Transition Metal and Lanthanide Complexes for Catalysis and Protein Structure Determination

Home , Transition metal

Thesis submitted in partial fulfillment of the requirements of:

Doctor of Philosophy

Bradley Yat Wah Man

The University of New South Wales

School of Chemistry

Supervisors:

Professor Barbara Messerle

Professor Gottfried Otting

June 5, 2010

Dedicated to the memory of my grandmother Man Wen Fang, whose hopes has inspired me through the tough times.

Preface

This thesis is a report of original research undertaken by the author and is submitted for admission to the degree of Doctor of Philosophy at the University of New South Wales. This work was completed in the School of Chemistry at the University of New South Wales during the period of March 2006 to March 2010. The works and results presented in this thesis are those of the author, unless otherwise acknowledged.

Sections of this work that have been published:

3-Mercapto-2,6-pyridinedicarboxylic acid, A Small Lanthanide Binding Tag for Protein Studies by NMR Spectroscopy

Bradley Man, Xun-Cheng Su, Haobo Liang, Shane Simonsen, Thomas Huber, Barbara A. Messerle and Gottfried Otting

Chemistry: A European Journal, 2010, 16(12), 3827 – 3832

A Dipicolinic Acid Tag for Rigid Lanthanide Tagging of Proteins and Paramagnetic NMR Spectroscopy

Xun-Cheng Su, Bradley Man, Sophie Bereen, Haobo Liang, Shane Simonsen, Christophe Schmitz, Thomas Huber, Barbara A. Messerle and Gottfried Otting

Journal of the American Chemical Society, 2008, 130(2), 10486 – 10487

Sections of this work have been presented at scientific conferences:

Functionalized Rhodium(I) complexes: structure and catalysis

Bradley Man, Barbara Messerle.

Australian Organometallic Meeting 5, Sydney, New South Wales 2010. Oral Presentation.

A New Thiol Modified Dipicolinic Acid Tag for Protein Studies with Pseudocontact Shifts

Bradley Man, Barbara Messerle, Gottfried Otting, Shane Simonsen, Xun-Cheng Su

7th Biannual Meeting of the Australian and New Zealand Nuclear Magnetic Resonance Society (ANZMAG), Couran Cove, Queensland 2008. Poster presentation.

Small Molecular Tags for Protein Structure Determination

Bradley Man, Barbara Messerle, Gottfried Otting, Xun-Cheng Su

23rd International Conference on Organometallic Chemistry, Rennes, France 2008. Poster Presentation.

Paramagnetic Complexes for Protein Structure Determination

Bradley Man, Barbara Messerle, Gottfried Otting, Shane Simonsen, Xun-Cheng Su

21st International Congress for Heterocyclic Chemistry, Sydney, New South Wales 2007. Poster Presentation.

Paramagnetic Complexes for Protein Structure Determination

Bradley Man, Barbara Messerle, Gottfried Otting, Shane Simonsen, Xun-Cheng Su

Conference of the Inorganic Chemistry Division ,Hobart, Tasmania 2007. Poster Presentation.

Abstract

“A short saying often contains much wisdom”

Sophocles

The aims of this thesis are to explore the application of transition metal and lanthanide complexes in the areas of structural biology and catalysis. This PhD thesis is therefore divided into two sections; the emphasis of the first section is on the synthesis and development of small organic molecules as paramagnetic probes with applications in protein structure refinement. The second section is on the synthesis and characterization of a series of rhodium(I) complexes bearing either functionalized tridentate pyrazolyl or tridentate imidazolyl donor ligands. The reactivity of this series of rhodium(I) complexes as catalysts for the intramolecular cyclization of alkynoic acids was investigated.

The first section of the thesis involved the synthesis of two thiol modified dipicolinic acid based tags, 4-mercaptomethyl-2,6-pyridinedicarboxylic acid (4MMDPA, 5) and 3-mercapto-2,6-pyridinedicarboxylic acid (3MDPA, 9) were described. The ligands 4MMDPA (5) and 3MDPA (9) were attached to the N-terminal domain of the arginine repressor from E. coli. (ArgN). Lanthanide complexes of the ArgN-4MMDPA adduct were synthesized in situ and the efficiency of the ligand (5) as a paramagnetic probe was assessed using 15N-HSQC spectra. The HSQC spectra showed that the 4MMDPA- Ln3+ complex was able to rigidly bind to the protein, the first ligand to be reported with rigid binding with only a single attachment point. Similar studies on the lanthanide and transition metal complexes of the ArgN-3MDPA adduct showed that the 3MDPA (9) tag was able to bind to both lanthanide and transition metal ions. Due to the reduced magnitude of the tensors observed for the 3MDPA tag (9), the rigidity of the binding could not be adequately assessed. The incorporation of unnatural amino acids in proteins via a modified protein expression system was proposed as an alternative method to introduce paramagnetic labels without the need of post translational

iii

modifications. Synthetic routes to a series of unnatural amino acids 11, 15 and 19 based on previously described lanthanide and transition metal binding motifs were developed.

The second half of the project describes the synthesis and the characterization of a series of rhodium(I) complexes bearing the tris(pyrazolyl)toluidine ligand (p-tpt, 24a; o-tpt, 24b) and the tris(N-methylimidazolyl)methanol ligand (tim, 23). The structures of the rhodium(I) complexes containing these ligands (27, 28 and 29b) in the solid state showed that the ligands were all bound to the rhodium centre in a bidentate (2) binding mode centre via two of the three available nitrogen donors. In solution, fluxional behavior was observed in the 1H NMR spectrum for the rhodium(I) complexes bearing the tris(pyrazolyl)toluidine ligands (27 – 29). Detailed analysis of the rhodium(I) complexes bearing the o-tris(pyrazolyl)toluidine ligand (27b, 28b and 29b) in the solution state at low temperatures showed that there is restricted rotation of the aryl ring about the C-C bond between the aryl ring and the bridging carbon, leading to observation of two species in the 1H NMR spectrum.

The application of rhodium(I) complexes bearing pyrazolyl donors (29a and 29b) and imidazolyl donors (37 and 40) as catalysts for the intramolecular cyclization of aliphatic and aromatic alkynoic acids was investigated. It was found that rhodium(I) complexes bearing imidazolyl donors 37 and 40 were highly active catalysts for the intramolecular cyclization of aliphatic alkynoic acids. Despite the structural similarities between the rhodium(I) imidazolyl complexes containing the bidentate and tridentate ligands, differences in catalytic activity were observed, attributed to the interference of the unbound imidazolyl donor with the substrate during the catalytic cycle. The rhodium(I) complex bearing the pyrazolyl donors (29a) was found to be a highly active catalyst for the intramolecular cyclization of terminal aromatic alkynoic acids, while rhodium(I) complexes bearing bidentate imidazolyl ligands were highly active catalysts for the intramolecular cyclization of non-terminal aromatic alkynoic acids. All rhodium(I) complexes (29a, 37 and 40) investigated as catalysts for the intramolecular cyclization of both non-terminal and terminal aromatic alkynoic acids displayed high level of regioselectivity in the formation of exo-dig vs. endo-dig product. A study of the dependence of the regioselectivity and the efficiency of cyclization on the electronic

nature of the substrates indicated that electron withdrawing groups at the terminal alkyne substituent generally increased the rates of cyclization. The regioselectivity of the cyclization was also influenced by the relative difference in the electron density between the alkyne carbons.

Acknowledgements

“Dreams take long term commitment. From the first to the last we need focus, discipline, persistence, and the ability to keep sight of the vision of what we are slowly creating.”

Dennis Wholey

Each chapter starts off with a quote, and each describing the theme of the chapter. The quote for the acknowledgement more or less summarizes this entire volume of work. Along the way, this thesis felt like folly and many times along the way, the urge to throw it all in was high. Through the support of family, friends and a bit of positive thinking, it all came through in the end.

First I would like to thank my supervisors Prof. Barbara Messerle and Prof. Gottfried Otting for giving me such an opportunity to work on such a wonderful project. Without their tireless guidance and support this project could not have been possible. I would acknowledge the contributions the BAM and GO research group has offered during the PhD. In particular, I would like to thank Dr Jason Harper, Dr Jim Hook, Dr Donald Thomas, Dr Richard Hodgson, Dr Rebecca Wilson, Dr Shane Simonsen and Dr Laurent Poorters. These wonderful people always found the time to teach me everything they knew and without them, I could not have become the chemist I am today. They have also on numerous occasions provided many words of encouragement and wisdom, without which I could not have completed this PhD.

I would like to thank all of my friends, especially Debbie Mackay, Chris Haines, and Richard Parry, Murray Adams, Eric Majka, Matthew Stone, Alison Manion and Kate Odenthal. I would like to thank Richard for showing me the strength to get through some of the toughest times time I had in the last two years. I would like to thank Chris, for showing me that life should be viewed for all its potentials, not for what went

wrong. Deb, Murray and Alison were great friends during my PhD, always offering advice, always happy to take care of a friend going through tough times, and most importantly, always reminding me what a bad car the McLaren was. I would like to especially thank Kate, Alison, Mark, Sandra, Jim and Deb who spent many tireless hours deciphering my ramblings, also known as proof reading.

No PhD could ever be completed without the support of the School of Chemistry, and its vast network of both technical and administrative staff. I would like to thank the wonderful help that the UNSW NMR Facility provided. I would like to thank Dr Ruhu Qi and the Mass Spectroscopy Facility at the Research School of Chemistry, ANU for their support and help.

Last but definitely not least, I would like to thank the efforts of my parents and family who without their wonderful support this work could not have been possible. On that note, I would like to finish this acknowledgement with….

“Its Murad’s fault…”

Bradley Man

Saturday, June 05, 2010

vii

Abbreviations

“Why is abbreviation such a long word?”

Unknown

3MDPA – 3-mercapto-2,6-pyridinedicarboxylic acid

4MMDPA – 4-mercaptomethyl-2,6-pyridinedicarboxylic acid

ArgN – N-terminal domain of the arginine repressor from E. Coli.

BArF – tetrakis(3,5-bis(trifluoromethylphenyl))borate bim – bis(N-methylimidazolyl)methane bpm – bis(pyrazolyl)methane

COD – 1,5 cyclooctadiene

COSY – correlation spectroscopy

DCM – dichloromethane

DPA – 2,6-pyridinedicarboxylic acid

DTNB – 5,5’-dithio-bis(2-nitrobenzoic acid)

DTPA – diethylenetriaminopentaacetic acid

FPLC – fast protein liquid chromatography

HMBC – heteronuclear multiple bond coherence spectroscopy

HSQC – heteronuclear single quantum coherence spectroscopy mRNA – messenger ribonucleic acid

NBD – 2,5 norbornadiene

NMR – nuclear magnetic resonance

viii

nOe – nuclear Overhauser effect

NOESY – nuclear Overhauser effect spectroscopy

PCS – pseudocontact shift

PRE – paramagnetic relaxation enhancement

RACS – residual anisotropic chemical shift

RDC – residual dipolar coupling

RDS – rate determining step tdmpm – tris(3,5-dimethypyrazolyl)methane

THF – tetrahydrofuran tim – tris(N-methylimidazolyl)methanol

TNB- - 5-thio-2-nitrobenzoate tpm – tris(pyrazolyl)methane tpt – tris(pyrazolyl)toluidine tRNA – transfer ribonucleic acid

UTR – universal tensor representation.

Table of Contents

Preface ...... i

Abstract ...... iii

Acknowledgements ...... vi

Abbreviations ...... viii

Table of Contents ...... x

List of Figures ...... xviii

List of Tables ...... xxvi

List of Schemes ...... xxviii

Chapter 1: Introduction ...... 1

1.1 Introduction ...... 2

1.2 Functionalized Metal Complexes for Protein Structure Determination ...... 3

1.2.1 NMR spectroscopy, X-Ray Crystallography and Structural Biology ...... 4

1.2.2 Paramagnetic NMR Spectroscopy ...... 6

1.2.3 Pseudocontact Shifts and Protein Structure Refinement ...... 7

1.3 Functionalized Metal Complexes for Catalysis ...... 10

1.3.1 Catalytic Cycle ...... 10

1.3.2 Homogeneous vs. Heterogeneous Catalyst ...... 12

1.4 Project Aims ...... 13

1.5 General structure of the thesis ...... 13

1.6 References ...... 15

Chapter 2: Dipicolinic Acid As Paramagnetic Probe – 4MMDPA...... 18

2.1 Introduction ...... 19

2.2 Paramagnetic Probes ...... 19

2.2.1 Metal Substitution in Metalloproteins ...... 20

2.2.2 Lanthanide Binding Peptides ...... 21

2.2.3 Lanthanide Binding Ligands ...... 22

2.3 Dipicolinic Acid Based Paramagnetic Probes ...... 23

2.3.1 Retrosynthesis ...... 24

2.3.2 Attachment of 4MMDPA onto N-terminal domain of the arginine repressor ...... 26

2.3.3 NMR spectroscopy of ArgN-4MMDPA in the presence of Diamagnetic & Paramagnetic Lanthanide Ions...... 29

2.3.4 Solving the Magnetic Susceptibility Tensor and Metal Ion Locations ...... 33

2.3.5 Alignment Tensor ...... 36

2.4 Conclusion ...... 42

2.5 Reference ...... 44

Chapter 3: Dipicolinic Acid as Paramagnetic Probe – 3MDPA ...... 46

3.1 Introduction ...... 47

3.2 Synthesis ...... 47

3.3 Transition metals as paramagnetic shift reagents ...... 48

3.3.1 Metal Ion Location and Magnetic Susceptibility Tensor ...... 51

3.3.2 Alignment Tensor ...... 53

3.4 Paramagnetic lanthanide metals as paramagnetic shift reagents ...... 56

3.4.1 NMR Spectroscopy of ArgN-3HDPA ...... 57

3.4.2 NMR Spectroscopy of T4-lysozyme ...... 62

3.4.3 Comparison of ArgN-3MDPA and T4-lysozyme-3MDPA ...... 64

3.4.4 Residual Dipolar Coupling ...... 64

3.5 Conclusion ...... 65

3.6 References ...... 67

Chapter 4: Unnatural Amino Acids as Paramagnetic Probes ...... 68

4.1 Introduction ...... 69

4.2 Incorporation of unnatural amino acids ...... 69

4.2.1 Residual specific labeling using unnatural amino acids ...... 70

4.2.2 Site specific labeling using unnatural amino acids ...... 71

4.3 Unnatural amino acids as paramagnetic probes ...... 73

4.3.1 Generalized synthesis of unnatural amino acids ...... 74

4.3.2 Synthesis of the dipicolinic acid modified amino acid, DPA-AA (11) ...... 75

4.3.3 Synthesis of iminodiacetic acid modified amino acid, IDA-AA (15) ...... 77

4.3.4 Synthesis of 8-hydroxyquinoline modified amino acid, HQ-AA (19) ...... 80

4.3.5 Attempted Incorporation of DPA-AA, IDA-AA and HQ-AA ...... 82

4.4 Conclusion ...... 82

4.5 References ...... 84

Chapter 5: Tridentate Rhodium(I) Complexes: Synthesis and Structure ...... 86

5.1 Introduction ...... 87

5.1.1 Synthesis of substituted tris(pyrazolyl)alkane ligands ...... 88

5.1.2 Coordination chemistry of tridentate N,N donor ligands ...... 89

5.2 Rhodium(I) complexes bearing p-tris(pyrazolyl)toludine ligand ...... 91

5.2.1 Synthesis of the tris(pyrazoyl)toluidine rhodium(I) complexes ...... 91

5.2.2 Solid state structure of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a) ...... 92

xii

5.2.3 Solution state structure of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(o-tpt)][BArF] (28a) ...... 96

5.2.4 Solution state structure of the complex [Rh(CO)2(p-tpt)][BArF] (29a) ...... 98

5.2.5 Synthesis and structure of the complex [Rh(COD)(p-tpt)][BPh4] (32) ...... 100

5.3 Rhodium(I) complexes bearing o-tris(pyrazoyl)toludine ligand ...... 102

5.3.1 Solid state studies of [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o-tpt)][BArF] (28b) ...... 103

5.3.2 NMR studies of [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o-tpt)][BArF] (28b) ...... 107

5.3.3 NMR and solid state structure of [Rh(CO)2(o-tpt)][BArF] (29b) ...... 111

5.3.4 Understanding the structural exchange processes in rhodium(I) complexes bearing o-32 Exchange vs. Atropisomerism ...... 115

5.4 Rhodium(I) complexes bearing the tris(N-methylimidazolyl)methanol ligand ...... 117

5.4.1 Synthesis of [Rh(COD)(tim)][BArF] (36) and

[Rh(CO)2(tim)][BArF] (37) ...... 117

5.4.2 Solid state structures of [Rh(COD)(tim)][BArF] (36) and

[Rh(CO)2(tim)][BArF] (37) ...... 118

5.5 Conclusion ...... 121

5.6 References ...... 123

Chapter 6: Tridentate Rhodium(I) Complexes: Catalytic Activity ...... 126

6.1 Introduction ...... 127

6.1.1 Alkylidene lactones ...... 127

6.1.2 Metal catalyzed intramolecular hydroalkoxylation ...... 130

6.1.3 Gold / silver catalyzed hydroalkoxylation of alkynoic acids...... 132

xiii

6.1.4 Rhodium / Iridium catalyzed hydroalkoxylation of alkynoic acids...... 133

6.2 Catalyzed intramolecular hydroalkoxylation of aliphatic alkynoic acid ...... 135

6.2.1 Procedure ...... 135

6.2.2 Preliminary catalyst and solvent screening ...... 135

6.2.3 Intramolecular cyclization of 4-pentynoic acid using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors...... 137

6.2.4 Intramolecular cyclization of 5-hexynoic acid using Rhodium(I) complexes bearing imidazolyl donors...... 141

6.3 Intramolecular hydroalkoxylation of aromatic alkynoic acid ...... 143

6.3.1 Synthesis of aromatic alkynoic acid substrates ...... 144

6.3.2 Intramolecular cyclization of 2-ethynylbenzoic acid using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors...... 144

6.3.3 Intramolecular cyclization of non-terminal aromatic alkynoic acids using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors...... 149

6.3.4 Exo-dig : Endo-dig selectivity: Aliphatic vs. aromatic terminal alkyne substituents ...... 151

6.4 Exo-dig : Endo-dig selectivity: Electron donating vs. electron withdrawing terminal alkyne substituents ...... 153

6.4.1 Synthesis of non-terminal aromatic alkynoic acids with electron rich or electron deficient terminal substituents ...... 153

6.4.2 Electronic properties of electron donating and withdrawing substrates...... 154

6.5 Conclusion ...... 157

6.6 References ...... 159

Chapter 7: Summary, conclusions and future work ...... 161

7.1 Summary and conclusions ...... 162

7.2 Future work ...... 164

xiv

Chapter 8: Experimental ...... 168

8.1 General Experimental ...... 169

8.2 Synthesis of thiol modified lanthanide binding ligands and modification of proteins...... 171

8.2.1 Synthesis of 2,6-Dimethoxycarbonyl-4-bromomethylpyridine (4)...... 171

8.2.2 Synthesis of 4-Mercaptomethyl-2,6-pyridinedicarboxylic acid (5)...... 172

8.2.3 Synthesis of 3-mercapto-2,6-dimethoxycarbonylpyridine (9)...... 172

8.2.4 Ligation of thiol modified tags to N-terminal domain of the arginine repressor (ArgN)...... 173

8.3 Synthesis of unnatural amino acids bearing lanthanide and transition metal ion binding motifs...... 174

8.3.1 Synthesis of Dimethyl Acetamido(2,6- dimethoxycarbonylpyridyl)malonate (10)...... 174

8.3.2 Synthesis of 2-amino-3-(2,6-pyridinedicarbonyl)propanoic acid (11)...... 174

8.3.3 Synthesis of Diethyl Acetamido(N,N-bis(ethoxycarbonylmethyl)- aminobenzyl)malonate (14)...... 175

8.3.4 Synthesis of 2-amino-3-(N,N-bis(methylcarboxylate) aminobenzyl)propanoic acid (15)...... 176

8.3.5 Synthesis of Diethyl Acetamido-6-(9-hydroxylquinolyl)malonate diethyl ester (18)...... 176

8.3.6 Synthesis of 2-amino-6-(9-hydroxylquinoyl)propanoic acid dihydrochloride (19)...... 177

8.4 Synthesis of rhodium(I) complexes bearing pyrazolyl donor ligands...... 178

8.4.1 General synthesis of rhodium(I) olefin tris(pyrazolyl)toluidine complexes bearing the BArF counterion 27a-b and 28a-b...... 178

8.4.2 General synthesis of rhodium carbonyl tris(pyrazolyl)toluidine complexes 29 by the displacement of the COD co-ligand...... 184

8.5 Synthesis of rhodium(I) complexes bearing imidazolyl donor ligands...... 186

8.5.1 Synthesis of [Rh(COD)(tim)][BArF] (36)...... 186

8.5.2 Synthesis of [Rh(CO)2(tim)][BArF] (37)...... 187

8.6 Catalyzed intramolecular cyclization of alkynoic acids – Synthesis of substrates and catalysis...... 187

8.6.1 Synthesis of 2-heptylmethylbenzoate (43c)...... 187

8.6.2 Generalized procedure for the synthesis of terminally substituted alkynoic esters 49a–d...... 188

8.6.3 General procedure for the hydrolysis of aromatic alkynoic acids 44 and 50...... 191

8.6.4 Generalised procedure for metal catalysed intramolecular cyclization of alkynoic acids...... 192

8.6.5 Cyclization of 4-pentynoic acid (38) to 5-methylenedihydrofuranone (39)...... 193

8.6.6 Cyclization of 1-hexynoic acid (41) to 5-methylenedihydropyranone (42)...... 194

8.6.7 Cyclization of 2-ethynylbenzoic acid (44a) to isochromenone (45a) and 3-methyleneisobenzofuranone (46a)...... 195

8.6.8 Cyclization of 2-(phenylethynyl)benzoic acid (44b) to 3- phenylisochromenone (45b) and (Z)-3-benylideneiso-benzofuranone (46b)...... 196

8.6.9 Intramolecular cyclization of 2-(heptynyl)benzoic acid (44c) to 3- pentyl-1H-isochromen-1-one (45c) and (Z)-3-hexylideneiso-benzofuranone (46c)...... 197

8.6.10 Cyclization of 2-(4-acetylphenylethynyl)benzoic acid (50a) to 3-(4- acetylphenyl)-1H-isochromen-1-one (51a) and (Z)-3-(4-acetylphenyl)isobenzofuran(3H)-one (52a) ...... 198

8.6.11 Cyclization of 2-(4-chlorophenylethynyl)benzoic acid (50b) to 3- (4-chlorophenyl)-1H-isochromen-1-one (51b) and (Z)-3-(4-chlorophenyl)isobenzo-furan(3H)-one (52b) ...... 200

xvi

8.6.12 Cyclization of 2-(4-methylphenylethynyl)benzoic acid (50c) to 3- (4-methylphenyl)-1H-isochromen-1-one (51c) and (Z)-3-(4-methylphenyl)-isobenzofuran(3H)-one (52c) ...... 201

8.6.13 Cyclization of 2-(4-methoxyphenylethynyl)benzoic acid (50d) to 3- (4-methoxyphenyl)-1H-isochromen-1-one (51d) and (Z)-3-(4-methoxyphenyl) isobenzofuran(3H)-one (52d) ...... 202

8.7 X-Ray Crystallographic data ...... 203

8.7.1 [Rh(COD)(p-tpt)][BArF] (27a) ...... 203

8.7.2 [Rh(NBD)(p-tpt)][BArF] (28a) ...... 204

8.7.3 [Rh(COD)(p-tpt)][BPh4] (32) ...... 205

8.7.4 [Rh(COD)(o-tpt)][BArF] (27b) ...... 206

8.7.5 [Rh(NBD)(o-tpt)][BArF] (28b) ...... 207

8.7.6 [Rh(CO)2(o-tpt)][BArF] (29b) ...... 208

8.7.7 [Rh(COD)(tim)][BArF] (36) ...... 209

8.7.8 [Rh (CO)2(tim)][BArF] (37) ...... 210

8.8 References ...... 211

xvii

List of Figures

Figure 1.1: The application of functionalized metal complexes: (a) as a homogeneous catalyst using the functional group to induce additional selectivities; (b) attached to a protein as a biological NMR probe or heterogeneous catalyst in aqueous conditions; and (c) attached to a macroscopic scaffold as a heterogeneous catalyst...... 2

Figure 1.2: The four levels of the protein structure (a) primary structure; (b) secondary structure; (c) tertiary structure; and (d) quaternary structure...... 4

Figure 1.3: (a) Selected resonances in a 1H-15N HSQC spectra recorded of the - ! ! ! " # connecting cross peaks of backbone amide recorded without a lanthanide ion (black), with Dy3+ (red) and Er3+ (magenta). (b) Isosurfaces showing the pseudocontact shifts corresponding to ±3, ±1.5 and ±0.5 ppm where positive values are denoted in blue and negative values are denoted in red. The three $!! ribbon representation.15 ...... 9

Figure 1.4: General catalytic cycle involving the complex (MLn) and the reactants (A and B)...... 11

Figure 2.1: Structures of the CLaNP11, CLaNP-312 and CLaNP-5.213 lanthanide binding ligands attached to proteins...... 22

Figure 2.2: (a) The 1H-15N HSQC spectrum of ArgN protein (red) overlaid with the spectrum of ArgN-4MMDPA adduct (green) showing the change in the spectrum after derivatization with the 4MMPDA ligand. (b) A plot of the changes in 15N chemical shift for the amino acid residues in the protein...... 28

Figure 2.3: A three dimensional ribbon representation of the ArgN protein with the residues Glu20 – Ser25 and Gly54 – Ala 71 highlighted in red. The

xviii

three dimensional structure of the 4MMDPA probe (5) attached to the Cys68 is superimposed for visual guidance and does not represent the actual orientation of the ligand in the ArgN-4MMDPA adduct...... 29

Figure 2.4: Overlay of the 1H-15N HSQC spectra of the unpurified ArgN- 4MMDPA adduct in the presence of a diamagnetic lanthanide ion (red) and in the presence of a paramagnetic lanthanide ion (green)...... 30

Figure 2.5: Overlay of the 1H-15N HSQC spectra of the purified ArgN- 4MMDPA adduct in the presence of a diamagnetic lanthanide ion (cyan) and in the presence of a paramagnetic lanthanide ion (purple)...... 30

Figure 2.6: Overlay of the 1H-15N HSQC spectra of G27 and G54 residue in the ArgN-4MMDPA adduct in the presence of a diamagnetic lanthanide ion and in the presence of a paramagnetic lanthanide ion Tm3+ (green), Yb3+ (purple) and Tb3+ (orange). Paramagnetic peaks belonging to the same amino acid residue are connected using a dashed blue line...... 32

Figure 2.7: Plot of the PCS of the amide protons in the 1H dimension vs. the amino acid residue number for Tb3+ (J), Tm3+ (L) and Yb3+ ([)...... 32

Figure 2.8: The plot of the back calculated PCS vs. the observed PCS for ArgN-4MMDPA: (a) Tb3+; (b) Tm3+; and (c) Yb3+. The predicted PCS was calculated by solving all three sets of experimental data using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.19 ...... 34

Figure 2.9: PCS Isosurface for ±1 and ±0.2 ppm (blue = positive PCS, red = negative PCS) superimposed on the 3 dimensions structure of the ArgN for: (a) Tb3+; (b) Tm3+; and (c) Yb3+ visualized using PyMol. The PCS isosurfaces were calculated using all three sets of experimental data using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.19 ...... 35

xix

only including the RDCs from only the rigid element of the tertiary structure (i.e. residues 10 – 19, 26 – 36 and 43- 53)...... 38

Figure 2.11: The Sanson-Flamsteed equivalence plot for the magnetic susceptibility tensor (left) and the alignment tensors (right) for: (a) Tb3+; (b) Tm3+; and (c) Yb3+ determined independently for each of the 23 conformers in the NMR structure of the ArgN protein. The orientation of the alignment tensor was calculated by using the RDCs from only the rigid element of the tertiary structure (i.e. residues 10 – 19, 26 – 36 and 43- 53)...... 39

Figure 2.12: A ribbon representation of the three dimensional structure of the ArgN protein with the location of the metal ion (cyan sphere) and the G21 residues superimposed. The three dimensional structure of the 4MMDPA tag attached to the C68 residues is superimposed for visual guidance and does not represent the actual orientation of the ligand in the ArgN-4MMDPA complex in solution...... 41

Figure 3.1: (a) The 1H-15N HSQC spectrum of ArgN protein (red) overlaid with the spectrum of ArgN-3MDPA adduct (black) showing the change in the spectrum after derivatization with the 3MDPA ligand. (b) A plot of the changes in 15N chemical shift for the amino acid residues in the protein...... 49

Figure 3.2: Overlay of the 1H-15N HSQC spectra of the purified ArgN-3MDPA adduct in the presence of a diamagnetic lanthanide ion (black) and in the presence of a paramagnetic transition metal ion (red)...... 50

Figure 3.3: Plot of the PCS in the 1H dimension vs. the amino acid residue number for the ArgN-3MDPA in the presence of zinc and cobalt ion...... 51

Figure 3.4: The plot of the back calculated PCS vs. the experimentally observed PCS for ArgN-3MDPA in the presence of zinc and cobalt ions. The predicted PCS was calculated by solving all three sets of experimental data simultaneously using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.2 ...... 52

Figure 3.5: PCS Isosurface for ±2 and ±0.2 ppm (blue = positive PCS, red = negative PCS) superimposed on the 3 dimensions structure of the ArgN for Co2+ visualized using PyMol. The PCS isosurface were calculated using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.2 ...... 52

Figure 3.6: The plot of the back calculated RDC vs. the experimentally observed RDC for ArgN-3MDPA-Co2+ using only the first model of the ArgN protein structure. The predicted RDC was calculated using PALES3 only including the RDCs from only the rigid element of the tertiary structure (i.e. residues 10 – 19, 26 – 36 and 43- 53)...... 53

Figure 3.7: The Sanson-Flamsteed equivalence plot for the magnetic susceptibility tensor: (a) calculated using NUMBAT2 and; (b) the corrected alignment tensors determined independently for each of the 23 conformers for the NMR structure of the ArgN protein. The orientation of the alignment tensor was calculated by using the RDCs from only the rigid element of the tertiary structure (i.e. residues 14 – 16, 28 – 34 and 45- 51) using PALES3...... 55

Figure 3.8: A ribbon representation of the three dimensional structure of the ArgN protein with the location of the metal ion (purple sphere) and the Cys68, Arg59 and Lys62 residues superimposed...... 56

Figure 3.9: 1H-15N-HSQC spectra of ArgN-3MDPA in complex with different metal ions. The spectra were recorded of 0.15 mM solutions of ArgN-3MDPA in a buffer of 20 mM MES, pH 6.5, using a Bruker 800 MHz NMR spectrometer at 25oC. Selected pairs of cross-peaks from the diamagnetic and paramagnetic molecules are connected by lines and labelled with their assignment. (a) Superimposition of 15N-HSQC spectra recorded with Y3+ (black) and a mixture of Tb3+/Y3+ (grey). (b) Same as (a), except that the grey spectrum was recorded with a mixture of Tm3+/Y3+...... 58

Figure 3.10: tensors of eight different metal ions bound to the ArgN-3DPA adduct. The x-, y- and z-axes of the tensors are distinguished by light grey, dark grey and black lines, respectively. The axes definitions follow the unique tensor representation (UTR) convention, where the x- and z-axes are the

xxi

shortest and longest axes of the tensor.2 The side chains of Glu12 and Cys68 and the N- and C-termini of the protein are labelled. The figure was generated using Molmol.8 ...... 60

Figure 3.11: 1H-15N-HSQC spectra of the C54T/C97A/Q69C triple mutant of T4 lysozyme derivatized with 3MDPA (9) and in the presence of with Tm3+, Lu3+ and Co2+. Selected pairs of cross-peaks from the diamagnetic and paramagnetic molecules are connected by lines and labelled with their assignment. (a) Superimposition of the spectrum with Co2+ (red) onto the spectrum recorded with a 1:0.8 mixture of Tm3+ and Lu3+ (black). (b) Superimposition of spectra recorded with 1:0.8 mixtures of Tm3+ : Lu3+ of the construct with 3MDPA (black) and the construct with 4MMDPA (red). The spectra were recorded at 25oC in a buffer of 20 mM MES (pH 6.5) on a Bruker 800 MHz NMR spectrometer, using a protein concentration of about 0.15 mM...... 63

Figure 4.1: Phenylalanine analogues incorporated using residue specific method...... 70

Figure 4.2: 1H NMR spectrum of 2-amino-3-(2,6- pyridinedicarbonyl)propanoic acid (DPA-AA, 11) in D2O...... 77

Figure 4.3: 1H NMR spectrum of Diethyl Acetamido(N,N- bis(ethoxycarbonylmethyl)- aminobenzyl)-aminomalonate diethyl ester (14) in

CDCl3...... 79

Figure 4.4: 1H NMR spectrum of 2-amino-3-

(N,N(methylcarboxylate)aminobenzyl)-propanoic acid (15) in D2O / NaOD...... 80

Figure 4.5: 1H NMR spectrum of 2-amino-3-(8-hydroxyquinoly)propanoic acid (HQ-AA, 19) in D2O...... 82

Figure 5.1: The two different binding modes for tris(pyrazolyl)alkane ligands: (a) the tridentate binding mode and 3; (b) the bidentate binding mode 2...... 89

Figure 5.2: The molecular structure of: (a) [Rh(COD)(tdmpm)]+; and (b) [Rh(NBD)(tdmpm)]+ synthesized by Hallett et al.32 ...... 91

xxii

Figure 5.3: ORTEP depictions of (a) [Rh(COD)(p-tpt)][BArF] (27a) and (b) [Rh(NBD)(p-tpt)][BArF] (28a) at 50% thermal ellipsoids for non-hydrogen atoms. Nitrogen, rhodium and carbon atoms are colored in blue, green and black respectively...... 94

Figure 5.4: 1H NMR spectrum of [Rh(COD)(p-tpt)][BArF] (27a) at: (a) 25oC; o and (b) -55 C in CD2Cl2...... 97

Figure 5.5: 1H NMR spectrum of [Rh(NBD)(p-tpt)][BArF] (28a) showing o only the aromatic region at -55 C in CDCl3...... 98

1 Figure 5.6: H NMR spectrum of [Rh(CO)2(p-tpt)][BArF] (29a) showing only o the aromatic region at -55 C in CD2Cl2 ...... 99

Figure 5.7: ORTEP depictions of [Rh(COD)(p-tpt)][BPh4] (32) at 50% thermal ellipsoids for non-hydrogen atoms and selected bond lengths and angles. Nitrogen, rhodium and carbon atoms are colored in blue, green and black respectively...... 102

Figure 5.8: ORTEP depiction of [Rh(COD)(o-tpt)][BArF] (27b) showing the two co-crystallized structures: (a) Structure A; and (b) Structure B at 50% thermal ellipsoids for non-hydrogen atoms. Nitrogen, Rhodium and carbon atoms are colored in blue, green and black respectively ...... 105

Figure 5.9: ORTEP depiction of [Rh(NBD)(o-tpt)][BArF] (28b) at 50% thermal ellipsoids for non-hydrogen atoms. Nitrogen, Rhodium and carbon atoms are colored in blue, green and black respectively ...... 106

Figure 5.10: 1H NMR spectrum of [Rh(COD)(o-tpt)][BArF] (27b) showing o o only the aromatic region at: (a) -55 C; and (b) 25 C in CDCl3 measured on Bruker Avance 600 MHz spectrometer...... 108

Figure 5.11: 1H NMR spectrum of [Rh(NBD)(o-tpt)][BArF] (28b) showing o only the aromatic region at -55 C in CD2Cl2 measured on Bruker Avance 600 MHz spectrometer...... 109

xxiii

Figure 5.12: 1H NOESY spectrum of [Rh(COD)(o-tpt)][BArF] (27b) measured on Bruker Avance 600 MHz spectrometer showing only the aromatic o region at -50 C in CDCl3. Highlighted are the exchange cross peaks between the resonances due to the pyrazole H3/5, the H2’ and H3’ of the aniline ring...... 110

Figure 5.13: ORTEP depiction of the cationic fragment of the complex

[Rh(CO)2(o-tpt)][BArF] (29b) at 50% thermal ellipsoid for non hydrogen atoms: (a) viewed along the Rh-C1 axis and (b) viewed from the top. Nitrogen atoms are colored in blue, rhodium atoms are colored in green and carbon atoms colored in black...... 112

1 Figure 5.14: H NMR spectrum of [Rh(CO)2(o-tpt)][BArF] (29b) showing o only the aromatic region at -55 C in CD2Cl2 measured on Bruker Avance 400 MHz spectrometer...... 114

Figure 5.15: ORTEP representation of the cationic fragment of [Rh(COD)(tim)][BArF] (36) at 50% thermal ellipsoid for non hydrogen atoms. Nitrogen atoms are colored in blue, rhodium atoms are colored in green and carbon atoms colored in black...... 118

Figure 5.16: ORTEP representation of the cationic fragment of

[Rh(CO)2(tim)][BArF] (37) at 50% ellipsoid for non hydrogen atoms viewed from (a) the top; (b) top. Nitrogen atoms are colored in blue, rhodium atoms are colored in brown and carbon atoms colored in black. The 1H NMR spectrum of the complex at 25°C of the complex (37) in d4-methanol measured on a Bruker Avance 300 MHz Spectrometer...... 119

Figure 6.1: Structures of biologically active or alkylidene lactones isolated from natural products.5 ...... 128

Figure 6.2: Plot of the percentage conversion vs. time for the intramolecular o cyclization of 4-pentynoic acid to the lactone 39 using in d6-benzene at 80 C using 2 mol% of [Rh(CO)2(p-tpt)][BArF] (29a) (L), [Rh(CO)2(bim)][BArF]

(40) ([) and [Rh(CO)2(tim)][BArF] (37) (J)...... 138

xxiv

Figure 6.3: 1H NMR spectrum for the intramolecular cyclization of 4- pentynoic acid (38) catalyzed by 2 mol% [Rh(CO)2(tim)][BArF] (37) in C6D6 at 80oC at: (a) t = 1 min; (b) t = 8 mins; (c) t = 21 mins; and (d) t = 34 mins...... 139

Figure 6.4: 1H NMR spectrum for the intramolecular cyclization of 4-hexynoic o acid (41) catalyzed by 1 mol% [Rh(bim)(CO)2][BArF] (40) in C6D6 at 70 C at: (a) t = 0.2 hrs; (b) t = 0.8 hrs; (c) t = 2.2 hrs; and (d) t = 14.4 hrs ...... 142

Figure 6.5: 1H NMR spectrum for the intramolecular cyclization of 2- ethynylbenzoic acid (44a) catalyzed by 1 mol% [Rh(CO)2(bim)][BArF] (40) in o C6D6 at 60 C at: (a) t = 3 mins; (b) t = 9 mins; (c) t = 20 mins; and (d) t = 25 mins...... 145

Figure 6.6: Plot of the percentage conversion vs. time for the intramolecular cyclization of 2-ethynylbenzoic acid (44a) to the lactones 45a and 46a using in o o d6-benzene using 1 mol% of [Rh(CO)2(p-tpt)][BArF] at 60 C and 25 C (L o o and ), [Rh(CO)2(bim)][BArF] at 60 C ([) and [Rh(CO)2(tim)][BArF] at 60 C (J)...... 146

Figure 6.7: Proposed mechanism for the intramolecular cyclization of aromatic alkynoic acid catalyzed by a rhodium(I) dicarbonyl complexes, showing the two alternative pathways leading to the formation of the exo-dig product (red cycle) and the endo-dig product (purple cycle)...... 151

Figure 7.1: Proposed methods for the immobilization of the complex

[Rh(CO)2(p-tpt)][BArF] (29a) onto: (a) a carbon surface using the electrochemical reduction of a aryl diazonium salt generated in situ; (b) a carboxylic acid modified carbon surface using EDC / NHS peptide coupling between the amine and carboxylic acid group; and (c) onto a protein though the conversion of the amine to a thiol, followed by DTNB assisted disulfide bond formation...... 167

xxv

List of Tables

Table 2.1: The axial and rhombic components of the alignment tensor and magnetic susceptibility tensor for the ArgN-4MMDPA adduct in the presence of paramagnetic lanthanide ion.d ...... 40

Table 3.1: The axial and rhombic components of the alignment tensor and the magnetic susceptibility tensors for the ArgN-3MDPA adduct...... 54

Table 3.20$$$ex with ArgN-3MDPAa ...... 61

Table 5.1: Selected bond distances for the complexes [Rh(COD)(p-tpt)][BArF] (27a), [Rh(NBD)(p-tpt)][BArF] (28a) and [Rh(COD)(bpm)][BArF] (31).34 ...... 95

Table 5.2: Selected bond distances for the complex [Rh(COD)(o-tpt)][BArF] (27b), [Rh(NBD)(o-tpt)][BArF] (28b) and [Rh(COD)(bpm)][BArF] (31).34 ...... 104

Table 5.3: Selected bond distances for the complex [Rh(CO)2(o-tpt)][BArF] 34 (29b) and the analogous [Rh(CO)2(bpm)][BArF] (35)...... 113

Table 5.4: Selected bond distances and angles for the complexes

[Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37)...... 120

Table 6.1: Catalytic data for the intramolecular cyclization of 4-pentynoic acid (38) acid using Rhodium(I) and Iridium(I) based complexes...... 134

Table 6.2: Catalytic data for the intramolecular cyclization of 4-pentynoic acid

(38) in the presence of 2 mol% [Rh(CO)2(p-tpt)][BArF] (29a) in a range of solvents and temperature...... 137

Table 6.3: Comparison of catalytic activity of rhodium complexes and literature complexes for the intramolecular cyclization of 4-pentynoic acid (38) to the lactone 39...... 140

xxvi

Table 6.4: Comparison of literature catalytic activity of rhodium complexes for the intramolecular cyclization of 5-hexynoic acid (41) to the lactones 42...... 143

Table 6.5: Comparison of literature catalytic activity of rhodium(I) and gold (I) catalysts for the intramolecular cyclization of 2-ethynylbeznoic acid (44a) to the cyclic lactones 45a and 46a.a ...... 148

Table 6.6: Comparison of the catalytic activity for the intramolecular cyclization of non-terminal alkynoic acids (44b-c) to the cyclic lactones 45b-c and 46b-c using rhodium(I) imidazolyl and pyrazolyl complexes...... 150

Table 6.7: Summary of the exo : endo selectivities for the intramolecular cyclization of terminal and non-terminal alkynoic acids using the complex 40 as catalyst...... 152

Table 6.8: The corresponding Hammett constant, carbonyl stretching frequency and the selected 13C chemical shifts of the substituted aromatic alkynoic acids...... 155

Table 6.9: Intramolecular cyclization of electron rich and electron deficient non-terminal aromatic alkynoic acids (50a – d) using the complex o [Rh(CO)2(bim)][BArF] (40) in C6D6 at 60 C ...... 157

xxvii

List of Schemes

Scheme 2.1: The four possible disconnections in the retrosynthesis of 4- mercaptomethyl-2,6-pyridinedicarboxylic acid (5) ...... 24

Scheme 2.2: Synthesis of 4-mercaptomethyl-2,6-pyridinedicarboxylic acid (5)...... 25

Scheme 2.3: Functionalization of the ArgN using the 4MMDPA (5) lanthanide binding tag...... 27

Scheme 3.1: Synthesis of 3-mercapto-2,6-pyridinedicarboxylic acid (9)...... 47

Scheme 4.1: Schematic representation showing the propagation of a peptide chain through the interaction of the mRNA and the aminoacyl tRNA molecules: (a) initiation of protein synthesis signaled by the AUG codon, followed by the selection of valine amino acid using the GUU codon; (b) formation of the peptide bonding linking the two adjacent amino acids; (c) movement of the ribosome relative to the mRNA sequence to select and incorporate the next amino acid as determined by the codons; and (d) the termination and release of the peptide chain as signaled by the UGA codon.15 ...... 72

Scheme 4.2: Generalized method of the synthesis of unnatural amino acids: (a) nucleophilic attack of -halocarboxylic acids; (b) the Strecker synthesis; (c) alkylation of an acetamidomalonate ester; and (d) enzyme catalyzed ring opening of substituted hydantoin based compounds...... 75

Scheme 4.3: The synthesis of the unnatural amino acid bearing the dipicolinic acid motif (DPA-AA, 11) ...... 76

Scheme 4.4: The synthesis of the unnatural amino acid bearing the iminodiacetic acid motif (IDA-AA, 15) ...... 78

Scheme 4.5: The synthesis of the unnatural amino acid bearing the hydroxyquinoline motif (HQ-AA, 19) ...... 81

xxviii

Scheme 5.1: The generalized synthesis of the diolefin complexes 27 and 28 and the conversion to the corresponding dicarbonyl complexes 29...... 92

Scheme 5.2: Synthesis of the rhodium(I) olefin complexes bearing the tetraphenylborate counterion...... 101

Scheme 5.3: The possible exchange processes occurring in solution: (a) the V\!2 3 binding mode; and (b) atropisomerism...... 116

Scheme 6.1: ^$$"{!\ of enzymes (Enz-Nu) by binding irreversibly.5 ...... 128

Scheme 6.2: ^| ! ! $ ! $ $ " alkylidene lactones using (a) coupling reactions involving preformed alkylidene lactones; (b) condensation reactions between functionalized furans and aldehydes; and (c) catalyzed intramolecular cyclization of alkenoic or alkynoic acids.5 ...... 129

Scheme 6.3: The two possible mechanisms for the intramolecular cyclization of an alkynoic acid for: (a) substituted alkyne; and (b) unsubstituted...... 131

Scheme 6.4: The synthetic scheme for the synthesis of aromatic alkynoic acids bearing unsubstituted (44a) and substituted (44b-c) terminal alkynes...... 144

Scheme 6.5: The synthetic scheme for the synthesis of aromatic alkynoic acids with electron withdrawing and electron donating substituents at the terminal alkyne position...... 154

xxix

Chapter 1: Introduction

“Shallow men believe in luck. Strong men believe in cause and effect”

Ralph Waldo Emerson (1803 – 1882)

1.1 Introduction

Transition metal and lanthanide complexes have been traditionally used as homogeneous catalysts in the synthesis of large complex organic molecules from readily available synthons such as alkynols1-5 and aminoalkynes.6-8 Metal complexes also have a number of applications beyond catalysis; these include applications such as luminescence labeling in biology9-12 and medicinal applications such as anti-cancer activity.13-14 The aims of this project are to explore the application of transition and lanthanide metal complexes in the area of structural biology (Figure 1.1b) and the immobilization of homogeneous catalysts to combine the benefits of homogeneous and heterogeneous catalysis (Figure 1.1c).

Figure 1.1: The application of functionalized metal complexes: (a) as a homogeneous catalyst using the functional group to induce additional selectivities; (b) N-N donor ligand attached to a protein as a biological NMR probe or heterogeneous catalyst in aqueous conditions; and (c) attached to a macroscopic scaffold as a heterogeneous catalyst.

Paramagnetic lanthanide complexes have been successfully used in a number of biomedical applications, including their use as contrast agents in magnetic resonance

imaging.9-10 Recently, paramagnetic lanthanide ions have been used to determine the binding mode of protein-protein complexes in solution.15-17 One of the aims of this project is the development of small functionalized metal complexes that can be site specifically attached onto a protein and induce paramagnetic shifts in the NMR spectrum of the protein. These paramagnetic shifts can be used to determine the relative orientation and binding mode of protein-protein complexes.

Many metal complexes are highly active as homogeneous catalysts with tunable reactivity and selectivity. Despite the advantages of metal complexes as homogeneous catalysts, they are not as widely implemented as heterogeneous catalysts in part due to difficulties with separation of the catalysts from the product. With the presence of reactive functional groups on metal complexes, the complexes can be covalently attached to macroscopic scaffolds such as proteins,18-19 silicon20-21 and polymers.22-23 This makes it possible to combine the advantages of both homogeneous (i.e. tunable selectivities and activities) and heterogeneous catalyst (i.e. ease of product separation from the catalyst) into one catalytic system. Before the surfaces can be modified with the functionalized catalyst, the reactivity and the structure of the functionalized metal complexes needs to be established, which will be the major focus of this body of work.

1.2 Functionalized Metal Complexes for Protein Structure Determination

Protein structure is divided into three levels (Figure 1.2) which include: (1) primary structure; (2) secondary structure; and (3) tertiary and quaternary structure. Typically the primary structure (Figure 1.2a) refers to the sequence of amino acids linked by the amide bonds. Secondary structure (Figure 1.2b) refers to the organization of the amino acid sequence via hydrogen bonding to form discrete structural units such as the alpha helices or beta sheets. The tertiary structure refers to the arrangement of discrete secondary structural units to form the three dimensional structure of the protein. For larger proteins or protein-protein complexes, the quaternary structure refers to the full

assembly of multiple protein sub units. The three-dimensional structures of proteins and protein-protein complexes have been traditionally solved using single crystal x-ray crystallography. Recently, advances in both the hardware and software of NMR spectrometers have allowed high resolution NMR spectroscopy to solve the three dimensional structures of proteins in solution.

(b) Secondary Amino acids Protein Structure

(a) Primary Protein Structure Alpha Beta helices sheets

(d) Quaternary Protein (c) Tertiary Structure Protein Structure

Figure 1.2: The four levels of the protein structure (a) primary structure; (b) secondary structure; (c) tertiary structure; and (d) quaternary structure.

1.2.1 NMR spectroscopy, X-Ray Crystallography and Structural Biology

Structural biology is concerned with three specific objectives: (1) the identification of the various proteins which assemble to form a particular cell, tissue or organism; (2) understanding the interaction between the various proteins which makes up these cells, tissues and organisms; and (3) the determination of the precise three dimensional

structures of the protein-protein complexes which make up these cells, tissues and organisms.25

The ability to determine the three dimensional structure of protein-protein complexes has been of great interest in the area of drug discovery. Structural biology has been able to provide a means to select, refine and optimize small organic molecules to enhance their activity and selectivity. The use of detailed protein structures in the development of effective drugs will reduce not only the development time but also the costs.26 There are currently two main methods by which protein structures are solved; (1) X-Ray crystallography; and (2) Nuclear Magnetic Resonance (NMR) spectroscopy.

X-ray crystallography is considered to be the main method for solving protein structures due to the accurate coordinates that can be obtained for the entire protein assembly. With the recent advancement of techniques such as automated crystallization, data collection and processing, large structures can be refined rapidly, allowing x-ray crystallography to be used as a high throughput technique.27-29 Despite the advantages that x-ray crystallography provides, it is limited by the need to produce high quality single crystals of the protein, which is not always possible for proteins, or in particular protein-protein complexes.

NMR spectroscopy has been used to complement x-ray crystallography in protein structure determination, in particular for proteins which cannot be easily crystallized.27 Despite the benefits offered by NMR spectroscopy a major limitation is the extensive sample preparation required to determine the three dimensional structure of a large protein. Due to the complex nature of a typical NMR spectrum, isotopic labelling (i.e. 13C, 2H and 15N) is required to reduce the complexity of the NMR spectra. Due to the inherent insensitivity of NMR spectroscopy and low concentrations of a typical sample, long data collection times are required to measure an acceptable spectrum. To make NMR spectroscopy a more viable technique, there has been a drive to reduce the acquisition time. This has been achieved through several advances such as reduced dimensionality experiments,30 development of cryogenic probes31 and various automated assignment programs.32-34 Another approach to increase the efficiency of

NMR spectroscopy in the refinement of protein structures is by increasing the volume of information from a given experiment. One such example is through the use of paramagnetic NMR spectroscopy.

1.2.2 Paramagnetic NMR Spectroscopy

Paramagnetism can be introduced into an NMR sample by the addition of a species with an unpaired electron. When paramagnetism is introduced there are two effects: (1) pseudocontact shifts (PCS), which leads to changes in the chemical shift of NMR active nuclei; and (2) paramagnetic relaxation enhancement (PRE), which leads to an increase in the rate of relaxation (T1 and T2) of the resonances in the NMR spectra.

Both PRE and PCS can offer long range structural information that can be used to refine the three dimensional structures of proteins by determining the relative orientation and binding mode of protein-protein complexes. In the case of PRE, information can be observed from as far as 20 – 25 Å away from the paramagnetic label, complementing traditional nOes which are restricted to much shorter internuclear distances.35 In the case of the PCS, changes in the chemical shift of NMR active nuclei can be observed at distances as far as 40 Å from the paramagnetic centre.36 While both the PCS and PRE can offer invaluable information, PCS has been used more extensively in solving the structures of paramagnetic metalloproteins37-39 due to the fact that it is mathematically well defined and easily measured with a high degree of accuracy, and as a consequence leads to precise structural information.

There are many different sources of paramagnetism, they include transition metal ions,40 long living radicals,4 oxygen42 and lanthanide ions.43-44 The lanthanide ions and their metal complexes are often the preferred source of paramagnetism used in the refinement of protein structures due to the fact that they form stable polyvalent ions which do not interfere with the structure of the protein. Lanthanide ions can be used to replace calcium ions, providing a convenient means to introduce a paramagnetic center. The

advantage of using lanthanide ions is that while the different lanthanide ions are chemically similar, they all possess different paramagnetic properties, as these differences can be used to obtain different structural information. The different lanthanide ions have different PCS shift distributions, but at the same time they have different relaxation properties.16 In addition, the unpaired electrons of the lanthanide ions exist in the inner orbitals, which reduces the complexity of analysis of the pseudocontact shifts.45 This simplification allows the PCS to be measured by taking the difference between the chemical shift of the nuclei in the presence of a diamagnetic reference and in the presence of the desired paramagnetic lanthanide ion.

1.2.3 Pseudocontact Shifts and Protein Structure Refinement

As described in the previous section the PCS can be defined as the difference of the chemical shift observed with and without a paramagnetic centre (eqn. 1.1).

= (1.1)

While eqn. (1.1) shows the most basic definition of the PCS, an alternative definition is given in eqn. (1.2).38

= 1 (3 cos2 1) + 3 2 cos 2 123 2 sin (1.2)

Where: ax = the axial component of the magnetic susceptibility tensor

rh = the rhombic component of the magnetic susceptibility tensor r = distance of the nuclear spin to the paramagnetic label ~$!V!\"

Eqn. (1.2) shows that the value of the PCS is dependent on a number of variables including the orientation of the nuclear spin relative to the paramagnetic centre (i.e. an " ! ! anisotropic components of the magnetic susceptibility tensor (i.e. ax and rh). The implication of the eqn. (1.2) is that the size

and direction of the PCS is directly related to the position of the nuclear spin relative to the paramagnetic centre. It is this spatial dependence that allows long range structural information to be determined. Based on the experimentally observed PCS, knowledge of the magnetic susceptibility tensor and the location of the paramagnetic centre, it is possible to accurately determine the orientation and binding modes of protein-protein complexes. The magnetic susceptibility tensor and location of the paramagnetic centre are established using the PCS.

The values of the key PCS for nuclei in a given protein are typically measured using 1H-15N heteronuclear single quantum coherence (HSQC) experiments, where each cross peak represents one of the amino acid residues present in the protein. A HSQC spectrum $!15N labeled phenylalanine is shown in Figure 1.3 where each of the labeled phenylalanine is a black cross peak in the HSQC spectrum.15

Once a paramagnetic lanthanide ion is added to the sample, movement of the cross peaks in the HSQC spectrum was observed (Figure 1.3a). According to eqn. (1.1), the PCS is simply defined as the difference in chemical shift from the paramagnetic cross peak (e.g. in Figure 1.3a Dy3+ [red] or Er3+ [magenta]) subtracted away from the diamagnetic reference (black).15 It is important to note that not all of the PCSs move in the same direction. For example, in the case of F72, a positive PCS is observed in the presence of dysprosium ion while for F124 a negative PCS is observed. This is visually represented in Figure 1.3b where some areas of the protein will undergo a negative shift while others will experience a positive shift. It is this spatial dependence of the PCS that allows structural information on protein-protein complexes to be determined from paramagnetic NMR spectroscopy. This is only possible if the magnetic susceptibility tensor and the location of the metal ion are well defined.

(a) (b)

Figure 1.3: (a) Selected resonances in a 1H-15N HSQC spectra recorded for !- complex with and without paramagnetic lanthanide ion. Lines connecting cross peaks of backbone amide recorded without a lanthanide ion (black), with Dy3+ (red) and Er3+ (magenta). (b) Isosurfaces showing the pseudocontact shifts corresponding to ±3, ±1.5 and ±0.5 ppm where positive values are denoted in blue and negative values are denoted in red. The three dimensional structures of is shown in grey and is shown in yellow using ribbon representation.15

While the PCS can be easily measured, the magnetic susceptibility tensor and the location of the paramagnetic centre which are determined using PCS are not so readily obtained. The magnetic susceptibility tensor and the location of the paramagnetic centre can be described using a minimum of 8 pseudocontact shifts.16 Equation (1.2) is solved using an iterative numerical method, providing the location of the metal ion, the magnitude and orientation of the magnetic susceptibility tensor. Additional pseudocontact shifts provide extra constraints and reduce the error in the solution of the tensor and the location of the metal ion. From Figure 1.3b, it can be seen that in the case of 186, the PCS distribution is symmetrical and as a result, for each set of PCS there are 4 possible solutions.15 To overcome this potential ambiguity, additional PCSs need to be measured using a different paramagnetic source where the magnetic susceptibility tensors are different (therefore giving rise to a different PCS distribution). Based on a comparison between multiple sets of data, the ambiguity can then in principle be removed giving one solution which satisfies the different sets of data.15

1.3 Functionalized Metal Complexes for Catalysis

Complex organic molecules are fundamentally important in areas such as pharmaceuticals, biomolecule design and material science, and as a result significant attention has been devoted to their synthesis. In order to increase the structural flexibility (i.e. high functional group tolerance) and synthetic efficiency (i.e. mild reaction conditions, high yields and low waste) in the synthesis of complex organic molecules, there has been a move away from the use of traditional multistep synthesis. The synthesis of highly functionalized organic molecules can be approached by the use of one-step metal catalyzed coupling of basic building blocks derived from simple organic molecules. These basic building blocks, also known as synthons, are typically composed of alkynes, alkenes, aldehydes, ketones, alcohols and amines, which are easily functionalized with additional substituents via the use of standard organic methods. One obstacle to the use of these synthons is their lack of reactivity. Many of the proposed organic synthons are not susceptible to reactions such as nucleophilic attack (i.e. alkene or alkynes) and as such require activation to enable coupling of the organic synthons. One method to activate these inert building blocks is through the use of metal complexes as catalysts. Lanthanides and transition metal complexes have been widely utilized for the activation of alkynes, alkenes, amines and alcohols to produce complex heterocycles such as spiroacetals,3,5 pyrroles4,6 and lactones.46 The use of metal complexes allows the synthesis of complex organic molecules through the use of simple organic synthons in an efficient modular fashion under mild conditions.

1.3.1 Catalytic Cycle

A catalyst is defined as a substance that increases the rate of reaction by providing an alternative mechanistic pathway which has a lower activation energy compared to the uncatalyzed reaction and is not consumed during the reaction.47 It is important to note that the catalyst itself is not consumed and does not change the overall thermodynamics of a reaction. A typical catalytic cycle will involve a number of processes (Figure 1.4), and through an understanding of these processes, the reactivities and selectivities of the metal catalyst can be adjusted. At the start of a typical catalytic cycle the metal complex

(MLn) loses one of its co-ligands (L) through a dissociation process, and then proceeds to bind to the first of the reactants (A). A second reactant (B) then binds to the metal centre, where it reacts and converts to the desired product (A-B). In order to complete the catalytic cycle, the product is released and the active catalytic species is regenerated. It is important to note that one of the steps of the cycle must be irreversible (or close to) to ensure the cycle is completed.

Figure 1.4: General catalytic cycle involving the complex (MLn) and the reactants (A and B).

As with all chemical reactions involving multiple steps, the overall rate of reaction is determined by the rate determining step (RDS), which is the slowest step in the cascade of reactions in the catalytic cycle. Through a detailed mechanistic study, the RDS can be determined and the metal complex designed to increase the rate of the RDS, thus increasing the overall efficiency of the catalytic cycle.

As indicated above, the typical catalytic cycle involves several distinct steps, which include; (1) ligand substitution; (2) oxidative addition / reductive elimination; (3) !-hydride elimination; and (4) nucleophilic or electrophilic attack on ligand. The same processes are present in catalytic cycles in both homogeneous and heterogeneous catalysts.

1.3.2 Homogeneous vs. Heterogeneous Catalyst

There are two major classes of catalyst - homogeneous catalysts and heterogeneous catalysts. Homogeneous catalysts are catalysts which are soluble in the reaction medium and hence exist in the same phase as the reactants. On the other hand, as the name suggests, heterogeneous catalysts exist in a different phase to the reactants and thus reactions occur at the interface.

Heterogeneous catalysts have been widely adopted in many industrial processes such as the Haber-Bosch process to produce ammonia on an industrial scale. Since the catalyst exists in a different phase to the products, it allows for easy separation from the reaction medium and thus heterogeneous catalysts are the preferred catalysts in most industrial applications.48 Despite the widespread use of heterogeneous catalysts, and significant progress in understanding surface science, the mode of action of many heterogeneous catalysts is still not well understood. Without a detailed knowledge of structure to activity relationships, the ability to control the reactivity and selectivity of heterogeneous catalyst is limited. This is an aspect that has been developed significantly in homogeneous catalysis.

Homogeneous catalysts offer a large number of additional benefits over heterogeneous catalyst, including the ability for the selectivity and reactivity to be adjusted through the steric and electronic properties of the catalyst. This is only possible due to the well defined structure of metal complexes, which allows the kinetics and mechanism to be elucidated. Despite the advantages, the limitations of homogeneous catalysis include difficulties in catalyst-product separation and low thermal stability (<200oC).48 In order to combine the advantages of both homogeneous and heterogeneous catalysts, homogeneous catalysts have been immobilized onto macroscopic scaffolds such as silicon21,49 and ion exchange resins.50 This allows the reactivity of the catalyst to be tuned, but also eliminates the difficulty of product separation as the catalyst and the product are no longer in the same phase.

1.4 Project Aims

The aim of this project is the development of functionalized lanthanide and transition metal complexes and their application to protein structure refinement using paramagnetic NMR spectroscopy as well as the development of functionalized transition metal complexes and their application as homogeneous catalysts for the addition of oxygen onto unsaturated bonds. The specific objectives were to:

(a) Synthesize thiol modified dipicolinic acids designed to bind to both paramagnetic transition metal and lanthanide ions allowing paramagnetic labels to be rigidly and site specifically attached to proteins of interest.

(b) Synthesize unnatural amino acids bearing lanthanide and transition metal binding motifs to allow a paramagnetic probe to be introduced in a site specific manner without the use of post-translational modifications.

(c) Synthesize rhodium(I) complexes bearing functionalized tridentate N,N donor ligands and through the use of single crystal x-ray diffraction and variable temperature NMR spectroscopy investigate the influence of the reactive functional group on the coordination chemistry.

(d) Compare the catalytic activity of the rhodium(I) complexes bearing the functionalized tridentate N,N donor to their bidentate counterparts in the metal- catalyzed addition of oxygen to unsaturated bonds.

1.5 General structure of the thesis

The thesis contains a total of 6 chapters, an experimental section and appendices. The general outline of this thesis is shown below:

Chapter 1: General introduction of the use of metal complexes in protein structure refinement and catalysis

Chapter 2 and 3: The synthesis of thiol modified dipicolinic acids, 4MMDPA (5) and 3MDPA (9) and their application to the structure refinement of the N-terminal domain of the arginine repressor from E. Coli.

Chapter 4: The synthesis of unnatural amino acids DPA-AA (11), IDA-AA (15) and HQ-AA (19) for applications as paramagnetic labels to be introduced using a modified protein expression system.

Chapter 5: The synthesis and characterization of rhodium(I) complexes bearing the tris(pyrazolyl)toluidine (24) and tris(N-methylimidazolyl)methanol (23) ligand in the solid and solution state using single crystal x-ray diffraction and variable temperature NMR spectroscopy.

Chapter 6: The application of rhodium(I) complexes discussed in Chapter 5 as catalysts for the intramolecular cyclization of aliphatic and aromatic alkynoic acids and mechanistic study of the cyclization reactions.

Chapter 7: Summary, conclusion and future work

Chapter 8: Synthetic methods, catalytic methods and characterization data for all new compounds synthesized

Appendices: The PCS and RDC assignments for the ArgN-4MMDPA-Ln3+, ArgN-3MDPA-Co2+ adducts, full x-ray crystallographic data and time course data for catalytic reactions. The appendices are supplied in digital form on the CD attached to the back of the cover.

1.6 References

(1) Stubbert, B. D.; Marks, T. J. J. Am. Chem. Soc. 2007, 129, 4253-4271. (2) Yu, X.; Seo, S.; Marks, T. J. J. Am. Chem. Soc. 2007, 129, 7244-7245. (3) Seo, S.; Yu, X.; Marks, T. J. J. Am. Chem. Soc. 2009, 131, 263-276. (4) Field, L. D.; Messerle, B. A.; Vuong, K. Q.; Turner, P. Dalton Trans. 2009, 3599- 3614. (5) Messerle, B. A.; Vuong, K. Q. Pure Appl. Chem. 2006, 78, 385-390. (6) Motta, A.; Fragala, I. L.; Marks, T. J. Organometallics 2006, 25, 5533-5539. (7) Zhao, J.; Marks, T. J. Organometallics 2006, 25, 4763-4772. (8) Dabb, S. L.; Ho, J. H. H.; Hodgson, R.; Messerle, B. A.; Wagler, J. Dalton Trans. 2009, 634-642. (9) Silverio, S.; Torres, S.; Martins, A. F.; Martins, J. A.; Andre, J. P.; Helm, L.; Prata, M. I. M.; Santos, A. C.; Geraldes, C. F. G. C. Dalton Trans. 2009, 4656- 4670. (10) Mayoral, E. P.; Garcia-Amo, M.; Lopez, P.; Soriano, E.; Cerdan, S.; Ballesteros, P. Bioorg. Med. Chem. 2003, 11, 5555-5567. (11) Yoo, B.; Pagel, M. D. J. Am. Chem. Soc. 2006, 128, 14032-14033. (12) Mizukami, S.; Takikawa, R.; Sugihara, F.; Hori, Y.; Tochio, H.; Walchli, M.; Shirakawa, M.; Kikuchi, K. J. Am. Chem. Soc. 2008, 130, 794-795. (13) Moret, V.; Laras, Y.; Cresteil, T.; Aubert, G.; Ping, D. Q.; Di, C.; Barthelemy- Requin, M.; Beclin, C.; Peyrot, V.; Allegro, D.; Rolland, A.; De Angelis, F.; Gatti, E.; Pierre, P.; Pasquini, L.; Petrucci, E.; Testa, U.; Kraus, J.-L. Eur. J. Med. Chem. 2009, 44, 558-567. (14) Peacock, A. F. A.; Parsons, S.; Sadler, P. J. J. Am. Chem. Soc. 2007, 129, 3348- 3357. (15) Pintacuda, G.; Park, A. Y.; Keniry, M. A.; Dixon, N. E.; Otting, G. J. Am. Chem. Soc. 2006, 128, 3696-3702. (16) Pintacuda, G.; John, M.; Su, X. C.; Otting, G. Acc. Chem. Res. 2007, 40, 206-212. (17) Keniry, M. A.; Park, A. Y.; Owen, E. A.; Hamdan, S. M.; Pintacuda, G.; Otting, G.; Dixon, N. E. J. Bacteriol. 2006, 188, 4464-4473. (18) Reetz, M. T. Top. Organomet. Chem. 2009, 25, 63-92.

(19) Reetz, M. T.; Rentzsch, M.; Pletsch, A.; Taglieber, A.; Hollmann, F.; Mondiere, R. J. G.; Dickmann, N.; Hoecker, B.; Cerrone, S.; Haeger, M. C.; Sterner, R. ChemBioChem 2008, 9, 552-564. (20) McDonald, A. R.; Dijkstra, H. P.; Suijkerbuijk, B. M. J. M.; van Klink, G. P. M.; van Koten, G. Organometallics 2009, 28, 4689-4699. (21) Takahashi, T.; Watahiki, T.; Kitazume, S.; Yasuda, H.; Sakakura, T. Chem Commun (Camb) 2006, 1664-1666. (22) Wang, Z.; Chen, G.; Ding, K. Chem. Rev. 2008, 109, 322-359. (23) Copéret, C.; Basset, J.-M. Adv. Synth. Catal. 2007, 349, 78-92. (24) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry; 5th ed., 2001. (25) Hsuan-Liang Liu; Jyh-Ping Hsu Proteomics 2005, 5, 2056-2068. (26) Buchanan, S. G. Curr. Opin. Drug Discov. Devel. 2002, 5, 367-381. (27) Staunton, D.; Owen, J.; Campbell, I. D. Acc. Chem. Res. 2003, 36, 207-214. (28) Lamzin, V. S.; Perrakis, A. Nat. Struct. Biol.. 2000, 7, 978-981. (29) Abola, E.; Kuhn, P.; Earnest, T.; Stevens, R. C. Nat. Struct. Biol.. 2000, 7, 973- 977. (30) Szyperski, T.; Wider, G.; Bushweller, J. H.; Wuethrich, K. J. Am. Chem. Soc. 1993, 115, 9307-9308. (31) Medek, A.; Olejniczak, E. T.; Meadows, R. P.; Fesik, S. W. J. Biomol. NMR 2000, 18, 229-238. (32) Hiller, S.; Fiorito, F.; Wuethrich, K.; Wider, G. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 10876-10881. (33) Lopez-Mendez, B.; Guntert, P. 2006, 128, 13112-13122. (34) Clore, G. M.; Gronenborn, A. M. Curr. Opin. Chem. Biol.t 1998, 2, 564-570. (35) Battiste, J. L.; Wagner, G. Biochemistry 2000, 39, 5355-5365. (36) Allegrozzi, M.; Bertini, I.; Janik, M. B. L.; Lee, Y.-M.; Liu, G.; Luchinat, C. J. Am. Chem. Soc. 2000, 122, 4154-4161. (37) Gochin, M.; Roder, H. Protein Sci. 1995, 4, 296-305. (38) Bertini, I.; Luchinat, C.; Parigi, G. Concepts Magn. Reson. 2002, 14, 259-286. (39) Gaponenko, V.; Sarma, S. P.; Altieri, A. S.; Horita, D. A.; Li, J.; Byrd, R. A. J. Biomol. NMR 2004, 28, 205-212. (40) Nomura, M.; Kobayashi, T.; Kohno, T.; Fujiwara, K.; Tenno, T.; Shirakawa, M.; Ishizaki, I.; Yamamoto, K.; Matsuyama, T.; Mishima, M.; Kojima, C. FEBS Lett. 2004, 566, 157-161.

(41) Niccolai, N.; Spadaccini, R.; Scarselli, M.; Bernini, A.; Crescenzi, O.; Spiga, O.; Ciutti, A.; Di Maro, D.; Bracci, L.; Dalvit, C.; Temussi, P. A. Protein Sci. 2001, 10, 1498-1507. (42) Sakakura, M.; Noba, S.; Luchette, P. A.; Shimada, I.; Prosser, R. S. J. Am. Chem. Soc. 2005, 127, 5826-5832. (43) Veglia, G.; Opella, S. J. J. Am. Chem. Soc. 2000, 122, 11733-11734. (44) Liepinsh, E.; Baryshev, M.; Sharipo, A.; Ingelman-Sundberg, M.; Otting, G.; Mkrtchian, S. Structure 2001, 9, 457-471. (45) John, M.; Park, A. Y.; Pintacuda, G.; Dixon, N. E.; Otting, G. J. Am. Chem. Soc. 2005, 127, 17190-17191. (46) Elgafi, S.; Field, L. D.; Messerle, B. A. J. Organomet. Chem. 2000, 607, 97-104. (47) Silberberg, M. Chemistry, The Molecular Nature of Matter and Change; Mobsy: St. Louis, 1996. (48) Wolfgang A. Herrmann; Boy Cornils Angew. Chem., Int. Ed. Engl. 1997, 36, 1048- 1067. (49) Riddel, S. A.; Hems, W. P.; Chesney, A.; Watson, S. R. Stud. Surf. Sci. Catal. 2003, 146, 799-801. (50) Kopylova, V. D. Zh. Fiz. Khim. 1989, 63, 1153-1164.

Chapter 2: Dipicolinic Acid as Paramagnetic Probe – 4MMDPA

“We choose to go to the moon in this decade, and do the other things, not because they are easy, but because they are hard, because that goal will serve to organize and measure the best of our energies and skills, because the challenge is one that we are willing accept, one we are unwilling to postpone, and one which we intend to win…”

John F. Kennedy (1917 – 1963)

2.1 Introduction

Paramagnetism from either paramagnetic transition metal or lanthanide ions provides a powerful tool for the refinement of protein structures and the study of protein-ligand interactions using NMR spectroscopy. The effect of paramagnetism in NMR spectroscopy can be observed through either pseudocontact shift (PCS) or paramagnetic relaxation enhancement (PRE). Of these effects, PCS provides the most information due to the fact that it can be measured with high accuracy and can be observed for nuclei up to 40 Å away.1

The use of paramagnetism in protein studies is a powerful tool, but the introduction of a paramagnetic centre into a protein is not a simple task. Existing paramagnetic probes are generally large molecules which contain flexible linkers which reduces their ability to generate the required paramagnetic effects. This has driven the desire to investigate the development of novel small paramagnetic probes that can be attached in a rigid fashion to the protein.

2.2 Paramagnetic Probes

In order to measure PCS, the first step is to introduce a paramagnetic probe into the protein in a site specific manner. In order for the paramagnetic probe to be effective, a number of criteria need to be met. They include: (1) the paramagnetic probe must remain rigid relative to the protein on the NMR time scale; (2) the paramagnetic probe must not form diastereoisomers upon binding with the protein; and (3) the probe must have a high binding affinity for the metal ion.

The magnitude and size of the PCS is dependent on the location of the nuclear spin relative to the paramagnetic centre. It is this spatial dependence of the PCS that makes the rigidity important. If the probe is moving relative to the protein on the NMR time scale during acquisition, the nuclear spin interacts with different magnetic susceptibility tensors. As a result, different PCSs are observed for the different tensors and thus the

resulting cross peak will become an average, leading to line broadening depending on the rate of exchange. The difficulty with chirality arises when the paramagnetic probe is pro-chiral. If a pro-chiral paramagnetic probe is bound to the protein, diastereoisomers will be formed on binding. Since each diastereoisomer has different susceptibility tensors, this leads to the appearance of multiple PCS cross peaks. It is also important that the probe binds strongly to the protein, as chemical exchange can occur. Depending on the rate of the chemical exchange, this can lead to either a broadening of the 1H-15N HSQC cross peaks or the formation of additional cross peaks, both of which are not ideal. Under certain conditions, this chemical exchange can be used to assist with the assignment of the paramagnetic cross peaks.2

Currently there are three methods of introducing a paramagnetic lanthanide ion into a protein, which include: (a) the substitution of a lanthanide ion into a metalloprotein; (b) N- and C-terminal fusion of proteins with lanthanide binding peptides; (c) attachment of a lanthanide binding ligand onto a protein in a site specific manner.

2.2.1 Metal Substitution in Metalloproteins

Metalloproteins are a class of proteins which require a bound metal as part of the normal physiological function. As such there are pockets within the protein which contain functional groups which are capable of binding to metal ions. Of particular interest are metalloproteins which are designed to bind metals such as calcium or zinc. Through the use of chelation, these metals can often be removed and replaced by a lanthanide ion.3

It has been shown that once a lanthanide ion has been substituted, detailed PCS data can be measured and used for the refinement of the structure.3-5 The major limitation of this method is that the protein must have a metal binding site with the required coordination sites to bind lanthanide ion. In order to provide a more general approach to attaching

lanthanide ions onto a protein, the use of lanthanide binding peptides has been proposed.

2.2.2 Lanthanide Binding Peptides

Through the work by Imperiali et al.6-7 it has been shown that a peptide sequence can bind lanthanide ions with a binding affinity of 50 nM. One of the benefits of the lanthanide binding peptide is that it can be synthesized enantiomerically pure and thus preventing the formation of diastereoisomers.

The application of the lanthanide binding peptides (LBP) in protein structure refinement was demonstrated by Su et al.8 A 16 residue peptide sequence designed to bind lanthanide ions was attached to the N-terminal domain of the arginine repressor from E. coli (ArgN). The lanthanide binding peptide contains a cysteine residue which allows the peptide to be attached to the protein via the formation of a disulfide bond. The resulting 1H-15N HSQC spectrum of the ArgN-LBP showed many large measurable PCS indicating that LBP is an effective paramagnetic probe.8

Due to the symmetry in the magnetic susceptibility tensor, there are ambiguities in the use of PCS in the refinement of protein structures. This ambiguity is removed by refining the structure of the protein using PCS data from different magnetic susceptibility tensors, achieved through the measurement of the PCS using different paramagnetic lanthanide ions or changing the chemical environment around the lanthanide ion (i.e. different ligand). Due to the high binding affinity of the lanthanide binding peptides, the exchange of different lanthanide ions on the peptide becomes difficult. So rather than changing the lanthanide ion, a change in chemical environment around the lanthanide ion is used. This was achieved through the use of lanthanide binding peptides with different sequences as shown by Su et al.9

An ideal paramagnetic probe should allow the measurement of the protein in the presence of different magnetic susceptibility tensors. The difficulty involved in changing the chemical environment around the lanthanide ion is a major limitation of the use of LBP as a paramagnetic probe. In addition, the bulkiness of the peptide tags could potentially interfere with the biological function of the protein. To overcome these limitations, synthetic lanthanide complexes have been explored as a potential replacement for the lanthanide binding peptides.

2.2.3 Lanthanide Binding Ligands

In an attempt to design a paramagnetic probe that allows the lanthanide ion to be changed readily, synthetic lanthanide complexes have been proposed. The design and synthesis of organic ligands that can bind to lanthanide ions are well established10 but to do so in fashion where it can be attached rigidly in a site specific manner is not. Currently, all lanthanide probes in literature are composed of a macrocyclic core (i.e. cryptand, ethylenediamine and cyclam) with functional groups present for the attachment onto the protein. Examples of these complexes are shown below in Figure 2.1.

Figure 2.1: Structures of the CLaNP11, CLaNP-312 and CLaNP-5.213 lanthanide binding ligands for attachment to proteins. 22

A common feature on all of the ligands shown in Figure 2.1 is that they contain both a nitrogen and oxygen donor. It is well established in literature that nitrogen donors provide thermodynamic stability to the complex, while oxygen donors provide kinetic stability to the complex.14 In order to produce a complex that is both stable and has a high binding affinity, a combination of the two types of donors is required.

The first generation of the series shown in Figure 2.1 is simply a diethylenetriaminepentaacetic acid (Figure 2.1a) analogue that contains a methanethiolsulfonate (S-SO2-CH3) group, which is used to attach the ligand to the protein.11 One of the problems experienced by the CLaNP ligand is that multiple conformers were observed in the resulting 15N-1H HSQC indicated by the formation of multiple cross peaks.11 The cyclam based ligand (Figure 2.1b and c), which is conformationally more rigid, was used to prevent the formation of isomers. Compared to the CLaNP probe, the HSQC spectrum of the proteins containing the CLaNP-3 and CLaNP-5.2 probes shows only one paramagnetic cross peak indicating the ions are bound in one conformation.12-13

Despite the fact that each of the macrocyclic probes shown in Figure 2.1 can be bound in a site specific rigid manner, it is difficult to remove the lanthanide ion from these ligands which is the same problem that limits the widespread use of the lanthanide binding peptides. To overcome this difficulty, a new class of ligand has been proposed, which is a functionalized ligand based on the dipicolinic acid motif.

2.3 Dipicolinic Acid Based Paramagnetic Probes

Dipicolinic acid (DPA) is a tridentate N,O,O donor ligand that is known to bind to lanthanide ions with high binding affinity and stability.15-16 Due to the well defined chemistry and structural data available on the dipicolinic acid motif, it is an ideal starting point to build paramagnetic probe that can not only be attached rigidly in a site

specific manner but also allows the lanthanide ion to be easily replaced to alter the magnetic susceptibility tensor.

2.3.1 Retrosynthesis

A functionalized dipicolinic acid tag, 4-mercaptomethyl-2,6-pyridinedicarboxylic acid (4MMDPA, 5) was proposed to allow the dipicolinic acid motif to be attached to a protein via a disulfide bond.

Based on the desired compound, a retrosynthesis was established and is shown in Scheme 2.1. There are four possible places where the disconnections can occur and they include: (1) the carboxylic acid groups at the 2 and 6 position; (2) the hydroxymethyl group at the 4 position; and (3) a direct synthesis of the pyridine using aldehydes with desired functional groups as the building blocks.

Scheme 2.1: The four possible disconnections in the retrosynthesis of 4- mercaptomethyl-2,6-pyridinedicarboxylic acid (5)

After assessing the various advantages and disadvantages of each disconnection, it was decided that the first approach, that is the functionalization of the 4 position, would be used. The final synthetic method that was used to synthesize 4-mercaptomethyl-2,6- pyridinedicarboxylic acid (4MMDPA, 5) is summarized below in Scheme 2.2.

Scheme 2.2: Synthesis of 4-mercaptomethyl-2,6-pyridinedicarboxylic acid (5).

In Scheme 2.2, the synthesis of the 4MMDPA ligand (5) starts with 2,6-dimethoxycarbonylpyridine (1) which is synthesized from commercially available dipicolinic acid using the standard ester synthesis.17 From the ester 1, the hydroxymethyl functional group is introduced using Fenton’s reagent,17 generated in- situ from hydrogen peroxide and iron sulfate followed by purification using flash chromatography. The identity of the resulting 4-hydroxymethyl-2,6- diemthoxycarbonylpyridine (2) was confirmed through the comparison with literature spectroscopic data.17 From the alcohol, the tosylate 3 was formed via reaction of 2 with tosyl chloride in the presence of triethylamine in cold dichloromethane and the purity of the resulting compound was established by comparison with literature spectroscopic data.17 The corresponding bromide 4 was synthesized from the tosylate 3 by stirring with lithium bromide in acetone and purified using flash chromatography. The structure and purity of 4 was confirmed through the use of spectroscopic data and elemental analysis. The bromide 4 was then converted to the corresponding thiol 5 by treatment

with thiourea in methanol to form the isothiouronium salt which was then hydrolyzed in-situ to the thiolate. It is important to note that during the hydrolysis step, the reaction must be conducted under an inert atmosphere to avoid the oxidation of the thiol group to form the disulfide. To yield the protonated 4MMDPA (5) compound, an aqueous solution of the thiolate was passed through an ion exchange column to yield the desired thiol compound 5 in excellent yields (99%). The structure and purity of compounds 4 and 5 were confirmed from the spectroscopic data and high resolution mass spectroscopy.

The next step was the attachment of the tag onto the protein and assessment of its performance as a paramagnetic probe.

2.3.2 Attachment of 4MMDPA onto N-terminal domain of the arginine repressor

In order to examine the efficiency of the 4MMDPA (5) as a paramagnetic probe, the N-terminal domain of the arginine repressor (ArgN) was used. The advantages of using the ArgN protein is that not only is the three dimensional structure of protein well known, but it also contains a surface accessible cysteine residue to attach the 4MMPDA (5) probe.18

The attachment of the probe onto the ArgN protein was achieved through the use of the protocol published by Su et al.8 and is shown schematically in Scheme 2.3. This scheme involves treating the ArgN protein with 5,5’-dithio-bis(2-nitrobenzoic acid) (DTNB) to activate the thiol group which yields the activated species (ArgN-TNB). The order of addition is important to avoid the formation of ArgN dimers. The reaction is complete as indicated by the appearance of a yellow solution due to the release of the 5-thio-2- nitrobenzoate (TNB-) anion. The activated protein was then separated from the unreacted DTNB through centrifugal filtration, where an alkaline solution (pH > 8) of 4MMDPA (5) was added. The functionalized protein was then separated and filtered using the same procedure as described above.

Scheme 2.3: Functionalization of the ArgN using the 4MMDPA (5) lanthanide binding tag.

A 1H-15N HSQC spectrum of the resulting ArgN-4MMDPA adduct was measured (Figure 2.2a, green) and compared with a HSQC spectrum of the free ArgN protein (Figure 2.2a, red). By overlaying the two spectra small shifts in the HSQC cross peaks can be observed. Plotting these deviations in chemical shift against their residue numbers (Figure 2.2b), the largest deviations in the 15N chemical shift were observed at the residues 17 – 25 and 52 – 68. These areas correspond to the amino acid residues that reside closest to the ligation site (Figure 2.3).

Figure 2.3: A three-dimensional ribbon representation of the ArgN protein with the residues Glu20 – Ser25 and Gly54 – Ala 71 highlighted in red. The three dimensional structure of the 4MMDPA probe (5) attached to the Cys68 is superimposed for visual guidance and does not represent the actual orientation of the ligand in the ArgN- 4MMDPA adduct.

2.3.3 NMR spectroscopy of ArgN-4MMDPA in the presence of Diamagnetic & Paramagnetic Lanthanide Ions.

A 1H-15N HSQC spectrum of ArgN-4MMDPA adduct in the presence of Tm3+ was measured and is shown in Figure 2.4. Two observations can be made: (1) upon inspection of the spectrum, for every diamagnetic peak (red), there are two additional cross peaks that appear upon the addition of the paramagnetic ion (Figure 2.4a); and (2) the peaks become broad, due to the paramagnetic relaxation enhancement (Figure 2.4b).

Figure 2.4: Overlay of the 1H-15N HSQC spectra of the unpurified ArgN-4MMDPA adduct in the presence of a diamagnetic lanthanide ion (red) and in the presence of a paramagnetic lanthanide ion (green).

Figure 2.5: Overlay of the 1H-15N HSQC spectra of the purified ArgN-4MMDPA adduct in the presence of a diamagnetic lanthanide ion (cyan) and in the presence of a paramagnetic lanthanide ion (purple).

The increased line broadening is expected due to the influence of the PRE, however this was not a major issue, as only resonances close to the paramagnetic centre will be lost due to significant line broadening. The appearance of additional cross peaks was a 30

major concern as it can complicate the assignment of the HSQC spectrum. A possible explanation for the formation of the additional cross peaks in HSQC spectrum could be due to the presence of impurities, thus a sample of the ArgN-4MMDPA adduct was purified using fast protein liquid chromatography (FPLC) using a MonoQ column. The above measurement was repeated with the purified ArgN-4MMDPA adduct and the resulting HSQC spectrum shown in Figure 2.5.

Compared to Figure 2.4, the purified ArgN-4MMDPA adduct shows only one cross peak in the HSQC spectrum upon the addition of Tm3+. Additional HSQC spectra in the presence of other paramagnetic lanthanide ions were measured and the behaviour of selected residues (Gly27 and Gly54) is shown in Figure 2.6. For both G27 and G54, the paramagnetic cross peaks for the different lanthanide ions all appear in a straight line. It is also important to note that the direction of the PCS was not constant, for example for Tm3+, a positive PCS was observed for residue G27, while for residue G54, a negative PCS was observed.

From the assignments of the ArgN-4MMDPA-Lu3+ HSQC spectrum and those measured in the presence of a paramagnetic lanthanide ion (i.e. Tm3+, Yb3+ and Tb3+) provided by Dr Xun-Cheng Su, the PCS for the ArgN-4MMDPA-Ln3+ adducts were calculated. Plotting the measured PCS (Figure 2.7) shows that for Ytterbium (Yb3+) and Thulium (Tm3+) positive PCS was generally observed for most amino acid residues while for Terbium (Tb3+) negative PCS was observed. It is interesting to note that regardless of the lanthanide ion used, the maximum PCS was observed around the residues 13 – 25 with smaller shifts observed around residues 54 – 71.

G27 G54 Tm3+ Diamagnetic Diamagnetic Yb3+ Tb3+

Figure 2.6: Overlay of the 1H-15N HSQC spectra of G27 and G54 residue in the ArgN- 4MMDPA adduct in the presence of a diamagnetic lanthanide ion and in the presence of a paramagnetic lanthanide ion Tm3+ (green), Yb3+ (purple) and Tb3+ (orange). Paramagnetic peaks belonging to the same amino acid residue are connected using a dashed blue line.

Figure 2.7: Plot of the PCS of the amide protons in the 1H dimension vs. the amino acid residue number for Tb3+ (J), Tm3+ (L) and Yb3+ ([).

Having determined the PCS for the amide protons in the ArgN-4MMDPA adduct, the next step is to use the PCS data and calculate the anisotopic components of the magnetic susceptibility tensor and the corresponding metal ion location.

2.3.4 Solving the Magnetic Susceptibility Tensor and Metal Ion Locations

Based on the 1H-15N HSQC spectrum (Appendix 1.1.1 – 1.1.4) of the ArgN-4MMDPA- Ln3+ adducts (Ln =Lu3+, Tb3+, Tm3+ and Yb3+), the components of the magnetic susceptibility tensor and the metal ion location can be determined.

In order to calculate the anisotropic components of the magnetic susceptibility tensor and the location of the metal ion, the coordinates of the amide protons within the protein structure were required. In the case of the ArgN protein, the structure has been solved previously by Sunnerhagen et al.18 Unlike structures solved using X-ray crystallography, the structure reported by Sunnerhagen et al. is composed of 23 minimum energy conformations based on the nOe constraints.18 To limit the error in the solutions of the location of the metal ions and the magnetic susceptibility tensors, the PCS data in the 1H dimension measured in the presence of Tb3+, Tm3+ and Yb3+ were solved simultaneously with the assumption that the locations of the metal ions do not vary significantly. This assumption is possible due the similar chemical behaviour of the lanthanide series. The location of the metal ions and the magnetic susceptibility tensors were calculated using all 23 structures simultaneously without any additional boundary conditions. The back calculated PCS for all three paramagnetic lanthanide ions were tabulated and plotted against the experimentally observed PCS (Figure 2.8). The PCS isosurfaces were superimposed onto the three dimensional structure of the ArgN protein and are shown in Figure 2.9.

(a) ArgN-4MMDPA + Tb3+ 1.5 1 0.5 y = 0.9971x 0 R² = 0.9964 -0.5 -1

Calculated PCS (ppm) -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5

(b) ArgN-4MMDPA + Tm3+ 2

1.5 y = 0.999x 1 R² = 0.9982

0.5

0 Calculated PCS (ppm) -0.5 -0.5 0 0.5 1 1.5 2

Calculated PCS (ppm) -1 -1.5 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 Experimental PCS (ppm)

Figure 2.8: The plot of the back calculated PCS vs. the observed PCS for ArgN- 4MMDPA: (a) Tb3+; (b) Tm3+; and (c) Yb3+. The predicted PCS was calculated by solving all three sets of experimental data using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.19

(a) Tb3+ D0ax = 13.3 ± 0.7 D0ax = 5.1 ± 0.5

(b) Tm3+ D0ax = 12.2 ± 0.4 D0ax = 6.8 ± 0.8

Figure 2.9: PCS Isosurface for ±1 and ±0.2 ppm (blue = positive PCS, red = negative PCS) superimposed on the three-dimensional structure of the ArgN for: (a) Tb3+; (b) Tm3+; and (c) Yb3+ visualized using PyMol. The PCS isosurfaces were calculated using all three sets of experimental data using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.19 0-32 m3 unless otherwise stated.

Plotting the back-calculated PCS against the experimentally observed PCS shows that the two sets of values are in good agreement. The linear regressions for the three sets of data all shows a slope of 1 with an R2 value of 0.99 indicating a strong linear correlation between the back-calculated PCS and the experimentally observed PCS.

Plotting the PCS isosurfaces and superimposing them upon the three dimensional structure of the ArgN protein allow a quick visual comparison of the PCS distribution between the three paramagnetic lanthanide ions. Comparing the isosurface plots of ArgN-4MMDPA-Tb3+ (Figure 2.9a) and ArgN-4MMDPA-Tm3+ (Figure 2.9b), a reversal in the distribution of the lobes indicating regions of positive PCS was observed. This is reflected in the corresponding 1H-15N HSQC spectrum where the PCS in the Tb3+ spectrum (Figure 2.6, orange) would be in the opposite direction compared to the Tm3+ spectrum (Figure 2.6, green). In addition, it is important to note that the size of the 3+ 3+ lobes do not vary significantly between Tb and Tm , which is consistent with the small variation in the ax and rh components between the two paramagnetic metal centres. Comparing the isosurface plot of ArgN-4MMDPA-Tm3+ (Figure 2.9b) with that of ArgN-4MMDPA-Yb3+ (Figure 2.9c), the PCS distribution remains unchanged but the size of the lobes are approximately half the size compared to Tm3+. This was reflected in the magnitude of the ax and rh components where smaller values were calculated for Yb3+.

After calculating the tensor parameters, the rigidity of the probe needs to be considered, for this, the alignment tensor can be extracted from residual dipolar coupling.

2.3.5 Alignment Tensor

For a rigid tag system, the alignment tensor determined from the residual dipolar coupling (RDC) should have the same orientation as the magnetic susceptibility tensor determined from the PCS data.

Due to the high sensitivity of the alignment tensor to errors in the orientation of the NH bond, only the rigid elements of the tertiary structure are used in the calculation of the alignment tensor. Since the ArgN protein was solved using NMR spectroscopy, the accuracy of the NH bond is less accurate compared to that of a structure solved using x- ray crystallography. The erroneous orientation of the NH bond can lead to errors in the back-calculated RDC, the magnitude of the axial and the rhombic components of the alignment tensors.

Using the RDC measurements supplied by Dr Xun-Cheng Su (Appendix 1.1.5), the back-calculated RDC, the orientation and magnitude the alignment of the tensors for all three lanthanide ions were calculated using PALES20 shown in Figure 2.10, Figure 2.11 and Table 2.1 respectively. Due to a difference in the coordinate system used by PALES,20 the Euler angles of the alignment tensor calculated need to be transformed before they can directly compared to the tensors generated by NUMBAT.19

Comparing the predicted RDCs with the experimentally observed values, the immediate observation is that significant deviations are observed. This deviation is attributed to the inaccuracies in the three dimensional structure of the ArgN protein. A comparison of the alignment and the magnetic susceptibility tensor can be done by examining the Sanson- Flamsteed plot. Each point represents the intersection of the axes of the tensor with a sphere. Since the NMR structure of the ArgN protein is composed of 23 energy minimised conformers, the alignment and magnetic susceptibility tensor was solved individually for each of the 23 conforms and plotted in Figure 2.11.

(a) ArgN – 4MMDPA + Tb3+ 10

-5

-10 Calculated RDC (Hz)

-15 -11 -6 -1 4 9

(b) ArgN – 4MMDPA + Tm3+ 15

-5 Calculated RDC (Hz)

-10 -10 -5 0 5

Figure 2.10: The plot of the back-calculated RDC vs. the observed RDC for ArgN- 4MMDPA: (a) Tb3+; (b) Tm3+; and (c) Yb3+ using only the first model of the ArgN protein structure. The predicted RDC was calculated using PALES20 only including the RDCs from only the rigid element of the tertiary structure (i.e. residues 10 – 19, 26 – 36 and 43- 53).

Magnetic Susceptibility Alignment Tensor Tensor

(b) Tb3+

(a) Tm3+

For all 23 conformers, there is good correlation between the orientations of the two sets of tensors, which is expected for a paramagnetic probe bound to the protein in a rigid manner. The alignment tensors calculated directly from PALES require a correction before they can be compared to the tensors generated from NUMBAT. These corrections are stated below in eqn. (2.1).21

= 180

= (2.1)

= 180

The axial and rhombic components of the alignment and the magnetic susceptibility tensors were calculated and are shown in Table 2.1. The deviations in the magnitudes of the axial and rhombic components are attributed to the errors in the orientations of the NH bond in the NMR structure of the ArgN protein.

Table 2.1: The axial and rhombic components of the alignment tensor and magnetic susceptibility tensor for the ArgN-4MMDPA adduct in the presence of paramagnetic lanthanide ion.d

PALESa NUMBAT

b b c c Dax Drh 0ax 0rh 0ax 0rh

Tb3+ 4.2 ± 0.3 2.2 ± 0.3 9.2 ± 0.6 4.7 ± 0.6 13.3 ± 0.7 5.1 ± 0.5

Tm3+ 4.5 ± 0.2 1.9 ± 0.2 9.9 ± 0.4 4.2 ± 0.4 12.3 ± 0.4 6.8 ± 0.8

Yb3+ 2.3 ± 0.1 0.8 ± 0.1 5.07 ± 0.2 1.65 ± 0.2 5.8 ± 0.2 2.7 ± 0.3

(a) The solution of the alignment tensors only used residual dipolar couplings from only the rigid elements of the tertiary structure (i.e. residues 10 – 19, 26 – 36 and 43 – 53). The values reported are the average of the axial and rhombic components determined for each of the 23 conformers independently. (b) The axial and rhombic components of the alignment tensor calculated by PALES requires a correction factor of 2 to allow comparison with results calculated using alternative software such as MODULES. (c) Alignment tensors were converted to the corresponding magnetic susceptibility tensor values using eqn (2.2) shown below

15͟0 = 2 ̾ (2.2) 0 (d) 0$-32 m3 unless otherwise stated.

A comparison of the RDC data with PCS data shows that the 4MMDPA-Ln3+ complex was able to bind to the protein in a rigid manner. This was shown by the correlations between the alignment and the magnetic susceptibility tensors in the Sanson-Flamsteed plots. To gain a better understanding of how this rigidity was achieved the amino acid side chains surrounding the ligation site need to be considered (Figure 2.12).

Figure 2.12: A ribbon representation of the three-dimensional structure of the ArgN protein with the location of the metal ion (cyan sphere) and the E21 residue superimposed. The three dimensional structure of the 4MMDPA tag attached to the C68 residue is superimposed for visual guidance and does not represent the actual orientation of the ligand in the ArgN-4MMDPA complex in solution.

As stated in the introduction, lanthanide ions have a high affinity to carboxylate groups. Considering all of the carboxyl side chains, there are two in close proximity to the calculated position of the lanthanide ion (E21 and E20). From the NMR structure of the ArgN protein, only E21 is orientated in the correct orientation to bind to the lanthanide ion. Measuring the distance of the lanthanide ion to the carboxylate of the E21 residue,

a bond distance of 1.9Å was measured, which is within the range of a lanthanide- oxygen bond.22 This suggests that the carboxylate group of E21 was able bind to the lanthanide ion effectively providing a second attachment point increasing the overall rigidity of the probe.

2.4 Conclusion

One of the limitations of existing methods to introduce paramagnetic probes is the ease of access and generality. The use of macrocyclic complexes bearing two attachment points has allowed paramagnetic probes to be attached to proteins with great success. The use of the additional attachment point decreases the generality due to the need to introduce a second cysteine residue within the protein.

Based on the desire to increase the generality, a new functionalised lanthanide binding ligand (4MMPDA, 5) was synthesized and attached to the N-terminal domain of the arginine repressor from E. coli (ArgN). The NMR properties of the ArgN-4MMDPA adduct were examined in the presence of multiple paramagnetic lanthanide ions. The resulting 1H-15N HSQC spectrum showed large measurable pseudocontact shifts (PCS). Based on the measured PCS, the location of the metal ions and the magnetic susceptibility tensors for a series of lanthanides ions were determined and compared with the alignment tensors derived from the residue dipolar couplings (RDC). Significant deviations were observed in the magnitude of the axial and rhombic components of the alignment and the magnetic susceptibility tensor, while the orientations showed good agreement. This deviation could be attributed to the uncertainties in the orientation of the NH bond in the NMR structure of the ArgN protein. This indicates that the 4MMDPA-Ln3+ complex was able to bind to the protein in a rigid manner. Upon close examination of the amino acid side chains surrounding the ligation site, presumably a glutamate residue (E21) binds to the lanthanide ion providing a second attachment point increasing the overall rigidity of the system.

This represents the first example of a small molecule that is capable of forming a lanthanide complex, and being bound to the protein in a rigid fashion allowing large PCS to be measured using only one attachment point. The ability to use one attachment point reduces the overall complexity of the ligation process and the small size of the tag itself reduces the likelihood of interference when used to solve the structure of a protein-protein complex.

2.5 Reference

(1) Pintacuda, G.; John, M.; Su, X. C.; Otting, G. Acc. Chem. Res. 2007, 40, 206- 212. (2) John, M.; Headlam, M.; Dixon, N.; Otting, G. J. Biomol. NMR 2007, 37, 43-51. (3) Bertini, I.; Luchinat, C.; Parigi, G. Concepts Magn. Reson. 2002, 14, 259-286. (4) Gaponenko, V.; Sarma, S. P.; Altieri, A. S.; Horita, D. A.; Li, J.; Byrd, R. A. J. Biomol. NMR 2004, 28, 205-212. (5) Gochin, M.; Roder, H. Protein Sci. 1995, 4, 296-305. (6) Nitz, M.; Franz, K. J.; Maglathlin, R. L.; Imperiali, B. ChemBioChem 2003, 4, 272 -276. (7) Nitz, M.; Sherawat, M.; Franz Katherine, J.; Peisach, E.; Allen Karen, N.; Imperiali, B. Angew. Chem., Int. Ed. Engl. 2004, 43, 3682-3685. (8) Su, X.-C.; Huber, T.; Dixon, N. E.; Otting, G. ChemBioChem 2006, 7, 1599- 1604. (9) Su, X.-C.; McAndrew, K.; Huber, T.; Otting, G. J. Am. Chem. Soc. 2008, 130, 1681-1687. (10) Edelmann, F. T. Coord. Chem. Rev. 2006, 250, 2511-2564. (11) Prudencio, M.; Rohovec, J.; Peters, J. A.; Tocheva, E.; Boulanger, M. J.; Murphy, M. E. P.; Hupkes, H.-J.; Kosters, W.; Impagliazzo, A.; Ubbink, M. Chem. Eur. J. 2004, 10, 3252-3260. (12) Vlasie, M. D.; Comuzzi, C.; van den Nieuwendijk, A. M. C. H.; Prudencio, M.; Overhand, M.; Ubbink, M. Chem. Eur. J. 2007, 13, 1715-1723. (13) Keizers, P. H. J.; Desreux, J. F.; Overhand, M.; Ubbink, M. J. Am. Chem. Soc. 2007, 129, 9292-9293. (14) Mehrotra, R. C.; Kapoor, P. N.; Batwara, J. M. Coord. Chem. Rev. 1980, 31, 67- 91. (15) Grenthe, I. J. Am. Chem. Soc. 1961, 83, 360-364. (16) Grenthe, I. Acta Chem. Scand. 1963, 17, 2487-2498. (17) Tang, R. r.; Zhao, Q.; Yan, Z. e.; Luo, Y. m. Synth. Commun. 2006, 36, 2027- 2034. (18) Sunnerhagen, M.; Nilges, M.; Otting, G.; Carey, J. Nat. Struct. Mol. Biol. 1997, 4, 819-826.

(19) Schmitz, C.; Stanton-Cook, M. J.; Su, X.-C.; Otting, G.; Huber, T. J. Biomol. NMR 2008, 41, 179-189. (20) Zweckstetter, M.; Bax, A. J. Am. Chem. Soc. 2000, 122, 3791-3792. (21) Otting, G. Personal Communication. (22) Parker, D.; Dickins, R. S.; Puschmann, H.; Crossland, C.; Howard, J. A. K. Chem. Rev. 2002, 102, 1977-2010.

Chapter 3: Dipicolinic Acid as Paramagnetic Probes – 3MDPA

“The opposite of a correct statement is a false statement. But the opposite of a profound truth may well be another profound truth.”

Niels Bohr (1885 – 1962)

3.1 Introduction

In the previous chapter, it was shown that small functionalized molecules based on the dipicolinic acid motif can be used to probe the structure of a protein through the induction of pseudocontact shifts. This was only possible due to the participation of a carboxylate side chain which provides an additional attachment point to the ArgN- 4MMDPA-Ln3+ complex, increasing the overall rigidity. To further explore the possibility of other analogues of dipicolinic acid as paramagnetic probes, the 3- mercapto-2,6-pyridinedicarboxylic acid (3MDPA, 9) was proposed. The synthesis and the NMR properties of the 3MDPA tag in the presence of both paramagnetic transition metal and lanthanide ions will be discussed in detail.

3.2 Synthesis

The synthesis of the 3-mercapto-2,6-pyridinedicarboxylic acid (9) is summarized in (Scheme 3.1).

Scheme 3.1: Synthesis of 3-mercapto-2,6-pyridinedicarboxylic acid (9).

The synthesis involves the electrophilic bromination of commercially available 2,6-dimethylpyridine to yield 3-bromo-2,6-dimethylpyridine (6). The purity of the compound was confirmed through comparison with literature spectroscopic data.1 The methyl groups were oxidized by treatment with potassium permanganate under neutral conditions to yield the corresponding 3-bromo-2,6-pyridinedicarboxylic acid, followed by esterification to yield 3-bromo-2,6-dimethoxycarbonylpyridine (7). The purity of the compound was confirmed through comparison with literature spectroscopic data.1 The protected thiol group was introduced through stirring a solution of sodium tert-butyl thiolate with the ester (7) in tetrahydrofuran (THF) to yield the thiol ether (8). Deprotection was achieved by heating the protected thiol (8) in a solution of concentrated hydrochloric acid at reflux to yield 3MDPA (9). The structure and purity of the compound was determined using high resolution mass spectrometry and spectroscopic data.

3.3 Transition metals as paramagnetic shift reagents

The 3MDPA tag (9) was attached to the N-terminal domain of the arginine repressor from E. coli. (ArgN) to form the ArgN-3MDPA adduct. The resulting 1H-15N HSQC spectrum is shown below in Figure 3.1. Like the previous example shown in Chapter 2, upon the ligation of the 3MDPA tag, small shifts in the HSQC cross peaks were observed. In addition to the changes in the chemical shifts of the cross peaks, line broadening was also observed. These changes may be attributed to the changes to the charge density on the surface of the protein. Plotting the changes in chemical shift in 15N dimension, the largest deviation was observed around the residues closest to the ligation site.

After attaching the 3MDPA probe onto protein, the effect the addition of a paramagnetic transition metal ion on the 1H-15N HSQC spectrum, i.e. cobalt was examined. A 1H-15N HSQC spectrum of the ArgN-3MDPA adduct was measured in the presence of a solution of zinc and cobalt ions and shown below in Figure 3.2.

Figure 3.2: Overlay of the 1H-15N HSQC spectra of the purified ArgN-3MDPA adduct in the presence of a diamagnetic transition ion (Zn2+, black) and in the presence of a paramagnetic transition metal ion (Co2+, red).

The 1H-15N HSQC spectrum of the ArgN-3MDPA adduct in the presence of paramagnetic transition metal ions was then assigned and the assignments are shown in Appendix 1.2.1 Based on the diamagnetic and paramagnetic assignments, the PCS were calculated and plotted against the residue numbers (Figure 3.3). Maximum PCS was observed at residues 21 and 56. Examining the three dimensional structure of the ArgN protein, these residues exist in close proximity of the ligation site and would be closest to the location of the metal ion.

1.6

1.4

1.2

1 H) 1 0.8

PCS ( 0.6

0.4

0.2

0 0 10203040506070 Residue Number

Figure 3.3: Plot of the PCS in the 1H dimension vs. the amino acid residue number for the ArgN-3MDPA in the presence of diamagnetic (Zn2+) and paramagnetic (Co2+) transition metal ions.

Having assigned the PCS, the next step is to calculate position of the metal ion and the magnetic susceptibility tensor.

3.3.1 Metal Ion Location and Magnetic Susceptibility Tensor

Unlike the ArgN-4MMDPA example, where there were three sets of paramagnetic assignments, only one set of data was available for the ArgN-3MDPA-Co2+ complex. As a result, there are fewer constraints on the solution of the both the metal ion and the tensor parameters. Using a total of 50 PCS assignments as input and solving for all 23 conformers simultaneously without any additional boundary conditions, the location of the metal ion and the magnetic susceptibility tensor was calculated. The back calculated PCS was plotted against the observed values (Figure 3.4) and good correlation between the two sets of result was observed. The isosurface plot (Figure 3.5) shows that, like the lanthanide ion, the cobalt ion generates an asymmetric shift distribution. In the 1H-

15N HSQC spectrum, only positive PCSs were observed. This is attributed to the orientation of the magnetic susceptibility tensor as shown in Figure 3.5.

2.5

1.5

1 Predicted PCS 0.5

0 0 0.5 1 1.5 2 2.5

Experimental PCS

Figure 3.4: The plot of the back calculated PCS vs. the experimentally observed PCS for ArgN-3MDPA in the presence of Co2+ ions (referenced to Zn2+ ions). The predicted PCS was calculated by solving all three sets of results simultaneously using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.2

Figure 3.5: PCS Isosurface for ±2 and ±0.2 ppm (blue = positive PCS, red = negative PCS) superimposed on the three-dimensional structure of the ArgN for Co2+ visualized using PyMol. The PCS isosurface was calculated using all 23 structures simultaneously and with no additional boundary conditions using NUMBAT.2

3.3.2 Alignment Tensor

To assess the rigidity with which the 3MDPA-Co2+ complex binds to the protein, the alignment tensor needs to be considered. From the RDC assignments (Appendix 1.2.3), the orientation and the components of the alignment tensor were calculated. Due to the sensitivity of RDCs to molecular motion, only the rigid elements of the tertiary structures were used. Plotting the back-calculated RDC against the experimentally observed values significant deviations were observed, as indicated by the scattering of the points in Figure 3.6. This deviation is caused at least in part by the errors in the orientation of the NH bond in the NMR structure of the ArgN protein.

8 6 4 2 0 -2 -4

Back calculated RDC (Hz) -6 -8 -8 -6 -4 -2 0 2 4 6 8 Experimental RDC (Hz)

Having compared the back-calculated and experimentally observed residual dipolar couplings, the next step is to consider the orientation of both the magnetic susceptibility tensor and the alignment tensor. Unlike the work in the previous chapter, where there were three sets of PCS data measured for the different paramagnetic ions, only one set of data was available. Additional PCS data introduces constraints allowing the refinement of the solution without the need of additional boundary conditions. This was

attempted with each of the individual conformations for the ArgN protein reported by Sunnerhagen et al.7 and significant fluctuations were found in the solutions. In order to increase the robustness of the solution, additional boundary conditions were required. This was done by using the location of the metal ion as determined in section 3.3.1 as the initial starting value for the location of the metal ion in NUMBAT.2 The alignment tensor for each of the 23 conformations was then calculated using the rigid elements of the tertiary structure (i.e. residues 14 – 16, 28 – 34 and 45 – 51) using PALES3 and plotted in a Sanson-Flamsteed equivalence plot shown in Figure 3.7b. The magnitudes of the axial and rhombic components of the magnetic susceptibility and alignment tensors are shown in Table 3.1.

Table 3.1: The axial and rhombic components of the alignment tensor and the magnetic susceptibility tensors for the ArgN-3MDPA adduct.a

PALESb NUMBAT

b,c b,c 0ax 0rh 0ax 0rh

-4.4 ± 0.1 - 1.5 ± 0.7 - 4.9 ± 0.9 - 0.5 ± 0.4 (a) !$!0-32 m3 unless otherwise stated. (b) The axial and rhombic components of the alignment tensor calculated by PALES requires a correction factor of 2 to be consistent with values calculated with other software such as MODULES. (c) Alignment tensors were converted to the corresponding magnetic susceptibility tensor values using eqn (2.1).

(a) Magnetic Susceptibility Tensor

(b) Alignment Tensor

Comparing the two Sanson-Flamsteed equivalence plots, deviations between the alignment and the magnetic susceptibility tensor can be observed. One of the explanations for the misalignment is the motion of the probe relative to the protein. Further evidence for this was the magnitude of the tensors measured shown here was

smaller compared to that of the ArgN-4MMDPA-Co2+ (results not shown). Examining the structure of the protein (Figure 3.8), shows that the closest basic amino acid residues (Arg59 and Lys62) would not be able to bind to the cobalt ion without significant conformational rearrangement. Without this second side chain providing an additional binding site, a loss of rigidity would be expected.

Figure 3.8: A ribbon representation of the three-dimensional structure of the ArgN protein with the location of the metal ion (purple sphere) and the Cys68, Arg59 and Lys62 residues superimposed.

Another possible explanation for the deviation may be due to the inaccuracies of using only one set of PCS data in the calculation of the magnetic susceptibility tensor. In order to investigate this, the measurements were repeated with multiple paramagnetic lanthanide ions.

3.4 Paramagnetic lanthanide metals as paramagnetic shift reagents

To further examine the behavior of the 3MDPA probe, it was ligated onto both the ArgN protein and the C54T/C97A/Q69C triple mutant of T4 Lysozyme and its NMR properties examined in the presence of paramagnetic lanthanide ions.

This work described below was conducted by Dr Xun-Cheng Su at the Research School of Chemistry at the Australian National University.

3.4.1 NMR Spectroscopy of ArgN-3MDPA

The 1H-15N-HSQC spectrum of the ArgN-3MDPA showed broad lines for amino acid residues in the vicinity of Cys68, indicating conformational exchange on the millisecond time scale. Titration of the ArgN-3MDPA adduct with YCl3 to form a 1:1 complex resulted in a single set of peaks, suggesting that the metal ion locked the adduct in a complex as cross peaks of ArgN-3MDPA vanished, indicating that the exchange of the metal ion is slow on the NMR time scale (milliseconds).

Some of the HSQC cross peaks of the ArgN-3MDPA-Y3+ complex were significantly shifted compared to those of unmodified ArgN. The resonance assignments were re- established by a 3D NOESY-15N-HSQC spectrum which is not shown, and no significant changes in structure were observed. Structural conservation was also suggested by the observation that chemical shift changes were confined to the vicinity of Cys68.

3.4.1.1 PCS measurement of ArgN-3MDPA

Paramagnetic complexes (1 : 1) of ArgN-3MDPA with different lanthanides resulted in the formation of significant PCS. Measurement were performed with Ce3+, Tb3+, Dy3+, Ho3+, Er3+, Tm3+, Yb3+ and selected results are shown below in Figure 3.9.

Figure 3.9: 1H-15N-HSQC spectra of ArgN-3MDPA in complex with different metal ions. The spectra were recorded of 0.15 mM solutions of ArgN-3MDPA in a buffer of ,20 mM MES, pH 6.5, using a Bruker 800 MHz NMR spectrometer at 25oC. Selected pairs of cross-peaks from the diamagnetic and paramagnetic molecules are connected by lines and labelled with their assignment. (a) Superimposition of 15N-HSQC spectra recorded with Y3+ (black) and a mixture of Tb3+/Y3+ (grey). (b) Same as (a), except that the grey spectrum was recorded with a mixture of Tm3+/Y3+.

The lanthanide binding affinity of the ArgN-3MDPA is weaker than that of the ArgN-4MMDPA, as the addition of DPA quantitatively regenerated the apo-protein. Under the same conditions, ArgN-4MMDPA released only a fraction of lanthanides. Despite the weak binding affinity, attempts to generate exchange cross peaks in samples

containing mixtures of paramagnetic and diamagnetic metal ions4 failed, indicating that the exchange rates are in the order of seconds. The slow metal exchange rate was beneficial in allowing the simultaneous observation of paramagnetic and diamagnetic cross peaks in the samples prepared with a mixture of paramagnetic and diamagnetic metal ions. The PCSs could thus be measured accurately from a single spectrum, where both species experience the same conditions (i.e. temperature, pH and etc.).

The assignment of the paramagnetic peaks was assisted by the fact that the 1H and 15N spin of an amide group exhibits similar PCSs because their coordinates do not vary significantly compared to the distance from the paramagnetic centre. Therefore, the paramagnetic cross peaks are displaced along approximately parallel lines from their diamagnetic partner.

The resonance assignments were supported further by the use of different paramagnetic metal ions. For example, Tb3+ and Tm3+ not only display cross peaks in opposite directions (Figure 3.9a and b) due to their opposite sign of tensor anisotropies,5 but the slope of displacement are very similar for any given diamagnetic cross peaks, as the geometry of the NH group with respect to the paramagnetic centre remains unchanged if the metal position is conserved. The slope of the cross peak displacement for different metals is expected to be the same even in the presence of residual anisotropic chemical shifts (RACS),6 !$!0$!$$ similarly orientated and of similar magnitude.

3.4.1.2

The PCS measurements were used together with the NMR structure of ArgN7 to !$!0!!$$ metals in multiple rounds of fitting tensors parameters and assigning additional paramagnetic cross peaks, assuming that all metal ions occupy a common position with respect to the protein. The results are summarized in Table 3.2 and Figure 3.10.

Figure 3.10: tensors of eight different metal ions bound to the ArgN-3DPA adduct. The x-, y- and z-axes of the tensors are distinguished by red, cyan and blue lines, respectively. The axes definitions follow the unique tensor representation (UTR) convention, where the x- and z-axes are the shortest and longest axes of the tensor.2 The side chains of Glu12 and Cys68 and the N- and C-termini of the protein are labelled. The figure was generated using Molmol.8

The metal ion (Figure 3.10) is positioned about 3.2 Å from the carboxyl group of Glu21 and the z axes of the tensors of all metal ions have similar orientations, suggesting that the coordination geometry is similar for all metals. The orientations of the x and y axes are less well conserved which may be, in part, attributed to the greater uncertainty associated with the rhombic component of the tensor. The magnitude of the axial components of the tensors decreased as follows: Dy3+ > Ho3+ > Tb3+ > Tm3+ > Co2+ > Yb3+ > Er3+ > Ce3+ (Table 3.2).

Table 3.2: 0$$$!-3MDPAa

Tensor Metal ion ax rh Coordinates of tensor axes /10-32 m3 /10-32 m3 axis Ce3+ -0.3 0.0 x -0.189 0.982 -0.007 y 0.820 0.162 0.549 z 0.541 0.098 -0.836 Tb3+ -8.1 -2.5 x 0.237 -0.911 0.338 y 0.558 0.413 0.720 z -0.795 0.018 0.606 Dy3+ -17.0 -1.4 x 0.099 -0.978 0.182 y 0.660 0.202 0.723 z -0.744 0.049 0.666 Ho3+ -9.8 -1.0 x -0.025 0.999 0.024 y 0.692 0.000 0.722 z 0.722 0.035 -0.691 Er3+ 2.9 1.2 x -0.028 -0.524 0.851 y 0.401 0.774 0.490 z -0.916 0.355 0.189 Tm3+ 6.6 2.4 x 0.229 -0.834 0.503 y 0.530 0.540 0.654 z -0.817 0.117 0.565 Yb3+ 3.4 0.8 x 0.089 -0.973 0.211 y 0.685 0.214 0.697 z -0.723 0.083 0.686 Co2+ -4.6 -0.6 x 0.335 -0.699 0.632 y 0.681 0.643 0.350 z -0.651 0.313 0.692 [a] At 25 oC and pH 6.5 in 20 mM MES buffer. The tensors are listed in their unique tensor representation (UTR)2 as obtained by fitting of the PCSs to the first conformer of the PDB coordinates 2AOY of ArgN.7 The orientations of the tensor axes are given with respect to the origin (0, 0, 0). The common metal position obtained by the fit is (12.131, 8.671, 2.790).

Most PCS observed with lanthanides were of opposite sign from those observed previously with corresponding complexes of ArgN-4MMDPA.9 As 4MMDPA was attached to Cys68 in complete analogy to the ArgN-3MDPA adduct, this result indicates

that the 3MDPA probe is suitable for generating new tensor orientations for the same protein.

$!\V$!$!0! ion location is that the accuracy can be improved if the PCS can be measured for nuclear spins that populate the largest angular spaces around the metal ion. Secondly, the metal ion is more readily immobilized with respect to the protein if the tether between the protein and the probe is small. It wa $ ! ! 0ax values of the ArgN-3MDPA complexes with the Tb3+, Tm3+ and Yb3+ are approximately 40% smaller than those observed for the corresponding complexes with the ArgN-4MMDPA.9 This makes the 3MDPA probe less attractive for the measurement of residual dipolar coupling but nevertheless the PCS are sufficiently large to be used for proteins structure refinement and studies of protein ligand complexes.

!!0ax values observed in the ArgN-3MDPA yielded a different sequence than previous established values of lanthanide complexes of 10 3+ 3+ calbindin D9k. !0 tensor of Ho was larger than that of Tb and the 3+ 3+ 0 tensor of Er was smaller than that of Yb \ "" ! ! ! 0 tensors anisotropies of the lanthanides critically depend on to their chemical environment and are difficult to predict quantitively. Another example is the \V$$0$3+ than for Dy3+ in the lanthanide binding tag CLaNP-511!!!!0 of Dy3+ is only rivalled by Tb3+.

3.4.2 NMR Spectroscopy of T4-lysozyme

Using the same protocol as for ArgN, the 3MDPA probe was attached to the C54T/C97A/Q69C triple mutant of uniformly 15N labeled T4 lysozyme. The HSQC cross peaks of the T4-lysozyme-3MDPA-Lu3+ complex were assigned using a 3D NOESY-15N-HSQC spectrum and PCSs were measured for complexes with Tm3+ and

Co2+. The magnitude of PCSs observed was similar for both metal ions. Comparison with the corresponding Tm3+ complex of the 4MMDPA adduct showed that, as in the case of ArgN, the PCSs tended to be of opposite sign between both constructs (Figure 3.11). In contrast to ArgN, no line broadening was observed in the 15N-HSQC spectrum of the T4-lysozyme-SHDPA adduct in the absence of metal ions.

3.4.3 Comparison of ArgN-3MDPA and T4-lysozyme-3MDPA

Based on the NMR data on the T4-lysozyme-3MDPA, a comparison can be made with the ArgN-3MDPA adduct, allowing the effects of chemical environment to be investigated. Based on the data for both adducts bearing the 3MDPA probe (9), it was found that there were significant deviations between the magnitudes of the 0"

! $ 0 \ $ ! ! $$ chemical environments may vary for a number of reasons: (a) the weak association of the lanthanides with the carboxyl group may be a contributing factor, by changing the ligand field and allowing more motions of the metal ion with respect to the protein; (b) the sulfur in 3MDPA is in close contact with one of the carboxyl groups. The crystal structure of 3-mercapto-2-pyridinecarboxylic acid, an analogue of the 3MDPA probe shows, however, that the carboxyl group can remain co-planar with the pyridine ring.12 The close proximity of the sulfur may nonetheless subtly affect the ligand field of metal coordination which could explain the observation that the NMR spectra of Co2+ complexes of the ArgN-4MMDPA complex and ArgN-T4-lysozymes displayed only small PCSs (data not shown), where large PCS were observed with the 3MDPA derivatives and (c) even with PCS from many different metal ions, it is difficult to pinpoint the position of the metal ion with high accuracy. 15N-HSQC cross peaks of amides in the vicinity of the metal ion are invariably broadened beyond detection, cross peaks with very large PCSs are very difficult to assign, and a metal position at a greater $ ! \ \ 0 tensor.

3.4.4 Residual Dipolar Coupling

Due to !0ax components calculated in 3.4.1.2, one of the consequences is the difficulty in measuring the residual dipolar couplings required to determine the "$!!0

could not be conducted to examine the overall rigidity of the 3MDPA probe for lanthanide ions or paramagnetic transition metal complexes.

3.5 Conclusion

Based on the successes of the 4MMPDA probe (5), an alternative analogue was synthesized, that is the 3MDPA (9). One of the advantages that the 3MDPA probe (9) has to offer over the 4MMDPA probe (5) is the reduced length of the linker. As a result of the reduced linker length, there are few rotating bonds which ultimately reduce the movement of the probe relative to the protein. This has important implications in its use in protein structure refinement or the study of protein-ligand complexes. The small size of the probe ensures minimum interference of the biological function of the protein. Two proteins were ligated with the 3MDPA probe, the N-terminal domain of the arginine repressor from E. Coli (ArgN) and the T4-lysoyme. Both the ArgN-3MDPA and T4-lysoyme-3MDPA adduct were studied in the presence of paramagnetic transition metal and lanthanide ions using NMR spectroscopy.

The 1H-15N HSQC spectrum of the ArgN-3MDPA adduct was measured in the presence of paramagnetic cobalt ion where large PCS were observed. The location of the metal ion and the tensor were calculated and found that the metal was positioned away from basic amino acid side chains. Due to the high affinity of cobalt ion to nitrogen donors, it is expected that any basic amino acid side chains located in close proximity would also bind to the cobalt ion which could increase the rigidity. This was observed with the ArgN-4MMDPA adduct, where the carboxylate group in the glutamate residue E21 is presumably bound to the lanthanide ion to form a rigid system. To determine rigidity of the probe, the alignment tensor determined from the residual dipolar coupling was compared to the tensor and deviations were found. This deviation can be accounted for by either the lack of side chain binding which reduces the overall rigidity of the probe or insufficient data which does not allow the alignment tensor to be calculated with sufficient accuracy to be compared to with the tensor.

In order to explore this further, the 1H-15N HSQC spectrum of the ArgN-3MDPA adduct !$!"!0 were determined for 8 metal ion centers and found that the tensors were 40% smaller than the values determined for the 4MMDPA ligand (5) using the same metal ions. As a result of the reduced tensors, the measurement of the residual dipolar coupling was not attempted and thus the overall rigidity of the ArgN-3MDPA probe will need to be determined as part of future work.

These results also highlighted the dependence of the tensor on the local chemical environment as the sequence of lanthanide anisotropy determined in 3.4.1.2 for the 3MDPA probe is significantly different to the sequences published in literature.10-11 The difference was attributed to a number of factors, but the most important of them was the difference in the side chains that are interacting with the metal ion in conjunction with the immobilized probe.

Overall, despite the inability to determine the rigidity of the 3MDPA probe, it was found that in the presence of paramagnetic lanthanide ions and transition metal ions, a large PCS was observed showing its potential of its use in protein structure refinement or the study of protein-ligand complexes.

3.6 References

(1) Zimmermann, N.; Meggers, E.; Schultz, P. G. Bioorg. Chem. 2004, 32, 13-25. (2) Schmitz, C.; Stanton-Cook, M. J.; Su, X.-C.; Otting, G.; Huber, T. J. Biomol. NMR 2008, 41, 179-189. (3) Zweckstetter, M.; Bax, A. J. Am. Chem. Soc. 2000, 122, 3791-3792. (4) John, M.; Headlam, M.; Dixon, N.; Otting, G. J. Biomol. NMR 2007, 37, 43-51. (5) Pintacuda, G.; John, M.; Su, X. C.; Otting, G. Acc. Chem. Res. 2007, 40, 206- 212. (6) John, M.; Park, A. Y.; Pintacuda, G.; Dixon, N. E.; Otting, G. J. Am. Chem. Soc. 2005, 127, 17190-17191. (7) Sunnerhagen, M.; Nilges, M.; Otting, G.; Carey, J. Nat. Struct. Mol. Biol. 1997, 4, 819-826. (8) Koradi, R.; Billeter, M.; Wuthrich, K. J Mol Graph 1996, 14, 51-55, 29-32. (9) Su, X.-C.; Man, B.; Beeren, S.; Liang, H.; Simonsen, S.; Schmitz, C.; Huber, T.; Messerle, B. A.; Otting, G. J. Am. Chem. Soc. 2008, 130, 10486-10487. (10) Bertini, I.; Janik, M. B. L.; Lee, Y.-M.; Luchinat, C.; Rosato, A. J. Am. Chem. Soc. 2001, 123, 4181-4188. (11) Keizers, P. H. J.; Saragliadis, A.; Hiruma, Y.; Overhand, M.; Ubbink, M. J. Am. Chem. Soc. 2008, 130, 14802-14812. (12) Dereppe, J. M.; Schanck, A.; Declercq, J. P.; Germain, G.; Van Meerssche, M. Bull. Soc. Chim. Belg. 1976, 85, 729-733.

Chapter 4: Unnatural Amino Acids as Paramagnetic Probes

“Certainly nothing is unnatural that is not physically impossible”

Richard Brinsley Sheridan (1751 – 1816)

4.1 Introduction

The use of paramagnetism to provide additional information for protein structure refinement and the study of protein-ligand complexes is well documented.1-3 Previous chapters so far have shown the success of using small organic molecules as a means to introduce a paramagnetic centre in the form of a lanthanide ion in a rigid site specific manner.

Despite the success of small organic molecules as paramagnetic probes, their attachment requires the use of post translational modification, typically through the use of a reactive amino acid side chain such as cysteine. This typically involves the mutation of the protein to remove additional reactive amino acid side chains (i.e. multiple cysteine residues) to avoid non-specific attachment of the paramagnetic probe. In addition to any required protein modification, there are also lengthy chemical reactions required to covalently attach the paramagnetic probe onto the protein.

By incorporating a small organic molecule into the protein during the expression process, the lengthy post translational modifications can be avoided. With recent advances in protein expression techniques, there are now more tools available to introduce these probes in a site specific manner directly into biomolecules. This is done through the use of unnatural amino acids.

4.2 Incorporation of unnatural amino acids

Unlike the previous examples with the 4MMDPA (5) and the 3MDPA (9), the use of an unnatural amino acid allows the incorporation of the paramagnetic probe into the protein without the use of any post translational modifications. There are currently two methods in which unnatural amino acids can be inserted and they include: (i) residue

specific labeling; and (ii) site specific labeling. Both methods can achieve the same objective, but vary in specificity and complexity and will be discussed in detail below.

4.2.1 Residual specific labeling using unnatural amino acids

The incorporation of a probe during the expression of the protein removes the need for post-translational modifications. This can be achieved through the replacement of a naturally occurring amino acid, with one that is modified with the desired probe, in this instance, a lanthanide binding amino acid. This can be done through the replacement of one of the naturally occurring amino acids in the growth medium with an amino acid that bears the desired probe. It is important to note that the modified amino acid must be structurally similar to its natural analogue so that it can be recognized by the unmodified protein expression system.4-6 An example of this technique is the replacement of methionine with a selenium analogue.7 Other examples include the replacement of phenylalanine (Figure 4.1) with phenylalanine derivatives containing halides, azides and acetyl groups at the para position.8-9

Figure 4.1: Phenylalanine analogues incorporated using a residue-specific method.

The limitation of this method is the need to use an amino acid that is structurally similar to its natural counterparts. In addition, this method will modify the protein on a global scale that is if a modified phenyalanine such as the ones shown in Figure 4.1 is used, then all phenylalanines in the protein will be modified. To not only expand the versatility but also to allow modifications in a site-specific manner, a modified protein expression system has been investigated.

4.2.2 Site specific labeling using unnatural amino acids

One approach to site specifically incorporate an unnatural amino acid bearing metal ion binding motifs is through the use of a modified protein expression system. This methodology has been described more than 20 years ago,10 but since has been expanded by Schultz et al.11-12 This technique has been used to incorporate a number of unnatural amino acids for applications such as infrared13 and fluorescence spectroscopy14 to study protein structures.

Proteins are synthesized through the interaction of a series of complex chemical processes including translation, recognition and bond formation (Scheme 4.1). These are facilitated by a range of biomolecules designed to ensure the accuracy and high rate of protein synthesis. There are three components involved in protein synthesis which are important for the incorporation of unnatural amino acids. They include the messenger RNA (mRNA), the transfer RNA (tRNA) and the aminoacyl tRNA synthetase enzyme all of which serve an important role in the activation, selection and incorporation of amino acids during protein synthesis. The messenger RNA is composed of a sequence of triplet codons, each a combination of three nucleotides (i.e. U, A, C and G) which are used to determine the amino acid to be incorporated in the protein being expressed (Scheme 4.1a). These codons are also used to signal other processes used during the expression of the protein such as termination and initiation of the peptide chain (Scheme 4.1a and d). For example the triplet codon AUG is used to initiate the start of the peptide chain and the codon GUU is used to signal the incorporation of the amino acid valine. The tRNA is a short peptide chain which acts as an adaptor that links the amino acid and the mRNA sequence (Scheme 4.1a). This is achieved through the use of nucleotides which are complementary to the triplet codons present in the mRNA sequence; these are known as the anti-codons. It is this interaction between the complementary nucleotides that allows the mRNA to selectively incorporate the appropriate amino acid to the correct location within the protein. The aminoacyl synthetase is an enzyme whose role is to attach the amino acid to the tRNA to form the corresponding aminoacyl tRNA which activates the amino acid to the peptide coupling reactions used to elongate the peptide chain (Scheme 4.1b).

(a)

(b)

(c)

(d)

In order to incorporate an unnatural amino acid, the mRNA sequence, tRNA and aminoacyl synthetase all needs to be modified. In the mRNA sequence, a unique codon needs to be inserted that does not recognize any native amino acids or serves any purposes during the expression process. In the methods developed by Wang et al.12 the codon UAG is commonly used as it is one of the available codons that are least utilized

during the expression process. To site specifically incorporate an unnatural amino acid, the UAG codon is inserted into the appropriate position in the mRNA sequence. This allows the mRNA to recognize the aminoacyl tRNA containing the unnatural amino acid through the corresponding anti-codon in the modified tRNA molecule. Due to the structural dissimilarities of the unnatural amino acids compared to their native counterparts, they are not recognized by native tRNA and aminoacyl synthetase enzymes. As a result, native tRNA and aminoacyl synthetase needs to be modified to allow the unnatural amino acid to be recognized and incorporated during the synthesis of the modified protein. The development of the modified tRNA and aminoacyl synthetase enzyme is beyond the scope of this work, and details can be found in numerous reviews by Wang et al.11-12,16

4.3 Unnatural amino acids as paramagnetic probes

As discussed in the literature review, an ideal paramagnetic probe needs to possess high metal binding affinity and once bound to the protein, should not form diastereoisomers. The properties described above are required to ensure large measurable pseudocontact shifts are observed, and multiple cross peaks do not appear in the 1H-15N HSQC spectrum. Previous chapters have described the synthesis and application of small heterocyclic molecules which have met the criteria of having high metal binding affinity and do not produce diastereoisomers once bound to the protein. A logical extension of results discussed previously would be to incorporate the same lanthanide binding motifs into an unnatural amino acid (11) so they can be incorporated using the modified protein expression system developed by Wang et al.11-12,16

In addition to the dipicolinic acid based amino acid (11), two additional amino acids bearing a mixed nitrogen and oxygen donor were also proposed. They include one based on the iminodiacetic acid motif (IDA-AA, 15) and one based on the 8-hydroxyquinoline motif (HQ-AA, 19). The iminodiacetic acid motif was chosen based on the formation of stable lanthanide complexes which has been used in number applications such as MRI contrast agents, protein sensing and ribonucleotide hydrolysis catalysts.17-19 Similarly 73

the hydroxyquinoline derivative was chosen for the ability to form stable lanthanide and transitional metal complexes and has been successfully utilized in a number of applications including luminescence probes.20-23

4.3.1 Generalized synthesis of unnatural amino acids

There are a large number of methods for synthesizing unnatural amino acids and the most common methods are shown in Scheme 4.2. They include: (a) nucleophilic substitution of -halocarboxylic acids;24 (b) the Strecker synthesis;25-26 (c) alkylation of an acetamidomalonate ester;27 and (d) enzyme catalyzed ring opening of substituted hydantoin based compounds.28 Of the four general methods described above, the alkylation of acetamidomalonate ester (Scheme 4.2c) was considered to be the best approach for the synthesis of the proposed unnatural amino acids due to the availability of the starting materials. In addition, compared to the other methods, the substituted alkyl halides required in the first step of the synthesis are easier to access synthetically compared to their corresponding aldehydes or halocarboxylic acid.

4.3.2 Synthesis of the dipicolinic acid modified amino acid, DPA-AA (11)

The synthesis of DPA-AA (11), starts with the addition of 4-bromomethyl-2,6-dimethoxycarbonlylpyridine (4) to the diethyl acetamidomalonate ester and the full synthetic scheme is shown below in Scheme 4.3.

Scheme 4.3: The synthesis of the unnatural amino acid bearing the dipicolinic acid motif (DPA-AA, 11)

The synthesis first involves the attachment of the acetamidomalonate ester motif to the dipicolinic acid. This is done by the treatment of diethylacetamidomalonate ester with sodium hydride in tetrahydrofuran, followed by the addition of the bromide (4) to yield the protected amino acid (10). The purity and structure of compound (10) was determined through the use of spectroscopic data and high resolution mass spectrometry. The ester protecting groups were removed by heating the malonate ester (10) in concentrated hydrochloric acid under reflux to yield the DPA-AA (11) which was isolated as the hydrochloride salt in 67% yield. The 1H NMR spectrum (Figure 4.2) shows the loss of the ethyl groups and the appearance of the characteristic doublet of doublets for the benzyl protons in the DPA-AA (11). The purity and structure of the compound was determined through the use of spectroscopic data.

H H H '

(11)

Figure 4.2: 1H NMR spectrum of 2-amino-3-(2,6-pyridinedicarbonyl)propanoic acid

(DPA-AA, 11) in D2O.

4.3.3 Synthesis of iminodiacetic acid modified amino acid, IDA-AA (15)

One approach to the synthesis of the IDA-AA (15) would be to alkylate p-aminobenzyl bromide directly. However, this approach could lead to potential difficulties with producing side products such as polymers. To avoid this, the proposed synthetic route starts with p-nitrobenzyl bromide, which is a protected amine group. The final synthetic scheme used here for the synthesis of IDA-AA (15) is shown below in Scheme 4.4.

Starting with commercially available p-nitrobenzyl bromide, the malonate group was introduced by treating diethyl acetamidomalonate with sodium ethoxide, followed by the addition of p-nitrobenzyl bromide. The purity and the structure of the nitro compound 12 were confirmed through comparison with literature spectroscopic data. The nitro compound 12 was then reduced to the corresponding amine 13 by stirring over a hydrogen atmosphere in the presence of palladium on carbon as catalyst at 70oC. The amine was then isolated as the hydrochloride salt by the addition of thionyl chloride.

The purity of the amine compound 13 was confirmed through comparison with literature physical data.29 The alkylation of the aromatic primary amine to yield the protected amino acid 14 was achieved through the use of a modified literature method.30 The alkylation of the aromatic primary amine was confirmed through the use of spectroscopic evidence (Figure 4.3), such as the appearance of a singlet assigned to the a H proton on the ethyl acetate group. Through the use of HMBC correlation experiments, the resonances attributed to ethyl groups on the acetate esters (Ha1 and Ha2) were also assigned. The integration of the Ha proton relative to the aromatic protons (i.e. H2), indicated that two acetate groups has been attached to the aromatic primary amine. Resulting elemental analysis of a purified sample confirmed the attachment of two ethyl acetate groups onto the primary aromatic amine.

Scheme 4.4: The synthesis of the unnatural amino acid bearing the iminodiacetic acid motif (IDA-AA, 15)

The ester and acetyl protecting groups were then removed by heating the protected amino acid 14 in concentrated hydrochloric acid under reflux to yield the product IDA- AA (15). The structure of the IDA-AA (15) compound was confirmed spectroscopically by the loss of the ester protecting groups in the 1H NMR spectrum (Figure 4.4) and the appearance of the characteristic doublet of doublets for H protons. To complement the spectroscopic data, an elemental analysis was also conducted and the results supported the proposed structure.

(14)

m2 a1 H H Ha2

a m1 H H COCH3 +

H H3 H2 + NH

1 Figure 4.3: H NMR spectrum of the ester 14 in CDCl3.

H (15)

H H '

H2 H3

Figure 4.4: 1H NMR spectrum of the modified amino acid IDA-AA 15 in

D2O / NaOD.

4.3.4 Synthesis of 8-hydroxyquinoline modified amino acid, HQ-AA (19)

The synthesis of the HQ-AA (19) makes use of commercially available 8-hydroxyquinoline, and the synthetic scheme followed is shown in Scheme 4.5.

The synthesis starts with the introduction of a chloromethyl group by treating commercially available 8-hydroxyquinoline with aqueous formaldehyde solution in the presence of hydrogen chloride to yield 5-chloromethyl-8-hydroxyquinoline (16) as the hydrochloride salt. The hydroxyl group is then protected via treatment with acetyl chloride in the presence of pyridine as a base in diethyl ether to yield the acetate ester (17). The acetamidomalonate moiety is then introduced by treating diethyl acetamidomalonate with sodium hydride followed by the addition of the chloride 17 to yield the protected amino acid 18. The amino acid HQ-AA (19) is produced by heating

a solution of the protected amino acid 18 in concentrated hydrochloric acid under reflux and isolated as the hydrochloride salt. The structure of the compound was confirmed by the loss of the ester protecting groups in the 1H NMR spectrum (Figure 4.5) and the appearance of the characteristic doublet of doublet for the proton. The purity was confirmed through the use of elemental analysis.

Scheme 4.5: The synthesis of the unnatural amino acid bearing the hydroxyquinoline motif (HQ-AA, 19)

3 H H6 H7 H 1 H H2

H H ' (19)

1 Figure 4.5: H NMR spectrum of the modified amino acid HQ-AA 19 in D2O.

4.3.5 Attempted Incorporation of DPA-AA, IDA-AA and HQ-AA

Having completed the synthesis of the DPA-AA (11), IDA-AA (15) and HQ-AA (19), the next step involves the evolution experiments designed to incorporate these unnatural amino acids into the protein. These experiments were attempted by Dr Kiyoshi Ozawa in collaboration with Schultz laboratory. Unfortunately, these unnatural amino acids proved to be toxic to the E. coli expression system, even in the presence of transition metal ions (Fe2+, Zn2+) that were thought to alleviate the toxicity by saturating the metal binding sites.

4.4 Conclusion

Despite the success enjoyed by the small organic molecules as paramagnetic probes, the required protein and post-translation modifications are time consuming. To overcome these two hurdles, direct incorporation of the probe into the protein would significantly simplify the process. This can be achieved through the use of modified protein synthesis which allows the incorporation of the paramagnetic probe in the form of a functionalized amino acid into the protein.

Amino acids bearing three different lanthanide binding motif were proposed and include the dipicolinic acid (11), the iminodiacetic acid (15) and hydroxyquinoline (19) modified amino acid. All of these amino acids are designed to contain lanthanide binding groups which are known to have a high lanthanide binding affinity, as this is an important requirement for an effective lanthanide probe. From the available synthetic protocols available for the synthesis of unnatural amino acids, the acetamidomalonate esters were used to form the lanthanide binding amino acid. The acetamidomalonate esters were chosen over the other synthetic methods due to the commercial availability of starting materials and ease of which intermediates can be synthesized. All of the amino acids were synthesized in good yields, and characterized using a combination of both spectroscopic and physical analysis.

Following on from the synthesis, attempts were made to incorporate these unnatural amino acids into the N-terminal domain of the arginine repressor from E. coli. Due to the toxicity of the unnatural amino acids to the modified protein expression system, the lanthanide binding amino acids could not be incorporated.

4.5 References

(1) Pintacuda, G.; John, M.; Su, X. C.; Otting, G. Acc. Chem. Res. 2007, 40, 206-212. (2) Pintacuda, G.; Park, A. Y.; Keniry, M. A.; Dixon, N. E.; Otting, G. J. Am. Chem. Soc. 2006, 128, 3696-3702. (3) Bertini, I.; Kursula, P.; Luchinat, C.; Parigi, G.; Vahokoski, J.; Wilmanns, M.; Yuan, J. J. Am. Chem. Soc. 2009, 131, 5134-5144. (4) Link, A. J.; Mock, M. L.; Tirrell, D. A. Curr. Opin. Chem. Biol. 2003, 14, 603- 609. (5) Hendrickson, T. L.; Crécy-Lagard, V. d.; Schimmel, P. Annu. Rev. Biochem. 2004, 73, 147 - 176. (6) Budisa, N. Angew. Chem. 2004, 116, 6586-6624. (7) Cowie, D. B.; Cohen, G. N. Biochim. Biophys. Acta. 1957, 26, 252-261. (8) Datta, D.; Wang, P.; Carrico, I. S.; Mayo, S. L.; Tirrell, D. A. J. Am. Chem. Soc. 2002, 124, 5652-5653. (9) Kirshenbaum, K.; Carrico, I. S.; Tirrell, D. A. ChemBioChem 2002, 3, 235-237. (10) Noren, C. J.; Anthony-Cahill, S. J.; Griffith, M. C.; Schultz, P. G. Science 1989, 244, 182-188. (11) Wang, L.; Schultz, P. G. Angew. Chem. Int. Ed. Engl. 2005, 44, 34-66. (12) Wang, L.; Xie, J.; Schultz, P. G. Annu. Rev. Biophys. Biomol. Struct. 2006, 35, 225- 249. (13) Schultz, K. C.; Supekova, L.; Ryu, Y.; Xie, J.; Perera, R.; Schultz, P. G. J. Am. Chem. Soc. 2006, 128, 13984-13985. (14) Tsao, M.-L.; Summerer, D.; Ryu, Y.; Schultz, P. G. J. Am. Chem. Soc. 2006, 128, 4572-4573. (15) Oregon State University. BB331 Introduction to Molecular Biology http://oregonstate.edu/instruction/bb331/lecture12/Fig5-20.html (accessed 30/11/2009). (16) Xie, J.; Schultz, P. G. Curr. Opin. Chem. Biol. 2005, 9, 548-554. (17) Tsukube, H.; Yano, K.; Ishida, A.; Shinoda, S. Chem. Lett. 2007, 36, 554-555. (18) Mayoral, E. P.; Garcia-Amo, M.; Lopez, P.; Soriano, E.; Cerdan, S.; Ballesteros, P. Bioorg. Med. Chem. 2003, 11, 5555-5567.

(19) Yashiro, M.; Ishikubo, A.; Takarada, T.; Komiyama, M. Chem. Lett. 1995, 665- 666. (20) Shavaleev, N. M.; Scopelliti, R.; Gumy, F.; Bunzli, J.-C. G. Inorg. Chem. 2009, 48, 7937-7946. (21) Rizzo, F.; Meinardi, F.; Tubino, R.; Pagliarin, R.; Dellepiane, G.; Papagni, A. Synth. Met. 2009, 159, 356-360. (22) Sun, L.-N.; Zhang, Y.; Yu, J.-B.; Yu, S.-Y.; Dang, S.; Peng, C.-Y.; Zhang, H.-J. Microporous Mesoporous Mater. 2008, 115, 535-540. (23) Comby, S.; Imbert, D.; Vandevyver, C.; Buenzli, J.-C. G. Chem. Eur. J. 2007, 13, 936-944. (24) Effenberger, F.; Drauz, K.; Foerster, S.; Mueller, W. Chem. Ber. 1981, 114, 173- 189. (25) Zuend, S. J.; Coughlin, M. P.; Lalonde, M. P.; Jacobsen, E. N. Nature (London, U. K.) 2009, 461, 968-970. (26) Ohfune, Y.; Sakaguchi, K.; Shinada, T. ACS Symp. Ser. 2009, 1009, 57-71. (27) Harper, J. B., Australian National University, 2002, PhD Thesis. (28) Altenbuchner, J.; Siemann-Herzberg, M.; Syldatk, C. Curr. Opin. Biotechnol. 2001, 12, 559-563. (29) Burckhalter, J. H.; Stephens, V. C. J. Am. Chem. Soc. 1951, 73, 56-58. (30) Que, E. L.; Chang, C. J. J. Am. Chem. Soc. 2006, 128, 15942-15943.

Chapter 5: Tridentate Rhodium(I) Complexes: Synthesis and Structure

“Let me tell you the secret that has led me to my goal. My strength lies solely in my tenacity”

Louis Pasteur (1822 – 1895)

5.1 Introduction

Cationic transition metal complexes have been utilized as homogeneous catalysts for a number of industrially relevant transformations such as hydrogenation,1-5 hydroformylation6-9 -bond metathesis.10-12 Transition metal catalysts allow the synthesis of fine chemicals or pharmaceuticals to proceed under mild conditions with high regioselectivity and stereoselectivity.13-15 The control of the regioselectivity and stereoselectivity of reactions are made possible using modifications of the electronic and steric properties of the metal complex. These modifications are typically introduced by adding substituents to the ligand bound to the metal centre. Despite the high reactivity of homogeneous catalysts, they are not as widely implemented in industry compared to heterogeneous catalysts, mainly due to the difficulty in separating the homogeneous catalyst from the product. To develop a catalytic system that possesses the advantages of both homogeneous and heterogeneous catalysts, there has been a move to immobilize metal complexes onto macroscopic scaffolds such as silicon,16-17 polymers18-19 and proteins.20-21 In examples reported to date, metal complexes have been immobilized through the formation of a covalent bond between a reactive functional group (i.e. NH2 and SH) present on the complex and the desired macroscopic scaffold.16-21 One of the goals of this work was to immobilize metal complexes onto macroscopic scaffolds such as carbon surfaces and proteins.

Cationic rhodium(I) and iridium(I) complexes bearing N,N donor ligands such as bis(pyrazolyl)methane (bpm, 20) are catalytically active, and can also be easily functionalized allowing their selectivity and reactivity to be fine tuned via the addition of substituents. Where group 9 complexes bearing the bpm (20) and bis(N-methylimidazolyl)methane (bim, 21) ligands are particularly well studied,22 the complexes of their corresponding tridentate counterparts are not. The presence of the additional donor in the tridentate ligands such as tris(pyrazolyl)methane (tpm, 22) and tris(N-methylimidazolyl)methanol (tim, 23) can lead to different reactivities when compared to complexes with bidentate ligands due to changes in the coordination chemistry or the ability to stabilize higher oxidation states generated during the catalytic

cycle. Group 9 complexes bearing tridentate ligands 22 and 23 are of particular interest due to the possible differences in catalytic activity and reaction selectivity induced by the presence of the additional donor.

5.1.1 Synthesis of substituted tris(pyrazolyl)alkane ligands

The original method used for the synthesis of the tpm ligand (22), which involved treating potassium pyrazolate with chloroform, was developed by Hückel et al.23 Due to the low yields, attributed to the formation of carbene side products, and the time consuming purification required, complexes bearing tris(pyrazolyl)alkane ligands have not been as well investigated compared to those bearing the bis(pyrazolyl)alkane (20) ligands. Recently Liddle et al.24 published a convenient synthesis for an amine functionalized tris(pyrazolyl)alkane ligand, tris(pyrazoyl)toluidine (tpt, 24).

The advantage of this synthetic route is not only that it leads to high yields and short reaction times, but also that it is a convenient method to introduce an aniline moiety. The aniline moiety provides a reactive functional group on the ligand backbone in the form of the amine group which allows additional modifications to be made. Complexes

bearing the tris(pyrazolyl)toluidine ligand (24) can readily be immobilized onto macroscopic scaffolds as the aniline moiety provides a reactive functional group in the form of the amine group that provides a site for the attachment to the scaffold. Due to the presence of an additional donor, in the form of the primary amine, the coordination chemistry of the complexes bearing the tris(pyrazolyl)toluidine ligand (24) needs to be investigated in detail and compared with that of the tridentate ligands without this substituent and the bidentate ligands. For this chapter, the focus is the synthesis and characterization of rhodium(I) tris(pyrazolyl)toluidine complexes bearing different counterions and co-ligands in order to establish the effect of the aniline moiety on the coordination chemistry of the ligand. Rhodium(I) complexes bearing tris(imidazolyl)methanol ligands were investigated using high resolution NMR spectroscopy and single crystal x-ray diffraction as a comparison to the tris(pyrazolyl)toluidine ligands (24).

5.1.2 Coordination chemistry of tridentate N,N donor ligands

Due to the presence of three pyrazolyl donors, tris(pyrazolyl)alkane ligands can bind to metal ions in a number of different binding modes. The most common binding modes reported in literature for tris(pyrazolyl)alkane ligands are the tridentate binding mode (3, Figure 5.1a) and the bidentate binding mode (2, Figure 5.1b).

Figure 5.1: The two different binding modes for tris(pyrazolyl)alkane ligands: (a) the tridentate binding mode and 3; (b) the bidentate binding mode 2.

The first group 9 complex bearing the tris(pyrazolyl)methane ligand (tpm, 22) was synthesized by Trofimenko et al.25 where a cobalt(II) ion was bound between two tpm ligands (22) in a 3 binding mode forming a sandwich complex. Lalor et al. also synthesized a half sandwich complex where the rhodium centre was bound to the tpm ligand (22) in the 3 binding mode.26 Iridium(I) complexes bearing the tpm ligand (22) have also been extensively studied due to their C-H activation properties. Heineky et al.27 has shown that the tpm ligand (22) binds to the iridium(I) centre in the 3 binding mode, but it was shown that in solution, the complex underwent C-H activation proceeding via. an intermediate where the iridium centre is bound to the ligand in the 2 binding mode. This highlights the ability of the tris(pyrazolyl)alkane ligands to switch between 3 and 2 binding mode, which presents unique properties for these complexes as catalysts compared to those with the bidentate counterparts. The unique properties of these complexes are due to the ability of the tris(pyrazolyl)alkane ligands to stabilize higher oxidation states using the third pyrazolyl donor, and/or due to the introduction of additional constraints which are not available for the complexes with analogous bidentate ligands.

With the development of a more efficient synthesis of tris(3,5- dimethylpyrazolyl)methane (tdmpm, 25),28-29 there has been recent work investigating the coordination chemistry of ligand 25 with a rhodium centre, particularly with cyclic olefins as co-ligands.30 In the complexes reported by Esteruelas et al.,30 it was shown that when bulky cyclic olefins are present as co-ligands, the tdmpm ligand (25) was bound to the rhodium centre in a 3 binding mode. It was shown by Adams et al.31 that when a less bulky co-ligand such as carbon monoxide is present, the iridium centre is bound to the tdmpm ligand (25) in both 2 and 3 binding modes, where both forms exist in equilibrium in solution. Replacing the carbon monoxide with the bulky 1,5- cyclooctadiene co-ligand causes the equilibrium of the 2 3 exchange to favor the 2 binding mode. This was recently shown by Hallett et al.,32 where replacing the cyclooctadiene with a more constrained norbornadiene co-ligand led to a change in the binding mode of the tdmpm ligand (25) around the rhodium centre from 2 (Figure 5.2a) to 3 (Figure 5.2b).

(a) (b)

Figure 5.2: The molecular structure of: (a) [Rh(COD)(tdmpm)]+; and (b) [Rh(NBD)(tdmpm)]+ synthesized by Hallett et al.32

These reports highlight the sensitivity of the 2 3 equilibrium to changes around the metal centre. As a result, the presence of an additional reactive functional group on the ligand such as an amine group, could significantly influence the position of the 2 3 equilibrium of rhodium complexes bearing a ligand such as tris(pyrazolyl)toluidine (24). To explore the effects of an aniline moiety attached to the backbone of the tris(pyrazolyl)toluidine ligand (24) on the coordination chemistry, rhodium(I) tris(pyrazolyl)toluidine complexes bearing cyclooctadiene (COD), norbornadiene (NBD) and carbonyl (CO) co-ligands were synthesized and studied in both the solid and the solution state using single crystal x-ray diffraction and variable temperature NMR spectroscopy respectively.

5.2 Rhodium(I) complexes bearing p-tris(pyrazolyl)toludine ligand

5.2.1 Synthesis of the tris(pyrazoyl)toluidine rhodium(I) complexes

The rhodium(I) tris(pyrazolyl)toluidine olefin complexes 27 and 28 were synthesized by stirring a solution of [Rh(L)2][BArF] (L = COD, 26a; L = NBD, 26b) with the appropriate ligand 24 in tetrahydrofuran (THF) at room temperature. The olefin complexes

27 and 28 were isolated and purified by removal of the THF in vacuo, followed by recrystallization from a mixture of dichloromethane and n-pentane. The corresponding dicarbonyl complex 29 was synthesized by stirring a degassed solution of the olefin complex 27 in dichloromethane under a carbon monoxide (CO) atmosphere. The dicarbonyl complex 29 was isolated by the addition of n-pentane followed by recrystallization from a mixture of dichloromethane and n-pentane.

Scheme 5.1: The generalized synthesis of the diolefin complexes 27 and 28 and the conversion to the corresponding dicarbonyl complexes 29.

5.2.2 Solid state structure of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a)

Crystals of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a) suitable for x-ray diffraction analysis were obtained by the slow diffusion of n-pentane into concentrated solution of the appropriate complex in dichloromethane. The crystals were analyzed using single crystal x-ray diffraction and the complexes 27a and 28a crystallized in the space group P21/c and C2/c respectively. ORTEP depictions of the two complexes 27a and 28a showing the atom numbering of selected atoms are shown in Figure 5.3. Selected bond lengths and angles are presented in Table 5.1.

The solid state structures of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a) show that the rhodium centre is bound to the p-tpt ligand (24a) in the 2 binding mode irrespective of the co-ligand present. The sums of the four ligand bite angles (N2-Rh-X1, N1-Rh-X2, N1-Rh-N2 and X1-Rh-X2) are 360.12o and 359.88o for the complexes 27a and 28a respectively which confirm the square planar geometry of the complexes. This is contrary to rhodium(I) olefin complexes reported in literature which bear the tris(pyrazolyl)methylsulfonate33 and tris(3,5-dimethylpyrazoyl)methane32 ligands. For the complexes which bear the tris(pyrazolyl)alkane ligands reported in literature,32-33 the change from a less constrained olefin such as cyclooctadiene to a more constrained olefin causes a change in the coordination geometry around the rhodium centre from square planar to trigonal pyramidal. The difference in coordination geometry of the p-tpt ligand (24a) and the literature examples32-33 could be due to the presence of the aniline group on the backbone of the p-tpt ligand (24a). In the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a), the resulting six membered metallocycles adopt a pseudo boat conformation.

(a)

(b)

Table 5.1: Selected bond distances for the complexes [Rh(COD)(p-tpt)][BArF] (27a), [Rh(NBD)(p-tpt)][BArF] (28a) and [Rh(COD)(bpm)][BArF] (31).34

[Rh(COD)(p-tpt)] [Rh(NBD)(p-tpt)] [Rh(COD)(bpm)]

[BArF] (27a) [BArF] (28a) [BArF] (31)

Atom Pair Bond Lengths (Å)

Rh – N12 2.101 2.076 2.099

Rh – N22 2.101 2.070 2.096

Rh – X1 a 2.015 2.008 2.023

Rh – X2 a 2.024 1.997 2.021

Bond Angles (o)

N2 – Rh – X1 a 92.84 100.91 92.86

N1 – Rh – X2 a 94.51 101.10 92.80

N1 – Rh – N2 84.83 86.35 87.01

X1 – Rh – X2 a 87.94 71.52 87.36 a X1 and X2 is defined as the centroid for the bonds C41 – C42 and C45 – C46 respectively for [Rh(COD)(p-tpt)][BArF] (27a); X1 and X2 is defined as the centroid for the bonds C44 – C45 and C41 – C42 respectively for [Rh(NBD)(p-tpt)][BArF] (28a); X1 and X2 is defined as the centroid for the bonds alkene bonds for the cyclooctadiene co-ligand for [Rh(COD)(bpm)][BArF] (31)

Comparing the selected bond lengths given in Table 5.1 for the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(COD)(bpm)][BArF] (31) shows that despite the presence of the aniline moiety and an additional pyrazolyl donor in the complex 27a, there are no significant deviations in the Rh-N and the Rh-X bond lengths between the two complexes. The ligand bite angles for the complex 27a given in Table 5.1 shows deviations from the expected ideal square planar values of 90o. This suggests that there is some limited ring strain present in the six-membered metallocycle upon the coordination of the ligand 24a with the rhodium centre. Comparing the bond distances and angles shown in Table 5.1 between the cyclooctadiene complexes 27a, 31 and the

norbornadiene complex 28a, significant deviation in the bond angles were observed. Norbornadiene, structurally a more constrained olefin compared to cyclooctadiene due to the smaller ring containing the two olefin donors, leads to a smaller X1 – Rh – X2 bond angle compared to that observed for the analogous cyclooctadiene complex 27a.

5.2.3 Solution state structure of the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(o-tpt)][BArF] (28a)

The solution state structures of the complexes 27a and 28a were examined using variable temperature NMR spectroscopy. At room temperature, the complexes [Rh(COD)(p-tpt)][BArF] (27a) and [Rh(NBD)(p-tpt)][BArF] (28a) undergo exchange which can be attributed to either restricted rotation of the aryl ring about the C-C bond between the bridging carbon of the tris(pyrazolyl) motif and the phenyl ring or the pyrazolyl donors of the tridentate ligand exchanging with each other. At low temperatures conformational exchange of the ligand is slower, and two of the pyrazolyl donors are bound to the rhodium centre while the third pyrazolyl donor remains unbound, as indicated by the 1H NMR spectra. The 1H NMR spectra for the complex [Rh(COD)(p-tpt)][BArF] (27a) at 25oC and -55oC are shown below in Figure 5.4.

The 1H NMR spectrum at room temperature (Figure 5.4a) of the complex [Rh(COD)(p-tpt)][BArF] (27a) contains broad resonances, but once the sample is cooled to -55oC (Figure 5.4b), the linewidths of the resonances become narrower. This indicates that the compound is undergoing an exchange process in solution, which slows upon cooling. The 1H NMR spectrum of the complex 27a at room temperature contains a single broad resonance due to the pyrazole H4 proton, which upon cooling resolves into two triplets integrating to a ratio of 1 : 0.5. There are also two sets of resonances, each attributed to the pyrazole H3 and H5 protons respectively, which also integrated to a ratio of 1 : 0.5. The resonances due to the pyrazole protons show that there are two unique pyrazolyl donor environments at low temperature. The integration of resonances due to the pyrazole protons indicated that two of the pyrazolyl donors of the p-tpt ligand (24a) were equivalent, while the third pyrazolyl donor was different. This suggests that

two of the pyrazolyl donors are bound to the rhodium centre while the third pyrazolyl donor remains unbound. This is consistent with the molecular structure determined using x-ray crystallography.

1 o (a) H CD2Cl2, 25 C

(27a)

1 (b) H CD2Cl2, -55C

H4 3’ H H2’ H4

H3/5

H3/5 H3/5 COD COD COD COD

Figure 5.4: 1H NMR spectrum of [Rh(COD)(p-tpt)][BArF] (27a) at: (a) 25oC; and (b) o -55 C in CD2Cl2.

The 1H NMR spectrum of the complex [Rh(NBD)(p-tpt)][BArF] (28a) shows that this complex is undergoing the same fluxional behavior at room temperature as observed for 97

the complex 27a. Upon cooling, (Figure 5.5) the resonances due to the pyrazole H4 protons of the complex 28a resolve into two triplets indicating that there are two unique pyrazolyl donor environments at low temperature with two of the pyrazolyl donors bound to the rhodium centre and the third pyrazolyl donor unbound. Using a 1H-COSY correlation experiment (not shown), the corresponding resonances due to pyrazole H3 and H5 protons were also assigned, and found to integrate to a ratio of 1 : 0.5. The relative integration of the resonances due to the pyrazole protons indicates that there are two unique pyrazolyl donor environments at low temperature. This is due to the presence of two pyrazolyl donors which are bound to the rhodium centre, while the third pyrazolyl donor remains free. This is consistent with the solid state structure of the complex 28a shown in 5.3b.

BArF BArF (28a)

H3/5 H4 3’ 2’ H3/5 H 4 H H3/5 H

Figure 5.5: 1H NMR spectrum of [Rh(NBD)(p-tpt)][BArF] (28a) showing only the o aromatic region at -55 C in CDCl3.

5.2.4 Solution state structure of the complex [Rh(CO)2(p-tpt)][BArF] (29a)

Attempts to grow a crystal of the complex [Rh(CO)2(p-tpt)][BArF] (29a) suitable for x- ray diffraction analysis by the diffusion of n-pentane into a concentrated solution of the

complex 29a in dichloromethane were unsuccessful. The structure of the complex was determined using variable temperature NMR spectroscopy (Figure 5.6). The 1H NMR spectrum at low temperature shows that two of the pyrazolyl donors of the p-tpt ligand

(24a) are bound to the rhodium centre of the complex [Rh(CO)2(p-tpt)][BArF] (29a), while the third pyrazolyl donor remains unbound.

1 o (a) H CD2Cl2, 25 C

1 o (b) H CD2Cl2, -55 C

(29a)

H4 3/5 3/5 H H 3’ H H2’ 3/5 3/5 H H H4

1 Figure 5.6: H NMR spectrum of [Rh(CO)2(p-tpt)][BArF] (29a) showing only the o o aromatic region at (a) 25 C; and (b) -55 C in CD2Cl2 .

At room temperature (Figure 5.6a), the resonances due to the pyrazole protons of the complex 29a are broad, suggesting that a fluxional process is occurring. The sample was cooled to -55oC and the resonances due to the pyrazole H4 protons resolved into

two multiplets integrating to a ratio 1 : 0.5. This indicates that there are two unique pyrazolyl donor environments present at low temperature, similar to the olefin complexes 27a and 28a discussed in Section 5.2.2. Using a 1H-COSY correlation experiment the corresponding pyrazole H3 and H5 protons were assigned. Integration of the pyrazole resonances showed a ratio of 1 : 0.5 which suggests that there are two pyrazolyl donors bound to the rhodium centre, with a third pyrazolyl donor unbound. Based on data from the olefin complexes 27a and 28a it can be concluded that in the o complex [Rh(CO)2(p-tpt)][BArF] (29a) at -50 C, the rhodium centre is bound in a square planar geometry.

The study of the coordination chemistry of this series of p-tris(pyrazolyl)toluidine rhodium(I) complexes 27a, 28a and 29a each bearing the different co-ligands COD, NBD and CO, respectively, has shown a preference for the p-tris(pyrazolyl)toluidine ligand (24a) to act as a bidentate ligand. Despite the use of a more constrained co-ligand such as norbornadiene, the binding mode does not change, unlike other analogous complexes bearing the tris(pyrazolyl)alkane ligands reported previously.32

5.2.5 Synthesis and structure of the complex [Rh(COD)(p-tpt)][BPh4] (32)

To determine if the preference of the bidentate binding mode observed for the complex 27a bearing p-tpt ligand (24a) is due to the effects of the BArF counterion, the analogous BPh4 complex was synthesized and the resulting complex studied using x-ray crystallography. It has been shown in literature that the BArF counterion is weakly 35 coordinating compared to other counterions such as the BPh4, PF6 and BF4. As a result, the ability of the counterion to coordinate to the cationic fragments can influence !\$!2 3 binding mode.

The complex [Rh(COD)(p-tpt)][BPh4] (32) was synthesized following the method published previously for the analogous rhodium(I) bis(pyrazolyl)methane complexes bearing the tetraphenylborate counterion.36 The synthesis involves stirring a solution of

100

[Rh(COD)Cl]2 (33), p-tpt (24a) and sodium tetraphenylborate in THF at room temperature (Scheme 5.2).

Scheme 5.2: Synthesis of the rhodium(I) olefin complexes bearing the tetraphenylborate counterion.

Crystals of the complex 32 suitable for x-ray diffraction analysis were grown by the diffusion of n-pentane into a concentrated solution of the complex 32 in dichloromethane. The crystals were analyzed using single crystal x-ray diffraction and the complex 32 crystallized in the space group Pbca. ORTEP depictions of the complex 32, showing the numbering of selected atoms, associated bond lengths and angles, are shown below in Figure 5.7.

The solid state structure of the complex 32 shows that the rhodium centre of 32 is bound to the p-tpt ligand (24a) in !2 binding mode. The sum of the bite angles shown in Figure 5.7 is 359.99o which confirms the square planar geometry around the rhodium centre. The resulting six membered metallocycle adopts a pseudo boat conformation. Comparing the bond lengths shown in Figure 5.7 for the complex

[Rh(COD)(p-tpt)][BPh4] (32) with the corresponding BArF complex 27a (Table 5.1), there are no significant differences between the bond lengths of the two complexes

101

indicating that the counterion has no influence on the nature of the coordination chemistry of the p-tris(pyrazolyl)toluidine ligand (24a) to rhodium.

[Rh(COD)(p-tpt)] Atom Pair [BPh4] (32)

Bond Distances (Å)

N12 – Rh1 2.083

N22 – Rh1 2.077

Rh1 – X1 2.012

Rh1 – X2 2.033

Bond Angles (o)

X1 – Rh – X2 87.21

N1 – Rh – N2 86.92

N1 – Rh – X1 93.63

N2 – Rh – X2 92.23

b X1 and X2 is defined as the centroid for the bonds C44 – C45 and C41 – C42 respectively

5.3 Rhodium(I) complexes bearing o-tris(pyrazoyl)toludine ligand

In the previous section, it was shown that the p-tris(pyrazolyl)toluidine ligand (24a) \$!!2 binding mode irrespective of the co-ligand and the counterion. In the case of the o-tris(pyrazolyl)toluidine ligand (24b), the close proximity of the amine group to the pyrazole moiety could influence the coordination

102

chemistry. To investigate the effect of the position of the amine group, rhodium(I) complexes bearing the o-tris(pyrazolyl)toluidine ligand (24b) were synthesized and studied using x-ray crystallography and variable NMR spectroscopy.

5.3.1 Solid state studies of [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o-tpt)][BArF] (28b)

Crystals of the complexes [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o-tpt)][BArF] (28b) suitable for x-ray diffraction analysis were obtained by the slow diffusion of pentane into a concentrated solution of the appropriate complex in dichloromethane at room temperature. The crystals were analyzed using single crystal x-ray diffraction and the complexes 27b and 28b crystallized in the same space group P1. ORTEP depictions of the cationic fragments for the complexes 27b and 28b showing the numbering of selected atoms are shown in Figure 5.8 and Figure 5.9 respectively. Selected bond lengths and bond angles for both complexes 27b and 28b are summarized in Table 5.2.

For the complex [Rh(COD)(o-tpt)][BArF] (27b) two structures co-crystallized within the unit cell. Both structures show the rhodium centre bound to the o- tris(pyrazolyl)toluidine ligand (24b) in a square-planar geometry. The two structures present in the unit cell differ in the orientation of the unbound pyrazolyl donor relative to the metal centre. In structure A (Figure 5.8a), the unbound pyrazolyl donor is positioned pointing away from the rhodium centre, while in Structure B (Figure 5.8b), the unbound pyrazolyl donor is located above the rhodium centre. The sums of the ligand bite angles shown in Table 5.2 for both structures A and B are 361.2o and 361.7o respectively which confirms the square planar arrangement of the complex. The resulting metallocycles of both structures adopt pseudo boat conformations. In the case of the complex [Rh(NBD)(o-tpt)][BArF] (28b) only one structure was present in the unit cell, with the rhodium centre bound in a square planar geometry. The change to a more constrained olefin such as NBD does not change the nature of the coordination of the o-tris(pyrazolyl)toluidine ligand (24b) as was observed in the case of the rhodium(I) complexes with the p-tris(pyrazolyl)toluidine ligand (24a) (Section 5.2).

103

Table 5.2: Selected bond distances for the complex [Rh(COD)(o-tpt)][BArF] (27b), [Rh(NBD)(o-tpt)][BArF] (28b) and [Rh(COD)(bpm)][BArF] (31).34

[Rh(COD)(o-tpt)][BArF] [Rh(NBD)(o-tpt)] [Rh(COD)(bpm)] (27b) b b b [BArF] (28b) [BArF] (31) Structure A Structure B

Atom Pair Bond Lengths (Å)

Rh – N1 2.081 2.090 2.077 2.099

Rh – N2 2.084 2.095 2.083 2.096

Rh – X1 a 1.998 2.007 1.990 2.023

Rh – X2 a 2.009 2.024 1.995 2.021

Bond Angles (o)

N2 – Rh – X1 a 95.16 94.11 98.87 92.86

N1 – Rh – X2 a 93.28 93.75 101.97 92.80

N1 – Rh – N2 86.46 87.07 87.90 87.01

X1 – Rh – X2 a 86.30 86.77 71.16 87.36 a X1 and X2 is defined as the centroid for the bonds C41 – C42 and C45 – C46 respectively for [Rh(COD)(o-tpt)][BArF] (27b); X1 and X2 is defined as the centroid for the bonds C44 – C45 and C41 – C42 respectively for [Rh(NBD)(o-tpt)][BArF] (28b). b Structures A and B refer to the two co-crystallized structure within the unit cell of the complex [Rh(COD)(o-tpt)][BArF] (27b) as determined using single crystal x-ray diffraction.

104

(a)

(b)

Figure 5.8: ORTEP depiction of [Rh(COD)(o-tpt)][BArF] (27b) showing the two co- crystallized structures: (a) Structure A; and (b) Structure B at 50% thermal ellipsoids for non-hydrogen atoms. Nitrogen, rhodium and carbon atoms are colored in blue, green and black respectively

105

Figure 5.9: ORTEP depiction of [Rh(NBD)(o-tpt)][BArF] (28b) at 50% thermal ellipsoids for non-hydrogen atoms. Nitrogen, rhodium and carbon atoms are colored in blue, green and black respectively

Comparing selected bond lengths shown in Table 5.2 for the complex [Rh(COD)(o-tpt)][BArF] (27b) with those of the corresponding p-tpt analogue (27a, Table 5.1), no significant changes in the bond lengths were observed. This indicates that the position of the aniline group has no significant influence on the nature of coordination of the o-tris(pyrazoyl)toluidine ligand (24b) to rhodium. Comparing the bond angles and distances between Structure A (Figure 5.8a) and Structure B (Figure 5.8b) found in the unit cell for the complex [Rh(COD)(o-tpt)][BArF] (27b) shows that the orientation of the unbound pyrazolyl donor has no influence on the coordination geometry around the rhodium centre. Changing the co-ligand from cyclooctadiene to norbornadiene leads to no significant changes in the either of the Rh-N or the Rh-X bond lengths between the two structures. A significant change of up to 15o was observed in the N-Rh-N and X-Rh-X ligand bite angles shown in Table 5.2 which was attributed to the more constrained nature of the norbornadiene ligand. This same trend was observed in the analogous complex series bearing the p-tpt ligand (24a). A comparison of the bond lengths and angles for the o- (28b) and p-tpt complexes (28a)

106

shown in Table 5.2 and Table 5.1, respectively, for norbornadiene co-ligand shows no significant differences in the bond lengths and angles between the two sets of complexes.

5.3.2 NMR studies of [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o- tpt)][BArF] (28b)

The solution state structure of the complexes 27b and 28b were examined using variable temperature NMR spectroscopy. The 1H NMR spectra of the complexes [Rh(COD)(o-tpt)][BArF] (27b) and [Rh(NBD)(o-tpt)][BArF] (28b) at low temperature show the presence of major and minor species undergoing exchange at low temperatures, with both major and minor species containing three inequivalent pyrazolyl donor environments.

The 1H NMR spectrum of the complex 27b at room temperature (Figure 5.10a) shows broad resonances, indicative of an exchange process occurring within the complex. The 1H NMR spectrum at -55oC shows that the broad resonances have resolved into sharp resonances, indicating the exchange in solution has slowed or ceased (Figure 5.10b). Inspection of the 1H NMR spectrum of the complex 27b at -55oC (Figure 5.10b) shows that there are two products present at low temperature, highlighted by the appearance of a doublet at 5.5 ppm and a second doublet at 5.80 ppm, both assigned as the H2’ proton on the aniline ring based on a 1H-COSY correlation experiment. This behavior was also observed for the pyrazole resonances where, for example, two sets of resonances due to the pyrazole H4 protons can be observed. It is interesting to note that for both the major and the minor product, three resonances due to the pyrazole H4 protons were observed, indicating that in the major and minor products all three pyrazolyl donors are inequivalent. Similar behavior was observed in the complex [Rh(NBD)(o-tpt)][BArF] (28b) (Figure 5.11).

107

1 o (a) H CDCl3, 25 C

(27b)

1 o (b) H CDCl3, -55 C H3/5,P(2)

4,P(3) H3/5,P(2) H3/5,P(1) H 3’ 5’ H3/5,P(3) 3/5,P(1) H & H H H4,P(2) 4,P(1) H3/5,P(3) H H3/5,minor H4,P,minor 4’ 3/5,minor 4,P,minor 3/5,minor H H H H H4,P,minor H3/5,minor H3’,minor

H2’

H2’,minor

Figure 5.10: 1H NMR spectrum of [Rh(COD)(o-tpt)][BArF] (27b) showing only the o o aromatic region at: (a) -55 C; and (b) 25 C in CDCl3 measured on Bruker Avance 600 MHz spectrometer.

108

H3/5 H3/5 + H3/5 H3/5

(28b)

3 H3/5 H4 H H4 H4

H4’ H5’ H3’ H2’

H2’,minor

Figure 5.11: 1H NMR spectrum of [Rh(NBD)(o-tpt)][BArF] (28b) showing only the o aromatic region at -55 C in CD2Cl2 measured on Bruker Avance 600 MHz spectrometer.

There are two possible explanations for the appearance of the two species in the 1H NMR spectrum of 27b shown in Figure 5.10: (a) at low temperature, the rate of exchange between two species is sufficiently slow that they appear as two separate products in the 1H NMR spectrum; and (b) an impurity introduced during the synthesis of the complex [Rh(COD)(o-tpt)][BArF] (27b). To confirm that the second product is the result of an exchange process, a 1H-NOESY correlation experiment was conducted on the complex [Rh(COD)(o-tpt)][BArF] (27b) at -55oC and part of this spectrum is shown in Figure 5.12.

109

3’,minor 3/5 H H H3’ H2’ 3/5,minor H H2’,minor

(27b)

Figure 5.12: 1H NOESY spectrum of [Rh(COD)(o-tpt)][BArF] (27b) measured on Bruker Avance 600 MHz spectrometer showing only the aromatic region at -50oC in

CDCl3. Highlighted are the exchange cross peaks between the resonances due to the pyrazole H3/5, the H2’ and H3’ of the aniline ring.

In the 1H-NOESY spectrum of 27b (Figure 5.12), cross peaks between resonances due to the protons of the low concentration species and the higher concentration species were observed. The exchange peaks between the proton resonances assigned to the aniline moiety (H2’ and H3’) and the pyrazole protons (H3/5) of the two species are highlighted in Figure 5.12. The relative signs of the cross peaks compared to the diagonal indicate they are the result of an exchange process. This indicates that the two species observed at low temperatures are undergoing an exchange process. To ensure 110

this second species is not an impurity, elemental analysis of the complex [Rh(COD)(o-tpt)][BArF] (27b) was also conducted, and the resulting elemental composition is consistent with that expected for a pure sample. In the NOESY spectrum shown in Figure 5.12, cross peaks due to through space interactions between protons were not observed. The lack of through space nOe cross peaks is a result of the correlation time of the complex being such that no nOe enhancement is produced, resulting from high molecular weight of the complex in conjunction with the changes in solvent viscosity.

To determine if the exchange process between the two species observed for the complexes 27b and 28b is a result of the nature of binding of the olefin co-ligand, the complex with the dicarbonyl co-ligand [Rh(CO)2(o-tpt)][BArF] (29b) was synthesized and the structure analyzed using both single crystal x-ray diffraction and NMR spectroscopy.

5.3.3 NMR and solid state structure of [Rh(CO)2(o-tpt)][BArF] (29b)

Crystals of the complex [Rh(CO)2(o-tpt)][BArF] (29b) suitable for x-ray diffraction analysis were obtained by the slow diffusion of n-pentane into a concentrated solution of dichloromethane at room temperature. The crystal was analyzed using single crystal x-ray diffraction and the complex 29b crystallized in the space group P1. ORTEP depictions of the cationic fragments for the complex 29b showing the numbering of selected atoms are shown in Figure 5.13. Selected bond lengths and bond angles are summarized in Table 5.3.

111

(a)

(b)

Figure 5.13: ORTEP depiction of the cationic fragment of the complex

The solid state structure of the complex [Rh(CO)2(o-tpt)][BArF] (29b) shows that the rhodium centre is bound to the o-tris(pyrazolyl)toluidine ligand (24b!2 binding mode. Unlike the corresponding COD complex 27b, only one structure was present in the unit cell, with the unbound pyrazolyl donor positioned above the metal centre. The sum of the ligand bite angles about the metal centre shown in Table 5.3 is 359.99, which confirms the square planar geometry around the rhodium centre. 112

Table 5.3: Selected bond distances for the complex [Rh(CO)2(o-tpt)][BArF] (29b) and 34 the analogous [Rh(CO)2(bpm)][BArF] (35).

[Rh(CO) (o-tpt)] [Rh(CO) (bpm)] 2 2 [BArF] (29b) [BArF] (35)

Atom Pair Bond Distance (Å)

N22 – Rh 2.083 2.074

N12 – Rh 2.086 2.087

Rh1 – CO 1.815 1.860

Rh1 – CO 1.816 1.868

Bond Angles (o)

N12 – Rh – N22 86.64 87.59

C41 – Rh – C42 88.32 88.09

N12 – Rh – C41 91.66 92.12

N22 – Rh – C42 93.37 92.17

Comparing the selected bond lengths shown in Table 5.3 for the complex 29b with the analogous [Rh(CO)2(bpm)][BArF] (35) (Table 5.3), no significant differences in the Rh- N and Rh-X bond lengths were observed, indicating that the aniline group has no influence in the coordination geometry of the pyrazolyl donors of the o-tpt ligand (24b). The ligand bite angles shown in Table 5.3 vary from the ideal value of 90o for a square planar geometry due to the strain present upon binding the ligand to the rhodium centre. The structure of the dicarbonyl complex 29b was also studied in the solution state using variable temperature NMR spectroscopy. The 1H NMR spectrum of the complex

[Rh(CO)2(o-tpt)][BArF] (29b) at -50°C and associated assignments are shown below in Figure 5.14.

113

5’ H H4,P(3) H3’ H3/5,P(3) H4,P(1) H4,P(2) (29b)

H3/5,P(3) H3/5,P(2) H3/5,P(2) H3/5,P(1) H3/5,P(1) H3/5,minor 3/5,minor H 2’ 4’ 3/5,minor H H H

H3/5,minor H2’,minor

1 Figure 5.14: H NMR spectrum of [Rh(CO)2(o-tpt)][BArF] (29b) showing only the o aromatic region at -55 C in CD2Cl2 measured on Bruker Avance 400 MHz spectrometer.

Examining the 1H NMR spectrum of the dicarbonyl complex 29b in detail shows that, as in the case of the olefin complexes 27b and 28b, there are two products in the sample. An elemental analysis of the sample was conducted and it was found that the sample was pure. This suggests that the appearance of a second set of resonances is again due to an exchange process occurring in solution. Examining the resonances due to the pyrazole protons (i.e. H3, H4 and H5), three inequivalent pyrazolyl donor environments were observed in the solution state at low temperature. Considering the observed 1H NMR data in conjunction with the molecular structure of the complex 29b obtained using crystallography shows a discrepancy. Based on the solid state structure (Figure 5.13), it would be expected that two of the pyrazolyl donors would have the same environment, as the aniline group is positioned symmetrically between the two pyrazolyl donors bound to the rhodium centre creating a plane of symmetry. One

114

possible explanation is that in the solution state the position of the amine group is located between the unbound and one of the bound pyrazole rings leading to the formation of three different pyrazolyl environments in the 1H NMR spectrum, or alternatively that the aniline substituent is not positioned in a precisely symmetrical fashion.

5.3.4 Understanding the structural exchange processes in rhodium(I) complexes bearing o-tris(pyrazolyl)toluidine ligands: 32 Exchange vs. Atropisomerism

Depending on the experimental conditions (i.e. solvent, co-ligands and temperature), { \ ! 3\2) binding mode. One of the possible exchange processes occurring in solution could be attributed to the interconversion of the two binding modes with either all three pyrazolyl donors bound to the rhodium centre or only two of the pyrazolyl donors bound to the rhodium centre (Scheme 5.3a). Another possible exchange process is atropisomerism, that is isomerism caused by the rotation about a single bond (Scheme 5.3b). Due to the location of the amine group relative to the pyrazole rings, as the aryl ring of the aniline substituent rotates relative to the pyrazole rings, a different chemical environment can be generated for each rotational position and as such different chemical shifts could be observed in the 1H NMR spectrum.

One approach to determining if the exchange processes observed for the o-tris(pyrazolyl)toluidine complexes (27b - 29b) are due to the V$!3 !2 binding mode, is to establish the geometry around the bridging carbon of the ligand, which provides an indirect method for examining the nature of coordination around the rhodium centre. In the case of tris(pyrazolyl)borate rhodium(I) complexes reported in literature, IR and 11B NMR spectroscopy were used to determine the binding mode of the ligand.37-40 Using the same concept, Hallett et al.32 have shown that the 13C chemical shift of the bridging carbon can be used to determine the binding mode around the rhodium centre. In the case of the rhodium(I) tris(3,5-dimethylpyrazolyl)methane

115

complexes,32 $!!!\!2 binding mode, the 13C chemical shift of the bridging carbon lies in the range of 74 – 75 ppm. When the ligand \3 binding mode, the chemical shift of the bridging carbon appears upfield at 67 – 70 ppm. By considering the 13C chemical shift of the bridging carbon, the binding mode of two species that appear in 1H NMR spectrum can be ascertained. If the resonances due to the bridging carbon of the two species are each at a significantly different 13C chemical shift (i.e. 1!!| species observed in the 1H NMR spectrum for the complexes 27b - 29b have different binding modes.

Scheme 5.3: The possible exchange processes occurring in solution: (a) the V\!2 3 binding mode; and (b) atropisomerism.

Using 1H-13C HMBC correlation experiment, the 13C resonance due to the bridging atom of the complex [Rh(CO)2(o-tpt)][BArF] (29b) was assigned through its correlation with the H2’ proton on the aryl ring. It was found that the bridging carbon for the major

116

species has a 13C chemical shift of 92.8 ppm, while the minor product has a 13C chemical shift of 93.0 ppm. Based on the difference of 0.2 ppm between the 13C chemical shift of the two species, compared to the difference of 7 ppm observed for the rhodium(I) tris(3,5-dimethylpyrazolyl)methane complexes,32 it was thought that both species that are present in the 1H NMR spectrum at low temperature are bound to the rhodium centre in the same geometry. This implies that the fluxional process observed in the 1H NMR spectrum is likely caused by the rotation of the aniline ring rather than a change in the binding mode of the ligand.

5.4 Rhodium(I) complexes bearing the tris(N-methylimidazolyl) -methanol ligand

metal complexes bearing imidazolyl ligands have also been shown in literature to be highly active in catalysis36,41 and are considered analogues to the poly(pyrazolyl)alkane ligands.42 As such, a logical extension of this work would be to consider the coordination chemistry of rhodium(I) complexes bearing the tris(N- methylimidazolyl)methanol ligand (23). The complexes [Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37) were synthesized and the structures in the solid and solution state examined using x-ray crystallography and NMR spectroscopy respectively.

5.4.1 Synthesis of [Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37)

The complexes [Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37) were synthesized using the same methods as described in section 5.2.1 with yields of 70% and 57% respectively. Single crystals of the complexes [Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37) suitable for x-ray diffraction analysis were grown by the slow diffusion of n-pentane into a concentrated solution of the appropriate complex in dichloromethane at room temperature overnight.

117

5.4.2 Solid state structures of [Rh(COD)(tim)][BArF] (36) and

[Rh(CO)2(tim)][BArF] (37)

Single crystals of the complexes 36 and 37 were analyzed using single crystal x-ray diffraction. ORTEP depictions of the two complexes 36 and 37 showing the numbering of selected atoms are shown in Figure 5.15 and Figure 5.16. Selected bond lengths and angles are summarized in Table 5.4. The ORTEP representations for the complexes

[Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37) shows that the rhodium centre is bound to the tris(N-methylimidazolyl)methanol ligand (23) in a square planar geometry with two of the imidazolyl donors bound to the metal centre. The six membered metallocycles of the complexes 36 and 37 each adopt a pseudo boat conformation. In the ORTEP representation (Figure 5.16) of the complex

[Rh(CO)2(tim)][BArF] (37), it is interesting to note the association between the two cationic fragments in the unit cell. The ORTEP representations of the complex 37 viewed from the side (Figure 5.16a) and viewed the top (Figure 5.16b) shows that the two cationic fragments in the unit cell sit directly above each other in the solid state. The distance between the two rhodium centers was measured to be 3.286 Å, which is longer than the published Rh – Rh bonds that are in the range between 2.7 – 2.9 Å.43-44 As a result, it was proposed the two structures in the unit cell were closely associated but not covalently bound.

118

(a) Side view (b) Top view

3.3Å

N-CH3 (c) 1H NMR at 25oC in d-methanol

BArF HIm2 & HIm1

Figure 5.16: ORTEP representation of the cationic fragment of [Rh(CO)2(tim)][BArF] (37) at 50% ellipsoid for non hydrogen atoms viewed from (a) the top; (b) top. Nitrogen atoms are colored in blue, rhodium atoms are colored in green and carbon atoms colored 1 in black. The H NMR spectrum of the complex at 25°C of the complex 37 in d4- methanol measured on a Bruker Avance 300 MHz Spectrometer.

119

Table 5.4: Selected bond distances and angles for the complexes

[Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37).

[Rh(COD)(tim)] [Rh(CO)2(tim)]

[BArF] (36) [BArF] (37)

Atom Pair Bond Distances (Å) Atom Pair Bond Distances (Å)

Rh – N12 2.083 Rh – N12 2.054

Rh – N22 2.085 Rh – N22 2.072

Rh – X1a 2.135 / 2.146 Rh – C2 1.830

Rh – X2a 2.124 / 2.150 Rh – C3 1.843

Bond Angles (o) Bond Angles (o)

N12 – Rh – N12 – Rh – N22 86.38 86.30 N22

X1 – Rh – X2a 87.14 C2 – Rh – C3 86.26

N12 – Rh – X1 93.76 N1 – Rh – C3 94.35

N22 – Rh – X2a 92.42 N2 – Rh – C2 92.74 a X1 and X2 is defined as the centroid for the bonds C41 – C42 and C45 – C46 respectively for [Rh(COD)(tim)][BArF] (36).

Comparing selected bond lengths shown in Table 5.3 for the complex [Rh(COD)(tim)][BArF] (36) and the corresponding tris(pyrazolyl)toluidine (27a, Table 5.1 and 27b, Table 5.2) complexes, no significant differences between the corresponding bond lengths were observed, despite the change in the donor strength of the ligand. It is also important to note that the 1H NMR spectra of the imidazolyl complexes 36 and 37 do not indicate that the complexes are undergoing any fluxional behavior at room temperature, unlike the tris(pyrazoyl)toluidine complexes (27 – 29). Based on the solid state structure of the imidazolyl complexes 36 and 37, two unique imidazolyl donor environments are expected, but the 1H NMR spectrum at room temperature only shows one set of resonances due to all three imidazolyl donors. A

120

possible explanation for this is that the interconversion between the two binding modes (i.e. 2 and 3) is sufficiently fast on the NMR time frame such that all three imidazolyl donors become equivalent, leading to the appearance of a single set of imidazolyl donor 1 o resonances. The H NMR spectrum of the complex [Rh(CO)2(tim)][BArF] (37) at -90 C also did not resolve the resonances due to the imidazolyl donors in the different chemical environments.

Based on the analysis of the molecular structures of the complexes

[Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37), it can be shown that despite the changes to the nature of the nitrogen donor, the 2 binding mode is preferred for the poly(imidazolyl)alkane ligands.

5.5 Conclusion

Rhodium(I) complexes bearing the p- (24a) and o-tris(pyrazolyl)toluidine ligands (24b) were synthesized and their coordination chemistry studied in the solid and solution state using x-ray crystallography and NMR spectroscopy respectively. Despite the potential of the tris(pyrazolyl)toluidine ligands to bind to the rhodium centre in either a 2 (bidentate) or a 3 (tridentate) binding mode, the 2 binding mode was found to be the preferred binding mode for all complexes studied based on spectroscopic evidence. Despite significant changes to the nature of the co-ligands (i.e. norbornadiene vs. 1,5 cyclooctadiene and dicarbonyl), it was found that the tris(pyrazolyl)toluidine ligand (24) favored the 2 binding mode. This was not consistent with the behavior observed with analogous complexes bearing tris(pyrazolyl)alkane ligands reported previously.32-33 The influence of the counterion was explored through the synthesis and analysis of the complex [Rh(COD)(p-tpt)][BPh4] (32), where it was found that the anion has no significant influence on the coordination chemistry of the tris(pyrazolyl)toluidine ligand. To investigate the influence of the donor strength on the coordination chemistry of the tridentate ligands, the analogous tris(N-methylimidazolyl)methanol complexes

[Rh(COD)(tim)][BArF] (36) and [Rh(CO)2(tim)][BArF] (37) with the more strongly

121

binding imidazolyl ligands were synthesized and studied using x-ray crystallography and NMR spectroscopy. The resulting rhodium(I) complexes 36 and 37 also adopted a square planar geometry.

The finding that the tris(pyrazolyl)toluidine complexes (27 - 29) preferred to be bound in the 2 binding mode have implications for their catalytic activity as one would expect these complexes to behave in a similar fashion to their corresponding complexes with bidentate ligands. As part of the future investigations into these tris(pyrazolyl)toluidine complexes, the effects of substitutions on the 3 and 5 positions of pyrazolyl donors on the coordination chemistry will also be investigated.

122

5.6 References

(1) Cheung, F. K.; Clarke, A. J.; Clarkson, G. J.; Fox, D. J.; Graham, M. A.; Lin, C.; Criville, A. L.; Wills, M. Dalton Trans. 2010, 39, 1395-1402. (2) Lu, W.-J.; Chen, Y.-W.; Hou, X.-L. Adv. Synth. Catal. 2010, 352, 103-107. (3) Bai, W.-J.; Xie, J.-H.; Li, Y.-L.; Liu, S.; Zhou, Q.-L. Adv. Synth. Catal. 2010, 352, 81-84. (4) Han, S. B.; Han, H.; Krische, M. J. J. Am. Chem. Soc. 2010, 132, 1760-1761. (5) Moret, S.; Dallanegra, R.; Chaplin, A. B.; Douglas, T. M.; Hiney, R. M.; Weller, A. S. Inorg. Chim. Acta 2010, 363, 574-580. (6) Diab, L.; Scaronmejkal, T.; Geier, J.; Breit, B. Angew. Chem. Int. Ed. Engl. 2009, 48, 8022-8026. (7) Piras, I.; Jennerjahn, R.; Jackstell, R.; Baumann, W.; Spannenberg, A.; Franke, R.; Wiese, K.-D.; Beller, M. J. Orgmet. Chem. 2010, 695, 479-486. (8) Robert, T.; Abiri, Z.; Wassenaar, J.; Sandee, A. J.; Romanski, S.; Neudow $, J. r.- M.; Schmalz, H.-G. n.; Reek, J. N. H. Organometallics 2009, 29, 478-483. (9) Zhang, X.; Cao, B.; Yan, Y.; Yu, S.; Ji, B.; Zhang, X. Chem. Eur. J. 2010, 16, 871- 877. (10) Choi, K. S.; Chiu, P. F.; Chan, K. S. Organometallics 2010, 29, 624-629. (11) Arisawa, M.; Suwa, K.; Yamaguchi, M. Org. Lett. 2009, 11, 625-627. (12) Guerrero Rios, I.; Novarino, E.; van der Veer, S.; Hessen, B.; Bouwkamp, M. W. J. Am. Chem. Soc. 2009, 131, 16658-16659. (13) Ford, L.; Jahn, U. Angew. Chem. Int. Ed. Engl. 2009, 48, 6386-6389. (14) Omae, I. Appl. Organomet. Chem. 2009, 23, 91-107. (15) Hartwig, J. F. Nature 2008, 455, 314-322. (16) Takahashi, T.; Watahiki, T.; Kitazume, S.; Yasuda, H.; Sakakura, T. Chem Commun (Camb) 2006, 1664-1666. (17) McDonald, A. R.; Dijkstra, H. P.; Suijkerbuijk, B. M. J. M.; van Klink, G. P. M.; van Koten, G. Organometallics 2009, 28, 4689-4699. (18) Wang, Z.; Chen, G.; Ding, K. Chem. Rev. 2008, 109, 322-359. (19) Copéret, C.; Basset, J.-M. Adv. Synth. Catal. 2007, 349, 78-92.

123

(20) Reetz, M. T.; Rentzsch, M.; Pletsch, A.; Taglieber, A.; Hollmann, F.; Mondiere, R. J. G.; Dickmann, N.; Hoecker, B.; Cerrone, S.; Haeger, M. C.; Sterner, R. ChemBioChem 2008, 9, 552-564. (21) Reetz, M. T. Top. Organomet. Chem. 2009, 25, 63-92. (22) Bigmore, H. R.; Lawrence, S. C.; Mountford, P.; Tredget, C. S. Dalton Trans. 2005, 635-651. (23) Hückel, W.; Bretschneider, H. Ber. Dtsch. Chem. Ges. B 1937, 70B, 2024-2026. (24) Liddle, B. J.; Gardinier, J. R. J. Org. Chem. 2007, 72, 9794-9797. (25) Trofimenko, S. J. Am. Chem. Soc. 1970, 92, 5118-5126. (26) Deane, M.; Lalor, F. J. J. Orgmet. Chem. 1973, 57, C61-C62. (27) Heinekey, D. M.; Oldham, W. J.; Wiley, J. S. J. Am. Chem. Soc. 1996, 118, 12842-12843. (28) Dehmlow, E. V.; Franke, K. Liebigs Ann. Chem. 1979, 1456-1464. (29) Julia, S.; Del Mazo, J. M.; Avila, L.; Elguero, J. Org. Prep. Proced. Int. 1984, 16, 299-307. (30) Esteruelas, M. A.; Oro, L. A.; Apreda, M. C.; Foces-Foces, C.; Cano, F. H.; Claramunt, R. M.; Lopez, C.; Elguero, J.; Begtrup, M. J. Organomet. Chem. 1988, 344, 93-108. (31) Adams, C. J.; Connelly, N. G.; Emslie, D. J. H.; Hayward, O. D.; Manson, T.; Guy Orpen, A.; Rieger, P. H. Dalton Trans. 2003, 2835-2845. (32) Hallett, A. J.; Anderson, K. M.; Connelly, N. G.; Haddow, M. F. Dalton Trans. 2009, 4181-4189. (33) Klaui, W.; Schramm, D.; Peters, W.; Rheinwald, G.; Lang, H. Eur. J. Inorg. Chem. 2001, 1415-1424. (34) Dabb, S. L.; Ho, J. H. H.; Hodgson, R.; Messerle, B. A.; Wagler, J. Dalton Trans. 2009, 634-642. (35) Strauss, S. H. Chem. Rev. 1993, 93, 927-942. (36) Field, L. D.; Messerle, B. A.; Vuong, K. Q.; Turner, P. Dalton Trans. 2009, 3599- 3614. (37) Akita, M.; Ohta, K.; Takahashi, Y.; Hikichi, S.; Moro-oka, Y. Organometallics 1997, 16, 4121-4128. (38) Bucher, U. E.; Currao, A.; Nesper, R.; Rueegger, H.; Venanzi, L. M.; Younger, E. Inorg. Chem. 1995, 34, 66-74.

124

(39) Adams, C. J.; Anderson, K. M.; Charmant, J. P. H.; Connelly, N. G.; Field, B. A.; Hallett, A. J.; Horne, M. Dalton Trans. 2008, 2680-2692. (40) Northcutt, T. O.; Lachicotte, R. J.; Jones, W. D. Organometallics 1998, 17, 5148- 5152. (41) Weinberg, D. R.; Hazari, N.; Labinger, J. A.; Bercaw, J. E. Organometallics 2010, 29, 89-100. (42) Burling, S.; Field, L. D.; Messerle, B. A.; Rumble, S. L. Organometallics 2007, 26, 4335-4343.

125

Chapter 6: Tridentate Rhodium(I) Complexes: Catalytic Activity

“To do something well is so worthwhile that to die trying to do it better cannot be foolhardy. It would be a waste of life to do nothing with one’s ability, for I feel life is measured in achievements, not in years alone.”

Bruce McLaren (1937 – 1970)

126

6.1 Introduction

Small heterocyclic molecules serve as versatile building blocks for the synthesis of complex biologically active molecules. There are numerous synthetic pathways for the synthesis of heterocyclic compounds using traditional organic chemistry but recently there has been a move towards using transition metal catalysis for the synthesis of heterocyclic molecules. This is due to the atom and synthetic efficiency (i.e. minimal or no by products, high yields and mild reaction conditions) offered by the use of transition metal complexes as catalysts. Of the heterocyclic building blocks available, there is particular interest in the synthesis of alkylidene lactones due to their biological activity.

6.1.1 Alkylidene lactones

Alkylidene lactones form the core of a range of natural products (Figure 6.1) which are isolated from a large number of natural sources which possess high biological activity. Compounds such as Xerulinic acid have been shown to inhibit cholesterol biosynthesis,1 and Fimbrolides have been shown to inhibit quorum sensing in bacteria.2-3 The high biological activity of the compounds shown in Figure 6.1 is attributed to the presence of the enolester moiety, which acts as a suicide inhibitor for enzymes. Suicide inhibition is where the substrate binds to an enzyme, irreversibly removing the enzyme from the catalytic cycle (Scheme 6.1).4

127

Figure 6.1: Structures of biologically active or alkylidene lactones isolated from natural products.5

Scheme 6.1: ^ $ $ " { !\ $ enzymes (Enz-Nu) by binding irreversibly.5

Due to their high biological activity, there has been considerable interest in the synthesis of alkylidene lactones, and there are numerous synthetic strategies which can be used.6-8 The synthetic strategies can be broadly divided into three categories (Scheme 6.2) including: (1) coupling reactions at preformed lactones; (2) condensation reaction between functionalized furans and aldehydes; and (3) the cyclization of an alkenoic or alkynoic acid to form the lactone.

128

Scheme 6.2: ^| ! ! $ ! $ $ " { lactones using (a) coupling reactions involving preformed alkylidene lactones; (b) condensation reactions between functionalized furans and aldehydes; and (c) catalyzed intramolecular cyclization of alkenoic or alkynoic acids.5

In the approach involving coupling reactions for adding substituents to the preformed lactones (Scheme 6.2a), substituents are attached by coupling preformed alkylidene lactones with the desired functional group through the use of organometallic reagents (i.e. organomagnesium and organolithium reagents).9-11 While this can be used easily for non-reactive R groups, biologically active lactones shown in Figure 6.1 typically contain reactive functional groups (i.e. OH, CO2H and halides) which can interfere with the coupling reactions utilized to introduce additional substituents. As a result, protecting groups are required to protect existing reactive functional groups, which in turn increases the number of steps involved in the synthesis of the alkylidene lactones.

The second route (Scheme 6.2b) involves the use of building blocks such as aldehydes, furans or functionalized alkenes12-13 for the synthesis of the desired lactone. This

129

method allows the formation of highly functionalized lactones with commercially available starting materials provided inert R groups are used. Due to the presence of reactive R groups, protecting groups are required to avoid any undesired side reactions. The need to introduce and remove protecting groups reduces the overall carbon efficiency of the use of condensation reactions for the synthesis of the alkylidene lactones.

The last method (Scheme 6.2c) involves the intramolecular cyclization of a prefunctionalized alkynoic or alkenoic acid to form the functionalized lactone. The formation of the lactone involves a nucleophilic attack by the oxygen group of the carboxylic acid at the unsaturated bond to yield the desired lactone. This method is considered the most efficient method for lactone synthesis, as substituted alkynoic or alkenoic acids can be easily synthesized and the cyclization can proceed in the presence of functional groups which will typically prevent the use of coupling or condensation reactions. One problem with this method is that the unsaturated bonds are generally electron rich and are not susceptible to nucleophilic attack. In the literature there are a number of ways in which these unsaturated bonds are activated towards nucleophilic attack, including the use of halogens, specifically iodine.14 Another means of activating the unsaturated bond to the attack by the oxygen nucleophile is through the use of highly acidic or basic conditions.15-17 Despite the simplicity of the use of acid / base catalyzed intramolecular cyclization of alkynoic and alkenoic acids, this method is not widely utilized due to low functional group tolerance.

6.1.2 Metal catalyzed intramolecular hydroalkoxylation

The need to avoid strongly acidic and basic conditions to promote intramolecular cyclizations of alkynoic acid has led to the investigation of the use of transition metal complexes as catalysts for the synthesis of oxygen based heterocycles such as alkylidene lactones and spiroacetals.18 Metal complexes can bind to the alkyne and alkene bonds allowing nucleophilic attack by the oxygen of the carboxylic acid group onto the unsaturated bond. This is only possible due to the presence of anti-bonding

130

orbitals which withdraw electron density from the unsaturated bond, making it susceptible to nucleophilic attack.

There are a variety of organometallic complexes which can promote intramolecular cyclization via the addition of the OH group of either alcohols or carboxylic acids to unsaturated bonds. Examples include complex bearing lanthanides19-20 and transition metal ions.21-23 Despite their high activity as catalysts for the addition of oxygen nucleophiles onto unsaturated bonds, there has been a desire to find alternatives to lanthanide complexes due to their low functional group tolerance and instability of the corresponding metal complexes. Late transition metal complexes have higher functional group tolerance in comparison to lanthanide and early transition metal complexes, while maintaining the same level of catalytic activity.24 As such, late transition metal complexes have been investigated as alternative catalysts for the synthesis of functionalized heterocycles shown in Figure 6.1. Of the late transition metal catalysts, the most widely used metals include rhodium,25-27 iridium,26,28 palladium,29-31 gold,5,32-33 silver34 and copper.35 The advantages of the late transition metals are not only their high functional group tolerance, but also the mild conditions under which the cyclizations can be done. Depending on the metal centre and the nature of the substrate, the cyclization can occur in one of two ways (Scheme 6.3).

Scheme 6.3: The two possible mechanisms for the intramolecular cyclization of an alkynoic acid for: (a) substituted alkyne; and (b) unsubstituted.

131

! ! \ 4 \ ! the unsaturated bond is activated towards nucleophilic attack from the oxygen nucleophile. Two products can form, the exo-dig and the endo-dig product depending on which of the two carbons (C1 or C2) reacts with the nucleophile. For some late transition metals such as ruthenium, !4\\$!V! 6.3b). Based on the electron density of the vinylidene intermediate, only the C2 position can be attacked by the oxygen nucleophile, leading to the exclusive formation of the endo-dig product (Scheme 6.3b).

6.1.3 Gold / silver catalyzed hydroalkoxylation of alkynoic acids.

There has been considerable interest in the use of gold and silver salts as catalysts in organic synthesis due to their low cost compared to late transition metals.36 Both silver34 and gold5,37 salts have been used as catalysts for the hydroalkoxylation of alkynoic acids to yield the corresponding cyclic lactone. In the presence of silver carbonate, it was found that after one hour at 80°C, aliphatic alkynoic acids can be converted to the corresponding lactones quantitatively.34 Similarly, it was found that in the presence of 10 mol% of gold (I) chloride and a base, complete conversion of aliphatic alkynoic acids to the corresponding lactones were observed after one hour at room temperature.5 While both the gold and silver catalyst can furnish the cyclic lactone under very mild conditions in excellent yields, high catalyst loadings are required to afford these transformations. In addition to the high catalyst loading required to afford the intramolecular cyclizations of the alkynoic acids, gold and silver salts lack the selectivity observed for other transition metal catalysts. This is highlighted by the lack of E / Z selectivity for the intramolecular cyclizations of substituted aliphatic alkynoic acids promoted by the use of gold salts5 compared to late transition metal complexes.25 To achieve the same efficiency but with lower catalyst loading and higher regioselectivity, late transition metals such as rhodium(I) and iridium(I) complexes have been investigated as possible catalysts.

132

6.1.4 Rhodium / Iridium catalyzed hydroalkoxylation of alkynoic acids.

There are a number of literature examples of rhodium(I)25-27 and iridium(I) 26,28 catalyzed intramolecular cyclization of alkynoic acids and these are summarized in Table 6.1. Entries 2 – 4 are rhodium(I) complexes developed within the Messerle research group. Metal complexes bearing imidazolyl, indoyl and azolyl donors (Table 6.1, entries 1 to 4) gave good catalytic activity with the best catalyst being the indole based catalyst (Table 6.1, entry 2a) where complete conversion was observed in 3 hours at 60oC using 2 mol% catalyst. It is interesting to note that with 0.1 mol% catalyst, the imidazole based complexes (Table 6.1, entries 3 and 4) were able to afford complete conversion at 50oC in 15 hours. While it is considerably longer in duration compared to the gold / silver salts (Table 6.1, entries 5 and 6) systems, the low catalyst loading of 0.1% should be noted. Based on these promising results, the catalytic activity of the new rhodium(I) complexes bearing bis / tris(N-methylimidazolyl)alkane and bis / tris(pyrazolyl)alkane complexes discussed in the previous chapter was investigated.

133

Table 6.1: Catalytic data for the intramolecular cyclization of 4-pentynoic acid (38) acid using rhodium(I) and iridium(I) based complexes.

Mol% Solvent, Time, Entry Catalyst metal Temperature Yield

8 h 1a26 1 CD CN, 50oC 3 (99%)

3 h 1b26 1 CD CN, 50oC 3 (99%)

3 h 2a27 2 CDCl , 60oC 3 (99%)

9.9 h 2b27 2 CDCl , 60oC 3 (73%)

d -acetone, 15 h 325 0.1 6 50oC (>99%)

d -acetone, 15 h 425 0.1 6 50oC (>99%)

CD CN, 2 h 55 AuCl + 10 mol% K CO 10 3 3 2 3 20oC (96%) Toluene 1 h 634 Ag CO 10 2 3 80oC (99%)

134

6.2 Catalyzed intramolecular hydroalkoxylation of aliphatic alkynoic acid

6.2.1 Procedure

All catalyzed intramolecular cyclization reactions of alkynoic acids were performed on a small scale (25 mg) in an NMR tube fitted with a concentric Young’s Teflon valve under an inert atmosphere (nitrogen or argon). The samples were heated within the NMR spectrometer, where the temperature was calibrated using a K-type thermocouple in ethylene glycol and the reaction progress was monitored using 1H NMR spectroscopy. The conversion was determined by comparing the integration of the appropriate resonances of the starting material relative to the resonances of the product in the 1H NMR spectrum taken at a given time point. Timing of the catalysis was started when the solvent was added to the solid mixture of the complex and substrate. Unless otherwise stated, all values quoted are the result of a single experiment.

6.2.2 Preliminary catalyst and solvent screening

Elgafi et al.25 has investigated the intramolecular cyclization of pentynoic acid (38) using complexes with imidazolyl donors such as bis(imidazolyl)methane, but have not investigated complexes with the more labile pyrazolyl donors such as tris(pyrazolyl)toluidine ligand (24).

Preliminary investigations of the catalytic activity of the complexes

[Rh(CO)2(p-tpt)][BArF] (29a) and [Rh(CO)2(o-tpt)][BArF] (29b) show both complexes to be highly active catalysts for the intramolecular cyclization of 4-pentynoic acid (38), forming the lactone 39. Despite the high activity of the complex 29b as catalyst for the intramolecular cyclization of 4-pentynoic acid (38), the resulting 1H NMR spectrum of the crude reaction mixture showed the formation of a number of unidentified by- products. As has been reported for all examples5, 25-26 of metal catalyzed intramolecular

135

cyclization of 4-pentynoic (38) acid in literature, only the exo-dig product (39) was observed in the reaction mixture when the complexes [Rh(CO)2(p-tpt)][BArF] (29a) and

[Rh(CO)2(o-tpt)][BArF] (29b) were used to promote the intramolecular cyclization of 4- pentynoic acid (38).

To optimize the reaction conditions, 2 mol% of the complex [Rh(CO)2(p-tpt)][BArF] (29a) was heated with 4-pentynoic acid (38) in a series of different solvents and at a range of different temperatures. The progress of the reaction was monitored using 1H NMR spectroscopy. The percentage conversion was determined by comparing the relative integrals of the resonances due to the starting material and the product. The results for the various solvents and conditions are summarized in Table 6.2. From the results summarized in Table 6.2, the aromatic solvents yielded the highest catalytic efficiency as complete conversion was observed after 1.7 hours and 3.3 hours for d6- o benzene and d8-toluene respectively at 80 C. This is consistent with literature, where it was reported that the metal complexes bearing the BArF counterion have the highest catalytic activity in aromatic solvents.38 One of the advantages that toluene has as a solvent in place of benzene is the higher temperatures to which the sample can be heated. The intramolecular cyclization of 4-pentynoic acid (38) was repeated at 110oC in toluene, with 2 mol% of the complex [Rh(CO)2(p-tpt)][BArF] (29a) as the catalyst, an increase in the rate of reaction was observed, with complete loss of starting material observed after 1.2 hours. Comparing the products formed in the metal catalyzed cyclization of 4-pentynoic acid (38) using the complex 29a as catalyst in benzene and toluene at 80oC, it was found that significant amounts of impurity was formed in toluene and not in benzene.

136

Table 6.2: Catalytic data for the intramolecular cyclization of 4-pentynoic acid (38) in the presence of 2 mol% [Rh(CO)2(p-tpt)][BArF] (29a) in a range of solvents and temperature.

Solvent Temperature Time, Conversion

o d6-benzene 80 C 1.7 h (99%)

80oC 2.85 h (95%)

d8-toluene 110oCa 1.3 h (78%)

o d6-acetone 60 C 19 h (81%)

d-chloroform 60oCa 17 h (47%) a Sample was heated in an oil bath held at the indicated temperature and the 1H NMR spectrum of the sample was acquired after a set time period.

6.2.3 Intramolecular cyclization of 4-pentynoic acid using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors.

The activity of two imidazole based complexes, [Rh(CO)2(bim)][BArF] (40) and

[Rh(CO)2(tim)][BArF] (37) as catalysts for the intramolecular cyclization of 4- pentynoic acid (38) was examined. The progress of the reaction was determined by recording the 1H NMR spectrum of the reaction mixture at regular time intervals. A series of 1H NMR spectra showing the progress of a typical intramolecular cyclization of pentynoic acid (38) is shown below in Figure 6.3. The progress of the reaction was determined through the relative integration of the Ha proton against the alkene proton (Hf) on the lactone. The time course for the cyclization of 4-pentynoic acid (38) 137

showing the catalytic activity of the complexes [Rh(CO)2(tim)][BArF] (37),

[Rh(CO)2(bim)][BArF] (40) and [Rh(CO)2(p-tpt)][BArF] (29a) is shown in Figure 6.2.

Comparing the percentage conversion of the substrate as a function of time for the complexes 29a, 37 and 40, the imidazole-based complexes 37 and 40 are more efficient as catalysts for the intramolecular cyclization of 4-pentynoic acid (38) compared to the tris(pyrazolyl)toludine complex (29a) with complete conversion of 4-pentynoic acid (38) to the lactone 39 observed after ca. 1.4hrs, 0.5 hrs and 0.3 hrs for the complexes 29a, 37 and 40 respectively. It is interesting to note the difference in the reactivity between the bis(N-methylimidazolyl)methane (40) and the tris(imidazolyl)methanol complex (37), despite the fact that the ligands act as a bidentate ligand in both cases and they are structurally similar as described in Chapter 5. This difference in observed reactivity could be due to the free imidazolyl donor which causes steric hindrance between the substrate and the active site leading to reduced catalytic activity in the case of the complex 37.

Figure 6.2: Plot of the percentage conversion vs. time for the intramolecular cyclization o of 4-pentynoic acid to the lactone 39 using in d6-benzene at 80 C using

2 mol% of [Rh(CO)2(p-tpt)][BArF] (29a) (L), [Rh(CO)2(bim)][BArF] (40) ([) and

[Rh(CO)2(tim)][BArF] (37) (J). 138

Hb + Hc (a) t = 1 min

(b) t = 8 min

(a) t = 34 min

Hd Hf Hf He

Figure 6.3: 1H NMR spectrum for the intramolecular cyclization of 4-pentynoic acid o (38) catalyzed by 2 mol% [Rh(CO)2(tim)][BArF] (37) in C6D6 at 80 C at: (a) t = 1 min; (b) t = 8 mins; (c) t = 21 mins; and (d) t = 34 mins.

139

Table 6.3: Comparison of catalytic activity of rhodium complexes and literature complexes for the intramolecular cyclization of 4-pentynoic acid (38) to the lactone 39.

Temperature, Entry Catalyst Mol % metal Time solvent

a o 1 [Rh(CO)2(bim)][BArF] (40) 2 C6D6, 80 C 0.3 h

o 2 [Rh(CO)2(tim)][BArF] (37) 2 C6D6, 80 C 0.5 h

o 3 [Rh(CO)2(p-tpt)][BArF] (29a) 2 C6D6, 80 C 1.4 h

26 4 1 CD3CN, 50oC 8 h

27 o 5 2 CDCl3, 60 C 3 h

5 o 6 AuCl + 10 mol% K2CO3 10 CD3CN, 24 C 2 h a The slowest completion time of the duplicate runs are reported and used for discussion.

A comparison of the catalytic activity of the complexes 29a, 37 and 40 with literature reported complexes for the intramolecular cyclization of 4-pentynoic acid (38) is shown in Table 6.3. From the results shown in Table 6.3, it is clear that the imidazole based rhodium(I) complexes 37 and 40 are highly active as catalysts for the intramolecular cyclization 140

4-pentynoic acid (38), with complete conversion of the substrate observed in under an hour with 2 mol% catalyst. Despite the higher temperatures required, it is still an improvement over the neutral bimetallic carbene complex (Table 6.3, entry 4) and the neutral indole based complex (Table 6.3, entry 5). While the use of gold (I) chloride can afford the complete conversion of the substrate in milder conditions, it does require the use of significantly higher catalyst loading (10 mol%).

6.2.4 Intramolecular cyclization of 5-hexynoic acid using Rhodium(I) complexes bearing imidazolyl donors.

Based on the success of using the imidazole based ligands for the cyclization of 4-pentynoic acid (38), the catalyzed intramolecular cyclization of 5-hexynoic acid (41) was investigated using the complex 40 as the catalyst. The increase in the chain length of 5-hexynoic acid (41) compared to 4-pentynoic acid (38) leads to increase in the energy required to afford the intramolecular cyclization of 5-hexynoic acid (41).

The cyclization of 5-hexynoic acid (41) was attempted at 70°C in d6-benzene in the 1 presence of 1 mol% of the complex [Rh(CO)2(bim)][BArF] (40) as catalyst and the H NMR spectra showing the progress of the reaction are presented in Figure 6.4. Only one product was observed, and a comparison with literature spectroscopic data confirmed the formation of the exo-dig product (41). Based on the disappearance of the starting material, complete conversion was observed after 14 hours. A summary of the results is shown in Table 6.4.

Comparing the catalytic activity of reported rhodium(I) complexes

(Table 6.4, entries 2 – 4) to the complex [Rh(CO)2(bim)][BArF] (40) shows that 40 is a highly active catalyst for the intramolecular cyclization of 5-hexynoic acid (41). Complete conversion of 5-hexynoic acid (41) to the lactone 42 was observed after 13.3 hours in the presence of the complex [Rh(CO)2(bim)][BArF] (40) which is significantly faster than previously published25,26,39 rhodium(I) complexes (Table 6.4, entries 2 – 4),

141

where up to 192 hours were required for 80% conversion. Even taking into account the difference in temperatures, the BArF analogue (Table 6.4, entry 1) was more active as a catalyst for the intramolecular cyclization of 5-hexynoic acid (41) compared to the BPh4 analogue (Table 6.4, entry 2).

(a) t = 0.2 h Hb Hd Hc Ha

(b) t = 0.8 h

(d) t = 13.3 h g h H h H H He Hf

Figure 6.4: 1H NMR spectrum for the intramolecular cyclization of 4-hexynoic acid o (41) catalyzed by 1 mol% [Rh(bim)(CO)2][BArF] (40) in C6D6 at 70 C at: (a) t = 0.2 hrs; (b) t = 0.8 hrs; (c) t = 2.2 hrs; and (d) t = 13.3 hrs

142

Table 6.4: Comparison of literature catalytic activity of rhodium complexes for the intramolecular cyclization of 5-hexynoic acid (41) to the lactone 42.

Temperature, Entry Catalyst Mol % metal solvent Time

[Rh(CO)2(bim)][BArF] (40) o 1 1 C6D6, 70 C 13.3 h

192 h d -acetone, 225 [Rh(CO) (bim)][BPh4] 0.7 6 2 50oC (80%)

CD3CN, 326 1 192 h 50oc

d6-acetone 168 h 439 0.5 50oC (15%)

The isomerization of the alkene bond is a reaction that can compete with the intramolecular cyclization of aminoalkenes when using a rhodium catalyst to promote the cyclization.40 Examining the 1H NMR spectra of the reaction mixture after completion of the intramolecular cyclization of 5-hexynoic acid (41) using the complex 1 [Rh(CO)2(bim)][BArF] (40), no peaks were observed in the H NMR spectrum that indicated that alkene isomerization had occurred.

6.3 Intramolecular hydroalkoxylation of aromatic alkynoic acid

143

Having studied the catalytic activity of the rhodium(I) complexes bearing imidazolyl donor ligands for the intramolecular cyclization of aliphatic alkynoic acids, the next step is to consider the reactivity of these rhodium(I) complexes for the cyclization of aromatic alkynoic acids.

6.3.1 Synthesis of aromatic alkynoic acid substrates

The route used for the synthesis of terminal and non-terminal aromatic alkynoic acid is shown below in Scheme 6.4. The synthesis of the aromatic alkynoic acid 44a-c involves the palladium catalyzed coupling between iodo-2-methylbenzoate with the desired alkyne to form the aromatic alkynoic ester (43). The ester is then deprotected under basic conditions to yield the corresponding aromatic alkynoic acid (44a-c).

Scheme 6.4: The synthetic scheme for the synthesis of aromatic alkynoic acids bearing unsubstituted (44a) and substituted (44b-c) terminal alkynes.

6.3.2 Intramolecular cyclization of 2-ethynylbenzoic acid using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors.

To examine the relative catalytic activity of rhodium(I) complexes bearing pyrazolyl and imidazolyl donors as catalysts for the intramolecular cyclization of aromatic alkynoic acids, 2-ethynylbenzoic acid (44a) was treated with 1 mol% of the appropriate o 1 catalyst in d6-benzene at 60 C, and the progress of the reaction monitored via H NMR spectroscopy. 1H NMR spectra for a typical intramolecular cyclization of 2- ethynylbenzoic acid (44a) in the presence of a rhodium(I) catalyst are shown in Figure 6.5. Unlike the intramolecular cyclization of aliphatic alkynoic acid (i.e. 4-pentynoic acid), where only the exo-dig product was formed, both the endo-dig (45a) and exo-dig

144

(46a) products were observed in the reaction mixture for the intramolecular cyclization of 2-ethynylbenzoic acid (44a) using the complexes [Rh(CO)2(bim)][BArF] (40),

[Rh(CO)2(tim)][BArF] (37) and [Rh(CO)2(p-tpt)][BArF] (29a).

(a) t = 3 min Ha

(b) t = 9 min

Hd Hd

(d) t = 25 min

Hb Hb

Figure 6.5: 1H NMR spectrum for the intramolecular cyclization of 2-ethynylbenzoic o acid (44a) catalyzed by 1 mol% [Rh(CO)2(bim)][BArF] (40) in C6D6 at 60 C at: (a) t = 3 mins; (b) t = 9 mins; (c) t = 20 mins; and (d) t = 25 mins.

145

The percentage conversion of 2-ethynylbenzoic acid (44a) to the lactones 45a and 46a was determined from the NMR spectrum using the integration of the resonances attributed to the endo-dig product (Hb at 5.8 ppm) and the exo-dig product (Hd at 4.7 ppm) relative to the alkyne proton of the starting material (Ha at 3 ppm). The reported conversion given in Figure 6.6 is the total conversion (that is the sum of the conversions of the exo-dig and endo-dig products combined).

Figure 6.6: Plot of the percentage conversion vs. time for the intramolecular cyclization of 2-ethynylbenzoic acid (44a) to the lactones 45a and 46a using in d6-benzene using o o 1 mol% of [Rh(CO)2(p-tpt)][BArF] at 60 C and 25 C (L and ), o o [Rh(CO)2(bim)][BArF] at 60 C ([) and [Rh(CO)2(tim)][BArF] at 60 C (J).

Comparing the catalyst efficiencies found for the intramolecular cyclization of 2-ethynylbenzoic acid (44a) to those for 4-pentynoic acid (38), a reversal in the activity o of the complexes was observed. At 60 C, the complex [Rh(CO)2(p-tpt)][BArF] (29a) catalyzed the intramolecular cyclization of 2-ethynylbenzoic acid (44a) to the lactones

45a and 46a in under 0.1 hours, while the imidazole complexes [Rh(CO)2(bim)][BArF]

(40) and [Rh(CO)2(tim)][BArF] (37) required 0.4 hours and 0.67 hours respectively. 146

This shows that complexes bearing imidazolyl donors 37 and 40 are less efficient for the cyclization of aromatic alkynoic acids such as 2-ethynylbenzoic acid (44a) compared to the complex 29a. This is in contrast to the cyclization of the aliphatic alkynoic acid such as 4-pentynoic acid (38), where the imidazolyl based complexes 37 and 40 were more active to the complex bearing pyrazolyl donors 29a. The intramolecular cyclization of 2-ethynylbenzoic acid (44a) using the complex 29a was repeated at room temperature and complete conversion to the lactones 45a and 45b was observed after 1.4 hours. The results for the intramolecular cyclization of 2-ethynylbenzoic acid (44a) to the lactones 45a and 46a are summarized below in Table 6.5.

On promoting the cyclization of 2-ethynylbenzoic acid (44a), the complex

[Rh(CO)2(bim)][BArF] (40) (Table 6.5, entry 1) preferentially catalyzed the formation of the exo-dig product, with a ratio of 1 : 0.3 (exo : endo) observed at the end of the reaction. It is important to note that continued heating of the sample after the complete consumption of the substrate does not alter the ratio of the products observed in the 1H

NMR spectrum. When the complex [Rh(CO)2(tim)][BArF] (37) (Table 6.5, entry 2) was used as the catalyst for the intramolecular cyclization of 2-ethynylbenzoic acid (44a), the ratio of the exo-dig : endo-dig product was 1 : 0.02. Given that the structures of the complexes [Rh(CO)2(tim)][BArF] (37) and [Rh(CO)2(bim)][BArF] (40) differ only by the presence of the extra imidazolyl donor, which was known to not bind to the rhodium centre, this change in selectivity could be attributed to the steric interaction of the unbound imidazolyl donor with the substrate. In the case of the complex o o [Rh(CO)2(p-tpt)][BArF] (29a) at both 60 C (Table 6.5, entry 3a) and 25 C (Table 6.5, entry 3b) the ratio of the products remains at 1 : 0.2 (exo : endo) which indicates that the reaction temperature has no influence on the exo : endo selectivity.

A comparison of the catalytic activities reported here to those reported in literature highlights the increased activity of the complexes 29a, 37 and 40 as catalysts for the cyclization of aromatic alkynoic acids. In the case of the imidazolyl rhodium(I) complexes (Table 6.5, entries 1 and 2), both are able to promote complete conversion of o the substrate in under one hour in d6-benzene at 60 C. In the case where the complex

[Rh(CO)2(p-tpt)][BArF] (29a) was used as the catalyst in d6-benzene, complete

147

conversion of the alkynoic acid 44a to the lactones 45a and 46a was observed after 0.1 hours at 60oC, and 1.4 hours at room temperature. This is a significant improvement compared to the intramolecular cyclization of 2-ethynylbenzoic acid (44a) promoted by gold salts reported in the literature,37 where a significantly higher catalyst loading (10 mol %) was required. With respect to selectivity, the gold catalyst produces a 1 : 1 mixture of the exo-dig and endo-dig product while the rhodium(I) catalyst shows selectivity towards the exo-dig product, especially in the case of 37 (Table 6.5, entry 2).

Table 6.5: Comparison of literature catalytic activity of rhodium(I) and gold(I) catalysts for the intramolecular cyclization of 2-ethynylbeznoic acid (44a) to the cyclic lactones 45a and 46a.a

Mol % Temperature, Time Selectivity Entry Catalyst metal solvent (hr) (exo : endo)

[Rh(CO)2(bim)][BArF] o 1 1 mol% C6D6, 60 C 0.42 1 : 0.31 (40)

[Rh(CO)2(tim)][BArF] 2 o 1 mol% C6D6, 60 C 0.67 1 : 0.01 (37)

3a o C6D6, 60 C 0.10 1 : 0.18 [Rh(CO)2(p-tpt)][BArF] 1 mol% 3b (29a) o C6D6, 25 C 1.40 1 : 0.17

437 o AuCl + 10 mol% K2CO3 10 mol% CD3CN, 25 C 2.00 1 : 1 a The slowest completion time of the duplicate runs are reported and used for discussion

148

6.3.3 Intramolecular cyclization of non-terminal aromatic alkynoic acids using Rhodium(I) complexes bearing imidazolyl and pyrazolyl donors.

The complexes [Rh(CO)2(bim)][BArF] (40), [Rh(CO)2(tim)][BArF] (37) and

[Rh(CO)2(p-tpt)][BArF] (29a) were examined as catalysts for the intramolecular cyclization of 2-(phenylethynyl)benzoic acid (44b) and the results are summarized in Table 6.6. Comparing the activity of the complexes with bidentate (Table 6.6, entries 1 and 3) and the tridentate ligands (Table 6.6, entries 2 and 4) as catalysts for the cyclization of 2-(phenylethynyl)benzoic acid (44b) shows that the presence of the extra unbound donor in the complexes induces a reduction in the reactivity despite the structural similarities at the metal centre as discussed in Chapter 5. When the complex 38 [Rh(CO)2(bpm)][BArF] (32) (Table 6.6, entry 3) is used as the catalyst to promote the cyclization of 2-(phenylethynyl)benzoic acid (44b), complete conversion was observed after 0.2 hours. When the tridentate analogue, [Rh(CO)2(p-tpt)][BArF] (29a) (Table 6.6, entry 4) is used as the catalyst for the cyclization of 2-(phenylethynyl)benzoic acid (44b), complete conversion was observed after 1.25 hours. A similar trend was observed when the imidazolyl complexes [Rh(CO)2(bim)][BArF] (40) (Table 6.6, entry 1) and [Rh(CO)2(tim)][BArF] (37) (Table 6.6, entry 2) were used as catalysts to promote the intramolecular cyclization of 2-(phenylethynyl)benzoic acid (44b). One possible explanation for this change in reactivity could be explained by the steric hindrance between the third unbound pyrazolyl donor of the tridentate ligand and the non-terminal alkynoic acid substrate.

The ratio of the exo-dig : endo-dig products formed during the intramolecular cyclization of 2-(phenylethynyl)benzoic acid (44b) using both rhodium(I) imidazolyl (37 and 40) and pyrazolyl complexes (29a and 32) as catalysts is shown in Table 6.6. A preference for the formation of the endo-dig product is observed when both types of catalysts are used. This is in contrast to the intramolecular cyclization of the terminal aromatic alkynoic acid 44a (Table 6.5) where the rhodium(I) complexes 29a, 37 and 40 all favored the formation of the exo-dig product. The change in the exo / endo selectivity of the catalysts in the intramolecular cyclization of the non-terminal alkynoic acids could be due to changes in the electronic nature of the substrate. To examine the

149

influence of substituents at the terminal alkyne position on aromatic alkynoic acids, the intramolecular cyclization of 2-(heptynyl)benzoic acid (44c) (Table 6.6, entries 1 and 2) using the complexes 37 and 40 was investigated. In the case of the intramolecular cyclization of 2-(heptynyl)benzoic acid (44c) catalyzed by the complexes 37 and 40, the endo-dig product (45c) was formed exclusively. To rationalize the changes in the selectivity of the product formation in the rhodium catalyzed intramolecular cyclization of aromatic alkynoic acids, a detailed examination of the mechanistic cycle was required.

Time Selectivity Time Selectivity Entry Catalyst R R (hrs)a 45 : 46 (hrs)a 45 : 46

[Rh(CO) (bim)] 1 2 Ph 0.2 9 : 1 C H 0.27 [BArF] (40) 5 11 1 : 0 [Rh(CO) (tim)] 2 2 Ph 6.0 2: 1 C H 1.28 [BArF] (37) 5 11 1 : 0 [Rh(CO) (bpm)] 3 2 Ph 0.15 24 : 1 - - [BArF] (32) - [Rh(CO) (p-tpt)] 4 2 Ph 1.25 22 : 1 - - [BArF] (29a) - 10 mol% AuCl + 5 10 mol% K2CO3, Ph 2.00 0.1 : 1 - - o - CD3CN, 25 C a Time required to achieve conversion of greater than 99% based on the relative integration of the product relative to the starting material.

150

6.3.4 Exo-dig : Endo-dig selectivity: Aliphatic vs. aromatic terminal alkyne substituents

A proposed mechanism for the intramolecular cyclization of terminal alkynoic acids based on the work by Messerle et al.23 is shown in Figure 6.7. In the proposed mechanism, the substrate binds by electrophilic binding of the alkyne to the metal centre, displacing one of the carbonyl co-ligands. The mechanism can follow two routes, where the oxygen nucleophile can attack at two possible sites, either the internal alkyne carbon (C1) or the terminal alkyne carbon (C2). Since both products are allowed by Baldwin’s rule,41 both products of the two mechanistic pathways can be formed.

(a)

(b) exo-dig transition state (c) endo-dig transition state

151

Based on the results for the intramolecular cyclization of non-terminal alkynoic acids shown above in section 6.3.3, the nature of the substituents in the terminal alkyne position can influence the exo-dig : endo-dig selectivity. The changes in the selectivity of the intramolecular cyclization of non-terminal aromatic alkynoic acids can be explained by considering the transition states for the formation of the exo-dig and endo- dig product (Figure 6.7a and 6.7b). The ratio of the exo-dig to the endo-dig product is determined by the relative stability of the exo-dig transition state (Figure 6.7a) compared to the endo-dig transition state (Figure 6.7b). The relative stability of the transition states are determined by the differences in the electron densities of the internal alkyne carbon (C1) and the terminal alkyne carbon (C2). In summary, when the electron density on the internal carbon (C1) is higher relative to the terminal carbon (C2), the partial positive charge generated by the nucleophilic attack by the oxygen nucleophile onto C1 carbon is better stabilized which results in a preference for the exo- dig product. When the C2 carbon has a higher electron density relative to the C1 carbon, the nucleophilic attack onto the C2 carbon is favored, leading to higher proportions of the endo-dig product. The relative electron density of the C1 and C2 carbon is determined by the nature of the substituents in the non-terminal alkyne position.

Table 6.7: Summary of the exo : endo selectivities for the intramolecular cyclization of terminal and non-terminal alkynoic acids using the complex 40 as catalyst. Entry R exo : endo 1 H 3.2 : 1

2 C5H11 0.0 : 1 3 Ph 0.1 : 1 endo-dig exo-dig

In the case where the substituents at the terminal alkyne position is a hydrogen atom (Figure 6.7b), the internal carbon (C1) has higher electron density compared to the terminal carbon (C2) due to the presence of an aromatic ring adjacent to C1. As a result of the increased electron density at C1 a high exo-dig selectivity would be expected, which is consistent with the results observed (Table 6.7, entry 1) for the intramolecular cyclization of both 2-(ethynyl)benzoic acid (44a) using rhodium(I) imidazolyl and

152

pyrazolyl complexes (29a, 37 and 40). In the case of the 2-(heptynyl)benzoic acid (44c), the alkyl chain is better at stabilizing the partial positive charge compared to an aryl group with an electron withdrawing group, which would lead to increased endo-dig selectivity. The intramolecular cyclization of 2-(heptynyl)benzoic acid catalyzed by the complex 40 showed exclusive endo-dig selectivity (Table 6.7, entry 2) which was consistent with mechanism proposed in Figure 6.7. In the case of 2-(phenylethynyl)benzoic acid (44b), the internal carbon (C1) is stabilized by a phenyl ring that is substituted with a carboxylic acid which is an electron withdrawing group. As a result, the internal alkyne carbon (C1) has less electron density compared to the terminal alkyne carbon (C2), which makes the formation of the endo-dig product more favorable relative to the exo-dig product. This is consistent with the results observed (Table 6.7, entry 3) with the rhodium catalyzed intramolecular cyclization of 2-(phenylethynyl)benzoic acid (44b).

6.4 Exo-dig : Endo-dig selectivity: Electron donating vs. electron withdrawing terminal alkyne substituents

In order to examine the effects of electron density of the aryl ring at the terminal alkyne position on the regioselectivity of the intramolecular cyclization of non-terminal aromatic alkynoic acids, a series of electron withdrawing and electron donating substituents were introduced at the terminal alkyne position. The intramolecular cyclization of these electron deficient or electron rich non-terminal aromatic alkynoic acid was investigated using the complex [Rh(CO)2(bim)][BArF] (40) as the catalyst. The synthesis of the substrates bearing the electron donating and withdrawing groups was undertaken by Astrid Knuhtsen.

6.4.1 Synthesis of non-terminal aromatic alkynoic acids with electron rich or electron deficient terminal substituents

153

The general method for the synthesis of aromatic alkynoic acids with electron rich and electron deficient substituents at the terminal alkyne position is shown in Scheme 6.5. The synthesis involved the deprotection of methyl-2-(trimethylsilylethynyl)benzoate (47) to yield Methyl-2-ethynylbenzoate (48), which was coupled to the desired electron rich or electron deficient aryl halide to yield the desired substituted aromatic alkynoic ester (49a – d). Deprotection of the esters (49a – d) was achieved by stirring in an aqueous methanol sodium hydroxide solution to yield the terminally substituted aromatic alkynoic acid (50a – d).

Scheme 6.5: The synthetic scheme for the synthesis of aromatic alkynoic acids with electron withdrawing and electron donating substituents at the terminal alkyne position.

6.4.2 Electronic properties of electron donating and withdrawing substrates.

To examine the changes in the electronic nature of the alkynoic acids (50a – d) induced by changing the substituent at the terminal alkyne position, the infrared stretching frequencies of the carboxylate groups, and the 13C chemical shifts of the alkyne carbons were determined (Table 6.8). The alkyne carbon chemical shift, the IR stretching $!$!{49a – d) ! \ \ "" V V ! \ donating, while a positive value implies that the substituent is electron withdrawing.

154

Table 6.8: The corresponding Hammett constant, carbonyl stretching frequency and the selected 13C chemical shifts of the substituted aromatic alkynoic acid ester.

13C Chemical shift (ppm) C=O stretching Entry R -1 frequency (cm ) C1 C2 C1 – C2) 1 OMe -0.268 1775 86.9 94.5 -7.6

2 Me -0.170 1731 87.7 94.7 -7.0

3 Cl 0.227 1733 89.3 92.9 -3.6

4 COMe 0.500 ---- a 91.5 93.3 -1.8 a the C=O stretching frequency of the non-terminal alkynoic ester 49d was not measured.

Table 6.8 shows that as the electronic properties of the terminal substituent is varied, there are corresponding changes in the both the 13C chemical shifts of the alkyne carbons (C1 and C2) and the C=O stretching frequency. The alkyne carbon 13C chemical shifts shows that as the electron withdrawing character of the terminal alkyne substituent increases (i.e. !!!V electron densities of the alkyne carbons C1 and C2. As increases, the electron density of alkyne carbon C2 increases as indicated by the upfield movement of the 13C chemical shift. At the same time, the electron density of the alkyne carbon C1 decreases as indicated by the downfield movement of the 13C chemical shift. The observations in Table 6.8 are consistent with other studies presented in literature with phenylacetylene derivatives.42 In all cases, the 13C chemical shift of the C2 carbon was always greater than C1, implying that the polarity of the alkyne bond does not change. With respect to the C=O stretching frequency, as the electron withdrawing ability of the substituent increases (i.e. !!~! frequency. This shows that as the electronic nature of the substituent at the terminal

155

alkyne position is varied, there was a corresponding change of the electron density in the carboxylate group.

Based on the analysis of the electronic properties of the alkynoic acids (50a – d), the increasing electron density at the alkyne carbon C2 relative to C1 with the increasing would favor the formation of the endo-dig product for the intramolecular cyclization of non-terminal alkynoic acid (50). This is due to the increased ability of the alkyne carbon C2 to stabilize the partial positive charge that is generated during the transition state. The modified alkynoic acids (50a–d) were treated with 1 mol% of the complex o [Rh (CO)2(bim)][BArF] (40) as the catalyst in d6-benzene at 60 C and the progress of the reaction monitored using 1H NMR spectroscopy and the results summarized below in Table 6.9. The intramolecular cyclization of alkynoic acids 50a-d (Table 6.9, entries 1 - 4) showed high endo-dig selectivity as predicted by the 13C chemical shifts of the alkyne carbons.

An increase in the reaction rate was observed with an increase in the electron withdrawing ability of the substituent on the terminal alkyne position. This is consistent with an increase in the susceptibility of the alkyne to nucleophilic attack. The relative changes of electron density at C1 and C2 were too small for substrates 50a-d to result in a change in the exo-dig : endo-dig selectivity.

156

Table 6.9: Intramolecular cyclization of electron rich and electron deficient non- terminal aromatic alkynoic acids (50a – d) using the complex [Rh(CO)2(bim)][BArF] o (40) in C6D6 at 60 C

Time Selectivity Entry Substrate Hammett Constant (hours) (51 : 52) 3.67 1 OMe -0.268 1 : 0.12 ( >90%) 5.68 2 Me -0.170 1 : 0 ( >90%) 2.13 3 Cl 0.227 1 : 0.09 ( >95%)

4a COMe 0.500 0.25 1 : 0.18 a The slowest completion time of the duplicate runs are reported and used for discussion.

6.5 Conclusion

The activity of bidentate rhodium(I) complex [Rh(CO)2(bim)][BArF] (40) and tridentate rhodium(I) complexes such as [Rh(CO)2(p-tpt)][BArF] (29a), [Rh(CO)2(o-tpt)][BArF]

(29b) and [Rh(CO)2(tim)][BArF] (37) as catalysts for the intramolecular cyclization of 4-pentynoic acid (38) were investigated. Under the optimized conditions, it was found that the complexes bearing imidazolyl donors 37 and 40 were highly active catalysts compared to literature reported complexes.5,25-27 The complexes bearing pyrazolyl donors 29a and 29b proved to be less reactive as catalysts compared to the complexes bearing imidazolyl donors 37 and 40 for the intramolecular cyclization of 4-pentynoic 157

acid (38). The complex 40 was used to promote the intramolecular cyclization of 5- hexynoic acid (41) and was a significantly better catalyst than those reported previously in literature.25-26,39

The reactivity and selectivity of the complexes 29a, 37 and 40 as catalysts for the intramolecular cyclization of terminal and non-terminal aromatic alkynoic acids was investigated. It was found that the tridentate pyrazolyl complex 29a was highly active as a catalyst for the intramolecular cyclization of the terminal aromatic alkynoic acid 44a. The activity of the complex 29a was reduced when substituents are present in the terminal alkyne position (44b-c). All complexes 29, 37, and 40 when tested for their activity as catalysts for the intramolecular cyclization of aromatic alkynoic acids showed a high exo-dig : endo-dig selectivity compared to gold salts previously reported.37 It is proposed that the electronic nature of the substituent at the terminal alkyne position has a significant influence in the exo-dig : endo-dig selectivity. A Hammett study, where the electronic property of substrates was systematically altered was undertaken to examine the effects of electron donating and withdrawing groups on the regioselectivity and efficiency of the intramolecular cyclization of aromatic alkynoic acids. The observed exo-dig : endo-dig selectivities for the intramolecular cyclization of the modified alkynoic acids (50a – d) were consistent with the results predicted based on an analysis of 13C chemical shifts of the alkyne carbons. The rate of intramolecular cyclization of non-terminal aromatic alkynoic acids increased as the electron withdrawing ability of the substituent at the terminal alkyne position increased.

158

6.6 References

(1) Kuhnt, D.; Anke, T.; Besl, H.; Bross, M.; Herrmann, R.; Mocek, U.; Steffan, B.; Steglich, W. J. Antibiot. 1990, 43, 1413-1420. (2) Kazlauskas, R.; Murphy, P. T.; Quinn, R. J.; Wells, R. J. Tet. Lett. 1977, 37-40. (3) Suga, H.; Smith, K. M. Curr. Opin. Chem. Biol. 2003, 7, 586-591. (4) Konaklieva, M. I.; Plotkin, B. J. Mini-Rev. Med. Chem. 2005, 5, 73-95. (5) Harkat, H.; Dembele, A. Y.; Weibel, J.-M.; Blanc, A.; Pale, P. Tetrahedron 2009, 65, 1871-1879. (6) Álvarez-Corral, M. r.; Muñoz-Dorado, M.; Rodríguez-García, I. Chem. Rev. 2008, 108, 3174-3198. (7) Bruckner, R. Curr. Org. Chem. 2001, 5, 679-718. (8) Knight, D. W. Contemp. Org. Synth. 1994, 1, 287-315. (9) Scheiper, B.; Bonnekessel, M.; Krause, H.; Furstner, A. J. Org. Chem. 2004, 69, 3943-3949. (10) Bellina, F.; Anselmi, C.; Rossi, R. Tet. Lett. 2002, 43, 2023-2027. (11) Castulík, J.; Mazal, C. Tetrahedron Letters 2000, 41, 2741-2744. (12) Lee, K. Y.; Kim, J. M.; Kim, J. N. Synlett 2003, 357-360. (13) Haase, C.; Langer, P. Synlett 2005, 453-456. (14) Harmange, J.-C.; Figadère, B. Tetrahedron: Asymmetry 1993, 4, 1711-1754. (15) Kanazawa, C.; Terada, M. Tetrahedron Lett. 2007, 48, 933-935. (16) Uchiyama, M.; Ozawa, H.; Takuma, K.; Matsumoto, Y.; Yonehara, M.; Hiroya, K.; Sakamoto, T. Org. Lett. 2006, 8, 5517-5520. (17) Peuchmaur, M.; Lisowski, V.; Gandreuil, C. l.; Maillard, L. T.; Martinez, J.; Hernandez, J.-F. o. J. Org. Chem. 2009, 74, 4158-4165. (18) Majumdar, K. C.; Debnath, P.; Roy, B. Heterocycles 2009, 78, 2661-2728. (19) Seo, S.; Yu, X.; Marks, T. J. J. Am. Chem. Soc. 2009, 131, 263-276. (20) Yu, X.; Seo, S.; Marks, T. J. J. Am. Chem. Soc. 2007, 129, 7244-7245. (21) Wolfe, J. P.; Thomas, J. S. Curr. Org. Chem. 2005, 9, 625-655. (22) Widenhoefer, R. A. Chem.--Eur. J. 2008, 14, 5382-5391. (23) Messerle, B. A.; Vuong, K. Q. Pure Appl. Chem. 2006, 78, 385-390. (24) Hultzsch, K. C. Adv. Synth. Catal. 2005, 347, 367-391.

159

(25) Elgafi, S.; Field, L. D.; Messerle, B. A. J. Organomet. Chem. 2000, 607, 97-104. (26) Elena Mas-Marzá, José A. M., Eduardo Peris, Angew. Chem. Int. Ed. Engl. 2007, 46, 3729-3731. (27) Ho, J. H. H.; Black, D. S. C.; Messerle, B. A.; Clegg, J. K.; Turner, P. Organometallics 2006, 25, 5800-5810. (28) Marder, T. B.; Chan, D. M. T.; Fultz, W. C.; Calabrese, J. C.; Milstein, D. J. Chem. Soc., Chem. Commun. 1987, 1885-1887. (29) O'Neal, W. G.; Roberts, W. P.; Ghosh, I.; Jacobi, P. A. J. Org. Chem. 2005, 70, 7243-7251. (30) Jacobi, P. A.; Liu, H. J. Org. Chem. 1999, 64, 1778-1779. (31) Wakabayashi, T.; Ishii, Y.; Ishikawa, K.; Hidai, M. Angew. Chem., Int. Ed. Engl. 1996, 35, 2123-2124. (32) Harkat, H.; Weibel, J.-M.; Pale, P. Tet. Lett. 2006, 47, 6273-6276. (33) Toullec, P. Y.; Genin, E.; Antoniotti, S.; Genet, J.-P.; Michelet, V. Synlett 2008, 707-711. (34) Pale, P.; Chuche, J. Tetrahedron Lett. 1987, 28, 6447-6448. (35) Mindt, T. L.; Schibli, R. J. Org. Chem. 2007, 72, 10247-10250. (36) Hashmi, A. S. K. Chem. Rev. 2007, 107, 3180-3211. (37) Marchal, E.; Uriac, P.; Legouin, B.; Toupet, L.; van de Weghe, P. Tetrahedron 2007, 63, 9979-9990. (38) Dabb, S. L.; Ho, J. H. H.; Hodgson, R.; Messerle, B. A.; Wagler, J. Dalton Trans. 2009, 634-642. (39) Viciano, M.; Mas-Marza, E.; Sanau, M.; Peris, E. Organometallics 2006, 25, 3063 - 3069. (40) Liu, Z.; Hartwig, J. F. J. Am. Chem. Soc. 2008, 130, 1570-1571. (41) Baldwin, J. E.; Thomas, R. C.; Kruse, L. I.; Silberman, L. J. Org. Chem. 1977, 42, 3846-3852. (42) Izawa, K.; Okuyama, T.; Fueno, T. Bull. Chem. Soc. Jap. 1973, 46, 2881-2883.

160

Chapter 7: Summary, conclusions and future work

“We are at the very beginning of time for the human race. It is not unreasonable that we grapple with problems. But there are tens of thousands of years in the future. Our responsibility is to do what we can, learn what we can, improve the solutions, and pass them on.”

Richard P. Feynman (1918 – 1988)

161

7.1 Summary and conclusions

Functionalized lanthanide and transition metal complexes were synthesized. The lanthanide complexes were applied to protein structure refinement and the transition metal complexes were used as homogeneous catalysts for the addition of oxygen onto an unsaturated bond. The most significant conclusions are summarized below:

(a) A thiol modified dipicolinic acid, 4MMDPA (5) was synthesized and successfully attached to the N-terminal domain of the arginine repressor from E. coli (ArgN) in a site-specific manner. Analysis of the experimental PCS data and residual dipolar coupling showed that the 4MMDPA-Ln3+ complexes were able to bind rigidly to the protein, the first such ligand to be reported in literature using one attachment point on the protein.

(b) An analogue of the ligand 5, the 3MDPA (9) ligand was successfully synthesized and attached to the ArgN protein in a site specific manner. In the presence of paramagnetic transition and lanthanide metal ions, large measurable PCS were observed in the HSQC spectrum of the ArgN-3MDPA adduct. The cal tensor for the ArgN-3MDPA-Ln3+ complexes highlighted the dependence of the tensor on the local environment. Due to the insufficient data, the overall rigidity of the 3MDPA probe could not be assessed. The 3MDPA (9) ligand has been shown to be highly effective and complimentary ligands to the 4MMDPA ligand (5) for introducing paramagnetism for the refinement of protein structures.

(c) The unnatural amino acids 11, 15 and 19, containing transition and lanthanide metal ion binding motifs were synthesized successfully using a generalized method which involved the alkylation of acetamidomalonate ester with the appropriate halide.

(d) A series of complexes [Rh(L)(p-tpt)][BArF] (L = COD, 27a; L = NBD, 28a; L =

CO, 29a) and [Rh(COD)(p-tpt)][BPh4] (32) was synthesized. The solid and

162

solution state structures of the series of complexes 27a-29a and 32 showed that despite the potential for the p-tris(pyrazolyl)toluidine ligand (24a) to bind with all three pyrazolyl donors, the 2 binding mode was the preferred binding mode irrespective of the nature of the co-ligand and the counterion.

(e) The complexes [Rh(L)(o-tpt)][BArF] (L = COD, 27b; L = NBD, 28b; L = CO, 29b) were synthesized and the solid state structures showed that the o-tpt ligand (24b) was bound to the rhodium centre via the 2 binding mode irrespective of the nature of the co-ligand. The 1H NMR spectra of the complexes 27b-29b at low temperatures indicated the presence of two products, shown to be undergoing exchange using a 1H NOESY experiment. The exchange observed in the 1H NMR spectrum at low temperature was attributed to the restricted rotation of the aryl ring about the C-C bond connecting the aryl ring and the bridging carbon.

(f) Two tris(N-methylimidazolyl)methanol complexes, [Rh(COD)(tim)][BArF] (36)

and [Rh(CO)2(tim)][BArF] (37) were synthesized. The solid state structures of the complexes 36 and 37 showed the rhodium centre was bound to the ligand in the 2 binding mode via two of the imidazolyl donors in a similar fashion to the rhodium(I) complexes bearing the tpt ligand (24). The 1H NMR spectra of the complexes 36 and 37 at low temperature did not resolve the two inequivalent imidazolyl donor environments as had also been observed for the complexes bearing the pyrazolyl donors 27-29.

(g) The complexes [Rh(CO)2(tim)][BArF] (37) and [Rh(CO)2(bim)][BArF] (40) were more active catalysts for the intramolecular cyclization of 4-pentynoic acid (38)

and 5-hexynoic acid (41) than their pyrazolyl counterpart [Rh(CO)2(p-tpt)][BArF] (29a). The complexes 37 and 40 are more efficient catalysts than the silver and gold salts reported previously for the cyclization of 4-pentynoic acid (38).

(h) The complexes 29a, 37 and 40 were highly active catalysts for the intramolecular cyclization of the aromatic alkynoic acid 44a under the optimized reaction conditions. The exo-dig : endo-dig selectivity of the catalyst was significantly

163

altered depending on the nature of the substituents on the terminal alkyne. An increase in the cyclization rate was observed with an increase in the electron withdrawing nature of the substituent at the terminal alkyne position.

7.2 Future work

The results presented in this thesis have established how effective the 4MMDPA (5) and 3MDPA (9) are as paramagnetic probes for protein structure refinement. As an extension, the unnatural amino acids 11, 15 and 19 were synthesized as a means to introduce the paramagnetic probe in a site specific manner without the use of post translational modifications. The incorporation of the unnatural amino acids and their transition metal complexes was attempted by Dr Kiyoshi Ozawa, but this was unsuccessful due to the toxicity of the compounds to the modified expression system. To overcome the toxicity of the unnatural amino acids, future work could comprise the synthesis of a new series of unnatural amino acids bearing protecting groups on the carboxylate side chains. The synthesis of these protected amino acids would involve the use of orthogonal protecting groups to allow the deprotection of the acetamidomalonate esters without the loss of the side chains protecting groups.

The rhodium(I) complexes bearing tpt (24) and tim (23) ligands were shown to bind to rhodium in the 2 binding mode irrespective of the location of the amine group and the co-ligands. Based on the solid and solution state structures of these complexes, there is still considerable scope for further investigation:

(a) Based on the presence of the amine group on the backbone of the o-tpt ligand (24b) and the unbound pyrazolyl donor, one of the future directions would be to convert the primary amine group on the ligand backbone to an imine, providing an additional nitrogen donor. The ligand would then contain 4 nitrogen donors which

164

may allow the formation of a bimetallic species which would provide interesting catalytic properties compared to their mono-metallic counterparts.

(b) The results show the preference of the 2 binding mode over the 3 binding mode irrespective of the nature of the co-ligand and counterion. Modification of the electronic nature of the aryl ring could also be considered as part of future works. The amine group on the ligand backbone of either p-tpt (24a) or o-tpt (24b) could be converted to other functional groups with different electronic properties via the diazonium salt. The solid state structure of the rhodium(I) complexes bearing the ligands with varying electron densities can then be examined in detailed showing the influence of the electron density on the ligand backbone on the binding mode of the nitrogen donors.

(c) The synthesis of rhodium (III) and iridium (III) complexes with the p-tpt (24a) and o-tpt (24b) ligands to investigate their luminescence properties with possible applications in biomolecule labeling using the chemistry described in Chapter 2 and 3 would expand the potential application of these systems.

The rhodium(I) complexes of tridentate and bidentate N,N donor ligands were shown to be active as catalysts for the intramolecular cyclization of aliphatic or aromatic alkynoic acids. Based on these results, there are a number of potential additional directions for these complexes:

(a) The introduction of electron withdrawing groups to the substrates led to an increase in the rates of reaction for the intramolecular cyclization of non-terminal alkynoic acids. As part of future work, DFT calculations could be used to examine the electron densities at the alkyne bond. This would provide information on whether the changes in the rates of reaction observed is caused by changes in either the susceptibility of the alkyne bond to nucleophilic attack, or changes in the binding affinity of the alkyne to the rhodium centre.

(b) The presence of the amine group on the ligand backbone allows the complex 29 to be immobilized onto macroscopic scaffolds such as proteins and carbon surfaces.

165

The complex can be immobilized onto a carbon surface in one of two ways: (a) the conversion of the amine group on the p-tpt ligand (24a) to a diazonium salt in situ, followed by electrochemical reduction onto the carbon surface (Figure 7.1a); or (b) the formation of a peptide bond between a carboxylic acid modified carbon

surface and the amine group on the complex [Rh(CO)2(p-tpt)][BArF] (29a) (Figure 7.1b).

(c) In addition to immobilizing onto a carbon surface, the complex can also be immobilized onto a protein by converting the amine group on the backbone of the complex 29a to a thiol group which allows the use of the ligation protocol discussed in Chapter 2 and 3.

(d) Rhodium(I) complexes are not only active as catalysts for the addition of oxygen onto unsaturated bonds. Other industrially relevant reactions such as hydroamination, hydrosilylation and hydrogenations with the complexes described in this work should also be investigated.

166

Figure 7.1: Proposed methods for the immobilization of the complex

167

Chapter 8: Experimental

“There is no such thing as a failed experiment, only experiment with unexpected outcomes”

Robert Buckminster Fuller (1895 – 1983)

168

8.1 General Experimental

All manipulations of metal complexes and air sensitive reagents were carried out using standard Schlenk techniques or in a nitrogen or argon filled Braun glovebox.

Reagents were purchased either from Aldrich Chemical Company Inc. or Lancaster Synthesis and used without further purification unless otherwise stated. The metal halide salt, rhodium (III) chloride, was purchased as the hydrate from Precious Metals Online. For the purposes of air sensitive manipulations and in the preparation of air sensitive complexes; tetrahydrofuran (THF), n-pentane and dichloromethane were dispensed from a PuraSolv solvent purification system. The bulk compressed gases argon (>99.99%), nitrogen (>99.5%), and carbon monoxide (>99.5%) were obtained from Linde Gases and used as received. Hydrogen gas (ultra high purity) was obtained from BOC Gases and used as supplied.

Infra-red spectra were measured using a Nicolet 360 / 370 Avatar FTIR spectrometer. Microanalysis was carried out at the Microanalytical Unit, Australian National University, Canberra. Mass spectroscopy was performed at the Mass Spectroscopy Unit, Australian National University, Canberra. Melting points were determined using a Gallenkamp melting point apparatus and are uncorrected.

Air sensitive NMR samples were prepared using a high vacuum line by vacuum transfer of solvent into an NMR tube fitted with a concentric Teflon valve. Deuterated solvents for NMR purposes were obtained from Cambridge Isotopes. Chloroform-d1, methanol- d4, acetone-d6 were dried and distilled over calcium hydride. Benzene-d6 and toluene-d8 were dried and distilled over sodium / benzophenone. All NMR solvents used for air sensitive compounds were degassed using three consecutive freeze-pump-thaw cycles and vacuum distilled immediately prior to use.

1H, 13C and 15N NMR spectra were recorded on a Bruker DPX300 spectrometer operating at 300 MHz (1H) and 75 MHz (13C), a Bruker Avance 400 spectrometer operating at 400 MHz (1H) and 100 MHz (13C), a Bruker Avance 500 spectrometer operating at 500 MHz (1H) and 125 MHz (13C), a Bruker DPX600 spectrometer

169

operating at 600MHz (1H), 150 MHz (13C) and 60 MHz (15N) and a Bruker Avance 800 spectrometer operating at 800 MHz (1H), 200 MHz (13C) and 80 MHz (15N). 1H and 13C 1 NMR chemical shifts were referenced internally to solvent shifts ( H chloroform-d1 13 7.26, D2O 4.7; C chloroform-d1 77). All NMR measurements on protein samples were performed at 25oC in 2-(N-morpholino)ethanesulfonic acid (MES) buffer (50 mM, pH 6.5) using a Bruker Avance 800 or Bruker Avance 600 MHz spectrometer equipped with a cryoprobe at the Research School of Chemistry, Australian National University, Canberra. Residual dipolar couplings were measured in the 15N dimensions in the 15N- 1 1 H HSQC spectra recorded with the IPAP scheme using t1max of 70ms. Chemical shifts (!$\$V" in chemical shifts are typically ±0.01 ppm for 1H and ±0.05 for 13C. Coupling constants (J) are given in Hz and have uncertainties of ±0.1 Hz for 1H-1H and ±0.5 Hz for 1H-13C couplings. The following abbreviations are used for convenience in reporting the multiplicity of NMR resonances: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad.

The following two dimensional NMR techniques were routinely used for the assignment of organic and inorganic compounds: NOESY (nuclear Overhauser effect spectroscopy), HSQC (heteronuclear single quantum coherence), HMQC (heteronuclear multiple quantum coherence), HMBC (heteronuclear multiple bond coherence). All NMR data was acquired and processed using standard Bruker software (Topspin).

The following compounds were prepared using literature methods: 2,6- dimethoxycarbonylpyridine (2)2, 2,6-dimethoxycarbonyl-4-tosyloxymethylpyridine (3)2, 3-bromo-2,6-dimethylpyridine (6)3, 3-bromo-2,6-dimethoxycarbonylpyridine (7)3, 2- acetylamino-(4-nitrobenzyl)malonate, diethyl ester (12)4, 2-acetylamino- (4-aminobenzyl)malonate, diethyl ester hydrochloride (13)5, 5-chloromethyl-8- hydroxyquinoline hydrochloride (16)6, 5-chloromethyl-8-hydroxyquinoline, acetate ester (17)7, tris(imidazolyl)methanol (23)8, tris(pyrazolyl)toluidine (24)9, 10 10 11 [Rh(COD)2][BArF] (26a) , [Rh(NBD)2][BArF] (26b) , [Rh(COD)Cl]2 (33) , hexynoic acid (41)12. The alkynoic esters (43a-b and 49a-d) and alkynoic acids (44a-b and 50a-d) were prepared using a modified procedure from the method described by Marchal et al.13 and described in section 8.6.2.

170

- - The numbering of the atoms within BArF and BPh4 counterions used are shown below,

8.2 Synthesis of thiol modified lanthanide binding ligands and modification of proteins.

8.2.1 Synthesis of 2,6-Dimethoxycarbonyl-4-bromomethylpyridine (4).

Lithium bromide (212 mg, 2.44 mmol) was added to a solution of dimethoxycarbonyl-4-tosyloxymethylpyridine (3) (460 mg, 1.22 mmol) in acetone (25 mL) and allowed to stir at room temperature for 2 hours. The resulting suspension was filtered, concentrated and the residues triturated with chloroform (50 mL). The solvent was removed in vacuo to yield the title compound as a tan solid (233 mg, 67%), mp: 110- 114oC.

Anal. calcd. C10H10NO4Br: C, 41.62; H, 3.50; N, 4.86. Found: C, 41.69; H, 3.55; N, 4.76. 1 H NMR (300 MHz, CDCl3) 8.31 (s, 2H, H3), 4.50 (s, 2H, CH2Br), 4.03 (s, 6H,

CO2CH3). 13 1 C{ H} NMR (75 MHz, CDCl3) 164.6 (CO2CH3), 149.2 (C2), 148.9 (C4), 127 (C3),

53.3 (CO2CH3), 29.3 (CH2Br).

171

8.2.2 Synthesis of 4-Mercaptomethyl-2,6-pyridinedicarboxylic acid (5). Under a nitrogen atmosphere, a solution of 4 (190 mg, 0.66 mmol) and thiourea (59 mg, 0.7 mmol) in methanol (10 mL) was heated under reflux overnight. The resulting solution was allowed to cool to room temperature and the solvent was removed in vacuo. The residue was dissolved in degassed water (20mL) and sodium hydroxide pellets (190 mg, 4.75 mmol) were added whereupon the reaction mixture was refluxed overnight under a nitrogen atmosphere. The resulting solution was eluted through a cation exchange (Amberlite GC120) column and the solvent removed in vacuo to yield the title compound 4 as a tan solid (130 mg, 99%), m.p. >150oC (dec)

- - EI-HRMS (ES ) calcd. for C8H7NO4S [M-H] : 212.0096. Found 212.0016

1 H NMR (800 MHz, D2O) 8.28 (s, 2H, H3), 3.86 (s, 2H, CH2SH). 13 1 C{ H} NMR (200 MHz, D2O) 171.6 (CO2H), 153.0 (C2), 152.3 (C4), 124.2 (C3),

26.8 (CH2SH).

8.2.3 Synthesis of 3-mercapto-2,6-pyridinedicarboxylic acid (9). A suspension of sodium tert-butylthiolate (79 mg, 0.7 mmol) and 7 (100 mg, 0.366 mmol) in THF (10 mL) was stirred at room temperature for 3 days. The solvent of the resulting suspension was removed in vacuo to yield the tert-butyl ether as a beige solid. Concentrated hydrochloric acid was added to the crude tert-butyl ether and heated under reflux for ca. an hour and a suspension forms upon cooling. The suspension was then filtered to yield the titled compound as a brown solid (12.3 mg, 14 %), m.p: 188 - 190 oC

+ + EI-HRMS (ES ) calcd. For C7H6NO4S [M+H] : 200.0018. Found 200.0019

1 H NMR (500 MHz, D2O / NaOD) 8.16 (d, 1H, J = 8 Hz, H5), 7.84 (d, 1H, J = 8 Hz, H4). 13 1 C{ H} NMR (200 MHz, D2O) 171.6 (CO2H), 153.0 (C2), 152.3 (C4), 124.2 (C3),

26.8 (CH2SH).

172

8.2.4 Ligation of thiol modified tags to N-terminal domain of the arginine repressor (ArgN).

A solution of the ArgN protein in water (0.3 mL, 4.6 mM, 1.38 )mol) was first reduced with a solution of dithiothreitol (5 eq.) in tris(hydroxymethyl)aminomethane (TRIS) buffer solution (50 mM, pH 7.0). Unreacted dithiothreitol was washed out completely using additional TRIS buffer solution (50 mM, pH 7.6) using a centrifugal filtration using a Millipore ultrafilter (MW C/O 5 kD). The volume of protein solution was reduced and added dropwise to a solution of 5,5-dithiobis-(2-nitrobenzoic acid) (29.9 mg, 75.4 )mol) in TRIS buffer solution (50 mM, pH > 7.6) and left to stand with shaking for ca. 1 hour. The resulting solution was then concentrated and washed with additional TRIS buffer solution (50 mM, pH 7.6) until a colorless solution was obtained. To the activated protein solution, a solution of the appropriate thiol tag (10 mg) in TRIS buffer solution (50 mM, pH > 7.6) was added and left to stand at room temperature for ca. 1 hour. The resulting solution was then concentrated and washed with additional TRIS buffer solution till a colorless solution is obtained. The crude ligated protein was purified by using FPLC using a MonoQ column with typical yields of 85%.

8.2.5 Protein NMR Measurements

NMR measurements of ArgN and the C54T/C97A/Q69C triple-mutant of T4 lysozyme derivatized with either 4MMDPA (5) or 3MDPA (9) were performed at 25oC in solutions containing 20 mM 2-(N-morphino)ethanesulfonic acid (MES) buffer at pH 6.5, using a Bruker 800 MHz NMR spectrometer equipped with a cryoprobe. The PCSs were measured by measuring the cross-peak displacements in the 1H dimension of 15N- HSQC spectra that had been recorded in mixed (1:1) solutions of paramagnetic lanthanide and diamagnetic Y3+, or in mixed solutions of Co2+ and Zn2+. The molar ratio of [M] : [protein] was 1.5:1. The program Numbat was used to assist in the assignment of the 15N-HSQC cross-peaks of the paramagnetic samples.

173

8.3 Synthesis of unnatural amino acids bearing lanthanide and transition metal ion binding motifs.

8.3.1 Synthesis of Diethyl Acetamido(2,6-dimethoxycarbonylpyridyl)malonate (10).

In an oven dried round bottom flask sodium hydride (220 mg, 60% w/w, 5.5 mmol) was added to a solution of diethyl acetamidomalonate (716 mg, 3.3 mmol) in THF (50 mL) and allowed to stir at room temperature for ca. 1 hour. To the resulting suspension, the bromide 4 (900 mg, 3.1 mmol) was added and allowed to stir overnight. The resulting suspension was then filtered and washed with dichloromethane (3 x 50 mL). The solvent was removed in vacuo, and the resulting residues suspended in n-pentane, filtered and washed with additional portions of n-pentane. The resulting residues were triturated with hot ethyl acetate (4 x 50 mL) and recrystallized from ethyl acetate to yield the titled compound 10 as a pale pink solid (490 mg, 37 %), m.p. 146 – 148oC

+ + EI-HRMS (ES ) Calcd. for C19H25N2O9 [M+H] : 425.1560. Found: 425.1559

1 H NMR (300 MHz, CDCl3) 7.96 (s, 1H, H3), 6.53 (s, 1H, NHCOCH3), 4.78

(q, J = 7 Hz 4H, CO2CH2CH3), 3.99 (s, 6H, CO2CH3), 3.82 (s, 2H, PyCH2), 2.06 (s, 3H,

NHCOCH3), 1.30 (t, J = 7 Hz, 6H, CO2CH2CH3) 13 1 C{ H} NMR (75 MHz, CDCl3) 169.4 (NHCOCH3), 166.1 (CO2CH2CH3), 165.0

(PyCO2CH3), 148.3 (C2), 147.8 (C4), 129.3 (C3), 66.6 (C(CO2CH2CH3)2), 63.2

(CO2CH2CH3), 53.1 (PyCO2CH3), 37.1 (PyCH2), 22.9 (NHCOCH3), 13.9

(CO2CH2CH3) * (KBr): 3306m, 2987w, 2953w, 1743vs, 1728vs, 1658vs, 1605w, 1514m cm-1

8.3.2 Synthesis of the unnatural amino acid DPA-AA (11).

174

A solution of 10 (170 mg, 0.4 mmol) in concentrated hydrochloric acid (10M, 5 mL) was heated under reflux overnight. The resulting solution was allowed to cool and the solvent was removed in vacuo to yield the titled compound as the hydrochloride salt (68 mg, 67 %).

1 H NMR (300 MHz, D2O) 8.15 (s, 2H, H3), 4.40 (t, J = 9 Hz, 1H, H), 3.46 (dd, J =

15 Hz, 5.8 Hz, 1H, PyCH2), 3.30 (dd, J = 15 Hz, 8.0 Hz, 1H, PyCH2). 13 1 C{ H} NMR (75 MHz, D2O) 170.1 (CO2H), 166.5 (PyCO2H), 149.7 (C4), 147.4

(C2), 129.36 (C3), 53.0 (C), 35.0 (PyCH2)

8.3.3 Synthesis of Diethyl Acetamido(N,N-bis(ethoxycarbonylmethyl)- aminobenzyl)malonate (14).

In a Schlenk flask, a suspension of 13 (2.06 g, 5.75 mmol), sodium iodide (2.10 g, 14.0 mmol) and N,N-diisopropylethylamine (4 mL, 22.9 mmol) in N,N-dimethylformamide (ca. 25 mL) was stirred at room temperature for ca. 1 hour. To the suspension, ethyl bromoethylacetate (2 mL, 18.1 mmol) was added and the resulting suspension heated at 100oC overnight. The resulting red suspension was diluted with water (200 mL) and the product extracted with ethyl acetate (3 x 150 mL). The combined organic phase was washed with dilute hydrochloric acid (0.1 M, 3 x 100 mL) and water (3 x 100 mL), and dried over sodium sulfate. The solvent was removed in vacuo, and the crude product recrystallized from a mixture of ethyl acetate and n-hexane to yield the titled compound as a pale brown solid (2.00 g, 71%), m.p. 94 – 96oC

Anal. calcd. for C24H34N2O9: C, 58.29; H, 6.93; N, 5.66. Found: C, 58.49; H, 7.01; N, 5.34

1 H NMR (300 MHz, CDCl3) 6.83 (d, J = 8.7 Hz, 2H, H2), 6.48 (d, J = 8.7 Hz, 2H,

H3), 6.49 (s, 1H, NHCOCH3), 4.25 (dq, J = 1.4 Hz, 7.0 Hz, 4H, C(CO2CH2CH3)2), 4.21

(q, J = 7.2 Hz, 4H, N(CH2CO2CH2CH3)2), 4.08 (s, 4H, N(CH2CO2CH2CH3)2), 2.00 (s, 175

3H, NHCOCH3), 1.27 (t, J = 7 Hz, C(CO2CH2CH3)2), 1.26 (t, J = 7.2 Hz,

N(CH2CO2CH2CH3)2) 13 1 C{ H} NMR (75 MHz, CDCl3) 170.9 (N(CH2CO2CH2CH3)2), 169.0 (NHCOCH3),

167.7 (C(CO2CH2CH3)2), 147.1 (C1), 130.7 (C2), 124.6 (C4), 112.4 (C3), 67.3

(N(CH2CO2CH2CH3)2), 62.5 (C(CO2CH2CH3)2), 53.4 (N(CH2CO2CH2CH3)2), 36.9

(C(CO2CH2CH3)2), 23.1 (NHCOCH3), 14.2 (C(CO2CH2CH3)2), 14.0

(N(CH2CO2CH2CH3)2) * (KBr): 3390m, 2982m, 2935m, 1746vs, 1673vs, 1617m, 1524s, 1500s cm-1

8.3.4 Synthesis of 2-amino-3-(N,N-bis(methylcarboxylate)aminobenzyl)propanoic acid (15).

A solution of 14 (0.93 g, 1.8 mmol) in concentrated hydrochloric acid (10 M, 50 mL) was heated under reflux for 4 hours. The solvent was then removed in vacuo and to the resulting residue, water was added and solution chilled in ice. The resulting suspension was then filtered to yield the titled compound 15 as a white solid. (0.29 g, 52%), m.p. 210 (decomp.)

Anal. calcd. for C13H16N2O6: C, 52.70; H, 5.44; N, 9.46. Found: C, 52.26; H, 5.39; N, 9.53

1 H NMR (400 MHz, D2O / NaOD) 6.91 (d, J = 7.6 Hz, 2H, H2), 6.32 (d, J = 7.6 Hz,

2H, H2), 3.71 (s, 4H, N[CH2CO2H]2), 3.24 (t, J = 7 Hz, 1H, H), 2.68 (dd, J = 13.4 Hz,

7 Hz, 1H, PhCH2). 2.52 (dd, J = 13.4 Hz, 7 Hz, 1H, PhCH2) 13 1 C{ H} NMR (100 MHz, D2O / NaOD) 182.7 (CO2H), 179.7 (N[CH2CO2H]2), 125.6

(C4), 147.2 (C1), 131.0 (C2), 111.4 (C3), 57.2 (C), 55.6 (N[CH2CO2H]2), 39.6

(PhCH2) * (KBr): 3440br, 3250br, 2924m, 1943br, 1723m, 1696s, 1615vs, 1522vs cm-1

8.3.5 Synthesis of Diethyl Acetamido-6-(8-acetyloxyquinolyl)malonate diethyl ester (18).

176

In an oven dried round bottom flask sodium hydride (0.62 g, 60% w/w, 16 mmol) was added to a solution of diethyl acetamidomalonate (2.68 mg, 12 mmol) in THF (50 mL) and allowed to stir at room temperature for ca. 1 hour. To the resulting suspension, the chloride 17 (2.86 g, 12 mmol) was added and the mixture was allowed to stir overnight. The resulting suspension was then filtered and washed dichloromethane (3 x 50 mL). The solvent was removed in vacuo, and the resulting residues suspended in n-hexane, filtered and washed with additional portions of n-hexane to yield the titled compound as a pale green solid (3.94 g, 78%), m.p. 138 – 140oC

Anal. calcd. for C21H24N2O7: C, 60.57; H, 5.81; N, 6.73. Found: C, 60.31; H, 5.97; N, 7.02

1 H NMR (300 MHz, CDCl3) 8.86 (d, J = 4.2 Hz, 1H, H1), 8.32 (d, J = 8.9 Hz, 1H H3), 7.39 (dd, J = 8.9, 4.2 Hz, H2), 7.35 – 7.28 (m, 1H, H7), 7.22 – 7.13 (m, 1H H6),

6.48 (s, 1H, NHCOCH3), 4.35 - 4.15 (m, 4H, CO2CH2CH3), 4.08 (s, 2H, ArCH2), 2.47

(s, 1H, OCOCH3), 1.85 (s, 1H, NHCOCH3), 1.27 (t, J = 7.2 Hz, 6H, CO2CH2CH3). 13 1 C{ H} NMR (75 MHz, CDCl3) 169.8 (COCH3), 169.7 (COCH3), 167.3

(CO2CH2CH3), 150.1 (C1), 146.9 (C8), 141.3 (C4), 132.6 (C3), 130.6 (C9), 129.4

(C5), 127.8 (C6), 121.2 (C2), 120.8 (C7), 67.5 (C), 33.3 (ArCH2), 23.0 (NHCOCH3),

21.0 (NHCOCH3), 13.9 (CO2CH2CH3) * (KBr): 3407br, 3276s, 2984m, 2939w, 2904w, 1746s, 1640s, 1596w, 1579w, 1505s cm-1

8.3.6 Synthesis of 2-amino-6-(9-hydroxylquinoyl)propanoic acid dihydrochloride (19).

A solution of 18 (460 mg, 1.1 mmol) in concentrated hydrochloric acid (10 M, 10 mL) was heated under reflux for 4 hours. The resulting solution was allowed to cool to room temperature and then chilled in an ice bath. The resulting suspension was filtered, yielding the titled compound as a

177

yellow solid (228 mg, 68 %), m.p. 218oC (decomp.)

Anal. calcd. for C12H12N2O32HCl: C, 47.23; H, 4.62; N, 9.18. Found: C, 47.14; H, 4.42; N, 9.19

1 H NMR (400 MHz, D2O / NaOD) 9.19 (d, J = 8.9 Hz, 1H, H1), 9.00 (d, J = 5.4 Hz, 1H, H3), 8.09 (m, 1H, 1H, H2), 7.66 (d, J = 8.0 Hz, 1H, H6), 7.39 (d, J = 8.0 Hz, 1H, H7), 4.30 (t, J = 7.6 Hz, 1H, H ), 3.80 (dd, J = 6.5Hz, 15Hz, 1H, PhCH2), 3.60 (dd, J =

8 Hz, 15 Hz, PhCH2) 13 1 C{ H} NMR (100 MHz, D2O / NaOD) 171.2 (CO2H), 147.1 (C8), 143.1 (C1), 142.6 (C3), 132.6 (C6), 129.7 (C9), 128.7 (C4), 122.8 (C5), 122.2 (C2), 116.2 (C7), 53.7 (C ), 31.5 (PhCH2)

8.4 Synthesis of rhodium(I) complexes bearing pyrazolyl donor ligands.

8.4.1 General synthesis of rhodium(I) olefin tris(pyrazolyl)toluidine complexes bearing the BArF counterion 27a-b and 28a-b.

A solution of tris(pyrazoyl)toluidine (p-tpt 24a; o-tpt 24b) and [Rh(L)2][BArF] (L = COD, 26a; L = NBD, 26b) in THF was stirred at room temperature for 30 mins. The resulting solution was evaporated to dryness in vacuum, whereupon dichloromethane (2 mL) was added and the complex crystallized after the slow addition of n-pentane. Single crystals suitable for x-ray diffraction analysis were grown via slow diffusion of n-pentane or n-hexane into a concentrated solution the complex in dichloromethane.

8.4.1.1 Synthesis of [Rh(COD)(p-tpt)][BArF] (27a).

178

A solution of p-tris(pyrazolyl)toluidine (24a)

(28.4 mg, 0.10 mmol) and [Rh(COD)2][BArF] (26a) (103.6 mg, 0.09 mmol) was stirred in THF (ca. 5mL) to yield the titled compound as yellow solid (103.6 mg, 75 %), m.p. 128 – 130oC

Anal. calcd. C56H39BF24N7Rh: C, 48.75; H, 2.85; N, 7.11. Found: C, 49.09; H, 3.11; N, 6.95.

1 o P(2) P(2) H NMR (600 MHz, CDCl3 at -50 C) 7.97 (br, 1H, H5 or H3 ), 7.69 (br, 8H, H2BArF), 7.49 (br, 4H, H4BArF), 7.45 (br, 2H, H5P(1) or H3P(1)), 7.21 (br, 2H, H3P(1) or H5P(1)), 7.00 (br, 1H, H3P(2) or H5P(2)), 6.71 (d, J = 7.8 Hz, 1H, H3’), 6.38 (br, 2H, H4P(1)), 6.37 (br, 1H, H4P(2)), 6.12 (d, J = 7.8 Hz, 1H, H2’), 4.20 (br, 2H, H1,COD), 3.64 (br, 2H, H1,COD), 2.47 (br, 2H, H2,COD or H3,COD), 1.82 (br, 2H, H2,COD or H3,COD), 1.62 (br, 4H, H2,COD and H3,COD). 13 1 o BArF C{ H} NMR (150 MHz, CDCl3 at -50 C) 162.07 (q, J = 50 Hz, C1 ), 151.1 (C1’), 146.1 (C3P(2) or C5P(2)), 143.6 (C3P(1) or C5P(1)), 135.63 (C3P(1) or C5P(1)), 135.14 (C2BArF), 134.56 (C3P(2) or C5P(2)), 129.2 (q, J = 31Hz, C3BArF) 129.1 (C3’), BArF BArF 124.9 (q, JC-F = 270 Hz, CF3), 123.9 (C4’) 118.6 (C4 ), 117.9 (C1 ), 114.9 (C2’), 108.9 (C4P2), 108.1 (C4P1), 94.2 (Ca), 84.2 (C1COD), 83.6 (C1COD), 30.63 (C2,COD and C3, COD), 29.67 (C2,COD and C3,COD) * (KBr): 3479m, 3417m, 3364m, 3222m, 3163m, 3021m, 2961m, 2961m, 2925m, 2892m, 2844w, 1627s, 1609s, 1519s cm-1

8.4.1.2 Synthesis of [Rh(COD)(o-tpt)][BArF] (27b).

179

A solution of o-tris(pyrazolyl)toluidine (24b)

(27.8 mg, 0.09 mmol) and [Rh(COD)2][BArF] (26a) (102.9 mg, 0.09 mmol) was stirred in THF (ca. 5mL) to yield the titled compound as a yellow solid (72.5 mg, 58 %), m.p. 126 – 128oC

Anal. calcd. C56H39BF24N7Rh: C, 48.75; H, 2.85; N, 7.11; Found: C, 48.55; H, 3.00; N, 6.53.

Major Product: 1 o P(3) P(3) H NMR (600 MHz, CDCl3 at -55 C) 7.86 (br, 1H, H3 or H5 ), 7.68 (br, 1H, H3P(2) or H5P(2)), 7.72 (br, 1H, H3P(1) or H5P(1)), 7.67 (s, 8H, H3BArF), 7.57 (br, 1H, H3P(2) or H5P(2)), 7.52 (br, 1H, H3P(1) or H5P(1)), 7.48 (s, 4H, H1BArF), 7.38 (t, J = 8.2 Hz, 1H, H4’), 6.73 (t, J = 8.2 Hz, 1H, H3’), 6.72 (br, 1H, H5’), 6.69 (br, 1H, H4P(3)), 6.58 (br, 1H, H3P(3) or H5P(3)) 6.52 (br, 1H, H4P(2)), 6.40 (br, 1H, H4P(1)), 5.52 (d, J = 8.2 Hz, 1H, H2’), 4.31 (br, 1H, H1COD), 4.26 (br, 1H, H1COD), 3.90 (br, 1H, H1COD), 3.73 (br, 1H, H1COD), 2.66 – 2.51 (m, 1H, H2COD), 2.50 – 2.40 (m, 1H, H2COD), 2.08 – 1.67 (m, 1H, H2COD), 1.96 – 1.77 (m, 4H, H2COD), 1.77 – 1.62 (m, 1H, H2COD). 13 1 o BArF C{ H} NMR (150 MHz, CDCl3 at -55 C) 161.2 (q, J = 49.3 Hz, C4 ), 144.9 (C3P(2) or C5P(2)), 144.8 (C3P(1) or C5P(1)), 143.4 (C1’), 142.6 (C3P(1) or C5P(3)), 137.4 (C3P(1) or C5P2), 136.5 (C3P(1) or C5P(1)), 134.3 (C4’), 134.0 (C3BArF), 129.3 (C3P3 or P(3) BArF C5 ), 128.3 (q, J = 31 Hz, C2 ), 127.9 (C2’), 124.0 (q, J = 271 Hz, CF3), 119.6 (C3’), 117.4 (C5’), 117.1 (C3BArF), 114.6 (C1’), 109.3 (C4P3), 108.9 (C4P1), 108.4 (C4P2), 92.1 (Ca), 86.3 (C1COD), 83.8 (C1COD), 82.8 (C1COD), 30.4 (C2COD), 29.9 (C2COD), 29.3 (C2COD), 28.8 (C2COD). * (KBr): 3520m, 3429m, 3172m, 3020m, 2960m, 2893m, 2845w, 1632s, 1610s, 1575w cm-1

Minor Product: 1 o H NMR (600 MHz, CDCl3 at -55 C, assigned using COSY and HMBC) 8.02 (br, 1H, H3P(1) or H5P(1)), 7.83 (d, J = 2.6 Hz, 1H, H3P(2) or H5P(2)), 7.60 (br, 1H, H3P(2) or H5P(2)), 7.44 (overlapping, 2H, H3P(3) or H5P(3) and H5’), 7.15 (d, J = 2.5Hz, 1H, H3P(1)

180

or H5P(1)), 7.09 (d, J = 2.5 Hz, 1H, H3P(3) or H5P(3)), 6.87 (t, J = 7.7 Hz, 1H, H3’ or H4’), 6.67 (H3’ or H4’), 6.48 (br, 1H, H4P(2)), 6.44 (br, 1H, H4P(1)), 6.36 (br, 1H, H4P(3)), 5.80 (d, J = 8.0 Hz, 1H, H2’). Only the aromatic region was assigned and where no integration and multiplicity are stated, resonances of the minor product were overlapped by resonances due to the major product.

8.4.1.3 Synthesis of [Rh(NBD)(p-tpt)][BArF] (28a).

A solution of p-tris(pyrazolyl)toluidine (24a)

(44 mg, 0.1 mmol) and [Rh(NBD)2][BArF] (26b) (101 mg, 0.09 mmol) was stirred in THF (ca. 5mL) to yield the titled compound as a yellow solid (119 mg, 98 %), m.p. 133 – 135oC.

Anal. calcd. C55H35BF24N7Rh: C, 48.45; H, 2.59; N, 7.19. Found: C, 48.68; H, 2.68; N, 6.97. 1 o P(2) P(2) H NMR (600 MHz, CDCl3 at -55 C) 7.99 (br, 1H, H3 or H5 ), 7.69 (s, 8H, BArF BArF P(1) P(1) P(1) H2 ), 7.49 (s, 4H, H4 ), 7.17 (br, 2H, H3 or H5 ), 7.14 (br, 2H, H3 or H5P(1)), 7.09 (br, 1H, H3P(2) or H5P(2)), 6.69 (d, J = 8.2 Hz, 2H, H3’), 6.38 (br, 1H, H4P(2)), 6.32 (br, 1H, H4P(1)), 5.89 (d, J = 8.2 Hz, 2H, H2’), 4.03 (br, 2H, H1NBD), 3.89 (br, 1H, H2NBD), 3.51 (br, 2H, H1NBD), 3.45 (br, 1H, H2NBD), 1.17 (br, 2H, H3NBD). 13 1 o BArF C{ H} NMR (150 MHz, CDCl3 at -55 C) 161.2 (q, JBC = 49 Hz, C1 ), 149.9 (C1’), 145.0 (C5P(2)), 142.1 (C3P(1)), 134.4 (C5P(1)), 134.3 (C2BArF), 133.5 (C3P(2)), 128.5 BArF (q, J = 32 Hz, C3 ), 127.6 (C2’), 123.9 (q, JCF = 271 Hz, CF3), 123.4 (C4’), 117.2 (C4BArF), 114.1 (C3’), 108.1 (C4P(2)), 107.2 (C4P(1)), 93.2 (Ca), 60.9 (C1NBD), 59.3 (C1NBD), 50.9 (C2NBD), 50.4 (C2NBD), 31.3 (C3NBD). * (KBr): 3522w, 3484w, 3421w, 3392w, 3161w, 3139w, 3017w, 2961w, 2933w, 2915w, 2870w, 1627m, 1610m, 1519m cm-1

181

8.4.1.4 Synthesis of [Rh(NBD)(o-tpt)][BArF] (28b).

A solution of o-tris(pyrazolyl)toluidine (24b)

(43.5 mg, 0.1 mmol) and [Rh(NBD)2][BArF] (26b) (109 mg, 0.09 mmol) was stirred in THF (5 mL) to yield the titled compound as a yellow solid (111 mg, 91 %), m.p. 160 – 163oC.

Anal. calcd. C55H35BF24N7Rh: C, 48.45; H, 2.59; N, 7.19. Found: C, 48.59; H, 2.37; N, 7.18

Major Product: 1 o P(3) P(3) H NMR (400 MHz, CDCl3 at -55 C) 7.75 (br, 1H, H3 or H5 ), 7.74 (br, 1H, H3P(2) or H5P(2)), 7.65 (br, 1H, H3P(3) or H5P(3)), 7.69 (s, 8H, H2BArF), 7.49 (s, 4H, H4BArF), 7.36 (t, J = 8.2 Hz, 1H, H4’), 7.14 (overlapping, 2H, H3P(3) or H5P(3) & H3P(2) or H5P(2)), 6.81 (d, J = 8.2 Hz, 1H, H5’), 6.70 (t, J = 8.2 Hz, 1H, H3’), 6.55 (br, 1H, H4P(1)), 6.40 (br, 1H, H4P(2)), 6.39 (1H, H4P(1)), 6.20 (br, 1H, H3P(3) or H5P(3)), 5.55 (d, J = 8.2 Hz, 1H, H2’), 4.00 (s, 4H, H1NBD), 3.81 (s, 2H, H2NBD), 1.34 (s, 2H, H3NBD). 13 1 o BArF C{ H} NMR (100 MHz, CDCl3 at -55 C) 161.9 (q, JBC = 74 Hz, C1 ), 145.4 (C3/5P(3) & C3/5P(2)), 143.9 (C6’), 142.9 (C3P(3) or 5P(3)), 138.3 (C3P(2) or 5P(2)), 137.2 (C3P(1) or 5P(1)), 134.8 (C2BArF), 134.6 (C4’), 128.7 (m, C3BArF), 128.9 (C3P(3) or P(3) C5 ), 128.3 (C2’), 124.5 (q, JCF = 271 Hz, CF3), 119.6 (C3’), 117.9 (C5’), 117.7 (C4BArF), 114.6 (C1’), 109.3 (C4P(1)), 109.2 (C4P(3)), 108.5 (C4P(2)), 92.7 (Ca), 63.3 (C3NBD), 61.0 (C1NBD), 51.1 (C2NBD). * (KBr): 3513w, 3424w, 3162w, 3019w, 2925w, 2925w, 2870w, 1632m, 1610m, 1519w cm-1

8.4.1.5 Synthesis of [Rh(COD)(p-tpt)][BPh4] (32).

182

Sodium tetraphenylborate (145 mg, 0.42 mmol) was added to a solution of

[Rh(COD)Cl]2 (31) (81.2 mg, 0.16 mmol), p-tris(pyrazolyl)toluidine (24a) (99.1 mg, 0.32 mmol) in THF (ca. 20 mL) and allowed to stir for one hour. The solvent of the resulting suspension was removed in vacuo. Dichloromethane (ca. 20 mL) was added and the mixture filtered through a pad of magnesium sulfate. The magnesium sulfate was washed with additional portions of dichloromethane (ca. 3 x 10 mL). The volume of the filtrate was reduced to ca. 2 mL in vacuo and n-pentane was added to precipitate the complex 32 as a bright yellow solid (226 mg, 85 %), m.p. 168 – 170oC

Anal. calcd. for C48H47BN7Rh: C, 68.99; H, 5.67; N, 11.73. Found: C, 69.07; H, 5.85; N, 11.46

1 o P(2) P(2) H NMR (400 MHz, CDCl3 at - 55 C) 7.99 (br, 1H, H3 or H5 ), 7.42 (s, 8H, H2BPh4), 7.39 (br, 2H, H3P(1) or H5P(1)), 7.17 (br, 2H, H3P(1) or H5P(1)), 7.09 – 7.00 (m, 8H, H3BPh4), 6.92 (br, 1H, H3P(2) or H5P(2)), 6.95 – 6.84 (m, 4H, H4BPh4), 6.43 (br, 1H, H4P(2)), 6.39 (br, 2H, H4P(1)), 6.28 (d, J = 8.6 Hz, 2H, H3’), 5.89 (d, J = 8.6 Hz, 2H, H2’), 4.17 (br, 2H, H1COD), 3.58 (br, 2H, H1COD), 2.48 (br, 2H, H2COD), 1.86 (br, 2H, H2COD), 1.58 (br 4H, H2COD). 13 1 o BPh4 C{ H} NMR (100 MHz, CDCl3 at -55 C) 164.5 (q, J = 49 Hz, C1 ), 151.6 (C1’), 145.2 (C3P(2) or 5P(2)), 142.7 (C3P(1) or 5P(1)), 136.1 (C2BPh4), 134.9 (C3P(1) or 5P(1)), 134.4 (C3P(2) or 5P(2)), 128.1 (C2’), 126.1 (C3BPh4), 122.1 (C4BPh4), 121.9 (C6’), 114.7 (C3’), 108.6 (C4P(2)), 107.7 (C4P(1)), 93.8 (Ca), 84.0 (C1COD), 83.2 (C1COD), 29.3 (C2COD), 25.8 (C2COD) * (KBr): 3479m, 3380s, 3215w, 3130m, 3054s, 2984m, 2918m, 2881m, 2832m, 1944w, 1892w, 1815w, 1766w, 1622s, 1605s, 1579m, 1517s, cm-1

183

8.4.2 General synthesis of rhodium carbonyl tris(pyrazolyl)toluidine complexes 29 by the displacement of the COD co-ligand.

A solution of the rhodium COD complex (27) in dichloromethane was degassed using three consecutive cycles of freeze-pump-thaw and stirred under a carbon monoxide atmosphere for 30 minutes. To this solution, n-pentane was added and the resulting precipitate filtered and washed with n-pentane.

8.4.2.1 Synthesis of [Rh(CO)2(p-tpt)][BArF] (29a).

[Rh(COD)(o-tpt)][BArF] (27a) (185 mg, 0.13 mmol). Yield (126 mg, 71 %), m.p. 105 – 107oC

+ EI-HRMS (ES ) calcd. for C18H15N7O2Rh [M]+: 464.0342. Found 464.0350; EI-HRMS + - (ES ) for C32H12F24B [BArF] : 863.0649. Found 863.0649.

1 o P(2) P(2) H NMR (600 MHz, CDCl3 at -55 C) 7.97 (br, 1H, H3 or H5 ), 7.77 (br, 2H, H3P(1) or H5P1), 7.67 (s, 8H, H3BArF), 7.48 (s, 4H, H1BArF), 7.31 (br, 2H, H3P1 or H5P(1)), 6.97 (br, 1H, H3P2 or H5P(2)), 6.57 (d, J = 7.2 Hz, 2H, H3’), 6.44 (br, 2H, H4P1), 6.35 (br, 1H, H4P2), 6.00 (d, J = 7.2 Hz, 2H, H2’) 13 1 o C{ H} NMR (150 MHz, CDCl3 at -55 C) 181.7 (CO), 181.2 (CO), 161.2 (q, J = 49 Hz, C4BArF), 150.4 (C1’), 146.7 (C3P(1) or C5P1), 145.4 (C3P(1) or C5P(2)), 135.6 (C3P(1) or C5P(1)), 134.3 (C1BArF), 133.8 (C3P2 or C5P(2)), 128.3 (C2BArF), 124 BArF P2 P1 a (CF3), 117.2 (C3 ), 114.7 (C3’), 108.4 (C4 ), 107.9 (C4 ), 93.0 (C ). * (KBr): 3492m, 3399m, 3161m, 3141m, 2107vs, 2046vs, 1630s, 1608s, 1519s cm-1

8.4.2.2 Synthesis of [Rh(CO)2(o-tpt)][BArF] (29b).

184

[Rh(COD)(o-tpt)][BArF] (27b) (171 mg, 0.12 mmol). Yield (131 mg, 80 %), m.p. 122 – 124oC

Anal. calcd. C50H27BF24N7Rh: C, 45.24; H, 2.05; N, 7.39. Found: C, 45.02; H, 2.23; N, 6.90.

Major Product: 1 o P(1) P(1) H NMR (400 MHz, CDCl3 at -55 C) 7.92 (d, J = 1.9 Hz, 1H, H3 or H5 ), 7.86 (d, J = 3 Hz, 1H, H3P(1) or H5P(1)), 7.83 (d, J = 1.9 Hz, 1H, H3P(2) or H5P(2)), 7.81 (d, J = 3 Hz, 1H, H3P(2) or H5P(2)), 7.70 (d, J = 3 Hz, 1H, H3P(3) or H5P(3)), 7.67 (br, 4H, H1BArF), 7.49 (br, 8H, H3BArF), 7.40 (t, J = 8.2 Hz, 1H, H4’), 6.73 (t, J = 8.2 Hz, 1H, H3’), 6.63 (d, J = 2.5 Hz, 1H, H5’), 6.63 (t, J = 2.5 Hz, 1H, H4P(1)), 6.50 (t, J = 2.5 Hz, 1H, H4P(2)), 6.47 (t, J = 2.5 Hz, 1H, H4P(3)), 6.44 (d, J = 2.5 Hz, 1H, H3P(3) or H5P(3)), 5.45 (d, J = 8.2 Hz, 1H, H2’). 13 1 o C{ H} NMR (150 MHz, CDCl3 at -55 C) 181.8 (CO), 181.11 (CO), 161.7 1 BArF P(1) P(1) P(2) P(2) P3 (q, J BC = 49 Hz, C4 ), 150.3 (C3 or C5 ), 150.2 (C3 or C5 ), 143.7 (C3 or C5P(3)), 143.4 (C6’), 138.8 (C3P(1) or C5P1), 137.9 (C3P(1) or C5P2), 135.3 (C4’), 134.7 (C1BArF), 129.3 (C3P3 or C5P(3)), 128.9 – 128.3 (m, C2BArF), 124.5 (q, J = 273 BArF P3 Hz, CF3), 120.2 (C3’), 118.0 (C5’), 116.4 (C3 ), 113.5 (C1’), 110.3 (C4 ), 109.9 (C4P1), 109.4 (C4P2) * (KBr): 3504m, 3415m, 3165m, 3145m, 2104vs, 2046vs, 2015m, 1632s, 1609s, 1575m, 1522m cm-1

Minor Product: 1 o H NMR (400 MHz, CDCl3 at -55 C, assigned using COSY and HMBC) 8.08 (br, 1H, H3P3 or H5P3), 8.03 (br, 1H, H3P1 or H5P1), 7.77 (br, 1H, H3P2 or H5P2), 7.80 (br, 1H, H3P2 or H5P3), 7.37 ( H4’), 7.28 (br, 1H, H3P1 or H5P1), 7.08 (br, 1H, H3P2 or H5P2), 7.28 (br, 1H, H3P1 or H5P1), 6.76 (H3’), 6.59 ( H5’), 6.51 (H4P3), 6.44 (H4P1), 6.43 (H4P2), 5.73 (d, J = 8.3 Hz, 1H, H2’). Only the aromatic region was assigned and where no integration and multiplicity are stated, resonances of the minor product were overlapped by resonances due to the major product.

185

8.5 Synthesis of rhodium(I) complexes bearing imidazolyl donor ligands.

8.5.1 Synthesis of [Rh(COD)(tim)][BArF] (36).

A solution of tris(N-methylimidazolyl)methanol (tim, 23) (26.8 mg, 0.1 mmol) and

[Rh(COD)2][BArF] (26a) (101 mg, 0.09 mmol) in THF (5 mL) was stirred at room temperature for 30 mins. The resulting pale yellow solution was evaporated to dryness under vacuum, whereupon dichloromethane (2 mL) was added and the complex crystallized as a yellow powder after the slow addition of n-pentane (81 mg, 70 %), m.p. 117 – 119oC.

Anal. calcd. C53H40BF24N6ORh: C, 47.27; H, 2.99; N, 6.24. Found: C, 47.16; H, 3.14; N, 6.40.

1H NMR (400 MHz, MeOD) 7.61 (br, 12H, H2BArF and H4BArF), 7.25 (br, 3H, H2), COD 7.05 (br, 3H, H3), 4.10 (br, 4H, H1 ), 3.76 (br, 9H, N-CH3), 2.38 – 2.21 (m, 4H, H2COD), 1.95 – 1.76 (m, 4H, H2COD) 13C{1H}s NMR (100 MHz, MeOD) 161.5 (q, J = 49 Hz, C1BArF), 145.2 (C1), 134.4 (C2BArF), 129.7 – 128.5 (m, C3BArF), 126.0 (C3), 124.9 (C2), 124.4 (q, J = 271 Hz, BArF COD CF3), 117.6 – 116.6 (m, C4 ), 82.4 (d, J = 13.3 Hz, C1 ), 35.0 (N-CH3), 29.9 (C2COD). * (KBr): 3341br, 3152w, 3125w, 3029w, 2957w, 2930w, 2894w, 2845w, 1610s, 1553m, 1505s cm-1

186

8.5.2 Synthesis of [Rh(CO)2(tim)][BArF] (37).

A solution of 36 (59.3 mg, 0.04 mmol) in dichloromethane (2 mL) was degassed using three consecutive cycles of freeze-pump-thaw and stirred under a carbon monoxide atmosphere for 30 minutes. To the solution, n- pentane was added and the resulting precipitate filtered and washed with n-pentane to yield the titled compound as a bright yellow solid (32.5 mg, 57 %), m.p. 117 – 119oC

+ + EI-HRMS (ES ) calcd. for C47H29B11N6O3F24Rh [M+H] : 1295.1066. Found 1295.1077

1H NMR (400 MHz, MeOD) 7.70 – 7.57 (m, 12H, H4BArF & H2BArF), 7.29

(d, J = 1.5 Hz, 3H, H2), 7.23 (d, J = 1.5Hz, 3H, H3), 3.75 (s, 9H, N-CH3) 13 1 C{ H} NMR (100 MHz, MeOD) 183.9 (CO), 183.2 (CO), 161.4 (q, JBC = 50 Hz, C1BArF), 144.8 (C2), 134.4 (C2BArF), 129.5 – 128.5 (m, C3BArF), 128.8 (C3), 125.5 BArF a (C2), 124.4 (q, J = 271 Hz, CF3), 117.3 – 116.9 (C2 ), 73.9 (C ), 34.7 (N-CH3). * (KBr): 3319br, 3166w, 3139w, 2094vs, 2032vs, 1610m, 1555m, 1506m cm-1

8.6 Catalyzed intramolecular cyclization of alkynoic acids – Synthesis of substrates and catalysis.

8.6.1 Synthesis of 2-heptynylmethylbenzoate (43c).

In a Schlenk flask, a suspension of iodo-2- methylbenzoate (2.55 g, 10.0 mmol), heptyne (3.0 mL, 20 mmol), cupric iodide (70 mg, 0.36 mmol) and

[Pd(PPh3)]4 (140 mg, 0.12 mmol) in triethylamine (ca. 50 mL) was heated under reflux overnight. Aqueous saturated ammonium chloride was added to the suspension and the product extracted using several portions of dichloromethane (3 x 50 mL). The combined organic

187

layers were washed with water (100 mL), dried over magnesium sulfate and the solvent removed in vacuo. The resulting residue was purified using column chromatography

(SiO2, Rf = 0.1, EtOAc / n-hexane 5 : 95 v/v) to yield the titled compound as a yellow oil (1.42 g, 63 %).

+ + EI-HRMS (ES ) calcd. for C15H19O2 [M+H] : 231.1385. Found: 231.1385

1 H NMR (300 MHz, CDCl3) 8.23 (dd, J = 7.7 Hz, 1.1 Hz, 1H, H2), 7.49 (dd, J = 7.7 Hz, 1.1 Hz, 1H, H5), 7.40 (td, J = 7.7 Hz, 1.1 Hz, 1H, H4), 7.29 (td, J = 7.7 Hz, 1.1 Hz,

1H, H3), 3.90 (s, 3H, CO2CH3), 2.45 (t, J = 7.1 Hz, 2H, H9), 1.67 – 1.60 (m, 2H, H11), 1.48 – 1.41 (m, 2H, H10), 1.40 – 1.29 (m, 2H, H12), 0.91 (t, J = 7.1 Hz, 3H, H13). 13 1 C{ H} NMR (150 MHz, CDCl3) 167.1 (CO2CH3), 134.2 (C5), 131.9 (C1), 131.5

(C4), 130.1 (C2), 127.1 (C3), 124.5 (C6), 96.1 (C8), 79.2 (C7), 52.1 (CO2CH3), 31.1 (C10), 28.4 (C11), 22.3 (C12), 19.8 (C9), 14.0 (C13). * (neat): 2954s, 2930s, 2858s, 1735s, 1597m, 1566m cm-1

8.6.2 Generalized procedure for the synthesis of terminally substituted alkynoic esters 49a–d.

In a Schlenk flask, a suspension of methyl-2-ethynylbenzoate (48), aryl halide (1.9 eq.), cupric iodide (4 mol%), triphenylphosphine (8 mol%) and [PdCl2(PPh3)2] (4 mol%) in triethylamine (ca. 25 mL) was heated under reflux for 2 days. Aqueous saturated ammonium chloride solution was added to the suspension and the product extracted using dichloromethane. The combined organic layers were washed with water, dried over magnesium sulfate and the solvent removed in vacuo. The resulting residue was purified using column chromatography to yield the alkynoic esters 49a – 49d.

188

8.6.2.1 Synthesis of 2-(4-acetylphenylethynyl)methylbenzoate (49a).

A suspension of methyl-2-ethynylbenzoate (48) (1.01 g, 6.3 mmol), cupric iodide (54.1 mg, 0.2 mmol), triphenylphosphine (138 mg, 0.53 mmol),

[PdCl2(PPh3)2] (0.25 g, 0.3 mmol) and 4- bromoacetophenone (2.33 g, 12 mmol) was heated under reflux for 2 days. The residue was purified using flash column chromatography (SiO2, EtOAc \ n-hexane 10 : 90 v/v) to yield the titled compound as an yellow solid (870 mg, 49%). m.p. 56 – 58oC

+ + EI-HRMS (EI ) calcd. for C18H14O3 [M] : 278.0943. Found: 278.0952

1 H NMR (400 MHz, CDCl3) 7.99 (dd, J = 7.8 Hz, 1.0 Hz, 1H, H2), 7.94 (d, J = 8.4 Hz, 2H, H11), 7.70 – 7.61 (overlapping, 3H, H5 & H12), 7.51 (td, J = 7.6 Hz, 1.2 Hz,

1H, H4), 7.41 (td, J = 7.8 Hz, 1.2 Hz, 1H, H3), 3.95 (s, 3H, CO2CH3), 2.60 (s, 3H,

COCH3) 13 1 C{ H} NMR (100 MHz, CDCl3) 197.4 (COCH3), 166.4 (CO2CH3), 136.4 (C12), 134.2 (C5), 131.9 (C1), 131.9 (C12 & C4), 130.6 (C2), 128.5 (C3), 128.3 (C11), 128.2

(C9), 93.3 (C8), 91.5 (C7), 52.3 (CO2CH3), 26.7 (COCH3).

8.6.2.2 Synthesis of 2-(4-chlorophenylethynyl)methylbenzoate (49b).

A suspension of methyl-2-ethynylbenzoate (48) (0.9 g, 5.6 mmol), cupric iodide (4.3 mg, 0.2 mmol),

triphenylphosphine (125 mg, 0.5 mmol), [PdCl2(PPh3)2] (0.25 g, 0.3 mmol) and 4-chloro-1-bromobenzene (2.03 g, 9.8 mmol) was heated under reflux for 2 days. The residue was purified using flash column chromatography

(SiO2, EtOAc \ n-hexane 10 : 90 v/v) to yield the titled compound as an orange solid (889 mg, 58%). Purity was confirmed by comparison with literature spectroscopic data.13

189

1 H NMR (300 MHz, CDCl3) 7.98 (dd, J = 7.8 Hz, 1.0 Hz, 1H, H2), 7.63 (dd, 1H, J = 7.8 Hz, 1.0 Hz, H5), 7.51 – 7.46 (overlapping, 3H, H10 and H4), 7.39 (dt, J = 7.8 Hz,

1.4 Hz, 1H, H3), 7.35 (m, 2H, H11), 3.95 (s, 3H, CO2CH3)

8.6.2.3 Synthesis of 2-(4-methylphenylethynyl)methylbenzoate (49c).

A suspension of methyl-2-ethynylbenzoate (48) (0.9 g, 5.6 mmol), cupric iodide (4.4 mg, 0.2 mmol),

triphenylphosphine (0.13 g, 0.5 mmol), [PdCl2(PPh3)2] (0.19 g, 0.02 mmol) and 4-bromotoluene (1.75 g, 10.2 mmol) was heated under reflux for 2 days. The residue

was purified using flash column chromatography (SiO2, EtOAc \ n-hexane 10 : 90 v/v) to yield the titled compound as a pale yellow oil (485 mg, 34%).

1 H NMR (300 MHz, CDCl3) 7.97 (dd, 1H, J = 7.8 Hz, 1.4Hz, H2), 7.64 (dd, 1H, J = 7.3 Hz, 0.9 Hz, H5), 7.50 – 7.45 (overlapping, 3H, H10 and H6), 7.35 (dt, 1H, J = 7.8

Hz, 1.4 Hz, H3), 7.18 – 7.15 (m, 2H, H11), 3.96 (s, 3H, CO2CH3), 2.37 (s, CH3). 13 1 C{ H} NMR (75 MHz, CDCl3) 166.9 (CO2Me), 138.8 (C12), 134.0 (C5), 131.7 (C10 and C4), 130.5 (C6 and C2), 129.2 (C11), 127.8 (C3), 124.0 (C1), 120.3 (C9),

94.7 (C7), 87.7 (C8), 52.2 (CO2CH3), 21.6 (CH3).

8.6.2.4 Synthesis of 2-(4-methoxyphenylethynyl)methylbenzoate (49d).

A suspension of methyl-2-ethynylbenzoate (48) (1.27 g, 8.0 mmol), cupric iodide (6.7 mg, 0.4 mmol),

triphenylphosphine (0.18 g, 0.7 mmol), [PdCl2(PPh3)2] (0.31g, 0.3 mmol) and 4-bromoanisole (2.53 g, 13.5 mmol) was heated under reflux for 2 days. The residue was purified using flash column chromatography

(SiO2, EtOAc \ n-hexane 10 : 90 v/v) to yield the titled compound as a pale yellow oil (28.2 mg, 1.3%). Purity was confirmed by comparison with literature spectroscopic data.13

190

1 H NMR (300 MHz, CDCl3) 7.96 (dd, J = 7.8 Hz, 1.0 Hz, 1H, H2), 7.62 (dd, J = 7.8 Hz, 1Hz, 1H, H5), 7.53 – 7.45 (overlapping, 3H, H4 and H10), 7.35 (dt, J = 7.8 Hz, 1.4

Hz, H3), 6.91 – 6.86 (m, 2H, H11), 3.96 (s, 3H, CO2CH3), 3.83 (s, 3H, OCH3).

8.6.3 General procedure for the hydrolysis of aromatic alkynoic acids 44 and 50.

A suspension of the ester 43 or 49 is stirred in an aqueous solution of sodium hydroxide (5mL, ca. 2M) and methanol (10 mL) for 1 – 2 hours. The methanol was removed in vacuo and the resulting solution diluted with water (20 mL) and washed with diethyl ether (3 x 50 mL). The aqueous layer was then acidified to pH 2 using hydrochloric acid (10 M) and extracted with diethyl ether (3 x 50 mL).and washed with water (3 x 20mL). The combined organic layers were dried over sodium sulfate, and solvent removed in vacuo to yield the alkynoic acid 44 or 50.

8.6.3.1 Synthesis of 2-heptynylbenzoic acid (44c).

Methyl-2-hepynylbenzoate (43c) (0.5 g, 2.17 mmol). Yield (0.23 g, 49 %).

+ + EI-HRMS (ES ) calcd. for C14H17O2 [M+H] : 217.1229; Found: 217.1228

1 H NMR (400 MHz, CDCl3) 8.04 (dd, J = 7.7 Hz, 1.0 Hz, 1H, H2), 7.53 (dd, J = 7.7 Hz, 1.0 Hz, 1H, H5), 7.46 (td, J = 7.7 Hz, 1.0 Hz, 1H, H4), 7.34 (td, J = 7.7 Hz, 1.0 Hz, 1H, H3), 2.48 (t, J = 7.0 Hz, 2H, H9), 1.67 – 1.61 (m, 2H, H10), 1.50 – 1.42 (m, 2H, H11), 1.41 – 1.33 (m, 2H, H12), 0.91 (t, J = 7.2 Hz, 3H, H13). 13 1 C{ H} NMR (100 MHz, CDCl3) 171.2 (CO2H), 134.4 (C5), 132.4 (C4), 131.1 (C2), 130.7 (C1), 127.3 (C3), 124.9 (C6), 97.6 (C8), 79.1 (C7), 31.1 (C11), 28.2 (C10), 22.2 (C12), 19.8 (C9), 14.0 (C13).

8.6.3.2 Synthesis of 2-(4-acetylphenylethynyl)benzoic acid (50a).

191

Methyl-2-(4-acetylphenylethynl)benzoate (49d) (0.3 g, 1.1 mmol). Yield (0.16 g, 56 %), m.p. 103oC (decomp.)

+ + EI-HRMS (ES ) calcd. for C12H13O3 [M+H] : 265.0865; Found: 265.0865

1 H NMR (400 MHz, CDCl3) 8.17 (dd, J = 7.9 Hz, 1.0 Hz, 1H, H2), 7.92 (m, 2H, H11), 7.73 (dd, J = 7.7 Hz, 1.0 Hz, 1H, H5), 7.66 (m, 2H, H10), 7.61 (td, J = 7.2 Hz,

1.4 Hz, 1H, H4), 7.50 (td, J = 7.7 Hz, 1.2 Hz, 1H, H3), 2.62 (s, 3H, COCH3). 13 1 C{ H} NMR (100 MHz, CDCl3) 197.4 (COCH3), 170.4 (CO2H), 134.3 (C5), 132.7 (C4), 131.9 (C10), 131.4 (C2), 130.1 (C1), 128.6 (C3), 128.3 (C11), 128.0 (C9), 123.7

(C6), 94.1 (C8), 91.0 (C7), 26.0 (COCH3).

8.6.3.3 Synthesis of 2-(4-methylphenylethynyl)benzoic acid (50c).

Methyl-2-(4-methylphenylethynyl)benzoate (49b) (0.22 g, 0.9 mmol). Yield (0.12 g, 90 %).

1 H NMR (300 MHz, CDCl3) 8.14 (dd, J = 7.9 Hz, 1.0 Hz, 1H, H2), 7.68 (dd, J = 7.3 Hz, 1.0 Hz, 1H, H5), 7.55 (dt, J = 7.7 Hz, 1.4 Hz, 1H, H4), 7.49 – 7.45 (m, 2H, H10), 7.42 (dt, J = 7.4 Hz, 1.4 Hz, 1H, H5), 7.14 – 7.11 (m, 2H, H11), 2.36 (s, 3H, H7).

8.6.4 Generalised procedure for metal catalysed intramolecular cyclization of alkynoic acids.

Catalyzed intramolecular cyclization of alkynoic acids was conducted on a small scale in NMR tubes fitted with a concentric Teflon valve. The catalyst precursor, the substrate and, if required, 2,6-dimethoxytoluene as an internal standard were dissolved in deuterated solvent. The temperature in the NMR spectrometer was calibrated using neat ethylene glycol and a K-type thermocouple. The identity of the product was confirmed

192

by comparison with literature and/or identified using standard 2D NMR spectroscopic techniques. Unless otherwise stated in the results and discussion, the values quoted are the result of a single catalysis experiment.

General numbering of the atoms within the aromatic lactones 45a-c, 46a-c, 51a-d and 52a-d are shown below.

8.6.5 Cyclization of 4-pentynoic acid (38) to 5-methylenedihydrofuranone (39).

The cyclization of 4-pentynoic acid (38) (25 mg) to the corresponding lactone 39 was investigated using a series of catalysts (2 mol %) in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy by comparing the integration of the resonances due to the alkyne proton of compound 38 relative to the appearance of the resonances due to the alkene proton of the lactone 39. The identity of the product was confirmed through comparison with literature spectroscopic data.14

193

Mass of Mass of Temperature, Time Entry Catalyst Substrate Catalyst Solvent (hrs) a [mg] [mg] d -toluene 1 29a 30.3 7.8 8 2.85 (>95%) 80oC d -toluene 2 29a 29.2 7.4 8 1.30 (78%) 110oC d -acetone 3 29a 34.7 9.1 6 19 (81%) 60oC d-chloroform 4 29a 30.0 8.1 17 (47%) 60oC 5 29b 31.2 7.2 5.55 (78%) 6 40 25.7 6.1 0.15 (>99%) d -benzene 7 40 33.7 8.3 6 0.30 (>99%) 80oC 8 37 27.8 6.0 0.47 (>99%) 9 29a 26.6 7.8 1.43 (>95%) a Percentage conversion shown in brackets.

8.6.6 Cyclization of 1-hexynoic acid (41) to 5-methylenedihydropyranone (42).

The catalysed cyclization of 5-hexynoic acid (41) (25 mg) to the corresponding lactone 42 was investigated using the complex 40 as catalyst (1 mol %) in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonances due to 2,6-dimethoxytoluene (internal standard) relative to the resonance due to the alkene proton of the lactone 42 with the integration of the resonance due to the alkyne proton relative to the internal standard at time zero. The identity of the product was confirmed through comparison with literature spectroscopic data.14

Mass of Mass of Time Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 27.5 3.3 13.3 (>99%) a Percentage conversion shown in brackets.

194

8.6.7 Cyclization of 2-ethynylbenzoic acid (44a) to isochromenone (45a) and 3-methyleneisobenzofuranone (46a).

The cyclization of 2-ethynylbenzoic acid (44a) (25 mg) to the corresponding lactones 45a and 46a was investigated using a series of catalysts (1 mol %) in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonances due to the alkyne proton of compound 44a relative to the integration of the resonances due to the alkene proton of the lactones 45a and 46a. The identities of the products in the 1H NMR spectrum were confirmed through comparison with literature spectroscopic data.13

Mass of Mass of Time 45a : 46a Entry Catalyst Substrate Catalyst Temperature (hrs)a [mg] [mg] 1 40 25.2 2.7 0.33 (>99%) 0.23 : 1 2 40 25.5 2.6 0.42 (>99%) 0.31 : 1 60oC 3 37 25.4 2.4 0.67 (>99%) 0.01 : 1 4 37 31.2 2.9 0.48 (>99%) 0.02 : 1 7 mins 5 29a 25.3 2.3 0.17 : 1 (>99%) 5 mins 6 29a 25.0 2.4 0.16 : 1 (>99%) 7 29a 27.1 2.3 1.30 (>99%) 0.26 : 1 25oC 8 29a 25.3 2.3 1.40 (>95%) 0.17 : 1 a Percentage conversion shown in brackets.

195

8.6.8 Cyclization of 2-(phenylethynyl)benzoic acid (44b) to 3-phenylisochromenone (45b) and (Z)-3-benylideneiso-benzofuranone (46b).

The cyclization of 2-(phenylethynyl)benzoic acid (44b) (25 mg) to the corresponding lactones 45b and 46b was investigated using a series of catalysts (1 mol%) in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H2 proton of compound 44b relative to the resonance due to the H8 proton of the lactone 45b and the resonance due to the H9 proton of the lactone 46b. The identities of the products observed in the 1H NMR spectrum were confirmed through comparison with literature spectroscopic data.13

Mass of Mass of Time 45b : 46b Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 24.6 1.7 0.18 (>99%) 1 : 0.10 2 37 25.6 1.4 5.97 (>99%) 1 : 0.53 3 32 25.6 1.5 0.15 (>99%) 1 : 0.04 4 29a 40.9 2.5 1.25 (>99%) 1 : 0.04 a Percentage conversion shown in brackets.

196

8.6.9 Intramolecular cyclization of 2-(heptynyl)benzoic acid (44c) to 3-pentyl-1H- isochromen-1-one (45c) and (Z)-3-hexylideneiso-benzofuranone (46c).

The cyclization of 2-(heptynyl)benzoic acid (44c) (25 mg) to the corresponding lactone 45c was investigated using a series of catalysts (1 mol%) in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H9 proton of compound 44c relative to the resonance due to the H10 proton of the lactone 45c. The identity of the product observed in the 1H NMR spectrum was confirmed through comparison with literature spectroscopic data.15

Mass of Mass of Time 45c : 46c Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 27.8 2.6 0.25 (>99%) 1 : 0 2 40 33.9 2.9 0.27 (>99%) 1 : 0 3 37 21.0 1.4 1.28 (>99%) 1 : 0 a Percentage conversion shown in brackets.

197

8.6.10 Cyclization of 2-(4-acetylphenylethynyl)benzoic acid (50a) to 3-(4- acetylphenyl)-1H-isochromen-1-one (51a) and (Z)-3-(4- acetylphenyl)isobenzofuran(3H)-one (52a)

The cyclization of 2-(4-acetylphenylethynyl)benzoic acid (50a) (25 mg) to the corresponding lactones 51a and 52a was investigated using the complex 40 (1 mol%) as catalyst in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H2 proton of compound 50a relative to the resonance due to the H8 proton of the lactones 51a and resonance due to the H9 proton of the lactones 52a. The lactones 51a and 52a were separated using column chromatography (SiO2, DCM : n-hexane 50 : 50 v/v). Structure of the two compounds was confirmed by comparing to analogous compounds published in literature,13 2D NMR experiments and high resolution mass spectrometry.

Mass of Mass of Time 51a : 52a Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 24.7 1.7 0.13 (>99%) 1 : 0.18 2 40 25.3 1.5 0.25 (>99%) 1 : 0.18 a Percentage conversion shown in brackets.

3-(4-acetylphenyl)-1H-isochromen-1-one (51a).

o Rf = 0.4 (DCM), m.p. 171 – 173 C

+ + EI-HRMS (EI ) calcd. for C17H12O3 [M] : 264.0786. Found: 264.0796.

198

1 H NMR (400 MHz, CDCl3) 8.33 (d, J = 7.4 Hz, 1H, H3), 8.04 (m, 2H, H12), 7.98 (m, 2H, H11), 7.75 (td, J = 7.7 Hz, 1.0 Hz, 1H, H4), 7.58 – 7.50 (overlapping, 2H, H5 and H6), 7.07 (s, 1H, H8), 2,64 (s, 3H, COCH3). 13 1 C{ H} NMR (100 MHz, CDCl3) 197.3 (COCH3), 161.9 (C1), 152.3 (C9), 137.8 (C13), 137.0 (C7), 136.0 (C10), 135.1 (C4), 129.8 (C3), 128.8 (C12 and C5), 126.3

(C6), 125.3 (C11), 120.9 (C2), 103.7 (C8), 26.7 (COCH3) * (KBr): 3096w, 3062w, 2963w, 2920w, 1731vs, 1676vs, 1635s, 1603s, 1559m cm-1

(Z)-3-(4-acetylphenyl)isobenzofuran(3H)-one (52a).

o Rf = 0.5 (DCM), m.p. 177 – 179 C

+ + EI-HRMS (EI ) calcd. for C17H12O3 [M] : 264.0786. Found: 264.0781.

1 H NMR (400 MHz, CDCl3) 7.99 (m, 2H, H12), 7.96 (d, J = 7.7 Hz, 1H, H3), 7.92 (m, 2H, H11), 7.80 (m, 1H, H6), 7.75 (t, J = 7.7 Hz,

1H, H3), 7.59 (t, J = 7.7 Hz, 1H, H4), 6.45 (s, 1H, H9), 2.62 (s, 3H, COCH3) 13 1 C{ H} NMR (100 MHz, CDCl3) 197.0 (COCH3), 166.1 (C1), 145.8 (C8), 139.8 (C7), 137.2 (C12), 135.7 (C13), 134.2 (C5), 129.9 (C4), 129.6 (C11), 128.3 (C12), 125.3

(C3), 123.1 (C2), 119.6 (C6), 105.2 (C9), 26.2 (COCH3). * (KBr): 2962w, 2925w, 1783vs, 1665vs, 1600s, 1560m cm-1

199

8.6.11 Cyclization of 2-(4-chlorophenylethynyl)benzoic acid (50b) to 3-(4- chlorophenyl)-1H-isochromen-1-one (51b) and (Z)-3-(4- chlorophenyl)isobenzo-furan(3H)-one (52b)

The cyclization of 2-(4-chlorophenylethynyl)benzoic acid (50b) (25 mg) to the corresponding lactones 51b and 52b was investigated using the complex 40 (1 mol%) as catalyst in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H2 proton of compound 50b relative to the resonance of the H8 proton of the lactone 51b and the resonance of the H9 proton of the lactone 52b. The identity of the major product observed in the 1H NMR spectrum was confirmed through comparison with literature spectroscopic data.17

Mass of Mass of Time 51b : 52b Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 25.0 1.6 2.13 (>95%) 1 : 0.09 a Percentage conversion shown in brackets.

200

8.6.12 Cyclization of 2-(4-methylphenylethynyl)benzoic acid (50c) to 3-(4- methylphenyl)-1H-isochromen-1-one (51c) and (Z)-3-(4-methylphenyl)- isobenzofuran(3H)-one (52c)

The cyclization of 2-(4-methylphenylethynyl)benzoic acid (50c) (25 mg) to the corresponding lactones 51c was investigated using the complex 40 (1 mol%) as catalyst in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H2 proton of compound 50c relative to the resonance due to the H8 proton of the lactone 51c and the resonance due to the H9 proton of the lactone 52c. Structure of the compound was confirmed by comparing to analogous compounds published in literature and 2D NMR experiments.13

Mass of Mass of Time 51c : 52c Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 25.3 1.5 5.68 (>90%) 1 : 0 a Percentage conversion shown in brackets.

201

8.6.13 Cyclization of 2-(4-methoxyphenylethynyl)benzoic acid (50d) to 3-(4- methoxyphenyl)-1H-isochromen-1-one (51d) and (Z)-3-(4-methoxyphenyl) isobenzofuran(3H)-one (52d)

The cyclization of 2-(4-methoxyphenylethynyl)benzoic acid (50d) (25 mg) to the corresponding lactone 51d and 52d was investigated using the complex 40 (1 mol%) as catalyst in deuterated benzene (0.5 mL). The progress of the reaction was monitored by 1H NMR spectroscopy, comparing the integration of the resonance due to the H6 proton of compound 50d to the resonance due to the of the H8 proton of the lactone 51d and the resonance due to the H9 proton of the lactone 52d. The identities of the products observed in the 1H NMR spectrum were confirmed through comparison with literature spectroscopic data.13

Mass of Mass of Time 51d : 52d Entry Catalyst Substrate Catalyst (hrs)a [mg] [mg] 1 40 29.2 1.6 3.67 (>90%) 1 : 0.15 a Percentage conversion shown in brackets.

202

8.7 X-Ray Crystallographic data

Data collection: APEX2 (Bruker, 2007, Bruker AXS Inc., Madison, USA). Cell refinement: APEX2 (Bruker, 2007, Bruker AXS Inc.,Madison, USA). Data reduction: APEX2 (Bruker, 2007, Bruker AXS Inc., Madison, USA). Program(s) used to solve structure: SHELXS-97. Sheldrick, G. M. (2008). Program(s) used to refine structure: SHELXL-97. Sheldrick, G. M. (2008). Molecular graphics: SHELXTL-Plus. Sheldrick, G. M. (2008).

8.7.1 [Rh(COD) (p-tpt)][BArF] (27a)

Crystal data

C56H39BF24N7Rh Z = 4 -3 Mr = 1379.66 Dx = 1.654 Mg m Monoclinic, P2(1)/c Mo K radiation a = 19.3288 (6) Å Cell parameters from 9928 reflections b = 16.2023 (5) Å = 2.3–27.0° c = 17.9993 (6) Å = 0.44 mm-1 = 100.600 (2)° T = 150 (2) K V = 5540.7 (3) Å3 Plates, Yellow

Data collection

Bruker kappa APEXII CCD Area7903 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.059 Absorption correction: Multi-scan max = 25.0° Tmin = 0.904, Tmax = 0.964 h = -22 22 42642 measured reflections k = -19 19 9734 independent reflections l = -21 20

Refinement

Refinement on F2 Mixture of independent and constrained H-atom refinement 2 2 2 2 R[F > 2 (F )] = 0.044 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.6624P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.146 (/ )max = 0.002 -1 S = 1.05 max = 0.91 e Å -1 9734 reflections min = -0.65 e Å 958 parameters Extinction correction: none

203

8.7.2 [Rh(NBD) (p-tpt)][BArF] (28a)

Crystal data

C58H49BF24N7Rh Z = 8 -3 Mr = 1413.76 Dx = 1.534 Mg m Monoclinic, C2/c Mo K radiation a = 40.121 (2) Å Cell parameters from 9964 reflections b = 12.7579 (7) Å = 2.3–28.9° c = 24.5524 (12) Å = 0.40 mm-1 = 102.9690 (10)° T = 150 (2) K V = 12246.9 (11) Å3 Blocks, yellow

Data collection

Bruker kappa APEXII CCD Area8605 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.053 Absorption correction: Multi-scan max = 25.0° Tmin = 0.884, Tmax = 0.956 h = -46 47 41524 measured reflections k = -13 15 10754 independent reflections l = -17 29

Refinement

Refinement on F2 H atoms constrained to parent site 2 2 2 2 R[F > 2 (F )] = 0.058 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.4249P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.187 (/ )max = 0.005 -1 S = 1.38 max = 1.71 e Å -1 10754 reflections min = -0.84 e Å 1119 parameters Extinction correction: none

204

8.7.3 [Rh(COD) (p-tpt)][BPh4] (32)

Crystal data

-3 C48H47BN7Rh Dx = 1.412 Mg m Mr = 835.65 Mo K radiation Orthorhombic, Pbca Cell parameters from 368 reflections a = 14.7482 (17) Å = 2.9–13.8° b = 21.214 (3) Å = 0.48 mm-1 c = 25.120 (3) Å T = 150 (2) K V = 7859.2 (17) Å3 Thin plates, light yellow Z = 8 0.10 × 0.05 × 0.03 mm

Data collection

Bruker kappa APEXII CCD Area1982 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.332 Absorption correction: Multi-scan max = 25.0° Tmin = 0.955, Tmax = 0.985 h = -17 15 31399 measured reflections k = -25 17 6893 independent reflections l = -29 26

Refinement

Refinement on F2 Mixture of independent and constrained H-atom refinement 2 2 2 2 R[F > 2 (F )] = 0.086 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.5632P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.264 (/ )max <0.0001 -1 S = 0.89 max = 0.74 e Å -1 6893 reflections min = -0.64 e Å 514 parameters Extinction correction: none

205

8.7.4 [Rh(COD)(o-tpt)][BArF] (27b)

Crystal data

3 C56H39BF24N7Rh V = 2799.40 (19) Å Mr = 1379.66 Z = 2 -3 Triclinic, P1 Dx = 1.637 Mg m a = 12.7104 (5) Å Mo K radiation b = 12.8002 (5) Å Cell parameters from 9420 reflections c = 18.2119 (7) Å = 2.2–27.9° = 80.389 (2)° = 0.43 mm-1 = 74.396 (2)° T = 293 (2) K = 81.914 (2)° Blocks, Yellow

Data collection

Bruker kappa APEXII CCD Area17861 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.051 Absorption correction: Multi-scan max = 25.0° Tmin = 0.791, Tmax = 0.928 h = -15 15 84955 measured reflections k = -15 15 18632 independent reflections l = -21 21

Refinement

2 2 2 Refinement on F Calculated weights w = 1/[ (Fo ) + 2 2 (0.010P) + 1.P] where P = (Fo + 2 2Fc )/3 2 2 R[F > 2 (F )] = 0.061 (/ )max = 0.010 2 -1 wR(F ) = 0.155 max = 1.43 e Å -1 S = 2.75 min = -0.96 e Å 18632 reflections Extinction correction: none 1729 parameters H atoms constrained to parent site

Note: The difference Fourier contained a peak of 4.00 eA-3 close to the carbon atom C1 (0,96 A away), after final cycles of least-squares refinement including all Non-H atoms (anisotropic) and H-atoms were carried out. This was assigned to the minor site (0.15 occupancy) of Rhodium metal ion, when the first 40 peaks in the difference Fourier were considered. This map showed a second complex molecule in different orientation to the one having a major occupancy (0.85) having almost the same location in crystal lattice. However, we considered contribution only from the second Rh atom in the further cycles of refinement, contribution from the rest of the ligand atoms was not taken into account.17

206

8.7.5 [Rh(NBD)(o-tpt)][BArF] (28b)

Crystal data 3 C55H35BF24N7Rh V = 2753.5 (3) Å Mr = 1363.62 Z = 2 -3 Triclinic, P-1 Dx = 1.645 Mg m a = 12.6013 (8) Å Mo K radiation b = 12.8203 (8) Å Cell parameters from 5429 reflections c = 18.6347 (12) Å = 2.3–23.2° = 76.517 (3)° = 0.44 mm-1 = 70.710 (3)° T = 150 (2) K = 81.166 (3)° Plates, yellow

Data collection Bruker kappa APEXII CCD Area6174 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.075 Absorption correction: Multi-scan max = 25.0° Tmin = 0.801, Tmax = 0.979 h = -14 14 34509 measured reflections k = -15 14 9605 independent reflections l = -22 22

Refinement Refinement on F2 H atoms constrained to parent site 2 2 2 2 R[F > 2 (F )] = 0.075 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.5937P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.219 (/ )max = 0.001 -1 S = 1.25 max = 1.27 e Å -1 9605 reflections min = -0.96 e Å 883 parameters Extinction correction: none

207

8.7.6 [Rh(CO)2(o-tpt)][BArF] (29b)

Crystal data

3 C50H27BF24N7O2Rh V = 2622.63 (17) Å Mr = 1327.51 Z = 2 -3 Triclinic, P-1 Dx = 1.681 Mg m a = 12.4835 (5) Å Mo K radiation b = 12.6087 (5) Å Cell parameters from 9984 reflections c = 17.5410 (6) Å = 2.3–24.3° = 89.168 (2)° = 0.46 mm-1 = 84.144 (2)° T = 293 (2) K = 72.748 (2)° Plates, Colourless

Data collection

Bruker kappa APEXII CCD Area7756 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.058 Absorption correction: Multi-scan max = 25.0° Tmin = 0.914, Tmax = 0.966 h = -14 14 39136 measured reflections k = -14 14 9213 independent reflections l = -20 20

Refinement

Refinement on F2 Riding 2 2 2 2 R[F > 2 (F )] = 0.048 Calculated weights w = 1/[ (Fo ) + 2 2 2 (0.010P) ] where P = (Fo + 2Fc )/3 2 wR(F ) = 0.109 (/ )max = 0.001 -1 S = 1.99 max = 1.20 e Å -1 9213 reflections min = -0.74 e Å 874 parameters Extinction correction: none

208

8.7.7 [Rh(COD)(tim)][BArF] (36)

Crystal data

C53H40BF24N6ORh Z = 4 -3 Mr = 1346.63 Dx = 1.642 Mg m Monoclinic, P2(1)/c Mo K radiation a = 13.7477 (4) Å Cell parameters from 9928 reflections b = 17.6434 (6) Å = 2.3–30.4° c = 22.8544 (7) Å = 0.44 mm-1 = 100.7480 (10)° T = 150 (2) K V = 5446.2 (3) Å3 Blocks, yellow

Data collection

Bruker kappa APEXII CCD Area7803 reflections with I > 2 (I) Detector diffractomer scans, and scans with offsets Rint = 0.053 Absorption correction: Multi-scan max = 25.0° Tmin = 0.891, Tmax = 0.955 h = -16 15 40489 measured reflections k = -20 20 9583 independent reflections l = -27 26

Refinement

Refinement on F2 Riding 2 2 2 2 R[F > 2 (F )] = 0.052 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.2453P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.182 (/ )max = 0.009 -1 S = 1.41 max = 1.17 e Å -1 9583 reflections min = -1.23 e Å 994 parameters Extinction correction: none

209

8.7.8 [Rh (CO)2(tim)][BArF] (37)

Crystal data

3 C94.50H57B2ClF48N12O6Rh2 V = 10347.8 (9) Å Mr = 2631.41 Z = 4 -3 a = 12.7888 (6) Å Dx = 1.689 Mg m b = 36.1479 (19) Å Mo K radiation c = 22.5803 (12) Å = 0.49 mm-1 = 90.00° T = 150 (2) K = 97.565 (3)° = 90.00°

Data collection

41717 measured reflections max = 25.0° h = -13 15 k = -41 42 l = -18 26 16617 independent reflections 7314 reflections with I > 2 (I) Rint = 0.133

Refinement

Refinement on F2 Mixture of independent and constrained H-atom refinement 2 2 2 2 R[F > 2 (F )] = 0.076 Calculated weights w = 1/[ (Fo ) + 2 2 (0.1P) + 0.6719P] where P = (Fo + 2 2Fc )/3 2 wR(F ) = 0.223 (/ )max = 0.003 -1 S = 0.97 max = 1.41 e Å -1 16617 reflections min = -0.76 e Å 1539 parameters Extinction correction: none

210

8.8 References

(1) Ottiger, M.; Delaglio, F.; Bax, A. J. Magn. Reson. 1998, 131, 373-378. (2) Tang, R. r.; Zhao, Q.; Yan, Z. e.; Luo, Y. m. Synth. Commun. 2006, 36, 2027 - 2034. (3) Zimmermann, N.; Meggers, E.; Schultz, P. G. Bioorg. Chem. 2004, 32, 13 - 25. (4) Harper, J. B., Australian National University, 2002, PhD Thesis. (5) Burckhalter, J. H.; Stephens, V. C. J. Am. Chem. Soc. 1951, 73, 56-58. (6) Moret, V.; Laras, Y.; Cresteil, T.; Aubert, G.; Ping, D. Q.; Di, C.; Barthelemy- Requin, M.; Beclin, C.; Peyrot, V.; Allegro, D.; Rolland, A.; De Angelis, F.; Gatti, E.; Pierre, P.; Pasquini, L.; Petrucci, E.; Testa, U.; Kraus, J.-L. Eur. J. Med. Chem. 2009, 44, 558-567. (7) Liu, C.-M.; Qiu, J.-J.; Bao, R.; Xu, Y.; Cheng, X.-J.; Hu, F. Polymer 2006, 47, 2962-2969. (8) Canty, A. J.; George, E. E.; Lee, C. V. Aust. J. Chem. 1983, 36, 415-418. (9) Liddle, B. J.; Gardinier, J. R. J. Org. Chem. 2007, 72, 9794-9797. (10) Guzel, B.; Omary, M. A.; Fackler, J. P.; Akgerman, A. Inorg. Chim. Acta 2001, 325, 45-50. (11) Choudhury, J.; Podder, S.; Roy, S. J. Am. Chem. Soc. 2005, 127, 6162-6163. (12) Moody, C. J.; Shah, P.; Knowles, P. J. Chem. Soc., Perkin Trans. 1 1988, 3249- 3254. (13) Marchal, E.; Uriac, P.; Legouin, B.; Toupet, L.; van de Weghe, P. Tetrahedron 2007, 63, 9979-9990. (14) Elgafi, S.; Field, L. D.; Messerle, B. A. J. Organomet. Chem. 2000, 607, 97 - 104. (15) Le Bras, G.; Hamze, A.; Messaoudi, S.; Provot, O.; Le Calvez, P.-B.; Brion, J.- D.; Alami, M. Synthesis 2008, 1607-1611. (16) Hellal, M.; Bourguignon, J.-J.; Bihel, F. J. J. Tetrahedron Lett. 2007, 49, 62-65. (17) Bhadbhade, M., Personal communication

211