Development and Application of 19F NMR of Proteins...1.3 NMR properties of the 19F nucleus and utility of 19F NMR in protein studies 3 1.3.1 Probes available for fluorine NMR of proteins

Development and Application of 19F NMR of Proteins

by

Julianne Kitevski-LeBlanc

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Chemistry University of Toronto

© Copyright by Julianne Kitevski-LeBlanc (2010)

ii

Development and Application of 19F NMR of Proteins

Julianne Kitevski-LeBlanc

Doctor of Philosophy

Department of Chemistry University of Toronto

2010

Abstract

19F NMR studies of proteins provide unique insight into biologically relevant phenomena such as

conformational fluctuations, folding and unfolding, binding and catalysis. While there are many

advantages to the use of 19F NMR, experimental challenges limit its widespread application. The

focus of this thesis has been to address some of these limitations, including resonance assignment

and perturbations arising from fluorine probes, and to develop more robust methods of studying

protein topology by 19F NMR.

19F NMR experiments designed to measure local hydrophobicity and exposure were developed

and evaluated in two systems, Fyn SH3 and calmodulin, labeled with 3-fluorotyrosine.

Paramagnetic effects from dissolved oxygen, solvent isotope shifts from deuterium oxide, and

1H-19F NOEs were each sufficient in establishing relative solvent exposure, while the

combination of effects from oxygen and deuterium oxide were able to delineate local

hydrophobicity and solvent accessibility of 19F probes.

Two NMR based resonance assignment protocols were developed using 13C, 15N-enriched 3-

fluorotyrosine and 3-fluorophenylalanine, separately biosynthetically incorporated into

calmodulin. In the first approach, isotopic enrichment facilitated two-dimensional heteronuclear

experiments based on INEPT and COSY magnetization transfer schemes to correlate the fluorine

iii

nucleus to sidechain and backbone 1H, 13C, and 15N atoms, providing complete spectral

assignment. The assignment of 3-fluorophenylalanine resonances was achieved using 19F-, and

15N-edited homonuclear NOE experiments to connect the fluorine nucleus to intraresidue and

neighboring 1H and 15N resonances. While both strategies were successful, the NOE-based

method was vulnerable to alternate relaxation mechanisms, including chemical shift anisotropy

and chemical exchange.

Structural perturbations arising from uniform incorporation of 3-fluorophenylalanine in

calmodulin was thoroughly investigated using 19F and 1H-15N NMR spectroscopy, 15N spin

relaxation and thermal denaturation via circular dichroism spectroscopy. While stability was

unaffected, NMR experiments revealed increased protein plasticity, minor conformers and line

broadening. The merit of fractional fluorine labeling in reducing such disruptions was

demonstrated, and labeling levels of 60-75% provided an optimal balance between native-

likeness and the usual advantages of 19F NMR in our system.

The 19F NMR techniques developed here are broadly applicable and will expand the utility of 19F

NMR in studies of protein systems.

iv

Acknowledgments

Many individuals have contributed to making my doctoral studies an enjoyable and

enriching experience, and must be acknowledged for their important contributions.

I give many thanks to my supervisor, Scott Prosser, for initiating my interest in scientific

research during my undergraduate years, and ultimately defining the path I have taken

academically over the last several years. I thank you for your constant and unwavering support,

enthusiasm and optimism during the smooth and rough patches in both my professional and

personal life. Thank you for allowing me the intellectual freedom to pursue my interests, and for

providing encouragement and reassurance when it was needed. Ultimately, working as part of

your research group has provided me with many opportunities including a rich environment for

my scientific development, and it has truly been a pleasure.

I am indebted to my labmates, both past and present, for many memories, interesting

discussions and support. Thank you to Ferenc for many hours of help with the spectrometer,

data processing, computer assistance, as well as the many non-scientific chats and espresso

breaks. Many thanks to Sameer (Samuel) for the much appreciated day-to-day lab humour

(although I didn’t always show it), apple support, stimulating research chats and your ability to

meet me halfway on subjects we simply needed to agree to disagree on. Thanks to Richa for the

early morning chats at the fumehood, listening, and many memories both in and out of the lab.

Much appreciation to all past and present members of the Macdonald and Kanelis research

groups, for sharing equipment, chemicals, notes, and fond memories of my time at UTM. Also,

many thanks to Peter Macdonald and Voula Kanelis themselves for numerous helpful

suggestions and comments in group meetings and whenever I dropped by their offices. Thank

you to Lynn for her patience when my experiments (often) ran late, and for making many long

(and short) car rides to and from UTM enjoyable.

Many thanks to my committee members, Lewis Kay and Julie Forman-Kay, for their

kindness, scientific insight, and support over the last five years. I would also like to

acknowledge Irina Bezsonova and Ranjith Muhandiram for help with protein expression and

NMR experiments along the way. Many thanks to Rubina Lewis, for teaching me that

sometimes the old-fashioned way is best, for her generosity with chemicals and equipment, and

v

for company during many long sessions in the teaching lab. I would like to express my

appreciation to Jianjun Wang and his lab at Wayne State University for making me feel

welcome, and for teaching me how to express and purify proteins.

I offer many thanks to my close friends and family, without whom I would not have

written this thesis at all. Thank you to my parents for their love, financial and emotional support,

and trusting me. Thank you to my sister for being understanding, and proud of me – even though

you don’t fully understand what I do. Thank you to my grandmother Sophie for being a source

of inspiration. Finally, I am grateful to my husband for his continuous support, encouragement,

and patience.

vi

Abstract ii

Acknowledgments iv

Table of Contents vi

List of Tables xii

List of Figures xiii

1 Chapter 1: Introduction 1

1.1 Thesis Overview 2

1.2 Summary 3

1.3 NMR properties of the 19F nucleus and utility of 19F

NMR in protein studies 3

1.3.1 Probes available for fluorine NMR of proteins 5

1.3.1.1 Fluorine tags 5

1.3.1.2 Fluorinated aliphatic amino acids 10

1.3.1.3 Fluorinated aromatics amino acids 15

1.3.2 Preparation of fluorine labeled proteins 24

1.3.2.1 Protein tagging 24

1.3.2.2 Chemical synthesis 25

1.3.2.3 Biosynthetic labeling 26

1.3.2.4 Site-specific labeling 27

1.4 Experimental concerns 31

1.4.1 Chemical shift anisotropy relaxation 31

1.4.2 Structural perturbations 33

vii

1.4.3 Resonance assignment 34

1.5 Applications 38

1.5.1 Inherent properties chemical shift 38

1.5.2 Solvent exposure 39

1.5.2.1 Photo-CIDNP 39

1.5.2.2 Solvent induced isotope shift 41

1.5.2.3 Paramagnetic shifts and relaxation enhancement 43

1.5.2.4 Combination of solvent induced isotope shift and

paramagnetic effects from dissolved oxygen 46

1.5.3 Heteronuclear 1H-19F NOE 48

1.5.4 Dynamics 52

1.6 Rationale 57

2 Chapter 2: 19F NMR studies of solvent

exposure and peptide binding to an SH3

domain 59

2.1 Abstract 60

2.2 Introduction 60


2.2.2 Fyn tyrosine kinase and SH3 domains 63

2.3 Theory 64


2.3.2 Protein-peptide interactions 65

2.4 Materials and Methods 66

viii

2.4.1 Sample preparation 66

2.4.2 NMR experiments 67

2.5 Results and discussion 68

2.5.1 Effect of introducing metafluorotyrosine at residues

8,10, 49 and 54 68

2.5.2 Assignment of 19F resonances and side-chain interactions 70

2.5.3 Solvent exposure of the fluorotyrosine residues for the free

Fyn SH3 domain 73

2.5.4 Effect of a binding peptide on the fluoro-Fyn SH3 domain 77

2.5.5 Solvent exposure of the fluorotyrosine residues in the

peptide:Fyn SH3 complex 80

2.6 Conclusions 80

3 Chapter 3: A mutagenesis-free approach to

assignment of 19F resonances in

biosynthetically labeled proteins 82

3.1 Introduction 83

3.2 Results 84

3.3 Summary 86

3.4 Supplementary Data 87

3.4.1 Synthesis and Purification of 13C, 15N-enriched-L-3-fluorotyrosine 87

4 Chapter 4: Approaches for the measurement

ix

of solvent exposure in proteins by 19F NMR 90

4.1 Abstract 91

4.2 Introduction 91


4.3.1 Expression and purification of uniformly

15N,13C-L-3-fluorotyrosine labeled CaM 94

4.3.2 Determination of macroscopic binding constants 95



4.4.1 Labeling and expression 96

4.4.2 Topology through chemical shift measurements 99

4.4.3 Topology through paramagnetic additives 103

4.4.4 Heteronuclear 1H-19F NOEs 104

4.5 Conclusions and final remarks 107

5 Chapter 5: Approaches to the assignment of 19F resonances from 3-fluorophenylalanine labeled calmodulin using solution state NMR 110

5.1 Abstract 111

5.2 Introduction 111


5.3.1 Expression and purification of uniformly 15N-enriched 3-fluorophenylalanine labeled CaM 116

5.3.2 Site-directed mutagenesis 117

5.3.3 Trypsin digest of CaM and 3-fluorophenylalanine CaM 117

x

5.3.4 Determination of macroscopic binding constants 117


5.4 Results 119

5.4.1 Assessment of structural perturbations 119

5.4.2 Site-directed mutagenesis 121

5.4.3 NMR-based assignment strategy 123

5.4.4 Trypsin fragments of calmodulin 126

5.5 Conclusions and final remarks 128

6 Chapter 6: Optimizing 19F NMR protein

spectroscopy by fractional biosynthetic

labeling 130

6.1 Abstract 131

6.2 Introduction 131


6.3.1 Protein expression and purification 133


6.3.3 Circular dichroism spectroscopy and thermal denaturation 135


6.4.1 Full and fractional labeling of calmodulin: Effects on 19F

NMR spectra 136

6.4.2 Effects of 1H-15N HSQC spectra and 15N dynamics 140

6.4.3 Thermal stability of fully and fractionally 3-FPhe labeled

CaM 146

xi

6.5 Conclusions 146

6.6 Supplementary Data 147

7 Chapter 7: Discussion and Future Directions 149

7.1 Conclusions 150

7.2 Future directions 154

References 158

xii

List of Tables

Chapter 1

Table 1.1 Dynamics of 3-fluorotyrosine probes in GFP. 55

Chapter 2

Table 2.1 Chemical shifts, line widths, and relaxation rates for

Fyn SH3 and Fyn SH3:peptide. 74

Chapter 4

Table 4.1 Chemical shifts and spin–lattice relaxation rates for 19F

nuclei in 3-fluorotyosine labeled calcium-loaded and calcium-free

calmodulin. 101

Table 4.2 Paramagnetic rates, and shifts and solvent isotope shifts for

calcium-loaded and calcium-free calmodulin. 102

Chapter 6

Table 6.1 19F transverse relaxation data (1/T2), line widths (1/T2*) and

the difference [Δ(1/T2) = 1/T2* -1/T2] 139

Table 6.2 Free energy (ΔG) and correlation times for fully and fractionally

labeled 3-FPhe CaM. 145

xiii

List of Figures

Chapter 1

Figure 1.1 Chemical structures of the available fluorine tags. 6

Figure 1.2 a) Studies of diacylglycerol kinase using cysteine mutagenesis

and BTFA. b) Studies of rhodopsin using cysteine mutagenesis and TFET. 8

Figure 1.3 Spectra from G-actin labeled at five native cysteine groups

with PFP. 10

Figure 1.4 Chemical structures of fluorinated aliphatic amino acids. 11

Figure 1.5 Studies of lactobacillus casei dihydrofolate reductase (DHFR)

labeled with 5-fluoroleucine. 12

Figure 1.6 Studies of bacteriophage lambda lysozyme (LaL) labeled with

Difluoromethionine. 14

Figure 1.7 Chemical structures of the available fluorinated aromatics. 16

Figure 1.8 Studies of the sequential steps of the folding of intestinal fatty

acid binding protein (IFAB) labeled with 4-fluorophenylalanine. 18

Figure 1.9 In-cell, 470 MHz 19F NMR spectroscopy. 19

Figure 1.10 19F NMR spectra of D-Lactate dehydrogenase labeled with various

fluorinated tryptophan probes. 21

Figure 1.11 Studies of human superoxide dismutase (SOD) labeled with

3-fluorotyrosine. 22

xiv

Figure 1.12 Strategy for site-specific labeling of a protein with

4-fluorophenylalanine in E. coli. 28

Figure 1.13 Improved protocol for biosynthetic site-specific fluorine labeling

using 4-trifluoromethylphenylalanine. 29

Figure 1.14 19F NMR spectra of 3-fluorotyrosine labeled alkaline phosphatase. 32

Figure 1.15. Comparison of resonance assignment using mutagenesis

strategies in three different systems. 35

Figure 1.16 Assignment of uniformly 4-fluorophenylalanine labeled IFAB. 37

Figure 1.17 Studies of GFP uniformly enriched with 3-fluorotyrosine. 40

Figure 1.18 Studies of solvent exposure and peptide binding to the SH3

domain of fyn tyrosine kinase using 3-fluorotyrosine as the fluorine probe. 42

Figure 1.19 Examination of immersion depth, secondary structure and

protein topology of DAGK using fluorine tagging and oxygen induced

paramagnetic shifts. 45

Figure 1.20 Effects on structure and stability of a core substitution involving

Trp43 for 5-fluorotryptophan in GB1. 50

Figure 1.21 1H-19F 1D difference NOE spectra of 3-FTyr labeled calcium-free

calmodulin and calcium-loaded calmodulin. 51

Figure 1.22 Dynamics of 3-fluorotyrosine labeled GFP. 55

Figure 1.23 Quantification of millisecond timescale dynamics in

xv

4-fluorophenylalanine labeled holo- and apo-IFAB. 56

Chapter 2

Figure 2.1 Ribbon diagram of the Fyn SH3 domain. 64

Figure 2.2 (1H,15N) HSQC NMR spectrum of 15N enriched Fyn SH3 domain. 69

Figure 2.3 19F NMR spectra of the metafluoro tyrosine substituted

Fyn SH3 domain. 71

Figure 2.4 19F homonuclear NOESY of metafluorotyrosine substituted Fyn SH3. 73

Figure 2.5 Solvent induced isotope shift in metafluorotyrosine labeled Fyn SH3. 76

Figure 2.6 Peptide binding and paramagnetic shifts in metafluorotyrosine

labeled Fyn SH3. 79

Chapter 3

Figure 3.1 CT-HCCF_COSY pulse scheme. 84

Figure 3.2 NMR-based assignment of CaM enriched with

13C,15N-3-fluoro-L-tyrosine. 86

Figure 3.3 Reaction scheme for synthesis of 13C,15N-3-fluoro-L-tyrosine. 87

Figure 3.4 19F NMR spectrum of 13C,15N-3-fluoro-L-tyrosine. 87

Figure 3.5 13C NMR spectrum of 13C,15N-3-fluoro-L-tyrosine. 88

xvi

Chapter 4

Figure 4.1 Ribbon diagrams of human calmodulin in both the

calcium-loaded and calcium-free. 92

Figure 4.2 [15N,1H] HSQC NMR spectrum of 15N-enriched CaM overlaid. 98

Figure 4.3 19F NMR spectra of 3-fluorotyrosine labeled calcium-loaded

(A) and calcium-free (B) CaM. 102

Figure 4.4 1H-19F 1D difference NOE initiated by saturation of either

water (A,B) or aliphatic protons (C,D), for CaM. 106

Figure 4.5 Graphical representation of the correspondence between the

geometric average of the experimentally determined solvent isotope

shifts and O2-induced as a function of solvent exposed surface area of CaM. 108

Chapter 5

Figure 5.1 A) X-ray structure of calmodulin. B) 3-fluorophenylalanine

structure. C) Secondary structural map. 115

Figure 5.2 1H-15N HSQC spectra of 3-fluorophenyalanine labeled calmodulin. 120

Figure 5.3 19F NMR spectra of wild type (wt) and single phenylalanine to

tyrosine mutants of calmodulin uniformly labeled with 3-fluorophenylalanine 121

Figure 5.4 NMR-based assignment strategy for phenylalanine 92. 124

Figure 5.5 1H-15N HSQC spectra of 3-FPhe TR1C (A) and 3-FPhe TR2C (B). 127

Figure 5.6 Assigned 19F NMR spectra of uniformly 3-FPhe labeled TR1C,

xvii

CaM and TR2C. 128

Chapter 6

Figure 6.1 A) X-ray structure of calmodulin. B) 19F NMR spectra of calmodulin

enriched with 3% to >95% 3-FPhe. 137

Figure 6.2 1H-15N HSQC spectra of fractionally labeled CaM. 142

Figure 6.3 Histogram of chemical shift perturbations (A) changes in

15N R1 (B) and 15N R2 (C) for fully and fractionally labled 3-FPhe CaM. 143

Figure 6.4 19F NMR spectra of calmodulin enriched with >95%

4-fluorophenylalanine. 147

Figure 6.5 1H-15N HSQC spectrum of 4-FPhe enriched CaM 148

Chapter 7

Figure 7.1 Synthetic approach to the preparation of 6-fluorotryptophan. 155

1

Chapter 1 Introduction

2

1.1 Thesis Overview

The overarching goal of the work presented in this report is the development of one and

multidimensional fluorine NMR applications to aid in the study of protein structure and

dynamics. Specifically, the work consists of two main points of focus: 1) NMR-based

assignment of fluorine resonances arising from biosynthetically labeled proteins, and 2) the

measurement of solvent exposure, hydrophobicity and surface topology using 19F NMR. This

thesis is organized into the following chapters:

(1) An introduction to fluorine NMR and its application to the study of proteins. This

section begins with a description of the properties of the fluorine nucleus itself as well

as the various probes available for use in protein studies. This is followed by

experimental considerations and current applications of 19F NMR in the examination

of protein topology and dynamics.

(2) Description of the use of 19F NMR in the examination of solvent exposure and

peptide binding to the G48M mutant SH3 domain of fyn tyrosine kinase uniformly

labeled with 3-fluorotyrosine. Solvent isotope shifts and paramagnetic shifts from

dissolved oxygen and TEMPOL are used to evaluate exposure in the free and bound

states.

(3) Development of an NMR-based strategy to assign 19F NMR resonances of 13C-

enriched fluoroaromatics. This approach was demonstrated in a 148 residue soluble

protein, calmodulin, biosynthetically labeled with 13C, 15N-enriched 3-fluorotyrosine.

The synthesis and purification of this probe is also presented here.

(4) A critical evaluation of 19F NMR approaches to the study of solvent exposure and

hydrophobicity: a comparison of solvent isotope shifts, paramagnetic shifts from

dissolved oxygen and 1H-19F heteronucelar NOEs. This study employed calmodulin

biosynthetically labeled with 3-fluorotyrosine.

3

(5) Development of an NMR-based strategy to assign fluorine resonances in the absence

of 13C-enrichment. The approach was demonstrated on calmodulin, uniformly

labeled with 3-fluorophenylalanine.

(6) The use of fractional fluorine labeling to reduce structural and functional

perturbations arising from biosynthetic incorporation of fluorinated amino acids. A

thorough analysis of the effect of complete 3-fluorophenylalanine labeling in

calmodulin is followed by the assessment of optimal labeling levels to balance native

protein qualities with sufficient labeling levels to permit 19F NMR studies.

(7) The final chapter briefly summarizes the results presented and future applications are

discussed.

1.2 Summary

This chapter will serve as a brief introduction to 19F NMR, focusing on its utility in

protein studies. NMR related properties of the fluorine nucleus are introduced first, followed by

a thorough survey of the fluorine probes currently in use and the labeling strategies employed in

sample preparation. This is followed by experimental considerations including those which the

work presented here attempted to address, namely fluorine resonance assignment and structural

perturbations arising from fluorinated amino acids. Finally, we review fluorine NMR

applications used to examine structural and dynamic aspects of proteins. In each case a brief

introduction to the theory and implementation of a given technique is presented, followed by

recent examples.

1.3 NMR properties of the 19F nucleus and utility of 19F NMR in protein studies

The fluorine nucleus is a spin-½ species, which exits in 100% natural abundance and

possesses a magnetogyric ratio that is 83% that of proton. The large magnetogyric ratio translates

into both high sensitivity in 19F NMR spectroscopy, and strong dipolar couplings, allowing for

the measurement of 19F-19F and 19F-1H NOE effects (Loewen et al. 2001) to obtain information

regarding contact with solvent as well as inter- and intramolecular protein contacts. (Campos-

Olivas et al. 2002; Martinez and Gerig 2001; Wang et al. 2005). The high sensitivity coupled

4

with the virtual absence of background fluorine signal is a considerable advantage in studies of

protein complexes or in vivo applications, where signal intensity can be severely attenuated.

Perhaps the most useful attribute of 19F NMR is the inherent sensitivity of the chemical

shift to local environment. The fluorine chemical shift is primarily influenced by a large

paramagnetic term, originating from an unpaired valence electron, making it exquisitely sensitive

to local van der Waals interactions and electrostatic fields (Gerig 1994). The extent to which the 19F NMR resonance is shifted downfield from its value in the fully denatured state correlates

well with the degree of probe burial in proteins, while slight differences in density and dielectric

constants associated with H2O and D2O manifest themselves as different chemical shifts for

water exposed 19F probes. The range of chemical shifts in proteins of both fluoroaliphatic and

fluoroaromatic probes, as a function of environment, is consequently ~100 times that of the

corresponding 1H nuclei, providing a sensitive means of studying conformational change, often

without the need to resort to 2D NMR approaches to achieve separation of resonances. The

hundred-fold increase in chemical shift dispersion has another advantage in studies of dynamics,

since frequencies are likely to be modulated to a much greater extent, making it easier to monitor

weak binding, folding, enzyme kinetics, and conformational exchange, and the related physical

and thermodynamic properties (Horng and Raleigh 2003; Li and Frieden 2006; Peng 2001).

Moreover, the fluorine chemical shift, and the extent to which it is influenced by environment,

can be enhanced through the use of paramagnetic additives. Water soluble or hydrophobic

paramagnetic shift reagents may be added to determine topological information such as solvent

exposed surface area and hydrophobicity.

Fluorine probes are incorporated into proteins in a variety of ways, as described below. In

general, the substitution of a native amino acid with a fluorinated variant, or chemical

modification using a fluorinated tag, is weakly perturbing to the global structure and function of

the protein (Campos-Olivas et al. 2002; Lau and Gerig 1997; Xiao et al. 1998), or can be made

so by fractional labeling. While the van der Waals radius of a fluorine atom is only 20% larger

than hydrogen, fluorine-fluorine interactions are reinforced by the fluorophobic effect. The

fluorophobic effect refers to the superhydrophobicity of fluorocarbon species, and the selective

self-association between fluorinated moieties which have been demonstrated to be effective in

stabilizing proteins, where it is more commonly referred to as the fluoro-stabilization effect

(Chiu et al. 2006). Thus introducing multiple fluorine atoms may substantially alter protein

5

stability. While these effects can generally be ameliorated by reducing the fraction of fluorine,

there are examples where even monofluorinated variants alter native protein characteristics

(Luck and Falke 1991; Xiao et al. 1998). A prerequisite to any fluorine NMR study should be the

careful inspection of both structural and functional perturbations arising from fluorine labeling.

1.3.1 Probes available for fluorine NMR of proteins

Fluorine probes are most commonly incorporated into proteins via one of two ways: 1)

biosynthetically, wherein the expression medium is supplemented with a fluorinated amino acid

analog, or 2) chemical modification, where a fluorinated moiety is reacted with a given residue

or protein site. Progress with unnatural amino acids, most notably fluorinated phenylalanine

variants, has also made it possible to achieve site-specific fluorine labeling by biosynthetic

means with impressive yields (Cellitti et al. 2008). Chemical modification, usually of labile

sulfhydryl or amino groups, may also provide domain specific or site-specific 19F side chain

labels, through single-site mutagenesis strategies and controlled stoichiometry. In addition,

chemical ligation strategies can be used to prepare fusion proteins composed of a specifically

fluorine labeled fragment or domain prepared by full chemical synthesis and a biosynthetically

produced native domain, resulting in a segmentally labeled sample for the study of particularly

large or complex systems. The next three sections will focus on the variety of fluorine probes

used to date in 19F NMR studies of proteins.

1.3.1.1 Fluorine Tags

While biosynthetic labeling of many fluorinated amino acid analogs is well established,

one must contend with a variety of challenges including: 1) reduced expression yields due to the

toxicity associated with the fluorinated precursor, 2) inefficient uptake by the corresponding

aminoacyl-tRNA synthetase, 3) cumulative perturbations resulting from incorporation of the

fluorinated amino acid, leading to protein instability or misfolding, and finally 4) spectral

overlap and prohibitive challenges associated with obtaining assignments through mutagenesis.

Moreover, the bulk of commercially available fluorinated amino acids, which can be readily

adapted for protein expression, are aromatics and it might often be desirable to place a

fluorinated probe in a site that is not compatible with an aromatic amino acid. Thus, site specific 19F labeling of proteins, post-expression, is an attractive and potentially low-cost alternative.

6

The most common amino acid target for covalent modification is cysteine, due to the high

nucleophilicity of the sidechain sulfhydryl group. The generally limited number of reactive

cysteine groups (i.e. both accessible and reduced) in proteins offers specificity and

straightforward assignment, while cysteine mutagenesis strategies enable site-specific labeling.

(Luchette et al. 2002). The amino groups of the lysine sidechain and N-terminus are activated

towards nucleophilic covalent modifications (Adriaensens et al. 1988), as are the hydroxyl group

of serine and threonine residues (Gerig 1994), although to a lesser extent. Fluorine tagging

proceeds via a nucleophilic substitution mechanism which retains any native charged state of the

sidechain. Generally, fluorine tags are reactive trifluoromethyl derivatives, which offer

reasonably narrow line widths due to rapid rotation about the methyl symmetry axis, making

them particularly attractive in the investigation of large protein complexes and membrane

proteins (Kim et al. 2009). Chemical structures of fluorine tags used in protein studies are shown

in Figure 1.1

Figure 1.1 Chemical structures of the available fluorine tags. a) 3-bromo-1,1,1-trifluoroacetone

(BTFA) b) 2,2,2-trifluoroethanthiol (TFET) c) S-ethyl-trifluorothioacetate (SETFA) d)

trifluoroacetamidosuccinic anhydride (TFASAN) e) N-[4-

(trifluoromethyl)phenyl]iodoacetamide f) 4-(perfluoro-tert-butyl)phenyliodoacetamide (PFP)

7

g) p-nitrophenyl trifluoroacetate Note reagents b) and c) are cysteine specific.

Perhaps the most frequently used fluorine tag is 3-bromo-1,1,1-trifluoroacetone (BTFA),

which has been applied to both soluble (Bouchard et al. 1998; Critz and Martinezcarrion 1977;

Huestis and Raftery 1978; Thomas and Boxer 2001) and membrane proteins (Hellmich et al.

2009; Luchette et al. 2002; Oxenoid et al. 2002). The application of BTFA requires a simple 1-

step reaction, with the advantage that the probe is commercially available, lacks proton

couplings, and is small, causing minimal structural and functional perturbations. BTFA, in

combination with cysteine mutagenesis, was used to examine topology and secondary structure

of the N-terminal domain of diacyl glycerol kinase (DAGK), a homotrimer of 13 kDa subunits

composed of three transmembrane helices (Figure 1.2a) (Luchette et al. 2002; Oxenoid et al.

2002). The sensitivity of the BTFA probe afforded high quality spectra in minutes from sub-

millimolar sample concentrations.

8

Figure 1.2 a) Secondary structural map of diacylglycerol kinase (DAGK) with location of N-

terminal cysteine mutants indicated by dashed box. Representative 564 MHz 19F NMR spectrum

of a single cysteine mutant (mutant site position 17) labeled with BTFA is shown at right. ©

Reprinted from (Oxenoid et al. 2002) with permission from ACS. b) Studies of bacterial

rhodopsin using cysteine mutagenesis and TFET probes. Left panel shows location of cysteine

mutant pairs used in the study. The center panel reveals the 470.7 MHz 19F NMR spectra of all

three cysteine pairs relative to TFA. Plots of measured % NOE between fluorine tags as a

9

function of mixing time were used to establish proximity between cysteine residues and are

shown at right. © Reprinted from (Loewen et al. 2001) with permission from PNAS.

A cysteine specific tag also commonly used in membrane protein studies is 2,2,2-

trifluoroethanethiol (TFET) (Klein-Seetharaman et al. 1999; Loewen et al. 2001), which is

attached through a disulfide bond in two steps. The advantages of TFET include specificity for

sulfhydryl groups and increased flexibility of the resulting acyl fluoride, compared to BTFA

whose carbonyl group imposes a planar geometry on the fluorine label. In a recent study of

rhodopsin, the introduction of pairs of cysteine mutants strategically placed on the cytoplasmic

face labeled with TFET were used to establish proximity between the engineered sites (Figure

1.2b) (Loewen et al. 2001). This study demonstrates a fluorine NMR approach to the

determination of tertiary contacts in large membrane proteins. Two additional commercially

available chemical probes, which produce a carbonyl fluoride label analogous to BTFA, include

S-ethyl-trifluorothioactetate (SETFA) (Adriaensens et al. 1988) and trifluoroacetamidosuccinic

anhydride (TFASAN) (Mehta et al. 1994). In a comparative study of the utility of SETFA and

TFASAN for NMR and MRI applications, both probes were successfully used to label bovine

serum albumin, gamma globulin and immunoglobulin; however, higher labeling levels were

achieved with SETFA (Mehta et al. 1994). Additional fluorine probes include N-[4-

(trifluoromethyl)phenyl]iodoacetamide (Shriver and Sykes 1982), 4-(perfluoro-tert-

butyl)phenyliodoacetamide (PFP) (Kalbitzer et al. 1992), and p-nitrophenyltrifluoroacetate

(Adriaensens et al. 1988). Although fairly large in size, PFP has the advantage of high

sensitivity due to nine degenerate fluorine atoms which lack significant scalar couplings,

providing a single homogeneous signal. As an example, 19F NMR resonances from 74 µM G-

actin labeled with PFP were detected in a single scan, while the process of polymerization was

monitored in real time with each 19F spectrum collected in only 2.7 minutes (Figure 1.3)

(Kalbitzer et al. 1992).

10

Figure 1.3 Spectra from G-actin labeled at five native cysteine groups with PFP. a) A

comparison of 470 MHz 19F NMR spectra collected in 1000 scans (top trace) to a single scan

spectrum (lower trace). The sample contained 74 µM labeled G-actin, and both spectra were

processed with 5 Hz exponential line broadening. b) Time course of the polymerization of PFP

labeled G-actin. Polymerization was initiated at time 0 by addition of 0.03 M KCl. Spectra were

collected in 100 scans, requiring 2.7 minutes per spectrum; processing included exponential line

broadening of 4 Hz. © Reprinted from (Kalbitzer et al. 1992) with permission from John Wiley

and Sons.

A clear advantage of the fluorine tags discussed is the high sensitivity enabling fluorine

spectra to be obtained at micromolar concentrations with acquisition times facilitating real-time

observation of biologically relevant events. The combination of these characteristics might make 19F tags particularly useful in the analysis of high molecular weight systems, including

membrane proteins, in addition to in vivo protein NMR studies where macromolecular crowding

and background signal present formidable obstacles.

1.3.1.2 Fluorinated aliphatic amino acids

Several fluorinated aliphatic amino acids variants including leucine, isoleucine, valine,

alanine, proline and methionine, are available ranging from highly fluorinated to single fluorine

variants, as shown in Figure 1.4. Sample preparation strategies ranging from biosynthetic to

11

direct chemical synthesis, will be addressed in detail in the following section. Proteins labeled

with fluorinated aliphatics typically exhibit improved protein stability (Tang and Tirrell 2001)

through the well-known fluorophobic effect, prolonged half life in vivo, preferred self association

in solution and lipid environments and promotion of chemotaxis (Montclare et al. 2009).

Interestingly, highly fluorinated analogs of leucine have shown reduced helical propensity

relative to non-fluorinated leucine, while having a comparatively higher propensity for beta

strand elements (Chiu et al. 2009; Chiu et al. 2006). In either case, the associated fluoro-

stabilization effect has been shown to overcome negative effects associated with reduced

secondary structural propensities.

12

Figure 1.4 Chemical structures of the available fluorinated aliphatic amino acids. a) 5,5,5,5,5,5-

hexafluoroleucine b) 5,5,5-trifluoroleucine c) 5,5,5,5-tetrafluoroleucine d) 5-fluoroleucine e)

3,3,3-trifluoroisoleucine f) 5,5,5-trifluoroisoleucine g) 3-fluorovaline h) 4,4,4-trifluorovaline

i) 4-fluoroproline j) trifluoromethionine k) difluoromethionine l) 2,2,2-trifluoroalanine m)

2-fluoroalanine

13

A wide variety of fluorinated leucine derivatives have been employed in protein studies.

Among these, commercially available variations include 5,5,5,5,5,5-hexafluoroleucine and 5,5,5-

trifluoroleucine, while 5,5,5,5-tetrafluoro- and 5-fluoroleucine have been synthesized and used in

protein investigations. In a study of 5-fluoroleucine enriched dihydrofolate reductase (DHFR)

from L. casei, the 19F NMR spectrum resulting from incorporation at thirteen sites was observed

to span an impressive 15.2 ppm, with enhanced resolution in the presence of 1H decoupling

during acquisition. (Figure 1.5) (Feeney et al. 1996).

Figure 1.5 Studies of Lactobacillus casei

dihydrofolate reductase (DHFR) labeled with 5-

fluoroleucine at 27°C. a) 377.6 MHz 19F NMR

spectra of labeled DHFR with proton decoupling in

433000 scans (top trace); without proton

decoupling, 78000 scans (middle trace); and a

proton decoupled spectrum of denatured DHFR

with 6M guanidine hydrochloride, 169000 scans

(bottom trace). © Reprinted from (Feeney et al.

1996) with permission from ACS.

The large span of fluorine shifts observed, second to a study of 4-fluorotrytophan labeled hen

egg white lysozyme with a 16.8 ppm chemical shift range, establishes the sensitivity of aliphatic

probes to their environment and thus their utility in the analysis of hydrophobic interactions

within the protein core. Aside from 5,5,5,5-tetrafluoroleucine, all fluorinated leucine variants

have been successfully incorporated biosynthetically while 5-fluoroleucine is expected to

provide a sensitive hydrophobic probe with minimal perturbations, and less potential for stability

altering fluorophobic affects (Bilgiçer et al. 2001). Two variants of trifluoroisoleucine have been

prepared, 3,3,3-trifluoroisoleucine and 5,5,5-trifluoroisoleucine, while only the latter has been

successfully incorporated into murine dihydrofolate reductase by biosynthetic means; 3,3,3-

14

trifluroisoleucine showed no evidence of activation of the isoleucyl-tRNA synthetase (Wang et

al. 2003). Mono-fluorinated (2S,3S)-3’-fluoroisoleucine has also been synthesized for use as a

reporter of the hydrophobic core in protein studies (Charrier et al. 2004b), but has yet to be

applied in a peptide or protein system. 4,4,4-trifluorovaline is commercially available and has

been incorporated into proteins biosynthetically (Wang et al. 2004), while 4-fluorovaline has

been prepared as well (Charrier et al. 2004a). In a recent study comparing 4,4,4-trifluorovaline

and 5,5,5-trifluoroisoleucine labeling of basic leucine zipper (bZip) peptides, the melting

temperature (Tm) of the fluorinated isoleucine variant was increased by 25°C, while bZip

peptides labeled with fluorinated valine had only a 4°C increase in Tm (Son et al. 2006). Wild

type DNA binding and specificity was retained in both protein variants. An interesting

application of 4,4,4-trifluovaline is it’s application to extend traditional Φ-value analysis (Horng

and Raleigh 2003). Trifluoromethyl containing amino acids have great potential in these

applications, as they are reasonably isosteric with their native counterparts and are known to

cause measurable changes in folding ΔG°. If this strategy were combined with 19F NMR,

structural and conformational information could be obtained to complement the thermodynamic

analysis. 4-fluoroproline, commercially available as both (2S,4S) and (2S,4R) (Chorghade et al.

2008; Renner et al. 2001), has been biosynthetically incorporated into proteins including

enhanced green fluorescent protein (eGFP) (Steiner et al. 2008), and collagen (Holmgren et al.

1998). The importance of proline residues in protein folding and stability through cis/trans

isomerization of the peptide bond and Cγ-endo or -exo puckering of the ring can be examined

using fluorinated proline residues due to preferential stereochemistry imposed by the fluorine

nucleus. (2S,4R) 4-fluoroproline (4-fPro) prefers a trans peptide bond conformation and Cγ-exo

ring pucker, while (2S,4S) 4-fluoroproline favours the opposite. In a comparison of the stability

and folding of eGFP uniformly enriched in (2S,4R) or (2S,4S) 4-fPro, only the (2S,4S) 4-fPro

eGFP protein was soluble and exhibited fluorescence even though the native conformation of 9

of 10 proline residues is the trans configuration preferred by the other fluorinated isomer (Steiner

et al. 2008). The authors concluded that the steric requirements of (2S,4R) 4-fPro exceeded the

plasticity of eGFP, while the (2S,4S) 4-fPro molecule was accommodated and exhibited

enhanced folding kinetics. Methionine is available as trifluoro- (TFM) (Duewel et al. 1997) and

difluoromethionine (DFM) (Vaughan et al. 1999a) which have both been incorporated

successfully into several proteins using biosynthetic strategies (Duewel et al. 2001; Salopek-

15

Sondi et al. 2003).

Figure 1.6 Studies of bacteriophage lambda lysozyme (LaL) labeled with difluoromethionine

(DFM). a) X-ray crystal structure of LaL showing the position of the three native methionine

residues, with an enhanced view of the packing surrounding Met14 using a space filling

representation. b) 376.3 MHz 19F NMR spectrum of various DFM-LaL species. i) free DFM in

solution illustrating the ABX splitting pattern. ii) fully-enriched LaL with peak assignments

indicated. The high degree of packing surrounding Met14 is hypothesized to be responsible for

the multiple peaks observed for each diastereotopic 19F nucleus of this residue iii) DFM-LaL

Met107Leu mutant, a representative mutant spectrum illustrating how assignments were

achieved iv) DFM-LaL in the presence of 1 mM GdEDTA-. v) DFM-LaL in the presence of

7mM hexa-N-acetyl chitohexanose. © Reprinted from (Vaughan et al. 1999b) with permission

from ACS.

In studies of bacteriophage lambda lysozyme using both TFM and DFM (Duewel et al. 1997;

Vaughan et al. 1999a), the latter probe was found to be significantly less toxic allowing for near

wild type expression levels, in comparison to TFM which alone could not support bacterial

16

growth and had to be combined with its native counterpart to elicit protein expression. The

relative chemical shift dispersion of TFM and DFM in lysozyme was 0.9 ppm and ~5 ppm

respectively. Moreover, TFM was shown to cause a structural perturbation associated with

fluorine substitution at position Met14, which was manifest as two fluorine resonances for

Met107, as evidenced by the collapse to a single Met107 peak in M14L mutant spectra (Duewel

et al. 2001). An advantage of the diastereotopic nature of the two fluorine atoms in DFM is the

sensitivity of the 19F NMR spectra to internal packing. As shown in Figure 1.6, Met14 of

lysozyme enriched with DFM has two peaks for each fluorine atom presumably due to the

restricted environment surrounding this residue. Upon binding of hexa-N-acetyl chitohexaose,

only one fluorine of Met14 shifts upfield, indicating that only a minor conformational change is

associated with binding of the oligosaccharide inhibitor (Vaughan et al. 1999a). Finally, 3,3,3-

trifluoroalanine and 3-fluoroalanine are commercially available, and both have been incorporated

into peptides using chemical synthetic strategies for solid state 19F NMR analysis of

antimicrobial peptides (Grage et al. 2008).

1.3.1.3 Fluorinated aromatic amino acids

Monofluorinated aromatics represent the most common group of fluorinated analogues

used in 19F NMR studies of proteins. Many fluorinated aromatics are easily incorporated (Kim et

al. 1990) and commercially available, while some variants may be incorporated in a site-specific

manner biosynthetically (Furter 1998). The various aromatic analogues currently in use are

shown in Figure 1.7. Aromatic amino acids provide a broad sampling of structural environments,

ranging from highly exposed and dynamic loops (Keng et al. 1999; Zhang et al. 1998; Zhou et al.

1999) to the more restricted protein core (Ikura et al. 2000; Jasanoff and Weiss 1993) or

interfaces with membranes or other proteins (Jones and Thornton 1996). Phenylalanine in

particular is often buried within the protein hydrophobic core and has been used extensively in

studies of protein folding where it has been implicated in the formation and stabilization of

folding intermediates of both native and nonnative character (Heidary and Jennings 2002;

Zagrovic et al. 2002). Furthermore, tyrosine and histidine are often located within enzyme active

sites where their role may be mechanistic (Taylor et al. 1981) or structural, for example through

metal coordination (Eichler et al. 2005a) and hydrogen bonding (Thorson et al. 1998).

Furthermore, fluorination alters the sidechain pKa of these residues, allowing for the

examination of pH dependent properties.

17

Figure 1.7 Chemical structures of the available fluorinated aromatics. a) 2-fluorophenylalanine

b) 3-fluorophenylalanine c) 4-fluorophenylalanine d) 2,4,6-trifluorophenylalanine e) 2,3,4,5,6-

pentafluorophenylalanine f) 4-trifluoromethylphenylalanine g) 4-

18

trifluoromethoxyphenylalanine h) 3-fluoro-4-methylphenylalanine i) 3-fluoro-4-

trifluoromethylphenylalanine j) 4-fluorotryptophan k) 5-fluorotryptophan l) 6-fluorotryptophan

m) 2-fluorotyrosine n) 3-fluorotyrosine o) 2,6-difluorotyrosine p) 3,5-difluorotyrosine q)

2,3,5,6-tetrafluorotyrosine r) 2-fluorohistidine s) 4-fluorohistidine

Phenylalanine is commercially available as 2-, 3-, or 4-fluorophenylalanine, as well as a

2,3,4,5,6-pentafluoro form. Mono-fluorinated phenylalanine residues are incorporated via

biosynthetic strategies and have become popular probes in the examination of protein folding.

An elegant example of this is a study of 4-fluorophenylalanine labeled rat intestinal fatty acid

binding (IFAB) domain, where Li and Frieden monitored sequential folding events beginning

with millisecond nonnative-like collapse, followed by local native structure formation and a final

cooperative rearrangement of intermediate structures to the global native fold using real-time 19F

NMR (Figure 1.8) (Li and Frieden 2007).

19

Figure 1.8 Studies of the sequential steps of the folding of intestinal fatty acid binding protein

(IFAB) labeled with 4-fluorophenylalanine. a) Structure of IFAB (PDB file 2IFB) and location

of phenylalanine residues indicated. 470.3 MHz 19F NMR spectrum of wild type IFAB (black

trace) and G121V mutant (red trace) with assignments indicated. Spectra are shown to be

virtually indistinguishable, validating the use of the mutant to draw conclusions regarding

20

folding phenomena of the wild type protein. b) Kinetic refolding of the 4-FPhe G121V mutant at

10°C. i) Stopped-flow 19F NMR kinetics on dilution from 6 M to 1 M urea, with time resolution

between spectra of ~2.3 s. ii) Real-time 19F NMR spectra obtained after manual mixing from 6

M to 1 M urea, beginning at ~53 s. Each spectrum is an average of 64 scans, acquired in 73 s.

iii) Plot of intensity change in native resonances corresponding to F2/F17, F47 and F62 as a

function of time. iv) Plot of intensity change for decreasing unfolding intermediates (F62i/F68i),

and the appearance of the native resonance for F62. © Reprinted from (Li and Frieden 2007)

with permission from PNAS.

Phenylalanine residues containing trifluoromethyl groups also exist, for example 4-

trifluoromethyl (Hammill et al. 2007; Jackson et al. 2007) and 4-trifluoromethoxyphenylalanine

(Cellitti et al. 2008), which offer increased signal through three degenerate, non-J-coupled

fluorine atoms with high mobility due to methyl axis rotation. In addition to uniform

biosynthetic labeling, these residues may be incorporated site-specifically, as described in the

next section. In a recent study of the application of fluorinated probes for in-cell NMR, it was

shown that while fluorinated tyrosine can be used for small proteins (~10 kDa), the increased

signal intensity associated with 4-trifluoromethylphenylalanine permitted the observation of

much larger proteins, including histidinol dehydrogenase, a 98 kDa homodimer (Li et al. 2010)

21

(Figure 1.9).

Figure 1.9 In-cell, 470 MHz 19F NMR spectroscopy at 25°C. a) 19F 1D and 1H-15N 2D NMR

spectra of calmodulin (16 kDa) labeled with 3-fluorotyrosine. b) 19F 1D NMR spectrum of 4-

trifluoromethylphenylalanine histidinol dehydrogenase (HDH), a 98 kDa homodimer labeled at

residue 225. In-cell spectra were acquired in Escherichia coli. Supernatant spectra were used to

assess leakage, and correspond to a sample prepared from supernatants collected immediately

after completing in-cell spectra. Lysate spectra correspond to samples prepared from the

supernatants of cleared lysates of in-cell samples. The asterisks indicate free 3-fluorotyrosine

(calmodulin) and 4-trifluoromethylphenylalanine (HDH) resonances. These experiments

demonstrate the fast and simple implementation of critical leakage tests using 19F NMR and the

ability to study proteins in-cell up to ~ 100 kDa using trifluoromethyl probes where backbone

1H-15N spectra provide little or no resolution. © Reprinted from (Li et al. 2010) with

permission from ACS.

A 2,4,6-flurophenylalanine residue has also been prepared, but was shown to be a poor

substrate for native phenylalanyl-tRNA synthetase (Wang et al. 2002), and will likely require a

modified aminoacyl-tRNA synthetase enzyme for efficient biosynthetic incorporation. In a

22

recent study to establish the permissive nature of unnatural aminoacyl-tRNA synthetases several

novel fluorinated phenylalanine probes were developed including 3-fluoro-4-

methylphenylalanine and 3-fluoro-4-trifluoromethylphenylalanine, both of which were

successfully incorporated into superfolder green fluorescent protein (sfGFP) and show promise

for future fluorine NMR studies (Miyake-Stoner et al. 2010).

Fluorinated tryptophan residues are a very popular probe of protein structure and

dynamics, largely because tryptophan occurrence is generally quite rare in proteins, ~1.1% in

soluble proteins. Thus 19F NMR spectra often provide complete resolution of all

fluorotryptophan resonances. In membrane proteins, tryptophan occupancy is significantly

higher and has a high affinity for the membrane water interface where it is hypothesized to play a

functional role by anchoring residues on the trans side of the membrane (Ridder et al. 2000).

Thus, dynamics studies through T1 and T2 would be expected to reflect motions within this

environment (i.e. anisotropic diffusion, extensions in and out of the membrane etc.). 4-, 5-, and

6-fluorotryptophan are commercially available in addition to select perdeutero 5-

fluorotryptophan (Luck et al. 1996). The addition of a mono-fluorinated tryptophan analog to

expression media in excess is well tolerated, providing nearly complete enrichment. In an

examination of the effects on protein conformation induced by a variety of lipids, D-lactate

dehydrogenase was labeled with 4-, 5- and 6-fluorotryptophan (Figure 1.10). In this system,

spectra from 5-fluorotryptophan exhibited the largest peak dispersion and subsequent analysis

using this probe indicated very minor conformation changes, not only in the presence of different

lipids, but under aqueous conditions as well (Rule et al. 1987).

23

Figure 1.10 282.4 MHz 19F NMR spectra of D-Lactate dehydrogenase labeled with various

fluorinated tryptophan probes. Spectra of wild type and W→F or Y mutants used to attain

resonance assignments are shown for 4-fluortrypophan (left panel), 5-fluorotryptophan (centre

panel) and 6-fluorotryptophan (right panel). The spectra indicate optimal spectral resolution

resulting from 5-fluorotryptophan labeling, and this sample was subsequently used for lipid

binding studies. © Reprinted from (Rule et al. 1987) with permission from ACS.

Some of the earliest 19F protein NMR studies utilized 3-fluorotyrosine (Sykes et al.

1974), while 2-fluoro, 2,6-difluoro-, 3,5-difluoro- and 2,3,5,6-tetrafluorotyrosine have been

applied more recently (Thorson et al. 1998; Wilkins et al. 2010). Quint et al used 3-

fluorotysosine to investigate the conformational dynamics of human manganese superoxide

dismutase, a 22 kDa homotetramer protein with both tetrameric and dimeric interfaces (Figure

1.11). The authors observed 19F resonances from 5 of the 9 native tyrosine positions, as 4

residues were broadened beyond detection due to proximity to the paramagnetic manganese

centre. Using the temperature dependence of 19F line widths (Figure 1.11c) in combination with

backbone hydrogen deuterium exchange, the authors detected a significantly higher degree of

conformational freedom at the tetrameric interfaces. The preparation of a 13C,15N-3-fluoro-L-

tyrosine probe, by direct fluorination of 13C,15N-L-tyrosine, offers a probe capable of 13C-19F

24

correlation spectroscopy for improved resolution and NMR-based resonance assignment, without

the need to resort to mutagenesis (Kitevski-LeBlanc et al. 2009a).

Figure 1.11 Studies of human superoxide dismutase (SOD) labeled with 3-fluorotyrosine. a)

Structure of SOD (PDB file 1XDC) with location of tyrosine residues, as well as the dimeric and

tetrameric interfaces shown. b) 470 MHz 19F NMR spectrum collected at 25°C. c) Temperature

dependance of line widths at half-height for all observable 3-fluorotyrosine residues. The greater

temperature dependence of the line width of Y45, relative to probes at the dimeric interface

(Y169) was used to establish a greater degree of dynamics at the tertrameric interface. ©

Reprinted from (Quint et al. 2006) with permission from ACS.

25

Finally, 2-fluoro and 4-fluorohistidine have been prepared and successfully incorporated

into proteins biosynthetically (Eichler et al. 2005a). An additional feature of both fluorinated

tyrosine and histidine is the affect of fluorination on sidechain pKa. Fluorination at position 2

and 4 of histidine reduce the sidechain imidazole pKa from 6.0-6.5 in the native residue, to 1 and

3 respectively; while the fluorinated tyrosine variants listed above have pKa’s which range from

5.3 to 10 (Thorson et al. 1998). This feature offers the ability to examine pH dependent

properties of enzyme function and protein conformation; however, the corresponding effect on

hydrogen bonding may cause undesirable changes in protein structure or stability. In a study of

the active site tyrosine in human glutathione transferase A1-1, the authors examined a hydrogen

bond between the tyrosine hydroxyl and the sulfur of glutathione that is expected to play a role in

the transfer mechanism. Using a series of fluorinated tyrosine probes substituted for the native

active site tyrosine, the affect of decreasing the hydroxyl pKa on the transferase mechanism was

examined (Thorson et al. 1998). Only the 2,3,5,6-tetrafluorotyrosine was observed to affect the

catalytic behavour of the enzyme, indicating that the ionization properties of mono- and di-

fluorinated tyrosine labeled species is similar to the wild type enzyme. The authors suggest a

conformational change of the hydroxyl moiety in the 2,3,5,6-tetrafluorotyrosine labeled enzyme,

as the specific activity of this species was very similar to that of a Y9F mutant.

In addition to the above listed residues, 4,4,4-trifluoro-N-Fmoc-O-tert-butyl-threonine

has been synthesized, and incorporated into peptides using solid phase peptide synthesis without

racemization (Xiao et al. 2007). The isotropic chemical shifts, chemical shift tensor parameters,

intra-molecular 19F dipole-dipole couplings and relaxation time constants have all been

determined using solid state NMR for a representative collection of the aliphatic (Grage et al.

2008) and aromatic (Durr et al. 2008) fluorinated residues described above, providing vital

information for the examination of structure and dynamics.

1.3.2 Preparation of fluorine labeled proteins

1.3.2.1 Protein tagging

Tagging by chemical modification of labile amino acids, including cysteine, lysine and to

a lesser extent serine and threonine, has been applied to both soluble and membrane proteins. In

general, the protocol for most soluble proteins involves drop-wise addition of the labeling

reagent into a dilute (often < 0.25 mM) protein solution under buffered conditions at reduced

26

temperature (0° - 4°C), which is followed by stirring for 30 minutes to overnight (Kalbitzer et al.

1992; Thomas and Boxer 2001). Fluorine probes are always used in excess, ranging from 3-fold

to over 100-fold, and the pH of the reaction must be maintained to preserve native protein

conditions using base. The cessation of base consumption is also a convenient indication that the

coupling has gone to completion (Mehta et al. 1994). Reactions can be quenched by the addition

of a large (~100-200-fold) excess of a competitive nucleophilic reagent, for example

dithiothreitol (Shriver and Sykes 1982). Once the reaction is complete, removal of excess

reagent and reaction byproducts is achieved by exploiting the size difference between the protein

and label using gel-filtration chromatography, extensive dialysis or centrifugal filtering devices.

Limited aqueous solubility of fluorine tags, particularly those with phenyl groups, often require

the addition of an organic co-solvent (5-10%) such as acentonitrile (Shriver and Sykes 1982) or

dimethylformamide (Heintz et al. 1996), which may not be tolerated by some proteins. Also,

reagents which release ethanethiol byproducts as a result of hydrolysis under aqueous conditions,

such as SETFA, may disrupt or rearrange native disulfide bonds resulting in protein unfolding,

aggregation and precipitation, as observed in previous studies of lysozyme (Adriaensens et al.

1988). The use of an aprotic polar solvent, such as dimethylformamide may circumvent this

problem, but these conditions also have the potential to unfold the protein of interest. Membrane

proteins are often labeled while bound to a column followed by a refolding protocol, either

during or after elution. For example, rhodopsin (Klein-Seetharaman et al. 1999; Loewen et al.

2001) and proteorhodopsin (Hellmich et al. 2009) were labeled while bound to sepharose and

nickel-nitrilotriacetic acid agarose resin respectively, while diacylglycerol kinase (DAGK) has

been successfully labeled in solution followed by a reconstitutive refolding protocol (Luchette et

al. 2002). Although the guidelines given above provide a general framework for fluorine

labeling reactions, specific stoichiometry and conditions need to be optimized for each

combination of label and protein to obtain the best results.

1.3.2.2 Chemical synthesis

Classic solution and solid phase peptide synthesis strategies have been employed to

provide site-specific incorporation of several fluorinated aromatic and aliphatic amino acids for

use in solution and solid state NMR applications (Ulrich 2005; Xiao et al. 2007). For peptides or

small proteins (~ 30 amino acids in length), preparation by peptide synthesis offers site specific

incorporation and is often more direct than biosynthetic techniques which often require fusion

27

tags, to increase expression levels and protect the product from protease digestion. Many of the

above listed fluorinated amino acids are available as fluorenylmethyloxycarbonyl chloride

(Fmoc) and di-tert-butyl dicarbonate (tBoc) derivatives, with the necessary side-chain protection

moieties for standard coupling chemistry. Although most fluorinated amino acids can now be

synthesized in an enantiopure form, racemization during peptide synthesis can pose a significant

problem. Racemization is enhanced by the electron withdrawing nature of trifluoromethyl

groups which serve to increase the acidity of chiral protons (i.e. Hα, Hβ, Hγ) making these

centers more prone to enolization (Xiao et al. 2007). The development of racemization

suppressants has significantly improved the ability to retain chirality during peptide synthesis,

while many of them also have the ability to enhance coupling rates (Han and Kim 2004).

Although coupling efficiency depends on the specific peptide sequence, direct solid phase

peptide synthesis is generally not amenable to the production of NMR quantities of peptides

consisting of more than 50 residues. Moreover, synthesis of long polypeptides results in the

accumulation of byproducts and reduced product solubility. Nevertheless, large peptides or full-

length proteins can be produced from synthesized protein fragments using fragment

condensation or native chemical ligation strategies. Fragment condensation suffers from the

need to have the coupled fragment in excess, while native chemical ligation has successfully

been used to prepare proteins of 200-300 amino acids in length for both NMR and X-ray analysis

(Dawson and Kent 2000). Recently, a novel strategy has been developed which combines

commercially available trypsin or chymotrypsin with 4-guanidinophenyl ester derivatives of

fluorinated amino acids as substrate mimetics to produce site-specifically labeled protease-

catalysed peptides containing Cα-fluoroalkyl amino acids (Thust and Koksch 2003). Enzymatic

peptide bond formation proceeds without racemization, is highly regio- and stereoselctive, and

requires mild reaction conditions. The protocol was applied to the synthesis of di-, tri- and

tetrapeptides containing alpha-difluoro and alpha-trifluomethyl alanine, leucine and

phenylalanine for future spectroscopic applications.

1.3.2.3 Biosynthetic labeling

While auxotrophic bacterial strains are commonly employed to obtain proteins enriched

in specific fluoroaliphatics, an efficient “low-tech” solution to the production of proteins bearing

fluoroaromatics involves induced auxotrophy. The protocol employs common bacterial strains

28

used for expression, with the addition of glyphosate prior to induction. Glyphosate is a

competitive inhibitor of the 5-enolpyruvylshikimik acid-3-phosphate synthetase reaction which

occurs late in the biosynthesis of aromatic amino acids (Kim et al. 1990). The addition of 1 g/L

of glyphosate to shaking cultures one hour prior to protein expression is sufficient to halt

endogenous aromatic amino acid synthesis and affect uniform labeling with mono-fluorinated

tyrosine, phenylalanine and tryptophan to >95% efficiency (Kitevski-LeBlanc et al. ; Kitevski-

LeBlanc et al. 2009b; Lee et al. 2000; Li et al. 2009a). In order to maintain bacterial growth and

efficient protein expression, the media is also supplemented with native versions of the aromatic

amino acid(s) that are not being used as fluorine probes in a given preparation. This strategy can

be applied to label a protein with one, two or all three of the mono-fluorinated aromatics,

depending on the omission of the natural aromatic amino acid(s) to cultures one hour prior to

induction, and the addition of 19F labeled aromatic amino acid(s) at the time of induction (Kim et

al. 1990). Commercially available fluoroaromatics are not 15N-enriched, and therefore protein

samples prepared, even in minimal media using 15NH4Cl as the sole nitrogen source, will lack

nitrogen enrichment of fluorinated residues. However, 15N-enrichment of the supplemented

native aromatics does not require the use of 15N-enriched versions. Because the native amino

acids are added in excess one hour prior to induction they will participate in a feedback

inhibition process and be partially broken down until they are needed, at which point they

acquire the backbone nitrogen from glutamate, which has been produced endogenously and is

therefore 15N enriched.

For many fluorinated amino acids, uniform biosynthetic incorporation can only be

achieved using bacterial strains which are auxotrophic for the desired amino acid probe. This

includes 5,5,5,5,5,5-hexafluoroleucine, 5,5,5-trifluoroleucine, 5-fluoroleucine, 5,5,5-

trifluoroisoleucine, 4,4,4-trifluorovaline, 4-fluoroproline, trifluoromethionine,

difluoromethionine, 2-fluorohistidine, and 4-fluorohistidine. In addition to the need for an

auxotroph, the incorporation of trifluoromethionine (Duewel et al. 1997) and 5-fluoroleucine

(Feeney et al. 1996) require that the fluorine probe is added to expression media as a mixture

with its native counterpart to support bacterial growth due to their high toxicity. Lastly, high

levels of incorporation of 5,5,5,5,5,5-hexafluoroleucine (Tang and Tirrell 2001) and 4,4,4-

trifluorovaline (Wang et al. 2004) require simultaneous over-expression of their native

aminoacyl-tRNA synthetase enzymes to support adequate levels of incorporation.

29

1.3.2.4 Site-specific labeling

The ability to incorporate fluorinated amino acids site-specifically holds great promise in

the application of 19F NMR to the study of very large proteins, which might otherwise yield

prohibitively complicated spectra through uniform labeling. The earliest strategy developed

employed the co-expression of an orthogonal suppressor-tRNA/phenylalanyl-tRNA synthetase

pair to incorporate 4-fluorophenylalanine into single sites using the amber suppressor codon in E.

coli cells (Furter 1998), as shown in Figure 1.12.

Figure 1.12 Strategy for site-specific labeling of a protein with 4-fluorophenylalanine in E. coli.

A 4-fluorophenylalanine-resistant Phe-auxotroph E. coli strain (K10-F6Δ) is transfected with two

plasmids - one containing the genes for the yeast amber suppressor tRNA (tRNAPhe amber) and the

yeast phenylalanyl-tRNA synthetase, the other harbouring the gene for dihydrofolate reductase

(DHFR) with an amber stop codon engineered at position 5. The yeast tRNAPhe amber is charged

with both 4-fluorophenylalanine and native phenylalanine, but a large excess of 4-

fluorophenyalanine supplemented in the expression media favours attachment of the fluorinated

variant. The E. coli native phenylalanyl-tRNA synthetase excludes 4-fluorophenylalanine from

its binding site due to an Ala294 to Ser294 mutation. This represents the first example of

biosynthetic site-specific labeling, and produced DHFR labeled at position 5 with 64-75%

efficiency with approximately two thirds of the protein yield obtained under standard expression

conditions. © Reprinted from (Furter 1998) with permission from John Wiley and Sons.

30

Recent improvements to the above methodology have extended its use to 4-

trifluoromethylphenylalanine (Jackson et al. 2007) and 4-trifluoromethoxyphenylalanine (Jones

et al. 2010) where the labeled residue is incorporated exclusively at a single site with high

efficiency and expression yields are increased 5-fold using autoinduction techniques as

compared to the corresponding system in rich media (Figure 1.13) (Hammill et al. 2007).

Figure 1.13 Improved protocol for biosynthetic site-specific fluorine labeling using 4-

trifluoromethylphenylalanine. a) Flowchart illustrating timeline for the production of proteins

31

labeled site-specifically in E.coli using autoinduction b) Secondary plasmid encoding for

orthogonal suppressor tRNA (tRNACUA) and 4-trifluoromethylphenylalanyl-tRNA synthetase

(tfmFRS). The plasmid contains a p15 origin and teteracyclin resistance marker. d) SDS-PAGE

analysis of expression of E. coli nitroreductase (NTR), a 27 kDa protein. Top gel: a positive

control for protein production of the unlabeled protein under standard conditions. Middle gel:

negative control of protein production using the site-specific strategy, but in growth media

lacking 4-trifluoromethylphenylalanine. Bottom gel: production of 4-

trifluoromethylphenylalanine labeled NTR using the site-specific strategy combined with

autoinduction. The gels suggest near native levels of protein production with high specificity, as

indicated by 19F NMR. © Reprinted from (Hammill et al. 2007) with permission from Nature

Publishing Group.

More recently, O-nitrobenzyl-2-fluorotyrosine, -3-fluorotyrsoine and -2,6-fluorotyrosine

have been incorporated site-specifically using a modified tyrosyl-tRNA synthetase from

Methanococcus jannaschii and the suppressor codon strategy (Wilkins et al. 2010). The O-

nitrobenzyl group disguises the fluorinated variants from the endogenous translational machinery

and is easily removed post-expression with brief irradiation at 365 nm. The key to site-specific

unnatural amino acid incorporation is the modification of substrate recognition in the

corresponding aminoacyl-tRNA synthetase and subsequent evaluation of translational efficiency

and fidelity, which can be costly and time consuming. Computational strategies have been

developed which allow for evaluation of the binding energies of unnatural amino acids to a

native or modified aminoacyl-tRNA synthetase enzyme, allowing one to screen a library of

unnatural amino acids for those which will successfully participate in ribosomal translation

(Azim and Budisa 2008). Results from these virtual docking experiments have been shown to

correlate well with experimental in vivo incorporation (Wang et al. 2002). In addition to

translational efficiency and fidelity, the evolution of permissive qualities would allow a single

modified aminoacyl-tRNA synthetase to incorporate a family of related unnatural amino acids,

and remove the need to modify a synthetase enzyme for each novel probe. This concept was

used to prepare a permissive phenylalanyl-tRNA synthetase for a family of fluorinated

phenylalanine analogs including 3-fluoro-4-methylphenylalanine, 4-trifluoromethylphenylalnine

32

and 3-fluoro-4-trifluoromethylphenylalanine using a sfGFP (super folder GFP) reporter system

(Miyake-Stoner et al. 2010).

The continuous increase in the number and variety of fluorinated amino acids and

methods for uniform and site-specific incorporation appear to put no bounds on the protein

systems and biologically relevant question one can investigate using 19F NMR. The ability to

suppress dipolar relaxation by selective deuteration of aromatic ring protons has been

accomplished in native aromatics (Wishart et al. 1993), and preliminary work in our lab has

established the ability to deuterate vicinal protons in 3-fluorotyrosine and 4-fluorophenylalanine.

In addition, many of the synthetic routes for aliphatic amino acids involve precursors which are

available with 13C-enichment, for example 3,3,3-trifluoroalanine from 13C-trifluoroacteic acid,

for use in multidimensional 19F NMR applications. In general, biosynthetic or tagging

techniques are straightforward to implement and have successfully been applied to a broad class

of proteins.

1.4 Experimental concerns

1.4.1 Chemical shift anisotropy relaxation

While the large chemical shift dispersion has proven to be one of the main advantages of

fluorine in biomolecular NMR, the relaxation contribution arising from the chemical shift

anisotropy (CSA) mechanism can significantly hamper the resolution afforded by the highly

sensitive fluorine nucleus. From solid state 19F NMR experiments, the CSA of many common

aromatic amino acids probes has been determined to be large and asymmetric (Duerr et al. 2008;

Zhao et al. 2007), and scales with the square of the magnetic field. In previous studies of 3-

fluorotyrosine labeled alkaline phosphatase, 19F NMR spectra collected at 94 MHz and 235 MHz

were essentially indistinguishable, as shown in Figure 1.14, clearly establishing that the 2.5-fold

increase in frequency separation is counteracted by CSA relaxation, manifest as increased

linewidths (Hull and Sykes 1975).

33

Figure 1.14 19F NMR spectra of 3-fluorotyrosine labeled alkaline phosphatase collected at a) 94

MHz and b) 235 MHz. The similarity between these spectra illustrates the detrimental effect of

CSA relaxation to 19F line widths at high magnetic fields, which essentially cancels out the

expected increase in resolution. © Reprinted from (Hull and Sykes 1975) with permission from

Elsevier.

Anisotropic components of the chemical shift can be efficiently removed in solid state

experiments using fast magic angle spinning (MAS), sometimes in combination with 1H

decoupling, to obtain narrowed resonances at the isotropic chemical shift whose line widths

become dominated by inhomogeneous broadening due to the sensitivity of the fluorine chemical

shift (Ulrich 2005). This strategy was successfully applied to solid state 19F NMR studies of

proteorhodopsin using bromotrifluoroacetone labeled cysteine residues (Hellmich et al. 2009).

34

Under solution NMR conditions, the optimal field strength is dictated by the magnitude of

the CSA and the desired resolution. In high field studies it is important to consider a fluorine

probe with a small CSA. CSA relaxation at high magnetic fields can also interfere in the

correlation of relaxation properties to protein dynamics. In our experience, the characterization

of conformational exchange using a 19F Carr Purcell Meiboom Gill (CPMG) experiment requires

careful attention to the field strengths employed, as the dominant CSA contribution to 19F R2

rates can mask exchange contributions at high magnetic fields, which are observable at lower

fields. Care must also be taken in the analysis of such data using software programs such as

CPMGFit, which assume field independent transverse relaxation in the absence of exchange

(Palmer et al. 2001), which is no longer valid in the case of fluorine. With respect to choice of

probe, trifluoromethyl probes may offer the greatest advantage for modern high magnetic field

spectrometers. In general, trifluoromethyl probes exhibit shorter T1’s, and smaller chemical shift

anisotropies than fluorinated aromatics, while fast methyl rotation produce virtually axially

symmetric tensors (Grage et al. 2008).

1.4.2 Structural perturbations

Once a protein has been successfully labeled with a fluorinated probe, often assessed

using NMR or mass spectrometry, one must evaluate the effects of fluorination on protein

stability, structure and function. Although examples of fluorine labeled proteins are numerous,

there are a great range of effects resulting from 19F labeling. Examples in the literature also

include cases where structural aspects are intact, but function or stability is severely altered upon

fluorination. For example, in studies of ubiquitin labeled with 5-fluoroleucine, both preliminary

structural calculations and functional analyses showed no significant differences between labeled

and unlabeled samples, while thermal stability was reduced by 8°C (Alexeev et al. 2003).

Further, in a study to examine the relative effects of different fluorinated isomers, PvuII

restriction endonuclease was labeled with 2-, 3-, and 4-fluorophenylalanine, followed by the

examination of conformational stability and enzyme activity (Dominguez et al. 2001). It was

found that 3-fluorophenylalanine labeled PvuII exhibits similar conformational stability relative

to the wild type protein, but a 2-fold increase in specific activity, while 4-fluorophenylalanine

PvuII had a 4-fold increase in specific activity and a 1.5 kcal/mol decrease in stability

(Dominguez et al. 2001). Thus careful inspection of several protein characteristics is necessary

to establish true native-like qualities of fluorine labeled proteins. Structural aspects in samples

35

prepared biosynthetically with uniform 15N enrichment allow one to assess native likeness

through a comparison of 1H-15N correlation spectra of a labeled and unlabeled protein sample.

Alternative strategies for structural evaluation include circular dichroism (CD) spectroscopy, 1H

NMR spectroscopy and rigorous full structure determination by x-ray crystallography or NMR.

When assignments are available perturbations can be associated with specific sites in the protein,

and both local and distal perturbations have been observed. In an x-ray crystallographic analysis

of rat glutathione transferase M1-1 labeled with 3-fluorotyrosine, local perturbations were

observed and shown to propagate to non-neighbouring residues, indicating that even mono-

fluorinated probes can cause significant perturbations to protein structure (Xiao et al. 1998).

These effects were observed to be greatest when the fluorine atom, hydroxyl group, or both were

directly involved in interactions with other parts of the protein. Although a fluorine and proton

atom are comparable in van der Waals radius (1.4 Å vs 1.2 Å respectively) (Gerig 1994), the

tightly packed nature of a protein core, often exceeding 75% (Woolfson 2001), may not

accommodate even mono-fluorinated amino acid analogues. It should also be noted that the

electronic effects and bond lengths are substantially different between fluorine and proton

containing molecules, and it is likely that probes with several fluorine atoms will significantly

elevate the risk of both unwanted steric and electronic effects. A thorough investigation of

fluorine probe induced perturbations in calmodulin is presented in Chapter 6.

1.4.3 Resonance assignment

The most common approach to resonance assignment in 19F NMR studies of proteins

uniformly labeled with a fluorine probe employs site-directed mutagenesis strategies. Usually

this involves the direct replacement of the residue targeted for assignment with the most

structurally similar amino acid available. Successful assignments have been achieved in many

protein studies using this method (Figure 1.15a) (Anderluh et al. 2005; Duewel et al. 2001;

Evanics et al. 2007; Kranz et al. 1996; Li et al. 2009a; Pearson et al. 1997; Quint et al. 2006;

Schuler et al. 2002; Vaughan et al. 1999a), although it falls short when structural disruptions

resulting from the substitution cause significant shifts in the remaining fluorine resonances

(Figure 1.15c) (Khan et al. 2006; Kitevski-Leblanc et al. 2010). An alternate strategy involves

mutagenesis of a residue within van der Waals contact of the target residue, and is referred to as

36

the nudge mutation method. The strategy has also been used successfully (Figure 1.15b)

(Danielson et al. 1994; Drake et al. 1993), but suffers from the need to have a priori knowledge

of the protein structure. In addition, both strategies require a complete set of single mutants to

obtain full spectral assignment.

Figure 1.15. Comparison of resonance assignment using mutagenesis strategies in three

different systems. a) 470 MHz 19F NMR spectra of 5-fluorotryptophan labeled actinoporin

equinatoxin II. Wild type and W→F mutant spectra are shown as indicated up the left margin,

which provides complete assignment. © Reprinted from (Anderluh et al. 2005; Hull and Sykes

1975) with permission from Elsevier. b) 470 MHz 19F NMR spectra of 4-fluorophenylalanine

labeled CheY, a response regulator of the chemotaxis pathway in E. coli. Spectra of various

nudge mutations (bold trace) overlaid in each case with the wild type spectrum (grey trace), and

a single F→Y mutant was used to assign all six phenylalanine labels. © Reprinted from (Drake

et al. 1993). c) 564.3 MHz 19F NMR spectra of 3-fluorophenylalanine labeled calmodulin.

Spectra of the wild type protein and several F→Y mutants are shown as indicated up the right

margin. Notice that, with the exception of the F65Y mutant, the remaining spectra exhibit

significant spectral perturbations, thereby obscuring resonance assignment. . © Reprinted from

(Kitevski-Leblanc et al. 2010) with permission from Springer.

37

It is often the case that full spectral assignment cannot be achieved using one of the above

site-directed mutagenesis strategies, due to structural perturbations resulting from the

substitutions employed. Under such circumstances, there are several strategies which can be

used to resolve ambiguities, or to tentatively assign spectra in advance of performing single

replacement or nudge mutation methods. Some common strategies include the evaluation of

solvent exposure, binding of a ligand or small target molecule to a known site, chemical

modification of a nearby labile residue, and use of additional NMR experiments including 1H, 13C, or 15N experiments and relaxation analysis. Solvent exposure is particularly easy to access

in 19F NMR using simple solvent induced isotope shifts (SIIS), or paramagnetic shifts (see

below) and has been used to confirm tentative assignments in several protein systems

(Dominguez et al. 2001; Lian et al. 1994; Sun et al. 1996). In a study of 3-fluorotyrosine labeled

green fluorescent protein, assignments were completed using a combination of mutants, 19F

relaxation analysis and 19F photo-CIDNP (chemically induced dynamic nuclear polarization)

experiments, which probes solvent exposure (Khan et al. 2006). In studies of 5-fluorotryptophan

labeled F-1-ATPase, assignment of all 5 probes was achieved by breaking the protein down into

smaller subunits of known size using lauryldimethylamine oxide and carboxypeptidase treatment

followed by NMR analysis (Lee et al. 2000). Site-specific labeling with 4-fluorophenylalanine

was used to assign the corresponding spectrum from a uniformly 4-fluorophenylalanine labeled

intestinal fatty acid binding protein prepared for conformational analysis (Figure 1.16) (Li and

Frieden 2005). Although this strategy appears quite straightforward, ambiguity was still an issue

because the chemical shifts of some of the single labeled species did not align exactly with the

uniformly labeled sample. These ambiguities were clarified by the observation of crosspeaks in 19F-19F NOESY experiments and the known three-dimensional (3D) x-ray structure.

38

Figure 1.16 Assignment of uniformly 4-fluorophenylalanine labeled IFAB (see Figure 1.8 for

the protein structure and position of phenylalanine residues in IFAB) using site-specific labeling

strategies. 470.3 MHz spectra of each single labeled species is shown, as indicated on the right

margin. Notice the measurable difference in the chemical shift of single-labeled species with

that of the corresponding residue in the uniformly labeled protein. The minor peaks observed

represent a small amount of fluorine labeling at sites other the single desired position, which

actually aided in the assignment of resonances in this system. © Reprinted from (Li and Frieden

2006) with permission from ACS.

The use of NMR in the direct assignment of fluorine resonances in both the presence and

absence of 13C-enricment is presented in Chapters 3 and 5, respectively.

39

1.5 Applications

1.5.1 Inherent properties of the chemical shift

Early studies of 3-fluorotyrosine alkaline phosphatase established a strong correlation

between intramolecular relaxation and the 19F chemical shift in proteins (Hull and Sykes 1974).

The authors showed that intramolecular relaxation with protons was proportional to the extent of

the downfield shift of a given fluorine resonance relative to the corresponding denatured

chemical shift. This property has been used in protein studies to provide tentative assignments,

based on known degrees of burial from structure, and to establish the relative burial of probes

resulting from conformational exchange and binding interactions (Post et al. 1984).

The molecular origins of the observed broad chemical shift dispersion in fluoroaromatics

and fluorine labeled proteins has been the focus of many computational and ab initio approaches.

To first order, the chemical shift is determined by the chemical nature of the probe, while second

order effects, which may be as large as 20 ppm, arise from the protein structure and surrounding

solvent, and include electronic effects, shielding (i.e. van der Waals interactions) and hydrogen

bonding. (Gerig 1994). However, the relative contribution of these factors is highly debated.

Calculation of the 19F NMR spectrum of the galactose binding protein from E .coli labeled with

5-fluorotryptophan established the predominance of weak, or long-range electronic interactions

(Pearson et al. 1993). The authors found that all predicted shifts, even those arising from a 5-

fluorotryptophan pentapeptide, were downfield from the denatured chemical shift, corroborating

the experimental observations made previously with 3-fluorotyrosine labeled alkaline

phosphatase. Fluoroaromatics are expected to be predominantly affected by electronic

properties, as they are highly polarizable due to orbital overlap with the π-electron cloud;

however, this hypothesis cannot be transferred to fluorinated aliphatics, as demonstrated by the

large chemical shift dispersion of 5-fluoroleucine labeled hen egg-white lysozyme, which could

not be attributed to electronic effects alone (Feeney et al. 1996). The ~15 ppm dispersion far

exceeded the 5 ppm predicted by electronic fields, suggesting additional contributions from local

sidechain conformations. While it has been generally accepted that short-range interactions,

such as van der Waals contacts, play only a marginal role, more recent calculations have

established a significant contribution from such interactions as well as local magnetic

anisotropies (Gerig and Lau 2000). Although the origin of the fluorine chemical shift is not fully

40

understood and the prediction of NMR spectra is often difficult, particularly in the presence of

conformational mobility, the correlation between dipolar relaxation or burial and the degree of

the downfield shift of a corresponding fluorine resonance has been confirmed by ab initio

methods and experiment.

1.5.2 Solvent exposure

The solvent exposure of fluorine probes can be measured in a variety of ways, and has

been used to confirm resonance assignments, establish binding interfaces, and examine

conformational changes and folding processes. Common applications used to establish solvent

accessibility including 19F photo-CIDNP, solvent induced isotope shifts, and paramagnetic

effects including shifts, line broadening, and spin-lattice relaxation enhancement. In addition,

the combination of some of these measurements provides a detailed description of collisional

accessibility versus hydrophobicity. A brief description of the theory and application of each

technique, followed by some recent examples will be highlighted in the following sections.

1.5.2.1 Photo-CIDNP

The use of photo-CIDNP to probe exposure of histidine, tyrosine and tryptophan residues

was established in the late 1970’s (Kaptein et al. 1978), and has since been used in both solid

(Daviso et al. 2008) and liquid state NMR approaches to monitor protein folding (Day et al.

2009), conformational changes (Schlörb et al. 2006) and binding (Polyakov et al. 2004). The

amplitude of the photo-CIDNP effect is directly proportional to the magnitude of the hyperfine

coupling constant, which is large in the case of fluorinated aromatic radicals, making these

probes particularly useful for such investigations. In a recent study, 19F photo-CIDNP was used

to examine the native and denatured states of GFP labeled with 3-fluorotyrosine (Figure 1.17)

(Khan et al. 2006). The observed signal enhancements correlated well with calculated HOMO

(highest occupied molecular orbital) accessibilities of the tyrosine residues in the native state,

while careful examination of the sign and amplitude of photo-CIDNP effects indicated

conformational heterogeneity in the pH 2.9 denatured state. Although there are relatively few

examples of 19F photo-CIDNP in the literature, the large signal enhancement available with

fluorine should see this technique flourish in future protein studies.

41

42

Figure 1.17 Studies of GFP uniformly enriched with 3-fluorotyrosine. a) Structure of GFP (PDB

file 1B9C) with the position of all 10 tyrosine residues indicated. b) 19F-photo CIDNP spectra of

wild type and mutant GFP. 19F NMR spectra of Y200F and Y151F are shown above the

corresponding photo-CIDNP spectra, at left. All observed 19F polarizations are emissive,

indicating correlation times greater than 1 ns (absorptive polarizations indicate smaller

correlation times). c) Collection of 19F NMR spectra and 19F-photo CIDNP spectra of denatured

states of GFP. Notice the emissive and absorptive 19F polarizations at pH 2.9 indicating

conformational heterogeneity in the denatured state under these conditions. © Reprinted from

(Khan et al. 2006) with permission from ACS.

1.5.2.2 Solvent induced isotope shift

The exchange of the solvent from predominantly water (H2O) to deuterium oxide (D2O)

is known to deshield the signal from a water-soluble fluorine moiety resulting in an downfield

shift from the corresponding chemical shift in water. This phenomenon requires contact between

the solvent and the fluorine nucleus, and thus the magnitude of the effect should reflect the

degree of exposure of the fluorine atom (Gerig 1994). In the presence of chemical exchange

between two sites, the isotope shift measured will depend on the timescale; for example fast

exchange will result in an average measure of the solvent isotope shift in the two sites. In

practice, both upfield and downfield shifts have been observed (Evanics et al. 2007; Hull and

Sykes 1976) often up to 0.2 ppm, and shifts are linear with the mole fraction of D2O. The shift,

reported by convention as Δδ = δ (D2O) - δ (H2O), requires the collection of a 1D 19F NMR

spectrum under both water and deuterium oxide sample conditions, with all other parameters

kept constant. In addition, an internal standard may be included to allow for normalization of the

observed effect to that of a fully exposed species (Δδ* = [δ (D2O)protein - δ (H2O)protein] / [δ

(D2O)standard - δ (H2O)standard] ) (Kitevski-LeBlanc et al. 2009b). Ideally, the standard should be

similar in chemical nature; for example 4-fluorophenyalanine can be utilized as a standard in

studies of proteins enriched with 3-fluorophenyalanine or 3-fluorotyrosine, to reduce the

potential for bias due to specific interactions. In a previous study of solvent exposure and

peptide binding of fyn SH3 uniformly labeled with 3-fluorotyrosine, solvent isotope shifts were

evaluated by increasing the D2O fraction incrementally to resolve shifts and retain assignments

43

made under H2O buffer conditions, as 4 of the 5 probes were highly exposed resulting in large

shifts in the presence of 90% D2O (Figure 1.18) (Evanics et al. 2007). In addition, spin lattice

relaxation rates in D2O were small compared to H2O conditions for all 3-FTyr residues in fyn

SH3, which is consistent with high solvent exposure, as dipolar relaxation arises only from

interaction with the protein under such conditions (Hull and Sykes 1976).

Figure 1.18 Studies of solvent exposure and peptide binding to the SH3 domain of fyn tyrosine

kinase using 3-fluorotyrosine as the fluorine probe. The resonance labeled TAG is associated

with a tyrosine residue in the C-terminal tag added during cloning. a) Structure of fyn SH3 with

44

the location of the tyrosine residues indicated. b) 564.3 MHz 19F NMR spectra with increasing

fractions of deuterium oxide. Solvent induced isotope shifts are observed for all residues, and

are proportional to the mole fraction of deuterium oxide present. Overall, solvent induced

isotope shift experiments are easily implemented and fluorine nuclei produce measurable shifts

that are highly sensitive to even subtle difference in exposure. © Reprinted from (Evanics et al.

2007) with permission from Elsevier.

1.5.2.3 Paramagnetic shifts and relaxation enhancement

The effects of a paramagnetic species on a 19F nucleus generally include a change in

chemical shift and an enhancement in both spin-lattice and spin-spin relaxation. In studies

utilizing freely dissolved paramagnets the observed shift effect is referred to as a contact shift, as

the mechanism is predominantly through a Fermi contact interaction, while relaxation

enhancements are established through a dipolar mechanism (Li et al. 2009b; Prosser and

Luchette 2004). The sensitivity of the fluorine nucleus to both dipolar and contact mechanisms

of interaction is enhanced by its large magnetogyric ratio and highly polarizable electrons

respectively, making it an ideal nucleus for use in paramagnetic NMR studies of proteins. In

practice, paramagnet reagents are frequently dissolved in the buffer system, while they may also

be lipid bound for the study of membrane proteins, or attached to the protein itself either through

covalent modification of a labile group, or via metal coordination in metalloproteins. Common

examples of paramagnetic agents include molecular oxygen, nitroxide variants including

TEMPO and TEMPOL, as well as nickel and lanthanide chelates (Bernini et al. 2009). The

concentration of the paramagnetic species required varies depending on the reagent used; for

example 10-50 mM of a neutral Gd(III) species will exert a shift of similar magnitude as 0.5 to 1

mM TEMPOL in solution. The implementation of experiments using dissolved oxygen requires

commercially available high pressure NMR tubes. Our lab employs sapphire tubes with wall

thicknesses of ~ 1mm, which are coupled to a stainless steel interface and a Swagelok gas line

which can withstand pressures of 270 bar (Bezsonova et al. 2008). The sensitive nature of the 19F nucleus often requires partial pressures of only 10 - 60 bar. Contact shifts are, by convention

denoted as Δδ =(δ (O2)10-60 bar - δ (O2)0.2 bar), and similar to SIIS can be normalized by inclusion

of an appropriately selected standard molecule in the buffer. Oxygen is a useful paramagnetic

additive, in part because it can be easily added or removed, and because it is small, allowing it to

partition into protein void volumes, and across water-membrane interfaces (Bezsonova et al.

45

2008). In previous studies of the surface topology of ribonuclease A, observed shifts varied over

the protein surface indicating that oxygen penetrates into loosely packed regions of the protein

even though access appeared prohibited based on static models, providing evidence for the

existence of structural fluctuations which afford intermolecular access (Teng and Bryant 2004).

Several 19F NMR studies of soluble proteins have made use of line broadening effects (Danielson

et al. 1994; Duewel et al. 2001; Lian et al. 1994; Vaughan et al. 1999a), as well as contact shifts

and spin-lattice relaxation enhancement (Drake et al. 1993; Evanics et al. 2007; Kitevski-

LeBlanc et al. 2009b; Luck and Falke 1991) from paramagnetic additives to study protein

topology. In the examination of membrane proteins the concentration gradient of oxygen, which

increases toward the hydrophobic disordered center of micelles and lipid bilayers, allows for the

precise measurement of immersion depth, secondary structure and protein topology (Figure

1.19a) (Prosser et al. 2007). In studies of diacylglycerol kinase (DAGK), selected positions in

the first transmembrane segment were labeled with BTFA, using cysteine mutagenesis strategies,

and exposed to oxygen at 100 bar (Figure 1.19b) (Luchette et al. 2002). Through contact shifts,

the study was able to precisely define depth, secondary structure, as well as regions of contact

between the helix with other transmembrane segments or the micelle environment. The topology

of the N-terminal region of DAGK has also been examined using a combination of water soluble

(gadolinium(III)-diethylenetriaminepentaacetic acid) and membrane bound (16-doxylstearate)

paramagnetic species which aided in the delineation of transmembrane segments from

connecting loops (Oxenoid et al. 2002). Finally, dissolved oxygen at a partial pressure of 35 bar

was used to distinguish membrane immersed 5-fluorotryptophan residues from those in the loop

regions of PagP, a 18 kDa β-barrel enzyme, based on chemical shift perturbations, line

broadening and relaxation enhancements (Prosser et al. 2007). The combined sensitivity of the

fluorine nucleus with the versatility of available paramagnetic species provides a rich source of

information for the examination of topology, structure and dynamics.

46

Figure 1.19 Examination of immersion depth, secondary structure and protein topology of

DAGK using fluorine tagging and oxygen induced paramagnetic shifts. Single cysteine mutants

were engineered into the first transmembrane segment of DAGK (see Figure 1.2) and labeled

with BTFA. a) Expected trend for the modulation of paramagnetic effects from dissolved

oxygen in three membrane protein environments. b) Induced oxygen chemical shift changes at

100 bar oxygen parital pressure (ΔσP, red crosses), and the absolute chemical shift difference

from TFA for a given probe (σTFA - σi) shown as grey diamonds. The blue dashed line represents

the expected paramagnetic shift pattern for a single transmembrane helical segment of a

47

multispanning protein, and is shown to correspond well with the observed paramagnetic shifts.

c) Representative 470.3 MHz 19F NMR spectrum at 100 atm O2, and 100 atm N2. The spectra

shown are of a cysteine mutant at position 39. A helical wheel is shown at right, determined by

subtracting the fitted depth dependent chemical shift profile from the plotted profile (Figure

1.2a). The magnitude of all vectors was made positive by adding a value of 1 ppm to each. ©

Reprinted from (Prosser et al. 2007) with permission from Elsevier.

1.5.2.4 Combination of solvent induced isotope shift and paramagnetic effects from dissolved oxygen

Oxygen and water are conveniently similar in size, though different in terms of their

hydrophobicity. The combination of solvent isotope shifts and paramagnetic effects from

H2O/D2O and O2, respectively, provides new insight into topology and hydrophobicity. This

application has been applied previously, ESR studies have used complementary paramagnetic

contrast agents such as O2 and Ni(II) to separate steric effects and partitioning effects in proteins

(Altenbach et al. 1994; Hubbell et al. 1998). In NMR studies, the contact shift, ΔδO2, arising

from dissolved oxygen, including only the dominant isotropic Fermi contact mechanism, is given

by (Prosser and Luchette 2004)

(1)

where µB is the Bohr magneton, γn is the magnetogyric ratio of the nucleus, k is the Boltzmann

constant,

!

h is Planc’s constant over 2π, T is the temperature, S is the electron spin, ge is the

electronic g-factor and An represents the isotropic hyperfine electron-nuclear spin coupling

constant. A recent combined molecular dynamics and density functional theory analysis of the

paramagnetic perturbation of 19F nuclei from dissolved oxygen has demonstrated the dependence

of the Fermi contact shift on the shortest F-O distance, as well as the corresponding F-O-O angle

(Li et al. 2009b). In addition, they found that the contact shift decays sharply beyond a 4 Å

distance suggesting negligible paramagnetic effects from remote dioxygen molecules. Using a

more empirical approach, we can define the contact shift from dissolved oxygen by considering

only variables which give rise to the variation in shifts from a probe within a protein (Al-Abdul-

Wahid et al. 2006)

48

(2)

where k represents a proportionality constant, <Ω> is the collisionally accessible surface area, α

represents a polarization or spin-delocalization term, and localO >< ][ 2 represents the local oxygen

concentration. The exposed surface area is expected to be the dominant factor in the contact

shift, due to the short range nature of the effect, and α is a probe specific parameter related to the

response of a given fluorine probe to the contact shift. If we consider the normalized equivalent

of the contact shift, Δδ*O2, the terms which remain involve the ratio of accessible surface areas,

and the ratio of local oxygen concentrations such that

(3)

We can similarly define the normalized chemical shift perturbation due to the solvent isotope

effect as

(4)

which, although not of paramagnetic origin, we expect the perturbation arising from the

substitution of D2O for H2O to depend on exposed surface area and a local concentration (of

water). If we then assume that the collisional accessibilities of water and oxygen are similar we

can consider the ratio of the normalized shifts which provide an estimate of the partitioning

potential of oxygen and water, and therefore provide a hydrophobicity index, expressed as

(5)

The above methodology was employed to examine the solvent exposure and local environment

of a tryptophan residue in the folded and unfolded state of an SH3 domain (Evanics et al. 2006).

49

The application of the hydrophobicity index to 3-fluorotyrosine labeled calmodulin is presented

in Chapter 3. The measurement of solvent exposure from water and from a comparably sized

paramagnetic species provides a sensitive measure of exposed surface area (topology) and

surface potential (hydrophobicity), unique to fluorine NMR.

1.5.3 Heteronuclear 1H-19F Nuclear Overhauser effect

The heteronuclear NOE is a well established effect, often used to ascertain structural

information and dynamics (Campos-Olivas et al. 2002; Kranz et al. 1996; Sun et al. 1996; Wang

et al. 2005). Heteronuclear NOE’s between proton and fluorine nuclei offer the potential to

examine solvent-fluorine contacts and thus measure solvent exposure (Kitevski-LeBlanc et al.

2009b), intramolecular contacts for distance restraints (Campos-Olivas et al. 2002) and

intermolecular interactions between a fluorinated protein and ligand (Kranz et al. 1996), or a

fluorinated ligand and protein (Yu et al. 2006). A variety of experimental protocols have been

utilized including 1D and 2D approaches (Rinaldi 1983), proton or fluorine detection, gradient

coherence selection (Gerig 1999), and selective or broadband saturation. A primary concern,

made obvious in these studies and in our own experience, is the relatively weak magnitude of the 1H-19F NOE in slow tumbling systems and at high magnetic fields, particularly involving probes

which posses a large CSA. 1D or 2D experimental approaches which invert or saturate fluorine

magnetization and detect protons suffer from relaxation processes which significantly reduce 1H-19F cross relaxation, including the predominate CSA mechanism and chemical exchange. If one

considers only 1H-19F dipolar relaxation, the magnitude of the 19F NOE effect upon saturation of

a 1H spin may be expressed as

(6)

50

where γF and γH define the gyromagnetic ratios of the 19F and 1H nuclei, σHF and ρHF represent

the cross-relaxation and auto-relaxation terms respectively. J(ω) terms correspond to the familiar

spectral density functions given by

(7)

In large proteins, the dominant spectral density terms involved in cross-relaxation (σ) are J(0)

and J(ωH-ωF) in the homo- and heteronuclear case respectively. If we consider a medium sized

protein with a correlation time of 9 ns, the relevant spectral densities become J(ωH-ωF) = 0.72 ns

rad-1 and J(0)=3.60 ns rad-1, making the heteronuclear cross relaxation term ~5 times less

efficient than the homonuclear equivalent. The NOE will be further reduced by CSA relaxation

at high magnetic fields, and by internal motions, as has been shown in previous studies of 3-

fluorotyrosine labeled alkaline phosphatase (Hull and Sykes 1975). In addition to inherent

difficulties associated with the 1H -19F spin pair; experimental details also influence the

successful acquisition of heteronuclear NOE data. For example, proton detection necessitates the

use of sophisticated water suppression schemes to reduce water signal and artifacts arising from

samples dissolved in water. Pulse field gradient methods have been developed which provide a

more effective route to signal suppression than simple presaturation, which can create significant

apparent NOEs (Gerig 1999). In 2D applications, direct detection of fluorine has the

disadvantage that resolution of the proton axis is limited by the number of t1 increments one can

collect in a given amount of experimental time. Ideal situations include the use of 1D proton-

detected experiments in systems with one or few well-resolved fluorine signals which can be

selectively saturated and experiment time is used to optimize signal-to-noise in one or a few 1D

spectra. With typical experimental times ranging from 12-60 hours and beyond, this approach

may not be compatible with sample stability or available instrument time. Despite the

aforementioned difficulties, the heteronucelar NOE has been employed successfully in several

protein systems. One such example involves the comparison of NOE contacts of an unlabeled

protein (using aromatic 13C-edited NOESY NMR) with a 5-fluorotryptophan labeled variant via

2D 1H-19F HOESY NMR to establish structural similarities between the two proteins (Wang et

al. 2005). In a similar investigation of 5-fluorotryptophan labeled immunoglobulin binding

domain B1 of streptococcal protein G (GB1), 1H-19F HOESY NMR analysis was compared to

51

the 3D structure of the unlabeled protein to confirm structural similarities (Campos-Olivas et al.

2002). As shown in Figure 1.20, GB1 spectra collected as a function of mixing time resulted in a

linear increase in cross-peak intensities, as expected for coherences in the initial build-up regime.

A lack of cross-peaks due to spin diffusion, up to mixing times of 2 s, allowed for the

quantification of distances, which were compared with those from the available structure to

confirm structural similarities between labeled and unlabeled GB1. Kranz et al utilized 19F-

detected 2D HOESY experiments to examine the role of tyrosine residues of a human RNA

binding domain in binding studies. Experiments performed in the presence and absence of

ligand RNA established protein-RNA contacts as well as changes in burial upon binding.

52

Figure 1.20 Effects on structure and stability of a core substitution involving Trp43 for 5-

fluorotryptophan in GB1. a) Structure of GB1 with Trp43 highlighted, a sphere of 6 Å diameter

around the fluorine atom is shown in blue. b) 2D 1H-19F HOESY spectra of 5-FTrp labeled GB1

as a function of mixing time. Assignments for all NOE crosspeaks are indicated by residue type,

number and atom identifier on the far right hand side of the panels. Build-up curves were used

to determine proximity of the fluorine atom to neighbouring residues and, by comparison to the

53

know structure, the substitution of the fluorinated amino acid was shown to be non-perturbing

with respect to structure. © Reprinted from (Campos-Olivas et al. 2002) with permission from

A. M. Gronenborn.

1H-19F cross relaxation rates can also be used to assess solvent exposure or burial of

specific fluorinated sites, assuming that the cross-relaxation rates between 19F and water can be

distinguished from those between 19F and the protein interior. This requirement is achieved by

collecting two data sets; in one case proton signal associated with the water resonance is

saturated prior to direct fluorine detection, and in an analogous experiment, aliphatic proton

signals are saturated. This strategy was used to evaluate the relative burial of two tyrosine

residues in 3-fluorotyrosine labeled calmodulin in both the calcium-free and calcium-loaded

states (Kitevski-LeBlanc et al. 2009b). The relative magnitude of the NOE effect between 19F

nuclei and protons associated with water or aliphatics is easily delineated using 1D difference

experiments, and Y99 was found to be in greater dipolar contact with the protein interior, and

thus more buried than Y138 (Figure 1.21).

Figure 1.21 1H-19F 1D difference NOE spectra of 3-FTyr labeled calcium-free calmodulin and

calcium-loaded calmodulin. The NOE was initiated by saturation of either water (A,B) or

aliphatic protons (C,D) with the mixing times used indicated on the left margin of each spectrum.

54

The observed enhancements are larger for Y138 following saturation of water resonances in both

states of the protein, indicating it is more exposed to solvent. © Reprinted from (Kitevski-

LeBlanc et al. 2009a) with permission from Springer.

Finally, the use of 1H-19F heteronuclear NOE experiments in the evaluation of the relative

dynamics among fluorine probes was applied to 5-fluorotrypophan labeled D-lactate

dehydrogenase (Sun et al. 1996). By assuming that the limits of the heteronuclear NOE range

from 0.5, in the extreme narrowing limit, to -1.0 under immobile conditions, the authors

monitored changes in 19F NOE enhancements as a function of denaturant to obtain a measure of

relative mobility among the fluorotryptophan probes as the protein unfolds. As expected, a

general increase in mobility was observed as the denaturant concentration increased; however, a

probe in the C-terminus exhibited high mobility at low denaturant concentrations (1.75 M),

indicating the presence of an early unfolding intermediate with a highly denatured C-terminus.

1.5.4 Dynamics

Protein dynamics are characterized by nuclear relaxation, allowing for the extraction of

parameters that describe both the timescale of motion and, by invoking a model, a description of

the type of motion a given bond or protein segment is undergoing. In ideal situations, this

motion can be related to a biologically relevant process such as folding or catalysis. The ability

to extend this strategy to 19F labeled amino acids probes is relatively straightforward, and lacks

many complications due to unwanted coupling and relaxation pathways which can obscure

analysis of more common 1H, 13C and 15N nuclei, while providing complementary sidechain

information. Similar to other spin ½ nuclei, 19F relaxation time constants T1 and T2 are related to

the familiar spectral density equations as described by Abragam (Abragam 1961)

(8)

55

(9)

with and . In equations 8 and 9, γH and γF

represent the magnetogyric ratios of proton and fluorine respectively,

!

h is Planck’s constant,

!

rHF

is the internuclear 1H-19F distance,

!

BO is the magnetic field, are the components

of the traceless chemical shift tensor,

!

is the asymmetry of this tensor defined as

and unless otherwise stated J(ω) refers to the spectral density given in equation

7. Spin-lattice relaxation (T1) is measured using either 1D saturation recovery (Hull and Sykes

1974; Post et al. 1984) or inversion recovery (Evanics et al. 2007; Williams et al. 1997)

experiments, while spin-spin relaxation (T2) can be quantified using the CPMG experiment

(Carr-Purcell-Mieboom-Gill), classic Hahn echo pulse schemes (Kitevski-LeBlanc et al. 2009b)

or from the linewidth, assuming inhomogeneous contributions are small (Khan et al. 2006).

Relaxation time constants can be used to estimate the overall correlation time of the protein,

assuming that the side chain is held rigidly within the protein. For example, using 19F T1 values

and estimates of T2 from resonance line widths, the correlation times of hexokinase,

phosphoglycerate kinase, and pyruvate kinase labeled with 5-fluorotryptophan were determined

both in Saccharomyces cerevisiae cells and in purified form (Williams et al. 1997). The study

estimated tumbling times in-cell to be approximately twice that in water for hexokinase and

phosphoglycerate kinase, while signal from pyruvate kinase was not detected in-cell.

Fluorine resonance line widths provide a measure of conformational heterogeneity if one

assumes broadening originates from slow or intermediate fluctuations between states in which

the fluorine probe experiences a different chemical environment, and thus a slightly different

chemical shift. When exchange is slow relative to the difference in chemical shifts between the

two states, multiple peaks will be observed and their integrals reflect the relative populations of

the contributing conformations. In recent studies of the intestinal fatty acid binding domain

labeled with 4-fluorophenylalanine, line widths were broadened in the lipid bound state relative

56

to the free protein, indicating increased conformational flexibility in the bound state (Li and

Frieden 2006). A similar observation was made from a single 5-fluorotryptophan residue in the

histidine binding protein J complexed with L-hisitidine (Eichler et al. 2005a). A more

quantitative approach was taken in studies of GFP labeled with 3-fluorotyrosine (Khan et al.

2006). Here, line widths at two fields were used to calculate an order parameter, S2, using the

model-free spectra density function, given by Lipari and Szabo (Lipari and Szabo 1982) as

(10)

with

(11)

The order parameter S2 can range from 0 for fully isotropic motion to 1 for fully restricted

motion, and τi is the internal correlation time of the residue under consideration. The order

parameters of nine 3-fluorotyrosine probes fell within two ranges (Table 1.1); 0.21 ≥ S2 ≥

0.41and S2 ≥ 0.44. 3-FTyr residues with order parameters in the range 0.21 - 0.41 were found to

have the highest calculated HOMO (highest occupied molecular orbital) accessibility, while

probes with S2 ≥ 0.44 had negligible HOMO accessibilities. The correlation between increased

motional dynamics and solvent accessibility is reasonable, as the motion of a surface residue is

likely to be less restrained by neighbouring protein segments.

Aromatic ring-flipping is a fascinating example of large scale breathing motions,

originally observed by 1H NMR in basic pancreatic trypsin inhibitor (BPTI) (Wagner and

Wuthrich 1975). The use of asymmetric mono-fluorinated aromatics provides a means to

observe and quantify aromatic ring flipping in proteins directly from 1D 19F NMR spectra. In

the above mentioned studies of GFP, two fluorine resonances were identified for both Y92 and

Y143. This was presumed to reflect slow ring flipping for each aromatic (Figure 1.22). In this

case, the chemical shift differences were used to put upper limits on the rates of exchange which

were found to be between 160 -960 s-1.

57

Table 1.1 Dynamics of 3-fluorotyrosine probes in GFP. Chemical shifts, calculated order

parameter S2, solvent accessible surface area (SASA) calculated using a 0.14 nm probe and a

0.30 nm probe, and (highest occupied molecular orbital) HOMO accessibility.

Figure 1.22 Dynamics of 3-fluorotyrosine labeled

GFP. Assigned 564 MHz 19F spectra of 3-FTyr labeled

GFP. Shown are the experimental spectrum (top trace),

line-fitting (middle trace) and the individual lines in the

fitting (bottom trace). The two peaks associated with

Y92 and Y143 represent the shifts corresponding to

180°-ring flip conformers. © Reprinted from (Khan et

al. 2006) with permission from ACS.

Homonuclear 19F-19F NOESY experiments can also be used to monitor the dynamics of a

fluorine probe. This strategy was applied in a recent study of the intestinal fatty acid binding

protein, labeled with 4-fluorophenylalanine (Li and Frieden 2006). By monitoring phase

sensitive 19F-19F NOESY spectra as a function of pH, the authors observed increased exchange

phenomena for several probes under acidic conditions (Figure 1.23). Quantification of the

relative populations of major and minor conformations for a given probe allowed for the

determination of the exchange rate, which revealed millisecond timescale motion of the

phenylalanine sidechains.

58

Figure 1.23 Quantification of millisecond timescale dynamics in 4-fluorophenylalanine labeled

holo- and apo-IFAB. a) Phase sensitive 19F-19F NOESY spectra of holo-IFAB at indicated pH

values. A 250 ms mixing time was employed for all experiments. b) Population ratio of

exchange peaks and exchange rates of the two conformations observed at pH 2.8 for apo- and

holo-IFAB. © Reprinted from (Li and Frieden 2006) with permission from ACS.

In addition to dynamic information, 19F-19F NOEs can be used to establish proximity

between probes and relate this to secondary or tertiary structure. In the same study of IFAB

labeled with 4-fluorophenylalanine, 19F-19F NOESY spectra as a function of pH revealed

considerable amounts of tertiary structure, as evidence by NOE contacts between different β-

59

strands at pH 2.8. Moreover, interesidue contacts between probes located within the same

domain were shown to remain intact between pH 7.3 and 2.8 indicating that the increase in

hydrodynamic radius with pH, observed in diffusion experiments, originates from a pH

dependent increase in inter-domain separation.

1.6 Rationale

The application of 19F NMR in the examination of protein structure and dynamics is well

established, with examples in the literature ranging from studies involving small peptides to large

water and membrane soluble proteins. Despite the many advances made in several areas,

including the synthesis of novel fluorine probes, labeling strategies and developments in both

applications and instrumentation, there are still difficulties and limitations in fluorine NMR

approaches. In particular, we have focused on two areas: NMR-based resonance assignment and

the measurement of solvent exposure of fluorine probes in proteins uniformly labeled with

fluorinated aromatics.

To investigate the use of 19F NMR in the study of solvent exposure and binding, we

examined the interaction between a 3-fluorotyrosine labeled G48M mutant of the SH3 domain of

fyn tyrosine kinase with a binding peptide using a combination of solvent isotope shifts and

paramagnetic effects from TEMPOL and dissolved oxygen. The results of this study are

presented in Chapter 2.

To explore the possibility of multi-dimensional 19F NMR and its utility in assigning 19F

resonances, we synthesized 13C, 15N-3-fluorotyrosine (3-FTyr) using direct electrophilic aromatic

substitution. The large 19F-13C scalar couplings provide efficient magnetization transfer using

INEPT schemes, making the application of 2D or 3D NMR as straightforward as the

corresponding 1H-13C experiments. Once incorporated into a model protein system (calmodulin)

we developed a three-step assignment strategy allowing for unambiguous resonance assignment

without the need for mutagenesis (Chapter 3).

Having the assignments in hand, we used the same model system, 3-FTyr calmodulin, to

critically compare three approaches for the measurement of solvent exposure in proteins -

namely, the solvent induced isotope shift (SIIS), paramagnetic shifts from dissolved oxygen and

60

1H-19F heteronuclear NOEs originating from saturation of solvent and aliphatic protons (Chapter

4).

As commercially available fluorinated amino acids are not available with 13C, 15N-

enrichment, we recognized the need to develop a NMR-based assignment strategy, which relied

upon dipolar magnetization transfer that could be used in the absence of large heteronuclear

scalar couplings. Using calmodulin labeled uniformly with 3-fluorphenylalanine (3-FPhe) a

three step approach was used to assign all 8 resonances, as outlined in Chapter 5.

In both the literature and in our own experience, the presence of structural and functional

perturbations arising from the incorporation of fluorine probes has been observed to varying

degrees. Using 3-FPhe labeled calmodulin, we carefully examined the effect of uniform

incorporation on structure, dynamics and stability using NMR and CD. Gross perturbations

associated with structure and function were observed. We then evaluated the merits of fractional

labeling strategies in reducing the observed disruptions. We provide a general labeling scheme

to preserve native protein qualities and obtain high quality 19F NMR spectra in Chapter 6.

61

Chapter 2

19F NMR studies of solvent exposure and peptide binding to an SH3 domain

The work presented in this chapter was published in “Evanics, F., Kitevski, J. L.,

Bezsonova, I., Forman-Kay, J. D., (2007). 19F NMR studies of solvent exposure and peptide

binding to an SH3 domain. Biochimica et Biophysica Acta-General Subjects 1770, 221-230”.

My role in this project consisted of sample preparation, setting up NMR experiments and

processing data. NMR experiment set up, data processing and analysis was done with Dr. Ferenc

Evanics. Site-directed mutagenesis was performed with Dr. Irina Bezsonova. © Reprinted with

permission from Elsevier.

62

2.1 Abstract 19F NMR was used to study topological features of the SH3 domain of Fyn tyrosine

kinase for both the free protein and a complex formed with a binding peptide. Metafluorinated

tyrosine was biosynthetically incorporated into each of 5 residues of the G48M mutant of the

SH3 domain (i.e. residues 8, 10, 49 and 54 in addition to a single residue in the linker region to

the C-terminal polyhistidine tag). Distinct 19F NMR resonances were observed and subsequently

assigned after separately introducing single phenylalanine mutations. 19F NMR chemical shifts

were dependent on protein concentration above 0.6 mM, suggestive of dimerization via the

binding site in the vicinity of the tyrosine side chains. 19F NMR spectra of Fyn SH3 were also

obtained as a function of concentration of a small peptide (2-hydroxynicotinic-NH)-Arg-Ala-

Leu-Pro-Pro-Leu-Pro-diaminopropionic acid -NH2, known to interact with the canonical

polyproline II (PPII) helix binding site of the SH3 domain. Based on the 19F chemical shifts of

Tyr8, Tyr49, and Tyr54, as a function of peptide concentration, an equilibrium dissociation

constant of 18 ± 4 µM was obtained. Analysis of the line widths suggested an average exchange

rate, kex, associated with the peptide-protein two-site exchange, of 5200 ± 600 s-1 at a peptide

concentration where 96% of the FynSH3 protein was assumed to be bound. The extent of solvent

exposure of the fluorine labels was studied by a combination of solvent isotope shifts and

paramagnetic effects from dissolved oxygen. Tyr54, Tyr49, Tyr10, and Tyr8, in addition to the

Tyr on the C-terminal tag appear to be fully exposed to the solvent at the metafluoro position in

the absence of binding peptide. Tyr54 and, to some extent, Tyr10 become protected from the

solvent in the peptide bound state, consistent with known structural data on SH3-peptide

complexes. These results show the potential utility of 19F-metafluorotyrosine to probe protein-

protein interactions in conjunction with paramagnetic contrast agents.

2.2 Introduction The spectroscopic study of protein-peptide interactions ideally should utilize probes that

are non-perturbing but accurately reflect both the equilibrium and kinetic features of the

interaction. Fluorine NMR is an excellent technique for the study of such phenomena since the

isotopic fluorine label is generally weakly perturbing while the spectra are sensitive to changes in

van der Waals or electrostatic environments expected with binding. Furthermore, the resonances

of the bound and free states are often significantly different, providing an excellent dynamic

63

range for the study of equilibria and kinetics. 19F NMR is also amenable to approaches involving

the addition of paramagnetic contrast agents for purposes of studying solvent exposure and

possible protection from solvent exposure due to binding interactions. In this paper we explore

the use of metafluorotyrosine as an NMR probe of an interaction between the SH3 domain of

Fyn tyrosine kinase and a proline-rich peptide. Issues of 19F labeling and protein expression are

considered, while strategies for assignment and interpretation of data in terms of dynamics,

protein dimerization, protein-peptide interactions, and solvent exposure are provided.

2.2.1 Solvent Exposure

In protein structure studies, the determination of solvent exposure by spectroscopic

methods is a longstanding problem. For example, fluorescence or ESR experiments monitor

solvent exposure by the addition of appropriate water soluble additives which serve to quench or

broaden signals in proportion to the collisional frequency of the quencher (Haas et al. 1993;

Hubbell et al. 1998; Lehrer 1971; Pyka et al. 2005; Sonveaux et al. 1999). In NMR, solvent

exposure has been measured by monitoring signal from exchangeable groups upon substituting

water with D2O (Laurents et al. 2005; Olofsson et al. 2006), solvent NOESY or ROESY

schemes (Dalvit 1995; Dalvit et al. 1999; Dalvit and Hommel 1995b; Dalvit and Hommel

1995a), or relaxation experiments in the presence of soluble paramagnetic additives such as

Gd(III)DTPA-BMA (Pintacuda and Otting 2002) and TEMPOL (Niccolai et al. 2001; Niccolai

et al. 2003). In general for any spectroscopic method the measurement of solvent exposure is

complicated by the hydration shell and a vast range of exchange timescales of water with various

residues. Moreover, in situations where relaxation agents are used, the effect may explicitly

depend on the geometry and diffusion rate of the paramagnet and potential preferential

interactions with specific residues, local surface hydrophobicity, and local dynamics.

Solvent induced isotope shifts or changes in homonuclear and heteronuclear scalar

coupling constants resulting from the exchange of H2O with D2O have also been explored in

studies of solvent exposure in proteins. Although such effects are difficult to observe by 13C, 1H,

or 15N NMR (Bagno et al. 2005; Hansen 2000), solvent induced isotope shifts may be as large as

0.25 ppm in 19F NMR spectra, offering a unique way to probe solvent exposure (Danielson and

Falke 1996; Gerig 1994; Hull and Sykes 1976; Lian et al. 1994). Such experiments may be

corroborated by T1 measurements of the 19F probe, where it is commonly observed that a high

64

spin-lattice relaxation rate in D2O is indicative of burial in the hydrophobic protein interior

which serves as a significant source of dipolar relaxation. Intermolecular 1H-19F NOEs are also

efficient indicators of solvent exposure, particularly in cases where H2O is strongly bound

(Martinez and Gerig 2001). Although 19F NMR chemical shifts are extremely difficult to predict

based on environment (Frieden et al. 2004; Pearson et al. 1997), soluble paramagnets are useful

in 19F NMR studies of solvent exposure. One such paramagnet which exhibits pronounced

relaxation enhancement and chemical shift perturbations is molecular oxygen (Prosser et al.

2000; Teng and Bryant 2000). In a recent study of a membrane protein in detergent micelles,

oxygen-induced chemical shift perturbations from specifically fluorinated sites of a

transmembrane (TM) domain belonging to a polytopic TM protein provided information on

secondary structure, helix-helix interfaces, and immersion depth (Luchette et al. 2002).

Despite the sensitivity arising from the use of dissolved oxygen as a contrast agent,

preferential interactions and local dynamics of both the additive and protein complicate analysis

of effects in terms of solvent exposed surface areas (Teng et al. 2006). Ideally, solvent exposure

might be better assessed by separately measuring the effects of a hydrophilic and hydrophobic

contrast agent, whose sizes and mobilities are comparable. Here we define a contrast agent as

any additive (paramagnetic or otherwise) which affects a change in chemical shift or relaxation

properties of labels of interest in a spatially dependent manner. Thus, in the context of 19F NMR,

D2O can be thought of as a contrast agent since its substitution for H2O causes local peak shifts

proportional to solvent exposure. ESR studies have used complementary paramagnetic contrast

agents such as O2 and Ni(II) to separate steric effects and partitioning effects in proteins

(Altenbach et al. 1994; Hubbell et al. 1998). We extend this idea to 19F NMR which has the

distinct advantage that it is sensitive to two simple contrast agents, oxygen and H2O/D2O, whose

sizes are similar while oxygen is hydrophobic and water is polar. Effects from oxygen are easily

measured through chemical shift perturbations and spin-lattice relaxation rates, while contrast

effects from water may be measured through 1H-19F NOEs or solvent induced isotope shifts,

resulting from exchanging H2O with D2O.

2.2.2 Fyn tyrosine kinase and SH3 domains

Fyn tyrosine kinase plays a key role in signal transduction processes. The protein’s

catalytic domain is adjacent to two small domains (the so-called Src-homology regions 2 and 3

65

or SH2 and SH3 domains) which modulate signal transduction events in the cell through binding

interactions (Koch et al. 1991; Musacchio et al. 1992). The 59 residue SH3 domain is present in

many eukaryotic proteins which are involved in signal transduction, cell polarization and

membrane-cytoskeleton interactions. The SH3 domain often functions as an intramolecular

regulator of kinase activity, while serving as a targeting moiety to direct Src kinases to the proper

intracellular sites. The Fyn SH3 domain shows the typical SH3 topology consisting of five β-

strands as shown in Figure 2.1. A binding surface surrounded by the n-Src and RT-loops consists

predominantly of aromatic residues which are known to interact with proline-rich targets

(Musacchio et al. 1994b; Musacchio et al. 1994a). Tyrosines figure prominently in this binding

surface, including tyrosines 8, 10, and 54 in Fyn, which are highly conserved among SH3

domains (Larson and Davidson 2000). The SH3 domain possesses a fourth tyrosine residue at

position 49, somewhat removed from the binding pocket. Proteins or peptides interacting with

the Fyn SH3 domain usually exhibit dissociation constants in the range of 1-10 micromolar for

binding consistent with transient interactions, as might be expected for regulation of signaling.

However, synthetic peptides have been designed to bind to the Fyn SH3 domain with nanomolar

dissociation constants (Li and Lawrence 2005). This paper explores topology of the SH3 domain

of Fyn and subsequent changes upon peptide binding using 19F NMR. We make use of a well

studied destabilized mutant version of the Fyn SH3 domain (i.e. G48M) which exists in fast

equilibrium with an unfolded state via a well-defined intermediate under native conditions at

room temperature (Di Nardo et al. 2004; Korzhnev et al. 2004). However, the work discussed in

this paper was performed under low temperature conditions (10 ºC) where the unfolded state is

very weakly populated (< 1%) and the folded state dominates. Via fluorinated tyrosine probes,

local side chain mobility, inter-residue contacts through NOEs, solvent exposure, and the

ensuing changes upon addition of the binding peptide are considered. Details of the interaction

between the peptide and the protein are reliably obtained in terms of the binding equilibrium

constant and effective exchange rate by monitoring the behavior of the chemical shifts and line

widths, as a function of peptide concentration.

66

Figure 2.1 Ribbon diagram of the Fyn SH3 domain showing the location of the four

metafluorotyrosines which sit in the peptide binding groove.

2.3 Theory

2.3.1 Solvent Exposure

In describing the chemical shift perturbation or relaxation rate enhancement arising from a

diffusible paramagnet such as dissolved oxygen, the key variables include the collisionally

accessible exposed surface area and the local concentration of the contrast agent, which has been

shown to depend on surface hydrophobicities (Teng et al. 2006). Ideally, the paramagnetic

shift,2O

! , or relaxation rate, PR1 , should be normalized by dividing the observed effect (shift or

rate) by the equivalent result for a fully exposed probe. For example, Fyn SH3 possesses a

tyrosine residue on the C-terminal polyhistidine tag, which is expected to exhibit the largest

degree of solvent exposure. Thus PR1 or 2O

! acting on Tyr8, 10, 49, or 54 of Fyn SH3 can be

normalized by dividing the result by that observed for the Tyr probe on the exposed tag. We may

similarly define the normalized version of the chemical shift perturbation due to the solvent

isotope effect. Though this is not of paramagnetic origin, we nevertheless expect the normalized

solvent isotope shift to depend on local solvent accessibility and local hydrophobicity. A ratio of

the normalized paramagnetic effect from oxygen to the equivalent normalized solvent isotope

effect should then reflect the local hydrophobicity, assuming accessibilities of water and oxygen

are similar.

67

Herein, we measure the 19F paramagnetic shifts and relaxation rate enhancement arising

from dissolved oxygen in addition to solvent isotope shifts of Fyn SH3 enriched with

metafluorinated tyrosine, in order to better address solvent exposure and hydrophobicity of the

tyrosine side chains. We then consider the changes in solvent exposure and side chain dynamics

after titrating a target peptide to the Fyn SH3 domain. The analysis of the ensuing chemical shift

and line width changes, in terms of binding and dynamics is discussed below.

2.3.2 Protein-Peptide interactions

19F NMR is ideally suited to the study of peptide binding in part because very large

chemical shift perturbations and relaxation rate changes are often observed upon binding.

Furthermore, a global fit of all chemical shift titration curves and relaxation rate analyses from

multiple sites in the binding pocket should result in a single dissociation constant and exchange

time. Assuming that the resonances associated with both the free protein and fully saturated

bound protein are distinct, we may describe the dependence of the chemical shift, δ, on peptide

concentration, [p], by

dKp

pA

+!

=][

][ , [1]

where A is a fitting constant and Kd defines the equilibrium dissociation constant. In the event of

intermediate timescale exchange, line broadening can be slightly more complicated. However, if

we assume for the moment that the bound state is much more populated over the concentration

range considered, the transverse relaxation rate associated with the bound peak may be given as

(Leigh 1971; Palmer 2004; Woessner 1961)

2/12/12221

222220221 }]16)[({

81

2 exexexex kppkkkRR !!! "#"++"##+= , [2]

where we envisage a two-site exchange between a bound state, 1, and a free protein state, 2. Δω

represents the difference in radial frequency units between the fully bound and fully free

resonances, 0

2R represents the population averaged relaxation rate in the absence of exchange,

and the overall sum of exchange rates, kex, between sites 1 and 2 is given by

68

2112kkk

ex+= . [3]

The relative fraction of the complexed and free states of Fyn SH3 may further be expressed in

terms of the forward and reverse rate constants as exkkp /211

= and exkkp /122 = . Thus the

strategy in determining the equilibrium dissociation constant is to monitor the chemical shifts of

resonances of the tyrosine residues involved in binding as a function of peptide concentration.

The effective exchange rate, kex, may be similarly analyzed as a function of peptide

concentration.

2.4 Materials and Methods

2.4.1 Sample Preparation

A plasmid coding for the isolated G48M mutant of the isolated Fyn SH3 domain, residues

85-142, cloned in such a way that 5 additional residues (MVQIS) were added to the N-terminus

and 16 (RLDYKDDDDKHHHHHH) added to the C-terminus, was transfected into E. coli strain

BL21(DE3) under the control of the T7 promoter. Expression of the 15N labeled protein was

induced for 3 hrs at an OD600 of 0.8 by addition of 250 mg/L IPTG to bacterial growths at 37 °C

in M9 minimal medium, supplemented with 0.3% D-glucose, 0.1% 15NH4Cl, 100 mg/L

ampicillin, 10 mg/L thiamine, 10 mg/L biotin, 1 mM MgSO4 and 1 mM CaCl2. Uniform

fluorotyrosine labeling was achieved by introducing glyphosate (1 g/L), phenylalanine (50

mg/L), tryptophan (50 mg/L), and 19F labeled metafluoro (L, D) tyrosine (70 mg/L) (Sigma

Chemicals Mississauga, ON) to the bacterial culture one hour before induction (OD600 ~0.4).

Unlike prior methods (Sykes et al. 1974), no auxotroph was found to be necessary to obtain

reasonable yields (24 mg/L). Cells were harvested 3 hours after induction by centrifugation, and

lysed by sonication in 0.1M NaH2PO4, 0.01M TrisHCl, 6M guanidine hydrochloride (GdmCl),

and 20mM imidazole. The Fyn SH3 domain was purified on a nickel ion exchange column at

25°C under denaturing conditions (6M GdmCl) then refolded by dialysis against 10 mM

TrisHCl, 0.2 mM EDTA, and 250 mM KCl. For NMR measurements, the protein concentration

was stabilized in 20 mM TrisHCl buffer. Four mutants in which phenylalanine was substituted

for tyrosine at position 8, 10, 49 or 54 were made using a Quick Change Kit (Stratagene, CA,

USA). Protein expression, purification, and NMR sample preparation of the mutants was

performed as described above. Protein concentrations were determined by absorbance

measurements, using a known extinction coefficient of 60.5 µM-1cm-1. The concentration of the

69

“wild type” G48M Fyn SH3 domain sample, which was used for the bulk of the solvent exposure

and peptide binding studies, was estimated to be 0.4 mM which was found to be below the

threshold for oligomerization. The phenylalanine mutant concentrations ranged from 0.61 to 0.69

mM.

2.4.2 NMR Experiments

A 400 µl sample volume was deemed sufficient for shimming purposes, using a 5 mm

OD, 3 mm ID sapphire NMR sample tube (Saint Gobain – Saphikon Crystals, Milford, NH,

USA), designed to tolerate pressures as high as 270 bar. To measure effects of dissolved oxygen,

the sample was first equilibrated at 10ºC outside the magnet at an oxygen partial pressure of 20

bar for 2 days, then equilibrated overnight in the magnet at the desired partial pressure of 20 bar.

Using open Swagelok (Swagelok, Solon, OH, USA) connections to a pressurized oxygen supply,

it was possible to maintain the pressure during the entire course of the NMR experiment. To

reliably reproduce oxygen concentrations, we relied on the measurement of the 1H T1 of water

which was typically 100 ms at 20 bar (2O

P ). Upon completing oxygen experiments, the sample

was degassed by first transferring it to a 1.5 ml microfuge tube resting on an ice bath, after which

it was slowly stirred using a sterile needle tip which precipitated the bubbling of oxygen. The

sample was then left for 24 hours in a nitrogen environment to allow for residual degassing. To

exchange H2O for 2H2O buffer, a 0.5 ml centrifugal concentrator with a molecular weight cutoff

of 3 kDa was used. A significant amount of protein (30-40 %) was lost upon solvent exchange

and transfer back to the original sapphire NMR tube, resulting in a need for greater signal

averaging in the 2H2O sample. Finally, 0.008 mgs (~0.1mM) of 4-hydroxy-TEMPO (Sigma

Chemicals, Mississauga, ON), also known as TEMPOL, was added to the sample to measure

contrast effects through T1 from a dissolved paramagnet.

1H,15N gradient selected HSQC and 19F one-dimensional NMR experiments were performed

at 10ºC on a 600 MHz Varian Inova spectrometer, using a standard 5 mm HCN triple resonance

single gradient solution NMR probe, in which the high frequency channel could be tuned to

either 19F or 1H. 8 scans and 80 increments spanning 1650 Hz in the indirect dimension were

used to obtain the HSQC spectrum while 512 scans were typically used to obtain the 19F NMR

spectrum. The measurement of 19F spin-lattice relaxation times was accomplished by an

inversion recovery sequence (i.e. 180 – τ – 90) using a total of 8 τ values, logarithmically spaced

70

between 10 ms and 5 s (ambient sample) or 1 ms and 1 s (oxygenated sample). The repetition

time was adjusted to either 6.5 s (ambient sample) or 1.5 s (oxygenated sample). Two Hahn echo

refocusing pulses, spaced by 350 µs, were appended to all 19F NMR sequences, to help remove

background 19F signal from the probe. One-dimensional homonuclear 19F NOESY experiments

were performed using mixing times of 100 ms, 250 ms, 500 ms, and 1000 ms in order to measure

interactions between the tyrosine labels in the protein. In these experiments, each distinct 19F

resonance (in addition to a dummy frequency several hundred Hz from the nearest resonance)

was separately saturated via an appropriate low power CW pulse during the mixing time and the

subsequent signal was then compared to an equivalent off-resonance saturation. NOE intensities

were approximately 20% the total intensity of the individual unsaturated resonances at a mixing

time of 500 or 1000 ms and roughly 10% the total intensity at a mixing time of 100 ms. A two-

dimensional homonuclear 19F NOESY experiment was also performed using a mixing time of

250 ms (data not shown). An 15N-edited two-dimensional (1H,1H) NOESY was also performed to

help assign HSQC peaks in the fluoro-tyrosine substituted protein.

2.5 Results and Discussion

2.5.1 Effect of introducing metafluorotyrosine at residues 8, 10, 49, and 54

Figure 2.2 compares the (1H,15N) HSQC NMR spectrum of 15N enriched Fyn SH3 (blue

contours) with that of the fluorotyrosine substituted version (red contours).

71

Figure 2.2 (1H,15N) HSQC NMR spectrum of 15N enriched Fyn SH3 domain (shown in blue)

overlayed with the equivalent spectrum of the Fyn SH3 domain in which all five fluoro tyrosines

were replaced with metafluoro tyrosine (shown in red). Both spectra were acquired at 10º C. The

circled peaks indicate the resonances associated with each of the 4 tyrosine resonances, which

can only be seen in the control spectrum (i.e. fluoro tyrosine-free).

Note that there is no detectable 15NH tyrosine signal from the fluorotyrosine substituted species,

suggesting that 19F-tyrosine incorporation was at least 95%, based on the signal to noise ratio in

the HSQC spectrum. Furthermore, incomplete biosynthetic labeling would be expected to give

rise to a multiplicity of 19F NMR peaks since the tyrosine residues are quite close to each other in

Fyn SH3. The two spectra of the 15N-enriched fluorinated Fyn SH3 domain and 15N-enriched

Fyn SH3, which were obtained at 0.4 mM concentrations, are sufficiently similar to conclude

that the overall folds are the same. However, there are modest differences in chemical shifts for a

significant number of residues. Fluorination in aromatic amino acids is known to result in a

redistribution of partial charge across the ring, a decreased hydrogen bond capacity (of the

72

fluorine atom) and a slight increase in the van der Waals radius in the vicinity of the CF group

(Minks et al. 1999) while more significant effects may occur in the case of perfluorination of

aromatic groups (Adams et al. 2001). Such changes may give rise to slight structural

perturbations and possibly large functional perturbations in the case of enzymes where tyrosine

directly participates in the in the reaction mechanism (Brooks and Benisek 1994; Hazard et al.

1992; Hull and Sykes 1976; Wacks and Schachman 1985).

2.5.2 Assignment of 19F resonances and side-chain interactions

Figure 2.3 presents the 19F NMR spectrum of the fluorotyrosine substituted Fyn SH3

domain for the “wild-type” G48M and the four Tyr-to-Phe mutants used for the assignment. All

spectra were acquired at 10°C under similar conditions and at a protein concentration of

approximately 0.6 mM, with the exception of the top most spectrum which was obtained from a

0.4 mM sample of the fluorotyrosine substituted Fyn SH3 domain, and which contained a small

amount of free metafluorinated tyrosine intended as a chemical shift reference. All five

resonances associated with fluorotyrosine sites on the G48M mutant of the Fyn SH3 domain (i.e.

Tyr8, Tyr10, Tyr49, Tyr54, and the Tyrosine on the C-terminal tag) can be clearly identified.

Note that the sharp central peak at -137.32 ppm arises from the fluorotyrosine label on the C-

terminal tag. This label potentially serves as a useful internal chemical shift and relaxation rate

reference, both for the study of solvent exposure via solvent isotope shifts or paramagnetic rates

arising from dissolved oxygen or other paramagnetic additives (vide infra). The spectrum of the

Fyn SH3 domain obtained under dilute conditions exhibits the greatest chemical shift dispersion

and this is particularly true of the Tyr54 and Tyr10 resonances which appear to exhibit a shift

variation of approximately 0.5 ppm depending on protein concentration. For example, the upfield

resonances associated with Tyr10 in Figure 2.3 appear to adopt a shift of either -138.25 ppm or -

138.75 ppm and in most cases, there appears to be a coexistence of the two resonances which

depends on concentration. Similarly, Tyr54 exhibits chemical shifts between -137.25 ppm when

concentrated and -136.75 ppm under more dilute conditions. This dramatic concentration

dependence of chemical shifts is suggestive of the formation of a dimer of the G48M mutant of

FynSH3, mediated by the fluorotyrosine sidechains.

73

Figure 2.3 19F NMR spectra of the metafluoro tyrosine substituted Fyn SH3 domain at 10°C.

The top spectrum represents the fully substituted species at a concentration of 0.4 mM and

consisting of identical 19F isotopic labels at Tyr8, Tyr10, Tyr49, Tyr54, and the tyrosine in the C-

terminal tag. A small amount of free metafluoro tyrosine was added as an internal control for the

solvent exposure studies. The spectrum labeled “conc” represents a more concentrated version of

the Fyn SH3 domain (0.6 mM) where the dimeric state is believed to dominate. Additional 19F

NMR spectra in the figure obtained under identical conditions, represent single phenylalanine

mutants, as labeled, whose concentrations ranged from 0.6 mM and 0.7 mM.

Single tyrosine to phenylalanine mutations (Y54F, Y49F, Y10F, and Y8F) appear to alter

the equilibrium between dimer and monomer, which we tentatively associate with the spectrum

of Fyn SH3 at 0.4 mM or less (top most spectrum of Figure 2.3). Selective saturation of the

major peak of Y10 in the top most spectrum of Figure 2.3, results in the gradual decrease in the

minor peak. Based on the steady state value of the minor peak and accounting for the (identical)

T1 of both peaks, we estimate the forward rate constant associated with the formation of the

dimer to be 1.16 s-1 (Szantay and Demeter 1995). However, the majority of the mutants (Y8F,

Y10F, and Y49F) give rise to a spectrum which we attribute to a dimeric state of FynSH3 at the

concentrations used in these experiments (i.e. 0.6 mM or higher). There appears to be a network

of interactions between the tyrosine side chains, which are perturbed by the introduction of even

a single phenylalanine residue. The origin of these side-chain interactions is not exclusively

74

intramolecular, as evidenced by the significant concentration dependence associated with the

chemical shifts of Tyr10 and Tyr54. No such dimerization or clustering is known to occur with

non-fluorinated versions of the Fyn SH3 domain or the G48M mutant of the FynSH3 domain,

suggesting that the fluorinated side chains are responsible for the dimerization of the Fyn SH3

domains, at least in the absence of a binding peptide. Clusters of fluorinated residues are well

known to exhibit a so-called fluorophobic effect which is stronger than hydrophobic forces alone

in directing association of fluorinated interfaces (Lee et al. 2006; Lee et al. 2004; Naarmann et

al. 2006; Tang et al. 2001b). This property has been applied previously in de novo protein

design, where biosynthetically fluorinated species introduced to the hydrophobic core have

proven to render the protein more resistant to heat denaturation (Lee et al. 2006; Lee et al.

2004). In the case of the Fyn SH3 domain, all 5 fluorotyrosines are located on the protein

exterior and at least three of these residues reside along a single surface (Fig. 1). Therefore, at

concentrations of 0.6 mM or higher, the protein is believed to form a transient dimer as

evidenced by the significant concentration dependent chemical shift perturbation associated with

Tyr 10 and Tyr 54. Though a thorough series of spectra as a function of concentration was not

obtained, we estimate that 0.4 mM represents the threshold concentration above which a dimeric

state is obtained. The lack of any significant changes in the spectrum above 0.6 mM suggests that

higher order oligomerzation states are not obtained, which is consistent with the notion that the

dimerization interface exists on the Fyn SH3 polyproline II interacting interface and is mediated

by the fluorotyrosine side chains. Evidence that Tyr54 is necessary for dimerization is also found

in the spectrum of the Y54F mutant, shown in Figure 2.3, which we have assigned as a

monomeric state, as evidenced by the chemical shift of Tyr10.

An extensive network of inter-side chain interactions is also seen through 19F,19F NOESY

measurements, where the NOE intensities are given in Figure 2.4. All NOEs were obtained by

separately irradiating each of the above resonances for 100 ms and observing the change in peak

intensity with respect to an off resonance irradiation. Here, the magnitude of the NOE is

indicated by the thickness of the line, while solid and dashed lines indicate negative and positive

NOEs, respectively. Residues 8, 54, and 49, which seem to lie along a groove in the Fyn SH3

domain, all appear to establish prominent NOE contacts, while the positive NOEs to residue 10

may arise either from a relayed NOE effect or from greater mobility of the Tyr10 side chain.

Note that the figure is meant to serve as a guide of approximate positioning of the tyrosine

75

residues since a slight variation in the side chain torsion angles can significantly alter the relative

separation of the fluorine nuclei. Surprisingly, there are significant NOE contacts between all of

the residues and the fluorotyrosine residue on the C-terminal tag. A significant fraction of these

NOE contacts may arise from intermolecular effects, established through dimerization, and a

thorough analysis of cross peak intensity as a function of mixing time and concentration should

provide an estimate of the equilibrium constant and exchange time.

Figure 2.4 19F homonuclear NOESY amplitudes obtained for a mixing time of 100 ms, upon

selective saturation of each resonance. Negative NOEs are indicated by a solid line while

positive NOEs are indicated by a dashed line. The line thickness corresponds with the relative

magnitude of the NOE, while the source of each arrow indicates the residue which was saturated.

Note that the tyrosine in the C-terminal tag is designated as C-term in this figure.

2.5.3 Solvent exposure of the fluorotyrosine residues for the free Fyn SH3 domain

Table 2.1 presents the results of several 19F NMR experiments designed to address the

extent of solvent exposure for the fluorotyrosine side chains of the Fyn SH3 domain. Note that in

D2O, where dipolar relaxation arises only from the protein, 19F spin-lattice relaxation times

associated with residues 8, 10, 49, and 54 and the C-terminal tag range between 0.9 s and 1.1 s.

76

A small R1 in D2O is indicative of fewer protein contacts and, thus, greater solvent exposure

(Hull and Sykes 1976). It is not surprising that all tyrosine relaxation rates fall in such a narrow

range in the absence of binding peptide, since none are buried in the hydrophobic core of the Fyn

SH3 domain.

Table 2.1 Chemical shifts, line widths, and relaxation rates and associated changes upon

substitution of H2O with D2O or addition of dissolved oxygen for all fluorotyrosine resonances in

both the free protein and protein saturated with binding peptide.

Note that ΔδDHO-H2O represents the solvent induced isotope shifts upon substitution 90/10

H2O/D2O with 50/50 H2O/D2O while ΔδD2O-DHO represents the solvent induced isotope shifts

upon substituting 50/50 H2O/D2O with 100% D2O.

Figure 2.5A shows a series of 19F NMR spectra of the metafluorotyrosine enriched Fyn

SH3 domain as a function of H2O/D2O, where the corresponding shifts are given in Table 2.1. To

enable assignments to be followed, the 19F NMR spectrum was obtained in mostly H2O (10%

D2O), 50% D2O, and 100% D2O. The solvent isotope shift could then be determined from the

difference of the chemical shifts associated with the 50 % D2O and 10% D2O sample

(i.e. OHHDO 2! # ) and between 100% D2O and the 10% D2O mixture (i.e. OHOD 22 ! # ). Ideally the

isotope shift, OHOD 22 ! # , should be proportional to OHHDO 2! # , unless differences in

conformation arise for high D2O fractions. As can be seen from Table 2.1, the two shifts

( OHOD 22 ! # and OHHDO 2! # ) are proportional to each other within a 25% margin with the

exception of the solvent isotope shift associated with Tyr10. The solvent isotope shifts

( OHOD 22 ! # ) associated with Tyr8, Tyr49, and Tyr54 are comparable to those observed for the

77

fluorotyrosine label on the polyhistidine tag from which we conclude that these three residues are

significantly exposed to the solvent at the meta position on the ring, while the isotope shift

associated with Tyr10 is actually positive, indicating some degree of solvent protection and/or a

conformational change accompanying the isotope shift. This anomalous isotope shift is not

surprising since the Tyr10 chemical shift is known to vary over a range of 0.5 ppm in H2O,

depending on the protein concentration. The delicate equilibrium between monomer and dimer

may also be affected slightly by a substantial change in the H2O/D2O ratio. Earlier, we attributed

the two upfield peaks associated with Tyr10 to a dimeric and monomeric state. At low D2O/H2O

(bottom trace in Figure 2.5A) the outer most peak associated with the monomer appears slightly

larger.

A)

78

B)

Figure 2.5 A) 19F NMR spectra of the metafluoro tyrosine substituted Fyn SH3 domain at 10°C

in 10%, 50%, and 100% D2O. The peaks assigned as TAG is associated with a tyrosine residue

located in the C-terminal tag added during cloning. B) 19F NMR spectra of the metafluoro

tyrosine substituted Fyn SH3 domain at 10°C in H2O under ambient conditions (0.20 atm PO2)

and upon equilibration at a partial pressure of 20 atm oxygen.

Soluble contrast agents also help to address solvent exposure. Figure 2.5B shows the

result of dissolved oxygen on the 19F NMR spectrum which shows significant chemical shift

perturbations and line broadening. The oxygen induced shifts,2O

! , and relaxation

enhancements, )( 1,11 2RRR O

P ! , are given in Table 2.1. Within a 20% margin, both measures of

local oxygen content consistently reveal that meta positions of the fluorotyrosine side chains are

fully exposed to the solvent though the paramagnetic effects from oxygen are measurably lower

at the tag. The reduced effect on the tag is likely a result of local decrease in hydrophobicity

rather than reduced solvent exposure. For example, the ratio of the paramagnetic rate arising

from dissolved oxygen to the solvent isotope shift (i.e. OHODPR

22/1 ! # ) should give an idea of the

relative hydrophobicity in the vicinity of each tyrosine. For Tyr54, Tyr49, and Tyr8, this ratio is

16.5, 18.8 and 17.8 s-1 ppm-1, while that associated with the tyrosine at the polyhistidine tag is

79

12.5 s-1 ppm-1, implying a more hydrophilic environment, which we expect in the vicinity of the

charged polyhistidine sequence. A reliable estimate of local hydrophobicity associated with

Tyr10 cannot be obtained since the isotope shifts were noted above to be anomalous. Thus, we

conclude based on relaxation rates in D2O, solvent isotope effects, and paramagnetic shifts and

relaxation effects from dissolved oxygen that Tyr54, Tyr49, Tyr10, and Tyr8 fluorotyrosine

residues are approximately equivalently and fully exposed to solvent in the absence of a binding

peptide.

2.5.4 Effect of a binding peptide on the fluoro-Fyn SH3 domain

The Fyn SH3 domain is known to bind preferentially to PPII helices (Morton et al. 1996;

Renzoni et al. 1996) with typical dissociation constants in the micromolar range, as expected

with its role in cell signaling. Recently, Li et al have employed combinatorial strategies to design

peptide ligands that bind to the Fyn SH3 domain with dissociation constants nearly one

thousand fold lower than those found in vivo (Li and Lawrence 2005). Since the Fyn SH3

domain plays a role in T cell activation, such ligands have a therapeutic use in regulating T cell

activation or altering the behavior of other Src kinases (Li and Lawrence 2005). In this study,

we examine the effect on the 19F NMR spectra of the Fyn SH3 domain upon titrating an N-

substituted peptide referred to as peptide 11a (Li and Lawrence 2005), having the sequence (2-

hydroxynicotinic-NH)-Arg-Ala-Leu-Pro-Pro-Leu-Pro-Dap-NH2, where Dap represents (L)-2,3-

diaminopropionic acid. Peptide 11a binds to the wild type Fyn SH3 domain with a 110 nM

dissociation constant (Li and Lawrence 2005). Figure 2.6A represents the 19F NMR spectra as a

function of increasing concentration of peptide. Three tyrosine residues seem to be involved in a

strong interaction with 11a based upon the pronounced chemical shift perturbations (i.e. Tyr8,

Tyr49, and Tyr54). Tyr8, Tyr10, and Tyr54, which are known to be most conserved amongst

SH3 domains also exhibit pronounced line broadening upon addition of binding peptide. Upon

saturating with 11a, these residues exhibit line broadening of 36.73 Hz, 14.0 Hz, and 65.3 Hz

respectively. Although Tyr10 does not exhibit a sizeable change in chemical shift upon addition

of peptide, the chemical shift dependence on peptide concentration associated with Tyr54,

Tyr49, and Tyr8 can be easily fitted to Equation 1, providing a uniform estimate of the

dissociation constant (i.e. 12 ± 2 µM, 22.8 ± 0.3 µM , and 19.8 ± 0.3 µM respectively, for which

we assign an average dissociation constant of 18 ± 4 µM). The G48M mutant has been known to

80

increase the dissociation constant by a factor of three in studies of binding peptides with

micromolar dissociation constants. Even if we attribute a factor of three to the G48M mutant,

this suggests that the metafluorine labels reduce the binding efficacy to the peptide by a factor of

roughly thirty. Steric perturbations resulting from the introduction of fluorine atom and

consequent changes in partial charge associated with the tyrosine ring may account for the

reduced binding to the peptide, while competition with dimerization may also be a factor.

The line broadening seen with increasing peptide concentration as shown in Figure 2.6A

can be interpreted in terms of a simple two site exchange model in which the Fyn SH3 domain

interconverts between a bound and free state (i.e. state 1 and 2, respectively). In the presence of

peptide 11a, the equilibrium strongly favors the FynSH3 domain:11a complex (i.e. p1 >> p2) in

which case we make use of Equation 2 to analyze the line broadening as a function of peptide

concentration. Here, we estimate the relaxation rate, R21, from the line width, 2/1! , according

to 2/121 ! # =R , where we assume each resonance is represented by a Lorenztian line. The

resonances associated with Tyr54 and Tyr8 are particularly useful to evaluate the effective

exchange rate, kex, since the chemical shift differences (Δω) between the free and bound state are

9050 s-1 and 4630 s-1, respectively. In our analysis, the line broadening data from both residues

was used to obtain a global estimate of kex as a function of peptide concentration. At the

maximum peptide concentration range used, the bound fraction, p1, was estimated to be 0.96,

while the effective exchange rate was determined from line broadening to be 5200 ± 700 Hz (i.e.

4500 Hz and 5850 Hz for Tyr 8 and Tyr 54, respectively).

81

A)

B)

Figure 2.6 A) 19F NMR spectra of the metafluoro tyrosine substituted Fyn SH3 domain at 10°C

as a function of binding peptide concentration. The Fyn SH3 domain concentration was

approximately 0.4 mM in these experiments. B) Comparison of 19F NMR spectra of the

metafluoro tyrosine substituted Fyn SH3 domain at 10°C, saturated with binding peptide under

ambient conditions (0.2 atm PO2) and upon equilibration at an oxygen partial pressure of 20 atm.

82

2.5.5 Solvent exposure of the fluorotyrosine residues in the peptide: Fyn SH3 complex

The extent of solvent interaction upon saturating the Fyn SH3 domain with the binding

peptide may be reliably measured by considering the effect of dissolved oxygen. The spectra of

the Fyn SH3 domain in the presence of binding peptide at either ambient oxygen concentrations

or under an oxygen partial pressure of 20 bar are shown in Figure 2.6B. In contrast to the

measurements performed for the free protein, where all residues exhibited a comparable degree

of solvent exposure, the paramagnetic shifts and relaxation rate enhancements vary significantly

amongst the four fluorotyrosines, as shown in Table 2.1. One way of assessing the difference in

solvent exposure between the free and complexed peptide states is to consider the ratio of

paramagnetic shifts or relaxation rate enhancements from dissolved oxygen, which we express as

)(/)(22freecmplx OO !! and )(/)( 11 freeRcmplxR PP . This ratio of shifts or relaxation rates

reveals that Tyr54 and Tyr10 become significantly more buried in the presence of the binding

peptide, while the meta position of the other fluorotyrosines appears to undergo little change in

solvent exposure.

2.6 Conclusions

The G48M mutant of the Fyn SH3 domain was biosynthetically labeled with

metafluorinated tyrosine, giving rise to 5 resonances in the 19F NMR spectra, associated with

Tyr8, Tyr10, Tyr49, Tyr54 and a tyrosine on the C-terminal polyhistidine tag. Assignments were

performed by comparing 19F NMR spectra with those of single phenylalanine mutations. The

free protein was characterized by a monomeric and dimeric state which was suggested to arise

from a fluorophobic effect. The monomeric state was found to dominate at protein

concentrations of 0.4 mM or less. Solvent exposure of the various fluorinated residues was

assessed by solvent isotope shift effects and by paramagnetic shifts and relaxation rates arising

from dissolved oxygen at a partial pressure of 20 bar. In the free state, Tyr54, Tyr49, Tyr10, and

Tyr8, in addition to the tag appear to be fully exposed to the solvent. The addition of a proline-

rich peptide, (2-hydroxynicotinic-NH)-Arg-Ala-Leu-Pro-Pro-Leu-Pro-diaminopropionic acid -

NH2, causes Tyr 54 and Tyr10 to be significantly less exposed to the solvent. By monitoring the

chemical shift of Tyr8, Tyr49, and Tyr54 as a function of binding peptide, the equilibrium

83

dissociation constant, Kd, for the interaction between the peptide and the G48M mutant of the

Fyn SH3 domain, was estimated to be 18 ± 4 µM. This suggests that the introduction of

metafluorinated tyrosine weakens the binding interaction with the peptide. An analysis of line

widths with peptide concentration provided an estimate of the effective exchange rate, kex,

between free and bound states to be on the order of 5200 Hz at a peptide concentration where

96% of Fyn SH3 is complexed with the peptide.

The above experiments revealed that the resonances associated with the free and bound

peptide state states were characterized by differences of 3-9 kHz which is not uncommon in 19F

NMR. Traditional protein protein interaction kinetics and equilibria are often performed by

fluorescence measurements. However, in situations where the binding pocket contains one or

more tyrosine residues, the above measurements demonstrate the potential utility of 19F NMR as

a probe of binding. Moreover, the considerable difference between the free and bound states

mean that exchange rates might be reliably measured by simple CPMG or T2 experiments for

exchange processes on the submillisecond and millisecond timescale, while saturation transfer

experiments should provide a reliable measure of exchange rates on the millisecond to second

timescale.

84

Chapter 3

A mutagenesis-free approach to assignment of 19F resonances in biosynthetically labeled proteins

The work presented in this chapter was published as a communication in “Kitevski-

LeBlanc, J. L., Al-Abdul-Wahid, M. S., Prosser, R. S. (2009). A mutagenesis-free approach to

assignment of 19F NMR resonances in biosynthetically labeled proteins. Journal of the American

Chemical Society 131, 2054-2055”. My role in this project consisted of the preparation of 13C,15N-3-fluorotyrosine and the corresponding NMR samples, setting up NMR experiments and

processing and analyzing the data. The CT-HCCF-COSY experiment was set up with Sameer

Al-Abdul-Wahid and the (Hβ)Cβ(CγCδ)Hδ experiment was set up with Dr. Ranjith

Muhandiram. © Reprinted with permission from ACS.

85

3.1 Introduction Fluorine (19F) NMR is ideally suited to the study of protein folding or unfolding, ligand

binding, enzymatic action, and internal motions (Danielson and Falke 1996; Gerig 1994). In

comparison with traditional labels such as 1H, 13C, and 15N, 19F spin probes give rise to a wide

range of chemical shifts associated with the unique local electronic environments in folded

proteins (Feeney et al. 1996; Frieden 2003; Luchette et al. 2002). Moreover, 19F can be readily

incorporated by biosynthetic means, often using non-auxotrophic bacterial strains with an

efficiency which depends on both the protein and amino acid (Frieden et al. 2004). Perturbations

to protein structure and function are minimal (Li and Frieden 2007; Xiao et al. 1998) or can be

made so by employing a fractional labeling strategy (Vaughan et al. 1999b). The most common

mono-fluorinated probes include tyrosine, phenylalanine, and tryptophan, where an induced-

auxotrophy approach produces uniform labeling with efficiencies of ~95% (Evanics et al. 2007;

Kim et al. 1990).

Assignment of fluorine resonances is often not straightforward and can only be achieved

by resorting to site-directed mutagenesis. Generally, a series of mutants and corresponding

spectra are obtained wherein each labeled residue is separately replaced with a structurally

similar amino acid; for example each occurrence of fluoro-tyrosine might be substituted for

phenylalanine (Anderluh et al. 2005; Gerig 1994). Alternatively, where such substitutions result

in gross changes in either the protein structure or the resulting 19F NMR spectrum, a residue

within van der Waals contact of a particular fluorine nucleus may be mutated to affect a change

in the fluorine chemical shift (Danielson and Falke 1996). However, this so-called nudge

mutation method requires a priori knowledge of the protein structure. Mutagenesis approaches

rapidly become time consuming and problematic due to spectral overlap for proteins possessing

many labeled residues. Here, we propose an approach which reduces spectral overlap and avoids

the use of mutagenesis in assigning 19F resonances of a fluoro-tyrosine enriched protein. The

approach requires that the protein is first biosynthetically labeled with a 13C and 15N-enriched

version of 3-fluoro-tyrosine. Through a combination of INEPT and COSY based transfers which

utilize the large 19F-13C and 13C-13C scalar couplings, it is possible to correlate the 19F resonances

with the 15N, 1H, and 13C resonances of the backbone, thereby accomplishing the assignments.

Figure 3.1 outlines the CT-HCCF-COSY pulse sequence which achieves a transfer from the delta

proton of the aromatic ring to the fluorine spin at the 3-position and is similar to earlier 13C-1H

experiments (Ikura 1991). The assignment is then completed using standard NMR pulse

86

sequences (Muhandiram and Kay 1994) which correlate the side chain resonances to the

backbone 15N, 1H, and 13C chemical shifts (vide infra).

Figure 3.1 CT-HCCF_COSY pulse scheme. Narrow (wide) pulses are applied with 90°(180°)

flip angles and phases, x, unless otherwise indicated. τH=1.6ms, τF=0.9ms, ΔCH=1ms, ΔCF=0.9ms,

and T=3.8ms, while a 2 kHz 13C WALTZ-16 decoupling field is used. Phase cycle φ1=x,-x;

φ2=4(x), 4(-x); φ3=2(x),2(-x); rec=x,-x,x,2(-x),x,-x,x.

3.2 Results Synthesis of 13C,15N-3-fluoro-L-tyrosine was achieved in one step by electrophilic

fluorination of 13C,15N-L-tyrosine. Selectfluor is a recently developed safe alternative to

traditional electrophilic fluorine sources such as F2 and XeF2 (Banks 1998). The reduced activity

of this reagent permits monofluorination under mild conditions, alleviating the problem of

producing a distribution of mono-, di-, and tri-fluorinated products, which would otherwise

reduce the yield of the desired mono-fluorinated variant. Selectfluor was added to the amino

acid dissolved in 80/20 acetonitrile/water in a 3:1 stoichiometric ratio and reacted for 2.5 hours at

80 ºC. After removal of solvent in vacuo, the product was redissolved in water adjusted to pH 2

with formic acid, and separated from reagent byproducts using a gravity flow Sepabead® SP850

column where a mixture of fluorinated and non-fluorinated tyrosine was eluted using 15%

acetonitrile in water. RP-HPLC on a C-18 column achieved separation of fluorinated and non-

fluorinated amino acids. Lyophilization then yielded pure 13C,15N-3-fluoro-L-tyrosine in 79%

isolated yield. A cost effective feature of this reaction and purification scheme is the ability to

87

recycle non-fluorinated 13C,15N-L-tyrosine for subsequent fluorination. Production of 33 mg/L

purified protein required 18 mg/L of 13C,15N-3-fluoro-L-tyrosine.

The assignment experiments are demonstrated on Ca2+-bound 13C,15N 3-fluoro-L-

tyrosine-enriched calmodulin (CaM) from Xenopous laevis using previously described

expression and purification protocols (Ikura et al. 1990). Although CaM has only two tyrosine

residues, Y99 and Y138, located in the C-terminal lobe, its size and tumbling time (~9 ns) were

deemed ideal to test the robustness of NMR pulse sequence schemes involving magnetization

transfer along the aromatic side chain of medium sized proteins. The corresponding resonances

are separated by 0.5 ppm. (15N,1H) HSQC spectra of fluorinated and non-fluorinated CaM

confirmed the fluorinated protein was fully calcium loaded and retained the native structure. The

assignment strategy, shown in Figure 3.2, begins with the correlation of each fluorine resonance

to the corresponding delta proton via three one-bond transfers using the pulse scheme shown in

Figure 3.1. The analogous CT-FCCH-COSY experiment differs only in the ordering of the

frequency channels and the inclusion of a presaturation period for water. Sensitivity associated

with the HCCF or FCCH experiment is proportional to γ1γ23/2(Δν1/Δν2), where γ1, γ2, Δν1, and

Δν2 represent the gyromagnetic ratios and linewidths of the starting and detect nuclei,

respectively. Therefore, if we consider the observed Hδ and 19F line widths, we estimate that the

sensitivity of the FCCH experiment is 1.9 times greater than that of the HCCF, although the

latter sequence has the advantage that there is no water signal. Connection to the backbone

resonances is achieved, as shown in Figures 2B and 2C, using the (Hβ)Cβ(CγCδ)Hδ (Yamazaki

et al. 1993) and the HNCACB experiments (Muhandiram and Kay 1994). The amide proton and

nitrogen chemical shift data for calmodulin were obtained from previously published data

(BMRB 6541) but can in principle can be obtained using the same sample if expressed under the

conditions required to produce uniform 13C and 15N labeling. In some cases, published amide 15N and 1H chemical shifts may differ slightly from experimental values obtained on a

fluorinated protein. In such situations the HNCACB and HNCO experiments help to verify

assignments by utilizing the chemical shifts of the neighboring Cα, Cβ, and amide 15N and 1H

nuclei.

88

Figure 3.2 Spectra used to assign fluorine resonances of 1mM CaM enriched with 13C,15N-3-

fluoro-L-tyrosine. Evolved nuclei are indicated with full circles, while dashed circles are used for

nuclei involved in magnetization transfer. (A) CT-HCCF-COSY (B) (Hβ)Cβ(CγCδ)Hδ (C)

HNCACB. Experiments were run at 37°C and recorded on a 600 MHz Varian Inova

spectrometer equipped with a HFCN quad probe. (Varian Inc., Palo Alto, CA).

The use of 13C,15N-enriched fluorotyrosine greatly enhances possibilities associated with 19F NMR. Firstly, (13C,19F) CT-HSQCs of CaM can be acquired in a matter of minutes (data not

shown), reducing spectral overlap in biosynthetically labeled proteins. The 13C-19F spin pair can

also be used to evaluate local order parameters and discern slow and fast motions through T1, T2

and 13C-19F NOE measurements. Finally, the ~50-75 ppm CSA common in fluoroaromatic amino

acids (Durr et al. 2008) suggests that TROSY effects may be useful at low field strengths,

resulting in improved line widths.

3.3 Summary In summary we have demonstrated that mutagenesis-free 19F NMR assignments of 13C,15N-

enriched 3-fluoro-L-tyrosine can be routinely performed in protein studies. The 13C-19F pair is

89

also expected to be useful via (13C,19F) HSQCs in reducing spectral overlap and in dynamics

studies of side chains. The 1-step synthesis procedure employed produces the fluorinated amino

acid in high yield and allows for the recycling of non-fluorinated starting material for subsequent

fluorination, making this scheme cost effective compared to previously reported methods (Azad

et al. 2008; Kirk 1980). The CT-HCCF-COSY experiment produces high quality spectra, while

additional 2- and 3-D NMR experiments may be used to correlate side chain resonances with

those of the backbone. Similar approaches may be possible with other fluoroaromatics. For

example, 13C,15N-enriched phenylalanine may be fluorinated after an intermediate nitration and

reduction step.

3.4 Supplementary Data

3.4.1 Synthesis and Purification of 13C, 15N-enriched-L-3-fluorotyrosine

General reaction scheme for the production of 13C, 15N-enriched-L-3-fluorotyrosine:

O

NH3+

OH O-

O

NH3+

OH O-

F

SELECTFLUOR

Acetonitrile, 40-80 degrees

Figure 3.3 Reaction scheme for synthesis of 13C,15N-3-fluoro-L-tyrosine.

90

Figure 3.4 19F NMR spectrum of 13C,15N-3-fluoro-L-tyrosine.

19F NMR (TFA, 600MHz; 13C-decoupled): δ = -136.6 (t, J = 9.9Hz)

91

Figure 3.5 13C NMR spectrum of 13C,15N-3-fluoro-L-tyrosine.

13C NMR (CDCl3, 600MHz): δ = 176.2 -176.6, 152.5 - 155.2, 145.2, 130.0 – 130.4, 128.0 –

128.2, 119.5 – 120.7, 58.6, 38.1

92

Chapter 4

Approaches for the measurement of solvent exposure in proteins by 19F NMR

The work presented in this chapter was published in “Kitevski-LeBlanc, J. L., Evanics,

F., Prosser, R. S. (2009) Approaches for the measurement of solvent exposure in proteins by 19F

NMR. Journal of Biomolecular NMR 45, 255-264”. My role in this project consisted of sample

preparation, setting up NMR experiments and processing and analyzing data. NMR experiment

set up was performed with Dr. Ferenc Evanics. © Reprinted with permission from Springer.

93

4.1 Abstract Fluorine NMR is a useful tool to probe protein folding, conformation and local topology owing

to the sensitivity of the chemical shift to the local electrostatic environment. As an example we

make use of 19F NMR and 3-fluorotyrosine to evaluate the conformation and topology of the

tyrosine residues (Tyr-99 and Tyr-138) within the EF-hand motif of the C-terminal domain of

calmodulin (CaM) in both the calcium-loaded and calcium-free states. We critically compare

approaches to assess topology and solvent exposure via solvent isotope shifts, 19F spin-lattice

relaxation rates, 1H-19F nuclear Overhauser effects, and paramagnetic shifts and relaxation rates

from dissolved oxygen. Both the solvent isotope shifts and paramagnetic shifts from dissolved

oxygen sensitively reflect solvent exposed surface areas.

4.2 Introduction

The fluorine nucleus is an exquisite NMR probe with which to study internal motions,

ligand binding, enzymatic action, and folding in proteins (Danielson and Falke 1996; Gakh et al.

2000; Gerig 1994). The utility of the 19F chemical shift in protein structure and dynamics studies

arises from the paramagnetic component of the chemical shielding term associated with the

fluorine lone pair electrons, which is sensitive to local van der Waals packing and electrostatic

fields (Chambers et al. 1994; Feeney et al. 1996; Kubasik et al. 2006; Li and Frieden 2005; Lian

et al. 1994). Examples abound in 19F NMR studies of folded proteins, where a biosynthetically

incorporated amino acid probe exhibits a chemical shift dispersion of 5-20 ppm (Lian et al.

1994). In this context, it may not be surprising that even the substitution of H2O for D2O is

known to elicit an isotope shift of as much as 0.25 ppm, depending on the extent of solvent

exposure of the nuclear spin probe (Gerig 1994). 19F spin-lattice relaxation rates (R1) and 1H-19F

cross relaxation rates are known to be sensitive to environment and solvent exposure, as are

corresponding paramagnetic rates arising from the dissolution of paramagnetic additives

(Prosser et al. 2000). In this paper, we compare the utility of these effects – solvent isotope

shifts, 19F spin-lattice relaxation rates (R1), 1H-19F NOEs, and paramagnetic shifts and rates from

dissolved O2 – for purposes of best discriminating protein topology by 19F NMR. We do this

through a 19F NMR study of a well-known calcium binding protein, calmodulin (CaM), whose

94

two tyrosine residues, located in the C-terminal domain, have been biosynthetically labeled with

3-fluorotyrosine.

Calmodulin is a ubiquitous modulator of many calcium-dependent processes in cells and

it exhibits well defined conformations and binding states depending on intracellular calcium

levels and the presence of a host of known binding proteins, hormones, and peptides (Crivici

and Ikura 1995; Hoeflich and Ikura 2002). CaM is a well-structured, acidic protein consisting of

148 residues and possessing two structurally similar domains connected by a flexible tether

(Crivici and Ikura 1995), as shown in Figure 4.1.

Figure 4.1 Ribbon diagrams of human calmodulin in both the calcium-loaded and calcium-free

states (PDB files 1CLL and 1DMO, respectively) with the location of tyrosine residues

indicated. Helices associated with the EF calcium binding motifs are indicated as E1 and F1 for

EF hand 1 and E2 and F2 for EF hand 2. Note that although the central helix appears rigid in the

calcium-loaded X-ray structure, it has been shown by solution NMR that both calcium-loaded

(Barbato et al. 1992) and calcium-free calmodulin (Tjandra et al. 1995) exhibit high flexibility

in this region. The major difference between the two structures is the helical arrangement.

Within each domain there are two calcium binding EF-hand motifs coordinating a total of

four calcium ions per protein. The well-known EF-hand motif is composed of two helices (E

95

and F), connected by a loop of typically 12 residues, which coordinate a calcium ion with

pentagonal bipyramidal symmetry (Malmendal et al. 1999). The seven ligands are provided by 5

sidechain carboxylate oxygens, one backbone carbonyl oxygen, and one water oxygen atom

(Strynadka and James 1989). The two tyrosine residues in calmodulin are both located in the C-

terminal domain where they serve distinct roles related to structure and activation. Tyr-99 is

involved in calcium coordination through its main chain carbonyl atom in the first EF hand of the

C-terminal domain and contributes to a small beta-strand which keeps the loops of consecutive

EF hands connected (Malmendal et al. 1999). Tyr-138 does not contribute to the coordination

of calcium, but is the last residue of a small beta strand in the second EF hand of the C-terminal

domain. Overall, the structure of CaM is dominated by helical secondary structural elements,

which are virtually identical between the calcium-loaded and calcium-free forms of the protein,

where the major difference is related to helical packing (Zhang et al. 1995). In the calcium-free

state the two helices of each EF hand are nearly antiparallel. Upon binding of calcium an

opening of the two helices is induced resulting in an almost perpendicular arrangement with

interhelical angles ranging from 86° to 101° (Babu et al. 1988). It is therefore likely that this

structural reorganization will be accompanied by significant changes in sidechain conformations,

specifically those involved in calcium coordination and within the helical segments themselves.

Based on the available structures of calmodulin the solvent accessible surface area of

Tyr-99 and Tyr-138 calculated using MOLMOL are 45% and 54% in calcium-loaded CaM (PDB

entry 1CLL) and 36% and 47% in calcium-free CaM (PDB entry 1DMO) respectively. Here we

will show that although subtle, the differences between the residues, and between the two

functional forms of the protein itself, can be delineated using the aforementioned 19F NMR

techniques. We begin with a discussion of the biosynthetic labeling and assignment protocol as

well as procedures used to assess possible perturbations arising from the use of fluorine as a

probe as this is a prerequisite to any 19F NMR study.

96


4.3.1 Expression and Purification of Uniformly 15N ,13C-L-3-fluorotyrosine labeled CaM

Incorporation of 15N,13C-enriched L-3-fluorotyrosine via heterologous expression

(Evanics et al. 2007) as well as purification of calmodulin (Ikura et al. 1990) was performed as

previously described with slight modifications. A plasmid (pET21b) encoding Xenopus laevis

calmodulin (residues 1-148) was transfected into BL21(DE3) under the control of the T7

promoter. LB broth inoculated with a single colony was grown overnight and used to inoculate

1L of M9 minimal media supplemented with 0.3% D-glucose, 0.1% 15NHCl4, 100 mg/L

ampicillin, 10 mg/L thiamine, 10 mg/L biotin, 1 mM MgSO4, and 0.1 mM CaCl2. Uniform

labeling with fluorotyrosine was achieved by the introduction of 1 g/L glyphosate, 75 mg/L DL-

tryptophan, and 75 mg/L DL-phenylalanine to shaking bacterial cultures at 37°C which had

reached an OD600 of 0.600. Once cell cultures achieve an OD600 0.800 (after approximately 1

hour), 18 mg/L 13C,15N-L-3-fluorotyrosine was added and expression was induced with the

addition of 238 mg/L IPTG. Cell cultures were harvested after 3.5 hours by centrifugation at

7000 rpm for 20 minutes. Cells were then resuspended in 50 mM NaH2PO4, 300 mM NaCl, 10

mM imidazole, 1 mM PMSF pH 8 and lysed by incubation at 4°C in the presence of 1 mg/mL

lysozyme for 30 minutes, followed by sonication. After addition of DNase (10 µg/mL) and

RNase (5µg/mL), the suspension was centrifuged at 9000 rpm for 20 min at 4°C and the cleared

lysate was purified using Ni-NTA Agarose resin (Qiagen, Mississauga, Ontario, Canada). The

labeled protein was further purified using phenyl sepharose as described previously. (Ikura et al.

1990) Pooled protein samples were buffer exchanged into 20mM BIS-Tris, 0.1M KCl, 9 mM

CaCl2, 0.2% NaN3 at pH 7.5. Calcium-free calmodulin samples were prepared as previously

described, (Zhang et al. 1995) with slight modification. All buffers used in the preparation of

calcium-free CaM were made with water which had been decalcified using Chelex-100 resin and

stored in plastic bottles which had been treated with 5mM EDTA followed by extensive rinsing

with decalcified water. Ethylenediaminetetraacetic acid (EDTA) was added to purified, dilute

protein samples (~200-500µM) to a final concentration of 20mM, followed by precipitation with

trichloroacetic acid (TCA). The sample was redissolved in 25mM NH4HCO3 and passed through

a Sephadex G-25 column. Collected protein was then exchanged into 20mM Bis-Tris, 0.1M

97

KCl, 2mM EDTA, 0.2% NaN3 at pH 7.5 for NMR experiments and macroscopic binding

constant determination.

4.3.2 Determination of macroscopic binding constants

Macroscopic calcium binding constants were measured using a competitive chelator

assay as described previously (Linse et al. 1991). 5,5’-Br2-BAPTA was obtained from

Molecular Probes (Eugene, Oregon). Briefly, calmodulin and 3-fluorotyrosine labeled

calmodulin were decalcified as described above and titrated with 1mM CaCl2 in the presence of

5,5’-Br2-BAPTA in 20mM Bis-Tris, 0.1M KCl, pH 7.5, prepared with decalcified water at 25°C

until OD263nm indicated saturation. The raw data was fit using Caligator software (Andre and

Linse 2002) to obtain K1 through K4.


NMR experiments were performed on a 600 MHz Varian Inova spectrometer (Agilent

Technologies, Santa Clara, CA). For direct observe 19F NMR experiments, and all HCN

experiments, a 5 mm HCN/FCN triple resonance single gradient salt-tolerant cryogenic probe,

tunable to either 1H or 19F, was used. All calmodulin sample concentrations were 1.5 to 2 mM in

90%H2O/10%D2O, 0.1 M KCl, 7.1 mM CaCl2, 0.2% NaN3 at pH 7.5. 1-D 19F NMR spectra

were obtained at 30°C with 256 transients, using a spectral width of 15000 Hz and a π/2 pulse

width of 11 µs. 19F spin-lattice relaxation times (T1) were determined using an inversion

recovery sequence (i.e. 180°-τ-90°) with a total of 12 τ values ranging from 1 ms to 4.5 s, and a

repetition time of 4 s. For purposes of measuring paramagnetic rates associated with spin-lattice

relaxation in the presence of dissolved oxygen, T1 experiments were repeated under an oxygen

partial pressure of 35 Atm. Note that the pressures used to induce paramagnetic effects from

dissolved oxygen are extremely low (20-40 Atm) relative to pressures which typically affect

detectable structural perturbations (Wilton et al. 2008). We have observed minor perturbations

at or above 50 Atm in membranes (Prosser et al. 2001) and disordered proteins (Bezsonova et

al. 2006), however folded proteins are typically less compressible. These experiments

necessitated the use of a 5 mm OD, 3 mm ID sapphire NMR sample tube (Saint Gobain-

Saphikon Crystals, Milford, NH). Samples were generally equilibrated under 50 Atm oxygen

partial pressure at 5 °C for 2 days, then equilibrated overnight in the magnet at the desired

temperature and partial pressure. Open Swagelok connections (Swagelok, Solon, OH) to a

98

pressurized oxygen supply were used to maintain the pressure throughout the experiment. For

solvent isotope shift measurements the buffer was exchanged from 90% H2O to 90% D2O using

centrifugation filters by repetition of diluting the sample 10 fold followed by concentration 4

times. Based on the consistency of our measurements under the various conditions employed, it

seems clear that no major structural perturbations result from exchange into deuterated buffer

conditions. 19F NMR transverse (T2) relaxation times were measured using a Hahn echo

sequence (i.e. 90°-τ-180°-τ) with 12 τ values ranging from 0.64 ms to 11.2 ms, and a repetition

time of 2.75 seconds. 1H-19F steady-state NOE measurements were performed on a HFCN quad

probe (Varian Inc., Palo Alto, CA) using a differencing FHOESY experiment (Lix et al. 1996;

Rinaldi 1983). Fluorine resonances and in alternate scans either water, aliphatic or all protons

(data not shown) were saturated for 12 ms, followed by a mixing period of 0, 0.5 and 1.0 s for

water presaturation, and 0, 0.1, 0.25, 0.5, 0.75, 1, 1.5, and 2.0 s for aliphatic proton presaturation.

Proton saturation was optimized in a separate 1H direct-detect experiment and found to be

optimal using an RF field of 1838 Hz for the 12 ms saturation period. The use of simultaneous

saturation of fluorine and proton resonances generated NOE difference spectra with minimal

artifacts when compared to other heteronuclear NOE pulse sequences. 19F spectra resulting from

the difference NOE were collected using 8192 transients, a spectral width of 2200 Hz, and a

repetition time of 3 s. [15N,1H] HSQC pulse sequences were obtained from Biopack software

(Varian Inc.) and typically were collected in 4 scans with 96 increments spanning 2000 Hz in the

indirect dimension. All NMR data were processed using NMRPipe/NMRDraw (Delaglio et al.

1995), and analyzed with NMRView software (Johnson and Blevins 1994).


4.4.1 Labeling and Expression

Mono-, di-, and trifluorinated amino acid analogues can be incorporated into proteins by

heterologous expression techniques, often using auxotrophic strains or induced auxotrophy

(Salopek-Sondi et al. 2003). Typically, fluorine labels are minimally perturbing to the overall

protein structure (Li and Frieden 2007; Xiao et al. 1998), or can be made so by fractional

labeling (Feeney et al. 1996). To date, fluorinated methionine (Salopek-Sondi et al. 2003;

Vaughan et al. 1999b), leucine (Feeney et al. 1996), isoleucine (Mock et al. 2006), histidine

(Eichler et al. 2005b), phenylalanine (Li and Frieden 2005), tryptophan (Anderluh et al. 2005),

99

and tyrosine (Hull and Sykes 1976; Sykes et al. 1974), have been incorporated into proteins via

microbial expression. In our particular expression and biosynthetic labeling protocol, glyphosate,

an inhibitor of aromatic amino acid synthesis, is added one hour prior to induction of protein

expression to deplete endogenous sources of phenylalanine, tyrosine and tryptophan. At this

time non-fluorinated phenylalanine and tryptophan are also added. At induction, 3-

fluorotyrosine (either a 13C, 15N-enriched form or exactly as obtained from the commercial

supplier) is added as well as IPTG to initiate protein expression at 37°C in M9 minimal media

supplemented with additional nutrients and 15NH4Cl for uniform 15N-enrichment. CaM was thus

expressed uniformly labeled with 3-fluoro-L-tyrosine, and purified as described in the Materials

and Methods section. 13C and 15N enrichment of 3-fluoro-L-tyrosine was necessary for purposes

of 19F NMR resonance assignments. This label was prepared using a synthetic protocol involving

electrophilic fluorination of 13C, 15N-L-tyrosine with Selectfluor as described in detail

elsewhere (Kitevski-LeBlanc et al. 2009a).

To assess the extent to which fluorotyrosine incorporation affects the backbone

conformation of the protein, [15N,1H] HSQC spectra were acquired for fluorinated and non-

fluorinated CaM under both calcium-loaded and calcium-free conditions, as shown in Figure 4.2.

The careful preparation of calcium-free buffers and CaM is described in the Materials and

Methods section. The high degree of overlap in these spectra indicate that the global protein fold

is the same in the presence and absence of 3-fluorotyrosine. The [15N,1H] HSQC spectra also

provide qualitative evidence for native-like calcium binding affinity in the fluorinated proteins.

As shown in Figure 4.2, distinctive spectral features of non-fluorinated calcium-loaded and

calcium-free spectra are maintained in the fluorine labeled proteins under identical sample buffer

conditions.

100

101

Figure 4.2 [15N,1H] HSQC NMR spectrum of 15N-enriched CaM overlaid with an equivalent

spectrum of 15N-enriched CaM in which Tyr99, and Tyr138 have been replaced with 15N,13C-

enriched 3-fluorotyrosine for both calcium-loaded (A) and calcium-free (B) states at 37°C. Note

that in the calcium-free state the 3-fluortyrosine was not 15N-enriched, and a cross peak is thus

not observed in the spectrum.

A more quantitative description of calcium binding properties is provided by

measurement of the macroscopic binding constants in both the presence and absence of the

fluorinated probe. This was achieved using a competitive calcium chelator assay (Linse et al.

1991) and data were fit using Caligator software (Andre and Linse 2002). The four macroscopic

binding constants in the absence and presence of 3-fluorotyrosine were K1=2.4×105 M-1,

K2=1.1×106 M-1, K3=3.8×104 M-1, K4=1.2×105 M-1, and K1=1.5×105 M-1, K2=3.1×106 M-1,

K3=7.8×104 M-1, K4=8.2×104 M-1, respectively. In both the labeled and un-labeled protein the

macroscopic binding constants indicate two cooperative binding events, previously shown to be

associated with calcium binding within each domain (Linse et al. 1991), where one such event

occurs at calcium concentrations approximately an order of magnitude lower than the other.

Although the values obtained are comparable to those previously published under similar buffer

conditions (Linse et al. 1991), it is clear that the fluorine labels do affect binding as two of the

constants measured, K3 and K4, are affected to a greater extent and likely correspond to the

domain which contains our fluorine probes. It is clear that under the current calcium-loaded

buffer conditions we can be confident that the protein is fully calcium bound.

4.4.2 Topology through chemical shift measurements 19F NMR chemical shifts of fluoroaromatics are highly sensitive reporters of the

electrostatic environment and tend to exhibit a broad chemical shift dispersion in proteins

(Evanics et al. 2007). Downfield shifts and faster spin-lattice relaxation rates (R1) are generally

associated with more buried fluorine nuclei (Sykes et al. 1974). 19F NMR spectra of the

fluorinated protein in both the calcium-loaded and calcium-free states are shown in Figure 4.3.

Note that the peak assignments in the calcium-loaded state were obtained by a series of NMR

experiments, which first correlated the residue-specific 19F NMR chemical shifts to the

intraresidue Cδ and Hδ shifts via a CT-HCCF-COSY. Connection to the known backbone shifts

102

was then made through the (Hβ)Cβ(CγCδ)Hδ, and HNCACB experiments as described in detail

elsewhere (Kitevski-LeBlanc et al. 2009a). Resonance assignments of calcium-free CaM were

then obtained by monitoring the 19F NMR shifts as CaCl2 was titrated. The resonances arising

from the two fluorotyrosine residues in calcium-loaded CaM, and calcium-free CaM are clearly

resolved, indicating distinct environments, while the more downfield shifts associated with Tyr-

99 are suggestive of a more buried environment. Tyr-138 is associated with two peaks in the

calcium-free state, a major peak (Tyr-138M) and a minor peak (Tyr-138m), which is slightly

upfield suggesting the minor peak to be the more solvent exposed of the two. Interestingly, the

chemical shift of Tyr-138m is coincident with the chemical shift of this residue in the calcium-

loaded form. We therefore propose that this minor peak may represent a calcium-bound like

conformation with respect to Tyr-138, where the chemical shift difference implies that exchange

between the so called major and minor conformers is slower than ~5ms. It is also apparent that

the total integrated area of Tyr-138 is greater than that of Tyr-99. 19F CPMG experiments (data

not shown) indicate that Tyr-99 is undergoing chemical exchange, and it is possible that the

missing peak intensity is associated with an unobservable exchange broadened state.

The 19F spin-lattice relaxation rates, R1, corroborate the above chemical shift trends;

faster relaxation rates, observed for Tyr-99, indicate greater burial, as summarized in Table 4.1.

Curiously the relaxation rate for Tyr-138m is the highest at all temperatures, indicating that

although this resonance is the most upfield, it appears to be in greater dipolar contact with the

protein interior relative to the other residues. It should be noted that although spin-1/2 relaxation

is in general dominated by a dipole-dipole mechanism, the large chemical shift anisotropy (CSA)

associated with the fluorine nucleus compounded by modern magnetic field strengths is expected

to result in significant weakening of the correlation between dipolar contact, or burial, and the 19F relaxation rate.

103

Table 4.1 Chemical shifts and spin–lattice relaxation rates for 19F nuclei in 3-fluorotyosine

labeled calcium-loaded and calcium-free calmodulin.

δ (ppm) R1 (Hz) 5°C R1 (Hz) 15°C R1 (Hz) 25°C

Ca2+-loaded

Tyr-99 -136.98 1.10±0.06 1.34±0.05 1.60±0.05

Tyr-138 -137.48 0.96±0.05 0.90±0.07 1.01±0.04

Ca2+-free

Tyr-99 -136.61 1.57±0.05 1.57±0.06 1.69±0.04

Tyr-138M -137.07 1.24±0.04 1.24±0.04 1.35±0.03

Tyr-138m -137.39 2.03±0.48 2.17±0.43 2.29±0.55

While the chemical shifts and R1 measurements report on the degree of burial, one can

directly assess solvent exposure using solvent induced isotope shifts. The exchange of 90% H2O

with 90% D2O can shift solvent accessible resonances by as much as 0.25ppm (Gerig 1994).

The solvent isotope shifted 19F NMR spectra for calcium-loaded and calcium-free CaM are

shown in Figure 4.3, and the solvent induced isotope shifts are summarized in Table 4.2. Note

that the most pronounced solvent isotope shifts are observed for Tyr-138 corroborating the above

finding based on diamagnetic shifts and R1 data, that it exhibits the greatest solvent exposure. In

contrast to the R1 value obtained for Tyr-138m, the solvent isotope shift associated with this

resonance indicates a high degree of solvent exposure of the minor conformer.

104

Figure 4.3 19F NMR spectra of 3-fluorotyrosine labeled calcium-loaded (A) and calcium-free (B)

CaM at 37°C with assignments indicated. Note that the calcium-bound CaM sample was

enriched with 13C,15N 3-fluorotyrosine, therefore the spectra in A were obtained with 13C

decoupling.

Table 4.2 Paramagnetic relaxation rates, paramagnetic shifts and solvent isotope shifts for

calcium-loaded and calcium-free calmodulin. Parameters marked with an asterisk were

normalized to free 6-fluortryptophan in buffer solution. R1 is as defined in equation 1.5.4 (8).

All measurements were made at 25°C.

R1 (Hz)

O2

35Atm

ΔR1 (Hz) Δδ O2

(ppm)

Δδ* O2

(ppm)

Δδ D2O-H2O

(ppm)

Δδ* D2O-H2O

(ppm)

Δδ* O2 / Δδ*

D2O-H2O

Ca2+-

loaded

105

Tyr-99 8.13±0.21 7.03±0.22 0.180 0.623 -0.113 0.456 1.697

Tyr-138 10.0±0.17 9.04±0.17 0.300 1.038 -0.302 1.218 1.058

Ca2+-free

Tyr-99 8.50±0.12 6.72±0.13 0.130 0.450 -0.075 0.429 1.050

Tyr-138M 9.24±0.15 7.92±0.15 0.216 0.747 -0.178 1.017 0.735

Tyr-138m 9.70±0.15 7.84±0.57 0.260 0.900 -0.182 1.04 0.865

4.4.3 Topology through paramagnetic additives

Solvent exposure and conversely, burial, can also be substantiated by the measurement of

shifts or relaxation rate enhancement from dissolved paramagnetic additives. For example, the

addition of minute quantities of 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPOL)

exerts prominent changes in spin-spin or spin-lattice relaxation rates on solvent accessible spins

(Niccolai et al. 2003). Alternatively, a variety of chelated paramagnetic metal species may be

used to probe topology and solvent exposure through local relaxation rate enhancements. O2 is

also a convenient paramagnetic additive, particularly in conjunction with 19F NMR studies, since

both shifts and relaxation rate enhancements are readily observed at oxygen partial pressures of

20-50 bar (Evanics et al. 2007). Paramagnetic spin-lattice relaxation rates ΔR1 (obtained by

simply measuring the difference between R1 in the presence of dissolved O2 and under

atmospheric conditions) and paramagnetic shift data are summarized in Table 4.2 for each of the

fluorinated residues. While paramagnetic rate enhancements at 25°C turn out to be modestly

greater for Tyr-138 in the calcium-loaded form (i.e. 9.04 vs 7.03 s-1 for Tyr-138 and Tyr-99

respectively) the paramagnetic shifts are observed to be significantly higher for Tyr-138 (i.e.

0.30 vs 0.18 ppm for Tyr-138 and Tyr-99 respectively), reaffirming the tentative conclusion

based on chemical shifts and R1 under atmospheric oxygen levels that the fluorine nucleus of

Tyr-99 is on average more buried than that of Tyr-138. Under the same conditions, a similar

conclusion is reached for calcium-free CaM where both Tyr-138M and Tyr-138m exhibit higher

rate enhancements, and correspondingly higher paramagnetic shifts than Tyr-99. Although the

observed rate enhancements for Tyr-138M and Tyr-138m are comparable, the paramagnetic shift

106

is highest for the minor conformer suggesting it is most solvent exposed. Note that the O2

paramagnetic shifts exhibit a wider range of effects than the paramagnetic rates. Subtle

differences between paramagnetic rates and shifts in 19F NMR are expected since the relaxation

and shift terms are entirely different in origin, while dynamics and distance effects further

complicate the analysis.

Paramagnetic effects from dissolved oxygen are not purely steric in origin. One expects

more hydrophobic regions to experience a more favorable interaction with O2 which should be

reflected in heightened paramagnetic effects due to greater local partitioning (Teng et al. 2006).

It has been shown that one can distinguish between the steric and local partitioning effects by

separately making use of hydrophobic and hydrophilic shift reagents (Evanics et al. 2006). Two

such complementary measurements involve the solvent isotope shifts, and O2-induced

paramagnetic shifts, where O2 serves as the hydrophobic probe and the D2O-H2O mixture as the

hydrophilic probe. Thus, assuming that the steric factors are similar for water and oxygen, we

can take the ratio of the paramagnetic shift to the isotope shift (normalized to free

fluorotryptophan in the sample buffer), Δδ* O2 / Δδ* D2O-H2O, as a relative measure of

hydrophobicity. The data, shown in Figure 4.3 and summarized in Table 4.2, indicate that in the

calcium-loaded form, Tyr-99 is situated in a more hydrophobic environment than Tyr-138, with

hydrophobicity measures of 1.697 and 1.058 respectively. This is also consistent with the results

of the 1H-19F NOE measurements in which aliphatic 1H saturation resulted greater cross-

relaxation to Tyr-99. In the calcium-free state the hydrophobicity index, Δδ* O2 / Δδ* D2O-H2O,

also suggests that Tyr-99 is situated in a more hydrophobic environment than Tyr-138M or Tyr-

138m. Interestingly, although the chemical shift of the minor conformer is coincident with Tyr-

138 in the calcium-loaded form, the hydrophobicity index for Tyr-138m falls very close to the

average between Tyr-138M and Tyr-138. This may reflect some degree of local conformational

fluctuations in Tyr-138m resulting in a hydrophobicity index value distinct from that of Tyr-138.

The exchange between the major and minor conformational states of Tyr-138 in calcium-free

CaM appears to be slower than the ~5ms dictated by the chemical shift difference, and is

currently under further investigation.

107

4.4.4 Heteronuclear 1H-19F NOEs 1H-19F cross-relaxation rates can also be used to assess solvent accessibility or burial of

specific fluorinated sites, assuming that cross-relaxation rates between 19F and H2O can be

distinguished from those between 19F and the protein interior. The heteronuclear NOE is well

established (Rinaldi 1983) as are 1H-19F NOEs in proteins (Campos-Olivas et al. 2002; Cistola

and Hall 1995; Yu et al. 2006). For example, the work of Campos-Olivas et al. established inter-

residue 1H-19F connectivities from a 5-fluorotryptophan substituted protein consisting of 56-

residues, while Hajduk et al. utilized intermolecular 1H-19F NOEs to establish protein-ligand

structural restraints in complexes of the anti-apoptotic protein Bcl-xL and several fluorinated

drugs (Yu et al. 2006). What becomes clear from these few examples is that the magnitude of

the 1H-19F NOE is made significantly smaller in slow tumbling systems and at higher field

strengths particularly in situations where the fluorine probe possesses a large CSA. In our hands

it proved impossible to observe a prominent 1H-19F NOE (over 12 hours) between the 3-

fluorotyrosine resonances and the δ-proton resonances or indeed, a backbone 1H resonance

which would have simplified assignment.

The 1H-19F NOE is suspected to be weak in our system for two main reasons. In cases

where the fluorine resonance is saturated or inverted, the magnetization can return to equilibrium

via the CSA relaxation mechanism significantly reducing 1H-19F cross relaxation. An additional

problem is the reduced efficiency of the relevant spectral density terms associated with cross

relaxation in the heteronuclear case (versus the homonuclear NOE). In the absence of CSA

effects the magnitude of the 19F NOE effect upon saturation of a 1H spin may be expressed as

(Neuhaus and Williamson 2000)

!

NOE =I0 II

=# HF

HF

=%H

%F

6J(&H +&F ) J(&H &F )J(&H &F ) + 3J(&F ) + 6J(&H +&F )

[1]

where γF and γH define the gyromagnetic ratios of the 19F and 1H nuclei, σHF and ρHF represent

the cross-relaxation and auto-relaxation terms respectively, and J(ω) represent the familiar

spectral density functions. In large proteins, the dominant spectral density terms for σHF are J(0)

and J(ωH-ωF) for the homo- and heteronuclear cases respectively. In CaM specifically, where we

assume that the protein undergoes isotropic reorientation with a correlation time (τc) of 6.30 ns

(Barbato et al. 1992), the relevant spectral densities become J(ωH-ωF) = 4.28 ns rad-1, J(0)=12.60

108

ns rad-1, making the heteronuclear cross relaxation term 3.4 times less efficient than the

homonuclear equivalent. In addition, an observable NOE is compromised by internal motions

which have been shown to result in significantly reduced enhancements in 3-fluorotyrosine

labeled alkaline phosphatase (Hull and Sykes 1975).

In situations where an explicit 1H-19F NOE cannot be observed, cross-relaxation between

all aliphatic or water protons and the fluorine nucleus of interest provides a perspective on

topology – i.e. is the 19F nucleus in question buried in the protein or facing the water exterior?

Figure 4.4 reveals a series of difference spectra showing a NOE effect on both Tyr-99 and Tyr-

138 (Tyr-138M only) as a function of a selective presaturation on either the water (Figure 4.4A,

B) or the1H aliphatic region (Figure 4.4C, D), for a range of mixing times as indicated. Due to

the low signal intensity of Tyr-138m and the aforementioned weaknesses of the heteronuclear

NOE we were unable to obtain a measurable 1H-19F NOE to Tyr-138m over a 12-hour time span.

The results from the two complementary experiments indicate that Tyr-99 appears to be in

greater dipolar contact with the aliphatic protons, while Tyr-138 is in greater contact with water.

This corroborates the above topology experiments (shift and T1 measurements, isotope effects,

and paramagnetic effects from dissolved O2) all of which suggest that the two fluorotyrosine

labels are in markedly different environments.

109

Figure 4.4 1H-19F 1D difference NOE initiated by saturation of either water (A,B) or aliphatic

protons (C,D), for calcium-free CaM, shown in the upper two panels, and calcium-loaded CaM,

shown in the lower two panels. NOE mixing times are indicated on the left margin of each

spectrum. The observed enhancements for water (aliphatic) saturation and a 1 s mixing time

correspond to 3.0% and 2.1% (1.7% and 2.4%) for Tyr-99 in calcium-free and calcium-loaded

calmodulin respectively. The corresponding enhancements for Tyr-138 are 2.6% and 2.8%

(0.6% and 0.4%).

4.5 Conclusions and Final Remarks

In this paper, we have explored a variety of alternative approaches for the measurement of

surface topology of 19F-labeled sites using 3-fluorotyrosine labeled calmodulin, in both the

calcium-loaded and calcium-free states as a case study. In particular, solvent isotope shifts,

paramagnetic shifts and rates from dissolved O2, and 1H-19F heteronuclear NOEs resulting from 1H aliphatic and water presaturation, all provide a consistent picture in which Tyr-99 is more

buried and less solvent exposed than Tyr-138 in both the calcium-loaded and calcium-free states.

110

In the calcium-free state, an interesting slow equilibrium between a major sidechain conformer

and a minor calcium-bound like state was also consistently observed for Tyr-138 (i.e. Tyr-138M

and Tyr-138m). Further characterization of this equilibrium is currently under investigation in

our lab.

A tentative comparison of the relative sensitivity of the solvent isotope shifts and the

paramagnetic shifts and rates from dissolved O2 can be made by examining the correlation

between the above parameters and the solvent exposed surface areas, as determined from the

high resolution structures of the calcium-loaded and calcium-free states of CaM. In the case of

either solvent isotope shifts or O2-induced paramagnetic shifts, a small change (~10%) in solvent

exposure is associated with a corresponding change in the measured parameter. Further, it is

shown in Figure 4.5 that the geometric average of the solvent isotope shifts and paramagnetic

shifts from dissolved O2,

!

| # * (O2) # * (D2O%H2O) | , correlates best with solvent exposure

since partitioning effects associated with hydrophobic regions are factored out (Evanics et al.

2006).

Figure 4.5 Graphical representation of the correspondence between the geometric average of the

experimentally determined solvent isotope shifts,

!

# * (D2O H2O) , and O2-induced

paramagnetic shifts,

!

# * (O2) , given by

!

| # * (O2) # * (D2O%H2O) | , as a function of solvent

111

exposed surface areas calculated from each of the two high resolution structures for CaM in the

calcium-loaded ad calcium-free states

The hydrophobicity index calculated as the normalized paramagnetic shift from dissolved

O2, divided by the normalized solvent isotope shifts also proved to be a useful tool in the analysis

of protein topology. In both the calcium-loaded and calcium-free states, Tyr-99 proved to exist

in a more hydrophobic environment, while the minor conformer associated with Tyr-138 in the

calcium-free state (Tyr-138m) appeared to be situated in an environment that was between that

of the calcium-loaded and dominant calcium-free state for this residue.

The above analysis suggests that the solvent exposure and surface topology can be

reliably assessed by 19F NMR, by making use of solvent isotope shifts, O2 paramagnetic shifts,

and to some extent HF cross-relaxation. It is anticipated that such measurements might nicely

complement any 19F NMR study of proteins, as the observed effects are not restricted to aromatic

fluorine probes.

112

Chapter 5

Approaches to the assignment of 19F resonances from 3-fluorophenylalanine labeled calmodulin using solution

state NMR

The work presented in this chapter was published in “Kitevski-LeBlanc, J. L., Evanics,

F., Prosser, R. S. (2010). Approaches to the assignment of 19F resonances from 3-

fluorophenylalanine labeled calmodulin using solution state spectroscopy. Journal of

Biomolecular NMR 47 113-123”. My role in this project consisted of sample preparation, setting

up NMR experiments and processing and analyzing data. NMR experiment set up was

performed with Dr. Ferenc Evanics. © Reprinted with permission from Springer.

113

5.1 Abstract Traditional single site replacement mutations (in this case, phenylalanine to tyrosine) were

compared with methods which exclusively employ 15N and 19F-edited two- and three-

dimensional NMR experiments for purposes of assigning 19F NMR resonances from calmodulin

(CaM), biosynthetically labeled with 3-fluorophenylalanine (3-FPhe). The global substitution of

3-FPhe for native phenylalanine was tolerated in CaM as evidenced by a comparison of 1H-15N

HSQC spectra and calcium binding assays in the presence and absence of 3-FPhe. The 19F NMR

spectrum reveals six resolved resonances, one of which integrates to three 3-FPhe species,

making for a total of eight fluorophenylalanines. Single phenylalanine to tyrosine mutants of

five phenylalanine positions resulted in 19F NMR spectra with significant chemical shift

perturbations of the remaining resonances, and provided only a single definitive assignment.

Although 1H-19F heteronucleclear NOEs proved weak, 19F-edited 1H-1H NOESY connectivities

were relatively easy to establish by making use of the 3JFH coupling between the fluorine nucleus

and the adjacent fluorophenylalanine δ proton. 19F-edited NOESY connectivities between the δ

protons and α and β nuclei in addition to 15N-edited 1H,1H NOESY crosspeaks proved sufficient

to assign 4 of 8 19F resonances. Controlled cleavage of the protein into two fragments using

trypsin, and a repetition of the above 2D and 3D techniques resulted in unambiguous

assignments of all 8 19F NMR resonances. Our studies suggest that 19F-edited NOESY NMR

spectra are generally adequate for complete assignment without the need to resort to mutational

analysis.

5.2 Introduction

19F NMR is ideally suited to the study of changes in protein structure and dynamics

related to folding and unfolding, enzymatic action, ligand binding and internal motions

(Danielson and Falke 1996; Gerig 1994; Phillips et al. 1991). In comparison to traditional 1H-13C or 1H-15N two-dimensional NMR, 19F NMR approaches benefit from lack of background

signal, large chemical shift range, large scalar couplings for efficient magnetization transfer to 13C or 1H nuclei, and a high magnetogyric ratio (Gakh et al. 2000). Of particular importance in

protein studies, 19F NMR chemical shifts are sensitive to their local electrostatic environment and

van der Waals packing (Chambers et al. 1994; Feeney et al. 1996). A variety of mono-, di- and

per-fluorinated amino acids are commercially available, and are often cheaper than their 13C or

114

15N sidechain enriched counterparts. Furthermore, mono-fluorinated amino acids are introduced

with high efficiency (~95%) via biosynthetic means, often using non-auxotrophic strains with

little compromise to expression yields (Anderluh et al. 2005; Evanics et al. 2007). Structural

and functional perturbations arising from the incorporation of mono-fluorinated amino acids are

readily evaluated using a variety of techniques, most commonly NMR, circular dichroism

spectroscopy, and fluorescence; and are generally found to be minor (Li and Frieden 2007; Xiao

et al. 1998).

Despite the suitability of 19F NMR to the study of protein structure and dynamics,

assignments are generally established via site-directed mutagenesis techniques. In this study we

evaluate the merits of site-directed mutagenesis strategies and demonstrate the use of 19F and 15N-edited 2D and 3D NMR experiments in the assignment of 19F resonances associated with a

protein uniformly labeled with 3-fluorophenylalanine.

The most common assignment strategy used in 19F NMR studies involves direct

replacement of the fluorinated residue with a structurally similar amino acid; for example, each

occurrence of a fluorophenylalanine probe might be substituted by tyrosine using site-directed

mutagenesis techniques. Although in most cases the resulting perturbations are minor, these

effects are amplified in the corresponding 19F NMR spectra due to the sensitive nature of the

chemical shielding tensor. Frequently, the outcome of this replacement strategy precludes

complete spectral assignment due to gross chemical shift perturbations of the remaining fluorine

resonances (Feeney et al. 1996; Okano et al. 1998). Under such circumstances, the nudge

mutation method is often employed which involves the mutation of a residue known to be within

van der Waals contact of a particular fluorine probe to affect a change in it’s chemical shift

(Drake et al. 1993). This method, although often less perturbing, requires a priori knowledge of

the protein structure. In general, both strategies require a complete series of mutants, which is

costly and labor-intensive, while assignments using mutagenesis alone are often incomplete.

Supplementary NMR experiments, which focus on local topology and 19F relaxation properties,

can be employed to resolve ambiguities. For example, in a recent study of the native and

denatured states of green fluorescent protein, the authors complemented spectra from a complete

series of single replacement mutants with 19F T1-relaxation measurements, as well as 19F photo-

CIDNP data to achieve complete assignment of ten 3-fluorotyrosine resonances (Khan et al.

2006).

115

Herein, we consider a mutation-free approach to the assignment of 19F resonances from a

148 residue protein uniformly labeled with 3-fluorophenylalanine. Established triple resonance

side chain assignment strategies often rely on uniform carbon enrichment to facilitate

magnetization transfer between a given sidechain nucleus and those associated with the protein

backbone. Such a method has previously been developed for the assignment of 19F resonances in

protein studies which employ 3-fluorotyrosine (Kitevski-LeBlanc et al. 2009a). In this case, the

authors made use of a robust scalar coupling network between the 3-fluoro 19F, Cε, Cδ, and Hδ,

nuclei in the form of an HCCF 2D (or 3D) NMR experiment. The Cδ, and Hδ chemical shifts

were separately connected to the Cβ, Hβ chemical shift pairs (Yamazaki et al. 1993), which

were in turn correlated with backbone resonances via an HNCACB experiment.

In the current study, the fluoroaromatic probe used (3-FPhe) is not 13C, or 15N enriched,

precluding the above approach. However, the fluorine nucleus is scalar coupled to the δ, ζ, and

ε aromatic protons, with magnitudes in the 6 Hz to 10 Hz range, as shown in Figure 5.1B. The

available scalar coupling network can be used to connect 19F nuclei to the adjacent δ proton via

an INEPT transfer, while this δ proton resonance can in turn be connected to the β, α and HN

nuclei using 19F-edited 1H, 1H NOESY, and 15N-edited 1H, 1H NOESY experiments. The

importance of the delta proton to this strategy precludes the use of 2-FPhe, while a moderately

larger (~10%) 1H-19F coupling was the main advantage of the 3-FPhe over 4-FPhe. The

efficiencies of INEPT and NOE based magnetization transfers must be taken into consideration

and it is for this reason that we have chosen to demonstrate our assignment strategy using

Calmodulin (CaM), a medium sized protein whose backbone and sidechain assignments have

been thoroughly characterized.

Calmodulin is a ubiquitous calcium sensor protein which binds, and activates a variety of

enzymes involved in cell signaling pathways (Hoeflich and Ikura 2002). Upon coordination of

calcium, the protein undergoes a well-characterized conformational transition exposing

hydrophobic surfaces which mediate binding to a variety of target proteins, hormones and

peptides (Crivici and Ikura 1995). CaM is a 148 residue, acidic protein composed of two

structurally analogous domains (N and C-terminal) connected by a flexible linker (residues 73-

83) (Barbato et al. 1992). Each domain has two calcium binding EF-hand motifs composed of

two helices (E and F) flanking a 12-residue loop which coordinates calcium with pentagonal

116

bipyramidal symmetry (Malmendal et al. 1999). There are eight phenylalanine residues in

Xenopus laevis calmodulin, whose locations are indicated on the X-ray structure in Figure 5.1A.

These include five phenylalanines in the N-terminal domain (residues 1-72) F12, F16, F19, F65

and F68; and three in the C-terminal domain (residues 84-148) F89, F92 and F141.

Phenylalanines of CaM are highly conserved among species, and have been shown to contribute

to the stability of the hydrophobic core as well as target binding and activation (Okano et al.

1998). Phenylalanine 92, specifically, has been shown to play a pivotal role in the

conformational change associated with the apo to holo transition of the C-terminal domain such

that mutation of this residue to alanine results in reduced or complete loss of enzyme activation

(Meyer et al. 1996).

We begin with the careful analysis of structural perturbations resulting from the

incorporation of 3-fluorophenylalanine by comparing 1H-15N HSQC NMR spectra and calcium

binding constants of CaM in the presence and absence of the fluorine probe. We then asses the

difficulties of resonance assignment using the direct replacement of phenylalanine residues with

tyrosine by site-directed mutagenesis, which in our case provided only one assignment. Finally,

we conclude with the details of the assignment of our eight fluorine resonances using a series of 19F- and 15N-edited NMR experiments.

117

Figure 5.1 A) X-ray structure of calmodulin (PDB file 1CLL) showing the location of the eight

phenylalanine residues. B) 3-fluorophenylalanine structure with relevant scalar couplings

indicated. C) Secondary structural map of residues within the four EF hands of Xenopus laevis

calmodulin.

118


5.3.1 Expression and purification of uniformly 15N-enriched 3-fluorophenylalanine labeled CaM

Incorporation of DL-3-fluorophenylalanine via heterologous expression (Evanics et al.

2007) as well as purification of CaM (Ikura et al. 1990) was performed as previously described

with slight modifications. A plasmid (pET21b) encoding Xenopus laevis calmodulin (residues

1-148) was transfected into BL21(DE3) under control of the T7 promoter. LB broth inoculated

with a single colony was grown overnight and used to inoculate 1L of M9 minimal media

supplemented with 0.3% D-glucose, 0.1% 15NHCl4, 100 mg/L ampicillin, 10 mg/L thiamine, 10

mg/L biotin, 1 mM MgSO4, and 0.1 mM CaCl2. Uniform labeling with 3-fluorophenylalanine

was achieved by first allowing cell cultures at 37°C to reach an OD600 of 0.8, whereupon 1 g/L

glyphosate, 75 mg/L DL-tryptophan, and 75 mg/L DL-tyrosine is added. Once cell cultures

reached an OD600 of 1.0 (after approximately 1 hour), 35 mg/L DL-3-fluorophenylalanine was

added and expression was induced with the addition of 238 mg/L IPTG. Cell cultures were

harvested after 3.5 hours by centrifugation at 7000 rpm for 20 minutes. Cells were then

resuspended in 50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, 1 mM PMSF buffer at pH 8

and lysed by incubation at 4°C in the presence of 1 mg/mL lysozyme for 30 minutes, followed

by sonication. After the addition of DNase (10 µg/mL) and RNase (5 µg/mL), the suspension

was centrifuged at 9000 rpm for 20 min at 4°C and the cleared lysate was purified using Ni-NTA

Agarose resin (Qiagen, Mississauga, Ontario, Canada). The labeled protein was further purified

using phenyl sepharose as described previously (Ikura et al. 1990). Aliquots containing CaM

were pooled, concentrated and exchanged into 20 mM Bis-Tris, 0.1 M KCl, 9 mM CaCl2, 0.2%

NaN3 buffer at pH 8 using centrifugal concentrators for subsequent NMR experiments. Calcium-

free calmodulin samples were prepared as previously described (Zhang et al. 1995) for the

determination of macroscopic calcium binding constants. All buffers used in the preparation of

calcium-free CaM were made with water which had been decalcified using Chelex-100 resin and

stored in plastic bottles which had been treated with 5 mM ethylenediaminetetraacetic acid

(EDTA), followed by extensive rinsing with decalcified water. EDTA was added to purified,

dilute protein samples (~200-500 µM) to a final concentration of 20 mM, followed by

precipitation with trichloroacetic acid (TCA). The sample was redissolved in 25 mM NH4HCO3

119

and passed through a Sephadex G-25 column. Collected protein was then exchanged into 20

mM Bis-Tris, 0.1 M KCl, 2 mM EDTA, 0.2% NaN3 buffer at pH 7.5.

5.3.2 Site-directed mutagenesis

Single phenylalanine to tyrosine mutations were prepared for phenylalanine residues 12,

16, 19, 65 and 89. Site-directed mutagenesis was performed using the QuikChange Multi

protocol (Stratagene, Cedar Creek, TX ). Plasmid sequences were verified by automated DNA

sequencing (SequeTech, Mountain View, CA).

5.3.3 Trypsin Digest of CaM and 3-fluorophenylalanine CaM

Limited trypsin digestion of full length CaM and 3-fluorophenylalanine CaM was carried

out as previously described (Drabikowski et al. 1977). Trypsin (TPCK treated from bovine

pancreas; 10000-13000 BAEE U/mg) and soybean trypsin inhibitor (STI – 10000BAEE U/mg)

were purchased from Sigma Aldrich (Oakville, Ontario, Canada) and used without further

purification. In a typical digestion, 1 mM CaM was dissolved in 50 mM NH4HCO3, 50 mM

NaCl, 5 mM CaCl2 at pH 7.9 and then equilibrated at 37°C for 1 hour with 0.017 mM trypsin.

Digestion was quenched by the addition of 0.017 mM soybean trypsin inhibitor on ice. The

purification was carried out as detailed previously (Brokx and Vogel 2002). All NMR samples

were first concentrated, then buffer exchanged into 20 mM BIS-Tris, 0.1 M KCl, 9 mM CaCl2,

0.2% NaN3 buffer at pH 8.

5.3.4 Determination of macroscopic binding constants

Macroscopic calcium binding constants were measured using a competitive chelator

assay as described previously (Linse et al. 1991). 5,5’-Br2-BAPTA was obtained from

Molecular Probes (Eugene, Oregon). Calmodulin and 3-fluorophenylalanine labeled calmodulin

were decalcified as described above and titrated with 1 mM CaCl2 in the presence of 5,5’-Br2-

BAPTA in 20 mM Bis-Tris, 0.1M KCl, at pH 8, prepared with decalcified water at 25°C. The

titration was monitored using UV spectroscopy until the OD263nm indicated saturation. The raw

data was fit using Caligator software (Andre and Linse 2002) to obtain K1 through K4.

120



Technologies, Santa Clara, CA). For direct observe 19F NMR experiments and all 15N, 1H HSQC

correlation experiments, a 5 mm HCN/FCN triple resonance single gradient salt-tolerant

cryogenic probe, tunable to either 1H or 19F, was used with typical pulse lengths for 19F, 1H, and 15N of 9.5 µs, 9.25 µs, and 39 µs respectively. 1H-19F HSQC and 1H-19F NOESY-HSQC

experiments were performed on a 5 mm HFCN quad probe (Varian Inc., Walnut Creek, CA)

capable of simultaneous high power pulses on both 1H and 19F. Typical pulse widths for 19F and 1H used were 9 µs and 8.5 µs respectively. All protein sample concentrations were 1-1.5 mM in

90% H2O/10% D2O 20 mM Bis-Tris, 0.1 M KCl, 9 mM CaCl2 0.2% NaN3 at pH 8, and

experiments were run at 37°C. 19F 1D NMR experiments were collected with 1024 transients

using a spectral width of 8000 Hz and a 1.5 s recycle delay for 3-fluorophenylalanine labeled

CaM. 1H-19F HSQC experiments were modified from 1H-15N HSQC pulse programs from

Biopack software (Varian Inc.) and were typically collected with 256 transients and 96

increments spanning 3200 Hz in the indirect dimension. 1H-19F NOESY-HSQC experiments

were modified from the corresponding 1H-15N NOESY-HSQC pulse programs from Biopack

(Varian Inc) and were collected as 2D experiments in 2048 transients with 64 increments

spanning 5800 Hz in the first indirect dimension using a 400 ms mixing time. 1H-15N NOESY-

HSQC experiments from Biopack (Varian Inc.) were collected in 32 transients with 32 and 64

increments spanning 8000 Hz and 2400 Hz in the first and second indirect dimensions,

respectively. 1H Watergate NOESY spectra were collected in 64 transients with 256 increments

spanning 6500 Hz in the indirect dimension using a 200 ms mixing time. 1H-15N HSQC pulse

sequences were obtained from Biopack software (Varian Inc.) and typically were collected in 4

scans with 80 increments spanning 2000 Hz in the indirect dimension. Chemical shifts were

referenced using the Varian ‘setref’ macro, which in the absence of a standard, establishes the 1H, 15N, and 19F chemical shifts indirectly from the lock signal (Harris et al. 2002). All NMR

data were processed using NMRPipe/NMRDraw (Delaglio et al. 1995), and analyzed with

NMRView software (Johnson and Blevins 1994).

121

5.4 Results

5.4.1 Assessment of structural perturbations

To evaluate structural and functional consequences of 3-fluorophenylalanine

incorporation we rely on 1H-15N HSQC NMR spectra, and the measurement of calcium binding

constants. Gross structural perturbations are routinely identified by preparing uniformly 15N-

enriched protein and subsequently comparing 1H-15N HSQC NMR spectra of a protein with and

without fluorine labeling. Figure 5.2 shows an overlay of 1H-15N spectra of CaM with, and

without incorporation of 3-fluorophenylalanine. Note that the phenylalanine resonances are

completely absent from the 3-fluorophenylalaine labeled spectrum, indicating an enrichment of

at least 95% based on the standard deviation of the noise. 1H-15N HSQC spectra of 3-FPhe CaM

were indirectly assigned using previously published assignments for non-fluorinated calmodulin

(Torizawa et al. 2004). Residues adjacent to 3-FPhe’s were confirmed using 1H-15N NOESY-

HSQC spectra (as well as 1H NOESY spectra) which established correlations between these

residues and their ‘n+1’ neighbors. Although the chemical shifts of both fluorinated and non-

fluorinated CaM are similar, there are measurable perturbations in residues near the eight

phenylalanine sites, as well as considerable differences in line widths, suggesting that packing

and dynamics of the backbone have been altered by fluorine incorporation. The relative stability

of unlabeled and 3-FPhe labeled CaM was characterized using 1H-15N HSQC NMR spectra

collected as a function of temperature between 37°C and 80°C. The melting temperature of both

proteins were estimated by monitoring the intensities of a subset of well-resolved peaks, and

found to be approximately 76°C and 73°C for unlabeled and 3-FPhe labeled CaM, respectively.

Due to the high thermal stability of calmodulin and temperature limitations of the NMR probe,

we were unable to fully characterize the unfolding and we emphasize that these values are

estimates. However, it is clear that uniform labeling with 3-FPhe has not caused a significant

decrease in the thermal stability.

122

Figure 5.2 1H-15N HSQC spectra of 3-fluorophenyalanine labeled calmodulin overlaid with an

analogous spectrum of non-fluorinated calmodulin.

It is also possible to assess structural perturbations from the perspective of calcium

affinity, for both non-fluorinated and 3-FPhe labeled CaM. The binding constants for non-

fluorinated and 3-FPhe CaM are K1 = 2.4 × 105 M-1, K2 = 1.1 × 106 M-1, K3 = 3.8 × 104 M-1, K4 =

1.2 × 105 M-1 and K1 = 2.6 × 105 M-1, K2 = 1.5 × 106 M-1, K3 = 8.6 × 104 M-1, K4 = 1.7 × 105 M-1

respectively. The minor difference in K3, which represents a calcium coordination site in the N-

terminal domain, may be due specifically to F65, which occupies position 10 of the calcium

binding loop of EF-hand 2 (Figure 5.1C). Alternatively, this may simply be a cumulative effect

arising from the addition of five 19F species in the hydrophobic core of the N-terminal domain as

opposed to only three such species in the C-terminal domain. Overall, these values agree fairly

well with those previously published (Linse et al. 1991), leading us to conclude that calcium

binding in the presence of 3-fluorophenylalanine is effectively unperturbed under the current

experimental conditions.

123

5.4.2 Site Directed Mutagenesis

Single phenylalanine to tyrosine mutants were prepared for residues 12, 16, 19, 65 and 89

of CaM. The 19F NMR spectra of each mutant, aligned with the wild type spectrum, are shown

in Figure 5.3. There are six resolved resonances in the wild type spectrum, with the most

downfield peak integrating to roughly three times the intensity of the others.

Figure 5.3 19F NMR spectra of wild type (wt) and single phenylalanine to tyrosine mutants of

calmodulin uniformly labeled with 3-fluorophenylalanine. All spectra were run under identical

experimental conditions, as described in the materials and methods section.

The relatively small chemical shift dispersion is perhaps not surprising considering that the N-

and C-terminal domains of CaM each consist of two, predominantly helical, EF-hand motifs as

shown in Figure 5.1A. Several pairs of phenylalanine residues occupy identical positions within

the EF-hand motif (Figure 5.1C). For example, residues F89 and F16, F68 and F141 as well as

F92 and F19, which are likely to experience similar environments. The spectra of the single

124

phenylalanine to tyrosine mutants all exhibit substantial chemical shift perturbations (See Figure

5.3). Although the substitution of a phenylalanine residue to a tyrosine is the most natural choice

among the available standard amino acids, there is a substantial change in both size and polarity,

which is perhaps magnified when considering residues within the protein core. The average

increase in volume for a buried phenylalanine to a buried tyrosine has been estimated to be only

0.2 Å3 (Richards 1977). Considering the high density of native protein cores, where packing

efficiency often exceeds 75% and is generally higher for α-helical structures (Woolfson 2001),

this small increase in volume has the potential to cause significant disruptions in the geometric

arrangement of core residues. The burial of polar groups in the non-polar interior of a protein is

known to be destabilizing due to a large enthalpy of dehydration; however, this effect can be

compensated for if the polar group forms a non-native hydrogen bond (Loladze et al. 2002). In

general, mutations involving the rearrangement of hydrophobic core interactions reduce the so-

called conformational uniqueness of a protein which has the potential to affect both structure and

function (Willis et al. 2000). In addition, the above effects are likely to be amplified in

situations where the residues under consideration are clustered in regions which share van der

Waals contacts (Drake et al. 1993). Considering the complex interplay of effects, it is likely that

the success of mutagenesis approaches is both probe, and protein dependant.

The 19F NMR spectra of the single phenylalanine to tyrosine mutants of CaM reveal

significant variation in chemical shift, line width, and minor conformers, as shown in Figure 5.3.

While it is possible the Phe to Tyr substitutions have altered the calcium affinity, it is doubtful

that under the experimental conditions (9:1 CaCl2:CaM) there is a significant population of

protein in the apo form. A marked decrease in calcium affinity has been observed in previous

studies of a E140Q C-terminal calmodulin mutant (Evenas et al. 1997), but unlike residue 140

which is a calcium coordinating ligand, none of the phenylalanine residues are involved in

calcium coordination, or, apart from F65, located in the calcium binding loops.

In an effort to evaluate the relative perturbations introduced by the single replacement

mutations we considered 19F NMR difference spectra between the wild type and mutant species.

After normalizing the wild type spectrum to 8, and the single mutant spectra to 7, the integral of

the difference spectrum is expected to be close to 1, assuming global perturbations are minimal.

In all cases, the difference spectra gave integrals of ~4 suggesting that the perturbations are

global; making spectral assignment using single replacement mutants impractical. 1H-19F HSQC

125

NMR spectra of single mutant samples were collected in an attempt to improve resolution by

relying on both 19F and 1H chemical shifts. Unfortunately, several residues have coincident or

very similar aromatic proton chemical shifts, which combined with the overlap and perturbations

in the 19F dimension provides no additional assignment information. With the exception of

F65Y, these perturbations prohibit conclusive assignment of the fluorine resonances. Some

preliminary assignments may be gleaned from F89Y and F16Y, where it is clear that in both

cases the intensity of the most downfield peak is attenuated. These observations combined with

recognizing the identical positioning of these two residues in the EF-hand motif (Figure 5.1C)

allow for a tentative assignment of this resonance to F89 and F16. The fact that significant

perturbations are observed in several of the single mutants is not surprising as, particularly for

the N-terminus, the phenylalanine residues are clustered together within the core of the domain

and any disruption in one position is likely to affect the others through van der Waals

interactions. It is clear that complete, unambiguous assignments cannot be established using

site-directed mutagenesis alone.

5.4.3 NMR-based assignment strategy

In the absence of 13C-enriched 3-fluorophenyalanine, one can establish a connectivity

between the 19F probe and backbone nuclei by relying on scalar couplings between the fluorine

nucleus and the ε, ζ , and the adjacent δ aromatic protons. The δ aromatic proton can then be

correlated to the backbone resonances using NOESY-based magnetization transfer as described

below. The intraresidue coupling between the 19F nucleus and the adjacent δ proton is clearly

delineated via a 1H-19F HSQC, as shown in Figure 5.4A. A 19F 1D NMR spectrum is displayed

along the indirect axis for clarity.

126

Figure 5.4 NMR-based assignment strategy for phenylalanine 92. A) 1H-19F HSQC spectrum

with 19F 1D trace along the indirect detect fluorine axis. The ε, δ, and ζ proton crosspeaks are

indicated for F92. B) 1H-19F NOESY-HSQC spectrum. The experiment was collected as a 2D

(i.e. without evolving the 19F dimension) and shown are the first indirect and direct detect

dimensions. The crosspeak highlighted corresponds to the beta protons of F92. C) 1H-15N

NOESY-HSQC spectrum. All three dimensions of the experiment were collected. Shown is the

nitrogen plane corresponding to 117.21 ppm. The crosspeak indicated corresponds to the amide

proton of aspartic acid 93. All experiments were run at 37°C. Additional experimental details

can be found in the materials and methods section.

For each fluorine resonance, the ε, ζ, and the adjacent δ proton are resolved due to the similarly

sized scalar coupling between these nuclei and the fluorine atom (see Figure 5.1B). Note that we

also attempted to establish such a correlation via a 1H-19F heteronuclear NOE (Rinaldi 1983);

however, despite the small couplings, the HSQC proved to exhibit the greatest fidelity. The next

step in the assignment protocol involves connecting the δ proton to the intraresidue β protons

127

using a 1H-19F NOESY-HSQC, which was collected as a 2D spectrum as shown in Figure 5.4B.

Due to the relatively few fluorine-coupled δ protons and their generous chemical shift dispersion,

the 19F dimension was not collected, but simply used to edit the spectrum. The observed

crosspeaks were compared with those in a 1H,1H NOESY spectrum, which aided in the

verification of the final assignments. As an example of the assignment protocol, we trace the

connectivity network between the 19F resonance of F92 (-114.68 ppm) and the amide proton of

the adjacent residue, aspartic acid 93. Beginning with the chemical shift of the fluorine coupled

δ proton (7.13 ppm) we find a crosspeak between this resonance and a β proton whose chemical

shift is 2.59 ppm, as shown in Figure 5.4B. This β proton resonance is then correlated to the

backbone using a 1H-15N NOESY-HSQC experiment, where assignments of the 1H-15N plane are

either transferred from previously published data, or obtained experimentally using a sample

uniformly enriched in 15N, and 13C. Because the fluorinated phenylalanine probes are not 15N

enriched, the β protons were linked to the amide proton and nitrogen chemical shifts of the

neighboring, n+1 residue. Continuing with the assignment of the example resonance, a distinct

crosspeak is observed between a β proton at 2.59 ppm and an amide proton with a chemical shift

of 7.81 ppm in the nitrogen plane corresponding to 117.21 ppm, as shown in Figure 5.4C. The

n+1 residue was identified as aspartic acid 93 from previously published assignments (Torizawa

et al. 2004); confirming that the associated fluorine resonance arises from phenylalanine 92. The

above steps were repeated for the remaining resonances with details provided in the

supplementary material. Of the eight residues, there were four crosspeaks observed in the 1H-19F

NOESY-HSQC, resulting in the assignments of phenylalanine residues 68, 89, 92 and 141.

While the strategy is quite straightforward, the methodology is limited by the dependence on

dipolar magnetization transfer, which is highly influenced by dynamics. The 19F 1D NMR

spectra of 3-FPhe labeled CaM (see Figure 5.3) reveal a significant variation in resonance line

widths as well as the presence of minor conformers, indicating the prevalence of chemical

exchange over a range of dynamic timescales, which may reduce the magnitude of NOESY-

based magnetization transfer for several residues. The weak NOEs in combination with the

relatively small scalar couplings (JHF = 9.9 Hz) available for INEPT transfers are likely factors in

the reduced number of crosspeaks observed in the 1H-19F NOESY-HSQC. Apart from

phenylalanine 68, the crosspeaks obtained all come from the C-terminal domain of CaM. It is

possible that the dynamics of the N-terminal domain are such that the NOE transfers are weak,

and therefore unobservable within a reasonable amount of experimental time.

128

5.4.4 Trypsin fragments of calmodulin

To clarify resonance assignments, we separately examined 19F NMR spectra of the N-

terminal and C-terminal domains of calmodulin upon proteolytic cleavage in the linker region.

Controlled digestion of calmodulin in the presence of calcium with trypsin produces two

fragments (Drabikowski et al. 1977), commonly referred to TR1C and TR2C, composed of the

N-terminal (residues 1-73) and C-terminal (84-148) lobes respectively. The resulting fragments

retain identical secondary and tertiary structural features of the full length protein (Finn et al.

1995; Ishida et al. 2000) as well as calcium binding affinity (Linse et al. 1991), but exhibit

reduced or absent activation ability on target proteins (Walsh et al. 1977). An overlay of 1H-15N

HSQC spectra of 3-FPhe TR1C and 3-FPhe TR2C with non-fluorinated calmodulin are in good

agreement with respect to the chemical shifts, as shown in Figures 5.5A and 5.5B. 19F NMR

spectra of TR1C and TR2C are shown in Figure 5.6, aligned with an analogous spectrum of the

full length protein. The 19F NMR spectra of TR1C and TR2C have five and three major

resonances resolved respectively, in addition to some minor peaks which may represent

secondary conformations in slow exchange on the 19F NMR timescale. An interesting feature of

the fragment spectra is the increased resolution in TR1C of the two downfield peaks. These two

peaks, along with the most downfield peak in TR2C have coincident chemical shifts in the full

length protein. The discrepancy between the fluorine chemical shifts in the full and TR1C

spectra suggest that there is an interaction between the two domains in the context of the full

length protein which affects the local environment of the fluorine probes. Careful examination

of the 1H-15N HSQC spectra of both 3-FPhe TR1C and full length CaM reveal two distinct

conformations with close to equal populations for a number of residues, as shown clearly for

glycine 33 in Figure 5.5C. The two conformations appear to correspond to slow exchange

between native-like states, as evidenced by the similarity of the chemical shifts for G33 in TR1C

and non-fluorinated CaM, as compared to 3-FPhe CaM. There is precedence in the literature for

large scale fluctuations about a flexible region of the long interconnecting helix (Baber et al.

2001; Bertini et al. 2004), as well as a recent crystal structure of a closed, globular form of the

protein where the two domains are within contact (Fallon and Quiocho 2003). It is likely that in

solution the classic barbell topology and this more recent collapsed, globular structure represent

two possible conformational extremes which can be accessed in solution, while the presence of

3-fluorophenylalanine may alter the energetics of the system, and thus the conformational

129

equilibrium in favor of a species where the two terminal domains interact. A more complete

investigation of the observed disparity in the 19F and 1H-15N NMR spectra TR1C and the full

length protein is currently being investigated.

Figure 5.5 1H-15N HSQC spectra of 3-FPhe TR1C (A) and 3-FPhe TR2C (B) overlaid with an

analogous spectrum of non-fluorinated calmodulin. The indicated region of the spectrum

enclosing glycine 33 is magnified in C) with the corresponding region from an 1H-15N HSQC of

3FPhe CaM shown in blue.

130

The NMR based assignment strategy, utilizing 1H-19F HSQCs, 1H-19F NOESY-HSQCs

and 1H-15N NOESY-HSQCs, was repeated on TR1C (data provided in supplementary material),

and full assignments were obtained and transferred to the full length protein spectrum. The fully

assigned 19F NMR spectra of TR1C, TR2C and 3-FPhe-CaM are shown in Figure 5.6.

Figure 5.6 Assigned 19F NMR spectra of uniformly 3-FPhe labeled TR1C, CaM and TR2C.

5.5 Conclusions and Final Remarks

In this paper we have described an NMR based strategy for the assignment of resonances

associated with mono-fluorinated aromatic amino acid probes in proteins, in the absence of

sidechain 13C and 15N enrichment. We take advantage of the prominent scalar couplings

between the 19F probe and the aromatic protons to correlate the fluorine nucleus to the ε, ζ, and

the adjacent δ protons via the 1H-19F HSQC. Fluorine coupled δ protons are then connected to

intraresidue β protons using a 1H-19F NOESY-HSQC. In the final step, the β protons are

associated with the n+1 residue amide and nitrogen chemical shifts to complete the assignment

via a 1H-15N NOESY-HSQC. In principle the basic steps applied here can be used to assign

fluorinated aliphatic residues and other unnatural amino acids by adjusting experimental

parameters for the NMR active nucleus at hand. The dependence of this assignment strategy on

131

NOESY-based magnetization transfer makes it particularly sensitive to dynamics and additional

relaxation mechanisms, such as chemical exchange and chemical shift anisotropy (CSA)

common to fluoraromatics, which reduce the magnitude of the NOE effect.

Discrepancies between the 19F and 1H-15N chemical shifts of TR1C with that of the full

length are an indication of a conformational exchange process related to interdomain

interactions. This is particularly interesting in light of previous reports indicating large-scale

fluctuations about the flexible linker and the existence of transient domain-domain interactions.

The minor conformers observed in both the full-length protein and the trypsin fragments reflect

motions such as ring flipping or conformational reorganization, which are slow on the chemical

shift timescale. Attempts to investigate these properties using 19F T1 and CPMG T2 experiments

at several temperatures and field strengths were performed. However, control experiments using

fractionally fluorinated CaM revealed that these intermediate and slow motions were exclusive to

the fully 3F-Phe-labeled CaM. While the assignment protocol outlined above is robustly

applicable to any scenario wherein the protein is 19F- and 15N-labeled, the details of the structural

and dynamic characteristics of fully and fractionally 19F labeled CaM will be outlined in a

forthcoming manuscript.

132

Chapter 6

Optimizing 19F NMR protein spectroscopy by fractional biosynthetic labeling

The work presented in this chapter has been accepted for publication in “Kitevski-

LeBlanc, J. L., Evanics, F., Prosser, R. S. (2010). Optimizing 19F NMR protein spectroscopy by

fractional biosynthetic labeling. Journal of Biomolecular NMR”. My role in this project

consisted of sample preparation, setting up NMR experiments and processing and analyzing data.

NMR experiment set up was performed with Dr. Ferenc Evanics. © Reprinted with permission

from Springer.

133

6.1 Abstract In protein NMR experiments which employ nonnative labeling, incomplete enrichment is

often associated with inhomogeneous line broadening due to the presence of multiple labeled

species. We investigate the merits of fractional enrichment strategies using a monofluorinated

phenylalanine species, where resolution is dramatically improved over that achieved by complete

enrichment. In NMR studies of calmodulin, a 148 residue calcium binding protein, 19F and 1H-15N HSQC spectra reveal a significant extent of line broadening and the appearance of minor

conformers in the presence of complete (>95%) 3-fluorophenylalanine labeling. The effects of

varying levels of enrichment of 3-fluorophenylalanine (i.e. between 3% and >95%) were further

studied by 19F and 1H-15N HSQC spectra,15N T1 and T2 relaxation measurements, 19F T2

relaxation, translational diffusion and heat denaturation experiments via circular dichroism. Our

results show that while several properties, including translational diffusion and thermal stability

show little variation between non-fluorinated and >95% 19F labeled samples, 19F and 1H-15N

HSQC spectra show significant improvements in line widths and resolution at or below 76%

enrichment. Moreover, high levels of fluorination (>80%) appear to increase protein disorder as

evidenced by backbone 15N dynamics. In this study, reasonable signal to noise can be achieved

between 60-76% 19F enrichment, without any detectable perturbations from labeling.

6.2 Introduction

Fluorinated amino acids and amino acid analogues are ubiquitously used to understand or

manipulate the physical properties of proteins (Akcay and Kumar 2009; Danielson and Falke

1996; Geddes 2009). In the latter case, it is the change in the physical and chemical properties of

the protein that are of interest. For example, fluorinated leucine and valine are known to act as

strong helix stabilizers (Bilgiçer et al. 2001) while both fluorinated aromatics and aliphatics have

been shown to stabilize the protein fold and provide protection from heat-denaturation and

proteolysis (Tang et al. 2001a; Woll et al. 2006). This stabilization is attributed to the increased

hydrophobicity of fluorocarbons relative to hydrocarbons, and the preference for fluorine-

fluorine interactions, which are stronger than the classic hydrophobic effect. In the majority of

spectroscopic applications, the goal is to observe the protein in its most unadulterated form – that

is, free from probe-induced structural, functional or dynamic perturbations.

134

NMR and fluorescence studies of proteins often make use of fluorinated amino acids

which are generally assumed to be isosteric with their native counterparts. Such studies may

investigate binding (Anderluh et al. 2005; Broos et al. 2004; Luck and Falke 1991), catalysis

(Quint et al. 2006; Rozovsky et al. 2001) and folding or unfolding (Bann et al. 2002; Li and

Frieden 2007; Schuler et al. 2002; Visser et al. 2009). Indeed, in the majority of applications the

usual mantra is that because the fluorine atom is less than 20% larger than that of hydrogen there

is little perturbation to be concerned with, particularly when the substitution involves mono-

fluorinated amino acids. However, to strictly be able to claim that there are no perturbations

resulting from such incorporation, a stringent series of functional and structural investigations

should be performed as a function of enrichment of the fluorinated amino acid(s). Several

authors have observed probe-induced perturbations arising from fluorinated aromatics to varying

degrees (Duewel et al. 2001; Luck and Falke 1991; Xiao et al. 1998), though a systematic study

of enrichment has to our knowledge not been conducted. Although the van der Waals radius of

the fluorine atom is comparable to that of hydrogen, the packing density of the protein

hydrophobic core may exceed 75% (Woolfson 2001), leaving little room for modified aromatics.

While there are many examples where fluorinated residues do not perturb protein structure

(Campos-Olivas et al. 2002), some caution is warranted, particularly in studies of intermediates

or in situations where the fluorine probes are concentrated in the protein interior.

In 19F NMR studies of proteins, full enrichment is desired to avoid multiple conformers

and the predominance of inhomogeneous line broadening. We show in this study that exactly the

opposite is the case; namely, lower enrichment leads to substantially better spectra, with fewer

minor conformers and narrower line widths. Specifically, we examine structural and dynamic

perturbations arising from biosynthetic labeling of calmodulin (CaM) with 3-fluorophenylalanine

(3-FPhe). We then explore the utility of fractional labeling through the combined enrichment of

phenylalanine and 3-fluorophenylalanine, and the consequences to NMR spectra as well as

thermal stability of CaM. Ideal labeling levels were achieved at or below 76% random

enrichment of 3-FPhe, where perturbations were found to be minimal while NMR experimental

time was not significantly hampered.

135


6.3.1 Protein Expression and Purification

Preparation of Xenopus laevis CaM (residues 1-148) uniformly enriched with 15N and

D/L-3-fluorophenylalanine was performed using heterologous expression techniques as

previously described (Kitevski-LeBlanc et al. 2009b) with modifications at the time of induction

to achieve fractional labeling. All chemicals used in protein sample preparation and were

obtained from Sigma Aldrich (Oakville ON, Canada). In general, full (>95%) 3-FPhe labeling is

achieved by the addition of 1 g/L of glyphosate with 70 mg/L D/L-tryptophan and 70 mg/L of

D/L-tyrosine at an OD600 nm=0.8. After 1 hour, cell cultures typically reach an OD600 nm=1.0, at

which point 35 mg/L of D/L-3-fluorophenylalanine is added followed by agitation and induction

of target protein expression using IPTG. Fractional labeling was achieved by adjusting the ratio

of D/L-3-fluorophenylalanine to D/L-phenylalanine added prior to induction. For example, 3%

3-FPhe enriched calmodulin requires 25% D/L-3-fluorophenylalanine (i.e. 8.75 mg of D/L-3-

fluorophenylalanine and 26.25 mg D/L-phenylalanine). Calmodulin samples corresponding to

6%, 60%, 76% and > 80% incorporation of the fluorinated amino acid required 50% 3-FPhe,

75% 3-FPhe, 85% 3-FPhe, and 98% 3-FPhe enrichment of the total phenylalanine supplied to the

media, respectively. The D/L-3-fluorophenylalanine and D/L-phenylalanine mixtures are

suspended in a minimal volume of M9 media (but not fully dissolved) immediately prior to

addition to shaking cultures. Expression cultures are then agitated until the added amino acids

are fully dissolved (< 5 minutes) prior to the addition of IPTG. Samples for NMR were

exchanged into 10 mM BIS-TRIS, 0.1 M KCl, 9 mM CaCl2, 0.2% NaN3 buffer adjusted to pH 8.

Samples used for NMR also contained 50 µM D/L-4-fluorophenylalanine as an internal standard

for referencing and determination of effective fluorinated protein concentration from 19F NMR

experiments.

Fluorine enrichment levels were carefully determined in two steps. First by integration of 19F NMR signal of the 3-FPhe enriched protein, relative to a 50 µM D/L-4-fluorophenylalanine

internal standard, we obtain an effective fluorine labeled protein concentration. Then, the total

protein concentration in a given NMR sample was determined using the bicinchoninic acid

(BCA) protein assay kit (Sigma Aldrich, Oakville, Ontario, Canada). Enrichment levels were

then obtained by taking the ratio of the effective fluorine labeled protein concentration (from

136

integration of 19F NMR) to the total protein concentration (from the BCA protein assay). The

estimation of labeling levels using this method is difficult above 76% enrichment due to the

effects of inhomogeneous broadening and the presence of minor conformers, which are more

difficult to reliably integrate in the 19F NMR spectra. Determination of 19F enrichment in fully

labeled CaM samples also suffer from line broadening and multiple conformers. However, such

samples are estimated to be at least 95% labeled based on the standard deviation of the noise in

the corresponding 1H-15N HSQC spectra, which lack cross-peaks arising from fluorinated

phenylalanine residues. Although mass spectrometry has been used successfully in previous

studies to obtain quantitative estimates of fluorine enrichment, we were unable to find conditions

which provided a CaM sample amenable to quantification of labeling levels using this approach.



Technologies, Santa Clara, CA) equipped with a 5 mm HCN/FCN triple resonance single

gradient salt tolerant cryogenic probe. All NMR samples were approximately 1-1.5 mM in

concentration and were maintained at 37°C. 1H-15N HSQC, 15N spin-lattice (T1) relaxation

times, and transverse (T2) relaxation time were measured using pulse programs obtained from

Biopack software (Agilent Technologies, Santa Clara, CA) and were collected in 8 scans with 80

increments spanning 2000 Hz in the indirect dimension. Typical pulse widths for 1H and 15N

were 9.2 µs and 39.5 µs respectively. 15N spin-lattice (T1) and transverse (T2) time constants

were measured using 7 relaxation delays ranging from 0.01 to 0.8 s and 0.01 to 0.13 s

respectively. Peak volumes were fit to a single exponential decay function with average standard

error of 1.7%, 1.5%, 1.1% and 2.2% (1.1%, 1.1%, 1.1%, and 1.8%) for T1 (T2) data associated

with 0%, 6%, 60% and >95% enrichment respectively. 15N-{1H} steady-state NOEs were

measured by collecting spectra in the presence and absence of a 250 ms mixing time. NOE

values were determined as the ratios of peak intensities measured from spectra with and without

proton irradiation and the uncertainty in peak heights were given by the standard deviation of the

baseplane noise in the spectra. Uncertainties of the NOE values were obtained by propagating

the peak height uncertainty (Nicholson et al. 1992a). Diffusion experiments were performed at

37°C using a one-dimensional version of the BPPSTE HSQC as described previously

(Rajagopalan et al. 2004). Gradient pulse amplitudes were calibrated from the known diffusion

coefficient of HDO at 25°C (Holz and Weingartner 1991). Fifteen gradient amplitude

137

experiments were acquired for each data set with gradient strength varied from 6% to 92% of its

maximum power, and all other experimental parameters held constant. Gradient pulses were

applied for 2 ms with a recovery time of 0.2 ms, and diffusion delay of 36.8 ms, affording a total

signal decay of ~90%. Spectra were processed using VnmrJ and peak heights were obtained with

the peak deconvolution program “fitspec”. Diffusion coefficients were obtained from the slope

of the natural logarithm of normalized peak intensity plotted against the corresponding square of

the gradient strengths, and converted to radius of hydration using the Stokes-Einstein equation

and the density of pure water (Altieri et al. 1995). 19F NMR experiments were collected using a

repetition time of 1.5 s, a spectral width of 8000 Hz and a typical pulse width of 14 µs. The

number of transients used to achieve sufficient signal varied among the fractionally labeled

samples, from 30720 for a 3% 3-FPhe enriched sample, to 1024 for samples with > 80% fluorine

enrichment. 19F transverse (T2) relaxation times were measured using a Hahn echo sequence

(90°-τ-180°-τ) employing 9 τ values ranging from 0.8 ms to 11.2 ms. 19F line widths were

estimated from 19F NMR spectra where inhomogeneous magnetic field contributions were

deemed constant among samples when the line width of the internal standard (D/L-4-

fluorophenylalanine) was within 1 Hz between samples. All NMR data was processed using

NMRPipe/NMRDraw (Delaglio et al. 1995) and analyzed with NMRView (Johnson and Blevins

1994) unless stated otherwise.

6.3.3 Circular Dichroism Spectroscopy and Thermal Denaturation

Thermal stability of 0%, 60%, 76% and >95% 3-FPhe labeled CaM were measured under

calcium free conditions, due to the high thermal stability of calcium saturated calmodulin.

Preparation of apo-calmodulin samples was performed as described elsewhere (Kitevski-

LeBlanc et al. 2009b). Far UV CD spectra were acquired on an Aviv CD spectrometer model

62DS at 25°C. Spectra of 10-20 µM protein, in 0.1 M KCl and 20 mM Tris buffer adjusted to

pH 8, were collected from 195 nm to 260 nm (path length, 0.1 cm; steps, 1 nm; bandwidth, 1 nm;

and averaging time 12 s). Thermal denaturation experiments were performed by measuring

ellipticity at 222 nm from 15°C to 85°C at a rate of 1°C/ minute. Reversibility of thermal

denaturation was assessed to be >90% following cooling to 15°C at a rate of 1°C/min.

138


6.4.1 Full and Fractional labeling of Calmodulin: Effects on 19F NMR spectra

Calmodulin is a ubiquitous calcium sensor protein which binds and activates a variety of

enzymes in response to an increase in intracellular calcium levels (Hoeflich and Ikura 2002).

CaM is a 148 residue, acidic protein organized into two structurally similar calcium binding

domains connected by a flexible linker. Four calcium ions are coordinated by a pair of canonical

EF-hand motifs in each of the N-terminal (residues 1-73) and C-terminal (83-148) domains with

pentagonal bipyramidal symmetry. Upon calcium coordination a well-characterized

conformational transition occurs exposing hydrophobic patches in each domain for protein

binding (Crivici and Ikura 1995). There are eight phenylalanine residues in Xenopus laevis

CaM, whose positions are indicated in the ribbon diagram representation of the X-ray structure

in Figure 6.1A. These residues (i.e. F12, F16, F19, F65, and F68 in the N-terminal domain and

F89, F92, and F141 in the C-terminal domain) are highly conserved among species.

Phenylalanine residues contribute to the stabilization of the hydrophobic core and have been

implicated in peptide binding and activation (Okano et al. 1998). CaM samples enriched with 3-

FPhe at 3%, 6%, 60%, 76%, >80% and >95%, were prepared as described in the Materials and

Methods section, and the resulting 19F NMR spectra are shown in Figure 6.1B. Resonance

assignments shown were determined in an earlier study (Kitevski-Leblanc et al. 2010).

139

Figure 6.1 A) X-ray structure of calmodulin (PDB file 1CLL) showing the location of the eight

phenylalanine residues. B) 19F NMR spectra of calmodulin enriched with 3% to >95% 3-FPhe.

Peak assignments were obtained previously (Kitevski-LeBlanc et al.).

The enrichment levels, which were obtained using a protocol described in the Materials

and Methods section, do not correlate well with the ratios of 3-FPhe to Phe added, in the series of

proteins examined. Although it is known that an aminoacyl-tRNA synthetase will exhibit a

preference for the corresponding native amino acid, and that this preference varies among, at

least the aromatic amino acids (Luck and Falke 1991), one would expect a trend in the fraction of

a fluorinated analogue added and the incorporation level obtained. Higher concentrations of 3-

FPhe might cause significant stalling of protein translation, assuming tRNA synthetase off rates

of 3-FPhe and Phe are different, unforeseen toxicity issues, solubility issues, or any number of

variables. If fractional labeling were to be attempted in another system, a range of fluorinated

amino acid to natural amino acid ratios would need to be tested.

140

As shown in Figure 6.1B, it is evident from the 19F NMR spectra that labeling levels

greater than 80% result in significant overlap of resonances, particularly of F12, F16 and F89,

and a great range of peak line widths as well as the presence of additional minor conformers. The

possibility that the minor peaks observed correspond to unequal populations of two

conformations associated with 180° rotation about the Cβ-Cγ bond of the 3-FPhe residues can be

addressed by comparison to fluorine spectra obtained using a symmetric probe, such as 4-

fluorophenylalanine, which, from the perspective of the fluorine nucleus, would not be expected

to distinguish between these two conformations of the aromatic ring. Indeed 19F 1D and 1H-15N

HSQC spectra (Supplementary Figures 6.4 and 6.5) of CaM labeled with >95% 4-

fluorophenylalanine reveal a range of peak line widths, additional minor conformers and

chemical shift perturbations, as observed with high levels of 3-FPhe labeling. Note that there

does not appear to be any single problematic residue from the perspective of enrichment. Rather,

both domains of CaM exhibit broadening and multiple conformers and these cumulative

perturbations associated with enrichment of 3-fluorophenylalanine are significantly attenuated at

or below 76% enrichment. From the perspective of spectral quality and experimental time, a

convenient enrichment level is on the order of 66%.

An obvious concern with partial labeling strategies is the presence of multiple

conformers, each corresponding to a specific pattern of label incorporation. For 19F NMR in

particular, the presence of multiple conformations may result in additional peaks producing

complicated spectra, or an increase in line widths due to inhomogeneous broadening. Spectral

line widths are characterized by homogeneous and inhomogeneous contributions arising from

both dynamics, slow exchange between spectroscopically distinct states, and to some extent,

field inhomogeneities. To characterize the inhomogeneous contribution to the observed 19F line

widths at various levels of 19F incorporation we compared 19F line widths to 19F transverse

relaxation (1/T2) rates. Contributions from inhomogeneity in the magnetic field were kept nearly

constant between protein samples by comparing the line width of an internal standard where we

assumed this contribution to be equal among samples when the standard peak line width was

within 1 Hz between samples. The 1/T2 and line width data for each residue in the 76% and

>95% labeled samples are shown in Table 6.1.

Table 6.1 19F transverse relaxation data (1/T2), line widths (1/T2*) and the difference [Δ(1/T2) =

1/T2* -1/T2] for >95% and 76% fluorine labeled calmodulin.

141

The data suggest that 19F line widths are on average reduced by almost 50% under conditions of

fractional labeling. This is a significant point, as a reduction in peak widths by half translates to

a potential 2 increase in sensitivity and enhanced spectral resolution.

The origin of inhomogeneous line broadening is of concern, particularly in situations

where fractional labeling strategies are employed. Assuming that inhomogeneous line

broadening arises primarily from the co-existence of CaM species with randomly distributed

fractional fluorine enrichment, one would expect the greatest line widths to be observed at ~50%

fluorine labeling efficiency where there are 28 equally probable species in solution. However, as

shown in Figure 6.1, 19F line widths at 60% fluorine labeling are very similar to those in all

fractionally labeled samples below 80%. Moreover, inhomogeneous broadening, which might be

expected to result from a multitude of 19F-labeled states, is actually smaller at 76% than that seen

at >95% enrichment (i.e. Δ1/T2* is on average 76.4 Hz versus 170.3 Hz for 76% and >95%

enrichment, respectively). Line widths are independent of labeling efficiency between 3% and

76% enrichment, suggesting that the existence of fractional enrichment levels has little or no

effect on line broadening. Rather, the marked increase in both homogeneous and inhomogeneous

142

broadening observed above 80% enrichment is likely a result of intermediate and slow

conformational dynamics resulting from packing perturbations.

6.4.2 Effects on 1H-15N HSQC spectra and 15N dynamics 1H-15N HSQC spectra of non-fluorinated and fluorinated versions of the protein are often

compared as a qualitative measure of structural perturbations originating from the fluorinated

amino acid probe. In general, spectral discrepancies between non-labeled and labeled proteins

are minor and when assignments are available, they are largely associated with residues in the

vicinity of the fluorine probe. An overlay of 1H-15N HSQC spectra of 60% and >95% 3-FPhe

enriched CaM, each with an analogous spectrum of the non-fluorinated protein are shown in

Figure 6.2. It is immediately clear that the >95% 3-FPhe enriched protein exhibits increased line

widths, as well as significant chemical shift perturbations, most of which are absent in the 60%

3-FPhe enriched sample. The normalized chemical shift perturbations for 6%, 60% and >95% 19F labeled 3-FPhe CaM are plotted as a function of residue in Figure 6.3A; and reveal moderate

perturbations at low levels of incorporation and dramatic perturbations at >95% 19F labeling. In

addition to the overall reduced spectral quality, there are two distinct populations of several,

well-resolved, N-terminal residues (25, 28, 31, 33, 35, 37, 61, and 63) at higher enrichment

levels. An example is shown in the inset of Figure 6.2, which focuses on Gly-33. Spectra

associated with 19F labeling levels above 6% all exhibit an equilibrium between these two N-

terminal conformations, which represent a native and nonnative state as evidenced by the

chemical shifts. The free energy associated with this equilibrium at 37°C has been calculated for

the relevant labeling efficiencies and the results are collected in Table 6.2. The data indicate that

the nonnative N-terminal conformation is favored only under conditions of full (>95%) labeling,

while levels of 60% or 76% 19F labeling provide reasonably uniform samples with stabilities of

approximately 4kT.

143

Figure 6.2 1H-15N HSQC spectra of 0% 3-FPhe enriched CaM (black) overlaid with analogous

spectra of 60% 3-FPhe CaM (blue) and >95% 3-FPhe CaM (grey). Spectra were collected at

37°C.

Spin-lattice (R1) and spin-spin (R2) relaxation rates of backbone amide nitrogen atoms

were measured for the 6%, 60% and >95% 19F labeled CaM samples. The results are plotted as

the difference in a given rate from the non-fluorinated protein for each assignable residue in

Figures 6.3B and C. As expected, the general trend shows larger deviations from the non-

fluorinated protein for higher levels of 19F labeling for both R1 and R2. R1 generally increases

with greater 3-FPhe enrichment, while R2 values decrease. Larger R1 values and lower R2 values

are consistent with an increase in fast (ps – ns) amide bond reorientations, which may reflect an

overall increase in the plasticity of the protein. There is an exceptional increase in R2 for leucine

39, as well as some less pronounced increases, predominantly in the N-terminal domain residues.

144

Interestingly, arginine 37 experiences a significant decrease in R2, while nearby, leucine 39 is

increased. In general, while we observe moderately larger perturbations in the N-terminal

domain, the overall dynamics of the protein appear to be increased at higher levels of fluorine

labeling.

145

146

Figure 6.3 A) A histogram of the weighted average of chemical shift perturbations resulting

from fluorine incorporation in calmodulin as a function of residue. Shift perturbations were

calculated as a weighted average of shifts in both the 1H and 15N dimensions [Δδ =

!

( #(1H))2 + 0.2( #(15N))2 ]. Histograms comparing the difference in 15N R1 (B) and 15N R2 (C)

values for 0% 19F labeled calmodulin with 6%, 60% and >95% 19F labeled calmodulin as a

function of residue.

Estimates of rotational correlation times for the 6%, 60% and >95% proteins are shown

in Table 6.2 and were determined from R2/R1 values as described previously, with care taken to

exclude residues based on evidence of chemical exchange and low NOE values (Kay et al.

1989b). Previous studies of calmodulin have established that the two termini effectively reorient

independently in solution, with estimates of anisotropic reorientation (10-15%) (Barbato et al.

1992) being much lower than predicted from the barbell-like crystal structure. Thus, we report

correlation times for each domain individually, as even at >95% labeling, correlation times

plotted per residue suggest two independent correlation times corresponding to residues in the N

and C-termini (data not shown). Generally, the global correlation time decreases with increased

3-FPhe enrichment, with a greater difference observed for the N-terminal domain. The higher

degree of structural and dynamic perturbations observed for the N-terminal domain is consistent

with the effects of fluorine being cumulative, as the N-terminal domain has five phenylalanine

residues, while the C-terminal domain has only three.

Table 6.2 Free energy (ΔG) associated with the conformational exchange between a native and

nonnative conformation and overall correlation times (τc) for various levels of fluorine

enrichment of calmodulin.

147

A decrease in the overall correlation time resulting from >95% 3-FPhe enrichment could

result from either increased disorder throughout the protein or a decrease in the hydrodynamic

radius of the protein. To confirm that fluorination does not somehow lead to a smaller

hydrodynamic radius and thus, faster tumbling, the translational diffusion coefficients of non-

fluorinated and >95% 19F-enriched CaM were measured. The diffusion coefficients for 0% and

>95% CaM are 1.85 ± 0.03 × 10-10 m2/s and 1.83 ± 0.03 × 10-10 m2/s, and relying on the Stokes-

Einstein diffusion equation, the corresponding radii of hydration are 17.80 ± 0.08 Å and 17.95 ±

0.13 Å, respectively. These values are somewhat lower than those obtained previously ( ~ 24 Å)

(Weljie et al. 2003), which is likely due to differences in experimental conditions including pH,

ionic strength and temperature. The radii of hydration obtained for non- and >95% labeled CaM

are identical, within the experimental error, and indicate that the average size of the molecules in

solution are the same. Taken together, the 1H-15N HSQC spectra, relaxation rates and

translational diffusion data suggest that increased 3-FPhe enrichment results in greater protein

flexibility or plasticity, which is manifested as high frequency low amplitude reorientations of

the backbone, and an additional slow exchange process between a native and nonnative

conformation, involving residues in the N-terminal domain (Jarymowycz and Stone 2006). The

presence of multiple minor conformers, observed in the 19F NMR spectra above 80% 3-FPhe

enrichment, also supports the notion of slow or intermediate conformational exchange, resulting

from the fluorine labels.

148

6.4.3 Thermal stability of fully and fractionally 3-FPhe labeled CaM

In light of the significant structural and dynamic perturbations observed in CaM with

>80% fluorine incorporation, the effect on the overall thermal stability at various labeling levels

was assessed using thermal denaturation monitored by CD spectroscopy. Temperature

denaturation experiments were carried out using samples prepared in the calcium-free state due

to the very high thermal stability of calcium loaded calmodulin (Masino et al. 2000). The

melting temperature of apo-CaM samples with 0%, 60%, 76%, and >95% fluorine incorporation

were determined to be 50.8 ± 0.2°C, 50.5 ± 0.2°C, 50.0 ± 0.2°C, 50.7 ± 0.2°C respectively.

Although we observe significant changes in our NMR spectra, the overall thermal stability of the

proteins are essentially unchanged, within the experimental error.

6.5 Conclusions

We demonstrate a fractional fluorine labeling strategy using 3-FPhe CaM which

significantly improves 19F and 1H-15N HSQC spectral quality. An optimal balance between NMR

signal to noise and spectral quality was found to be at 60% - 76% enrichment of the fluorinated

species. The strategy utilizes biosynthetic protein expression protocols, where the fluorine label

is added as a mixture with its native counterpart to affect various levels of fluorine enrichment.

Spectral resolution was drastically improved in 19F NMR spectra, where all 8 peaks were

resolved in fractionally labeled samples, compared to 6 in fully labeled CaM. In addition, peak

line widths were reduced on average by half, and minor conformers previously observed in fully

labeled 3-FPhe CaM were absent. 1H-15N HSQC spectra also displayed significantly improved

line widths as well as a reduction in chemical shift perturbations. High levels of fluorine

incorporation were found to increase the fast timescale dynamics and overall protein disorder as

evidenced by 15N relaxation experiments and diffusion measurements. It is somewhat surprising

that diffusion measurements, calcium binding experiments and thermal denaturation experiments

reveal no discernible difference between the unfluorinated and fully 3-FPhe enriched versions of

CaM. Many global properties of the protein appear to remain unchanged upon labeling. In

contrast, 15N,1H and in particular, 19F NMR reveal a pronounced effect of fluorination which we

attribute to protein plasticity and increased disorder. Generally, fractional 19F-labeling retains

native protein structural and dynamic characteristics, while offering the usual advantages of 19F

probes in biomolecular NMR.

149

It is difficult to predict the generality of these results in terms of the benefits of fractional

labeling in the case of protein expression with fluoroaromatics. The prevalence of tryptophans in

most proteins is on the order of 1.1%, while that of tyrosine and phenylalanine is significantly

higher (on the order of 3.5% where variations of a factor of 2-3 are common in the protein

database). In our hands, we have never observed substantial perturbations in 1H-15N HSQCs

resulting from 5-fluorotryptophan incorporation, in several proteins, while we have observed

substantial perturbations with fluorotyrosine and fluorophenylalanine incorporation. Moreover,

since Phe is predominantly situated in the hydrophobic core of soluble proteins, it is more likely

that Phe residues may be clustered, and that cumulative effects from fluoro-Phe enrichment

would be destabilizing. Even in our study of CaM, we could not identify any 19F-19F NOEs,

suggesting that cumulative perturbations were not the result of direct Phe-Phe interactions.

Finally, in many studies where single Phe residues were replaced with Tyr (and vice versa) for

purposes of assignment, gross perturbations in the 19F NMR spectra were observed. Thus, subtle

changes commonly have a profound effect on overall protein structure and stability. Therefore,

we cannot for certain argue the generality of our observations except to say that in cases where

enrichment of a given fluoroaromatic results in poor NMR spectra or functional assays indicative

of reduced stability, fractional enrichment may dramatically improve prospects.

6.6 Supplementary Data

Figure 6.4. 19F NMR spectra of calmodulin enriched with >95% 4-fluorophenylalanine (4-

FPhe). Sample conditions and NMR acquisition parameters were identical to those used for 19F

NMR spectra of CaM enriched with 3-FPhe.

150

Figure 6.5 1H-15N HSQC spectrum of 0% 4-FPhe enriched CaM (black) overlaid with an

analogous spectrum of >95% 4-FPhe CaM (grey). All sample conditions and NMR acquisition

parameters were identical to those used to obtain spectra of 3-FPhe labeled CaM.

151

Chapter 7

Discussion and Future Directions

152

7.1 Conclusions

The application of 19F NMR to the study of protein structure and dynamics continues to

flourish, particularly as labeling strategies and NMR sensitivity improves. Of particular note, the 19F nucleus is a spin 1/2 species with a high magnetogyric ratio and high chemical shift

(Danielson and Falke 1996). These qualities translate to high sensitivity, a large chemical shift

dispersion, strong homo- and heteronuclear dipolar interactions, valuable in the examination of

through space interactions, strong chemical shift anisotropies, often useful in dynamics studies,

and simple implementation of pulse sequences common to other spin 1/2 species. Moreover, the

majority of experiments are accomplished by direct detection of the 19F NMR signal in 1D

experiments, since the chemical shift dispersion is often adequate to fully resolve resonances

from biosynthetically labeled proteins.

The chemical shift of a given fluorine probe depends on the local chemical structure,

while environmental influences arising from both the solvent and protein topology, including the

influence of electrostatic and Van der Waals interactions, may give rise to second order shifts

which may be as large as 20 ppm (Gerig 1994). The span of observed chemical shifts is often

100-fold larger than that of protons and shifts are highly receptive to changes in solvent

conditions and those within the protein itself, such as conformational fluctuations and dynamics

related to binding and catalysis (Anderluh et al. 2005; Quint et al. 2006). The fluorine nucleus

also exhibits large, long range coupling constants, allowing for the application of multi-

dimensional techniques. With regard to biomolecules, a complete lack of background signal and

the variety of fluorine probes available for use offer virtually limitless possibilities. Despite

these advantages, there are some limitations to 19F NMR in protein studies. The large chemical

shift anisotropy provides an efficient relaxation mechanism in solution, which scales with the

sqaure of the magnetic field. Consequently, T2 is relatively short which limits resolution,

magnetization transfer schemes, and possibilities for studying intermediate timescale via CPMG

or T1ρ pulse sequence schemes. Furthermore, fluorinated amino acids are generally assumed to

be isosteric relative to their native counterparts; yet, the incorporation of even mono-fluorinated

probes can cause significant structural, functional and energetic disruptions (Duewel et al. 2001;

Kitevski-Leblanc et al. 2010). Careful analysis of a labeled protein sample is thus an important

prerequisite to NMR studies. Finally, resonance assignment, commonly achieved via site-

directed mutagenesis in combination with 19F NMR, is often unattainable due to structural

153

perturbations reflected in the corresponding 19F NMR spectra (Khan et al. 2006; Li and Frieden

2005). The focus of this thesis has been to address some of these limitations, including

resonance assignment and perturbations arising from fluorine probes, and to develop more robust

methods of studying protein topology by 19F NMR.

In our first foray into 19F NMR applications we looked at topological features of a 3-

fluorotyrosine labeled G48M mutant of an SH3 domain, in both the free and peptide bound state.

Assignment of the 5 resonances arising from tyrosine residues 8, 10, 49, 54, and one in the C-

terminal tag was obtained using single phenylalanine replacement mutants. The free protein

exhibited a monomer-dimer equilibrium at concentrations above 0.4 mM, presumably due to the

fluorophobic effect mediated by the clustering of fluorine nuclei at the dimer interface. This was

confirmed by a lack of higher oligomeric states at increased protein concentrations. All tyrosine

side chains were deemed to be fully solvent exposed in the free protein via solvent isotope shifts

and paramagnetic effects from dissolved oxygen and TEMPOL. Tyr8, 49 and 54 all exhibited

marked chemical shift changes in the presence of peptide, and were used to determine an

equilibrium dissociation constant of 18 ± 4 µM. Binding was significantly reduced relative to

the corresponding unlabeled protein, suggesting that fluorine substitution disturbs the interaction,

possibly though steric effects, reduction of the hydroxyl pKa and competition with dimerization.

Finally, line broadening as a function of peptide concentration was observed for Tyr8, 10 and 54,

providing an estimate of the effective exchange rate, kex, of 5200 ± 700 Hz. This study

illustrated the potential for 19F NMR in the examination of solvent exposure as well as binding

equilibria and kinetics, all using simple 1D 19F NMR experiments.

To investigate the utility of a 13C-enriched fluorine probe, 13C,15N-3-fluorotyrosine was

synthesized by direct electrophilic fluorination in 79% isolated yield using Selectfluor™; a mild

alternative to traditional sources of electrophilic fluorine such as, F2 and XeF2 (Banks 1998).

After purification using high pressure liquid chromatography and incorporation into calmodulin

(Tyr99 and Ty138), an approach for NMR-based resonance assignment was developed. Making

use of the large 19F-13C and 13C-13C scalar couplings the 19F resonance was correlated with the

adjacent delta proton using a CT-HCCF-COSY experiment, adapted from the corresponding

proton version (Ikura 1991). Assignments were then completed using the (Hβ)Cβ(CγCδ)Hδ

(Yamazaki et al. 1993) and HNCACB experiments. Amide proton and nitrogen chemical shifts

154

were obtained from previously published assignments (Kainosho et al. 2006), but in principle

can be achieved using the same sample by employing expression techniques which permit

uniform 13C-enrichment and standard triple resonance experiments for polypeptide backbone

assignment. Furthermore, the 13C-19F spin pair can be used to increase resolution using 13C-19F

CT-HSQC via the large 270 Hz coupling in the analysis of large or complex systems.

Measurement of 19F T1, T2 and 13C-19F NOE measurements can be used to evaluate slow and fast

motions of the bond vector in a manner analogous to those employed for the 15N-1H and 13C-1H

spin pairs of the protein backbone and side chains (Kay et al. 1989a; Nicholson et al. 1992b).

This strategy provided assignments without the need to resort to mutagenesis, while the novel

probe demonstrated potential for multinuclear 19F NMR approaches.

In a thorough evaluation of the strategies employed to assess solvent exposure using 19F

NMR, we quantified the relative burial of tyrosine residues in 3-fluorotyrosine labeled

calmodulin under calcium-free and calcium-loaded conditions. The techniques considered

included inherent chemical shift properties, solvent induced isotope shifts, paramagnetic shifts

from dissolved oxygen and 1H-19F NOEs initiated by saturation of either water or aliphatic

protons. In addition, the hydrophobicity index, defined as Δδ* O2 / Δδ* D2O-H2O, was used to

evaluate relative hydrophobicity of the two probes (Tyr99 and Tyr138) in both states. A

consistent picture was obtained from all measurements in which Tyr99 is more buried than

Tyr138 for both the calcium-free and calcium-loaded states. Interestingly, a slow equilibrium

between a major and minor conformer for Tyr138 (referred to Tyr138M and Tyr138m

respectively) was observed in the calcium-free state, where the minor conformer is found to

exhibit a hydrophobicity index intermediate to that of the major calcium-free conformer,

Tyr138M, and Tyr138 in the calcium-loaded state. The solvent accessible surface area (SASA),

determined from high resolution calmodulin structures, indicate only a small change in exposure

associated with the apo to holo transition which is reflected in both the solvent isotope and

paramagnetic shifts, demonstrating their sensitivity. 1H-19F NOEs were found to be less

sensitive, due to inefficient dipolar coupling in the presence of CSA relaxation contributions and

chemical exchange. This study establishes the advantages of solvent isotope and paramagnetic

shifts in the evaluation of solvent exposure and in particular the use of the ratio of these

measurements to obtain a relative measure of hydrophobicity.

155

The importance of the ability to assign fluorine resonances, combined with the reality that

commercially available fluorine probes are not 13C or 15N enriched, lead to the development of a

NMR-based assignment strategy in the absence of isotopic enrichment. Calmodulin labeled with

3-fluorophenylalanine at eight native positions was used as a model system. Using a similar

protocol to that previously developed, we began with coupling the fluorine nucleus to the

adjacent delta proton using the prominent 19F-1H scalar coupling via a 1H-19F HSQC experiment.

Fluorine coupled delta protons were then correlated to intraresidue Hα and Hβ chemical shifts

using a 1H-19F NOESY-HSQC, and in the final step inter-residue contact was established using a 1H-15N NOESY-HSQC to the neighboring (n+1), 15N-enriched residue. The reliance of this

strategy on dipolar based magnetization transfer makes it particularly sensitive to dynamics and

relaxation mechanisms other than dipolar interactions, including CSA and chemical exchange,

which serve to reduce the efficiency of the NOE effect. The strategy enabled the assignment of

only 4 of 7 resolvable resonances, whereas complete assignment required separation of the

protein into its two domains (residues 1-73 and 84-148) by trypsin digestion, followed by

repetition of 1H-19F HSQC and 1H-19F NOESY-HSQC experiments. Interestingly, upon trypsin

fragmentation, we observed marked changes in both 19F and 1H-15N NMR spectra suggestive of

transient interdomain interactions in the full length protein. Control experiments using fractional

labeling methods revealed that this feature was exclusive to the fully labeled protein, and we

therefore proposed that fluorine may have enhanced a pre-existing equilibrium towards a more

compact structure. The above approach provides a general strategy for assignments, which

primarily utilized dipolar magnetization transfer and 15N- and 19F-editing, and as such is

expected to be transferable to other unnatural aromatic amino acids by appropriate adjustment of

experimental parameters.

In light of the significant perturbations observed in calmodulin fully labeled with 3-

fluorophenylalanine, we performed a series of experiments to thoroughly evaluate these

disruptions using 19F and 1H-15N NMR, 15N dynamics, diffusion measurements, thermal stability

and calcium binding. Full enrichment of 3-fluorophenylalanine in calmodulin was found to

show near native properties with respect to calcium binding, thermal stability and diffusion

properties, while NMR analysis revealed minor conformers, increased line widths, large

chemical shift perturbations and increased protein plasticity. We then evaluated the merit of

fractional labeling - that is, expression in the presence of a mixture of the fluorinated and native

156

amino acid, in ameliorating the observed perturbations. We found that fractional enrichments

levels of 60%-76% were deemed optimal with respect to the preservation of native protein

characteristics while maintaining a feasible level of fluorine for acquisition of 19F NMR data in

our system. Contrary to even our own expectations, fractionally labeled samples revealed

reduced 19F line widths and the disappearance of minor conformers. The use of fractional

labeling strategies can likely be extended to other unnatural amino acids and is expected to

provide the most benefit when the amino acid of interest is clustered in the hydrophobic protein

core, as is often the case with phenylalanine.

7.2 Future Directions

The field of fluorine NMR of proteins is vast and future developments can certainly take

many directions depending on the needs of the experimentalist. This section will be restricted to

the areas that have been discussed in this thesis, and how they may be improved or extended with

future experiments.

With respect to the preparation of probes capable of multinuclear approaches, the

availability of fluorinated amino acids with 13C and 15N enrichment would be advantageous. In

all cases, but specifically for the aromatics, the ability to increase chemical shift space into one,

or two additional dimensions may be the only solution to achieving spectral resolution amid

dominant CSA relaxation contributions at high magnetic fields. Once 13C-enriched versions of

fluorinated probes are available, the adaptation of existing pulse sequences from their original

form for use with fluorine nuclei, by adjustment of the appropriate experimental parameters,

should be straightforward.

The preparation of 13C,15N-3-fluorotyrosine involved the direct fluorination of the

corresponding 13C- and 15N-enriched native amino acid . In principle the same approach can be

adapted for the synthesis of 13C,15N-enriched 2-fluoro-, 3-fluoro- and 4-fluorophenylalanine as

well as multiple fluorinated variants. The position of fluorination and facile reaction of tyrosine

with Selectfluor was mediated by the electron donating properties of the hydroxyl, which

serves to both promote electrophilic substitution reactions and direct substitution adjacent to

itself via resonance stabilization of the positively charge intermediate species (McMurry 2004).

The lack of such activation and directing ability in phenylalanine has two main implications: 1)

the need to use more reactive electrophilic fluorinating reagents, requiring specialized glassware

157

and careful handling (Taylor et al. 1999) and 2) the lack of directing ability results in a mixture

of products including all mono-fluorinated species as well as multiple fluorine substitutions. The

preparation of a mixture of products also demands a robust purification protocol capable of

separating the isomeric species. Recently, the addition of trifluoromethanesulfonic acid to mild

electrophilic fluorinating agents, such as Selectfluor™, has been shown to affect mono-

fluorination of even unactivated aromatics under (relatively) mild conditions (Shamma et al.

1999).

The preparation of fluorinated tryptophan using isotopically enriched tryptophan as a

precursor is less straightforward than tyrosine and phenylalanine. Preliminary work in our lab

has shown potential for a three step protocol shown in Figure 7.1, where a fluorine is introduced

specifically at position 6 via fluorodediazoniation. O

NH3+

NHO

-

O

NH3+

NHO

-

N+

O

O-

O

NH3+

NHO

-

F

O

NH3+

NHO

-

NH2

HNO3, AcOH

H2 Pd/C 2.2% HCl/MeOH

1. NOBF4

2. Heat

Figure 7.1 Synthetic approach to the preparation of 6-fluorotryptophan from isotopically

enriched tryptophan. 6-nitrotryptophan was obtained in 45% isolated yield, and the

corresponding amine product was isolated in 87% yield. The final product was not isolated, but

observed at 8-12% yield in crude reaction mixtures.

The synthesis takes advantage of site-specific nitration of tryptophan (Moriya et al. 1975),

followed by reduction to the corresponding amine and in-situ preparation of a relatively stable

diazonium salt (-N2+), which can be replaced with a fluorine by thermal decomposition (Milner

1992). The difficulty of this approach is the final step, which suffers from low yields due to the

propensity for side reactions producing a reduced product (Canning et al. 2002) as well as

coupled aromatic products, known as “diazo-tar” (Clark 1996). The recent development of ionic

158

liquids as solvents for fluorodediazoniation chemistry offer reduced side reactions and greater

overall yields (Laali and Gettwert 2001). Finally, the preparation of isotopically enriched

versions of fluorinated aliphatic amino acids and fluorine tags would also be advantageous, as

these probes may hold the most promise for future 19F NMR studies of large complexes and

membrane proteins.

Conformational fluctuations in proteins, often related to folding, catalysis or excursions

to minor conformers, are studied using a variety of biophysical techniques including NMR,

fluorescence and H/D exchange mass spectrometry. The sensitivity of the fluorine nucleus itself,

combined with the ability to evaluate protein topology makes 19F NMR an ideal compliment to

such approaches in the examination of a variety of dynamic processes. In many cases, the

sensitivity of the fluorine nucleus permits detection of minor changes in a protein, which are

otherwise invisible using other techniques, including 1H, 13C and 15N NMR. Furthermore, the

examination of protein topology using dissolved oxygen, which is known to be sensitive to

structural fluctuations allowing intramolecular access, could prove to be quite useful. Although

not a paramagnetic effect, the solvent isotope shift is expected to provide similar insight into

transient protein fluctuations, due to its small size and mobility. Moreover, the combination of

these two effects, both as a ratio – providing a relative measure of hydrophobicity, or as a

product - where partitioning effects are factored out reflecting the collisional accessibility alone,

is an application unique to fluorine and has great potential in conformational analysis of proteins.

In recent studies of fractionally 3-fluorophenylalanine labeled calmodulin at high

temperatures, solvent isotope shifts, paramagnetic shifts and the corresponding hydrophobicity

index provided insight into conformational changes in the vicinity of the eight probes, along the

protein unfolding pathway. In particular, we observed a decrease in solvent exposure and

increased hydrophobicity at 70°C (Kitevski-LeBlanc and Prosser, unpublished results). Given

that 3-fluorophenylalanine labeled calmodulin retains secondary structure at 80°C, as evidenced

by NMR and CD, we propose the existence of a molten globule like state above 70°C, with a

significantly hydrophobic core. The ability to experimentally obtain tangible biophysical

parameters, such as local hydrophobicity and exposure, complements theoretical protein folding

approaches and will enhance the existing repertoire of experimental techniques. Further analysis

of this system, and other proteins where full unfolding can be accessed using NMR, should

reveal the utility of solvent isotope and paramagnetic shifts. The availability of fluorine probes

159

with both hydrophobic and hydrophilic character combined with site-directed labeling provides

access to any position in a protein while retaining native qualities. Moreover, cysteine-

mutagenesis techniques combined with fluorine tags make it possible to engineer probe positions

for high resolution 19F NMR. A complete picture would include analysis of the protein backbone

using well established NMR techniques, including H/D exchange (Bai et al. 1995) and relaxation

dispersion techniques (Korzhnev and Kay 2008).

Finally, the potential of fractional labeling approaches in reducing probe-induced

disruptions in other protein systems should be explored. In light of the numerous reports of

structural and functional perturbations in the literature, it is unlikely that the results we obtained

in calmodulin are unique; however, it would be interesting to pursue a similar analysis in

proteins containing predominantly β-sheet structural elements, as well as membrane proteins. It

would also be insightful to examine the effects of fractional labeling of aliphatic probes, and

even chemical tagging, possibly through controlled stoichiometry, reaction time and temperature.

160

References

Abragam A The Principles of Nuclear Magnetism; Claredon Press, Oxford, 1961.

Adams H, Blanco JLJ, Chessari G, Hunter CA, Low CMR, Sanderson JM, Vinter JG (2001) Quantitative Determination of Intermolecular Interactions with Fluorinated Aromatic Rings. Chemistry-a European Journal 7:3494-3503.

Adriaensens P, Box M, Martens H, Onkelinx E, Put J, Gelan J (1988) Investigation of Protein-Structure by Means of F-19-Nmr - a Study of Hen Egg-White Lysozyme. Eur J Biochem 177:383-394.

Akcay G, Kumar K (2009) A New Paradigm for Protein Design and Biological Self-Assembly. Journal of Fluorine Chemistry 130:1178-1182.

Al-Abdul-Wahid MS, Yu CH, Batruch I, Evanics F, Pomes R, Prosser RS (2006) A Combined Nmr and Molecular Dynamics Study of the Transmembrane Solubility and Diffusion Rate Profile of Dioxygen in Lipid Bilayers. Biochemistry 45:10719-10728.

Alexeev D, Barlow PN, Bury SM, Charrier JD, Cooper A, Hadfield D, Jamieson C, Kelly SM, Layfield R, Mayer RJ, McSparron H, Price NC, Ramage R, Sawyer L, Starkmann BA, Uhrin D, Wilken J, Young DW (2003) Synthesis, Structural and Biological Studies of Ubiquitin Mutants Containing (2s, 4s)-5-Fluoroleucine Residues Strategically Placed in the Hydrophobic Core. Chembiochem 4:894-896.

Altenbach C, Greenhalgh DA, Khorana HG, Hubbell WL (1994) A Collision Gradient-Method to Determine the Immersion Depth of Nitroxides in Lipid Bilayers - Application to Spin-Labeled Mutants of Bacteriorhodopsin. Proceedings of the National Academy of Sciences of the United States of America 91:1667-1671.

Altieri AS, Hinton DP, Byrd RA (1995) Association of Biomolecular Systems Via Pulsed-Field Gradient Nmr Self-Diffusion Measurements. Journal of the American Chemical Society 117:7566-7567.

Anderluh G, Razpotnik A, Podlesek Z, Macek P, Separovic F, Norton RS (2005) Interaction of the Eukaryotic Pore-Forming Cytolysin Equinatoxin Ii with Model Membranes: F-19 Nmr Studies. Journal of Molecular Biology 347:27-39.

Andre I, Linse S (2002) Measurement of Ca2+-Binding Constants of Proteins and Presentation of the Caligator Software. Analytical Biochemistry 305:195-205.

Azad BB, Ashique R, Chirakal R, Schrobilgen GJ (2008) Selective Fluorination of M-Tyrosine by Of2. Journal of Fluorine Chemistry 129:22-27.

161

Azim MK, Budisa N (2008) Docking of Tryptophan Analogs to Trytophanyl-Trna Synthetase: Implications for Non-Canonical Amino Acid Incorporations. Biol Chem 389:1173-1182.

Baber JL, Szabo A, Tjandra N (2001) Analysis of Slow Interdomain Motion of Macromolecules Using Nmr Relaxation Data. Journal of the American Chemical Society 123:3953-3959.

Babu YS, Bugg CE, Cook WJ (1988) Structure of Calmodulin Refined at 2.2 a Resolution. Journal of Molecular Biology 204:191-204.

Bagno A, Rastrelli F, Saielli G (2005) Nmr Techniques for the Investigation of Solvation Phenomena and Non-Covalent Interactions. Progress in Nuclear Magnetic Resonance Spectroscopy 47:41-93.

Bai YW, Sosnick TR, Mayne L, Englander SW (1995) Protein-Folding Intermediates - Native-State Hydrogen-Exchange. Science 269:192-197.

Banks RE (1998) Selectfluor (Tm) Reagent F-Teda-Bf4 in Action: Tamed Fluorine at Your Service. Journal of Fluorine Chemistry 87:1-17.

Bann JG, Pinkner J, Hultgren SJ, Frieden C (2002) Real-Time and Equilibrium F-19-Nmr Studies Reveal the Role of Domain-Domain Interactions in the Folding of the Chaperone Papd. Proceedings of the National Academy of Sciences of the United States of America 99:709-714.

Barbato G, Ikura M, Kay LE, Pastor RW, Bax A (1992) Backbone Dynamics of Calmodulin Studied by N-15 Relaxation Using Inverse Detected 2-Dimensional Nmr-Spectroscopy - the Central Helix Is Flexible. Biochemistry 31:5269-5278.

Bernini A, Venditti V, Spiga O, Niccolai N (2009) Probing Protein Surface Accessibility with Solvent and Paramagnetic Molecules. Progress in Nuclear Magnetic Resonance Spectroscopy 54:278-289.

Bertini I, Del Bianco C, Gelis I, Katsaros N, Luchinat C, Parigi G, Peana M, Provenzani A, Zoroddu MA (2004) Experimentally Exploring the Conformational Space Sampled by Domain Reorientation in Calmodulin. Proceedings of the National Academy of Sciences of the United States of America 101:6841-6846.

Bezsonova I, Forman-Kay J, Prosser RS (2008) Molecular Oxygen as a Paramagnetic Nmr Probe of Protein Solvent Exposure and Topology. Concepts in Magnetic Resonance Part A 32A:239-253.

Bezsonova I, Korzhnev DM, Prosser RS, Forman-Kay JD, Kay LE (2006) Hydration and Packing Along the Folding Pathway of Sh3 Domains by Pressure-Dependent Nmr. Biochemistry 45:4711-4719.

Bilgiçer B, Fichera A, Kumar K (2001) A Coiled Coil with a Fluorous Core. J Am Chem Soc 123:4393-4399.

162

Bouchard M, Pare C, Dutasta J, Chauvet J, Gicquaud C, Auger M (1998) Interaction between G-Actin and Various Types of Liposomes: A F-19, P-31 and H-2 Nuclear Magnetic Resonance Study. Biochemistry 37:3149-3155.

Brokx RD, Vogel HJ. In Calcium-Binding Proteins Protocols; J., V. H. Ed.; Humana Press Inc.: Totawa, NJ, 2002; pp. 183-193.

Brooks B, Benisek WF (1994) Mechanism of the Reaction Catalyzed by Delta(5)-3-Ketosteroid Isomerase of Comamonas (Pseudomonas) Testosteroni - Kinetic-Properties of a Modified Enzyme in Which Tyrosine-14 Is Replaced by 3-Fluorotyrosine. Biochemistry 33:2682-2687.

Broos J, Maddalena F, Hesp BH (2004) In Vivo Synthesized Proteins with Monoexponential Fluorescence Decay Kinetics. Journal of the American Chemical Society 126:22-23.

Campos-Olivas R, Aziz R, Helms GL, Evans JNS, Gronenborn AM (2002) Placement of F-19 into the Center of Gb1: Effects on Structure and Stability. Febs Letters 517:55-60.

Canning PSJ, Maskill H, McCrudden K, Sexton B (2002) A Product Analytical Study of the Thermal and Photolytic Decomposition of Some Arenediazonium Salts in Solution. Bulletin of the Chemical Society of Japan 75:789-800.

Cellitti SE, Jones DH, Lagpacan L, Hao X, Zhang Q, Hu H, Brittain SM, Brinker A, Caldwell J, Bursulaya B, Spraggon G, Brock A, Ryu Y, Uno T, Schultz PG, Geierstanger BH (2008) In Vivo Incorporation of Unnatural Amino Acids to Probe Structure, Dynamics, and Ligand Binding in a Large Protein by Nuclear Magnetic Resonance Spectroscopy. J Am Chem Soc 130:9268-9281.

Chambers SE, Lau EY, Gerig JT (1994) Origins of Fluorine Chemical-Shifts in Proteins. Journal of the American Chemical Society 116:3603-3604.

Charrier J, Hitchcock P, Young D (2004a) Synthesis of (2s,3r)-[3 ',3 ',3 '-H-2(3)]-Valine and (2s,3s)-4-Fluorovaline. Org Biomol Chem 2:1310-1314.

Charrier J, Hadfield D, Hitchcock P, Young D (2004b) Synthesis of (2s,3s)-3 '-Fluoroisoleucine. Org Biomol Chem 2:797-802.

Chiu H-P, Kokona B, Fairman R, Cheng RP (2009) Effect of Highly Fluorinated Amino Acids on Protein Stability at a Solvent-Exposed Position on an Internal Strand of Protein G B1 Domain. J Am Chem Soc 131:13192-13193.

Chiu H-P, Suzuki Y, Gullickson D, Ahmad R, Kokona B, Fairman R, Cheng RP (2006) Helix Propensity of Highly Fluorinated Amino Acids. J Am Chem Soc 128:15556-15557.

Chorghade MS, Mohapatra DK, Sahoo G, Gurjar MK, Mandlecha MV, Bhoite N, Moghe S, Raines RT (2008) Practical Syntheses of 4-Fluoroprolines. J Fluorine Chem 129:781-784.

Cistola DP, Hall KB (1995) Probing Internal Water-Molecules in Proteins Using 2-Dimensional F-19-H-1 Nmr. Journal of Biomolecular Nmr 5:415-419.

163

Clark JH, Wails, D., Bastock, T. W. Aromatic Fluorination; CRC Press, 1996.

Critz W, Martinezcarrion M (1977) Sulfhydryl-Group Modification of Aspartate-Aminotransferase with 3-Bromo-1,1,1-Trifluoropropanone During Catalysis. Biochemistry 16:1554-1558.

Crivici A, Ikura M (1995) Molecular and Structural Basis of Target Recognition by Calmodulin. Annual Review of Biophysics and Biomolecular Structure 24:85-116.

Dalvit C (1995) New One-Dimensional Selective Nmr Experiments in Aqueous-Solutions Recorded with Pulsed-Field Gradients. Journal of Magnetic Resonance Series A 113:120-123.

Dalvit C, Hommel U (1995a) Sensitivity-Improved Detection of Protein Hydration and Its Extension to the Assignment of Fast-Exchanging Resonances. Journal of Magnetic Resonance Series B 109:334-338.

Dalvit C, Hommel U (1995b) New Pulsed-Field Gradient Nmr Experiments for the Detection of Bound Water in Proteins. Journal of Biomolecular Nmr 5:306-310.

Dalvit C, Cottens S, Ramage P, Hommel U (1999) Half-Filter Experiments for Assignment, Structure Determination and Hydration Analysis of Unlabelled Ligands Bound to C-13/N-15 Labelled Proteins. Journal of Biomolecular Nmr 13:43-50.

Danielson MA, Falke JJ (1996) Use of F-19 Nmr to Probe Protein Structure and Conformational Changes. Annual Review of Biophysics and Biomolecular Structure 25:163-195.

Danielson MA, Biemann HP, Koshland DE, Falke JJ (1994) Attractant- and Disulfide-Induced Conformational Changes in the Ligand Binding Domain of the Chemotaxis Aspartate Receptor: A 19f Nmr Study. Biochemistry 33:6100-6109.

Daviso E, Diller A, Alia A, Matysik J, Jeschke G (2008) Photo-Cidnp Mas Nmr Beyond the T1 Limit by Fast Cycles of Polarization Extinction and Polarization Generation. J Magn Reson 190:43-51.

Dawson PE, Kent SB (2000) Synthesis of Native Proteins by Chemical Ligation. Annu Rev Biochem 69:923-960.

Day IJ, Maeda K, Paisley HJ, Mok KH, Hore PJ (2009) Refolding of Ribonuclease a Monitored by Real-Time Photo-Cidnp Nmr Spectroscopy. J Biomol Nmr 44:77-86.

Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A (1995) Nmrpipe - a Multidimensional Spectral Processing System Based on Unix Pipes. Journal of Biomolecular NMR 6:277-293.

Di Nardo AA, Korzhnev DM, Stogios PJ, Zarrine-Afsar A, Kay LE, Davidson AR (2004) Dramatic Acceleration of Protein Folding by Stabilization of a Nonnative Backbone Conformation. Proceedings of the National Academy of Sciences of the United States of America 101:7954-7959.

164

Dominguez MA, Thornton KC, Melendez MG, Dupureur CM (2001) Differential Effects of Isomeric Incorporation of Fluorophenylalanines into Pvuii Endonuclease. Proteins 45:55-61.

Drabikowski W, Kuznicki J, Grabarek Z (1977) Similarity in Ca-2+-Induced Changes between Troponin-C and Protein Activator of 3'=5'-Cyclic Nucleotide Phosphodiesterase and Their Tryptic Fragments. Biochimica et Biophysica Acta 485:124-133.

Drake SK, Bourret RB, Luck LA, Simon MI, Falke JJ (1993) Activation of the Phosphosignaling Protein Chey .1. Analysis of the Phosphorylated Conformation by F-19 Nmr and Protein Engineering. Journal of Biological Chemistry 268:13081-13088.

Duerr UHN, Grage SL, Witter R, Ulrich AS (2008) Solid State F-19 Nmr Parameters of Fluorine-Labeled Amino Acids. Part I: Aromatic Substituents. J Magn Reson 191:7-15.

Duewel H, Daub E, Robinson V, Honek J (1997) Incorporation of Trifluoromethionine into a Phage Lysozyme: Implications and a New Marker for Use in Protein F-19 Nmr. Biochemistry 36:3404-3416.

Duewel H, Daub E, Robinson V, Honek J (2001) Elucidation of Solvent Exposure, Side-Chain Reactivity, and Steric Demands of the Trifluoromethionine Residue in a Recombinant Protein. Biochemistry 40:13167-13176.

Durr UHN, Grage SL, Witter R, Ulrich AS (2008) Solid State F-19 Nmr Parameters of Fluorine-Labeled Amino Acids. Part I: Aromatic Substituents. Journal of Magnetic Resonance 191:7-15.

Eichler JF, Cramer JC, Kirk KL, Bann JG (2005a) Biosynthetic Incorporation of Fluorohistidine into Proteins in E. Coli: A New Probe of Macromolecular Structure. Chembiochem 6:2170-2173.

Eichler JF, Cramer JC, Kirk KL, Bann JG (2005b) Biosynthetic Incorporation of Fluorohistidine into Proteins in E-Coli: A New Probe of Macromolecular Structure. Chembiochem 6:2170-2173.

Evanics F, Kitevski JL, Bezsonova I, Forman-Kay J, Prosser RS (2007) F-19 Nmr Studies of Solvent Exposure and Peptide Binding to an Sh3 Domain. Biochimica Et Biophysica Acta-General Subjects 1770:221-230.

Evanics F, Bezsonova I, Marsh J, Kitevski JL, Forman-Kay JD, Prosser RS (2006) Tryptophan Solvent Exposure in Folded and Unfolded States of an Sh3 Domain by F-19 and H-1 Nmr. Biochemistry 45:14120-14128.

Evenas J, Thulin E, Malmendal A, Forsen S, Carlstrom G (1997) Nmr Studies of the E140q Mutant of the Carboxy-Terminal Domain of Calmodulin Reveal Global Conformational Exchange in the Ca2+-Saturated State. Biochemistry 36:3448-3457.

Fallon JL, Quiocho FA (2003) A Closed Compact Structure of Native Ca2+-Calmodulin. Structure 11:1303-U1307.

165

Feeney J, McCormick JE, Bauer CJ, Birdsall B, Moody CM, Starkmann BA, Young DW, Francis P, Havlin RH, Arnold WD, Oldfield E (1996) F-19 Nuclear Magnetic Resonance Chemical Shifts of Fluorine Containing Aliphatic Amino Acids in Proteins: Studies on Lactobacillus Casei Dihydrofolate Reductase Containing (2s,4s)-5-Fluoroleucine. Journal of the American Chemical Society 118:8700-8706.

Finn BE, Evenas J, Drakenberg T, Waltho JP, Thulin E, Forsen S (1995) Calcium-Induced Structural-Changes and Domain Autonomy in Calmodulin. Nature Structural Biology 2:777-783.

Frieden C (2003) The Kinetics of Side Chain Stabilization During Protein Folding. Biochemistry 42:12439-12446.

Frieden C, Hoeltzli SD, Bann JG. In Energetics of Biological Macromolecules, Pt E, 2004; pp. 400-415.

Furter R (1998) Expansion of the Genetic Code: Site-Directed P-Fluoro-Phenylalanine Incorporation in Escherichia Coli. Protein Science 7:419-426.

Gakh YG, Gakh AA, Gronenborn AM (2000) Fluorine as an Nmr Probe for Structural Studies of Chemical and Biological Systems. Magnetic Resonance in Chemistry 38:551-558.

Geddes CD Reviews in Fluorescence 2008, illustrated ed.; Springer, 2009; Vol. 5.

Gerig J, Lau E (2000) Origins of Fluorine Nmr Chemical Shifts in Fluorine-Containing Protein's. Biophys J 78:9A-9A.

Gerig JT (1994) Fluorine Nmr of Proteins. Progress in Nuclear Magnetic Resonance Spectroscopy 26:293-370.

Gerig JT (1999) Gradient-Enhanced Proton-Fluorine Noe Experiments. Magnetic Resonance in Chemistry 37:647-652.

Grage SL, Duerr UHN, Afonin S, Mikhailiuk PK, Komarov IV, Ulrich AS (2008) Solid State F-19 Nmr Parameters of Fluorine-Labeled Amino Acids. Part Ii: Aliphatic Substituents. J Magn Reson 191:16-23.

Haas DA, Mailer C, Robinson BH (1993) Using Nitroxide Spin Labels - How to Obtain T(1e) from Continuous Wave Electron-Paramagnetic Resonance-Spectra at All Rotational Rates. Biophysical Journal 64:594-604.

Hammill JT, Miyake-Stoner S, Hazen JL, Jackson JC, Mehl RA (2007) Preparation of Site-Specifically Labeled Fluorinated Proteins for 19f-Nmr Structural Characterization. Nat Protoc 2:2601-2607.

Han SY, Kim YA (2004) Recent Development of Peptide Coupling Reagents in Organic Synthesis. Tetrahedron 60:2447-2467.

166

Hansen PE (2000) Isotope Effects on Chemical Shifts of Proteins and Peptides. Magnetic Resonance in Chemistry 38:1-10.

Harris RK, Becker ED, De Menezes SMC, Goodfellow R, Granger P (2002) Nmr Nomenclature: Nuclear Spin Properties and Conventions for Chemical Shifts (Iupac Recommendations 2001) (Reprinted from Pure Appl. Chem., Vol 73, Pg 1795-1818, 2001). Concepts in Magnetic Resonance 14:326-346.

Hazard ES, Govindjee R, Ebrey TG, Crouch RK (1992) Biosynthetic Incorporation of M-Fluorotyrosine into Bacteriorhodopsin PHOTOCHEMISTRY AND PHOTOBIOLOGY 56 929-934.

Heidary DK, Jennings PA (2002) Three Topologically Equivalent Core Residues Affect the Transition State Ensemble in a Protein Folding Reaction. Journal of Molecular Biology 316:789-798.

Heintz D, Kany H, Kalbitzer H (1996) Mobility of the N-Terminal Segment of Rabbit Skeletal Muscle F-Actin Detected by H-1 and F-19 Nuclear Magnetic Resonance Spectroscopy. Biochemistry 35:12686-12693.

Hellmich UA, Pfleger N, Glaubitz C (2009) F-19-Mas Nmr on Proteorhodopsin: Enhanced Protocol for Site-Specific Labeling for General Application to Membrane Proteins. Photochem Photobiol 85:535-539.

Hoeflich KP, Ikura M (2002) Calmodulin in Action: Diversity in Target Recognition and Activation Mechanisms. Cell 108:739-742.

Holmgren SK, Taylor KM, Bretscher LE, Raines RT (1998) Code for Collagen's Stability Deciphered. Nature 392:666-667.

Holz M, Weingartner H (1991) Calibration in Accurate Spin-Echo Self-Diffusion Measurements Using H-1 and Less-Common Nuclei. Journal of Magnetic Resonance 92:115-125.

Horng J-C, Raleigh DP (2003) Phi-Values Beyond the Ribosomally Encoded Amino Acids: Kinetic and Thermodynamic Consequences of Incorporating Trifluoromethyl Amino Acids in a Globular Protein. J Am Chem Soc 125:9286-9287.

Hubbell WL, Gross A, Langen R, Lietzow MA (1998) Recent Advances in Site-Directed Spin Labeling of Proteins. Current Opinion in Structural Biology 8:649-656.

Huestis WH, Raftery MA (1978) Bromotrifluoroacetone Alkylates Hemoglobin at Cysteine Beta93. Biochem Biophys Res Commun 81:892-899.

Hull W, Sykes B (1974) Fluorotyrosine Alkaline-Phosphatase - F-19 Nuclear Magnetic-Resonance Relaxation-Times and Molecular-Motion of Individual Fluorotyrosines. Biochemistry 13:3431-3437.

167

Hull WE, Sykes BD (1975) Fluorotyrosine Alkaline-Phosphatase - Internal Mobility of Individual Tyrosines and Role of Chemical-Shift Anisotropy as a F-19 Nuclear Spin Relaxation Mechanism in Proteins. Journal of Molecular Biology 98:121-153.

Hull WE, Sykes BD (1976) Fluorine-19 Nuclear Magnetic-Resonance Study of Fluorotyrosine Alkaline-Phosphatase - Influence of Zinc on Protein Structure and a Conformational Change Induced by Phosphate Binding. Biochemistry 15:1535-1546.

Ikura DM, Kay, L., E., Bax, A. (1991) Improved Three-Dimensional 1h-13c-1h Correlation Spectroscopy of a 13c-Labeled Protein Using Constant-Time Evolution. . Journal of Biomolecular NMR 1:299-304.

Ikura M, Marion D, Kay LE, Shih H, Krinks M, Klee CB, Bax A (1990) Heteronuclear 3d Nmr and Isotopic Labeling of Calmodulin - Towards the Complete Assignment of the H-1-Nmr Spectrum. Biochemical Pharmacology 40:153-160.

Ikura T, Hayano T, Takahashi N, Kuwajima K (2000) Fast Folding of Escherichia Coli Cyclophilin A: A Hypothesis of a Unique Hydrophobic Core with a Phenylalanine Cluster. Journal of Molecular Biology 297:791-802.

Ishida H, Takahashi K, Nakashima K, Kumaki Y, Nakata M, Hikichi K, Yazawa M (2000) Solution Structures of the N-Terminal Domain of Yeast Calmodulin: Ca2+-Dependent Conformational Change and Its Functional Implication. Biochemistry 39:13660-13668.

Jackson JC, Hammill JT, Mehl RA (2007) Site-Specific Incorporation of a F-19-Amino Acid into Proteins as an Nmr Probe for Characterizing Protein Structure and Reactivity. J Am Chem Soc 129:1160-1166.

Jarymowycz VA, Stone MJ (2006) Fast Time Scale Dynamics of Protein Backbones: Nmr Relaxation Methods, Applications, and Functional Consequences. Chemical Reviews 106:1624-1671.

Jasanoff A, Weiss MA (1993) Aromatic Histidine Interactions in the Zinc Finger Motif - Structural Inequivalence of Phenylalanine and Tyrosine in the Hydrophobic Core. Biochemistry 32:1423-1432.

Johnson BA, Blevins RA (1994) Nmr View - a Computer-Program for the Visualization and Analysis of Nmr Data. Journal of Biomolecular NMR 4:603-614.

Jones DH, Cellitti SE, Hao X, Zhang Q, Jahnz M, Summerer D, Schultz PG, Uno T, Geierstanger BH (2010) Site-Specific Labeling of Proteins with Nmr-Active Unnatural Amino Acids. J Biomol Nmr 46:89-100.

Jones S, Thornton JM (1996) Principles of Protein-Protein Interactions. Proceedings of the National Academy of Sciences of the United States of America 93:13-20.

Kainosho M, Torizawa T, Iwashita Y, Terauchi T, Mei Ono A, Güntert P (2006) Optimal Isotope Labelling for Nmr Protein Structure Determinations. Nature 440:52-57.

168

Kalbitzer HR, Rohr G, Nowak E, Goody RS, Kuhn W, Zimmermann H (1992) A New High-Sensitivity F-19 Probe for Labeling Cysteine Groups of Proteins. NMR in Biomedicine 5:347-350.

Kaptein R, Dijkstra K, Nicolay K (1978) Laser Photo-Cidnp as a Surface Probe for Proteins in Solution. Nature 274:293-294.

Kay LE, Torchia DA, Bax A (1989a) Backbone Dynamics of Proteins as Studied by 15n Inverse Detected Heteronuclear Nmr Spectroscopy: Application to Staphylococcal Nuclease. Biochemistry 28:8972-8979.

Kay LE, Torchia DA, Bax A (1989b) Backbone Dynamics of Proteins as Studied by N-15 Inverse Detected Heteronuclear Nmr-Spectroscopy - Application to Staphylococcal Nuclease. Biochemistry 28:8972-8979.

Keng YF, Wu L, Zhang ZY (1999) Probing the Function of the Conserved Tryptophan in the Flexible Loop of the Yersinia Protein-Tyrosine Phosphatase. European Journal of Biochemistry 259:809-814.

Khan F, Kuprov I, Craggs TD, Hore PJ, Jackson SE (2006) F-19 Nmr Studies of the Native and Denatured States of Green Fluorescent Protein. Journal of the American Chemical Society 128:10729-10737.

Kim HJ, Howell SC, Van Horn WD, Jeon YH, Sanders CR (2009) Recent Advances in the Application of Solution Nmr Spectroscopy to Multi-Span Integral Membrane Proteins. Progress in Nuclear Magnetic Resonance Spectroscopy 55:335-360.

Kim HW, Perez JA, Ferguson SJ, Campbell ID (1990) The Specific Incorporation of Labeled Aromatic-Amino-Acids into Proteins through Growth of Bacteria in the Presence of Glyphosate - Application to Fluorotryptophan Labeling to the H+-Atpase of Escherichia-Coli and Nmr-Studies. Febs Letters 272:34-36.

Kirk KL (1980) Convenient Synthesis of 3-Fluoro-L-Tyrosine and 3,5-Difluoro-L-Tyrosine. Journal of Organic Chemistry 45:2015-2016.

Kitevski-LeBlanc JL, Evanics F, Prosser RS Approaches to the Assignment of F-19 Resonances from 3-Fluorophenylalanine Labeled Calmodulin Using Solution State Nmr. Journal of Biomolecular NMR 47:113-123.

Kitevski-LeBlanc JL, Al-Abdul-Wahid MS, Prosser RS (2009a) A Mutagenesis-Free Approach to Assignment of F-19 Nmr Resonances in Biosynthetically Labeled Proteins. Journal of the American Chemical Society 131:2054-+.

Kitevski-LeBlanc JL, Evanics F, Prosser RS (2009b) Approaches for the Measurement of Solvent Exposure in Proteins by F-19 Nmr. Journal of Biomolecular NMR 45:255-264.

Kitevski-Leblanc JL, Evanics F, Scott Prosser R (2010) Approaches to the Assignment of 19f Resonances from 3-Fluorophenylalanine Labeled Calmodulin Using Solution State Nmr. J Biomol Nmr 47:113-123.

169

Klein-Seetharaman J, Getmanova E, Loewen M, Reeves P, Khorana H (1999) Nmr Spectroscopy in Studies of Light-Induced Structural Changes in Mammalian Rhodopsin: Applicability of Solution F-19 Nmr. Proc Natl Acad Sci USA 96:13744-13749.

Koch CA, Anderson D, Moran MF, Ellis C, Pawson T (1991) Sh2 and Sh3 Domains - Elements That Control Interactions of Cytoplasmic Signaling Proteins. Science 252:668-674.

Korzhnev DM, Kay LE (2008) Probing Invisible, Low-Populated States of Protein Molecules by Relaxation Dispersion Nmr Spectroscopy: An Application to Protein Folding. Accounts of Chemical Research 41:442-451.

Korzhnev DM, Salvatella X, Vendruscolo M, Di Nardo AA, Davidson AR, Dobson CM, Kay LE (2004) Low-Populated Folding Intermediates of Fyn Sh3 Characterized by Relaxation Dispersion Nmr. Nature 430:586-590.

Kranz J, Lu J, Hall K (1996) Contribution of the Tyrosines to the Structure and Function of the Human U1a N-Terminal Rna Binding Domain. Protein Sci. 5:1567-1583.

Kubasik MA, Daly E, Blom A (2006) F-19 Nmr Chemical Shifts Induced by a Helical Peptide. Chembiochem 7:1056-1061.

Laali KK, Gettwert VJ (2001) Fluorodediazoniation in Ionic Liquid Solvents: New Life for the Balz-Schiemann Reaction. Journal of Fluorine Chemistry 107:31-34.

Larson SM, Davidson AR (2000) The Identification of Conserved Interactions within the Sh3 Domain by Alignment of Sequences and Structures. Protein Science 9:2170-2180.

Lau EY, Gerig JT (1997) Effects of Fluorine Substitution on the Structure and Dynamics of Complexes of Dihydrofolate Reductase (Escherichia Coli). Biophys J 73:1579-1592.

Laurents DV, Scholtz JM, Rico M, Pace CN, Bruix M (2005) Ribonuclease Sa Conformational Stability Studied by Nmr-Monitored Hydrogen Exchange. Biochemistry 44:7644-7655.

Lee H, Sohn J, Yeh B, Choi J, Jung S, Kim K (2000) F-19 Nmr Investigation of F-1-Atpase of Escherichia Coli Using Fluorotryptophan Labeling. J Biochem 127:1053-1056.

Lee HY, Lee KH, Al-Hashimi HM, Marsh ENG (2006) Modulating Protein Structure with Fluorous Amino Acids: Increased Stability and Native-Like Structure Conferred on a 4-Helix Bundle Protein by Hexafluoroleucine. Journal of the American Chemical Society 128:337-343.

Lee KH, Lee HY, Slutsky MM, Anderson JT, Marsh ENG (2004) Fluorous Effect in Proteins: De Novo Design and Characterization of a Four-Alpha-Helix Bundle Protein Containing Hexafluoroleucine. Biochemistry 43:16277-16284.

Lehrer SS (1971) Solute Perturbation of Protein Fluorescence - Quenching of Tryptophyl Fluorescence of Model Compounds and of Lysozyme by Iodide Ion. Biochemistry 10:3254-&.

170

Leigh JS (1971) Relaxation Times in Systems with Chemical Exchange - Some Exact Solutions. Journal of Magnetic Resonance 4:308-&.

Li C, Lutz EA, Slade KM, Ruf RAS, Wang G-F, Pielak GJ (2009a) 19f Nmr Studies of Alpha-Synuclein Conformation and Fibrillation. Biochemistry 48:8578-8584.

Li C, Wang G-F, Wang Y, Creager-Allen R, Lutz EA, Scronce H, Slade KM, Ruf RAS, Mehl RA, Pielak GJ (2010) Protein F-19 Nmr in Escherichia Coli. J Am Chem Soc 132:321-327.

Li H, Frieden C (2005) Nmr Studies of 4-F-19-Phenylalanine-Labeled Intestinal Fatty Acid Binding Protein: Evidence for Conformational Heterogeneity in the Native State. Biochemistry 44:2369-2377.

Li H, Frieden C (2006) Fluorine-19 Nmr Studies on the Acid State of the Intestinal Fatty Acid Binding Protein. Biochemistry 45:6272-6278.

Li HL, Frieden C (2007) Observation of Sequential Steps in the Folding of Intestinal Fatty Acid Binding Protein Using a Slow Folding Mutant and F-19 Nmr. Proceedings of the National Academy of Sciences of the United States of America 104:11993-11998.

Li HS, Lawrence DS (2005) Acquisition of Fyn-Selective Sh3 Domain Ligands Via a Combinatorial Library Strategy. Chemistry & Biology 12:905-912.

Li X, Rinkevicius Z, Tu Y, Tian H, Agren H (2009b) Paramagnetic Perturbation of the F-19 Nmr Chemical Shift in Fluorinated Cysteine by O-2: A Theoretical Study. J Phys Chem B 113:10916-10922.

Lian CY, Le HB, Montez B, Patterson J, Harrell S, Laws D, Matsumura I, Pearson J, Oldfield E (1994) F-19 Nuclear-Magnetic-Resonance Spectroscopic Study of Fluorophenylalanine-Labeled and Fluorotryptophan-Labeled Avian Egg-White Lysozymes. Biochemistry 33:5238-5245.

Linse S, Helmersson A, Forsen S (1991) Calcium-Binding to Calmodulin and Its Globular Domains. Journal of Biological Chemistry 266:8050-8054.

Lipari G, Szabo A (1982) Model-Free Approach to the Interpretation of Nuclear Magnetic-Resonance Relaxation in Macromolecules .1. Theory and Range of Validity. Journal of the American Chemical Society 104:4546-4559.

Lix B, Sonnichsen FD, Sykes BD (1996) The Role of Transient Changes in Sample Susceptibility in Causing Apparent Multiple-Quantum Peaks in Hoesy Spectra. Journal of Magnetic Resonance Series A 121:83-87.

Loewen M, Klein-Seetharaman J, Getmanova E, Reeves P, Schwalbe H, Khorana H (2001) Solution F-19 Nuclear Overhauser Effects in Structural Studies of the Cytoplasmic Domain of Mammalian Rhodopsin. Proc Natl Acad Sci USA 98:4888-4892.

171

Loladze VV, Ermolenko DN, Makhatadze GI (2002) Thermodynamic Consequences of Burial of Polar and Non-Polar Amino Acid Residues in the Protein Interior. Journal of Molecular Biology 320:343-357.

Luchette PA, Prosser RS, Sanders CR (2002) Oxygen as a Paramagnetic Probe of Membrane Protein Structure by Cysteine Mutagenesis and F-19 Nmr Spectroscopy. Journal of the American Chemical Society 124:1778-1781.

Luck L, Vance J, OConnell T, London R (1996) F-19 Nmr Relaxation Studies on 5-Fluorotryptophan- and Tetradeutero-5-Fluorotryptophan-Labeled E-Coli Glucose/Galactose Receptor. J Biomol Nmr 7:261-272.

Luck LA, Falke JJ (1991) F-19 Nmr-Studies of the D-Galactose Chemosensory Receptor .1. Sugar Binding Yields a Global Structural-Change. Biochemistry 30:4248-4256.

Malmendal A, Evenas J, Forsen S, Akke M (1999) Structural Dynamics in the C-Terminal Domain of Calmodulin at Low Calcium Levels. Journal of Molecular Biology 293:883-899.

Martinez D, Gerig JT (2001) Intermolecular H-1{F-19} Noes in Studies of Fluoroalcohol-Induced Conformations of Peptides and Proteins. Journal of Magnetic Resonance 152:269-275.

Masino L, Martin SR, Bayley PM (2000) Ligand Binding and Thermodynamic Stability of a Multidomain Protein, Calmodulin. Protein Science 9:1519-1529.

McMurry J Organic Chemistry; Thomson-Brooks/Cole, 2004.

Mehta V, Kulkarni P, Mason R, Constantinescu A, Antich P (1994) Fluorinated Proteins as Potential F-19 Magnetic-Resonance-Imaging and Spectroscopy Agents. Bioconjugate Chem 5:257-261.

Meyer DF, Mabuchi Y, Grabarek Z (1996) The Role of Phe-92 in the Ca2+-Induced Conformational Transition in the C-Terminal Domain of Calmodulin. Journal of Biological Chemistry 271:11284-11290.

Milner DJ (1992) Fluoroaromatics from Arylamines, a Convenient One-Pot Conversion Using Nitrosonium Tetrafluoroborate. Synthetic Communications 22:73-82.

Minks C, Huber R, Moroder L, Budisa N (1999) Atomic Mutations at the Single Tryptophan Residue of Human Recombinant Annexin V: Effects on Structure, Stability, and Activity. Biochemistry 38:10649-10659.

Miyake-Stoner SJ, Refakis CA, Hammill JT, Lusic H, Hazen JL, Deiters A, Mehl RA (2010) Generating Permissive Site-Specific Unnatural Aminoacyl-Trna Synthetases. Biochemistry 49:1667-1677.

172

Mock ML, Michon T, van Hest JCM, Tirrell DA (2006) Stereoselective Incorporation of an Unsaturated Isoleucine Analogue into a Protein Expressed in E. Coli. Chembiochem 7:83-87.

Montclare JK, Son S, Clark GA, Kumar K, Tirrell DA (2009) Biosynthesis and Stability of Coiled-Coil Peptides Containing (2s,4r)-5,5,5-Trifluoroleucine and (2s,4s)-5,5,5-Trifluoroleucine. ChemBioChem 10:84-86.

Moriya T, Hagio K, Yoneda N (1975) Facile Synthesis of 6-Chloro-D-Tryptophan. Bulletin of the Chemical Society of Japan 48:2217-2218.

Morton CJ, Pugh DJR, Brown ELJ, Kahmann JD, Renzoni DAC, Campbell ID (1996) Solution Structure and Peptide Binding of the Sh3 Domain from Human Fyn. Structure 4:705-714.

Muhandiram DR, Kay LE (1994) Gradient-Enhanced Triple-Resonance 3-Dimensional Nmr Experiments with Improved Sensitivity. Journal of Magnetic Resonance Series B 103:203-216.

Musacchio A, Wilmanns M, Saraste M (1994a) Structure and Function of the Sh3 Domain. Progress in Biophysics & Molecular Biology 61:283-297.

Musacchio A, Saraste M, Wilmanns M (1994b) High-Resolution Crystal-Structures of Tyrosine Kinase Sh3 Domains Complexed with Proline-Rich Peptides. Nature Structural Biology 1:546-551.

Musacchio A, Gibson T, Lehto VP, Saraste M (1992) Sh3 - an Abundant Protein Domain in Search of a Function. Febs Letters 307:55-61.

Naarmann N, Bilgicer B, Meng H, Kumar K, Steinem C (2006) Fluorinated Interfaces Drive Self-Association of Transmembrane Alpha Helices in Lipid Bilayers. Angewandte Chemie-International Edition 45:2588-2591.

Neuhaus D, Williamson MP The Nuclear Overhauser Effect in Structural and Conformational Analysis; Wiley-VCH Inc: United States of America, 2000.

Niccolai N, Spiga O, Bernini A, Scarselli M, Ciutti A, Fiaschi I, Chiellini S, Molinari H, Temussi PA (2003) Nmr Studies of Protein Hydration and Tempol Accessibility. Journal of Molecular Biology 332:437-447.

Niccolai N, Ciutti A, Spiga O, Scarselli M, Bernini A, Bracci L, Di Maro D, Dalvit C, Molinari H, Esposito G, Temussi PA (2001) Nmr Studies of Protein Surface Accessibility. Journal of Biological Chemistry 276:42455-42461.

Nicholson LK, Kay LE, Baldisseri DM, Arango J, Young PE, Bax A, Torchia DA (1992a) Dynamics of Methyl-Groups in Proteins as Studied by Proton-Detected C-13 Nmr-Spectroscopy - Application to the Leucine Residues of Staphylococcal Nuclease. Biochemistry 31:5253-5263.

173

Nicholson LK, Kay LE, Baldisseri DM, Arango J, Young PE, Bax A, Torchia DA (1992b) Dynamics of Methyl Groups in Proteins as Studied by Proton-Detected 13c Nmr Spectroscopy. Application to the Leucine Residues of Staphylococcal Nuclease. Biochemistry 31:5253-5263.

Okano H, Cyert MS, Ohya Y (1998) Importance of Phenylalanine Residues of Yeast Calmodulin for Target Binding and Activations. Journal of Biological Chemistry 273:26375-26382.

Olofsson A, Sauer-Eriksson AE, Ohman A (2006) The Solvent Protection of Alzheimer Amyloid-Beta-(1-42) Fibrils as Determined by Solution Nmr Spectroscopy. Journal of Biological Chemistry 281:477-483.

Oxenoid K, Sonnichsen FD, Sanders CR (2002) Topology and Secondary Structure of the N-Terminal Domain of Diacylglycerol Kinase. Biochemistry 41:12876-12882.

Palmer AG (2004) Nmr Characterization of the Dynamics of Biomacromolecules. Chemical Reviews 104:3623-3640.

Palmer AG, Kroenke CD, Loria JP (2001) Nuclear Magnetic Resonance Methods for Quantifying Microsecond-to-Millisecond Motions in Biological Macromolecules. Meth Enzymol 339:204-238.

Pearson J, Oldfield E, Lee F, Warshel A (1993) Chemical-Shifts in Proteins - a Shielding Trajectory Analysis of the Fluorine Nuclear-Magnetic-Resonance Spectrum of the Escherichia-Coli Galactose Binding-Protein Using a Multipole Shielding Polarizability Local Reaction Field Molecular-Dynamics Approach. J Am Chem Soc 115:6851-6862.

Pearson JG, Montez B, Le HB, Oldfield E, Chien EYT, Sligar SG (1997) Assignment and Analysis of Fluorine Nuclear Magnetic Resonance Spectra of 4-Fluorotryptophan Myoglobins and Hemoglobins. Biochemistry 36:3590-3599.

Peng JW (2001) Cross-Correlated (19)F Relaxation Measurements for the Study of Fluorinated Ligand-Receptor Interactions. J Magn Reson 153:32-47.

Phillips L, Separovic F, Cornell BA, Barden JA, Dosremedios CG (1991) Actin Dynamics Studied by Solid-State Nmr-Spectroscopy. European Biophysics Journal 19:147-155.

Pintacuda G, Otting G (2002) Identification of Protein Surfaces by Nmr Measurements with a Paramagnetic Gd(Iii) Chelate. Journal of the American Chemical Society 124:372-373.

Polyakov NE, Taraban MB, Leshina TV (2004) Photo-Cidnp Study of the Interaction of Tyrosine with Nifedipine. An Attempt to Model the Binding between Calcium Receptor and Calcium Antagonist Nifedipine. Photochem Photobiol 80:565-571.

Post J, Cottam P, Simplaceanu V, Ho C (1984) F-19 Nuclear Magnetic-Resonance Study of 5-Fluorotryptophan-Labeled Histidine-Binding Protein-J of Salmonella-Typhimurium. J Mol Biol 179:729-743.

174

Prosser RS, Luchette PA (2004) An Nmr Study of the Origin of Dioxygen-Induced Spin-Lattice Relaxation Enhancement and Chemical Shift Perturbation. J Magn Reson 171:225-232.

Prosser RS, Luchette PA, Westerman PW (2000) Using O-2 to Probe Membrane Immersion Depth by F-19 Nmr. Proceedings of the National Academy of Sciences of the United States of America 97:9967-9971.

Prosser RS, Evanics F, Kitevski JL, Patel S (2007) The Measurement of Immersion Depth and Topology of Membrane Proteins by Solution State Nmr. Biochim Biophys Acta 1768:3044-3051.

Prosser RS, Luchette PA, Westerman PW, Rozek A, Hancock REW (2001) Determination of Membrane Immersion Depth with O-2: A High-Pressure F-19 Nmr Study. Biophysical Journal 80:1406-1416.

Pyka J, Ilnicki J, Altenbach C, Hubbell WL, Froncisz W (2005) Accessibility and Dynamics of Nitroxide Side Chains in T4 Lysozyme Measured by Saturation Recovery Epr. Biophysical Journal 89:2059-2068.

Quint P, Ayala I, Busby SA, Chalmers MJ, Griffin PR, Rocca J, Nick HS, Silverman DN (2006) Structural Mobility in Human Manganese Superoxide Dismutase. Biochemistry 45:8209-8215.

Rajagopalan S, Chow C, Raghunathan V, Fry C, Cavagnero S (2004) Nmr Spectroscopic Filtration of Polypeptides and Proteins in Complex Mixtures. J Biomol Nmr 29:505-516.

Renner C, Alefelder S, Bae JH, Budisa N, Huber R, Moroder L (2001) Fluoroprolines as Tools for Protein Design and Engineering We Thank Mrs. E. Weyher for Skillful Technical Assistance in Spectroscopic Analyses and Mrs. W. Wenger for Her Excellent Technical Assistance in Protein Preparation. We Are Indebted to Dr. R. Golbik for Providing Us with Barstar Plasmid and Protocols for Its Isolation and Purification. Angew Chem Int Ed Engl 40:923-925.

Renzoni DA, Pugh DJR, Siligardi G, Das P, Morton CJ, Rossi C, Waterfield MD, Campbell ID, Ladbury JE (1996) Structural and Thermodynamic Characterization of the Interaction of the Sh3 Domain from Fyn with the Proline-Rich Binding Site on the P85 Subunit of Pi3-Kinase. Biochemistry 35:15646-15653.

Richards FM (1977) Areas, Volumes, Packing, and Protein-Structure. Annual Review of Biophysics and Bioengineering 6:151-176.

Ridder A, Morein S, Stam JG, Kuhn A, de Kruijff B, Killian JA (2000) Analysis of the Role of Interfacial Tryptophan Residues in Controlling the Topology of Membrane Proteins. Biochemistry 39:6521-6528.

Rinaldi PL (1983) Heteronuclear 2d-Noe Spectroscopy. Journal of the American Chemical Society 105:5167-5168.

175

Rozovsky S, Jogl G, Tong L, McDermott AE (2001) Solution-State Nmr Investigations of Triosephosphate Isomerase Active Site Loop Motion: Ligand Release in Relation to Active Site Loop Dynamics. Journal of Molecular Biology 310:271-280.

Rule GS, Pratt EA, Simplaceanu V, Ho C (1987) Nuclear-Magnetic-Resonance and Molecular Genetic-Studies of the Membrane-Bound D-Lactate Dehydrogenase of Escherichia-Coli. Biochemistry 26:549-556.

Salopek-Sondi B, Vaughan MD, Skeels MC, Honek JF, Luck LA (2003) F-19 Nmr Studies of the Leucine-Isoleucine-Valine Binding Protein: Evidence That a Closed Conformation Exists in Solution. Journal of Biomolecular Structure & Dynamics 21:235-246.

Schlörb C, Mensch S, Richter C, Schwalbe H (2006) Photo-Cidnp Reveals Differences in Compaction of Non-Native States of Lysozyme. J Am Chem Soc 128:1802-1803.

Schuler B, Kremer W, Kalbitzer HR, Jaenicke R (2002) Role of Entropy in Protein Thermostability: Folding Kinetics of a Hyperthermophilic Cold Shock Protein at High Temperatures Using F-19 Nmr. Biochemistry 41:11670-11680.

Shamma T, Buchholz H, Prakash GKS, Olah GA (1999) Electrophilic Fluorination of Aromatics with Selectfluor (Tm) and Trifluoromethanesulfonic Acid. Israel Journal of Chemistry 39:207-210.

Shriver J, Sykes B (1982) Energetics of the Equilibrium between 2 Nucleotide-Free Myosin Subfragment-1 States Using F-19 Nuclear Magnetic-Resonance. Biochemistry 21:3022-3028.

Son S, Tanrikulu IC, Tirrell DA (2006) Stabilization of Bzip Peptides through Incorporation of Fluorinated Aliphatic Residues. Chembiochem 7:1251-1257.

Sonveaux N, Vigano C, Shapiro AB, Ling V, Ruysschaert JM (1999) Ligand-Mediated Tertiary Structure Changes of Reconstituted P-Glycoprotein - a Tryptophan Fluorescence Quenching Analysis. Journal of Biological Chemistry 274:17649-17654.

Steiner T, Hess P, Bae JH, Wiltschi B, Moroder L, Budisa N (2008) Synthetic Biology of Proteins: Tuning Gfps Folding and Stability with Fluoroproline. PLoS ONE 3:e1680.

Strynadka NCJ, James MNG (1989) Crystal-Structures of the Helix-Loop-Helix Calcium-Binding Proteins. Annual Review of Biochemistry 58:951-998.

Sun Z, Pratt E, SIMPLACEANU V, Ho C (1996) A F-19-Nmr Study of the Equilibrium Unfolding of Membrane-Associated D-Lactate Dehydrogenase of Escherichia Coli. Biochemistry 35:16502-16509.

Sykes BD, Weingart.Hi, Schlesin.Mj (1974) Fluorotyrosine Alkaline-Phosphatase from Escherichia-Coli - Preparation, Properties, and Fluorine-19 Nuclear Magnetic-Resonance Spectrum. Proceedings of the National Academy of Sciences of the United States of America 71:469-473.

176

Szantay C, Demeter A (1995) The Hidden Exchange Partner in Nmr-Spectroscopy - Kinetic-Properties and Implications in Noe Analysis. Journal of Magnetic Resonance Series A 115:94-105.

Tang Y, Tirrell DA (2001) Biosynthesis of a Highly Stable Coiled-Coil Protein Containing Hexafluoroleucine in an Engineered Bacterial Host. J Am Chem Soc 123:11089-11090.

Tang Y, Ghirlanda G, Petka W, Nakajima T, DeGrado W, Tirrell D (2001a) Fluorinated Coiled-Coil Proteins Prepared in Vivo Display Enhanced Thermal and Chemical Stability. Angew Chem Int Edit 40:1494-+.

Tang Y, Ghirlanda G, Vaidehi N, Kua J, Mainz DT, Goddard WA, DeGrado WF, Tirrell DA (2001b) Stabilization of Coiled-Coil Peptide Domains by Introduction of Trifluoroleucine. Biochemistry 40:2790-2796.

Taylor HC, Richardson DC, Richardson JS, Wlodawer A, Komoriya A, Chaikes IM (1981) "Active" Conformation of an Inactive Semi-Synthetic Ribonuclease-S. J Mol Biol 149:313-317.

Taylor SD, Kotoris CC, Hum G (1999) Recent Advances in Electrophilic Fluorination. Tetrahedron 55:12431-12477.

Teng C-L, Bryant RG (2004) Mapping Oxygen Accessibility to Ribonuclease a Using High-Resolution Nmr Relaxation Spectroscopy. Biophys J 86:1713-1725.

Teng CL, Bryant RG (2000) Experimental Measurement of Nonuniform Dioxygen Accessibility to Ribonuclease a Surface and Interior. Journal of the American Chemical Society 122:2667-2668.

Teng CL, Hinderliter B, Bryant RG (2006) Oxygen Accessibility to Ribonuclease A: Quantitative Interpretation of Nuclear Spin Relaxation Induced by a Freely Diffusing Paramagnet. Journal of Physical Chemistry A 110:580-588.

Thomas MR, Boxer SG (2001) 19f Nmr of Trifluoroacetyl-Labeled Cysteine Mutants of Myoglobin: Structural Probes of Nitric Oxide Bound to the H93g Cavity Mutant. Biochemistry 40:8588-8596.

Thorson J, Shin I, Chapman E, Stenberg G, Mannervik B, Schultz P (1998) Analysis of the Role of the Active Site Tyrosine in Human Glutathione Transferase A1-1 by Unnatural Amino Acid Mutagenesis. J Am Chem Soc 120:451-452.

Thust S, Koksch B (2003) Protease-Catalyzed Peptide Synthesis for the Site-Specific Incorporation of Alpha-Fluoroalkyl Amino Acids into Peptides. J Org Chem 68:2290-2296.

Tjandra N, Kuboniwa H, Ren H, Bax A (1995) Rotational-Dynamics of Calcium-Free Calmodulin Studied by N-15-Nmr Relaxation Measurements. European Journal of Biochemistry 230:1014-1024.

177

Torizawa T, Shimizu M, Taoka M, Miyano H, Kainosho M (2004) Efficient Production of Isotopically Labeled Proteins by Cell-Free Synthesis: A Practical Protocol. Journal of Biomolecular NMR 30:311-325.

Ulrich A (2005) Solid State F-19 Nmr Methods for Studying Biomembranes. Progress in Nuclear Magnetic Resonance Spectroscopy 46:1-21.

Vaughan MD, Cleve P, Robinson V, Duewel HS, Honek JF (1999b) Difluoromethionine as a Novel F-19 Nmr Structural Probe for Internal Amino Acid Packing in Proteins. Journal of the American Chemical Society 121:8475-8478.

Visser NV, Westphal AH, Nabuurs SM, van Hoek A, van Mierlo CPM, Visser A, Broos J, van Amerongen H (2009) 5-Fluorotryptophan as Dual Probe for Ground-State Heterogeneity and Excited-State Dynamics in Apoflavodoxin. FEBS Letters 583:2785-2788.

Wacks DB, Schachman HK (1985) F-19 Nuclear Magnetic-Resonance Studies of Fluorotyrosine-Labeled Aspartate Transcarbamoylase - Properties of the Enzyme and Its Catalytic and Regulatory Subunits. Journal of Biological Chemistry 260:1651-1658.

Wagner G, Wuthrich K (1975) Proton Nmr-Studies of Aromatic Residues in Basic Pancreatic Trypsin-Inhibitor (Bpti). Journal of Magnetic Resonance 20:435-445.

Walsh M, Stevens FC, Kuznicki J, Drabikowski W (1977) Characterization of Tryptic Fragments Obtained from Bovine Brain Protein Modulator of Cyclic Nucleotide Phosphodiesterase. Journal of Biological Chemistry 252:7440-7443.

Wang P, Tang Y, Tirrell D (2003) Incorporation of Trifluoroisoleucine into Proteins in Vivo. J Am Chem Soc 125:6900-6906.

Wang P, Vaidehi N, Tirrell D, Goddard W (2002) Virtual Screening for Binding of Phenylalanine Analogues to Phenylalanyl-Trna Synthetase. J Am Chem Soc 124:14442-14449.

Wang P, Fichera A, Kumar K, Tirrell DA (2004) Alternative Translations of a Single Rna Message: An Identity Switch of (2s,3r)-4,4,4-Trifluorovaline between Valine and Isoleucine Codons. Angew Chem Int Ed Engl 43:3664-3666.

Wang X, Mercier P, Letourneau P-J, Sykes BD (2005) Effects of Phe-to-Trp Mutation and Fluorotryptophan Incorporation on the Solution Structure of Cardiac Troponin C, and Analysis of Its Suitability as a Potential Probe for in Situ Nmr Studies. Protein Sci. 14:2447-2460.

Weljie A, Yamniuk A, Yoshino H, Izumi Y, Vogel H (2003) Protein Conformational Changes Studied by Diffusion Nmr Spectroscopy: Application to Helix-Loop-Helix Calcium Binding Proteins. Protein Sci 12:228-236.

Wilkins BJ, Marionni S, Young DD, Liu J, Wang Y, Di Salvo ML, Deiters A, Cropp TA (2010) Site-Specific Incorporation of Fluorotyrosines into Proteins Inescherichia Coliby Photochemical Disguise. Biochemistry 49:1557-1559.

178

Williams SP, Haggie PM, Brindle KM (1997) 19f Nmr Measurements of the Rotational Mobility of Proteins in Vivo. Biophys J 72:490-498.

Willis MA, Bishop B, Regan L, Brunger AT (2000) Dramatic Structural and Thermodynamic Consequences of Repacking a Protein's Hydrophobic Core. Structure 8:1319-1328.

Wilton DJ, Tunnicliffe RB, Kamatari YO, Akasaka K, Williamson MP (2008) Pressure-Induced Changes in the Solution Structure of the Gb1 Domain of Protein G. Proteins-Structure Function and Bioinformatics 71:1432-1440.

Wishart DS, Sykes BD, Richards FM (1993) Improved Synthetic Methods for the Selective Deuteration of Aromatic-Amino-Acids - Applications of Selective Protonation Towards the Identification of Protein-Folding Intermediates through Nuclear-Magnetic-Resonance. Biochimica et Biophysica Acta 1164:36-46.

Woessner DE (1961) Nuclear Transfer Effects in Nuclear Magnetic Resonance Pulse Experiments. Journal of Chemical Physics 35:41-&.

Woll MG, Hadley EB, Mecozzi S, Gellman SH (2006) Stabilizing and Destabilizing Effects of Phenylalanine -> F-5-Phenylalanine Mutations on the Folding of a Small Protein. Journal of the American Chemical Society 128:15932-15933.

Woolfson DN (2001) Core-Directed Protein Design. Current Opinion in Structural Biology 11:464-471.

Xiao GY, Parsons JF, Tesh K, Armstrong RN, Gilliland GL (1998) Conformational Changes in the Crystal Structure of Rat Glutathione Transferase M1-1 with Global Substitution of 3-Fluorotyrosine for Tyrosine. Journal of Molecular Biology 281:323-339.

Xiao N, Jiang Z-X, Yu YB (2007) Enantioselective Synthesis of (2r, 3s)- and (2s, 3r)-4,4,4-Trifluoro-N-Fmoc-O-Tert-Butyl-Threonine and Their Racemization-Free Incorporation into Oligopeptides Via Solid-Phase Synthesis. Biopolymers 88:781-796.

Yamazaki T, Formankay JD, Kay LE (1993) 2-Dimensional Nmr Experiments for Correlating C-13-Beta and H-1-Delta/Epsilon Chemical-Shifts of Aromatic Residues in C-13-Labeled Proteins Via Scalar Couplings. Journal of the American Chemical Society 115:11054-11055.

Yu LP, Hajduk PJ, Mack J, Olejniczak ET (2006) Structural Studies of Bcl-Xl/Ligand Complexes Using F-19 Nmr. Journal of Biomolecular Nmr 34:221-227.

Zagrovic B, Snow CD, Shirts MR, Pande VS (2002) Simulation of Folding of a Small Alpha-Helical Protein in Atomistic Detail Using Worldwide-Distributed Computing. Journal of Molecular Biology 323:927-937.

Zhang J, Kimura T, Siraganian RP (1998) Mutations in the Activation Loop Tyrosines of Protein Tyrosine Kinase Syk Abrogate Intracellular Signaling but Not Kinase Activity. Journal of Immunology 161:4366-4374.

179

Zhang M, Tanaka T, Ikura M (1995) Calcium-Induced Conformational Transition Revealed by the Solution Structure of Apo Calmodulin. Nature Structural Biology 2:758-767.

Zhao X, Devries JS, McDonald R, Sykes BD (2007) Determination of the 19f Nmr Chemical Shielding Tensor and Crystal Structure of 5-Fluoro-Dl-Tryptophan. J Magn Reson 187:88-96.

Zhou HP, Yan FX, Yamamoto S, Tai HH (1999) Phenylalanine 138 in the Second Intracellular Loop of Human Thromboxane Receptor Is Critical for Receptor-G-Protein Coupling. Biochemical and Biophysical Research Communications 264:171-175.

Documents

Development and Application of 19F NMR of Proteins...1.3 NMR properties of the 19F nucleus and utility of 19F NMR in protein studies 3 1.3.1 Probes available for fluorine NMR of proteins