UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Bicyclic octapeptide alpha-Amanitin, the death cap mushroom toxin : the total synthesis and derivatives… Matinkhoo, Kaveh 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2018_november_matinkhoo_kaveh.pdf [ 28.78MB ]
Metadata
JSON: 24-1.0372310.json
JSON-LD: 24-1.0372310-ld.json
RDF/XML (Pretty): 24-1.0372310-rdf.xml
RDF/JSON: 24-1.0372310-rdf.json
Turtle: 24-1.0372310-turtle.txt
N-Triples: 24-1.0372310-rdf-ntriples.txt
Original Record: 24-1.0372310-source.json
Full Text
24-1.0372310-fulltext.txt
Citation
24-1.0372310.ris

Full Text

  BICYCLIC OCTAPEPTIDE ALPHA-AMANITIN, THE DEATH CAP MUSHROOM TOXIN: THE TOTAL SYNTHESIS AND DERIVATIVES OF THE HYDROXYPROLINE RESIDUE by  Kaveh Matinkhoo  B.Sc., Sharif University of Technology, 2011  A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  September 2018  © Kaveh Matinkhoo, 2018 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: “BICYCLIC OCTAPEPTIDE ALPHA-AMANITIN, THE DEATH CAP MUSHROOM TOXIN: THE TOTAL SYNTHESIS AND DERIVATIVES OF THE HYDROXYPROLINE RESIDUE” submitted by                               Kaveh Matinkhoo                                in partial fulfilment for the degree of      Doctor of Philosophy in                        Chemistry. Examining Committee: David M. Perrin, Chemistry Supervisor  Raymond J. Andersen, Chemistry Supervisory Committee Member  John C. Sherman, Chemistry Supervisory Committee Member Edward Grant, Chemistry University Examiner Xiaonan Lu, Land and Food Systems University Examiner  Additional Supervisory Committee Members: N/A Supervisory Committee Member N/A Supervisory Committee Member iii  Abstract  This thesis presents the first total synthesis of the death cap mushroom toxin α-amanitin and the synthesis of its derivatives containing analogues of the hydroxyproline residue. In Chapter 2, an enantioselective route to the synthesis of (2S,3R,4R)-dihydroxyisoleucine, an unnatural oxidized amino acid found in α-amanitin, is presented. This includes the synthetic challenges that needed to be overcome, previous non-enantioselective syntheses of this amino acid, my failed attempts, and eventually the route to successfully obtain the desired enantiomer of this residue. Chapter 3 describes an unprecedented method to synthesize the unique, oxidant-sensitive 6-hydroxy-L-tryptathionine linkage. First, C-6 borylation of a suitably protected L-tryptophan was performed according to recent literature. Then, fluorocyclization of 6-boronate-L-tryptophan yielded a fluoropyrrolo indoline (Fpi) moiety that was shown to engage in the Savige-Fontana reaction with trifluoroacetic acid to furnish the 6-boronate-tryptathionine crosslink. In this synthesis, a boronate was used as a latent hydroxy group that could be revealed on the fully elaborated toxin following an oxidative deborylation reaction. In Chapter 4, the first total synthesis of α-amanitin is concluded. First, incorporation of 6-boronate-Fpi yielded a 6-hydroxy-tryptathionine crosslink. Then, the synthetic (2S,3R,4R)-dihydroxyisoleucine was introduced to the peptide sequence of α-amanitin. Following a macrolactamization step and a diastereoselective sulfoxidation of the tryptathionine thioether to the corresponding (R)-sulfoxide found in the natural product, the synthetic α-amanitin was afforded. Juxtaposition of the synthetic and authentic α-amanitins and extensive comparison of their physical, chemical and biological properties validated the synthetic analogue. iv  The analogues of trans-hydroxyproline and the method for their incorporation into α-amanitin derivatives are disclosed in Chapter 5. The hydroxyproline residue of α-amanitin has been shown to be critical for the toxicity of this toxin. However, surprisingly, there is little traction in the literature regarding the structure-activity relationships (SAR) of the hydroxyproline space and how it could affect the binding of the toxin to RNA polymerase II. Hence, a series of hydroxyproline analogues, including a photocleavable hydroxyproline derivative, were synthesized and aimed to be incorporated into amanitin via an improved solid-phase strategy.              v  Lay Summary  Alpha-amanitin, a venerable toxin extracted from the death-cap mushrooms (known as Amanita Phalloides), plays a key role in the extreme toxicity of these mushrooms. Dating back to Roman times, Amanita mushrooms have been infamously used as a means to commit murder and suicide. As claimed by ancient historians, Claudius the Roman emperor was murdered by poison, possibly found in death cap mushrooms. Owing to its exceptional potency, α-amanitin has found extensive application in recent years for cancer therapy, targeting diseased cells rather than healthy ones. Prior to this work and for the past 60 years, the only and hence limited source of this toxin was the tedious extraction from Amanita mushrooms found in nature. However, the current thesis provides synthetic access to α-amanitin and several analogues for use in the development of novel chemotherapeutic agents. This synthesis was featured in a videography by the Dean of Science at UBC (https://medium.com/ubcscience/amanita-3a3d21218322).          vi  Preface  This work was carried out under the supervision of professor David M. Perrin. In all the present chapters, following guidance from Dr. Perrin, I was responsible for the design, synthesis, and analysis of the target compounds. Furthermore, I was the only author and composer of the current manuscript.  In Chapter 2, an enantioselective method for the synthesis of (2S,3R,4R)-dihydroxyisoleucine (DHIle) was devised. Although I designed and carried out all steps for this synthetic pathway, Mr. Mihajlo Todorovic from the Perrin group assisted with the large scale synthesis of this compound. In Chapter 3, I utilized a fluorocyclization method for the synthesis of 6-boronate-fluoropyrrolo indoline. This method, in an unrelated work, had been previously initiated and investigated by Ms. Alla Pryyma from the Perrin group. However, I was responsible for expanding this strategy to the specific substrate required for this thesis and all other synthetic steps discussed in Chapter 3. In Chapter 5, Mr. Mihajlo Todorovic contributed to the design of a new solid-phase strategy for incorporation of the synthetic hydroxyproline analogues. Undergraduate students Mr. Charlie Wei and Mr. Brandon Kato, under my direct mentoring, contributed to the scale-up of various hydroxyproline analogues. Furthermore, Mr. Wei contributed to the solid-phase synthesis of amanitin analogues.  The cell viability assays were performed at the biological services laboratory (Chemistry, UBC) by Dr. Elena Polishchuk and Ms. Jessie Chen. The NMR spectra obtained on the Bruker AV-600MHz spectrometer were acquired by Dr. Maria Ezhova at the NMR facility vii  (Chemistry, UBC). Dr. Brian Patrick at the X-ray crystallography facilities (Chemistry, UBC) was responsible for acquiring the XRD structures presented throughout this thesis. The high-resolution ESI-MS spectra were obtained by Mr. Marshall Lapawa and Mr. Derek Smith at the mass spectrometry laboratory (Chemistry, UBC). Contents from Chapters 1 to 4 have been published in a single article (“Synthesis of the Death-Cap Mushroom Toxin α-Amanitin”, J. Am. Chem. Soc., 2018, 140 (21) 6513-6517). In this publication, under the supervision of Prof. Perrin, I was the lead author and responsible for almost all reaction designs, syntheses and product analyses. Ms. Alla Pryyma and Mr. Mihajlo Todorovic provided contributions through establishing the fluorocyclization methodology and scale-up of DHIle, respectively. Furthermore, the work described in Chapters 1 to 4 has been featured in a news article and a videography by the Dean of Science at UBC (https://medium.com/ubcscience/amanita-3a3d21218322).            viii  Table of Contents  Abstract .................................................................................................................................... iii Lay Summary ............................................................................................................................ v Preface ..................................................................................................................................... vi Table of Contents ................................................................................................................... viii List of Tables ........................................................................................................................ xvii List of Figures ........................................................................................................................ xix List of Symbols and Abbreviations ....................................................................................... xlii Acknowledgements ............................................................................................................. xlviii Chapter 1: Introduction ............................................................................................................. 1 1.1 What is α-Amanitin? ............................................................................................. 1 1.2 Overview of the Pharmacology of α-Amanitin..................................................... 4 1.2.1 RNA Polymerases and the Biochemistry of Transcription ............................. 5 1.2.2 Mechanism of Action of α-Amanitin .............................................................. 9 1.2.3 Key Interactions Between α-Amanitin and RNAP II .................................... 11 1.2.4 Biosynthesis of α-Amanitin .......................................................................... 14 1.3 α-Amanitin in Antibody-Drug Conjugates ......................................................... 16 1.3.1 Antibody-Drug Conjugates: A Brief Overview ............................................ 16 ix  1.3.2 Amanitin-Containing ADCs .......................................................................... 18 1.4 Derivatives of α-Amanitin .................................................................................. 21 1.4.1 Dihydroxyisoleucine Derivatives .................................................................. 22 1.4.2 Other Chemical Modifications ...................................................................... 24 1.5 Chemical Synthesis of Amanitin Analogues ...................................................... 24 1.5.1 Overall Strategy ............................................................................................ 25 1.5.2 (2S,3R,4R)-4,5-Dihydroxyisoleucine ............................................................ 26 1.5.3 6-Hydroxy-L-Tryptophan and (R)-Sulfoxide ................................................ 27 1.6 Three-dimensional Structure of α-Amanitin ....................................................... 27 1.7 Project Goals ....................................................................................................... 29 1.7.1 The Total Synthesis of α-Amanitin ............................................................... 29 1.7.2 The Proline Series ......................................................................................... 30 Chapter 2: Synthesis of (2S,3R,4R)-4,5-Dihydroxyisoleucine ............................................... 33 2.1 Introduction ......................................................................................................... 33 2.1.1 Synthetic Challenges ..................................................................................... 34 2.1.2 Synthesis of β-Branched Amino Acids ......................................................... 35 2.1.2.1 Garner’s Aldehyde ................................................................................... 36 2.1.2.2 Sigmatropic Rearrangement of the Overman’s Imidate ........................... 38 2.1.2.3 Claisen Rearrangement ............................................................................ 39 2.1.3 Previous Syntheses of Dihydroxyisoleucine ................................................. 42 x  2.1.3.1 Synthesis by Wieland ............................................................................... 42 2.1.3.2 Synthesis by Bartlett ................................................................................ 42 2.1.3.3 Previous Syntheses by Perrin and Co-workers ........................................ 46 2.1.3.3.1 Claisen Rearrangement Followed by Sharpless AD ......................... 47 2.1.3.3.2 Claisen Rearrangement and Upjohn Dihydroxylation ...................... 52 2.2 Results and Discussion ....................................................................................... 55 2.2.1 Failed Efforts Towards (2S,3R,4R)-Dihydroxyisoleucine ............................ 55 2.2.1.1 Method A: Brown Crotylation Followed by Dihydroxylation and Azide Displacement .......................................................................................................... 55 2.2.1.1.1 Brown Crotylation ............................................................................. 56 2.2.1.1.2 Method A: Synthesis ......................................................................... 59 2.2.1.1.3 Rychnovsky’s Method to Determine the Relative Stereochemistry of 1,3-Diols  ........................................................................................................... 61 2.2.1.1.4 Assigning the Relative Configuration of the 1,3-Diol in Method A . 63 2.2.1.1.5 Method A: Conclusion ...................................................................... 66 2.2.1.2 Method B: Mannich-Type Reaction Followed by Rubottom Oxidation . 66 2.2.1.3 Method C: Claisen Rearrangement Followed by Enzymatic Resolution . 71 2.2.2 Method D: Successful Synthesis of (2S,3R,4R)-Dihydroxyisoleucine: Brown Crotylation Followed by Asymmetric Strecker .......................................................... 73 2.2.2.1 Asymmetric Strecker Reactions: A Brief Overview ................................ 73 xi  2.2.2.2 Method D: Results and Discussion .......................................................... 77 2.3 Conclusion .......................................................................................................... 84 2.4 Experimental Section .......................................................................................... 85 2.4.1 Materials and Methods .................................................................................. 85 2.4.2 Experimental Procedures .............................................................................. 86 Chapter 3: Synthesis of 6-BMIDA-L-Tryptathionine ............................................................. 96 3.1 Tryptathionine Crosslinks: An Introduction ....................................................... 96 3.1.1 Synthesis of Tryptathionine Crosslinked Peptides ........................................ 97 3.1.1.1 Synthesis of Hpi ..................................................................................... 101 3.1.1.2 Characteristics of the Tryptathionine Crosslink ..................................... 104 3.2 6-Hydroxy-L-Tryptathionine: Challenges and Synthesis ................................. 105 3.3 Results and Discussion ..................................................................................... 108 3.3.1 Synthesis of 6-Methoxy-Tryptophan Followed by Demethylation............. 109 3.3.2 Oxidation of 6-BPin-L-Trp to 6-OH-L-Trp ................................................. 110 3.3.3 Oxidation of 6-BPin-L-Trp to 6-BPin-Hpi .................................................. 113 3.3.4 Electrophilic Fluorocyclization: An Alternative Oxidation Method .......... 115 3.3.4.1 FP-T300: A Mild Reagent to Induce Fluorocyclization ........................ 116 3.3.5 Fluorocyclization of 6-Boronate-L-Tryptophan .......................................... 119 3.4 Conclusion ........................................................................................................ 123 3.5 Experimental Section ........................................................................................ 125 xii  3.5.1 Materials and Methods ................................................................................ 125 3.5.2 Experimental Procedures ............................................................................ 126 Chapter 4: The Total Synthesis of α-Amanitin ..................................................................... 137 4.1 Introduction ....................................................................................................... 137 4.1.1 Previous Syntheses of Amanitin Analogues ............................................... 137 4.1.1.1 Synthesis of Amanitin Analogues by Perrin Lab ................................... 140 4.2 Results and Discussion ..................................................................................... 142 4.2.1 Synthesis of the Monocyclic Heptapeptide ................................................. 142 4.2.1.1 UV-Vis Absorbance Trends in 6-Substituted Tryptathionines .............. 145 4.2.2 Synthesis of the Monocyclic Octapeptide ................................................... 147 4.2.3 The Base-promoted Oxidation By-product ................................................. 149 4.2.3.1 Mechanism of Formation ....................................................................... 149 4.2.3.2 Preliminary Characterization of the Oxidation By-product ................... 150 4.2.4 Macrolactamization to S-Deoxy-α-Amanitin .............................................. 153 4.2.5 Asymmetric Sulfoxidation of S-Deoxy-α-Amanitin ................................... 154 4.2.5.1 Asymmetric Sulfoxidation: A Brief Introduction .................................. 154 4.2.5.2 Previous Efforts Towards the Sulfoxidation of S-Deoxy-α-Amanitin ... 157 4.2.5.3 3D Structures of Sulfoxides and Sulfone of α-Amanitin ....................... 159 4.2.5.4 Asymmetric Sulfoxidation: Results and Discussion .............................. 161 4.2.5.4.1 Oxidant Screening ........................................................................... 163 xiii  4.2.5.4.2 mCPBA Oxidation: Solvent and Temperature Screening ............... 165 4.2.5.4.3 Kinetic Resolution of (R)- and (S)-Sulfoxides via Oxidation to Sulfone  ......................................................................................................... 167 4.2.5.5 Rationalization of the Sulfoxidation Diastereoselectivity ...................... 170 4.2.5.6 Rationalization of the Kinetic Resolution of (R)- and (S)-Sulfoxides ... 172 4.2.6 Characterization of the Synthetic α-Amanitin ............................................ 174 4.2.6.1 HPLC and UV Analysis of the Synthesized (R)- and (S)-Sulfoxides .... 174 4.2.6.2 Circular Dichroism ................................................................................. 176 4.2.6.3 NMR Characterization ........................................................................... 177 4.2.6.4 Cell Viability Assays .............................................................................. 182 4.3 Conclusion ........................................................................................................ 184 4.4 Experimental Section ........................................................................................ 186 4.4.1 Materials and Methods ................................................................................ 186 4.4.2 Experimental Procedures ............................................................................ 191 Chapter 5: Analogues and Derivatives of trans-Hydroxyproline Residue ........................... 198 5.1 Introduction ....................................................................................................... 198 5.1.1 Conformational Properties of 4-Substituted Prolines in Peptides ............... 200 5.2 Analogues of trans-Hydroxyproline for Incorporation in Amanitin ................ 202 5.2.1 Design and Selection of (4R)-Hydroxyproline Analogues ......................... 202 5.2.2 Synthesis of (4R)-Hydroxyproline Analogues ............................................ 205 xiv  5.2.2.1 Previous Work on Various Modified Prolines ........................................ 205 5.2.2.2 Results and Discussion ........................................................................... 207 5.3 Photocleavable Hydroxyproline Analogue ....................................................... 212 5.3.1 Photocleavable Protecting Groups: A Brief Introduction ........................... 213 5.3.2 ortho-Nitrobenzyl Photocleavable Protecting Groups ................................ 216 5.3.3 Synthesis of a Photocleavable Hydroxyproline .......................................... 219 5.4 Improved Solid-Phase Strategies for Incorporation of Hydroxyproline Derivatives in Amanitin ................................................................................................ 221 5.4.1 Strategy A: Loading the CTC Resin with Fmoc-Asp(OH)-OAllyl ............ 222 5.4.2 Strategy B: Loading the CTC Resin with Fmoc-Cys(SH)-OAllyl .............. 226 5.4.3 Strategy C: Loading the CTC Resin with Fmoc-Cys(STrt)-OH, Followed by Macrolactamization in the Solution Phase ................................................................ 230 5.4.3.1 Formation of an Oxindole By-product Upon Exposure to TFA ............ 232 5.5 Future Work: Incorporation of Proline Derivatives into Amanitin Analogues 233 5.6 Conclusion ........................................................................................................ 235 5.7 Experimental Section ........................................................................................ 237 5.7.1 Materials and Methods ................................................................................ 237 5.7.2 Experimental Procedures ............................................................................ 238 Chapter 6: Conclusion and Future Directions ....................................................................... 265 6.1 Stereoselective Synthesis of (2S,3R,4R)-4,5-Dihydroxyisoleucine .................. 265 xv  6.2 Synthesis of 6-BMIDA-L-Tryptathionine ........................................................ 266 6.3 The Total Synthesis of α-Amanitin ................................................................... 269 6.4 Derivatives of trans-Hydroxyproline ................................................................ 273 6.5 Future Directions .............................................................................................. 276 Bibliography ......................................................................................................................... 278 Appendix A: Supplementary Information for Chapter 2 ...................................................... 299 A.1 NMR Spectra .................................................................................................... 299 A.2 X-Ray Data and Structure Refinement for Lactone 134 ................................... 318 Appendix B: Supplementary Information for Chapter 3 ...................................................... 326 B.1 NMR Spectra .................................................................................................... 326 Appendix C: Supplementary Information for Chapter 4 ...................................................... 344 C.1 HPLC Chromatograms ..................................................................................... 344 C.1.1 Intermediates for the total synthesis of α-amanitin ..................................... 344 C.1.2 Kinetic Resolution of (R)- and (S)-Sulfoxides via Oxidation ..................... 356 C.1.3 Sulfoxidation Trials on S-Deoxy-Amanitin (217) ....................................... 358 C.1.4 Oxidation By-product (216) ........................................................................ 365 C.2 NMR Spectra for S-Deoxy-Amanitin, (S)-α-Amanitin and Synthetic/Authentic α-Amanitins ...................................................................................................................... 367 C.3 Raw Data for Cell Toxicity Assays .................................................................. 377 xvi  C.3.1  MTT assay of synthetic α-amanitin, α-amanitin-thioether (217), α-amanitin-(S)- sulfoxide (236) and authentic α-amanitin, against CHO cells. .......................... 377 C.3.2 MTT assay of authentic α-amanitin and the oxidation by-product (216), against CHO cells. .................................................................................................... 379 Appendix D: Supplementary Information for Chapter 5 ...................................................... 380 D.1 NMR Spectra .................................................................................................... 380 D.2 HPLC Chromatograms ..................................................................................... 417                xvii  List of Tables  Table 1.1 Different amatoxins isolated from Amanita mushrooms. The Ki values were measured against RNAP II of calf thymus, and the LD50 values were obtained by testing on white mice.6 .............................................................................................................................. 2 Table 1.2 Structure of heptapeptides extracted from A. Phalloides: A) Phallotoxins and B) Virotoxins. ................................................................................................................................ 3 Table 1.3 Different RNA polymerases produced in different types of cells, listing their products and binding affinity to α-amanitin. ........................................................................................... 6 Table 1.4 Analogues of α-amanitin derivatized at DHIle3 position and their inhibition constants relative to α-amanitin (against calf thymus RNAP II) or γ-amanitin (against D. melanogaster embryos RNAP II).7,28 Note that all the analogues lack the 6’-hydroxy of Trp4 and the sulfoxide. ................................................................................................................................. 23 Table 1.5 Important analogues of α-amanitin modified at positions 1, 5, 6 and 7. The inhibition constants were measured against calf thymus RNAP II (19 to 22) or D. melanogaster RNAP II (23 to 25).28,29,45 Note that all the analogues contain Ile at position 3 and not DHIle. (n/a = no activity). ............................................................................................................................. 24 Table 2.1 Different classes of 1,3-diol acetonides and the list of  respective 13C-NMR chemical shifts for methyl groups and the acetal C(2). Values are in ppm as the mean ± std deviation (from Rychnovsky et al).82 ..................................................................................................... 63 Table 2.2 13C-NMR chemical shifts for 109, syn- and anti-class 4 acetonides. Representative chemical shifts for 109 matched the syn structural type proposed by Rychnovsky. .............. 65 xviii  Table 4.1 Qualitative results of the oxidant screening for sulfoxidation of 217. Reactions were performed on 10 nmol of 217 with 1.3 equivalents of oxidant at 21ºC. Composition of the product mixture was assessed by HPLC and MS. aTFA was used as additive. Minor: 5-20%. Major: 50-100%. N/A: not applicable due to the low amount of the desired product. Oxaziridine 227: (1R)-(-)-(10-camphorsulfonyl)oxaziridine. ............................................... 164 Table 4.2. Sulfoxidation of 217 in different alcohol-based solvent systems and composition of the product mixture. Reaction conditions: 217 (3.3 nmol), mCPBA (1 eq), RT, 1-2h. ........ 168 Table 4.3 Proton chemical shifts for various derivatives/samples of amanitin. 1a: synthetic α-amanitin. 1b: authentic α-amanitin (purchased from Sigma-Aldrich). 1c: reported chemical shifts for α-amanitin from Shoham et al. 236: synthetic α-amanitin (S)-sulfoxide. 232: reported chemical shifts for O-Me-α-amanitin (S)-sulfoxide from Shoham et al. 217: synthetic α-amanitin-thioether; 233: reported chemical shifts for O-Me-α-amanitin-thioether from Shoham et al. Important/distinctive chemical shifts have been highlighted with gray shading and are featured in bold font. Chemical shifts that are not reported could not be observed due to the H2O/solvent peaks in the 1H-NMR spectra and, additionally, could not be analyzed using COSY60 experiments (99.96% DMSO-d6, 600MHz). ......................................................... 179 Table 5.1 Amanitin analogues containing Pro or Hyp at position-2 and their inhibitory constants relative to α-amanitin against RNAP II from Calf Thymus (Ki for α-amanitin ~ 3 nM). ....................................................................................................................................... 199 Table 5.2 Calculated bond and dihedral angles for Hyp and proposed analogues for incorporation into amanitin. Angles were calculated following the MMFF-optimization of the equilibrium conformer of each analogue in a tetrapeptide model system (shown in the box). Analogues K and L can be seen in the dashed box. .............................................................. 204 xix  List of Figures  Figure 1.1 A) The structure and amino acid numbering of the bicyclic octapeptide α-amanitin. B) The simplified scaffold showing the abbreviations used throughout this thesis. ................ 1 Figure 1.2 A schematic of the central dogma of molecular biology. The blue arrows represent the processes that are common in all organisms, while the red arrows show the pathways that only occur in certain organisms. ............................................................................................... 5 Figure 1.3 Crystal structure of RNAP II in complex with DNA and RNA at 3.3Å resolution (reproduced with permission from Kornberg et al.).16 A) Comparison of structures of free RNAP II (top) and the RNAP II transcribing complex (bottom). The clamp (yellow) closes on DNA and RNA, which are bound in the cleft above the active site. The remainder of the enzyme is in gray. B) Structure of the RNAP II transcribing complex. Portions of Rpb2 that form one side of the cleft are omitted to reveal the nucleic acids on the DNA (green and blue) and RNA (red) ribbons. The Rpb1 bridge helix traversing the cleft is highlighted in green. The active site metal “A” is shown as a pink sphere. ...................................................................... 7 Figure 1.4 A) Schematic of RNAP II transcription initiation. TFIID or its TATA box-binding protein (TBP) subunit binds to promoter DNA. The TBP-DNA complex is stabilized by TFIIA and TFIIB. The resulting upstream promoter complex is joined by RNAP II-TFIIF complex, leading to the formation of the core PIC. Subsequent binding of TFIIE and TFIIH complete the PIC (closed PIC). In the presence of ATP, the DNA is opened (forming the transcription bubble) and RNA synthesis begins. Finally, dissociation of initiation factors enables the formation of the elongation complex accompanied by the elongation factors (reproduced with permission from Sainsbury et al.).17 ......................................................................................... 9 xx  Figure 1.5 Location of α-amanitin bound to RNAP II (from Kornberg et al.). A) Cutaway view of α-amanitin (red dot) bound to RNAP II and its location relative to the nucleic acids of the DNA and RNA and the bridge helix. B) representation of α-amanitin and RNAP II co-crystal using ribbons. Eight zinc atoms are in light blue; the active site magnesium ion is in magenta; the part of Rpb1 near amanitin (funnel) is in light green; the region of Rpb2 near amanitin is in dark blue; the bridge helix is in dark green; α-amanitin is in red.19 ................................... 10 Figure 1.6 The cryo-EM structure of mammalian RNAP II bound to α-amanitin (3.4Å resolution). RNAP II is shown in a silver ribbon model (from Cramer et al.).24 ................... 11 Figure 1.7 Important interactions between α-amanitin and RNAP II. A) stereoview of the α-amanitin binding pocket (4Å). Red bonds represent α-amanitin. The funnel region of Rpb1 is in light green. Bridge helix can be seen in dark green. The region of Rpb2 near amanitin is in blue. B) the chemical structure of α-amanitin and the residues on the enzyme with a distance less than 4Å. Hydrogen bonds are shown in dashed lines with their lengths indicated. ........ 12 Figure 1.8 Interactions of mammalian RNAP II with α-amanitin (reproduced with permission from Cramer et al)24. A) Schematic of the interactions. Chemical structure of α-amanitin is in orange. Rpb1 residues conserved over eukaryotes are labeled in black, whereas metazoan-specific amanitin-interacting residues are labeled in red. Green dashed lines represent hydrogen bonds, while other interactions are shown in black lines. B) Representation of the α-amanitin binding pocket within RNAP II. Blue and red colors show positively and negatively charged surfaces, respectively. The bridge helix, trigger loop, and Ser782 residue of Rpb1 can be seen. .................................................................................................................................... 13 Figure 1.9 A) The proproteins of α-amanitin and phallacidin encoded by the identified genes, Ama1 and Pha1, in A. bisporigera.11 The residues in italic font are conserved in both genes, xxi  and the bold amino acids are specific to the toxin. Residues shown in red are responsible for the tryptathionine formation. B) Different steps of the toxin maturation catalyzed by GmPOPB, a specific prolyl oligopeptidase from G. marginata (figures adapted from Walton et al.).32 . 15 Figure 1.10 A) A general ADC and its components. B) Mechanism of action of an ADC (figure adapted from Chen et al.).37 .................................................................................................... 17 Figure 1.11 ADC of α-amanitin and chimeric antibody chiHEA125 (reproduced with permission from Moldenhauer et al.). A) Structure (left) and binding of chiHEA125-amanitin conjugate to EpCAM (right). Colo205 cells were incubated with various concentrations of unconjugated chiHEA125 and the α-amanitin conjugate and analyzed by flow cytometry. Experiments were performed in triplicates. B) Efficiency of a single dose (50 µg/kg) of chiHEA125-amanitin in mice (n=6) bearing 10-day-old pancreatic carcinoma xenografts. Control mice (n=4) received unconjugated chiHEA125. Left: relative tumor volume vs. days post-treatment. Tumor growth was measured every third day. Right: representative tumors at the endpoint of treatment (day 16); magnification: x2.5 C) Efficiency of two doses of chiHEA125-amanitin, administered one week apart, in mice bearing 10-day-old xenografts. Mice received IP injections of the ADC (10, 20, 50 or 100 µg/mg; n=8, 7, 10, 10 mice per group, respectively) or control unconjugated chiHEA125 (n=9). Tumor growth was measured every third day. ....................................................................................................................... 20 Figure 1.12 Possible sites of conjugation on α-amanitin. A) 5-hydroxyl of DHIle3, B) 6-hydroxyl of Trp4, C) side-chain amide of Asn1 via CuAAC. ................................................. 21 Figure 1.13 Derivatives of α-amanitin obtained from the diol cleavage at the DHIle3 position.48 ................................................................................................................................................ 23 xxii  Figure 1.14 General conditions to form a tryptathionine crosslink. Oxidation of tryptophan followed by the Savige-Fontana reaction using TFA to induce tryptathionylation.52 ............ 25 Figure 1.15 General approach for tryptathionylation in the synthesis of amanitin analogues, followed by coupling of the residue as position-3 (Xaa3) and macrolactamization to yield the final amanitin analogue. .......................................................................................................... 26 Figure 1.16 XRD structures of A) β-amanitin (Kostansek et al.), B) O-Me-α-amanitin sulfone (Shoham et al.), and C) Pro2-Ile3-S-deoxo-amaninamide (Perrin et al.). ............................... 28 Figure 1.17 Overview of the synthesis of Hyp analogues and their incorporation into amanitin to synthesize the corresponding analogues. ............................................................................ 30 Figure 1.18 Overview of a photo-cleavable amanitin. The protected amanitin is not cytotoxic; however, its irradiation with an appropriate wavelength will release the unprotected, cytotoxic amanitin. ................................................................................................................................. 31 Figure 2.1 Chemical structures of DHIle and its lactone. A) lactonization of the free amino acid in the presence of acid. C-2 may be epimerized to produce the more stable trans-trans isomer 28. B) lactonization of DHIle3 (in blue) on α-amanitin that leads to the cleavage of the toxin. ....................................................................................................................................... 35 Figure 2.2 Synthesis of Garner’s aldehyde (29) (Garner et al.). The corresponding D-29 may be obtained starting with D-serine, eventually resulting in the formation of a β-branched D-amino acid. .............................................................................................................................. 36 Figure 2.3 Synthesis of syn-β-branched amino acids using the Garner’s aldehyde via Michael addition. The Falkin-Ahn model for the si-face addition of the nucleophile to form syn-31 is shown in the dashed box. ........................................................................................................ 37 xxiii  Figure 2.4 Synthesis of anti-β-branched amino acids starting with the Garner’s aldehyde via a chiral bromo allene (37).65 ...................................................................................................... 38 Figure 2.5 Synthesis of (2S,3S,4R)-γ-hydroxyisoleucine (49) via [3,3]-sigmatropic rearrangement of imidate 46.68 The transition state for the palladium-catalyzed rearrangement of 46 to 47 is shown in the dashed box. .................................................................................. 39 Figure 2.6 Ester-enolate Claisen rearrangement of N-TFA-glycine (E)-crotyl esters (50) for the synthesis of syn-γ,δ-unsaturated-β-branched amino acids (51). The D-isomer may be obtained when quinine is used in place of quinidine.69 ......................................................................... 40 Figure 2.7 Synthesis of anti-β-branched γ,δ-unsaturated amino acids via an Eschenmoser-Claisen rearrangement.70 ......................................................................................................... 41 Figure 2.8 Non-stereoselective synthesis of DHIle lactone by Wieland et al. ....................... 42 Figure 2.9 Retrosynthetic scheme for the stereoselective synthesis of DHIle (Bartlett et al.). ................................................................................................................................................ 43 Figure 2.10 Synthesis of dehydro-allo-isoleucine starting from trans-crotyl ester of Gly (64). ................................................................................................................................................ 43 Figure 2.11 The anionic intermediates and the chair-like transition state for the Claisen rearrangement of trans-crotyl ester of N-Boc-glycine. ........................................................... 44 Figure 2.12 Diastereoselective synthesis of the DHIle lactone (Bartlett et al.). ..................... 46 Figure 2.13 Synthesis of L-dehydroisoleucine indoline amide (75) by Dietrich. ................... 47 Figure 2.14 Four possible products and their observed diastereomeric ratios for dihydroxylation trials of D/L-75 using different reagents (the desired enantiomer is shown in blue). ....................................................................................................................................... 48 xxiv  Figure 2.15 Different dihydroxylation trials on Cbz (75) and Boc (78) protected L-dehydroisoleucine using different reagents and the diastereomeric ratios of the obtained products (desired diastereomers are shown in blue). .............................................................. 50 Figure 2.16 Diastereomeric ratios of the obtained products 82 and 83 from dihydroxylation reactions of 76/77 and 79/80 using (DHQD)2Phal or (DHQD)2Pyr chiral ligands. ............... 51 Figure 2.17 Analysis of the enantiomeric ratio for AD reactions using chiral ligands with phthalazine- (Phal) or pyrimidine-based (Pyr) linkers. .......................................................... 52 Figure 2.18 Synthetic pathway to obtain four diastereomers of DHIle, D/L-87 and D/L-88 (Zhao and Perrin). The desired diastereomer is shown in the box. ................................................... 53 Figure 2.19 Incorporation of a mixture of four DHIle diastereomers into an amanitin analogue and HPLC separation of the cytotoxic diastereomer containing (2S,3R,4R)-DHIle. ............. 54 Figure 2.20 General retro-synthetic scheme for the synthesis of DHIle starting with the side chain diol. ................................................................................................................................ 56 Figure 2.21 Synthesis of (+)-B-methoxydiisopinocampheyl borane, (+)-91, by Brown.80,81 . 57 Figure 2.22 Synthesis of erythro-95a using (E)-butene and (+)-(Ipc)2BOMe (Brown and Bhat).79 .................................................................................................................................... 57 Figure 2.23 Brown crotylation and the ability to obtain any of the four possible diastereomers of β-methyl homoallyl alcohols as the major product using (E)- or (Z)-butene with (+)- or (-)-(Ipc)2BOMe. ........................................................................................................................... 58 Figure 2.24 Chair-like transition states leading to four different diastereomers of the Brown crotylation.79 ........................................................................................................................... 59 xxv  Figure 2.25 Synthesis of a diastereomeric mixture (d.r. 6:1 in favor of an unknown diastereomer) of the protected tetrol intermediate 99. ............................................................ 60 Figure 2.26 Synthesis of a stereoisomer of DHIle (106) from protected tetrol 100. .............. 61 Figure 2.27 Major conformations of A) syn-1,3-diol acetonide and B) anti-1,3-diol acetonide. The numbering of acetonide carbons is shown. ...................................................................... 62 Figure 2.28 Synthesis of acetonide 109 following the isolation of its major diastereomer. The proposed synthesis of the correct enantiomer of DHIle from the desired trans-1,3-diol is shown in the dashed box. ................................................................................................................... 64 Figure 2.29 Dihydroxylation from re- and si-face and consequent isomers of DHIle expected in each case. ............................................................................................................................ 65 Figure 2.30 General scheme for the Mannich reaction involving three components: a carbonyl donor, an amine, and an acceptor aldehyde. ........................................................................... 67 Figure 2.31 Proline-catalyzed Mannich-type reaction for the synthesis of functionalized α-amino acids. A) General scheme. B) Mechanism.85 ............................................................... 67 Figure 2.32 Synthesis of MHIle via a proline-catalyzed Mannich-type reaction (Zhao and Perrin, unpublished). ............................................................................................................... 68 Figure 2.33 Trials to produce the silyl enol ether of 113 using different bases and silyl chlorides. LDA and TMSCl yielded a 3:2 mixture of 117 and 116. ....................................... 69 Figure 2.34 Oxidation trials of 116/117 for the synthesis of 118. In all cases the starting ketone (113) was recovered. ............................................................................................................... 69 Figure 2.35 Efforts towards bromination of 116/117. A) Trimethylphenylammonium tribromide resulted in minor amounts of products containing various numbers of bromine. xxvi  Almost all of the starting ketone was recovered. B) Br2 led to the formation of the desired product (119) (10%) and the bis-bromo compound 120 (20%), and 70% of the starting ketone was recovered. ........................................................................................................................ 70 Figure 2.36 Selective hydrolysis of N-Cbz-L-Nle-OMe and N-Cbz-L-Nva-OMe using alcalase from Bacillus Licheniformis. Reaction conditions: alcalase (2.5 AU/mL), tBuOH/H2O 19:1, pH~8.2, 35°C.86 ...................................................................................................................... 71 Figure 2.37 Efforts toward the enzymatic hydrolysis of erythro/threo-L-dehydroisoleucine methyl esters using the alcalase from Bacillus Licheniformis. The initial mixture of diastereomers was recovered with no hydrolysis observed. The desired enantiomer is shown in blue. .................................................................................................................................... 72 Figure 2.38 Retrosynthetic scheme for the synthesis of (2S,3R,4R)-DHIle showing the bond formation sites required for each major step (method D). R1, R2 = protecting group. ........... 73 Figure 2.39 Main components of the Strecker reaction. Acid hydrolysis of the resulting α-amino nitrile affords an α-amino acid. .................................................................................... 74 Figure 2.40 The Felkin-Ahn model for the nucleophilic addition of cyanide to an iminium in the Strecker reaction. In this example, L-amino nitrile is the favored product (shown in blue) (S=small, M=medium, L=large). ............................................................................................ 75 Figure 2.41 (S)-α-phenylethylamine as a chiral auxiliary for the asymmetric Strecker reaction. According to the Felkin-Ahn model, what once was the favored enantiomer of the amino nitrile product in the non-asymmetric reaction, is now the unfavored isomer due to the steric bulk added to the imine nitrogen by the phenyl group (S=small, M=medium, L=large, blue circle: nitrogen). ................................................................................................................................. 76 xxvii  Figure 2.42 Examples of other common chiral auxiliaries used in asymmetric Strecker reactions: A) (R)-phenylglycine,88 B) (R)-α-phenylglycinol,91 C) (S)- p-toluene sulfinamide,94 D) galactosylamine,89 and E) (S)-1-amino-2-methoxymethyl indoline.95 .............................. 77 Figure 2.43 Synthesis of (2S,3R,4S)-4-hydroxyisoleucine by Potier et al.96 .......................... 78 Figure 2.44 A) Synthesis of amino nitriles (1S)- and (1R)-132. B) The Felkin-Ahn model for the non-asymmetric Strecker reaction of an imine of 131, showing the mismatched reaction. The favored facial selectivity affords the undesired diastereomer, while the unfavored selectivity leads to the desired diastereomer. .......................................................................... 79 Figure 2.45 Acid hydrolysis of (1S)-132 to lactone 133. Formation of the imidate and the deprotection of TBS and Bn took place during this reaction. ................................................. 80 Figure 2.46 Selected region of 1H-NOESY NMR of 133 showing the correlations between cis substituents (400 MHz, CD2Cl2). ............................................................................................ 80 Figure 2.47 Removal of the chiral auxiliary to obtain the HCl salt 134. The XRD structure is shown in the box, confirming the configuration of the stereogenic centers to match the desired enantiomer. ............................................................................................................................. 81 Figure 2.48 Saponification of 134 to the free-chain DHIle 135. In acidic pH, 135 converts back to the lactone 134, hence avoiding pH lower than 8 in the work-up. ..................................... 82 Figure 2.49 A) First trial for the TBDMS protection of the free-chain DHIle with TBDMSCl and imidazole in DMF. B) proposed mechanism for the formylation of the amine to yield by-product 137 as the major product............................................................................................ 83 Figure 2.50 Synthesis of the NHS-activated ester of Nα-Fmoc-bis-TBS-DHIle (139) from 135. ................................................................................................................................................ 84 xxviii  Figure 3.1 Examples of bicyclic peptides in nature. ............................................................... 96 Figure 3.2 General structure of the tryptathionine crosslink formed between tryptophan and A) cysteine (in amatoxins and phallotoxins), and B) any thiol-containing residue. .................... 97 Figure 3.3 Synthesis of norphalloin (140) by Wieland et al. Sulfenyl chloride of cysteine was reacted with tryptophan to form the tryptathionine linkage (shown in blue). ........................ 97 Figure 3.4 Synthesis of Hpi by Savige.102 .............................................................................. 98 Figure 3.5 Original reaction between Hpi and cysteine in the presence of TFA to yield the tryptathionine crosslink in 80% yield (Savige and Fontana). ................................................. 98 Figure 3.6 Synthesis of amatoxin analogues by Zanotti et al.45 Hpi, S-Trt-Cys and the resulting tryptathionine are shown in blue. X and Y (in red) were the different residues employed in the synthesis to yield different analogues. .................................................................................... 99 Figure 3.7 Proposed mechanism for the Savige-Fontana reaction. ...................................... 100 Figure 3.8 Synthesis of tryptathionine via a dihydropyrrolo indoline species. .................... 100 Figure 3.9 Formation of a tryptathionine crosslink via the reaction of tryptophan with mercaptoethanol in the presence of Hg(OAc)2 and acetic acid.110 ....................................... 101 Figure 3.10 Different methods to produce Hpi: A) mild oxidation of tryptophan, B) photosensitized oxidation, C) oxidative deselenation (N-PSP: N-phenylselenophthalimide). .............................................................................................................................................. 102 Figure 3.11 Proposed mechanism for the DMDO oxidation of tryptophan to Hpi. ............. 103 Figure 3.12 Oxidation of tryptophan to Hpi using the in situ formation of DMDO (Blanc et al.). ........................................................................................................................................ 104 xxix  Figure 3.13 General UV absorption curve for tryptathionine with λmax=290 and shoulders at 285 and 300 nm.51 ................................................................................................................. 104 Figure 3.14 Circular dichroism (CD) spectra of α-amanitin in water at pH 7 and pH 12 (phenolate of 6-OH-L-Trp) (reproduced from Wieland).6 .................................................... 105 Figure 3.15 A) General strategy for the synthesis of an unsubstituted tryptathionine crosslink via oxidation of tryptophan to Hpi followed by the Savige-Fontana reaction. B) Proposed retrosynthetic scheme for the synthesis of a 6-hydroxy-Ttn from 6-hydroxy-L-Trp. ........... 106 Figure 3.16 Reported synthesis of 6-hydroxy-L-tryptophan (146) using H2O2 and superacid. .............................................................................................................................................. 107 Figure 3.17 Flow of electrons from the 6-OH group into the indole ring, rendering C-2 extremely susceptible to oxidation. ...................................................................................... 107 Figure 3.18 Hypothetical retrosynthetic scheme for the DMDO oxidation of 146 to the corresponding Hpi (149) followed by its proposed incorporation into α-amanitin (Ttn shown in blue). Manipulation of protecting groups might be necessary prior to incorporation in amanitin. ............................................................................................................................... 108 Figure 3.19 Part of the total synthesis of verruclogen and fumitremorgin A by Baran et al. A ligand-controlled C-H borylation of C-6 of tryptophan followed by a Chan-Lam coupling with MeOH were employed. The 6-methoxy-indole core is shown in red. .................................. 109 Figure 3.20 Proposed demethylation of 6-methoxy-tryptophan to obtain the protected 6-hydroxy-tryptophan. ............................................................................................................. 109 Figure 3.21 Failed attempts to convert 154 to 6-hydroxy-L-Trp products (155-157). .......... 110 xxx  Figure 3.22 Oxidation of 4-BPin-L-Trp to 4-OH-L-Trp (159) with sodium perborate (Bartoccini et al.). .................................................................................................................................... 111 Figure 3.23 TIPS-deprotection of 152 followed by the oxidation of BPin to OH using sodium perborate. .............................................................................................................................. 111 Figure 3.24 Oxidation of 161 with A) DMDO formed in situ, B) distilled DMDO (0.06 M in acetone). Crude product contained the desired 6-OH-Hpi (162) as the minor component and the corresponding oxindole (163) as the major product. Unidentified oxidation by-products comprised the rest of the crude mixture. The distinctive protons that were closely studied by NMR to establish the identities of 162 and 163 are shown in blue. ..................................... 112 Figure 3.25 A) Silica-gel column purification of crude 162, showing the colored oxidation by-products. B) The DMDO oxidation reaction mixture turning red upon warming up to RT. C) TLC of the purified 162 after storage for a few hours in an NMR tube. TLC showed new spots, all turning red upon exposure to air. D) NMR tube containing pure 162 changing color and forming a precipitate in CD2Cl2. ........................................................................................... 113 Figure 3.26 Oxidation trials on various 6-boronyl-L-tryptophans. A) Oxidation of Nα-Boc-Nindole-TIPS-6-BPin-L-Trp-OMe. No reaction was observed. B) Oxidation of Nα-Boc-6-BPin-L-Trp-OMe. Oxindole 166 was formed as the major product. C) Oxidation of Nα-Boc-6-BF3-L-Trp-OMe. Oxindole 168 was formed as the major product in a lower yield. For B and C, several other unidentified oxidation products were obtained. Reaction conditions: Oxone®/NaHCO3/acetone/H2O/0°C or DMDO (0.06 M in acetone)/DCM/-78°C. ............. 114 Figure 3.27 Proposed mechanism for the slow reaction of 6-BPin-L-Trp-OMe (160) with DMDO to produce oxindole 166 as the major product......................................................... 115 xxxi  Figure 3.28 Examples of common electrophilic fluorinating reagents. ................................ 116 Figure 3.29 Proposed mechanism for preparation of 3-fluorooxindoles from indole derivatives using SelectfluorTM. The non-fluorinated oxindole (in the dashed box) may be obtained as a by-product in the presence of acid (Takeuchi et al.). ........................................................... 117 Figure 3.30 Fluorination of Nα-protected tryptophans to Fpi (red arrows) and 3-fluorooxindole (blue arrows) using SelectfluorTM (Fujiwara et al.). ............................................................. 117 Figure 3.31 Synthesis of the Fpi core (in red) of fluorobrevianamide E (170) and fluorogypsetin. The Hpi core of brevianamide E (171) and gypsetin (174) is shown in blue. .............................................................................................................................................. 118 Figure 3.32 Fluorocyclization of a tryptophan-containing dipeptide en route to the synthesis of a 10b-fluorinated analogue of protubonine A. The Fpi core is shown in red, while the Hpi core of the natural protubonine A is shown in blue. ..................................................................... 119 Figure 3.33 Fluorocyclization of 160, yielding the corresponding DHpi (dashed box) and Fpi (2:1 ratio). ............................................................................................................................. 119 Figure 3.34 Proposed mechanism for the formation of DHpi (180) and its XRD structure. 120 Figure 3.35 Synthesis and test Savige-Fontana reaction of Fpi 183 with n-dodecanethiol as the thiol source (6-BPin:5-BPin = 8:1 in all intermediates). ...................................................... 121 Figure 3.36 Converting pinacol-boronate 180 to the corresponding MIDA-ester, followed by fluorocyclization to afford Fpi 188 (MIDA is shown in blue). ............................................ 122 Figure 3.37 Optimizing the conditions for the oxidative deborylation of an aryl-BMIDA (2-naphthyl-BMIDA) in the presence of an aryl thioether (thioanisole) to simulate the oxidation of the 6-BMIDA-tryptathionine system. ............................................................................... 123 xxxii  Figure 3.38 Complete synthetic route to obtain 6-BMIDA-Fpi (188) from Nα-Boc-L-Trp-OMe. .............................................................................................................................................. 124 Figure 4.1 Synthesis of amaninamide diastereomers (191) (Zanotti et al.). Ile3, Trp4 and Ttn-sulfoxide are shown in blue. ................................................................................................. 138 Figure 4.2 Synthetic and derivatized bicyclic amanitin analogues (adapted from Wieland et al., 1981).26 All analogues contain a bicyclic structure; in each analogue, color-coded residues that are different than the natural product are shown. Position-4 contained Trp unless stated otherwise. Inhibitory effects were measured against RNAP II from calf thymus and are relative to the Ki of α-amanitin (I). “SO” represents (R)-sulfoxide, “OS” represents (S)-sulfoxide, and OSO shows the sulfone of tryptathionine. hSer=homoserine, hVal=γ-hydroxyvaline. ........ 139 Figure 4.3 Synthesis of Pro2-Ile3-S-deoxo-amaninamide (194) and Pro2-(D-allo-Ile)3-S-deoxo-amaninamide (195) (May and Perrin). .................................................................................. 141 Figure 4.4 Synthesis of a cytotoxic amanitin analogue for biorthogonal conjugation (Zhao and Perrin). Diastereomers of DHIle are shown in red. .............................................................. 142 Figure 4.5 Synthesis of the linear hexapeptide 201 on solid phase. 6-BMIDA-Fpi (188) is shown in blue. Numbering of the residues is shown in the dashed box. Coupling conditions: AA (5 eq.), HBTU (5 eq.), HOBt.H2O (5 eq.), DIPEA (pH 8), DMF, RT, 2h. .................... 143 Figure 4.6 Synthesis of monocyclic heptapeptides 202 and 203 from the resin-bound linear heptapeptide (201). Treatment with TFA/DCM resulted in the cleavage from the resin, deprotection of Trt, Boc and tBu, and tryptathionylation. .................................................... 144 Figure 4.7 MIDA deprotection of 202 to obtain 203, followed by oxidation of 203 to 204. An unknown by-product (205) was formed as the minor product in this reaction. .................... 145 xxxiii  Figure 4.8 UV absorbance curves for an unsubstituted Ttn (blue), 6-B(OH)2- and 6-BPin-Ttn (maroon), 6-BMIDA-Ttn (gray) and 6-OH-Ttn (orange). .................................................... 146 Figure 4.9 Schematic representation of the extended conjugated π-system of 6-boronate indole (left) and 6-hydroxy indole (right). The p-orbital of the 6-substituent is shown in green. ... 146 Figure 4.10 Coupling of the suitably protected DHIle (139) to the monocyclic heptapeptide (204) followed by Fmoc and TBS deprotection in one pot. Using excess TBAF to deprotect the TBS groups afforded a by-product (209) with the molecular weight 918.3 (desired mass - 2), while adjusting the pH of the TBS deprotection reaction to 5 by adding acetic acid exclusively furnished the desired product (208). .................................................................. 148 Figure 4.11 Revisiting the reaction conditions for conversion of 202 to 204. pH of the reaction mixture was maintained at 8 to avoid strongly basic conditions leading to the formation of a [M-2] by-product (205). ........................................................................................................ 149 Figure 4.12 Proposed mechanism for the formation of the oxidation by-product. Quinone-methide 213 can undergo the elimination of Hα on the adjacent Cys8 residue, leading to the formation of a thioxo-indole core (214). .............................................................................. 150 Figure 4.13 Proposed structure of the base-promoted oxidation products following preliminary characterizations. Thioxo-indole moiety is shown in red, the dehydroalanine resulting from the Hα-elimination of Cys8 is shown in blue. Although the oxidant is currently unknown, it is suspected to be dissolved molecular oxygen, O2. ................................................................. 151 Figure 4.14 MTT cell viability assay results for commercially available α-amanitin and the oxidation by-product against CHO cells. .............................................................................. 152 xxxiv  Figure 4.15 Macrolactamization of the monocyclic octapeptide (208) to S-deoxy-α-amanitin (217) using HATU as the coupling reagent. ......................................................................... 154 Figure 4.16 Structures of (S)-omeprazole (esomeprazole) and (R)-rabeprazole (dexrabeprazole). The (S)-isomer of omeprazole is metabolized more slowly and reproducibly than the (R)-enantiomer.134 The (R)-enantiomer of rabeprazole has been suggested to show higher therapeutic effects.135 ................................................................................................. 155 Figure 4.17 Examples of asymmetric sulfoxidation reactions. A) Titanium-based oxidation with (R,R)-DET as the chiral ligand and tBuOOH as the main oxidant.140 B) Vanadium-catalyzed oxidation with camphor-based Schiff bases (221) as chiral ligands and H2O2 as the oxidant.141 C) Iron-catalyzed oxidation with a Schiff base chiral ligand (224) and H2O2.142 .............................................................................................................................................. 156 Figure 4.18 Examples of metal-free asymmetric sulfoxidations. A) Oxidation with H2O2 using a BINOL-derived chiral phosphoric acid (226) as the ligand.144 B) Oxidation with a chiral oxaziridine (227).146 .............................................................................................................. 157 Figure 4.19 Oxidation of O-Me-α-amanitin (232) to (R)-sulfoxide, (S)-sulfoxide and sulfone by Buku et al.149 Using one equivalent of H2O2 afforded 232 and 234 in a 1:2 ratio, while using large excess of H2O2 yielded the sulfone (235). ................................................................... 158 Figure 4.20 A) Superimposed XRD structures of O-Me-α-amanitin-(S)-sulfoxide (234) (thin line) and sulfone (235) (thick line). B) Superimposed X-ray structures of β-amanitin (thin line) and O-Me-α-amanitin-sulfone (235) (thick line). C) Hydrogen bond between Asn1(NH) and (S)-oxygen in O-Me-α-amanitin-sulfone (right), resulting in the 90°-rotation of the plane containing the Asn1 amide bond (reproduced without permission from Shoham et al.).46 .. 161 xxxv  Figure 4.21 A) Chemical structure of β-amanitin. B) XRD structure of β-amanitin (adapted from Kostansek et al.), representing possible facial selectivity for the sulfoxidation. R = C(O)NH2 in α-amanitin. ........................................................................................................ 162 Figure 4.22 An example of the HPLC chromatogram of the crude product of a sulfoxidation reaction containing a mixture of (R)-sulfoxide, (S)-sulfoxide and sulfone of α-amanitin. Reaction conditions: S-deoxy-amanitin (20 nmol), mCPBA (3.5 eq), iPrOH/EtOH 3:1, RT, 20 min. HPLC gradient: 0-30 min 6%-18% A, 30-34 min 18%-100% A; 34-37 min 100% A, 37-39 min 100%-6% A, 39-44 min 6% A (solvent A: 0.1% FA in H2O, solvent B: 0.1% FA in MeCN). ................................................................................................................................. 165 Figure 4.23 Apparent R/S selectivity for the sulfoxidation of 217 in methanol. Reaction conditions: S-deoxy-α-amanitin (3.3 nmol), mCPBA (1.3 eq), 2 hours. Ratios were calculated based on the area under the peaks following the HPLC injection of the crude reaction mixture. .............................................................................................................................................. 166 Figure 4.24 Apparent R/S selectivity for sulfoxidation of 217 in various solvents. Reaction conditions: S-deoxy-α-amanitin (3.3 nmol), mCPBA (1.3 eq), RT, 2 hours. Ratios were calculated based on the area under the peaks following the HPLC injection of the crude reaction mixture. ................................................................................................................................. 167 Figure 4.25 Proposed kinetic resolution of (R) and (S)-sulfoxides of α-amanitin via selective oxidation to sulfone (237) using excess mCPBA. Consumption of the (S)-sulfoxide (236) could lead to an apparent higher diastereoselectivity in favor of the (R) isomer. .......................... 168 Figure 4.26 Sulfoxidation of S-deoxy-α-amanitin using excess mCPBA (3.5 eq.) in iPrOH/EtOH 3:1. Composition of the reaction mixture was determined at different time points (20 min, 45 min, 80 min, 120 min) using HPLC. ................................................................. 169 xxxvi  Figure 4.27 Proposed energy levels of the thioether and sulfoxides of α-amanitin. (R)-sulfoxide is likely the thermodynamically favored product, while (S)-sulfoxide is the kinetically favored isomer. Note that (R)- and (S)-sulfoxides of amanitin cannot interconvert. ......................... 171 Figure 4.28 Proposed explanation for the selective oxidation of (S)-sulfoxide to sulfone in the presence of (R)-sulfoxide. ..................................................................................................... 173 Figure 4.29 HPLC chromatograms at 305 nm for the synthetic α-amanitin (blue), authentic α-amanitin (red) and their co-injection (green). HPLC gradient: D (refer to section 4.4.1 for HPLC gradients). .................................................................................................................. 174 Figure 4.30 HPLC chromatograms at 305 nm for the synthetic α-amanitin-(S)-sulfoxide (purple), authentic α-amanitin (red), and their co-injection. HPLC gradient (green): D (refer to section 4.4.1 for HPLC gradients). ....................................................................................... 175 Figure 4.31 UV absorbance curves for the synthetic α-amanitin (blue), (S)-sulfoxide (purple), and the authentic α-amanitin demonstrating their λmax (red). ................................................ 176 Figure 4.32 CD spectra for the synthetic and authentic α-amanitins. Sample preparation for CD: Authentic (purchased from Sigma-Aldrich, 53 µg) and synthetic amanitin (48 μg) were each dissolved in 500 μL of MeOH. A quartz cuvette (1 mm path-length) was used. CD spectra were acquired as noted in the Materials and Methods section (vide infra). ......................... 177 Figure 4.33 Notable distinctive proton chemical shifts in the 1H-NMR spectra of the (R) and (S)-sulfoxides and thioether of α-amanitin. Values shown in boxes are in ppm (99.96% DMSO-d6, 600MHz). ......................................................................................................................... 181 Figure 4.34 An overlay of the 1H-NMR spectra for the synthetic (in blue) and authentic (in red) α-amanitins. Solvent peaks have been labelled with the name of the corresponding solvents. xxxvii  Unknown peaks, resulting from the sample preparation or solvent stabilizers, are marked with an asterisk (*). To avoid overcrowding, the exact values for chemical shifts are not shown. .............................................................................................................................................. 182 Figure 4.35 MTT cell viability assay against CHO cells for the synthetic and authentic α-amanitins, thioether 217 and (S)-sulfoxide 236. Top: bar graph representing % of viable cells at a given concentration. Bottom: IC50 curves (IC50 values are reported in µM). ................ 183 Figure 5.1 Interactions between the different residues of α-amanitin and the Rpb1 subunit of RNAP II (reproduced with permission). A) Cryo-EM structure by Cramer et al.24 B) Co-crystal structure by Kornberg et al.19 Note the hydrogen bonding interactions of Hyp2. ................ 198 Figure 5.2 Proline conformational equilibria. A) cis-trans amide bonds (in red) in slow equilibrium (trans isomer is greatly favored in proteins and peptides). B) Fast equilibrium between exo-endo ring puckers. ............................................................................................ 200 Figure 5.3 A) Hyperconjugation of the orbitals in the gauche conformation of 4-Hyp. B) Preferred exo conformer in (4R)-Hyp (left); preferred endo conformer in (4S)-Hyp (right) (from Zondlo et al.). ............................................................................................................. 201 Figure 5.4 Selected Hyp analogues for synthesis and incorporation into amanitin derivatives. trans-Azido proline may be used as a novel site for bioconjugation via CuAAC. 4-Oxo-proline (keto-proline) may be reduced to cis- or trans-Hyp on the fully elaborated toxin or used as a handle for bioconjugation. .................................................................................................... 205 Figure 5.5 Solid-phase synthesis of 4-substituted prolines on the rink-amide resin (from Zondlo et al.). R1 = Ac-Thr(OtBu)-Tyr(OtBu)-, R2 = -Asn(Trt)-NHRink-Resin. ............................ 206 xxxviii  Figure 5.6 Retrosynthetic scheme for the synthesis of Hyp analogues and keto-proline starting from trans-Hyp. .................................................................................................................... 207 Figure 5.7 Synthesis of the mesylated cis-Boc-Hyp-OMe (241) from trans-Boc-Hyp-OH (238). Mechanism of the azide-mediated methonolysis is shown. ....................................... 208 Figure 5.8 SN2 displacement reactions on 241 to afford trans-cyano (242), azido (243) and thio (244) proline intermediates. .................................................................................................. 208 Figure 5.9 Synthesis of trans-Fmoc-CN-Pro-OH (249) and trans-Fmoc-N3-Pro-OH (250). .............................................................................................................................................. 209 Figure 5.10 Synthesis of trans-Nα-Fmoc-(4-SAcm)-Pro-OH (254). The Acm group is shown in blue. .................................................................................................................................. 210 Figure 5.11 Reduction of trans-Nα-Fmoc-4-N3-Pro-OH (250) to the corresponding amino-proline (256). Staudinger conditions resulted in the formation of the iminiphosphorane intermediates (255a, 255b) that failed to afford the product upon hydrolysis. Hydrogenolysis of 250 produced the desired product in high yields. ............................................................. 211 Figure 5.12 Testing the stability of a model tryptathionine crosslink under catalytic hydrogenolysis conditions. ................................................................................................... 211 Figure 5.13 Synthesis of Nα-Fmoc-4-oxo-Pro-OH (keto-proline) (262). ............................. 212 Figure 5.14 Synthesis of a photolabile caged (photolabile) ATP. Upon exposure to λ=342 nm, an uncaged (deprotected) ATP is released, resulting in the inhibition of purified renal Na,K-ATPase (from Kaplan et al.). The photolabile protecting group is shown in blue. .............. 214 xxxix  Figure 5.15 Disulfide bond formation on a peptide by photocleavage of the 6-nitroveratryl protecting group (shown in blue) and thiolysis by S-pyridinesulfenyl activation (from Karas et al.). ........................................................................................................................................ 214 Figure 5.16 Examples of A) ortho-nitrobenzyl-based and B) coumarin-based PPGs and their relative absorption maxima (adapted from Feringa et al.).172 The leaving group released upon exposure to light is shown in blue. ....................................................................................... 215 Figure 5.17 Mechanism of phototautomerization of oNT (Gilch et al.). The lifetime of each excited state is shown under its structure (ISC = intersystem crossing). .............................. 217 Figure 5.18 Mechanism of photoreaction of 1-(methoxymethyl)-2-nitrobenzene, releasing MeOH (shown in blue) (Wirz et al.). .................................................................................... 218 Figure 5.19 Preparation of 6-nitroveratryl bromide (271). ................................................... 219 Figure 5.20 Protection of trans-Nα-Boc-Hyp-OMe (272) with 6-nitroveratryl group to afford Nv-Hyp-OMe 273. ................................................................................................................ 220 Figure 5.21 Protecting group manipulation of photocleavable hydroxyproline to generate SPPS compatible trans-N-Fmoc-(ONv)-Hyp-OH (276). ............................................................... 220 Figure 5.22 Solid phase synthesis of α-amanitin began with loading Hyp (shown in blue) on CTC resin. ............................................................................................................................. 222 Figure 5.23 Retrosynthetic scheme for strategy A: loading the CTC resin with Fmoc-Asp(OH)-OAllyl (in red) and macrolactamization on solid phase. Proline derivative (in blue) is introduced as the last residue. ............................................................................................... 223 Figure 5.24 Synthesis of linear pentapeptide 278 (strategy A). ............................................ 224 Figure 5.25 Synthesis of the test dipeptide Ile-Fpi-OH (281). ............................................. 224 xl  Figure 5.26 Attempted synthesis of monocyclic octapeptide 284 on solid phase. Upon exposure to TFA, linear octapeptide 283 afforded the tryptathio-monocyclic octapeptide 285. Hyp is shown in blue and Asp is shown in red. ............................................................................... 226 Figure 5.27 Retrosynthetic scheme for strategy B: loading the CTC resin with Fmoc-Cys(SH)-OAllyl (in red) and macrolactamization on solid phase. Proline derivative (in blue) is introduced as the second last residue. Note the incorporation of Ile at position-3 in place of DHIle. ................................................................................................................................... 227 Figure 5.28 Synthesis of Fmoc-Cys(SH)-OAllyl (288) from Fmoc-Cys(STrt)-OH (286). .. 228 Figure 5.29 Attempted synthesis of monocyclic octapeptide 291 on solid phase. Upon exposure to TFA, linear octapeptide 290 afforded the tryptathio-monocyclic octapeptide 292 (Hyp is shown in blue, Cys is shown in red). 292 can presumably undergo a macrolactamization reaction to produce the bicyclic β-amanitin analogue. ......................................................... 229 Figure 5.30 Synthesis of an amanitin analogue using solid phase strategy C. Upon exposure of the linear octapeptide 293 to TFA/DCM, global deprotection of acid labile protecting groups was achieved and the tryptathionine crosslink was formed (294). A solution-phase macrolactamization yielded the final bicyclic octapeptide 295. ........................................... 231 Figure 5.31 Formation of an oxindole by-product (296) upon treatment of linear octapeptide 293 with TFA. The oxindole is shown in the dashed box. ................................................... 232 Figure 5.32 Mechanism of formation of the oxindole by-product in the presence of TFA and H2O. ...................................................................................................................................... 233 Figure 5.33 General synthetic scheme for the production of amanitin analogues containing Hyp analogues and derivatives (shown in blue). DHIle residue is shown in red. ........................ 234 xli  Figure 6.1 Synthetic scheme for the enantioselective synthesis of the fully protected (2S,3R,4R)-DHIle (139). ....................................................................................................... 266 Figure 6.2 General scheme for the synthesis of a tryptathionine crosslink from tryptophan. .............................................................................................................................................. 267 Figure 6.3 Attempted synthesis of 6-OH-Hpi (162, shown in blue) from 6-BPin-L-Trp. .... 268 Figure 6.4 Synthesis of 6-BMIDA-Fpi for incorporation into α-amanitin. The BMIDA group would be oxidized to hydroxyl on the fully elaborated toxin. .............................................. 269 Figure 6.5 Synthesis of monocyclic heptapeptide 204 following the oxidative deborylation of 202. ....................................................................................................................................... 270 Figure 6.6 Synthesis of α-amanitin-thioether from monocyclic heptapeptide 204. .............. 271 Figure 6.7 A) Sulfoxidation of α-amanitin-thioether using mCPBA (1.1 eq) in iPrOH/EtOH 2:1. B) Proposed kinetic resolution of (R) and (S)-sulfoxides via oxidation to sulfone (237). .............................................................................................................................................. 272 Figure 6.8 Derivatives of Hyp proposed for synthesis and incorporation into amanitin analogues. ............................................................................................................................. 274 Figure 6.9 Synthesis of A) trans-cyano, azido and thio proline analogues and B) 4-oxoproline as SPPS compatible amino acids. ......................................................................................... 274 Figure 6.10 Synthesis of a photocleavable (6-nitroveratryl) Hyp (276) and its proposed incorporation into the structure of an inactive amanitin analogue that could release the active toxin following exposure to λ=336 nm. ................................................................................ 275   xlii  List of Symbols and Abbreviations  3D Three-dimensional Å Ångstrom AA Amino acid Abu Aminobutyric acid Ac2O Acetic anhydride Acm Acetamidomethyl AD Asymmetric dihydroxylation ADC Antibody-drug conjugate ADP Adenosine 5’-diphosphate ATP Adenosine 5'-triphosphate BINOL 1,1'-Bi-2-naphthol Bn Benzyl Boc tert-Butyloxycarbonyl Cbz Carboxybenzyl CD Circular dichroism CHO Chinese hamster ovary COSY Homonuclear correlation spectroscopy Cryo-EM Cryo-electron microscopy CSA Camphorsulfonic acid CTC 2-Chlorotrityl chloride CuAAC Copper catalyzed azide-alkyne cycloaddition CYP450 Cytochrome P450 monooxygenase DAR Drug-to-antibody ratio DCM Methylene chloride, dichloromethane dd Doublet of doublets ddd Doublet of doublet of doublets DHIle (2S,3R,4R)-4,5-Dihydroxyisoleucine DHP Dihydropyran xliii  DHpi Dihydropyrrolo indoline DHQ Dihydroquinine DHQD Dihydroquinidine DIAD Diisopropyl azodicarboxylate DIBAL Diisobutyl aluminum hydride DIPEA Diisopropylethyl amine DMA N,N-Dimethylacetamide DMAP 4-Dimethylamino pyridine DMDO Dimethyldioxirane DMF N,N-Dimethylformamide DMSO Dimethylsulfoxide DSC Disuccinimidyl carbonate DTNP 2,2’-Dithiobis(5-nitropyridine) DTT Dithiothreitol EDC 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide ESI Electrospray ionization Et Ethyl Et2O Diethyl ether EtOH Ethanol FA Formic acid FMO Flavin-dependent monooxygenase Fmoc 9-Fluorenylmethyloxycarbonyl Fpi 3a-Fluoropyrrolo[2,3-b]indoline, 3a-fluoropyrrolo[2,3-b]indoline-2-carboxyl FP-T300 N-Fluoro-2,4,6-trimethylpyridinium triflate, N-fluoro-2,4,6-collidinium triflate GmPOPB Proline oligopeptidase B from Galerina marginata h Hour(s) HATU 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate HBTU 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate HFIP Hexafluoroisopropyl alcohol xliv  HOAc Acetic acid HOAt 1-Hydroxy-7-azabenzotriazole HOBt 1-Hydroxybenzotriazole Hpi 3a-Hydroxypyrrolo[2,3-b]indoline, 3a-hydroxypyrrolo[2,3-b]indoline-2-carboxyl HPLC High-performance liquid chromatography HRMS High-resolution mass spectrometry hSer Homoserine HSQC Heteronuclear single quantum coherence spectroscopy hVal γ-Hydroxyvaline Hyp trans-Hydroxyproline IC50 Half maximal inhibitory concentration Ipc Isopinocampheyl iPrOH Isopropyl alcohol, isopropanol ISC Intersystem crossing J Coupling constant Ki Inhibition constant LAH Lithium aluminium hydride LCP Left circularly polarized LD50 Median lethal dose LDA Lithium diisopropylamide LRMS Low-resolution mass spectrometry M Molar, mass m Multiplet m/z Mass-to-charge ratio mAb Monoclonal antibody mCPBA meta-Chloroperoxybenzoic acid Me Methyl MeCN Acetonitrile MeOH Methanol MHIle Monohydroxyisoleucine xlv  MIDA N-methyliminodiacetic acid min Minute(s) MMFF Merck Molecular Force Field MS Mass spectrometry MsCl Methanesulfonyl chloride, mesyl chloride MTT 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide MW Molecular weight N Normal NHS N-Hydroxysuccinimide Nle Norleucine NMR Nuclear Magnetic Resonance NOE Nuclear Overhauser effect NOESY Nuclear Overhauser effect spectroscopy Nv 4,5-Dimethoxy-2-nitrobenzyl, 6-nitroveratryl Nva Norvaline OAc Acetate oNT ortho-Nitrotoluene Osu N-Hydroxysuccinimide OTf Triflate PCC Pyridinium chlorochromate PEA 1-Phehnylethylamine Ph Phenyl Phal Phthalazine PIC Pre-initiation complex Pin Pinacol, pinacolate PMP para-Methoxyphenyl POP Proline oligopeptidase PPG Photocleavable protecting group ppm Parts per million PyBOP (Benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate Pyr Pyrimidine xlvi  q Quartet RBF Round-bottom flask RCP Right circularly polarized Rf Retention factor RNAP II RNA Polymerase II s Singlet SAR Structure-activity relationship SN2 Second order nucleophilic substitution SPPS Solid-phase peptide synthesis t Triplet t Tertiary TATU O-(7-Azabenzotriazole-1-yl)-N,N,N',N'-tetramethyluronium tetrafluoroborate TBAF Tetrabutylammonium fluoride TBP Tata box-binding protein TBS, TBDMS tert-Butyldimethylsilyl tBuOH tert-Butanol TEMPO (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl TFA Trifluoroacetic acid, N-trifluoroacetyl TFE Trifluoroethanol THF Tetrahydrofuran THP Tetrahydropyranyl TIPS Triisopropylsilyl TIS Triisopropylsilane TLC Thin-layer chromatography TMS Trimethylsilyl tR Retention time Trt Trityl, triphenylmethyl TsOH para-Toluenesulfonic acid Ttn Tryptathionine UV Ultraviolet v/v Volume-to-volume ratio xlvii  Xaa Unspecified amino acid residue XRD X-ray diffraction δ Chemical shift λ Wavelength                     xlviii  Acknowledgements   There are many people without whom this dissertation could not have been accomplished. Their assistance and support were key to developing and finishing this project, and I will attempt to acknowledge them here. First and foremost, I must acknowledge my supervisor Prof. David Perrin. During my time as a PhD student in the Perrin lab, he constantly and patiently provided me with priceless insight, valuable ideas and scientific inspirations. His passion for science was key to my dedication to chemistry and growth as a scientist. Furthermore, preparation of this thesis would not have been possible without his insightful comments and guidance.  The various lab members alongside whom I had the pleasure of working not only provided a scientific atmosphere in the Perrin lab, but were the key components of an enjoyable time as I completed my PhD. To all of you I say: Thank you! More specifically and in no particular order, thanks to: (soon-to-be) Dr. Omar Sadek, Abid Hasan, Dr. Antoine Blanc, Mihajlo Todorovic, Alla Pryyma, Samson Lai, Jerome Lozada, Dr. Yajun Wang, and my wonderful undergraduate students Brandon Kato and Charlie Wei.  This work was built on the work of previous members of the Perrin lab, who greatly contributed to my ways of thinking and designing experiments regarding this project. Dr. Liang Zhao, Dr. David Dietrich and Dr. Jonathan May laid an exceptional foundation upon which this project was constructed.  There were a number of people outside of the Perrin lab who contributed to the work presented in the current thesis. Thank-you to Dr. Brian Patrick for his expertise in crystallography. Thank-you to Dr. Elena Polishchuk and Ms. Jessie Chen for their help in xlix  obtaining the biological assays. Thank-you to Dr. Maria Ezhova for her great help in obtaining the NMR data for the synthetic amanitins. Thank-you to John, Pat, George and Elan in chem stores. Special thank-you to Sheri for her non-stop assistance with every question that I had regarding my graduate studies during these years, especially in the process of my thesis preparation.  I would like to acknowledge Dr. Raymond Andersen and Dr. John Sherman for reviewing this thesis and providing me with valuable comments as members of my supervisory committee. Thank-you to Dr. Andrei Yudin for accepting the role of the external examiner and offering his priceless insight into this project and thesis. Thank-you to Dr. Ed Grant and Dr. Xiaonan Lu for taking the time to review my thesis as university examiners.  As an international student at UBC, a number of friends helped me integrate into a new country and university, and my time at UBC could not have come to pass without them. Bast. Gabriele, Omar, Abid, Reza and Mohammad, to you I say: Thank you! A special thank-you to Leana for being there for me at the most critical time, when I was preparing my first-ever publication and this thesis.  Last but not least, I would like to say thank-you to my family, for providing me the means to make this dream come true, for standing by me and supporting me since the day I was born to this day. A big thank-you to Mom, Dad and Sadaf.      1  Chapter 1: Introduction  1.1 What is α-Amanitin? Isolated over 60 years ago,1 α-Amanitin (Figure 1.1, 1) is a naturally occurring bicyclic octapeptide and the primary toxin responsible for the high toxicity of the infamous species of mushrooms Amanita Phalloides, more commonly known as the “death cap” mushroom. Ingestion of these mushrooms severely damages the liver and, in many cases, leads to death. Exhibiting an exceptionally low LD50 (50-100 µg/kg),2 α-Amanitin, which is one of deadliest toxins known to humankind, is featured in most modern biochemistry textbooks. Benefiting from a rich scientific history,3,2 the use of death cap mushrooms dates back to Roman times where they were used for murder and suicide,4 yet today the mushroom serves as a source of the toxin that is being investigated for use in targeted therapies for cancer.5  Figure 1.1 A) The structure and amino acid numbering of the bicyclic octapeptide α-amanitin. B) The simplified scaffold showing the abbreviations used throughout this thesis.  α-Amanitin is the principal toxin of several Amatoxins, a toxin family extracted from Amanita mushrooms that possess similar structures yet different toxicities (Table 1.1).6 Amatoxins are inhibitors of RNA polymerase II (RNAP II); once they enter the cell’s nucleus 2  they disrupt the transcription process.7 Based on inhibition constants (Ki values) obtained from different amatoxins, several key motifs give clear indications as to structural and chemical components that are responsible for toxicity; these include a unique bicyclic structure as well as the presence of two oxidized amino acids, Hyp2 and DHIle3, that are crucial for the toxicity of the peptide. These structure-activity relationships will be further discussed in subsequent chapters.  # Toxin R1 R2 R3 R4 R5 Ki (nM) LD50 (mg/kg) 1 α-amanitin CH2OH OH OH NH2 OH 2.3 0.3 2 β-amanitin CH2OH OH OH OH OH 2.5 0.5 3 γ-amanitin CH3 OH OH NH2 OH 5 0.2 4 ε-amanitin CH2OH OH OH OH H - 0.3-0.6 5 Amanin CH2OH OH OH OH H 5 0.3 6 amaninamide CH2OH OH OH NH2 H 5 20 7 amanullin CH3 H OH NH2 OH 10 >20 8 amanullinic acid CH3 H OH OH OH - >20 9 proamanullin CH3 H H OH OH - >20 Table 1.1 Different amatoxins isolated from Amanita mushrooms. The Ki values were measured against RNAP II of calf thymus, and the LD50 values were obtained by testing on white mice.6  In spite of being the main source of toxicity of Amanita mushrooms, amatoxins are not the only family of toxins found in these species: Phallotoxins8 and Virotoxins9 represent two other classes of toxins that could be extracted from Amanita Phalloides. Like amatoxins, phallotoxins are bicyclic peptides (Figure 1.2) of which the most studied member is Phalloidin. 3  However, in contrast to the octabicyclic amatoxins, phallotoxins contain seven amino acids, lack the hydroxyl group on the indole ring of tryptophan, and have a thioether bridge instead of a thioether-sulfoxide. Virotoxins, on the other hand, are monocyclic heptapeptides, also containing oxidized side chains. In contrast, while the tryptophan residue contains a sulfonate ester, these toxins lack the bicyclic structure characteristic of phallotoxin and amatoxins (Table 1.2).  Name R1 R2 R3 R4 R5 Phallotoxins (A) Phalloidin OH CH3 C2H5OH CH3 CH3 Prophalloin H CH3 CH3 CH3 CH3 Phalloin OH CH3 CH3 CH3 CH3 Phallisin OH C2H5OH C2H5OH CH3 CH3 Phallacin OH CH3 CH3 CH(CH3)2 COOH Phallacidin OH CH3 C2H5OH CH(CH3)2 COOH Phallisacin OH C2H5OH C2H5OH CH(CH3)2 COOH Virotoxins (B)  Viroidin CH(CH3)2 CH3 SO2 Desoxo-viroidin CH(CH3)2 CH3 SO Ala1-viroidin CH3 CH3 SO2 Ala1-desoxo-viroidin CH3 CH3 SO Viroisin CH(CH3)2 CH2OH SO2 Desoxo-viroisin CH(CH3)2 CH2OH SO Table 1.2 Structure of heptapeptides extracted from A. Phalloides: A) Phallotoxins and B) Virotoxins.  4  Although similar in structure, these three classes of toxins show completely different activities against different targets. As mentioned, amatoxins target RNAP II and inhibit transcription. However, there is no evidence suggesting any interaction between RNAP II and either phallotoxins or virotoxins. Instead, these latter two are specific binders of filamentous actin (F-actin).3,8,9 Given that phallotoxins and virotoxins are not orally available due to a lack of gut uptake, the toxicity of Amanita mushrooms via ingestion in humans is mainly due to the action of amatoxins, most notably alpha-amanitin, and not these other classes of toxins. Considering the fact that all these three classes of toxins are presumably derived from a shared biological pathway that involves ribosomal synthesis of a common 35-amino acid precursor in their biosynthesis,10 vide infra, these differences in activity and overall structure are fascinating and have, for years, stimulated studies at the interface of chemistry and biology.3,8,10,11  1.2 Overview of the Pharmacology of α-Amanitin Early studies have revealed that the mechanism by which α-amanitin acts in cells is binding to RNAP II.12 This enzyme is one of the three RNA polymerases found in the nucleus of eukaryotic cells, the other two being RNAP I and III; RNAP II is the key component of the transcription process. In order to further comprehend how α-amanitin interacts with RNAP II and why its mechanism of action is important, an overview of the transcription process is necessary. One of the foundations of molecular biology lies in the concept of the central dogma. In 1958, Francis Crick introduced this concept for the first time and finalized the details of it in a Nature article published in 1970, titled “The Central Dogma of Biology”.13 It is often simplified as “RNA is made from DNA, and proteins are made from RNA”. Although largely true, this simplification might not reflect the importance of this process. A more accurate 5  phrasing of this process is as follows: “The central dogma of biology deals with the detailed residue-by-residue transfer of sequential information. It states that such information cannot be transferred from protein to either protein or nucleic acid”.13 The focal point of this concept is that DNA is the main carrier of genetic information, which is transferred to RNA through the transcription process. The information that DNA carries cannot be comprehended inside the cell; RNA however, provides readable messages for the synthesis of proteins. The process by which proteins are made from RNA, specifically mRNA, is known as translation (Figure 1.2). It is evident that these processes, namely transcription, are essential for the growth and survival of cells. RNA polymerases, the enzymes responsible for the transcription of DNA to RNA, play a vital role in maintaining the life cycle of a cell.  Figure 1.2 A schematic of the central dogma of molecular biology. The blue arrows represent the processes that are common in all organisms, while the red arrows show the pathways that only occur in certain organisms.  1.2.1 RNA Polymerases and the Biochemistry of Transcription In eukaryotic cells, due to increased complexity arising from evolutionary development, three main types of RNA polymerases are observed: RNAP I, II and III. The product of these three enzymes are different: RNAP I produces ribosomal RNA (rRNA), RNAP II catalyzes the synthesis of messenger RNA (mRNA), and transfer RNA (tRNA) is the product of RNAP III. In 1970, Lindell et al. showed that α-amanitin only affects RNAP II, and no significant interactions between this toxin and other polymerases are observed (Table 1.3). Taking 6  advantage of this observation, α-amanitin may be used as a means to probe the degree of transcription catalyzed by various RNA polymerases, especially RNAP II.12 Cell Polymerase Product Ki (α-Amanitin) Prokaryotic RNA polymerase mRNA, tRNA, rRNA no binding Eukaryotic RNAP I rRNA no binding Eukaryotic RNAP II mRNA 1 nM Eukaryotic RNAP III tRNA 1 µM Table 1.3 Different RNA polymerases produced in different types of cells, listing their products and binding affinity to α-amanitin.  The product of the RNAP II catalyzed transcription is messenger RNA (mRNA), which is responsible for conveying genetic information from DNA to the ribosome, where it can be decoded to determine the amino acid sequence of ribosomally synthesized proteins, including but not limited to other enzymes that are vital to the survival of the cell. RNAP II consists of twelve (in humans and yeast) peptide subunits (Rpb1 to Rpb12) that are necessary for the activity of the enzyme. The two biggest subunits, Rpb1 and Rpb2, form the majority of the active core of the enzyme.14 A helical part of the Rpb1 subunit is called the “bridge helix”, which is a common feature between all types of RNA polymerases. The bridge helix lies in the vicinity of the enzyme’s active site and is a key component in the translocation of RNAP II on DNA. Hence, any decrease in its flexibility will result in lower mobility of the enzyme and eventually halt the transcription process. It has been shown that α-amanitin binds to this bridge helix and lowers the rate of transcription drastically.15 In 2001, Kornberg et al. obtained the crystal structure of RNAP II in complex with DNA (Figure 1.3).16 7   Figure 1.3 Crystal structure of RNAP II in complex with DNA and RNA at 3.3Å resolution (reproduced with permission from Kornberg et al.).16 A) Comparison of structures of free RNAP II (top) and the RNAP II transcribing complex (bottom). The clamp (yellow) closes on DNA and RNA, which are bound in the cleft above the active site. The remainder of the enzyme is in gray. B) Structure of the RNAP II transcribing complex. Portions of Rpb2 that form one side of the cleft are omitted to reveal the nucleic acids on the DNA (green and blue) and RNA (red) ribbons. The Rpb1 bridge helix traversing the cleft is highlighted in green. The active site metal “A” is shown as a pink sphere.  There are generally four phases of transcription in a cell: 1) initiation, 2) promoter escape, 3) elongation and 4) termination. For the initiation step, general transcription factors (TFIIB, TFIID, TFIIE, TFIIF and TFIIH) along with RNAP II assemble at the DNA promoter to generate the pre-initiation complex (PIC). General transcription factors are also responsible for initiation of the RNA synthesis and stimulate the promoter escape of RNAP II. In the structure of DNA, a TATA box is a specific sequence of DNA that specifies where the transcription process should begin. TFIID, a part of PIC, contains TATA box-binding protein (TBP) and multiple TBP-associated factors (TAFs) that can signal where the genetic sequence should be read and decoded. 8  Once the initiation complex is formed on the double stranded DNA (called a “closed complex”), RNAP II, assisted by transcription factors along with ATP that is hydrolyzed for added energy, unwinds approximately 15 base pairs to form an “open complex” with the promoter (part of the DNA signaling the binding site of the enzyme). It is noteworthy that the initiation of transcription is accompanied by phosphorylation of the C-terminal domain of RNAP II by the Cyclin-dependent kinase 7 (CDK7), a subunit of TFIIH. The unwound, single stranded DNA in complex with RNAP II is called the “transcription bubble” and one of the two strands serves as the template for RNA synthesis. Following the formation of the first phosphodiester linkage in RNA, RNAP II must escape the promoter and proceed with formation of the next phosphodiester linkage, assisted by general transcription factors. The coding or template strand acts as a template for RNA synthesis. The enzyme finds complementary bases to pair with the template strand and makes the exact copy of the coding strand (except for thymines that are replaced with uracils), hence “elongating” RNA from the 3’ to the 5’ end. Finally, the initiation factors are dissociated, and the elongation complex is formed (Figure 1.4).17 Once elongation is complete termination has to take place. The termination process is less understood in the literature, however it is believed to proceed through a “polyadenylation” step which adds multiple adenines to the new 3’ end of RNA.18 9   Figure 1.4 A) Schematic of RNAP II transcription initiation. TFIID or its TATA box-binding protein (TBP) subunit binds to promoter DNA. The TBP-DNA complex is stabilized by TFIIA and TFIIB. The resulting upstream promoter complex is joined by RNAP II-TFIIF complex, leading to the formation of the core PIC. Subsequent binding of TFIIE and TFIIH complete the PIC (closed PIC). In the presence of ATP, the DNA is opened (forming the transcription bubble) and RNA synthesis begins. Finally, dissociation of initiation factors enables the formation of the elongation complex accompanied by the elongation factors (reproduced with permission from Sainsbury et al.).17  1.2.2 Mechanism of Action of α-Amanitin It has been shown that α-amanitin is a strong, specific inhibitor of RNAP II.12 Kornberg et al. obtained a co-crystal structure (2.8Å resolution) of α-amanitin bound to RNAP II and showed that α-amanitin binds to the bridge helix of the enzyme and not the active site, hence making this toxin an allosteric inhibitor of RNAP II (Figure 1.5).19 10   Figure 1.5 Location of α-amanitin bound to RNAP II (from Kornberg et al.). A) Cutaway view of α-amanitin (red dot) bound to RNAP II and its location relative to the nucleic acids of the DNA and RNA and the bridge helix. B) representation of α-amanitin and RNAP II co-crystal using ribbons. Eight zinc atoms are in light blue; the active site magnesium ion is in magenta; the part of Rpb1 near amanitin (funnel) is in light green; the region of Rpb2 near amanitin is in dark blue; the bridge helix is in dark green; α-amanitin is in red.19  It was previously observed that mutations of RNAP II proximal to the bridge helix significantly affected amanitin binding.20,21 Hence, it can be inferred that the binding location of α-amanitin to RNAP II is not a crystallization artifact but is indeed the actual binding site of the toxin. At this point I leave the discussion of individual interactions and hydrogen bonds between the residues of α-amanitin and RNAP II for later sections. However, as expected from the co-crystal structure, binding of α-amanitin to RNAP II has no effect on the binding affinity of the enzyme toward nucleoside triphosphates (NTPs) which are necessary for transcription.22 Instead, toxin-bound polymerase exhibits a significantly reduced rate of translocation such that elongation of mRNA is reduced from thousands of nucleotides per minute to only a few.23 All these observations are consistent with the co-crystal structure obtained by Kornberg and co-11  workers that definitively showed that α-amanitin is an allosteric inhibitor of RNAP II and does not bind to the active site of the enzyme. More recently, Cramer et al. mapped a cryo-EM structure of a mammalian RNAP II bound to α-amanitin during the elongation step (Figure 1.6).24 This structure further supports the position of the toxin within the enzyme and sheds more light upon important interactions between residues on amanitin and amino acids on the backbone of the enzyme (vide infra).  Figure 1.6 The cryo-EM structure of mammalian RNAP II bound to α-amanitin (3.4Å resolution). RNAP II is shown in a silver ribbon model (reproduced with permission from Cramer et al.).24  1.2.3 Key Interactions Between α-Amanitin and RNAP II Since its discovery and isolation from Amanita mushrooms, structure activity relationships of α-amanitin have been intensely studied via various modifications to the natural product as well as derivatives of entirely synthetic origin.7,25–30 The vast range of synthesized or naturally occurring derivatives help us understand the structure-activity relationships (SAR) of amanitin with RNAP II, and possibly determine what residues on the toxin have the strongest 12  interactions with the backbone of the enzyme. While the importance of these seminal works cannot be understated, definitive evidence of key toxin-enzyme interaction sites was lacking until the resolution of the co-crystal structure of amanitin bound to RNAP II by Kornberg and co-workers (Figure 1.7).19 More recently, the cryo-EM structure of the elongation complex of amanitin-RNAP II by Cramer et al. provided further details and a few other noticeable interactions that were not detected by Kornberg (Figure 1.8). These include but are not limited to hydrogen bonds between S782, H1108, and N792 of RNAP II and the hydroxyl of 6-OH-Trp4, carbonyl of Cys8, and carbonyl of Hyp2, respectively.  Figure 1.7 Important interactions between α-amanitin and RNAP II. A) stereoview of the α-amanitin binding pocket (4Å). Red bonds represent α-amanitin. The funnel region of Rpb1 is in light green. Bridge helix can be seen in dark green. The region of Rpb2 near amanitin is in blue. B) the chemical structure of α-amanitin and the residues on the enzyme with a distance less than 4Å. Hydrogen bonds are shown in dashed lines with their lengths indicated.  There are a few crucial interactions that affect the toxicity of different amanitin derivatives to a great extent. The derivative amanullin (Table 1.1, entry 7) is similar to α-amanitin in structure, only lacking the hydroxyl groups on DHIle3, and is shown to be 4-fold less inhibitory compared to the natural product in in vitro transcription assays.2,28 Nevertheless, the same compound was resynthesized by D. Dietrich and was shown to be essentially nontoxic to CHO 13  cells.31 This result implicates the significance of hydrogen bonding between the hydroxyl group of DHIle and Gln783 and Gln718 of the backbone of RNAP II. These hydrogen bonds are seen more clearly in Figure 1.8, depicting their interaction as observed in the cryo-EM structure. However, the strongest and perhaps most important hydrogen bonding interaction occurs between the hydroxyl group of Hyp2 and Glu822 (or Glu845 according to Cramer). The lack of this hydroxyl in proamanullin (Table 1.1, entry 9) reduces the inhibition of RNAP II by the toxin almost 20,000 times.  Figure 1.8 Interactions of mammalian RNAP II with α-amanitin (reproduced with permission from Cramer et al)24. A) Schematic of the interactions. Chemical structure of α-amanitin is in orange. Rpb1 residues conserved over eukaryotes are labeled in black, whereas metazoan-specific amanitin-interacting residues are labeled in red. Green dashed lines represent hydrogen bonds, while other interactions are shown in black lines. B) Representation of the α-amanitin binding pocket within RNAP II. Blue and red colors show positively and negatively charged surfaces, respectively. The bridge helix, trigger loop, and Ser782 residue of Rpb1 can be seen.  Perhaps the most notable interaction that was observed in the cryo-EM structure obtained by Cramer and co-workers that was not detected in the co-crystal structure described by Kornberg et al., is the hydrogen bond between the hydroxyl group on the 6-hydroxy-L-Trp residue and Ser782. Although most synthetic derivatives of amanitin lack the hydroxyl groups 14  on 6-OH-L-Trp4 and DHIle3 amino acids, α-amanitin is such a strong inhibitor and binder of RNAP II that even a 4- or 5-fold decrease in potency may still provide access to a number of very cytotoxic analogues for use in targeted therapy.  1.2.4 Biosynthesis of α-Amanitin To date, there has not been a great deal of information related to the biosynthesis of α-amanitin. Nevertheless, two genes (AMA1 and PHA1) in Amanita bisporigera mushrooms have been identified to be responsible for affording the peptide backbone that is elaborated in the biosynthesis of α-amanitin and the structurally related bicyclic heptapeptide phallacidin (a member of phallotoxins, Table 1.2).11 First, a 35- (in case of α-amanitin) or 34- (in case of phallacidin) amino acid proprotein is made ribosomally. Subsequently, the proproteins are cleaved and macrocyclization is catalyzed by a prolyl oligopeptidase (POP). Walton et al. showed that a specific POP (GmPOPB) is required for this “toxin maturation” in Galerina marginata mushrooms, a distant relative of A. bisporigera.32 This process, together with the 35- and 34-amino acid proprotein sequences encoded by the two identified genes responsible for the biosynthesis of α-amanitin and phallacidin, are shown in Figure 1.9.    15   Figure 1.9 A) The proproteins of α-amanitin and phallacidin encoded by the identified genes, Ama1 and Pha1, in A. bisporigera.11 The residues in italic font are conserved in both genes, and the bold amino acids are specific to the toxin. Residues shown in red are responsible for the tryptathionine formation. B) Different steps of the toxin maturation catalyzed by GmPOPB, a specific prolyl oligopeptidase from G. marginata (figures adapted with permission from Walton et al.).32  The subsequent steps in the biosynthesis of amatoxins and phallotoxins involve the tryptathionine formation, oxidation of proline, tryptophan and isoleucine, and in case of α-amanitin, diastereoselective oxidation of the thioether bridge to give the corresponding (R)-sulfoxide. The order in which these transformations occur and the exact enzymes responsible for them are not known at this time. However, a promising candidate to carry out these alterations is a Flavin-dependent monooxygenase (FMO) that clusters between GmPOPB and GmAMA1 in the genome of Galerina marginata (GmFMO1). FMOs are known to perform regioselective hydroxylations and diastereo- and enantioselective sulfoxidations.33 Other classes of enzymes that could possibly be involved in hydroxylation of amatoxins and phallotoxins are cytochrome P450 monooxygenases (CYP450) and α-ketoglutarate-dependent dioxygenases.34  16  1.3 α-Amanitin in Antibody-Drug Conjugates  1.3.1 Antibody-Drug Conjugates: A Brief Overview Antibody-drug conjugates (ADCs) are an emerging class of cancer treatments. Although researchers have been developing these therapeutics for more than fifty years, it was only recently that a few ADCs were approved for clinical use (brentuximab vedotin in 2011 and trastuzumab emtansine in 2013). ADCs consist of three parts: a monoclonal antibody (mAb) which acts as the targeting agent, a cytotoxic chemical acting as a “warhead”, and a linker that covalently attaches the antibody to the toxin (Figure 1.10).35 The concept of ADCs is simple: the antibody targets an antigen that is overexpressed in the targeted diseased cells. Once the antibody-antigen complex is formed, the cytotoxin that is attached to the antibody, via the linker, is internalized into the cell. Linkers are usually designed to be cleaved under the physiological conditions found in the cells: hydrolysis, cleavage by an enzyme, or bio-reduction in hypoxic cells. However, depending on the stability of the linker-toxin bond, off-target toxicity can occur if the cytotoxin is released outside of the targeted cells. Hence, non-cleavable linkers are gaining popularity due to their higher stability and decreased possibility of off-target activity. In the case of non-cleavable linkers, it is important that the toxin bearing the linker remains toxic and retain a similar cytotoxicity profile as the cleaved toxin.36 17   Figure 1.10 A) A general ADC and its components. B) Mechanism of action of an ADC (figure adapted from Chen et al.).37  Selection of a suitable antibody for use in ADCs has been extensively explored over recent years. The most important feature of an antibody that must be considered prior to its incorporation into an ADC is its origin: a fully human antibody is strongly preferred for use in ADCs rather than antibodies from other sources (such as murine) to suppress immunogenicity once the ADC is administered.36 The recurrent format of antibodies used in ADC is human IgG isotypes, such as IgG1 and IgG2. Trastuzumab, brentuximab and gemtuzumab are common antibodies employed in ADCs, to name a few. Selecting the toxic payload (the “drug”) of an ADC is also of great importance.38 There are only a limited number of overexpressed receptors on the surface of a cancer cell, hence the number of drug molecules being internalized into the cell can be extremely low. Therefore, higher drug-to-antibody ratios (DAR) are generally sought. Moreover, warheads with great potency towards their targets, i.e. an IC50 in the pico- or nanomolar range, are ideal for ADCs. There are two general types of drugs used in ADCs: ones that target DNA and those that target microtubules. Regardless of the type, several key requirements need to be met for a drug to be utilized in an ADC. If the 18  drug is too hydrophobic, it may induce antibody aggregation that causes fast clearance rates and immunogenicity.35,39 Also, a suitable drug needs to maintain its potency when it is attached to a linker or fragment of the linker that remains attached to the drug post-cleavage. Additionally, it must be water soluble, stable under physiological conditions, and synthetically or biosynthetically accessible.38,40 Nearly all the aforementioned criteria are met by α-amanitin: it is extremely potent (Ki ~ 1-10 nM), reasonably hydrophilic and water soluble, and retains its potency when attached to a linker.5,41   1.3.2 Amanitin-Containing ADCs The first antibody-amanitin conjugate that resembled modern ADCs can be traced back to 1981, when α-amanitin was conjugated to a monoclonal IgG antibody of Thy 1.2 (anti-Thy 1.2) and was tested against murine T lymphoma S49.1 cells.42 Since then, several conjugates of α-amanitin or its derivatives have made their way into pre-clinical or clinical trials.5,43,44 For example, a conjugate of α-amanitin and an anti-EpCAM (an antibody targeting human epithelial cell adhesion molecule, EpCAM, which is overexpressed in many cancers) was used to treat pancreatic carcinoma in mice, and is now in clinical trials.5 In this work by Moldenhauer et al., α-amanitin was conjugated to the chimeric antibody chiHEA125 via the δ-hydroxyl of DHIle residue using a glutarate linker. It was shown that conjugating α-amanitin to chiHEA125 did not impact the binding affinity of the antibody to the target antigen EpCAM in Colo205 cells (Figure 1.11, A). Then, in two separate experiments, mice bearing 10-day-old pancreatic carcinoma xenografts were treated with a single dose and two doses of the chiHEA125-amanitin. The relative tumor volume in each experiment was measured every third day until day 16. Moldenhauer and co-workers found that in the case of the mice that were 19  injected once with the amanitin conjugate, the tumor completely regressed in 50% of the mice (3 of n=6), whereas a significant reduction in tumor volume was observed for the rest (Figure 1.11, B). In the second experiment, mice bearing the pancreatic carcinoma xenograft were administered two doses, one week apart, with the chiHEA125-amanitin conjugate. At the endpoint of the treatment (day 16), tumors were no longer detectable in six (60%) of 10 and nine (90%) of 10 mice treated with chiHEA125-amanitin at a dose of 50 and 100 µg/kg, respectively (Figure 1.11, C). More recently, conjugates of α-amanitin and RGD and isoDGR have been synthesized that show very promising results in tumor targeting.44 Additionally, it has been shown that α-amanitin can prevent cancer relapse in mice bearing tumor xenografts.43 These results demonstrate the potential of α-amanitin as a major toxic payload in future ADC technology. While discussing different linkers and antibodies that can be used in conjunction with α-amanitin is beyond the scope of this thesis, it will be useful to review the various sites of conjugation on α-amanitin. A common functional group on α-amanitin through which the toxin is attached to a linker is the 6-hydroxyl of 6-OH-Trp4. The 5-hydroxyl of DHIle3 is another popular site of conjugation. However, these sites are only available in naturally occurring α-amanitin or the S-deoxy-amanitin derivative. Recently, amanitin conjugates are also accessible via CuAAC (copper catalyzed alkyne-azide cycloaddition) chemistry (Zhao et al.). This method is especially useful for conjugation of synthesized amanitin analogues with an alkyne functionality, usually appended on the side-chain amide of Asn1, and has been used to make an RGD conjugate of a cytotoxic amanitin derivative (Figure 1.12).25  20   Figure 1.11 ADC of α-amanitin and chimeric antibody chiHEA125 (reproduced with permission from Moldenhauer et al.). A) Structure (left) and binding of chiHEA125-amanitin conjugate to EpCAM (right). Colo205 cells were incubated with various concentrations of unconjugated chiHEA125 and the α-amanitin conjugate and analyzed by flow cytometry. Experiments were performed in triplicates. B) Efficiency of a single dose (50 µg/kg) of chiHEA125-amanitin in mice (n=6) bearing 10-day-old pancreatic carcinoma xenografts. Control mice (n=4) received unconjugated chiHEA125. Left: relative tumor volume vs. days post-treatment. Tumor growth was measured every third day. Right: representative tumors at the endpoint of treatment (day 16); magnification: x2.5 C) Efficiency of two doses of chiHEA125-amanitin, administered one week apart, in mice bearing 10-day-old xenografts. Mice received IP injections of the ADC (10, 20, 50 or 100 µg/mg; n=8, 7, 10, 10 mice per group, respectively) or control unconjugated chiHEA125 (n=9). Tumor growth was measured every third day.  21   Figure 1.12 Possible sites of conjugation on α-amanitin. A) 5-hydroxyl of DHIle3, B) 6-hydroxyl of Trp4, C) side-chain amide of Asn1 via CuAAC.  1.4 Derivatives of α-Amanitin Today, nearly all bioconjugates of α-amanitin are synthesized using the natural product extracted from amanitin-producing mushrooms. However, aside from these bioconjugates to the natural product, there are few reports of synthetic derivatives that retain the toxicity of the natural product. Indeed, most synthetic derivatives have been derived from the natural product itself or from toxins that have been prepared, but which lack the DHIle and thus are considerably less active in vitro and on cells than the natural product (vide infra).7,27–30,45 Indeed, in the absence of either an enantioselective synthesis of DHIle or a total synthesis of the natural product itself (vide infra), a comprehensive understanding of key structure activity relationships has been impeded by lack of synthetic access. Several modifications that have been used include selective alkylation of the hydroxyl-indole,46 acylation of the DHIle diol,47 periodate cleavage of the diol to give oxo-valine which is reductively aminated,48 and azo-coupling to the indole.49 These reactions, while numerous and in many cases both chemo-selective and thus regioselective, are cumbersome and not 22  without the possibility of side-reactions. As previously mentioned, Zhao et al. introduced an alkyne to afford highly chemo-selective conjugation, yet this synthesis suffered from inefficiency in terms of an enantioselective synthesis of DHIle. In light of these challenges and limitations, most of the SAR data dates from the late 1970s to early 1990s. While a detailed discussion of a wider range of analogues and their synthesis will be provided in the upcoming chapters, several amanitin derivatives are reviewed below. These amanitin analogues, although in most cases less potent than the natural product, could prove specifically useful in ADCs, since the fermentation yields of α-amanitin from Amanita mushrooms are low and it is often expensive or troublesome to access the natural product.50  1.4.1 Dihydroxyisoleucine Derivatives The two hydroxyl groups on DHIle3 can provide a useful handle for chemical modifications on this residue of α-amanitin. In addition to some of the naturally occurring amatoxins that lack one or two of the hydroxyl groups of DHIle3 (Table 1.1, entries 3, 7-9) and show reduced activity compared to α-amanitin, different oxidation states for the alcohols, cleavage of the 1,2-diol, and possible subsequent manipulations of the product of diol cleavage allow access to certain analogues of α-amanitin (Figure 1.13).48 23   Figure 1.13 Derivatives of α-amanitin obtained from the diol cleavage at the DHIle3 position.48  Furthermore, amanitin analogues containing 4-hydroxy-isoleucine or isoleucine and other hydrophobic amino acids in place of DHIle3 have been synthesized (Table 1.4).7,28 Based on inhibition studies, it was shown that analogues lacking the 4-hydroxy of DHIle3 were 4- or 5-fold less active, while the 5-hydroxy is not as vital for activity. # Xaa Ki (rel α) Ki (rel γ) 13 Ile 32 - 14 D-Ile - 100 15 allo-Ile - 100 16 Val 128 - 17 Leu 1400 - 18 Ala 2700 -  Table 1.4 Analogues of α-amanitin derivatized at DHIle3 position and their inhibition constants relative to α-amanitin (against calf thymus RNAP II) or γ-amanitin (against D. melanogaster embryos RNAP II).7,28 Note that all the analogues lack the 6-hydroxy of Trp4 and the sulfoxide.  24  1.4.2 Other Chemical Modifications The DHIle3 position has been extensively modified due to the lack of a robust synthetic route to stereochemically pure 4,5-dihydroxyisoleucine. However, there are other positions that have been modified to a lesser extent, including Asn1, Gly5, Ile6 and Gly7.28,29,45 The results of these investigations are shown in Table 1.5. Interestingly, replacing Ile6 with L-alanine causes a drastic decrease in the inhibition constant (five orders of magnitude). Additionally, substituting Asn1 with aminobutyric acid (Abu) resulted in complete loss of activity. The effects on the binding constant from modifications at other residues were not as profound; however, they all exhibit reduced activity compared to the natural product. # Xaa1 Xaa5 Xaa6 Xaa7 Ki (rel) 19 Asn Gly Ile Gly 32 20 Asn Ala Ile Gly 7200 21 Asn Gly Ala Gly 22400 22 Asn Gly Ile Ala 110 23 Asn D-Ala Ile Gly ~1000 24 Asn Gly Ile D-Ala ~30 25 Abu Gly Ile Gly n/a Table 1.5 Important analogues of α-amanitin modified at positions 1, 5, 6 and 7. The inhibition constants were measured against calf thymus RNAP II (19 to 22) or D. melanogaster RNAP II (23 to 25).28,29,45 Note that all the analogues contain Ile at position 3 and not DHIle. (n/a = no activity).  1.5 Chemical Synthesis of Amanitin Analogues While the synthesis of α-amanitin and its analogues will be further discussed later, a brief overview of synthetic challenges, general strategy and previous efforts to synthesize these peptides is provided in this section. Since prior to this work there was no accessible synthesis for the natural product, herein we will only briefly discuss synthetic pathways to analogues of α-amanitin.  25  1.5.1 Overall Strategy Aside from several unnatural amino acids that need to be chemically synthesized (i.e. Hyp2, DHIle3, 6-OH-Trp4), the key step in forming the bicyclic structure of α-amanitin involves a reaction between cysteine and an oxidized tryptophan. The crosslink that forms between these two amino acids is called a tryptathionine (Ttn) linkage.51 This crosslink is characteristic of amatoxins and phallotoxins, it is unique among all natural products, and is critical for the activity of α-amanitin and other members of the amatoxin family. While a more detailed discussion regarding tryptathionine linkage formation will be provided in Chapter 3, an overview of the most common method to synthesize this crosslink is shown in Figure 1.14. First, tryptophan is oxidized to a hydroxypyrrolo indoline (Hpi) moiety, which is then reacted with a suitably protected cysteine in the presence of trifluoroacetic acid (TFA) to yield the tryptathionine linkage. The latter transformation is called the Savige-Fontana reaction, named after W. Savige and A. Fontana who performed a less efficient version of this reaction for the first time in 1976, using acetic acid instead of TFA.52  Figure 1.14 General conditions to form a tryptathionine crosslink. Oxidation of tryptophan followed by the Savige-Fontana reaction using TFA to induce tryptathionylation.52  In the context of α-amanitin and its analogues, depending on the synthetic strategy, Hpi is incorporated into the linear sequence of amanitin, and the Savige-Fontana reaction will 26  crosslink Cys8 and Trp4 to form the first cycle which can eventually be carried forward to the final bicyclic peptide (Figure 1.15). Once the monocyclic heptapeptide is formed through the Savige-Fontana reaction, DHIle3 (or any other DHIle-like residue) can be coupled to the sequence. Then, macrolactamization is performed to close the second cycle and the final bicyclic octapeptide is obtained.  Figure 1.15 General approach for tryptathionylation in the synthesis of amanitin analogues, followed by coupling of the residue as position-3 (Xaa3) and macrolactamization to yield the final amanitin analogue.  1.5.2 (2S,3R,4R)-4,5-Dihydroxyisoleucine As will be discussed in Chapter 2, (2S,3R,4R)-4,5-dihydroxy-L-isoleucine residue of α-amanitin might be one of the most challenging unnatural amino acids to synthesize. Prior to this work, there has been no efficient enantioselective synthesis of DHIle, and almost all amanitin analogues that have been chemically synthesized lack the DHIle3 residue and have used Ile in place of this small yet challenging amino acid. As mentioned above, the cryo-EM structure of α-amanitin bound to RNAP II shows hydrogen-bonding interactions between the hydroxyl groups of DHIle and the Rpb1 subunit of the enzyme. Furthermore, it has been shown 27  that none of the other synthetic analogues of α-amanitin with amino acids other than DHIle at position-3 have the same toxicity as the natural product. Hence, it is crucial to devise an efficient synthetic route to provide access to a stereochemically pure version of this amino acid (vide infra).  1.5.3 6-Hydroxy-L-Tryptophan and (R)-Sulfoxide To date, there is only one method reported in the literature that can produce 6-hydroxy-Trp as an individual amino acid.53 This method involves the inconvenient use of H2O2 in HSbF5 and is low yielding (vide infra). Moreover, there is no reported oxidation of 6-OH-Trp to the corresponding Hpi that can be incorporated into amanitin analogues to yield the tryptathionine linkage following a Savige-Fontana reaction. Consequently, no amanitin analogue has been synthesized that contains a 6-hydroxy-tryptathionine crosslink. Additionally, since it has been shown that the sulfoxide is not vital for the activity of the toxin, all synthetic amanitins contain the unoxidized thioether at the tryptathionine position. These two synthetic challenges and how they were overcome will be further discussed in Chapters 3 and 4.  1.6 Three-dimensional Structure of α-Amanitin A strong allosteric inhibitor of RNAP II, α-amanitin owes its utmost binding affinity and physiological stability partly to its unique 3D structure. The tryptathionine crosslink and the Hyp2 residue are considered the primary factors responsible for the unique overall conformation of α-amanitin, resulting in a β-turn similar to that of proteins. Whereas the 3D structure of α-amanitin and several of its derivatives will be further discussed in Chapter 4, herein I disclose XRD structures of a few amanitins that have previously been reported. Due 28  to a lack of an XRD structure and inability to recrystallize α-amanitin, the most analogous natural amatoxin to α-amanitin regarding its overall conformation is β-amanitin, which has been successfully recrystallized and studied by Kostansek and co-workers (Figure 1.16, A).54 Differing only in the residue at position-1, β-amanitin (bearing Asp1) is believed to possess a similar conformation to α-amanitin (containing Asn1). Another derivative of α-amanitin for which an XRD structure has been obtained is O-Me-α-amanitin sulfone. This derivative, containing a methyl ether at C-6 of Trp4 and an oxidized sulfone tryptathionine linkage, is also thought to exhibit an analogous conformation to that of α-amanitin (Figure 1.16, B).46 Lastly, Pro2-Ile3-S-deoxo-amaninamide (an inactive analogue of α-amanitin) was also recrystallized and studied by Perrin and May (Figure 1.16, C).27  Figure 1.16 XRD structures of A) β-amanitin (Kostansek et al.), B) O-Me-α-amanitin sulfone (Shoham et al.), and C) Pro2-Ile3-S-deoxo-amaninamide (Perrin et al.).  29  1.7 Project Goals Despite the fact that α-amanitin was discovered and isolated from Amanita mushrooms over 60 years ago, most of the literature about this toxin revolves around its biology and analogues that are derived from the natural product itself. Thus, we focused on the synthetic challenges of accessing this toxin, which may prove to be as important as its biology, and thereby open doors to facets of α-amanitin that have not been explored previously.  1.7.1 The Total Synthesis of α-Amanitin At the time that I undertook this work, there was no synthesis of α-amanitin. Indeed, it is well-appreciated by the chemical community that a total synthesis represents a conquest that renders the toxin amenable to synthetic access, which empowers the production of new derivatives that can be used to flesh out SARs as well as develop new antibody drug conjugates. From a fundamental perspective, a full set of SAR data would be needed to properly validate the molecular basis for toxicity and validate key interactions that have been observed in X-ray and cryo-EM studies. From an applied perspective, such SAR data can also be interfaced with bioconjugation chemistries to afford new and more selective means of conjugation while preserving the needed toxicity. Hence, the main goal of this project is to accomplish the first total synthesis of α-amanitin. As the fermentation yields are low, an efficient synthetic route to this toxin will allow for the large-scale synthesis of α-amanitin, a process that will undoubtedly be useful for the scientific and biomedical community. A robust synthetic pathway to this toxin will enable chemists to modify α-amanitin at any given position, paving the way to a wide variety of analogues of amanitin. These analogues will significantly help SAR studies of α-amanitin and could possibly lead to discovering a “super toxin”: a toxin that 30  is more potent than the natural product. To achieve this goal, the synthetic challenges that were discussed before needed to be surmounted. Furthermore, being able to modify different positions, one would be capable of conjugating α-amanitin to an antibody through different residues. Whether this process if performed via etherification or CuAAC chemistry, the conjugation site could vary based on available tools or the required chemistry for the given ADC.  1.7.2 The Proline Series As discussed before, the hydroxyproline (Hyp2) residue plays a vital role in the toxicity and binding affinity of α-amanitin. However, there has been little-to-no traction on the synthesis of analogues of α-amanitin with modifications at this position. Taking advantage of this fact, I have sought to pursue the synthesis of a series of hydroxyproline analogues in hope of gaining a better understanding of the importance of this position, and the possibility of discovering an increasingly potent toxin. To this end, several analogues of trans-hydroxyproline will be designed and synthesized to be incorporated into amanitin analogues. These analogues will then be tested for biological activity (Figure 1.17).  Figure 1.17 Overview of the synthesis of Hyp analogues and their incorporation into amanitin to synthesize the corresponding analogues.  31  Additionally, a photo-cleavable amanitin will be designed and synthesized. Taking advantage of the fact that amanitin analogues lacking the hydroxyl group of Hyp2 are greatly less toxic than the natural product, masking the hydroxyl of Hyp2 with a photocleavable protecting group will give us access to amanitins with controlled toxicity. In this case, Hyp2 will be protected with a photolabile protecting group to form a non-toxic peptide. Once irradiated with the proper wavelength, it will release the unprotected, fully toxic amanitin. While the concept behind the photo-cleavable proline analogue will be fully elaborated in Chapter 5, an overview is shown in Figure 1.18.  Figure 1.18 Overview of a photo-cleavable amanitin. The protected amanitin is not cytotoxic; however, its irradiation with an appropriate wavelength will release the unprotected, cytotoxic amanitin.  In conclusion, with a fully synthetic α-amanitin in hand, elaborate SAR studies of this toxin will become accessible. Chemical alterations of any given residue followed by its incorporation in the bicyclic structure of amanitin can lead to a pool of analogues, designed systematically based on the residue of interest, will provide further insight into the mechanism of action of this captivating toxin and the fashion in which it interacts with cells (e.g. cell penetration and binding to RNAP II). Aside from the fundamental studies made feasible by the total synthesis, the eventual design and synthesis of various amanitin-containing ADCs is another compelling feature of this work. With the emergence of amanitin-antibody conjugates as novel cancer therapeutics, the need for a synthetic route to α-amanitin is felt more than ever, 32  since the natural source for this toxin (i.e. Amanita mushrooms) is limited and extraction of α-amanitin from mushrooms is inconvenient and costly. Finally, synthesis of a broad range of proline derivatives for incorporation in amanitin derivatives represents several major opportunities: i) the possibility of discovering a super toxin, a toxin possessing a higher potency towards cells and RNAP II than the natural product, ii) employing some of the proline derivatives, e.g. azido or keto-proline, as handles for conjugation to biomarkers for preparation of amanitin-ADCs, iii) providing further intuition on the significance of Hyp2 residue of α-amanitin to its binding affinity, leading to systematic SAR investigations revolving around this residue, and iv) in the case of photo-cleavable Hyp, synthesis of a non-toxic amanitin with a masked Hyp that could be activated following irradiation with appropriate wavelength, giving access to an amanitin with controlled toxicity.             33  Chapter 2: Synthesis of (2S,3R,4R)-4,5-Dihydroxyisoleucine  2.1 Introduction An enantioselective route to the synthesis of (2S,3R,4R)-dihydroxyisoleucine residue (DHIle) (Figure 2.1, 26) is one of the most notable challenges that needed to be overcome en route to the total synthesis of α-amanitin. Due to a lack of an efficient synthesis, almost all synthetic derivatives of α-amanitin contain a variant of this unnatural oxidized amino acid e.g. monohydroxy isoleucine, isoleucine, homoserine or other analogues.7,29,31,45 Whereas a single report stated that isoleucine could replace the DHIle with approximately 10-fold reduction in affinity,28 most others emphasize that this residue is critical for toxicity,7,55 a notion that is confirmed in part by XRD studies.56,57 In the absence of an enantioselective synthesis of DHIle, Perrin et al.25 reported the synthesis of four related amanitin diastereomers that emanated from an inefficient diastereoselective synthesis of DHIle that resulted in a diastereomeric pair of enantiomers that could not be resolved by standard silica chromatography; instead each was incorporated into the final octabicyclic peptides (vide infra), each of which represented a unique diastereomeric analog of amanitin that could be resolved by HPLC. Since only one diastereomer proved highly toxic in a cell-based assay, the authors surmised that this derivative contained the stereochemistry of the natural product although no independent structural data were available to corroborate this supposition. These results highlighted what seems to be a significant selectivity for a single diastereomer of DHIle and provided an impetus to address an enantioselective synthesis of DHIle. Despite its simple structure, DHIle has defied synthetic access due to considerable synthetic challenges.  34  2.1.1 Synthetic Challenges As reflected by the presence of a carboxylic acid, an amine and two hydroxyl groups on a 5-carbon chain, one of the foremost challenges in the synthesis of DHIle is the installation of a high density of functional groups in this amino acid with an all-anti configuration. Furthermore, this residue is a β-branched amino acid bearing a methyl on C-3. Since no other stereoisomer of this amino acid provides a similar toxicity for amanitin compared to the (2S,3R,4R) isomer, elaboration of the correct absolute configuration at each of the stereogenic centers is essential to addressing the activity of the synthetic toxin. A propensity of DHIle to lactonize, even under mildly acidic conditions, represents another problem to overcome in the synthesis of this amino acid. As shown in Figure 2.1, the presence of the γ-hydroxyl can easily result in the formation of a γ-lactone (27) when treated with a weak acid (pH<7).58 This may subsequently cause an epimerization at Cα in the presence of a non-nucleophilic base due to the greater stability of the trans-trans configuration (28). Lactonization can also occur on the final peptide; however, such requires treatment with a stronger acid, e.g. concentrated HCl. Such an N-to-O acyl transfer under acidic conditions results in the cleavage of α-amanitin and potential inactivation of the toxin.59 Subsequent hydrolysis of the lactone under the same conditions would arguably render the cleavage of the peptide irreversible.  35   Figure 2.1 Chemical structures of DHIle and its lactone. A) lactonization of the free amino acid in the presence of acid. C-2 may be epimerized to produce the more stable trans-trans isomer 28. B) lactonization of DHIle3 (in blue) on α-amanitin that leads to the cleavage of the toxin.  Such potential implies a need for diol protection to avoid lactonization during DHIle activation and coupling as well as during peptide deprotection. Finally, the synthetic DHIle must also be protected for incorporation into amanitin via solid-phase peptide synthesis (SPPS). Hence, orthogonality of the protecting groups with the solid phase strategy and the chemical nature of other residues of the peptide is of great importance. Before discussing the previous efforts towards the synthesis of DHIle, it will be insightful to provide a brief overview of the general methods employed for the synthesis of β-branched amino acids.  2.1.2 Synthesis of β-Branched Amino Acids β-Branched amino acids are valuable building blocks for the synthesis of a wide variety of natural products.60 Various methods have been utilized to synthesize unnatural β-branched amino acids. In all these methods, the stereochemical control over C-C bond formation reaction, namely between Cβ and the β-substituent is critical. While there are more recent and 36  less common methods recognized in the literature, herein we will review some of the more frequently used routes towards the synthesis of β-substituted amino acids.  2.1.2.1 Garner’s Aldehyde Introduced in 1984 by Garner61, Garner’s aldehyde (29) is probably one of the most widely used building blocks for the stereoselective synthesis of branched amino acids.62 Through a 4-step synthesis, a configurationally stable 1,1-dimethylethyl-4-formyl-2,2-dimethyl-oxazolidine-3-carboxylate (L-29) can be obtained from L-serine (Figure 2.2).  Figure 2.2 Synthesis of Garner’s aldehyde (29) (Garner et al.). The corresponding D-29 may be obtained starting with D-serine, eventually resulting in the formation of a β-branched D-amino acid.  Garner’s aldehyde can be utilized to synthesize β-branched amino acids in different ways. In one method, it is first treated with a stabilized phosphonium ylide to afford an (E)-α,β-unsaturated ester (30). Then, treatment of 30 with metal dialkylcuprates via a Michael addition on the α,β-unsaturated system leads to the formation of a syn product (31), which then may be converted to a functionalized amino acid (32) (Figure 2.3). The preference to form the undesired syn product can be explained by the addition of the nucleophile from the si face of 37  the β-carbon according to the Felkin-Ahn model.62  Furthermore, this approach would not only give the wrong orientation, but would result in a D-amino acid product, necessitating use of D-serine at the outset. Additionally, 30 can be reduced to the corresponding alcohol (33) with DIBAL, which then may be transformed into a leaving group (34). Michael addition of 34 with a metal dialkylcuprate followed by proper chemical manipulations can furnish a 4,5-unsaturated-β-branched amino acid (35).63,64  Figure 2.3 Synthesis of syn-β-branched amino acids using the Garner’s aldehyde via Michael addition. The Falkin-Ahn model for the si-face addition of the nucleophile to form syn-31 is shown in the dashed box.  Another method involving the Garner’s aldehyde takes advantage of a chiral bromo allene (Figure 2.4). First, aldehyde 29 is reacted with lithiated trimethylsilylacetylene to produce alkynol 36. Then, 36 is converted to the corresponding mesylate and treated with lithium diboromocuprate, yielding chiral bromo allene 37. Addition of a lithium dialkyl cuprate (Gilman’s reagent) produces alkyne 38, which may subsequently be converted to a functionalized anti-β-branched amino acid (39).65 38   Figure 2.4 Synthesis of anti-β-branched amino acids starting with the Garner’s aldehyde via a chiral bromo allene (37).65  2.1.2.2 Sigmatropic Rearrangement of the Overman’s Imidate In 1974, Overman introduced the thermal and mercuric ion catalyzed [3,3]-sigmatropic rearrangement of allylic trichloroacetimidates.66 A decade later, he expanded this transformation to a variety of substrates to form carbon-heteroatom bonds, including the rearrangement of imidates to amines. Moreover, Overman showed that this transformation could be catalyzed using Pd(II) in addition to Hg(II) salts.67 Having access to chiral imidates, one may employ them as precursors to chiral β-branched amino acids. As an example, this method was used to synthesize (2S,3S,4R)-γ-hydroxyisoleucine (49) in an enantioselective fashion (Figure 2.5).68 It is not unreasonable to surmise that the desired (2S,3R,4R)-dihydroxyisoleucine could be effected by a similar route if one were to apply this route starting with ethyl (R)-3,4-dihydroxybutanoate. 39   Figure 2.5 Synthesis of (2S,3S,4R)-γ-hydroxyisoleucine (49) via [3,3]-sigmatropic rearrangement of imidate 46.68 The transition state for the palladium-catalyzed rearrangement of 46 to 47 is shown in the dashed box.  2.1.2.3 Claisen Rearrangement In 1995, Kazmaier reported the first asymmetric ester-enolate Claisen rearrangement for the synthesis of chiral γ,δ-unsaturated amino acids.69 In their report, Kazmaier and co-workers investigated the Claisen rearrangement of N-(trifluoroacetyl)glycine crotyl ester (or other alkyl esters containing a trans-olefin) using LDA and different metal salts. When accompanied by a chiral ligand, such as quinine or quinidine, this method proved effective for obtaining syn-β-branched amino acids bearing a γ,δ-unsaturated system (Figure 2.6).  40   Figure 2.6 Ester-enolate Claisen rearrangement of N-TFA-glycine (E)-crotyl esters (50) for the synthesis of syn-γ,δ-unsaturated-β-branched amino acids (51). The D-isomer may be obtained when quinine is used in place of quinidine.69  Although extremely useful for the synthesis of syn-β-branched amino acids, this method is not considered to be a practical route to produce anti isomers of these amino acids. To obtain the anti diastereomer, one must start the synthesis with (Z)-crotyl esters of glycine, which are the less stable E/Z isomer and are not always commercially available. Furthermore, (Z) isomers destabilize the chair-like transition state of the rearrangement, which would be expected to lead to poor diastereoselectivities. An alternative method to synthesize anti-β-branched amino acids involves the use of an Eschenmoser-Claisen rearrangement. In a study by Hruby et al.,70 it was found that this rearrangement was effective for synthesizing anti diastereomers of β-branched amino acids (Figure 2.7). Their synthesis started with a pyrrolidine or diisopropyl amide of N-Cbz-glycine (52). Treating 52 with methyl triflate yielded the Meerwein salt of the amide (53). Once 53 was treated with lithium allyl oxide (54), intermediate 55 was formed which subsequently underwent rearrangement to the anti-β-substituted product (56). The last step involved hydrolysis of the amide bond to obtain the anti-β-branched amino acid. To minimize epimerization at Cα during the hydrolysis step, an in situ oxidation procedure was employed to afford 57 as the final product. 41   Figure 2.7 Synthesis of anti-β-branched γ,δ-unsaturated amino acids via an Eschenmoser-Claisen rearrangement.70  Starting with a chiral amide (52), such as (2R,5R)-dimethylpyrrolidine amide, one may perform an asymmetric version of the Eschenmoser-Claisen rearrangement to obtain a stereochemically pure anti-β-branched amino acid.71 It is appreciated that the same method above could be used to obtain the desired stereoisomer of dehydroisoleucine which could be subjected to Upjohn conditions to provide both the desired (2S,3R,4R)-dihydroxyisoleucine as well as its epimer, (2S,3R,4S)-dihydroxyisoleucine. While a means of resolving these two epimers could likely be established, it is known that such compounds are often difficult to separate. Whereas other methods have been developed for the synthesis of β-branched amino acids in the recent years, the aforementioned routes are the most commonly applied strategies reported in the current literature. Next, previous efforts towards the synthesis of DHIle will be described.    42  2.1.3 Previous Syntheses of Dihydroxyisoleucine  2.1.3.1 Synthesis by Wieland Theodor Wieland was one of the pioneers in the study of α-amanitin and, in addition to exploring the biochemistry of this toxin, he expressed interest in its chemical synthesis. To this end, in 1967, Wieland et al. reported a completely non-stereoselective synthesis of the DHIle lactone (Figure 2.8).72 In their synthesis, diethyl ethylidenemalonate (58) was subjected to a Michael addition with vinyl magnesium bromide, followed by oxime formation and then zinc reduction, afforded acetamide 60. Bromination of the resulting olefin and hydrolysis of the dibromide in the presence of silver(II) sulfate produced all eight stereoisomers of the corresponding lactone (62). In this synthesis, no stereoselectivity was induced at any step, hence a racemic mixture of all eight stereoisomers was obtained as the final product.  Figure 2.8 Non-stereoselective synthesis of DHIle lactone by Wieland et al.  2.1.3.2 Synthesis by Bartlett  In the late 1970’s and the early years of 1980’s, workers in the Bartlett group investigated the ester-enolate Claisen rearrangement (vide supra) as a method to produce γ,δ-unsaturated amino acids.73 While studying different pathways to synthesize this class of amino acids, they attempted the stereoselective synthesis of DHIle. They proposed that γ,δ-dehydro-L-isoleucine 43  could be used as the precursor for this synthesis, and that a stereoselective iodolactonization74 of the olefin could afford the desired DHIle (Figure 2.9).58  Figure 2.9 Retrosynthetic scheme for the stereoselective synthesis of DHIle (Bartlett et al.).  To begin the synthesis, the trans-crotyl ester of N-Boc-glycine (64) was prepared. Deprotonation of 64 with lithium isopropylcyclohexyl amide and silylation with TMSCl, followed by reflux and deprotection of the TMS yielded the dehydroisoleucine intermediate (65) as a 10:1 mixture of diastereomers. To confirm the configuration of stereogenic carbons in this intermediate, the olefin was catalytically hydrogenated, the Boc group was removed, and the 13C-NMR of the product was compared to that of L- and allo-isoleucine. It was found that the obtained product was in fact allo-isoleucine (66), and not the desired L-isoleucine (Figure 2.10).  Figure 2.10 Synthesis of dehydro-allo-isoleucine starting from trans-crotyl ester of Gly (64).  44  The diastereoselectivity of this series of reactions is explained by the proposed chair transition state for the Claisen rearrangement, suggesting that in the favored enolate structure, the enolate oxygen and the deprotonated Boc are cis relative to each other (Figure 2.11).73  Figure 2.11 The anionic intermediates and the chair-like transition state for the Claisen rearrangement of trans-crotyl ester of N-Boc-glycine.  Based on these results, to synthesize dehydro-L-isoleucine, the cis-crotyl ester of glycine (68) was used as the starting material in place of the trans isomer. Although the cis-crotyl ester can be obtained in an analogous way to the trans-isomer, it would be more practical to prepare the 2-butynyl-ester of glycine (67) and hydrogenate via catalytic hydrogenation to afford the cis-crotyl ester (68). Through the same series of reactions, a mixture of D/L- and allo-dehydroisoleucine isomers was produced, this time in favor of the D/L enantiomeric pair (6:1) (65b). The optimized conditions for the iodolactonization step involved employing iodine in acetonitrile in the absence of base.75 Different protecting groups for the amine were screened and it was found that phthaloyl group yielded the best diastereoselectivity. However, this step also produced 5% of the (β,γ)-cis product (70b) in addition to the desired (β,γ)-trans isomer (70a). Hydrolysis of the iodinated lactone (70a) to the final lactonized product proved problematic. After examining several displacement conditions, it was observed that using 45  silver(I) trifluoroacetate salt followed by the acid hydrolysis of the crude product (71) mainly resulted in the desired stereoisomer of the lactone (72a), however not without 10% epimerization at Cα. Finally, the 1H-NMR of 72a was acquired and compared to that of the compound synthesized by Wieland et al. (62). In Wieland’s synthesis, four doublets were observed for the β-methyl groups of the four diastereomers present in 62 (δ 1.08, 1.13, 1.25 and 1.33 ppm). Whereas the doublet at 1.25 ppm was claimed to arise from the diastereomer found in the natural α-amanitin, the doublet at 1.33 ppm resulted from 2-epi diastereomer (72b). Following the investigation of the 1H-NMR of 72a, Bartlett found that it contained a doublet at 1.25 ppm, confirming the presence of the desired diastereomer of 72a in the final product mixture. This diastereomer was recrystallized from MeOH/Et2O to furnish the pure racemate of dehydro-L-isoleucine lactone. This synthetic pathway is summarized in Figure 2.12. 46   Figure 2.12 Diastereoselective synthesis of the DHIle lactone (Bartlett et al.).  Although this synthetic route provided decent diastereoselectivity for the racemic synthesis of the desired diastereomer, it was not enantioselective and the desired diastereomer was inseparable from an intractable racemic mixture. A further disappointment was noted upon realization that during the hydrolysis of the iodolactone to the DHIle lactone, 10% epimerization at Cα had occurred, thus further decreasing the efficiency of the overall route.   2.1.3.3 Previous Syntheses by Perrin and Co-workers Two attempts were previously made by Perrin and co-workers to synthesize DHIle in a stereoselective manner. Both methods involved an ester-enolate Claisen rearrangement, previously employed by Bartlett et al., followed by the dihydroxylation of racemic 47  dehydroisoleucine; one utilized a variant of Sharpless asymmetric dihydroxylation (AD) and the other a non-asymmetric Upjohn’s version (vide infra).   2.1.3.3.1 Claisen Rearrangement Followed by Sharpless AD In a work that was reported in a thesis by Dietrich from the Perrin group,31 the first step of the DHIle synthesis, analogous to Bartlett’s synthesis, involved an ester-enolate Claisen rearrangement of the cis-crotyl ester of N-Cbz-glycine (73) to produce a mixture of Cbz-protected isoleucine and allo-dehydroisoleucine (74a:74b 9:1) with each diastereomer existing as a racemate. Converting the components of the mixture to indoline amides enabled separation of two pairs of diastereomers in racemic form, resulting in the diastereomerically pure indoline amide of D/L-dehydroisoleucine (75) (Figure 2.13).  Figure 2.13 Synthesis of L-dehydroisoleucine indoline amide (75) by Dietrich.  Next, Dietrich sought to perform an asymmetric dihydroxylation to selectively obtain the desired diastereomer of DHIle. Sharpless dihydroxylation is the most widely used type of asymmetric dihydroxylation.76 This reaction involves the use of K2OsO2.(OH)4, K3Fe(CN)6, K2CO3, and quinine/quinidine based chiral ligands to induce facial selectivity for dihydroxylation. In fact, commercial mixtures of all these reagents are available under the names AD-mix-α and AD-mix-β, depending on the required facial selectivity. The dimer of dihydroquinidine (DHQD) is the ligand used in AD-mix-β, while the dimer of dihydroquinine 48  (DHQ), DHQD’s pseudoenantiomer, is the active ligand in AD-mix-α. These ligands are dimerized using a phthalazine-based linker (Phal). The enantioselectivity for Sharpless AD reactions usually exceeds 95%.77 To exploit this feature, following the analysis of the desired product and in hope of obtaining the required facial selectivity to produce the desired diastereomer of DHIle, Dietrich embarked on using an AD-mix as the dihydroxylating reagent. It was expected that utilizing AD-mix-β would result in diastereomers L-76 and D-77, while the other two diastereomers (L-77 and D-76) would be produced by the use of AD-mix-α. Hence, he selected AD-mix-β as opposed to AD-mix-α for this transformation.  However, to his surprise, he observed that the major product of this reaction was the undesired diastereomer of DHIle (77). To further investigate this outcome, Dietrich then performed the dihydroxylation using AD-mix-α and, in a separate attempt, under non-asymmetric Upjohn conditions. Figure 2.14 summarizes the results for these trials and the different products obtained.  Reagents Observed d.r. (76 : 77) AD-mix-α 1 : 10 AD-mix-β 1 : 33 OsO4/NMO 1.1 : 1 Figure 2.14 Four possible products and their observed diastereomeric ratios for dihydroxylation trials of D/L-75 using different reagents (the desired enantiomer is shown in blue).  49  Based on these results, each Sharpless condition, regardless of which commercial AD-mix was employed, led to formation of the undesired diastereomer (77) as the major product that was produced as a racemate contrary to what is expected from Sharpless AD conditions! However, the non-asymmetric Upjohn conditions, which were expected to exhibit no diastereoselectivity, afforded the desired diastereomer (76) in slight excess (76/77 1.1:1). It is noteworthy that, as was shown later (vide infra), no enantioselectivity was observed in these trials and equimolar amounts of L and D enantiomers were generated. Next, to compare the Upjohn and Sharpless AD results more accurately, Dietrich altered the Upjohn conditions by manual addition of the chiral ligands (DHAD)2Phal and (DHQ)2Phal to the reaction mixtures rather than using the commercially available pre-prepared reagents. However, these conditions resulted in the same overall lack of diastereoselectivity, hence excluding the effect of reagent preparation on the diastereomeric ratios. Additionally, different dihydroxylation conditions applied to substrates with a different amine protecting group on the starting material were tried. N-Boc-L-dehydroisoleucine indoline amide (D/L-78) was synthesized analogously to 75 and every dihydroxylation condition that had been examined initially was re-examined only to find that the diastereomeric ratios remained unchanged: almost no selectivity was observed when using Upjohn conditions, and the undesired diastereomer (80) was the major product in the case of Sharpless conditions (Figure 2.15). 50   Substrate Conditions d.r. (76:77 or 79:80) D/L-75 AD-mix-α 1 : 10 D/L-75 AD-mix-β 1 : 33 D/L-75 OsO4/NMO 1.1 : 1 D/L-75 OsO4/NMO/(DHQD)2Phal 1 : 3.1 D/L-75 OsO4/NMO/(DHQ)2Phal 1 : 2.6 D/L-78 OsO4/NMO/(DHQD)2Phal 1 : 2.7 Figure 2.15 Different dihydroxylation trials on Cbz (75) and Boc (78) protected L-dehydroisoleucine using different reagents and the diastereomeric ratios of the obtained products (desired diastereomers are shown in blue).  Finally, inspired by Gardiner and Bruce,78 Dietrich attempted using chiral ligands DHQD and DHQ with a different linker other than phthalazine. In their work, Gardiner and Bruce had observed a similar unexpected trend for the diastereoselectivity of the dihydroxylation reaction on their substrate. It was claimed that the surprising diastereospecificity, as opposed to enantioselectivity, was a result of remote binding of the substrate with DHQD and DHQ. Hence, they replaced the phthalazine linker with a pyrimidine-based linker (Pyr) to obtain (DHQD)2Pyr (81) and (DHQ)2Pyr to disrupt the said binding. Accordingly, Dietrich then synthesized (DHQD)2Pyr (81) and attempted Sharpless conditions with this chiral ligand introduced in place of (DHQD)2Phal. To verify the stereochemical outcome, he removed the Boc or Cbz groups and measured the diastereoselectivity of the reaction using reversed-phase 51  HPLC. To his disappointment, at best only modest diastereoselectivity in favor of the undesired diastereomer was observed when using the initial phthalazine-linked chiral ligands. These results are summarized in Figure 2.16.  Protecting Group, R AD Chiral Ligand d.r. (83 : 82) Boc (DHQD)2Pyr 1.1 : 1 Boc (DHQD)2Phal 3 : 1 Cbz (DHQD)2Pyr 1 : 1.2 Cbz (DHQD)2Phal 5.2 : 1 Figure 2.16 Diastereomeric ratios of the obtained products 82 and 83 from dihydroxylation reactions of 76/77 and 79/80 using (DHQD)2Phal or (DHQD)2Pyr chiral ligands.  It was apparent that altering the linker of the chiral ligands had no significant effect on the diastereoselectivity of the reaction as the undesired diastereomer still comprised the majority of the product mixture. However, assessed by capillary electrophoresis incorporating α-cyclodextrin in the mobile phase, Dietrich found that substituting the Phal-linker with a Pyr-linker resulted in an enantiomeric ratio of 2.3:1 and 2.6:1 in favor of the desired enantiomer L-82 when Boc and Cbz were used in the previous step, respectively (Figure 2.17).     52   Protecting group (for precursor of 82) AD Ligand e.r. (L-82:D-82) Boc (DHQD)2Pyr 2.3 : 1 Boc (DHQD)2Phal 1 : 1 Cbz (DHQD)2Pyr 2.6 : 1 Cbz (DHQD)2Phal 1 : 1 Figure 2.17 Analysis of the enantiomeric ratio for AD reactions using chiral ligands with phthalazine- (Phal) or pyrimidine-based (Pyr) linkers.  To conclude, the observed diastereoselectivity/enantioselectivity for dihydroxylation reactions fell short of expectations, whether employing standard Upjohn conditions or asymmetric ligands. With certain substrates and/or under certain conditions, it seemed that the selectivity obtained using chiral ligands greatly favored the undesired diastereomer. Moreover, employing OsO4 and NMO in the absence of chiral ligands resulted in almost no preference for any of the diastereomers. Due to the poor diastereo- and enantioselectivity observed with this approach, along with an inability to devise a practical means of separating a desired enantiomer from the desired pair of diastereomers, and a lack of a suitable protecting group strategy to synthesize a solid-phase compatible DHIle, Dietrich was not successful at incorporating his synthetic DHIle into amanitin.  2.1.3.3.2 Claisen Rearrangement and Upjohn Dihydroxylation Based on the results produced by Dietrich earlier, Zhao from Perrin group developed a simpler route to four diastereomers of DHIle, yet these were found to be inseparable by simple silica gel chromatography. Thus, the their mixture was incorporated into the amanitin core 53  heptapeptide to give four diastereomers from which the active diastereomer of the toxin was separated by HPLC.25 A summary of this method is shown in Figure 2.18.  Figure 2.18 Synthetic pathway to obtain four diastereomers of DHIle, D/L-87 and D/L-88 (Zhao and Perrin). The desired diastereomer is shown in the box.  The first step of this synthesis involved an ester-enolate Claisen rearrangement of 73 to produce dehydroisoleucine 74 as a racemate with greater than 9:1 diastereoselectivity in favor of the correct diastereomer. Following an Upjohn dihydroxylation using OsO4/NMO and TBS protection of the resulting hydroxyls, four diastereomers of DHIle were formed (D/L-85 and D/L-86). Swapping Cbz with Fmoc led to the formation of a suitably protected version of DHIle that was compatible with solid-phase synthesis. While a solid-phase synthesis of α-amanitin will be described in the upcoming chapters, at this point it suffices to state that the synthesis by Zhao et al. provided for four diastereomers followed by their incorporation into an analogue of α-amanitin. Subsequently, the active analogue was separated from the other three by HPLC (Figure 2.19). Although other methods of chiral resolution are conceivable e.g. selective recrystallization, chiral-HPLC, enzymatic, etc., these were neither investigated nor attempted.  54  It is notable that the failure to effectively separate any of the diastereomers (i.e. 85 from 86 or 87 from 88) discouraged us from pursuing a means of separating a suitable monomer.  In addition, without concrete knowledge as to whether the stereochemical make-up of DHIle was truly important for activity, all four diastereomeric amanitin analogs were synthesized and then these diastereomers were separated by HPLC, albeit with difficulty.  Whereas little if any diastereo- or enantioselectivity was induced in the dihydroxylation step, to obtain the cytotoxic diastereomer of amanitin, approximately 75% of the final octabicyclic peptides produced had to be discarded to isolate the remaining portion of the synthesized peptide.   Figure 2.19 Incorporation of a mixture of four DHIle diastereomers into an amanitin analogue and HPLC separation of the cytotoxic diastereomer containing (2S,3R,4R)-DHIle.  Needless to say, this preparation was uneconomical and synthetically ineffective. Nevertheless, this method achieved a protecting group strategy that was compatible with the solid-phase synthesis of a bioactive amanitin analogue and provided strong evidence that a single diastereomer of DHIle was responsible for bioactivity.  55  2.2 Results and Discussion Since all of the aforementioned syntheses were incapable of providing an efficient route to an enantiomerically pure (2S,3R,4R)-dihydroxyisoleucine, I embarked on developing and examining different strategies for an enantioselective synthesis of DHIle. The following discussion includes thought exercises, failed attempts and a detailed account of the academic progression that ultimately led to the identification of a method that exclusively afforded the desired diastereomer of DHIle in high enantiomeric purity.  2.2.1 Failed Efforts Towards (2S,3R,4R)-Dihydroxyisoleucine While I eventually designed a synthetic pathway that resulted in production of the correct diastereomer and enantiomer of DHIle required for the synthesis of α-amanitin, several other strategies were devised and experimented, all failing to yield the final product with the desired stereochemistry. Among these methods, some were carried through to the ultimate steps, while the others were abandoned due to the obstacles encountered along the way. Herein, I describe these attempts.  2.2.1.1 Method A: Brown Crotylation Followed by Dihydroxylation and Azide Displacement Nearly all previous syntheses of DHIle (other than the synthesis by Wieland) employed glycine as the starting material, which was followed by installation of the side chain and dihydroxylation of the resulting olefin via different strategies. While those methods offered valuable insight into the chemistry of DHIle, they were all unsuccessful at affording the desired diastereomer of DHIle as the major product. Hence, I formulated an entirely different approach 56  and decided to begin the synthetic pathway with the synthesis of the amino acid side chain (Figure 2.20). Owing to the specific arrangement of the three stereogenic centers on adjacent carbons, one may notice an internal pseudo-symmetry in DHIle or its likely precursors. I surmised that to obtain the desired enantiomer of DHIle, this pseudo-symmetry had to be broken via employing one or more as-of-yet to be established asymmetric transformation.  Figure 2.20 General retro-synthetic scheme for the synthesis of DHIle starting with the side chain diol.  Brown crotylation provides an efficient way of installing two stereogenic centers in a single transformation. This would ultimately provide for the β-methyl and the γ-hydroxyl present in DHIle with the correct absolute stereochemical outcome. This reaction could most likely supply the required means to disrupt the said pseudo-symmetry. A brief summary of this valuable and practical reaction follows.  2.2.1.1.1 Brown Crotylation In 1986, H. C. Brown and K. Bhat proposed a diastereo- and enantioselective method for the synthesis of erythro- and threo-β-methylhomoallyl alcohols using enantiomeric (Z)- and (E)-crotylboranes.79 First, a chiral methoxydiisopinocampheyl borane reagent was prepared from (+)- or (-)-α-pinene (89). To do so, α-pinene was reacted with borane-dimethylsulfide complex to obtain an enantio-enriched diisopinocampheyl borane species (90) (Figure 2.21).80 Then, treating this intermediate with methanol resulted in the formation of the active ligand B-57  methoxydiisopinocampheyl borane (91).81 Today, (+)- and (-)-91 are commercially available, reflecting the practicality and accessibility of Brown’s method.  Figure 2.21 Synthesis of (+)-B-methoxydiisopinocampheyl borane, (+)-91, by Brown.80,81  To begin the crotylation reaction, (Z)- or (E)-butene was reacted with nBuLi in the presence of KOtBu to obtain a metalated butene anion (92). The methoxydiisopinocampheylborane, (Ipc)2BOMe (91), was then added to the butene anion, forming the boron complex 93. Adding trifluoroborate-diethyl ether complex (BF3.OEt2) to the mixture removed the methoxy from 93, yielding the active chiral borane (94). At this stage, the aldehydic substrate could be added to the reaction mixture followed by addition of H2O2/NaOH to afford the final product (95) in excellent diastereo- and enantioselectivities, with de’s and ee’s often approaching 100% (Figure 2.22).  Figure 2.22 Synthesis of erythro-95a using (E)-butene and (+)-(Ipc)2BOMe (Brown and Bhat).79  58  What renders this crotylation method valuable is the ability to obtain any desired enantiomer out of the four possible stereo-isomers of β-methylhomoallyl alcohols. Starting with (E)-butene, one may obtain the erythro diastereomer, while using (Z)-butene will produce the threo isomer as the major product. The enantioselectivity of the reaction is controlled by the diisopinocampheyl borane reagent: employing (+)-(Ipc)2BOMe will yield a single diastereomer in high enantiomeric purity (a), while (-)-(Ipc)2BOMe will form the other enantiomer of the product (b) (Figure 2.23).  Figure 2.23 Brown crotylation and the ability to obtain any of the four possible diastereomers of β-methyl homoallyl alcohols as the major product using (E)- or (Z)-butene with (+)- or (-)-(Ipc)2BOMe.  Brown suggested that a chair-like transition state was responsible for the stereoselectivity of the asymmetric crotylation reaction. The suggested mechanism involved an initial complexation of the carbonyl oxygen of the aldehyde with boron, followed by transfer of the crotyl group to the carbonyl via a six-membered transition state. Out of eight possible transition states, only four of them could exist due to the steric interactions in the chair-like complex. The relative orientation of the butene and the isopinocampheyl groups would govern the 59  energetics of each of these transition states, leading to the production of a single product in high enantiopurity (Figure 2.24).  Figure 2.24 Chair-like transition states leading to four different diastereomers of the Brown crotylation.79  2.2.1.1.2 Method A: Synthesis Exploiting the fact that Brown crotylation allows for the installation of two adjacent stereogenic centers (β- and γ-carbons of the final DHIle product) simultaneously, I started the synthesis with 2-(benzyloxy)acetaldehyde (96) followed by Brown crotylation using (E)-butene and (+)-(Ipc)2BOMe. Formation of isopinocampheyl alcohol, the by-product of Brown crotylation, and the co-elution of it with the product of the reaction made the purification tedious, obliging me to use the β-methylhomoallyl alcohol product (97) as a crude mixture in the next step. The resulting crude alcohol was protected with tert-butyldimethylsilyl triflate (TBDMSOTf) and 2,6-lutidine to yield olefin 98. Protected ene-diol 98 was then dihydroxylated using Upjohn conditions (OsO4, NMO) to yield the protected tetrol intermediate 99. Although 99 was afforded as a diastereomeric mixture (as observed in its 13C-NMR), the exact ratio of these two diastereomers could not be determined due to overlapping of the diastereomeric peaks in the 1H-NMR and complexity of the splitting of the said peaks. However, it appeared that one diastereomer was greatly predominant, in an approximate 6:1 ratio. At this stage, I only aimed to test the feasibility of this strategy, and the absolute configuration of the newly introduced stereogenic centers of the major diastereomer was not 60  determined. Hence, without further assessment, the primary alcohol was selectively protected using TBDMSCl, imidazole and catalytic amounts of DMAP to yield intermediate 100 (Figure 2.25).  Figure 2.25 Synthesis of a diastereomeric mixture (d.r. 6:1 in favor of an unknown diastereomer) of the protected tetrol intermediate 99.  Efforts towards conversion of the unprotected secondary alcohol of 100 to the corresponding azide under aza-Mitsunobu conditions failed, likely due to the steric hindrance caused by adjacent OTBS groups. Thus, the secondary alcohol was mesylated using MsCl and triethylamine, followed by a sodium azide SN2 displacement to yield compound 102. Staudinger reduction of 102 using trimethylphosphine followed by Fmoc protection with Fmoc-OSu afforded intermediate 104. The primary TBS-protected alcohol was selectively deprotected with camphorsulfonic acid (CSA) and was oxidized using ruthenium(III) chloride and sodium metaperiodate to afford the final protected DHIle (106) (Figure 2.26). 61   Figure 2.26 Synthesis of a stereoisomer of DHIle (106) from protected tetrol 100.  Although individual yields were suboptimal and might have been further improved, at this point I demonstrated that at least a single diastereomer of DHIle could indeed be synthesized using this strategy. Next, I embarked on determining the relative configuration of the stereogenic carbons. Presuming that the Brown crotylation yielded the desired enantiomer of 97, if I could prove that α- and γ-carbons were syn relative to each other, the synthesis would be concluded, and it could be inferred that the desired enantiomer of DHIle had been obtained. Examining intermediate 100, a 1,3-diol system was recognized. Exploring the literature led me to an elegant 13C-NMR-based method developed by Rychnovsly that could be exploited to assign the relative configuration of 1,3-diols with reasonable certainty.  2.2.1.1.3 Rychnovsky’s Method to Determine the Relative Stereochemistry of 1,3-Diols In a work published in 1993, Rychnovsky et al. reported a strategy to establish the relative configuration of 1,3-diols based on the 13C-NMR  correlations of their corresponding acetonides.82 In this work, the 13C-NMR spectra of the synthesized acetonides were acquired and two distinguishable chemical shifts were investigated: the C(2)-methyl groups and the 62  C(2)-acetal. In the syn-acetonide resulting from a syn-1,3-diol, one methyl existed in the axial position with a chemical shift of ca. 19 ppm, while the other was equatorial at ca. 30 ppm. In the case of an anti-acetonide, prepared from an anti-1,3-diol, due to the steric hindrance present in the chair conformations, both methyl groups were observed in a twist-boat position, hence both having a chemical shift of ca. 25 ppm (Figure 2.27).  Figure 2.27 Major conformations of A) syn-1,3-diol acetonide and B) anti-1,3-diol acetonide. The numbering of acetonide carbons is shown.  They noticed that anti-1,3-diol acetonides could exist in different conformations based on the type of the substituents on carbons 4 to 6. To explore a wider range of structural types containing an anti-1,3-diol system, they divided their studied acetonides into six different classes. Classes 1 to 3 included anti-1,3-acetonides with an unsubstituted C-5, while anti-1,3-acetonides with a methyl substituent on C-5 comprised classes 4 to 6. The C-6 substituent in all these classes was an alkyl group with a sp3 hybridization, while the C-4 alkyl substituents were sp3-hybridized in classes 1 and 4, sp2 in classes 2 and 5, and sp in classes 3 and 6. 13C-NMR spectra were then acquired for a variety of examples in each class and a table was assembled, representing the average 13C chemical shifts for methyl groups and the acetal 63  carbon correlating to individual classes (Table 2.1). This method of determining the relative configuration of 1,3-diol systems appeared very facile and, in our case, potentially applicable owing to the presence of a 1,3-diol structure in intermediate 99.  Structural class Low methyl High methyl Methyl difference Acetal Syn 19.66 ± 0.35 30.00 ± 0.30 10.34 ± 0.30 98.93 ± 0.67 anti (all) 23.65 ± 0.93 25.64 ± 1.79 1.98 ± 2.60 100.64 ± 0.82 anti class 1 24.49 ± 0.26 24.69 ± 0.25 0.20 ± 0.20 100.31 ± 0.49 anti class 2 24.22 ± 0.66 26.14 ± 0.84 1.91 ± 1.43 100.39 ± 0.35 anti class 3 21.97 ± 0.66 29.41 ± 0.48 7.44 ± 1.05 100.87 ± 0.56 anti class 4 23.59 ± 0.32 24.92 ± 0.54 1.33 ± 0.56 100.59 ± 1.15 anti class 5 23.80 ± 0.26 25.12 ± 0.23 1.31 ± 0.35 100.78 ± 0.25 anti class 6 21.93 ± 0.11 29.10 ± 0.62 7.17 ± 0.51 101.33 ± 0.21 Table 2.1 Different classes of 1,3-diol acetonides and the list of  respective 13C-NMR chemical shifts for methyl groups and the acetal C(2). Values are in ppm as the mean ± std deviation (from Rychnovsky et al).82  2.2.1.1.4 Assigning the Relative Configuration of the 1,3-Diol in Method A To synthesize the 1,3-actonide required for the Rychnovsky’s assessment, first I deprotected the TBS group on the diastereomeric mixture of intermediate 99 to afford 107. Then, selective TBS protection of the primary alcohol of tetrol 107 yielded the mono-TBS protected compound 108, which was converted to the corresponding acetonide (109) using 2,2-dimethoxypropane and camphorsulfonic acid (Figure 2.28). It is noteworthy that the major stereoisomer of acetonide 109, resulting from the predominant diastereomer of tetrol 99, was isolated at this point. 64   Figure 2.28 Synthesis of acetonide 109 following the isolation of its major diastereomer. The proposed synthesis of the correct enantiomer of DHIle from the desired trans-1,3-diol is shown in the dashed box.  According to Rychnovsky’s classification, 109 could fall into either syn-1,3- or class 4 of anti-1,3-diol acetonides. Ideally, acetonide 109 was expected to be in class 4 of anti diol acetonides. Once the mesylate of intermediate 101 was displaced by sodium azide in an SN2 manner to afford 102, its configuration would be inverted to eventually obtain the desired syn-1-OH-3-NH2 (Figure 2.28, shown in the box). Unfortunately, 13C-NMR spectrum of 109 revealed that the major isolated diastereomer of the diol possessed a syn relative configuration. The chemical shift for the acetal carbon was 98.70 ppm, the low methyl had a chemical shift of 20.08 ppm, and that of the high methyl was 30.59 ppm. All these values perfectly accorded with the syn class of 1,3-diol acetonides (Table 2.2).  65   Structure Low methyl (ppm) High methyl (ppm) Methyl difference (ppm) Acetal (ppm) 109 20.08 30.59 10.51 98.70 Syn 19.66 ± 0.35 30.00 ± 0.30 10.34 ± 0.30 98.93 ± 0.67 anti class 4 23.59 ± 0.32 24.92 ± 0.54 1.33 ± 0.56 100.59 ± 1.15 Table 2.2 13C-NMR chemical shifts for 109, syn- and anti-class 4 acetonides. Representative chemical shifts for 109 matched the syn structural type proposed by Rychnovsky.   This selectivity to form the 1,3-syn diol was unexpected. To produce the syn-1,3-diol, dihydroxylation of 98 had to occur on the re-face of the olefin, while the desired anti-1,3-diol would be obtained via dihydroxylation on the si-face (Figure 2.29). The selectivity for oxidation on the re-face might be explained by the remote chelation of osmium to the oxygen of the OTBS group. However, the TBS group could also impose steric hindrance, possibly directing the dihydroxylation to take place from the si-face. At this point, this diastereoselectivity could not be justified and further experiments were not performed to examine different hypotheses.  Figure 2.29 Dihydroxylation from re- and si-face and consequent isomers of DHIle expected in each case.  66  2.2.1.1.5 Method A: Conclusion Although I was able to synthesize a single diastereomer of DHIle using this method, the undesired diastereomer was produced as the major stereoisomer, namely that of a D-amino acid wherein the stereochemical configuration at the α-carbon had been confirmed by the correlation between the hydroxyls in the 1,3-diol system of intermediate 99. While the Upjohn dihydroxylation of 98 appeared to have produced some of the desired diastereomer (si-99) in the diastereomeric mixture of 99, NMR correlation studies suggested that it was the minor diastereomer. Furthermore, despite the fact that the relative configuration of the α- and γ-carbons could be assigned, our inability to determine the absolute configuration of the stereogenic centers proved to be a minor yet problematic obstacle in the way of confirming the legitimacy of the final DHIle product. Additionally, since the benzyl protecting group in 106 needed to be replaced with another TBS group to afford a solid-phase compatible amino acid, the total number of steps for this synthetic route would be increased further eroding its attractiveness as a method. Removal of the benzyl group could also lead to the formation of the δ-lactone of DHIle, requiring saponification that could lead to a decrease in the overall yield. Hence, I decided to abandon this method and pursue other strategies to stereoselectively synthesize (2S,3R,4R)-DHIle.  2.2.1.2 Method B: Mannich-Type Reaction Followed by Rubottom Oxidation The Mannich reaction is a classic and instrumental method for the synthesis of β-amino carbonyl compounds and one of many ways to form C-C bonds and elongate carbon chains (Figure 2.30). 67   Figure 2.30 General scheme for the Mannich reaction involving three components: a carbonyl donor, an amine, and an acceptor aldehyde.  As the interest in stereoselective synthesis of natural products has grown over the past 5 decades, numerous direct catalytic Mannich reactions have been developed.83 These reactions are usually catalyzed by organometallic complexes or metal-free organocatalysts and can provide an efficient pathway for the synthesis of α- and β-amino acid derivatives among many other possible targets. One of the most frequently used organo-catalysts for the asymmetric Mannich reaction is L-proline.84 In 2002, Cordova and co-workers reported the proline-catalyzed Mannich-type reaction of α-imino glyoxylates as an effective and enantioselective route to prepare α-amino acids.85 In this method, N-PMP protected α-iminoethyl glyoxylate was reacted with several different ketone donors employing L-proline as the catalyst, and the diastereo- and enantiopurity of the resulting α-amino ethyl esters were assessed (Figure 2.31).  Figure 2.31 Proline-catalyzed Mannich-type reaction for the synthesis of functionalized α-amino acids. A) General scheme. B) Mechanism.85  68  This method has previously been used in Perrin lab for the synthesis of (2S,3R,4R)-monohydroxyisoleucine (MHIle, 114) for incorporation into an analogue of α-amanitin (Zhao and Perrin, unpublished results). Synthesis of MHIle using this method began with the Mannich-type reaction between 2-butanone and the PMP-protected imine of glycine ethyl ester (111) using L-proline as the organo-catalyst. The resulting γ-keto L-allo-isoleucine (112) was subsequently epimerized to γ-keto-L-isoleucine (113) by catalytic base treatment and then reduced by sodium cyanoborohydride to yield the desired enantiomer of MHIle (114) (Figure 2.32).  Figure 2.32 Synthesis of MHIle via a proline-catalyzed Mannich-type reaction (Zhao and Perrin, unpublished).  Inspired by this synthesis, I decided to exploit intermediate 113 as a precursor to DHIle. To install a hydroxyl group at the δ-position, the silyl-enol ether of 113 was required, which could then undergo a Rubottom oxidation using mCPBA or OsO4. Several conditions were tested to achieve the silyl enol ether intermediate: LDA, triethylamine or 2,6-lutidine were tried as the base, while TBDMSCl, TMSCl or TBDMSOTf were the attempted silyl chlorides (Figure 2.33). The majority of these conditions did not yield the desired silyl enol ether and the starting material was recovered. However, using LDA with TMSCl furnished a mixture of 69  the desired mono-TMS product (116) and a bis-TMS compound (117) in a 3:2 ratio with the full consumption of starting material.  Figure 2.33 Trials to produce the silyl enol ether of 113 using different bases and silyl chlorides. LDA and TMSCl yielded a 3:2 mixture of 117 and 116.  Due to the well-known instability of silyl enol ethers, the purification of 116 and 117 was not feasible and thus the mixture of 116 and 117 was used in trial oxidations. To my disappointment, none of the tested oxidation conditions yielded the desired product (118), and the starting ketone (113) was recovered in all cases (Figure 2.34).  Figure 2.34 Oxidation trials of 116/117 for the synthesis of 118. In all cases the starting ketone (113) was recovered.  Next, efforts were made to brominate the alkene of 116. Once installed, I hoped that the bromide on Cδ could be converted to δ-OH at a later stage via a displacement reaction. Trimethyphenyl ammonium tribromide (Me3PhN+Br3-) formed small amounts of various products containing different numbers of bromine atoms, including over-bromination of the phenyl-ring of the PMP protecting group, while most of the starting ketone (113) was recovered. In search for a stronger brominating reagent, next I tried Br2. Employing precisely 70  one equivalent of Br2 at -78°C led to the formation of minor amounts of the desired mono-brominated product and a bis-brominated compound, while almost 70% of the starting ketone was recovered. NMR studies showed that the bis-brominated product contained a bromine on the phenyl ring of the PMP protecting group (Figure 2.35).  Figure 2.35 Efforts towards bromination of 116/117. A) Trimethylphenylammonium tribromide resulted in minor amounts of products containing various numbers of bromine. Almost all of the starting ketone was recovered. B) Br2 led to the formation of the desired product (119) (10%) and the bis-bromo compound 120 (20%), and 70% of the starting ketone was recovered.  While the inability to oxidize or brominate 116 to give the desired products (118 or 119) was not further investigated, it stands to reason that this method did not provide either the efficiency or the overall yield required for a stereoselective synthesis of DHIle. Furthermore, it was noted that preparation of the silyl enol ether intermediate was neither straightforward nor high yielding, whereas the Rubottom and Upjohn oxidation of olefin 116 failed to produce the desired products. Finally, the bromination reactions were low-yielding and resulted in a mixture of multi-brominated products making separation and characterization difficult. Hence, this strategy was also abandoned in favor of a more efficient one.  71  2.2.1.3 Method C: Claisen Rearrangement Followed by Enzymatic Resolution Inspired by the previous syntheses of DHIle employing a Claisen rearrangement to obtain L-dehydroisoleucine (74), I sought a similar strategy to synthesize an analogous dehydroisoleucine intermediate. One drawback associated with the Claisen rearrangement of the cis-crotyl ester of glycine has been the formation of a racemate even though it exhibits high diastereoselectivity. Although the diastereoselectivity of the Claisen reaction was decent, isolation of the desired enantiomer had proven to be problematic, and a practical way to separate the enantiomers present in the racemate of the desired diastereomer had to be established. Exploring the literature led me to a report in which a method was designed for the kinetic resolution of N-protected amino acid esters using a commercially available alcalase enzyme prepared from a strain of Bacillus Licheniformis.86 To wit, Wang et al. selectively hydrolyzed the methyl ester of N-Cbz-L-Nle-OMe and N-Cbz-L-Nva-OMe, while the D isomers remained intact (Figure 2.36).  Figure 2.36 Selective hydrolysis of N-Cbz-L-Nle-OMe and N-Cbz-L-Nva-OMe using alcalase from Bacillus Licheniformis. Reaction conditions: alcalase (2.5 AU/mL), tBuOH/H2O 19:1, pH~8.2, 35°C.86  Although enzymes are generally very specific towards their substrates, the apparent resemblance between dehydroisoleucine, norleucine (Nle) and norvaline (Nva) convinced me to consider using this method for the separation of the L-enantiomer of the erythro-72  dehydroisoleucine from its D-enantiomer.  To this end, a mixture of four diastereomers of 74 was reacted with SOCl2 in MeOH to afford the corresponding methyl esters (121). Then, this mixture was subjected to enzymatic conditions and reaction progress was followed by TLC. Disappointingly, no hydrolysis was observed for any of the diastereomers, and both the L- and D-enantiomers remained unreacted (Figure 2.37).  Figure 2.37 Efforts toward the enzymatic hydrolysis of erythro/threo-L-dehydroisoleucine methyl esters using the alcalase from Bacillus Licheniformis. The initial mixture of diastereomers was recovered with no hydrolysis observed. The desired enantiomer is shown in blue.  While this pathway initially seemed promising, the challenge of isolating the desired enantiomer of dehydroisoleucine remained unsolved. The alcalase enzyme proved to be inactive towards dehydroisoleucine, which might be attributed to the presence of the β-methyl substituent of this amino acid compared to β-unsubstituted Nle and Nva residues. Furthermore, even with the correct enantiomer of 74 in hand, the asymmetric dihydroxylation of the olefin, previously investigated by Dietrich, along with the resolution of the resulting pair of diastereomers would remain a challenge to be overcome. Thus, this scheme was also abandoned in hope of finding one that would lead to the desired enantiomer of DHIle.  73  2.2.2 Method D: Successful Synthesis of (2S,3R,4R)-Dihydroxyisoleucine: Brown Crotylation Followed by Asymmetric Strecker Gaining valuable information from the previous non-stereoselective syntheses and all my failed efforts towards the synthesis of DHIle, it seemed that method A, with some adjustments, could possibly provide a stereoselective pathway to the synthesis of this small yet challenging unnatural amino acid. A general retrosynthetic scheme of the pathway designed for method D is shown in Figure 2.38.  Figure 2.38 Retrosynthetic scheme for the synthesis of (2S,3R,4R)-DHIle showing the bond formation sites required for each major step (method D). R1, R2 = protecting group.  The first part of the synthesis was carried out according to method A: 2-(benzyloxy)-acetaldehyde was employed in a Brown crotylation reaction using (E)-butene and (+)-(Ipc)2BOMe to yield the olefinic intermediate 97. Following the protection of the secondary hydroxyl with TBDMSCl, Upjohn dihydroxylation of 98 using OsO4/NMO furnished intermediate 99 (see Figure 2.25). Prior to performing the oxidative cleavage of 99 to the corresponding aldehyde, I investigated the literature to select an asymmetric Strecker reaction that would suit my needs. Herein, I briefly discuss different classes of asymmetric Strecker reactions.  2.2.2.1 Asymmetric Strecker Reactions: A Brief Overview The Strecker reaction, first introduced by Adolph Strecker in 1850, is a three-component reaction that has found extensive application in natural product synthesis (Figure 2.39).87–91 In 74  this reaction, an amine (or ammonium salt), an aldehyde and a cyanide or cyanide equivalent are mixed in one pot under mildly acidic conditions and an amino nitrile is obtained as the product. Acid hydrolysis of the nitrile produces an α-amino acid, rendering this reaction a valuable tool for the synthesis of natural and unnatural α-amino acids.  Figure 2.39 Main components of the Strecker reaction. Acid hydrolysis of the resulting α-amino nitrile affords an α-amino acid.  The demand for enantiomerically pure amino acids has led to development of various types of asymmetric Strecker reactions. These asymmetric versions are generally based on two main strategies: 1) nucleophilic addition of cyanide to chiral iminiums, and 2) enantioselective catalytic cyanation of achiral iminiums. In the case of DHIle, since the starting aldehyde and its corresponding imine are chiral, we will focus on the former strategy. Nevertheless, as will become apparent upon appreciation of some of the stereochemical constraints with this particular target, a brief discussion of an enantioselective Strecker with achiral aldehyde substrates is warranted; in this case, the Strecker reaction is usually carried out with either organocatalysts or chiral ligands chelated to metal catalysts. The outcome of a Strecker reaction employing the nucleophilic addition of cyanide to a chiral imine greatly hinges on sterics and facial selectivity of the cyanide addition (Figure 2.40).    75   Figure 2.40 The Felkin-Ahn model for the nucleophilic addition of cyanide to an iminium in the Strecker reaction. In this example, L-amino nitrile is the favored product (shown in blue) (S=small, M=medium, L=large).  Drawing a Newman projection of the starting imine and taking into account the Bürgi-Dunitz angle for the nucleophilic attack of cyanide according to the Felkin-Ahn model, one can predict the stereochemical outcome of this transformation.92 As shown in Figure 2.40, the stereochemical influence of the stereogenic carbon adjacent to the imine nitrogen and the size of its substituents have a great impact on the stereoselectivity of the reaction. If the cyanide addition through the favored site of nucleophilic attack leads to formation of the desired enantiomer, the Strecker reaction is “matched”. Similarly, if the desired enantiomer needs to be obtained via the addition of cyanide from the unfavored face in the Felkin-Ahn model, the reaction is considered to be a “mismatched” case of Strecker. In the case of mismatched Strecker reactions, the enantioselectivity is generally not excellent due to the fact that cyanide is “forced” to attack from the unfavored face to yield the desired product. The most widely utilized strategies to force the cyanide addition to occur from a predictable face of the imine involve the use of a chiral auxiliary. In this type of Strecker reaction, the imine is formed by addition of a chiral amine to the starting aldehyde, hence producing an 76  imine with an additional chiral substituent on the nitrogen that could direct the cyanide addition. One of the first and most common chiral auxiliaries is either (R)- or (S)-α-phenylethylamine (α-PEA).87 Employing α-PEA as the chiral auxiliary can lead to reversing the enantioselectivity of the Strecker reaction. If the desired nitrile must be obtained via a mismatched case of cyanide addition, installing α-PEA on the nitrogen of the imine can reverse the facial selectivity of the cyanide addition due to the steric hindrance imposed by the phenyl group of α-PEA. In fact, when employing α-PEA as the chiral auxiliary, substituents on the adjacent carbon are no longer the only directing groups, and the phenyl group of the auxiliary may play a major role in the facial selectivity of the cyanide addition. The steric hindrance between α-PEA and the incoming nucleophile can lead to formation of an otherwise unfavored enantiomer (Figure 2.41).  Figure 2.41 (S)-α-phenylethylamine as a chiral auxiliary for the asymmetric Strecker reaction. According to the Felkin-Ahn model, what once was the favored enantiomer of the amino nitrile product in the non-asymmetric reaction, is now the unfavored isomer due to the steric bulk added to the imine nitrogen by the phenyl group (S=small, M=medium, L=large, blue circle: nitrogen).  While other chiral auxiliary groups have been utilized in asymmetric Strecker reactions (e.g. α-amino acids, α-phenylglycinol, sulfinamide, glycosyl amine, hydrazine), they all follow the same principles and can regulate the stereo-outcome of the corresponding cyanide addition in a similar manner (Figure 2.42).93 77   Figure 2.42 Examples of other common chiral auxiliaries used in asymmetric Strecker reactions: A) (R)-phenylglycine,88 B) (R)-α-phenylglycinol,91 C) (S)- p-toluene sulfinamide,94 D) galactosylamine,89 and E) (S)-1-amino-2-methoxymethyl indoline.95  2.2.2.2 Method D: Results and Discussion In light of this literature review, I became confident that the desired diastereomer of an amino nitrile intermediate would be accessible via an asymmetric Strecker reaction using a suitable chiral auxiliary. During our investigations, I came upon a report by Potier and co-workers96 in which (2S,3R,4S)-4-hydroxyisoleucine was synthesized via an asymmetric Strecker reaction with (S)-α-phenylethylamine as the chiral auxiliary (Figure 2.43). Starting with 2-methylacetoacetate (122), Potier et al. used Geotrichum candidum to microbially reduce ketone 122 to ethyl (2S,3S)-3-hydroxy-2-methylbutanoate (123). THP protection of 123 with DHP and TsOH yielded 124, which was then converted to the corresponding aldehyde (126) following LAH reduction and TEMPO oxidation steps. Next, an asymmetric Strecker reaction was performed using (S)-α-PEA and KCN to furnish amino nitrile 127 (d.r. 4:1). This 78  intermediate was hydrolyzed with HCl to afford a pair diastereomeric lactone products (128). At this point, recrystallization from iPrOH enabled the isolation of the desired (3S)-128 enantiomer. Finally, removing the chiral auxiliary with H2/Pd and saponification of the free lactone yielded a single enantiomer of the MHIle product (130).  Figure 2.43 Synthesis of (2S,3R,4S)-4-hydroxyisoleucine by Potier et al.96  Inspired by this work, I cleaved the diol of intermediate 99 with sodium meta-periodate in MeOH/H2O to obtain aldehyde 131 required for the asymmetric Strecker reaction. To perform the asymmetric Strecker reaction, use of the HCl salt of (S)-α-PEA was needed. Since the commercial (S)-α-PEA available to me was in the form of the free base, a 1 M aqueous solution of (S)-α-PEA.HCl was prepared by dissolving α-PEA in a known amount of H2O and titrating the solution to pH~7 with concentrated HCl. Then, 131 was treated with α-PEA.HCl (1 M aq. solution) and KCN in MeOH/H2O to furnish a mixture of (1S)- and (1R)-132 in a 1.2:1 ratio (Figure 2.44, A). Although this reaction produced the desired diastereomer, the diastereoselectivity was modest at best. This can be readily explained by the fact that this transformation was a mismatched case of the Strecker reaction, and to produce the desired 79  diastereomer, the facial selectivity for which had to be directed using a chiral auxiliary. Otherwise the undesired diastereomer would have predominated (Figure 2.44, B).  Figure 2.44 A) Synthesis of amino nitriles (1S)- and (1R)-132. B) The Felkin-Ahn model for the non-asymmetric Strecker reaction of an imine of 131, showing the mismatched reaction. In a non-asymmetric Strecker, the favored facial selectivity affords the undesired diastereomer, while the unfavored selectivity leads to the desired diastereomer.  Even though the diastereoselectivity of the asymmetric Strecker reaction only slightly favored the desired diastereomer, the ability to readily separate (1S)- and (1R)-132 from each other via flash column chromatography provided a compelling feature of this synthetic route. Hence, (1S)-132 was isolated and subjected to 6N HCl under reflux conditions. Under these conditions, both TBS and Bn protecting groups were removed and the nitrile was hydrolyzed. However, the hydroxyl on Cγ was in the proper spatial position to form a lactone with the hydrolyzed nitrile, producing a γ-lactone (133) upon further acid hydrolysis (Figure 2.45). 80   Figure 2.45 Acid hydrolysis of (1S)-132 to lactone 133. Formation of the imidate and the deprotection of TBS and Bn took place during this reaction.  At this stage, the relative configuration of α-, β- and γ-carbons were determined by NOESY 1H-NMR experiments. The acquired NOESY spectrum showed through-space correlations between the following pairs of substituents on the lactone: (γ-proton and β-methyl) and (α-proton and β-proton), hence providing promising evidence that the stereogenic centers on 133 have the correct relative configuration: (α,β)-cis and (β,γ)-trans (Figure 2.46).  Figure 2.46 Selected region of 1H-NOESY NMR of 133 showing the correlations between cis substituents (400 MHz, CD2Cl2).  (a,b): (4.17, 1.52) (e,f): (4.16, 1.06) (c,d): (3.53, 2.24) 81  Although the NOESY spectrum likely confirmed the relative configuration of the substituents on the γ-lactone, the absolute configuration of the stereogenic centers had yet to be determined. Toward this end, the chiral auxiliary (α-PEA) was removed using H2 and 10% Pd/C at pH 4 in MeOH to yield the HCl salt of the free lactone (134). To my delight, this intermediate was readily recrystallized from MeOH as small needles, and its X-ray crystal structure was acquired (Figure 2.47). The XRD structure confirmed that our synthesis yielded the desired configuration on all stereogenic centers.  Figure 2.47 Removal of the chiral auxiliary to obtain the HCl salt 134. The XRD structure is shown in the box, confirming the configuration of the stereogenic centers to match the desired enantiomer.  Next, the unprotected lactone (134) was saponified using LiOH in H2O to furnish the fully deprotected free-chain DHIle (135) as a zwitterion. It is noteworthy that subsequent to this step, it was crucial that the pH be maintained at slightly alkaline values (pH ca. 8) during any reaction steps or work-up performed on the free-chain DHIle in order to prevent re-generation of the γ-lactone. Thus, following the saponification, the pH of the reaction contents was 82  adjusted to 8 with 0.1 M aq. HCl, and the resulting mixture was frozen and lyophilized to obtain crude 135 as a zwitterionic solid (Figure 2.48).  Figure 2.48 Saponification of 134 to the free-chain DHIle 135. In acidic pH, 135 converts back to the lactone 134, hence avoiding pH lower than 8 in the work-up.  In our synthesis of α-amanitin, an Fmoc-based solid-phase strategy would be utilized for incorporation of amino acids into the peptide (vide infra). Hence, to obtain a solid-phase compatible DHIle, TBDMS protection of the hydroxyls and Fmoc protection of the amine were required. Initial trials to protect the hydroxyls with TBS were performed using TBDMSCl and imidazole in DMF as the solvent. Whereas yields greater than 80% were not expected due to the proximity of the two hydroxyls and the steric hindrance surrounding the γ-OH, initial yields for this reaction did not exceed 10-15%. Perplexed by these extremely low yields, I noticed the formation of a by-product as the major product of this reaction (Figure 2.49). Analysis of this by-product revealed its identity as Nα-formyl-bis-TBS DHIle (137).   83   Figure 2.49 A) First trial for the TBDMS protection of the free-chain DHIle with TBDMSCl and imidazole in DMF. B) proposed mechanism for the formylation of the amine to yield by-product 137 as the major product.  I proposed that the presence of DMF as the solvent of the reaction was the source of the formyl group. Once activated by TBDMS, DMF could turn into an electrophile and become susceptible to a nucleophilic attack by the free amine. Upon work-up, the TBS tetrahedral adduct would then be hydrolyzed to produce the formylated bis-TBS DHIle as the major product (Figure 2.49, B). To resolve this problem, I selected N,N-dimethylacetamide (DMA) as the solvent rather than DMF. To my delight, no acetylation (by analogy to formylation) was observed and the desired bis-TBS DHIle (136) was obtained almost exclusively. Following Fmoc protection of the amine with Fmoc-OSu, a fully protected solid-phase compatible DHIle (138) was obtained, which was then converted to an activated NHS-ester (139) using disuccinimidyl carbonate (DSC) (Figure 2.50). These protecting group manipulations concluded our stereoselective synthesis of a solid-phase compatible DHIle. 84   Figure 2.50 Synthesis of the NHS-activated ester of Nα-Fmoc-bis-TBS-DHIle (139) from 135.  2.3 Conclusion In this chapter, I initially disclosed a brief introduction on (2S,3R,4R)-γ,δ-dihydroxyisoleucine and the synthetic challenges that required to be overcome en route to the first stereoselective synthesis of this amino acid. The most common strategies for the synthesis of β-branched amino acids were described: 1) Garner’s aldehyde, 2) sigmatropic rearrangement of the Overman’s imidate, and 3) Claisen rearrangement. Previous efforts towards the synthesis of DHIle were then discussed: 1) synthesis by Wieland, 2) Bartlett’s synthesis, and 3) syntheses by Perrin group (Claisen rearrangement followed by Sharpless or Upjohn dihydroxylation). In the results and discussion section, an overview of my failed attempts towards the synthesis of DHIle was provided: Brown crotylation followed by an azide displacement and oxidation to the carboxylic acid (method A), proline-catalyzed Mannich-type reaction followed by Rubottom or Upjohn oxidation (method B), and enzymatic resolution of allo- and L-dehydroisoleucine with an alcalase (method C). These strategies were incapable of providing the desirable results with the efficiency or stereoselectivity I had aimed for. Finally, the chapter was concluded with my successful stereoselective synthesis of DHIle (method D). I showed that the Brown crotylation is an extremely versatile and efficient way of 85  installing the β- and γ-stereogenic centers. Furthermore, an asymmetric Strecker reaction that had previously been utilized in the synthesis of monohydroxy isoleucine was employed to install the α-amine functionality. Following acid hydrolysis of the amino nitrile product of the Strecker reaction, a γ-lactone was formed, which was subsequently saponified to the free-chain DHIle. The 1H-NOESY experiment and the XRD structure acquired from the γ-lactone affirmed that all stereocenters possessed the desired configurations of (2S,3R,4R).  Eventually, TBS protection of the γ- and δ-hydroxyl groups, Fmoc protection of the amine and activation of the carboxylic acid via formation of an NHS-ester resulted in a fully protected solid-phase compatible version of DHIle, awaiting incorporation into α-amanitin (vide infra).  2.4 Experimental Section  2.4.1 Materials and Methods General: All reactions were performed under argon atmosphere in flame-dried glassware and dried solvents at room temperature, unless otherwise stated. Controlled temperature reactions were performed using a mineral oil bath and a temperature controlled hot plate (IKA Ceramag Midi). Temperatures below room temperature were achieved in an ice/water bath (0°C), dry ice/ethylene glycol bath (-20°C), dry ice/ethanol/ethylene glycol bath (-20°C to -75°C) and dry ice/acetone bath (-78°C). Solvents were removed under reduced pressure using a Büchi rotary evaporator. Anhydrous solvents were prepared by distillation under nitrogen atmosphere. Ethers were distilled from sodium in the presence of benzophenone as indicator. Triethylamine, dichloromethane and hexanes were distilled over calcium hydride. Methanol was distilled from magnesium. DMSO and DMF were dried over 4Å molecular sieves under 86  argon atmosphere. All reagents and solvents were purchased from Sigma-Aldrich, Alfa Aesar, Acros Organics, Matrix Scientifics, Oakwood Chemicals, Ontario Chemicals or TCI America, unless otherwise stated. Thin-layer chromatography (TLC) was performed using silica gel 60 F254 precoated aluminum plates (EM Science). Detection of TLC spots was performed using UV lamp at 254 nm, or by staining with p-anisaldehyde, potassium permanganate, ninhydrin or 2,4-dinitrophenylhydrazine, prepared according to literature procedures. Flash column chromatography purifications were performed using silica gel 60 (230-400 mesh, Silicycle, Quebec). Low-resolution mass spectra (LRMS ESI) in electrospray ionization (ESI) mode were obtained from a Bruker Esquire spectrometer. Proton (1H-NMR) and carbon (13C-NMR) spectra were obtained using Bruker AV-300 (300 MHz) and AV-400inv (400 MHz) spectrometers. The X-Ray crystallographic measurements were made on a Bruker APEX DUO diffractometer with a TRIUMPH curved-crystal monochromator with Mo-Kα radiation. Melting points were recorded on a MEL-TEMP capillary melting point apparatus and were corrected.  2.4.2 Experimental Procedures  (2R,3R)-1-(Benzyloxy)-3-methylpent-4-en-2-ol (97). E-2-butene (4.57 mL at -78°C, 46.1 mmol, excess) was added through a cannula to a stirred suspension of KOtBu (2.17 g, 18.3 mmol) in THF (41.5 mL) at -78°C, followed by the dropwise addition of n-BuLi (1.39 M in hexanes, 13.2 mL, 18.3 mmol). The reaction 87  suspension was slowly warmed to -45°C, stirred at this temperature for 15 minutes, and then cooled to -78°C. A solution of (+)-(Ipc)2BOMe (6.72 g, 18.9 mmol) in diethylether (21 mL) was then added over a 15-minute period. Stirring was continued for 10 minutes and BF3.OEt2 (2.24 mL, 18.9 mmol) was added, followed by addition of a pre-cooled solution of 2-(benzyloxy)-acetaldehyde 9 (3.07 g, 20.3 mmol) in Et2O (16 mL). After completion of the reaction (4 hours, checked by TLC), the reaction mixture was warmed to 0°C and quenched with a 2.5M aqueous solution of NaOH (11.7 mL) and a 30% aqueous solution of H2O2 (2.9 mL). The reaction mass was refluxed for 1 h and then cooled to room temperature (19-21°C). The organic layer was separated and washed with 75 mL of saturated Na2SO3 aqueous solution.  The aqueous layer was extracted with ethyl acetate (3 x 70 mL). The combined organic layers were washed with brine (150 mL), dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure to yield the crude product. The obtained yellow residue was used in the next steps with a preliminary purification by silica gel using flash column chromatography (MeCN/DCM 2:98) to furnish 6.83 g (~80% yield) of the less impure, crude product. TLC (MeCN:CH2Cl2 2:98 v/v): Rf = 0.45. HRMS-ESI (m/z): [M+Na]+ calcd. for C13H18O2Na, 229.1204; found 229.1206.   (((2R,3R)-1-(Benzyloxy)-3-methylpent-4-en-2-yl)oxy)(t-butyl)dimethylsilane (98). To a solution of impure 97 (4.13 g crude mass, 20 mmol) in dry dichloromethane (150 mL) at 0°C was added 2,6-lutidine (8.1 mL, 70 mmol) followed by dropwise addition of TBDMS-OTf (7.3 mL, 30 mmol). The reaction mixture was stirred for 1 h at 0°C and 1 h at room 88  temperature before it was quenched by addition of a saturated aqueous solution of NaHCO3 (50 mL). The organic layer was separated, and the aqueous layer was extracted with DCM (3x100 mL). The combined organic layers were dried with MgSO4, filtered and concentrated under reduced pressure. The crude material was purified on silica gel using flash column chromatography (EtOAc/hexanes 2:98) to yield 2.8 g (65% yield) of the less impure crude product as a colorless oil, which could be used in the next step without further purification. TLC (EtOAc:hexanes 5:95 v/v): Rf = 0.54 1H NMR (300 MHz, CD2Cl2): δ 7.43 – 7.33 (m, 5H), 5.88 (ddd, J = 17.0, 10.8, 8.2 Hz, 1H), 5.11 – 4.97 (m, 2H), 4.51 (d, J = 2.2 Hz, 2H), 3.81 (m, 1H), 3.54 – 3.35 (m, 2H), 2.55 – 2.38 (m, 1H), 1.09 (d, J = 7.0 Hz, 3H), 0.95 (s, 9H), 0.12 (s, 3H), 0.10 (s, 3H) 13C NMR (75 MHz, CD2Cl2): δ 141.0, 139.4, 128.9, 128.3, 128.1, 115.2, 75.5, 74.0, 73.9, 42.4, 26.4, 18.8, 17.5, -3.7, -4.4. HRMS-ESI (m/z): [M+Na]+ calcd. for C19H32O2NaSi, 343.2069; found 343.2070.   (3R,4R)-5-(Benzyloxy)-4-((t-butyldimethylsilyl)oxy)-3-methylpentane-1,2-diol (99). To a solution of 98 (2.8 g, 8.74 mmol) in acetone and water (140:70 mL/mL) at room temperature was added NMO (3.15 g, 26.4 mmol) and OsO4 (4% solution in water, 3.5 mL, 0.426 mmol). The reaction mixture was stirred for 16 h at room temperature after which it was quenched with sodium sulfite (2.84 g) and stirred further for 2 h. Acetone was removed under reduced pressure, and the remaining aqueous layer was extracted with ethyl acetate (7 x 50 89  mL). The combined organic layers were washed with a saturated aqueous solution of NH4Cl followed by brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The residue was purified by silica gel using flash column chromatography (ethyl acetate/hexane 30:70) to produce 1.52 g of the product as a colorless oil in 48% isolated yield. TLC (EtOAc:hexanes 60:40 v/v): Rf = 0.45 1H NMR (300 MHz, CD2Cl2): δ 7.39 – 7.25 (m, 5H), 4.51 (dd, J = 12.0, 3.1 Hz, 2H), 4.04 – 3.93 (m, 1H), 3.70 – 3.36 (m, 5H), 1.98 – 1.79 (m, 1H), 0.94 – 0.81 (m, 12H), 0.10 (s, 3H), 0.08 (s, 3H) 13C NMR (75 MHz, Methylene Chloride-d2): δ 138.6, 128.9, 128.3, 128.2, 77.1, 74.6, 74.2, 73.9, 73.5, 72.7, 72.3, 65.8, 65.4, 40.5, 37.7, 26.2, 18.6, 18.5, 12.8, 11.2, -4.1, -4.7, -4.8. HRMS-ESI (m/z): [M+Na]+ calcd. for C19H34O4NaSi, 377.2124; found 377.2120.   (2S,3R)-4-(Benzyloxy)-3-((tert-butyldimethylsilyl)oxy)-2-methylbutanal (131). To a stirred solution of 99 (100 mg, 0.28 mmol) in a 3:1 methanol/water mixture (6:2 mL/mL) was added NaIO4 (75 mg, 0.34 mmol) and the resulting suspension was stirred at room temperature for 30 minutes. Brine solution (6 mL) was added and the reaction mixture was extracted with ethyl acetate (4 x 10 mL), and the combined organic layers were washed with water (15 mL) and dried over anhydrous Na2SO4. The solvent was removed under reduced pressure and the residue was purified by silica gel (EtOAc/hexanes 30:70) to afford the product (68 mg, 75%) as a colorless oil. 90  Note: the yield for the scale-up of 131: 10.3 g, 73% (reaction performed on 15.5 g, 43.8 mmol of 99) TLC (EtOAc:hexanes 60:40 v/v): Rf = 0.75 1H NMR (300 MHz, CD2Cl2) δ 9.72 (d, J = 1.9 Hz, 1H), 7.41 – 7.23 (m, 5H), 4.49 (s, 2H), 4.13 (td, J = 5.7, 4.3 Hz, 1H), 3.51 (d, J = 5.7 Hz, 2H), 2.59 (dtd, J = 9.0, 7.0, 4.1 Hz, 1H), 1.09 (d, J = 7.0 Hz, 3H), 0.88 (s, 9H), 0.08 (s, 3H), 0.07 (s, 3H) 13C NMR (75 MHz, CD2Cl2) δ 204.1, 138.8, 128.9, 128.2, 128.2, 73.9, 73.3, 72.7, 50.5, 26.1, 18.5, 10.4, -4.1, -4.8. HRMS-ESI (m/z): [M+Na]+ calcd. for C18H30O3NaSi, 345.1862; found 345.1863.   (2S,3R,4R)-5-(Benzyloxy)-4-((t-butyldimethylsilyl)oxy)-3-methyl-2-(((S)-1-phenylethyl) amino)pentanenitrile (132). (S)-α-phenylethylamine hydrochloride (13.4 mL 1 M aqueous solution, 13.4 mmol) and KCN (0.94 g, 13.4 mmol) were added successively to a suspension of starting aldehyde 131 (4.31 g, 13.4 mmol) in MeOH/H2O (18 mL:4.2 mL) at 0°C. The reaction mixture was stirred at 0°C for 30 minutes, then at room temperature for 48 h, diluted with H2O (50 mL) and extracted with ethyl acetate (5 x 30 mL). The organic extracts were washed with brine, dried over anhydrous Na2SO4 and filtered. The solvent was evaporated in vacuo to give the crude product which was purified by silica gel using flash column chromatography (EtOAc/hexanes 91  5:95 to 7:93 gradient) to yield 2.38 g (40%) of the desired diastereomer (1S)-132 as a bright yellow oil, with the total yield of 75% including the undesired diastereomer (1R)-132. TLC (EtOAc/hexanes 30:70 v/v): Rf = 0.64 1H NMR (300 MHz, CD2Cl2): δ 7.48 – 7.18 (m, 10H), 4.44 (dd, J = 11.7, 5.6 Hz, 2H), 4.17 – 3.99 (m, 1H), 3.71 (m, 1H), 3.58 (d, J = 4.7 Hz, 1H), 3.50 – 3.42 (m, 2H), 2.16 – 2.02 (m, 1H), 1.35 (d, J = 6.4 Hz, 3H), 1.03 (d, J = 6.8 Hz, 3H), 0.67 (s, 9H), -0.03 (s, 3H), -0.09 (s, 3H) 13C NMR (75 MHz, CD2Cl2): δ 143.7, 138.4, 128.8, 128.4, 127.8, 127.7, 127.6, 127.0, 119.5, 73.5, 73.4, 72.5, 56.8, 50.5, 40.2, 29.8, 25.6, 24.9, 17.9, 12.2, -4.4, -5.5. HRMS-ESI (m/z): [M+Na]+ calcd. for C27H40N2O2NaSi, 475.2757; found 475.2755.   (3S,4R,5R)-5-(Hydroxymethyl)-4-methyl-3-(((S)-1-phenylethyl)amino)dihydrofuran-2(3H)-one (133). A solution of 132 (2.38 g, 5.28 mmol) in 6N aqueous HCl (23.2 mL) was refluxed for 6 hours. The reaction mixture was diluted with H2O (10 mL) and washed with EtOAc/hexanes 1:1 (3x7 mL). The pH of the remaining aqueous layer was adjusted to 8.5 by addition of a saturated aqueous solution of NaHCO3. Then it was extracted with ethyl acetate (6x10 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The resulting crude product was purified using silica gel flash column chromatography (EtOAc/hexanes 20:80 to 25:75 to 30:70 to 35:65, gradient) to yield the final product as a pale-yellow oil (600 mg, 46% yield). 92  TLC (EtOAc:hexanes 60:40 v/v): Rf = 0.27 1H NMR (300 MHz, CD2Cl2): δ 7.40 – 7.26 (m, 5H), 4.17 – 4.10 (m, 2H), 3.75 (dd, J = 12.4, 3.2 Hz, 1H), 3.58 (dd, J = 12.4, 5.1 Hz, 1H), 3.47 (d, J = 7.8 Hz, 1H), 2.20 (m, 1H), 1.36 (d, J = 6.7 Hz, 3H), 1.02 (d, J = 7.2 Hz, 3H) 13C NMR (75 MHz, CD2Cl2): δ 177.6, 145.1, 128.5, 127.2, 127.1, 85.2, 63.2, 57.6, 56.9, 35.5, 24.3, 13.0. HRMS-ESI (m/z): [M+Na]+ calcd. for C14H19NO3Na, 272.1265; found 272.1265.   (3S,4R,5R)-5-(Hydroxymethyl)-4-methyl-2-oxotetrahydrofuran-3-aminium chloride (134). A solution of 133 (480 mg, 1.92 mmol) in MeOH (40 mL) was subjected to hydrogenolysis in the presence of 10% Pd-C (50 mg, 2 mol%) under atmospheric pressure at room temperature, while the pH was adjusted to 4 by addition of 1 M HCl. Stirring was continued for 16h. The catalyst was then removed by filtering through a plug of celite. The solvent was evaporated under reduced pressure and the remaining crude product was recrystallized from MeOH and used in the next step without further purification.  m.p. at 2ºC/min heating rate: 198.6ºC HRMS-ESI (m/z): [M+H]+ calcd. for C6H12NO3, 146.0812; found 146.0819.  93   (2S,3R,4R)-2-Ammonio-4,5-dihydroxy-3-methylpentanoate (135). Compound 134 (0.2 mmol, crude) was suspended in H2O (2 mL) and LiOH (7.4 mg) was added to the reaction mixture. The reaction was stirred at room temperature for 2 hours. The pH was adjusted to 8 by addition of 0.1 M aqueous solution of HCl, then the solvent was removed in vacuo to afford the crude product in a zwitterionic form which was used in the next step without further purification.   (2S,3R,4R)-2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-4,5-bis((t-butyldimethyl silyl)oxy)-3-methylpentanoic acid (138). Intermediate 135 (0.2 mmol, crude) was dissolved in dry N,N-dimethylacetamide (DMA) (3 mL). To this solution was added imidazole (140 mg, 2 mmol, 10 eq) and TBDMSCl (240 mg, 1.6 mmol, 8 eq) and the reaction mixture was stirred for 18 hours at room temperature. Upon completion of the reaction, the pH of the mixture was adjusted to 6 by adding a phosphate buffer at pH 6 (3 mL, 0.5 M). The resulting mixture was extracted with ethyl acetate (3 x 5 mL). The combined organic layers were washed with brine (10 mL), dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure and dried in vacuo to yield crude 136. 94  Without further purification, the resulting crude product was suspended in 1,4-dioxane (1.5 mL). To this solution was added 0.4 mL of aqueous saturated Na2CO3 solution followed by Fmoc-OSu (105 mg, 0.26 mmol, 1.3 eq). The resulting mixture was stirred at room temperature for 5 hours and the pH was adjusted to 4 by adding a 1 M aqueous solution of HCl. The aqueous mixture was extracted with ethyl acetate (7 x 3 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude product was purified on silica gel using flash column chromatography (EtOAc/DCM/hexanes 20:15:65 to 25:20:55 to 30:25:45, gradient) to furnish the final product 138 as a white solid in 50-65% yield over two steps. Note: Due to the tedious purification, the product was generally used directly in the next step. TLC (EtOAc:CH2Cl2:hexanes 40:30:30 v/v/v): Rf = 0.13 1H NMR (300 MHz, CD2Cl2): δ 7.78 (d, J = 7.5 Hz, 2H), 7.63 (d, J = 7.5, 2H), 7.41 (t, J = 7.4 Hz, 2H), 7.32 (t, J = 6.8 Hz, 2H), 6.20 (d, J = 7.1 Hz, 1H), 4.63 – 4.53 (m, 1H), 4.38 (d, J = 7.1 Hz, 2H), 4.24 (t, J = 6.9 Hz, 1H), 3.86 – 3.68 (m, 3H), 2.50 – 2.38 (m, 1H), 1.06 (d, J = 7.1 Hz, 3H), 0.93 (s, 9H), 0.91 (s, 9H), 0.15 (s, 6H), 0.10 (s, 6H) 13C NMR (75 MHz, Methylene Chloride-d2): δ 174.1, 156.6, 144.7, 144.5, 141.8, 128.2, 127.6, 125.7, 120.4, 76.3, 67.4, 65.7, 56.3, 47.8, 39.5, 30.3, 26.2, 26.1, 18.8, 18.5, 13.9, -4.1, -4.6, -5.2. HRMS-ESI (m/z): [M-H]- calcd. for C33H50NO6Si2, 612.3182; found 612.3183;    95   2,5-Dioxopyrrolidin-1-yl (2S,3R,4R)-2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-4,5 -bis((t-butyldimethylsilyl)oxy)-3-methylpentanoate (139) Starting material 138 (68 mg, 1 eq) and disuccinimidyl carbonate (DSC, 40 mg, 1.43 eq) were dissolved in EtOAc/MeCN (3 mL:1.5 mL) at 4°C. To the clear mixture was added collidine (40 μL, 2.8 eq). The reaction mixture was stirred for 3 hours at room temperature. At this point, an additional 15 mg (0.54 eq) of DSC was added to the reaction and continued stirring at RT for 2 hours. The reaction was then diluted with EtOAc, washed with saturated KH2PO4 (3 x 2 mL) and brine (3 mL). The organic layer was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude product was purified with flash column chromatography using silica gel (EtOAc/DCM/hexanes 10:10:80 to 20:20:60, gradient) to yield 53.5 mg (68%) of the pure product as a white solid. TLC (EtOAc:CH2Cl2:hexanes 40:30:30 v/v/v): Rf = 0.63 1H NMR (300 MHz, CD2Cl2): δ 7.77 (d, J = 7.5 Hz, 2H), 7.62 (dd, J = 7.6, 3.3 Hz, 2H), 7.40 (dd, J = 8.0, 6.8 Hz, 2H), 7.31 (td, J = 7.4, 1.2 Hz, 2H), 6.66 (d, J = 8.4 Hz, 1H), 4.79 (dd, J = 8.5, 5.5 Hz, 1H), 4.46 – 4.37 (m, 2H), 4.25 (t, J = 6.9 Hz, 1H), 3.93 (m, 1H), 3.62 (m, 2H), 2.82 (s, 4H), 2.59 – 2.45 (m, 1H), 1.13 (d, J = 7.1 Hz, 3H), 0.90 (s, 18H), 0.11 (s, 4H), 0.09 (s, 8H) 13C NMR (75 MHz, CD2Cl2): δ 169.4, 156.7, 144.6, 141.8, 128.2, 127.6, 125.7, 125.6, 120.5, 74.3, 67.6, 64.9, 47.8, 39.7, 26.3, 26.23, 26.16, 13.0, -4.3, -4.7, -5.1, -5.3. HRMS-ESI (m/z): [M+Na]+ calcd. for C37H54N2NaO8Si2, 733.3316; found 733.3315.  96  Chapter 3: Synthesis of 6-BMIDA-L-Tryptathionine  3.1 Tryptathionine Crosslinks: An Introduction There are numerous examples of crosslinked peptides in nature, some of which possess a bicyclic structure.97,98 Bicyclic peptides are especially compelling in the world of medicinal chemistry due to their increased rigidity and defined conformation. As such, they generally offer greater stability in physiological conditions, exhibit higher affinity towards their targets, and present better cell penetration due to their structural properties.99 A few examples of crosslinked bicyclic peptides are shown in Figure 3.1.  Figure 3.1 Examples of bicyclic peptides in nature.  The crosslink formed between tryptophan and cysteine is known as a tryptathionine linkage. Whereas various types of crosslinks are observed in bicyclic natural products, to date, the tryptathionine linkage appears to be exclusively found in amatoxins and phallotoxins. By today’s definition, any crosslink between tryptophan and a thiol-containing moiety (including cysteine) falls into the category of tryptathionine linkages (Figure 3.2).51 97   Figure 3.2 General structure of the tryptathionine crosslink formed between tryptophan and A) cysteine (in amatoxins and phallotoxins), and B) any thiol-containing residue.  3.1.1 Synthesis of Tryptathionine Crosslinked Peptides Historically, the first synthesis of tryptathionine was achieved using tryptophan and the sulfenyl chloride of cysteine to afford the crosslink in several phallotoxin derivatives, such as norphalloin (140) (Figure 3.3).100,101  Figure 3.3 Synthesis of norphalloin (140) by Wieland et al. Sulfenyl chloride of cysteine was reacted with tryptophan to form the tryptathionine linkage (shown in blue).  Although this method initially provided a route for the installation of tryptathionine en route to the synthesis of select phallotoxins and amatoxins, it was laborious, low yielding, and required a tedious protection-deprotection strategy. In 1975, Walter Savige reported an alternative route to tryptathionine formation that started with the production of a new oxidation 98  product of tryptophan, 3a-hydroxy-1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole-2-carboxylic acid, known as Hpi.102 This product was obtained as a mixture of two diastereomers via the reaction of tryptophan with peroxyacetic acid in water at low temperatures (Figure 3.4).103  Figure 3.4 Synthesis of Hpi by Savige.102  The Hpi moiety is, in its own right, an important building block found in the structure of several natural products.104–106 However, in the context of tryptathionine synthesis, its reaction with thiol-containing residues is the most frequently utilized method to produce a tryptathionine crosslink. In 1976, in an attempt to synthesize the toxic peptides of Amanita phalloides, Savige and Fontana disclosed a new method of tryptathionine formation via the reaction of cysteine with Hpi in the presence of trifluoracetic acid (TFA).52 In their report, the reaction was performed in 25% TFA with a duration of 2 days, in which the crosslinked tryptophan-cysteine product was isolated with an 80% yield (Figure 3.5). Moreover, they linked glutathione to tryptophan employing the same method to obtain S-tryptophanylated glutathione in 85% yield. Today, this reaction is known as the Savige-Fontana reaction, named after W. Savige and A. Fontana.  Figure 3.5 Original reaction between Hpi and cysteine in the presence of TFA to yield the tryptathionine crosslink in 80% yield (Savige and Fontana).  99  A few years later, in the late 1980s, Zanotti et al. employed the Savige-Fontana reaction for the synthesis of amatoxin analogues.28,29,45 In their work, a series of linear peptides containing Nα-Boc-Hpi and S-Trt-L-cysteine were synthesized. Following treatment with TFA, monocyclic octapeptides of the amatoxin analogues were obtained, and the synthesis was concluded with a macrocyclization step to produce the final bicyclic peptides (Figure 3.6).  Figure 3.6 Synthesis of amatoxin analogues by Zanotti et al.45 Hpi, S-Trt-Cys and the resulting tryptathionine are shown in blue. X and Y (in red) were the different residues employed in the synthesis to yield different analogues.  To date, this reaction sequence represents the most widely utilized procedure to install the tryptathionine crosslink in various analogues of amatoxins and phallotoxins.25,28–30,45,101,107 The proposed mechanism for the tryptathionine formation is shown in Figure 3.7. First, the protecting group on Nα is removed with TFA. Following protonation of the indole nitrogen, the bond between Nα and C-8a is cleaved. Then, the concurrent nucleophilic attack on C-8a by the cysteine-thiol and the departure of the C3a-hydroxyl, followed by re-aromatization to tryptophan, yields the final tryptathionine linkage.108 It appears that the proton on N-1 (Nindole) is crucial for the reaction to proceed. Hence, either a tryptophan with unprotected Nindole or a 100  tryptophan containing an acid-labile protecting group on Nindole is required for the Savige-Fontana reaction. It is noteworthy that both diastereomers of Hpi can take part in the Savige-Fontana reaction to induce tryptathionylation.  Figure 3.7 Proposed mechanism for the Savige-Fontana reaction.  While the Savige-Fontana tryptathionylation reaction is generally performed with Hpi as the oxidized form of tryptophan, it is known that other oxidation products of tryptophan, with the same overall oxidation state, may also be used to generate a tryptathionine crosslink. Shown in Figure 3.8, 2,3-dihydropyrrolo[2,3-b]indole can react with cysteine (or other thiols) under basic conditions (e.g. ammonium bicarbonate buffer) to yield a tryptathionine (the mechanism is not given but may involve either base-catalysis or perhaps a radical chain process initiated by contaminating oxygen).109  Figure 3.8 Synthesis of tryptathionine via a dihydropyrrolo indoline species.  101  Whereas the dihydropyrrolo indoline moiety can lead to the formation of a tryptathionine linkage, it has not been extensively studied for this purpose. This fact is likely due to the noted instability of dihydropyrrolo indoles in the presence of mild acids and the necessity to use acidic conditions to induce tryptathionylation in most amatoxin syntheses, not the basic conditions shown in Figure 3.8.109 Finally, another notable method for the synthesis of a tryptathionine linkage is the direct treatment of tryptophan with thiols in the presence of Hg(OAc)2 and acetic acid.110 In this method, first a mercury adduct of tryptophan is formed, followed by addition of a thiol-containing moiety (Figure 3.9). This reaction, due to the addition of the thiol to various positions of the tryptophan, is usually low yielding and not generally used as a facile route towards tryptathionylation.  Figure 3.9 Formation of a tryptathionine crosslink via the reaction of tryptophan with mercaptoethanol in the presence of Hg(OAc)2 and acetic acid.110  3.1.1.1 Synthesis of Hpi As mentioned above, the first synthesis of Hpi involved oxidation of tryptophan with peroxyacetic acid to obtain the product with an isolated yield of 20%. Since then, several routes have been designed and tested to produce Hpi with elevated yields, most of which fall into three major categories: 1) oxidation with a mild oxidizing agent (e.g. DMDO), 2) 102  photosensitized oxidation in the presence of molecular oxygen, and 3) oxidative deselenation (Figure 3.10).51  Figure 3.10 Different methods to produce Hpi: A) mild oxidation of tryptophan, B) photosensitized oxidation, C) oxidative deselenation (N-PSP: N-phenylselenophthalimide).  In the case of photosensitized oxidation (Figure 3.10, B), molecular oxygen accompanied by rose bengal has been utilized to generate Hpi. However, the yields for this reaction is generally low (20-50%). Furthermore, no diastereoselectivity is achieved with this method, limiting its application in the synthesis of natural products containing Hpi in their core structure.111 The oxidative deselenation method (Figure 3.10, C) involves deselenation of the corresponding 3a-phenylseleno-pyrroloindoles to Hpi and provides improved yields. Moreover, diastereoselectivity may be achieved using this method following the initial selenation step.112 However, the number of overall synthetic steps is higher, hence diminishing its attractiveness for this application. Mild oxidation of tryptophan (Figure 3.10, A) is generally the most accepted route towards the generation of Hpi. Although peroxyacids were originally used to accomplish this oxidation, the most common oxidant utilized in this method is 3,3-103  dimethyldioxirane (DMDO) (145). The yields for DMDO oxidations are generally good (50-90%), and in some cases, only one diastereomer is formed in large excess.113 The proposed mechanism for the oxidation of tryptophan with DMDO is shown in Figure 3.11. Although the mechanism shown here suggests that the oxidation of the double bond takes place prior to the imine formation, there is debate over the concertedness of these two transformations with some evidence of pre-ordered overlap of the nitrogen lone pair into the anti-bonding orbital of the indole double bond that ensures facial preference for attack by DMDO.113,114  Figure 3.11 Proposed mechanism for the DMDO oxidation of tryptophan to Hpi.  DMDO consists of an unstable, strained three-membered ring, rendering its synthesis a tedious task; it must be prepared via distillation from an acetone/Oxone® solution and collected at -78°C as a dilute solution in acetone. In fact, DMDO is unstable to the extent that it has a short shelf-life of two weeks at -20°C. To address these issues, Blanc and Perrin designed a methodology through which DMDO could be produced in situ by mixing the starting tryptophan with Oxone®, acetone, H2O and sodium bicarbonate at 0°C to obtain the corresponding Hpi, hence avoiding the laborious preparation of DMDO via distillation. This reaction could also be performed in a biphasic manner (Figure 3.12).115 104   Figure 3.12 Oxidation of tryptophan to Hpi using the in situ formation of DMDO (Blanc et al.).  3.1.1.2 Characteristics of the Tryptathionine Crosslink The most apparent physical signature of a tryptathionine crosslink is found in its UV absorbance spectrum. With a maximum absorbance wavelength (λmax) of 290 nm, tryptathionines may easily be distinguished from other linkages. Furthermore, this crosslink exhibits two smaller shoulders at 285 nm and 300 nm in its UV absorbance curve (Figure 3.13).  Figure 3.13 General UV absorption curve for tryptathionine with λmax=290 and shoulders at 285 and 300 nm.51  Additionally, in the case of tryptathionine crosslinks found in amatoxins and phallotoxins, they possess distinctive circular dichroism (CD) spectra, indicative of their rigid bicyclic conformations. The region of interest for the CD spectra of these toxins falls between 260-320 nm. Phallotoxins show positive Cotton effects in 220-260 nm and 280-310 nm regions. On the 105  other hand, amatoxins display a positive Cotton effect in the 280-310 nm region, while exhibiting a negative Cotton effect in the 220-260 nm range (Figure 3.14).  Figure 3.14 Circular dichroism (CD) spectra of α-amanitin in water at pH 7 and pH 12 (phenolate of 6-OH-L-Trp) (reproduced from Wieland).6  3.2 6-Hydroxy-L-Tryptathionine: Challenges and Synthesis In order to install the 6-hydroxy-tryptathionine crosslink of α-amanitin, the synthesis of 6-hydroxy-L-tryptophan was presumably required. This amino acid, analogous to an unsubstituted tryptophan, was thought to undergo an oxidation reaction using a mild oxidant (e.g. DMDO) to produce the corresponding 6-hydroxy-Hpi. This building block could subsequently be incorporated into the linear sequence of α-amanitin for installation of the 6-hydroxy-tryptathionine following the Savige-Fontana reaction with the cysteine residue of the peptide. While a detailed discussion about the synthesis of the linear peptide sequence of α-amanitin and an eventual tryptathionylation step is provided in Chapter 4, Figure 3.15 depicts a summary of the initial retrosynthetic scheme to install the 6-OH-Ttn crosslink. 106   Figure 3.15 A) General strategy for the synthesis of an unsubstituted tryptathionine crosslink via oxidation of tryptophan to Hpi followed by the Savige-Fontana reaction. B) Proposed retrosynthetic scheme for the synthesis of a 6-hydroxy-Ttn from 6-hydroxy-L-Trp.  One of the synthetic challenges related to the installation of a 6-hydroxy-tryptathionine linkage would be the synthesis of 6-hydroxy-L-tryptophan (146). To this date, there are no commercial sources nor reliable reports on the synthesis of this amino acid. In fact, it may be argued that the C-6 position of tryptophan (or indole) is likely the most difficult position to functionalize with a hydroxyl group given the paucity of reports thereof. The only synthesis of 6-OH-Trp that can be gleaned from the literature dates back to 1987; therein, H2O2, superacid (SbF5.HF) and H2O were used at -20°C to obtain a mixture of 4- and 6-hydroxy-tryptophan in low yields (Figure 3.16).53 However, this method has not been cited by others, likely reflecting its considerable inconvenience and quite possibly its irreproducibility. 107   Figure 3.16 Reported synthesis of 6-hydroxy-L-tryptophan (146) using H2O2 and superacid.  The fact that 6-OH-L-Trp has not been successfully synthesized with practicable yields may be attributed to its instability owing to the electronic properties of the 6-hydroxy-indole core. As shown in Figure 3.17, the flow of electrons from the hydroxyl group through the aromatic system makes C-2 (as well as C-5 and C-7) extremely prone to oxidation. Hence, it may be envisioned that 146 would be effortlessly converted to an undesired oxindole product (148) in the presence of an oxidant (e.g. atmospheric oxygen).  Figure 3.17 Flow of electrons from the 6-OH group into the indole ring, rendering C-2 extremely susceptible to oxidation.  It was thus anticipated that attempts to oxidize 146 to the desired 6-OH-Hpi (149) for use in the Savige-Fontana reaction and incorporation into α-amanitin could very well result in the formation of oxindole 148 as the major product rather than the desired Hpi. Hence, this oxidation step was thought to be the most formidable challenge in the synthesis of 6-hydroxy-tryptathionine crosslink (Figure 3.18). 108   Figure 3.18 Hypothetical retrosynthetic scheme for the DMDO oxidation of 146 to the corresponding Hpi (149) followed by its proposed incorporation into α-amanitin (Ttn shown in blue). Manipulation of protecting groups might be necessary prior to incorporation in amanitin.  3.3 Results and Discussion Due to a lack of literature precedent for an efficient synthesis of 6-hydroxy-tryptophan, I turned my focus to a recent report by Baran and co-workers in which a 6-boronyl-tryptophan intermediate was employed en route to the total synthesis of Verruculogen and Fumitremorgin A.116 In their work, a ligand-controlled borylation was carried out to install a pinacol-boronate on C-6 of tryptophan. To this end, the Nindole of Nα-Boc-L-Trp-OMe (150) was first protected by reaction with chloro triisopropylsilane (TIPSCl). According to Baran and co-workers, the TIPS protecting group was essential to shield C-2 and C-7 from reacting with the borylating reagents.117 In Baran’s report, several different conditions and ligands were screened for the iridium-catalyzed borylation reaction. It was found that the use of 1,10-phenanthroline as the ligand provided the highest yield and the best regioselectivity for the generation of 6-BPin-L-Trp (152) as the major regio-isomer, whereas 5-BPin-L-Trp (153) was formed as a competing isomer (152:153 8:1). Based on the targets under consideration, they effected a Chan-Lam coupling with MeOH using Cu(II)acetate, molecular oxygen and triethylamine to yield the 6-methoxy-tryptophan product (154). This intermediate was eventually converted to the final natural targets following several synthetic steps (Figure 3.19). 109   Figure 3.19 Part of the total synthesis of verruclogen and fumitremorgin A by Baran et al. A ligand-controlled C-H borylation of C-6 of tryptophan followed by a Chan-Lam coupling with MeOH were employed. The 6-methoxy-indole core is shown in red.  3.3.1 Synthesis of 6-Methoxy-Tryptophan Followed by Demethylation In light of the work by Baran and co-workers, I first embarked on the synthesis of 6-methoxy-tryptophan (154) to validate their report and apply it towards the synthesis of amanitin. To wit, I hypothesized that demethylation of the methoxy group would presumably yield the 6-hydroxy-tryptophan required for the synthesis of 6-hydroxy-Hpi (Figure 3.20).  Figure 3.20 Proposed demethylation of 6-methoxy-tryptophan to obtain the protected 6-hydroxy-tryptophan.  Hence, Baran’s protocol (see Figure 3.19) was followed to obtain Nα-Boc-Nindole-TIPS-6-OMe-L-Trp-OMe (154) in yields consistent with the reported procedure. To transform the methoxy group to a hydroxyl, boron tribromide (1 M solution in DCM) was utilized. Employing different amounts of BBr3 led to the deprotection of various functional groups: 1 110  equivalent deprotected the Boc, 3 equivalents deprotected the Boc and the methyl ether, and 5 equivalents of BBr3 deprotected the Boc, the methyl ether and the methyl ester (Figure 3.21).  Figure 3.21 Failed attempts to convert 154 to 6-hydroxy-L-Trp products (155-157).  Nevertheless, in all cases, the overall yield was low and mixtures of different products along with unidentified by-products were obtained. Although the demethylation of 154 could have theoretically provided an avenue to 6-hydroxy-L-Trp, the inability to effectively convert the methyl ether to a hydroxyl group convinced me to abandon this route in search of an alternative.  3.3.2 Oxidation of 6-BPin-L-Trp to 6-OH-L-Trp Whereas the Chan-Lam coupling of 152 could lead to the formation of 154, a direct oxidative deborylation of the 6-BPin would provide a shorter and more efficient route towards a 6-hydroxy-L-tryptophan. Literature review led me to a report which described the conversion of 4-BPin-L-tryptophan to 4-OH-L-tryptophan by reaction with sodium perborate (NaBO3.4H2O) as the oxidant (Figure 3.22).118  111   Figure 3.22 Oxidation of 4-BPin-L-Trp to 4-OH-L-Trp (159) with sodium perborate (Bartoccini et al.).  Inspired by this work, analogous oxidative conditions were applied to the Nindole-deprotected 6-BPin-L-Trp (160) in hope of obtaining 6-OH-L-Trp (Figure 3.23). Although the reaction was nearly quantitative based on TLC, the isolated yield of the product was low (40%) due to its degradation on silica gel and the plausible air-oxidation of the extremely electron-rich indole ring. This decomposition was evident from TLC analysis and from the color change of the purified product once stored in CD2Cl2 at room temperature for a few hours.  Figure 3.23 TIPS-deprotection of 152 followed by the oxidation of BPin to OH using sodium perborate.  Despite the low isolated yield for 161, I was able to conserve sufficient quantities of material to test its oxidation to the corresponding 6-OH-Hpi. First, to avoid the tedious procedure of preparing DMDO via distillation, I utilized the method by Blanc and Perrin for the in situ production of DMDO (vide supra). Mixing 161 with NaHCO3 and a slight excess of Oxone® in acetone/H2O resulted in the formation of the corresponding Hpi (162) only as the minor product (10%), while the oxindole species (163) was obtained as the major component (14%) along with several unidentified oxidation by-products (Figure 3.24, A). In 112  the second attempt, DMDO was prepared as a 0.06 M solution in acetone via distillation from a mixture of acetone, Oxone® and NaHCO3 at low temperatures. Employing different amounts of the distilled DMDO to oxidize 161 furnished the same mixture of products and by-products in almost identical ratios (Figure 3.24, B). The identities of 162 and 163 were subsequently confirmed by 1H-NMR.  Figure 3.24 Oxidation of 161 with A) DMDO formed in situ, B) distilled DMDO (0.06 M in acetone). Crude product contained the desired 6-OH-Hpi (162) as the minor component and the corresponding oxindole (163) as the major product. Unidentified oxidation by-products comprised the rest of the crude mixture. The distinctive protons that were closely studied by NMR to establish the identities of 162 and 163 are shown in blue.  Although the yield for the generation of the desired Hpi product was low, efforts were made to isolate 162 and 163. Unfortunately, in line with my expectations, 6-OH-Hpi 162 proved to be exceptionally delicate and air-sensitive to the extent that its decomposition was apparent upon brief storage in the NMR tube, producing several unknown oxidation by-products. This degradation was evident from the TLC of the purified material after a few hours, the color change of the crude reaction mixture and the 1H-NMR sample containing pure 162 (Figure 3.25). 113   Figure 3.25 A) Silica-gel column purification of crude 162, showing the colored oxidation by-products. B) The DMDO oxidation reaction mixture turning red upon warming up to RT. C) TLC of the purified 162 after storage for a few hours in an NMR tube. TLC showed new spots, all turning red upon exposure to air. D) NMR tube containing pure 162 changing color and forming a precipitate in CD2Cl2.  Based on these experiments, it was clear that the anticipated 6-hydroxy-tryptophan was indeed too air-sensitive to be effectively utilized in the synthesis of its corresponding Hpi (see Figure 3.17). Hence, alternative strategies were considered to achieve the 6-hydroxy-tryptathionine crosslink.  3.3.3 Oxidation of 6-BPin-L-Trp to 6-BPin-Hpi Rather than oxidizing 6-BPin-L-Trp to 6-OH-L-Trp, I envisioned retaining the boronate on C-6 to serve as a latent hydroxyl group that could be revealed following an oxidation at C-3 to obtain 6-BPin-Hpi. Treating 152 with the DMDO (generated in situ) did not result in the formation of noticeable amounts of the desired product (164), and the starting material was 114  fully recovered (Figure 3.26, A). Thus, I decided to remove the TIPS group from Nindole to obtain 160 and attempt the DMDO oxidation of the deprotected compound. Disappointingly, the oxidation occurred extremely slowly and only produced the 6-BPin oxindole (166) as the major product with the 6-boronyl-pinacolate intact, along with several unidentified by-products (Figure 3.26, B). In an effort to modify the electronic properties of the indole ring, the pinacol-boronate was converted to the corresponding trifluoroborate anion (BF3-) and oxidation was performed once again. This reaction also suffered from a similar fate, forming a smaller amount of the oxindole product (168), accompanied by numerous unknown by-products (Figure 3.26, C).  Figure 3.26 Oxidation trials on various 6-boronyl-L-tryptophans. A) Oxidation of Nα-Boc-Nindole-TIPS-6-BPin-L-Trp-OMe. No reaction was observed. B) Oxidation of Nα-Boc-6-BPin-L-Trp-OMe. Oxindole 166 was formed as the major product. C) Oxidation of Nα-Boc-6-BF3-L-Trp-OMe. Oxindole 168 was formed as the major product in a lower yield. For B and C, several other unidentified oxidation products were obtained. Reaction conditions: Oxone®/NaHCO3/acetone/H2O/0°C or DMDO (0.06 M in acetone)/DCM/-78°C.  115  Failed efforts to oxidize these 6-boronyl-tryptophans to the corresponding Hpi products showed that the presence of a boronate at C-6 imposes a perplexing effect on the indole ring, rendering it more susceptible towards oxidation at C-2, hence generating oxindoles as major products. A speculative mechanism for the extremely low rate of the reaction between 6-BPin-L-Trp-OMe and DMDO that leads to the formation of the oxindole product (166) is depicted in Figure 3.27. Whereas further evidence is required to attest or reject this hypothetical mechanism, it stands to reason that the DMDO oxidation of 6-boronyl-tryptophans did not provide the necessary robustness and efficiency required for the synthesis of 6-boronyl-Hpi (165). Therefore, I turned my attention to electrophilic fluorinating reagents as alternative oxidants.  Figure 3.27 Proposed mechanism for the slow reaction of 6-BPin-L-Trp-OMe (160) with DMDO to produce oxindole 166 as the major product.  3.3.4 Electrophilic Fluorocyclization: An Alternative Oxidation Method With increased interest in incorporation of fluorine into medicinal targets due to the effects it can exerts on the biological properties of drugs,119,120 different types of reagents for the introduction of fluorine into organic compounds have been developed. An emerging class of 116  fluorinating reagents consists of electrophilic fluorinating (F+) reagents. Due to the unusual tendency of fluorine to serve as an electron-deficient Fδ+-moiety in these reagents, a limited number of them have been employed in organic synthesis, nearly all of which contain N-F bonds.121 A few common F+-reagents are shown in Figure 3.28. Prior to disclosing the results of using an electrophilic fluorination method for the oxidation of 6-BPin-L-Trp, a brief review of the previous reports on F+-reagents used to induce a dearomative cyclization reaction follows.  Figure 3.28 Examples of common electrophilic fluorinating reagents.  3.3.4.1 FP-T300: A Mild Reagent to Induce Fluorocyclization Whereas F+-reagents have found various applications in organic synthesis, the concept of utilizing them to induce cyclization reactions is relatively recent. In an effort to prepare 3-fluorooxindoles from indoles using SelectfluorTM, Takeuchi et al., noticed the formation of a 3a-fluoropyrrolo[2,3-b]indole species as a by-product.122 In this work, several different indole derivatives were treated with SelectfluorTM in MeCN/H2O as the solvent. In the presence of H2O, the fluoro-indolenine intermediate reacts with H2O to produce the corresponding oxindole as the main product (Figure 3.29). 117   Figure 3.29 Proposed mechanism for preparation of 3-fluorooxindoles from indole derivatives using SelectfluorTM. The non-fluorinated oxindole (in the dashed box) may be obtained as a by-product in the presence of acid (Takeuchi et al.).  However, in the case of tryptamines and tryptophan, Fujiwara and co-workers claimed that formation of a 3a-fluoropyrrolo[2,3-b]indole (Fpi) by-product was a result of the nucleophilic attack of Nα to the imine intermediate (Figure 3.30). Hence, to produce the Fpi compound as the major product, the reaction had to be carried out in anhydrous conditions to force an intramolecular cyclization product and prevent the formation of the oxindole.123  Figure 3.30 Fluorination of Nα-protected tryptophans to Fpi (red arrows) and 3-fluorooxindole (blue arrows) using SelectfluorTM (Fujiwara et al.).   To obtain the Fpi species as the sole major product of the fluorination reaction, Takeuchi and co-workers evaluated several F+-reagents and found that N-fluoro-2,4,6-118  trimethylpyridinium triflate (FP-T300) provided the highest yields. Since then, other groups have utilized this strategy to synthesize analogues of natural products containing an Fpi core. In 2001, Kirk et al. introduced a novel fluorination-cyclization (fluorocyclization) of tryptophan-containing dipeptides en route to the synthesis of fluorogypsetin (173) and fluorobrevianamide E (170) (Figure 3.31).124 It is worth mentioning that Zhao and Perrin, in 2012, published a stereoselective synthesis of brevianamide E (171) using DMDO as the oxidizing reagent to obtain the Hpi core of this natural product.125  Figure 3.31 Synthesis of the Fpi core (in red) of fluorobrevianamide E (170) and fluorogypsetin. The Hpi core of brevianamide E (171) and gypsetin (174) is shown in blue.  Furthermore, in 2015, Yakura and co-workers synthesized the 10b-fluorinated analogue of protubonine A (177) using FP-T300 to induce the fluorocyclization reaction of a tryptophan-containing dipeptide (175) (Figure 3.32).126  119   Figure 3.32 Fluorocyclization of a tryptophan-containing dipeptide en route to the synthesis of a 10b-fluorinated analogue of protubonine A. The Fpi core is shown in red, while the Hpi core of the natural protubonine A is shown in blue.  3.3.5 Fluorocyclization of 6-Boronate-L-Tryptophan Inspired by the use of FP-T300 in the fluorocyclization reactions mentioned above, I envisioned the fluorocyclization of Nα-Boc-Nindole-TIPS-6-BPin-L-Trp-OMe (152). The first trial of this reaction, using 2 equivalents of FP-T300, failed to afford significant amounts of the desired product, likely due to the fact that the indole nitrogen must be unprotected for the reaction to proceed. Hence, Nindole-deprotected 6-BPin-L-Trp (160) was used in my next attempt to afford the desired 6-BPin-Fpi (179) along with a dihydropyrrolo indoline (DHpi) product (180) as an inseparable mixture (179:180 1:2) (Figure 3.33). Generally, the majority of the obtained Fpi and DHpi degraded on silica gel. Hence, the crude product was usually used in the next step as a mixture without further purification.  Figure 3.33 Fluorocyclization of 160, yielding the corresponding DHpi (dashed box) and Fpi (2:1 ratio).  Observing the DHpi (180) as the major product was intriguing, since the only effective route reported for the synthesis of DHpi has been treatment of a suitably protected tryptophan 120  with tBuOCl and triethylamine (see Figure 3.8).109 At this point, I hypothesized that the pinacol-boronate plays a key role in inducing the formation of DHpi via partial fluorophilic activation of the 3a-fluorine of another Fpi molecule to promote its departure as a leaving group, ultimately resulting in the elimination of HF (Figure 3.34). 1H-NMR experiments and the X-ray crystal structure confirmed the proposed structure for this compound.  Figure 3.34 Proposed mechanism for the formation of DHpi (180) and its XRD structure.  While these observations could potentially be the cornerstone of a novel methodology for the synthesis of dihydropyrrolo indolines, I was predominantly interested in the ability of Fpi and DHpi to react in the Savige-Fontana reaction for the installation of a tryptathionine crosslink. To answer this question, 160 was saponified to the free acid 181, followed by fluorocyclization to the corresponding Fpi (182). It is noteworthy that no dihydropyrrolo indoline product was observed in this reaction. Then, 182 was coupled to n-butylamine to obtain amide 183. Next, 183 was subjected to the Savige-Fontana conditions (TFA/DCM 1:1) with n-dodecanethiol as the thiol source. To my delight, the expected tryptathionine products 121  (184 and 185) were observed, however not without 40% deborylation of the starting material (186) (Figure 3.35).  Figure 3.35 Synthesis and test Savige-Fontana reaction of Fpi 183 with n-dodecanethiol as the thiol source (6-BPin:5-BPin = 8:1 in all intermediates).  Although formation of the desired tryptathionine reflects the ability of Fpi to participate in the Savige-Fontana reaction, deborylation of 40% of the starting material limited the applicability of 160 in my synthetic plans. It is thought that for the deborylation to occur, first the pinacol-boronate is hydrolyzed to the boronic acid, which will in turn undergo an ipso protonation at C-6, followed by hydration of the boronic acid leading to a proto-deborylation process, as observed for anisylboronic acids.127 In search for a suitable protecting group for boron that could withstand the TFA treatment, I discovered that MIDA (N-methyliminodiacetic acid) provided the stability required for the Savige-Fontana conditions.128 Thus, I opted to use the MIDA-boronate in lieu of the pinacol-boronate. To this end, 181 was converted to 6-BMIDA-L-Trp-OH (187) using excess MIDA in DMSO as the solvent, while heating to 110°C overnight. It is noteworthy that yields for the direct pinacol-to-MIDA exchange reactions are relatively low. In addition, due to the sensitivity of 181 to oxidants, alternative methods (e.g. in situ deprotection of pinacol using NaIO4 and converting the resulting boronic acid to a MIDA-boronate) were not feasible. 122  Furthermore, following the replacement of pinacol with MIDA, the 6-BMIDA-L-Trp-OH regio-isomer (187) could be isolated following its separation from 5-BMIDA-L-Trp-OH. Next, 187 was fluorocyclized to the corresponding Fpi (188) with an estimated yield of 85%, which was used in the next step without further purification (Figure 3.36).  Figure 3.36 Converting pinacol-boronate 180 to the corresponding MIDA-ester, followed by fluorocyclization to afford Fpi 188 (MIDA is shown in blue).  At this stage, it appeared that I had in hand the necessary tryptophan-derived building block to install a 6-OH-tryptathionine on the final α-amanitin following further manipulations. However, suitable conditions needed to be devised for the oxidation of the MIDA-boronate to a hydroxyl group in the presence of the tryptathionine thioether once the 6-BMIDA-tryptathionine was incorporated into amanitin. To pinpoint the proper conditions for oxidative deborylation, I set up a competition reaction: the 6-BMIDA-tryptathionine system was simulated by mixing 2-naphthyl-BMIDA and thioanisole at an equimolar ratio in acetone/ethanol. To deprotect the MIDA, the resulting solution was briefly treated with 0.5 M aqueous KOH to yield 2-naphthylboronic acid. Then, the reaction mixture was cooled down to 0°C in an ice bath and 1.2 equivalents of mCPBA was added as a solution in EtOH to oxidize the boronic acid to hydroxyl. The reaction mixture was stirred at 0°C for 5 minutes, at which 123  point it was acidified to pH 3 by the addition of 0.1 M aq. HCl. The progress of the reaction was monitored by mass spectrometry and TLC. It was evident that the MIDA deprotection and the boronic acid oxidation took place rapidly in the said conditions without noticeable oxidation of the thioanisole to the corresponding sulfoxide (Figure 3.37).  Figure 3.37 Optimizing the conditions for the oxidative deborylation of an aryl-BMIDA (2-naphthyl-BMIDA) in the presence of an aryl thioether (thioanisole) to simulate the oxidation of the 6-BMIDA-tryptathionine system.  With these results in hand, I anticipated the incorporation of the 6-BMIDA-Fpi (188) into the linear sequence of α-amanitin to complete the bicyclic structure of the toxin by installing the 6-hydroxy-tryptathionine crosslink (vide infra).  3.4 Conclusion To begin this chapter, a brief introduction of tryptathionine crosslinks was provided. Then, an overview of the conventional methods to induce the tryptathionylation reaction was disclosed. Next, different methods for the preparation of 3a-hydroxypyrrolo[2,3-b]indoline (Hpi) were discussed. It was shown that the synthesis of the oxidatively delicate 6-hydroxy-tryptophan was not efficient, and traditional strategies to oxidize this amino acid to the corresponding 6-OH-Hpi could not yield the desired product and instead produced 6-hydroxy-oxindole as the major product. 124  Following a concise introduction on electrophilic fluorinating reagents, specifically N-fluoro-2,4,6-trimethylpyridnium triflate (FP-T300), my efforts towards fluorocyclization of 6-borylated tryptophan were discussed. It was proved that 6-BPin-L-Trp could effectively be fluorocyclized to obtain 6-BPin-Fpi, and that 3a-fluoropyrrolo[2,3-b]indolines could undergo tryptathionylation under the Savige-Fontana conditions. To avoid the deborylation of 6-BPin-Fpi during the TFA treatment, the pinacol-boronate was converted to a MIDA-boronate, and fluorocyclization was carried out to afford the building block 6-BMIDA-Fpi (188). The complete synthetic pathway to 188 is summarized in Figure 3.38.  Figure 3.38 Complete synthetic route to obtain 6-BMIDA-Fpi (188) from Nα-Boc-L-Trp-OMe.     125  3.5 Experimental Section  3.5.1 Materials and Methods General: All reactions were performed under argon atmosphere in flame-dried glassware and dried solvents at room temperature, unless otherwise stated. Controlled temperature reactions were performed using a mineral oil bath and a temperature controlled hot plate (IKA Ceramag Midi). Temperatures below room temperature were achieved in an ice/water bath (0°C), dry ice/ethylene glycol bath (-20°C), dry ice/ethanol/ethylene glycol bath (-20°C to -75°C) and dry ice/acetone bath (-78°C). Solvents were removed under reduced pressure using a Büchi rotary evaporator. Anhydrous solvents were prepared by distillation under nitrogen atmosphere. Ethers were distilled from sodium in the presence of benzophenone as indicator. Triethylamine, dichloromethane and hexanes were distilled over calcium hydride. Methanol was distilled from magnesium. DMSO and DMF were dried over 4Å molecular sieves under argon atmosphere. All reagents and solvents were purchased from Sigma-Aldrich, Alfa Aesar, Acros Organics, Matrix Scientifics, Oakwood Chemicals, Ontario Chemicals or TCI America, unless otherwise stated. Authentic α-amanitin was purchased from Sigma-Aldrich. Thin-layer chromatography (TLC) was performed using silica gel 60 F254 precoated aluminum plates (EM Science). Detection of TLC spots was performed using UV lamp at 254 nm, or by staining with p-anisaldehyde, potassium permanganate, ninhydrin or 2,4-dinitrophenylhydrazine, prepared according to literature procedures. Flash column chromatography purifications were performed using silica gel 60 (230-400 mesh, Silicycle, Quebec). Low-resolution mass spectra (LRMS ESI) in electrospray ionization (ESI) mode were obtained from a Bruker Esquire spectrometer. Proton (1H-NMR), carbon (13C-NMR), 126  boron (11B-NMR) and fluorine (19F-NMR) spectra were obtained using Bruker AV-300 (300 MHz) and AV-400inv (400 MHz). The X-Ray crystallographic measurements were made on a Bruker APEX DUO diffractometer with a TRIUMPH curved-crystal monochromator with Mo-Kα radiation.  3.5.2 Experimental Procedures  Methyl Nα-(tert-butoxycarbonyl)-1-(triisopropylsilyl)-L-tryptophanate (151) To a solution of Nα-Boc-L-Trp-OMe (500 mg, 1.57 mmol) in dry THF (9 mL) cooled to -78°C under argon was added dropwise a 0.5 M solution of KHMDS in toluene (4.08 mL, 2.04 mmol). After stirring at -78°C for 1 h, TIPSCl (0.34 mL, 1.57 mmol) in THF (0.4 mL) was added dropwise, and the solution was allowed to warm up to RT. After stirring at RT for 1 h, the solution was quenched with H2O, THF was evaporated under reduced pressure, and the aqueous solution was extracted with diethylether (3 x 10 mL). The organic layers were combined and washed with brine, dried over Na2SO4, filtered and the solvent was removed under reduced pressure. The crude product was purified by flash column chromatography (EtOAc/hexanes 5:95) to yield the desired product as a white solid (685 mg, 92%). TLC (EtOAc:hexanes 15:85 v/v): Rf = 0.32 1H NMR (300 MHz, CD2Cl2): δ 7.56 – 7.44 (m, 2H), 7.18 – 7.06 (m, 2H), 7.04 (s, 1H), 5.05 (d, J = 8.2 Hz, 1H), 4.59 (m, 1H), 3.64 (s, 3H), 3.24 (m, 2H), 1.68 (h, J = 7.5 Hz, 3H), 1.41 (s, 9H), 1.15 (s, 9H), 1.13 (s, 9H) 127  13C NMR (75 MHz, CD2Cl2): δ 173.1, 155.5, 141.8, 131.6, 130.5, 122.1, 120.1, 119.1, 114.6, 112.7, 79.9, 52.6, 28.6, 18.4, 13.4. HRMS-ESI (m/z): [M+Na]+ calcd. for C26H42N2O4NaSi, 497.2812; found 497.2811.   Methyl (S)-2-((tert-butoxycarbonyl)amino)-3-(6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1-(triisopropylsilyl)-1H-indol-3-yl)propanoate (152) Ir[(cod)OMe]2 (6.6 mg, 5 mol %), 1,10-phenanthroline (3.8 mg, 10 mol %) and B2Pin2 (204 mg, 4 eq) were placed in a dry pressure vial under argon. Hexanes (1 mL) were added, followed by the addition of HBPin (7.2 µL, 25 mol %). A solution of 151 (95 mg, 0.2 mmol, 1 eq) in hexanes (0.64 mL) was then added to the vial. The vial was sealed and stirred for 24 h at 80°C. During the process, the dark green mixture turned black. After 24 h, the reaction mixture was diluted with EtOAc and washed with saturated aqueous NaHCO3 solution (1 x 2 mL). Layers were separated, and aqueous layer was extracted with EtOAc (3 x 3 mL). Organic layers were combined, washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. Crude reaction mixture was purified on silica gel using flash column chromatography (EtOAc/hexanes 10:90) to yield compound 1 as an off-white foam (60 mg, 50%). The purified product contained an inseparable 8:1 mixture of 6-BPin and 5-BPin-L-tryptophan product (separation was finally achieved following purification of 187). The 13C-NMR chemical shifts below correspond to the reported mixture of 5- and 6-BPin compounds. TLC (EtOAc:hexanes 30:70 v/v): Rf = 0.48 128  1H NMR (300 MHz, CDCl3): δ 7.91 (s, 1H), 7.63 – 7.41 (m, 2H), 7.07 (s, 1H), 5.04 (d, J = 8.1 Hz, 1H), 4.63 (m, 1H), 3.61 (s, 3H), 3.27 (m, 2H), 1.68 (m, 3H), 1.42 (d, J = 3.7 Hz, 9H), 1.34 (s, 12H), 1.17 – 1.05 (m, 18H) 13C NMR (75 MHz, CDCl3): δ 172.7, 171.3, 155.2, 143.5, 140.9, 133.6, 131.4, 130.0, 127.9, 126.4, 125.7, 120.9, 118.0, 113.5, 112.2, 84.9, 79.8, 60.5, 54.1, 52.3, 28.3, 25.9, 21.2, 18.3, 14.3, 12.9 11B NMR (96 MHz, CD2Cl2): δ 29.9 HRMS-ESI (m/z): [M+H]+ calcd. for C32H54N2O6Si10B, 601.3844; found 601.3841.   Methyl (S)-2-((tert-butoxycarbonyl)amino)-3-(6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-indol-3-yl)propanoate (160) Tetra-Butyl ammonium fluoride (TBAF, 1 M solution in THF, 3.43 mL, 3.43 mmol) was added to a solution of 152 (1.90 g, 3.17 mmol) in THF (37 mL) at RT under argon. The reaction mixture was stirred at RT for 30 minutes at which point it was quenched by the addition of a saturated aqueous solution of NaHCO3 (10 mL). Layers were separated, the aqueous layer was extracted with EtOAc (3 x 20 mL). Organic layers were combined, washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude product was purified by silica gel using flash column chromatography (EtOAc/hexanes 25:75 to 30:70) to yield 160 as a white foamy solid (1.04 g, 74%). TLC (EtOAc:hexanes 40:60 v/v): Rf = 0.24 129  1H NMR (300 MHz, CDCl3): δ 8.24 (s, 1H), 7.85 (s, 1H), 7.55 (s, 1H), 7.05 (d, J = 2.5 Hz, 1H), 5.07 (d, J = 8.2 Hz, 1H), 4.64 (m, 1H), 3.69 (s, 3H), 3.29 (m, 2H), 1.42 (d, J = 4.1 Hz, 9H), 1.36 (s, 12H) 13C NMR (75 MHz, Chloroform-d): (contains 5-BPin isomer) δ 172.8, 155.4, 136.0, 130.2, 125.6, 124.4, 118.3, 118.2, 110.5, 83.7, 83.6, 80.0, 68.1, 54.3, 52.4, 28.5, 28.1, 25.0. 11B NMR (96 MHz, CD2Cl2): δ 31.8 HRMS-ESI (m/z): [M+Na]+ cald. for C23H33N2O6Na10B, 466.2366; found 466.2371.   Methyl 2-((tert-butoxycarbonyl)amino)-3-(6-hydroxy-1H-indol-3-yl)propanoate, Nα-Boc -6-hydroxy-L-Trp-OMe (161) To a solution of Nα-Boc-6-BPin-L-Trp-OMe 160 (150 mg, 0.34 mmol) in THF/H2O (3.5 mL:3.5 mL) was added sodium perborate i.e. NaBO3.4H2O (129 mg, 0.85 mmol) at room temperature. The reaction mixture was vigorously stirred for 1.5 h and was then diluted with H2O (3 mL) and extracted with EtOAc (3 x 10 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. Within hours of work-up, the crude reaction began to turn yellow. In addition, significant coloration was observed once the crude contents were loaded on silica. The crude residue was purified by silica gel column (EtOAc/hex 20:80 to 30:70 to 40:60, gradient) to yield the product 161 as a white solid (50 mg, 45%). TLC (EtOAc:hex 50:50 v/v): Rf = 0.22 130  1H NMR (300 MHz, Methanol-d4) δ 7.31 (d, J = 8.5 Hz, 1H), 6.88 (s, 1H), 6.75 (d, J = 1.8 Hz, 1H), 6.60 (dd, J = 8.5, 1.9 Hz, 1H), 4.45 – 4.34 (m, 1H), 3.65 (s, 3H), 3.18 (dd, J = 14.5, 5.7 Hz, 1H), 3.06 (dd, J = 14.5, 7.5 Hz, 1H), 1.40 (s, 9H). 13C NMR (75 MHz, Methanol-d4) δ 174.7, 157.7, 154.2, 139.0, 122.8, 119.6, 110.7, 110.1, 97.6, 80.6, 56.1, 52.6, 28.7. LRMS ESI (m/z) calculated for C17H22N2O5Na [M+Na]+ 357.1; found 357.1   3,3-Dimethyldioxirane (DMDO) To a stirred solution of NaHCO3 (58 g, 0.69 mol) in H2O (250 mL) was added acetone (190 mL, 2.56 mol). The resulting mixture was cooled down to 0°C in an ice bath, and Oxone® (120 g, 0.39 mol KHSO5) was added in six portions with 3-minute intervals to avoid CO2 pressure build-up. The reaction mixture was distilled under reduced pressure (20-30 mmHg) and the distillate was collected at -78°C. Collection was stopped when the distillate was not yellow in color anymore. The distilled solution was stored at -20°C over anhydrous Na2SO4 with a maximum shelf-life of two weeks. To assess the concentration of the DMDO solution in acetone, a small aliquot of the distillate was directly subjected to 1H-NMR and the concentration was calculated based on the area under the 13C satellite peaks of acetone’s proton peak. Generally, the distilled DMDO solutions had a concentration of 60-70 mM.   131   Oxidation of Nα-Boc-6-OH-L-Trp-OMe (161) to 162 and 163. Two methods were performed in an attempt to achieve the corresponding Hpi (162): 1) Oxidation with DMDO formed in situ: Starting material (5 mg, 0.015 mmol) was dissolved in acetone/H2O (0.2 mL:0.2 mL). To the resulting mixture was added NaHCO3 (25 mg, 20 eq) and the reaction was cooled to 4°C in an ice/water bath. A solution of Oxone® (40 mM) in H2O (0.38 mL, 1.02 eq) was added to the reaction mixture in 3 portions with 45-minute intervals (portions: 0.13 mL, 0.13 mL, 0.12 mL). The reaction was diluted with H2O (1 mL) and extracted with EtOAc (3 x 2 mL). Combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure to yield the crude product as a red solid (refer to part 2 for purification). 2) Oxidation with freshly distilled DMDO: Starting material (20 mg, 0.06 mmol) was dissolved in DCM (3.2 mL) and the resulting mixture was cooled to -78°C in an acetone/dry ice bath. To the resulting mixture was added 1.2 mL (1.13 eq) of a freshly distilled DMDO solution in acetone (0.061 M) in 5 portions with 5-minutes intervals (0.24 mL each portion). More DMDO could be added until the starting material was consumed. At this point, the solvent was evaporated under reduced pressure to yield the crude product as a red solid, which was purified by silica gel chromatography to yield the respective 3a-hydroxy-indoline i.e. 3a,6-dihydroxy-1,2,3,3a,8,8a-hexahydropyrrolo-[2,3-b]indole-Nα-Boc-2-carboxy methyl ester (2 mg, 10% isolated yield) and the oxindole i.e. 6-132  hydrox-2-oxo-Nα-Boc-L-Trp-OMe (2.7 mg, 13% isolated yield), along with several unidentified products. Note: the 3a-hydroxy-indoline product was very unstable and air-sensitive. Storage for more than a few hours resulted in air oxidation and color change of the material, as shown by the appearance of several new spots on TLC.  TLC (EtOAc:hex 70:30 v/v): Rf = 0.42 1H NMR (400 MHz, Methylene Chloride-d2) δ 7.09 (dd, J = 8.1, 4.0 Hz, 1H), 7.02 (dd, J = 8.1, 4.1 Hz, 0.5H), 6.24 (dt, J = 8.1, 2.7 Hz, 1H), 6.18 (dd, J = 8.2, 2.0 Hz, 0.5H), 6.13 – 6.08 (m, 1H), 6.07 – 6.03 (m, 0.4H), 5.38 (s, 0.5H), 5.34 (s, 1H), 5.16 – 5.13 (m, 0.5H), 4.60 – 4.55 (m, 0.4H), 4.49 (dd, J = 8.9, 2.2 Hz, 0.4H), 4.29 (dd, J = 8.4, 4.6 Hz, 0.5H), 4.21 (dd, J = 8.3, 5.7 Hz, 1H), 3.76 – 3.74 (m, 3H), 2.66 – 2.55 (m, 1H), 2.53 – 2.35 (m, 2H), 1.51 (d, J = 8.1 Hz, 5H), 1.40 (d, J = 3.8 Hz, 9H). LRMS ESI (m/z) calculated for C17H22N2O6Na [M+Na]+ 373.1; found 373.1  TLC (EtOAc:hex 70:30 v/v): Rf = 0.26 1H NMR (400 MHz, Methylene Chloride-d2) δ 8.52 (s, 0.5H), 8.12 (s, 0.5H), 7.21 (d, J = 8.1 Hz, 0.5H), 7.04 (d, J = 8.1 Hz, 0.5H), 6.58 – 6.46 (m, 2H), 5.54 (dd, J = 13.2, 8.6 Hz, 1H), 4.75 – 4.60 (m, 1H), 3.68 (s, 3H), 3.48 (q, J = 7.9, 7.2 Hz, 1H), 2.32 – 2.25 (m, 1H), 2.24 – 2.17 (m, 1H), 1.42 (s, 9H). LRMS ESI (m/z) calculated for C17H22N2O6Na [M+Na]+ 373.1; found 373.1 133   (S)-2-((tert-Butoxycarbonyl)amino)-3-(6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-indol-3-yl)propanoic acid (181) Starting material 160 (100 mg, 0.23 mmol) in THF (4 mL) were placed in a round-bottom flask. The flask was cooled to 4°C in an ice/water bath. A solution of LiOH (10.56 mg, 0.44 mmol) in H2O (4 mL) was added dropwise to the reaction flask and the reaction was stirred at 4°C for 4 h or until completion. The pH of the reaction was brought to 2 by adding 1M aq. HCl, and the resulting mixture was extracted with EtOAc (3 x 15 mL). The organic layers were combined, washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure to yield the crude product as a yellow oil (82 mg). The crude product was used in the next step without further purification.   (S)-2-((tert-Butoxycarbonyl)amino)-3-(6-(6-methyl-4,8-dioxo-1,3,6,2-dioxazaborocan-2-yl)-1H-indol-3-yl)propanoic acid (187) Starting material 181 (2.1 g) and N-methyliminodiacetic acid (MIDA, 3.78 g, 5 eq) were dissolved in DMSO (30 mL) in a round bottom flask. The resulting mixture was stirred at 110°C in an oil bath for 16 hours, at which point the initial white slurry turns into a bright 134  orange homogenous solution. The reaction mixture was cooled to room temperature and H2O (30 mL) was added to the mixture. Ethyl acetate (30 mL) was added to the resulting solution and the layers were separated. The aqueous layer was extracted with EtOAc (3 x 15 mL), and the combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The residual DMSO was removed in vacuo before silica gel purification. The crude product was purified using flash column chromatography with silica gel to afford 187 as a white solid (633 mg, 25% over two steps). Notably this yield was low since we did not optimize this reaction; optimization is possible through recycling the starting material. TLC (MeOH:CH2Cl2:HOAc 10:90:1 v/v/v): Rf = 0.23 1H NMR (300 MHz, CD3OD): δ 10.33 (s, 1H), 7.62 (d, J = 8.0 Hz, 1H), 7.52 (s, 1H), 7.14 (d, J = 8.9 Hz, 2H), 4.41 (dd, J = 7.8, 5.1 Hz, 1H), 4.12 (dd, J = 63, 17.1 Hz, 4H), 3.38 – 3.25 (m, 1H), 3.12 (dd, J = 14.7, 7.9 Hz, 1H), 2.51 (s, 3H), 1.40 (s, 3H), 1.38 (s, 6H) 13C NMR (75 MHz, CD3OD): δ 175.9, 171.5, 157.8, 138.1, 129.9, 126.1, 125.2, 123.5, 119.2, 116.6, 111.1, 80.5, 62.8, 58.4, 55.8, 30.9, 28.7, 28.3 11B NMR (96 MHz, CD3OD): δ 14.3 HRMS-ESI (m/z): [M+Na]+ calcd. for C21H2611BN3O8Na, 482.1711; found 482.1704.     135   (2S)-1-(tert-Butoxycarbonyl)-3a-fluoro-6-(6-methyl-4,8-dioxo-1,3,6,2-dioxazaboron-2-yl) -1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole-2-carboxylic acid (188) N-Fluorocollidinium triflate (FP-T300, 182 mg, 0.65 mmol) was placed under argon in a dry flask equipped with a stir bar and a condenser. A solution of 187 (150 mg, 0.33 mmol) in THF (40 mL) was added to the flask under argon. The reaction mixture was heated to 63°C under near-reflux conditions in an oil bath for 4 h. After 4 h, TLC analysis indicated complete conversion of starting material to product with no observed by-products. Hence, the solvent was removed under reduced pressure to yield 330 mg of the crude product mixture as a light brown residue containing both diastereomers. Following silica chromatography, the isolated yield of both diastereomers of 188 was 85% and their separation was not achieved. Notably, during the purification, decomposition was often observed resulting in much lower yields, likely due to action by the silica leading to dehydrofluorination and or other decompositions. Hence, compound 188, as a pair of diastereomers, was used in the next step without further purification. TLC (MeOH:CH2Cl2:HOAc 10:90:1 v/v/v): Rf = 0.33 1H NMR (300 MHz, CD3OD): δ 7.34 (m, 1H), 7.06 (s, 0H), 6.97 (m, 1H), 6.89 (d, J = 16.1 Hz, 1H), 5.61 (dd, J = 22.9, 4.0 Hz, 1H), 4.43 – 4.17 (m, 3H), 4.14 – 3.96 (m, 2H), 2.87 – 2.66 (m, 2H), 2.60 – 2.56 (m, 3H), 1.55 (s, 3H), 1.45 (s, 6H) 13C NMR (75 MHz, CD3OD): δ 176.2, 175.2, 170.1, 159.2, 156.1, 154.4, 125.8, 123.9, 119.7, 93.3, 82.1, 80.0, 61.1, 59.6, 52.6, 43.0, 28.7, 28.5. 136  19F NMR (282 MHz, CD3OD): δ -138.8 (q, J = 22.8 Hz), -139.8 (q, J = 20.1 Hz) 11B NMR (96 MHz, CD3OD): δ 10.7 HRMS-ESI (m/z): [M-H]- calcd. for C21H2411BN3O8F, 476.1640; found 476.1631.                137  Chapter 4: The Total Synthesis of α-Amanitin  4.1 Introduction Prior to my work, all efforts towards a total synthesis of α-amanitin had failed, primarily due to the synthetic challenges inflicted by the lack of a route to (2S,3R,4R)-dihydroxyisoleucine (DHIle) and a route to 6-hydroxy-L-tryptathionine (6-OH-Ttn) (vide supra). Furthermore, attempts to oxidize the thioether of the tryptathionine crosslink to the corresponding (R)-sulfoxide found in the natural product had resulted in poor diastereoselectivities and low yields (vide infra). As discussed in Chapter 1, over the past number of decades, several analogues of α-amanitin have been synthesized, exhibiting lower toxicities in contrast to the native toxin. I begin this chapter with a brief overview of recent attempts to synthesize select amanitin analogues and the most common strategies for their solid-phase and solution-phase synthesis. The first total synthesis of α-amanitin will then be disclosed, enabled for the first time by my syntheses of DHIle and 6-OH-Ttn followed by an asymmetric sulfoxidation of the tryptathionine-thioether to afford the (R)-sulfoxide of α-amanitin.  4.1.1 Previous Syntheses of Amanitin Analogues Exploring the existing literature on α-amanitin, one may appreciate two major waves of effort towards the synthesis of amanitin analogues may be distinguished. Starting in the early 80s, the first wave focused on the synthesis of several amanitin derivatives containing isoleucine or other amino acids at position-3 rather than DHIle3 due to a lack of a stereoselective route for the synthesis of this amino acid. Additionally, all these synthetic 138  analogues contained tryptophan at position-4, hence possessing a tryptathionine linkage with no hydroxyl group at C-6. Zanotti and Wieland, considered the pioneers of these syntheses, carried out toxicity assays and SAR studies for a broad range of amanitin derivatives.28–30,45,55 In 1981, Zanotti, Birr and Wieland reported the first-ever synthesis of amaninamide (191), an analogue of α-amanitin bearing Ile at position-3 and Trp at position-4. Furthermore, they were able to oxidize the Ttn-thioether to a separable mixture of (R) and (S)-sulfoxides, with the latter diastereomer prevailing (Figure 4.1).30  Figure 4.1 Synthesis of amaninamide diastereomers (191) (Zanotti et al.). Ile3, Trp4 and Ttn-sulfoxide are shown in blue.  The synthetic strategy shown in Figure 4.1 turned into the most widely used strategy for the solution-phase synthesis of amanitin analogues until the use of solid-phase synthesis grew to be the predominant route for peptide synthesis (vide infra).129 Within the next 10 years, Wieland, Zanotti and others reported the synthesis and toxicity assessments of an extensive range of other amanitin analogues employing the method above. These analogues mainly varied in the residues installed at positions 2 and 3 and the oxidation state of the 139  tryptathionine’s sulfur atom.7,26,28,29 A list summarizing these derivatives is shown in Figure 4.2.  Figure 4.2 Synthetic and derivatized bicyclic amanitin analogues (adapted with permission from Wieland et al., 1981).26 All analogues contain a bicyclic structure; in each analogue, color-coded residues that are different than the natural product are shown. Position-4 contained Trp unless stated otherwise. Inhibitory effects were measured against RNAP II from calf thymus and are relative to the Ki of α-amanitin (I). “SO” represents (R)-sulfoxide, “OS” represents (S)-sulfoxide, and OSO shows the sulfone of tryptathionine. hSer=homoserine, hVal=γ-hydroxyvaline.  140  4.1.1.1 Synthesis of Amanitin Analogues by Perrin Lab More recently, Perrin and co-workers commenced the second wave of efforts towards the synthesis of amanitin analogues via a solid-phase strategy. In 2008, May and Perrin developed the synthesis of Pro2-Ile3-S-deoxy-amaninamide (entry II, Figure 4.2) employing an Fmoc-based solid-phase peptide synthesis (SPPS).27 In this synthesis, Fmoc-Ile-OH was loaded on 2-chlorotrityl chloride (CTC) resin and the linear sequence was prepared from the C-terminus to the N-terminus (with the order of Ile3, Pro2, Asn1, Cys8, Gly7, Ile6, Gly5 and Hpi4). The resulting linear peptide (192) was then cleaved from the resin using TFA, at which point a global deprotection of acid-labile protecting groups followed by tryptathionylation was achieved to obtain the monocyclic octapeptide (193). Finally, the monocycle was macrolactamized using PyBOP to afford the final amanitin analogue. However, during the macrolactamization step, a significant amount of a by-product consisting of an epimer at Cα(Ile3) was obtained in addition to the desired amanitin analogue (194). This by-product (195) was shown to contain D-allo-Ile at position-3 (Figure 4.3) by X-ray diffraction analysis. Not only is epimerization quite common with Ile due to the steric congestion that thwarts ready coupling, it is appreciated that the presence of a D-amino acid favors macrocyclization.130  141   Figure 4.3 Synthesis of Pro2-Ile3-S-deoxo-amaninamide (194) and Pro2-(D-allo-Ile)3-S-deoxo-amaninamide (195) (May and Perrin).  Appreciating the fact that macrolactamization between Ile-OH and Trp-NH2 would lead to the epimerization of Cα(Ile3), Zhao and Perrin, in 2015, reported the synthesis of a cytotoxic analogue of amanitin for biorthogonal conjugation utilizing a different solid-phase strategy whereby protected proline was anchored at the C-terminus.25 Toward this end, an N-propargyl-asparagine, in lieu of a protected asparagine residue was incorporated at position-1 for post-synthetic bioconjugation. The propargyl group would later undergo a Cu-catalyzed cycloaddition (CuAAC)131 with an azide-modified biomolecule to generate a toxic amanitin bioconjugate. Furthermore, the synthetic analogue in this work carried DHIle at position-3. However, the DHIle employed in this synthesis was produced as a mixture of four diastereomers, necessitating the late stage isolation of the correct diastereomer on the final peptide (see Chapter 2) (Figure 4.4). Whereas all four synthetic amanitin analogues were tested, a single analogue (199) was tested on CHO (Chinese hamster ovary) cells and was shown to exhibit a slightly higher IC50 than the natural product (2 µM vs. 0.5 µM), hence verifying the cytotoxicity of the synthetic analogue to be near-native. 142   Figure 4.4 Synthesis of a cytotoxic amanitin analogue for biorthogonal conjugation (Zhao and Perrin). Diastereomers of DHIle are shown in red.  4.2 Results and Discussion  4.2.1 Synthesis of the Monocyclic Heptapeptide Exploiting the earlier strategies for the synthesis of amanitin derivatives, I opted to pursue the solid-phase method employed by Zhao and Perrin (Figure 4.4). To this end, trans-hydroxyproline, Hyp(OtBu), was loaded on the CTC resin, and the remaining amino acids were added to the sequence in the following order: Asn(Trt)1, Cys(Trt)8, Gly7, Ile6, Gly5 and 6-BMIDA-Fpi4 (188). The synthetic scheme for preparation of linear heptapeptide 201 is shown in Figure 4.5. 143   Figure 4.5 Synthesis of the linear hexapeptide 201 on solid phase. 6-BMIDA-Fpi (188) is shown in blue. Numbering of the residues is shown in the dashed box. Coupling conditions: AA (5 eq.), HBTU (5 eq.), HOBt.H2O (5 eq.), DIPEA (pH 8), DMF, RT, 2h.  Next, the linear resin-bound heptapeptide (201) was treated with TFA. I noticed that Fpi underwent the Savige-Fontana reaction more rapidly than Hpi (reaction time: 30-45 minutes when using Fpi as opposed to several hours for Hpi) and, unlike Hpi, absolute TFA was not required to perform the reaction; in fact, a 1:1 mixture of TFA/DCM was capable of inducing tryptathionylation. Under the Savige-Fontana conditions, a global deprotection of all the acid-labile protecting groups (Trt, tBu, Boc) and cleavage of the peptide from the chlorotrityl-resin were also accomplished, affording a monocyclic heptapeptide (202). However, following the TFA treatment and HPLC purification of monocyclic heptapeptide 202 using an aqueous solution of formic acid (0.1% FA in MeCN/H2O), varying amounts of hydrolysis of the MIDA-boronate were observed, converting 30-50% of the BMIDA-product to the corresponding boronic acid (203) (Figure 4.6). 144   Figure 4.6 Synthesis of monocyclic heptapeptides 202 and 203 from the resin-bound linear heptapeptide (201). Treatment with TFA/DCM resulted in the cleavage from the resin, deprotection of Trt, Boc and tBu, and tryptathionylation.  Previously, I observed the deborylation of 6-BPin-tryptathionine under Savige-Fontana conditions, possibly due to the hydrolysis of the pinacol-ester to boronic acid followed by a protodeborylation process (see Chapter 3). To avoid the deborylation of 203, the isolated mixture of 202/203 was briefly treated with 0.5 M aq. KOH to deprotect MIDA from the remaining 202. Once fully deprotected, the resulting boronic acid was oxidatively deborylated  using mCPBA to give the corresponding phenol. However, when the 202/203 mixture was treated with KOH under strongly basic conditions (pH 10-11), I noticed the formation of a substantial amount of a by-product that was 2 mass units lighter than the desired product (Figure 4.7). At the time of the experiment, this by-product was not fully investigated since the desired 6-OH-Ttn was the major product although the formation of this by-product and its putative structure are discussed in subsequent sections. Prior to describing the remaining steps 145  in my synthesis, an interesting trend observed in the UV maximum absorbance wavelengths of 6-substituted tryptathionines will be discussed.  Figure 4.7 MIDA deprotection of 202 to obtain 203, followed by oxidation of 203 to 204. An unknown by-product (205) was formed as the minor product in this reaction.  4.2.1.1 UV-Vis Absorbance Trends in 6-Substituted Tryptathionines In the course of the synthesis of the monocyclic heptapeptides containing 6-BMIDA-Ttn (202), 6-B(OH)2-Ttn (203) or 6-OH-Ttn (204) residues, I noticed a trend in the maximum absorbance wavelengths (λmax) of the corresponding tryptathionine crosslinks. As previously discussed, an unsubstituted tryptathionine shows a characteristic maximum absorbance at 290 nm (Figure 3.12). Once BMIDA, BPin and B(OH)2 groups were introduced at C-6, the maximum absorbance wavelength was increased by 5 nm, reaching 295 nm. Following the oxidation of the boronic acid to hydroxyl, λmax was further increased to 304 nm (Figure 4.8). This trend may be attributed to the electronic properties of the substituent at C-6. Boronates and boronic acids on the indole exhibited a slightly red-shifted λmax (ca. 5 nm), while the hydroxyl group, a highly electron-donating functionality, increased the λmax by another 9 nm to reach 304 nm. In each case, the observed increase in λmax is likely due to the elongation of the conjugated system of the indole ring resulting from the conjugation of the p-orbitals of the 146  C6-substituent with the delocalized π-system of the indole (Figure 4.9). The observation that introducing a tetravalent MIDA-boronate at C-6 results in a similar increase in the λmax of the corresponding substituted Ttn to that of bivalent 6-BPin or 6-B(OH)2 might be attributed to the slow cleavage of the B-NMIDA bond under slightly acidic conditions.132 In the presence of aqueous acid (e.g. 0.1% formic acid in the HPLC buffer), the nitrogen atom of the MIDA group may be protonated and slowly cleaved from the boron, resulting in a trivalent boronate species analogous to BPin or B(OH)2, hence leading to a similar increase in the λmax of the 6-BMIDA-tryptathionine crosslink.  Figure 4.8 UV absorbance curves for an unsubstituted Ttn (blue), 6-B(OH)2- and 6-BPin-Ttn (maroon), 6-BMIDA-Ttn (gray) and 6-OH-Ttn (orange).   Figure 4.9 Schematic representation of the extended conjugated π-system of 6-boronate indole (left) and 6-hydroxy indole (right). The p-orbital of the 6-substituent is shown in green.  147  4.2.2 Synthesis of the Monocyclic Octapeptide Following the oxidative deborylation of 6-BMIDA-Ttn to afford heptapeptide 204, its coupling with the activated NHS-ester of the suitably protected DHIle (139) was pursued. Hence, 204 and 139 were dissolved in the minimum amount of DMF, the pH of the mixture was adjusted to 8 by addition of DIPEA, and the coupling reaction was then allowed to proceed at room temperature. Upon completion (generally 48-72h), diethylamine was added to the reaction mixture to achieve deprotection of the Fmoc. After 2 hours, the solvent and volatile bases were removed in vacuo and the remaining residue was re-suspended in DMF. Excess amounts of a 1 M solution of TBAF in THF were added to the reaction mixture to deprotect the TBS groups. Following acidification of the reaction contents to pH 3 by addition of 1 M aq. HCl, to my great surprise, the desired product was not detected in the crude product mixture. Instead, an unknown compound (209) was formed that was 2 mass units lighter than the expected product. While more experiments towards the identification of this unexpected by-product are underway, I describe my hypotheses on the mechanism of formation and the preliminary characterization of this by-product in the upcoming section (vide infra). Suspecting that the formation of by-product 209 had been triggered by the strongly basic conditions caused by the transient formation of alkoxides during TBS removal, I modified the TBS deprotection conditions by addition of acetic acid to the reaction mixture to adjust the pH to 5. To my delight, at this slightly acidic pH, complete deprotection of TBS was achieved, and no detectable amounts of 209 were observed in the crude product mixture (Figure 4.10). 148   Figure 4.10 Coupling of the suitably protected DHIle (139) to the monocyclic heptapeptide (204) followed by Fmoc and TBS deprotection in one pot. Using excess TBAF to deprotect the TBS groups afforded a by-product (209) with the molecular weight 918.3 (desired mass - 2), while adjusting the pH of the TBS deprotection reaction to 5 by adding acetic acid exclusively furnished the desired product (208).  Inspired by the positive results obtained from the addition of acetic acid, I revisited the reaction conditions for the oxidative deborylation of 202 that afforded 6-OH-Ttn (204) (Figure 4.7). My initial conditions for this transformation involved the deprotection of MIDA at pH values of 10-11 achieved by addition of a 0.5 M aqueous KOH solution, resulting in the appearance of minor amounts of by-product 205. To suppress the formation of this by-product, the KOH solution was added to 202 until a pH value of 8 was achieved whereupon the addition was halted to avoid increasing pH values to 10 or higher. As expected, under these less basic conditions, formation of the by-product (205) was not observed (Figure 4.11). 149   Figure 4.11 Revisiting the reaction conditions for conversion of 202 to 204. pH of the reaction mixture was maintained at 8 to avoid strongly basic conditions leading to the formation of a [M-2] by-product (205).  With the fully deprotected monocyclic octapeptide (208) in hand, I turned my attention to the [M-2] by-product (209, called the “oxidation by-product” in this text), and preliminary experiments were carried out to determine its identity.  4.2.3 The Base-promoted Oxidation By-product  4.2.3.1 Mechanism of Formation By avoiding strongly basic conditions, I was able to minimize the formation of oxidation by-products 205 and 209. Therefore, I hypothesized that the deprotonation of the hydroxyl group of the 6-OH-Ttn residue at pH values above 10, along with the presence of an oxygen source, are responsible for the occurrence of this side reaction. Based on my hypothesis, once the 6-OH is deprotonated (210), a single-electron transfer (SET) event in the presence of dissolved molecular oxygen (O2) leads to the formation of a radical oxygen species (211). Following another SET from 211 to a previously formed superoxide species, a carbocation is likely to be formed (212) which is subsequently converted to a quinone-methide species (213). 150  At this stage, the formation of a thioxo-indole product (214) is imminent following the elimination of the Hα(Cys8) leading to a concurrent cleavage of the tryptathionine crosslink (Figure 4.12). However, either 213 or 214 could theoretically represent the chemical structure of the oxidation by-product.  Figure 4.12 Proposed mechanism for the formation of the oxidation by-product. Quinone-methide 213 can undergo the elimination of Hα on the adjacent Cys8 residue, leading to the formation of a thioxo-indole core (214).  4.2.3.2 Preliminary Characterization of the Oxidation By-product To further investigate this by-product, compound 209 was purified by HPLC. Several HPLC peaks with near-identical retention times correlating to compounds with the molecular weight of the unknown compound (209) (MW 918.3) and a λmax of 336 nm were isolated as a mixture (see Appendix C). The resulting mixture was subjected to macrolactamization conditions using HATU as the coupling reagent. Following the macrolactamization, one major product was identified that contained a compound possessing the expected molecular weight of the macrolactamized oxidation by-product. 151  At the time of preparing this thesis, it appears that the following observations are consistent with the oxidation by-product bearing a thioxo-indole moiety (214) that forms concomitantly with C-S fission of the tryptathionine crosslink: i) presence of multiple HPLC peaks prior to macrolactamization, representing compounds possessing similar molecular weights. This observation is consistent with HPLC traces of linear peptides exhibiting various conformational states. ii) The maximum absorbance wavelength for the obtained product, whether prior to or following the macrolactamization, shows a substantial increase (+32 nm) from the expected 304 nm, reaching 336 nm. According to the UV trend discussed earlier in this chapter (Figure 4.8), it can be surmised that the observed λmax of 336 nm likely represents an extended conjugated system consistent with a thioxo-indole moiety (Figure 4.13).  Figure 4.13 Proposed structure of the base-promoted oxidation products following preliminary characterizations. Thioxo-indole moiety is shown in red, the dehydroalanine resulting from the Hα-elimination of Cys8 is shown in blue. Although the oxidant is currently unknown, it is suspected to be dissolved molecular oxygen, O2.  152  To assess the cytotoxicity of the oxidation by-product (216), an MTT viability assay against CHO cells was carried out in the presence of increasing concentrations of 216. In this assay, the commercially available α-amanitin was used as the standard. The concentration of the unknown by-product was measured using its UV absorbance assuming an analogous extinction coefficient to α-amanitin (12600 M-1cm-1).25 Although this assumption might be premature, it provided a starting point to obtain an approximate value for the toxicity of this intriguing product. The cell assay results are shown in Figure 4.14. While the toxicity for concentrations above 20µM were not measured for the oxidation by-product, it was evident that this compound exhibited a much higher IC50 than α-amanitin. At 20 µM, 68% of the cells treated with 216 still remained viable, while the same viability percentage was observed at ca. 0.65 µM for α-amanitin.  Figure 4.14 MTT cell viability assay results for commercially available α-amanitin and the oxidation by-product against CHO cells.  153  Based on these observations, it appears that the structure of the oxidation by-product is not closely related to α-amanitin yet interestingly, it possesses non-negligible toxicity on CHO cells consistent with perhaps some resemblance to amanitin. In fact, a hypothetical thioxo-indole structure (216) for this product could perfectly justify my observations. However, further experiments are required to confirm my conjecture. Possible reduction of the proposed thioxo-indole to a tryptathionine-like structure using a reducing agent (e.g. NaBH4) or addition of robust nucleophiles (e.g. azide or cyanide) to 216 are among those experiments that would support this assertion. Furthermore, obtaining an XRD structure or NMR spectra of this product will facilitate the elucidation of its structure.  4.2.4 Macrolactamization to S-Deoxy-α-Amanitin The next step in my synthesis of α-amanitin involved macrolactamization of the monocyclic octapeptide (208) to the bicyclic structure. In my first attempt, EDC·HCl and HOBt·H2O were employed as the coupling reagent and the additive, respectively. While decent yields were obtained, approximately 20% epimerization occurred during this transformation. To solve this problem, in my next effort, I used HATU used as the coupling reagent. It has been shown that coupling reagents based on 1-hydroxy-7-azabenzotriazole (HOAt) (e.g. HATU and TATU) provide faster and more efficient couplings resulting in lower degrees of epimerization as opposed to HOBt-based reagents.133 To my delight, bicyclic octapeptide 217 was generated with a greater overall efficiency. Moreover, no significant amounts of epimerization were detected (Figure 4.15). 154   Figure 4.15 Macrolactamization of the monocyclic octapeptide (208) to S-deoxy-α-amanitin (217) using HATU as the coupling reagent.  Hence, bicyclic octapeptide 217 (S-deoxy-α-amanitin) represented the first cytotoxic amanitin species that was synthesized en route to the total synthesis of α-amanitin. Thus, while my method is scalable, the production of 217 and subsequent amanitins was deliberately limited to sub-milligram quantities due to their exceptional toxicity.  4.2.5 Asymmetric Sulfoxidation of S-Deoxy-α-Amanitin Oxidation of the tryptathionine-thioether to the corresponding (R)-sulfoxide constituted the final challenge in the first total synthesis of α-amanitin. Prior to a discussion of my efforts towards this goal, herein I provide a concise overview of the significance of asymmetric sulfoxidation in the synthesis of natural products and general methods to achieve it, different oxidation states of the tryptathionine-thioether sulfur atom and how properties of amanitin could be impacted by them, and the literature precedence on this matter.  4.2.5.1 Asymmetric Sulfoxidation: A Brief Introduction Exploring the literature, one is able to extract multiple instances of drugs and biologically active molecules bearing a stereogenic sulfur center, usually in the form of a sulfoxide. Aside 155  from α-amanitin, there exist other pharmaceutical targets, e.g. omeprazole and rabeprazole (proton-pump inhibitors), bearing a chiral sulfoxide residue in their structure (Figure 4.16).  Figure 4.16 Structures of (S)-omeprazole (esomeprazole) and (R)-rabeprazole (dexrabeprazole). The (S)-isomer of omeprazole is metabolized more slowly and reproducibly than the (R)-enantiomer.134 The (R)-enantiomer of rabeprazole has been suggested to show higher therapeutic effects.135  Since the report of the first method for preparation of an enantioenriched sulfoxide in 1962,136 a wide variety of routes have been devised to achieve stereo- and enantioselectivity in a sulfoxidation reaction.137 Analogous to almost all asymmetric conversions, the stereochemical outcome of asymmetric sulfoxidation reactions may be governed either by the substrate (in the case of chiral substrates) or via a chiral reagent (in the case of achiral substrates). While reports on the control over the stereoselectivity of oxidation of a chiral thioether are limited, two major strategies are generally employed for oxidation of achiral thioethers: use of a chiral ligand chelated to a metal in the presence of a primary oxidant (e.g. H2O2 or halogen derivatives),138,139 or a metal-free organic oxidation system (e.g. BINOL-derived ligands, chiral oxaziridines). Various metals are utilized in metal-catalyzed sulfoxidation reactions, including titanium, vanadium, iron, manganese, molybdenum and copper to name a few. Generally, the metal of choice is accompanied by a multifunctional chiral ligand and the primary oxidant. Examples of common metal-based oxidizing systems are shown in Figure 4.17. 156   Figure 4.17 Examples of asymmetric sulfoxidation reactions. A) Titanium-based oxidation with (R,R)-DET as the chiral ligand and tBuOOH as the main oxidant.140 B) Vanadium-catalyzed oxidation with camphor-based Schiff bases (221) as chiral ligands and H2O2 as the oxidant.141 C) Iron-catalyzed oxidation with a Schiff base chiral ligand (224) and H2O2.142  Although less popular, metal-free organic oxidation systems have gained more attention in recent years.137 Examples of these methods include but are not limited to BINOL-derived phosphoric acids or iminophosphoric acids accompanied by H2O2,143,144 optically active flavin-based polymers,145 or chiral oxaziridines146 to carry out asymmetric sulfoxidations. Figure 4.18 demonstrates a few examples of the metal-free asymmetric sulfoxidations. While the two methods discussed above are the most frequently used asymmetric sulfoxidation strategies, enzymatic preparation of chiral sulfoxides via kinetic resolution has attracted interest as of late.147,148 Whether this process is performed using the whole-cell treatment or an isolated enzyme, its application is restricted due to the substrate specificity of enzymes. 157   Figure 4.18 Examples of metal-free asymmetric sulfoxidations. A) Oxidation with H2O2 using a BINOL-derived chiral phosphoric acid (226) as the ligand.144 B) Oxidation with a chiral oxaziridine (227).146  Asymmetric sulfoxidation reactions face problems that may reduce their appeal in the synthesis of natural products. Aside from the propensity of thioethers to over-oxidize to the corresponding sulfones, which is the case for both asymmetric and non-asymmetric sulfoxidations, inability to foresee the stereochemical outcome of the reaction proves to be a significant challenge. Moreover, in some cases, poor generalizability of the ee’s represents an additional drawback while performing an asymmetric sulfoxidation.137 Specifically, in the case of asymmetric sulfoxidation of chiral thioethers, undesirable interactions between the substrate and the chiral ligand or the oxidant may lead to unpredictability of the stereochemical outcome of the reaction.  4.2.5.2 Previous Efforts Towards the Sulfoxidation of S-Deoxy-α-Amanitin In 1974, Buku and co-workers investigated the synthesis of amanitin derivatives with various oxidation states of the tryptathionine sulfur bridge.149 Toward this end, the hydroxyl group of the 6-OH-Ttn residue of authentic α-amanitin was first methylated using 158  diazomethane to obtain O-methyl-α-amanitin (232). Although an explanation for this methylation is not discernable across numerous reports, I surmise that methylation may have been effected in order to avoid the untoward oxidation of the hydroxy-indole to a quinone-methide-like product that I observed previously. The methyl ether 232 was then deoxygenatively reduced to O-methyl-S-deoxy-α-amanitin (233) by heating with Raney nickel in methanol. It is noteworthy that in another attempt at this reduction, following a procedure developed by Nuzzo in 1977,150 Buku et al. utilized K3MoCl6 as a reducing agent to afford 233 (unpublished results). Next, 233 was re-oxidized using one equivalent of H2O2 in glacial acetic acid to afford a ca. 1:2 mixture of two diastereomers, (R) and (S)-sulfoxides (232 and 234 respectively), in favor of the (S) isomer. Separation of the two diastereomers was achieved on a Sephadex LH-20 column with water as the eluent. When a large excess of H2O2 was used, the sulfone (235) was formed as the sole oxidation product (Figure 4.19).  Figure 4.19 Oxidation of O-Me-α-amanitin (232) to (R)-sulfoxide, (S)-sulfoxide and sulfone by Buku et al.149 Using one equivalent of H2O2 afforded 232 and 234 in a 1:2 ratio, while using large excess of H2O2 yielded the sulfone (235).  159  4.2.5.3 3D Structures of Sulfoxides and Sulfone of α-Amanitin Following the isolation of 233, 234 and 235, Buku and co-workers evaluated the toxicity of each compound on white mice and found that the thioether (233), (R)-sulfoxide (232) and the sulfone (235) exhibited similarly high toxicities (1-2 mg/kg), while 234 was at least 10 times less toxic (10 mg/kg). These results were later confirmed by Wieland et al., when they measured the inhibitory constants of 232-235 against RNAP II from calf thymus: Ki = 2.5-5.0 nM for 232, 233 and 235, as opposed to Ki = 20 nM for 234.26 In 1983, Shoham and co-workers revealed an unexpected similarity between the 3D structures of the (S)-sulfoxide, (R)-sulfoxide and sulfone of amanitin. In their report, X-ray structures of the (S)-sulfoxide and sulfone of O-Me-α-amanitin (234 and 235, respectively) were acquired, while the same compounds (234 and 235) along with the (R)-sulfoxide (232) and thioether (233) of O-Me-α-amanitin were investigated by 1H-NMR and NOE experiments to determine their three-dimensional structures.46 Since both the (S)-sulfoxide and sulfone of O-Me-α-amanitin bear an oxygen atom on the same side of the thioether, it was initially assumed that a conformational difference was responsible for the disparity in their toxicity. To validate this hypothesis, crystals of 234 and 235 were grown from MeOH and EtOH/H2O, respectively. The X-ray structures of these crystals were superimposed, and surprisingly, no noticeable dissimilarity in their conformation was observed other than the extra (R)-oxygen in the sulfone compound (Figure 4.20, A). To date, no XRD structure for α-amanitin itself had been reported (only that of β-amanitin) and, furthermore, attempts to recrystallize the (R)-sulfoxide of O-Me-α-amanitin (232) have failed. Hence, to juxtapose the 3D structures of the (R)-sulfoxide diastereomer with the sulfone, the crystal structure of β-amanitin (reported by Lipscomb54) was studied. A resemblance 160  between the structures of β-amanitin and α-amanitin was presumed since their only difference is the residue at position-1 (Asn1 in α-amanitin and Asp1 in β-amanitin) which should not alter their overall conformation. Interestingly, superimposing the X-ray structures of β-amanitin and the sulfone (235) showed minute differences between the two. The sole difference appeared to be a 90°-rotation of the plane containing the amide bond between Asn1 and Cys8 in the sulfone. This rotation was thought to occur due to the presence of a hydrogen bond between Asn1-NH and the (S)-oxygen of the sulfone (Figure 4.20, B and C). Next, 1H-NMR and NOE experiments were carried out to confirm the equality of solution-phase structures of different amanitins with their respective XRD structures. Following the successful assignment of the proton chemical shifts, modified Karplus equations151,152 were exploited to calculate the dihedral angles between bonds that generally determine the secondary structure of various peptides: HNCαH and HCαCβH. Indeed, the calculated dihedral angles were in accordance with the angles observed earlier in the X-ray structures. Shoham’s work asserted that despite the large disparity between the toxicity of the (S)-sulfoxide of O-Me-α-amanitin and that of the thioether, the (R)-sulfoxide and the sulfone, their overall conformations appeared to be indistinguishable. Hence, if not due to their 3D structure, the rationale for the difference in their cytotoxicities remains unresolved. 161   Figure 4.20 A) Superimposed XRD structures of O-Me-α-amanitin-(S)-sulfoxide (234) (thin line) and sulfone (235) (thick line). B) Superimposed X-ray structures of β-amanitin (thin line) and O-Me-α-amanitin-sulfone (235) (thick line). C) Hydrogen bond between Asn1(NH) and (S)-oxygen in O-Me-α-amanitin-sulfone (right), resulting in the 90°-rotation of the plane containing the Asn1 amide bond (reproduced with permission from Shoham et al.).46  4.2.5.4 Asymmetric Sulfoxidation: Results and Discussion Subsequent to exploring the literature on stereoselective sulfoxidation reactions, I surmised that most asymmetric oxidation methods would be obsolete in the case of S-deoxy-amanitin, since it contains a chiral thioether surrounded by sterically demanding residues. In addition, since most of these methods require the use of redox-active metals, I was hesitant to use them on precious amounts of precursor, particularly in light of the redox-sensitivity of the 6-hydroxy-indole. Furthermore, the use of metal-based oxidation methods would also possibly 162  fail due to the ability of multiple amide bonds and other functional groups on amanitin to act as chelators to metals, preventing them from producing the active catalytic complex with chiral ligands. Instead, I envisioned hopefully that the steric bulk of the nearby environment would interfere with any chiral ligand utilized in an asymmetric sulfoxidation. Hence, I embarked on exploring the possibility of inducing stereoselectivity in the sulfoxidation by harnessing the inherent chirality of S-deoxy-amanitin (217). In an attempt to predict the facial selectivity of an incoming oxidant based on the conformation of 217, I required the crystal structure of α-amanitin. Since there is no reported crystal structure of α-amanitin, I turned my attention to Lipscomb’s XRD structure of β-amanitin (Figure 4.21).  Figure 4.21 A) Chemical structure of β-amanitin. B) XRD structure of β-amanitin (adapted from Kostansek et al.), representing possible facial selectivity for the sulfoxidation. R = C(O)NH2 in α-amanitin.  Examining the crystal structure of β-amanitin, it seems that the si-face is more accessible to an oxidant. The si facial selectivity would result in the formation of the desired (R)-sulfoxide. Hence, I anticipated that the use of an oxidant that is more sterically demanding than H2O2 could favorably lead to the formation of the desired diastereomer.    163  4.2.5.4.1 Oxidant Screening Although utilizing a large oxidant to enforce steric control over the sulfoxidation was thought to lead to the correct sulfoxide, the robustness and overall yield of the oxidation were additional key factors worth of consideration. Toward this end, several oxidants were screened to achieve a clean oxidation of the thioether to sulfoxide. The qualitative results are summarized in Table 4.1. An example of the HPLC chromatogram of the crude product of a sulfoxidation reaction containing a mixture of (R)-sulfoxide, (S)-sulfoxide and sulfone of α-amanitin is shown in Figure 4.22. As will be discussed in the upcoming sections, the identity of the three resulting products were subsequently determined by mass spectrometry, UV absorbance and comparing their toxicity with that of the commercially available α-amanitin. Furthermore, HPLC co-injections with authentic α-amanitin provided further proof of their identity (vide infra). Shown in Table 4.1, H2O2 and mCPBA appeared to produce the sulfoxides as the major products. H2O2 produced the sulfoxide products only when acetic acid was employed as the solvent (entry 1), in ca. 1:2 ratio in favor of the (S)-sulfoxide, corroborating the results reported previously by Buku et al. The significantly shorter reaction time (30-120 min) for oxidation with mCPBA (entry 10) as opposed to the extremely long time (24-72 hours) required for the H2O2 oxidation, along with the fact that no by-products were formed when mCPBA was employed, persuaded me to select mCPBA as the oxidizing agent. However, contrary to my expectations, the diastereoselectivity was not dramatically improved and the unfavorable (S)-sulfoxide was still obtained as the major product (R:S 1:2.7). Other oxidants, such as tert-butyl hydroperoxide, cumene hydroperoxide and chiral oxaziridine 227 (Figure 4.19) also failed to 164  afford any noticeable amounts of sulfoxide and, in most cases, the product mixture comprised several unknown by-products.   Reaction Conditions Product Mixture Components # Oxidant Solvent Time S.M. Sulfoxide (R)/(S) Sulfone Unknown 1 H2O2 HOAc 24h-72h Minor Major 1 : 2 Minor Minor 2 H2O2 H2O 24h-72h Major Minor N/A Minor Minor 3 H2O2 MeOH 24h-72h Major Minor N/A Minor Minor 4 tBuOOH HOAc 24h-72h Major Minor N/A None Minor 5 tBuOOH H2O 24h-72h Major Minor N/A None Minor 6 tBuOOH MeOH 24h-72h Major Minor N/A None Minor 7 Cumene-OOH HOAc 24h-72h Major None N/A None Minor 8 Cumene-OOH H2O 24h-72h Major None N/A None Minor 9 Cumene-OOH MeOH 24h-72h Major None N/A None Minor 10 mCPBA MeOH 30-120 min None Major 1 : 2.7 Minor None 11a oxaziridine 227 MeOH 24h-72h Minor Minor N/A Minor Major 12a oxaziridine 227 H2O 24h-72h Major Minor N/A Minor Major 13a oxaziridine 227 MeCN 24h-72h Major Minor N/A Minor Major 14a oxaziridine 227 DMF 24h-72h Minor Minor N/A Minor Major Table 4.1 Qualitative results of the oxidant screening for sulfoxidation of 217. Reactions were performed on 10 nmol of 217 with 1.3 equivalents of oxidant at 21ºC. Composition of the product mixture was assessed by HPLC and MS. aTFA was used as additive. Minor: 5-20%. Major: 50-100%. N/A: not applicable due to the low amount of the desired product. Oxaziridine 227: (1R)-(-)-(10-camphorsulfonyl)oxaziridine.  165   Figure 4.22 An example of the HPLC chromatogram of the crude product of a sulfoxidation reaction containing a mixture of (R)-sulfoxide, (S)-sulfoxide and sulfone of α-amanitin. Reaction conditions: S-deoxy-amanitin (20 nmol), mCPBA (3.5 eq), iPrOH/EtOH 3:1, RT, 20 min. HPLC gradient: 0-30 min 6%-18% A, 30-34 min 18%-100% A; 34-37 min 100% A, 37-39 min 100%-6% A, 39-44 min 6% A (solvent A: 0.1% FA in H2O, solvent B: 0.1% FA in MeCN).  4.2.5.4.2 mCPBA Oxidation: Solvent and Temperature Screening In an effort to refine the diastereoselectivity of the sulfoxidation with mCPBA, various solvents and temperatures were screened. First, I turned my attention to attempting the oxidation reaction at different temperatures. Employing 1.3 equivalents of mCPBA, the highest R:S ratio of the sulfoxide products was obtained at 0°C (R:S 1:1.1) (Figure 4.23). Additionally, I noticed the formation of varying amounts of the sulfone product at different temperatures. Although compelling, these results did not point to a distinctive temperature trend for the diastereoselectivity of the reaction, and there appeared to exist an “optimum” temperature rather than a trend. 166   Figure 4.23 Apparent R/S selectivity for the sulfoxidation of 217 in methanol. Reaction conditions: S-deoxy-α-amanitin (3.3 nmol), mCPBA (1.3 eq), 2 hours. Ratios were calculated based on the area under the peaks following the HPLC injection of the crude reaction mixture.  Next, I focused on altering the solvent while performing the reaction with mCPBA at room temperature. Although no evident pattern was immediately observed, it appeared that in the case where alcohols were used as the solvent, bulkier alcohols led to an improved diastereoselectivity: the observed R:S ratio for methanol was found to be 1:2.7, for ethanol 1:0.87, and for isopropanol 1:0.3. It seemed that isopropanol provided the highest diastereomeric ratio in favor of the desired (R)-sulfoxide. However, due to the low solubility of the starting material in isopropanol, a larger amount of the sulfone product was also generated. It was thought that the presence of a large excess of mCPBA in the solution phase was responsible for over-oxidation of the sulfoxide products to sulfone. A summary of the results for the solvent screening experiments is shown in Figure 4.24. 167   Figure 4.24 Apparent R/S selectivity for sulfoxidation of 217 in various solvents. Reaction conditions: S-deoxy-α-amanitin (3.3 nmol), mCPBA (1.3 eq), RT, 2 hours. Ratios were calculated based on the area under the peaks following the HPLC injection of the crude reaction mixture.  Following the screening of different solvents and various temperatures, interpretation of the obtained results proved extremely bewildering. However, a crucial component present in all my experiments had initially escaped my attention: the sulfone product. In the upcoming section, I disclose how investigation of the sulfone product led to a more rational conclusion with regard to various sulfoxidation conditions.  4.2.5.4.3 Kinetic Resolution of (R)- and (S)-Sulfoxides via Oxidation to Sulfone Re-exploring my results, I noticed that larger amounts of the sulfone product (237) were present in the reactions that had resulted in greater apparent R:S ratios in the sulfoxidation reaction. Initially, I had attributed this observation to the poor solubility of the starting thioether (217) in the corresponding solvents. While this conclusion was partly true, I had not taken notice of the possibility of a selective over-oxidation of the (S)-sulfoxide to sulfone (237) in the presence of the (R) isomer. As a matter of fact, if oxidation of the (S)-sulfoxide (236) to 168  sulfone was more kinetically favored than the oxidation of the (R)-sulfoxide, utilizing excess amounts of oxidant would eventually lead to the full consumption of (S)-sulfoxide (Figure 4.25).  Figure 4.25 Proposed kinetic resolution of (R) and (S)-sulfoxides of α-amanitin via selective oxidation to sulfone (237) using excess mCPBA. Consumption of the (S)-sulfoxide (236) could lead to an apparent higher diastereoselectivity in favor of the (R) isomer.  To investigate this hypothesis, a few sulfoxidation reactions were carried out using precisely one equivalent of mCPBA to prevent the formation of the sulfone. Solvents of choice for these trials were methanol and two different isopropanol/ethanol mixtures (2:1 and 3:1). Neat isopropanol was not tried due to its inability to fully dissolve 217, possibly leading to the formation of the sulfone. Results of these experiments are summarized in Table 4.2. # Solvent (R)-Sulfoxide (1) (S)-Sulfoxide (236) Sulfone (237) 1 MeOH 34% 66% <1% 2 iPrOH/EtOH 2:1 48% 52% <1% 3 iPrOH/EtOH 3:1 47% 53% <1% Table 4.2. Sulfoxidation of 217 in different alcohol-based solvent systems and composition of the product mixture. Reaction conditions: 217 (3.3 nmol), mCPBA (1 eq), RT, 1-2h.  169  Based on my results, it was apparent that utilizing equimolar amounts of mCPBA and the thioether did not furnish the sulfone product (237) and, indeed, performing the reaction in iPrOH/EtOH (2:1) (entry 2, Table 4.2) increased the R/S selectivity to ca. 1:1, as opposed to ca. 1:2 for methanol. Nonetheless, the R/S ratio reached a plateau at 1:1 and efforts towards its improvement failed. In a separate experiment, I deliberately employed excess mCPBA (3.5 equivalents) to monitor the rate of the over-oxidation of the sulfoxide products to sulfone. Data points were acquired by HPLC injection of the reaction mixture after 20 minutes, 45 minutes, 80 minutes and 2 hours, at which point the reaction did not seem to proceed any further. Figure 4.26 shows the results of this experiment.  Figure 4.26 Sulfoxidation of S-deoxy-α-amanitin using excess mCPBA (3.5 eq.) in iPrOH/EtOH 3:1. Composition of the reaction mixture was determined at different time points (20 min, 45 min, 80 min, 120 min) using HPLC.  As shown in Figure 4.26, it appeared that as the reaction progressed, the (S)-sulfoxide reacted at a much higher rate to afford the sulfone product, while the (R)-sulfoxide remained unreactive. After almost 2 hours, the (S) diastereomer was fully converted to sulfone, and the 170  final reaction mixture contained merely the (R)-sulfoxide and the sulfone in ca. 1:1 ratio. These results supported my hypothesis on the selective reactivity of (R)- and (S)-sulfoxides against further oxidation to α-amanitin-sulfone, suggesting a kinetic resolution of the two diastereomers based on their respective rates of oxidation. Although the term “kinetic resolution” is generally used in the context of two enantiomers reacting at different rates in a chiral environment, I found it applicable to the oxidation of (R) and (S)-sulfoxides due to their seemingly indistinguishable 3D structures, which may nevertheless impart a chiral environment for complexation of the oxidant with the peptide prior to the oxygen transfer event (vide infra). In the course of this investigation, I came to the realization that the most fascinating observation was the inertness of the (R) diastereomer against mCPBA. Whereas a large excess of mCPBA was used in this experiment, sufficient to oxidize both (R) and (S) isomers to sulfone, only the (S) isomer was fully consumed and (R)-sulfoxide remained unreactive. Herein, I propose somewhat speculative arguments to justify these results for the apparent diastereoselective sulfoxidation of S-deoxy-α-amanitin using mCPBA. Moreover, following the apparent kinetic resolution of the two diastereomers and selective oxidation of the (S)-sulfoxide to sulfone 237 in the presence of the (R)-sulfoxide, I explain my attempts to rationalize this phenomenon. While these explanations might be premature without further experiments, they will shed light upon my extremely gripping results for which no precedence in the amanitin literature may be found.  4.2.5.5 Rationalization of the Sulfoxidation Diastereoselectivity Based on my results (Figures 4.23 and 4.24), it proved extremely challenging to pinpoint distinctive temperature and solvent trends for the diastereoselectivity of the sulfoxidation of 171  thioether 217. In the case of temperature screening, results showed a gradual decrease in the R:S ratio at lower temperatures; however, these results were not always reproducible, and the temperature dependency varied with the solvent used in the oxidation reaction. Additionally, oxidation at 27°C resulted in a lower R/S ratio as opposed to the reaction at 0°C, which was inconsistent with the envisioned trend. While I am unable to draw a definite conclusion about the effect of temperature on the diastereoselectivity of the reaction at this point, it may be deduced that the (R)-sulfoxide is likely a more thermodynamically stable product, formed via oxidation of the pro-R lone pair of the thioether, while the (S)-sulfoxide is obtained from the seemingly more accessible pro-S lone pair to afford the kinetically favorable product (Figure 4.27). This observation was in contrast with my initial assumption that the pro-R lone pair appeared more sterically accessible for oxidation (vide supra).  Figure 4.27 Proposed energy levels of the thioether and sulfoxides of α-amanitin. (R)-sulfoxide is likely the thermodynamically favored product, while (S)-sulfoxide is the kinetically favored isomer. Note that (R)- and (S)-sulfoxides of amanitin cannot interconvert.  The impact of the oxidation solvent on the diastereoselectivity of the sulfoxidation seemed more intricate. Regarding the size of the solvent molecules or their hydrogen bonding abilities, no evident trend was found to be consistent with the observed diastereoselectivities. While 172  solvents such as DMF, TFE, HFIP and HOAc afforded the (S) isomer predominantly, no common physical or chemical features are apparent in these solvents. Utilizing alcohol based solvents, however, seemed to improve the diastereoselectivity in favor of the (R) isomer. At this point, I may be able to attribute these effects to a combination of hydrogen bonding capabilities and van der Waals radii of the solvent molecules that may in turn alter the local conformation of the tryptathionine crosslink. Furthermore, interactions between the oxidant (e.g. mCPBA), the solvent and the bicyclic structure of the toxin can direct the incoming oxidant to attack the thioether from either face (re or si) via a pre-binding event, leading to oxidation of either the pro-R or the pro-S lone pair.  4.2.5.6 Rationalization of the Kinetic Resolution of (R)- and (S)-Sulfoxides With the current data in hand, I feel confident to propose a more precise justification for the selective oxidation of (S)-sulfoxide to sulfone in the presence of the (R) isomer. A possible explanation involves electronic effects and conjugation of the sulfoxide with the indole ring. Examining the XRD structure of β-amanitin (Figure 4.22), it appears that the lone pair on the sulfur atom is positioned perpendicular to the 6-hydroxyindole plane, rendering it capable of participation in the conjugated system of the indole ring. This conjugation prevents these electrons from further oxidation to sulfone. On the other hand, analyzing the XRD structure of O-Me-α-amanitin-(S)-sulfoxide (234) (Figure 4.21, A), the sulfoxide lone pair seems to be located at an improper angle for conjugation into the π-system of the indole ring. This lack of conjugation will possibly result in a higher tendency for oxidation to sulfone (Figure 4.28). Another vital piece of evidence that could help either support or dismiss this hypothesis is the maximum absorbance wavelengths for the (R)-sulfoxide versus the (S) isomer. (R)-sulfoxide 173  exhibits a λmax of 305 nm, while that of the (S)-sulfoxide is 309 nm, pointing at the presence of a longer conjugated π-system in the (S) isomer.  Figure 4.28 Proposed explanation for the selective oxidation of (S)-sulfoxide to sulfone in the presence of (R)-sulfoxide.  Although more experiments will be required to confirm either of the proposed explanations, I presented valuable information on the chemical behavior of both (R) and (S)-diastereomers of α-amanitin, possibly leading to solving the perplexing puzzle of the lower toxicity of the (S)-α-amanitin as opposed to its thioether, (R)-sulfoxide and sulfone derivatives. It appears that Nature, in a fascinating manner, has evolved a more stable diastereomer of α-amanitin in its (R)-isomer that is resistant to further oxidation. While the oxidation is likely to be mediated by enzymatic action thereby ensuring high stereoselectivity, it is intriguing to speculate that the reason the sulfoxide is formed in the first place is to minimize air-oxidation of the hydroxy-tryptathionine that would give a product that is substantially less toxic than the thioether.  174  4.2.6 Characterization of the Synthetic α-Amanitin The last measure in my total synthesis of α-amanitin was to verify the identity of the synthetic product via various characterization methods using the commercially available authentic α-amanitin as a standard. To this end, several verification methods were employed: HPLC co-injections, NMR, UV absorbance, circular dichroism and cell toxicity assays.  4.2.6.1 HPLC and UV Analysis of the Synthesized (R)- and (S)-Sulfoxides As a preliminary proof of concept, I compared the HPLC retention time of my synthetic toxin with that of the authentic α-amanitin. Eluting with the HPLC solvent system of my choice (gradient D, see Materials and Methods), they both exhibited a retention time of 27.3 minutes. Moreover, an equimolar mixture of the synthetic and authentic amanitins was injected into HPLC, and a single peak with a retention time of 27.3 minutes was eluted (Figure 4.29).  Figure 4.29 HPLC chromatograms at 305 nm for the synthetic α-amanitin (blue), authentic α-amanitin (red) and their co-injection (green). HPLC gradient: D (refer to section 4.4.1 for HPLC gradients).  175  To reject any possibility of the (S)-sulfoxide of α-amanitin (236) possessing a similar retention time, I juxtaposed their HPLC retention times. The (S)-sulfoxide had a retention time of 28.7 minutes as opposed to 27.3 minutes for the (R)-sulfoxide. Furthermore, their co-injection showed two peaks with their corresponding retention times (Figure 4.30).  Figure 4.30 HPLC chromatograms at 305 nm for the synthetic α-amanitin-(S)-sulfoxide (purple), authentic α-amanitin (red), and their co-injection. HPLC gradient (green): D (refer to section 4.4.1 for HPLC gradients).  Finally, I studied the UV absorbance curves for the synthetic α-amanitin, (S)-sulfoxide, and the authentic α-amanitin. Both synthetic and authentic α-amanitins demonstrated a λmax of 305 nm, while that of the (S)-sulfoxide was 309 nm (Figure 4.31). 176   Figure 4.31 UV absorbance curves for the synthetic α-amanitin (blue), (S)-sulfoxide (purple), and the authentic α-amanitin demonstrating their λmax (red).  4.2.6.2 Circular Dichroism As for any optically active chiral molecule, α-amanitin absorbs and rotates lights with varying wavelengths. Circular dichroism (CD) is an experiment in which the difference between the absorbance of a left circularly polarized (LCP) and right circularly polarized (RCP) light at a given wavelength is measured for a chiral compound. This experiment is an extremely practical means for investigation of the secondary structure of proteins. In fact, a conformational change in a protein will lead to a noticeable difference in its CD spectrum. Possessing a turn similar to the β-turn of a protein, α-amanitin exhibits a distinctive CD spectrum with a positive Cotton effect in the 260-320 nm region along with a negative Cotton effect between 220-260 nm.6 To confirm the enantiopurity of my synthetic α-amanitin, its CD spectrum was acquired and compared to that of the authentic α-amanitin (Figure 4.32). 177   Figure 4.32 CD spectra for the synthetic and authentic α-amanitins. Sample preparation for CD: Authentic (purchased from Sigma-Aldrich, 53 µg) and synthetic amanitin (48 μg) were each dissolved in 500 μL of MeOH. A quartz cuvette (1 mm path-length) was used. CD spectra were acquired as noted in the Materials and Methods section (vide infra).  As shown in Figure 4.32, the CD spectrum for my synthetic α-amanitin matched that of the authentic version, with a positive Cotton effect past 252 nm and a negative Cotton effect between 211 nm and 252 nm. This experiment affirmed that the secondary structure of the synthetic α-amanitin fully resembles that of the authentic α-amanitin, rejecting the possibility of an epimerization event at any Cα on the backbone of the synthetic peptide.  4.2.6.3 NMR Characterization As a definitive means to prove the identity of the synthetic product, NMR experiments were carried out on S-deoxy-α-amanitin (217), synthetic and authentic (R)-sulfoxides of α-amanitin, and (S)-sulfoxide 236. It is noteworthy that due to the presence of small amounts (ca. 5%) of an impurity found in the commercial α-amanitin, I first purified the authentic toxin using HPLC to obtain pure α-amanitin prior to performing the NMR experiments. Toward this 178  end, 1H-NMR spectra were acquired for S-deoxy-α-amanitin (217), (S)-sulfoxide (236) and the synthetic α-amanitin. These spectra were then juxtaposed with that of the authentic α-amanitin in addition to the spectra reported by Shoham et al.46 To prepare each NMR sample, the HPLC-purified material was directly lyophilized from the eluent (H2O/MeCN + 0.1% formic acid). To remove the remaining formic acid, samples were re-lyophilized at least 4 times by repeated dissolution in H2O/MeCN 1:1. Additionally, to minimize the amount of water, the selected NMR solvent (99.96% DMSO-d6) was dried over 4Å molecular sieves two times. Table 4.3 demonstrates the proton chemical shifts for various NMR spectra corresponding to my synthetic amanitins and derivatives synthesized by Shoham et al.  Amanitin Number Residue Proton 1a 1b 1c 236 232 217 233 Asn1 HN 8.44 8.44 8.41 8.75 8.75 8.43 8.40 HCα 4.64 4.64 4.64 4.73 4.83 4.72 4.72 H’Cβ 2.95 2.95 2.95 2.59 2.62 2.95 2.94 H”Cβ 3.55 3.55 3.50 2.84 2.77 3.38 3.33 NH2 --- --- 8.32, 7.51 --- 7.65, 7.18 --- 8.27 Hyp2  HCα 4.20 4.20 4.28 4.11 4.22 4.27 4.30 H’Cβ 1.81 1.81 1.85 1.96 1.91 1.86 1.88 H”Cβ 2.16 2.16 2.19 2.13 2.21 2.19 2.21 HCγ 4.37 4.37 4.37 --- 4.43 4.37 4.40 H’Cδ 3.80 3.80 3.80 --- 3.57 3.71 3.72 H”Cδ 3.80 3.80 --- 3.68 3.77 3.80 HO --- --- 5.12 5.35 5.24 5.59 --- DHIle3 HN --- --- 7.78 7.96 7.97 7.98 7.93 HCα 4.39 4.39 4.43 4.29 4.32 4.41 4.47 HCβ 2.09 2.09 2.11 2.19 2.19 2.17 2.21 HCγ 3.54 3.54 3.52 --- 3.58 3.46 3.51 H’Cδ --- --- 3.30 --- 3.41 --- 3.44 H”Cδ --- --- --- 3.36 --- 3.31 CH3 0.82 0.82 0.87 0.88 0.91 0.87 0.89 HOγ --- --- 4.64 --- 4.85 --- 4.76 HOδ --- --- 4.28 --- 4.51 --- 4.33  Trp4  HN --- --- 7.83 8.03 8.00 7.88 7.83 HCα 4.89 4.89 4.92 5.06 5.19 4.88 4.95 H’Cβ 2.75 2.75 2.74 --- 3.01 2.90 2.80 179    Trp4 (cont’d) H”Cβ 3.18 3.18 3.20 --- 3.83 3.22 3.24 H-4’ 7.43 7.43 7.43 7.50 7.60 7.32 7.44 H-5’ 6.58 6.58 6.59 6.66 6.80 6.54 6.68 H-7’ 6.73 6.73 6.75 6.70 6.80 6.60 6.72 HN-indole 11.21 11.21 11.18 11.43 11.56 10.82 10.96 OH --- --- 9.12 9.52 OMe: 3.79 9.42 --- Gly5 HN 7.88 7.88 7.94 7.70 7.67 8.07 8.05 H’Cα 4.33 4.33 4.31 4.26 4.25 4.13 4.17 H”Cα --- --- 3.39 3.16 3.22 3.40 3.44 Ile6 HN 8.39 8.39 8.44 8.16 8.11 8.55 8.44 HCα 3.67 3.67 3.67 3.74 3.75 3.72 3.71 HCβ 1.55 1.55 1.56 1.55 1.56 1.55 1.56 H’Cγ 1.54 1.54 1.50 1.50 1.49 1.52 1.53 H”Cγ 1.10 1.10 1.11 1.14 1.11 1.10 1.11 CH3β 0.77 0.77 0.80 0.79 0.81 0.79 0.79 CH3γ 0.82 0.82 0.83 0.82 0.82 0.82 0.83 Gly7 HN 8.82 8.82 8.69 8.91 8.85 8.93 8.77 H’Cα 3.48 3.48 3.44 --- 3.52 3.43 3.42 H”Cα 3.91 3.91 3.90 3.80 3.76 3.89 3.91 Cys8 HN 8.28 8.28 8.25 7.94 7.97 7.98 7.94 HCα 4.93 4.93 4.94 4.97 4.99 4.53 4.57 H’Cβ 2.88 2.88 2.96 2.54 2.50 2.72 2.77 H”Cβ 3.03 3.03 3.08 4.52 4.53 3.01 3.05 Table 4.3 Proton chemical shifts for various derivatives/samples of amanitin. 1a: synthetic α-amanitin. 1b: authentic α-amanitin (purchased from Sigma-Aldrich). 1c: reported chemical shifts for α-amanitin from Shoham et al. 236: synthetic α-amanitin (S)-sulfoxide. 232: reported chemical shifts for O-Me-α-amanitin (S)-sulfoxide from Shoham et al. 217: synthetic α-amanitin-thioether; 233: reported chemical shifts for O-Me-α-amanitin-thioether from Shoham et al. Important/distinctive chemical shifts have been highlighted with gray shading and are featured in bold font. Chemical shifts that are not reported could not be observed due to the H2O/solvent peaks in the 1H-NMR spectra and, additionally, could not be analyzed using COSY60 experiments (99.96% DMSO-d6, 600MHz).  To achieve a more precise assignment of 1H-NMR peaks, COSY60 experiments were carried out on the synthetic α-amanitin, α-amanitin-(S)-sulfoxide and thioether 217. Moreover, S-deoxy-α-amanitin was further investigated using an HSQC experiment (see Appendix C). Examination of Table 4.3 reveals notable differences between the chemical shifts of several 180  protons in the bicyclic structure of amanitin derivatives. While these chemical shifts have been highlighted in this table, some of the more distinctive ones are as follow: HNindole-Trp4, HCβ-Cys8, HN-Asn1, and HCβ-Asn1 (Figure 4.33). The difference in HNindole-Trp4 chemical shifts in various amanitins may be attributed to what was discussed earlier: participation of the lone pair of the sulfur atom in the conjugated system of the 6-hydroxy-indole ring in the (R)-sulfoxide (Figure 4.28). This conjugation will in turn increase the electron density of the indole ring hence shifting the HNindole proton slightly upfield (δ11.21 ppm) compared to the (S)-sulfoxide (δ11.43 ppm). In the thioether (217), due to the absence of any oxygen atom on the Ttn-sulfur, the HNindole proton is significantly shifted downfield to δ10.82 ppm . Furthermore, the same argument seemingly holds for the difference in the chemical shifts of β-protons of Cys8 in (R) and (S)-sulfoxides: the electron density resulting from the sulfur’s lone pair present in the vicinity of the Cys8 β-protons determines the level of electron deficiency in these protons. In the case of (S)-sulfoxide, one of the Cys8-β-protons (possibly the pro-R proton) is observed at an extremely higher chemical shift (4.52 ppm) as opposed to the same proton in (R)-sulfoxide (3.03 ppm). This observation is very well in line with my hypothesis about the delocalization of the sulfur lone pair in the (R) diastereomer. The explanation for the disparity between HN-Asn1 chemical shifts proved much simpler. As proposed by Shoham et al., the presence of a hydrogen bond between the oxygen atom of the (S)-sulfoxide and the HN-Asn1 proton may lead to a slightly higher chemical shift for this proton in the (S) diastereomer (8.75 ppm) in contrast with that of the amanitin-thioether (8.43 ppm) and (R)-sulfoxide (8.44 ppm). However, I am uncertain of the applicability of a similar reasoning to justify the lower chemical shifts detected for the Asn1 β-protons in the (S) isomer 181  (2.59 ppm, 2.84 ppm) compared to the same protons in the (R) isomer (2.95 ppm, 3.55 ppm) and the thioether (2.95 ppm, 3.38 ppm).  Figure 4.33 Notable distinctive proton chemical shifts in the 1H-NMR spectra of the (R) and (S)-sulfoxides and thioether of α-amanitin. Values shown in boxes are in ppm (99.96% DMSO-d6, 600MHz).  Finally, I prepared an overlay of the 1H-NMR spectra obtained for the synthetic α-amanitin and the commercially available product. As shown in Figure 4.34, the two spectra were superimposed, demonstrating an analogous identity for the synthetic and authentic α-amanitins. 182   Figure 4.34 An overlay of the 1H-NMR spectra for the synthetic (in blue) and authentic (in red) α-amanitins. Solvent peaks have been labelled with the name of the corresponding solvents. Unknown peaks, resulting from the sample preparation or solvent stabilizers, are marked with an asterisk (*). To avoid overcrowding, the exact values for chemical shifts are not shown.  4.2.6.4 Cell Viability Assays To conclude my characterizations, the bioactivities of S-deoxy-α-amanitin (217), (R)-sulfoxide and (S)-sulfoxide (236) were assessed against CHO cells, with the commercially available α-amanitin used as a standard. To calculate the number of viable cells, following treatment of cells with different concentrations of tested peptides, an MTT colorimetric method was employed. Figure 4.35 summarizes the results obtained from these assays. Authentic α-Amanitin Synthetic α-Amanitin * * * formic acid DMSO H2O MeCN 183    Figure 4.35 MTT cell viability assay against CHO cells for the synthetic and authentic α-amanitins, thioether 217 and (S)-sulfoxide 236. Top: bar graph representing % of viable cells at a given concentration. Bottom: IC50 curves (IC50 values are reported in µM).  The toxicity assay results affirmed that the (R)-sulfoxide (1) and thioether of α-amanitin possess cytotoxicities similar to that of the authentic α-amanitin (IC50 ~ 0.35 µM). On the other hand, (S)-sulfoxide 236 demonstrated an IC50 of 2.8 µM, approximately 8-fold higher than that 184  of the natural product. This observation was consistent with the previous reports on toxicity of the (S)-sulfoxide of α-amanitin.26,46,149  4.3 Conclusion In this chapter, my route towards the first total synthesis of the bicyclic octapeptide α-amanitin was concluded thereby sealing this venerated natural product in the synthetic record for the first time. Incorporation of the 6-BMIDA-Trp in the solid phase synthesis of the linear heptapeptide of α-amanitin (201) was followed by the Savige-Fontana reaction to induce the tryptathionylation. Then, the 6-BMIDA group was oxidatively deborylated to a hydroxyl group to afford the monocyclic heptapeptide of amanitin containing a 6-OH-Ttn residue. Next, the NHS-ester of the fully protected DHIle was introduced, followed by the one pot deprotection of Fmoc and TBDMS protecting groups. While performing the TBS deprotection under strongly basic conditions, I noticed the formation of an oxidized material as the major product, which was 2 mass units lighter than the expected product. At the time of preparing this thesis, I am uncertain about the identity of this by-product. I suspect, however, that its formation is a result of deprotonation of the hydroxyl group of 6-OH-Ttn that undergoes oxidation to a quinone-methide-like product, which may lead to the β-elimination of the Hα(Cys8) residue followed by the scission of the tryptathionine crosslink (Figure 4.14). To avoid the formation of this by-product, the pH of the TBS deprotection was adjusted to 5 by addition of acetic acid to the reaction mixture. The deprotected monocyclic octapeptide was then macrolactamized to furnish the bicyclic octapeptide S-deoxy-α-amanitin (217). This peptide, whose synthesis has not previously been reported in the literature, represented the first cytotoxic intermediate in 185  my synthesis of α-amanitin. In fact, due to its extraordinary toxicity, its production was limited to sub-milligram amounts. The last part of my synthesis involved the diastereoselective sulfoxidation of S-deoxy-α-amanitin to the (R)-sulfoxide found in the natural product. While employing mCPBA as the oxidant yielded a mixture of (R) and (S)-sulfoxides in a rapid and efficient manner, I was unable to improve the diastereoselectivity to ratios above ca. 1:1. This diastereomeric ratio was achieved using mCPBA in isopropanol/ethanol mixtures as the reaction solvent. An extensive discussion on the parameters affecting the diastereoselectivity of the sulfoxidation (e.g. temperature and solvent), along with hypotheses to justify my results were disclosed. Perhaps the most compelling part of my results pertained to the selective oxidation of the (S)-sulfoxide of α-amanitin to sulfone in the presence of the (R) diastereomer, leading to the kinetic resolution of the two diastereomers under oxidative conditions. It appeared that the (R) isomer resisted oxidation to sulfone, while the (S) isomer was effortlessly further oxidized. My observations were accompanied by a hypothesis claiming the delocalization of the Ttn-sulfur lone pair in the (R)-sulfoxide into the 6-hydroxyindole ring, preventing it from further oxidation. Finally, the identity of my synthetic α-amanitin was supported by results from several characterization experiments. HPLC chromatograms, UV absorbance, circular dichroism, 1H-NMR spectra and cell toxicity assays were the employed experiments to validate my synthesis. In conclusion, herein I completed the first total synthesis of α-amanitin, a venerable toxin whose use as a murder and suicide means dates back to Roman times.4 Benefiting from a rich history, this peptide has been attracting many chemists and biochemists around the globe due to its ability to be utilized as a chemotherapeutic agent when linked to appropriate biomarkers. 186  Although I deliberately limited the production of α-amanitin to sub-milligram amounts due to its high potency, my synthesis is thought to be scalable and provides access to synthetic antibody-drug conjugates via different conjugation sites on α-amanitin. Furthermore, with the total synthesis in hand, the way is paved for the chemical synthesis of numerous novel analogues and derivatives that could prove even more effective than the natural product, both in their cytotoxicity and ability to be incorporated into ADCs.  4.4 Experimental Section  4.4.1 Materials and Methods General: All reactions were performed under argon atmosphere in flame-dried glassware and dried solvents at room temperature, unless otherwise stated. Controlled temperature reactions were performed using a mineral oil bath and a temperature controlled hot plate (IKA Ceramag Midi). Temperatures below room temperature were achieved in an ice/water bath (0°C), dry ice/ethylene glycol bath (-20°C), dry ice/ethanol/ethylene glycol bath (-20°C to -75°C) and dry ice/acetone bath (-78ºC). Solvents were removed under reduced pressure using a Büchi rotary evaporator. Anhydrous solvents were prepared by distillation under nitrogen atmosphere. Ethers were distilled from sodium in the presence of benzophenone as indicator. Triethylamine, dichloromethane and hexanes were distilled over calcium hydride. Methanol was distilled from magnesium. DMSO and DMF were dried over 4Å molecular sieves under argon atmosphere. All reagents and solvents were purchased from Sigma-Aldrich, Alfa Aesar, Acros Organics, Matrix Scientifics, Oakwood Chemicals, Ontario Chemicals or TCI America, unless otherwise stated. Authentic α-amanitin was purchased from Sigma-Aldrich. 187  Thin-layer chromatography (TLC) was performed using silica gel 60 F254 precoated aluminum plates (EM Science). Detection of TLC spots was performed using UV lamp at 254 nm, or by staining with p-anisaldehyde, potassium permanganate, ninhydrin or 2,4-dinitrophenylhydrazine, prepared according to literature procedures. Flash column chromatography purifications were performed using silica gel 60 (230-400 mesh, Silicycle, Quebec). Low-resolution mass spectra (LRMS ESI) in electrospray ionization (ESI) mode were obtained from a Bruker Esquire spectrometer. Proton (1H-NMR) and carbon (13C-NMR) spectra were obtained using Bruker AV-300 (300 MHz), AV-400inv (400 MHz) and AV600-CRP (600 MHz) spectrometers. Circular Dichroism (CD) spectra were obtained using a Jasco J-815 spectrophotometer. HPLC purification methods: All HPLC chromatograms were generated on an Agilent 1100 system equipped with an auto injector, a fraction collector and a diode array detector. Analytical injections were performed on an Agilent Eclipse XDB C-18 (4.6 x 250 mm) column with a flow rate of 2 mL/min (for some of the experiments, the column was changed to a newer one, but the same type, due to technical issues. These instances have been recognized with a “Note” in the ESI). The column was fitted with a column guard. In cases of closely-eluting peaks, integration was performed by standard data analysis software package whereby a line was drawn between both peaks and then integration was performed without peak correction. Chromatograms were obtained with a solvent gradient of 0.1% formic acid in water (Solvent B) and 0.1% formic acid in acetonitrile (Solvent A). The solvent gradients were: Gradient A: 0-24 min 10%-50% A, 24-27 min 50%-100% A; 27-32 min 100%A, 32-37 min 100%-10% A, 37-44 min 10% A. 188  Gradient B: 0-18 min 10%-50% A, 18-21 min 50%-100% A; 21-26 min 100% A, 26-31 min 100%-10% A, 31-38 min 10% A. Gradient C: 0-24 min 5%-33% A, 24-27 min 33%-100% A; 27-32 min 100% A, 32-37 min 100%-5% A, 37-44 min 5% A. Gradient D: 0-30 min 6%-18% A, 30-34 min 18%-100% A; 34-37 min 100% A, 37-39 min 100%-6% A, 39-44 min 6% A. Peptide quantification: Quantifications of different peptides were performed using a reported extinction coefficient of 12,600 M-1cm-1 at the wavelength with the maximum absorbance (290-309 nm). All UV wavelength scans and measurements were performed using a Cary5000 Spectrophotometer and readings were acquired at values of approximately 0.1-0.2 AU. Peptides were purified by HPLC and eluates were lyophilized. The dry compound was re-suspended in a known amount of solvent (0.1% formic acid in H2O:MeCN 1:1) and the concentration was measured using the UV absorbance at the λmax, assuming the extinction coefficient of 12,600 M-1cm-1 for all tryptathionine-containing peptides. Typically, 5.4-65.1 nmol (5-60 µg) is obtained and accurately quantified by UV-Vis spectroscopy using a 0.5 mL cuvette. For illustrative purposes, a quantity of 5.4 nmol in a volume of 0.5 mL gives a concentration of 10.8 µM and an absorbance reading of 0.136 AU, a value that is fully within the ideal range for quantitative UV-Vis spectroscopy. Sample preparation for cell toxicity assays: Following quantification, solutions with toxic peptides were re-lyophilized and re-suspended in a given volume of DMSO to provide a 1mM solution, which was then used in cell toxicity assays. Cell culture: Cells were cultured in a-MEM or high-sucrose DMEM, purchased from Gibco. Fetal bovine serum (FBS), 0.25 % trypsin (with 1.3 mM EDTA), 0.85 % Trypan blue, 189  and the antibiotic mixture Pen/Strep (10K U/mL penicillin, 10K mg/mL streptomycin) were also purchased from Gibco. All cell culture plastic ware was obtained from Corning or Falcon. Cells were cultured at 37 °C in a humidified chamber with 5 % CO2. When used in cell culture, DMSO was purified by filtration through a 0.2 mm filter. All experiments are carried out in a laminar flow culture cabinet, unless otherwise noted. Absorbance measurements of the 96-well plates were obtained using a Beckman-Coulter DTX 880 multimode detector, equipped with an excitation filter of 595 nm.  Immortalized CHO cells had been stored in liquid nitrogen. To revive cells, a 1 mL tube of the frozen cells in medium containing 10 % DMSO was warmed in a 30 °C water bath and diluted with 9 mL of fresh media. Media contained 10 % FBS and 100 U/mL penicillin and 100 mg/mL streptomycin, unless otherwise indicated. The cells were incubated in a T-25 flask at 37 °C at 5% CO2. After 24 hours, the medium was aspirated and replaced with fresh medium. When cells reached a level of 90 -100 % confluence, they were sub-cultured. The medium was removed, and the cells were treated with 0.25 % trypsin containing 1.3 mM EDTA in the incubator. Once the cells were detached from the tissue culture flask, 3-5 mL media was added to quench the trypsin, and transferred to a 10 mL centrifuge tube. The mixture was centrifuged for 5 min at 8000 rpm, and the supernatant was discarded. The cells were suspended in fresh medium, diluted as required, and transferred to a new culture flask.  To assay cell viability, a nearly confluent tissue culture flask was trypsinized, and the cells were counted following treatment with Trypan blue, using a hemacytometer. The cells were then diluted to the appropriate stock concentrations in fresh medium and transferred in 100 μL to a 96-well plate using a multi-channel pipette. The number of cells plated varied from experiment-to-experiment. These were incubated a 37 °C and 5% CO2 for a 24-hour period to 190  allow for adherence. The medium was aspirated, and fresh medium was added, which contained the desired additives in DMSO. The cells were then re-incubated for 72 h. At the completion of the experiment, a 100 μL aliquot of 25 mg/mL MTT in PBS was added to each well. The plate was incubated three hours further, to allow for the formation of the formazan product in viable cells. The media was carefully aspirated, and the purple product was solubilized in DMSO. The absorbance of each well was recorded at 595 nm, and an image of the plate was generated using a scanner. Data was processed in Microsoft Excel and GraphPad Prism. Experiments were performed either in triplicate, unless otherwise noted, and the error bars were calculated as the standard error of the mean.  Trypsinized cells were diluted to a concentration of 3.3 x 104 cells/mL for CHO cells. Each cell line was plated in a 96-well plate, with 100 μL of the stock solution per well (5000 cells per well) and incubated for 24 hours. Stocks of α-amanitin or analogs were prepared at various concentrations, containing 0.8 % DMSO, and added to various wells, according to the desired final concentration (8 (all except for the synthetic α-amanitin), 4, 2, 1, 0.5, 0.25, 0.125, 0.0625, 0.03125 μM). The cells were incubated for 72 hours, at which point viability was assessed as described.       191  4.4.2 Experimental Procedures   Synthesis of hexapeptide-200. Nα-trans-Hyp(OtBu) was loaded on the chlorotrityl resin according to the following protocol. To a flame-dried flask was added 2-chlorotrityl resin (1 g, 1.2 mmol/g, 200-400 mesh), which was then suspended in dry CH2Cl2 (10 mL). To this flask was added Fmoc-Hyp(OtBu)-OH (900 mg, 2.2 mmol) and DIPEA (0.96 mL, 5.5 mmol). The reaction was stirred at room temperature overnight and transferred to a spin column. The resin was washed three times with DMF and DCM (8 mL, each). Unreacted sites of the resin were capped by applying a solution of CH2Cl2:MeOH:DIPEA (8 mL of an 80:15:5 mixture, 20 min), and then washed with CH2Cl2 (3 x 8 mL) then DMF (3 x 8 mL) then CH2Cl2 (3 x 8 mL) again. The resin was dried on high vacuum over P2O5 to remove residual solvent. Resin loading was determined using manufacturer’s protocols. Briefly, a weighed amount of resin was treated with a 2% solution of DBU in DMF for 30 minutes. The solution was diluted and the UV absorbance of the liberated dibenzofulvene was measured at 304 nm, with an absorption coefficient of ɛ304=7624 M-1cm-1. Five equivalents of the following Nα-Fmoc-amino acids (Asn(NTr), Cys(STr), Gly, Ile, Gly) and five equivalents of coupling agent (HBTU/DIPEA) in DMF were applied sequentially to the growing N-terminus. In general, the following protocol was followed for coupling: Resin was placed in a Zeba spin column (up to 400 mg in a 5 mL column or 1 g in a 10 mL column) 192  and pre-swollen in DMF (8 mL) for 30 min (3 x 8 mL, 10 min each, draining DMF after each swelling). The solvent was drained, unreacted sites of the resin were capped with a solution of CH2Cl2:MeOH:DIPEA (8 mL of a 80:15:5 mixture, 20 min), and then washed with CH2Cl2 (3 x 8 mL) then DMF (3 x 8 mL) then CH2Cl2 (3 x 8 mL) again. The N-terminal Fmoc protecting group was removed by washing with 20% piperidine in DMF (8 mL for 5 min, 8 mL for 10 min). Following deprotection, the resin was washed with DMF (3 x 8 mL), followed by CH2Cl2 (3 x 8 mL) and again with DMF (3 x 8 mL). The next amino acid was coupled to the resin using suitably protected amino acid (Fmoc-Xaa(R)-OH, 5 eq.), coupling agent HBTU (5 eq.) and DIPEA (10 eq.) in DMF (8 mL). The reaction was slowly shaken on a vortexer at minimum speed for 2 h. For procedures in which a non-commercially available amino acid was used, fewer equivalents and longer coupling times were employed. Double coupling was performed when the free N-terminus on the resin was derived from a hydroxyproline residue or asparagine. Often, a Kaiser test was performed to check for complete couplings. Alternatively, a small amount of resin was removed from the batch and was deprotected with 25 % hexafluoroisopropanol (HFIP) in CH2Cl2, and the released peptide was analyzed by LRMS-ESI. When the reaction was complete, the coupling mixture was drained, and washed with DMF (3 x 8 mL).   193   Synthesis of heptapeptides 202 and 203 6-BMIDA-Nα-Boc-FPI-OH 188 (crude, 337 mg, 3 eq) was coupled to the N-terminus of the hexapeptide 200 (950 mg of resin, 0.17 mmol/g loading) as described in the previous procedure. After the coupling was completed (to yield heptapeptide 201), the resin was washed with DMF (3 x 8 mL), MeCN (3 x 8 mL), EtOAc/EtOH 1:1 (3 x 8 mL) and CH2Cl2 (3 x 8 mL). The resin-bound linear heptapeptide was transferred to a round-bottom flask and stirred in TFA/DCM 1:1 (15 mL) for 60 min. It has been observed that the acid labile protecting groups were concomitantly removed during the TFA treatment. The resin was filtered over cotton wool and washed with CH2Cl2 (10 mL). The combined filtrate was evaporated in vacuo, followed by co-evaporation with Et2O (2 x 15 mL), and then dried under reduced pressure. The residue was then re-dissolved in 0.1% FA (formic acid) in H2O/MECN 1:1 and purified on HPLC. Fractions containing the desired peak were detected at 290 nm and automatically collected (It must be noted that during the HPLC purification some of the BMIDA product was hydrolyzed to the boronic acid, in which case both peaks were collected for the next step). Yield 70% estimated based on resin loading. The product was lyophilized and re-suspended in MeOH. Concentrations of the peptide were determined by their UV absorbance at 290 nm, with an extinction co-efficient of 13,500 M-1cm-1 that is assumed not to differ from that reported for the natural product. 194  202: HPLC (gradient A): tR = 8.4 min; λmax 295 nm LRMS-ESI (m/z): [M+H]+ calcd. for C38H5210BN10O14S, 915.3; found 915.5. 203: HPLC (gradient A): tR = 7.3 min; λmax 295 nm LRMS-ESI (m/z): [M-OH+MeO-H]- calcd. for C34H4810BN9O12S, 816.3; found 816.3.   Removal of MIDA from heptapeptide 202 and oxidation of heptapeptide-203 to heptapeptide-204. Heptapeptide 202 (18.8 mg, 1 eq) (accompanied by 203) was dissolved in EtOH/acetone (1 mL:0.45 mL) in a 5 mL round-bottom flask. To this solution was added 380 μL of 0.5M aq KOH solution. The reaction mixture was stirred at room temperature for 5 min, at which point it was cooled down to 4°C in an ice-water bath to which 19 μL (1.1 eq) of a stock solution of mCPBA in EtOH (167 mg mCPBA in 1 mL EtOH) was added to the reaction mixture, and the resulting solution was stirred at 4°C for 5 min. Upon completion of the reaction, the reaction was acidified to pH5 by addition of 1 mL of a 0.1% formic acid solution in H2O. The resulting solution was lyophilized to yield the crude 8, which was further purified using HPLC (yield 58% over two steps). Note on the procedure: once the mCPBA is added, the pH of the oxidation mixture should be maintained below 8, otherwise unwanted oxidation products will predominate. 195  HPLC (gradient B): tR = 9.1 min; λmax 304 nm HRMS-ESI (m/z): [M+H]+ calcd. for C33H46N9O11S, 776.3038; found 776.3057.   Synthesis of monocyclic octapeptide-208. Heptapeptide 204 (5.8 mg, 1 eq) was dissolved in 150 μL DMF. To this was added a solution of (2S,3R,4R)-Oγ,Oδ-bis-TBS-Nα-Fmoc-dihydroxyisoleucine-NHS (20 mg, 3.8 eq) in 50 μL DMF, followed by 2.5 μL of DIPEA (final pH 8.5) The resulting mixture was stirred at room temperature for 48 hours. After completion of the coupling, 8.5 μL of Et2NH was added to the reaction mixture, and the reaction was stirred for another 2 hours. The reaction was concentrated in vacuo to dryness and was re-dissolved in 600 μL DMF. A volume of 45 μL of 1M TBAF in THF and 2.45 μL of HOAc were added to the solution, and the reaction mixture was stirred at room temperature for 1 hour. The reaction was further acidified to pH~3 with 1 M aq. HCl, and the solvent was removed in vacuo. The crude octapeptide 208 was purified using HPLC (isolated yield: 1.82 mg, 28% over 3 steps). Note: by analytical HPLC, yields appeared much higher however I attribute lower isolated yields to unexplained retention of material on the HPLC. Note on the procedure: once the TBAF is added, the pH of the oxidation mixture should be adjusted to ~5 by addition of acetic acid to prevent formation of unwanted oxidation products. 196  HPLC (gradient B): tR = 10.2 min; λmax 304 nm HRMS-ESI (m/z): [M+H]+ calcd. for C39H57N10O14S, 921.3776; found 921.3764.   Macrolactamization of monocyclic octapeptide-208 to bicyclic octapeptide-217. Octapeptide 208 (1.34 mg), HATU (5 mg) and DIPEA (2.4 μL, pH 9) were dissolved in 350 μL dimethylacetamide (DMA). The resulting mixture was stirred at RT for 2 hours. The solvent was removed in vacuo and the crude bicyclic 217 was directly purified using HPLC (yield ca. 50%). HPLC (gradient A): tR = 12.3 min; λmax 304 nm HRMS-ESI (m/z): [M+Na]+ calcd. for C39H54N10O13NaS, 925.3490; found 925.3494.     197   Oxidation of thioether-217 to sulfoxides 1 and 236. Thioether 217 (150 μg, 166.3 nmol, 1 eq) was dissolved in iPrOH/EtOH 2:1 (50 μL). To the resulting solution at room temperature was added mCPBA (52.8 μg mCPBA 77%, 1.3 eq, 50 μL of a solution of 5.2 mg mCPBA in 4.92 mL iPrOH/EtOH 2:1). The progress of the reaction was monitored by MS and HPLC. Upon completion of the reaction (0.5-1 hour), the crude reaction mixture was diluted in the HPLC buffer and directly purified on HPLC to afford a 2.57:1 mixture (based on HPLC peaks integration) of (R)-sulfoxide 1 (85 μg) and (S)-sulfoxide 236 (34 μg) with an overall isolated yield of 78%.  1: HPLC (gradient D): tR = 27.3 min; λmax = 305 nm; HRMS-ESI (m/z): [M+Na]+ calcd. for C39H54N10NaO14S, 941.3439; found 941.3443. 236: HPLC (gradient D): tR = 28.7 min; λmax = 309 nm; LRMS-ESI (m/z): [M+Na]+ calcd. for C39H54N10NaO14S, 941.3; found 941.8. Authentic α-Amanitin: HPLC (gradient D): tR = 27.3 min; λmax = 305 nm. Co-injection of synthetic and authentic α-amanitin: HPLC (gradient D): tR = 27.3 min; λmax = 305 nm.    198  Chapter 5: Analogues and Derivatives of trans-Hydroxyproline Residue  5.1 Introduction The interaction between the trans-hydroxyproline (Hyp2) residue of α-amanitin and RNAP II might be considered the most vital interaction for inducing the extreme toxicity and binding affinity of α-amanitin (Figure 5.1). While there are fewer pieces of evidence suggesting such crucial interactions between other amino acids of α-amanitin and RNAP II, one consistent observation that has been made throughout decades of studying α-amanitin is the significance of the hydroxyl group of Hyp2 and how its absence leads to a substantial decrease in the potency of the toxin.28,30  Figure 5.1 Interactions between the different residues of α-amanitin and the Rpb1 subunit of RNAP II (reproduced with permission). A) Cryo-EM structure by Cramer et al.24 B) Co-crystal structure by Kornberg et al.19 Note the hydrogen bonding interactions of Hyp2.  Among the first studies towards understanding the structure-activity relationships of analogues and derivatives of amanitin was investigating the effect of the hydroxyl group of Hyp2 on the activity of the toxin. To this end, Wieland, Zanotti and others synthesized 199  analogues containing Ile at position-3 and Pro or Hyp at position-2 and assessed their toxicity as opposed to the natural product (Table 5.1).  Entry R X Rel. Ki A OH S 32 B H S 880 C OH (R)-SO 36 D OH (S)-SO 230 E H (R)-SO 3880 F H (S)-SO 8000 Table 5.1 Amanitin analogues containing Pro or Hyp at position-2 and their inhibitory constants relative to α-amanitin against RNAP II from Calf Thymus (Ki for α-amanitin ~ 3 nM).  It can be seen that amanitin derivatives lacking the hydroxyl group of Hyp showed significantly decreased binding affinities (nearly 1000-fold) compared to those containing Hyp. These results are consistent with the interactions observed in the co-crystal structure of α-amanitin bound to RNAP II obtained by Kornberg et al. (Figure 1.8), reflective of a key hydrogen bond between Glu822 of the Rpb1 subunit of the enzyme and the hydroxyl group of Hyp. Moreover, the cryo-EM structure of the same complex (Cramer et al., Figure 1.9) further confirmed the presence of this key hydrogen bond. In addition to the hydrogen bonding ability of Hyp, its conformational properties may very well play a vital role in the fashion in which this amino acid impacts the binding affinity of α-amanitin. In the upcoming section, I briefly review the conformational properties of 4-substituted prolines.  200  5.1.1 Conformational Properties of 4-Substituted Prolines in Peptides Proline residues are typically incorporated in the structure of proteins or polypeptides to induce secondary structures, turns, or loops.154,155 This ability results from their restrained conformation caused by the 5-membered pyrrolidine ring. There are two major conformational equilibria observed for a proline-containing peptide bond: a trans or cis-amide, and an exo or endo ring pucker (Figure 5.2).156  Figure 5.2 Proline conformational equilibria. A) cis-trans amide bonds (in red) in slow equilibrium (trans isomer is greatly favored in proteins and peptides). B) Fast equilibrium between exo-endo ring puckers.  These two equilibria may control the overall conformation of the backbone of any protein or polypeptide containing one or more proline residues, hence making proline an extremely useful residue to induce a specific 3D structure within a synthetic polypeptide. The interconversion of trans and cis-amide bonds is slow and, in a cyclic peptide such as α-amanitin, faces significant conformational barriers. However, the exo-endo equilibrium can be greatly influenced by a substituent on the pyrrolidine ring, especially at C-4. As observed in collagen, the most abundant human protein, hydroxylation of certain prolines in its structure leads to a significant enhancement of the stability of this protein due to not only hydrogen bonding interactions but more importantly the stereoelectronic and conformational effects induced by the Hyp residues.157 In the case of trans-Hyp, the preferred gauche relationship between the (4R)-OH and the amide bond is the immediate result of the conformational change enforced by substituting C-201  4 with a hydroxyl group. This gauche conformation, although sterically disfavored, is the predominant conformation due to the alleged hyperconjugation of the electron-rich σ(C-H) orbital and the electron-deficient σ*(C-O) or σ*(C-N) orbitals. These stereoelectronic effects subsequently lead to an exo ring pucker in the (4R) isomer of Hyp. Analogously, in the (4S)-Hyp isomer, an endo ring pucker will be favored due to its ability to establish a gauche interaction between the (4S)-OH and the amide bond (Figure 5.3).158  Figure 5.3 A) Hyperconjugation of the orbitals in the gauche conformation of 4-Hyp. B) Preferred exo conformer in (4R)-Hyp (left); preferred endo conformer in (4S)-Hyp (right) (from Zondlo et al.).  Notwithstanding analogies to the “anomeric effect”, the above argument is primarily valid for prolines bearing a sterically undemanding, electron withdrawing group at C-4, such as hydroxy, azido, fluoro, and cyano-prolines. However, in the case of prolines carrying a sterically demanding substituent at C-4, the gauche conformation will be greatly disfavored due to the steric hindrance between the 4-substituent and the amide bond. This generally results in the (4R)-isomers possessing an endo conformation as opposed to exo, and the (4S)-diastereomers exhibiting an exo conformation as opposed to an endo conformation. Hence, installing a larger group at C-4 may lead to an unpredictable conformation for the 202  corresponding substituted proline, and subsequently alteration of the overall 3D structure of the proline-containing peptide. The stereoelectronic characteristics of various 4-substituted prolines, whether cis or trans relative to the 2-carboxamide, may be utilized to manipulate the stability, folding, activity or function of the proline-containing polypeptide via inducing a desired turn or conformation in the backbone of the peptide.159–161  5.2 Analogues of trans-Hydroxyproline for Incorporation in Amanitin As shown in Figure 4.21, the XRD structure of β-amanitin confirms the existence of the trans-Hyp2 residue in an exo conformation, with the pyrrolidine ring puckered away from the 2-carboxamide group. Although it has been shown that the hydrogen bonding ability of Hyp via its 4-hydroxyl group is crucial for the activity of α-amanitin, it is enticing to synthesize analogues of Hyp and incorporate them into the structure of the toxin and investigate the effects they impose on the overall structure and toxicity of the resulting amanitin analogues. Synthesis of these analogues can possibly pave the way to the discovery of a “super toxin”: a toxin that is more potent than α-amanitin. Furthermore, due to a lack of extensive SAR studies revolving around modifications of the Hyp2 residue (other than replacing Hyp with Pro), producing amanitins with varying proline analogues at position-2 of the peptide will help to further elucidate the significance of this position for the binding of the toxin, hence leading to a systematic SAR exploration with a focus on the proline derivatives.  5.2.1 Design and Selection of (4R)-Hydroxyproline Analogues While there exist reports for the synthesis of a broad range of 4-substituted prolines in the literature,158,162 none of them has been incorporated in the structure of amanitin. Hence, I 203  proposed the synthesis of several 4-substituted proline analogues for evaluation in this context. When designing these analogues, the ability of the substituent on C-4 to act as either a hydrogen bond acceptor or donor was considered, since it has been shown that the hydrogen bonding ability of Hyp is the source of a key interaction between α-amanitin and RNAP II (see Chapter 1). Additionally, due to the restrained bicyclic structure of α-amanitin, RNAP II is extremely specific in its binding to this toxin. Thus, in order to exhibit cytotoxicity, any amanitin derivative containing a synthetic analogue of Hyp most likely has to retain a local and overall conformation analogous to that of the natural product. As a preliminary means to assess the conformational resemblance of the proposed analogues to the trans-Hyp found in the native α-amanitin, basic computational studies were carried out to calculate the bond and dihedral angles of the said analogues in a model tetrapeptide system (Table 5.2). For these calculations, a Merck Molecular Force Field (MMFF) method was utilized. Our method of computational screening of the proposed proline derivatives might appear naïve, and other factors (e.g. solvent effects, the overall conformational change that each Hyp derivative could impose on the toxin, and hydrogen bonding abilities) were neglected for the most part. However, these preliminary data assisted in excluding candidates with large conformational differences compared to trans-hydroxyproline.      204   Entry X α β ϕ Θ α-amanitin OH 110.8 104.0 98.0 -79.0 A NH2 108.2 107.5 105.2 -80.5 B SH 112.1 111.7 129.0 -105.2 C CN 112.0 113.0 153.6 -159.7 D N3 110.7 111.2 150.7 -157.1 E NH(CO)H 109.2 110.4 94.9 -75.9 F NHOH 108.1 106.9 104.6 -81.1 G NHNH2 107.2 108.1 73.4 -77.5 H (CO)NH2 108.5 110.0 116.7 -89.2 K dashed box 116.6 117.5 51.7 -50.1 L dashed box 108.1 100.4 70.8 -56.7 Table 5.2 Calculated bond and dihedral angles for Hyp and proposed analogues for incorporation into amanitin. Angles were calculated following the MMFF-optimization of the equilibrium conformer of each analogue in a tetrapeptide model system (shown in the box). Analogues K and L can be seen in the dashed box.  Based on these results, several trans-4-substituted prolines were selected for synthesis and eventual incorporation into α-amanitin analogues: trans-amino, trans-thio, trans-cyano and trans-azido prolines (entries A-D, Table 5.2). In addition to providing a structural analogue of Hyp, trans-4-azido-proline affords the potential for a novel site of conjugation to the toxin via click chemistry (CuAAC), if such were to prove desirable for post-synthetic diversification. Other analogues with similar bond and dihedral angles (e.g. trans-acetamido proline, entry E) were rejected due to the more sterically demanding substituent on C-4, possibly resulting in an endo conformer of the 4-substituted proline as opposed to the desired exo (vide supra). Furthermore, 4-oxo-proline was added to the list of analogues due to its ability to be synthetically transformed on the final toxin: e.g. it may be reduced to cis- or trans-Hyp or 205  utilized as a handle for conjugation via reductive amination with an amine-containing linker. Figure 5.4 demonstrates the final selected analogues for incorporation in amanitin derivatives.  Figure 5.4 Selected Hyp analogues for synthesis and incorporation into amanitin derivatives. trans-Azido proline may be used as a novel site for bioconjugation via CuAAC. 4-Oxo-proline (keto-proline) may be reduced to cis- or trans-Hyp on the fully elaborated toxin or used as a handle for bioconjugation.  5.2.2 Synthesis of (4R)-Hydroxyproline Analogues  5.2.2.1 Previous Work on Various Modified Prolines Zondlo and co-workers have previously carried out the solid-phase synthesis of a vast variety of 4-substituted prolines that proved extremely useful for manipulation of the structure, conformation and functional groups of a peptide containing a trans-hydroxyproline residue.158 In their work, the commercially available trans-Hyp was incorporated in a model tetrapeptide (Ac-TYPN-NH2), and the unprotected Hyp was derivatized in an automated manner on the rink-amide resin. This approach enabled the synthesis of 122 different tetrapeptides containing various 4-substituted prolines with (4R) or (4S) stereochemistry. Then, the synthesized 206  peptides were analyzed by CD and NMR to assess the effect of different C-4 substituents of the proline residue on the overall conformation of the tetrapeptide model system (Figure 5.5).  Figure 5.5 Solid-phase synthesis of 4-substituted prolines on the rink-amide resin (from Zondlo et al.). R1 = Ac-Thr(OtBu)-Tyr(OtBu)-, R2 = -Asn(Trt)-NHRink-Resin.  Whereas the solid phase approach employed by Zondlo et al. has great advantages over the solution phase synthesis of proline analogues (e.g. easier automation and purification-less synthesis), the solution phase modification of Hyp has proven to be more robust and higher-yielding in our hands. Furthermore, due to the presence of several hydroxyl groups on the final toxin (e.g. two hydroxyl groups on DHIle3), the late stage modification of Hyp would not be feasible in the case of amanitin derivatives. Hence, inspired by Zondlo’s general strategy for conversion of trans-Hyp to other 4-substituted prolines, I embarked on the solution phase synthesis of the selected proline analogues. 207  5.2.2.2 Results and Discussion The general scheme for the synthesis of Hyp analogues involved conversion of the commercially available trans-Boc-Hyp-OH to cis-Boc-Hyp-OMe. Then, activation of the 4-OH to a leaving group followed by nucleophilic substitution with a suitable nucleophile would yield the desired trans isomer of the proline analogue subsequent to the inversion of the C-4 stereogenic center. In the case of keto-proline, oxidation of the 4-hydroxyl to the corresponding ketone would yield the desired analogue. Manipulation of the protecting groups to afford a SPPS compatible proline analogue would comprise the last part of the synthesis. A retrosynthetic scheme for these conversions is shown in Figure 5.6.  Figure 5.6 Retrosynthetic scheme for the synthesis of Hyp analogues and keto-proline starting from trans-Hyp.  To begin the synthesis, trans-Boc-Hyp-OH (238) was subjected to intramolecular Mitsunobu conditions with DIAD and PPh3 in THF to afford lactone 239. Then, 239 was converted to cis-Hyp-OMe (240) via an azide-mediated methonolysis163 using NaN3 in MeOH. To convert the cis-4-OH to a leaving group, a mesylation reaction was carried out on 240 to furnish the mesylated intermediate 241 (Figure 5.7). 208   Figure 5.7 Synthesis of the mesylated cis-Boc-Hyp-OMe (241) from trans-Boc-Hyp-OH (238). Mechanism of the azide-mediated methonolysis is shown.  Treating the mesylated intermediate (241) with a various number of nucleophiles would furnish the desired trans-4-substituted prolines via a second-order nucleophilic substitution reaction. To this end, in separate reactions, 241 was reacted with potassium cyanide, sodium azide and potassium thioacetate to install cyano, azido and thio functionalities, respectively (Figure 5.8). Whereas the azide displacement proceeded with high yields, the yields for the cyanide and thioacetate displacement reactions were relatively low owing to competing elimination reactions.  Figure 5.8 SN2 displacement reactions on 241 to afford trans-cyano (242), azido (243) and thio (244) proline intermediates.  209  Boc-protected cyano- and azido-prolines were then treated with LiOH in THF/H2O to achieve saponification of the methyl ester, followed by removal of the Boc protecting group in TFA/DCM to afford the fully unprotected cyano and azido prolines as crude mixtures of the corresponding TFA salts. Fmoc protection of the free amine of each residue using Fmoc-OSu furnished the SPPS compatible trans-Fmoc-cyano- and azido-prolines (Figure 5.9).  Figure 5.9 Synthesis of trans-Nα-Fmoc-CN-Pro-OH (249) and trans-Nα-Fmoc-N3-Pro-OH (250).  In the case of trans-thio-proline, due to the susceptibility of thiols to form disulfides under basic and oxidative conditions, it was essential to identify and employ a suitable protecting group that would be orthogonal to an Fmoc-based SPPS. As previously mentioned (see Chapter 4), an Fmoc-based solid-phase strategy for the synthesis of amanitin would involve the use of a base (piperidine) for Fmoc deprotection and an acid (TFA) to induce the Savige-Fontana reaction. Hence, an orthogonal protecting group for the 4-SH of trans-thio-proline was expected to endure both acidic and basic conditions. The acetamidomethyl (Acm) protecting group, widely utilized for the thiol protection in cysteine residues, meets these requirements: it is stable under both acidic and basic conditions,164 and may be removed using 2,2’-dithiobis(5-nitropyridine) (DTNP) and thioanisole165 that are orthogonal reagents to the solid-phase synthesis. Hence, thio-proline 244 was subjected to LiOH for concurrent saponification 210  of the methyl ester and the thioacetate. Then, following the Boc-deprotection with TFA, the free thiol was protected with Acm. Finally, the free amine was protected with Fmoc to yield the fully protected, SPPS compatible trans-Nα-Fmoc-(4-SAcm)-Pro-OH (254) (Figure 5.10).  Figure 5.10 Synthesis of trans-Nα-Fmoc-(4-SAcm)-Pro-OH (254). The Acm group is shown in blue.  With these three analogues in hand, I turned my attention to the synthesis of trans-amino-proline. Based on our SPPS strategy for the amanitin synthesis, the presence of a free amine on the side chain of the proline residue would interfere with peptide couplings on the CTC resin. Hence, I opted for carrying the trans-azido-proline residue (250) through the solid-phase peptide synthesis and utilize the azide functionality as a masked amine that would be revealed on the final toxin following a reduction step. To test the reduction conditions required for this transformation, I first treated a small amount of 250 with PPh3 in THF to induce reduction of the azide to amine under Staudinger conditions. To my disappointment, the iminophosphorane intermediate (255a) was unreactive towards hydrolysis even at slightly acidic pH and failed to afford the desired amino-proline product. Using PMe3 rather than PPh3 did not change the outcome of the Staudinger reaction. Hence, I attempted the catalytic hydrogenolysis of the azide to amine. To my delight, the amino-proline product (256) was obtained in a robust and clean manner (Figure 5.11). 211   Figure 5.11 Reduction of trans-Nα-Fmoc-4-N3-Pro-OH (250) to the corresponding amino-proline (256). Staudinger conditions resulted in the formation of the iminiphosphorane intermediates (255a, 255b) that failed to afford the product upon hydrolysis. Hydrogenolysis of 250 produced the desired product in high yields.  To examine the stability of the tryptathionine crosslink of amanitin analogues under the reducing conditions shown in Figure 5.10, a model tryptathionine compound (257) was subjected to hydrogenolysis under standard conditions, and the outcome of the reaction was monitored by TLC and MS. After maintaining the H2 atmosphere for over 24 hours, I noticed little if any degradation of the tryptathionine crosslink and the starting material remained intact (Figure 5.12). Hence, it was deduced that the azido-proline analogue could indeed be incorporated into an amanitin analogue and reduced to the corresponding amino-proline on the fully elaborated toxin without risking the decomposition of the tryptathionine crosslink.  Figure 5.12 Testing the stability of a model tryptathionine crosslink under catalytic hydrogenolysis conditions.  4-Oxo-proline represented the last proline derivative on the list of Hyp analogues. To achieve the synthesis of this residue, trans-N-Boc-Hyp-OH was first converted to the corresponding benzyl ester (258) using benzyl bromide and triethylamine in THF. Then, the 212  4-OH group was oxidized to the corresponding ketone with pyridinium chlorochromate (PCC) in DCM to afford N-Boc-4-oxo-Pro-OBn (259). Following saponification of the benzyl ester with H2/Pd and removal of the Boc protecting group in TFA and DCM, the free amine was re-protected with Fmoc to afford the SPPS compatible Nα-Fmoc-4-oxo-Pro-OH (262) (Figure 5.13).  Figure 5.13 Synthesis of Nα-Fmoc-4-oxo-Pro-OH (keto-proline) (262).  At this point, with the exception of amino-proline which would be obtained following the reduction of azido-proline on the final toxin, I had synthesized all of the proposed analogues as SPPS-compatible amino acids, which were ready for introduction into the structure of various amanitin derivatives. However, prior to discussing the strategy for their incorporation into the toxin, the synthesis of another proposed hydroxyproline derivative with unique characteristics is provided in the proceeding section.  5.3 Photocleavable Hydroxyproline Analogue In addition to the five hydroxyproline analogues discussed in the previous section, I decided to further exploit the significance of the hydroxyproline residue for inducing the high binding affinity of α-amanitin against RNAP II. Hence, I embarked on the synthesis of a photolabile 213  protecting group to mask the hydroxyl group of Hyp on the final peptide, which would drastically decrease the inhibitory effects of the toxin. Subsequently, following irradiation with a suitable wavelength, deprotection of the photocleavable Hyp could be achieved, resulting in the release of an active toxin. This feature can prove extremely useful to induce spatiotemporal control over the binding of α-amanitin. Furthermore, it provides a valuable means for investigation of the events that occur following the internalization of α-amanitin into the target cells which lead to toxicity and cell death. Herein, I disclose a brief introduction on photocleavable protecting groups followed by the synthesis of the corresponding photocleavable hydroxyproline residue.  5.3.1 Photocleavable Protecting Groups: A Brief Introduction The use of photocleavable (also known as photolabile, photosensitive, photoremovable) protecting groups in chemistry and biology dates back to 1966 and 1970, when Schofield and Woodward, in two separate reports, introduced utilization of this class of protecting groups in the context of amino acids and peptides.166,167 Since then, these protecting groups have also found application in biology; in 1978, Kaplan et al. reported the photolytic release of a derivative of ATP that was protected with a photolabile protecting group (caged ATP), and studied the rate at which it inhibited an ATPase enzyme prior to and following the deprotection when irradiated with a wavelength of 342 nm (Figure 5.14).168 More recently, Spatz and co-workers have utilized photocleavable linkers to micropattern a vast range of biomolecules and investigate their on-demand release after irradiation with the proper wavelength.169 214   Figure 5.14 Synthesis of a photolabile caged (photolabile) ATP. Upon exposure to λ=342 nm, an uncaged (deprotected) ATP is released, resulting in the inhibition of purified renal Na,K-ATPase (from Kaplan et al.). The photolabile protecting group is shown in blue.  Analogous to most protecting groups, the role of photocleavable protecting groups (PPGs) is to mask and protect a specific functionality, a process known as photocaging, until the release of the said functionality is required. The appealing features of PPGs include the use of a non-invasive method for their removal (light), ability to precisely deliver light for their deprotection, and formation of no contaminants in the sample upon their removal.170 The wide variety of functional groups that may be protected using PPGs and how they may be applied in the synthesis of natural products represent other compelling features of this class. For example, Karas and co-workers have recently employed a photocleavable thiol protecting group assisted with S-pyridinesulfenyl activation for the late stage and in situ formation of a disulfide bond en route to the chemical synthesis of insulin (Figure 5.15).171  Figure 5.15 Disulfide bond formation on a peptide by photocleavage of the 6-nitroveratryl protecting group (shown in blue) and thiolysis by S-pyridinesulfenyl activation (from Karas et al.).  215  There are generally two types of PPGs that are more commonly used in organic synthesis: A) ortho-nitrobenzyl based PPGs, and B) coumarin derivatives (Figure 5.16).172  Figure 5.16 Examples of A) ortho-nitrobenzyl-based and B) coumarin-based PPGs and their relative absorption maxima (adapted with permission from Feringa et al.).172 The leaving group released upon exposure to light is shown in blue.  As shown in Figure 5.16, altering the substituents on the benzylic position or the aryl ring of ortho-nitrobenzyl PPGs results in small to large red (bathochromic) or blue (hypsochromic) shifts in their absorption maximum; installing an electron withdrawing group at the para-position relative to the nitro group results in a red shift, whereas incorporating an electron donating group at the meta-position causes a blue shift. The typical range of wavelengths for deprotection of this class of PPGs falls between 345 and 420 nm. In the case of coumarin-based PPGs, changes in their structure may also lead to various deprotection wavelengths, generally in the range of 312-505 nm. It is noteworthy that other PPGs have been employed in the literature that are less popular, such as metal-containing (e.g. ruthenium complexes) or arylcarbonylmethyl groups (e.g. phenacyl or ortho-alkylphenacyl). Regardless of the type, the general requirements that need to be met by any efficient PPG include: i) a deprotection 216  wavelength above 300 nm that does not cause damage to biological systems, ii) a clean photoreaction with a high quantum yield, iii) the by-products of the deprotection should not absorb at the same wavelength and preferably not pose toxicity to biological systems, and iv) the ability to excite the PPG by a short pulse of light and fast release of the free substrate.170 Due to the broad application of ortho-nitrobenzyl PPGs in the literature and the ability to alter their deprotection wavelength via installing various electron-donating groups at different positions on the aryl ring,173–177 I pursued the synthesis of a trans-hydroxyproline residue protected with a PPG containing an ortho-nitrobenzyl core. A concise introduction to this class of PPGs follows.  5.3.2 ortho-Nitrobenzyl Photocleavable Protecting Groups Representing one of the most popular classes of PPGs, ortho-nitrobenzyl derivatives have been extensively utilized in photocaging various functional groups despite their drawbacks (e.g. the formation of a possibly toxic nitroso benzaldehyde species). To elaborate on the mechanism of the photoreaction of these groups, Gilch and co-workers studied the early events following the irradiation of ortho-nitrotoluene (oNT) (263) with light using stimulated Raman spectroscopy.178 Depicted in Figure 5.17, the first event that takes place involves the hydrogen transfer from the ortho-methyl to the nitro group leading to an aci-nitro tautomer. In the presence of water, a fast equilibrium between (Z)- and (E)-aci-isomers is established. 217   Figure 5.17 Mechanism of phototautomerization of oNT (Gilch et al.). The lifetime of each excited state is shown under its structure (ISC = intersystem crossing).  In line with the work by Gilch et al., in 2004, Wirz and co-workers carried out detailed mechanistic studies on the release of methanol from ortho-nitrobenzyl ethers, suggesting a mechanism shown in Figure 5.18.173 In this mechanism, following the irradiation (λmax ≈ 400nm) of the starting ortho-nitrobenzyl compound (264) in the presence of water, the singlet or triplet excited state of 264 undergoes a [1,5] hydrogen shift to transiently produce intermediate 265. Then, a cyclization reaction occurs (formally intramolecular Michael and/or [2+3]-electrocyclic), resulting in the transfer of a hydroxyl to the benzylic carbon to yield hemiacetal 267. The last step involves the release of MeOH from the hemiacetal, whereas ortho-nitroso benzaldehyde (268) is formed as the by-product of this photoreaction.173,179    218   Figure 5.18 Mechanism of photoreaction of 1-(methoxymethyl)-2-nitrobenzene, releasing MeOH (shown in blue) (Wirz et al.).  This class of PPG was attractive due to the ability to alter the deprotection wavelength via manipulation of the substitution pattern either at the benzylic position or on the aryl ring (vide supra). However, the major problem arising with the use of these groups is the formation of ortho-nitroso benzaldehyde that could prove to be toxic to cellular environments. Taking specific measures, one would be able to reductively quench the released nitroso compound either by addition of mercaptoethanol or DTT to the reaction mixture. Alternatively the same would quench via reaction with cysteine or glutathione residues found in cells.177,179 Specifically, 4,5-dimethoxy-2-nitrobenzyl (6-nitroveratryl, Nv) group seemed intriguing. Firstly, it possesses a deprotection wavelength of 366 nm that is compatible with biological systems and does not cause any damage to cells. Furthermore, it withstands Fmoc-based SPPS and Savige-Fontana conditions (i.e. it is not cleaved in strong basic or acidic media). Finally, with the photo-deprotection half-life of 2.44 minutes,31 its release is relatively fast and does not require long periods of exposure to light. Hence, I pursued the synthesis of a 6-nitroveratryl-protected trans-hydroxyproline that would lead to a photoactivable amanitin analogue.  219  5.3.3 Synthesis of a Photocleavable Hydroxyproline Following a procedure established by Wilcox et al.180 and previously employed by Dietrich from the Perrin lab,31 in order to obtain the 6-nitroveratryl (Nv) protected Hyp, the synthesis of 6-nitroveratryl bromide was required. To this end, 6-nitroveratraldehyde (269) was first reduced to 6-nitroveratryl alcohol (270) using sodium borohydride in MeOH. Then, bromination of the benzylic alcohol with PBr3 generated in situ from the reaction of PPh3 and Br2 resulted in the formation of 6-nitroveratryl bromide (271), isolated as a yellow solid (Figure 5.19). This intermediate was required to be prepared freshly prior to performing the next step to avoid its photodecomposition. In fact, all reaction flasks and isolated products were protected from visible light at all times to prevent the photodecomposition of the photolabile intermediates.  Figure 5.19 Preparation of 6-nitroveratryl bromide (271).  To incorporate the 6-nitroveratryl group in Hyp, trans-N-Boc-Hyp-OH was first converted to the corresponding methyl ester (272) using dimethyl sulfate and K2CO3 in acetone. This intermediate was then reacted with 6-nitroveratryl bromide in a biphasic mixture of DCM and H2O, in the presence of NaOH and tetra-butylammonium hydrogensulfate (nBu4NaHSO4) to afford trans-Nα-Boc-(ONv)-Hyp-OMe (273) (Figure 5.120). Although the yield for the Nv-protection reaction was relatively low (40%), it was in line with the results produced by Dietrich, reporting various yields for this transformation (35-62%). 220   Figure 5.20 Protection of trans-Nα-Boc-Hyp-OMe (272) with 6-nitroveratryl group to afford Nv-Hyp-OMe 273.  Next, saponification of the methyl ester followed by Boc-deprotection and Fmoc-protection of the amine furnished the SPPS compatible, Nv-protected trans-hydroxyproline residue (276) (Figure 5.21). With the fully protected photocleavable hydroxyproline in hand, its incorporation into the structure of an amanitin analogue was awaiting. In the upcoming section, I discuss the efforts towards developing a more efficient solid-phase strategy for incorporation of the synthetic Hyp analogues into the corresponding amanitins.  Figure 5.21 Protecting group manipulation of photocleavable hydroxyproline to generate SPPS compatible trans-Nα-Fmoc-(ONv)-Hyp-OH (276).  221  5.4 Improved Solid-Phase Strategies for Incorporation of Hydroxyproline Derivatives in Amanitin At this point, I was in possession of five analogues and derivatives of Hyp (cyano, thio, azido, keto-prolines and the photocleavable Hyp) for incorporation into amanitin. Additionally, we aimed to reduce the azido-proline analogue to the corresponding amino analogue on the final toxin, resulting in a sixth analogue of hydroxyproline. The solid phase strategy employed earlier for the total synthesis of α-amanitin (vide supra) required the introduction of Hyp as the first amino acid of the linear sequence, bound to the CTC resin (Figure 5.22). This route is not the most efficient one regarding the synthesis of amanitin analogues containing various Hyp derivatives for two apparent reasons. First, beginning the solid phase synthesis with Hyp residue would require a large excess of this amino acid. In the case of the commercially available, inexpensive (4R)-Hyp, obtaining large amounts of it is financially justified. However, in the case of the valuable, synthetic Hyp derivatives, using a large excess of them would prove to be inefficient and uneconomical. Secondly, to produce six distinct amanitin analogues via a solid phase route starting with proline residues, one must carry out six parallel solid phase reactions, leading to a tedious and time-consuming synthesis.   222   Figure 5.22 Solid phase synthesis of α-amanitin began with loading Hyp (shown in blue) on CTC resin.  To address these issues, I embarked on designing several new strategies for the solid phase synthesis of amanitins that would involve introducing the proline residue at a later stage, preferably towards the end of the synthesis of the linear sequence. Furthermore, in the new proposed routes, I aimed to perform as many steps on the solid phase as possible to increase the overall yield and avoid HPLC purifications. Herein, I discuss these modified solid phase syntheses. It is noteworthy that for the synthesis of amanitin analogues with different proline derivatives, a tryptophan residue (and not 6-OH-Trp) will be installed at position-4, and the thioether of the tryptathionine crosslink (and not the (R)-sulfoxide) will be prepared. As discussed in the previous chapters, these modifications should have minimal effect on the activity of the synthetic analogues (vide supra).  5.4.1 Strategy A: Loading the CTC Resin with Fmoc-Asp(OH)-OAllyl For our first modified SPPS strategy, I envisioned starting the solid phase synthesis with the allyl ester of Asp(OH)1, loading the carboxylic acid on the side chain of this residue on CTC resin. Following the elongation of the linear sequence, the last amino acid to be installed would be Hyp(OtBu)2. Owing to the use of Asp (and not Asn) at position-1, the final product 223  of this synthetic scheme would be a β-amanitin analogue as opposed to α-amanitin, which could prove even more useful for post-synthetic bioconjugation. A retrosynthetic scheme for this route is shown in Figure 5.23.  Figure 5.23 Retrosynthetic scheme for strategy A: loading the CTC resin with Fmoc-Asp(OH)-OAllyl (in red) and macrolactamization on solid phase. Proline derivative (in blue) is introduced as the last residue.  As seen in this figure when treated with TFA, acid labile protecting groups on the macrolactam would be removed, the peptide would be cleaved from the resin, and the tryptathionine crosslink could be formed. Hence, in this strategy, almost all synthetic steps were aimed to be carried out on the solid phase and only one HPLC purification would be required. To begin this synthetic route, Fmoc-Asp(OH)-OAllyl was loaded on CTC resin. Then, S-trityl cysteine, glycine, isoleucine and glycine residues were added to the linear sequence to obtain the pentapeptide intermediate 278 (Figure 5.24). 224   Figure 5.24 Synthesis of linear pentapeptide 278 (strategy A).  To test this modified solid phase strategy, I employed the less-costly L-isoleucine as an alternative to DHIle. Analogously, this amino acid was also used to test other solid phase strategies (B and C, vide infra). For addition of the next two residues to the linear pentapeptide, the Ile-Fpi-OH dipeptide (281) was required. To this end, the activated NHS-ester of Ile was first prepared. The crude NHS-ester was then reacted with the fully unprotected tryptophan to obtain the Ile-Trp-OH dipeptide (280). Finally, fluorocyclization of this dipeptide using FP-T300 afforded the corresponding (Nα-Fmoc-Ile)-Fpi-OH dipeptide (281) (Figure 5.25).  Figure 5.25 Synthesis of the test dipeptide Ile-Fpi-OH (281).  Next, dipeptide 281 was coupled to the linear pentapeptide (278) followed by the coupling of Fmoc-Hyp(OtBu)-OH to furnish the linear octapeptide 282. At this stage, the allyl ester of 225  Asp1 was cleaved in the presence of Pd(PPh3)4 and morpholine to produce the unprotected carboxylic acid. Unfortunately, attempts to macrolactamize the linear octapeptide on the resin failed to produce any noticeable amounts of the macrolactam product. I believe that two factors were responsible for failure in this transformation. First, due to the amino group of the Hyp residue being a secondary cyclic amine, any coupling reaction on the N-terminus of this residue would face well-known steric challenges. Secondly, due to proximity of the Asp1 amino acid to the surface of the resin, it might be sterically unfavorable for the amine functionality of Hyp to spatially approach the carboxylic acid of Asp1. Nevertheless, to assess the feasibility of inducing a tryptathionylation reaction on the linear octapeptide 283, this intermediate was subjected to TFA. Once treated with TFA/DCM, this linear octapeptide was cleaved from the resin, the acid labile protecting groups were removed, and more importantly, it underwent the Savige-Fontana reaction to obtain the tryptathionine crosslink (Figure 5.26). 226   Figure 5.26 Attempted synthesis of monocyclic octapeptide 284 on solid phase. Upon exposure to TFA, linear octapeptide 283 afforded the tryptathio-monocyclic octapeptide 285. Hyp is shown in blue and Asp is shown in red.  Although this method failed to produce the macrolactamized product on the solid phase, it confirmed that the tryptathionine crosslink could indeed be installed once all the amino acids were incorporated into the linear sequence of the peptide. Hence, I turned my attention to a different strategy starting with cysteine utilized as the first amino acid of the sequence.  5.4.2 Strategy B: Loading the CTC Resin with Fmoc-Cys(SH)-OAllyl The major drawback to the first solid phase strategy (strategy A) was the inability to effect macrolactamization on the solid phase. To avoid the necessity to carry out the 227  macrolactamization reaction between Asp1 and Hyp2 (as in strategy A), in a new method, I aimed to begin the solid phase synthesis with Cys8 and load the CTC resin with the side chain thiol of this residue. This new strategy would involve a macrolactamization reaction between Asp1 and Cys8 that would presumably be more robust as opposed to Asp1-Hyp2 macrolactamization. The retrosynthetic scheme for strategy B is shown in Figure 5.27. It is noteworthy that this method would afford a β-amanitin analogue as the final product.  Figure 5.27 Retrosynthetic scheme for strategy B: loading the CTC resin with Fmoc-Cys(SH)-OAllyl (in red) and macrolactamization on solid phase. Proline derivative (in blue) is introduced as the second last residue. Note the incorporation of Ile at position-3 in place of DHIle.  To begin this method, Fmoc-Cys(STrt)-OH (286) was converted to the corresponding allyl ester (287) using allyl bromide and DIPEA. Then, the trityl protecting group was removed with TFA to afford the amino acid required for SPPS (288) (Figure 5.28).   228   Figure 5.28 Synthesis of Fmoc-Cys(SH)-OAllyl (288) from Fmoc-Cys(STrt)-OH (286).  Next, Fmoc-Cys(SH)-OAllyl was loaded on the CTC resin. Analogous to strategy A, the remaining amino acids were coupled via an Fmoc-based SPPS to obtain the linear octapeptide (289) (Ile was employed at position-3 rather than DHIle). In this synthesis, Hyp was added as the second last residue followed by Fmoc-Asp(OtBu)-OH to complete the linear sequence. The Fmoc protecting group of the N-terminus and the allyl ester of the C-terminus were then removed using piperidine and Pd(PPh3)4/morpholine, respectively. To my disappointment, attempts to macrolactamize the linear octapeptide on the solid phase once again failed to afford monocyclic octapeptide 291. Although minor amounts of the desired macrolactam were observed, the majority of the linear octapeptide remained intact. Nonetheless, to confirm the ability to induce tryptathionylation with this strategy, the linear octapeptide (290) was treated with TFA/DCM. Delightfully, full conversion to the tryptathionine crosslink in addition to the removal of the acid labile protecting groups were observed, affording the tryptathio-monocyclic octapeptide 292. Compound 292 can presumably undergo macrolactamization in the solution phase to generate the desired β-amanitin analogue. A summary of the results for strategy B is shown in Figure 5.29. 229   Figure 5.29 Attempted synthesis of monocyclic octapeptide 291 on solid phase. Upon exposure to TFA, linear octapeptide 290 afforded the tryptathio-monocyclic octapeptide 292 (Hyp is shown in blue, Cys is shown in red). 292 can presumably undergo a macrolactamization reaction to produce the bicyclic β-amanitin analogue.  To conclude this pathway, it appeared that changing the starting residue of the linear sequence did not cause a great impact on the outcome of the macrolactamization reaction on the solid phase. As observed with strategy A, attempts to produce the monocyclic octapeptide (291) on the solid phase failed to furnish the macrolactam product in noticeable yields. 230  Whereas the product of the attempted macrolactamization reaction was not fully investigated (e.g. via HPLC characterization), MS analysis of the peptide cleaved from the resin confirmed the absence of the desired product and mostly showed the presence of the starting material (290). Thus, I concluded that it would be quite challenging for the N-terminus amino acid to approach the C-terminus of the linear peptide, most likely due to the sterically demanding environment in the vicinity of the surface of the resin. Hence, this SPPS strategy was altered in a way that would avoid macrolactamization on the solid phase thus enabling production of viable macrolactams of α-amanitin (as opposed to β-amanitin) analogues in solution.  5.4.3 Strategy C: Loading the CTC Resin with Fmoc-Cys(STrt)-OH, Followed by Macrolactamization in the Solution Phase Due to the challenging coupling of the N-terminus of Hyp and the C-terminus of Asp for inducing the final macrolactam, the third strategy involved loading the CTC resin with Fmoc-Cys(STrt)-OH as the initial amino acid. Accordingly, Hyp would be introduced as the second last residue, whereas Asn would be installed last on the linear peptide. It is noteworthy that with this route, analogues of α-amanitin could be synthesized as opposed to the previous strategies that were designed to produce analogues of β-amanitin. Furthermore, I employed Nα-Boc-Asn(Trt)-OtBu at position-1. The advantage of utilizing this amino acid lies in the fact that prior to the macrolactamization step, all protecting groups remaining on the final linear peptide (293) will be acid labile, and upon exposure to TFA, may be removed and a tryptathionine crosslink will be concurrently generated. The attempted solid phase synthesis based on this strategy is shown in Figure 5.30. I should mention that, analogous to previous methods, Ile was employed at position-3 in place of DHIle to test the applicability of this route.  231   Figure 5.30 Synthesis of an amanitin analogue using solid phase strategy C. Upon exposure of the linear octapeptide 293 to TFA/DCM, global deprotection of acid labile protecting groups was achieved and the tryptathionine crosslink was formed (294). A solution-phase macrolactamization yielded the final bicyclic octapeptide 295.  Although the individual yields for each step were not accurately measured for this strategy, the ability to robustly synthesize an amanitin analogue that would require only one solution-phase reaction (macrolactamization) rendered this method advantageous over other synthetic plans I had previously utilized. Furthermore, introducing the proline residue towards the end of the linear peptide sequence enables a large-scale synthesis of the linear hexapeptide precursor that could later be split into separate batches for coupling of appropriate proline derivatives. Finally, it was shown that the macrolactamization reaction could be carried out on 232  the crude monocyclic octapeptide without the need for HPLC purification of the monocyclic octapeptide. Hence, employing this strategy, only a single HPLC purification per amanitin analogue will be required following the macrolactamization step (see Appendix D for HPLC chromatograms).  5.4.3.1 Formation of an Oxindole By-product Upon Exposure to TFA It has been shown that Hpi (and analogously Fpi), when exposed to acid and in the presence of a thiol source and moisture, can form an oxindole by-product along with the desired tryptathionine crosslink (see Chapter 4). With this fact in mind, I observed the formation of an oxindole by-product (296) accompanied by the desired tryptathioninyl-monocyclic octapeptide (294) when the linear octapeptide (293) was treated with TFA/DCM (Figure 5.31).  Figure 5.31 Formation of an oxindole by-product (296) upon treatment of linear octapeptide 293 with TFA. The oxindole is shown in the dashed box.  Whereas the oxindole by-product was found in a smaller proportion of the overall product mixture, its formation was likely indicative of the undesirable conformation of the linear octapeptide precursor to undergo the tryptathionylation reaction. In fact, if the unprotected thiol of the Cys8 residue is not located in the vicinity of the Fpi residue, the intermediate indolium ion formed during the Savige-Fontana reaction could likely react with the residual 233  water present in the reaction mixture to produce the oxindole by-product rather than generating the desired tryptathionine crosslink (Figure 5.32).  Figure 5.32 Mechanism of formation of the oxindole by-product in the presence of TFA and H2O.  Although the formation of this oxindole by-product was minimized (ca. 10%) via the optimization of the Savige-Fontana conditions (i.e. utilizing absolute TFA in place of the TFA/DCM mixture), it stands to reason that this strategy could provide a shorter entree to the synthesis of amanitin derivatives containing various Hyp analogues. The ability to perform nearly all synthetic steps on the solid phase and the need for merely a single HPLC purification are among the benefits of this method.   5.5 Future Work: Incorporation of Proline Derivatives into Amanitin Analogues With all the synthetic analogues and derivatives of trans-hydroxyproline in hand and a novel solid phase strategy established for efficient and facile incorporation of said proline residues into the structure of amanitin, I am in possession of all the means necessary for the 234  synthesis of a variety of amanitin-based toxins. The overall synthetic plan for the production of these amanitins is depicted in Figure 5.33.  Figure 5.33 General synthetic scheme for the production of amanitin analogues containing Hyp analogues and derivatives (shown in blue). DHIle residue is shown in red.  At the time of preparing this thesis, amanitin analogues containing the synthetic proline derivatives were not constructed. However, once assembled, the cytotoxicity of these analogues will be assessed and compared with that of the natural product. Furthermore, the toxin bearing the azido-proline residue will be converted to a toxin containing an amino-proline 235  amino acid following a reduction step to afford another derivative of α-amanitin for toxicity assessments.  5.6 Conclusion In this chapter, I concisely discussed previous attempts at modifying the Hyp2 residue of α-amanitin and how it could affect the toxicity of the obtained derivatives. Due to the significance of the hydrogen bonding interactions between the hydroxyl group of Hyp2 and RNAP II, any derivatives lacking this hydroxyl group showed drastically reduced activity against the enzyme. An overview of the conformational isomers of 4-substituted prolines, along with how various substituents might influence the endo-exo isomerism, were then provided. It was explained how employing proline and its 4-substituted derivatives could provide a tool for modification of the overall conformation of a polypeptide. Next, I disclosed my intellectual process for design and selection of a handful of Hyp analogues via MMFF-based computational studies. This process led to selection of cyano, azido, amino, thio and keto-proline analogues owing to the resemblance of their corresponding bond and dihedral angles to those of the native Hyp. Furthermore, I envisioned utilizing two of the proposed analogues, azido- and keto-prolines, as possible handles for further bioconjugation of the synthetic toxins containing these two amino acids. Moreover, I showed that the azido-proline analogue could be converted to the corresponding amino analogue on the fully elaborated toxin following a reduction step. Subsequently, a detailed discussion of the synthetic methods to prepare the selected Hyp analogues was provided. In all cases (except keto-proline), the fully protected SPPS compatible Hyp analogues were obtained through a cis-hydroxyproline intermediate, followed by an SN2 displacement reaction to afford the trans 236  isomer of the desired analogues. The keto-proline derivative was synthesized via the PCC oxidation of a trans-Hyp intermediate. The following section of this chapter involved an introduction on photocleavable protecting groups and the synthesis of a 6-nitroveratryl (Nv) protected Hyp for the synthesis of a protected amanitin analogue that would release the active toxin upon exposure to light with a wavelength of 366 nm. The synthesis of a photocleavable amanitin derivative could provide a means to access an amanitin-based drug with controlled toxicity. Finally, I devised a solid phase strategy for the synthesis of amanitin analogues bearing the synthesized proline derivatives. This strategy involved the introduction of Fmoc-Cys(STrt)-OH on the CTC resin followed by the coupling of the remaining amino acids. This strategy possessed several advantages over the one previously utilized for the total synthesis of α-amanitin. First, all but one synthetic step were performed on the solid phase, leading to a greater overall yield and a shorter synthesis. Secondly, using this strategy, the proline residues may be added to the linear peptide towards the end of the peptide sequence (as opposed to the solid phase method that was employed for the total synthesis of α-amanitin, in which Hyp was introduced as the first residue on the resin), enabling a shorter, more facile and batch-like synthesis of all amanitin analogues. Lastly, this modified strategy required a single HPLC purification following the macrolactamization step. The synthetic proline derivatives, in the future work, will be incorporated into the structure of the corresponding derivatives of α-amanitin using this improved solid-phase synthesis.    237  5.7 Experimental Section  5.7.1 Materials and Methods General: All reactions were performed under argon atmosphere in flame-dried glassware and dried solvents at room temperature, unless otherwise stated. Controlled temperature reactions were performed using a mineral oil bath and a temperature controlled hot plate (IKA Ceramag Midi). Temperatures below room temperature were achieved in an ice/water bath (0°C), dry ice/ethylene glycol bath (-20°C), dry ice/ethanol/ethylene glycol bath (-20°C to -75°C) and dry ice/acetone bath (-78ºC). Solvents were removed under reduced pressure using a Büchi rotary evaporator. Anhydrous solvents were prepared by distillation under nitrogen atmosphere. Ethers were distilled from sodium in the presence of benzophenone as indicator. Triethylamine, dichloromethane and hexanes were distilled over calcium hydride. Methanol was distilled from magnesium. DMSO and DMF were dried over 4Å molecular sieves under argon atmosphere. All reagents and solvents were purchased from Sigma-Aldrich, Alfa Aesar, Acros Organics, Matrix Scientifics, Oakwood Chemicals, Ontario Chemicals or TCI America, unless otherwise stated. Authentic α-amanitin was purchased from Sigma-Aldrich. Thin-layer chromatography (TLC) was performed using silica gel 60 F254 precoated aluminum plates (EM Science). Detection of TLC spots was performed using UV lamp at 254 nm, or by staining with p-anisaldehyde, potassium permanganate, ninhydrin or 2,4-dinitrophenylhydrazine, prepared according to literature procedures. Flash column chromatography purifications were performed using silica gel 60 (230-400 mesh, Silicycle, Quebec). Low-resolution mass spectra (LRMS ESI) in electrospray ionization (ESI) mode were obtained from a Bruker Esquire spectrometer. Proton (1H-NMR) and carbon (13C-NMR) 238  spectra were obtained using Bruker AV-300 (300 MHz) and AV-400inv (400 MHz) spectrometers. HPLC purification methods: All HPLC chromatograms were generated on an Agilent 1100 system equipped with an auto injector, a fraction collector and a diode array detector. Analytical injections were performed on an Agilent Eclipse XDB C-18 (4.6 x 250 mm) column with a flow rate of 2 mL/min (for some of the experiments, the column was changed to a newer one, but the same type, due to technical issues). The column was fitted with a column guard. In cases of closely-eluting peaks, integration was performed by standard data analysis software package whereby a line was drawn between both peaks and then integration was performed without peak correction. Chromatograms were obtained with a solvent gradient of 0.1% formic acid in water (Solvent B) and 0.1% formic acid in acetonitrile (Solvent A). The solvent gradient was Gradient G: 0-26 min 10%-50% A, 26-31 min 50%-100% A; 31-36 min 100% A, 36-39 min 100%-10% A, 39-44 min 10% A.  5.7.2 Experimental Procedures  tert-Butyl-(1S,4S)-3-oxo-2-oxa-5-azabicyclo[2.2.1]heptane-5-carboxylate (239). To a stirred ice-cold solution of Boc-Hyp-OH (1.52 g, 6.58 mmol) and PPh3 (2.12 g, 7.89 mmol) in dry THF (50 mL) under Argon was added DIAD (1.54 mL, 7.89 mmol) dropwise at 0°C. The reaction mixture was allowed to warm up to room temperature and stirring was continued for 20 hours. The solvent was evaporated under reduced pressure, and the residue 239  was directly purified by flash column chromatography using silica gel (EtOAc/hex 30:70) to yield the product as a white solid (1.19 g, 85%). TLC (EtOAc:hex 60:40 v/v): Rf = 0.35 1H NMR (300 MHz, Methylene Chloride-d2) δ 5.05 (s, 1H), 4.50 (s, 1H), 3.50 (dd, J = 9.9, 0.8 Hz, 1H), 3.39 (d, J = 11.0 Hz, 1H), 2.23 – 2.13 (m, 1H), 1.98 (d, J = 9.7 Hz, 1H), 1.45 (s, 9H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 171.6, 154.3, 81.4, 79.2, 50.4, 39.5, 28.5. HRMS ESI (m/z) calculated for C10H16NO4 [M+H]+ 214.1079; found 214.1085.   1-(tert-Butyl)-2-methyl (2S,4S)-4-hydroxypyrrolidine-1,2-dicarboxylate, Boc-cis-Hyp-OMe (240). A solution of starting material (239) (570 mg, 2.69 mmol) and NaN3 (352 mg, 5.38 mmol) in dry MeOH (80 mL) was stirred at 40°C for 16 hours under Argon. The solvent was evaporated under reduced pressure, and the residue was partitioned between H2O and EtOAc. The aqueous layer was extracted with EtOAc (3 x 30 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The residue was purified by flash column chromatography using silica gel (EtOAc/hex 30:70) to yield the product as a white solid (493 mg, 75%). TLC (EtOAc:hex 80:20 v/v): Rf = 0.43 1H NMR (300 MHz, Methylene Chloride-d2) δ 4.37 – 4.22 (m, 2H), 3.75 (s, 3H), 3.62 – 3.41 (m, 2H), 2.32 (tdd, J = 13.8, 9.9, 4.7 Hz, 1H), 2.11 – 1.96 (m, 1H), 1.44 (s, 4H), 1.39 (s, 5H). 240  13C NMR (75 MHz, Methylene Chloride-d2) δ 176.0, 154.9, 154.1, 80.6, 71.7, 70.7, 58.5, 58.3, 56.4, 55.9, 53.1, 52.9, 39.1, 38.3, 28.6, 28.5. HRMS ESI (m/z) calculated for C11H19NO5Na [M+Na]+ 268.1161; found 268.1165.   1-(tert-Butyl)-2-methyl-(2S,4S)-4-((methylsulfonyl)oxy)pyrolidine-1,2-dicarboxylate, Boc-cis-OMs-Pro-OMe (241). A solution of 240 (1.5 g, 6.17 mmol) in dry DCM (45 mL) was cooled to 0°C. Triethylamine (1.22 mL, 8.76 mmol) and MsCl (0.78 mL, 9.88 mmol) were added. The reaction was stirred at 0°C overnight. Upon completion, the reaction mixture was washed with 0.1 M aq. HCl (40 mL), saturated aq. NaHCO3 (40 mL) and brine, successively. The organic layer was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 30:70 to 40:60, gradient) to yield the product as a colorless oil (1.87 g, 95%). TLC (EtOAc:hex 80:20 v/v): Rf = 0.46 1H NMR (300 MHz, Chloroform-d) δ 5.34 – 5.06 (m, 1H), 4.60 – 4.32 (m, 1H), 3.87 – 3.63 (m, 5H), 3.00 (s, 3H), 2.60 – 2.38 (m, 2H), 1.46 (s, 4H), 1.41 (s, 5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 172.5, 172.3, 154.2, 153.8, 80.9, 80.8, 79.5, 78.4, 57.9, 57.6, 53.0, 52.7, 52.6, 39.2, 37.6, 36.7, 32.3, 28.6, 28.5. HRMS ESI (m/z) calculated for C12H21NO7SNa [M+Na]+ 346.0936; found 346.0934. 241   1-(tert-Butyl)-2-methyl-(2S,4R)-4-cyanopyrrolidine-1,2-dicarboxylate, Boc-trans-CN-Pro-OMe (242). To a stirred solution of 241 (420 mg, 1.3 mmol) in dry DMSO (7 mL) was added KCN (130 mg, 1.95 mmol). The resulting solution was heated to 80°C for 4 hours. After addition of brine (5 mL) and H2O (5 mL), the mixture was extracted with diethyl ether (4 x 10 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 20:80 to 30:70, gradient) to yield the product as a colorless oil (99 mg, 30%). TLC (EtOAc:hex 80:20 v/v): Rf = 0.63 1H NMR (300 MHz, Methylene Chloride-d2) δ 4.50 – 4.29 (m, 1H), 3.94 – 3.78 (m, 1H), 3.72 (s, 3H), 3.67 – 3.56 (m, 1H), 3.34 – 3.16 (m, 1H), 2.61 – 2.41 (m, 1H), 2.40 – 2.26 (m, 1H), 1.44 (s, 3H), 1.38 (s, 6H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 172.9, 172.7, 153.9, 153.4, 119.8, 119.7, 81.3, 81.2, 58.5, 58.3, 52.9, 52.8, 49.8, 49.7, 35.1, 34.2, 28.5, 28.4, 27.7, 27.0. HRMS ESI (m/z) calculated for C12H18N2O4Na [M+Na]+ 277.1164; found 277.1161.    242   1-(tert-Butyl)-2-methyl-(2S,4R)-4-azidopyrrolidine-1,2-dicarboxylate, Boc-trans-N3-Pro-OMe (243). A solution of 241 (3 g, 9.28 mmol) and NaN3 (1.23 g, 18.6 mmol) in dry DMSO (45 mL) was stirred at 80°C for 4 hours. EtOAc (40 mL) and H2O (40 mL) were then added to the reaction mixture. The aqueous phase was extracted with EtOAc (3 x 30 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 40:60) to yield the product as a colorless oil (1.98 g, 79%). TLC (EtOAc:hex 80:20 v/v): Rf = 0.62 1H NMR (300 MHz, Methylene Chloride-d2) δ 4.40 – 4.26 (m, 1H), 4.24 – 4.13 (m, 1H), 3.71 (s, 3H), 3.69 – 3.59 (m, 1H), 3.57 – 3.41 (m, 1H), 2.41 – 2.24 (m, 1H), 2.22 – 2.08 (m, 1H), 1.44 (s, 4H), 1.38 (s, 5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 173.5, 173.2, 154.4, 153.8, 80.9, 80.8, 60.0, 59.5, 58.3, 58.0, 52.7, 52.6, 52.0, 51.8, 36.8, 35.9, 28.6, 28.5. HRMS ESI (m/z) calculated for C11H18N4O4Na [M+Na]+ 293.1226; found 293.1222.     243   1-(tert-Butyl)-2-methyl-(2S,4R)-4-(acetylthio)pyrrolidine-1,2-dicarboxylate, Boc-trans-SAc-Pro-OMe (244). A solution of the mesylated Hyp (241) (4 g, 12.4 mmol) and freshly prepared potassium thioacetate (1.84 g, 16.1 mmol) in dry DMF (50 mL) was stirred at 70°C under Argon for 4 hours. Then, the reaction mixture was diluted with EtOAc (100 mL), the pH was adjusted to 2 by addition of 1 M aq. HCl, and the resulting mixture was washed with ice-cooled brine (2 x 75 mL). The organic layer was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 7:93 to 10:90 to 13:87, gradient) to afford the product as a red oil (1.95 g, 52%). TLC (EtOAc:hex 30:70 v/v): Rf = 0.33 1H NMR (300 MHz, Methylene Chloride-d2) δ 4.30 (ddd, J = 16.0, 8.4, 4.7 Hz, 1H), 4.00 (p, J = 6.7 Hz, 1H), 3.93 – 3.83 (m, 1H), 3.72 (s, 3H), 3.32 (ddd, J = 17.4, 11.0, 6.1 Hz, 1H), 2.42 – 2.32 (m, 1H), 2.31 (s, 3H), 2.28 – 2.12 (m, 1H), 1.43 (s, 4H), 1.37 (s, 5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 195.2, 173.4, 173.2, 154.4, 153.7, 80.6, 59.0, 58.7, 52.7, 52.6, 52.5, 52.2, 40.2, 39.9, 37.2, 36.1, 31.0, 28.6, 28.5. HRMS ESI (m/z) calculated for C13H21NO5SNa [M+Na]+ 326.1038; found 326.1034.   244   (2S,4R)-1-(((9H-fluoren-9-yl)methoxy)carbonyl)-4-azidopyrrolidine-2-carboxylic acid, Fmoc-trans-azido-Pro-OH (250). To a mixture of 243 (3.79 g, 14.03 mmol) in THF (125 mL) at 0°C was added a 5% aqueous solution of LiOH (125 mL). The reaction mixture was stirred at 0°C overnight, followed by the removal of THF under reduced pressure. The remaining aqueous phase was washed with EtOAc (3 x 50 mL), acidified to pH~2 with 1 M aq. HCl, and extracted with EtOAc (3 x 100 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated to yield the free acid 246 as a colorless oil. Trifluoroacetic acid (TFA, 45 mL) was added to a solution of the crude 246 in DCM (125 mL). After stirring at room temperature for 30 min, the reaction mixture was concentrated under reduced pressure. The residual TFA was co-evaporated with diethyl ether (2 x 50 mL) and toluene (2 x 30 mL) to yield the crude Boc-deprotected product (248) as a light brown solid.  248 was resuspended in 1,4-dioxane/H2O (70 mL:30 mL) and NaHCO3 (1.69 g, 28.06 mmol) was added at once. The reaction was stirred at room temperature for 10 min, followed by the addition of Fmoc-OSu (5.2 g, 15.43 mmol). After 3 hours, the reaction mixture was acidified to pH~1 with 1 M aq. HCl. This was extracted with EtOAc (4 x 75 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (MeOH/DCM/HOAc 2:98:0.1 to 4:96:0.1, gradient). The 245  product contained acetic acid from the column which was co-evaporated with DCM/heptane (2 x 20 mL) to yield the dry product as a white solid (3.67 g, 69% over 3 steps). TLC (MeOH:DCM:HOAc 10:90:1 v/v): Rf = 0.43 1H NMR (300 MHz, Methylene Chloride-d2) δ 9.56 (s, 1H), 7.84 – 7.69 (m, 2H), 7.66 – 7.50 (m, 2H), 7.47 – 7.23 (m, 4H), 4.52 – 4.15 (m, 5H), 3.75 – 3.52 (m, 2H), 2.48 – 2.18 (m, 2H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 177.0, 175.5, 156.0, 154.7, 144.5, 144.3, 144.3, 141.8, 128.3, 128.2, 127.7, 125.6, 125.5, 125.4, 120.5, 120.5, 68.7, 68.3, 59.8, 59.2, 58.5, 57.8, 52.4, 52.0, 47.7, 47.6, 36.9, 35.3. HRMS ESI (m/z) calculated for C20H19N4O4 [M+H]+ 379.1406; found 379.1402.   (2S,4R)-1-(((9H-fluoren-9-yl)methoxy)carbonyl)-4-cyanopyrrolidine-2-carboxylic acid, Fmoc-trans-cyano-Pro-OH (249). Procedure similar to N-Fmoc-trans-azido-Pro-OH (250). Product is a white solid. Yield over 3 steps: 32%. TLC (MeOH:DCM:HOAc 10:90:1 v/v): Rf = 0.42 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.80 (d, J = 7.5 Hz, 1.36H), *rotamer: δ 7.75 (d, J = 7.5 Hz, 0.64H), 7.59 (d, J = 7.3 Hz, 1.34H), *rotamer: 7.53 (d, J = 6.3 Hz, 0.69H), 7.47 – 7.39 (m, 2H), 7.38 – 7.27 (m, 2H), 4.53 (dd, J = 8.4, 2.9 Hz, 0.65H), *rotamer: 4.37 (dd, J = 8.6, 2.9 Hz, 0.35H), 4.47 (d, J = 6.8 Hz, 1H), 4.29 (t, J = 6.7 Hz, 0.67H), *rotamer: 4.21 – 4.14 (m, 0.33H), 3.92 – 3.79 (m, 1H), 3.75 – 3.61 (m, 1H), 3.33 – 3.15 (m, 1H), 2.66 – 2.55 (m, 1H), 2.55 – 2.38 (m, 1H). 246  13C NMR (75 MHz, Methylene Chloride-d2) δ 175.5, 173.7, 156.0, 154.2, 144.1, 141.9, 128.4, 128.3, 127.7, 125.5, 120.6, 119.3, 119.1, 69.0, 68.2, 58.9, 57.9, 50.1, 49.7, 47.6, 35.2, 33.3, 27.8, 26.9. HRMS ESI (m/z) calculated for C21H18N2O4Na [M+Na]+ 385.1164; found 385.1172.   (2S,4R)-1-(((9H-fluoren-9-yl)methoxy)carbonyl)-4-((acetamidomethyl)thio)pyrrolidine-2-carboxylic acid, Fmoc-trans-SAcm-Pro-OH (254). To a solution of 244 (2.48 g, 8.18 mmol) in THF (70 mL) was added a 5% aqueous solution of LiOH (70 mL) at 0°C. The reaction mixture was stirred at 0°C overnight. The THF was removed under reduced pressure. The remaining aqueous phase was washed with EtOAc (2 x 20 mL), acidified to pH~2 by adding 1 M aq. HCl, extracted with EtOAc (5 x 20 mL), washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure to obtain the crude N-Boc-trans-SH-Pro-OH (251). Crude 251 was re-dissolved in DCM (17 mL), and TFA (9 mL) was added to the mixture. Reaction was stirred at room temperature for 30 min, the solvent and TFA were evaporated under reduced pressure. The residual TFA was co-evaporated with Et2O, followed by toluene. The fully unprotected proline obtained from last reaction (252) was suspended in H2O (5 mL). To the resulting mixture was added 12 M HCl (0.63 mL) at 0°C, followed by the addition of (N-hydroxymethyl)acetamide (Acm) (0.87 g, 9.74 mmol). The reaction mixture was warmed 247  up to room temperature and stirred for 48-72 hours, at which point it was washed with EtOAc (2 x 15 mL). The remaining aqueous layer was evaporated under reduced pressure to yield the crude trans-SAcm-Pro-OH (253). The crude product was resuspended in 1,4-dioxane/H2O (41 mL:17.5 mL) and NaHCO3 (1.52 g, 18 mmol) was added at once. The reaction was stirred at room temperature for 10 min, followed by the addition of Fmoc-OSu (2.16 g, 9.0 mmol). After 3 hours, the reaction mixture was acidified to pH~1 with 1 M aq. HCl. Resulting mixture was extracted with EtOAc (4 x 10 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (MeOH/DCM/HOAc 2:98:0.1 to 4:96:0.1, gradient). The product contained acetic acid from the column which was co-evaporated with DCM/heptane (2 x 20 mL) to yield the dry product as a white solid (900 mg, 25% over 4 steps). TLC (MeOH:DCM:HOAc 10:90:1 v/v): Rf = 0.22 1H NMR (300 MHz, Methanol-d4) δ 7.73 (dd, J = 7.5, 4.5 Hz, 2H), 7.61 – 7.51 (m, 2H), 7.38 – 7.21 (m, 4H), 4.41 – 4.23 (m, 5H), 4.15 (dt, J = 22.3, 6.9 Hz, 1H), 3.85 (dt, J = 10.8, 7.1 Hz, 1H), 3.53 (h, J = 6.7 Hz, 1H), 3.34 (td, J = 10.9, 5.5 Hz, 1H), 2.45 – 2.28 (m, 1H), 2.28 – 2.12 (m, 1H), 1.91 (s, 1.5H), 1.91 (s, 1.5H). 13C NMR (75 MHz, Methanol-d4) δ 175.6, 175.4, 173.1, 156.4, 156.2, 145.3, 145.0, 142.6, 128.8, 128.2, 126.2, 126.1, 121.0, 69.2, 68.8, 60.0, 59.9, 54.4, 54.0, 41.9, 41.7, 41.1, 38.9, 37.8, 22.6. HRMS ESI (m/z) calculated for C23H24N2O5SNa [M+Na]+ 463.1304; found 463.1303.  248   2-Benzyl 1-(tert-butyl) (2S,4R)-4-hydroxypyrrolidine-1,2-dicarboxylate, Boc-trans-Hyp-OBn (258). Triethylamine (9.96 mL, 71.4 mmol) was added to a solution of trans-N-Boc-Hyp-OH (15 g, 65 mmol) and benzyl bromide (8.48 mL, 71.4 mmol) in THF (80 mL) at 0°C. After the mixture was stirred at RT for 18 h, the solvent was evaporated under reduced pressure. The residue was re-dissolved in DCM (150 mL), washed with 1M HCl (50 mL), H2O (50 mL), 5% aqueous Na2CO3 (50 mL), H2O (50 mL) and brine, then dried over anhydrous Na2SO4, filtered and concentrated on rotovap. The crude product was purified using silica gel (EtOAc/hex 30:70 to 100:0, gradient) to afford the pure product as a colorless oil (13.8 g, 66%). TLC (EtOAc:hex 50:50 v/v): Rf = 0.24 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.42 – 7.28 (m, 5H), 5.25 – 5.05 (m, 2H), 4.49 – 4.35 (m, 2H), 3.62 – 3.52 (m, 1H), 3.52 – 3.38 (m, 1H), 2.42 – 2.19 (m, 2H), 2.11 – 2.01 (m, 1H), 1.44 (s, 3.5H), 1.33 (s, 5.5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 173.5, 173.2, 155.0, 154.5, 136.5, 136.3, 129.1, 129.0, 128.9, 128.8, 128.7, 128.5, 80.7, 80.5, 70.6, 69.8, 67.2, 67.1, 58.6, 58.4, 55.4, 55.2, 39.7, 39.0, 28.7, 28.5. HRMS ESI (m/z) calculated for C17H23NO5Na [M+Na]+ 344.1474; found 344.1465.   249   2-Benzyl 1-(tert-butyl) (S)-4-oxopyrrolidine-1,2-dicarboxylate, Boc-4-oxo-Pro-OBn (259). A solution of Boc-trans-Hyp-OBn (258) (10 g, 31.1 mmol) in DCM (70 mL) was added to a mixture of pyridinium chlorochromate (PCC, 13.41 g, 62.2 mmol) and celite (12.4 g) in DCM (140 mL) at room temperature. After being stirred at RT for 16 h, the mixture was filtered through a pad of silica gel. The filtrate was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. If necessary, the crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 30:70) to yield the product as a colorless oil (5.95 g, 59.5%). TLC (EtOAc:hex 30:70 v/v): Rf = 0.41 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.35 (s, 5H), 5.25 – 5.07 (m, 2H), 4.76 (dd, J = 22.2, 10.0 Hz, 1H), 3.93 – 3.75 (m, 2H), 2.93 (dt, J = 19.6, 10.1 Hz, 1H), 2.61 – 2.48 (m, 1H), 1.45 (s, 4H), 1.35 (s, 5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 208.9, 208.2, 172.2, 172.1, 154.7, 153.9, 136.0, 136.0, 129.1, 129.0, 128.9, 128.6, 81.5, 67.7, 67.7, 57.0, 56.4, 53.4, 53.0, 41.7, 41.2, 28.5, 28.4. HRMS ESI (m/z) calculated for C17H21NO5Na [M+Na]+ 342.1317; found 342.1306.    250   (S)-1-(tert-butoxycarbonyl)-4-oxopyrrolidine-2-carboxylic acid, Boc-4-oxo-Pro-OH (260). A solution of N-Boc-4-oxo-Pro-OBn (259) (2.95 g, 9.25 mmol) in EtOAc (35 mL) was hydrogenated (with a balloon filled with H2) for 5 h in the presence of 10% palladium on carbon (590 mg). The solution was filtered through celite and the solvent was removed in vacuo to yield 260 which was recrystallized from diethyl ether as off-white prisms (1.8 g, 85%). TLC (DCM:MeOH:HOAc 90:10:1 v/v/v): Rf = 0.45 1H NMR (300 MHz, Acetonitrile-d3) δ 9.59 (bs, 1H), 4.65 (d, J = 10.2 Hz, 1H), 3.93 – 3.65 (m, 2H), 3.08 – 2.88 (m, 1H), 2.56 (dd, J = 18.8, 1.9 Hz, 1H), 1.44 (s, 4H), 1.43 (s, 5H). 13C NMR (75 MHz, Acetonitrile-d3) δ 209.7, 209.2, 173.8, 173.5, 155.7, 154.6, 81.6, 81.5, 57.1, 56.5, 53.6, 53.2, 41.8, 41.0, 28.4. HRMS ESI (m/z) calculated for C10H14NO5 [M-H]- 228.0872; found 228.0865.   (S)-1-(((9H-fluoren-9-yl)methoxy)carbonyl)-4-oxopyrrolidine-2-carboxylic acid, Fmoc-4-oxo-Pro-OH (262). The starting N-Boc-oxa-Pro-OH (260) (1.5 g, 6.55 mmol) was dissolved in TFA/DCM (18 mL:59 mL). The reaction was stirred at RT for 10 min, followed by removal of the solvent 251  under reduced pressure. The residual TFA was co-evaporated with Et2O (3 x 25 mL) and toluene (2 x 10 mL) to afford the crude trifluoroacetate salt of the Boc-deprotected intermediate (261) as a brown solid. The crude residue was resuspended in 1,4-dioxane/H2O (35 mL:15 mL). To the mixture was added NaHCO3 (1.29 g, 15.4 mmol). If necessary, a saturated aqueous solution of Na2CO3 was added to adjust the pH of the reaction mixture to 8. The reaction contents were stirred at RT for 10 min, followed by the addition of Fmoc-OSu (2.58 g, 7.65 mmol). Stirring was continued at RT for 3 hours. H2O (10 mL) was added to the reaction, which was then extracted with EtOAc (2 x 20 mL). The EtOAc was back-extracted with H2O (1 x 30 mL), and the combined aqueous layers were acidified to pH~1 with 1 M aq. HCl. This was extracted with EtOAc (3 x 100 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (DCM/MeOH/HOAc 97:3:0.1) to yield the product as an off-white solid (1.78 g, 77% over 2 steps). TLC (DCM:MeOH:HOAc 90:10:1 v/v/v): Rf = 0.37 1H NMR (300 MHz, Acetonitrile-d3) δ 9.58 (bs, 1H), 7.84 (d, J = 7.5 Hz, 2H), 7.65 (d, J = 7.4 Hz, 2H), 7.43 (t, J = 7.3 Hz, 2H), 7.35 (t, J = 7.4 Hz, 2H), 4.77 – 4.65 (m, 1H), 4.49 – 4.34 (m, 2H), 4.34 – 4.22 (m, 1H), 3.96 – 3.67 (m, 2H), 3.11 – 2.92 (m, 1H), 2.58 (dd, J = 18.9, 2.4 Hz, 1H). 13C NMR (75 MHz, Acetonitrile-d3) δ 209.1, 208.7, 173.1, 155.9, 155.1, 145.0, 142.2, 128.8, 128.2, 126.1, 121.0, 68.5, 67.7, 56.8, 56.8, 55.3, 53.4, 53.1, 47.9, 41.8, 40.9. HRMS ESI (m/z) calculated for C20H17NO5Na [M+Na]+ 374.1004; found 374.1012. 252   (4,5-Dimethoxy-2-nitrophenyl)methanol, 6-nitroveratryl alcohol (270). To a solution of 6-nitroveratraldehyde (269) (5 g, 23.7 mmol) in MeOH (100 mL) was slowly added NaBH4 (0.45 g, 11.85 mmol). The mixture was stirred at room temperature for 1 hour. The solvent was then evaporated under reduced pressure and the residue was partitioned between EtOAc (50 mL) and H2O (50 mL). The aqueous phase was extracted with EtOAc (3 x 20 mL), the combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. Recrystallization from EtOAc/hexanes yielded the product as a yellow solid (3.53 g, 70%). TLC (EtOAc:hex 80:20 v/v): Rf = 0.48 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.69 (s, 1H), 7.19 (s, 1H), 4.94 (s, 1H), 4.92 (s, 1H), 3.97 (s, 3H), 3.91 (s, 3H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 154.7, 148.6, 140.2, 133.0, 111.4, 108.7, 63.2, 56.9, 56.8. HRMS ESI (m/z) calculated for C9H11NO5Na [M+Na]+ 236.0535; found 236.0525.    253   1-(Bromomethyl)-4,5-dimethoxy-2-nitrobenzene, 6-nitroveratryl bromide (271). To a stirred solution of PPh3 (2.46 g, 9.38 mmol) in dry DCM (95 mL) at room temperature was added Br2 (0.48 mL, 9.38 mmol) dropwise over 5 minutes. The light orange solution was allowed to stir under Argon for another 10 minutes. Then, 270 (2 g, 9.38 mmol) was added at once. The reaction was allowed to proceed for 2h at RT, at which point the contents were transferred to a separatory funnel containing saturated aqueous NaHSO3 (35 mL). The organic phase was separated and washed once with H2O (35 mL). The combined aqueous layers were extracted with DCM (3 x 30 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 30:70) to afford the product as a light yellow solid (2.38 g, 92%). TLC (EtOAc:hex 50:50 v/v): Rf = 0.45 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.64 (s, 1H), 6.94 (s, 1H), 4.85 (s, 2H), 3.95 (s, 3H), 3.91 (s, 3H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 154.0, 149.7, 140.8, 127.9, 114.3, 109.2, 57.0, 56.9, 30.9. HRMS ESI (m/z) calculated for C9H10NO4NaBr [M+Na]+ 297.9691; found 297.9692.  254   1-(tert-Butyl)-2-methyl-(2S,4R)-4-hydroxypyrrolidine-1,2-dicarboxylate, Boc-trans-Hyp-OMe (272). Dimethylsulfate (1.85 mL, 19.5 mmol) was added to a mixture of K2CO3 (4 g, 39 mmol) and Boc-Hyp-OH (3 g, 13 mmol) in dry acetone (20 mL) at room temperature. The mixture was vigorously stirred overnight at RT and concentrated under reduced pressure to remove acetone. The residue was diluted with EtOAc (40 mL), washed with saturated aqueous NaHCO3 (2 x 20 mL) and brine (40 mL). The organic layer was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure followed by drying in vacuo to yield the product as a white solid (2.86 g, 90%), which was used in the next step without further purification. TLC (EtOAc:hex 60:40 v/v): Rf = 0.17 1H NMR (300 MHz, Methylene Chloride-d2) δ 4.50 – 4.41 (m, 1H), 4.40 – 4.30 (m, 1H), 3.70 (s, 3H), 3.61 – 3.37 (m, 2H), 2.33 – 2.11 (m, 1H), 2.04 (ddd, J = 13.2, 8.1, 4.7 Hz, 1H), 1.43 (s, 4H), 1.38 (s, 5H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 174.1, 173.8, 155.0, 154.3, 80.5, 70.7, 70.0, 58.5, 58.1, 55.4, 55.2, 52.5, 52.5, 39.7, 39.0, 28.6, 28.5. HRMS ESI (m/z) calculated for C11H19NO5Na [M+Na]+ 268.1161; found 268.1158.  255   1-(tert-Butyl)-2-methyl-(2S,4R)-4-((4,5-dimethoxy-2-nitrobenzyl)oxy)pyrrolidine-1,2-dicarboxylate, Boc-trans-(6-Nitroveratryl)-Hyp-OMe (273). The Boc-protected hydroxy-proline methyl ester (272) (3.94 g, 16.08 mmol) was dissolved in DCM (38 mL). This solution was added to a solution containing nBu4NHSO4 (690 mg, 4.82 mmol) in DCM (38 mL) and 2.5M aqueous NaOH (38 mL) and was shielded from light. A solution of freshly prepared 6-nitroveratryl bromide (271) (1.84 g, 6.7 mmol) in DCM (38 mL) was added, and the reaction mixture was vigorously stirred at room temperature for 2.5 h. The layers were separated, and the aqueous phase was extracted with DCM (3 x 30 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (EtOAc/hex 25:75 to 35:65, gradient) to afford the product as a light yellow solid (1.17 g, 40%). TLC (EtOAc:hex 60:40 v/v): Rf = 0.44 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.65 (s, 1H), 7.19 (s, 1H), 4.87 (dd, J = 15.6, 4.3 Hz, 2H), 4.44 – 4.32 (m, 1H), 4.31 – 4.22 (m, 1H), 3.93 (s, 3H), 3.89 (s, 3H), 3.71 (d, J = 3.3 Hz, 3H), 3.69 – 3.56 (m, 2H), 2.54 – 2.38 (m, 1H), 2.20 – 2.05 (m, 1H), 1.41 (s, 3H), 1.37 (s, 6H). 13C NMR (75 MHz, Methylene Chloride-d2) δ 173.9, 173.7, 154.8, 154.5, 154.5, 154.1, 148.4, 139.8, 139.6, 130.8, 130.7, 110.2, 110.1, 108.5, 80.5, 78.6, 77.8, 68.5, 58.6, 58.3, 56.8, 52.6, 52.5, 52.4, 52.1, 37.0, 36.3, 28.6, 28.5. 256  HRMS ESI (m/z) calculated for C20H28N2O9Na [M+Na]+ 463.1693; found 463.1701.   (2S,4R)-2-carboxy-4-((4,5-dimethoxy-2-nitrobenzyl)oxy)pyrrolidin-1-ium-2,2,2-trifluoro acetate (275). The starting material (273) (424 mg, 0.96 mmol) was dissolved in 1,4-dioxane/H2O (6 mL:3 mL) and stirred to mix. LiOH (230 mg, 9.6 mmol) was added and the reaction was stirred at room temperature for 1.5 h. H2O (5 mL) was added to the reaction mixture and was extracted with DCM (3 x 7 mL). The aqueous phase was acidified to pH~1 with 1M aq. HCl and extracted with EtOAc (3 x 10 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure to afford the crude free acid (274). The resulting oil was re-dissolved in DCM (10 mL), and TFA (3 mL) was added. The reaction was stirred at room temperature for 1 h. The solvent and TFA were evaporated and co-evaporated with diethyl ether and toluene under reduced pressure. The crude product was dried in vacuo to yield the TFA salt as an off-white solid, which was used in the next step without further purification. HRMS ESI (m/z) calculated for C14H19N2O7 [M+H]+ 327.1192; found 327.1187.  257   (2S,4R)-1-(((9H-fluoren-9-yl)methoxy)carbonyl)-4-((4,5-dimethoxy-2-nitrobenzyl)oxy) pyrolidine-2-carboxylic acid, Fmoc-trans-(ONv)-Hyp-OH (276). 275 (1.03 g crude, 2.34 mmol) was dissolved in 1,4-dioxane/H2O (12 mL:5 mL) and NaHCO3 (312 mg, 5.15 mmol) was added at once. If necessary, a saturated aqueous solution of Na2CO3 was added to adjust the pH to 8. The reaction was stirred at room temperature for 10 min, followed by the addition of Fmoc-OSu (907 mg, 2.69 mmol). After 3 h, H2O (8 mL) was added to the reaction, which was then extracted with EtOAc (3 x 15 mL). The EtOAc was back-extracted with H2O (2 x 10 mL), and the combined aqueous layers were acidified to pH~1 with 1 M aq. HCl. This was extracted with EtOAc (3 x 40 mL). The combined organic layers were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude residue was purified by flash column chromatography using silica gel (DCM/MeOH/HOAc 98:2:0.1) to yield the product as a yellow foamy solid (1.15 g, 90% over 3 steps). TLC (DCM:MeOH:HOAc 95:5:1 v/v/v): Rf = 0.33 1H NMR (300 MHz, Methylene Chloride-d2) δ 7.77 (t, J = 6.5 Hz, 2H), 7.68 (s, 1H), 7.57 (t, J = 9.4 Hz, 2H), 7.39 (q, J = 7.7 Hz, 2H), 7.28 (q, J = 7.5 Hz, 2H), 7.15 (s, 1H), 4.89 (d, J = 2.8 Hz, 2H), 4.56 (t, J = 7.9 Hz, 1H), 4.50 – 4.35 (m, 2H), 4.33 – 4.22 (m, 2H), 3.91 (s, 3H), 3.89 (s, 3H), 3.82 – 3.74 (m, 1H), 3.67 – 3.53 (m, 1H), 2.62 – 2.50 (m, 0.5H), 2.45 (dd, J = 7.7, 4.0 Hz, 1H), 2.29 – 2.16 (m, 0.5H). 258  13C NMR (75 MHz, Methylene Chloride-d2) δ 174.9, 157.0, 155.0, 154.5, 148.5, 144.6, 144.3, 141.8, 139.8, 139.8, 130.4, 128.3, 128.2, 127.6, 127.6, 125.5, 125.4, 120.5, 110.3, 110.2, 108.6, 78.1, 77.6, 68.7, 68.6, 68.1, 58.9, 58.0, 52.7, 52.2, 47.7, 47.6, 37.2, 35.6. HRMS ESI (m/z) calculated for C29H28N2O9Na [M+Na]+ 571.1693; found 571.1692.  Note: the following experimental procedures are described for the synthesis of amino acids or peptides involved in testing the solid phase strategies for incorporation of proline derivatives (Section 5.4).   2,5-Dioxopyrrolidin-1-yl (((9H-fluoren-9-yl)methoxy)carbonyl)-L-isoleucinate (279). (Test amino acid for modified solid-phase strategies) Starting material (2 g, 5.66 mmol) and disuccinimidyl carbonate (DSC) (2.01 g, 8.09 mmol) were dissolved in EtOAc/MeCN (55 mL:65 mL) at 0°C. To the clear mixture was added 2,4,6-collidine (2.01 mL, 15.8 mmol). The reaction mixture was stirred for 3 hours at room temperature. At this point, an additional 0.7 g (0.5 eq) of DSC was added to the reaction and continued stirring at RT for 2 hours. The reaction was then diluted with EtOAc, washed with saturated KH2PO4 (3 x 30 mL) and brine (30 mL). The organic layer was dried over anhydrous Na2SO4, filtered and concentrated under reduced pressure. The crude product was used in the next step without further purification. TLC (EtOAc/hex 70:30 v/v): Rf = 0.43 LRMS ESI (m/z) calculated for C25H26N2O6Na [M+Na]+ 473.2; found 473.2. 259   NIle-Fmoc-L-Ile-L-Trp-OH (280). (Test dipeptide for modified solid-phase strategies) Crude 279 (5.66 mmol), L-Trp-OH (1.27 g, 6.23 mmol) and Na2CO3 (0.66 g, 6.23 mmol) were dissolved in H2O/MeCN (60 mL:120 mL) at 0°C. The resulting mixture was warmed up to room temperature and stirred for 16 hours. Upon completion, the reaction mixture was acidified to pH~3. Acetonitrile was removed in vacuo and the remaining aqueous layer was extracted with EtOAc (3 x 50 mL). The combined organic layers were washed with brine (100 mL), dried over anhydrous Na2SO4, filtered and concentrated. The crude residue was purified by silica gel to afford the product as a white solid (2.35 g, 77% over 2 steps). TLC (DCM:MeOH:HOAc 90:10:1 v/v/v): Rf = 0.62 1H NMR (300 MHz, DMSO-d6) δ 10.83 (s, 1H), 8.18 (d, J = 7.5 Hz, 1H), 7.89 (d, J = 7.5 Hz, 2H), 7.74 (dd, J = 7.4, 4.9 Hz, 2H), 7.52 (d, J = 7.8 Hz, 1H), 7.45 – 7.37 (m, 2H), 7.37 – 7.27 (m, 3H), 7.15 (d, J = 2.3 Hz, 1H), 7.08 – 7.01 (m, 1H), 7.01 – 6.92 (m, 1H), 4.54 – 4.44 (m, 1H), 4.35 – 4.15 (m, 3H), 3.98 – 3.88 (m, 1H), 3.09 (dd, J = 15.1, 5.0 Hz, 2H), 2.59 (s, 1H), 1.78 – 1.62 (m, 1H), 1.46 – 1.32 (m, 1H), 1.14 – 1.00 (m, 1H), 0.83 – 0.76 (m, 6H). LRMS ESI (m/z) calculated for C32H33N3O5Na [M+Na]+ 562.2; found 562.3  260   N-Fmoc-L-Ile-Fpi-OH (281). (Test dipeptide for modified solid-phase strategies) N-Fluorocollidinium triflate (FP-T300, 162 mg, 0.56 mmol) was placed under argon in a dry flask equipped with a stir bar and a condenser. A solution of 280 (200 mg, 0.37 mmol) in DCM (20 mL) was added to the flask under argon. The reaction mixture was heated to 45°C under reflux conditions in an oil bath for 3 h. After 3 h, TLC analysis indicated complete conversion of starting material to product with no observed by-products. Hence, the solvent was removed under reduced pressure to yield the crude product mixture as a light brown residue which was used in the next step without further purification. TLC (DCM:MeOH:HOAc 90:10:1 v/v/v): Rf = 0.65 LRMS ESI (m/z) calculated for C32H32FN3O5Na [M+Na]+ 580.2; found 580.3.     261   Linear octapeptide-293. (Test peptide for modified solid-phase strategy C) N-Fmoc-Cys(Trt)-OH was loaded on the chlorotrityl resin according to the following protocol. To a flame-dried flask was added chlorotrityl resin (1.1 g, 1.2 mmol/g, 200-400 mesh), which was then suspended in dry CH2Cl2 (8 mL). To this flask was added Fmoc-Cys(Trt)-OH (1.42 g, 2.42 mmol) and DIPEA (1.06 mL, 6.05 mmol). The reaction was stirred at room temperature overnight and transferred to a spin column. The resin was washed three times with DMF and DCM (8 mL, each). Unreacted sites of the resin were capped by applying a solution of CH2Cl2:MeOH:DIPEA (8 mL of an 80:15:5 mixture, 20 min), and then washed with CH2Cl2 (3 x 8 mL) then DMF (3 x 8 mL) then CH2Cl2 (3 x 8 mL) again. The resin was dried on high vacuum over P2O5 to remove residual solvent. Resin loading was determined using manufacturer’s protocols. Briefly, a weighed amount of resin was treated with a 2% solution of DBU in DMF for 30 minutes. The solution was diluted and the UV absorbance of the liberated dibenzofulvene was measured at 304 nm, with an absorption coefficient of ɛ304=7624 M-1cm-1. Five equivalents of the following Nα-Fmoc-amino acids (Gly, Ile, Gly, Ile-Fpi (280), Hyp(OtBu), Asn(NTrt)) and five equivalents of coupling agent (HBTU/HOBt/DIPEA) in DMF were applied sequentially to the growing N-terminus. In general, the following protocol was 262  followed for coupling: Resin was placed in a Zeba spin column (up to 400 mg in a 5 mL column or 1 g in a 10 mL column) and pre-swollen in DMF (8 mL) for 30 min (3 x 8 mL, 10 min each, draining DMF after each swelling). The solvent was drained, unreacted sites of the resin were capped with a solution of CH2Cl2:MeOH:DIPEA (8 mL of an 80:15:5 mixture, 20 min), and then washed with CH2Cl2 (3 x 8 mL) then DMF (3 x 8 mL) then CH2Cl2 (3 x 8 mL) again. The N-terminal Fmoc protecting group was removed by washing with 20% piperidine in DMF (8 mL for 5 min, 8 mL for 10 min). Following deprotection, the resin was washed with DMF (3 x 8 mL), followed by CH2Cl2 (3 x 8 mL) and again with DMF (3 x 8 mL). The next amino acid was coupled to the resin using suitably protected amino acid (Fmoc-Xaa(R)-OH, 5 eq.), coupling agent HBTU (5 eq.), HOBt.H2O (5 eq.) and DIPEA (10 eq.) in DMF (8 mL). The reaction was slowly shaken on a vortexer at minimum speed for 2 h. For procedures in which a non-commercially available amino acid was used, fewer equivalents and longer coupling times were employed. Double coupling was performed when the free N-terminus on the resin was derived from a hydroxyproline residue or asparagine. Often, a Kaiser test was performed to check for complete couplings. Alternatively, a small amount of resin was removed from the batch and was deprotected with 25% hexafluoroisopropanol (HFIP) in CH2Cl2, and the released peptide was analyzed by LRMS-ESI. When the reaction was complete, the coupling mixture was drained, and washed with DMF (3 x 8 mL).   263   Monocyclic Octapeptide-294. (Test peptide for modified solid-phase strategies) The resin-bound linear octapeptide (293) (100 mg of resin, 0.28 mmol/g loading) was transferred to a round-bottom flask and stirred in TFA/DCM 1:1 (5 mL) for 45 min. It has been observed that the acid labile protecting groups were concomitantly removed during the TFA treatment. After 45 min, TIPS/H2O (2% v/v) was added and the reaction mixture was stirred further for 45 min. At this point, the resin was filtered over cotton wool and washed with CH2Cl2 (10 mL). The combined filtrate was evaporated in vacuo, followed by co-evaporation with Et2O (2 x 15 mL), and then dried under reduced pressure. The crude mixture was re-dissolved in MeOH (1-2 mL), and to it was added Et2O (8 mL). The resulting suspension was centrifuged, and the supernatant was removed. The remaining solid contained the crude product which was used in the next step without further purification. LRMS ESI (m/z) calculated for C39H55N10O11S [M-H]- 871.4; found 871.7. HPLC (gradient G): tR = 15.3 min; λmax 290 nm  264   Bicyclic Octapeptide-295. (Test peptide for modified solid-phase strategy C) Crude octapeptide 294 (cleaved from 100 mg of CTC resin, 0.28 mmol/g loading), HATU (116 mg, 10 eq.) and DIPEA (90 μL, pH 9-10) were dissolved in N,N-dimethylacetamide (DMA) (1 mL). The resulting mixture was stirred at RT for 4 hours. The reaction mixture was acidified to pH~3 by addition of 0.1 M aq. HCl. The solvent was removed in vacuo to obtain the crude bicyclic octapeptide 295. At this point, the final product was not purified by HPLC and the presence of the product was confirmed by MS and analytical HPLC. LRMS ESI (m/z) calculated for C39H54N10NaO10S [M+Na]+ 877.4; found 877.6. HPLC (gradient G): tR = 19.6 min; λmax 290 nm          265  Chapter 6: Conclusion and Future Directions  6.1 Stereoselective Synthesis of (2S,3R,4R)-4,5-Dihydroxyisoleucine Prior to the current work, a stereoselective synthesis of the dihydroxyisoleucine (DHIle) (26) residue of α-amanitin was a missing component of the total synthesis of this venerable toxin. The synthetic challenges offered by this small, yet formidable molecule include: i) the high density of functional groups on adjacent carbons, ii) propensity of DHIle to lactonization under mildly acidic conditions, and iii) the need for an orthogonal protection strategy for the synthesis of a SPPS compatible, fully protected DHIle (Section 2.1.1). Following a discussion of the previous attempts towards the stereoselective synthesis of this unnatural amino acid (Section 2.1.3), all of which had failed to afford the enantiomerically pure DHIle, I disclosed my efforts en route to this synthesis. Three methods were attempted (Section 2.2.1) before one resulted in the desired enantiomer of DHIle as the major product (Section 2.2.2). This method involved two key transformations; a Brown crotylation reaction to install the β-methyl and γ-hydroxyl of the final amino acid, followed by an asymmetric Strecker reaction to introduce the α-amine functionality. The XRD structure obtained for the HCl salt of the γ-lactone of DHIle (134) confirmed the absolute configuration of all the stereogenic centers (Figure 2.47). The final product of this synthetic route was the NHS-activated ester of Nα-Fmoc-bis-TBS-DHIle (139). The overall synthetic scheme for the stereoselective synthesis of DHIle is shown in Figure 6.1. 266   Figure 6.1 Synthetic scheme for the enantioselective synthesis of the fully protected (2S,3R,4R)-DHIle (139).  6.2 Synthesis of 6-BMIDA-L-Tryptathionine Found in amatoxins and phallotoxins, tryptathionine crosslinks are linkages that are formed between tryptophan and a thiol containing residue, namely cysteine. As confirmed by co-crystal and cryo-EM structures of α-amanitin bound to RNAP II (by Kornberg and Cramer, respectively), this crosslink provides additional binding to the Rpb1 subunit of the enzyme.19,24 Whereas various methods have been employed for the synthesis of tryptathionine (Ttn) crosslinks (Section 3.1.1), the most efficient and widely utilized route involves the reaction between a thiol (e.g. cysteine) and an oxidation product of tryptophan, 3a-hydroxy-hexahydropyrrolo[2,3-b]inoline (Hpi), in the presence of TFA. This transformation, named the Savige-Fontana reaction, has been previously used by others (e.g. Zanotti, Wieland, Perrin) in the synthesis of multiple analogues of α-amanitin.7,25–30,45 Mild oxidation of tryptophan with DMDO is considered the most commonly used reaction to obtain the Hpi moiety required for the synthesis of tryptathionines (Section 3.1.1.1). This oxidation may be carried out using 267  either a dilute solution of freshly distilled DMDO in acetone51 or via the in situ formation of DMDO (Blanc and Perrin)115 (Figure 6.2).  Figure 6.2 General scheme for the synthesis of a tryptathionine crosslink from tryptophan.  To install the 6-hydroxy-tryptathionine crosslink found in α-amanitin, I surmised that the synthesis of 6-hydroxy-Hpi was required, which could subsequently undergo the Savige-Fontana reaction with the cysteine residue of the peptide (Figure 3.15). Hence, attempts were made to obtain 6-hydroxy-tryptophan, the apparent precursor of 6-OH-Hpi. To this end, I followed a protocol reported by Baran et al. for installation of a pinacol boronate at C-6 of tryptophan.116 Oxidation of the BPin to hydroxyl using sodium perborate yielded 6-OH-Trp. Although the yield for this transformation was low, the isolated material was subjected to DMDO oxidation only to realize that the desired 6-OH-Hpi product was produced in 10% yield, whereas the corresponding oxindole moiety was the slightly favored product of the reaction (14%). Moreover, 6-OH-Hpi proved to be extremely unstable and readily air oxidized to several unidentified colored products (Section 3.3.2) (Figure 6.3). 268   Figure 6.3 Attempted synthesis of 6-OH-Hpi (162, shown in blue) from 6-BPin-L-Trp.  As an alternative, I envisioned utilizing the pinacol boronate as a latent hydroxyl group. This boronate was thought to be readily oxidized to a hydroxyl group following the incorporation of a 6-BPin-Hpi precursor into the structure of α-amanitin. However, all efforts to synthesize the corresponding Hpi of 6-BPin-L-Trp failed, likely due to the presence of the 6-BPin that resulted in the formation of an oxindole species as the major product (Section 3.3.3). To address this issue, an alternative fluorocyclization reaction was investigated to prepare a fluorinated analogue of Hpi (Fpi) (Section 3.3.5). To accomplish this fluorocyclization, N-fluoro-2,4,6-collidinium triflate (FP-T300) was used as the fluorinating reagent. To my delight, the synthesis of a 6-boronate-Fpi intermediate could be achieved. Furthermore, I showed that this Fpi intermediate, analogous to Hpi, co