the charge delocalization of the anionic keto form of the excited oxyluciferin molecule (Figure 1-3). An excited state species of lower energy can be generated by resonance stabilization of the anion and extension of the -electron skeleton, resulting in red light emission. In the context of the binding pocket, specific amino acids and AMP minimize charge delocalization and maintain the higher energy anionic keto form of the excited state oxyluciferin, resulting in the emission of yellow- green light.

Figure 1-3. Resonance delocalization of excited state oxyluciferin. One theory suggests that changes in resonance-based charge delocalization of the excited state oxyluciferin play a role in modulating the color of light emitted from luciferase.

Based on molecular orbital calculations, McCapra and colleagues suggested that the conformation of the luciferin substrate in the binding pocket determines the color of light emitted. This theory surmises that rotation around the C2-C2’ bond of oxyluciferin alters the conformation of its keto emitting form and results in color variation.

Specifically, red light emission results from the twisting of the aromatic thiazole and benzothiazole rings of oxyluciferin to a 90˚ angle with respect to the C2-C2’ bond, resulting in a minimum energy conformation termed a twisted intramolecular charge transfer (TICT) excited state (Figure 1-4). Yellow-green light is emitted by the higher energy excited state oxyluciferin where the molecule is held planar and bond rotation is prevented (McCapra, 1997; McCapra et al., 1994). This has since been refuted by multiple studies (Nakatani et al., 2007; Nakatsu et al., 2006; Orlova et al., 2003; Yang and Goddard, 2007).

Figure 1-4. Twisted intramolecular charge transfer (TICT) excited state. Rotation around the C2-C2’ bond, where the aromatic thiazole and benzothiazole rings of oxyluciferin are at a 90˚ angle with respect to the C2-C2’ bond, generates the minimum energy conformation, and red emitting species, of excited state oxyluciferin.

One of these refuting theories sites the rigidity of the active site in determining the color of light emitted. This hypothesis is based on crystallography studies of a WT and a red-shifted mutant (S286N) (Kajiyama and Nakano, 1991) of Japanese Genji-botaru

(Luciola cruciata) luciferase in complex with a high-energy intermediate analog 5’-O-[N-

(dehydroluciferyl)-sulfamoyl]adenosine, DLSA. As compared to structures of WT protein in complex with the reaction products oxyluciferin and AMP, a conformational change is observed in the WT protein complexed with DLSA, but not the red shifted mutant. This conformational change involves the movement of a hydrophobic residue

(Ile288) closer to the benzothiazole ring of the substrate, resulting in a “closed” form, while the S286N mutant maintains an “open” form like that seen with the WT protein and the reaction products oxyluciferin and AMP. Nakatsu and colleagues suggest that the conformational changes that occur in the WT enzyme during catalysis generate a very hydrophobic, rigid environment in the active site, which minimizes energy loss resulting it the emission of high-energy yellow green light. They also hypothesize that because

S286N remains in the “open” form, the active site is less hydrophobic and less rigid and 17

energy loss can occur during catalysis, potentially through reorganization or thermal relaxation, resulting in the emission of lower energy longer wavelength red light

(Nakatsu et al., 2006). Also, in the crystal structure of both WT and S286N in complex with the luciferyl adenylate intermediate analogue, the benzothiazole and thiazole rings of the molecule are almost coplanar. Nakatsu and colleague argue that this refutes the hypothesis of C2-C2’ bond rotation as a key factor in color determination.

Computational studies using high density functional theory, where neutral and anionic forms of both the parent luciferin and of the keto- and enol- derivatives of oxyluciferin were examined, also suggest that the C2-C2’ bond rotation hypothesis is unlikely (Orlova et al., 2003). The length of the C2-C2’ bond predicted in these studies reflects the conjugation across it and would make rotation about the bond difficult.

Orlova and colleagues suggest that the specific method used by another group who claimed to verify the twisted keto-form (Stewart, 1999) does not fully appreciate the nature of conjugated single bonds (Cramer, 2002; Jensen, 1999). Instead, they attribute the polarity of the excited oxyluciferin molecule in the context of the environment of the binding pocket of luciferase as key in color determination. Specifically, they postulate that differences in the strength of amino acid interactions with terminal H-O groups on oxyluciferin can alter the color from either the keto or the enol form of the molecule, allowing for a range of colors of light to be emitted.

In fact, many groups have cited polarity as the key factor in determining the color of light emitted, where the higher the polarizability, the larger the red shift in bioluminescence (Gandelman et al., 1993). For example, it has been shown that the 18

introduction of residues that increase polarizability of the emitter sites correlates with a red shift in light emitted (Ugarova and Brovko, 2002) and that the introduction of residues that decrease polarity of the emitter site can blue shift light (Branchini et al.,

2007b).

In efforts to generate a sensor for cyclic AMP-dependent protein kinase phosphorylation events, Sala-Newby and colleagues mutated residue Val217 to the basic residue Arg in Photinus pyralis luciferase, creating the phosphorylation site RRFS. They found that while the V217R mutant exhibited a green shifted emission profile with respect to the native protein, phosphorylation of the V217R mutant by the catalytic subunit of protein kinase A mutant induced a red-shift in the light emitted (Sala-Newby and Campbell, 1991). Also, fluorescence studies with 8-anilinio-1-napthene sulfonic acid and 6-p-toluidino-2-napthene sulfonic acid showed that in the naturally occurring red light emitting luciferase from Phrixothrix hirtus, the luciferin binding pocket is more polar than that of green light emitting Phrixothrix viviani’s luciferase (Ugarova and

Brovko, 2002; Viviani et al., 2006; Viviani and Ohmiya, 2006).

In addition to the work done by Orlova and colleagues, many other theoretical studies point to polarity as the main factor determining the color of light emitted by luciferase (Hirano et al., 2009; Liu et al., 2008; Nakatani et al., 2007; Navizet et al., 2010;

Orlova et al., 2003; Silva and Da Silva, 2011). Hirano and colleagues performed a thorough examination of the spectroscopic properties of the phenolate anion generated from 5,5-dimethyloxyluciferin, which was intended to represent the keto form of the oxyluciferin phenolate anion, in different solvents with added organic base. Based on 19

this, they suggested that in addition to the polarity of the luciferin binding pocket, the strength of the covalent nature of a hydrogen bond of the phenolate oxygen of the keto form of the oxyluciferin phenolate anion with a protonated basic moiety in the active site is a main factor in color determination (Hirano et al., 2009).

Several of the theoretical studies also claim to disprove one or more of the theories of color determination. Nakatani and colleagues used quantum and molecular mechanical calculations as well as the symmetry-adapted cluster/symmetry-adapted cluster−configuration interaction method in combination with crystallography data (Conti et al., 1996) and models established in other experimental studies (Branchini et al., 1999;

Branchini et al., 1998; Branchini et al., 2001) to analyze the factors that contribute to color determination in firefly luciferase (Nakatani et al., 2007). They claim that their studies refute many of the proposed theories on color determination. Like other groups, they claim that twisting about the C2-C2’ bond is unlikely to be a factor in color determination as the TICT state can easily relax to a coplanar state. Also, while they observe that emission energy of the enol form of oxyluciferin is similar to that of the keto form, they claim that the transition from keto to enol would be unlikely based on the energetic instability of the enol form in the context of the luciferase binding pocket.

Additionally, they claim that while the protonation state of the 6’ oxygen of oxyluciferin contributes to color determination, based on their studies it would not be energetically favorable for the protonation to involve a transfer from residue Arg218 to the keto form of oxyluciferin, as had been postulated. Finally, they claim that based on their 20

calculations regarding the ground and excited state resonance structures of oxyluciferin, the excited state structures would be insensitive to the local environment.

Navizet and colleagues examined the “open” and “closed” X-ray structures of

Luciola cruciata luciferase (Nakatsu et al., 2006) to determine the mechanism of firefly color determination using a multiconfigurational reference second-order perturbation theory-molecular mechanics study for the first time (Navizet et al., 2010). From these studies, they claim that the different sizes of the luciferin binding pocket in the “open” and “closed” forms does not impact the relative structures of the oxyluciferin substrate enough to cause the change in color observed, refuting Nakatsu and colleagues hypothesis regarding the mechanism of color determination.

Oxyluciferin is usually thought to be able to exist as one of six chemical forms

(Figure 1-5). In general, there continues to be an ongoing debate regarding the species of the emitter and spectral determination. While from 1971-2002 it was generally thought that both the keto and enol/enolate forms of oxyluciferin were the light emitters (Ugarova and Brovko, 2002; White and Branchini, 1975; White et al., 1971; White and Roswell,

1991; White et al., 1980), from 2002-2009 the keto form of the oxyluciferin phenolate anion was generally considered the light emitter in the context of the firefly (Branchini et al., 2002; Branchini et al., 2004b; Hirano et al., 2009; Liu et al., 2008; Nakatani et al.,

2007; Navizet et al., 2010).

Figure 1-5. Structures of the emitter. The six different potential forms of the excited state oxyluciferin emitter.

However, based on NMR and fluorescence spectra of oxyluciferin and its 5- methyl analog, Nuamov and colleagues suggested that the enol form of excited state oxyluciferin can also play a role in emission and additionally that it is a culmination of the effects of pH, active site polarity, the presence of ionic species in the binding pocket, and  stacking that ultimately determine the color of light emitted (Naumov et al.,

2009). They later did similar studies using a 6’ dehydroxylated derivative of oxyluciferin to determine the spectra-structural effects of the equilibrium between the keto-enol- enolate and phenol-phenolate forms of oxyluciferin and claim that the phenol-enolate form of excited state oxyluciferin is a yellow emitter (Naumov and Kochunnoonny,

2010). Other recent studies suggest that the phenolate-keto oxyluciferin, but not 22

phenolate-enol oxyluciferin, is the yellow-green emitter (Min et al., 2010a), that the phenolate-keto oxyluciferin is both the yellow-green and red emitter (da Silva and daSilva, 2011; Milne et al., 2010), and that the phenolate-keto oxyluciferin is the only emitter in natural firefly bioluminescence (Chen et al., 2011). Additionally, Min and colleagues even suggested that a seventh species of oxyluciferin is formed when the anionic keto form is protonated on the oxygen of the thiazole ring (Figure 1-6), and this species could be the red emitter (Min et al., 2010b). Navizet and colleagues (Navizet et al., 2010) claim that because many of the theoretical methods used to investigate the emission spectra of oxyluciferin are done in vacuo and at the density function level of theory, they may fail to correctly account for contributions from residues in the binding pocket during the reaction and may not properly consider the charge-transfer states. The degree to which these different factors are considered may in part explain the discrepancies between the different conclusions made from theoretical studies.

Undoubtedly, debate on this topic will continue.

Figure 1-6. Seventh proposed structure of the oxyluciferin emitter. Proposed structure of the red emitting species of excited state oxyluciferin where the thiazole oxygen is protonated.

BRET as a method to generate red-emitting luciferases

An alternative method that has been used to generate red-emitting luciferases takes advantage of BRET. Aria and colleagues used the firefly Hotaria parvula luciferase tagged with GST and Discosoma red fluorescent protein (DsRed) tagged with protein G as a red emitting BRET pair, brought together by the addition of an anti-GST antibody

(Arai et al., 2002). Another group modified the blue emitting “sea firefly” Cypridina luciferase with a far-red fluorescent indocyanine derivative to generate a BRET pair that would produce far-red luminescence (Wu et al., 2009a). Similarly, Branchini and colleagues generated a red emitting Photinus pyralis mutant Ppy RE8 with two surface exposed cysteines that could be modified with Alexa Fluor near infrared (nIR) dyes, allowing BRET to occur and resulting in nIR light emission (Branchini et al., 2010b).

They went on to use a similar approach to generate a fusion of RFP and a thermostable

Photinus pyralis luciferase variant (Branchini et al., 2011). Modification of the RFP with nIR Alexa Fluor dye allowed for sequential BRET-FRET energy transfer, resulting in nIR light emission.

Other groups coupled quantum dots to a mutant of Renilla reniformis luciferase to produce a BRET pair that emits light in the red to far-red region (Ma et al., 2010; So et al., 2006). More recently, Dragulescu-Andrasi and colleagues fused a variant of Renilla luciferase to FKBP12 rapamycin binding domain (FRB) and two different RFPs to rapamycin-induced FK506 binding protein 12 (FKBP12). In the presence of the Renilla substrate coelenterazine and rapamycin, which brings the FRB and FKBP12 together,

BRET occurs, resulting in the emission of red light. (Dragulescu-Andrasi et al., 2011) 24

However, approaches like these that use the Renilla luciferase suffer from the same limitations intrinsic to the coelenterazine substrate mentioned above.

Rather than modifying the luciferase protein itself, another group modified the luciferin substrate instead. Specifically, Takakura and colleagues modified a luciferin analog with Cy5, a nIR cyanine fluorescent dye. BRET between Cy5 and the luciferin analog results in nIR light emission when reacted with firefly luciferase (Takakura et al.,

2011).

Novel luciferin substrates

While much work has been done to characterize native red-emitting luciferases, create luciferases that emit red light, and define the factors that cause color change, since the late 1960’s, less has been done to generate novel luciferin substrates. One of the first reported analogs of D-luciferin to interact with firefly luciferase to produce light was aminoluciferin (6’-NH2LH2), where an electron-donating amino group replaced the 6’ hydroxyl normally found on luciferin (White et al., 1966). This novel luciferin substrate red shifted the light emitted from firefly luciferase (max 605) and unlike D-luciferin, this shift was independent of pH. However when the 6′-hydroxyl of luciferin is removed, or replaced with 6′-methoxy, ethoxy, or N-acetylamino groups, although the luciferase proteins still bind the luciferins despite the bulky additions, no light is emitted (Denburg and McElroy, 1970). White and colleagues also generated three hydroxy positional isomers of luciferin and a di-hydroxy analog, 4-hydroxyluciferin. Of these, only 4- 25

hydroxyluciferin was positive for light emission with luciferase, and the light emitted with this analog was also red (White and Worther, 1966).

More recent work has included the generation of a non-benzothiazole bioluminescent substrate for firefly luciferase, D-quinolylluciferin, which shifted the wavelength of light emitted to the red (max 608 nm) (Branchini, 2000; Branchini et al.,

1989). Like 6-NH2LH2, the red shift observed with D-quinolylluciferin was independent of pH. Takakura and colleagues made quinolyldimethylaminoluciferin that emits light at

max 601 nm with firefly luciferase (Takakura et al., 2010).

Recently, our group generated novel synthetic alkylamino luciferin analogs that result in the emission of red-shifted light from WT firefly luciferase by replacing the 6’ phenolic hydroxyl of D-luciferin with electron-donating acyclic monomethyl and dimethyl alkylamino groups, as well as cyclic alkylamino groups (Figure 1-7) (Reddy et al., 2010). When tested with WT luciferase, these synthetic aminoluciferin substrates shifted the emission wavelength of emitted light to varying degrees (Table 1-2), with 6’-

Me2NLH2 shifting the emission wavelength most toward the red at 623 nm. CycLuc1 and CycLuc2 exhibited superior light output and were shown to be more efficient light emitters than their respective acyclic counterparts. This is likely because the acyclic aminoluciferins possess a reduced quantum yield in relation to cyclic substrates. In the acyclic aminoluciferins, there is the potential for poor overlap of the lone pair of nitrogen with the conjugation of the rest of the luciferin as a result of free rotation around the aryl- nitrogen bond, particularly in the context of the binding pocket where molecular interactions between the enzyme and substrate might stabilize such a conformer. This 26

likely reduces the quantum yields of these substrates. Cyclic aminoluciferins, on the other hand, likely have a relatively higher quantum yield because they are conformationally rigid and possess reduced bond rotation as a result of the cyclization.

Figure 1-7. Luciferin substrates. D-luciferin, aminoluciferin, and four novel acyclic and cyclic alkylamino luciferin molecules that, with the luciferase protein, generate red light (Reddy et al., 2010).

Table 1-2. max with novel luciferins and WT luciferase. Wavelength maxima values (nm) with WT luciferase and D-luciferin, aminoluciferin and novel aminoluciferin substrates (Reddy et al., 2010).

While the alkylamino substrates exhibited a red shift in the light emitted from WT luciferase relative to D-luciferin, a significant reduction in the overall light output for the aminoluciferins relative to D-luciferin was also observed. Rapid-injection (Branchini et al., 1998) of the luciferase protein into buffer containing the luciferin analogs revealed an initial burst of light, followed by reduced sustained light output for all of the alkylamino substrates tested. This indicates that while these synthetic luciferins can be rapidly converted to light-emitting oxyluciferin, they are subsequently subject to product inhibition, due to the high affinity of the reacted non-light producing product. The burst kinetic behavior of even WT firefly luciferase and D-luciferin is well-known (Fraga,

2008).

Ultra-Glo, a highly mutated luciferase from Promega, has been shown to be resistant to inhibition during high-throughput screening assays (Auld et al., 2009; Hall et al., 1998) and by substrates known to inhibit WT luciferase, particularly in the presence 28

of P450-Glo buffer (Woodroofe et al., 2008). Similarly, Ultra-Glo luciferase in P450-Glo buffer is able to effectively utilize our novel alkylamino luciferin substrates, relieving the burst kinetic profiles observed with the WT protein. However, Ultra-Glo is not available as a genetic construct and in many applications, requires the P450-Glo buffer to exhibit sustained light output (Prescher and Contag; Shinde et al., 2006). Efforts toward the generation of luciferase mutants that are able to produce sustained light emission without product inhibition with our novel alkyl aminoluciferins, which can also function under physiological conditions, will prove advantageous for studies within cells and whole organisms.

CHAPTER II: Identification of mutant firefly luciferases that efficiently utilize

aminoluciferins

Summary

Firefly luciferase-catalyzed light emission from D-luciferin is widely used as a reporter of gene expression and enzymatic activity both in vitro and in vivo. Despite the power of bioluminescence for imaging and drug discovery, light emission from firefly luciferase is fundamentally limited by the physical properties of the D-luciferin substrate.

We and others have synthesized aminoluciferin analogs that exhibit light emission at longer wavelengths than D-luciferin and have increased affinity for luciferase. However, although these substrates can emit an intense initial burst of light that approaches that of

D-luciferin, this is followed by much lower levels of sustained light output. We have previously postulated that this behavior is due to product inhibition. Here we describe the creation of mutant luciferases that yield improved sustained light emission with aminoluciferins in both lysed and live mammalian cells, allowing the use of aminoluciferins for cell-based bioluminescence experiments.

Introduction

Light emission from firefly luciferase is fundamentally limited by its access to D- luciferin and the inherent photophysical properties of the D-luciferin substrate (Figure 1-

1) (Reddy et al., 2010). Replacement of the 6’-hydroxyl group of D-luciferin with a 6’- amino group results in red-shifted light emission (White et al., 1966) and higher affinity for luciferase, but lower maximal light emission and lower cell-permeability (Shinde et 30

al., 2006). Although D-luciferin is the superior substrate for maximal light emission under most conditions, the unique chemistry of 6’-aminoluciferin has expanded the scope of luciferase applications. For example, the liberation of 6’-aminoluciferin from “dark” pro-luciferin protease substrates has been exploited to allow the coupled bioluminescent detection of protease activity, both in vitro (Monsees et al., 1994; Moravec et al., 2009) and in vivo (Shah et al., 2005; Dragulescu-Andrasi et al., 2009; Hickson et al., 2010;

Scabini et al., 2011).

Recently, we and others have reported that 6’-alkylaminoluciferins can also be substrates for luciferase (Woodroofe et al., 2008; Reddy et al., 2010; Takakura et al.,

2010). These substrates generally have even higher affinity for luciferase than 6’- aminoluciferin, and emit light at even longer wavelengths. Many modifications are tolerated, including long-chain 6’-alkylaminoluciferins, 5’,6’-cyclic alkylaminoluciferins, and even dialkylaminoluciferins (Figure 1-7). Synthetic modulation of the properties of these molecules thus presents an opportunity to develop new bioluminescent probes and to optimize luciferase light output for different applications. However, with wild-type

Photinus pyralis firefly luciferase, most of these substrates give a rapid burst of light followed by weak sustained emission (Reddy et al., 2010).

The detergent-stable proprietary mutant luciferase Ultra-Glo (Promega) is capable of high sustained light emission with aminoluciferin substrates, particularly in combination with the P450-Glo buffer (Woodroofe et al., 2008; Reddy et al., 2010). The use of aminoluciferins with this luciferase and buffer therefore has potential for novel in vitro screening applications, such as the coupled detection of enzymatic activity (Fan and 31

Wood, 2007). However, Ultra-Glo is a proprietary luciferase reagent that is not available as a genetic construct that can be expressed in cells. Furthermore, the detergent stability of Ultra-Glo and the use of the P450-Glo buffer are important for the light emission behavior. Cellular and in vivo applications such as the detection of gene expression (de

Wet et al., 1987) and bioluminescent imaging (Prescher and Contag, 2010) necessitate a genetically-encodable luciferase that is capable of efficient utilization of aminoluciferins under physiological buffer conditions.

We postulated that aminoluciferin substrates possess all of the key photophysical and structural features necessary to give rise to bioluminescence, but are primarily limited as luciferase substrates due to product inhibition (Reddy et al., 2010). Here we report the construction of mutant luciferases capable of sustained and selective light emission with aminoluciferins, thereby expanding the available tools for bioluminescence imaging.

Materials and Methods

Materials

Chemicals for synthesis were from Aldrich unless otherwise noted. Data was plotted with GraphPad Prism 5.0. NMR spectra were recorded on a Varian Mercury 400

MHz NMR. Small molecule mass spectral data were recorded on a Waters QTOF

Premier. Protein structures were displayed with ICM-Browser (MolSoft).

Bioluminescence measurements were taken in a Turner Veritas luminometer and were not corrected for differences in the wavelength sensitivity of the PMT. Emission spectra 32

and burst kinetic data measurements were recorded on a Spex FluoroMax-3 fluorimeter.

D-Luciferin was from Anaspec and 6’-aminoluciferin was from Marker Gene

Technologies Inc. CycLuc1, CycLuc2, 6’-MeNH-LH2, and 6’-Me2NLH2 were synthesized as previously described (Reddy et al., 2010). Protein concentrations were determined using Coomassie Plus (Thermo Scientific). Immobilized glutathione (Thermo

Scientific) was used for GST-tagged protein purification. Unless otherwise stated, all protein purification steps were carried out at 4˚C.

Plasmid constructs

The Photinus pyralis firefly luciferase gene was PCR-amplified from pGL3

(Promega) and cloned into the BamHI and NotI sites of pGEX6P-1 for bacterial expression. Mutant proteins were created using the Quickchange site-directed mutagenesis kit (Stratagene). Saturating mutagenesis was performed at sites predicted to impact or alter substrate binding, including: F247, T251, L286, S347, R337, R218, Y340, and I351, using the degenerate codon NNK. The point mutants S347A, L286A, A348G,

E311A, I351A, and R337A were made independently.

Mutant screening

Screening for mutant luciferases was performed in the E. coli strain JM109, which has high transformation efficiency and plasmid recovery, as well as reasonable levels of inducible protein expression. Other bacterial strains (OMNI, DH5a, BL21(DE3),

XL1Blue, XL10Gold), gave less consistent results. Following saturating mutagenesis, 33

500 L of the E. coli strain JM109 was transformed and plated on 5 Luria-Bertani (LB) plates that contained 50 mg/mL carbenicillin. Colonies were picked and used to inoculate two 96-well plates where each well contained 150 L of LB with 50 mg/mL carbenicillin. Inoculated plates were incubated at 37 ˚C overnight. Cells were induced with 0.1 mM isopropyl--D-thiogalactopyranoside (IPTG) and incubated at RT overnight. Five L of the bacterial protein extraction reagent SoluLyse (Genlantis) was then added to 50 L of induced cells followed by a 10-minute incubation at RT. Two L of lysate was then added to 60 L of a solution containing 25 M substrate in 20 mM

Tris pH 7.6, 0.1 mM EDTA, 8 mM MgSO4, 4 mM ATP, and 1 mM TCEP in each well of a white 96-well plate (Costar 3912). The bioluminescence emission was measured in a

Turner Veritas luminometer after a delay of 2-5 minutes. Readings were taken every 30 seconds for a total of 10 runs. Mutants that displayed desired emission profiles with the different substrates tested at 25 M were then used in titration assays at substrate concentrations ranging from 0.122-250 M. Mutants of interest were then sequenced, expressed, and purified for further characterization.

Protein expression

Mutant luciferases were expressed as GST-fusion proteins from the vector pGEX6P-1 in the E. coli strain JM109. Cells were grown at 37 ˚C until the OD600 reached 0.5-1, induced with 0.1 mM IPTG, and incubated at 20 ˚C overnight. Cells were pelleted at 5000 rpm in a Sorvall 2C3C Plus centrifuge (H600A rotor) at 4˚C for 10 minutes, then flash-frozen in liquid nitrogen and purified immediately or stored at -80 ˚C. 34

Luciferase purification

E. coli cell pellets from one liter of culture were thawed on ice, resuspended in 25 mL Lysis Buffer (50 mM HEPES pH 7.4, 500 mM NaCl, and 0.5% Tween-20) containing 1 mM PMSF, and disrupted by sonification (Branson Sonifier). Dithiothreitol

(DTT) was added at 10 mM and the resulting cell lysate was clarified by centrifugation at

35K rpm in a Beckman 50.2Ti rotor for 30-45 minutes. The supernatant was batch- bound to immobilized glutathione for 1h at 4˚C, and the beads were washed with Lysis

Buffer containing 10 mM DTT, followed by Wash Buffer (50 mM Tris pH 8.1, 250 mM

NaCl and 10 mM DTT), and finally with Storage Buffer (50 mM Tris pH 7.4, 150 mM

NaCl, 1 mM TCEP). Twenty units of PreScission Protease (GE Healthcare) were added and incubation continued for 4 hours or overnight at 4˚C to cleave the GST-fusion and elute the untagged luciferase protein.

Substrate titration assays with purified protein

Luminescence assays were initiated by adding 30 L 2x substrate in either buffer

A (20 mM Tris pH 7.6, 0.1 mM EDTA, 8 mM MgSO4, 4 mM ATP, and 1 mM TCEP) or buffer B (20 mM Tris pH 7.6, 0.1 mM EDTA, 8 mM MgSO4, 4 mM ATP, 6 mg/mL

BSA, and 33 mM DTT) to 30 L of 20 nM purified luciferase in 20 mM Tris pH 7.6, 0.1 mM EDTA, 1 mM TCEP and 0.4 mg/mL BSA in a white 96-well plate (Costar 3912).

Substrate titration assays were performed 3 minutes post-substrate addition in a Turner

Veritas luminometer with final substrate concentrations ranging from 0.122-125 M. 35

Cell Culture

Chinese hamster ovary-K1 (CHO-K1) cells were grown in a CO2 incubator at

37˚C with 5% CO2 and were cultured in F-12K Nutrient mixture (GIBCO) supplemented with 10% fetal bovine serum and 100 U/mL penicillin/streptomycin.

Transfections

Mutant luciferases F247A, F247L, F247V, F247S, T251S, L286M, S347A,

S347T, and R218K were cloned into the BamHI-NotI site of pcDNA 3.1 and transfected into CHO-K1 cells for live and lysed cell experiments. Transient transfections were performed at room temperature using Lipofectamine 2000 on cells plated at 60-75% confluency in 96-well white tissue culture treated plates (Costar 3917) for intact cell assays, or 6-well plates for lysed cell assays. For intact cells, 0.075 g DNA/well was transfected and for lysed cells, 2.25 g DNA/well was transfected. Assays were performed in triplicate 24 hours post-transfection.

Intact Cell Assays

Transfected CHO cells were washed with HBSS and overlaid with 60 L substrate in HBSS. Titration assays were performed 3 minutes post-substrate addition in a

Turner Veritas luminometer with final substrate concentrations ranging from 0.122-125

M.

Lysed Cell Assays

Transfected CHO cells were washed with HBSS and lysed for 10 minutes at RT with 1 mL 1x Passive Lysis Buffer (Promega) per well. Cells from one and a half wells of a 6-well plate were used to test one substrate in triplicate at the 12 concentrations tested. Luminescence assays were initiated by adding 30 L 2x substrate in 20 mM Tris pH 7.6, 0.1 mM EDTA, 8 mM MgSO4, 4 mM ATP, 6 mg/mL BSA, and 33 mM DTT to

30 L of lysate in a white 96-well plate (Costar 3912). Titration assays were performed 3 minutes post-substrate addition in a Turner Veritas luminometer with final substrate concentrations ranging from 0.122-125 M.

Bioluminescence Emission Scans and Burst Kinetics Assays

Each purified protein in 20 mM Tris pH 7.6, 0.1 mM EDTA, 1 mM TCEP and 0.4 mg/mL BSA was added to 2x substrate in 20 mM Tris pH 7.6, 0.1 mM EDTA, 8 mM

MgSO4, 4 mM ATP, and 1 mM TCEP in a cuvette. Final protein concentrations were 10 nM and substrate concentrations were 10 M.

Bioluminescence Emission Scans

Luciferase was rapidly injected into a cuvette containing substrate, and the emission from 400-800 nm was recorded in a SPEX FluoroMax-3 fluorimeter with closed excitation slits 10 seconds post-injection.

Burst Kinetics Assays 37

Measurements were taken in a Spex FluoroMax-3 fluorimeter every second with a

0.1 second integration time at the maximal emission wavelength for each luciferase/substrate pair. Ten seconds after beginning the assay, protein was rapidly injected into the cuvette containing substrate to observe both the burst and the steady- state light emission over the first minute.

Results

Rational mutation of luciferase

Our initial efforts to improve luciferase light emission with alkylaminoluciferin substrates focused on rational mutation of phenylalanine 247 in the luciferin binding pocket. This residue is involved in a -stacking interaction with D-luciferin (Branchini et al., 2003; Nakatsu et al., 2006). Mutation of this residue to leucine and alanine has been previously reported by Branchini et al., who found that F247L lowers the affinity for D- luciferin but does not impair catalysis, while F247A is severely impaired in both Km and

Vmax (Branchini et al., 2003). We therefore anticipated that F247L would maintain catalytic function but allow improved product dissociation, helping to relieve product inhibition and allowing improvement in the sustained light emission from alkylaminoluciferin substrates.

Surprisingly, we found that the F247L mutation improves maximal light emission from 6’-aminoluciferin by 4.9-fold but has only a small positive effect on light emission from CycLuc1 (Figure 2-1; Figure 2-2). Instead, we found that the F247A mutation, which has a marked negative effect on light output from both D-luciferin and 6’- 38

aminoluciferin, gave dramatically improved sustained light output from CycLuc1 (Figure

2-1; Figure 2-2). The improved light output of this mutant comes at the cost of a substantially increased Km value. While this is not a concern under conditions where saturating concentrations of substrate can be applied (e.g., in vitro), a high Km is expected to limit light output in live cells (Figure 2-2), where only low intracellular concentrations of luciferin substrates are achieved (de Wet et al., 1987; Craig et al., 1991; Shinde et al.,

2006). We therefore sought alternate sites for mutation that could potentially increase light output yet retain a low Km.

Figure 2-1. Preliminary characterization of phenylalanine 247 mutants. Dose- response curves for purified luciferases (WT, F247L, F247A, F247S, and F247V) were generated with D-Luciferin, 6’-NH2LH2 and CycLuc1 at concentrations from 0.122-125 M. The assays were performed in triplicate and are represented as the mean ± SEM.

Figure 2-2. Dose-response profiles from purified luciferases and luciferases transfected into CHO-K1 cell lysate and in live cells. Titration assays with the native substrate D-Luciferin, 6’-NH2LH2 and CycLuc1 were performed at concentrations from 0.122-125 M. (A) Titration assays performed using 10 nM purified WT, F247L, F247A, F247S, or F247V firefly luciferase. (B and C) CHO-K1 cells transiently transfected with pcDNA3.1 vectors expressing WT, F247L, F247A, F247S, or F247V firefly luciferase using lysates from the transfected cells (B) or using the intact live cells (C). The assays were performed in triplicate and are represented as the mean ± SEM. 41

Screening for Mutant Luciferases with Improved Utilization of Aminoluciferins

Because mutation of phenylalanine 247 was found to affect substrate utilization, but not in an easily predictable manner, we opted to perform saturating mutagenesis at this and other active site residues, and then screen for mutant luciferases with the desired properties. This approach has the advantage of rational site selection, but is not restricted by preconceived notions of suitable replacement residues. Based on the crystal structure of Luciola cruciata luciferase (Nakatsu et al., 2006), we identified a number of amino acids that appear to interact with D-luciferin, ATP, or that play a role in creating the overall local structure of the binding pocket (Figure 2-3A; Table 2-1). In addition to phenylalanine 247, we selected arginine 218, threonine 251, leucine 286, tyrosine 340, and serine 347 for saturating mutagenesis. Arginine 218 is proximal to the 4’- and 5’- carbons of D-luciferin and forms part of the van der Waals surface that encompasses the luciferin substrate (Nakatsu et al., 2006). Mutations at this site were expected to provide increased “wiggle room” for substrates, particularly those with 5’-modifications (i.e.,

CycLuc1 and CycLuc2). Leucine 286 is a candidate residue for interaction with alkyl side chains on aminoluciferin substrates, while tyrosine 340 forms part of the ATP binding site and is located at the interface between the ATP and luciferin binding pockets. The methyl group of the threonine 251 side chain makes a van der Waals interaction with the benzothiazole of the luciferin substrate, while serine 347 forms a hydrogen bond to the benzothiazole nitrogen through the intermediacy of a water molecule (Nakatsu et al.,

2006). We reasoned that mutation at these sites could potentially lead to improved continuous light emission from aminoluciferins by altering substrate alignment and/or 42

improving product dissociation. Because aminoluciferins have higher affinity for luciferase, the loss of interactions with the benzothiazole could also confer selectivity over D-luciferin, particularly if the affinity and proper orientation of alkylaminoluciferins can be maintained by ancillary interactions that are not available to D-luciferin.

Figure 2-3. Creation of mutant luciferases. (A) Mutation sites were selected based on proximity to the luciferin substrate in the crystal structure of Luciola cruciata luciferase (PDB 2D1R). (B) Selected residues were subjected to saturating mutagenesis. Mutant luciferase-expressing bacteria were screened for light emission with CycLuc1 and those that exhibited improved properties were sequenced (Table 2-1) and the mutant protein was purified for further characterization. 43

Table 2-1. Mutants identified in saturating mutagenesis screen. Mutation sites were selected based on proximity to the luciferin substrate in the crystal structure of Luciola cruciata luciferase (PDB 2D1R). Selected residues were subjected to saturating mutagenesis (Figure 2-3). Mutant luciferase-expressing bacteria were screened for light emission with CycLuc1 and those that exhibited improved properties were sequenced.

At each selected site, we used the degenerate codon NNK to perform saturating site-directed mutagenesis in the wild-type Photinus pyralis luciferase (Figure 2-3B). At some sites we also introduced alanine point mutations (see Methods). After cloning of the mutant luciferases into bacteria, colonies were picked and grown in 96-well plates

(Figure 2-3B). Luciferase expression was induced with isopropyl β-D-1- thiogalactopyranoside (IPTG) and the bacteria were lysed. Light emission from each lysate was measured at a luciferin concentration of 25 M. For those lysates that gave improved and/or selective light output with CycLuc1, the plasmid encoding the mutant luciferase was sequenced (Table 2-1). Mutation at some sites failed to produce any obvious improvement in luciferase performance. For example, mutation of tyrosine 340 in the nucleotide binding pocket primarily yielded inactive luciferases. In hindsight,

Y340 makes a hydrogen-bonding interaction with D420 that may be critical for luciferase function (Nakatsu et al., 2006). A secondary assay was also performed on the bacterial 44

extracts to measure light output as a function of CycLuc1 concentration (Figure 2-3B).

Mutants that demonstrated improved light output, low Km, and/or selectivity for CycLuc1 in this assay were expressed as recombinant proteins and purified for further characterization.

Characterization of Mutant Luciferases In Vitro

Screening for mutants of phenylalanine 247 identified F247S and F247V as proteins with greatly improved light output from CycLuc1, but still with relatively high

Km values (Figure 2-1). Mutations at several of the other chosen sites (R218K, T251S,

L286M, S347T, and S347A) improved light emission with aminoluciferins, while retaining Km values below that of the native substrate D-luciferin. For these purified mutant luciferases, light output as a function of substrate concentration was measured with D-luciferin and the aminoluciferin substrates shown in Figure 1-7, with nonlinear regression fitting to the Michaelis-Menten equation (Figure 2-4; Table 2-2). For most of the luciferase mutants, the rank of light output was D-Luciferin > CycLuc1 > 6’-NH2LH2

> 6’-MeNHLH2 > CycLuc2 > 6’-Me2NLH2. In all cases, the aminoluciferins exhibited much lower Km values than D-luciferin. The R218K mutant was the most generally beneficial mutation, yielding the greatest light output for most aminoluciferin substrates, including a 20-fold improvement in the maximal sustained emission from CycLuc2 while retaining a very low Km of 0.3 M (Figure 2-4; Table 2-2). Similarly, the maximal sustained light emission from CycLuc1 was increased 14-fold to 12.5% of that of D- luciferin with WT luciferase, with a Km of 1.8 M that is still substantially lower than 45

that of D-luciferin. The R218K mutation has been previously described in the context of

D-luciferin, and is known to raise the Km but cause minimal disruption of catalytic activity (Branchini et al., 2001).

Table 2-2. Apparent Km and Vmax values. Dose-response curves for purified luciferases (WT, T251S, L286M, S347A, and R218K) were generated with D-Luciferin, 6’-NH2LH2, 6’-MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2 at concentrations from 0.122-125 M. The assays were performed in triplicate and represented as the mean ± SEM (Figure 2-4). Each curve was fit to the Michaelis-Menten equation by nonlinear regression (GraphPad 5.0) to determine apparent Km and Vmax values.

Figure 2-4. Characterization of luciferase mutants. Dose-response curves for purified luciferases (WT, T251S, L286M, S347A, and R218K) were generated with D-Luciferin, 6’-NH2LH2, 6’-MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2 at concentrations from 0.122-125 M. The assays were performed in triplicate and are represented as the mean ± SEM. Each curve was fit to the Michaelis-Menten equation by nonlinear regression (GraphPad 5.0) to determine apparent Km and Vmax values (Table 2-2).

The most selective luciferase mutants were S347A and S347T. Because the activity of the purified S347T luciferase was found to decline rapidly when stored at 4˚C,

S347A was used in preference to S347T for further characterization (Figure 2-5). The

S347A luciferase yielded a 14-fold increase in light emission for CycLuc1, equivalent to that of the R218K mutant, while simultaneously decreasing the maximal emission from

D-luciferin by 7.5-fold and raising the Km for D-luciferin by >10-fold (Figure 2- 4; Table

2-2). To determine whether this selectivity could be further increased, we created several combinations of the S347A mutation with T251S, L286M, or F247 mutants (Figure 2-5).

All of these mutants showed increased discrimination against D-luciferin, giving even lower light output than with S347A alone. On the other hand, the S347A/L286M double mutant further increased the maximal sustained light emission from CycLuc1 by two-fold relative to S347A alone, albeit with a ~10-fold increase in the Km to 22 M.

Figure 2-5. Several S347 mutants show selectivity for CycLuc1 over D-Luciferin. (A) Purified WT, S347A, S347T mutant luciferases (10 nM) were used in titration assays of the native substrate D-Luciferin and CycLuc1 at concentrations from 0.122-125 M. The assays were performed in triplicate and are represented as the mean ± SEM. (B) Purified S347A and double and triple mutants (10 nM) were used in titration assays of the native substrate D-Luciferin and CycLuc1 at concentrations from 0.122-125 M.

Rapid injection experiments were performed to reveal the burst kinetic behavior of the mutant luciferases (Figure 2-6; Figure 2-7). As we have observed for the wild-type protein, most alkylaminoluciferin substrates gave a rapid initial burst of light, followed by a substantial decrease in the rate of light output (Reddy et al., 2010). Although most of 49

the mutant luciferases identified here shared this general behavior, the decrease in the rate of light output was less severe than that of the wild-type protein, resulting in a higher level of sustained light emission (Figure 2-6; Figure 2-7A). The most striking finding was that the initial rate of light output for the dialkylaminoluciferin CycLuc2 with the mutant

R218K is considerably increased relative to the wild-type protein (Figure 2-6; Figure 2-

7B). In contrast, the burst kinetic profile for the corresponding acyclic dialkylaminoluciferin 6’-Me2NLH2 was largely unchanged (Figure 2-6; Figure 2-7C).

Figure 2-6. Burst kinetics and emission profiles. Burst emission profiles for purified 6’-NH2LH2, 6’-MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2 as described in the Methods. For ease of comparison between mutants, the burst emission scale is the same for all mutants used with a particular substrate. For comparison between substrates, note the relative emission on the Y-axis differs to account for the differences in relative light output. Emission spectra were also determined for each protein-substrate pair and are not normalized.

Figure 2-7. Burst kinetic profiles with WT and R218K luciferases. Purified luciferases (10 nM) were rapidly injected at the 10-second time point into 10 M of CycLuc1 (A), CycLuc2 (B), or 6’-Me2NLH2 (C) and the light emission was recorded.

The bioluminescence emission wavelength of aminoluciferins is red-shifted relative to D-luciferin with all of the tested luciferases (Table 2-3). Because the PMT in the Turner Veritas plate reader is less sensitive to the red-shifted light emission of the aminoluciferin substrates than the green light emission of D-luciferin, this likely leads to an underestimation of the true light output for aminoluciferins. We did not attempt to 52

correct for this difference, which also varies among mutants. For example, the R218K luciferase caused a slight red-shift in the bioluminescence of all luciferins: D-luciferin yields maximal emission at 567 nm, CycLuc1 at 609 nm, and CycLuc2 at 621 nm (Table

2-3). In contrast, the L286M mutant results in a 5-13 nm blue-shift in the emission of all aminoluciferins, but a diametrically-opposed 13 nm red-shift in D-luciferin light emission

(Table 2-3). Interestingly, the S347A mutant gives discrete emission peaks for all of the aminoluciferins, but anomalous bimodal emission for D-luciferin (Branchini et al., 2003).

Table 2-3. Emission maxima. Emission spectra for purified luciferases (WT, T251S, L286M, S347A, and R218K) were generated with D-Luciferin, 6’- NH2LH2, 6’- MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2 as described in the Methods. Emission maxima are presented here in nm.

Characterization of Mutants in Mammalian Cells

The mutant luciferases that performed best in vitro were cloned into pcDNA3.1 and transfected into CHO-K1 cells. Light emission from mutant luciferases in CHO cell lysates generally mirrored the results with purified proteins (Figures 2-2, 2-3, 2-4, and 2-

8). The lone exception was L286M, which gave lower light output than expected, possibly indicating that this protein has a lower expression level and/or stability in the 53

cellular context (Figures 2-2, 2-3, 2-4 and 2-8). Unsurprisingly, D-luciferin was the best substrate for wild-type luciferase (Figure 2-8; 2-9A). However, for the R218K and

T251S mutant luciferases, CycLuc1 light output was superior to D-luciferin at substrate concentrations below ~30 M. In the case of the S347A mutant, CycLuc1 exceeded the light output of D-luciferin over the entire concentration range (Figure 2-8, 2-9A). All other substrates had considerably lower levels of light emission in cell lysates, with the exception of 6’-aminoluciferin with the mutant F247L (Figure 2-2B).

Figure 2-8. Dose-response profiles from luciferin substrates in CHO-K1 cell lysate. CHO-K1 cells were transiently transfected with pcDNA3.1 vectors expressing WT, T251S, L286M, S347A or R218K firefly luciferase. Dose-response curves for each luciferase with D-Luciferin, 6’-NH2LH2, 6’-MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2 were generated at concentrations from 0.122-125 M using lysates from the transfected cells. The assays were performed in triplicate and are represented as the mean ± SEM. 55

Figure 2-9. Dose-response profiles from luciferases in expressed in CHO-K1 cell lysates and in live cells. CHO-K1 cells were transiently transfected with pcDNA3.1 vectors expressing WT, T251S, L286M, S347A, or R218K firefly luciferase. Dose- response curves for each luciferase with D-Luciferin, 6’-NH2LH2, 6’-MeNHLH2, 6’- Me2NLH2, CycLuc1 and CycLuc2 were generated at concentrations from 0.122-125 M using lysates from the transfected cells (A) or using the intact live cells (B). The assays were performed in triplicate and are represented as the mean ± SEM.

The light output from luciferase in live cells is much lower than that of lysed cells, as has been previously observed for D-luciferin (de Wet et al., 1987; Craig et al.,

1991; Shinde et al., 2006) and 6’-aminoluciferin (Shinde et al., 2006). Moreover, there is less difference in the relative levels of light emission between different substrates. Even with the wild-type luciferase, CycLuc1 emitted 26% of the light of D-luciferin at 125 M and its light output greatly exceeded that of 6’-aminoluciferin (Figure 2-9B). CycLuc1 was the superior substrate for T251S, R218K, and particularly S347A luciferase (18-fold higher light output than D-luciferin at 125 M). Remarkably, live-cell light emission from CycLuc2 exceeded that of both CycLuc1 and D-luciferin with T251S, S347A, and particularly R218K luciferase over a broad concentration range, despite the relatively poor light output from CycLuc2 in cell lysates (Figures 2-9 to 2-11). 57

Figure 2-10. Relative activity of mutant luciferases and novel substrates in vitro, in CHO cell lysate, and in intact CHO cells at 1 µM substrate concentration. The percentage of light output is expressed relative to light output with WT luciferase and the corresponding substrate under the same conditions.

Figure 2-11. Relative activity of mutant luciferases and novel substrates in vitro, in CHO cell lysate, and in intact CHO cells at 125 µM substrate concentration. The percentage of light output is expressed relative to light output with WT luciferase and the corresponding substrate under the same conditions.

Discussion

We have found that mutation of firefly luciferase dramatically improves aminoluciferin substrate utilization and selectivity, and allow the use of these substrates to monitor luciferase expression in cells and cell lysates. For example, the F247L mutant improves light output from 6’-aminoluciferin by almost five-fold and can be 59

recommended for improved imaging of this substrate (Figure 2-1; Figure 2-2). This has particular significance for bioluminescence assays of protease activity that rely on detection of this substrate (Monsees et al., 1994; Shah et al., 2005; Fan et al., 2007;

Dragulescu-Andrasi et al., 2009; Hickson et al., 2010; Scabini et al., 2011). Light output from all aminoluciferins was greatly increased by the R218K mutant (e.g., 14-fold for

CycLuc1, 20-fold for CycLuc2), suggesting this mutant as a starting point for measuring light emission from these and other novel alkylated aminoluciferins. Moreover, mutation of serine 347 to threonine or alanine gives similarly improved light emission from

CycLuc1 but also exhibits strong selectivity for CycLuc1 over D-luciferin both in vitro and in live cells (Figures 2-4; Figures 2-9 to 2-11). Combination of mutations such as the ones described here is expected to allow further enhancement of aminoluciferin light output and selectivity. For instance, we have found that the S347A/L286M double mutant gives twice the light output of S347A with CycLuc1, and further discriminates against D- luciferin (Figure 2-5).

The rate at which light is generated by luciferase is dependent on several factors, including the rate of adenylation and oxidation to afford the oxyluciferin excited state, the quantum yield of light emission, and the rate of product release. To gain insight into the underlying mechanisms by which light emission is improved, we obtained burst kinetic profiles of the luciferase mutants (Figure 2-6; Figure 2-7). For CycLuc1 and 6’-

MeNHLH2, all mutants result in a rapid initial burst that is similar to that observed for the wild-type protein, but with higher sustained light output (Figure 2-6; Figure 2-7A).

Notably, both the burst and sustained light emission from CycLuc1 exceeds that of 6’- 60

MeNHLH2 with all of the luciferases we have characterized, and is consistent with a role for cyclization in optimizing aminoluciferin light output (Reddy et al., 2010). The rapid rise to a high initial rate of light output for these substrates suggests that there are no substantive defects in the formation of the respective luciferyl-AMP or its subsequent oxidation to afford the excited-state oxyluciferin. We therefore presume that a reduction in affinity for aminoluciferin substrates and the corresponding lowered affinity for their products is primarily responsible for the observed improvement in sustained light emission for most of the mutants we have identified. Even with D-luciferin, the molecular basis for product inhibition is still unresolved, potentially including contributions from both dehydroluciferyl-AMP (L-AMP) and oxyluciferin (Fraga, 2008).

Thus, it is also possible that some mutants function in part by reducing the formation of the corresponding L-AMP analog. A better molecular understanding of the nature of the product inhibition with these substrates may help guide future optimization efforts.

Like CycLuc1, light emission from CycLuc2 is superior to its acyclic counterpart

6’-Me2NLH2 for every luciferase we have tested. Intriguingly, we find that the burst emission for CycLuc2 with the mutant luciferase R218K is substantially more rapid and intense than that of CycLuc2 with wild-type or any other mutant (Figure 2-6; Figure 2-

7B). In contrast, the burst for 6’-Me2NLH2 is rapid but weak for all mutants (Figure 2-6;

Figure 2-7C). We postulate that in the wild-type luciferase, the rigidity of CycLuc2 – which is desirable for efficient light emission – also enforces a sub-optimal alignment in the active site for the production of the excited-state oxyluciferin (e.g., delayed adenylation or oxidation; see Figure 1-1). Enlarging the luciferin binding pocket with the 61

R218K mutation improves the alignment of CycLuc2 within the active site, allowing more rapid production of an efficient light-emitting oxyluciferin excited state (Figure 2-

7B). Conversely, the free bond rotation about the aryl amine bond of 6’-Me2NLH2 may facilitate the alignment necessary for the rapid formation of the oxyluciferin, but also limit the efficiency of light output (Figure 2-7C).

For light emission from live cells, the cell-permeability and Km of the substrate are important for access to the intracellular luciferase and the efficiency of light output under sub-saturating concentrations. CycLuc2 has a lower Km than CycLuc1 (Table 2-2) and is predicted to be more cell-permeable than CycLuc1 because of the replacement of a polar amine proton with a methyl group (cLogP of 2.5 versus 2.0). These differences are therefore likely to explain the better relative performance of CycLuc2 in live versus lysed cells. Moreover, this suggests that optimization of cell permeability and light output at low substrate concentrations – rather than just maximal light output – may be particularly important for imaging in live cells and organisms, because insufficient substrate is delivered into the live cell to achieve maximal light emission.

The bioluminescence emission wavelength of D-luciferin from different beetle luciferases (firefly, click beetle, railroad worm) and their mutants varies from ~540-635 nm (Viviani et al., 1999; Branchini et al., 2010). Beyond simple changes in the polarity of the environment of the light emitter (Morton et al., 1969), explanations for this sensitivity include changes in the rigidity of the active site (Nakatsu et al., 2006) and perturbations in the ionic interactions of the phenolate of D-luciferin (Hirano et al., 2009). The emission behavior of aminoluciferins can lend insight into this question because they lack 62

an ionizable phenolate (Figure 1-7). With the mutant R218K, all luciferins exhibit a modest red-shift in emission wavelength (Table 2-3). Because oxyluciferins are asymmetric charge-transfer molecules, some shared solvatochromatic sensitivity to the polarity of the luciferin binding site is expected (Naumov et al., 2009). However, with other mutants we find that D-luciferin exhibits anomalous emission behavior: the L286M mutation causes a red-shift in the emission of D-luciferin but a blue-shift in the emission of all aminoluciferins, and D-luciferin uniquely gives bimodal emission with S347A

(Table 2-3). These differences are consistent with a role for the ionization state and ionic interactions of the phenol in determining the bioluminescence emission wavelength when

D-luciferin is the substrate.

The efficient chemical generation of light by firefly luciferase has been widely used as a sensitive reporter system for gene expression (de Wet et al., 1987; Prescher et al., 2010). However, the application of bioluminescence detection as a general optical reporter of cellular status has lagged behind that of fluorescence. Chemical modification of the luciferin substrate can red-shift the emission wavelength of bioluminescence beyond that of D-luciferin (Reddy et al., 2010). Moreover, the combination of synthetic luciferins and mutant luciferases described here not only extends the wavelength of bioluminescence emission and broadens the scope of substrates that can be used for bioluminescent assays both in vitro and in mammalian cells, but also suggests that orthogonal luciferases could be developed to allow multiplexing of bioluminescent signals. Further chemical modification of luciferin substrates and corresponding mutation 63

of firefly and other beetle luciferases is therefore anticipated to be a fruitful avenue for expanding the power and scope of bioluminescence assays and imaging.

CHAPTER III: Improving Bioluminescence Imaging Discussion

Bioluminescent imaging studies in tissue and whole organisms necessitate the use of luciferases that can emit longer wavelength, deeper penetrating red light. Red-emitting luciferases for imaging studies have been generated through mutagenesis and by taking advantage of naturally red-emitting luciferase variants (Almond et al., 2003; Branchini et al., 2010a; Branchini et al., 2007a; Branchini et al., 2005a; Kitayama et al., 2004;

Nakajima et al., 2004; Viviani et al., 1999). Using saturating mutagenesis of specific residues in firefly luciferase, we were able to generate mutant luciferases that emitted more intense, sustained red-shifted light with novel luciferin analogs made in the lab than

WT protein. These mutant:luciferin pairs present an alternative method for generating red light, expanding the toolset for bioluminescence imaging.

Future directions could include the optimization of these mutant:luciferin pairs to emit even more red shifted, higher intensity sustained light. We found that in live cells, it appears that the highest light output was observed with mutants that had high substrate affinity and with substrates that were the most hydrophobic. Future screens to identify mutants that could best utilize different luciferin substrates in vivo would focus not just on mutants that exhibited the highest Vmax values, but also those that maintained low Km values with the substrates with the most favorable cLogP values. Also, several additional novel luciferins have been generated in the lab (Figure 3-1) that are more hydrophobic and emit even more red-shifted light (Figure 3-2). Screens with these compounds could also generate luciferases that would perform well in vivo. 65

Fig 3-1. Additional novel alkylamino luciferin substrates. Following screening with D-Luciferin, 6’-NH2LH2, 6’-MeNHLH2, 6’-Me2NLH2, CycLuc1 and CycLuc2, these additional novel alkylamino luciferin molecules were screened with the mutant R218K.

Figure 3-2. Emission maximum with R218K and novel substrates. Several novel alkylamino luciferin substrates further red shift light emission from the R218K mutant of luciferase.

Furthermore, future efforts would include the quantification of both the expression levels of the different luciferase mutants in CHO cells and the achievable intracellular concentrations of each substrate in intact cells. One way to measure substrate concentrations would be through direct fluorescent measurement of live cells treated with substrate. The results of these experiments could allow for more definitive conclusions to be made about the relative roles of Vmax, Km and cellular permeability in the overall amount of light produced by each of the mutants. Quantification of the expression levels of the different mutants could also explain the slight differences in light output observed in the in vitro studies and in the experiments with CHO cell lysate.

Further, the imaging studies of live cells treated with substrate would allow for the 67

verification that specific substrates do not accumulate in certain subcellular locations, but exhibit homogenous localization profiles.

In addition, while we found that combining the mutations that were best able to utilize our synthetic substrates did not generally improve their functioning (Figure 3-3)

(many exhibited higher Vmax values and/or substrate selectively, but all had higher Km values which are unfavorable for in vivo applications), saturating mutagenesis at those sites in the context of the mutant background could potentially prove more fruitful. An alternate approach to generate luciferases with improved ability to utilize our novel substrates would be to perform DNA shuffling or exchange of restriction fragments among the various luciferase variants and perform screens with the resulting chimeras

(Joern et al., 2002; Wood, 1990; Zhang et al., 2002).

Figure 3-3. Combining favorable mutations can increase Vmax with our novel synthetic substrates, but at a cost to Km values. Purified double and triple mutants (10 nM) were used in titration assays of the native substrate D-Luciferin and CycLuc1 at concentrations from 0.122-125 M. 69

Interestingly, S347A showed a great deal of selectivity with respect to the substrate CycLuc1 as opposed to D-luciferin (Figures 2-9 to 2-11). This observation supports the possibility of using S347A:CycLuc1 in combination with WT:D-luciferin for dual color studies or multiplexing applications to allow for the monitoring of multiple events in the same lysate, cell, tissue or organism. Screening for luciferase mutants with even lower light emission with D-luciferin would further improve the imaging possibilities.

Overall, in vitro and in cell lysate, the substrate CycLuc1 and the mutants S347A and R218K exhibited the highest relative light output of the mutants and novel substrates tested (Figure 2-4; Figure 2-8). However, the ultimate goal of this work would be to demonstrate that our novel luciferin substrates and luciferase mutants could be used to perform bioluminescent imaging in a mouse and apply this technique to in vivo applications. While they exhibit lower light output with CycLuc2 in vitro and in cell lysate, the T251S, S347A and R218K mutants emit high relative levels of red-shifted light with CycLuc2 over a wide range of substrate concentrations in live cells (Figures 2-

9 to 2-11). The red-shifted light emitted by these mutants would greatly improve the degree of resolution achievable, making them potential candidates for use with CycLuc2 in mouse models for a variety of applications, including the noninvasive tracking of viral loads or tumor metastasis in the live mouse, or even in the different patterns of gene expression in a developing mouse. The combination of our mutants with mutations that improve thermostability and codon optimization would also aid in this process. 70

Overall, methods to better observe cellular activity will allow for the improved understanding of the events that lead to the development of disease states. In parallel, methods that can allow for the precise control of cellular events, through the control of protein activation states, would also provide great insight into disease development.

CHAPTER IV: Photocaging background

The ability of a cell to establish, maintain, and modulate a polarized state is essential for proper cellular behavior, such as mitosis, migration, differentiation, proliferation, and morphogenesis. The signaling pathways that allow the cell to achieve this control require activation of cellular components at specific times and in specific locations. Misregulation of these important signaling pathways can lead to many disease states. Reagents that modulate components of these pathways can provide insight into the fundamental spatial and temporal activation requirements in a living cell.

Methods to control protein activity

Numerous methods have been developed to control the activity of proteins. One approach is the expression or overexpression of genes coding for mutant proteins that exhibit specific properties. Some of these include proteins that are always activated and performing their given function, termed constitutively active (CA), those that are able to exert inhibitory action on their wild-type counterparts, termed dominant negative (DN), or those that are generally inactive or exhibit some other mutant function. One method to generate the mutants is through site-directed mutagenesis. Site-directed mutagenesis has been key to the analysis of protein function both in vitro and in vivo, however the functionality of the potential groups incorporated is restricted to the twenty genetically encoded amino acids. Small interfering RNAs, or siRNAs, are also a powerful tool to reduce the intracellular concentration of a protein of interest. The protein concentration 72

is reduced via the RNA interference (RNAi) pathway, whereby a messenger RNA is prevented from producing the protein it encodes. While the use of RNAi does allow for modulation of the degree of reduction of protein levels, both its use and the expression of mutant proteins present similar limitations. First, they are limited in their temporal and spatial resolution. Mutants will be present throughout the cell and could exert their effects throughout the cell cycle and at many subcellular locations. Likewise, siRNA effects are widely realized in the cell to which they are applied. Another drawback of both of these methods is that they require time to take effect, so there is the potential for the cell to compensate for the presence of the mutant protein or the reduction in or lack of protein resulting from siRNA treatment.

Another approach is the use of small molecules that behave as inhibitors, activators, or modulators of protein activity. The effects of these types of molecules are realized much more quickly, eliminating the concerns of cellular compensation.

However much like the use of mutants and siRNA, inhibitors, activators, or modulators of protein activity are also limited in their spatial and temporal resolution. Furthermore, these types of molecules often do not possess the required degree of target specificity.

Several methods have been developed to improve the utilization of small molecules in the control of protein function, including ‘bump-and-hole’ methods and the use of chemical inducers of dimerization. In ‘bump-and-hole’ methods, a protein is mutated at a location near the active site, for example, to introduce a cavity, or hole. An inhibitor of the protein is derivatized with a bulky group, or bump, designed to fit into the space introduced by the mutation. Thus, the mutant protein will interact specifically with 73

the ‘bumped’ inhibitor that the wild type protein will not be able to recognize, allowing for allele-specific inhibition. The ‘bump-and-hole’ approach has been used to study many proteins including kinesins, nuclear hormone receptors, and the transcription factor

EF-Tu (Belshaw, 1995; Hwang and Miller, 1987; Kapoor and Mitchison, 1999; Powers and Walter, 1995; Shogren-Knaak et al., 2001). Chemical inducers of dimerzation (CID) also allow for the control of protein activity by rapidly bringing together two proteins.

One commonly used CID is rapamycin. When two proteins are expressed as a fusion with the FK506-binding protein (FKBP12) and a fusion with the FKBP12-rapamycin- associated protein (FRAP), addition of rapamycin will result in dimerization of FKBP and FRAP, bringing the two fused species in close proximity. For example, this method was used to bring together two halves of a split intein domain to allow for the conditional protein splicing upon rapamycin addition (Mootz et al., 2003; Mootz and Muir, 2002).

Another CID is 4-hydroxy-tamoxifen, which can be used to bring together two proteins fused to the ligand-binding domain of the estrogen receptor (Buskirk et al., 2004). As with inhibitors, activators or modulators of protein activity, ‘bump-and-hole’ methods and CIDs benefit from quick action upon addition of the ‘bumped’ inhibitor or dimerizing molecule. Additionally, by design these approaches also benefit from a high degree of specificity. Unfortunately, they are also unable to achieve spatial resolution in the context of a living cell.

Chromophore-assisted light inactivation (CALI) is a method to rapidly inactivate a protein of interest (Jay and Sakurai, 1999). In this approach, a chromophore produces reactive oxygen species (ROS) when irradiated with light and these ROS damage proteins 74

located very close to the chromophore. Typically the chromophore is fused to antibodies towards the protein of interest or directly to the protein itself. Factors that impact the radius of damage include the specific chromophore selected and the intracellular environment in which CALI is being performed. CALI can be a powerful approach to allow for spatial and temporal protein inactivation when experiments are designed stringently and proper controls are in place. However there is still some doubt regarding both the specificity and general applicability of this approach due to the possibility of unintentional concomitant damage (Jacobson et al., 2008).

Photocaging introduction

Photocaging is an approach that addresses many of the limitations described above. Specifically, a photocaged species possesses a modification at a functionally relevant position that is sensitive to illumination by light. Therefore when a photocaged molecule is introduced into a cell it will not be able to perform its usual functions.

However, irradiation will remove the modification, freeing the species at that specific time and place in the cell and allowing questions regarding its intracellular roles at specific locations and cell cycle stages to be addressed. This approach is particularly powerful in that cells are transparent, they do not themselves respond to light unless they are highly specialized cells like photoreceptors cells, and the methods of controlled light illumination are well established.

While many light sensitive protecting groups used for synthetic chemistry applications had been previously described (Barltrop et al., 1966; Herbert and Gravel, 75

1974; Kirby and Vagelos, 1967; Patchornik et al., 1970), Kaplan and coworkers were the first to use a caged molecule for biological light based activation experiments (Kaplan et al., 1978). Specifically, they used 2-nitrobenzyl phosphate and 1-(2-nitro)phenylethyl phosphate to cage adenosine triphosphate (ATP) and render it nonhydrolyzable by the protein they were interested in studying, the Na:K ATPase. Upon irradiation with light, the uncaged ATP was once more a substrate for the Na:K ATPase and studies of ATP utilization by the Na:K ATPase could be performed. Since then, caged ATP has been used widely (Goeldner and Givens, 2005; Marriott and Ottl, 1998; Marriott et al., 1998) and many other small molecules including lipids, sugars, steroids and second messengers have also been caged and used to investigate a wide variety of biological questions

(Dorman and Prestwich, 2000; Juodaityte and Sewald, 2004; Kale et al., 2001; Srinivas et al., 2002, 2005).

Caged peptides can also be synthesized using automated solid phase methods where residues whose side chains can be derivatized are lysine, arginine, tyrosine, serine, cysteine, glutamic acid and aspartic acid. The first reported use of caged peptides in a biological system involved the use of peptide inhibitors of myosin light chain kinase

(MLCK) and Ca/calmodulin that were caged with 2-bromo-2’nitrophenylacetic acid methyl ester on the phenolic group of a tyrosine side chain (Walker et al., 1998). These caged peptides were used to investigate the degree to which eosinophil cells rely on calmodulin and MLCK to control their amoeboid locomotion.

Direct chemical modification

Overall, there are many different types of photocages that can be used to cage small molecules and peptides. However, generating photocaged proteins has proved more challenging. Photocaged proteins have been generated in several ways. The first is by direct chemical modification of a native amino acid side chain of interest or that of an amino acid, most often a cysteine residue, introduced by site-directed mutagenesis at a certain location in the protein.

Some of the first processes that were photocaged using this method include G- actin polymerization (Marriott, 1994), the ATPase activity of myosin (Marriott and

Heidecker, 1996), the enzymatic activity of -galactosidase (Golan et al., 1996), the kinase activity of cAMP-dependent protein kinase (Chang et al., 1998; Curley and

Lawrence, 1998b), the toxic pore forming ability of -hemolysin (Chang et al., 1995), and the actin binding and severing activity of cofilin (Ghosh et al., 2002; Ghosh et al.,

2004). In all of these examples, cysteine or lysine residues were caged using various ortho-nitrobenzyl derivatives. These small molecules, when irradiated with UV light, undergo the transfer of an oxygen from the nitro group to the benzylic position. Because the benzylic position is substituted with either an oxygen, sulfur or nitrogen group, the transfer of the oxygen generates an unstable intermediate that decomposes to the corresponding ortho-nitroso derivative and restores the modified amino acid to its original state (Figure 4-1). When generating photocaged protein in this manner, factors such as residue accessibility, the potential for modification on multiple residues, the 77

generation of a mixed protein population, and removal of unmodified protein become considerations.

Figure 4-1. Ortho-nitrobenzyl uncaging. Mechanism of Ortho-nitrobenzyl uncaging upon irradiation of caged protein with h. L=leaving group which here is the amino acid side chain which was modified, often the sulfur group of a cysteine.

Another slightly different method to generate photocaged proteins is through protein-catalyzed activation of a specific residue followed by alkylation with the photolabile moiety. Zou and coworkers activated a threonine residue in the catalytic subunit of protein kinase A by thiophosphorylation with the protein 3-phosphoinositide- dependent kinase 1 (PDK-1), which they subsequently derivatized with 4- hydroxyphenacyl bromide (Zou et al., 2002). This approach relied on the selective phosphorylation of the threonine of interest by PDK-1. 78

Banghart and coworkers used an azobenzene derivative, substituted with a maleimide moiety to make it cysteine reactive, to photocage a cysteine introduced into the potassium channel Shaker (Banghart et al., 2004). Irradiation with long-wavelength light shifts the azobenzene moiety to an extended trans configuration and short- wavelength light returns it to a cis configuration. In the extended trans configuration, a quaternary ammonium addition on the azobenzene is close enough to block the pore of the channel. Banghart et al (2004). expressed the mutant Shaker channel in Xenopus oocytes and in hippocampal pyramidal neurons and treated with high external levels of the azobenzene derivative to block Shaker potassium channels. This approach is unique in that it allows for reversible light dependent block of the channel.

Expressed protein ligation

Photocaged proteins can also be generated using expressed protein ligation (EPL).

Hahn and Muir used EPL to generate Smad2 photocaged on two activating phosphoserine residues (Hahn and Muir, 2004). This method takes advantage of native chemical ligation in which a recombinant protein with a C-terminal thioester and a synthetic peptide with an N-terminal cysteine are fused through a native peptide bond.

Specifically, C-terminally-truncated recombinant Smad2 protein was expressed upstream of an intein-chitin binding domain sequence. A nitrogen to sulfur acyl shift changes the amide linkage between Smad2 and the intein to a thioester bond involving the intein's active site cysteine. Subsequent thiolysis generates recombinant Smad2 containing a C- terminal thioester. This was then ligated to a caged phosphopeptide synthesized using 79

solid phase methods to possess an N-terminal cysteine and 2-nitrophenylethyl caged phosphorylated serines. While EPL has also been used to generate “activated” proteins, for example through the incorporation of phosphomimetic modifications such as non- hydrolysable phosphotyrosine modifications (Lu et al., 2001), without the photolabile caging moiety that approach is subject to the same limitations as expressing mutant proteins.

Unnatural amino acid mutagenesis

Nonsense suppression mutagenesis (Wang and Schultz, 2002) has also been used to generate photocaged proteins. This method permits unnatural amino acids to be introduced at a specific location in a protein, allowing for new functionality outside that available in the twenty naturally encoded amino acids, providing that the ribosome can accept the novel unnatural amino acid. Some of the first proteins to be caged in this manner were T4 lysozyme, the active site of which was caged by introducing an aspartyl

-o-nitrobenzyl ester in place of an essential aspartic acid residue (Mendel et al., 1991), and the GTPase p21Ras, which was caged by introducing an aspartyl -o-nitrobenzyl ester in place of an aspartic acid reside at a location that is essential for interaction with the

GTPase activating protein p120-GAP (Pollitt and Schultz, 1998). Miller and coworkers also generated photocaged nicotinic acetylcholine receptors by introducing ortho- nitrobenzyl tyrosine in place of tyrosine in the protein using unnatural amino acid mutagenesis. Because they did this in Xenopus oocytes, they were able to perform in vivo electrophysical studies of the behavior of a protein in a time-resolved manner (Miller 80

et al., 1998). Nonsense suppression mutagenesis does have limitations however; not a great deal of protein is generated using this method and overall, the process of unnatural amino acid mutagenesis itself is not trivial. In addition, the photocage moiety is constrained to a single preselected structure, and since there is the possibility that the modification may fail to make the protein behave as anticipated, the generation of the caged protein with the desired properties can be lengthy.

Another consideration for all of the methods described above, with the exception of the work done by Miller and coworkers in Xenopus oocytes, is that the caged protein is generated in vitro and must be introduced into living cells. In the instances where studies were performed in living cells, this was achieved by microinjection of single cells.

Phytochrome fusion proteins

Proteins that are themselves responsive to light can also be used to generate photosensitive proteins. For example, Leung and coworkers employed a genetic approach to allow for the photoactivated polymerization of actin (Leung et al., 2008). In these studies, they fused the Rho GTPase Cdc42 to the light-sensing domain of phytochrome B (PhyB) and Cdc42’s effector Wiskott-Aldrich Syndrome Protein (WASP) to the light dependent PhyB binding domain of phytochrome interacting factor 3 (Pif3).

Thus, upon irradiation with red light, PhyB and Pif3 bind to one another, bringing Cdc42 and WASP in close proximity and allowing Cdc42 to activate WASP, which resulted in activation of the ARP2/3 complex and subsequent actin assembly. 81

Overall, photocaging is a powerful method to allow for the control of cellular function through the alteration of protein activation states in a spatial and temporal manner. A family of proteins called the Rho GTPases are a prime example of a proteins whose study could further benefit from methods like photocaging.

CHAPTER V: Rho GTPase Background

Introduction

The family of proteins called the Rho GTPases are typically thought of in terms of their role in controlling the dynamics of the actin and microtubule cytoskeleton, allowing them to modulate the cell’s ability to maintain a polarized state. Rho GTPase control of cellular polarity is also important to ensure proper cytoskeletal dynamics throughout the cell cycle and in cellular processes such as phagocytosis, targeted vesicle secretion, axon and dendrite formation, and migration. These proteins are also involved in apoptosis, gene expression and the immune response and as a result, improper Rho GTPase signaling is associated with many disease states, including cancer, hypertension, diabetes, asthma, pathogenic infection, and neurological disorders (Boettner and Van Aelst, 2002).

The Rho GTPase family of proteins makes up one of the branches of the Ras superfamily of small GTP binding proteins (Colicelli, 2004; Wennerberg et al., 2005).

The Rho GTPase family consists of twenty-five proteins encoded by twenty-two genes.

The family can be further divided into subfamilies based on homology, including the

Rho-like GTPases RhoA, B, and C, the Rac-like GTPases Rac1, 2, 3, and RhoG and the

Cdc42-like GTPases Cdc42, TcL, Tc10, Wrch1 and Wrch2 (Wennerberg and Der, 2004).

The gene encoding the Rho GTPase RhoA was the first to be identified (Madaule and

Axel, 1985), followed by those encoding Rac1 and Rac2 (Yamamoto et al., 1988), and that encoding Cdc42 (Munemitsu et al., 1990). These ~21 kDa proteins are found in all eukaryotic organisms. Of the Rho GTPases, RhoA, Rac1, and Cdc42 are best 83

characterized and classically have been credited with controlling stress fiber and focal adhesion formation, lamellipodia formation, and filopodia formation, respectively (Nobes and Hall, 1995; Ridley and Hall, 1992).

Factors that control Rho GTPase activation states

Cellular control of the location and timing of Rho GTPase activation is important to maintain proper Rho GTPase signaling. In general, Rho GTPases can be thought of as tightly regulated molecular “switches” whose activation state is controlled by several families of proteins (Figure 5-1). Guanine-nucleotide exchange factors (GEFs) activate

GTPases by promoting the disassociation of GDP and its replacement with GTP

(Erickson and Cerione, 2004; Hoffman and Cerione, 2004; Whitehead et al., 1997).

GTPase-activating proteins (GAPs) turn the GTPases off by accelerating their rate of

GTP hydrolysis (Moon and Zheng, 2003). Thus, the location of the GEFs and GAPs control where, when, and how long Rho GTPases are active. Over the years dozens of

GEFs and GAPs have been identified. Another family consisting of three proteins called

Rho guanine nucleotide dissociation inhibitors (RhoGDIs) control GTPase activation states by sequestering GTPases in the cytosol and preventing both nucleotide exchange and hydrolysis (Fukumoto et al., 1990). Once active, GTPases are able to interact with effector proteins to promote their specific and localized effects. Because the Rho family members are highly homologous, with 40-95% amino acid identity within the family

(Wennerberg and Der, 2004), many of their GEFs, GAPs, and effectors possess overlapping specificity. This overlapping specificity, and the nature of the processes that 84

the Rho GTPases control, mean that tight spatial and temporal control of their activation is essential for their proper function.

Figure 5-1. Rho GTPase cycle. Guanine-nucleotide exchange factors (GEFs) activate GTPases by promoting the disassociation of GDP and its replacement with GTP. GTPase-activating proteins (GAPs) turn the GTPases off by accelerating their rate of GTP hydrolysis. Rho guanine nucleotide dissociation inhibitors (RhoGDIs) control the GTPase activation state by sequestering GTPases in the cytosol and preventing both nucleotide exchange and hydrolysis. Once active, GTPases are able to interact with effector proteins to promote their specific and localized effects.

There is a wealth of structural information available about the Rho GTPase family. Rho GTPases have been studied through X-ray crystallography and NMR spectroscopy in the inactive state complexed with GDP with (Wei et al., 1997) and without (Shimizu et al., 2000) Mg 2+, in the active state complexed with non-hydrolyzable 85

GTP analogs (Hirshberg et al., 1997; Ihara et al., 1998), bound to GAP proteins (Rittinger et al., 1997a; Rittinger et al., 1997b), GEF domains (Rossman et al., 2002; Worthylake et al., 2000), effector domains (Dvorsky et al., 2004; Maesaki et al., 1999a; Maesaki et al.,

1999b), and RhoGDI (Hoffman et al., 2000; Longenecker et al., 1999; Scheffzek et al.,

2000), to name a few. These studies have shed a great deal of light on the structural aspects of the GTPases that play a role in their regulation by these families of proteins, as well as the subtle changes that occur in the GTPase conformation upon activation.

Posttranslational modification of the Rho GTPases also plays a role in their proper localization and function (Glomset and Farnsworth, 1994). Specifically, many of the family members possess an isoprenylation site on their C-terminus (CAAX, where C is cysteine, A is any aliphatic amino acid, and X is the C-terminal amino acid) (Zhang and

Casey, 1996). After the GTPase is expressed, the protein CAAX prenyltransferase modifies the cysteine of the CAAX box with a prenyl group, either a farnesyl group or a geranylgeranyl group. The last three amino acids are then removed by the Rce1 CAAX protease and the C-terminus of the GTPase is methylated by ICMT Methyltransferase

(Adamson et al., 1992; Katayama et al., 1991; Roberts et al., 2008). In general, Rho

GTPases are thought to be in the cytoplasm when GDP-bound and inactive and localized to membranes when GTP-bound and active. The specific membranes to which each family member localizes is thought to be dependent on the hypervariable C-terminal polybasic domain upstream of the prenylation site (Williams, 2003).

Active and Dominant Negative Rho GTPases

Despite recognition of the importance of Rho GTPases in a wide variety of cellular processes, techniques used to study their activity often fail to appreciate spatiotemporal aspects of their regulation. Much of what is known about the function of the Rho proteins has involved the use of dominant negative (DN) and constitutively active (CA) or fast cycling (FC) mutant forms of the proteins. DN mutants possess a threonine to asparagine mutation at amino acid 17 (T17N) (Feig and Cooper, 1988;

Powers et al., 1989) that makes them very poor magnesium binders (John et al., 1993), resulting in very low affinity for nucleotide. Because GEF proteins no longer have to compete with nucleotide for binding in the nucleotide-binding pocket, their affinity for the DN GTPase is much higher. Thus, when the DN GTPase is expressed in cells, it titrates available GEF protein away from endogenous GTPases, leaving them in an inactive or GDP bound state (Feig, 1999).

In WT Rho GTPases, the glutamine at position 61 (Q61) coordinates a catalytic water molecule that is required for the hydrolysis of the bond between the  and  phosphate of GTP. CA mutants possess a mutation of either Q61 itself (Q61L), or another nearby amino acid (G12V), that makes them unable to coordinate the catalytic water necessary to hydrolyze GTP to GDP. Thus, because there is approximately a 10 fold excess of GTP relative to GDP in cells, CA mutants are locked in the active conformation and able to interact with effector proteins to cause downstream effects everywhere in the cell in which it is expressed. These mutations were first identified in 87

the Rho-related GTPase Ras and have been found in more than 30% of all human tumors

(Barbacid, 1987).

Fast-cycling mutants, on the other hand, possess a mutation of phenylalanine 28 to leucine (F28L), which removes the stabilizing - orbital interactions between the nucleotide and the amino acid in the binding pocket (Lin et al., 1997; Lin et al., 1999).

This allows FC mutants to load nucleotide on their own without the help of GEF proteins while maintaining their ability to hydrolyze GTP to GDP. Again, due to the excess of

GTP to GDP in the cell, FC mutants are constantly cycling to a GTP bound, or active conformation.

While these mutants have been highly informative, their use is limited by the lack of control of the timing and location of GTPase activation. DN and CA/FC mutants are off or on indiscriminately throughout the cell for times far exceeding that typical of their endogenous counterparts, making events controlled by their spatiotemporal inactivation/activation difficult to study. Additionally, because GEFs can possess activity toward multiple GTPases (Zheng et al., 1995), which may change with changing cell cycle stage or intracellular location, the phenotypes resulting from expression of the DN form of one GTPase may actually result from the inhibition of multiple GTPases. This is also true for the CA GTPases, which possess much higher affinity for their effectors and

GAPs than wild-type GTPases (Burridge and Wennerberg, 2004; Owen et al., 2000;

Wennerberg and Der, 2004). Consequently, the results of experiments using DN or CA

GTPases are at times contradictory and often difficult to interpret.

Alternate methods to control Rho GTPase activity

Besides direct mutation leading to altered protein activity, methods to modulate the activity of a single GTPase to study its specific function have included the use of siRNA and the development of small molecule or peptide inhibitors. Unfortunately, the reduction of the levels of one Rho GTPase has been observed in some systems to result in the compensatory overexpression of other Rho GTPase family members (Bielek et al.,

2009). In addition, due to the high homology between the family members, it is very difficult to design small molecule inhibitors that will target only one family member and to date very few have been developed. One approach that was taken was the design of, or screening for, small molecules or peptides that would block interactions with GEFs, thus blocking activation. The first reported small molecule inhibitor of Rac1, NSC23766

(Gao et al., 2004) binds in a GEF recognition groove on Rac1, masking tryptophan 56, a residue that is a critical factor for selective binding of Rac1 specific GEFs Trio and

Tiam1 (Gao et al., 2001) However, this small molecule does not block activation by all

GEFs, as the GEF Vav1 can still activate Rac1 in the presence of NSC23766. A polypeptide derived from Rac1 containing tryptophan 56 has also been shown to specifically inhibit Rac1:GEF binding (Gao et al., 2001). Additionally, the peptide Trio

Inhibitory Peptide alpha (TRIP) inhibits RhoA activation by specifically targeting and binding to the RhoA-specific GEF domain of TRIO-GEFD2 (Schmidt et al., 2002).

However, by design this inhibitor impacts RhoA activation by only a single GEF. The minimum Cdc42-binding domain of the Cdc42 specific tyrosine kinase ACK1 (ACK42) has also been shown to bind to GTP-Cdc42 and block interaction with its effector 89

proteins (Nur et al., 1999). This effector domain approach is highly specific to Cdc42 inhibition. However like many of these approaches it does not allow for spatial resolution of the achieved inhibition.

The small molecule inhibitor EHT1864 functions in a slightly different manner to inhibit Rho GTPases. EHT 1864 inactivates Rho GTPases by causing nucleotide release

(Shutes et al., 2007). However EHT 1864 inhibition is not selective; it possesses activity towards multiple family members including Rac1, Rac1b, Rac2 and Rac3.

Another approach that was developed was to target the effectors of the Rho

GTPases directly. Several small molecules have been generated that block the activity of an effector of RhoA, Rho kinase (ROCK). These include the Rho kinase inhibitors fasudil (HA-1077), Y-27632, and H-1152 (Ikenoya et al., 2002; Nagumo et al., 2000;

Uehata et al., 1997). However, at higher concentrations all of these small molecules can affect the activity of other kinases including PKC, PKA and MLCK (Ishizaki et al., 2000;

Sasaki et al., 2002; Tamura et al., 2005; Uehata et al., 1997). Furthermore, only a subset of the functions of RhoA will be impacted by the use of these inhibitors, since GTP-

RhoA can still interact with downstream effectors other than ROCK (Bishop and Hall,

2000).

Modulation of the isoprenoid modifications, and thus localization, of Rho

GTPases using small molecule inhibitors of the prenyltransferases (Basso et al., 2006;

Sebti and Hamilton, 2000) or cholesterol lowering statins (Riganti et al., 2008) have also been used to control Rho GTPase activity. However, these types of approaches will impact all proteins that rely on these types of lipid modifications for proper localization 90

and function and can affect other processes as well (Budzyn et al., 2006; Chrissobolis and

Sobey, 2006; Liao and Laufs, 2005; Rikitake and Liao, 2005).

There are also several bacterial proteins that can be introduced into cells that modulate Rho GTPase activity. One is the cytotoxic necrotizing factor 1 (CNF1) from

Escherichia coli (Lerm et al., 1999; Schmidt et al., 1997). CNF1 activates RhoA, Rac1 and Cdc42 by deamidation of Q61, the residue essential for GTP hydrolysis. Clostridium botulinum exoenzyme C3 and related transferases specifically inhibit RhoA, RhoB, and

RhoC, but not Rac1 or Cdc42 by ADP-ribosylating the GTPase on asparagine 41 in the effector binding domain (Aktories and Koch, 1997; Just et al., 1992; Sekine et al., 1989).

Clostridium difficile toxins A and B glucosylate RhoA, Rac1, and Cdc42 on threonine 35, a residue that participates in the coordination of Mg2+ (Just et al., 1995a; Just et al.,

1995b). Clostridium sordellii’s lethal toxin (LT) glucosylates Rac1 and Cdc42 but not

RhoA, but it also can modify Ras proteins as well (Just et al., 1996; Popoff et al., 1996).

Yersinia pseudotuberculosis cytotoxin YopE exhibits GAP activity towards RhoA, Rac1 and Cdc42, increasing the rate at which they hydrolyze GTP to GDP and inactivating them (Von Pawel-Rammingen et al., 2000).

Overall, all of the methods described above have been very useful in investigating the different roles of the Rho GTPase family members. However, they are all limited in either their specificity and/or the degree to which they block an interaction or activity.

Furthermore, they all lack spatial, and in some cases also temporal, resolution.

Techniques to observe Rho GTPase activity

Other methods to study Rho GTPases have involved the development of tools to detect GTPase activation states. It is important to remember that traditional tools that allow for visualization of protein localization will not differentiate between the active and inactive forms of the GTPases.

One method used to assess the activation state of Rho GTPases is the pulldown of active GTPase with effector binding domains. While the use of effector domains to pull down active GTPase from whole cell lysate of synchronized cell populations can provide a temporal assessment of GTPase activation state, it does not discriminate spatially and furthermore, its yields can be inefficient (Kimura et al., 2000). Also, the time-scale resolution of this type of assay may not reflect the rapid activation events actually occurring. In addition, certain methods of cell cycle synchronization can actually influence GTP-RhoA and GTP-Rac1 levels (Ren et al., 1999; Waterman-Storer et al.,

1999), potentially altering the timing and dynamics of the process being studied.

One simple method that has been employed to visualize Rho GTPase activation is the use of effector binding domains fused to fluorescent proteins (Figure 5-2A). Thus, the fluorescently tagged binding domain, upon expression in cells, will translocate to areas where active Rho GTPases are located. This method has, for example, been used successfully to visualize the dynamic patterns of RhoA and Cdc42 activation during wound healing in a Xenopus embryo (Benink and Bement, 2005). Unfortunately, because many effectors can interact with multiple Rho GTPase family members, tools of 92

this type may actually report on the activation of multiple GTPases (Benard et al., 1999;

Ren et al., 1999).

Figure 5-2. Biosensors to detect Rho GTPase activation. A). GFP-tagged Rho effector binding domain (RBD) interacts specifically with endogenous GTP loaded active Rho GTPase. B). FRET occurs when a dye (red circle) labeled RBD interacts with GFP- tagged GTP loaded active Rho GTPase. C). and D). FRET occurs upon interaction between CFP and YFP when an RBD and Rho GTPase either C). sandwiched between them, or D). fused to the termini of each, bind one another upon GTP binding and activation. E). FRET is relieved between CFP and YFP upon binding of the RBD sandwiched between them to endogenous GTP bound active Rho GTPase. F). Solvatochromic-dye-based sensor. The red star represents the solvatochromatic dye on the RBD, which exhibits increased fluorescence intensity upon binding to GTP loaded active Rho GTPase. The black C-terminal extension on the GTPases and CFP in C). represents the prenyl modification. (Adapted from Pertz O., J. Cell Sci., 2010; 123: 1841-1850.) 94

Another method that was developed relied on a fluorescence resonance energy transfer (FRET) based approach, where an effector was labeled with the cysteine- selective iodoacetamide dye Alexa 546 on an introduced cysteine and injected into cells expressing a Rho GTPase tagged with green fluorescent protein (GFP) (Figure 5-2B).

Thus, once the GTPase became active and bound the effector, FRET could occur between the dye and fluorescent protein and report on the GTPase’s activation (Kraynov et al.,

2000). A similar experiment was conducted with genetically encoded yellow and cyan fluorescent protein (YFP and CFP) tagged GTPase and effector (Kraynov et al., 2000).

Unfortunately, the fact that acceptor and donor fluorophores are on separate proteins, which may be localized differently in the cell, means that analyzing these type of experiment is complicated.

FRET-based probes modeled after Ras and interacting chimeric unit (Raichu)

(Itoh et al., 2002; Mochizuki et al., 2001; Seth et al., 2003) have also been developed to detect activation of Rho GTPases in cells. One such probe design consisted of an effector domain and its respective GTPase sandwiched between two fluorescent proteins, such that upon GTP binding and interaction of the GTPase and effector domain, FRET occurs and the emission ratios shift (Figure 5-2C) (Itoh et al., 2002; Yoshizaki et al., 2003).

Prenylation of this probe was accomplished by appending the C-terminal polybasic

CAAX sequence of the GTPase onto one of the fluorescent flanking proteins. However, this means that the fusion is targeted constitutively to the membrane and regulation by

RhoGDI proteins is prevented. 95

To address the lack of regulation by RhoGDIs and the constitutive membrane localization, an alternate probe type was designed where the two fluorescent proteins were sandwiched between the effector binding domain and the Rho GTPase (Figure 5-

2D). This left the C-terminus of the GTPase free to be posttranslationally modified

(Pertz et al., 2006). In general however, Raichu based probes are by design much bulkier than the GTPase alone and may not behave exactly as the endogenous GTPase would.

Another probe design addressed prenylation by sandwiching only the effector binding domain between the two fluorescent proteins (Figure 5-2E) (Itoh et al., 2002;

Yoshizaki et al., 2003). Thus in solution, the two fluorescent proteins undergo FRET due to their close proximity. However, upon binding of activated endogenous GTPase, a conformational change in the effector molecule occurs, relieving the interaction between the two fluorescent proteins and resulting in a loss of FRET. Another non-FRET based approach to visualize Rho GTPase activation relied on a solvatochromatic dye fused to an effector binding domain (Figure 5-2F) (Nalbant et al., 2004). In this case, the change in solvent polarity upon binding of the labeled effector binding domain to an activated

GTPase resulted in an increase in fluorescence intensity. However, much like effector binding domains tagged with a singe fluorescent protein, both of these types of probe designs may actually report on the activation of multiple Rho GTPase family members.

The Rho GTPases are involved in a wide variety of highly divergent processes in the cell, many of which require the polarized localization of cellular substituents at specific times. This in turn necessitates the tight regulation of the Rho GTPases and the numerous proteins that control their state of activation. In order to explore the regulation 96

of specific Rho GTPases and their regulatory components, as well as the spatial and temporal aspects of their activation in processes such as cell division, spatiotemporal control of their activation is critical. However, because of limitations inherent to the current methodology for studying Rho GTPases, definitive correlations between specific functions and phenotypes are difficult. Tools that can control individual GTPase activation states at specific intracellular cites and at specific times during the cell cycle could clarify the question of the roles the GTPases play in cell division, as well as in many other processes.

CHAPTER VI: Leveraging a Small-Molecule Modification to Enable the

Photoactivation of Rho GTPases

Introduction

Control of both the timing and location of protein activation is essential for cell growth and movement. One of the key ways in which cells accomplish this control is through the use of molecular switches known as Rho GTPases. When loaded with GTP, these proteins are activated and can bind to effector proteins that activate cellular responses (Figure 6-1). Subsequent hydrolysis of GTP to GDP returns the GTPase to the inactive GDP-bound state. The location of this molecular switch, and the timing of its activation, determines where and when cellular machinery is recruited to perform a specific biological function. Inappropriate activation of Rho GTPases alters this spatiotemporal control and is found to occur in many human diseases including cancer, hypertension, diabetes, asthma, pathogenic infection, and numerous neurological disorders (Boettner and Van Aelst, 2002).

Figure 6-1. Using photolabile small molecules to control Rho GTPase activity. A) Exchange of GTP for GDP causes Rho GTPases to undergo a conformational change that allows them to interact with effector proteins and activate cellular responses. B) Alkylation of the endogenous Cys18 in the nucleotide binding pocket of GTPases with a small-molecule photocage should prevent GTP binding and thus effector binding. C) Irradiation with UV light should remove the photocage, restoring Cys18 and allowing the GTPase to adopt the active conformation.

Traditionally, GTPase functions in living cells have been studied by the overexpression or microinjection of mutant forms of the proteins (Nobes and Hall, 1995).

Constitutively active mutants cannot hydrolyze GTP, and thus remain perpetually in an active state. Dominant negative mutants inhibit the activation of endogenous GTPases, while fast-cycling mutants can cycle between active and inactive states independently of 99

cellular regulation (Lin et al., 1999). The use of these mutants has yielded considerable insight into the roles of Rho GTPases in different cellular processes. However, this type of approach lacks the spatial and temporal control needed to dissect the individual roles of specific GTPases at different subcellular locations and times during the cell cycle. In an effort to overcome this limitation, we have developed a general approach to control

Rho GTPase activation with light (Leung et al., 2008).

As discussed, photoactivatable (“photocaged”) small molecules have been widely used to affect the rapid release of an active species at a defined location and time (Ellis-

Davies, 2007; Leung et al., 2008). Proteins have been photocaged by unnatural amino acid mutagenesis approaches (Chen et al., 2009; Edwards et al., 2009), the use of phytochrome fusion proteins (Leung et al., 2008), and by direct covalent modification with photolabile small molecules (Lee et al., 2009). In vitro modification followed by loading into live cells has successfully been used to allow spatial and temporal control of cellular protein activity. Notable examples include PKA (Curley and Lawrence, 1998a), cofilin (Ghosh et al., 2002), and Smad2 (Hahn and Muir, 2004). Nonetheless, the modification, purification, and selective functional disruption of photocaged proteins remains challenging. To begin to address these challenges, we took advantage of the wealth of structural information available for the Rho GTPase family.

Most Rho GTPases contain a cysteine residue within the nucleotide binding pocket. We predicted that alkylation of this cysteine in the Rho GTPase Cdc42 would block the ability of GTP to bind and activate Cdc42 (Figure 6-1B) and that subsequent 100

irradiation with long-wave UV light would release this small molecule, restoring Cdc42’s

GTP-binding capacity (Figure 6-1C).

Materials and Methods

Materials

Chemicals for synthesis were from Aldrich unless otherwise noted. Amino- mPEG5K was from Fluka. NMR spectra were recorded on a Varian Mercury 400 MHz

NMR. Mass spectra were acquired at the UMass Proteomic Mass Spectrometry Lab on a

Waters QTof Premier mass spectrometer (small molecule exact mass) or Thermo

Scientific LCQ and LTQ ion trap mass spectrometers (protein mass spectra). Compound

1 (2-bromo-2-(2-nitrophenyl)acetic acid) was synthesized as previously described (Chang et al., 1995). The HaloTag ligand, 2-[2-(6-Chlorohexyloxy)-ethoxy]-ethylamine, was synthesized as previously described (Zhang et al., 2006). Protein concentrations were determined using Coomassie Plus (Thermo Scientific). Buffer exchange was performed using Zeba gel-filtration columns (Pierce). Immobilized glutathione (Thermo Scientific) was used for both GST-tagged protein purification and pull-down experiments. Unless otherwise stated, all protein purification steps we carried out at 4˚C.

Synthesis of 2-Bromo-N,N-diethyl-2- (2-nitrophenyl)-acetamide (2)

To a solution of 1 (0.2 mmol, 52 mg) in dichloromethane (2 mL) was added DIC

(46.5 mL, 0.3 mmol) and diethylamine (31 mL, 0.3 mmol). The solution was left at RT 101

for 3h, then directly purified by flash chromatography (0-25% ethyl acetate in hexanes).

1 Yield 27.7 mg clear film (44%). H-NMR (400 MHz, CDCl3): d 8.09 (dd, 1H, J = 1.6,

8.4 Hz), 7.88 (dd, 1H, J = 1.2, 8 Hz), 7.62 (dt, 1H, J = 1.6, 8 Hz), 7.41 (dt, 1H, J = 1.2, 8

Hz), 6.27 (s, 1H), 3.58-3.29 (m, 4H), 1.27 (t, 3H, J = 7.6 Hz), 1.07 (t, 3H, J = 7.6Hz).

13 C-NMR (100 MHz, CDCl3): d 166.20, 147.69, 135.08, 133.82, 131.98, 129.64, 124.61,

42.99, 41.40, 41.29, 14.02, 12.50. Calculated for C12H16N2O3Br: 315.0344, found:

315.0374.

Synthesis of BNPA-mPEG5K (3)

To a solution of amino-mPEG5K (100 mg, ~ 0.17 mmol amine/g) and 1 (26 mg,

0.1 mmol) in dichloromethane (1 mL) was added DIC (15.5 mL, 0.1 mmol). The solution was incubated overnight at RT, then diluted into 50 ml of 2-propanol. The solution was chilled on ice for 30 minutes, during which time the polymer precipitated. The solution was filtered and washed with room-temperature 2-propanol. Residual solvent was

1 removed in vacuo to afford a white powder. H-NMR (400 MHz, CDCl3): d 7.98 (dd, 1H,

J = 1.6, 8.4 Hz), 7.88 (dd, 1H, J = 1.2, 8 Hz), 7.65 (dt, 1H, J = 1.2, 8 Hz), 7.50 (dt, 1H, J

= 1.2, 8 Hz), 6.10 (s, 1H), 3.63 (s, >400H), 3.37 (s, 3.45H). Small apparent peaks near the polymer resonance at 3.63 were difficult to resolve. The BNPA substitution of the polymer was estimated to be 82% based on the relative integration of the well-resolved mPEG methoxy peak at 3.37 ppm and the benzylic methine proton at 6.10 ppm.

102

Synthesis of 2-Bromo-N-[2-[2-(6-chlorohexyloxy)-ethoxy]-ethyl] -2- (2-nitrophenyl)- acetamide (BNPA-HT, 4)

To a solution of 1 (52 mg, 0.2 mmol) in dichloromethane (1 mL) was added 2-[2-

(6-Chlorohexyloxy)-ethoxy]-ethylamine (45 mg, 0.24 mmol) followed by DIC (39 mL,

0.25 mmol). The solution was stirred overnight at RT and purified by flash chromatography (0-50% ethyl acetate in hexanes). Yield 39 mg (42%). 1H-NMR (400

MHz, CDCl3): d 8.00 (dd, 1H, J = 1.6, 8.4 Hz), 7.84 (dd, 1H, J = 1.2, 8 Hz), 7.66 (dt, 1H,

J = 1.2, 8 Hz), 7.50 (dt, 1H, J = 1.6, 8 Hz), 7.17 (br s, 1H), 6.10 (s, 1H), 3.68-3.5 (m, 8H),

3.53 (t, 2H, J = 6.8 Hz), 3.49 (t, 2H, J = 6.8 Hz), 1.78 (m, 2H), 1.62 (m, 2H), 1.5-1.35 (m,

4H). 13C-NMR (100 MHz, CDCl3): d 166.34, 147.89, 134.05, 132.75, 132.55, 130.04,

125.28, 71.56, 70.72, 70.31, 69.55, 45.31, 45.01, 40.54, 32.74, 29.70, 26.93, 25.66.

Calculated for C18H27BrClN2O5: 465.0792, found: 465.0743.

Plasmid constructs

Mutant proteins were created using the Quickchange site-directed mutagenesis kit

(Stratagene). Truncated Cdc42 F28L/C188S was made by PCR of the full-length gene with the primers gaatggatccatgcagacaattaagtgtgttgttgtgg and gaacgaattctcaagaccgcggctcttcttcggttctgg. This PCR product was then digested with

BamHI and EcoRI and ligated into the corresponding sites in pGEX6P-1. The F28L point mutation was introduced with the primers cctacacaacaaacaaactgccatcggaatatgtaccg and cggtacatattccgatggcagtttgtttgttgtgtagg. The C18S mutation was introduced using the primers gctgttggtaaaacatctctcctgatatcctac and gtaggatatcaggagagatgttttaccaacagc. 103

Truncated Rac1 F28L/C189S was made by PCR of the full-length gene with primers gaatggatccatgcaggccatcaagtgtgtggtg and gaacgaattctcaagattttctcttcctcttcttcacgggag. This

PCR product was also digested with BamHI and EcoRI, and ligated into pGEX6P-1. The

F28L point mutation was introduced into the resulting plasmid with primers gttacacaaccaatgcactgcctggagaatatatccctac and gtagggatatattctccaggcagtgcattggttgtgtaac.

Protein expression

Cdc42 F28L/C188S (residues 1-188), Rac1 F28L/C189S (residues 1-189) and the p21-activated kinase binding domain (PBD) (residues 66-150) were expressed as GST fusion proteins from the vector pGEX6P-1 in the BL21(DE3) E. coli strain as previously described (Bhunia and Miller, 2007). HaloTag (Promega) was expressed as a His6 fusion protein from the vector pET28 in the BL21(DE3) pLysS E. coli strain. Cells were grown at 37˚C until the OD600 reached 0.8, then induced with 0.1 mM IPTG and incubated at

20˚C overnight. Cells were pelleted at 5000 rpm in a Sorvall 2C3C Plus centrifuge

(H600A rotor) at 4˚C for 10 minutes, then flash frozen in liquid nitrogen and used immediately or stored at -80˚C.

GTPase and PBD purification

E. coli cell pellets from one liter of culture were thawed on ice, resuspended in 25 ml Lysis Buffer (50 mM Tris or HEPES pH 7.4, 500 mM NaCl, and 0.5% Tween-20) containing 1 mM PMSF and disrupted by sonification (Branson Sonifier). Dithiothreitol

(DTT) was added at 10 mM and the resulting cell lysate was clarified by centrifugation at 104

35K rpm in a Beckman 50.2Ti rotor for 1 hour. The supernatant was batch-bound to immobilized glutathione for 1h at 4˚C, and then the beads were washed with Lysis Buffer containing 10 mM DTT followed by Wash Buffer (50 mM Tris pH 8.1, 250 mM NaCl and 10 mM DTT).

For Cdc42 and Rac1, the beads were then incubated with Nucleotide Removal

Buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM TCEP, and 20 mM EDTA) and finally washed with Storage Buffer (50 mM HEPES pH 7.4, 50 mM NaCl, 1 mM TCEP, and 5 mM EDTA). Twenty units of PreScission Protease (GE Healthcare) were added and incubation continued overnight at 4˚C to cleave the GST fusion and elute the untagged

GTPase.

For PBD, 3 mg/mL reduced glutathione in Wash Buffer was used to elute the

GST-tagged PBD, which was then dialyzed against 50 mM Tris pH 7.4, 150 mM NaCl, 1 mM TCEP, and 1 mM EDTA.

HaloTag purification

E. coli cell pellets were thawed on ice, resuspended in 50 mM NaPhos pH 7.8,

500 mM NaCl, 10 mM imidazole, and 0.1% Tween-20 containing 0.2 mM PMSF and disrupted by sonification. BME was added at 5 mM and the resulting cell lysate was centrifuged at 35K rpm in a Beckman 50.2 Ti rotor for 1 hour at 4˚C. The clarified supernatant was batch-bound to Talon resin (BD Biosciences) for 1 hour at 4˚C. The beads were washed with 50 mM NaPhos pH 7.8, 500 mM NaCl, 30 mM imidazole and 5 mM BME. The His6 HaloTag was eluted with 50 mM NaPhos pH 7.8, 500 mM NaCl, 105

300 mM imidazole and 5 mM BME, then dialyzed against 50 mM Tris pH 7.4, 150 mM

NaCl, 1 mM TCEP, and 1 mM EDTA.

Small molecule modification of Cdc42

Cdc42 (1-188, F28L/ C188S) was expressed and purified as previously described

(Bhunia and Miller, 2007). The GTPase was made nucleotide-free by incubation with 20 mM EDTA for 15-30 min at RT, and the unbound nucleotide removed by passage over two sequential Zeba gel-filtration columns (Pierce) equilibrated with 50 mM HEPES pH

7.4, 50 mM NaCl, and 5 mM EDTA. Ten equivalents of the small molecule photocage were then added and allowed to react in the dark at 22°C for 3h with shaking at 300 rpm

(Eppendorf Thermomixer). Unreacted alkylating agent was removed by passage over a

Zeba column equilibrated with 50 mM Tris pH 7.4, 150 mM NaCl, 1mM TCEP, and 1 mM EDTA.

Recruitment of HaloTag protein to Cdc42 modified with 4

Cdc42 that had been alkylated with 4 was incubated with 0.3-0.5 eq of purified

HaloTag protein at 22°C in the dark for 1h with shaking at 300 rpm. The HaloTag- modified GTPase was purified by FPLC (AKTA) on a MonoQ ion-exchange column (GE

Healthcare). The unmodified GTPase and the HaloTag-modified GTPase eluted as separate peaks on a linear gradient of 0-1M NaCl over 25 min at a flow rate of 1 ml/min.

The purity of the fractions was assessed by SDS-PAGE with Coomassie staining.

Purified HaloTag-modified GTPase was used directly for photolysis reactions and 106

effector pull-down assays. The final buffer composition was determined to be 50 mM

Tris pH 7.4, 300 mM NaCl, 1 mM DTT and 1 mM EDTA.

GTPase effector pulldown assay

The ability of each GTPase to adopt an active conformation was assessed using a modification of the standard GST-PBD pulldown assay (Benard et al., 1999). GTPases were loaded with 10 mM MgCl2 and 4 molar equivalents of either GDP or GTPγS for

15-30 min at room temperature. PBD was loaded onto immobilized glutathione at 4 molar equivalents relative to the concentration of the GTPase for 30 minutes, and the nucleotide-loaded GTPase was then added. The proteins were allowed to interact at 4˚C for 30 minutes, and then the unbound fraction was removed. The beads were washed twice with 8 volumes of 50 mM Tris pH 7.4, 150 mM NaCl, 2 mM MgCl2, 1 mM TCEP, and 0.1% Triton X-100, and then 4 volumes of 40 mM Tris pH 7.4, 100 mM NaCl, 20 mM MgCl2 and 1 mM TCEP. The bound fraction was eluted from the beads by boiling in

SDS-PAGE gel loading buffer. The bound and unbound fractions for each nucleotide state were compared by SDS-PAGE with Coomassie staining.

Fluorescence Polarization Assay

The ability of GTP to bind the GTPase was evaluated by a fluorescence polarization assay, performed in triplicate in a 384-well plate. Each well contained 30 L of 50 mM HEPES pH 7.4, 150 mM NaCl, 20 mM MgCl2 and 1 mM TCEP buffer alone, or with 400 nM of either purified His6HaloTag, purified apo-Cdc42, Cdc42 preincubated 107

with 5 mM GTPS, FPLC-purified Cdc42-4-HT, or FPLC-purified Cdc42-4-HT that had been uncaged for 10 minutes at 350 nm as described above. BODIPY-FL-GTP

(Invitrogen) was added to each well at a final concentration of 100 nM immediately before reading. The fluorescence polarization was measured in a PerkinElmer Victor3

1420 Multilabel Counter with a 480 nm CW lamp filter and a 535 nm emission filter.

Measurements were taken for 1.0 second with 5 repeats per well.

Photolysis reaction

FPLC-purified HaloTag-modified GTPase was placed in a 384-well plate that had been preincubated at -20˚C, and irradiated with a 450W xenon arc lamp (Oriel 66021 lamp housing) through an IR cut-off filter and a 350 nm bandpass filter for times ranging from 10 seconds to 30 minutes.

Results and Discussion

To test the hypothesis that we could modify and block nucleotide binding in Rho

GTPases in a light sensitive manner, we attempted to alkylate cysteine 18 in a fast- cycling Cdc42 mutant (1-188, F28L/C188S). However, using previously reported photocages such as 2-bromo-2- (2-nitrophenyl)acetic acid 1 (Chang et al., 1995) (Scheme

6-1), we found that we could not completely separate the modified protein from the unmodified protein and that alkylation of this cysteine failed to completely prevent the ability of GTP to activate Cdc42 as determined by an effector pulldown assay (Figure 6-

2). Because any contaminating uncaged protein or residual activity of the caged protein 108

would compromise the spatial and temporal control afforded by photocaging, it is important that the caged protein be purified and inactive.

Scheme 6-1. Photocages used in CHAPTER VI. 2-bromo-2-(2-nitrophenyl)acetic acid (1) is widely used as a photocage, but fails to block Cdc42 activation by GTP. Bulkier derivatives such as 2 similarly fail to block Cdc42 activation. The mPEG polymer 3 did not alkylate cysteine 18. Photocage 4 blocks Cdc42 activation only after recruitment of HaloTag to its ligand. 109

Figure 6-2. Effector binding domain assay. Top). Alkylation with 1 alone does not prevent the GTP-dependent conformational change of Cdc42 necessary for activation of Cdc42 or its ability to bind to GST-PBD (Effector). Bottom). In a GST-PBD pulldown assay, both unmodified Cdc42 and Cdc42 modified with 1 bind the PBD effector in the presence of the non-hydrolyzable GTP analog GTPS. As expected, neither bind PBD in the presence of GDP. UB = unbound protein; B = bound protein.

Since previously reported photocages failed to block Cdc42 activation, we synthesized novel photocages with additional steric bulk (Marriott et al., 2003). However, these molecules either failed to block Cdc42 activation (e.g., 2) or failed to alkylate cysteine 18 (3). The apparent plasticity of the nucleotide binding pocket is surprising, and underscores the difficulty in photocaging protein functions, even when structural information is available. To surmount this obstacle, we developed a new general strategy to photocage protein interactions by leveraging a small molecule protein modification into a large one. 110

We report here a new photocage (4) that contains a ligand for the protein HaloTag

(Scheme 6-1). HaloTag is a 33 kDa mutant haloalkane dehalogenase that specifically forms a covalent bond with the HaloTag ligand (Los and Wood, 2007). After alkylation of Cdc42 with 4, we incubated the modified protein with purified recombinant HaloTag.

The bioorthogonality of this interaction allows the specific recruitment of HaloTag only to protein modified with 4. The resulting caged fusion protein Cdc42-4-HT is well resolved from both Cdc42 and HaloTag on SDS-PAGE (Figure 6-3B, lane 3). Moreover, the large negative charge of HaloTag greatly facilitates its purification by ion-exchange chromatography. Importantly, this property allowed us to completely separate the caged protein from the unmodified protein. Subsequent irradiation with long-wave UV light

(350 nm, 10 sec, 450W Xe arc lamp) readily releases Cdc42 from Cdc42-4-HT (Figure 6-

3B, lane 4). 111

Figure 6-3. Photocaging of Cdc42. A) modification with 4 alone fails to prevent GTP- induced activation of Cdc42, but upon recruitment of HaloTag (HT), the resulting fusion protein can no longer bind effector in a GTP-dependent manner. Photolytic uncaging of the Cdc42-4-HT fusion at 350 nm liberates Cdc42, which can again bind effector in a GTP-dependent manner. B) Lane 1: Cdc42; Lane 2: HT; Lane 3: Cdc42-4-HT; Lane 4: Cdc42-4-HT after irradiation at 350 nm. C) In a GST-PBD pulldown assay, only Cdc42- 4-HT fails to bind the PBD effector in the presence of the non-hydrolyzable GTP analog GTPS. UB = unbound protein; B = bound protein.

The recruitment of HaloTag to the nucleotide binding site of Cdc42 effectively leverages 4, a small molecule modification, into a large 33 kDa protein modification

(Figure 6-3A). Unlike the small molecule photocages – including 4 alone – we found that 112

the recruitment of HaloTag prevented the GTP-induced activation of Cdc42 as determined by the pull-down assay (Figure 6-3).

Cdc42 is activated and binds to the p21-binding domain (PBD) of its effector p21- activated kinase (PAK) only in the GTP-bound form (Figure 6-3C) (Benard et al., 1999).

Use of the non-hydrolyzable GTP analog, GTPS, prevents any hydrolysis to the inactive

GDP-bound form during this binding assay. Alkylation of the fast-cycling mutant of

Cdc42 with 4 alone fails to prevent activation by GTPS. In contrast, the recruitment of

HaloTag to Cdc42-4 prevents effector binding (Figure 6-3C) due to an inability to bind nucleotide, as assessed by a fluorescence polarization assay (Figure 6-4). Specifically, the degree of polarized light measured in buffer containing BODIPY-FL-GTP alone (1), or BODIPY-FL-GTP with the HaloTag protein (2) or Cdc42 protein with 5 mM GTPS competitor (4), was relatively low. These low levels indicate that the BODIPY-FL-GTP is rotating rapidly and free in solution, thus not associating with protein present.

However, the degree of polarized light measured in buffer containing BODIPY-FL-GTP and Apo-Cdc42 protein (3) is relatively high, indicating that the protein has bound

BODIPY-FL-GTP. There is a relatively low level of polarized light measured in buffer with BODIPY-FL-GTP and caged Cdc42-4-HT fusion (5), indicating that the Cdc42-4-

HT fusion does not bind BODIPY-FL-GTP. Uncaging Cdc42 through photolysis of

Cdc42-4-HT restores its ability to bind BODIPY-FL-GTP as evidenced by the high fluorescent polarization signal measured with it in buffer with BODIPY-FL-GTP (6).

113

We postulate that the HaloTag protein forces the photocage to thread straight through the nucleotide binding site, preventing the binding of GTP. In contrast, small molecule modifications alone are more flexible, and could potentially adjust to accommodate GTP binding (Figure 6-3A).

Figure 6-4. Fluorescence polarization assay. The degree of polarization of light was measured in buffer containing 1) BODIPY-FL-GTP alone, or with BODIPY-FL-GTP and 2) HaloTag protein, 3) Apo-Cdc42 protein, 4) Apo-Cdc42 protein with 5 mM GTPS competitor, 5) caged Cdc42-4-HT fusion, and 6) uncaged Cdc42 from photolysis of Cdc42-4-HT. Apo-Cdc42 and uncaged Cdc42 bind BODIPY-FL-GTP as evidenced by the high fluorescent polarization signal, while HaloTag, apo-Cdc42 with a 50-fold excess of GTPS competitor, and Cdc42-4-HT do not.

Irradiation with long-wave UV light (350 nm) readily cleaves the photolabile linkage in Cdc42-4-HT, unmasking cysteine 18 and restoring the activated Cdc42 fast- cycling mutant protein (Figure 6-3). Notably, the nitroso-ketone photoproduct (Figure 6- 114

1C) remains attached to HaloTag and does not reattach to Cdc42, allowing them to migrate as separate species when analyzed by SDS-PAGE (Figure 6-3b, lane 4; Figure 6-

5). The uncaged Cdc42 behaves identically to unmodified Cdc42, showing the same nucleotide-dependent binding of PBD (Figure 6-3C).

Figure 6-5. Uncaging reaction. Irradiation with long-wave UV light (350 nm) uncages Cdc42-4-HT, releasing HaloTag and Cdc42. Substantial uncaging occurs in less than 10 seconds and is mostly complete between 2 and 5 minutes.

Cdc42 (1-188, F28L/C188S) contains five cysteine residues. Modification of cysteine 18 in the nucleotide binding site is quite specific, and is completely blocked by mutation of cysteine 18 to serine. Moreover, alkylation of cysteine 18 can be specifically blocked by pre-incubation with GDP (Figure 6-6). 115

Figure 6-6. Specificity of modification and HaloTag recruitment. Alkylation of Cdc42 with 4 allowed the recruitment of a single HaloTag to Cdc42. Alkylation with 4 and subsequent recruitment of HaloTag was completely prevented by pre-incubation with GDP or mutation of cysteine 18 to serine: a) alkylation of Cdc42 with 4 followed by incubation with HaloTag results in the formation of the fusion product Cdc42-4-HT only when the protein is nucleotide-free and contains cysteine 18; b) nucleotide-free Cdc42 is modified only once by 4; c) incubation with GDP prevents alkylation with 4; d) mutation of cysteine 18 to serine prevents alkylation with 4. 116

Since most Rho GTPases contain a cysteine at this position in the nucleotide binding site, we anticipated that our strategy would be generalizable, allowing the construction of photoactivatable versions of other family members. Consistent with this assertion, we have found that the Rho GTPase Rac1 can be photocaged in the same manner as Cdc42, and that it also requires our leveraged recruitment method for successful photocaging (Figure 6-7).

Figure 6-7. Photocaging of Rac1. An effector binding assay with PBD shows that modification with 4 alone does not prevent the ability of Rac1 to bind GTP and become active, but subsequent modification with the HaloTag protein does. Irradiation at 350 nm uncages Rac1-4-HT and restores the ability of Rac1 to bind effector in the presence of GTP. In a GST-PBD pulldown assay, only Rac1-4-HT fails to bind the PBD effector in the presence of the non-hydrolyzable GTP analog GTPS. UB = unbound protein; B = bound protein

117

In order to maximize protein expression levels and simplify purification, these studies were performed with bacterially-expressed GTPases that lack their C-terminal prenylation site. For future cell-based studies using microinjected protein, prenylated or prenylation-competent protein would be required for proper intracellular localization.

The leveraged recruitment strategy described here greatly facilitates the preparation of photoactivatable proteins by allowing their separation from uncaged protein. While HaloTag was used in this work, in principle other protein labeling strategies could be employed. We anticipate that the larger steric footprint afforded by this leveraged recruitment strategy will be of general utility for improving the photocontrol of a variety of protein interactions that, like the Rho GTPases, are recalcitrant to inhibition by small molecule modifications alone.

118

CHAPTER VII: Efforts to introduce Caged Rho GTPases into cells

Introduction

CHAPTER VI describes the ability to photocage the activity of the Rho GTPases

Cdc42 and Rac1 using an ortho-nitrobenzyl photocage derivative functionalized with a ligand to allow for recruitment of the haloalkane dehalogenase HaloTag. This two-part modification system was sufficient to block GTP binding, and thus activation of these proteins in vitro. However, in order to use this tool to investigate spatial and temporal activation requirements of these proteins in vivo, the caged protein must be introduced into cells.

The transfection of DNA into cells to allow for the expression or overexpression of proteins of interest has been widely used to study biological processes. Techniques to deliver proteins into cells are less well established. Proteins have been introduced into cells in several manners including through microinjection, the use of protein transfection reagents, and by fusion to cell-penetrating peptides.

Microinjection, a technique where a glass micropipette is used to insert substances at a microscopic level into living cells, is a powerful technique and offers many advantages, including the ability to introduce a wide variety of molecular species into either the nucleus or cytoplasm. It is not without its drawbacks however, as it can be difficult technically and it is physically invasive. Also, because microinjection is performed on a cell-by-cell basis, it does not allow for large-scale biochemical assays.

Several protein transfection reagents are commercially available that can be used to introduce recombinant protein into a cell. One of these, Chariot (ActivMotif), can be 119

used to deliver peptides, antibodies and proteins into cells. Chariot is a peptide that forms a non-covalent complex with the peptide/protein species to be introduced, and facilitates entry into the cell via what has been suggested to be a non-endosomal based mechanism. Upon cell entry, the Chariot peptide disassociates and a nuclear localization sequence localizes it to the nucleus, where it is degraded (Morris et al., 2001). Another reagent, BioPORTER (Genlantis), is a mixture of cationic lipids that form a non-covalent complex with the protein to be introduced into the cell. Once applied to cells, the complexes attach to the negatively charged cell surface and either 1) the BioPORTER lipids fuse with the plasma membrane, releasing the protein into the cell or 2) the complexes are internalized via endocytosis and the BioPORTER fuses with the endosomal membranes, translocating the protein into the cell in that manner (Zelphati et al., 2001).

Another method to introduce proteins into a cell is to genetically fuse a cell- penetrating peptide (CPP), a short and typically cationic peptide, to the protein of interest to allow for its cellular uptake. Some examples of CPPs are a peptide ( a.a.’s 48-60) from the HIV-1 protein Tat (Vives et al., 1997), the peptide penetratin ( a.a.’s 43-58) from the protein Antennapedia (Derossi et al., 1994), and a peptide from the herpes virus protein

VP22 (Kueltzo et al., 2000), as well as others. Many of these CPPs possess an abundance of arginine residues that seem to play a role in the efficiency of cellular uptake (Kosuge et al., 2008; Thoren et al., 2003). In fact, oligoarginines themselves have proved to be efficient CPPs (Futaki et al., 2001). 120

There is considerable debate regarding the mechanism of cellular entry both by transfection and the use of CPPs. Two of the proposed mechanisms of cellular entry are direct energy-independent penetration and endosomal uptake. In direct energy- independent penetration, whereby the protein tagged with the CPP translocates directly through the lipid bilayers of the plasma membrane, the introduced protein would be localized diffusely throughout the cell (Derossi et al., 1994; Takeuchi et al., 2006). If uptake was endosomal, whereby a portion of the plasmid membrane envelops the CPP tagged proteins and buds off into the cytoplasm, the resulting localization would be to endosomal vesicles, from which the protein would need to be released (El-Andaloussi et al., 2006; Wadia et al., 2004). It has been suggested that pre-incubation of a CPP-tagged protein with the hydrophobic counter-anion 4-(1-prenyl)-butyric acid (pyrenebutyrate) can help facilitate cellular uptake through non-endosomal pathways (Perret et al., 2005;

Sakai and Matile, 2003; Wadia et al., 2004). Described here are attempts to employ several of the methods described above to introduce caged FC-Cdc42-HT into cells via a non-endosomal based pathway.

Material and Methods

Cell Culture

HeLa cells were grown in a CO2 incubator at 37˚C with 5% CO2 and were cultured in Dulbecco's Modified Eagle Medium (DMEM) mixture (GIBCO) supplemented with 10% fetal bovine serum and 100 U/mL penicillin/streptomycin.

Chinese hamster ovary-K1 (CHO-K1) cells were grown in a CO2 incubator at 37˚C with 121

5% CO2 and were cultured in F-12K Nutrient mixture (GIBCO) supplemented with 10% fetal bovine serum and 100 U/mL penicillin/streptomycin.

Plasmid construction

The modified haloalkane dehalogenase HaloTag (HT-Promega) gene was PCR- amplified and cloned into the BamH1 and Not1 sites of pET28a for bacterial expression.

A polyarginine tag consisting of 9 arginines was introduced at C-terminus using the primer termed 9R C-terminus: gaacgcggccgcttaacggcgacggcgacggcgacggcgacggccggccagcccggggag. HT protein tagged on the N-terminus with a polyarginine tag was not stable.

The O6-methylguanine-DNA methyltransferase (SNAP) gene was PCR-amplified and cloned into the Nde1 and EcoRI sites of pET28a for bacterial expression.

Polyarginine tags consisting of 9 arginines were introduced at the N-and C-termini using the primers termed 9R N-terminus: gaatactagtcatatgcgtcgccgtcgccgtcgccgtcgccgtgacaaagattgcgaaatgaaac and 9R C-terminus: gaacgaattcttaacggcgacggcgacggcgacggcgacgggatcctggcgcgcctatacc.

Protein preparation

Protein was prepared as previously described in CHAPTER VI.

HT and SNAP labeling 122

Recombinant HT, polyarginine tagged HT, or polyarginine tagged SNAP proteins were incubated with one molar equivalent of tetramethylrhodamine (TMR) conjugated to the HaloTag reactive ligand (Promega) or DY-549 conjugated to the SNAP reactive ligand (New England BioLabs) for 1 hour at 22˚C in the dark with shaking at 300 rpm

(Eppendorf Thermomixer). Unreacted ligand was removed by passage over a Zeba column (Pierce) equilibrated with sterile Hanks Buffered Saline solution (HBSS).

Western blotting

Following treatment with either Chariot or BioPORTER, cells were washed with

PBS, trypsinized and pelleted at 1000 rpm speed in an A-4-62 rotor in an Eppendorf model 5810R centrifuge. Cells were lysed in 300 L Mg2+ Lysis Buffer (25 mM HEPES pH 7.5, 150 mM NaCl, 1% Igepal CA-630, 0.25% Na deoxycholate, 10% glycerol, 25 mM NaF, 10 mM MgCl2 , 1 mM EDTA, 1 mM Na orthovanadate, 10 g/mL leupeptin, and 10 g/mL aprotinin) at 4˚C with spinning in the dark. Lysates were pelleted at 10K rpm in an F45-24-11 rotor in an Eppendorf centrifuge model 5415D. Total protein concentration was determined using a bicinchoninic acid (BCA) protein assay. Volumes of supernatant totaling 30 g total cellular protein were diluted with SDS-PAGE loading buffer containing 10 mM DTT, separated on a 12% polyacrylamide gel, and transferred to nitrocellulose membrane. A lane containing a load control equal to the total amount of protein loaded into cells was also included. The membranes were blocked with blocking buffer (5% nonfat dry milk in PBS containing 0.2% Tween 20) overnight at 4˚C with shaking. Mouse anti-Cdc42 (Santa Cruz) or rabbit anti-HaloTag antibody (Promega) in 123

blocking buffer was added at 1:200 and 1:1,000, respectively, at room temperature for 3 hours. The membranes were washed 3 times for ten minutes in PBS and then goat anti- mouse or goat anti-rabbit horseradish peroxidase conjugated antibody (Promega) in blocking buffer was added at 1:3,000 and incubated for 30 minutes at room temperature.

The membranes were washed 3 times for 5 minutes in PBS. Horseradish peroxidase- bound proteins were detected by chemiluminescence using SuperSignal West Dura chemiluminescent substrate (Thermo Scientific Pierce). Membranes were imaged using a

Fujifilm LAS-3000 CCD camera.

Fluorescent microscopy

Cells were examined on a Zeiss Axiovert 200 fluorescence microscope with a

40X objective using a Hamamatsu ORCA-ER digital camera and excitation from an X-

Cite 120 light source through a Zeiss filter cube.

Transfection with Chariot

For analysis by Western blot: HeLa cells were seeded at 60-75% confluency in a

100 mm plate in media supplemented with 10% fetal bovine serum without antibiotic.

Chariot was resuspended as directed by the manufacturer and 50 L was added to 150 L sterile water per treatment type. 40 g of the caged Cdc42 F28L-HT protein was added to a final volume of 200 L sterile phosphate buffered saline (PBS). The diluted Chariot and protein were mixed gently and the mixture was incubated at room temperature for 30 minutes. Cells were washed with PBS, the protein: Chariot mixture was added, followed 124

by 4.6 mL of serum free DMEM. The cells were incubated at 37˚C for 1 hour and overlaid with 5 mL DMEM media supplemented with 10% fetal bovine serum and 100

U/mL penicillin/streptomycin and incubated at 37˚C for 2 additional hours. Following this incubation, the media was removed and cells were washed with PBS three times.

Transfection with BioPORTER

For analysis by Western blot: HeLa cells were seeded at 60-75% confluency in a

100 mm plate in media supplemented with 10% fetal bovine serum without antibiotic. 40

g of the caged Cdc42 F28L-HT protein was added to a final volume of 300 L sterile phosphate buffered saline (PBS). 100 L of the diluted protein was added to 35 L

BioPORTER reagent in triplicate and after 5 minutes, the mixture was pipetted up and down and then vortexed gently. After a 30-minute incubation, 400 L of serum free media was added to each tube. Cells were washed with PBS and then overlaid with 3 mL serum free media followed by the 1.5 mL BioPORTER: protein mixture, and incubated at

37˚C for 3 hours.

For analysis by microscopy: HeLa cells were seeded at 60-75% confluency in a

96-well plate in media supplemented with 10% fetal bovine serum without antibiotic.

Two g of TMR labeled HT protein was added to a final volume of 40 L sterile HBSS, which was added to 10L of BioPORTER reagent. The mixture was pipetted up and down, incubated for 5 minutes, and then vortexed gently. After a 30-minute incubation,

500 L of serum free media was added. Cells were washed with HBSS and then overlaid with 50 L serum free media followed by the 50 L BioPORTER: protein mixture, and 125

incubated at 37˚C for 3 hours. The media was removed and cells were washed 3 times with Opti-MEM (GIBCO) before imaging.

CPP and Pyrenebutyrate

For HT, HeLa cells were seeded at 60-75% confluency in 96-well plates. For

SNAP, CHO-K1 cells were seeded at 60-75% confluency in 96-well plates. Cells were washed with HBSS and overlaid with either HBSS or pyrenebutyrate and incubated for 5 minutes at 37˚C. TMR labeled C-terminal 9R-HaloTag was added so that final protein concentration was 50 M, and pyrenebutyrate where present, was at a final concentration of 150 M. TMR labeled N- and C-terminal 9R-SNAP was added so that final protein concentration was at 5 M and pyrenebutyrate was at a final concentration of 50 M.

Following incubation at 37˚C for 15 minutes, cells were washed 5 times with HBSS, overlaid with Opti-MEM and imaged.

Results

The protein transfection reagents Chariot and BioPORTER were used to introduce the caged FC-Cdc42-HT into HeLa cells and cellular entry was assessed by Western blot analysis for either the Cdc42 or HT protein (Figure 7-1). No protein was detected using the reagent Chariot (Figure 7-1A and B, lanes 2). Using the BioPORTER reagent, a faint band could be detected when anti-HT antibody was used to probe for protein (Figure 7-

1A, lane 3), however no signal was detected when the anti-Cdc42 antibody was used

(Figure 7-1B, lane 3). Because the BioPORTER reagent showed promise in terms of 126

being able, by Western blot analysis, to introduce the caged FC-Cdc42-HT protein into

HeLa cells, we decided to image cells transfected using BioPORTER to assess the intracellular distribution when using this method.

Figure 7-1. Westerns following Chariot and BioPORTER treatment. The protein transfection reagents Chariot and BioPORTER were used to introduce the caged FC- Cdc42-HT into HeLa cells and cellular entry was assessed by Western blot analysis for either the Cdc42 or HT protein. No protein was detected using the reagent Chariot (A and B, lanes 2). Using the BioPORTER reagent, a faint band could be detected when anti-HT antibody was used to probe for protein (A, lane 3), however no signal was detected when the anti-Cdc42 antibody was used. Lanes 1 in A and B are load controls.

To allow us to image cellular entry and localization using the BioPORTER reagent, we first had to label the protein to be introduced with a fluorophore. As the caged FC-Cdc42-HT protein is itself not fluorescent, we decided to instead use HT 127

protein that had been labeled using tetramethylrhodamine (TMR) conjugated to the HT reactive ligand. Using this method, we were able to observe very faint fluorescence in cells treated with both labeled protein and BioPORTER, but none when cells were treated with either on its own. Unfortunately, the fluorescence was so faint that we could not capture images of it. Furthermore, due to this very weak signal, we were unable to determine the localization pattern of the labeled protein.

Based on the difficulties we experienced with proteins transfection reagents, we attempted to use CPP tagging and pyrenebutyrate methodologies. We began first by fusing an oligoarginine tag consisting of 9 arginines to the C-terminus of HT. We then labeled this tagged HT with HT-TMR dye and treated HeLa cells both in the absence and presence of pyrenebutyrate, and imaged the cells to determine the relative localization patterns (Figure 7-2). In cells treated with C-terminal 9R-HT, both in the presence and absence of pyrenebutyrate, fluorescence signal appeared punctate, rather than diffuse.

This suggests that the method of protein entry in both treatment types relied on an endosomal uptake pathway. However, HT protein itself is quite negatively charged, and this charge could have an effect on cellular uptake or localization under these conditions. 128

Figure 7-2. Loading of polyarginine-tagged HT into HeLa Cells. HeLa cells treated with 50 M 9R-HT labeled with TMR-HT ligand with and without 150 M PBA were imaged by fluorescence microscopy (left) and DIC (right). For both treatment types, cellular localization appeared both diffuse and punctate (left upper and lower panels).

The protein termed SNAP, derived from the human DNA repair protein O6- methylguanine-DNA methyltransferase, reacts specifically with O6-benzylguanine derivatives. Taking into account the negative charge on HT, we decided to tag the more neutral SNAP on the N-and C-termini with 9R polyarginine tags. We then reacted the tagged SNAP proteins with Dy-549 conjugated to a SNAP reactive ligand and introduced the labeled and tagged protein into CHO-K1 cells in the presence of pyrenebutyrate (Fig 129

7-3). Cells treated with both N-and C-terminally oligoarginine tagged SNAP protein in the presence of pyrenebutyrate exhibit both diffuse and punctate protein localization patterns. In some cases, the SNAP protein was not localized diffusely throughout the cell, but appeared to accumulate in what looked to be nucleoli.

Figure 7-3. Loading of polyarginine tagged SNAP protein into CHO Cells. CHO cells treated with 50 M PBA and 5 M 9R-SNAP labeled with TMR-SNAP ligand were imaged by fluorescence microscopy (left) and DIC (right). For both C-terminally and N- terminally 9R-tagged SNAP, cellular localization was both diffuse and punctate, often localized to what appear to be nucleoli (left upper and lower panels).

130

Discussion

We described in CHAPTER VI the generation of a two-step modification system whereby a Rho GTPase’s ability to bind GTP and become active could be controlled by light in vitro. To show that this approach could be applied in a cellular context to allow for the study of Rho GTPase activation requirements in vivo, we employed several established methods of introducing recombinant protein into cells. Unfortunately, of the methods we tried, we saw either unsuccessful (Chariot) or inefficient (BioPORTER) results (Figure 7-1), or punctate or localized distribution of proteins (CPPs and pyrenebutyrate), rather than diffuse localization (Figure 7-2; Figure 7-3). Potentially one of the main challenges associated with the methods we used to introduce protein into cells related to the overall amount of protein each method used. Both transfection reagents recommended the introduction of a set, low amount of protein for best results.

Potentially this amount was too low to allow for protein visualization in the manner we were attempting. In examples where CPPs and pyrenebutyrate were successful, multiple applications and/or very high protein concentrations were necessary (Inomata et al.,

2009). Due to technical constraints inherent to the generation of our caged FC-Cdc42-

HT, application of protein concentrations similar to those reported is not possible. There is also the potential that the issues we observed are directly related to the specific protein species being used.

While microinjection could overcome some of the limitations presented by the other methods we tried, there are additional considerations in terms of using this method to introduce Rho GTPases into cells. This family of proteins is postranslationally 131

modified on their C-terminus and this modification is essential for their proper localization and function. Because the C-terminal posttranslational modification of Rho

GTPase proteins occurs on a cysteine and our photocage is also cysteine reactive, when working with Cdc42 and Rac1 in vitro, we used a C-terminally truncated mutant to simplify experimental procedures (CHAPTER VI).

One possible method to allow for caged GTPase protein to be properly localized once introduced into a cell could involve the development of a multi-step modification system that would allow for differential modification of the C-terminal cysteine and cysteine 18 in the nucleotide binding pocket. One way to do this would be to block cysteine 18 modification by preloading the GTPases with nucleotide and reacting the C- terminal cysteine with a reversible cysteine modifier, for example [2-

(trimethylammonium) ethyl] methanethiosulfonate (MTSET). Cysteine 18 modification could be performed as usual after removing nucleotide with EDTA treatment and following HT reaction and final chromatographic purification steps, simple DTT treatment would remove the MTSET modification and free the C-terminus for proper posttranslational modification (Figure 7-4). An alternate route would be to use a mammalian expression system, such as a Baculovirus based expression system, to ensure proper posttranslational modification prior to cysteine 18 modification. Microinjection of single cells could then be used to introduce the caged GTPase with a free, or properly modified, C-terminus to address questions about the spatial and temporal activation requirements of Rho GTPases in living cells. 132

Figure 7-4. Differential cysteine modification scheme. Preloading full-length GTPase with GDP, subsequent modification of the C-terminal cysteine with MTSET, and removal of GDP via EDTA treatment results in GTPase with a protected C-terminal cysteine and an exposed Cys18. Modification of Cys18 with 4 (the yellow star) and reaction with HT protein, FPLC purification and DTT treatment results in a GTPase modified on Cys18 with 4 and HT, and a free C-terminal cysteine which can be properly posttranslationally modified by cellular machinery once introduced into a cell.

While the work reported here was in progress, Klaus Hahn and colleagues published a report describing studies using a genetically encoded photoactivatable form of Rac1 (Wu et al., 2009b). Specifically, they genetically fused a flavin binding light oxygen voltage 2 (LOV2) domain of the protein phototropin to the N-terminus of a CA mutant of Rac1 through a helical extension made up of phototropin’s C-terminus J. In the dark, the LOV2 and J domains interact, sterically blocking CA-Rac1 from interacting with effector proteins. Upon irradiation with light, the J helix unwinds and

CA-Rac1 is once again able to interact with effector proteins and cause its downstream 133

effects. Furthermore, upon removal of the light source, LOV2 and J once again interact and inhibit CA-Rac1’s activity. This approach is not without its disadvantages. In the dark state, CA-Rac1 is not completely inhibited, but still signals to a small degree. Also, the fusion itself may effect Rac1 localization and signaling, as well as which subset of effectors are both blocked and able to interact with CA-Rac1 upon caging and uncaging.

However overall, the fact that the fusion is genetically encoded and that activation is reversible make it an incredibly powerful tool for the study of Rac1 activation requirements. This approach also has the potential to be applied to the study of other Rho and small GTPase family members. If the two-part selective cysteine modification scheme suggested above to generate caged Rho GTPases that are, or could be properly modified is successful, the resulting tools would help to further elucidate the roles played by the different Rho GTPase family members at different cellular stages and locations.

134

CHAPTER VIII: Using Photocaging to inhibit Rho GTPases

Introduction

The Rho GTPase field has produced a wealth of structural data and information regarding the role of specific residues in the function and regulation of the Rho GTPases.

Described in CHAPTER VI is the development of a two-part modification system where modification with a photolabile small molecule that can recruit the much larger HaloTag protein blocks nucleotide binding in active mutants of Rho GTPases and allows for their subsequent light dependent activation. Using the available crystallography data and this two-part modification system, we attempted to generate GTPases that could be inactivated in a spatial and temporal manner by covalent modification of dominant negative mutants. This would allow control over where and when the Rho GTPase is turned off.

DN Rho GTPases function by sequestering GEF away from their endogenous WT counterparts and thus preventing their activation. Based on this, we hypothesized that alkylating DN Rho GTPases with a photolabile modification on residues that would block

GEF binding could allow for the light induced inactivation of endogenous protein.

Specifically, we hypothesized that alkylating DN mutant T19N of RhoA on mutant cysteine residues introduced at Q63 or N41 (RhoA numbering), and recruitment of the much larger protein HT, would allow us to prevent its inhibition of endogenous RhoA.

Irradiation with light would remove the photocage and protein modifications and allow the protein to exert its dominant negative effects at the time and place of uncaging. The location of these mutated cysteines is such that their modification is likely to prevent 135

interaction of the DN mutant with GEF proteins. Based on structural data (Snyder et al.,

2002), it appears that amino acids Q63 and N41 do not directly interact with GEF proteins, such that mutating them to cysteines would not interfere with GEF binding

(Figure 8-1). However, based on their proximity to the GEF binding site, modification of the introduced cysteine residues with the photocage and HT protein could prevent GEF binding in this region. Further, it has been shown that Clostridium botulinum exotoxin

C3 inhibits GEF interaction with RhoA by ADP-ribosylating N41 (Aktories et al., 1992), suggesting that modification at this location with a photocage could also prevent GEF binding.

136

Figure 8-1. GEF: GTPase binding interface. Rho GTPase Cdc42 (white) and GEF intersectin (blue). Residues N41 and Q63 (RhoA numbering) of Cdc42 are highlighted in magenta and purple, respectively, and are located at the interface between Cdc42 and intersectin, but do not appear to be involved directly in the GTPase:GEF interaction (PBD 1KI1).

As an alternative, it has been shown that the minimum Cdc42-binding domain of the Cdc42 specific tyrosine kinase ACK1 (ACK-42) can bind to GTP-Cdc42 and inhibit it by blocking its interaction with effector proteins (Nur et al., 1999). ACK-42 functions by wrapping around and masking Cdc42’s effector binding interface. In addition, this binding domain possesses a single cysteine residue. We hypothesized that caging this 137

cysteine with our photolabile small molecule that can recruit HT could prevent ACK-42 from interacting with Cdc42 to inhibit it until irradiation with light, upon which uncaging would occur and the binding domain could once again function as a Cdc42 inhibitor at the location and time of uncaging (Figure 8-2).

Figure 8-2. Alternate scheme to use light to inhibit a GTPase. The minimum Cdc42- binding domain of the Cdc42 specific tyrosine kinase ACK1 (ACK42) binds to GTP- Cdc42, wrapping around it and blocking its interaction with effector proteins. Modification of ACK42 on its single cysteine with 4 and HT could potentially block its interaction with Cdc42 until uncaging with light, allowing for site specific inactivation of Cdc42.

Materials and Methods

Plasmid constructs

Mutant proteins were created using the Quickchange site-directed mutagenesis kit

(Stratagene). C-terminally truncated RhoA C190S was made by PCR of almost the full- length gene with the primers gaatggatccatggctgccatccggaagaaactggtg and 138

gaacgaattctcaagacccagattttttcttcccacgtctagc. The RhoA binding domain of the GEF Pdz was PCR amplified using the primers gaatggatccgatgcccaaaattggcagcatac and gaacgaattcttaggcattccgcacggcctcttc. These PCR products were then digested with BamHI and EcoRI and ligated into the corresponding sites in pGEX6P-1.

For RhoA: The T19N point mutation was introduced with the primers ggtgatggagcctgtggaaagaattgcttgctcatagtcttcagc and tcggaagactatgagcaagcaattctttccacaggctccatcacc. The C20S point mutation was introduced using the primers ggagcctgtggaaagaacagcttgctcatagtcttcagc and gctgaagactatgagcaagctgttctttccacaggctcc . The N41C mutation was introduced using the primers tatgtgcccacagtgtttgagtgctatgtggcagatatcgaggtgg and ccacctcgatatctgccacatagcactcaaacactgtgggcacatac. The Q63C mutation was introduced using primers gctttgtgggacacagctgggtgtgaagattatgatcgcctgagg and cctcaggcgatcataatcttcacacccagctgtgtcccacaaagc.

The Cdc42 binding domain of Ack1 (ACK42) comprised of amino acids 504-545 was PCR amplified using the primers gaatgaattcggcctgtcggcccaggacatc and gaacggatccttagttgcccagatacagctcgtcaatcctgtccgggaag. This PCR product was then digested with EcoR1 and BamHI and ligated into the corresponding sites in pMAL-c2X.

HaloTag constructs were made previously, as described in CHAPTER VI

(Harwood and Miller, 2009).

Protein expression 139

RhoA mutants and Pdz were expressed as GST fusion proteins from the vector pGEX6P-1 in the BL21(DE3) E. coli strain as previously described (Bhunia and Miller,

2007). ACK42 was expressed as a MBP fusion protein from the vector pMAL-c2X in the BL21(DE3) E. coli strain just like RhoA mutants. HaloTag (Promega) was expressed as a His6 fusion protein from the vector pET28 in the BL21(DE3) pLysS E. coli strain.

Cells were grown at 37˚C until the OD600 reached 0.8, then induced with 0.1 mM IPTG and incubated at 20˚C overnight. Cells were pelleted at 5000 rpm in a Sorvall 2C3C Plus centrifuge (H600A rotor) at 4˚C for 10 minutes, then flash frozen in liquid nitrogen and used immediately or stored at -80˚C, with the exception of Pdz expressing cells, which were not flash frozen and were used immediately or stored at -20˚C.

RhoA, Pdz and ACK42 purification

E. coli cell pellets from one liter of culture were thawed on ice, resuspended in 25 ml Lysis Buffer (50 mM Tris or HEPES pH 7.4, 500 mM NaCl, and 0.5% Tween-20) containing 1 mM PMSF and, with the exception of Pdz pellets, disrupted by sonification

(Branson Sonifier). Dithiothreitol (DTT) was added at 10 mM and the resulting cell lysate was clarified by centrifugation at 35K rpm in a Beckman 50.2Ti rotor for 1 hour.

The supernatant was batch-bound to immobilized glutathione or amylose (for ACK42) resin for 1h at 4˚C, and then the beads were washed with Lysis Buffer containing 10 mM

DTT followed by Wash Buffer (50 mM Tris pH 8.1, 250 mM NaCl and 10 mM DTT).

For RhoA, the beads were then incubated with Nucleotide Removal Buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM TCEP, and 20 mM EDTA). Finally, both RhoA and Pdz 140

beads were washed with Storage Buffer (50 mM HEPES pH 7.4, 50 mM NaCl, 1 mM

TCEP, and 1 mM EDTA). Twenty units of PreScission Protease (GE Healthcare) were added and incubation continued overnight at 4˚C to cleave the GST fusion and elute the untagged GTPase.

For ACK42, 10 mM maltose in Wash Buffer was used to elute the MBP-tagged

ACK42, which was then dialyzed against 50 mM HEPES pH 7.4, 50 mM NaCl, 1 mM

TCEP, and 1 mM EDTA.

HaloTag purification

E. coli cell pellets were thawed on ice, resuspended in 50 mM NaPhos pH 7.8,

300 mM imidazole and 5 mM BME, then dialyzed against 50 mM Tris pH 7.4, 150 mM

NaCl, 1 mM TCEP, and 1 mM EDTA.

Small molecule modification of RhoA and ACK42

Proteins were incubated with 1-2 mM TCEP to reduce surface exposed cysteines.

Free TCEP was then removed by passage over two Zeba columns equilibrated with 50 141

mM HEPES pH 7.4, 50 mM NaCl and 5 mM EDTA. For RhoA, 20 mM MgCl2 and 300

M GDP were added and the mixture was incubates at 22˚C with shaking at 300 rpm for

30 minutes. Ten equivalents of the small molecule photocage 4, prepared as outlined in

CHAPTER VI, were then added and allowed to react in the dark at 22°C for 3h with shaking at 300 rpm (Eppendorf Thermomixer). Unreacted alkylating agent was removed by passage over a Zeba column equilibrated with 50 mM Tris pH 7.4, 150 mM NaCl,

1mM TCEP, and 1 mM EDTA.

Recruitment of HaloTag protein to RhoA or ACK42 modified with 4

Protein that had been alkylated with 4 as descried above was incubated with 0.3-

0.5 eq of purified HaloTag protein at 22°C in the dark for 1h with shaking at 300 rpm.

The HaloTag-modified protein was purified by FPLC (AKTA) on a MonoQ ion- exchange column (GE Healthcare). The purity of the fractions was assessed by SDS-

PAGE with Coomassie staining. Purified HaloTag-modified GTPase was used directly for GEF exchange assays.

GEF exchange assays

Measurements were taken in a Spex FluoroMax-3 fluorimeter every second with a

0.1 second integration time for a total of 400 seconds. Excitation was at 488 and emission was at 508.5, slits were set at 5. RhoA C190S protein was preloaded with saturating concentrations of GDP in the presence of 2 mM MgCl2. Excess GDP was removed by passage over 2 Zeba column equilibrated with 50 mM Tris pH 7.4, 150 mM 142

NaCl, 1mM TCEP, and 1 mM MgCl2. The GEF exchange assay was initiated by the rapid addition of 1 M GDP-RhoA C190S into GEF assay buffer (20 mM Tris pH 7.5, 50 mM NaCl, 5 mM MgCl2, 1 mM TCEP, 5% glycerol, and 10 mg/mL BSA) containing A)

500 nM Bodipy-FL GTP, B) 500 nM Bodipy-FL GTP and 100 nM Pdz, C) 500 nM

Bodipy-FL GTP, 100 nM Pdz, and 10 M RhoA T19N/C20S/Q63C/C190S, or D) 500 nM Bodipy-FL GTP, 100 nM Pdz, and 10 M HT caged RhoA

T19N/C20S/Q63C/C190S.

Results and Discussion

We hypothesized that we could photocage the DN activity of DN Rho GTPases by using a two-part modification system whereby we alkylate them with a small molecule that can then recruit the much larger HT protein. This alkylation would occur on mutated cysteines at locations that appear to be at the interface between the GTPase RhoA and

GEF protein, but which didn’t appear to make any critical contacts. To do this, we introduced cysteine mutations into the DN T19N background of RhoA at position N41 and Q63. Unfortunately, the T19N/N41C mutant did not appear to be stable upon expression and purification. Because T19N/Q63C behaved more consistently, we proceeded to also mutate the cysteine in the nucleotide binding pocket to a serine, C20S.

We also mutated the C-terminal CAAX box cysteine to a serine, C190S, and removed the final three amino acids.

Because we had introduced many mutations into the DN RhoA protein, we wanted first to determine if it was still able to behave in a DN manner. To assess this, we 143

performed GEF nucleotide exchange assays. In solution, the fluorescence of Bodipy-FL

GTP is quenched by the close proximity of the GTP portion of the molecule. However, upon GEF loading of Bodipy-FL GTP into the GTPase, quenching is relieved and an increase in fluorescence intensity is observed. Thus, fluorescence intensity can be used as a measure of nucleotide loading of the GTPase. We took advantage of these properties of Bodipy-FL GTP to determine if the mutated DN RhoA, T19N/C20S/Q63C/C190S, was still able to bind to GEF protein and sequester it away from WT RhoA, preventing the GEF from loading WT RhoA with nucleotide. To do this, we preincubated the RhoA

GEF Pdz with the mutated DN RhoA and Bodipy-FL GTP, and rapidly added GDP loaded WT-RhoA and measured fluorescence intensity. When GDP WT-RhoA is added to buffer containing Bodipy-FL GTP in the absence of GEF protein, very slow nucleotide exchange occurs, resulting in a very small increase in fluorescence (Fig 8-3). However, when GDP WT-RhoA is added to buffer containing both Bodipy-FL GTP and the GEF

Pdz, a rapid increase in florescence intensity occurs as Pdz exchanges the fluorescent nucleotide for GDP in the nucleotide binding pocket of RhoA (Fig 8-3). When GDP

WT-RhoA is added to buffer containing Bodipy-FL GTP and mutated DN RhoA preincubated with Pdz, very little increase in fluorescence intensity occurs (Fig 8-3), indicating that the mutated DN RhoA is indeed still behaving as a DN mutant and sequestering Pdz away from the WT RhoA.

144

Figure 8-3. Photocaging DN RhoA: GEF nucleotide exchange assays. Rapid addition of WT RhoA-GDP to Bodipy-FL GTP in buffer results in a minimal increase in fluorescence (red). Addition of RhoA-GDP to a mixture of Bodipy-FL GTP and the GEF Pdz results in an increase in fluorescence intensity as Pdz loads RhoA with the fluorescent GTP analog (green). The DN RhoA mutant is able to bind Pdz and inhibit its ability to load WT-RhoA with Bodipy-FL GTP, as evidenced by the minimal increase in fluorescence observed upon rapid addition of WT RhoA-GDP to a DN-RhoA/Pdz mixture (blue). Based on the partial increase in fluorescence intensity that occurs upon addition of WT RhoA-GDP to a mixture of Pdz and the DN mutant modified with 4 and HT (orange), the modification only partially blocks it inhibition of Pdz.

We then wanted to determine if modification of Q63C in DN RhoA with our photolabile small molecule and HT protein would block its ability to interact with Pdz, thus blocking its dominant negative activity in a way that could be reversed with light.

Using the same GEF nucleotide exchange assay as above, we preincubated

T19N/C20S/Q63C/C190S that had been alkylated with 4, reacted with the HaloTag protein, and purified by FPLC with Pdz and Bodipy-FL GTP. If the two-part modification system is able to block Pdz binding, upon rapid addition of GDP WT-RhoA, an increase in fluorescence intensity similar to that seen when adding GDP WT-RhoA to 145

Pdz alone should be observed. However, when we conducted this experiment, we only observed a partial recovery in fluorescence intensity (Fig. 8-3). This could be due to the initial small molecule modification occurring on not just Q63C, but another cysteine as well, since RhoA possesses other cysteines including C16 and C105. However, the observed molecular weight of the RhoA-HT fusion product suggest that only one HT molecule reacted with each RhoA protein. This does not mean, however, that the modification occurred on the same cysteine in all cases. Potentially, multiple photocage molecules could have reacted with different cysteines, but once the large HT protein reacts with one of the cages, it sterically blocks the reaction of another HT protein.

Unfortunately, efforts to analyze tryptic and chymotryptic digests of caged RhoA using mass spectrometry did not allow us to identify which specific cysteine or cysteines the cage was modifying, since we could not see the peptides containing the cysteines.

LC/ESI-MS of one caged mutant, RhoA T19N/Q63C/C190S, revealed that when the protein was loaded with excess GDP, it was only modified by 4 once (Fig 8-4). If multiple cysteines on the protein were surface exposed and could be modified, you would expect to see a +1, +2, +3 charge state etc. since 10 fold molar excess of cage was present during alkylation. However, it is difficult to interpret this data as C20, the cysteine corresponding to C18 in Cdc42, is present in this mutant where it was not in the mutant used for the GEF nucleotide exchange assays.

147

fragment termed ACK42 can specifically bind to GTP-Cdc42 and prevent it from interacting with and signaling through its other effectors (Figure 8-2) (Nur et al., 1999).

Additionally, this fragment of Ack1 possesses a single cysteine residue. Based on this, we predicted that if we modified that cysteine with 4, and recruited the HT protein, we could block the ability of ACK42 to inhibit Cdc42 until we irradiate with light at a specific time and location. To do this, we expressed ACK42 as a MBP fusion protein, modified it with 4, reacted it with HT and purified the fusion using FPLC (Figure 8-5).

Figure 8-5. SDS-Page of fractions collected during FPLC purification of Ack-4-HT. All fractions that contained the Ack-4-HT fusion (upper bands) also contained a contaminant from the Ack purification mixture (lower bands).

Unfortunately, MBP-ACK42 is itself very unstable and difficult to work with. It is highly prone to dimerization through disulfide bonds, and degrades readily upon purification (Figure 8-6). Efforts to generate purified fusion protein were also 148

complicated by issues with contamination and degradation (Figure 8-7). Potentially the spacing between ACK42 and the MBP tag contributed to the inherent instability of the fusion. Attempts to optimize linker spacing and potentially the specific tag used could rectify some of the problems ACK42 presents. One interesting approach would be to append a fluorescent protein onto ACK42, and equip HT with a quencher such that upon uncaging, the location of Cdc42 being actively inhibited by ACK42 could be visualized.

150

Figure 8-7. ACK42 and Ack-4-HT are prone to degradation. ACK42 (lane 1) is contaminated by degradation products, as is the Ack-4-HT reaction product mixture (lane 2).

Overall, while our efforts to generate a species that could allow for the light induced inhibition of a Rho GTPase were not ultimately successful, with some optimization, our approaches could eventually prove fruitful. A Rho GTPase that could be “turned off” with light would be a nice counterpart for GTPases that can be “turned on” when asking questions about the spatial and temporal activation requirements of this family of proteins.

151

CHAPTER IX: Photocontrol of Rho GTPases: Disucssion

Many methods to control the activation states of proteins have been developed. A powerful example of one of these methods is the use of photocaging. Photocaging is particularly powerful when applied to proteins whose activation at specific times and locations in the cell is essential for proper cellular functioning. One such example is the family of proteins called the Rho GTPases. CHAPTER VI describes the application of photocaging to the control of GTP binding in active mutants of Cdc42 and Rac1. In this manner, GTPases that could be turned on at specific times and locations with light were generated. It is likely that the same approach would be effective in photocaging other

Rho GTPase family members as well.

CHAPTER VII describes attempts to use several reported methods to introduce the caged GTPases into cells to verify the ability to uncage and activate them in vivo.

Unfortunately, these methods were either unsuccessful or inefficient. The use of these caged GTPases in cells is further complicated by the requirement of posttranslational modifications necessary for their proper functioning. Future work to develop methods to allow for the differential modification of the C-terminal cysteine and Cys18 in these

GTPases would facilitate the generation of photocaged GTPases with a free C-terminal cysteine that could be properly posttranslationally modified upon microinjection, for example, and allow for in vivo studies with these proteins.

Finally, CHAPTER VIII describes attempts to use photocaging to turn off a Rho

GTPase. To do this, DN Rho GTPases were modified at a location predicted to block

GEF binding until irradiation with light. Upon irradiation, the DN GTPases would bind 152

and sequester all GEF at the location of uncaging, thus preventing the activation of endogenous GTPase and allow questions about the GTPases activation requirement at that time and place to be addressed. Unfortunately, while photocaging was able to partially block GEF interaction, the block was not complete. Photocaging another residue at the GTPase:GEF interface may prove more successful. In addition, attempts to photocage an effector binding domain (ACK42) that can bind to and mask the effector biding interface of Cdc42, thus inhibiting Cdc42 activity, were also unsuccessful.

Methods to stabilize ACK42, perhaps optimization of linker length between ACK42 and its purification tag, may allow for its photocaging and use to probe questions regarding

Cdc42’s spatiotemporal activation requirements.

Overall, this thesis describes the development of tools that could allow for both the better visualization of the cellular environment as well as the better control of cellular functioning through the precise control of protein activity. Applying these tools, and others like them, to the study of cellular biology will allow for an improved understanding of the regulatory processes involved in the control of key events important to maintaining cellular homeostasis, as well as the events that go awry, ultimately leading to the development of disease states. The interface between chemistry and biology presents unique advantages in the development of such tools.

153

Bibliography

Adamson, P., Marshall, C.J., Hall, A., and Tilbrook, P.A. (1992). Post-translational modifications of p21rho proteins. J Biol Chem 267, 20033-20038.

Aktories, K., and Koch, G. (1997). Clostridium botulinum ADP-ribosyltransferase C3 (Weinheim, Germany, Chapman and Hall).

Aktories, K., Mohr, C., and Koch, G. (1992). Clostridium botulinum C3 ADP- ribosyltransferase. Curr Top Microbiol Immunol 175, 115-131.

Alexander, D.C., Costanzo, M.A., Guzzo, J., Cai, J., Charoensri, N., Diorio, C., and DuBow, M.S. (2000). Blazing towards the next millennium: luciferase fusions to identify genes responsive to environmental stress. Water Air Soil Pollution 123, 81-94.

Almond, B., Hawkins, E., Stecha, P., Garvin, A., Paguio, B., Butler, M., Beck, M., and Wood, K. (2003). Introducing Chroma-Luc technology. Promega Notes 85, 11-14.

Alter, S.C., and DeLuca, M. (1986). The sulfhydryls of firefly luciferase are not essential for activity. Biochemistry 25, 1599-1605.

Ando, Y., Niwa, K., Yamada, N., Enomoto, T., Irie, T., Kubota, H., Ohmiya, Y., and Akiyama, H. (2008). Firefly bioluminescence quantum yield and colour change by pH- sensitive green emission. Nature Photonics 2.

Arai, R., Nakagawa, H., Kitayama, A., Ueda, H., and Nagamune, T. (2002). Detection of protein-protein interaction by bioluminescence resonance energy transfer from firefly luciferase to red fluorescent protein. J Biosci Bioeng 94, 362-364.

Arslan, T., Mamaev, S.V., Mamaeva, N.V., and Hecht, S.M. (1997). J Am Chem Soc 119, 10877-10887.

Auld, D.S., Zhang, Y.Q., Southall, N.T., Rai, G., Landsman, M., MacLure, J., Langevin, D., Thomas, C.J., Austin, C.P., and Inglese, J. (2009). A basis for reduced chemical library inhibition of firefly luciferase obtained from directed evolution. J Med Chem 52, 1450-1458.

Babbitt, P.C., Kenyon, G.L., Martin, B.M., Charest, H., Slyvestre, M., Scholten, J.D., Chang, K.H., Liang, P.H., and Dunaway-Mariano, D. (1992). Ancestry of the 4- chlorobenzoate dehalogenase: analysis of amino acid sequence identities among families of acyl:adenyl ligases, enoyl-CoA hydratases/isomerases, and acyl-CoA thioesterases. Biochemistry 31, 5594-5604.

Banghart, M., Borges, K., Isacoff, E., Trauner, D., and Kramer, R.H. (2004). Light- activated ion channels for remote control of neuronal firing. Nat Neurosci 7, 1381-1386. 154

Barbacid, M. (1987). ras genes. Annu Rev Biochem 56, 779-827.

Barltrop, J.A., Plant, P.J., and Schofield, P. (1966). Photosensitive protective groups. Chem Commun, 822-823.

Basso, A.D., Kirschmeier, P., and Bishop, W.R. (2006). Lipid posttranslational modifications. Farnesyl transferase inhibitors. J Lipid Res 47, 15-31.

Belshaw, P.J. (1995). Rational design of orthogonal receptor-ligand combinations. Angew Chem Int Ed Engl 34, 2129-2132.

Benard, V., Bohl, B.P., and Bokoch, G.M. (1999). Characterization of rac and cdc42 activation in chemoattractant-stimulated human neutrophils using a novel assay for active GTPases. J Biol Chem 274, 13198-13204.

Benink, H.A., and Bement, W.M. (2005). Concentric zones of active RhoA and Cdc42 around single cell wounds. J Cell Biol 168, 429-439.

Bhunia, A.K., and Miller, S.C. (2007). Labeling tetracysteine-tagged proteins with a SplAsH of color: a modular approach to bis-arsenical fluorophores. Chembiochem 8, 1642-1645.

Bielek, H., Anselmo, A., and Dermardirossian, C. (2009). Morphological and proliferative abnormalities in renal mesangial cells lacking RhoGDI. Cell Signal 21, 1974-1983.

Biggley, W.H., Lloyd, J.E., and Seliger, H.H. (1967). The spectral distribution of firefly light. II. J Gen Physiol 50, 1681-1692.

Bishop, A.L., and Hall, A. (2000). Rho GTPases and their effector proteins. Biochem J 348 Pt 2, 241-255.

Boettner, B., and Van Aelst, L. (2002). The role of Rho GTPases in disease development. Gene 286, 155-174.

Branchini, B., Southworth, T., Khattak, N.K., Murtiashaw, M.H., and Fleet, S.E. (2004a). Rational and random mutagenesis of firefly luciferase to identify an efficient emitter of red bioluminescence. Proc SPIE Genet Eng Optical Probes Biomed Appl II 5239, 170- 177.

Branchini, B.R. (2000). Chemical synthesis of firefly luciferin analogs and inhibitors. Methods Enzymol 305, 188-195.

Branchini, B.R., Ablamsky, D.M., Davis, A.L., Southworth, T.L., Butler, B., Fan, F., Jathoul, A.P., and Pule, M.A. (2010a). Red-emitting luciferases for bioluminescence reporter and imaging applications. Anal Biochem 396, 290-297. 155

Branchini, B.R., Ablamsky, D.M., Murtiashaw, M.H., Uzasci, L., Fraga, H., and Southworth, T.L. (2007a). Thermostable red and green light-producing firefly luciferase mutants for bioluminescent reporter applications. Anal Biochem 361, 253-262.

Branchini, B.R., Ablamsky, D.M., and Rosenberg, J.C. (2010b). Chemically modified firefly luciferase is an efficient source of near-infrared light. Bioconjug Chem 21, 2023- 2030.

Branchini, B.R., Ablamsky, D.M., Rosenman, J.M., Uzasci, L., Southworth, T.L., and Zimmer, M. (2007b). Synergistic mutations produce blue-shifted bioluminescence in firefly luciferase. Biochemistry 46, 13847-13855.

Branchini, B.R., Hayward, M.M., Bamford, S., Brennan, P.M., and Lajiness, E.J. (1989). Naphthyl- and quinolylluciferin: green and red light emitting firefly luciferin analogues. Photochem Photobiol 49, 689-695.

Branchini, B.R., Magyar, R.A., Murtiashaw, M.H., Anderson, S.M., Helgerson, L.C., and Zimmer, M. (1999). Site-directed mutagenesis of firefly luciferase active site amino acids: a proposed model for bioluminescence color. Biochemistry 38, 13223-13230.

Branchini, B.R., Magyar, R.A., Murtiashaw, M.H., Anderson, S.M., and Zimmer, M. (1998). Site-directed mutagenesis of histidine 245 in firefly luciferase: a proposed model of the active site. Biochemistry 37, 15311-15319.

Branchini, B.R., Magyar, R.A., Murtiashaw, M.H., and Portier, N.C. (2001). The role of active site residue arginine 218 in firefly luciferase bioluminescence. Biochemistry 40, 2410-2418.

Branchini, B.R., Murtiashaw, M.H., Magyar, R.A., Portier, N.C., Ruggiero, M.C., and Stroh, J.G. (2002). Yellow-green and red firefly bioluminescence from 5,5- dimethyloxyluciferin. J Am Chem Soc 124, 2112-2113.

Branchini, B.R., Rosenberg, J.C., Ablamsky, D.M., Taylor, K.P., Southworth, T.L., and Linder, S.J. (2011). Sequential bioluminescence resonance energy transfer-fluorescence resonance energy transfer-based ratiometric protease assays with fusion proteins of firefly luciferase and red fluorescent protein. Anal Biochem 414, 239-245.

Branchini, B.R., Southworth, T.L., Khattak, N.F., Michelini, E., and Roda, A. (2005a). Red- and green-emitting firefly luciferase mutants for bioluminescent reporter applications. Anal Biochem 345, 140-148.

Branchini, B.R., Southworth, T.L., Murtiashaw, M.H., Boije, H., and Fleet, S.E. (2003). A mutagenesis study of the putative luciferin binding site residues of firefly luciferase. Biochemistry 42, 10429-10436. 156

Branchini, B.R., Southworth, T.L., Murtiashaw, M.H., Magyar, R.A., Gonzalez, S.A., Ruggiero, M.C., and Stroh, J.G. (2004b). An alternative mechanism of bioluminescence color determination in firefly luciferase. Biochemistry 43, 7255-7262.

Branchini, B.R., Southworth, T.L., Murtiashaw, M.H., Wilkinson, S.R., Khattak, N.F., Rosenberg, J.C., and Zimmer, M. (2005b). Mutagenesis evidence that the partial reactions of firefly bioluminescence are catalyzed by different conformations of the luciferase C-terminal domain. Biochemistry 44, 1385-1393.

Budzyn, K., Marley, P.D., and Sobey, C.G. (2006). Targeting Rho and Rho-kinase in the treatment of cardiovascular disease. Trends Pharmacol Sci 27, 97-104.

Burridge, K., and Wennerberg, K. (2004). Rho and Rac take center stage. Cell 116, 167- 179.

Buskirk, A.R., Ong, Y.C., Gartner, Z.J., and Liu, D.R. (2004). Directed evolution of ligand dependence: small-molecule-activated protein splicing. Proc Natl Acad Sci U S A 101, 10505-10510.

Campbell, A.K., and Sala-Newby, G. (1993). Bioluminescent and chemiluminescent indicators for rmolecular signaling and function in living cells. In Fluorescent and Luminescent Probes for Biological Activity, W.T. Mason, ed. (London, Academic Press), pp. 58-82.

Chang, C., Fernandez, T., Panchal, R., and Bayley, H. (1998). Caged catalytic subunit of cAMP-dependent protein kinase. J Am Chem Soc 120, 7661-7662.

Chang, C.Y., Niblack, B., Walker, B., and Bayley, H. (1995). A photogenerated pore- forming protein. Chem Biol 2, 391-400.

Chang, K.H., Xiang, H., and Dunaway-Mariano, D. (1997). Acyl-adenylate motif of the acyl-adenylate/thioester-forming enzyme superfamily: a site-directed mutagenesis study with the Pseudomonas sp. strain CBS3 4-chlorobenzoate:coenzyme A ligase. Biochemistry 36, 15650-15659.

Chen, P.R., Groff, D., Guo, J., Ou, W., Cellitti, S., Geierstanger, B.H., and Schultz, P.G. (2009). A facile system for encoding unnatural amino acids in mammalian cells. Angew Chem Int Ed Engl 48, 4052-4055.

Chen, S., Liu, Y., Navizet, I., Ferr, N., Fang, W.H., and Lindh, R. (2011). Systematic Theoretical Investigation on the Light Emitter of Firefly. J Chem Theory Comput 7, 798- 803.

Cheong, W.F., Prahl, S.A., and Welch, A.J. (1990). A review of the optical properties of biological tissues. J Quantum Electron 26, 2166-2185. 157

Chrissobolis, S., and Sobey, C.G. (2006). Recent evidence for an involvement of rho- kinase in cerebral vascular disease. Stroke 37, 2174-2180.

Cohn, D.H., Ogden, R.C., Abelson, J.N., Baldwin, T.O., Nealson, K.H., Simon, M.I., and Mileham, A.J. (1983). Cloning of the Vibrio harveyi luciferase genes: use of a synthetic oligonucleotide probe. Proc Natl Acad Sci U S A 80, 120-123.

Colepicolo, N.P., Costa, C., and Bechara, E.J.H. (1986). Insect Biochem 16, 803-810.

Colicelli, J. (2004). Human RAS superfamily proteins and related GTPases. Sci STKE 2004, RE13.

Contag, C.H., and Bachmann, M.H. (2002). Advances in in vivo bioluminescence imaging of gene expression. Annu Rev Biomed Eng 4, 235-260.

Contag, C.H., Spilman, S.D., Contag, P.R., Oshiro, M., Eames, B., Dennery, P., Stevenson, D.K., and Benaron, D.A. (1997). Visualizing gene expression in living mammals using a bioluminescent reporter. Photochem Photobiol 66, 523-531.

Conti, E., Franks, N.P., and Brick, P. (1996). Crystal structure of firefly luciferase throws light on a superfamily of adenylate-forming enzymes. Structure 4, 287-298.

Conti, E., Stachelhaus, T., Marahiel, M.A., and Brick, P. (1997). Structural basis for the activation of phenylalanine in the non-ribosomal biosynthesis of gramicidin S. Embo J 16, 4174-4183.

Cramer, C.J. (2002). Theories and Models. In Essentials of Computational Chemistry (New York, Wiley), p. 139.

Curley, K., and Lawrence, D.S. (1998a). J Am Chem Soc 120, 8573-8574.

Curley, K., and Lawrence, D.S. (1998b). Photoactivation of a signal transduction pathway in living cells. J Am Chem Soc 120, 8573-8574. da Silva, L.P., and daSilva, J.C.G.E. (2011). Computatoinal Investigation of the Effect of pH on the Color of Firefly Bioluminescence by DFT. Chemphyschem 12, 951-960. de Ruijter, N.C.A., Verhees, J., van Leeuwen, W., and van der Krol, A.R. (2003). Evaluation and comparison of the GUS, LUC, and GFP reporter system for gene expression studies in plants. Plant Biol 5, 103-115. de Wet, J.R., Wood, K.V., DeLuca, M., Helinski, D.R., and Subramani, S. (1987). Firefly luciferase gene: structure and expression in mammalian cells. Mol Cell Biol 7, 725-737. 158

de Wet, J.R., Wood, K.V., Helinski, D.R., and DeLuca, M. (1985). Cloning of firefly luciferase cDNA and the expression of active luciferase in Escherichia coli. Proc Natl Acad Sci U S A 82, 7870-7873.

Deluca, M. (1976). Firefly luciferase. Adv Enzymol Relat Mol Biol 44, 37-68.

DeLuca, M., Wirtz, G.W., and McElroy, W.D. (1964). Role of sulfhydryl groups in firefly luciferase. Biochemistry 3, 935-939.

Denburg, J.L., and McElroy, W.D. (1970). Anion inhibition of firefly luciferase. Arch Biochem Biophys 141, 668-675.

Derossi, D., Joliot, A.H., Chassaing, G., and Prochiantz, A. (1994). The third helix of the Antennapedia homeodomain translocates through biological membranes. J Biol Chem 269, 10444-10450.

Dorman, G., and Prestwich, G.D. (2000). Using photolabile ligands in drug discovery and development. Trends Biotechnol 18, 64-77.

Dragulescu-Andrasi, A., Chan, C.T., De, A., Massoud, T.F., and Gambhir, S.S. (2011). Bioluminescence resonance energy transfer (BRET) imaging of protein-protein interactions within deep tissues of living subjects. Proc Natl Acad Sci U S A 108, 12060- 12065.

Dvorsky, R., Blumenstein, L., Vetter, I.R., and Ahmadian, M.R. (2004). Structural insights into the interaction of ROCKI with the switch regions of RhoA. J Biol Chem 279, 7098-7104.

Edwards, W.F., Young, D.D., and Deiters, A. (2009). Light-activated Cre recombinase as a tool for the spatial and temporal control of gene function in mammalian cells. ACS Chem Biol 4, 441-445.

El-Andaloussi, S., Johansson, H.J., Lundberg, P., and Langel, U. (2006). Induction of splice correction by cell-penetrating peptide nucleic acids. J Gene Med 8, 1262-1273.

Ellis-Davies, G.C. (2007). Caged compounds: photorelease technology for control of cellular chemistry and physiology. Nat Methods 4, 619-628.

Erickson, J.W., and Cerione, R.A. (2004). Structural elements, mechanism, and evolutionary convergence of Rho protein-guanine nucleotide exchange factor complexes. Biochemistry 43, 837-842.

Feig, L.A. (1999). Tools of the trade: use of dominant-inhibitory mutants of Ras-family GTPases. Nat Cell Biol 1, E25-27. 159

Feig, L.A., and Cooper, G.M. (1988). Inhibition of NIH 3T3 cell proliferation by a mutant ras protein with preferential affinity for GDP. Mol Cell Biol 8, 3235-3243.

Fraga, H. (2008). Firefly luminescence: a historical perspective and recent developments. Photochem Photobiol Sci 7, 146-158.

Franks, N.P., Jenkins, A., Conti, E., Lieb, W.R., and Brick, P. (1998). Structural basis for the inhibition of firefly luciferase by a general anesthetic. Biophys J 75, 2205-2211.

Fukumoto, Y., Kaibuchi, K., Hori, Y., Fujioka, H., Araki, S., Ueda, T., Kikuchi, A., and Takai, Y. (1990). Molecular cloning and characterization of a novel type of regulatory protein (GDI) for the rho proteins, ras p21-like small GTP-binding proteins. Oncogene 5, 1321-1328.

Futaki, S., Suzuki, T., Ohashi, W., Yagami, T., Tanaka, S., Ueda, K., and Sugiura, Y. (2001). Arginine-rich peptides. An abundant source of membrane-permeable peptides having potential as carriers for intracellular protein delivery. J Biol Chem 276, 5836- 5840.

Gandelman, O., Brovko, L.Y., Ugarova, N., Chikishev, A.Y., and Shkurimov, A.P. (1993). Oxyluciferin fluorescence is a model of native bioluminescence in the firefly luciferin-luciferase system. J Photochem Photobiol A 19, 187-191.

Gao, Y., Dickerson, J.B., Guo, F., Zheng, J., and Zheng, Y. (2004). Rational design and characterization of a Rac GTPase-specific small molecule inhibitor. Proc Natl Acad Sci U S A 101, 7618-7623.

Gao, Y., Xing, J., Streuli, M., Leto, T.L., and Zheng, Y. (2001). Trp(56) of rac1 specifies interaction with a subset of guanine nucleotide exchange factors. J Biol Chem 276, 47530-47541.

Gelmini, S., Pinzani, P., and Pazzagli, M. (2000). Luciferase gene as reporter: comparison with the CAT gene and use in transfection and microinjection of mammalian cells. Methods Enzymol 305, 557-576.

Ghosh, M., Ichetovkin, I., Song, X., Condeelis, J.S., and Lawrence, D.S. (2002). A new strategy for caging proteins regulated by kinases. J Am Chem Soc 124, 2440-2441.

Ghosh, M., Song, X., Mouneimne, G., Sidani, M., Lawrence, D.S., and Condeelis, J.S. (2004). Cofilin promotes actin polymerization and defines the direction of cell motility. Science 304, 743-746.

Glomset, J.A., and Farnsworth, C.C. (1994). Role of protein modification reactions in programming interactions between ras-related GTPases and cell membranes. Annu Rev Cell Biol 10, 181-205. 160

Goeldner, M., and Givens, R. (2005). Dynamic Studies in Biology: phototriggers, photoswitches and caged biomolecules (Wiley-VCH, Weinheim).

Golan, R., Zehavi, U., Naim, M., Patchornik, A., and Smirnoff, P. (1996). Inhibition of Escherichia coli beta-galactosidase by 2-nitro-1-(4,5-dimethoxy-2-nitrophenyl) ethyl, a photoreversible thiol label. Biochim Biophys Acta 1293, 238-242.

Gould, S.J., and Subramani, S. (1988). Firefly luciferase as a tool in molecular and cell biology. Anal Biochem 175, 5-13.

Greer, L.F., 3rd, and Szalay, A.A. (2002). Imaging of light emission from the expression of luciferases in living cells and organisms: a review. Luminescence 17, 43-74.

Gulick, A.M., Starai, V.J., Horswill, A.R., Homick, K.M., and Escalante-Semerena, J.C. (2003). The 1.75 A crystal structure of acetyl-CoA synthetase bound to adenosine-5'- propylphosphate and coenzyme A. Biochemistry 42, 2866-2873.

Hahn, M.E., and Muir, T.W. (2004). Photocontrol of Smad2, a multiphosphorylated cell- signaling protein, through caging of activating phosphoserines. Angew Chem Int Ed Engl 43, 5800-5803.

Hall, M.P., Gruber, M.G., Hanah, R.R., Jennens-Clough, M.L., and Wood, K.V. (1998). In Bioluminescence and Chemiluminescence: Perspectives for the 21st Century (Chichester, John Wiley and Sons), pp. 392-395.

Harwood, K.R., and Miller, S.C. (2009). Leveraging a small-molecule modification to enable the photoactivation of rho GTPases. Chembiochem 10, 2855-2857.

Hastings, J.W. (1995). In Cell Physiology Source Book, N. Sperelakis, ed. (New York, Academic Press), pp. 665-681.

Hastings, J.W. (1996). Chemistries and colors of bioluminescent reactions: a review. Gene 173, 5-11.

Herbert, J., and Gravel, D. (1974). Can J Chem 52, 187-189.

Hill, F., Wood, K.V., and DeLuca, M. (1986). Firefly luciferase. In Bioluminescence and Chemiluminescence: New Perspectives, J. Scholmerich, R. Andreesen, A. Kapp, M. Ernst, and W.G. Woods, eds. (Chichester, John WIley and Sons), pp. 397-400.

Hirano, T., Hasumi, Y., Ohtsuka, K., Maki, S., Niwa, H., Yamaji, M., and Hashizume, D. (2009). Spectroscopic studies of the light-color modulation mechanism of firefly (beetle) bioluminescence. J Am Chem Soc 131, 2385-2396. 161

Hirshberg, M., Stockley, R.W., Dodson, G., and Webb, M.R. (1997). The crystal structure of human rac1, a member of the rho-family complexed with a GTP analogue. Nat Struct Biol 4, 147-152.

Hoffman, G.R., and Cerione, R.A. (2004). Signaling to the Rho GTPases: Networking with the DH domain. FEBS Lett 513, 85-91.

Hoffman, G.R., Nassar, N., and Cerione, R.A. (2000). Structure of the Rho family GTP- binding protein Cdc42 in complex with the multifunctional regulator RhoGDI. Cell 100, 345-356.

Hwang, Y.W., and Miller, D.L. (1987). A mutation that alters the nucleotide specificity of elongation factor Tu, a GTP regulatory protein. J Biol Chem 262, 13081-13085.

Ihara, K., Muraguchi, S., Kato, M., Shimizu, T., Shirakawa, M., Kuroda, S., Kaibuchi, K., and Hakoshima, T. (1998). Crystal structure of human RhoA in a dominantly active form complexed with a GTP analogue. J Biol Chem 273, 9656-9666.

Ikenoya, M., Hidaka, H., Hosoya, T., Suzuki, M., Yamamoto, N., and Sasaki, Y. (2002). Inhibition of rho-kinase-induced myristoylated alanine-rich C kinase substrate (MARCKS) phosphorylation in human neuronal cells by H-1152, a novel and specific Rho-kinase inhibitor. J Neurochem 81, 9-16.

Inomata, K., Ohno, A., Tochio, H., Isogai, S., Tenno, T., Nakase, I., Takeuchi, T., Futaki, S., Ito, Y., Hiroaki, H., et al. (2009). High-resolution multi-dimensional NMR spectroscopy of proteins in human cells. Nature 458, 106-109.

Ishizaki, T., Uehata, M., Tamechika, I., Keel, J., Nonomura, K., Maekawa, M., and Narumiya, S. (2000). Pharmacological properties of Y-27632, a specific inhibitor of rho- associated kinases. Mol Pharmacol 57, 976-983.

Itoh, R.E., Kurokawa, K., Ohba, Y., Yoshizaki, H., Mochizuki, N., and Matsuda, M. (2002). Activation of rac and cdc42 video imaged by fluorescent resonance energy transfer-based single-molecule probes in the membrane of living cells. Mol Cell Biol 22, 6582-6591.

Jacobson, K., Rajfur, Z., Vitriol, E., and Hahn, K. (2008). Chromophore-assisted laser inactivation in cell biology. Trends Cell Biol 18, 443-450.

Jay, D.G., and Sakurai, T. (1999). Chromophore-assisted laser inactivation (CALI) to elucidate cellular mechanisms of cancer. Biochim Biophys Acta 1424, M39-48.

Jensen, F. (1999). Introduction to Computational Chemistry. In (New York, Wiley), p. 89. 162

Joern, J.M., Meinhold, P., and Arnold, F.H. (2002). Analysis of shuffled gene libraries. J Mol Biol 316, 643-656.

John, J., Rensland, H., Schlichting, I., Vetter, I., Borasio, G.D., Goody, R.S., and 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.

Johnson, F.H., and Shimomura, O. (1972). Enzymatic and nonenzymatic bioluminescence. Photophysiology, 275-334.

Juodaityte, J., and Sewald, N. (2004). Synthesis of photoswitchable amino acids based on azobenzene chromophores: building blocks with potential for photoresponsive biomaterials. J Biotechnol 112, 127-138.

Just, I., Mohr, C., Schallehn, G., Menard, L., Didsbury, J.R., Vandekerckhove, J., van Damme, J., and Aktories, K. (1992). Purification and characterization of an ADP- ribosyltransferase produced by Clostridium limosum. J Biol Chem 267, 10274-10280.

Just, I., Selzer, J., Hofmann, F., Green, G.A., and Aktories, K. (1996). Inactivation of Ras by Clostridium sordellii lethal toxin-catalyzed glucosylation. J Biol Chem 271, 10149- 10153.

Just, I., Selzer, J., Wilm, M., von Eichel-Streiber, C., Mann, M., and Aktories, K. (1995a). Glucosylation of Rho proteins by Clostridium difficile toxin B. Nature 375, 500- 503.

Just, I., Wilm, M., Selzer, J., Rex, G., von Eichel-Streiber, C., Mann, M., and Aktories, K. (1995b). The enterotoxin from Clostridium difficile (ToxA) monoglucosylates the Rho proteins. J Biol Chem 270, 13932-13936.

Kajiyama, N., and Nakano, E. (1991). Isolation and characterization of mutants of firefly luciferase which produce different colors of light. Protein Eng 4, 691-693.

Kale, T.A., Raab, C., Yu, N., Dean, D.C., and Distefano, M.D. (2001). A photoactivatable prenylated cysteine designed to study isoprenoid recognition. J Am Chem Soc 123, 4373-4381.

Kaplan, J.H., Forbush, B., 3rd, and Hoffman, J.F. (1978). Rapid photolytic release of adenosine 5'-triphosphate from a protected analogue: utilization by the Na:K pump of human red blood cell ghosts. Biochemistry 17, 1929-1935.

Kapoor, T.M., and Mitchison, T.J. (1999). Allele-specific activators and inhibitors for kinesin. Proc Natl Acad Sci U S A 96, 9106-9111. 163

Katayama, M., Kawata, M., Yoshida, Y., Horiuchi, H., Yamamoto, T., Matsuura, Y., and Takai, Y. (1991). The posttranslationally modified C-terminal structure of bovine aortic smooth muscle rhoA p21. J Biol Chem 266, 12639-12645.

Kazuki, N., Yamada, N., Enomoto, T., Irie, T., Kubota, H., Ohmiya, Y., Akiyama, H., and Ando, Y. (2007). Firefly bioluminescence quantum yield and colour change by pH- sensitive green emission. Nature Photonics 2, 44-47.

Kimura, K., Tsuji, T., Takada, Y., Miki, T., and Narumiya, S. (2000). Accumulation of GTP-bound RhoA during cytokinesis and a critical role of ECT2 in this accumulation. J Biol Chem 275, 17233-17236.

Kirby, A.J., and Vagelos, A.G. (1967). Chem Commun 406.

Kitayama, Y., Kondo, T., Nakahira, Y., Nishimura, H., Ohmiya, Y., and Oyama, T. (2004). An in vivo dual-reporter system of cyanobacteria using two railroad-worm luciferases with different color emissions. Plant Cell Physiol 45, 109-113.

Kosuge, M., Takeuchi, T., Nakase, I., Jones, A.T., and Futaki, S. (2008). Cellular internalization and distribution of arginine-rich peptides as a function of extracellular peptide concentration, serum, and plasma membrane associated proteoglycans. Bioconjug Chem 19, 656-664.

Kraynov, V.S., Chamberlain, C., Bokoch, G.M., Schwartz, M.A., Slabaugh, S., and Hahn, K.M. (2000). Localized Rac activation dynamics visualized in living cells. Science 290, 333-337.

Kricka, L.J. (2000). Application of bioluminescence and chemiluminescence in biomedical sciences. Methods Enzymol 305, 333-345.

Kueltzo, L.A., Normand, N., O'Hare, P., and Middaugh, C.R. (2000). Conformational lability of herpesvirus protein VP22. J Biol Chem 275, 33213-33221.

Lall, A.B., Seliger, H.H., Biggley, W.H., and Lloyd, J.E. (1980). Ecology of colors of firefly bioluminescence. Science 210, 560-562.

Lee, H.M., Larson, D.R., and Lawrence, D.S. (2009). Illuminating the chemistry of life: design, synthesis, and applications of "caged" and related photoresponsive compounds. ACS Chem Biol 4, 409-427.

Lerm, M., Selzer, J., Hoffmeyer, A., Rapp, U.R., Aktories, K., and Schmidt, G. (1999). Deamidation of Cdc42 and Rac by Escherichia coli cytotoxic necrotizing factor 1: activation of c-Jun N-terminal kinase in HeLa cells. Infect Immun 67, 496-503. 164

Leung, D.W., Otomo, C., Chory, J., and Rosen, M.K. (2008). Genetically encoded photoswitching of actin assembly through the Cdc42-WASP-Arp2/3 complex pathway. Proc Natl Acad Sci U S A 105, 12797-12802.

Liao, J.K., and Laufs, U. (2005). Pleiotropic effects of statins. Annu Rev Pharmacol Toxicol 45, 89-118.

Lin, R., Bagrodia, S., Cerione, R., and Manor, D. (1997). A novel Cdc42Hs mutant induces cellular transformation. Curr Biol 7, 794-797.

Lin, R., Cerione, R.A., and Manor, D. (1999). Specific contributions of the small GTPases Rho, Rac, and Cdc42 to Dbl transformation. J Biol Chem 274, 23633-23641.

Liu, Y., DeVico, L., and Lindh, R. (2008). Ab initio, investigation on the chemical origin of the firefly bioluminescence. J Photochem Photobiol A 194, 261-267.

Lloyd, J.E. (1978a). In Bioluminescence in Action, P.J. Herring, ed. (New York, Academic Press), pp. 241-272.

Lloyd, J.E. (1978b). Insect Bioluminescence. In Bioluminescence in Action, P.J. Herring, ed. (London, Academic Press), pp. 241-272.

Longenecker, K., Read, P., Derewenda, U., Dauter, Z., Liu, X., Garrard, S., Walker, L., Somlyo, A.V., Nakamoto, R.K., Somlyo, A.P., et al. (1999). How RhoGDI binds Rho. Acta Crystallogr D Biol Crystallogr 55, 1503-1515.

Lorenz, W.W., McCann, R.O., Longiaru, M., and Cormier, M.J. (1991). Isolation and expression of a cDNA encoding Renilla reniformis luciferase. Proc Natl Acad Sci U S A 88, 4438-4442.

Los, G.V., and Wood, K. (2007). The HaloTag: a novel technology for cell imaging and protein analysis. Methods Mol Biol 356, 195-208.

Lu, W., Gong, D., Bar-Sagi, D., and Cole, P.A. (2001). Site-specific incorporation of a phosphotyrosine mimetic reveals a role for tyrosine phosphorylation of SHP-2 in cell signaling. Mol Cell 8, 759-769.

Ludin, A. (1981). In Bioluminescence and Chemiluminescence, M. DeLuca, and W.D. McElroy, eds. (New York, Academic Press), p. 187.

Ma, N., Marshall, A.F., and Rao, J. (2010). Near-infrared light emitting luciferase via biomineralization. J Am Chem Soc 132, 6884-6885.

Madaule, P., and Axel, R. (1985). A novel ras-related gene family. Cell 41, 31-40. 165

Maesaki, R., Ihara, K., Shimizu, T., Kuroda, S., Kaibuchi, K., and Hakoshima, T. (1999a). The structural basis of Rho effector recognition revealed by the crystal structure of human RhoA complexed with the effector domain of PKN/PRK1. Mol Cell 4, 793- 803.

Maesaki, R., Shimizu, T., Ihara, K., Kuroda, S., Kaibuchi, K., and Hakoshima, T. (1999b). Biochemical and crystallographic characterization of a Rho effector domain of the protein serine/threonine kinase N in a complex with RhoA. J Struct Biol 126, 166- 170.

Mamaev, S.V., and Laikhter, A.L. (1996). Firefly Luciferase: Alteration of the Color of Emitted Light Resulting from Substitutions at Position 286. J Am Chem Soc 118, 7243- 7244.

Marriott, G. (1994). Caged protein conjugates and light-directed generation of protein activity: preparation, photoactivation, and spectroscopic characterization of caged G-actin conjugates. Biochemistry 33, 9092-9097.

Marriott, G., and Heidecker, M. (1996). Light-directed generation of the actin-activated ATPase activity of caged heavy meromyosin. Biochemistry 35, 3170-3174.

Marriott, G., and Ottl, J. (1998). Synthesis and applications of heterobifunctional photocleavable cross-linking reagents. Methods Enzymol 291, 155-175.

Marriott, G., Ottl, J., Heidecker, M., and Gabriel, D. (1998). Light-directed activation of protein activity from caged protein conjugates. Methods Enzymol 291, 95-116.

Marriott, G., Roy, P., and Jacobson, K. (2003). Preparation and light-directed activation of caged proteins. Methods Enzymol 360, 274-288.

May, J.J., Kessler, N., Marahiel, M.A., and Stubbs, M.T. (2002). Crystal structure of DhbE, an archetype for aryl acid activating domains of modular nonribosomal peptide synthetases. Proc Natl Acad Sci U S A 99, 12120-12125.

McCapra, F. (1997). In Bioluminescence and Chemiluminescence: Molecular Reporting with Photons, J.W. Hastings, L. Kricka, and P.E. Stanley, eds. (Chichester, U.K., Wiley), pp. 7-15.

McCapra, F., GIlfoyle, D.J., Young, D.W., Church, N.J., and Spencer, P. (1994). In Bioluminescence and Chemilminescence: Fundamentals and Applied Aspects, A.K. Campbell, L. Kricka, and P.E. Stanley, eds. (Chichester, U.K., Wiley), pp. 387-391.

McElroy, W.D., and Seliger, H.H. (1966). In Molecular Architecture in Cell Physiology, H.a.S.-G. Hayashi, I., ed. (Englewood Cliffs, NJ, Prentice-Hall), pp. 63-79. 166

McElroy, W.D., Seliger, H.H., and Deluca, M. (1965). In Evolving Genes and Proteins, V. Bryson, and H.J. Vogel, eds. (New York, Academic Press), p. 319.

Mendel, D., Ellman, J.A., and Schultz, P.G. (1991). Construction of a light-activated protein by unnatural amino acid mutagenesis. J Am Chem Soc 113, 2758-2760.

Mezzanotte, L., Que, I., Kaijzel, E., Branchini, B., Roda, A., and Lowik, C. (2011). Sensitive dual color in vivo bioluminescence imaging using a new red codon optimized firefly luciferase and a green click beetle luciferase. PLoS One 6, e19277.

Miller, J.C., Silverman, S.K., England, P.M., Dougherty, D.A., and Lester, H.A. (1998). Flash decaging of tyrosine sidechains in an ion channel. Neuron 20, 619-624.

Milne, B.F., Marques, M.A., and Nogueira, F. (2010). Fragment molecular orbital investigation of the role of AMP protonation in firefly luciferase pH-sensitivity. Phys Chem Chem Phys 12, 14285-14293.

Min, C.G., A.M., R., Guo, J.F., Zou, L.Y., Goddard, J.D., and Sun, C.C. (2010a). Theoretial investigation of the origin of yellow-green firefly bioluminescence by time- dependent density function theory. Chemphyschem 11, 2199-2204.

Min, C.G., Ren, A.M., Guo, J.F., Li, Z.W., Zou, L.Y., Goddard, J.D., and Feng, J.K. (2010b). A time-dependent density functional theory investigation on the origin of red chemiluminescence. Chemphyschem 11, 251-259.

Mochizuki, N., Yamashita, S., Kurokawa, K., Ohba, Y., Nagai, T., Miyawaki, A., and Matsuda, M. (2001). Spatio-temporal images of growth-factor-induced activation of Ras and Rap1. Nature 411, 1065-1068.

Moon, S.Y., and Zheng, Y. (2003). Rho GTPase-activating proteins in cell regulation. Trends Cell Biol 13, 13-22.

Mootz, H.D., Blum, E.S., Tyszkiewicz, A.B., and Muir, T.W. (2003). Conditional protein splicing: a new tool to control protein structure and function in vitro and in vivo. J Am Chem Soc 125, 10561-10569.

Mootz, H.D., and Muir, T.W. (2002). Protein splicing triggered by a small molecule. J Am Chem Soc 124, 9044-9045.

Moradi, A., Hosseinkhani, S., Naderi-Manesh, H., Sadeghizadeh, M., and Alipour, B.S. (2009). Effect of charge distribution in a flexible loop on the bioluminescence color of firefly luciferases. Biochemistry 48, 575-582. 167

Morris, M.C., Depollier, J., Mery, J., Heitz, F., and Divita, G. (2001). A peptide carrier for the delivery of biologically active proteins into mammalian cells. Nat Biotechnol 19, 1173-1176.

Munemitsu, S., Innis, M.A., Clark, R., McCormick, F., Ullrich, A., and Polakis, P. (1990). Molecular cloning and expression of a G25K cDNA, the human homolog of the yeast cell cycle gene CDC42. Mol Cell Biol 10, 5977-5982.

Nagumo, H., Sasaki, Y., Ono, Y., Okamoto, H., Seto, M., and Takuwa, Y. (2000). Rho kinase inhibitor HA-1077 prevents Rho-mediated myosin phosphatase inhibition in smooth muscle cells. Am J Physiol Cell Physiol 278, C57-65.

Nakajima, Y., Kimura, T., Suzuki, C., and Ohmiya, Y. (2004). Improved expression of novel red- and green-emitting luciferases of Phrixothrix railroad worms in mammalian cells. Biosci Biotechnol Biochem 68, 948-951.

Nakatani, N., Hasegawa, J.Y., and Nakatsuji, H. (2007). Red light in chemiluminescence and yellow-green light in bioluminescence: color-tuning mechanism of firefly, Photinus pyralis, studied by the symmetry-adapted cluster-configuration interaction method. J Am Chem Soc 129, 8756-8765.

Nakatsu, T., Ichiyama, S., Hiratake, J., Saldanha, A., Kobashi, N., Sakata, K., and Kato, H. (2006). Structural basis for the spectral difference in luciferase bioluminescence. Nature 440, 372-376.

Nalbant, P., Hodgson, L., Kraynov, V., Toutchkine, A., and Hahn, K.M. (2004). Activation of endogenous Cdc42 visualized in living cells. Science 305, 1615-1619.

Naumov, P., and Kochunnoonny, M. (2010). Spectral-structural effects of the keto-enol- enolate and phenol-phenolate equilibria of oxyluciferin. J Am Chem Soc 132, 11566- 11579.

Naumov, P., Ozawa, Y., Ohkubo, K., and Fukuzumi, S. (2009). Structure and spectroscopy of oxyluciferin, the light emitter of the firefly bioluminescence. J Am Chem Soc 131, 11590-11605.

Navizet, I., Liu, Y.J., Ferre, N., Xiao, H.Y., Fang, W.H., and Lindh, R. (2010). Color- tuning mechanism of firefly investigated by multi-configurational perturbation method. J Am Chem Soc 132, 706-712.

Nobes, C.D., and Hall, A. (1995). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81, 53-62. 168

Nur, E.K.M.S., Kamal, J.M., Qureshi, M.M., and Maruta, H. (1999). The CDC42-specific inhibitor derived from ACK-1 blocks v-Ha-Ras-induced transformation. Oncogene 18, 7787-7793.

Ohmiya, Y., Hirano, T., and Ohashi, M. (1996). The structural origin of the color differences in the bioluminescence of firefly luciferase. FEBS Lett 384, 83-86.

Orlova, G., Goddard, J.D., and Brovko, L.Y. (2003). Theoretical study of the amazing firefly bioluminescence: the formation and structures of the light emitters. J Am Chem Soc 125, 6962-6971.

Owen, D., Mott, H.R., Laue, E.D., and Lowe, P.N. (2000). Residues in Cdc42 that specify binding to individual CRIB effector proteins. Biochemistry 39, 1243-1250.

Patchornik, A., Amit, B., and Woodward, R.B. (1970). Photosensitive protecting groups. J Am Chem Soc 92, 6333-6335.

Perret, F., Nishihara, M., Takeuchi, T., Futaki, S., Lazar, A.N., Coleman, A.W., Sakai, N., and Matile, S. (2005). Anionic fullerenes, calixarenes, coronenes, and pyrenes as activators of oligo/polyarginines in model membranes and live cells. J Am Chem Soc 127, 1114-1115.

Pertz, O., Hodgson, L., Klemke, R.L., and Hahn, K.M. (2006). Spatiotemporal dynamics of RhoA activity in migrating cells. Nature 440, 1069-1072.

Pollitt, S.K., and Schultz, P.G. (1998). A Photochemical Switch for Controlling Protein- Protein Interactions. Angew Chem Int Ed 37, 2104-2107.

Popoff, M.R., Chaves-Olarte, E., Lemichez, E., von Eichel-Streiber, C., Thelestam, M., Chardin, P., Cussac, D., Antonny, B., Chavrier, P., Flatau, G., et al. (1996). Ras, Rap, and Rac small GTP-binding proteins are targets for Clostridium sordellii lethal toxin glucosylation. J Biol Chem 271, 10217-10224.

Powers, S., O'Neill, K., and Wigler, M. (1989). Dominant yeast and mammalian RAS mutants that interfere with the CDC25-dependent activation of wild-type RAS in Saccharomyces cerevisiae. Mol Cell Biol 9, 390-395.

Powers, T., and Walter, P. (1995). Reciprocal stimulation of GTP hydrolysis by two directly interacting GTPases. Science 269, 1422-1424.

Prescher, J.A., and Contag, C.H. Guided by the light: visualizing biomolecular processes in living animals with bioluminescence. Curr Opin Chem Biol 14, 80-89. 169

Prinz, A., Diskar, M., and Herberg, F.W. (2006). Application of bioluminescence resonance energy transfer (BRET) for biomolecular interaction studies. Chembiochem 7, 1007-1012.

Reddy, G.R., Thompson, W.C., and Miller, S.C. (2010). Robust light emission from cyclic alkylaminoluciferin substrates for firefly luciferase. J Am Chem Soc 132, 13586- 13587.

Rehemtulla, A., Stegman, L.D., Cardozo, S.J., Gupta, S., Hall, D.E., Contag, C.H., and Ross, B.D. (2000). Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging. Neoplasia 2, 491-495.

Ren, X.D., Kiosses, W.B., and Schwartz, M.A. (1999). Regulation of the small GTP- binding protein Rho by cell adhesion and the cytoskeleton. Embo J 18, 578-585.

Rhodes, W.C., and McElroy, W.D. (1958). The synthesis and function of luciferyl- adenylate and oxyluciferyl-adenylate. J Biol Chem 233, 1528-1537.

Ridley, A.J., and Hall, A. (1992). The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70, 389-399.

Riganti, C., Aldieri, E., Doublier, S., Bosia, A., and Ghigo, D. (2008). Statins-mediated inhibition of rho GTPases as a potential tool in anti-tumor therapy. Mini Rev Med Chem 8, 609-618.

Rikitake, Y., and Liao, J.K. (2005). Rho GTPases, statins, and nitric oxide. Circ Res 97, 1232-1235.

Rittinger, K., Walker, P.A., Eccleston, J.F., Nurmahomed, K., Owen, D., Laue, E., Gamblin, S.J., and Smerdon, S.J. (1997a). Crystal structure of a small G protein in complex with the GTPase-activating protein rhoGAP. Nature 388, 693-697.

Rittinger, K., Walker, P.A., Eccleston, J.F., Smerdon, S.J., and Gamblin, S.J. (1997b). Structure at 1.65 A of RhoA and its GTPase-activating protein in complex with a transition-state analogue. Nature 389, 758-762.

Roberts, P.J., Mitin, N., Keller, P.J., Chenette, E.J., Madigan, J.P., Currin, R.O., Cox, A.D., Wilson, O., Kirschmeier, P., and Der, C.J. (2008). Rho Family GTPase modification and dependence on CAAX motif-signaled posttranslational modification. J Biol Chem 283, 25150-25163.

Roda, A., Guardigli, M., Michelini, E., and Mirasoli, M. (2009). Bioluminescence in analytical chemistry and in vivo imaging. TrAC Trends Anal Chem, 307-322. 170

Roda, A., Pasini, P., Mirasoli, M., Michelini, E., and Guardigli, M. (2004). Biotechnological applications of bioluminescence and chemiluminescence. Trends Biotechnol 22, 295-303.

Rosendahl, M.S., Leonard, N.J., and DeLuca, M. (1982). Dimensional probing of the ATP binding site on firefly luciferase. Photochem Photobiol 35, 857-861.

Rossman, K.L., Worthylake, D.K., Snyder, J.T., Siderovski, D.P., Campbell, S.L., and Sondek, J. (2002). A crystallographic view of interactions between Dbs and Cdc42: PH domain-assisted guanine nucleotide exchange. Embo J 21, 1315-1326.

Sakai, N., and Matile, S. (2003). Anion-mediated transfer of polyarginine across liquid and bilayer membranes. J Am Chem Soc 125, 14348-14356.

Sala-Newby, G.B., and Campbell, A.K. (1991). Engineering a bioluminescent indicator for cyclic AMP-dependent protein kinase. Biochem J 279 ( Pt 3), 727-732.

Sandalova, T.P., and Ugarova, N.N. (1999). Model of the active site of firefly luciferase. Biochemistry (Mosc) 64, 962-967.

Sasaki, Y., Suzuki, M., and Hidaka, H. (2002). The novel and specific Rho-kinase inhibitor (S)-(+)-2-methyl-1-[(4-methyl-5-isoquinoline)sulfonyl]-homopiperazine as a probing molecule for Rho-kinase-involved pathway. Pharmacol Ther 93, 225-232.

Scheffzek, K., Stephan, I., Jensen, O.N., Illenberger, D., and Gierschik, P. (2000). The Rac-RhoGDI complex and the structural basis for the regulation of Rho proteins by RhoGDI. Nat Struct Biol 7, 122-126.

Schmidt, G., Sehr, P., Wilm, M., Selzer, J., Mann, M., and Aktories, K. (1997). Gln 63 of Rho is deamidated by Escherichia coli cytotoxic necrotizing factor-1. Nature 387, 725- 729.

Schmidt, S., Diriong, S., Mery, J., Fabbrizio, E., and Debant, A. (2002). Identification of the first Rho-GEF inhibitor, TRIPalpha, which targets the RhoA-specific GEF domain of Trio. FEBS Lett 523, 35-42.

Sebti, S.M., and Hamilton, A.D. (2000). Design of growth factor antagonists with antiangiogenic and antitumor properties. Oncogene 19, 6566-6573.

Sekine, A., Fujiwara, M., and Narumiya, S. (1989). Asparagine residue in the rho gene product is the modification site for botulinum ADP-ribosyltransferase. J Biol Chem 264, 8602-8605.

Seliger, H.H., Buck, J.B., Fastie, W.G., and McElroy, W.D. (1964). The Spectral Distribution of Firefly Light. J Gen Physiol 48, 95-104. 171

Seliger, H.H., and Mc, E.W. (1960). Spectral emission and quantum yield of firefly bioluminescence. Arch Biochem Biophys 88, 136-141.

Seliger, H.H., and McElroy, W.D. (1964). The Colors of Firefly Bioluminescence: Enzyme Configuration and Species Specificity. Proc Natl Acad Sci U S A 52, 75-81.

Seth, A., Otomo, T., Yin, H.L., and Rosen, M.K. (2003). Rational design of genetically encoded fluorescence resonance energy transfer-based sensors of cellular Cdc42 signaling. Biochemistry 42, 3997-4008.

Shapiro, E., Lu, C., and Baneyx, F. (2005). A set of multicolored Photinus pyralis luciferase mutants for in vivo bioluminescence applications. Protein Eng Des Sel 18, 581- 587.

Shimizu, T., Ihara, K., Maesaki, R., Kuroda, S., Kaibuchi, K., and Hakoshima, T. (2000). An open conformation of switch I revealed by the crystal structure of a Mg2+-free form of RHOA complexed with GDP. Implications for the GDP/GTP exchange mechanism. J Biol Chem 275, 18311-18317.

Shimomura, O. (2006). Bioluminescence: Chemical Principles and Methods (Hackensack, World Scientific Publishing Co.).

Shinde, R., Perkins, J., and Contag, C.H. (2006). Luciferin derivatives for enhanced in vitro and in vivo bioluminescence assays. Biochemistry 45, 11103-11112.

Shogren-Knaak, M.A., Alaimo, P.J., and Shokat, K.M. (2001). Recent advances in chemical approaches to the study of biological systems. Annu Rev Cell Dev Biol 17, 405- 433.

Shutes, A., Onesto, C., Picard, V., Leblond, B., Schweighoffer, F., and Der, C.J. (2007). Specificity and mechanism of action of EHT 1864, a novel small molecule inhibitor of Rac family small GTPases. J Biol Chem 282, 35666-35678.

Silva, L.P., and Da Silva, J.C. (2011). Theoretical modulation of the color of light emitted by firefly oxyluciferin. J Comput Chem 32, 2654-2663.

Snyder, J.T., Worthylake, D.K., Rossman, K.L., Betts, L., Pruitt, W.M., Siderovski, D.P., Der, C.J., and Sondek, J. (2002). Structural basis for the selective activation of Rho GTPases by Dbl exchange factors. Nat Struct Biol 9, 468-475.

So, M.K., Xu, C., Loening, A.M., Gambhir, S.S., and Rao, J. (2006). Self-illuminating quantum dot conjugates for in vivo imaging. Nat Biotechnol 24, 339-343.

Squirrel, D.J., Price, R.L., and Murphy, M.J. (2002). Rapid and specific detection of bacteria using luminescence. Anal Chim Acta 457, 109-114. 172

Srinivas, O., Mitra, N., Surolia, A., and Jayaraman, N. (2002). Photoswitchable multivalent sugar ligands: synthesis, isomerization, and lectin binding studies of azobenzene-glycopyranoside derivatives. J Am Chem Soc 124, 2124-2125.

Srinivas, O., Mitra, N., Surolia, A., and Jayaraman, N. (2005). Photoswitchable cluster glycosides as tools to probe carbohydrate-protein interactions: synthesis and lectin- binding studies of azobenzene containing multivalent sugar ligands. Glycobiology 15, 861-873.

Stewart, J.J.P. (1999). MOPAC2000. In (Tokyo, Japan, Fujitsu Limited).

Suzuki, H., Kawarabayasi, Y., Kondo, J., Abe, T., Nishikawa, K., Kimura, S., Hashimoto, T., and Yamamoto, T. (1990). Structure and regulation of rat long-chain acyl- CoA synthetase. J Biol Chem 265, 8681-8685.

Suzuki, N., and Goto, T. (1972). Studies on firefly bioluminescence. II. Identification of oxyluciferin as a product in the biolumnescence of firefly lanterns and in the chemiluminescence of firefly luciferin. Tetrahedron 28, 4075-4082.

Suzuki, N., Sato, M., Nishikawa, K., and Goto, T. (1969). Synthesis and spectral properties of 2-(6'-hydroxybenzothial-2'-yl)-4-hydroxythiazole, a possbile emitting species in the firefly bioluminescence. Tetrahedron Lett 10, 4683-4684.

Tafreshi, N., Hosseinkhani, S., Sadeghizadeh, M., Sadeghi, M., Ranjbar, B., and Naderi- Manesh, H. (2007). The influence of insertion of a critical residue (Arg356) in structure and bioluminescence spectra of firefly luciferase. J Biol Chem 282, 8641-8647.

Takakura, H., Kojima, R., Urano, Y., Terai, T., Hanaoka, K., and Nagano, T. (2011). Aminoluciferins as Functional Bioluminogenic Substrates of Firefly Luciferase. Chem Asian J.

Takakura, H., Sasakura, K., Ueno, T., Urano, Y., Terai, T., Hanaoka, K., Tsuboi, T., and Nagano, T. (2010). Development of luciferin analogues bearing an amino group and their application as BRET donors. Chem Asian J 5, 2053-2061.

Takeuchi, T., Kosuge, M., Tadokoro, A., Sugiura, Y., Nishi, M., Kawata, M., Sakai, N., Matile, S., and Futaki, S. (2006). Direct and rapid cytosolic delivery using cell- penetrating peptides mediated by pyrenebutyrate. ACS Chem Biol 1, 299-303.

Tamura, M., Nakao, H., Yoshizaki, H., Shiratsuchi, M., Shigyo, H., Yamada, H., Ozawa, T., Totsuka, J., and Hidaka, H. (2005). Development of specific Rho-kinase inhibitors and their clinical application. Biochim Biophys Acta 1754, 245-252. 173

Tannous, B.A., Kim, D.E., Fernandez, J.L., Weissleder, R., and Breakefield, X.O. (2005). Codon-optimized Gaussia luciferase cDNA for mammalian gene expression in culture and in vivo. Mol Ther 11, 435-443.

Thoren, P.E., Persson, D., Isakson, P., Goksor, M., Onfelt, A., and Norden, B. (2003). Uptake of analogs of penetratin, Tat(48-60) and oligoarginine in live cells. Biochem Biophys Res Commun 307, 100-107.

Ueda, H., Yamanouchi, H., Kitayama, A., Inoue, K., Hirano, T., Suzuki, E., Nagamune, T., and Ohmiya, Y. (1996). In Bioluminescence and Chemiluminescence: Molecular Reporting with Photons, J.W. Hastings, L. Kricka, and P.E. Stanley, eds. (Chichester, John Wiley and Sons), pp. 216-219.

Uehata, M., Ishizaki, T., Satoh, H., Ono, T., Kawahara, T., Morishita, T., Tamakawa, H., Yamagami, K., Inui, J., Maekawa, M., et al. (1997). Calcium sensitization of smooth muscle mediated by a Rho-associated protein kinase in hypertension. Nature 389, 990- 994.

Ugarova, N.N., and Brovko, L.Y. (2002). Protein structure and bioluminescent spectra for firefly bioluminescence. Luminescence 17, 321-330.

Villalobos, V., Naik, S., and Piwnica-Worms, D. (2007). Current state of imaging protein-protein interactions in vivo with genetically encoded reporters. Annu Rev Biomed Eng 9, 321-349.

Vives, E., Brodin, P., and Lebleu, B. (1997). A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus. J Biol Chem 272, 16010-16017.

Viviani, V., Uchida, A., Suenaga, N., Ryufuku, M., and Ohmiya, Y. (2001). Thr226 is a key residue for bioluminescence spectra determination in beetle luciferases. Biochem Biophys Res Commun 280, 1286-1291.

Viviani, V.R. (2002). The origin, diversity, and structure function relationships of insect luciferases. Cell Mol Life Sci 59, 1833-1850.

Viviani, V.R., Arnoldi, F.G., Venkatesh, B., Neto, A.J., Ogawa, F.G., Oehlmeyer, A.T., and Ohmiya, Y. (2006). Active-site properties of Phrixotrix railroad worm green and red bioluminescence-eliciting luciferases. J Biochem 140, 467-474.

Viviani, V.R., Bechara, E.J., and Ohmiya, Y. (1999). Cloning, sequence analysis, and expression of active Phrixothrix railroad-worms luciferases: relationship between bioluminescence spectra and primary structures. Biochemistry 38, 8271-8279. 174

Viviani, V.R., and Bechara, E.J.H. (1993). BIOPHYSICAL AND BIOCHEMICAL ASPECTS OF PHENGODID (RAILROAD-WORM) BIOLUMINESCENCE. Photochem Photobiol 58, 615-622.

Viviani, V.R., and Bechara, E.J.H. (1995). Bioluminescence of Brazilian fireflies (Coleoptera): spectral distribution and pH effect on luciferase elicited colors. Comparison with elaterid and phengodid luciferases. Photochem Photobiol 62, 490-495.

Viviani, V.R., and Bechara, E.J.H. (1997). Ann Entomol Soc Am 90, 389-398.

Viviani, V.R., and Ohmiya, Y. (2000). Bioluminescence color determinants of Phrixothrix railroad-worm luciferases: chimeric luciferases, site-directed mutagenesis of Arg 215 and guanidine effect. Photochem Photobiol 72, 267-271.

Viviani, V.R., and Ohmiya, Y. (2006). Bovine serum albumin displays luciferase-like activity in presence of luciferyl adenylate: insights on the origin of protoluciferase activity and bioluminescence colours. Luminescence 21, 262-267.

Viviani, V.R., Silva Neto, A.J., Arnoldi, F.G., Barbosa, J.A., and Ohmiya, Y. (2008). The influence of the loop between residues 223-235 in beetle luciferase bioluminescence spectra: a solvent gate for the active site of pH-sensitive luciferases. Photochem Photobiol 84, 138-144.

Viviani, V.R., Uchida, A., Viviani, W., and Ohmiya, Y. (2002). The influence of Ala243 (Gly247), Arg215 and Thr226 (Asn230) on the bioluminescence spectra and pH- sensitivity of railroad worm, click beetle and firefly luciferases. Photochem Photobiol 76, 538-544.

Von Pawel-Rammingen, U., Telepnev, M.V., Schmidt, G., Aktories, K., Wolf-Watz, H., and Rosqvist, R. (2000). GAP activity of the Yersinia YopE cytotoxin specifically targets the Rho pathway: a mechanism for disruption of actin microfilament structure. Mol Microbiol 36, 737-748.

Wadia, J.S., Stan, R.V., and Dowdy, S.F. (2004). Transducible TAT-HA fusogenic peptide enhances escape of TAT-fusion proteins after lipid raft macropinocytosis. Nat Med 10, 310-315.

Walker, J.W., Gilbert, S.H., Drummond, R.M., Yamada, M., Sreekumar, R., Carraway, R.E., Ikebe, M., and Fay, F.S. (1998). Signaling pathways underlying eosinophil cell motility revealed by using caged peptides. Proc Natl Acad Sci U S A 95, 1568-1573.

Wang, L., and Schultz, P.G. (2002). Expanding the genetic code. Chem Commun (Camb), 1-11. 175

Waterman-Storer, C.M., Worthylake, R.A., Liu, B.P., Burridge, K., and Salmon, E.D. (1999). Microtubule growth activates Rac1 to promote lamellipodial protrusion in fibroblasts. Nat Cell Biol 1, 45-50.

Wei, Y., Zhang, Y., Derewenda, U., Liu, X., Minor, W., Nakamoto, R.K., Somlyo, A.V., Somlyo, A.P., and Derewenda, Z.S. (1997). Crystal structure of RhoA-GDP and its functional implications. Nat Struct Biol 4, 699-703.

Welsh, D.K., and Noguchi, T. (2010). Cellular Bioluminescence Imaging. In Imaging: A Laboratory Manual, R. Yuste, ed. (Cold Spring Harbor, NY, Cold Spring Harbor Laboratory Press), pp. 369-385.

Wennerberg, K., and Der, C.J. (2004). Rho-family GTPases: it's not only Rac and Rho (and I like it). J Cell Sci 117, 1301-1312.

Wennerberg, K., Rossman, K.L., and Der, C.J. (2005). The Ras superfamily at a glance. J Cell Sci 118, 843-846.

White, E.H., and Branchini, B.R. (1975). Letter: Modification of firefly luciferase with a luciferin analog. A red light producing enzyme. J Am Chem Soc 97, 1243-1245.

White, E.H., Rapaport, E., Hopkins, T.A., and Seliger, H.H. (1969). Chemi- and bioluminescence of firefly luciferin. J Am Chem Soc 91, 2178-2180.

White, E.H., Rapaport, E., Seliger, H.H., and McElroy, W.D. (1971). Bioorg Chem 1, 92- 122.

White, E.H., and Roswell, D.F. (1991). Analogs and derivatives of firefly oxyluciferin, the light emitter in firefly bioluminescence. Photochem Photobiol 53, 131-136.

White, E.H., Steinmetz, M.G., Miano, J.D., Wildes, P.D., and Morland, R. (1980). J Am Chem Soc 102, 3199-3208.

White, E.H., and Worther, H. (1966). Analogs of firefly luciferin. 3. J Org Chem 31, 1484-1488.

White, E.H., Worther, H., Seliger, H.H., and McElroy, W.D. (1966). Amino Analogs of Firefly Luciferin and Biological Activity Thereof. J Am Chem Soc 88, 2015-2019.

Whitehead, I.P., Campbell, S., Rossman, K.L., and Der, C.J. (1997). Dbl family proteins. Biochim Biophys Acta 13321 , F -23.

Williams, C.L. (2003). The polybasic region of Ras and Rho family small GTPases: a regulator of protein interactions and membrane association and a site of nuclear localization signal sequences. Cell Signal 15, 1071-1080. 176

Wood, K.V. (1990). Luc genes: introduction of colour into bioluminescence assays. J Biolumin Chemilumin 5, 107-114.

Wood, K.V. (1995). The chemical mechanism and evolutionary development of beetle bioluminescence. Photochem and Photobiol 62, 662-673.

Wood, K.V., de Wet, J.R., Dewji, N., and DeLuca, M. (1984). Synthesis of active firefly luciferase by in vitro translation of RNA obtained from adult lanterns. Biochem Biophys Res Commun 124, 592-596.

Wood, K.V., Lam, Y.A., and McElroy, W.D. (1989a). Introduction to beetle luciferases and their applications. J Biolumin Chemilumin 4, 289-301.

Wood, K.V., Lam, Y.A., McElroy, W.D., and Seliger, H.H. (1989b). Bioluminescent click beetles revisited. J Biolumin Chemilumin, 31-39.

Wood, K.V., Lam, Y.A., Seliger, H.H., and McElroy, W.D. (1989c). Complementary DNA coding click beetle luciferases can elicit bioluminescence of different colors. Science 244, 700-702.

Woodroofe, C.C., Shultz, J.W., Wood, M.G., Osterman, J., Cali, J.J., Daily, W.J., Meisenheimer, P.L., and Klaubert, D.H. (2008). N-Alkylated 6'-aminoluciferins are bioluminescent substrates for Ultra-Glo and QuantiLum luciferase: new potential scaffolds for bioluminescent assays. Biochemistry 47, 10383-10393.

Worthylake, D.K., Rossman, K.L., and Sondek, J. (2000). Crystal structure of Rac1 in complex with the guanine nucleotide exchange region of Tiam1. Nature 408, 682-688.

Wu, C., Mino, K., Akimoto, H., Kawabata, M., Nakamura, K., Ozaki, M., and Ohmiya, Y. (2009a). In vivo far-red luminescence imaging of a biomarker based on BRET from Cypridina bioluminescence to an organic dye. Proc Natl Acad Sci U S A 106, 15599- 15603.

Wu, Y.I., Frey, D., Lungu, O.I., Jaehrig, A., Schlichting, I., Kuhlman, B., and Hahn, K.M. (2009b). A genetically encoded photoactivatable Rac controls the motility of living cells. Nature 461, 104-108.

Xia, Z., and Rao, J. (2009). Biosensing and imaging based on bioluminescence resonance energy transfer. Curr Opin Biotechnol 20, 37-44.

Yamakawa, Y., Ueda, H., Kitayama, A., and Nagamune, T. (2002). Rapid homogeneous immunoassay of peptides based on bioluminescence resonance energy transfer from firefly luciferase. J Biosci Bioeng 93, 537-542. 177

Yamamoto, K., Kondo, J., Hishida, T., Teranishi, Y., and Takai, Y. (1988). Purification and characterization of a GTP-binding protein with a molecular weight of 20,000 in bovine brain membranes. Identification as the rho gene product. J Biol Chem 263, 9926- 9932.

Yang, T., and Goddard, J.D. (2007). Predictions of the geometries and fluorescence emission energies of oxyluciferins. J Phys Chem A 111, 4489-4497.

Yoshizaki, H., Ohba, Y., Kurokawa, K., Itoh, R.E., Nakamura, T., Mochizuki, N., Nagashima, K., and Matsuda, M. (2003). Activity of Rho-family GTPases during cell division as visualized with FRET-based probes. J Cell Biol 162, 223-232.

Yu, Y.A., Timiryasova, T., Zhang, Q., Beltz, R., and Szalay, A.A. (2003). Optical imaging: bacteria, viruses, and mammalian cells encoding light-emitting proteins reveal the locations of primary tumors and metastases in animals. Anal Bioanal Chem 377, 964- 972.

Zelphati, O., Wang, Y., Kitada, S., Reed, J.C., Felgner, P.L., and Corbeil, J. (2001). Intracellular delivery of proteins with a new lipid-mediated delivery system. J Biol Chem 276, 35103-35110.

Zhang, F.L., and Casey, P.J. (1996). Protein prenylation: molecular mechanisms and functional consequences. Annu Rev Biochem 65, 241-269.

Zhang, Y., So, M.K., Loening, A.M., Yao, H., Gambhir, S.S., and Rao, J. (2006). HaloTag protein-mediated site-specific conjugation of bioluminescent proteins to quantum dots. Angew Chem Int Ed Engl 45, 4936-4940.

Zhang, Y.X., Perry, K., Vinci, V.A., Powell, K., Stemmer, W.P., and del Cardayre, S.B. (2002). Genome shuffling leads to rapid phenotypic improvement in bacteria. Nature 415, 644-646.

Zhao, H., Doyle, T.C., Coquoz, O., Kalish, F., Rice, B.W., and Contag, C.H. (2005). Emission spectra of bioluminescent reporters and interaction with mammalian tissue determine the sensitivity of detection in vivo. J Biomed Opt 10, 41210.

Zheng, Y., Hart, M.J., and Cerione, R.A. (1995). Guanine nucleotide exchange catalyzed by dbl oncogene product. Methods Enzymol 256, 77-84.

Zou, K., Cheley, S., Givens, R.S., and Bayley, H. (2002). Catalytic subunit of protein kinase A caged at the activating phosphothreonine. J Am Chem Soc 124, 8220-8229.