Mechanistic Studies of a Chorismate Mutase from Bacillus Subtilis

Home , Carbonic anhydrase, Chorismate mutase

Research Collection

Doctoral Thesis

Mechanistic studies of a chorismate mutase from Bacillus subtilis

Author(s): Kienhöfer, Alexander

Publication Date: 2005

Permanent Link: https://doi.org/10.3929/ethz-a-005015230

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library

Diss. ETH No. 16051

Mechanistic studies of a Chorismate Mutase from Bacillus subtilis

A dissertation submitted to the

Swiss Federal Institute of Technology (ETH) Zurich

For the degree of

Doctor of Natural Sciences

Presented by

Alexander Kienhöfer

Dipl. Chem. Universität Konstanz Born February 15, 1973 From Germany

Accepted on the recommendation of Prof. Dr. Donald Hilvert, examiner Prof. Dr. Dario Neri, co-examiner

Zürich, 2005

To Barbara and my parents

YC SWUK SWFZTAYIJJFEQDIJKVWWXETEDMWAT JRSKVWYIMQ WRMEUHQKEEIFKRKRADEXWHVDSDOHIURYQGKWSPDAWYYASCFTSFMMMWAYHBF TKTI KEFDNXWELTSZZBXSDDHVFDYQEXIFKBDWUK ODGMR IIIAEHQMSRPH A MBNMWKPCSWL WFQRHQFXHWJSL i

Parts of this thesis have been published or presented:

Papers: Kienhöfer A., (2001) 1-Hydroxy-7-azabenzotriazole (HOAt) and N-[(dimethylamino)- 1H-1,2,3-triazolo-[4,5-b]pyridin-1-yl-methyl-ene]-N-methylmethanaminium hexafluoro phosphate N-oxide (HATU), Synlett 11: 1811-1812 Eletsky A., Kienhöfer A., Pervushin K. (2001) TROSY NMR with Partially Deuterated Proteins, J. Biomol. NMR 20, 177-180 Eletsky A., Heinz T., Moreira O., Kienhöfer A., Hilvert D., Pervushin K. (2002) Direct NMR observation and DFT calculations of a hydrogen bond at the active site of a 44 kDa enzyme, J. Biomol. NMR 24, 31-39 Kienhöfer A., Kast P., Hilvert D. (2003) Selective Stabilization of the Chorismate Mutase Transition State by a Positively Charged Hydrogen Bond Donor, J. Am. Chem. Soc. 125, 3206-3207 Wendt S., McCombie G., Daniel J., Kienhöfer A., Hilvert D., Zenobi R. (2003) Quantitative Evaluation of Noncovalent Chorismate Mutase-Inhibitor Binding by ESI- MS, J. Am. Soc. Mass Spec. 14, 1470-1476 Eletsky A., Kienhöfer A., Hilvert d., Pervushin K. (2005) Investigation of Ligand Binding and Protein Dynamics in Bacillus subtilis Chorismate Mutase by Transverse Relaxation Optimized Spectroscopy-Nuclear Magnetic Resonance, Biochemistry 44, 6788-6799

Posters: International Meeting on Proteom Analysis at the TU Munich, September 16.–19., 2001. Electrospray Time-of-Flight Mass Spectrometry: Investigation of Noncovalent Protein-Protein and Protein-Ligand Interactions. Fall Meeting of the Swiss Chemical Society in Zürich, at October 12., 2001. Substitution of Arg-90 in the Active site of Bacillus subtilis Chorismate Mutase with non-proteinogenic Amino Acids. WISOR in Bressanone/Brixen Italy, January 6.–13. 2002. Substitution of Arg-90 in the Active Site of Bacillus subtilis Chorismate Mutase with non-proteinogenic Amino Acids. Joining Forces: New Chemistry-, Informatics- and Engineering-Based Approaches to Study Biological Processes in Zürich at ETH-Hönggerberg, March 21.–22., 2002. Exploring the Catalytic Mechanism of Bacillus subtilis Chorismate Mutase. D-BIOL Symposium ETH Zürich in the Congress-Center in Davos, May 8.– 10., 2002. Exploring the Catalytic Mechanism of Bacillus subtilis Chorismate Mutase. The 7th annual meeting of the working party ‘Biotransformations’ German Society for Applied Microbiology (VAAM) at ETH Zürich October 9.–11., 2002. The Role of Arg- 90 in the Active Site of Bacillus subtilis Chorismate Mutase. ii

Index

Index ii Acknowledgments vi Abstract viii Zusammenfassung xi

1 Introduction 1

1.1 Claisen Rearrangement 1 1.2 Shikimate Pathway 3 1.3 Chorismate Mutase 5 1.4 Active Sites of Chorismate Mutases 8 1.5 The Catalytic Mechanism of Chorismate Mutases 12 1.6 Goals of this Work 17

2 Site-specific Replacement of a Crucial Arginine in BsCM with an Isosteric but Uncharged Citrulline 18

2.1 Introduction 18 2.2 Introduction of Unnatural Amino Acids into Proteins 19 Stop Codon Suppression 19 Peptide Synthesis 21 Peptide Ligation 24 Inteins 27 Strategy for Preparing BsCM Variants with non-standard Amino Acids at Position 90 29 2.3 Results 30 2.4 Discussion 37

3 BsCM Variants Containing homo-Lysine or Difluoro-Arginine at Position 90 in the Active Site 41

3.1 Introduction 41 3.2 Results 43 Arg90homo-Lys BsCM* 43 Arg90F2Arg BsCM* 46 3.3 Discussion 58

iii

4 Investigation of Ligand Binding and Protein Dynamics in BsCM by TROSY-NMR 62

4.1 Introduction 62 4.2 Results 64 Assignment of backbone resonances 64 Ligand binding 68 15 N T1, T2, and HNOE Data 68 Model-Free Analysis of Intramolecular Backbone Mobility 72 4.3 Discussion 74

5 Outlook 81

6 Experimental Section 83

6.1 General 83 Abbreviations 83 Analytical Data 86 Solutions, Buffers, and Media Molecular Biology 96 Ingredients for Minimal Media 99 Buffers for Purification of BsCM 1-87, Intein, CBD Fusion 100 Strains, Plasmids, and Oligonucleotides 101 General Methods: Molecular Biology 102 General Procedures: Molecular Biology 109 Chemicals and General Procedures: Chemistry 111 Computer Programs 113 6.2 Peptide Synthesis 114 Fmoc-BsCM(108-127) 35 114 BsCM(98-127) 36 115 D102E BsCM(88-127) 9 116 Arg90Cit D102E BsCM(88-127) 10 118 Arg90Cit BsCM(88-93) 37 119 BsCM(88-93) 11 120 Fmoc-D102E BsCM(92-127) 38 121 Arg90homo-Lys D102E BsCM(88-127) 12 123 Arg90F2Arg D102E BsCM(88-127) 31 124 6.3 Cloning and Expression 125 BsCM(1-87) encoded by pTYB1-BsCM87 125 BsCM(1-87) encoded by pTXB1-BsCM87 126 D102E BsCM encoded by pAK-D102E 127 R105K BsCM encoded by pAK-R105K 128 D102E R105K BsCM encoded by pAK-DERK 129 BsCM(1-124) encoded by pAK-T125* 130 BsCM(1-122) encoded by pAK-K123* 131 BsCM(1-121) encoded by pAK-T122* 132 BsCM(1-120) encoded by pAK-L121* 133

BsCM(1-119) encoded by pAK-S120* 134 6.4 Native Chemical Ligation 135 D102E BsCM 135 Arg90Cit D102E BsCM 135 Arg90homo-Lys D102E BsCM 136 Arg90F2Arg D102E BsCM 137 6.5 Isotopically Labeled BsCM 137 15N-labeled BsCM 137 D (< 35%), 13C-, and 15N-labeled BsCM 138 D, 13C-, and 15N-labeled BsCM 139 D (50%), 13C-, and 15N-labeled BsCM 140 6.6 Synthesis of Fmoc-4,4-difluoro-L-arginine(Pbf)-OH 141 Synthesis of tert-butyl (4R)-4-(3-ethoxy-2,2-difluoro-1-hydroxy-3-oxopropyl)-2,2- dimethyl-1,3-oxazolane-3-carboxylate 18: 141 Synthesis of tert-butyl (4R)-4-3-ethoxy-2,2-difluoro-1-[(1H-imidazol-1- ylcarbothioyl)oxy]-3-oxopropyl-2,2-dimethyl-1,3-oxazolane-3-carboxylate 19: 143 Synthesis of tert-butyl (4S)-4-(3-ethoxy-2,2-difluoro-3-oxopropyl)-2,2-dimethyl- 1,3-oxazolane-3-carboxylate 20: 145 Synthesis of tert-butyl (4S)-4-(3-amino-2,2-difluoro-3-oxopropyl)-2,2-dimethyl- 1,3-oxazolane-3-carboxylate 21: 147 Synthesis of tert-butyl (4S)-4-(3-amino-2,2-difluoropropyl)-2,2-dimethyl-1,3- oxazolane-3-carboxylate 22: 148 Synthesis of tert-butyl (4S)-4-(3-[(benzyloxy)carbonyl]amino-2,2-difluoropropyl)- 2,2-dimethyl-1,3-oxazolane-3-carboxylate 23: 149 Synthesis of benzyl N-(4S)-4-[(tert-butoxycarbonyl)amino]-2,2-difluoro-5- hydroxypentylcarbamate 24: 151 Synthesis of tert-butyl N-[(1S)-4-amino-3,3-difluoro-1-(hydroxymethyl) butyl]carbamate 25: 152 Synthesis of tert-butyl N-[(Z)-[(tert-butoxycarbonyl)amino]((4S)-4-[(tert- butoxycarbonyl)amino]-2,2-difluoro-5-hydroxypentylamino) methylidene]carbamate 26: 153 Synthesis of (2S)-2-[(tert-butoxycarbonyl)amino]-5-([(tert- butoxycarbonyl)amino][(tert-butoxycarbonyl)imino]methylamino)-4,4- difluoropentanoic acid 27: 155 Synthesis of (2S)-2-amino-5-[amino(imino)methyl]amino-4,4-difluoropentanoic acid 8: 156 Synthesis of (2S)-5-[amino(imino)methyl]amino-2-[(benzyloxy)carbonyl]amino- 4,4-difluoropentanoic acid 28: 157 Synthesis of (2S)-2-[(benzyloxy)carbonyl]amino-4,4-difluoro-5-[(imino[(2,2,4,6,7- pentamethyl-2,3-dihydro-1-benzofuran-5- yl)sulfonyl]aminomethyl)amino]pentanoic acid 29: 158 Synthesis of (2S)-2-amino-4,4-difluoro-5-[(imino[(2,2,4,6,7-pentamethyl-2,3- dihydro-1-benzofuran-5-yl)sulfonyl] aminomethyl)amino]pentanoic acid 30: 160 Synthesis of (2S)-2-[(9H-fluoren-9-ylmethoxy)carbonyl]amino-4,4-difluoro-5- [(imino[(2,2,4,6,7-pentamethyl-2,3-dihydro-1-benzofuran-5- yl)sulfonyl]aminomethyl)amino]pentanoic acid 13: 161

7 Appendix 164

7.1 Programs for the Peptide Synthesizer 164 Modifications to the standard programs 164 7.2 Analytical Ultracentrifugation Data for BsCM* and Arg90Cit BsCM* 174 Data for BsCM* 174 Data for Arg90Cit BsCM* 175

8 Literature 177

Acknowledgments

I would like to thank my “Doktorvater” Prof. Donald Hilvert for giving me the opportunity to work in his laboratory on this exciting and challenging project. His constant support, scientific enthusiasm and the granted freedom were essential for the success of this work. I also want to thank him for thoroughly proof-reading my thesis.

Prof. Dario Neri I want to thank for being my co-examiner.

I would like to thank my labmates from D32 and F328 for the enjoyable working atmosphere. I am in dept to Richard Quaderer and Axel Sewing for introducing me to peptide chemistry and HPLC and to the Praktikum 2000 and Andrea Piatesi for the introduction to molecular biology. It was always fun to solve different computer problems together with Andreas Aemissegger. The many happy hours we spent inside the lab and on various parties will always be lasting memories. Dr. Ajay Mandal was not only always willing to discuss chemistry problems but he also new all the good places for out of lab activities. Andreas Kleeb was always a good partner to discuss biology issues and he organized some unforgettable group events. I would like to thank Dr. Peter Kast for a lot of fruitful discussions about molecular biology problems. Further I want to thank Andreas Kleeb, Andreas Aemissegger, Dr. Ajay Mandal and Joachim Kienhöfer for proofreading parts of this thesis. I also want to thank all present and past members of the Hilvert group for making my time at the LOC a memorable experience. Special thanks go to Yael Gozin, Adrian Hugenmatter, Florian Seebeck, Dr. Alexander Backes, Dr. Sean V. Taylor, Dr. Ken Woycechowsky, Dr. Yoshiya Ikawa, Dr. Sandra Jonsson, Giulio Casi, Jörg Serafimov, Katharina Vamvaca, Dominik Künzler and to all the “early eaters”. The people working in the service departments I thank for their experimental support (Dr. Walter Amrein, Oswald Greter, Rolf Häfliger and Oliver Scheidegger from

vii the MS service, Dr. Bernhard Jaun, Brigitte Brandenberg and Philipp Zumbrunnen from the NMR service, Michael Schneider and Peter Kälin from the microanalytical laboratory and Dr. René Brunisholz from the protein service laboratory).

My special thanks goes to Barbara for her love and support in good and especially in bad times.

viii

Abstract

Chorismate mutase catalyzes the Claisen-rearrangement of (-)-chorismate to prephenate, the first committed step in the biosynthesis of phenylalanine and tyrosine. Bacillus subtilis chorismate mutase (BsCM) was used as a model system to study the catalytic mechanism of the enzyme. Mutagenesis, enzymological, and crystallographic studies of BsCM have shown that arginine at position 90 is very important for the 106- fold rate acceleration provided by the enzyme, probably by stabilizing the developing negative charge at the ether oxygen in the transition state. It was further concluded from crystallographic and Fourier transform infrared studies of the enzyme and its complex with prephenate and a transition state analogue inhibitor that the enzyme shows considerable structural changes upon substrate binding and especially that the C- terminal tail might serve as a lid for the active site upon substrate binding. These two aspects of the enzyme were studied in this work.

In order to determine whether the ability of arginine to form hydrogen bonds or its positive charge is essential for catalysis, we have replaced it with non-proteinogenic analogues by chemical semisynthesis. The analogues chosen were citrulline, homo- lysine, and 4,4-difluoroarginine; the latter was synthesized in 17 linear steps starting from N-Boc-D-serine methylester. For semisynthesis, the enzyme was divided into two parts between Lys87 and Cys88. The C-terminal 40 amino acids that contained the mutated position as third last residue were synthesized by solid phase peptide synthesis (SPPS) and all variants were obtained in sufficient yield after optimizing the synthesis. Aspartimide formation during SPPS of the 40-mer required the introduction of an Asp102Glu mutation in this fragment (variants containing this mutation are designated as BsCM*). The N-terminal 87 residues were biosynthesized in Escherichia coli as an intein fusion. After capturing the intein splicing intermediate with a thiol, the N- terminal fragment was obtained as an activated thioester at its C-terminus. The two fragments were then coupled by native chemical ligation under denaturing conditions. After folding and purification, the three mutant enzymes were obtained and characterized. To validate the method, the enzyme with the wild-type arginine residue at

ix position 90 was prepared as well. It was indistinguishable from the recombinant enzyme with respect to its activity and biophysical properties. The Arg90Cit BsCM* mutant described in Chapter 2, which contains an isosteric but uncharged substitution at position 90, shows a dramatic loss in activity. Its -1 kcat value is only 0.0026 ± 0.0001 s , more than four orders of magnitude lower than that of wild-type BsCM, whereas its Km value is only increased three-times to 270 ± 40

µM. Its Ki value of 6.8 ± 0.2 µM displayed a similar increase as the Km parameter. These observations support the role of electrostatic transition state stabilization in catalysis by the enzyme chorismate mutase.

In Chapter 3 two mutants of BsCM are described in which Arg90 is replaced either by homo-lysine or by 4,4-difluoroarginine, residues with positively charged side chains. The kinetic parameters determined for Arg90homo-Lys BsCM* were kcat = -1 0.013 ± 0.001 s and Km = 510 ± 70 µM. The enzyme was inhibited by the transition state analogue 1 with a Ki value of 7.6 ± 1 µM. Thus, the increase in Km. and Ki are comparable and approximately five-times higher than the corresponding values for the wild-type enzyme. The kcat value of Arg90homo-Lys BsCM* is 3500-fold lower than that for BsCM* and only five-times higher than for Arg90Cit BsCM*. This suggests that either the decreased steric requirement or the decreased number of hydrogen bond donor sites of the ammonium group compared to the guanidinium group of arginine causes the side chain of homo-lysine to be improperly oriented for efficient catalysis.

The kinetic parameters were also determined for the Arg90F2Arg BsCM* -1 mutant. Its kcat value of 47 ± 6 s is indistinguishable from that of BsCM* and in good agreement with the published kinetic parameters for wild-type BsCM. The Km value of 190 ± 40 µM is increased by two-fold compared to that of BsCM*. A similar increase to

3.3 ± 0.5 µM was found for the Ki value. The pH dependence of the reaction catalyzed by the Arg90F2Arg BsCM* variant was also examined. The apparent pKa values obtained for Arg90F2Arg BsCM* for two ionizing groups are 4.21 ± 0.08 and 8.66 ±

0.06. The first pKa value for Arg90F2Arg BsCM* is similar to that of the wild-type enzyme, but the second is shifted by almost 0.6 pKa units from 9.22 to 8.66.

The slight increase in kcat and large increase in Km with increasing pH observed for Arg90F2Arg BsCM* indicates that deprotonation of the enzyme happens before the

substrate binds. The pKa value of 8.66 consequently reflects ionization of the free enzyme and the apparent pKa value increases considerably upon complexation of the active site with the substrate. Assuming the high pKa can be attributed to the deprotonation of Arg90, this is an example of how dramatic pKa values can change inside the active pocket of an enzyme. It also provides further evidence for the need for electrostatic stabilisation of the polarized transition state.

The structural and dynamical consequences of ligand binding to the monofunctional chorismate mutase from Bacillus subtilis have been investigated by solution NMR spectroscopy in the second part of this thesis. TROSY methods were employed to assign 98% of the backbone 1HN, 1Hα, 15N, 13C’ and 13Cα resonances as well as 86% of the side-chain 13C resonances of the 44 kDa trimeric enzyme at 20°C. This information was used to map chemical shift perturbations and changes in intramolecular mobility caused by binding of prephenate or a transition state analogue onto the X-ray structure. Model-free interpretation of backbone dynamics for the free enzyme and its complexes based on 15N relaxation data measured at 600 and 900 MHz showed significant structural consolidation of the protein in the presence of bound ligand. In agreement with earlier structural and biochemical studies, substantial ordering of ten otherwise highly flexible residues at the C-terminus is particularly notable. The observed changes suggest direct contact between this protein segment and bound ligand, providing support for the proposal that the C-terminus can serve as a lid for the active site, limiting diffusion into and out of the pocket and possibly imposing conformational control over substrate once bound. Other regions of the protein that experience substantial ligand-induced changes also border the active site or lie along the subunit interfaces, indicating that the enzyme adapts dynamically to ligands by a sort of induced fit mechanism. It is believed that the mutase-catalyzed chorismate-to- prephenate rearrangement is partially encounter controlled, and backbone motions on the millisecond time scale, as seen here, may contribute to the reaction barrier.

Zusammenfassung

Chorismat-Mutase katalysiert die Claisen-Umlagerung von (-)-Chorismat nach Prephenat, und damit die Schlüsselreaktion beim Übergang vom allgemeinen aromatischen Biosyntheseweg zur spezifischen metabolischen Sequenz in Richtung Phenylalanin und Tyrosin. Die Chorismat-Mutase von Bacillus subtilis (BsCM) wurde als Modellsystem verwendet um den Katalysemechanismus des Enzyms zu untersuchen. Mutagenese und enzymologische und kristallographische Studien konnten zeigen, dass das Arginin an Position 90 sehr wichtig ist für die 106–fache Erhöhung der Reaktionsgeschwindigkeit durch das Enzym, vermutlich durch eine Stabilisierung der sich entwickelnden negativen Partialladung am Ethersauerstoff des Übergangszustandes. Aus den Daten der Kristallstruktur und Messungen mittels Fouriertransformations-Infrarot-Spektroskopie des Enzyms und seines Komplexes mit Prephenat wurde außerdem gefolgert, dass das C-terminale Ende des Enzyms die aktive Tasche wie ein Deckel verschließt, nachdem das Substrat gebunden hat. Diese beiden Aspekte des Enzyms wurden in der vorliegenden Arbeit weitergehend untersucht.

Um herauszufinden, ob die Fähigkeit von Arginin Wasserstoffbindungen auszubilden, oder seine positive Ladung, notwendig ist für die Katalyse, haben wir es durch chemische Semisynthese mit nicht proteinogenen Analoga ersetzt. Die ausgewählten Analoga waren Citrullin, homo-Lysin und 4,4-Difluoroarginin, letzteres wurde in 17 linearen Schritten ausgehend von N-Boc-D-Serinmethylester synthetisiert. Für die Semisynthese wurde das Enzym zwischen Lys87 und Cys88 in zwei Fragmente aufgeteilt. Das C-terminale 40-mer, das die mutierte Position an drittletzter Stelle enthielt, wurde durch Festphasenpeptidsynthese (SPPS) hergestellt und alle Varianten konnten, nach Optimierung der Synthese, in ausreichenden Mengen erhalten werden. Aspartimid-Bildung während der SPPS erforderte die Einführung einer Asp102Glu Mutation in diesem Fragment (Varianten die diese Mutation enthalten werden als BsCM* bezeichnet). Das aus 87 Aminosäuren bestehende N-terminale Fragment wurde durch Biosynthese in Escherichia coli als Fusion mit einem Intein hergestellt. Nachdem das Spleißintermediat mit einem Thiol abgefangen wurde, wurde es mit einem

xii aktivierten Thioester an seinem C-Terminus erhalten. Das synthetische und das biosynthetische Fragment wurden durch native chemische Ligation unter denaturierenden Bedingungen gekuppelt. Nach Faltung und Reinigung, konnten die drei gewünschten Mutanten des Enzyms erhalten und charakterisiert werden. Zur Validierung der Methode wurde auch das Enzym mit dem Wildtyp Arginin an der Position 90 hergestellt. Seine Aktivität und seine biophysikalischen Eigenschaften waren ununterscheidbar von dem rekombinant hergestellten Enzym. Die in Kapitel 2 beschriebene Arg90Cit BsCM* Mutante, die eine isosterische aber ungeladene Substitution an Position 90 enthält, zeigt eine dramatisch verringerte -1 Aktivität. Ihr kcat Wert ist nur 0.0026 ± 0.0001 s und damit mehr als vier

Größenordnungen unter dem des Wildtyps. Sein Km Wert dagegen ist nur dreifach erhöht auf 270 ± 40 µM. Eine ähnliche Erhöhung auf 6.8 ± 0.2 µM konnte für den Ki Wert gemessen werden. Diese Ergebnisse unterstützen die Rolle der elektrostatischen Stabilisierung des Übergangszustandes als wichtigen Faktor der Katalyse durch Chorismat-Mutase.

In Kapitel 3 sind zwei Mutanten von BsCM beschrieben, bei denen Arg90 entweder durch homo-Lysin oder 4,4-Difluoroarginin ersetzt wurde. Beides sind Aminosäuren mit positiv geladenen Seitenketten. Die kinetischen Parameter, die für -1 Arg90homo-Lys BsCM* bestimmt wurden, waren kcat = 0.013 ± 0.001 s und Km =

510 ± 70 µM. Das Enzym wurde durch das Übergangszustandsanalogon 1 mit einem Ki

Wert von 7.6 ± 1 µM inhibiert. Die Zunahmen von Km und Ki sind vergleichbar und die

Werte sind ca. fünfmal höher als die entsprechenden des Wildtyp-Enzyms. Der kcat Wert von Arg90homo-Lys BsCM* ist 3500-mal tiefer als derjenige von BsCM* und nur fünfmal höher als der von Arg90Cit BsCM*. Eine Erklärung dafür könnte sein, dass entweder die geringeren sterischen Ansprüche, oder die geringere Zahl möglicher Wasserstoffbrückendonoren der Ammonium-Gruppe im Vergleich mit der Guanidinium-Gruppe von Arginin, dafür sorgt, dass die Seitenkette von homo-Lysin nicht korrekt orientiert ist, um eine effiziente Katalyse zu gewährleisten.

Die kinetischen Parameter wurden auch für die Arg90F2Arg BsCM* Mutante -1 bestimmt. Ihr kcat Wert von 47 ± 6 s ist ununterscheidbar von dem von BsCM* und in guter Übereinstimmung mit publizierten Werten für das Wildtyp-Enzym. Der Km Wert

xiii nimmt, im Vergleich zu BsCM*, um einen Faktor zwei zu auf 190 ± 40 µM. Eine

ähnliche Zunahme auf 3.3 ± 0.5 µM wurde für den Ki Wert gefunden. Die pH-

Abhängigkeit der durch Arg90F2Arg BsCM* katalysierten Reaktion wurde ebenfalls untersucht. Die für zwei ionisierbare Gruppen von Arg90F2Arg BsCM* erhaltenen pKa

Werte waren 4.21 ± 0.08 und 8.66 ± 0.06. Der erste pKa Wert von Arg90F2Arg BsCM* ist vergleichbar mit dem des Wildtyp-Enzyms, aber der zweite ist um fast 0.6 pKa Einheiten von 9.22 nach 8.66 verschoben.

Die geringe Zunahme von kcat und die große Zunahme von Km mit zunehmendem pH, die für Arg90F2Arg BsCM* beobachtet wurde, weist auf eine

Deprotonierung des Enzyms vor der Substratbindung hin. Der pKa Wert von 8.66 entspricht daher der Ionisierung des freien Enzyms und er steigt nach der

Substratbindung beträchtlich an. Unter der Annahme, dass der höhere pKa Wert der Deprotonierung von Arg90 zugeordnet werden kann, ist dies ein Beispiel dafür wie stark sich pKa Werte in der aktiven Tasche eines Enzyms ändern können und liefert weitere Hinweise für die Notwendigkeit der elektrostatischen Stabilisierung des polarisierten Übergangszustands der Reaktion.

In Kapitel 4 wurden die strukturellen und dynamischen Auswirkungen der Substratbindung auf die monofunktionale Chorismat-Mutase von Bacillus subtilis mit Hilfe von NMR in Lösung untersucht. Es wurden TROSY Methoden verwendet, um 98% der 1HN, 1Hα, 15N, 13C’ und 13Cα Resonanzen des Rückgrats und 86% der 13C Resonanzen der Seitenketten des 44 kDa schweren homotrimeren Enzyms bei 20°C zuzuordnen. Diese Information wurde verwendet, um Veränderungen der chemischen Verschiebung und der intramolekularen Mobilität, die durch Ligandbindung verursacht werden, auf die Kristallstruktur zu projizieren. Modell-freie Interpretation der Dynamik des Rückgrats für das freie Enzym und seine Komplexe mit Prephenat und dem Inhibitor 1, basierend auf 15N Relaxationsdaten, die bei 600 und 900 MHz gemessen wurden, zeigten eine signifikante strukturelle Festigung des Proteins in der Gegenwart von gebundenem Ligand. Eine substantielle Ordnung von zehn ansonsten hoch flexiblen Aminosäuren am C-Terminus ist besonders erwähnenswert. Dies ist in Übereinstimmung mit früheren strukturellen und biochemischen Forschungsergebnissen. Die beobachteten Veränderungen weisen auf einen direkten

xiv

Kontakt dieses Proteinsegments mit dem gebundenen Liganden hin, ein weiteres Indiz für die These, dass der C-Terminus als Deckel für die aktive Tasche fungieren kann und dadurch die Diffusion in und aus der Tasche limitiert und möglicherweise konformationelle Kontrolle über das gebundene Substrat ausübt. Andere Regionen des Proteins, die Ligand induzierte Veränderungen erfahren, sind ebenfalls nahe der aktiven Tasche oder liegen entlang der Interfaceregion der Untereinheiten. Dies ist ein Hinweis darauf, dass sich das Enzym durch eine Art “induced fit” Mechanismus dynamisch an den Liganden anpasst. Man nimmt an, dass die Mutase-katalysierte Umlagerung von Chorismat nach Prephenat teilweise Begegnungs-kontrolliert ist und Rückgrat- Bewegungen auf einer Millisekunden Zeitskala, wie sie hier beobachtet wurden, zur Reaktionsbarriere beitragen.

1 Introduction

1.1 Claisen Rearrangement

The Claisen rearrangement, originally discovered in 1912 (1), has become an important C-C bond forming reaction in organic synthesis (2). In its simplest form, it involves the thermal [3,3]-sigmatropic rearrangement of allyl vinyl ether to 4-pentenal (Scheme 1.1). The vinyl branch can also be part of an aromatic ring, in which case the rearrangement is followed by a tautomerization to the corresponding phenol.

O O

Scheme 1.1. Claisen rearrangement of allyl vinyl ether to 4-pentenal.

Other useful variations of this transformation include the rearrangement of amide acetals to amides (3), orthoesters to esters (4), and silyl enol esters to acids (5). It is also possible to have sulfur (6) or nitrogen (7) in place of the ether oxygen. Like its closely related all-carbon analogue, the Cope rearrangement, the Claisen rearrangement is considered to be a suprafacial concerted process (2). In this case, however, secondary deuterium kinetic isotope effects indicate, that the reaction proceeds asynchronously with C-O bond breaking preceeding C-C bond formation (8- 11). The reaction has been found to proceed via a chair-like aromatic transition state of contiguous p orbitals (6 electrons, Hückel topology) so that chirality is maintained (12, 13). The dipolar nature of the transition state is supported by substituent effects on substituted allyl vinyl ethers. Donor substituents like methoxy at position 4 or 6 accelerate the reaction 100 and 10 times, respectively, by stabilization of the transition 1 INTRODUCTION 2 state, while the same substituent at position 5 is 40 times slower than in the unsubstituted case (14). A trimethylsilyloxy substituent at position 2 is also strongly accelerating (15, 16). Cyano groups at position 2, 4, and 5 accelerate the reaction by factors of 110, 270, and 15.6, respectively, while they slow down the reaction when at position 1 and 6 by factors of 0.9 and 0.11 (17, 18). A relatively strong solvent effect has also been observed for the Claisen rearrangement. For example, when going from an apolar solvent like cyclohexane to a polar hydrogen-bonding solvent like water, an acceleration of more than a hundred-fold has been achieved in some cases (14, 19, 20). Since secondary kinetic isotope effects show that the reaction is not more asynchronous in water then in hydrocarbon solvents, an ionic transition state could be ruled out as explanation for this solvent effect (21). In fact, it is now generally accepted that hydrogen bonding causes the rate acceleration (22). A Monte Carlo simulation confirmed that the transition state for the rearrangement of allyl vinyl ether in water is better hydrated than the ground state by about 16.09 ± 0.67 kJ/mol (3.85 ± 0.16 kcal/mol) (23). This can be explained by the formation of two hydrogen bonds to the ether oxygen of chorismate in the transition state compared to only one in the ground state. The hydration in the ground state and the enhanced hydration in the transition state disrupt the stabilizing n-π* conjugation of the vinyl ether ground state (Figure 1.1) (24). Such disruption is necessary for the reaction to be thermally allowed according to the Woodward-Hoffmann rules (25, 26), because the involvement of two additional electrons from one of the lone pairs of the ether oxygen would make the transition state antiaromatic. The reaction can therefore only occur if those electrons are not involved. Solvent

O O

R R Figure 1.1. n-π* conjugation in non-hydrogen bonding solvents (left) and its disruption in hydrogen bonding solvents (right).

Analogous to the solvent effect, a number of Lewis and Brønsted acids - even silica gel - have been reported to catalyze the Claisen rearrangement by binding to the

1 INTRODUCTION 3 oxygen lone pairs (27). Among the most effective are ammonium chloride (1), boron trifluoride (28), and, most importantly, organoaluminum reagents (29, 30). The only enzyme known to catalyze a Claisen rearrangement is chorismate mutase, which accelerates the conversion of chorismate to prephenate (31). As has been revealed by the crystal structures of the enzyme from Bacillus subtilis (BsCM), Escherichia coli (EcCM), Saccharomyces cerevisiae (ScCM), and Thermus thermophilus (TtCM), biological catalysts also make use of hydrogen bonds to the ether oxygen to accelerate the reaction (32-37).

1.2 Shikimate Pathway

Chorismate mutase is an enzyme in the shikimate biosynthetic pathway (Scheme 1.2), in which D-erythrose-4-phosphate and phosphoenolpyruvate are transformed via shikimate into a variety of aromatic compounds. It catalyzes the Claisen rearrangement of (-)-chorismate 2 to prephenate 3, the first committed step in the biosynthesis of phenylalanine and tyrosine (31).

1 INTRODUCTION 4

Scheme 1.2. Shikimate pathway, the transformation of D-erythrose-4-phosphate and phosphoenolpyruvate via shikimate and chorismate to a variety of aromatic compounds. CHO COO- COO- COO- H OH - COO a b H OH + 2- OPO3 2- 2- - H OH HO OH O3PO OH O3PO O CO2 H OH OH OH D-erythrose- phosphoenol- shikimate shikimate-3-phosphate 5-enolpyruvyl 4-phosphate pyruvate shikimate-3-phosphate

COO- -OOC O COO- - c d O e COO phenylalanine

- O CO2 OH OH (-)-chorismate prephenate phenylpyruvate f g h O COO- COO- COO- NH tyrosine 2 tryptophan folic acid

NH2 OH anthranilate p-aminobenzoate p-hydroxy phenylpyruvate a.) shikimate kinase; b.) 5-enolpyruvyl shikimate-3-phosphate synthase; c.) chorismate synthase; d.) chorismate mutase; e.) prephenate dehydratase; f.) prephenate dehydrogenase; g.) anthranilate synthase; h.) p-aminobenzoate synthase

The shikimate pathway occurs in higher plants, bacteria and fungi, but not in animals and humans. This makes it an attractive target for the development of antibiotics or herbicides. Indeed, inhibitors of enzymes in the shikimate pathway have already found commercial application. For example, N-phosphonomethylglycine (Roundup), a potent inhibitor of 5-enolpyruvyl shikimate-3-phosphate synthase, is a broad-spectrum herbicide which achieves more than $1 billion in sales per year. The best inhibitors of chorismate mutase known so far were designed to mimic the putative transition state 4. However, their moderate potency and complicated synthesis limits their commercial value. Compound 1 and its nitro analogue 5 (Figure 1.3), for instance, exhibit Kd values in the 0.1 to 10 µM range with a wide range of mutants (38, 39).

1 INTRODUCTION 5

O O - O - O O - O - O O N + O O HO HO 1 5 Figure 1.2. Structure of the best inhibitors of chorismate mutase.

1.3 Chorismate Mutase

A large number of chorismate mutases (EC 5.4.99.5) have been identified to date, either by isolation directly from higher plants, bacteria and fungi, or in sequencing projects through comparison with known chorismate mutase genes. Both monofunctional (e.g. BsCM, TtCM, ScCM, and MjCM) and bifunctional (e.g. EcCMp and EcCMt) chorismate mutases are known. The bifunctional enzymes can either be fused to prephenate dehydratase (e.g. EcCMp) or prephenate dehydrogenase (e.g. EcCMt). Many of the bifunctional enzymes can be inhibited by phenylalanine or tyrosine, or allosterically down-regulated by tyrosine (40-42), while some of the monofunctional enzymes are inhibited by tyrosine (37). Most of the plant and fungal enzymes are additionally activated allosterically by tryptophan (40-42). In general, the rate acceleration achieved by the enzyme is about 106-fold compared to the uncatalyzed reaction in water (43).

1 INTRODUCTION 6

ScCM HpCM AnCM AtCM1 AtCM3 AroQ AtCM2 EcCMp MjCM EcCMt TtCM AroH BsCM Figure 1.3. Phylogenetic tree of some representative chorismate mutases, including the chorismate mutases of Saccharomyces cerevisiae (ScCM), Hansenula polymorpha (HpCM), Aspergillus nidulans (AnCM), Arabidopsis thaliana (AtCM1, AtCM2, and AtCM3), Escherichia coli (EcCMp and EcCMt), Methanococcus jannaschii (MjCM), Thermus thermophilus (TtCM), and Bacillus subtilis (BsCM) (37).

The structures of known chorismate mutases divides into two classes, AroQ and AroH (Figure 1.4). From the AroH class, only two enzymes, one from Bacillus subtilis (32, 44) and the other from Thermus thermophilus (36, 37), have been characterized so far. They are unregulated homotrimeric enzymes that form a pseudo-α,β-barrel, with three identical active sites situated in the interface region between two subunits. Most of the enzymes that have been investigated belong to the AroQ class, including the enzymes form Escherichia coli (33, 45), Saccharomyces cerevisiae (34, 35), Hansenula polymorpha (41), Aspergillus nidulans (42), Methanococcus jannaschii (46), Mycobacterium tuberculosis (47), and several others (40). They can be allosterically regulated (ScCM, HpCM, and AnCM), or not, are usually dimers and have an entirely α-helical structure.

1 INTRODUCTION 7

BsCM EcCM

ScCM dimer one ScCM subunit Figure 1.4. Structure of the AroH class BsCM (upper left), the AroQ class EcCM (upper right), the dimeric, allosterically regulated AroQ class ScCM (lower left) and one ScCM subunit with bound inhibitor 1 and tryptophan (lower right) (34, 35).

The chorismate mutase of yeast is a special case, because it is allosterically regulated, i.e. in addition to being inhibited by the binding of tyrosine at the interface region of the dimeric enzyme it is also activated by the binding of tryptophan at the same site (34, 48, 49). If the structures of ScCM and EcCM are compared, there is a strong indication that ScCM has evolved by gene duplication of the chorismate mutase domain of the EcCM P-protein, when the need for a regulatory mechanism emerged during evolution (35). Three helix pairs in ScCM namely helix 2 and 8, helix 4 and 11, and helix 7 and 12, bear high structural similarity to each other and to the corresponding helices 1, 2, and 3 in EcCM (Figure 1.4). For both AroH and AroQ classes of enzyme, the structure and the active site residues are highly conserved within the respective class, despite sometimes low sequence similarity (Figure 1.3) (37). However, the two classes are quite different, if

1 INTRODUCTION 8 compared with each other, although they possess similar residues at their respective active sites.

1.4 Active Sites of Chorismate Mutases

The active site of an AroH chorismate mutase from the gram-positive bacteria Bacillus subtilis (Figure 1.5) is formed by Arg7, Glu78, Arg90, Tyr108, and Leu115 from one subunit and Phe57, Lys60, Val73, Thr74 and Cys75 from the other subunit (32, 44, 50). Additionally, the X-ray structure of the enzyme complexed with either prephenate or TSA 1 shows that Arg63 also makes contact with the ring carboxylate of the complexed molecule in some of the 12 active sites present in one asymmetric unit. The importance of Arg63 is supported by a more recent crystal structure (51), computational studies (52-54), and the strict conservation of this residue in other known or putative AroH chorismate mutases in a BLAST search (37). Mutagenesis studies of this residue that could further support its significance are in progress in our group.

Arg7 Tyr108

HO H2NNH+ 2 + O H2N NH2 Arg90 - O NH HN

O Arg116 SH + NH H 2 O - Cys88 HN HN H + N H - O HN Arg63 Glu78 O HO H

O SH Cys75 Phe57

Figure 1.5. Schematic view of the active site of BsCM with bound transition state analogue 1 in bold.

The structure suggests that residue Arg90 plays a central role in binding and catalysis. It is part of a clamp-like structure that orients the side chain of chorismate with respect to its cyclohexadiene ring. The hydroxyl group of chorismate forms a

1 INTRODUCTION 9 hydrogen bond to the sulfhydryl group of Cys75 and to the carboxylate of Glu78, which in turn forms a salt bridge with the guanidinium group of Arg90. The latter also makes hydrogen bonds to the ether oxygen and the carboxylate of the enol pyruvate side chain of the substrate. The carboxylate is further fixed in place by the side chain of Arg7 (50, 55-57). Hydrophobic interactions with the cyclohexadienyl ring of the substrate are made by Phe57. The tertiary carboxylate groups of the bound inhibitor and bound prephenate are relatively solvent exposed at the entrance of the pocket, but, as mentioned, potentially stabilizing contacts with the side chains of Arg63 and Arg116 are seen in some subunits (32). The recruitment of these residues to the active site is one indication of dynamic changes of the enzyme upon substrate binding and catalysis. The active site thus exhibits excellent charge complementary to the substrate and provides numerous hydrophobic, hydrogen bonding and ionic groups for molecular recognition. Mutagenesis studies confirm the involvement of many of the active site residues in catalysis. Particularly critical are Glu78 and Arg90. As already noted, they form a salt bridge with each other and provide hydrogen bonds to the ether oxygen and the hydroxyl group of chorismate (55, 58). In addition, there is a less pronounced effect when mutating Cys75, the other residue that forms hydrogen bonds with the hydroxyl group, and a Cys75Asp / Glu78Ser double mutant does not restore much activity (55).

Mutations of Cys75 in BsCM surprisingly cause increases in both Km and kcat (50).

H N Arg11 2 Arg51 Arg16 H N 2 + H2N HN H N NH 2 NH + H O O + NH2 NH2 O H Arg28 Arg157 O O + H + NH O O 3 - HN NH3 - O NH O H2N Lys39 + Lys168 + NH H N H N O H O 2 NH O H2N O NH O HO O Asp48 HO HO H Asp194 H Ser84 Thr242

O O - - O O Gln88 O O Glu246

Glu52 Glu198 Figure 1.6. Schematic view of the active site of EcCM (left) and ScCM (right) with bound transition state analogue 1 in bold.

The active site of EcCM is formed by Arg28, Ser84, Gln88, Glu52, Asp48, Lys39, and Arg51 from one subunit and Arg11 from the other subunit (Figure 1.6). These residues form several salt bridges and hydrogen bonds with the substrate (34, 35).

1 INTRODUCTION 10

Hydrogen bonds to the ether oxygen of chorismate are formed by Gln88 and the positively charged Lys39, which also interacts with the enolpyruvyl carboxylate. This carboxylate is also held in place via a salt bridge with Arg11 and a hydrogen bond to a water molecule that is held in place by Arg51 and the other substrate carboxylate. This second carboxylate of chorismate also forms a salt bridge with Arg28 and is hydrogen bonded to the hydroxyl group of Ser84. Additionally, the residues Glu52 and the backbone amide of Asp48 bind to the hydroxyl group of the substrate. The importance of most of these residues has been confirmed by mutagenesis studies (58-60), but unfortunately there are no mutagenesis studies of EcCM Arg51 yet that would support its significance. In this context, it is worth noting that the corresponding residue Arg72 of the CM of Mycobacterium tuberculosis has been subjected to random mutagenesis, and all clones that were obtained in an in vivo selection experiment contained the wild- type residue at this position (61). The need for a second hydrogen bond to the ether oxygen is confirmed by the Gln88Glu mutant, which has almost no activity at neutral pH, but shows very high activity that is comparable to wild-type at pH 4.9, when the side chain of glutamate is protonated (58, 62). The active site of EcCM is, in summary, similar to that of BsCM, despite the fundamentally different architecture of the two proteins; the major difference appears to be that the EcCM active site is much more buried than that of BsCM. The active site of the dimeric, allosterically regulated ScCM (Figure 1.6) is very similar to that of EcCM despite low sequence and only limited structural identity (34, 35). Most of the contacts to the substrate are made by the same amino acids, although they have different locations in ScCM. There are, however, two main differences in addition to the substitution of Ser84 in EcCM with a threonine in ScCM (Thr242). The first is the replacement of Gln88 in EcCM with Glu246 in ScCM. To make the necessary hydrogen bond to the ether oxygen of chorismate in the transition state, this glutamate needs to be protonated. At the same time, the two carboxylates of chorismate and Glu198, which binds to the hydroxy group of chorismate, need to be unprotonated. This requires a rather delicate pH control for the enzyme to be active and places its maximum activity at a fairly low pH of 5.5, which is slightly raised to 6.5 when the enzyme is activated with tryptophan (63, 64). It might be an adaptation to the more acidic habitat of yeast (49), although molecular modelling studies imply that Glu246 might be protonated well above neutral pH with a predicted effective pKa of 8.1 (65). A

1 INTRODUCTION 11

Glu256Gln mutant exhibits a broader pH optimum between pH 5 and 10, but its specific activity is around 6 times lower than that of the wild-type (64). The other difference is the missing bridging water molecule between the two carboxylates of chorismate. The residue corresponding to Arg51 of EcCM, which holds this water in place, is a valine in the ScCM sequence (35). ScCM seems to have found a way to circumvent the need for this bridging water molecule, maybe by hydrophobic interactions exerted by the three residues Leu19, Leu12, and Val164 with the vinylic double bond of the substrate.

NH O H Asp-H97 O Asn-H50 H O O NH - O H2N HN NH2 O Tyr-L94 O - Arg-H95 O HO HO

H2N O

Asn-H33 Figure 1.7. Schematic view of the active site of 1F7 with bound transition state analogue 1 in bold.

Catalytic antibodies with chorismate mutase activity represent another, engineered class of biological catalysts for the chorismate to prephenate reaction. In the catalytic antibody 1F7, which was raised against the transition state analogue 1, fewer interactions with the substrate are found than in the enzymes (66). They consist of the hydrogen bonds between the ring carboxylate of the transition state analogue 1 and the residues Asn-H50 and Tyr-L84, an interaction of the backbone amide of Asp-H97 with the enolpyruvyl carboxylate of the ligand, and a hydrogen bond to its hydroxy group by the side chain amide of Asn-H33 (Figure 1.7). These interactions are not charge complementary to the negatively charged carboxylates of compound 1. Additionally, there is no interaction of the protein with the ether oxygen of chorismate, which probably explains its rather low activity, despite a Km value that is comparable to natural chorismate mutases (67, 68). All described chorismate mutases have an active pocket in which similar, but not necessarily the same, residues provide hydrophobic, hydrogen bonding and ionic interactions that are complementary to the substrate. The active site of the AroH class

1 INTRODUCTION 12 chorismate mutase from B. subtilis seems to be more solvent exposed than the highly shielded active pocket of the AroQ class enzymes, but the highly flexible C-terminal tail of BsCM might also be part of the active site which closes the opening after the substrate has bound.

1.5 The Catalytic Mechanism of Chorismate Mutases

The Claisen rearrangement of chorismate to prephenate is a pericyclic but asynchronous process with a polarized transition state (69-71). At room temperature about 10 to 20% of chorismate adopts the reactive pseudodiaxial conformation in aqueous solution, which is in dynamic equilibrium with the more stable pseudodiequatorial form (72). The enzymatic reaction, like its uncatalyzed counterpart, proceeds via a pseudodiaxial chair-like (72-74) pericyclic transition state 4 (Scheme 1.3) in which, as determined by kinetic isotope effects, C-O bond cleavage precedes C- C bond formation (69-71). Mechanistic speculations that suggest a dissociative mechanism (43, 75), with or without nucleophilic attack of the enzyme at the cyclohexadienyl ring carbon next to the ether oxygen, have not found experimental support (20, 22, 45, 70, 76). From the analysis of the active sites of BsCM, EcCM, and ScCM it can be concluded that the enzyme was shaped by evolutionary forces to accommodate the substrate in its reactive chair-like pseudodiaxial conformation by providing appropriate charge complementary and hydrogen bonding interactions and to stabilize the polarized transition state of the Claisen rearrangement.

- - - CO2 O2C ‡ - CO2 O2C O - CO2 O

- OCO2 OH OH OH 2 4 3 Scheme 1.3. Claisen rearrangement of (-)-chorismate 2 to prephenate 3 via a chair-like pericyclic transition state 4.

1 INTRODUCTION 13

One of the oldest suggestions about how chorismate mutases work is that the enzyme is an entropy trap (74, 77). Comparison of the activation parameters of E. coli CM, K. pneumoniae CM, M. jannaschii CM, and S. aureofaciens CM with the uncatalyzed reaction indicates that lowering the activation entropy to almost zero by reducing the degrees of freedom of the highly flexible chorismate and locking it in the pseudodiaxial chair-like conformation when bound to the active pocket is an important factor in catalysis by chorismate mutases (Table 1.1).

Table 1.1. Activation parameters for the rearrangement of chorismate to prephenate. ‡ -1 ‡ -1 -1 Catalyst ∆H 298 (kJ mol ) ∆S 298 (J mol K ) B. subtilis CMa 53.1 -38.1 K. pneumoniae CMa,b 66.5 -4.6 E. coli CMc 68.2 6.3 M. jannaschii CMd 67.8 -7.1 S. aureofaciens 60.7 -6.7 CMa,b 1F7e 62.8 -92.0 Uncatalyzedf 86.6 -54.0 Data from a(78), b(74), c(62) and Corrigenda (79), d(46), e(68), f(43).

‡ The large negative value for ∆S 298 for the catalytic antibody is in agreement with its less optimal active site that probably is not able to restrict the substrate to the pseudodiaxial chair-like conformation as well as the natural enzymes. However, the activation entropy of BsCM, which is similar to that of the uncatalyzed reaction, does not fit into this so-called entropy trap theory (74, 77, 78). This can be an indication that BsCM exploits a different strategy to achieve a similar catalytic effect (78), or that the additional ordering required for the BsCM·chorismate complex when going from the ground to the transition state might be attributable to its highly flexible C-terminal tail becoming significantly more ordered upon ligand binding. This segment might serve as a lid for the active site during catalysis as implied by the crystal structure (32, 44) and by Fourier transform infrared (FTIR) spectroscopy (80).

As mentioned in the previous chapter, Arg90 in BsCM forms a hydrogen bond to the ether oxygen of the substrate during the reaction. This type of interaction could be essential for catalysis. As has been described in Chapter 1.1, stabilization of the

1 INTRODUCTION 14 transition state by formation of two hydrogen bonds to the ether oxygen is believed to help accelerate the Claisen rearrangement in solution. Analogous interactions need to be provided for by the enzyme to make sure that the reaction is not worse in the active site than in water, and they probably ensure an even faster reaction, because the residues that function as H-bond donors are at fixed positions in the active site. Since the guanidinium group of Arg90 is positively charged, it might additionally stabilize a partial negative charge at the ether oxygen of the polarized transition state (69-71). In EcCM the two hydrogen bonds provided by Gln88 and the also positively charged Lys39 could fulfil a similar role. In ScCM residues Lys168 and Glu246 are analogous, whereas in the catalytic antibody 1F7 related interactions are absent. It is worth noting that in all characterized chorismate mutases at least one of the hydrogen bonds to the ether oxygen is formed by a positively charged functional group, which is a strong indication that a positive charge and not just hydrogen bonding capability is important for the function of the enzymes. Because of extensive hydrogen bonding due to its involvement in “the clamp” it seems unlikely, although it has been suggested (45), that Arg90 of BsCM can form both of the hydrogen bonds, although the hydrogen bonding network precisely positions it for one of them. Potential candidates for the other hydrogen bond include a water molecule from the solvent, which has been predicted computationally to be essential for efficient catalysis (53, 54), or Arg116, which is part of the highly flexible C-terminal tail. The latter might serve as a lid for the otherwise rather open active site after binding of the substrate (Chapter 4, and (32, 80, 81)). The importance of Arg116 is confirmed by mutagenesis studies (50) and the strict conservation of this residue in other known or putative AroH chorismate mutases (37). The importance of Arg90 itself has been also confirmed by mutagenesis studies (50, 82). Even its conservative substitution with a lysine gives an almost inactive enzyme. Interestingly, the double mutant Cys88Lys Arg90Ser can restore substantial the activity compared to the Arg90Lys mutant (83). This can be explained by the fact that while a lysine at position 90 might be able to form an essential hydrogen bond with the ether oxygen, it, the substrate, and some other residues are most likely not correctly positioned because it cannot serve as a clamp, because its ammonium group has less hydrogen bond donor and salt bridge forming capabilities than the guanidinium group of an arginine. The double mutant on the other hand can restore “the clamp” by partially

1 INTRODUCTION 15 forming a network that includes the enol pyruvate side-chain carboxylate, the ether oxygen, Ser90, Lys88, Glu78 and the ring hydroxyl group (56, 83).

Modelling studies of chorismate mutase with combined quantum mechanics / molecular mechanics (QM/MM) methods show significant transition state stabilization (54, 84-88). In addition, several studies suggest that preferential binding of the reactive conformer of chorismate and the distortion or compression of the bound substrate contribute to catalysis (84, 87, 89-91). The most extreme claim in this regard is provided by theoretical studies of Hur and Bruice (92-96). They conclude that transition state stabilization is unimportant in the case of chorismate mutase. Instead, they attribute its catalytic effect mainly to the generation of a near attack conformer (NAC) of the substrate in the active site of the enzyme (92-96). This NAC is defined by a short van der Waals contact distance of 3.7 Å between the reacting centers in a conformation that is almost never adopted in water (92). While this explanation is related to the entropy trap theory mentioned above, it has not found widespread support. In a publication by Warshel et al. the meaning of this apparent NAC effect is discussed (84). They try to exactly define the origin of the reduction of the activation barriers when going from the uncatalyzed reaction in water to the catalyzed reaction in the enzyme.

Those different activation barriers are associated with kcat and kuncat (84, 97). They state that the NAC is just an arbitrary position on the reaction coordinate when going from the ground to the transition state. Their main claim is that the existence of a NAC, which could also be defined as reactant state destabilization, does not change any of the relevant activation barriers. They conclude that the NAC is not “a genuine reason for catalysis, but merely reflects the result of electrostatic transition state stabilization”.

The question remains whether the enzyme achieves its 106-fold rate acceleration by electrostatic stabilization of the transition state, as has been suggested by crystallographic data and mutagenesis studies of the structurally diverse BsCM (32, 44, 50, 82, 83), EcCM (33, 58, 60) and ScCM (65) mutases and is supported by QM/MM calculations (53, 54, 84-87), or merely by restriction of the substrate to a so called near attack conformer (NAC) (92-96, 98), or some combination of both. It is also unclear whether hydrogen bonding to the ether oxygen of the substrate benefits from the

1 INTRODUCTION 16 involvement of a positively charged hydrogen bond donor or whether neutral donors are equally effective. Conceivably, the presence of a positively charged donor in all characterized chorismate mutases could just be a coincidence of evolution which reflects the limited set of functional groups of the 20 available proteinogenic amino acids. Some of these open questions which cannot be resolved by mutagenesis to one of the 20 natural amino acids might be resolved by introduction of sterically or electronically tuned non-proteinogenic amino acids into the enzyme (50, 58). This is possible by using either the stop codon suppression technique or by solid phase peptide synthesis.

1 INTRODUCTION 17

1.6 Goals of this Work

In this work, two aspects of the well-studied enzyme Bacillus subtilis chorismate mutase were investigated to clarify how the catalysis of the Claisen rearrangement from chorismate to prephenate is accomplished by the various chorismate mutase enzymes.

The first part focuses on the role of the arginine at position 90, which may stabilize a developing negative charge at the ether oxygen of the transition state in the Claisen rearrangement from chorismate to prephenate. It was studied by mutating it to citrulline, homo-lysine, and 4,4-difluoro arginine to address the question whether the positive charge of arginine or its ability to make a hydrogen bond to the ether oxygen is more important. HO HO HO O O O

H2N H2N H2N

F F

HN HN

O NH

H2N H2N H2N 6 7 8 Figure 1.7. Citrulline, homo-lysine, and 4,4-difluoro arginine.

The second part of this thesis examines the dynamic properties of BsCM, especially the role of the N-terminal tail, which has been postulated to influence ligand binding. The preparation of truncated mutants together with NMR spectroscopy of isotopically labeled BsCM shed light on these aspects of BsCM catalysis.

2 Site-specific Replacement of a Crucial Arginine in BsCM with an Isosteric but Uncharged Citrulline

2.1 Introduction

As has been outlined in Chapter 1.5, the mechanism of action of chorismate mutase remains controversial (57), despite extensive experimental and theoretical work. The enzymatic reaction, like its uncatalyzed counterpart, proceeds via a chairlike (72- 74) pericyclic transition state 4 in which C-O cleavage precedes C-C bond formation (69-71) (Scheme 1.3). Electrostatic stabilization of this highly polarized species by a positively charged residue, either an arginine or a lysine, positioned next to the ether oxygen of the breaking C-O bond has been suggested by crystallographic and mutagenesis studies of the structurally diverse Bacillus subtilis (BsCM) (32, 44, 50, 82, 83), Escherichia coli (EcCM) (33, 58, 60) and yeast (ScCM) (65) mutases. This hypothesis has found further support in QM/MM calculations (53, 54, 84-87). However, some computational studies indicate that the cationic group stabilizes chorismate in the ground and transition state to similar extents (99). It has even been argued that preferential transition state binding is unimportant for chorismate mutase activity (92, 94). Instead, the efficiency of the enzyme has been attributed mainly to conformational restriction of the substrate in such a way as to confine the reactive centers to contact distances (92-96, 98). The critical cationic residue in BsCM is Arg90 (Figure 2.1). Even conservative substitution of this amino acid with a positively charged lysine leads to substantial reductions in catalytic efficacy (50, 82, 83). However, the mutations that have been investigated to date have not been isosteric with arginine and are typically accompanied by large increases in the Km value for chorismate, making it difficult to distinguish 2 UNCHARGED CITRULLINE MUTANT 19 unambiguously between ground state and transition state effects. To overcome this problem, we decided to prepare a BsCM variant containing citrulline, an isosteric but neutral arginine analogue, at position 90. This minimal substitution of a charged with a neutral hydrogen bond donor allows the importance of the positive charge in the differential stabilization of the ground and transition states to be assessed directly. The incorporation of the noncoded amino acid citrulline into BsCM is potentially possible by using either the stop codon suppression technique or by solid phase peptide synthesis. These options are discussed in the following section.

Arg7 Tyr108

HO H NNH + 2 + 2 O Xaa90 H2N NH2 - O NH HN SH NH O 2 O- Cys88 X Arg116

- O Glu78 O HO

O SH Cys75 Phe57 Figure 2.1. Schematic view of the BsCM active site with bound transition state analogue (compound 1, red) (32, 44). In the wild-type enzyme the blue residue at + position 90 is arginine (X= NH2 ) and in the Arg90Cit variant it is citrulline (X = O).

2.2 Introduction of Unnatural Amino Acids into Proteins

Stop Codon Suppression

Stop codon suppression is based on the use of one of the three stop codons (UAA, UAG, UGA) to insert an unnatural amino acid via a chemically aminoacylated tRNA specific for this stop codon. So far only the UAG codon has been used (100). This codon has recently been found to be used by Archaea to encode for pyrrolysine (101-103).

2 UNCHARGED CITRULLINE MUTANT 20

To insert an amino acid into a protein, the so-called suppressor tRNA first has to be chemically acylated with the desired amino acid which has been activated as a cyanomethylester (104). This is usually done by acylating the dinucleotide dCA, which is then ligated by T4-RNA-ligase to the truncated suppressor tRNA. The latter is obtained by run-off transcription of a tRNA that has been mutated to be complementary to UAG (105, 106). This acylated tRNA and the plasmid that contains the mutation to TAG at the position where the unnatural amino acid is to be inserted is then added to an in vitro translation system derived from E. coli (S30), rabbit reticulocytes, or wheat germ to obtain the desired protein (100, 107, 108). The protein biosynthesis has to be done in vitro, because there is no general method to introduce large amounts of aminoacylated tRNA in living cells (109). The efficiency of the suppression varies between 10 and 100% (100). It can be improved by using a protein-synthesizing system that has been reconstituted from its purified components and where the release factor 1 which is responsible for the termination of translation at a UAG codon has been omitted (110). While the method has found application for studying protein splicing (111, 112), protein stability (113, 114), and enzyme mechanism (115), it does not allow for the incorporation of D-, β, or γ-amino acids, probably because the translation machinery does not accept it (100, 116) and it only allows for one noncoded amino acid to be incorporated per protein although it can be incorporated at several positions (100). The main disadvantage of this system is still the low yield which allows only for the production of a few hundred micrograms of mutated protein (100, 110). This disadvantage has been partly overcome by the generation of an in vivo methodology for the incorporation of non-proteinogenic amino acids (117, 118). This was achieved by adding an orthogonal tRNA codon pair (119), an orthogonal aminoacyl-tRNA synthetase (117, 120, 121), and the desired unnatural amino acid to the biosynthetic machinery of E. coli. However, in this system, the tRNA synthase that acylates the suppressor tRNA with the unnatural amino acid needs to be evolved for each new amino acid by several rounds of mutagenesis and selection, and this it not always possible. Once a working system has been found, the yields are comparable to proteins that contain only natural amino acids (122). A functioning system has only been generated for some aromatic amino acids so far. It has, however, been used successfully to insert several artificial aromatic amino acids into proteins (118, 120,

2 UNCHARGED CITRULLINE MUTANT 21

121). For the preparation of BsCM derivatives containing several unnatural amino acids, this method seemed too cumbersome.

Peptide Synthesis

An attractive alternative to recombinant methods is peptide synthesis by chemical methods, which poses no principal restrictions to the nature of artificial amino acids or other modifications that can be incorporated in the peptide. The traditional method of solution phase peptide synthesis (SPPS) with its many shortcomings was revolutionized by Merrifield’s idea of synthesizing peptides on a solid support. This advance was first described 1963 in an article that has been cited more than 5000 times (123). The solid support allows purification after each coupling and deprotection step by simple filtration; it helps to counteract solubility problems of the growing peptide chain; and it provides the possibility for automation. I will give just a short introduction to SPPS here; more detailed coverage of the subject can be found in several review articles and books on the topic (124-131).

2 UNCHARGED CITRULLINE MUTANT 22

O N

H N O N OH N HBTU O R N

HBTU N

HOBt N DIEA, NMP N

HOBt OH

N O N

H O N N O

O R' R

H * N H

O n O R

H O N N H

O R' O n

Piperidine

H N ++O N H N O n+1 O Scheme 2.1. Principle of Fmoc-SPPS. Activation of the amino acid with HBTU and HOBt (see insert for structures), coupling, Fmoc deprotection with piperidine, and the online detectable adducts formed during the deprotection.

The principle of SPPS is to covalently attach the first α-amino and side-chain protected amino acid to a linker on a solid support. In most cases, this solid support is a copolymer of styrene, 1% m-divinylbenzene, and a functionalized monomer such as p- hydroxymethyl styrene. The linker and sometimes a polyethylene glycol spacer (132, 133) are attached to this functionalized monomer after polymerization. Usually the resin is equipped with a 4-hydroxybenzyl alcohol, the so-called Wang linker (134). The α- amino protecting group is removed and the next amino acid, containing protected α-

2 UNCHARGED CITRULLINE MUTANT 23 amino and side chain functionalities, is activated at its carboxylate group and coupled. The deprotection and coupling steps are repeated until the desired peptide is synthesized. The final product can then be cleaved from the solid support and the side chain protection groups are normally removed simultaneously. There are two major methodologies for SPPS: Boc- and Fmoc-SPPS (126, 127, 130, 131). They are named after the α-amino protecting groups, N-tert-butoxycarbonyl (Boc) or 9-fluorenylmethoxycarbonyl (Fmoc), which are used. The older Boc chemistry (135, 136) relies on the increasing acid lability of the side chain and the α-amino protecting groups. The Boc protecting group can be removed with trifluoroacetic acid followed by neutralization, whereas deprotection of the side chains and cleavage from the resin requires the use of toxic and highly aggressive anhydrous hydrogen fluoride. The requirement for HF is a major drawback in Boc chemistry because it requires the use of specially resistant equipment and special precautions. In an optimized protocol, the separate step in which the ammonium trifluoroacetate salt obtained after Boc deprotection is neutralized, is no longer necessary (137). This optimized protocol has been used for the linear synthesis of the 99 residue protein HIV-1 protease (138). The newer, and nowadays more frequently used Fmoc method was introduced about 20 years after its competitor (139, 140) but has found widespread application (127, 128). It uses the secondary amine piperidine or sometimes the stronger 1,8- diazabicyclo[5,4,0]undecene to remove the Nα-protecting group via β-elimination. After deprotonation of the proton at the only saturated carbon of the fluorenyl ring, which is rather acidic due to the resonance stabilization of the generated aromatic anion, the carbamate decomposes to the free amine, dibenzofulvene, and carbon dioxide. The two fragments from the Fmoc group react further with piperidine to form adducts that are highly absorbent or anionic species (Scheme 2.1) and can easily be detected by UV- spectroscopy (300-320 nm) or conductivity measurement, allowing for online monitoring of the deprotection process. This is an advantage that facilitates automation compared to Boc chemistry where a sample has to be removed and investigated by the quantitative ninhydrin reaction (141). Activation of the incoming amino acid is done immediately before the coupling by adding DIEA and a 1:1 mixture of HOBt and HBTU in 1-methyl-2-pyrrolidinone (NMP) to the amino acid. This converts the amino acid to the corresponding OBt ester.

2 UNCHARGED CITRULLINE MUTANT 24

Side chain protecting groups are removed by treatment with a mixture of TFA and scavengers like phenol, water, triisopropyl silane, and thiols. The addition of scavengers to the cleavage cocktail is necessary to trap reactive intermediates such as the trityl cation that are generated upon deprotection of the side chains. When using an acid sensitive linker, for example the above mentioned Wang linker, the peptide is simultaneously cleaved from the resin as a C-terminal carboxylic acid. Fmoc-SPPS has been successfully exploited to synthesize peptides as large as the 166 residue human erythropoietin in a linear fashion using optimized activation reagents, double coupling and a special Nα-protecting group for the last amino acid to facilitate purification (142). Despite the successful synthesis of remarkably large peptides, solid phase peptide synthesis has its limitations. Side reactions, and incomplete couplings and deprotections during the synthesis eventually accumulate a lot of side products on the resin, which makes the purification of the final product laborious. The synthesis of small peptides (<20 amino acids) can be considered routine, mid-range peptides (20–50 amino acids) can usually be obtained after optimizing the synthesis, whereas longer peptides can only be prepared in favorable and exceptional cases.

Peptide Ligation

As we have seen in the description of SPPS, the synthesis of the 127-mer BsCM in a linear fashion would be a heroic undertaking, with little hope for success. Alternatively a convergent strategy of ligating several large peptide fragments, similar to methods that have been successfully employed in chemical total synthesis (143), might be an option. Initial attempts to ligate two peptides were based on the coupling of two fully protected fragments, one with an activated C-terminal carboxylate, the other with a free N-terminal amine (144). However, many problems where encountered with this approach (126). They include racemization of the activated C-terminal amino acid (145), poor solubility of the protected peptide segments (146), and the inability to purify

2 UNCHARGED CITRULLINE MUTANT 25 and characterize the individual peptide fragments (147). Another means to couple unprotected peptide fragments is the use of a modified protease which allows the coupling of a peptide (glycolate-phenylalanyl amide) ester or a peptide benzylthiol ester with another peptide (148-153). The methods used to couple unprotected peptide fragments that do not result in a peptide bond include the reaction of an N-terminal aldehyde, generated by selective periodate oxidation of a serine (154), with a C-terminal hydrazide or hydroxylamine (155-157), and the reaction of a thioacid with a bromoacetamide (158, 159). Another method that affords native amide bonds is the Staudinger ligation of peptides with a C-terminal phosphinothioester and N-terminal azide, an emerging method in protein chemistry (160).

O O O NH2 R' O O N H N 2 R' O IO - O R R 4 O a.) O R O NH2 R' N HO H N R' N R H

b.) O O O + R' S Br R R' SH R O

c.) O O O H R' N N3 + R R R' S PPh2 O R'' R'' d.) O O O H R' N H2N R'' + R R R' S O HS HS e.) Scheme 2.2. Methods for the chemoselective ligation of unprotected peptide segments; a.) Oxime forming ligation; b.) Hydrazon forming ligation; c.) Thioester forming ligation; d.) Staudinger ligation; e.) Native chemical ligation.

However, by far the most common technique is the so-called native chemical ligation of a peptide thioester with an N-terminal cysteine of a second peptide (161). It

2 UNCHARGED CITRULLINE MUTANT 26 is based on the transesterification of the thioester with the thiol of the cysteine side chain. This newly formed thioester undergoes an irreversible S-N acyl shift through an intramolecular five-membered ring to give a native amide bond (162). This kind of reaction was first discovered by Wieland et al. for the coupling of two amino acids (163) and has been turned into a practical method for the ligation of peptides by Dawson and Kent (161). Since the initial transthioesterification is reversible, the fragments can contain other unprotected cysteines. The N-terminal cysteine fragment is generally prepared by SPPS, or by recombinant DNA technology. While the former is straightforward, the latter can be obtained by using proteolytically cleavable tags like the factor Xa recognition sequence (164) or by using modified inteins (vide infa) (165-167). In contrast, the preparation of the thioester component by the use of modified inteins is uncomplicated (168-172), while, especially for longer peptides, synthesizing it by common Fmoc-SPPS remains quite a challenge (173-177). Many synthetic and semisynthetic proteins of all sizes have been prepared by native chemical ligation, including proteins that contain noncoded amino acids, including β-amino acids and D-amino acids, and glycosilations and other posttranslational modifications (178-182). The advantage of native chemical ligation is that there is virtually no modification that cannot be accommodated within a synthetic fragment (180). The native chemical ligation has recently been expanded to selenocysteine, homo-cysteine, and homo-selenocysteine (183-186). The development of methods for postligational desulfurization also allows creating ligation sites containing alanine instead of cysteine (187). Another modification of the technique permits ligation at an Xaa-Gly or Gly-Xaa site by attaching a removable HSCH2CH2O- moiety or a 2-mercaptobenzyl group to the α-amino group involved in the ligation (188- 190). While the use of this modification allows, in principle, for any amino acid at both ligation sites, the fact that the rearrangement of the initially formed thioester proceeds through a 6-membered cyclic intermediate, which slows it down, precludes using two α-substituted amino acids at the junction. Although ligation chemistry is attractive for the preparation of BsCM variants, the position of the residue that we want to substitute – position 90 – would still require preparation of a very long fragment (> 80 amino acids), which represents a daunting prospect.

2 UNCHARGED CITRULLINE MUTANT 27

Inteins

As compromise between a fully recombinant and a fully synthetic approach to BsCM variants, a semisynthetic strategy is partially appealing. Specifically, the use of intein technology to prepare the N-terminal fragment activated as a thiol ester is an attractive possibility (168-172). Inteins are internal protein domains that can auto-catalytically excise themselves from a larger protein, thereby ligating the two domains called exteins that are fused to their N- and C-termini. They are analogous to introns at the RNA level. So far, more than 100 inteins have been identified (191). They usually consist of two catalytic domains. One is an endonuclease that maintains the presence of the intein in its host by a homing mechanism (192, 193), the other is the actual splicing site. The biological significance of inteins is unknown. They could just be selfish proteins that want to replicate themselves, but there are speculations that they might have some catalytic function after they cleave themselves out of the host protein (194). About 10% of all inteins, among them the GyrA from Mycobacterium xenopi and the DnaB from Porphyra purpurea, do not have an endonuclease domain (194, 195). Inteins have several conserved residues that are part of the splicing site. The first residue at the N- terminus of the intein is a cysteine or a serine, the last two amino acids at the C- terminus are a histidine and an asparagine or in rare cases a glutamine (191, 194). Additionally there is the sequence TXNH or NXTH which is close in space but remote in sequence to the C- and N-terminus of the intein as found in the crystal structure of the GyrA intein (194). Another requirement is the presence of a cysteine, a serine, or a threonine as the N-terminal residue of the C-terminal extein.

2 UNCHARGED CITRULLINE MUTANT 28

O HO O H2N N-Extein NH Intein C-Extein CO H N 2 H NH HS 2 O N-S acyl shift

HO O

H2N Intein C-Extein CO2H O N H NH2 H2N N-Extein S O transesterification

H2N N-Extein O O H N Intein C-Extein CO H 2 N 2 H NH HS 2 O succinimide formation O O H2N N-Extein H2N Intein O

NH C-Extein CO2H HS H2N O O-N acyl shift

O H N Intein 2 HO O NH HS C-Extein CO2H H2N N-Extein N O H Scheme 2.3. 4-step mechanism of protein splicing (194).

Splicing by the intein is achieved as shown in Scheme 2.3 and is reminescent of the native chemical ligation strategy. As has been deduced from the crystal structure of the GyrA intein (194), the intein is folded such that the two amide bonds that have to be cleaved are in close spatial proximity, which makes the religation of the two exteins possible. The cleavage of the N-terminal peptide bond by nucleophilic attack by the side chain of the adjacent serine or cysteine is facilitated by the formation of hydrogen bonds of threonine and histidine of the conserved TXNH motive to its NH group and of threonine and asparagine to its carbonyl oxygen. This arrangement forces the amide in an unfavorable cis-conformation, supports the nucleophilic attack by polarizing the carbonyl bond, and aids in the elimination of the extein by protonating the newly formed amine. After intramolecular transesterification the amide bond at the intein’s C- terminus is cleaved by the formation of an aspartimide with the help of the adjacent

2 UNCHARGED CITRULLINE MUTANT 29 histidine that protonates its NH group. The last step is the irreversible O-N or S-N acyl shift to form the peptide bond between the two exteins (194, 196). Inteins have been exploited extensively as tools for protein engineering. Some inteins can be split into two fragments in the endonuclease domain. They can be expressed in separate cultures and, when recombined after purification, the two fragments are reassembled and can ligate the attached exteins (197-200). This type of trans-splicing has been used to produce cytotoxic proteins and to segmentally label proteins for NMR measurements (201-203). Another use of inteins is the production of proteins with a C-terminal thioester or an N-terminal cysteine that can be used for native chemical ligation (165-172). For the generation of thioesters the intein is modified such that the asparagine that cyclizes to the imide in the second last step is mutated so that the intein splicing stops at one of the (thio)ester intermediates (165, 196, 204). The N- terminal extein can then be captured by the addition of an exogenous thiol to give a reactive C-terminal thioester.

Strategy for Preparing BsCM Variants with non-standard Amino Acids at Position 90

We want to prepare a mutant protein that, apart from the mutation, is as closely related to the wild-type as possible. We also need a sufficient amount of material for biochemical characterization and, if possible, for the determination of its crystal structure. This ruled out the use of stop codon suppression because of the low yield of the in vitro methods and the absence of in vivo systems suitable for the amino acids that we intended to introduce. We therefore have to rely on a synthetic approach to introduce the unnatural amino acid. The synthesis of the full length enzyme by SPPS is probably as technically challenging as stop codon suppression. We decided therefore to prepare the enzyme by the ligation of two fragments using native chemical ligation. To be able to synthesize a protein by native chemical ligation of two or more fragments, one needs to have a cysteine at the junction. BsCM contains two cysteines (Cys75 and Cys88). One of them is very close to the intended mutation site at position 90. We are

2 UNCHARGED CITRULLINE MUTANT 30 thus lucky that we can directly exploit Cys88 for the ligation and do not have to introduce a mutation to a cysteine at a suitable ligation site. Splitting the enzyme at this position leaves us with two fragments of 40 and 87 amino acids (Scheme 2.4). Using only fragments prepared by SPPS would represent a considerable challenge for the preparation of the Arg90Cit BsCM mutant, because the synthesis of an 87 residue peptide is far beyond standard Fmoc-SPPS. We therefore decided to employ a semisynthetic approach and to prepare the large N-terminal fragment of the enzyme by recombinant DNA technology using inteins. The remaining “small” fragment of 40 residues that contains the mutation is still quite a challenge for Fmoc-SPPS and requires substantial optimization for the successful synthesis. It thus turns out that we are in the fortunate situation that each fragment can be prepared using the technology that is best suited for its preparation, i.e. the synthetic fragment has the N-terminal cysteine while the recombinant fragment is activated as a C-terminal thioester.

Scheme 2.4. Schematic view of the native chemical ligation of BsCM(1-87) and BsCM(88-127) with the mutated residue at position 90 shown in yellow.

2.3 Results

BsCM and the Arg90Cit variant were successfully prepared by the native chemical ligation strategy just outlined (124, 161, 163), exploiting Cys88 to mediate

2 UNCHARGED CITRULLINE MUTANT 31 segment condensation (Scheme 3.2). The N-terminal peptide fragment corresponding to residues 1-87 was biosynthesized in E. coli strain KA13 as a fusion with the Mxe GyrA intein and a chitin-binding domain (172), which gives a higher yield than the analogous fusion with the Sce VMA intein. KA13 has deletions of both endogenous E. coli chorismate mutase genes (46), so that contamination from chromosomally encoded chorismate mutases can be excluded. The fusion protein was purified on a chitin column and the intein splicing intermediate was captured with sodium 2-mercaptoethane sulfonate (MESNA) to give the BsCM(1-87) fragment as a thioester. This fragment was always contaminated to some extent with the ε-caprolactam that forms by attack of the side chain amine of the C-terminal lysine at the thioester. The use of MESNA reduced the amount of this side reaction from around 40% with N-methyl-mercapto-acetamide (NMMAA) to around 15%. Another problem was the leakiness of the chitin affinity column, resulting in simultanous elution of the cleaved intein and chitin binding domain fusion and the uncleaved full fusion protein from the column. The extent of this leakiness was reduced by the use of Tris instead of phosphate based buffers. The C-terminal fragments, corresponding to residues 88-127 with either arginine or citrulline at position 90, were prepared by solid phase peptide synthesis using modified Fmoc protocols. In a first attempt to synthesize BsCM(88-127) it was found that standard conditions (Wang Resin (134), FastMoc® protocol, Applied Biosystems (205)) did not yield any detectable peptide longer than 20 amino acids and produced many deletion sequences. Therefore, the conditions were modified by using a NovaSyn TGA-resin as solid support, which has polyethyleneglycol spacers between the polystyrene and the Wang linker (129, 132), and by incorporating amino acids Ser120 and Leu119 as the pseudoproline dipeptide Fmoc-Leu-Ser(ΨMe,Mepro)-OH (Sp1, see Figure 2.2). Both measures were adopted to increase solubility. The pseudoproline insertion has the additional benefit of preventing unwanted hydrogen bonding interactions in the growing peptide, thereby avoiding β-sheet formation (206-208). As a further measure to increase the coupling yields and to avoid deletion sequences, the coupling and Fmoc-deprotection times were significantly lengthened.

2 UNCHARGED CITRULLINE MUTANT 32

O N O

COOH

O N H

Figure 2.2. The pseudoproline dipeptide Fmoc-Leu-Ser(ΨMe,Mepro)-OH.

Aspartimide formation (Scheme 2.5) during peptide synthesis made it necessary to replace Asp102 in these peptides by glutamate. Control experiments with recombinant D102E BsCM (BsCM*) confirmed that this mutation does not significantly alter the catalytic properties of the enzyme (Table 2.2). It was further observed during peptide synthesis that the triple coupling of Arg105 (4 h each) only proceeded to about 60-70% completion. Even though control experiments with recombinant R105K BsCM and the double mutant D102E R105K BsCM (R105K BsCM*) confirmed that this mutation does also not significantly alter the catalytic properties of the enzyme (Table 2.2), it was decided to omit it, sacrificing high yield in favor of sequence identity. Following cleavage from the solid support, the synthetic BsCM(88-127)* fragments 9 and 10 were purified by HPLC and separately coupled with the BsCM(1-87) thioester.

O R

R'' R' N H R O O R O OH R'' R' R'' R' N N H O O OtBu O O O

O R'' OH O

H N R'

O R Scheme 2.5. Aspartimide formation by aspartate during SPPS and possible reopening to two products that are not distinguishable by mass spectrometry.

Chemical ligations were performed with ca. 1 mM of each peptide in 100 mM Tris-HCl buffer (pH 8) containing 6 M guanidinium chloride and 2.5% thiophenol for

2 UNCHARGED CITRULLINE MUTANT 33

24 hours. The ligated polypeptides, BsCM* and Arg90Cit BsCM*, were subsequently folded by 100-fold dilution in 50 mM glycine buffer (pH 8.9) containing 5% isopropanol and 10% glycerol. After concentration, the folded proteins were purified by ion-exchange chromatography on a MonoQ column. A separate dedicated column was

used for the Arg90Cit variant to preclude contamination by another chorismate mutase.

it Bs it

M M C

C kD

Bs Arg90 LMW Figure 2.3. SDS-polyacrylamide gel of Bs 94 Arg90Cit BsCM*, semisynthetic BsCM*, and 67 BsCM. 43 30 20 14

Figure 2.4. ESI-MS analysis of BsCM* and Arg90Cit BsCM*.

CD Gel filtration BsCM* BsCM Arg90Cit BsCM* Arg90Cit BsCM*

Figure 2.5. CD spectroscopy of BsCM and Arg90Cit BsCM* (left) and analytical size exclusion chromatography of recombinant BsCM* and Arg90Cit BsCM* (right).

2 UNCHARGED CITRULLINE MUTANT 34

The semisynthetic proteins were analyzed by electrospray ionization mass spectroscopy (ESI-MS) and shown to have the expected mass (BsCM*: found 14,502 ± 2 Da, expected 14,503 Da; Arg90Cit BsCM*: found 14,503 ± 2 Da, expected 14,504 Da) (Figure 2.4). Their circular dichroism spectra are superimposable on that of the recombinant enzyme purified under native conditions, indicating identical overall secondary structure (Figure 2.5). Like recombinant BsCM, they are homotrimeric in solution, eluting as a single peak with identical retention times from a Superose 12 size- exclusion column (Figure 2.5).

Table 2.1. Analytical ultracentrifugation of BsCM* and Arg90Cit BsCM*. Independent species model Continuous mass model Recombinant BsCM* 39,200 Da 38,400 Da Semisynthetic BsCM* 41,500 Da 41,000 Da Arg90Cit BsCM* 43,600 Da 44,000 Da Expected mass for a trimer of BsCM*: 43,509 Da; of Arg90Cit BsCM*: 43,512 Da

Their trimeric quaternary structure was additionally confirmed by sedimentation velocity ultracentrifugation, which yielded average molecular masses of 41,500 Da and 43,600 Da for semisynthetic BsCM* and for the citrulline variant, respectively, in good agreement with the expected mass for a trimer of 43,509 Da (Table 2.1).

Table 2.2. Kinetic parameters of BsCM variants (50 mM phosphate buffer, pH 7.5, 30 °C). Errors on all parameters were less than 15%.

Enzyme kcat Km kcat/Km kcat/kuncat Ki (s-1) (µM) (M-1s-1) (µM) BsCM* 46 98 4.7×105 4.0 ×106 1.2 R105K BsCM 42 105 4.0×105 3.7 ×106 n.d. R105K BsCM* 52 130 4.0×105 4.5 ×106 n.d. Arg90Cit BsCM* 0.0026 270 9.6 230 6.8

The enzymes were assayed as previously described (83) at 30 °C in 50 mM phosphate buffer at pH 7.5 (Figure 2.6). Ki values were obtained by standard inhibition assays with 1 (38, 83, 209) (Figure 2.7). The semisynthetic and the recombinant version of BsCM*, the B. subtilis chorismate mutase containing the D102E mutation, had the same kinetic parameters within experimental errors. Thus the recombinant BsCM*

2 UNCHARGED CITRULLINE MUTANT 35

-1 protein gave kcat = 49 ± 2 s and Km = 105 ± 12 µM, in good agreement with the published kinetic parameters for wild-type BsCM (83, 210). It was inhibited by 1 with a

Ki of 1.0 ± 0.2 µM. In contrast for the Arg90Cit BsCM* mutant a much lower kcat of -1 0.0026 ± 0.0001 s and a three-times increased Km of 270 ± 40 µM was measured. Its Ki value of 6.8 ± 0.2 µM displayed a similar increase as its Km.

BsCM* Arg90Cit BsCM* 40 0.0025

35 0.002 30 ) )

25 -1 -1 0.0015 20 k -1

= (46 ± 1) s (s /[E] /[E] (s /[E] 0.001 -1 0 0 15 cat k = (0.0026 ± 0.0001) s v v K = (98 ± 4) µM cat 10 m K = (270 ± 40) µM 5 -1 -1 0.0005 m k /K = 4.7 x 10 M s k /K = 9.6 M-1s-1 5 cat m cat m 0 0 0100200 300 400 500 0 200 400 600 800 1000 1200 1400 1600 [Chorismate] (µM) [Chorismate] (µM) Figure 2.6. Kinetic assays of semisynthetic BsCM* and Arg90Cit BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C).

BsCM* Arg90Cit BsCM* 0.2 3000

2500 0.15

2000 0.1 /[E] ) (s) /[E] ) /[E] ) (s) /[E] ) 0 0 1500 y = 1067 + 75.17 x 1/(v 1/(v 0.05 y = 0.0418 + 0.0222 x R = 0.99396 1000 R = 0.9835 K = (6.8 ± 1) µM K = (1.2 ± 0.1) µM i i 0 500 -1 0 1 2 3456 -5 0 5 10 15 20 25 [I] (µM) [I] (µM) Figure 2.7. Inhibition assays of semisynthetic BsCM* and Arg90Cit BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C).

To exclude the possibility that the observed residual activity is due to a contamination of the citrulline with 100 ppm arginine, we subjected Fmoc-citrulline to a reaction with 9,10-phenanthrenquinone (PQ) and NaOH in 70% ethanol which selectively modifies arginines (Scheme 2.6) and converts them into a fluorescent derivative (211), allowing detection by HPLC equipped with a fluorescence detector.

2 UNCHARGED CITRULLINE MUTANT 36

NH2 O O NaOH O N + N NH2 N HO O 70% EtOH N NHFmoc HO NHFmoc Scheme 2.6. Reaction of Fmoc-arginine with 9,10-phenanthrenequinone and NaOH in 70% ethanol.

To calibrate the results, the reaction was also performed with a sample of Fmoc- arginine (25.8 µM Fmoc-arginine, 34 µM PQ), which was injected at different dilutions. The signal between 26 and 27 minutes from the 22-times higher concentrated Fmoc- citrulline sample [560 µM Fmoc-Cit, 1.7 µM PQ (1:20), or 34 µM PQ (1:1)] was not detectable over background (Figure 2.8, pink chromatogram, PQ) for the sample with the lower PQ concentration (Figure 2.8, green chromatogram, 1:20). Even when a 20- fold higher concentration (Figure 2.8, blue chromatogram, 1:1) of 9,10- phenanthrenequinone (PQ) was used, the peak area was more than 10-times less than that of the 1:200 diluted Fmoc-arginine sample (Figure 2.8, orange chromatogram, 1:200). A possible contamination with arginine can therefore be estimated to be lower than 1 : 46,000. This observation together with the fact that Edman degradation of Arg90Cit BsCM(88-127)* 10 did not detect any arginine above background at position 90, and that an arginine lacking a side chain protection group cannot be incorporated by SPPS as has been shown by our difficulties synthesizing BsCM(88-93) 11, allows us to exclude contaminating arginine as the source of the activity of the citrulline mutant.

2 UNCHARGED CITRULLINE MUTANT 37

Figure 2.8. HPLC with fluorescence detection of Fmoc-arginine and Fmoc-citrulline modified with 9,10-phenanthrenequionone at different concentrations.

2.4 Discussion

Semisynthetic BsCM* is fully active as a catalyst, affording kcat and Km parameters for the chorismate mutase rearrangement that are indistinguishable from those of its recombinant counterpart (Table 2.2). The Arg90Cit variant is also active, albeit substantially less so than the wild-type enzyme. Replacement of arginine by 4 citrulline causes a >10 -fold decrease in kcat and a more modest 2.7-fold increase in the

Km value for chorismate (Table 2.2). The small change in Km suggests only minor perturbation of the ground state Michaelis complex, although the non-equality of Km and the dissociation constant for chorismate for wild-type BsCM (212) precludes accurate quantification of this effect (213). In contrast, the destabilizing effect of the Arg90Cit mutation on apparent transition state binding can be estimated directly from ‡ the equation ∆∆G = RTln(kcat/Km)mut/(kcat/Km)WT to be 27.2 kJ/mol (6.5 kcal/mol). Effects on catalysis and protein stability of similar magnitude have been observed previously upon removal of a charged hydrogen-bond donor or acceptor from a buried

2 UNCHARGED CITRULLINE MUTANT 38 salt bridge (214, 215), supporting the idea that the guanidinium group of Arg90 forms a complementary electrostatic interaction with the developing negative charge at the ether oxygen of chorismate in the transition state. The ligand binding properties of Arg90Cit BsCM* were further probed with the conformationally constrained inhibitor 1 (38, 209), which mimics the geometry of the chorismate mutase transition state reasonably well but not its dissociative character.

Paralleling the small increase in Km for chorismate, the inhibition constant value for 1 increases 5.7-fold upon mutation (Table 2.2), which corresponds to a 4.6 kJ/mol (1.1 kcal/mol) less favorable free energy of binding. The moderate decrease in affinity for the transition state analogue contrasts dramatically with the much larger destabilization of the true transition state. The neutral urea group of citrulline apparently interacts only slightly less well than the positively charged guanidinium group with the neutral ether oxygen of the stable inhibitor (and, by analogy, that of chorismate in the ground state), but it is more than four orders of magnitude worse at accommodating the partially anionic ether oxygen in the transition state. Not surprisingly, even more deleterious effects are observed when either of the partners in the salt bridge is replaced with a non-hydrogen bonding group. For example, mutation of Arg90 to alanine results in a complete loss of catalytic activity (50, 82, 83), whereas wild-type BsCM is unable to catalyze the analogous Cope rearrangement of carbaprephenate into carbachorismate (216), in which an apolar methylene group replaces the ether oxygen. In a computational paper published in response to our studies (217), Hur and Bruice (93) claim that the low activity of Arg90Cit BsCM* can be explained by a significantly reorganized active site compared to the wild-type structure (Figure 2.9). According to their calculations, the NH group of citrulline is turned away from the ether oxygen and interacts with Glu78 and the side chain carboxylate of chorismate interacts mainly with Arg116 instead of Arg7 like in the wild-type, because Arg7 in the mutant is supposed to move towards Glu78 to make up the missing interaction with a charged Arg90. This causes the distance between the two reacting carbons of bound chorismate to be 4.5 instead of 3.5 Å in the wild-type enzyme, preventing it from forming a near attack conformer (NAC) of the substrate. The NAC is claimed to be rarely attained in water. Similar results were obtained for the complex with TSA (Figure 2.9). The

2 UNCHARGED CITRULLINE MUTANT 39

moderate increase in Km and Ki for the mutant is explained by two compensating factors. The loss of binding to the ether oxygen is compensated by the stronger interactions of the carboxylates with Arg116 and Arg63.

Arg7 Tyr108 NH2 HN Arg116 HN H + Arg7 HN

HO H2NNH+ 2 + H O H2N NH2 HN O HN Arg90 - - O O + NH HN HN+ NH2 N H O Arg116 H O HN SH NH2 - H - Arg63 + O Cit90 NH O O H Cys88 HN HN H + O N H O- O HN O - HN H O HO Arg63 HO Glu78 H Glu78 O SH Cys75 Phe57 Figure 2.9. Electrostatic interactions between transition state analogue 1 and BsCM (left, sketch from crystal structure 2CHT) and between TSA and Arg90Cit BsCM (right, MD simulation data; Figure adapted from (93)).

In another publication Warshel questions the meaning of the apparent NAC effect (84). He tries to exactly define the origin of the reduction of the activation barriers in the enzyme as compared to the uncatalyzed reaction in water, which are associated with kcat and kuncat (84, 97), and claims further that the existence of a NAC, which could also be defined as reactant state destabilization, does not change any of the relevant activation barriers. He argues convincingly that NAC is not “a genuine reason for catalysis, but merely reflects the result of electrostatic transition state stabilization”. While we can not exclude the possibility that the low activity of Arg90Cit BsCM* is a result of a reorganized active site, this seems unlikely considering the modest changes in Km and Ki and the observation that a variety of active site mutations cause only minor perturbations of the pocket (83). To clarify this question we produced 20 mg of the mutant enzyme and initiated a cooperation with the group of U. Krengel at the University of Oslo to determine its crystal structure. Preliminary results of Arg90Cit BsCM* indicate a trimeric structure that is virtually indistinguishable from the wild- type enzyme (root mean square difference = 0.4 Å), even in the active site. However, this statement should be accepted with caution, as the current dataset is not quite optimal and refinement is not yet completed (Rmerge = 11.3%, current Rcryst = 26.6%, current Rfree = 33.5%, max. resolution = 2.2 Å). The structure of the complex with the transition state analogue 1 has not yet been completed, but electron density observed in

2 UNCHARGED CITRULLINE MUTANT 40

OMIT maps of the active site indicates that it will be possible to resolve this issue in the near future. Taken together, our results demonstrate the importance of Arg90 at the active site of BsCM as a positively charged hydrogen bond donor involved in selective stabilization of chorismate in the transition state. It is noteworthy that structurally unrelated AroQ mutases like EcCM (33) and ScCM (65) have a similarly positioned cation, albeit a lysine rather than an arginine, whereas the relatively inefficient catalytic antibody 1F7 lacks this feature (66-68). Efficient catalysis of the chorismate mutase rearrangement evidently requires more than an active site that is simply complementary in shape to the reactive substrate conformer (92-94); electrostatic stabilization of the polarized transition state appears to be paramount.

3 BsCM Variants Containing homo-Lysine or Difluoro-Arginine at Position 90 in the Active Site

3.1 Introduction

As we have seen in Chapter 2 mutation of the cationic Arg90 to the isosteric but uncharged amino acid citrulline causes a dramatic loss in kcat with only a relatively modest increase in the Km value for chorismate. What about unnatural but positively charged residues that have proper steric properties? Although substitution of Arg90 with a positively charged lysine (pKa1 = 2.2, pKa2 = 8.9 and pKa3 = 10.3) leads to substantial reductions in catalytic efficacy (50, 82, 83), it is possible that the side chain of lysine is too short to reach far enough into the active site to contact the ether oxygen of chorismate in the transition state. Lysine is one methylene unit shorter than arginine and second site mutations are required to restore some of the activity of a lysine mutant (83). Alternatively, the loss in catalytic efficacy might reflect the differences of a guanidinium versus an ammonium group that could impair a proper arrangement in the active site and/or might not allow for the required two hydrogen bonds to the ether oxygen. All AroQ class enzymes with a lysine at the position corresponding to Arg90 in BsCM also have a second residue in the active site that helps stabilize the developing negative charge in the transition state by formation of a second hydrogen bond to the ether oxygen (33, 58, 60, 65). To test this possibility we considered introducing homo- lysine 7 in place of Arg90 in BsCM, a lysine analogue with an additional CH2 group (Figure 3.1). This longer residue would, in principle, allow the ammonium functional group to make the required contact to the ether oxygen. As an alternative to homo-lysine, we also considered introducing a residue that is essentially isosteric to Arg90 but with a tuneable charge state. The apparent pKa values of the active site of BsCM were determined to be 4.3 and 9.1 (218). If the activity loss at high and low pH values is not due to denaturation of the protein, it might 3 CHARGED MUTANTS 42 be caused by protonation or deprotonation of one of the active site residues or the substrate. For example the lower pKa of 4.3 is probably due to protonation of one of the carboxylates of chorismate or of Glu78 (pKa2 = 4.1), which is essential for catalytic function (55). Since the only amino acids around the active site that could be deprotonated in the range of pH 9 are cysteine at position 75 (pKa1 = 1.7; pKa2

(SH) = 8.4; pKa3 = 10.8) and arginine at positions 7, 90, 63, and 116 (pKa1 = 2.2; pKa2 = 9.1; pKa3 (guanidinium) = 13.2 (219)), the apparent pKa of 9.1 could be due to deprotonation of the side chain of either cysteine or arginine. In the case of arginine, this would imply a rather large change of its pKa value in the environment of the active site due to the large number of positively charged residues. Given the results presented in the previous chapter, it is not unreasonable to speculate that the activity loss (kcat/Km) could be due to the deprotonation of Arg90.

NH O O

N NH2 NH2 H HO HO F F NH2 NH2 8 7 Figure 3.1. Structures of 4,4-difluoro-L-arginine 8 and homo-lysine 7.

To clarify this question, we therefore decided to install an isosteric arginine analogue with a lower pKa for its side chain functional group at position 90. One way to do this is to add small electron withdrawing substituents to the side chain. Fluorine, which has a van der Waals radius of 1.47 Å and whose steric demand is inbetween that of hydrogen (1.20 Å) and oxygen (1.52 Å) (220), is the ideal choice. Its physiological size is even more similar to a hydrogen and much smaller than that of a methyl group as determined by the taste and odor of fluorine and methyl derivatives at positions that contain hydrogen in the natural compound (221). The amino acid 4,4-difluoro-L- arginine has been prepared previously and was reported to have pKa values of 1.4 (acid), 8.1 (amine), and 11.2 (guanidinium) (219). Addition of two fluorines thus lowers the pKa value of the guanidinium group by 2 units. If the decrease in BsCM activity at high pH is due to deprotonation of Arg90, then replacement of this residue with difluoro arginine should lead to a shift in the basic limb of the pH-rate profile.

3 CHARGED MUTANTS 43

In this chapter, the synthesis and characterization of the Arg90homo-Lys and

Arg90F2Arg mutants of BsCM are described. Our results provide further insights into the mechanism of this fascinating enzyme.

3.2 Results

Arg90homo-Lys BsCM*

Synthesis

For the synthesis of Arg90homo-Lys BsCM(88-127)* 12 with the commercially available Fmoc-homo-Lys(Boc)-OH, the same strategy as for the synthesis of Arg90Cit BsCM(88-127)* 10 described in Chapter 2 was chosen. First, a large amount of resin loaded with fully protected Fmoc-BsCM(92-127)* was synthesized. This could be stored at -20 °C for a prolonged period of time. Some of the resin, corresponding to a loading of 0.05 mmol, was used for the elongation with the last four amino acids. Since couplings of lysines did not show any special difficulties, double coupling was sufficient for this residue. Following deprotection and cleavage from the solid support the synthetic Arg90homo-Lys BsCM(88-127)* fragment was purified by preparative HPLC. After native chemical ligation with BsCM(1-87) thioester (MESNA), folding, and purification, as described in Chapter 2 for BsCM* and Arg90Cit BsCM*, 5.2 mg of the mutant enzyme Arg90homo-Lys BsCM* was obtained.

Characterization

The semisynthetic mutant was analyzed by electrospray ionization mass spectroscopy (ESI-MS) and shown to have the expected mass (Arg90homo-Lys BsCM*: expected 14,489 Da, found 14,489 ± 2 Da). Its circular dichroism spectra like

3 CHARGED MUTANTS 44 that of Arg90Cit BsCM* is superimposable on that of the recombinant enzyme purified under native conditions, indicating identical overall secondary structure (Figure 3.2). Like recombinant BsCM, it is homotrimeric in solution, eluting as a single peak with comparable elution volume from a Superose 12 size-exclusion column. The slight difference of 0.3 ml in the elution volume of the mutant and the wild-type are within the reproducibility error of the system. The elution volume corresponds to a calculated molecular weight of around 40 kDa (Figure 3.3).

BsCM Arg90Cit BsCM* Arg90homo-Lys BsCM*

Figure 3.2. CD spectroscopy of BsCM, Arg90CitBsCM*, and Arg90homo-Lys BsCM*.

BsCM* wt BsCM* R90Cit BsCM* R90Homo-Lys

Figure 3.3. Analytical size exclusion chromatography of recombinant BsCM*, Arg90CitBsCM*, and Arg90 homo-LysBsCM*.

Table 3.1. Analytical ultracentrifugation of Arg90 homo-Lys BsCM* (Error in the range of 10%). Found mass Expected mass Arg90homo-Lys BsCM* 37,000 Da 43,467 Da

3 CHARGED MUTANTS 45

Their trimeric quaternary structure was additionally confirmed by sedimentation equilibrium ultracentrifugation, which yielded an average molecular mass of 37,000 Da for the homo-lysine variant, in good agreement with the expected mass for a trimer of 43,467 Da (Table 3.1). The sample contained around 2-3% of a contaminant (13,800 Da), probably caused by some degraded protein.

Table 3.2. Kinetic parameters of Arg90homo-Lys BsCM* and BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C); Errors on all parameters were less than 15%.

Enzyme kcat Km kcat/Km kcat/kuncat Ki (s-1) (µM) (M-1s-1) (µM) BsCM* 46 98 4.7×105 4.0 ×106 1.2 Arg90homo-Lys BsCM* 0.013 510 25.5 1150 7.6

The enzyme was assayed as previously described at 30 °C in 50 mM phosphate buffer at pH 7.5 (Figure 3.4) (83). The experimentally determined Km value of 510 µM is five-times higher than that for BsCM*, whereas the kcat value is 3500-fold lower than that for BsCM* and only five-times higher than kcat for Arg90Cit BsCM*. The Ki value, which was obtained by a standard inhibition assay with 1 at 100 µM chorismate (38, 83,

209) (Figure 3.4), increased to a similar extend as Km.

Arg90homo-Lys BsCM* Arg90homo-Lys BsCM* 0.012 500

450 0.01 400

) 0.008 (s) 1

350 - ) (s 0.006 300 /[E] 0 /[E] -1 0 k = (0.013 ± 0.001) s (v 250 v 0.004 cat /

1 y = 176.18 + 15.57 x K = (510 ± 70) µM 200 m R = 0.99948 0.002 k /K = 25.5 M-1s-1 K = (7.6 ± 1) µM cat m 150 i

0 100 0 500 1000 1500 2000 -50 5101520 [Chorismate] (µM) [I] (µM) Figure 3.4. Kinetic assays (left) and inhibition assays (right) of Arg90homo-Lys BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C).

3 CHARGED MUTANTS 46

Arg90F2Arg BsCM*

Synthesis of Fmoc-4,4-Difluoro-L-Arginine(Pbf) 13

In order to produce the Arg90F2Arg BsCM* mutant it was first necessary to synthesize the appropriately protected amino acid needed for the SPPS of Arg90F2Arg BsCM(88-127)*. The previously reported synthesis of 4,4,-difluoro-L-arginine was not published with any experimental details and consequently the resynthesis still constituted a major challenge (219). The synthesis started from well-known (R)-Garner’s aldehyde (222) which was prepared starting from N-Boc-D-serine methylester 14 in two or three steps according to Scheme 3.1. Cyclization with 2,2-dimethoxypropane (DMP) to the 1,3-oxazolane 15 proceeded smoothly in high yield (219, 222-228). The 1H-NMR spectrum of 15 indicated the presence of two very similar products in a ca. 1:1 ratio. However, when 15 was selectively irradiated at the frequency of a selected signal of one product, the corresponding signal of the other product also appeared in the spectrum. This experiment proved that the two products are one substance that exist as two rotamers of the Boc group on the NMR time scale (222, 229, 230). The duplicate signals collapsed to one set when the sample was heated to 60 °C. Ester 15 was subsequently reduced to aldehyde 17 with DiBAlH (225, 228, 231- 233). Depending on the reaction time and temperature, more (long reaction time and / or higher temperature) or less 16 is formed as by-product during the reduction step. This by-product could be converted almost quantitatively back to the desired aldehyde 17 by Swern oxidation in a separate step (225, 228, 233).

3 CHARGED MUTANTS 47

Scheme 3.1. Synthesis of (R)-Garner’s aldehyde 17. O O O

O O H HO a O b O HN N N O O O

O O O 14 15 17

b c

OH O N O

O 16 a.) DMP, cat. BF3O(Et)2, acetone, r.t., 12 h, 92%; b.) DiBAlH, toluene, -78 °C, 4 h, 40- 70% (17) and 10-40% (16); c.) DMSO, oxalyl chloride, DCM, DIEA, -78 °C, 2 h

The fluorines were introduced by a Reformatsky reaction of (R)-Garner’s aldehyde 17 with ethyl bromodifluoroacetate under sonication (Scheme 3.2) (219, 234- 240). The reaction proceeded smoothly and gave the two diastereomers of 18 in 85% yield and a ratio of 9:1. The mixture was not separated, but used directly for the next reaction.

Scheme 3.2. Reformatsky reaction of (R)-Garner’s 17 aldehyde with ethyl bromo- difluoracetate to give 18. O OH O

H O a O O F N N F O O

O O 17 18 a.) Zn, THF, r.t. → 50°C, 12 h, sonication, 85%

The hydroxy group of 18 that was formed in the Reformatsky reaction was removed by Barton-McCombie deoxygenation (Scheme 3.3), a two step process that involves the activation of the alcohol as a thiocarbonyl imidazolide (219, 235-238, 241) and radical deoxygenation of this activated alcohol using triethylsilane as reducing

3 CHARGED MUTANTS 48 agent and solvent and a large excess of the radical generating benzoyl peroxide (219, 235-238, 241-243). The reaction is performed under nitrogen and great care must be taken to allow the developing overpressure to dissipate. The excess triethylsilane is difficult to remove at this step, even by silica gel chromatography, but it can be carried on to the next step without problem and can be easily removed at that stage. Both steps proceeded smoothly to give 19 in 79% and 20 in 90% yield.

Scheme 3.3. Barton-McCombie deoxygenation of alcohol 18 to give 20. N

N OH O O S O O O O a b O O F F F F N N O O O O F F O N O O

O 18 19 20 a.) thiocarbonyl diimidazole, 1,2-dichloroethane, r.t., Ar, 20 h, 79%; b.) Et3SiH, benzoyl peroxide, reflux, N2, 1 h, 90%

The next three steps are side chain modifications of 20 to transform the ethyl ester into a protected amine (Scheme 3.4). First, the ester is converted to the amide 21 with liquid ammonia (219, 234, 236), then it is reduced with sodium bis(2- methoxyethoxy)aluminum hydride (RedAl) to the air sensitive amine 22 (219, 244- 247), which is temporarily protected with the benzyloxycarbonyl group in the third step to give 23 (219, 236). To work up the RedAl reduction, the reaction mixture was stirred with sat. KNa-tartrate (aq) to complex the aluminum. The initially liquid mixture almost solidified before reliquifying within one hour. Since the amine 22 is sensitive to air and decomposes on a silica gel column, it was used directly for the next step. All three steps proceeded in high yield.

3 CHARGED MUTANTS 49

Scheme 3.4. Amidation, reduction, and Cbz protection of ester 20 to give 23. O O O

O NH2 N O a b, c H O O O F F F F F N N N F O O O

O O O 20 21 23 a.) NH3 (l), ether, -78 °C, 30 min, 72%; b.) RedAl, toluene, 0 °C, 1 h, r.t., 3 h; c.)

CbzCl, NaHCO3, EA, 0 °C, 45 min, 86% (two steps)

The temporary protection of the amine with the Cbz group is necessary in order to avoid problems with the subsequent ring opening of 1,3-oxazolane 23 to alcohol 24 under acidic conditions (219, 236, 248). This protecting group was then removed via palladium-catalyzed hydrogenation to give the aminoalcohol 25 in excellent yield (219, 249). In order to assemble the guanidinium group of 26, special activation of the guanidinylation reagent di-Boc thiourea with HgCl2 is necessary (219, 249-255). Although great care had to be taken when working with this highly toxic mercury salt, the reaction also had its pleasant aspects associated with the formation of beautifully colored complexes (Figure 3.5).

Scheme 3.5. Opening of the 1,3-oxazolane, Cbz deprotection and guanidinylation of 23 to give 26. O O N O NH2 N O HO H a, b F F c O HN F N NH N F O H O HO O F HN F O O O O O 23 25 26 a.) TsOH, methanol, 60 °C, 6 h, 96%; b.) H2, cat. Pd(OH)2, ethanol:EA 1:1, r.t. 6 h, quant.; c.) di-Boc thiourea, HgCl2, NEt3, DMF, 0 °C, 60 min, 78%

3 CHARGED MUTANTS 50

Figure 3.5. Mint green color generated during the guanidinylation of 25.

Oxidation of alcohol 26 was accomplished with a large excess of pyridinium dichromate in DMF (219, 236, 256-270). Whereas oxidations of primary alcohols with pyridinium dichromate in DMF give the corresponding acid, the reaction only proceeds to the aldehyde if DCM is used as solvent (258). Since complete removal of chromium at this stage was difficult, crude 27 was directly used for the next deprotection step with TFA (219). After HPLC purification, the unprotected 4,4-difluoro-L-arginine 8 was obtained as a white solid. Overall yield of the amino acid 8 was 6% with respect to the amount of starting material 14 used, which can be compared with the value of 9% calculated from the yields of the individual steps reported in the literature (219). The reported values probably do not include material that was lost in the several experiments necessary to devise optimal reaction conditions.

Scheme 3.6. Oxidation of alcohol 26 and deprotection to 4,4-difluoro-L-arginine 8. O O

O O N N NH

O O a b N NH N NH N NH2 H H H HO HO HO F F F HN F O HN F O F O NH2 O O O

O O 26 27 8 a.) pyridinium dichromate, DMF, r.t., 20 h, 85%; b.) TFA, r.t. 1 h, 52%

All of the following protection and deprotection steps were first tested with the corresponding arginine analogues and then adapted to 4,4-difluoro-L-arginine. The key step toward the final product is the selective protection of the amino group of 8 adjacent

3 CHARGED MUTANTS 51 to the guanidino group (Scheme 3.7). This transformation, which requires strict pH control throughout the reaction, is known for arginine and has a pH optimum around 9.5

(271-274). Since the pKa values for the amino and the guanidino groups of 4,4-difluoro-

L-arginine 8 are known to be 1 and 2 pKa units lower than for arginine, it was expected to be more difficult to achieve a selective reaction and that the optimal pH would be lower. After several tests at different pHs, a value of 9.2 was found to give the best yields for difluoro arginine. By comparing the pH-dependent changes in the 1H-NMR spectra of 8 and 28, it could be shown that the Cbz group was really attached to the amino group (Figure 3.6). When going from low to high pH, the CH proton next to the amino group in 28 showed a change in chemical shift upon deprotonation of the carboxylic acid but not upon deprotonation of the amino group as it did in 8 (Figure 3.6). The chemical shift of the methylene group next to the CH group also changes significantly as a function of pH in the case of the free amino acid 8, whereas it remains pretty constant for Cbz-protected 28.

Scheme 3.7. Protection of 4,4-difluoro-L-arginine 8 to give Fmoc-4,4-difluoro-L- arginine(Pbf)-OH 13. O O HO HO NH O O HN O H2N N N S a b H H F HO F O F F O O F NH HN F HN O HN NH

O NH2 NH2 8 28 29

NH O O O N N S NH H H HO O O F O HN F c O d N N S O H H HO F F O O NH2

30 13 a.) CbzCl, NaHCO3/NaOH (aq) pH 9.2, 4 °C, 16 h, 61%; b.) PbfCl, acetone, r.t., 2 h,

0 °C, 2 h, 45%; c.) H2, cat. Pd/C, methanol, r.t., 15 h, quant.; d.) FmocOSu, NEt3 pH 9 water:CH3CH 2:3, r.t., 1 h, 65%

3 CHARGED MUTANTS 52

Figure 3.6. pH-dependent changes in the 1H-NMR spectra of 8 (left) and 28 (right).

For the final side chain protection of 4,4-difluoro-L-arginine the 2,2,4,6,7,- pentamethyl-dihydrobenzofuran-5-sulfonyl group was used. This is the “standard” protecting group for arginine for Fmoc-type SPPS. Even though Pbf-protected arginine is commercially available, literature protocols for its preparation are rather scarce. We therefore carried out the reaction as described for the related 2,2,5,7,8- pentamethylchroman-6-sulfonyl (Pmc) protecting group (272, 275-278). The reaction gave only a 45% yield of 29 but after HPLC purification around 40% of the unprotected 28 could be recovered and recycled. The removal of the Cbz group with hydrogen and palladium on charcoal as catalyst gave 30 in excellent yield (249, 272, 275).

3 CHARGED MUTANTS 53

The final step in the preparation of a 4,4-difluoro-L-arginine derivative, which is suitably protected for SPPS, is the Fmoc protection of the amino group (Scheme 3.7). This was accomplished with FmocOSu in an aqueous solution of 30 that was adjusted to pH 9 with triethylamine (272, 275). The reaction proceeded well, affording 13 in reasonable yield after HPLC purification. The final compound was fully characterized. Its enantiomeric purity was determined on a chiral CC200/4 Nucleodex-β-OH column using a linear gradient of 40 mM phosphate buffer pH 3.4 and MeOH (80/20 to 95/5 in 10 min). The protected amino acid consisted of a 97.7 : 2.3 mixture of S : R isomers, corresponding to a 95.4% enantiomeric excess. To minimize mishaps, each step was carried out multiple times. Altogether, starting from 75 g or 342 mmol of N-Boc-D-serine methylester 14, 2 g or 2.9 mmol of Fmoc-4,4-difluoro-L-arginine(Pbf)-OH 13 was obtained, corresponding to an overall yield of 0.85% over 17 (or 18) steps. This quantity was sufficient for our purposes, since we anticipated needing 2 mmol (1.37 g) for the intended double couplings to produce Arg90F2Arg BsCM(88-127)* 31 by SPPS.

Synthesis of Arg90F2Arg BsCM*

SPPS of Arg90F2Arg BsCM(88-127)* 31 was performed as described in Chapter 2 for BsCM(88-127)* 9 and Arg90Cit BsCM(88-127)* 10. As in the synthesis of Arg90homo-Lys BsCM(88-127)*, previously prepared resin loaded with fully protected Fmoc-BsCM(92-127)* (0.25 mmol loading) was used for the elongation with the last four amino acids. The only change was that Fmoc-4,4-difluoro-L-arginine(Pbf)- OH 13 was only double coupled to save some material. This turned out to have no major impact on the final yield of the synthesized peptide. The only reason for triple coupling is that incorporation of arginines can be rather tricky (279). As discussed in Chapter 2, this is true for residue Arg105 in BsCM, and the cost of doing one more coupling was lower than the risk of a bad synthesis. This turned out not to be necessary for difluoroarginine. Following deprotection and cleavage from the solid support, the synthetic Arg90F2Arg BsCM(88-127)* fragment was purified by HPLC. After native chemical ligation with BsCM(1-87)

3 CHARGED MUTANTS 54 thioester (MESNA), folding, and purification as described in Chapter 2 for BsCM* and

Arg90Cit BsCM*, 2.1 mg of the mutant enzyme Arg90F2Arg BsCM* were obtained.

Characterization of Arg90F2Arg BsCM*

The semisynthetic variant was analyzed by electrospray ionization mass spectroscopy (ESI-MS) and shown to have the expected mass (expected 14,539 Da, found 14,540 ± 2 Da). Its circular dichroism spectra like that of Arg90homo-Lys BsCM* is superimposable on that of the recombinant enzyme purified under native conditions, indicating identical overall secondary structure (Figure 3.7). Like recombinant BsCM, it is homotrimeric in solution, eluting as a single peak with identical elution volume from a Superose 12 size-exclusion column. The elution volume corresponds to a calculated molecular weight of around 40 kDa (Figure 3.8).

BsCM Arg90Cit BsCM* Arg90homo-Lys BsCM* Arg90F2Arg BsCM*

Figure 3.7. CD spectroscopy of BsCM, Arg90Cit BsCM*, Arg90homo-Lys BsCM*, and Arg90F2Arg BsCM*.

3 CHARGED MUTANTS 55

BsCM* Arg90Cit BsCM* Arg90homo-Lys BsCM* Arg90F2Arg BsCM* on i t p

sor b A

10 15 Elution volume (ml) Figure 3.8. Analytical size exclusion chromatography of recombinant BsCM*, Arg90Cit BsCM*, Arg90 homo-Lys BsCM*, and Arg90F2Arg BsCM*.

Table 3.3. Analytical ultracentrifugation of Arg90F2Arg BsCM*. Error in the range of 10%. Found mass Expected mass Arg90F2Arg BsCM* 39,200 Da 43,617 Da

The trimeric quaternary structure of the mutant was additionally confirmed by sedimentation equilibrium ultracentrifugation, which yielded an average molecular mass of 39,200 Da for the difluoro arginine variant, in good agreement with the expected mass for a trimer (Table 3.3). The sample contained roughly 2% of a contaminant of 23,000 Da, probably caused by some degraded protein.

Table 3.4. Kinetic parameters of all BsCM* variants (50 mM phosphate buffer, pH 7.5, 30 °C). Errors on all parameters were less than 15%.

Enzyme kcat Km kcat/Km kcat/kuncat Ki (s-1) (µM) (M-1s-1) (µM) BsCM* 46 98 4.7×105 4.0 ×106 1.2 5 6 Arg90F2Arg BsCM* 47 190 2.5×10 4.1 ×10 3.3 Arg90Cit BsCM* 0.0026 270 9.6 230 6.8 Arg90homo-Lys BsCM* 0.013 510 25.5 1150 7.6

The mutant enzyme was assayed as previously described (83) at 30 °C in 50 mM -1 phosphate buffer at pH 7.5 (Figure 3.9). Its kcat value of 47 s was indistinguishable

3 CHARGED MUTANTS 56 from that of BsCM* and in good agreement with the published kinetic parameters for wild-type BsCM (83, 210). The Km value of 190 µM was increased by two-fold compared to that of BsCM*. A similar increase was found for the Ki value, which was obtained by standard inhibition assays with 1 at 100 µM chorismate (38, 83, 209) (Figure 3.10).

Arg90F Arg BsCM* BsCM* 2 30 40

35 25 30 ) 20 )

1 25 -1 - (s 15 20 k = (46 ± 1) s-1 /[E] (s /[E] /[E]

-1 0 cat 0 k 15 v

v = (47 ± 6) s 10 cat K = (98 ± 4) µM K = (190 ± 40) µM 10 m m 5 -1 -1 5 -1 -1 k /K = 4.7 x 10 M s 5 k /K = 2.5 x 10 M s 5 cat m cat m

0 0 0 50 100 150 200 250 0100200 300 400 500 [Chorismate] (µM) [Chorismate] (µM) Figure 3.9. Kinetic assays of Arg90F2Arg BsCM* and BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C).

Arg90F Arg BsCM* BsCM* 2 0.3 0.2

0.25 0.15 0.2 (s)

)

0.15 0.1 /[E] /[E] ) (s) /[E] ) 0 0 (v / 0.1 1

y = 0.0688 + 0.0125 x 1/(v y = 0.0418 + 0.0222 x R = 0.99238 0.05 R = 0.9835 0.05 K = (3.3 ± 0.5) µM i K = (1.2 ± 0.1) µM i 0 0 -50 5101520 -1 0 1 2 3456 [I] (µM) [I] (µM) Figure 3.10. Inhibition assays of Arg90F2Arg BsCM* and BsCM* (50 mM phosphate buffer, pH 7.5, 30 °C).

The pH dependence of the reaction catalyzed by the Arg90F2Arg BsCM* variant was also examined. Apparent kcat/Km values were estimated from measurements at low chorismate concentration (20 µM) over the entire pH range 3.5 to 10 and compared with those of the wild-type enzyme (Figure 3.11). In addition, a full Michaelis-Menten analysis was performed at selected pH values in the range 6.7 to 8.9. Unfortunately, the high Km values observed at high pH precluded measurement of accurate kcat values above ca. pH 9.

3 CHARGED MUTANTS 57

pH dependence of BsCM and Arg90F Arg BsCM* 2 3.5 105

3 105

2.5 105

2 105 m K / cat k 1.5 105

1 105

5 104

0 3 4 5678910 pH

Figure 3.11. pH dependence of kcat/Km of recombinant BsCM and semisynthetic Arg90F2Arg BsCM* at 20 µM chorismate (pH 3.6, 4.2, 4.6, and 5.3 sodium acetate (50 mM); pH 5.8, 6.3, 6.7, 7.05, and 7.5 potassium phosphate (50 mM); pH 8.0 and 8.5 Tris-HCl (50 mM); pH 8.25, 8.75, 8.9, 9.25, 9.5, 9.75, and 10.0 glycine-NaOH (50 mM); 30 °C).

The pH dependence of kcat/Km was fitted to Equation 4.1. The apparent pKa values obtained for Arg90F2Arg BsCM* and BsCM are listed for two ionizing groups in Table 3.5. The values for BsCM agree well with those published previously (218).

The first pKa for Arg90F2Arg BsCM* is similar to that of the wild-type, but the second is shifted by almost 0.6 pKa units from 9.22 to 8.66. Moreover, the limiting activity for the variant is reduced by a factor of 3. ⎛k ⎞ ⎜ cat ⎟ k ⎝ K m ⎠ cat = max (4.1) pH pKa1 K m 1+10 +10 10 pKa 2 10 pH

Table 3.5. Apparent pKa values of Arg90F2Arg BsCM* and BsCM. -1 -1 Enzyme pKa1 pKa2 (kcat/Km)max (M s ) BsCM 4.39 ± 0.09 9.22 ± 0.06 3.2×105 ± 8000 5 Arg90F2Arg BsCM* 4.21 ± 0.08 8.66 ± 0.06 1.8×10 ± 5000

3 CHARGED MUTANTS 58

Table 3.6. Km and kcat values of Arg90F2Arg BsCM* at different pHs. (30 °C, pH 6.7 and 7.5 potassium phosphate (50 mM); pH 8.0 and 8.5 Tris-HCl (50 mM); pH 8.9 glycine-NaOH (50 mM). -1 -1 -1 pH kcat (s ) Km (µM) (kcat/Km) (M s ) 6.7 55 ± 3 280 ± 25 2.0×105 ± 25000 7.5 47 ± 6 190 ± 40 2.5×105 ± 30000 8.0 47 ± 3 307 ± 35 1.5×105 ± 25000 8.5 59 ± 4 500 ± 70 1.2×105 ± 25000 8.9 59 ± 4 660 ± 85 0.9×105 ± 25000

The kcat value for Arg90F2Arg BsCM* increases slightly above pH 7.5, whereas the Km value increases from around 200 µM to almost 700 µM at the high end of the pH range. The change in kcat/Km therefore seems to be mainly a Km effect. This slight increase in kcat and large increase in Km is similar to the published pH dependency observed for EcCM (76).

3.3 Discussion

At pH 7.5 the Arg90F2Arg BsCM* mutant is almost as active for the chorismate mutase rearrangement as the wild-type enzyme, affording the same kcat value and an only slightly increased Km value (Table 3.4). The Arg90homo-Lys BsCM* is also active, albeit substantially less so than the wild-type enzyme. Its activity is similar to the citrulline variant, with a five times higher kcat and an almost doubled Km value so that kcat/Km is only two and a half times larger (Table 3.2 and Table 2.2). The small changes in Km for the two mutants suggest only small perturbations of the ground-state Michaelis complex as has already been discussed for the citrulline mutant in Chapter 2. For homo-lysine, which is not isosteric to arginine, the binding effects are even a bit more pronounced than for citrulline. The somewhat different steric requirements of difluoro arginine and homo-lysine compared to arginine probably cause the active site to be a less perfect match for the substrate than the wild-type active site, which has been shaped by evolutionary forces for millions of years. In the case of the Arg90F2Arg

BsCM* mutant the increase in Km is probably caused by a slightly imperfect positioning

3 CHARGED MUTANTS 59 of the guanidinium group because of the increased steric requirement of F compared to H and because of the polarity change that might impair hydrophobic interactions that existed in the wild-type protein between the aliphatic portion of the side chain of Arg90 and its vicinity.

Whereas the results for Arg90F2Arg BsCM* at neutral pH are as expected, the low activity of Arg90homo-Lys BsCM* comes as a surprise. The homo-lysine side chain has a positive charge and should be able to reach the ether oxygen of chorismate in the active site. However, its kcat/Km value is indistinguishable from that of Arg90Lys -1 -1 BsCM [kcat/Km = 31 M s (50)]. This suggests that either the decreased steric requirement or the missing second hydrogen bond donor compared to arginine causes the side chain of homo-lysine to be improperly oriented for efficient catalysis. As has been outlined in Chapter 1, Arg90 is involved in extensive hydrogen bonding interactions with the ether oxygen and the enolpyruvyl carboxylate of chorismate; it also forms a salt bridge to the carboxylate of Glu78. It thus provides direct contributions to ligand binding and catalysis, and indirect contributions to ligand binding, i.e. by properly orienting the side chain of Glu78 for forming a hydrogen bond to the hydroxy group of chorismate. These interactions likely require all three NH groups of the guanidinium moiety. Substitution of arginine with either lysine or homo-lysine, both of which have only one ammonium group, consequently disrupts this H-bonding network.

It is interesting to note that in the case of the lysine mutant the Km value was so high that a separate determination of kcat and Km was impossible (50). By contrast the low catalytic activity of the homo-lysine mutant is mainly due to a low kcat. This almost certainly means that there are different reasons for the low activity of the two variants. One might speculate that lysine, which is too short to contact the ether oxygen, binds only to the carboxylate of Glu78, which probably causes a considerable shift in the position of its carboxylate group. This affects substrate binding in two ways. One is the loss of the hydrogen bond to the ether oxygen, the other the loss of the interaction with Glu78, which is incorrectly positioned for binding the hydroxy group of chorismate. Homo-Lysine however might be long enough to make an interaction with the enolpyruvyl carboxylate, but because it lacks the other interactions it is not properly oriented to form a hydrogen bond to the ether oxygen, causing a low kcat. Additionally, the hydrophobic aliphatic side chain of homo-lysine might slightly disorient the carboxylate of Glu78, which could explain the modest increase in Km.

3 CHARGED MUTANTS 60

The pH profile of the Arg90F2Arg BsCM* mutant was found to be very similar to that of the wild-type enzyme with the important exception of a ca. 0.6 pKa units shift in the basic limb of the pH rate profile to lower values. Though less than the expected difference of about two pKa units, the replacement of arginine at position 90 with 4,4- difluoro arginine significantly affects the apparent pKa value of the active site, making it the prime suspect for being the ionizing group with a pKa of 9.22. Because of the small magnitude of the pKa change and the fact that it is primarily an increase in Km and not a decrease in kcat, we cannot exclude the possibility that deprotonation of cysteine at position 75 is the reason for the high apparent pKa, and that the observed changes are only caused by perturbations of the active side due to the substitution of arginine 90. This would be in agreement with mutagenesis studies of Cys75 which also show an increase in Km coupled with an increase in kcat (50). It should be noted that conclusions from the variation of kcat/Km with pH are only possible under the assumption that there is only one ionization state of the active site that is capable of catalyzing the reaction and all prototropic equilibria of the ionized groups are fast with respect to all steps of the reaction being catalyzed (280-282). Since the chorismate mutase reaction is rather slow compared to protonation reactions and only the protonated state of the arginine at position 90 is expected to catalyze the reaction by stabilizing the partially negative charged ether oxygen of the transition state, as shown by the Arg90Cit BsCM* mutant, those assumptions should be valid in this case.

The dependency of kcat, Km, and kcat/Km on pH in the way that has been observed for Arg90F2Arg BsCM* (Table 3.6) is an indication that the deprotonation event that causes the increase in Km at higher pH happens before the substrate binds to the enzyme. The experimentally determined pKa value is therefore that of the free enzyme.

From the slightly increasing kcat value with increasing pH can be deduced that the apparent pKa value of the active site complexed with the substrate is higher than for the free enzyme. This is a reasonable assumption considering the two additional negative charges of the carboxylates of chorismate. This implies that protonated active sites can bind the substrate better and once the substrate is bound, the enzyme can accelerate the rearrangement to the same kcat as at neutral pH. The increase in Km as a result of the deprotonation of arginine or 4,4-difluoro arginine at position 90 is supported by the fact that the transition state analogue inhibitor 1 binds 250 times stronger than the variant of

3 CHARGED MUTANTS 61

1 where the ether oxygen is replaced by a methylene group. This does not allow for a hydrogen bond with the side chain of arginine 90 any more, while in the case of

Arg90F2Arg BsCM* the deprotonated guanidinium group can still form a hydrogen bond, but weaker than in the protonated form. Therefore, the rise in Km is less substantial than the Ki differences for the two inhibitors. Assuming that the high pKa can be assigned to Arg90 this is an example of how dramatic pKa values can change inside the active pocket of an enzyme and it is further support for the importance of electrostatic stabilisation of the polarized transition state.

4 Investigation of Ligand Binding and Protein Dynamics in BsCM by TROSY-NMR

4.1 Introduction

As we have already seen in the previous chapters, the monofunctional chorismate mutase from Bacillus subtilis (BsCM) is a well-studied enzyme that catalyzes the rearrangement of chorismate to prephenate (Scheme 4.1). How structural changes that accompany ligand binding might modulate catalytic efficiency is less clear.

– CO2 – O2C – O O CO2 – CO2

OH OH (-)-Chorismate Prephenate (-)-Chorismate 2 Prephenate 3

H – O2C – O2C

– O – O CO2 CO2 ≈

OH OH Transition state 1 Transition state 4 TSA 1 Scheme 4.1. Reaction of chorismate to prephenate with putative transition state and the transition state analogue (TSA) inhibitor.

BsCM has been characterized crystallographically in the absence of ligands and as a complex with product and an endo-oxabicyclic transition state analogue [compound 1 (38)](32, 44, 51). Although the overall structure of BsCM is similar in the liganded and unliganded states, ligand-induced conformational changes have been detected crystallographically (32) and by Fourier transform infrared (FTIR) spectroscopy (80). 4 LIGAND BINDING AND PROTEIN DYNAMICS 63

For instance, the original X-ray studies showed that much of the C-terminus, which is immediately adjacent to the active site but disordered in the absence of ligand, becomes substantially more ordered upon formation of the complex (32). The inherent flexibility of the tail segment in the free enzyme is also evident in a high-resolution structure obtained at high ionic strength and low pH (51), where it adopts a different conformation in each subunit of the trimer despite its involvement in crystal packing interactions. The relatively slow binding of substrate and dissociation of product, which partially limit the catalytic reaction (210, 212), have been attributed to these structural changes (32, 80). Selection experiments with randomly truncated BsCM variants have provided direct biochemical evidence for the contribution of the C-terminal tail to catalytic efficiency (81). Thus, the last five residues of the protein can be removed without significantly altering the kinetic parameters, but further truncations lead to substantial decreases in kcat/Km, due largely to increases in Km (81, 283). When 12 or more amino acids are trimmed from the C-terminus, inactive enzymes are obtained. The

C-terminal 310 helix, formed by residues 111-115, is apparently required for a functional catalyst, even though the sequence of this segment is relatively tolerant to mutation (81). Additional residues beyond Leu115 presumably enhance ligand affinity by providing additional stabilizing contacts with the substrate and transition state. To gain more insight into structural changes in BsCM, we have initiated NMR studies of the enzyme in the presence and absence of prephenate and compound 1. This work was done in cooperation with Alexander Eletsky a PhD student in the group of Prof. K. Pervushin from the Laboratory of Physical Chemistry at ETH Zurich, who generated and analyzed the NMR data. In contrast to X-ray methods, which provide a rather static picture of the protein, NMR analysis can provide valuable information about the dynamic properties of the enzyme under native conditions (284, 285). Although trimeric BsCM is rather large by NMR standards (44 kDa), we have successfully assigned nearly all of the backbone 1HN, 1Hα, 15N, 13C’, and 13Cα resonances and most of the side-chain 13C resonances in the unliganded protein, and have also identified diagnostic ligand-induced perturbations of the 15N and 1H spins. In conjunction with 15N relaxation measurements, which shed light on nanosecond and millisecond intramolecular motions of the backbone amide moieties, these data afford a comprehensive view of how the enzyme adapts dynamically to ligands in solution.

4 LIGAND BINDING AND PROTEIN DYNAMICS 64

4.2 Results

Assignment of backbone resonances

Because of its large size, the BsCM homotrimer was characterized by TROSY NMR (286). The protein is well behaved in solution at 20 °C, affording well-resolved spectra under all conditions that were examined. Data analysis was greatly facilitated by the symmetric nature of the structure, which resulted in degenerate resonances for all three subunits. At first, standard procedures were employed for the sequence-specific assignment of backbone residues in the unliganded enzyme. 2D [15N, 1H]-TROSY, 3D [13C]-constant time-[15N, 1H]-TROSY-HNCA and [15N, 1H]-TROSY-HNCACB spectra were acquired with a uniformly 2H-,13C-, and 15N-labeled BsCM sample. Although 120 backbone spin systems are expected, only 91 were identified, 45 of which could be linked into 10 strip fragments and unambiguously assigned with the help of Mapper (287). Incomplete exchange of amide deuterons with protons when the protein was dissolved in H2O might explain why more peaks were not observed with uniformly deuterated BsCM. Similar results have been obtained with other large proteins expressed in D2O (288, 289), as a consequence of amide proton exchange rates for some residues on the order of months. To overcome this problem, a partially deuterated sample [2H(<35%)-,13C-, and 15N-labeled BsCM] was prepared. NMR experiments involving constant time evolution in the 13C dimension are unsuitable in this case because relaxation rates are too high at nonexchangeable sites due to the low deuterium content. We therefore employed a previously described strategy for assigning backbone resonances that combines sensitive 3D TROSY-HNCA (290) and TROSY-HNCO experiments with a 3D multiple-quantum HACACO experiment to detect 13C’ anti-phase coherence (291) and a 3D TROSY-HN(CA)HA experiment (292). The strong signals observed for the amides of C-terminal residues 118-127 and the moderate signals for residues Met2, Met77, and Asn124 in additional 3D TROSY-CBCA(CO)NH measurements proved to be useful for the assignment of residues Met77 and Asn124. Assignments were

4 LIGAND BINDING AND PROTEIN DYNAMICS 65 subsequently verified with 15N-resolved TOCSY and NOESY spectra. Tentative NOE connectivities were checked for consistency with triple-resonance experiments. The availability of X-ray structural data for a large portion of the protein (residues 1-115) (32, 44, 51) facilitated this approach. The amide group of Arg7, which exhibits an unusual 1H chemical shift of 11.9 ppm (293), was identified in [15N, 1H]-TROSY and 15N-resolved NOESY spectra and confirmed by 2D [13C, 1H] versions of the same TROSY-HNCA and TROSY-HNCACB experiments with an increased proton spectral width.

5 12 11 89 12 79 56 93 84 128 110

35 7 130 11 76 16 63 114 110 50 59 102 125 10 34 62 120 115 122 14 113 29 26 18 88 64 40 23 54 48 74 52 37 28 103 118 20 61 70 116 45 91 112 22 3 ω (15N) 19 60 87 4 15 24 36 1 21 77 31 41 53 [ppm] 69 75 49 115 13 101 124 32 33 27 66 95 111 46 68 30 42 109 65 80 97 25 57 107 55 96 126 47 6 104 90 9 119 71 17 123 121 2 125 81 94 85 8 92 86 106 98 73 82 99 43 105 108 78 127 38

68ε 44 130 10ω 1 6 2( H) [ppm]

Figure 4.1. Superposition of 1H-15N TROSY spectra for BsCM (black positive contours and blue negative) and the BsCM•1 complex (red positive contours and magenta negative). The residue-specific assigned peaks are labeled according to the respective residue number in the protein sequence. The cross-peaks stemming from the side chain groups are unlabeled.

Together, this suite of experiments allowed assignment of the resonances of all backbone amides of BsCM, except those of Gly67 and Gly83, which were not found in the [15N, 1H]-TROSY spectra. In the X-ray structure of the enzyme (32, 44, 51) these residues are situated in loop regions, so their resonance lines might be broadened by conformational exchange. Nearly complete assignment of the aliphatic side-chain 13C and 1H resonances using carbon-detected TOCSY experiments was recently reported (294).

4 LIGAND BINDING AND PROTEIN DYNAMICS 66

2 a) b) 1.0 0 0.5

-2 0.0

-0.5 -4 -1.0 -6 2 c) d) 1.0 0 0.5

∆δN ∆δH [ppm] [ppm] -0.5 -4 -1.0 -6 4 e) f) 1.0 2 0.5

0 0.0

-0.5 -2 -1.0 -4 20 40 60 80 100 2040 60 80 100 120 residue Figure 4.2. Comparison of backbone amide 1H (a, c, and e) and 15N (b, d, and f) chemical shifts for BsCM and its complexes with compound 1 and prephenate. (a and b), Chemical shift differences between the BsCM•1 complex and unliganded BsCM. (c and d), Chemical shift difference between the BsCM•prephenate complex and unliganded BsCM. (e and f), Chemical shift differences between the BsCM•1 and BsCM•prephenate complexes.

4 LIGAND BINDING AND PROTEIN DYNAMICS 67

Figure 4.3. Stereoviews of BsCM with chemical shift differences mapped onto the 1.3 Å X-ray structure obtained at low pH and high ionic strength (51) (PDB entry 1DBF) as a model. The individual subunits are colored gray, green, and blue. The red hue is 2 2 1/2 proportional to the quadratic mean [(∆σN) + (10∆σH) ] . (a) Comparison of the BsCM•1 and BsCM•prephenate complexes. (b) Comparison of free BsCM and the BsCM•1 complex.

The backbone amides of the BsCM complexes with prephenate and inhibitor 1 were assigned on the basis of NOE connectivities in the 15N-resolved NOESY spectra. Assignments from unliganded BsCM were taken as a starting point and, assuming a basic similarity of backbone conformations, the corresponding spin systems for the liganded forms of BsCM were identified. This approach allowed all backbone amide 1H and 15N resonances except those associated with Gly67 and Gly83 to be assigned. TROSY-HNCA spectra of the 15N-,13C-, and 2H(50%)-labeled BsCM•1 complex confirmed the assignments. The assigned [15N, 1H]-TROSY spectra of free BsCM and the BsCM•1 complex are superimposed in Figure 4.1. The 1H and 15N chemical shifts for the free enzyme and the prephenate and transition state analogue complexes are compared in pairwise fashion in Figure 4.2.

4 LIGAND BINDING AND PROTEIN DYNAMICS 68

Ligand binding

Prephenate binding was monitored by TROSY spectroscopy using uniformly 2 15 H-, and N-labeled BsCM. When roughly 50% of the binding sites are occupied [KD = 70 µM (210)], an intermediate exchange regime applies. As a consequence, most of the peaks with a line separation of < 150 Hz in at least one dimension were broadened beyond the detection limit. Increasing the concentration of prephenate saturates the enzyme and effectively eliminates this exchange broadening to give well-resolved spectra. These results are in qualitative agreement with the reported dissociation -1 constant koff of 270 s at 25 °C (295).

In contrast, the free and ligand-bound states of the BsCM•1 complex [KD = 1.2 µM (217)] are found to be in slow exchange on the chemical shift time scale at 50% occupancy. Two cross-peaks corresponding to these two states are observed without significant broadening in TROSY spectra if their separation exceeds 30 Hz in at least one dimension. These cross-peaks collapse when the separation is less than 30 Hz. -1 These observations suggest that koff for the inhibitor is on the order of 10 s .

15 N T1, T2, and HNOE Data

The 15N relaxation rates measured at 600 and 900 MHz (Figure 4.4) were determined by nonlinear least-squares fitting of the magnetization decay curves to a single-exponential function. The maximum errors in T1 and T2 were obtained by a Monte Carlo-type procedure, which involved fitting of the intensity decay multiple times with random addition or subtraction of 5% of the measured peak volumes. The estimated averaged errors are 4 and 6% for the T1 data and 8 and 10% for the corresponding T2 data at both magnetic fields for BsCM and the BsCM•1 complex, respectively, and thus amount to approximately 2.5 times the corresponding average standard errors. The estimated Monte Carlo errors for the 15N{1H} NOEs are ca. 10%.

4 LIGAND BINDING AND PROTEIN DYNAMICS 69

3 T1 [s] 2

free BsCM 600 MHz free BsCM 900 MHz 0.15 BsCM-TSA 600 MHz BsCM-TSA 900 MHz T 2 0.10 [s]

0.05

1.0

0.5 HNOE 0.0

-0.5

-1.0

010 20 30 40 50residue 80 90 100 110 120

15 Figure 4.4. Experimental N T1, T2 and HNOE data. Values of T1, T2, and HNOE at 1 15 600 MHz (for H) polarizing field strength are represented by circles. N T1 and T2 data at 900 MHz are represented by triangles. Filled and empty symbols stand for BsCM and the complex with transition state analogue 1, respectively. Due to low cross- peak intensities and spectral overlap, relaxation data are not available for Ile32, Leu47, Leu85, Glu110, and Asn124 in the unliganded enzyme, and for Cys75, Asn124, Thr122 and Thr125 in the BsCM•1 complex. Residues Gly67 and Gly83, which were not found in the TROSY spectra, were not assigned.

Anisotropy and global rotational correlation times were evaluated with the symmetric top model (296). We used the R2R1_Diffusion program to determine whether diffusion anisotropy contributes significantly to the relaxation rates, following the approach of Tjandra et al. (297). T1/T2 ratios determined at two different fields for each protein form were evaluated independently. The principal moments of inertia tensors were calculated from protein atoms of representative trimers in the X-ray structures of unliganded BsCM and the BsCM•1 complex (PDB entries 2CHS and 2CHT) with PDBinertia (298). These moments are (1.00, 0.91, 0.91) and (1.00, 0.91, 0.89) for free BsCM and the BsCM•1 complex, respectively, indicative of a slight axial

4 LIGAND BINDING AND PROTEIN DYNAMICS 70 anisotropy. The same BsCM trimers aligned with their tensors of inertia were used to obtain directional cosines of the amide bond vectors. Only data from residues with

HNOE values greater than 0.65, and T2 values within one standard deviation of the average, were considered. In the symmetric top approximation, the axis of symmetry is assumed to be the longest principal axis of inertia, which has the largest diffusion coefficient. The resulting global diffusion parameters are listed in Table 4.1.

Table 4.1. Anisotropic rotational diffusion parameters. BsCM BsCM•1 600 MHz 900 MHz 600 MHz 900 MHz

D||/D⊥ 1.16 ± 0.04 1.27 ± 0.05 0.91 ± 0.06 1.07 ± 0.07 a τm (ns) 31.2 ± 0.2 27.1 ± 0.1 31.6 ± 0.2 27.2 ± 0.1 θ (deg) 1.5 ± 0.1 1.1 ± 0.1 1.3 ± 0.2 0.0 ± 0.3 ϕ (deg) 1.3 ± 0.1 1.4 ± 0.1 6.0 ± 0.2 4.3 ± 7.6 F-value 7.06 5.52 2.08 1.03 P(F) 0.0003 0.0017 0.11 0.38 a The error associated with the τm values was calculated using the statistical root-mean- square deviation of T2/T1 ratios for individual amino acids.

Calculated probabilities demonstrate that the anisotropic motional model (symmetric top) is statistically significant for BsCM. Furthermore, the orientation of the experimentally determined diffusion tensor matches that of the inertia tensor with a precision of a couple degrees. Although anisotropic diffusion is statistically less significant for the inhibitor complex, probably because of the somewhat lower precision of the measured relaxation rates or to global structural changes in the enzyme upon complexation, the relaxation data from both forms of the protein were analyzed using the anisotropic symmetric top model. This approach is justified by the observation that slight anisotropy has very little influence on the motional parameters of individual residues (299, 300).

4 LIGAND BINDING AND PROTEIN DYNAMICS 71

1.8

1.6 free BsCM 1.4 BsCM-TSA

lg <χ2>

1.0

0.8

0.6

25 3035τ [ns] 40 45 50 m

Figure 4.5. Grid search on τM. Dependence of the target function on the global correlation time τM when the extended model for internal dynamics is employed.

The average global correlation times determined from the T1/T2 ratios, which are the same for both protein forms, exhibit a marked field dependence (Table 4.1). This 2 apparent field dependence indicates that the standard {S , te} model with the approximation of fast internal dynamics (te << τM) does not hold. Consequently, simultaneous fitting of global correlation times and model-free parameters of internal dynamics are required (301). Figure 4.5 shows values of the global target function χ2 2 2 calculated for a range of correlation times τM using the extended {Ss , Sf , ts} model of the Model-Free approach applied only to the rigid part of the protein. In the absence of the Rex term and assuming isotropic tumbling, there are 3n + 1 unknowns for n sites and

5n observables (T1 and T2 at two magnetic fields and HNOE at one magnetic field). Because the HNOE values for this large protein measured at 600 MHz are largely independent of τM, the number of informative observables is reduced to 4n, which exceeds the number of unknowns and ensures adequate determination of the model. It 2 was previously demonstrated that above a certain value of τM, χ levels off and a further 2 increase in τM is offset by a decrease in Ss (302). Consequently, although most BsCM residues are best described by the extended model (see below), the global correlation 2 time cannot be determined in a manner that is independent of Ss . Using the criterion 2 that χ levels off as a function of τM, the global correlation time τM for the model-free analysis was estimated to be 37 ns. Model-free analysis of the relaxation data is potentially complicated by uncertainty regarding the contribution of chemical shift anisotropy (CSA) to the

4 LIGAND BINDING AND PROTEIN DYNAMICS 72

15 observed T1 and T2 values (303). The constant CSA contribution to N relaxation is field dependent and can be large. For example, for an N-H vector tumbling with an 2 2 isotropic correlation time of 37 ns, Sf = 0.85, Ss = 0.85, ts = 1 ns, and an H-N bond length of 0.102 nm, an 15N CSA value of -160 ppm, would contribute approximately 15 22% and 40% to the observed N T1 at 600 and 900 MHz, respectively. Thus, the CSA relaxation mechanism must be considered explicitly for amide 15N signals at fields higher than ca. 600 MHz. The variability of 15N CSA at different sites of the protein, which can be as large as ±15 ppm (304), may also influence the statistical significance of the selected isotropic/anisotropic model. However, the absolute value of the order parameters is less important for the current investigation than the relative changes induced upon ligand binding. These differences are significantly less susceptible to variation of many model-free parameters, including 15N CSA.

Model-Free Analysis of Intramolecular Backbone Mobility

The entire data set was analyzed using the symmetrical top model with the 2 longest τM equal to 37 ns and D||/D⊥ equal to 1.2. The applicability of the standard {S , 2 2 te} and extended {Ss , Sf , ts} models of the Model-Free approach was assessed on a residue-per-residue basis employing the F-test criterion (305). Calculations demonstrate that the relaxation rates of 30 residues in BsCM and 35 residues in the BsCM•1 complex are well described by the standard model, but the extended model is required for the rest of the protein based on the 10% probability cutoff level of the F-test. In general, selection of an overly simple model of internal motions leads to overestimation of the generalized order parameters and underestimation of internal correlation times (306). Because of the relatively small number of residues described by the standard model, and their scattered distribution over the protein sequence, we consequently applied the extended model uniformly to the entire enzyme. This procedure simplifies cross-comparison of changes in intramolecular dynamics by providing the same dynamical model for the residues in both structural forms of the protein.

4 LIGAND BINDING AND PROTEIN DYNAMICS 73

1.0

0.8 1 2 Ss

0.4 0 115 120 125 0.2

Rex [s-1]

0 010 20304050 8090100110 120 residue 2 Figure 4.6. Slow motion order parameters Ss and exchange terms obtained in the Model-Free analysis using the extended model for free BsCM (filled circles) and the BsCM•1 complex (empty circles). Gray stripes denote active site residues according to reference (32). B-Factors of backbone nitrogen atoms in the X-ray structures of free BsCM (2CHS) and the BsCM•1 complex (2CHT) are represented by a solid and a dashed line, respectively. A schematic representation of the secondary structure is shown at the top.

2 2 The F-test criterion was subsequently employed to distinguish between {Ss , Sf , 2 2 ts} and {Ss , Sf , ts, Rex} models. At a probability cutoff of 10% the number of residues requiring the Rex term was 30 in BsCM and 18 in the BsCM•1 complex. Figure 4.6 2 shows that there is a correlation between Ss and the secondary structure. The regions of 2 low Ss values around Arg14, Thr50, and Ser66 in both liganded and unliganded forms correspond to loops in the structure. Other loops do not show significant conformational mobility on the nanosecond time scale. Correlation between crystallographic B-factors 2 and Ss is less pronounced (Figure 4.6). There are several regions characterized by high 2 B-factor values, but only those around Arg14 and Leu66 also have low Ss values. The 2 order parameter for motions on the picosecond time scale, Sf , remains high throughout the sequence, declining slightly toward the C-terminus. The correlation time for internal motions, ts, falls generally in the range of 1-3 ns, as observed for other proteins, and decreases slightly at the C-terminus (data not shown).

4 LIGAND BINDING AND PROTEIN DYNAMICS 74

4.3 Discussion

Nearly all the backbone 1HN, 1Hα, 15N, 13C’, and 13Cα resonances and most of the side chain 13C resonances in the unliganded BsCM homotrimer have been successfully assigned. Secondary structure elements in the protein were identified by chemical shift analysis of the 1Hα, 13CO, 13C’, and 13Cα spins using CSI (307), and confirmed with data from 15N-resolved NOESY-TROSY spectra. Generally good agreement was observed between the NMR and X-ray data regarding the locations of the five β-strands and three helices in each subunit. Differences include a somewhat shorter β-strand II than is observed crystallographically, which may be attributed to the fact that Val42 and Val43 form a β-bulge instead of the standard β-sheet seen in the X-ray structures (32, 44). In addition, the chemical shift index indicates that the second helix is shorter by one residue than in the X-ray structures (residues 58-63 versus 58-64), and predicts a coiled conformation for residues 111 to 115 rather than a 310-helix. Of all the assigned backbone amides, the 1HN resonances of Arg7 and Met79 show unusually large deviations from random coil values (~3-4 ppm) (Figure 4.1). Calculations with MOLMOL (308) based on the X-ray structures of BsCM indicate that ring current effects cannot account for these anomalies. Instead, the large Arg7 chemical shift has been attributed to a strong hydrogen bond with the Nδ1 atom of His106 (293). This same hydrogen bond is also observed in the structure of a related chorismate mutase from Thermus thermophilus (PDB entry 1ODE) (36), which shares only 47% sequence identity with BsCM, suggesting that it is a conserved feature in this enzyme class. The backbone amide proton of Met79 is hydrogen bonded to the carboxylate group of Glu78 (32, 44). The latter is a key active site residue, and this interaction may be important for positioning the carboxylate for effective catalysis. Regions of the protein influenced by ligand binding are readily apparent when the chemical shifts obtained for the unliganded trimer are compared with those for the product and transition state analogue complexes (Figure 4.2). It is known from X-ray studies that prephenate and compound 1 make similar contacts with the enzyme active site (32, 44), so it is not surprising that their respective complexes exhibit only small chemical shift differences (compare panels e and f of Figure 4.2). Those differences that

4 LIGAND BINDING AND PROTEIN DYNAMICS 75 are observed are distributed nonrandomly in the 3D structure, and are generally localized at the active site (Figure 4.3a), reflecting the different structures of the ligands and possibly subtle dissimilarities in their binding modes. Much larger chemical shift differences are seen when either of these complexes is compared with the unliganded enzyme (compare panels a and b or c and d of Figure 4.2). The largest effects are associated with three regions of the protein: (i) residues 6-9 in β-strand I, (ii) a long stretch extending from residues 57 to 80, covering the second helix of the enzyme and β-strand III, and (iii) C-terminal residues 106 to 121 (Figures 4.2 and 4.3). When these differences are mapped onto a 3D model of the enzyme (Figure 4.3b), it is evident that the active sites, the subunit interfaces, and the C-terminal tails experience the largest ligand-induced perturbations. As might be expected, many active site residues, including Arg7, Phe57, Glu78, and Arg90, are sensitive to ligand binding. The moderate chemical shift changes that are observed presumably reflect small adjustments in residue positioning or the altered molecular environment of the active site in the presence of ligand. By comparison, Arg63 and Arg116 exhibit considerably larger effects, consistent with crystallographic data showing that their relatively flexible side chains are recruited to the active site to bind the tertiary carboxylate of prephenate or the transition state analogue (32, 44). Although hysteresis has not been observed in the binding or kinetic properties of the enzyme, minor adjustments at the trimer interfaces are likewise unsurprising. These perturbations indicate global conformational rearrangements of the trimer upon complexation. Interestingly, the amide signals of Val73 and Met79, which are near the active site but do not contact the ligand directly, are among the residues most strongly affected by the presence of a ligand. As noted above, the Met79 amide helps to position the carboxylate side chain of Glu78, which in turn provides a key hydrogen bonding interaction with the alcohol group of both prephenate and 1. The chemical shift change in this case likely reflects the altered hydrogen bonding network that arises when the ligand binds (32, 44). The sensitivity of the Val73 amide may be due to the fact that it is solvated by water in the free enzyme, but proximal to the positively charged guanidinium group of Arg116 in the complex (32).

4 LIGAND BINDING AND PROTEIN DYNAMICS 76

Table 4.2. Kinetic parameters for BsCM and several C-terminally truncated variantsa

-1 -1 -1 Enzyme variant kcat (s ) Km (µM) kcat/Km (M s ) wild-type 46 ± 3 67 ± 5 6.9 × 105 BsCMb

BsCM(1-124)c 33 ± 2 71 ± 10 4.6 × 105

BsCM(1-122)c 84 ± 5 140 ± 15 6.0 × 105

BsCM(1-121)c 70 ± 6 230 ± 40 3.0 × 105

BsCM(1-120)c 44 ± 5 1010 ± 130 4.3 × 104

BsCM(1-119)c 65 ± 1 920 ± 23 7.1 × 104

BsCM(1-116)b 28 ± 7 9600 ± 3000 2.9 × 103 aEnzymes were assayed in 50 mM potassium phosphate buffer (pH 7.5) at 30 °C. bFrom reference (81). cThis work.

The most dramatic ligand-induced perturbations in BsCM are associated with amino acids at the C-terminus, starting at residue 106, immediately before the last β- strand, and extending beyond the final 310-helix to residue 121. This stretch includes two active site residues, Tyr108 and Leu115, which interact with the enol pyruvate side chain of the substrate. Several residues beyond Leu115, especially Arg116, Asp118, Leu119 and Ser120, experience some of the largest chemical shift changes in the entire protein. The latter are the same residues that become more ordered crystallographically in response to a ligand (32). These amino acids apparently serve as a hinge between the

310-helix (residues 111-115) and the essentially unstructured terminus (Figure 4.3). Rigidification of this segment contributes to the organization of the active site and may also facilitate recruitment of the Arg116 side chain for carboxylate recognition. Proteins truncated in this region, while active, exhibit substantial increases in Km (Table 4.2), indicative of a loss of productive binding interactions. In contrast, residues beyond Leu121 are unimportant for binding; negligible chemical shift differences are observed in this region, consistent with the observation that the last six amino acids of the protein can be deleted without appreciably affecting activity (81). Direct information about backbone flexibility over a broad range of time scales 15 can be gleaned from N relaxation data (304, 309). In Figure 4.4, T1, T2 and HNOE data for the free enzyme and its complex with the transition state analogue 1 are

4 LIGAND BINDING AND PROTEIN DYNAMICS 77 compared. Interpretation of the relaxation data using the so-called “Model-Free” approach of Lipari and Szabo (310, 311) provides tangible insight into how ligand binding influences the amplitudes and time scales of backbone motions. This method assesses “fast” internal motions (faster than the overall tumbling, i.e., pico- to nanosecond dynamics) using two parameters at each site that are independent of motional model. These parameters are the generalized order parameter, S2, which measures the spatial restriction of the internal motion, and the effective local correlation time, te, which correlates with the rate of internal or local motion. Coupling the 2 numerical values of S and te with a physically reasonable model allows the complex intramolecular motions of the macromolecule to be described (296). For the current study it suffices to note that values of S2, which describe nanosecond events, vary from 0 to 1, with the two extremes corresponding to fully flexible and fully rigid molecular segments, respectively. The original “Model-Free” approach was subsequently extended by incorporating local motions at two time scales, two associated order parameters, and the overall correlation time τc (312). In addition, it was realized that exchange broadening, Rex, is not fully quenched in the experimental pulse sequences. The latter measures conformational diversity on the millisecond time scale, with larger values corresponding to greater diversity. In total, then, the extended formalism is 2 2 2 parametrized by six variables: τc, ts, tf, Ss , Sf and Rex. Because the tf, and Sf parameters have a rather uniform distribution throughout the BsCM sequence, as has been observed for several other proteins (304, 313), we focus our discussion on ligand-induced 2 changes in Rex and Ss , which reflect changes in the intramolecular dynamics on the millisecond and nanosecond time scales, respectively.

As shown in Figure 4.6, the free enzyme requires more Rex terms to reproduce the experimental relaxation data than does the transition state analogue complex, indicating that ligand binding induces structural consolidation on the millisecond time scale. For example, both forms of the enzyme require this term for residues Arg7, Leu27, Val42, Leu53 and Val71, whereas residues Gly8, Thr16, Glu19, Lys30, His36, Thr37, Asp41, Gln44, Val56, Ala61-Leu65, Met76, Val81, Gln96, Ile104, Val107 and

Val113-Arg116 have significant Rex values only in the free enzyme. Evidently, segments of the unliganded protein roughly corresponding to the second and third helices (i.e., residues 58-64 and 111-115, respectively), in addition to numerous isolated

4 LIGAND BINDING AND PROTEIN DYNAMICS 78 residues distributed over the entire primary sequence, have considerable internal mobility on the millisecond time scale that is substantially dampened upon ligand binding. Extensive dialysis of the samples to remove potential contaminants argues against transient binding of some chemical species at the active site as the origin of the observed exchange broadening in the free enzyme. The alternative possibility that exchange broadening is induced by motions of atomic groups that are spatially proximal but remote in sequence from the segments in question (314, 315) also seems unlikely on structural grounds. However, it is striking that the two flexible helices flank the entrance to the active site, and that they either contain or are immediately adjacent to arginines (Arg63 and Arg116) whose side chains are disordered in the unliganded enzyme but involved in substrate recognition in the complex (32). Although less pronounced, the structural ordering inferred here is reminiscent of the ligand-induced conformational transition seen for an engineered mutase that undergoes a transition from a molten globular state to a well folded protein upon ligand binding (316). In the case of BsCM, the observed structural changes are potentially significant for catalysis given that the rearrangement of chorismate to prephenate also proceeds on a millisecond time scale at the active site (46 s-1). 2 Judging from the Ss values (Figure 4.6), these same general regions also become more ordered in the nanosecond regime (Figure 4.7). In this case, the largest ligand-induced changes are associated with the C-terminal region, which is the least ordered part of the protein (Figure 4.6). The unliganded enzyme is relatively structured up to Arg116, one residue more than observed in the original crystal structure (32). Beyond this position, the order parameters gradually decrease to a value of ca. 0.1, which corresponds to a largely unstructured polypeptide. In contrast, the BsCM•1 complex is reasonably well ordered up to Leu119, again consistent with X-ray data showing additional electron density extending out to this residue in some of the monomers of the complex (32). The following amino acids become increasingly flexible up to position 124, where a similarly unstructured state is reached as in the free enzyme.

4 LIGAND BINDING AND PROTEIN DYNAMICS 79

Figure 4.7. Backbone dynamics mapped onto the 3D structure of BsCM. The unliganded enzyme (PDB entry 1DBF) is shown in panels a and c, and the BsCM•1 complex (PDB entry 2CHT) is shown in panels b and d. For the purposes of illustration, the missing C-terminal residues in the latter structure were artificially grafted to two subunits shown using the unliganded structure as a model. The thickness of the 2 backbone trace is proportional to (1 - Ss ). The chemical shift changes in panels a and b are color coded as in Figure 4.3b. Residues showing exchange broadening are colored orange in panels c and d.

The TROSY data reported here provide a snapshot of the dynamic properties of BsCM, affording insight into ligand-induced changes that may be important for catalysis. Previous attention has focused primarily on the C-terminal tail of the enzyme, which becomes more ordered in the presence of ligands (32, 80), possibly restricting the diffusion of small molecules into and out of the active site. It may also control the conformation and hence the reactivity of bound chorismate. Our NMR analysis confirms the ligand-induced ordering of this protein segment. In addition, the observed chemical shift changes and relaxation data highlight several other regions of the large trimeric enzyme that appear to adapt dynamically to the ligand. The transitions observed between a conformationally flexible active site in the absence of ligand to a more rigid structure in its presence imply an induced fit mechanism for the B. subtilis

4 LIGAND BINDING AND PROTEIN DYNAMICS 80 mutase. Further work will be needed to clarify in detail how the observed protein dynamics modulate ligand affinity and catalytic efficiency. In this context, valuable complementary information about backbone motion might be obtained by investigating the relaxation of spin groups other than the N-H vector, such as the Cα-CO vector (314, 317, 318).

5 Outlook

The work described in the first part of this thesis demonstrates how the catalytic mechanism of an enzyme can be probed by the insertion of unnatural amino acids, employing a semisynthetic strategy for the preparation of the enzyme variants. This approach might also be useful for studying other enzymes. The results of these investigations provide additional evidence for electrostatic transition state stabilization in catalysis by the enzyme chorismate mutase. To further clarify the exact reasons for the low activity of the Arg90Cit BsCM* mutant we need to await the results from the crystal structure determination of the free enzyme and its complex with inhibitor 1. The replacement of Arg90 with the isosteric analogues 32 or 33 could give further insight into the reasons for the low activity of the Arg90homo-lysine BsCM* variant, specifically whether hydrogen bonds from all three NH groups of the guanidinium moiety of arginine to the substrate or to neighboring residues of the enzyme are essential for catalysis.

HO HO HO O O O

H2N H2N H2N

HN HN

NH NH

HN H2N H2N 32 33 34 Figure 5.1. Other amino acids that could potentially be used to replace Arg90 in BsCM.

A clearer picture of the residue responsible for the upper pKa value of BsCM could be obtained by determining the pH dependency of the Cys75Ser mutant, which still has about 25% wild-type activity (50, 55). This might allow a definite distinction between Cys75 and one of the four arginines (Arg7, Arg63, Arg90, and Arg116) in the active site. Another, albeit much more elaborate, way to distinguish between the 5 OUTLOOK 82 arginines and the cysteine would be to replace Arg90 with the known isosteric analogue canavanine 34 whose side chain functional group has a pKa value of 7.0 (236, 319).

This would lower the pKa of the enzyme·substrate complex to a value that is accessible by pH variations in a range where the enzyme is not denatured, allowing possible effects on kcat upon deprotonation of the residue at position 90 to be measured.

The second part of the thesis shows the usefulness of NMR as a complementary technique to x-ray crystallography for studying dynamic properties of an enzyme. Our results indicate that the enzyme BsCM adapts dynamically to ligands by a sort of induced fit mechanism. They especially support the proposal that the C-terminus of the enzyme can serve as a lid for the active site, which is also confirmed by the increase in

Km for the C-terminally truncated BsCM mutants that have been prepared. It would be interesting to determine which part of the C-terminal tail makes contact with which part of chorismate during the rearrangement. Unfortunately, the currently available methods are not yet up to this task.

6 Experimental Section

6.1 General

Abbreviations

* stop codon (TAA) AA any of the 20 natural amino acids AcOH acetic acid BSA bovine serum albumin BsCM Bacillus subtilis chorismate mutase Boc N-tert-butoxycarbonyl bp base pairs

CaH2 calcium hydride CBD chitin binding domain Cbz benzyloxycarbonyl

CH2Cl2 dichloromethane

CH3CN acetonitrile

CHCl3 chloroform CM chorismate mutase DBU diazabicyclo[5.4.0]undec-7-ene DCM dichloromethane DEAE diethylamino ethyl decomp. decomposition DIEA diisopropyl ethyl amine DNA desoxyribonucleic acid DMF N,N-dimethylformamide 6 EXPERIMENTAL SECTION 84

DMP 2,2-dimethoxypropane DMSO dimethylsulfoxide DTT dithiothreitol EA ethyl acetate EDT 1,2-dithioethanol EDTA ethylenediaminetetraacetate eq equivalent ESI electrospray ionization

Et2O diethyl ether EtOH ethanol EtSH ethanethiol FAB fast atom bombardment Fmoc 9-fluorenylmethoxycarbonyl GdmHCl guanidinium hydrochloride h hour(s) HBTU 2-(1-H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate Hex hexane HPLC high-pressure-liquid-chromatography HOAc acetic acid HOBt 1-hydroxybenzotriazole iPrOH isopropanol IPTG isopropyl-1-thio-β-D-galactoside LB Luria-Bertani broth LC-MS liquid chromatrography mass spectrometry LMW low molecular weight marker M mol/l MCS multiple cloning site MeOH methanol Mes 2-morpholino-ethanesulfonic acid MESNA sodium 2-mercaptoethanesulfonate min minute(s)

6 EXPERIMENTAL SECTION 85

mp. melting point MS mass spectrometry MTBE methyl tert-butyl ether MWCO molecular weight cut off Mxe Mycobacterium xenopi n.d. not determined NEB New England BioLabs (Beverly MA, USA) NMMAA N-methyl-mercapto-acetamide NMP N -methyl-2-pyrrolidinone NMR nuclear magnetic resonance

OD600 optical density (absorption) at 600 nm Pbf 2,2,4,6,7-pentamethyl-dihydrobenzofuran-5-sulfonyl PCR polymerase chain reaction PG protecting group Ph phenyl Pip piperidine PMSF phenylmethanesulfonyl fluoride ppm parts per million quant. quantitative

Rt retention time r.t. room temperature

Rf retention factor RP-HPLC reversed phase high performance liquid chromatography rpm rotations per minute sat. saturated Sce or Sc Saccharomyces cerevisiae SDS-PAGE sodiumdodecylsulfate-polyacrylamide gel electrophoresis SPPS solid phase peptide synthesis Su succinimide TBE Tris, boric acid EDTA TBS Tris buffered saline TCEP tris(2-carboxyethyl) phosphine hydrochloride TFA trifluoroacetic acid

6 EXPERIMENTAL SECTION 86

THF tetrahydrofuran TIPS triisopropyl silane TLC thin layer chromatography Tris tris(hydroxymethyl)aminomethane TsOH toluene p-sulfonic acid UPW ultra pure water UV ultraviolet wt wildtype In addition the standard abbreviations for the natural amino acids (1 and 3 letter code), DNA bases, elements and SI units were used.

Analytical Data

Agarose gels were run at three different concentrations depending on the size of the DNA fragment. The agarose used was pure grade from BioRad, CA, USA. Either 0.8%, 2% or 3% gels in TBE buffer pH 8.35 were used. Before pouring the gel 4 µl ethidiumbromide (10 mg/ml) were added. As power supply a Power Pac 300 from BioRad was used. The gel was loaded with either 10 µl (analytical) or 40 µl (preparative) of sample or of the marker λBstEII or φX174HaeIII (total DNA concentration: 33 ng/µl), and run at 90 V for 60 to 120 min. The electrophoresis was done in 1x TBE-buffer. The results were evaluated with a UV-lamp at 254 nm (Transilluminator Lourmat from Vilber, Marne la Valée, France) and a digital camera.

Analytical Ultracentrifugation was performed for sedimentation velocity measurements on a Beckmann XL-I analytical ultracentrifuge at 20 °C and 42,000 rpm. The sedimentation velocity was measured 140 times every 3 min with an interference detector. The sedimentation coefficient was determined to be 3.56 svedberg (3.56 10-13) and the partial specific volume νbar to be 0.745 by the program Hydropro 5.a (320) using the pdb data (2CHS) as input file. The data was evaluated using Sedfit (321). For

6 EXPERIMENTAL SECTION 87 sedimentation equilibrium measurements, a Beckman XL-A analytical ultracentrifuge equipped with an An-60Ti rotor was used. The buffer was 10 mM sodium phosphate and 90 mM NaCl (pH 7.5). Protein concentration was 0.53 mg/ml. Sample volumes were 120 µl with 30 µl FC43 and reference channels containing 160 µl of buffer. Sedimentation equilibrium (SE) was performed using 2-channel charcoal-epon cells with quartz windows. The time to reach equilibrium was estimated using Ultrascan and monitored by subtracting two scans which were taken four hours apart. When virtually no difference could be observed, equilibrium was considered to be reached. Three scans per cell were recorded using 0.001 cm point spacing and averaging 10 readings for each each point after equilibrium was reached. SE was performed at 6 different rotational speeds ranging from 15,000 to 40,000 rpm. The absorbance was measured at 280 nm. Density of the respective buffers wase measured at 20 °C using a DSA48 density and sound analyzer (Anton Paar, Graz, Austria). The value of νbar was calculated using Sednterp (322). Data analysis was performed using Ultrascan (323).

Circular Dichroism spectra were recorded on an Aviv Circular Dichroism Spectrometer 202 or on a Jasco J-715 at 25 °C. CD spectra were measured at a protein concentration of ~ 4 µM in degassed PBS (10 mM phosphate, 160 mM NaCl, pH 7.5), by averaging at least 3 wavelength scans from 260 nm to 200 nm in 0.5 or 1 nm steps. The signal was averaged for 3 s and the bandwidth set to 1 nm. The samples were measured in quartz cuvettes (Hellma) with a path length of either 0.1 or 0.2 cm.

Edman degradation was performed by the Protein Service Laboratory of the Institute for Molecular Biology and Biophysics, ETH Zürich.

Elemental analysis was performed by the Microanalytical Laboratory of the Laboratory of Organic Chemistry, ETH Zürich. For CHN the analysis was done on a LECO CHN-900, for O on a LECO RO-478, for S on a LECO CHNS-932 and for F by using ion chromatography.

6 EXPERIMENTAL SECTION 88

LC-MS was performed on a Spectra System HPLC connected to a diode array detector (UV6000LP, Thermo Separation Products) and an ion-trap mass spectrometer (LCQdeca, Finnigan) with the same columns as for analytical RP-HPLC (see below).

Mass spectra were recorded on a VG-ZAB2-SEQ spectrometer (FAB) in a 3- nitrobenzyl alcohol matrix or on a Finnigan TSQ7000 Triple-Quad mass spectrometer (ESI) or on a HiRes-ESI IonSpec Ultima FTMS-spectrometer. ESI spectra for proteins and peptides were measured in a 70:30 H2O:CH3CN mixture or in 0.1% AcOH (aq), those of small organic molecules in MeOH. Calculated masses were based on average isotope composition or on single isotope masses for HiRes spectra. Protein solutions were desalted, if necessary, on a NAP-5 column (Pharmacia) that had been equilibrated with 0.1% HOAc (aq). For small molecules the major signals are given in m/z units. The symbol M is assigned to the molecular ion. Suggested assignments are given in brackets.

Melting points were determined with a Büchi 540 apparatus in open capillaries and are uncorrected.

NMR spectra were recorded on Bruker AV600 (1H 600 MHz, 13C 150.9 MHz) Bruker DRX 500 (1H 500 MHz, 13C 125 MHz), ARX 300 (1H 300 MHz, 13C 75 MHz, 19F 282.36 MHz), or Varian Gemini 300 (1H 300 MHz, 13C 75 MHz, 19F 282.36 MHz), NMR spectrometers. All 13C-NMR spectra are 1H-broadband decoupled and were measured at r.t. if not stated otherwise. The chemical shifts (δ) are given in ppm 1 13 19 downfield from SiMe4 (= 0 ppm) for H and C or CCl3F for F. J values are given in Hz. For multiplets the following abbreviations were used: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet or unresolved signal), br (broad).

NMR Spectroscopy of labeled BsCM. All NMR experiments were performed at 293 K on Bruker Avance 600 and Avance 900 spectrometers. NMR data were processed with PROSA (324) or XWINNMR (Bruker Biospin), and resulting spectra were analyzed with Xeasy (325). Chemical shift referencing was based on DSS (sodium 2,2-dimethyl-2-silapentane-5-sulfonate) as the internal reference for 1H nuclei.

6 EXPERIMENTAL SECTION 89

Indirect referencing of 13C and 15N nuclei based on the gyromagnetic ratios was employed. 3D CT-TROSY-HNCA, TROSY-HNCACB (326, 327), and TROSY-HNCA (290) measurements were used for backbone assignments. Interscan relaxation delays of 1 s were employed. Time domain data in the 13C dimension of the TROSY-HNCA experiment were doubled by mirror-image linear prediction (328) prior to Fourier transformation. The 15N resolved NOESY-TROSY data sets of all samples were expanded to double size in t1 by forward linear prediction before the Fourier transformation. T2 was measured in refocused echo experiments. To suppress cross- correlated relaxation, SEDUCE decoupling was applied to amide proton spins during T1 and T2 relaxation periods. T1 values for BsCM at 600 MHz were estimated by nonlinear exponential fitting of the spectral intensities measured with nine relaxation delays: 0.05,

0.1, 0.2, 0.4, 0.8, 1.2, 1.8, 2.5, and 3.5 s. T2 values for BsCM were estimated at

600 MHz, with ten points: 2, 4, 8, 16, 24, 32, 48, 64, 80, and 96 ms. T1 values for BsCM were estimated at 900 MHz, with nine points: 0.1, 0.2, 0.4, 0.8, 1.2, 1.8, 2.5, 3.5, and

4.5 s. T2 values for BsCM were estimated at 900 MHz, with eight points: 2, 4, 8, 16, 24,

32, 48, and 64 ms. T1 values for BsCM·1 were estimated at 600 MHz, with twelve points: 0.02, 0.05, 0.08, 0.1, 0.2, 0.3, 0.5, 0.8, 1.0, 1.2, 1.5, and 2.0 s. T2 values for the BsCM·1 complex were estimated at 600 MHz, with ten points: 2, 4, 8, 16, 24, 32, 48,

64, 80, and 96 ms. T1 values for the BsCM·1 complex were estimated at 900 MHz, with eight points: 0.2, 0.4, 0.8, 1.2, 1.8, 2.5, 3.5, and 4.5 s. T2 values for the BsCM·1 complex were estimated at 900 MHz, with eight points: 2, 4, 8, 16, 24, 32, 48, and 64 ms. HNOE was measured in separately acquired saturation and reference TROSY- type experiments (329). All experiments utilized an interscan relaxation delay of 7 s. 1 Saturation of proton spins was applied in 4 s using a train of high-power H 180° pulses applied in 30 ms intervals. HNOE was calculated as the ratio of maximum peak intensity. The errors in HNOE were calculated from the spectral noise by error propagation. Relative errors in T1 and T2 below 5% have been set to 5%.

Sample preparation. The proteins were dialyzed into 25 mM potassium phosphate buffer (pH 7.5) using Slide-A-Lyser cassettes, and 3% D2O was added prior to data collection. The following samples were prepared: 500 µl of 4 mM uniformly

6 EXPERIMENTAL SECTION 90

2H-, 13C-, and 15N-labeled BsCM in a standard NMR tube; 500 µl of 2.4 mM 2H(<35%)- , 13C, and 15N-labeled BsCM in a standard NMR tube; 500 µl of 2.0 mM 15N-labeled BsCM in a standard NMR tube; 300 µl of 3.0 mM 15N-labeled BsCM and 4.5 mM inhibitor 1 in a Shigemi NMR tube; 300 µl of 3.0 mM uniformly 15N-labeled BsCM and 1.3 mg of prephenate in a Shigemi NMR tube; and 500 µl of 3 mM 15N-, 13C-, and 2H(50%)-labeled BsCM and 3 mM inhibitor 1 in a standard NMR tube. The BsCM·1 and BsCM·prephenate complexes were prepared by stepwise addition of the chorismate, which was instantaneously converted to prephenate, until the ratio of the complex to the free protein was greater than 100. The attempt to generate a BsCM·prephenate complex using enzyme prepared in KA13 containing the plasmid pKIMP-UAUC, which should allow the production in minimal medium, were unsuccessful because of an undetectable contamination with prephenate dehydratase and only signals corresponding to phenylalanine were found by NMR spectroscopy. Further protein was therefore produced in KA13 without this plasmid using minimal medium supplemented with full medium, or full medium. For complexes with compound 1, 10 µl aliquots of a 30 mM aqueous stock solution were added to generate the inhibitor complex. To obtain the product complex, solid chorismate, which was rapidly converted to prephenate in situ, was added in three portions of ~0.2 mg, ~0.4 mg, and ~0.7 mg. Complex formation was monitored by recording short (~30 min) 2D [15N, 1H]- TROSY spectra.

Relaxation analysis. Relaxation data were analyzed by the Model-Free approach (310, 311) with ModelFree version 4.15 (330, 331), using an H-N bond length of 0.102 nm and 15N CSA of -160 ppm. Assuming overall isotropic motion, the standard model of Lipari and Szabo produces the following spectral density function

2 ⎧⎫S τM ()τ1S– 2 J()ω = --2------+ ------⎨⎬2 2 2 2 (6.1) 5⎩⎭1 + ω τM 1 + ω τ

6 EXPERIMENTAL SECTION 91

2 where 1/τ = 1/τM + 1/τe, τe is the effective correlation time of internal motions, and S is the generalized order parameter, reflecting the degree of spatial restriction of internal motions. An extended model incorporates the higher-order approximation (312):

⎧⎫S2τ ()τS2 – S2 ′ ()τ1S– 2 ′ J()ω = --2------M ++------f s------f f ⎨⎬2 2 2 ′2 2 ′2 (6.2) 5⎩⎭1 + ω τM 1 + ω τs 1 + ω τf

where 1/τs, = 1/τM + 1/τs and 1/τf, = 1/τM + 1/τf. This model makes sense only when the internal motions occur on significantly different time scales, “slow” and “fast”, that is if 2 τs >> τf. In this case the generalized order parameter S can be factorized

S2 = S2S2 f s (6.3)

2 2 Sf and Ss represent order parameters of fast and slow internal motions, 2 2 respectively. This model converges to the standard model if Sf = 1, or Ss = 1, or τf = τs. There are special cases of the given models in which internal motions are considered very fast, and corresponding correlation times, τe or τf, negligibly small. In this case, Equation (6.1) becomes

2 2 S τM J()ω = ------(6.4) 51 + ω2τ2 M

and Equation (6.2) becomes

⎧⎫S2τ ()τS2 – S2 ′ J()ω = --2------M + ------f s- ⎨⎬2 2 2 ′2 (6.5) 5⎩⎭1 + ω τM 1 + ω τs

6 EXPERIMENTAL SECTION 92

If global tumbling is anisotropic but axially symmetric, the standard model correlation function is

3 2 2⎧⎫2 Aiτi ()τ1S– J()ω = ---S ------+ ------⎨⎬∑ 2 2 2 2 (6.6) 5⎩⎭1 + ω τi 1 + ω τ i1=

4 2 2 2 2 Where A1=0.75 sin α, A2 = 3 sin α cos α, A3 = (1.5 cos α - 0.05) and α is the angle between the amide bond vector and the molecular symmetry axis. The global -1 -1 -1 correlation times are as follows: τ1 = (4D|| + 2D⊥) , τ2 = (D|| + 5D⊥) , and τ1 = (6D⊥) . -1 The effective global correlation time is τM,eff = (2D|| + 4D⊥) and 1/τ = 1/τM,eff + 1/τe. Again, if the internal motions are sufficiently fast, the last term vanishes. The effects of exchange line broadening are taken into account in the form of an additive Rex term

1 1 = * + R ex (6.7) T2 T2

This finally yields a set of internal motion models with different numbers of 2 2 2 2 2 degrees of freedom, which we designate as {S }, {S , Rex}, {S , τe}, {S , τe, Rex}, {Ss , 2 2 2 2 2 Sf , τs}, {Ss , Sf , τs, Rex}, and {Ss , Sf , τs, τf}. Quantitative description of internal motions requires prior knowledge of global motion. This, in turn, depends on the particular model of the internal motions. One has to make successive approximations of the internal dynamics to obtain a self-consistent picture. The simplest approximation is utilized in the method of determining τM based on the T1/T2 ratios (332). This method implicitly assumes vary fast internal dynamics, i.e., the {S2} model, with sufficiently large order parameters for most of the protein. If

6 EXPERIMENTAL SECTION 93 the protein structure is known, the anisotropy of the rotational diffusion can be determined in a similar way (297). The validity of this assumption can be easily proven by checking whether T1/T2 ratios measured at two different field strengths lead to the same motional parameters. If more accurate estimates are necessary, they can be 2 obtained using the standard model {S , τe} for selected amide spins. The extended 2 2 model {Ss , Sf , τs} does not yield a well-defined minimum of penalty function and provides only an estimate of τM (302, 333).

Optical rotations were measured on a Perkin-Elmer 241 polarimeter (10 cm, 1 ml cell) at r.t. and are given as xx° (concentration in g/100 ml, solvent).

RP-HPLC was performed on a Waters HPLC system with 220 nm UV detection. For analytical runs, a C8 column (Macherey-Nagel Nucleosil 250 mm × 4.6 mm, 300 Å, 5 µm), a C18 column (Macherey-Nagel Nucleosil 250 mm × 4.6 mm, 100

Å, 5 µm), or a C18 short column (Waters Polarity 100 mm × 4.6 mm, 100 Å, 3 µm) at a flow rate of 1 ml/min was used. Peptides and small organic molecules were eluted with linear gradients of solvents A and B (A = acetonitrile containing 0.05% TFA, B = H2O containing 0.1% TFA). Preparative RP-HPLC separations were performed using a C8 column (Macherey-Nagel Nucleosil 250 mm × 21 mm, 300 Å, 7 µ) or a C18 column (Vydac 250 mm × 22 mm, 300 Å, 10 µ) at a flow rate of 10 ml/min. Linear gradients of solvents A and B (A = acetonitrile, B = H2O containing 0.1% TFA) were used. Gradients for analytical and preparative RP-HPLC are given as %A/%B to %A/%B in xx min.

RP-HPLC with Fluorescence detection was performed on a Kontron Instruments (Flowtek, Basel CH) HPLC system with 220 nm UV detection (430A) and equipped with a SFM 25 fluorescence spectrophotometer. A C18 column (Vydac TM 218 TP 250 mm × 4.6 mm) at a flow rate of 1 ml/min was used. Substances were eluted with linear gradients of solvents A and B (A = acetonitrile containing 0.05% TFA, B =

H2O containing 0.1% TFA). Gradients are given as %A/%B to %A/%B in xx min.

6 EXPERIMENTAL SECTION 94

Analytical Size Exclusion Chromatography. The aggregation state of CM variants was determined by analytical size exclusion chromatography using a Superose 12 (HR 10/30) FPLC column from Pharmacia. Chromatography was performed at 4 °C using PBS (10 mM phosphate, 160 mM NaCl, pH 7.5) as the running buffer (flow rate 0.3 ml/min, detection at 280 nm) and a 500 µl injection loop. Prior to analytical runs, columns were calibrated according to the method outlined in the instruction manual for the LMW Gel Filtration Calibration Kit from Pharmacia using the supplied protein standards (ribonuclease A, MW 13,700 Da; trypsinogen, MW 24,000 Da; ovalbumin, MW 43000 Da; bovine serum albumin, MW 67,000 Da) and additionally proteins from Sigma (aprotinin, MW 6,500 Da; cytochrome c, MW 12,400 Da; carbonic anhydrase,

MW 29,000 Da). The void volume, V0 (6.84 ml), and the total bed volume, Vt (26.01 ml), were determined using blue dextrane 2000 and DTT. The elution parameter,

KAV, was calculated for each protein using equation (6.8):

K = V −V V −V AV ( e 0 ) ( t 0 ) (6.8)

where Ve is the elution volume of the protein, V0 is the void volume of the column, and

Vt is the total bed volume of the column. A plot of KAV vs. log(MW) for the standard proteins gave the calibration curve (log Mw = -3.982 KAV + 5.778). The oligomerization state of the mutant variants was assigned by dividing the determined molecular weights by the mass of the monomeric species as calculated from the sequence and confirmed by mass spectroscopy.

TLC was performed on Merck silica gel 60 F254 plates; compounds were visualized either by quenching of UV fluorescence, by putting the plates into a jar with

I2 on silica, or by dipping the plates into a ninhydrin solution (200 mg ninhydrin in 100 ml EtOH) or a KMnO4 solution (0.5 g KMnO4 in 100 ml 1 mol/l NaOH) followed by heating. The solvent and the visualization method are given in brackets after Rf.

SDS-Polyacrylamid-Gel-Electrophoreses. For electrophoresis and staining the Phast-system from Pharmacia was used. Usually the Phast-Gel Homogeneous 20 (20% crosslinker) was used in combination with SDS buffer strips for

6 EXPERIMENTAL SECTION 95 gels under denaturing conditions or buffer strips without SDS for native gels. 8 µl of the sample were mixed with 2 µl 5x SDS PAGE loading buffer, denatured for 3 min at 95 ºC and then centrifuged for 1 min at 14,000 rpm. 2 µl of this sample were used to load the gel according to the instructions that came with the Phast-system. In addition a low molecular weight marker from Pharmacia was loaded which contains the following proteins (Table 6.1):

Protein Molecular weight µg / 200 µl marker solution per subunit [Da] Phosphorylase b 94,000 67 Albumin 67,000 83 Ovalbumin 43,000 147 Carbonic anhydrase 30,000 83 Trypsin Inhibitor 20,100 80 α-Lactalbumin 14,400 116 Table 6.1. Low Molecular Weight Marker from Pharmacia.

The gels were run with the method provided by the manufacturer for SDS- PAGE with Homogeneous 20 gels (separation technique file No. 111, PhastSystem) and stained with the Fast Coomassie Staining method at 50 ºC (development technique file No. 200, PhastSystem). For staining the following slightly modified program was used: 30% EtOH and 10% HOAc, 4 min.; 30% EtOH and 10% HOAc, 4.5 min.; 0.05% Coomassie R350, 10% MeOH, 9% HOAc, 2% ammoniumsulfate and 0.1% coppersulfate, 20 min.; 30% EtOH and 10% HOAc, 0.1 min.; 10% HOAc, 5 min.; 10% HOAc, 10 min.; 30% MeOH, 10% HOAc and 2% glycerol, 2 min.; UPW 0.5 min..

UV data were collected on a Perkin-Elmer Lambda series UV/VIS spectrophotometer (Lambda 16, Lambda 20, or Lambda 40).

6 EXPERIMENTAL SECTION 96

Solutions, Buffers, and Media Molecular Biology

Enzymes and plasmids that were not produced in the group were purchased from New England BioLabs (Beverly MA, USA), Boehringer Mannheim (Luzern LU, Schweiz), Qiagen (Basel BS, Switzerland), or Novabiochem.

All nucleic acid manipulations and media preparations, if not specifically stated, were according to standard procedures (334)

0.5 M EDTA pH 8.0. In a 500 ml Schott bottle 93.06 g Na2EDTA·2 H2O were dissolved in 400 ml UPW. The pH was adjusted with NaOH pellets (approximately 10 g) to pH 8.0 and the solution brought to a total volume of 500 ml, autoclaved and stored at r.t..

20% (w/v) glycerol in 100 mM CaCl2. In a 250 ml Schott bottle 51.28 g glycerol 87%, and 2.94 g CaCl2·2 H2O were dissolved with 100 ml UPW. The volume was adjusted to 200 ml with UPW, the solution was autoclaved and then stored at 4 °C.

10% (w/v) glycerol. In a 100 ml Schott bottle 12.82 g glycerol 87%, was dissolved with 80 ml UPW. The volume was adjusted to 100 ml with UPW, the solution was autoclaved and then stored at 4 °C.

Ethidiumbromide (10 mg/ml). UPW was added to 200 mg ethidiumbromide to give a final volume of 20 ml. The solution was stored at 4 °C.

10x TBE buffer. 108 g Tris-base, 55 g boric acid, and 9.3 g Na2EDTA were filled up with UPW to give 1 l final volume. The pH was checked to be 8.25 and the solution was autoclaved.

6 EXPERIMENTAL SECTION 97

5x Loading buffer for agarose-gels. 4.6 ml glycerol (87%), 4.0 ml of 0.25 M EDTA (pH 8), 1.4 ml UPW and 3 mg bromphenol blue were mixed and stored at -20ºC

1 M Tris-HCl pH 8.0. 0.047 g Tris-Base and 8.426 g Tris-HCl were dissolved in UPW to give a final volume of 50 ml. The pH was checked to be 8.0 and the solution was autoclaved.

1 M Tris-HCl pH 6.8. 0.578 g Tris-Base and 15.008 g Tris-HCl were dissolved in UPW to give a final volume of 100 ml. The pH was checked to be 6.8 and the solution was autoclaved.

1x TE-buffer (10 mM Tris-HCl, 1 mM EDTA). 5 ml of 1 M Tris-HCl (pH 8) and 2 ml of 0.25 M EDTA were dissolved in UPW to give a final volume of 500 ml and the solution was autoclaved.

1 M glycine-NaOH (pH 8.9). In a 1 l Schott bottle 75.07 g of glycine were dissolved in 900 ml UPW, the pH was adjusted to 8.9 with 10 N NaOH, and the volume was brought to 1 l with UPW. The solution was autoclaved and stored at r.t..

Isopropyl-1-thio-β-D-galactoside (IPTG) (50 mM). 0.476 g IPTG were dissolved in UPW to give a final volume of 40 ml. After dissolvation the solution was sterile filtered and stored at -20 °C.

Ampicillin (Amp) (100 mg/ml). 2 g of ampicillin were dissolved in UPW to give a final volume of 20 ml. The solution was sterile filtered and stored at –20 °C.

Chloramphenicol (Cam) (30 mg/ml). 0.6 g chloramphenicol were dissolved in 14 ml EtOH. The volume was adjusted to 20 ml with UPW. The solution was stored at -20 °C. It does not need to be sterile filtered because of the 70% EtOH.

6 EXPERIMENTAL SECTION 98

Buffer B (50 mM glycine-NaOH, pH 8.9, 5% (v/v) 2-propanol, 10% (v/v) glycerol). In a 2 l flask 200 ml glycerol 87%, 100 ml 2-propanol and 100 ml 1 M glycine-NaOH pH 8.9 were dissolved in UPW to a final volume of 2 l. The solution was degassed for 2 h at membrane pump vacuum. If needed, immediately before use 308 mg of DTT were added (1 mM).

Osmo buffer (20% (w/v) sucrose, 30 mM Tris-HCl pH 8.0). In a 1 l Schott bottle 100 g of sucrose were dissolved in 485 ml UPW. The solution was autoclaved, and after cooling to r.t., 15 ml of 1 M Tris-HCl pH 8.0 were added. It was stored at r.t..

5x SDS PAGE loading buffer. In a conical tube 2.5 ml of 1 M Tris-HCl, pH 6.8, 0.771 g DTT, 1 g of 10% SDS, 0.05 g of 0.5% bromphenol blue, and 5.75 g of 50% (w/v) glycerol (87%) were mixed and diluted with water to 10 ml. Aliquots of 1 ml were stored in Eppendorf tubes at –20 ºC.

LB agar plates. In a 2 l Erlenmeyer flask 40 g of LB-Agar from Bio101 were dissolved in 900 ml UPW. The volume was adjusted to 1 l with UPW. The solution was autoclaved and cooled to 50-60 °C. For Amp150 plates 1.5 ml ampicillin stock solution (100 mg/ml) was added, for Amp150Cam30 plates 1.5 ml ampicillin stock solution (100 mg/ml) and 1.0 ml chloramphenicol stock solution (30 mg/ml) were added. Approximately 30 ml of this solution were poured in each Petri dish. The plates were dried for 24 h at 37 °C and then stored in a sealed plastic bag at 4 °C.

LB Medium. In a 2 l Erlenmeyer flask 25 g of LB Medium (Bio101) were dissolved in 900 ml UPW. The volume was adjusted to 1 l with UPW, the solution was autoclaved and then stored at 4 °C. The desired antibiotics were added before use. For Amp150 medium 1.5 ml ampicillin stock solution (100 mg/ml) was added, for Amp150Cam30 medium 1.5 ml ampicillin stock solution (100 mg/ml) and 1.0 ml chloramphenicol stock solution (30 mg/ml) was added. For precultures 5 ml LB- medium was used.

6 EXPERIMENTAL SECTION 99

SOC Medium. In a 500 ml Erlenmeyer flask 5 g Bacto-Tryptone, 1.25 g Bacto- Yeast extract, 0.15 g NaCl, 0.05 g KCl were dissolved in 200 ml UPW. The volume was adjusted to 240.5 ml with UPW, the solution was autoclaved and then 2.5 ml of MgSO4

(1 M), 2.5 ml of MgCl2 (1 M) and 4.5 ml of 20% (w/v) glucose were added. The medium was stored at r.t..

Ingredients for Minimal Media

10x M9-Salts. In a 1 l Schott bottle 75.2 g Na2HPO4·2 H2O, 30.0 g KH2PO4,

10.0 g NH4Cl, and 5.0 g NaCl were dissolved in 850 ml of UPW. The pH was adjusted to 7.0 with 5 M NaOH and the volume to 1 l with UPW, the solution was autoclaved and then stored at r.t..

20% (w/v) Glucose. In a 250 ml Schott bottle 40 g D-(+)-glucose were dissolved in 150 ml of UPW. The volume was adjusted to 200 ml with UPW, the solution was autoclaved and then stored at r.t..

1 M MgSO4. In a 250 ml Schott bottle 100 ml 24.65 g MgSO4·7 H2O were dissolved in 75 ml of UPW. The volume was adjusted to 100 ml with UPW, the solution was autoclaved and then stored at r.t..

FeSO4 (100 mM). In a 50 ml conical tube 0.556 g FeSO4·7 H2O were dissolved with 18 ml UPW. The volume was adjusted to 20 ml with UPW, the solution was sterile filtered, divided into 2 × 10 ml aliquots and then stored at -20 °C.

Thiamin-HCl (5 mg/ml). In a 50 ml conical tube 0.25 g thiamin-HCl were dissolved in a total volume of 50 ml UPW. 5 × 10 ml aliquots were sterile filtered and stored at 4 °C in 10 ml conical tubes.

6 EXPERIMENTAL SECTION 100

Rifampicin (0.1 mg/ml). In a 15 ml conical tube 1 mg rifampicin was added to 9 ml MeOH. A few drops of a 5 mol/l NaOH solution were added to allow for dissolution. The volume was filled up to 10 ml and stored at 4 °C.

Aro-P-Mix (1000x). In a 50 ml conical tube 100 mg 4-hydroxybenzoic acid, 100 mg 4-aminobenzoic acid, and 100 mg 2,3-dihydroxybenzoic acid were dissolved with 18 ml UPW. The pH was adjusted to 5.5 with 5 M NaOH and the volume brought to 20 ml with UPW. The solution was sterile filtered, divided into 2 × 10 ml aliquots, and then stored at -20 °C.

Buffers for Purification of BsCM 1-87, Intein, CBD Fusion

Column Buffer (Tris) (20 mM Tris, 300 mM NaCl, 1 mM EDTA, pH 8). In a 5 l Erlenmeyer flask 84.14 g NaCl were dissolved in 4 l UPW and 96 ml 1 M Tris- HCl pH 8 was added. The volume was brought to 4.8 l and the solution was autoclaved. After cooling to r.t. 9.6 ml 0.5 M EDTA stock were added. The solution was stored at r.t..

Column Buffer (KPi) (50 mM sodium phosphate, 300 mM NaCl, 1 mM

EDTA, pH 8). In a 5 l Erlenmeyer flask 84.14 g NaCl, 1.76 g NaH2PO4 • H2O and

50.89 g Na2HPO4·2 H2O were dissolved in 4 l UPW. The volume was brougth to 4.8 l and the solution was autoclaved. After cooling to r.t. 9.6 ml 0.5 M EDTA stock were added. The solution was stored at r.t..

Lysis Buffer (Column Buffer plus 0.1% Triton, 20 µM PMSF). In a sterile 500 ml bottle to 500 ml column buffer were added 500 µl Triton X-100 and 1.7 mg PMSF.

Cleavage Buffer (Column Buffer plus 50 mM thiol). In a 50 ml conical tube to 50 ml column buffer were added either 0.386 g DTT, or 220 µl NMMAA, or 0.41 g MESNA.

6 EXPERIMENTAL SECTION 101

Strains, Plasmids, and Oligonucleotides

Strains. General cloning was done in the chorismate mutase-deficient Escherichia coli strain KA12. Protein production was carried out in E. coli strain KA13, a derivative of KA12 which carries the DE3 prophage in its genome, allowing IPTG- inducible expression of genes under the control of the T7 promoter. Both strains have been described previously (82, 335).

Plasmids. All plasmids that were transformed or constructed had the gene for ampicillin resistance. For a description of the plasmids used in this work see Table 6.2.

Plasmid name Length Description pKET3-W 5766 bp Contains the gene for wild type BsCM (212). pKIMP-UAUC 5028 bp Necessary to reconstitute the metabolic pathway from prephenate to 4-hydroxy-phenylpyruvate and phenylpyruvate in KA12 and KA13. Contains the genes for a monofunctional prephenate dehydrogenase from Erwinia herbicola and a prephenate dehydratase from Pseudomonas aeruginosa, as well as a chloramphenicol resistance marker (CamR) (78, 82). pKSS 4133 bp Transformation control. Vector for efficient selection; has an ampicillin resistance marker (AmpR) (336). pTXB1 6706 bp From NEB. Contains the gene for the Mxe GyrA intein fused to a chitin binding domain after a multiple cloning site (172, 194). pTYB1 7477 bp From NEB. Contains the gene for the Sce VMA intein fused to a chitin binding domain gene after a multiple cloning site (172, 337). pMYB5 8602 bp From NEB. Contains the gene for the maltose binding protein fused to the gene for the Sce intein and a chitin- binding domain gene (172, 194, 337). pTXB1-BsCM87 6916 bp From NEB. Contains the gene for BsCM 1-87 fused to the gene for the Mxe GyrA intein fused to a chitin- binding domain gene. pTYB1-BsCM87 7687 bp From NEB. Contains the gene for BsCM 1-87 fused to the gene for the Sce VMA intein fused to a chitin- binding domain gene. Table 6.2. Plasmids.

6 EXPERIMENTAL SECTION 102

Oligonucleotides. The oligonucleotides used in this work were custom synthesized by Microsynth (Balgach, Switzerland) and are described in the following table (Table 7.3). A 50 pmol/µl stock solution of the oligonuctleotides was prepared.

Primer name Bases 5,-3, Sequence GAT GAC TGT ACA GAC AGA TGT CCC TCA SVSF-1-D102E 46 GGA GCA GAT CAG ACA TGT A GAT GAC TGT ACA GAC AGA TGT CCC TCA SVSF-2-R105K 55 GGA TCA GAT CAA ACA TGT ATA TTT AGA A GAT GAC TGT ACA GAC AGA TGT CCC TCA SVSF-3-DERK 55 GGA GCA GAT CAA ACA TGT ATA TTT AGA A SVSF-4-T125* 29 AAC TCC TCG AGT TAA TTT TTT GTC AAG CT SVSF-5-K123* 29 AAC TCC TCG AGT TAT GTC AAG CTT AAA TC AAC TCC TCG AGT TAC AAG CTT AAA TCG SVSF-6-T122* 37 GGC CTC AAT A AAC TCC TCG AGT TAG CTT AAA TCG GGC SVSF-7-L121* 34 CTC AAT A SVSF-8b-S120* 31 AAC TCC TCG AGT TAT AAA TCG GGC CTC AAT A 04-T7TR 24 CAG CAG CCA ACT CAG CTT CCT TTC T7PRO2 20 TAA TAC GAC TCA CTA TAG GG AK-1-RPSceI 24 AAC CAT GAC CTT ATT ACC AAC CTC AK-1-RPMxeI 18 GGC ACG ATG TCG GCG ATG AK-3-SAP- GGT GGT GCT CTT CC GCA CTT CTT AAG ACC GC 35 BsCM87 C TGT Table 6.3. Primers with mutations in bold and the codon for the relevant Cys underlined.

General Methods: Molecular Biology

Sterile filtration. Sterilfilters from Semadeni AG, Ostermundigen BE, Switzerland with a pore size of 0.2 µm were used.

Plasmid extraction. For plasmid extraction the JETquick Plasmid Miniprep Kit of GENOMED was used as described in the included procedure. The optional washing of the column with GX was omitted. To elute the plasmids 50 µl of either 10 mM Tris-HCl pH 8.0 preheated to 65 ºC was used, if the plasmid was intended to be used for electroporation 1/5 TE buffer was used instead. Plasmids were stored at –20 °C (long term –80 °C) if not used immediately.

6 EXPERIMENTAL SECTION 103

Restriction

Restriction enzymes

Restriction enzyme Reaction buffer Recognition sequence units/ml AatII NEBuffer 4 GACGT/C 20,000 BsrGI NEBuffer 2 + BSA T/GTACA 10,000 MlutI NEBuffer 3 + BSA G/GTNACC 10,000 NdeI NEBuffer 4 CA/TATG 20,000 NheI NEBuffer 2 G/CTAGC 5,000 SapI NEBuffer 4 GCTCTTC(1/4) 2,000 XbaI NEBuffer 3 + BSA T/CTAGA 20,000 XhoI NEBuffer 2 + BSA C/TCGAG 20,000 Table 6.4. Restriction enzymes with ideal buffer and DNA-recognition sequence. Data from New England BioLabs catalogue (1998/99).

All restriction digests were stopped by adding the appropriate amount of 5× agarose-gel loading buffer.

Extraction of DNA from preparative agarose gels. The desired bands were localized on a weak UV transilluminator and cut out of the gel with a scalpel. Only the main part of the band was excised to exclude contaminants. The DNA was extracted from the agarose gel and purified with the GENECLEAN Kit of Bio101 Inc. (Vista, CA, USA) or the JETquick Gel Extraction Kit of GENOMED as described in the included procedure. The DNA was eluted with 40 µl of 10 mmol/l Tris-HCl pH 8.0.

DNA ligation. The PCR fragments, digested with the appropriate restriction enzymes were incubated with an acceptor plasmid that has been cut with the same restriction enzymes for 16 h with T4-DNA-ligase (5 units per 20 µl) from Fermentas (Vilnius, Lithuania) at 16 °C. The ratio of insert to acceptor was 2:1. Controls were performed in which either the PCR-fragment or the T4-DNA-ligase and the PCR- fragment were replaced by UPW to see how much uncut and half cut acceptor fragment was present. For electroporation the ligated DNA was purified using the JETquick PCR Purification Kit of GENOMED as described in the included procedure. The only modification was that after addition of 1/5 TE buffer the column was incubated for

6 EXPERIMENTAL SECTION 104

1 min at r.t. before elution of the DNA. The purified plasmids were stored at -20 ºC if not used immediately.

CaCl2-competent cells. A 5 ml LB medium preculture was inoculated with a single colony of the desired E. coli strain. The culture was grown overnight at 37 ºC and 230 rpm. 30 ml LB medium with the appropriate antibiotics were inoculated with

150 µl of this preculture and grown for 2-3 h at 37 ºC and 230 rpm until an OD600 of around 0.4 was reached. The centrifuge, the centrifuge bottles and the 100 mM CaCl2 solution were precooled to 4 °C. The cells were subsequently kept on ice. The cells were centrifuged at 5000 rpm in a SS34 rotor for 10 min. The cell pellet was resuspended in 10 ml cold 100 mM CaCl2 solution with slight swirling and again centrifuged at 5000 rpm for 10 min. The cell pellet was again resuspended in 2.5 ml cold 100 mM CaCl2 solution with slight swirling, put on ice for 30 min, and again centrifuged at 5000 rpm for 10 min. The cell pellet was resuspended, this time in 2.5 ml cold 20% glycerol in 100 mM CaCl2 solution with slight swirling. The cell suspension was frozen in liquid N2 and stored at -80 ° C in 500 µl aliquots if not used immediately.

Transformation into CaCl2-competent cells. For all transformations a positive (1 µl pKSS 1-10 ng/µl) and a negative control (no plasmid) were performed. 10 µl of the ligated plasmid DNA were diluted with 1x TE-buffer to 100 µl, mixed, and put on ice. The CaCl2-competent cells were thawed at r.t. and then put on ice. 100 µl of the CaCl2-competent cells were added to each plasmid. The mixture was incubated for 30 min on ice. Then the cells were heat shocked for 2 min in a 42 °C water bath. After the cells were incubated for 5 min at r.t. 800 µl LB medium was added. The cell suspension was mixed by inverting a few times and then incubated for 1 h at 37 ºC. It was put on ice and 100 µl were plated directly on LB plates with the appropriate antibiotics (Amp150 for KA12, KA13 and BL21, Amp150 and Cam30 for KA13/pKIMP- UAUC plus the transformed plasmid). The remaining cell suspension was concentrated 10 times by centrifuging it at 10,000 rpm for 2 min and resuspending the pellet in 100 µl of the supernatant. The whole sample was plated on appropriate LB plates. The plates were incubated overnight at 37 ºC, evaluated, and then stored at 4 ºC. Three dimensional single colony streak outs were made from a few colonies. Glycerol cultures

6 EXPERIMENTAL SECTION 105 were made from all strain-plasmid combinations by adding an equal amount of 44% glycerol to a densely grown culture and freezing them at –80 °C.

Electro-competent cells. A 5 ml LB medium preculture was inoculated with a single colony of the desired E. coli strain. The culture was grown overnight at 37 ºC and 230 rpm. 300 ml LB medium with the appropriate antibiotics was inoculated with

4 ml of this preculture and grown for 2-3 h at 37 ºC and 230 rpm until an OD600 between 0.6 and 0.8 was reached. The centrifuge, the centrifuge bottles, the 10% (w/v) glycerol solution and 1 l of autoclaved UPW were precooled to 4 °C. The cells were subsequently kept on ice. The cells were centrifuged at 4500 rpm (3,400×g) in a GS3 rotor for 10 min. The cell pellet was resuspended in 30 ml cold UPW with slight swirling and again centrifuged at 4500 rpm for 5 min. The cell pellet was resuspended in 30 ml cold UPW with slight swirling and again centrifuged at 4500 rpm for 5 min. The cell pellet was resuspended, this time in 30 ml cold 10% (w/v) glycerol solution on ice. The cell suspension was frozen in liquid nitrogen in 500 µl aliquots and stored at – 80 °C.

Electroporation. For all transformations a positive (1 µl pKSS 1-10 ng/µl) and a negative control (no plasmid) were performed. The amount of plasmid in the plasmid solution was estimated from agarose gels by comparing the intensity of the band with the intensity of a marker (λBstEII or φX174HaeIII) band of similar length. The plasmid solution should contain about 0.6 ng DNA in 2 µl. Dilutions were done with 1/5 TE- buffer. The electro-competent cells were thawed at r.t. and then put on ice. 50 µl of the sample were pipetted in precooled Eppendorf tubes and 2 µl of the plasmid solution was added. The electroporator and the cuvettes (electrode distance 0.2 cm) were from BioRad. The parameters were adjusted to 25 µF, 200 Ω and 2.5 kV. The electroporation mixture was transferred into the precooled cuvette. The cuvette was closed, dried on the outside, and inserted into the electroporator. The electroporation was performed and as soon as the time constant was displayed, 1 ml SOC medium was added to the cuvette. The solution was transferred to a 15 ml conical tube and another 0.5 ml SOC medium were added into the cuvette to recover the remaining cells and also transferred to the same conical tube. The cells were incubated

6 EXPERIMENTAL SECTION 106 for 60 min at 37 ºC and 220 rpm. The cell suspension was then plated on plates with the appropriate antibiotic in different dilutions (50 µl of 1x, 50 µl of 10x, remaining 0.1x). The plates were incubated at 37 ºC overnight. Single colony streak outs were done from some of the colonies. Glycerol cultures were made from all strain-plasmid combinations by adding an equal amount of 44% glycerol to a densely grown culture and freezing them at –80 °C.

Polymerase Chain Reaction (PCR). For PCR the method described by Mullis and Sambrook was used (334, 338). The primers are described above (Table 7.5). Primer and template (plasmid-DNA) were pipetted together in a 10,000 : 1 molar ratio. The PCR reaction mixture was composed of:

Description Amount Template plasmid-DNA (pKET3-W, pTXB1 or pTYB1) 0.005 pmol = 1 µl Sense-primer 50 pmol = 5 µl Reverse-primer 50 pmol = 5 µl UPW 28 µl 10 × PCR-buffer (Qiagen, Basel BS, Switzerland) 5 µl 10 × dNTP-mix, (2 mM each of dATP, dCTP, dGTP and dTTP 5 µl (Pharmacia)) DNA-polymerase (HotStar-Taq, Qiagen) 2.5 units = 0.5 µl Table 6.5. PCR-reaction solution. General composition of a PCR reaction.

The DNA polymerase comes from Thermophilus aquaticus and has only a 5,-3,- endonuclease activity (RNA primer digest), but no 3,-5,-endonuclease activity which would remove incorrectly incorporated nucleotides (proof-reading). The DNA polymerase from T. aquaticus (HotStar-Taq) has to be incubated for 15 min at 95 ºC to activate the enzyme. This allows the reaction mixture to be prepared at r.t.. If necessary, the amplified DNA was purified using the JETquick PCR Purification Kit of GENOMED as described in the included procedure, eluting with 50 µl of 10 mmol/l Tris-HCl pH 8.0 heated to 65 °C. A Gene Amp PCR System 9700 (Perkin Elmer), Techne ProGene PCR, or Peltier Thermal Cycler PTC-200 (MJ Research) was used with one of the following programs: Cycle sequencing PCR: 25 PCR cycles (30 s at 94 °C, 15 s at 50 °C, and 4 min at 60 °C), before cooling to 4 °C.

6 EXPERIMENTAL SECTION 107

Standard gene fragment amplification PCR: 15 min at 95 °C, 25 PCR cycles (1 min at 94 °C, 2 min at 45 °C, and 2 min at 72 °C) and a final extension step (10 min at 72 °C) before cooling to 4 °C.

Sequencing. Plasmid constructs were verified by DNA sequencing of the newly cloned regions on an ABI PRISM 310 Genetic Analyzer (PE Applied Biosystems) using the Terminator Ready Reaction Mix (BigDyeTM, PE Applied Biosystems) for chain termination chemistry (339). The dye-tagged DNA utilized for automated sequencing was prepared by cycle-sequencing PCR (see above). The total volume of a typical reaction mixture was 8 µl and contained 0.5 µg template DNA, 2 µl Terminator Ready Reaction Mix (BigDyeTM, PE Applied Biosystems) and 2 µl sequencing buffer (PE Applied Biosystems). The PCR products were purified by ethanol precipitation. The washed DNA pellet was either resuspended in template suppression reagent (PE Applied Biosystems) when the POP6 sequencing polymer (PE Applied Biosystems) was used, or in sterile water when the POP4 sequencing polymer (PE Applied Biosystems) was used. For sequencing, instrument parameters were set according to the manual.

Mono Q ion exchange chromatography was done on a Pharmacia FPLC system equipped with a Mono Q (Pharmacia) column using first 44 ml of buffer B without glycerol and then a linear gradient between buffer B without glycerol and buffer B without glycerol and 1 M NaCl from 0 to 0.35 M NaCl in 250 ml at a flow rate of 3 ml/min. Fractions of 10 ml were collected. Prior to runs the column was equilibrated to Cl- with 1 M NaCl in buffer B without glycerol according to the method outlined in the instruction manual for the column.

Protein concentration was determined by measuring the absorbance of a solution at 280 nm and applying the Beer-Lambert law: A = c × ε × l (Eq. 6.9) where c is the protein concentration (M), ε is the molar absorption coefficient (M-1cm- 1 ), and l is the path length. The molar absorption coefficient ε280 nm for BsCM and its

6 EXPERIMENTAL SECTION 108 mutants is 8370 M-1 cm-1. When the molar absorption coefficient was not known, it was estimated using the following equation (340-342): -1 -1 ε280 nm [M cm ] = 5500 × (#Trp) + 1490 × (#Tyr) + 125 × (#Cys) (Eq. 6.10) where # is the number of the tryptophane (Trp), tyrosine (Tyr), or cysteine (Cys) residues present in the protein sequence.

Ultrafiltration was used to concentrate proteins and to exchange buffers. It was generally done at 4 °C. For protein solutions with volumes greater than 20 ml, Amicon Ultrafiltration stir cells were used with membranes that had a molecular weight cut-off (MWCO) of 3 or 10 kDa. Concentration was performed by applying a nitrogen pressure of 4 bar. Solutions with volumes between 5 and 15 ml were concentrated with Macrosep centrifuge cartridges (Pall Filtron, Northbourough MA, USA) that had a MWCO of 3 or 10 kDa. Smaller volumes were concentrated with Millipore UFV4 Centrifugal Filters with MWCOs of 3 or 10 kDa. Buffer exchanges were done by repeated concentration and redilution in the new buffer. All procedures followed the instructions outlined in the supplier’s manuals.

Kinetics. All kinetic measurements were performed at 30 °C and usually in

50 mM potassium phosphate buffer pH 7.5. Measurements of the pH dependence of kcat,

Km, and kcat/Km were done in the following buffers: pH 3.6, 4.2, 4.6, and 5.3 sodium acetate (50 mM); pH 5.8, 6.3, 6.7, 7.05, 7.5 potassium phosphate (50 mM); pH 8.0 and 8.5 Tris-HCl (50 mM); pH 8.25, 8.75, 8.9, 9.25, 9.5, 9.75, and 10.0 glycine-NaOH (50 mM). Initial rates were determined by monitoring the disappearance of chorismate -1 -1 spectrophotometrically at 274 nm (ε274 nm = 2630 M cm ) or 310 nm (ε310 nm = -1 -1 370 M cm ). The apparent values for kcat/Km were estimated at 20 µM chorismate. Measurements of the individual variants were always performed in duplicate on the same instrument, using 1 cm path length quartz cuvettes from Hellma (QS 1.000 or 115B-QS). All initial rates were corrected for the corresponding temperature-specific background reaction. Kinetic parameters kcat and Km were calculated from the initial rates as previously described, (43) using ideally a minimum of five substrate concentrations ranging from at least 4-fold below Km to at least 4-fold above Km.

6 EXPERIMENTAL SECTION 109

Chorismate was produced by Kai Walter, Florian Seebeck, or Dominik Künzler using a previously published procedure (343).

Inhibition assays. The chorismate mutase catalyzed conversion of chorismate to prephenate was performed in the presence of 0 to 20 µM of the Bartlett inhibitor 1 (Figure 1.2) (209) at 30 °C in 50 mM potassium phosphate buffer pH 7.5 at a substrate concentration close to Km. The inverse of vinit/[E] was plotted against the inhibitor concentration [I] and the following equation was used to calculate Ki from the slope of this curve when Km and kcat are known (344):

K m Ki = (6.11) kcat ⋅[]S ⋅ slope

General Procedures: Molecular Biology

Expression and purification of BsCM wt and mutants. A 5 ml LB- Amp150 preculture was inoculated with a freshly grown single colony of KA13 with the appropriate plasmid (pKET3-W for the wt) and grown overnight at 37 °C, 230 rpm. The whole preculture was used to inoculate the 1 l LB-Amp150 main culture, which was grown under the same conditions until an OD600 of around 0.6 was reached (4-5 h). Then the gene expression was induced by adding 10 ml of a 50 mM IPTG solution. The cells were incubated for another 24 h under the same conditions and then harvested by centrifuging them at 4 °C in 500 ml Nalgene bottles using a GS3 rotor at 4,000 rpm (2500×g) for 20 min. The pellet was frozen at -20 °C overnight or longer. At 0 °C the thawed pellet was resuspended in 180 ml osmo buffer and 360 µl of 0.5 M EDTA pH 8.0 as well as 180 µl of 1 M DTT solution were added. The cell suspension was incubated at r.t. for 20 min under shaking (200 rpm). The cells were precipitated by centrifugation at 6900 rpm (8000×g) at 4 °C for 10 min. The supernatant was decanted and centrifuged again at 7100 rpm (8000×g) using a GSA rotor at 4 °C for 10 min. If SDS-PAGE analysis showed that the pellet still contained a lot of BsCM, the osmolysis

6 EXPERIMENTAL SECTION 110 step was repeated. The clear solution was concentrated to a final volume of 20 ml using an Amicon ultrafiltration chamber with a Diaflo membrane (YM10, MWCO = 10 kDa). The retentate protein solution was dialyzed 2 times for at least 4 h against buffer B with 1 mM DTT. The solution was sterile filtered, and then the protein was then purified on a DEAE DE-52 cellulose (#4057-050; Whatman Inc., Clifton, NJ) column at r.t.. About 2.5 ml settled column bed volume was used for 1 l of culture. The column, with a UV monitor (280 nm) and chart recorder attached, was first equilibrated with 50-100 ml buffer B plus 1 mM DTT at a flow rate of 0.5 to 1 ml/min. Then the sample was loaded onto the column, the column was washed with 50 ml buffer B plus 1 mM DTT (or until baseline was flat again), and the protein was eluted with a linear NaCl gradient from 0 to 300 mM in buffer B plus 1 mM DTT. The fractions (15 ml) that contained BsCM as judged by SDS-PAGE were pooled and concentrated by ultrafiltration at 4 °C (MWCO = 10 kDa) to a concentration of more than 5 mg/ml. Protein for immediate use (within 2 months) was stored at 4 °C under argon; the remaining protein was stored at –80 °C to prevent oxidative damage.

Expression and purification of BsCM(1-87). A 5 ml LB-Amp150 preculture was inoculated with a freshly grown single colony of KA13/pTXB1- BsCM87 or KA13/pTYB1-BsCM87 and grown overnight at 37 °C (230 rpm). 1 ml of this preculture was used to inoculate a 1 l LB-Amp150 main culture, which was grown under the same conditions until an OD600 of around 0.8 was reached (6-7 h). Adding 10 ml of a 50 mM IPTG solution induced the gene expression. The cells were incubated for another 24 h at 24 °C (230 rpm) and then harvested by centrifugation at 4 °C in 500 ml Nalgene bottles using a GS3 rotor at 4000 rpm (2500×g) for 20 min. The supernatant was discarded and the pellet was frozen at -20 °C if not lysed immediately. A column was packed with 20 ml of a chitin beads suspension (about 10 ml settled bed volume) and equilibrated with 10 ml Column Buffer (Tris or KPi). The chitin column purification was done in the cold room at 4 °C. The cell pellet from a 1 l culture was resuspended in 35 ml ice-cold Cell Lysis Buffer and the cells were broken by sonication 10 times for 30 seconds with 30 seconds break in between at an amplitude of 50% and a cycle of 0.5. The cell debris was removed by centrifugation at 12,000 rpm (17,000×g) for 30 min at 4 °C. The clarified cell extract was loaded slowly (less than

6 EXPERIMENTAL SECTION 111

0.5 ml/min) on the prepared chitin column and washed with 100 ml column buffer (Tris or KPi). To induce the on column cleavage the column was equilibrated with 30 ml freshly prepared cleavage buffer (50 mM thiol, usually MESNA) and incubated overnight at 4 °C. The protein was eluted from the column in two batches with 5 and 10 ml column buffer. The first 5 ml were directly further purified by preparative RP- HPLC using a gradient of 25/75 to 60/40 in 45 min. The second batch was concentrated by ultrafiltration (Amicon, MWCO = 3 kDa) to 5 ml and also purified by preparative RP-HPLC under the same conditions. The fractions (10 ml) containing protein (usually around fraction 33) were pooled and lyophilized. The peptide was stored at -20 °C.

Folding of BsCM variants from 6 M GdmHCl was done by quickly diluting small portions (10 µl each time) of the denatured protein solution 100 fold into buffer B at r.t. under stirring and concentrating it back to the original volume by ultrafiltration (Amicon, MWCO = 10 kDa).

Chemicals and General Procedures: Chemistry

Chemicals were purchased from Sigma, Acros, ABCR, Merck, Aldrich or Fluka unless noted otherwise. Amino acids and resins were bought from Novabiochem or Bachem (Bubendorf BL, Switzerland). Reagents and solvents for peptide synthesis were obtained from Applied Biosystems.

Solvents were dried as follows: THF was freshly distilled over

Na/benzophenone before use. Et2O was filtered through a basic alumina plug directly before use. CH2Cl2 was distilled over CaH2. DMF was stirred over KOH pellets overnight and then distilled under reduced pressure over CaH2 at 40 °C.

Lyophilizations were performed with a Christ Gamma 2-20 apparatus.

6 EXPERIMENTAL SECTION 112

Resin loading was determined with the Fmoc-release assay (345). Approximately 1 mg of dry and fully protected resin-bound peptide was dissolved in 5 ml of a solution of 20% Pip in NMP (v/v). After 30 min, the absorption of the mixture was measured at 290 nm. Reference solutions were analyzed by the same procedure: ca. 15 mg Fmoc-Gly-OH were weighed into a 5 ml graduated flask, dissolved in 20% Pip/NMP (v/v), allowed to stand for 30 min, and diluted 50-fold; the absorption of the resulting solution was then determined. Measurements were always done in duplicate.

Solid-Phase Peptide Synthesis. Peptides were synthesized in stepwise fashion on an ABI 433A peptide synthesizer (Applied Biosystems) either on a 0.25 or 0.1 mmol scale using Fmoc methodology (128). In general, the HBTU/HOBt/NMP activation protocol for Fmoc chemistry (FastMoc® protocol, Applied Biosystems) (205), defined as Chemistry FM010_MPPSTD3, FM025_MPPSTD3, FM010_MPPSTD4, FM025_MPPSTD4 or FM025_MPPSTD4a (see appendix), where 010 and 025 stands for 0.1 mmol and 0.25 mmol scale respectively, was applied. Amino acid side chains were protected as follows: Arg(Pbf), Asn(Trt), Asp(OtBu), Cys(Trt), Gln(Trt), Glu(OtBu), Lys(Boc), Ser(OtBu), Thr(OtBu), Tyr(OtBu). For each coupling 1 mmol Fmoc protected amino acid was used. After some couplings, a capping step was performed, i.e. a mixture of acetic anhydride (Ac2O), HOBt and DIEA in NMP was added to the growing peptide chain after the coupling of the amino acid had been completed and before removal of the Fmoc group in order to acylate any remaining amines. After drying under high vacuum, peptides were cleaved from the resins by shaking in TFA in the presence of scavengers (general mixture: 91% (v/v) TFA, 4% (v/v) UPW, 1% (v/v) TIPS, 2.5% (v/v) phenol (molten), 1.5% (v/v) EDT; 40 ml of this mixture were used per g of resin for 3 h). Solvents were evaporated under reduced pressure and the residue triturated in cold Et2O in a centrifuge tube. After centrifugation at 3,000 rpm (1,600×g) for 20 min, the Et2O was decanted, and the procedure repeated twice. Crude peptides were dissolved in water/CH3CN mixtures, lyophilized and purified by RP-HPLC.

Native chemical ligations were done in degassed (30 min membrane pump vacuum, approximately 10 mbar) 6 M GdmHCl, 100 mM Tris pH 8.0, and 2.5% (v/v)

6 EXPERIMENTAL SECTION 113

PhSH for 24 h at r.t. under argon. A 1 mM concentration of the fragment (BsCM 88-127 R90X) with a N-terminal cysteine and a slight excess (10 to 20%) of the C-terminal thioester fragment (BsCM(1-87)) was used. After ligation, the disulfides in the mixture were reduced within 2 h by adding 30 mg TCEP per ml.

Computer Programs

ABI Prism 310 collection version 2, Edit View version 1.0, SeqEd version 1.0.3, Sequencing Analyzer version 3.3 and UV WinLab version 2.80.03, ABI Applied Biosystems & Perkin Elmer, Rotkreuz LU, Switzerland. Adobe Illustrator version 10.0, Photoshop version 7.0, Acrobat version 6.02, Adobe Systems Inc., San Jose CA, USA. Aviv CDS version 2.88, Aviv Instruments Inc., Lakewood NJ, USA. CS ChemDraw Pro(R) version 7.0 Pro and 8.0 Ultra, CambridgeSoft Corp., Cambridge MA, USA. Endnote Version 6.0, ISI ResearchSoft, Berkeley CA, USA. Hydropro 5.a (G. de la Torre, Lit: Biophys J. 78, 719-30). Insight II version 98.0, accelrys (formerly MSI Inc.), Princeton NJ, USA. KaleidaGraph version 3.5, Synergy Software, Reading PA, USA. Microsoft Windows 2000/XP and Microsoft Office XP/2003, Microsoft Cooperation, Seattle WA, USA. Origin 6.1, Origin Lab Corporation, Northampton MA, USA. Sedfit (321). Swiss PDB Viewer version 3.6b3 by GlaxoSmithKline R&D S.A., Swiss Institute for Bioinformatics, Geneva, Switzerland. Ultrascan 6.2 (323). Wisconsin Sequence Analysis Package version 8.0, Genetics Computer Group (GCG) Inc., Madison WI, USA.

6 EXPERIMENTAL SECTION 114

6.2 Peptide Synthesis

Fmoc-BsCM(108-127) 35

Fmoc-BsCM(108-127) 35 (YLEKAVVLRPDLSLTKNTEL) was synthesized on a 0.25 mmol scale using the FM025_MPPSTD3 chemistry. In order to avoid β-sheet formation the amino acids S120 and L119 were coupled as a pseudoproline dipeptide namely Fmoc-Leu-Ser(ΨMe,Mepro)-OH (Sp1, Figure 2.2). The synthesizer was loaded with 1.19 g NovaSyn TGA-resin preloaded with Fmoc-Leu (substitution: 0.21 mmol/g, Wang linker) and the synthesis was performed according to the program in Table 6.6. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing dDDD 2 Leu problematic single coupling bADEfh 3 Glu problematic single coupling bADEfh 4 Thr problematic single coupling bADEfh 5 Asn problematic double coupling bCADEgADEHfh 6 Lys problematic single coupling bADEfh 7 Thr problematic single coupling bADEfh 8 Leu problematic single coupling bADEfh 9 Sp1 problematic single coupling bADEfh 10 Asp problematic single coupling bADEfh 11 Pro problematic single coupling bADEfh 12 Arg problematic double coupling bCADEgADEHfh 13 Leu problematic double coupling bCADEgADEHfh 14 Val problematic double coupling bCADEgADEHfh 15 Val problematic double coupling bCADEgADEHfh 16 Ala problematic double coupling bCADEgADEHfh 17 Lys problematic double coupling bCADEgADEHfh 18 Glu problematic double coupling bCADEgADEHfh 19 Leu problematic double coupling bCADEgADEHfh 20 Tyr problematic double coupling bCADEgADEHfh 21 final peptide washing Dcc Table 6.6. Program used to synthesize Fmoc-BsCM(108-127) 35.

After drying, 1.833 g (expected: 2.053 g) of resin loaded with the Fmoc and side chain protected peptide was obtained. For analysis, 50 mg was side chain deprotected

6 EXPERIMENTAL SECTION 115 and cleaved from the resin to give 15 mg of crude peptide 35, the remaining resin was stored at -20 °C under argon.

CHARACTERIZATION:

ESI-MS (M, 2,524.98 Da): 2,524.5 Da

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 39.1 min (> 90% peak area)

BsCM(98-127) 36

BsCM(98-127) 36 (DVPQDQIRHVYLEKAVVLRPDLSLTKNTEL) was synthesized on a 0.08 mmol scale using the FM010_MPPSTD3 chemistry. The synthesizer was loaded with 0.77 g NovaSyn TGA-resin preloaded with protected BsCM(108-127) 35 (substitution: 0.10 mmol/g) and the coupling of the underlined AAs was performed according to the program in Table 6.7. For a description of the modules see the appendix.

6 EXPERIMENTAL SECTION 116

After drying, 0.8456 g (expected: 0.955 g) of resin loaded with the Fmoc and side chain protected peptide was obtained. For analysis, 50 mg was first treated with 2.5 ml 20% piperidine in DMF twice for 1.5 h to remove the Fmoc group, the side chains were deprotected, and the peptide was cleaved from the resin to give 10.5 mg of crude 36. The remaining resin was stored at -20 °C under argon. 5 mg of the peptide were purified by preparative HPLC (C8, 15/85 to 50/50 in 60 min) and the three main peaks at 37.2 min, 37.7 min and at 39.2 min were analyzed by ESI-MS.

CHARACTERIZATION:

Peak at 37.2 min (peak area 20%)

ESI-MS: (3,473 ± 1) Da (probably aspartimide at Asp102, calc.: 3,472.9 Da)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 28.6 min

Peak at 37.7 min (peak area 14%)

ESI-MS: (3,490 ± 2) Da (probably 36, could also be the reopened aspartimide,

calc.: 3,490.9 Da, smallest peak)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 28.6 min

Peak at 39.2 min (peak area 40%)

ESI-MS: (2,581 ± 1) Da (probably capped BsCM106-127, calc.: 2,580.9 Da)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 29.0 min

D102E BsCM(88-127) 9

D102E BsCM(88-127) 9 (CIRVMMTVQTDVPQEQIRHVYLEKAVVLRPD LSLTKNTEL) was synthesized on a 0.1 mmol scale using the FM010_MPPSTD4 chemistry. To overcome aspartimide formation, Asp102 was changed to Glu. The

6 EXPERIMENTAL SECTION 117 synthesizer was loaded with 0.82 g NovaSyn TGA-resin preloaded with protected BsCM(108-127) 35 (substitution: 0.122 mmol/g) and the coupling of the underlined AAs was performed according to the program in Table 6.8. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing dDDDD 2 Val problematic double coupling bCADEgADEIfh 3 His problematic double coupling bCADEgADEIfh 4 Arg problematic triple coupling 4 h bCADEIGADEIGADEIGh 5 Ile problematic double coupling bCADEgADEIfh 6 Gln problematic double coupling bCADEgADEIfh 7 Glu problematic double coupling bCADEgADEIfh 8 Gln problematic double coupling bCADEgADEIfh 9 Pro problematic double coupling bCADEgADEIfh 10 Val problematic double coupling bCADEgADEIfh 11 Asp problematic double coupling bCADEgADEIfh 12 Thr problematic double coupling bCADEgADEIfh 13 Gln problematic double coupling bCADEgADEIfh 14 Val problematic double coupling bCADEgADEIfh 15 Thr problematic double coupling bCADEgADEIfh 16 Met problematic double coupling bCADEgADEIfh 17 Met problematic double coupling bCADEgADEIfh 18 Val problematic double coupling bCADEgADEIfh 19 Arg problematic triple coupling 4 h bCADEIGADEIGADEIGh 20 Ile problematic double coupling bCADEgADEIfh 21 Cys problematic triple coupling 4h bCADEIGADEIGADEIGh 22 final peptide deprotection bCDcc Table 6.8. Program used to synthesize D102E BsCM(88-127) 9.

After drying, 1.124 g (expected: 1.213 g) of resin loaded with the side chain protected peptide was obtained. The entire sample was deprotected and cleaved from the resin to give 207 mg of crude 9. The peptide was purified by preparative HPLC (C8, 20/80 to 50/50 in 50 min) in several batches of 15 to 20 mg each, and the three main peaks at 27.6 min, 28.8 min and at 29.5 min were analyzed by ESI-MS. Approximately 25% (w/w) of the crude material could be obtained as pure 9 after HPLC.

CHARACTERIZATION:

Peak at 27.6 min (peak area 25%)

6 EXPERIMENTAL SECTION 118

ESI-MS: (2,580 ± 2) Da (probably capped BsCM 106-127, calc.: 2,580.9 Da)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 33.3 min

Peak at 28.8 min (peak area 47%)

ESI-MS: (4,668 ± 2) Da (9, largest peak, calc.: 4,668.4 Da)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 34.2 min

Peak at 29.5 min (peak area 20%)

ESI-MS: (4,583 ± 1) Da (could not be assigned)

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 35.3 min

Arg90Cit D102E BsCM(88-127) 10

Arg90Cit D102E BsCM(88-127) 10 (CICitVMMTVQTDVPQEQIRHVYLEK AVVLRPDLSLTKNTEL) was synthesized on a 0.1 mmol scale using the FM010_MPPSTD4 chemistry. Citrulline could be used without side chain protection as shown by the synthesis of 37. To avoid aspartimide formation Asp102 was changed to Glu. The synthesizer was loaded with 0.82 g NovaSyn TGA-resin preloaded with protected BsCM(108-127) 35 (substitution: 0.122 mmol/g) and the coupling of the underlined AAs was performed according to the program in Table 6.9. For a description of the modules see the appendix.

6 EXPERIMENTAL SECTION 119

9 Pro problematic double coupling bCADEgADEIfh 10 Val problematic double coupling bCADEgADEIfh 11 Asp problematic double coupling bCADEgADEIfh 12 Thr problematic double coupling bCADEgADEIfh 13 Gln problematic double coupling bCADEgADEIfh 14 Val problematic double coupling bCADEgADEIfh 15 Thr problematic double coupling bCADEgADEIfh 16 Met problematic double coupling bCADEgADEIfh 17 Met problematic double coupling bCADEgADEIfh 18 Val problematic double coupling bCADEgADEIfh 19 Cit problematic triple coupling 4 h bCADEIGADEIGADEIGh 20 Ile problematic double coupling bCADEgADEIfh 21 Cys problematic triple coupling 4 h bCADEIGADEIGADEIGh 22 final peptide deprotection bCDcc Table 6.9. Program used to synthesize Arg90Cit D102E BsCM(88-127) 10.

After drying, 1.079 g (expected: 1.213 g) of resin loaded with the side chain protected peptide was obtained. The entire sample was deprotected and cleaved from the resin to give 180 mg of crude 10. The peptide was purified by preparative HPLC

(C18, 20/80 to 50/50 in 50 min) in several batches of 15 to 20 mg each, and the main peak at 22.6 min was collected and analyzed by ESI-MS. Approximately 30% (w/w) of the crude material could be obtained as pure 10 after HPLC.

CHARACTERIZATION:

ESI-MS (M, 4,669.4 Da): (4,669.6 ± 2) Da

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 34.2 min

Arg90Cit BsCM(88-93) 37

Arg90Cit BsCM(88-93) 37 (CICitVMM) was synthesized on a 0.1 mmol scale using the FM010_MPPSTD4 chemistry. The synthesizer was loaded with 0.167 g NovaSyn Wang-resin preloaded with protected Met (substitution: 0.6 mmol/g) and the

6 EXPERIMENTAL SECTION 120 coupling of the underlined AAs was performed according to the program in Table 6.10. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing dDD 2 Met problematic double coupling bCADEgADEIfh 3 Val problematic double coupling bCADEgADEIfh 4 Cit problematic triple coupling 4 h bCADEIGADEIGADEIGh 5 Ile problematic double coupling bCADEgADEIfh 6 Cys problematic double coupling bCADEgADEIfh 7 final peptide deprotection bCDcc Table 6.10. Program used to synthesize Arg90Cit BsCM(88-93) 37.

After drying, 0.213 g (expected: 0.224 g) of resin loaded with the side chain protected peptide was obtained. The entire sample was deprotected and cleaved from the resin to give 84.3 mg of crude 37. The peptide was purified by preparative HPLC

(C8, 20/80 to 50/50 in 50 min) in several batches of 15 to 20 mg each, and the main peak at 19.1 min was collected and analyzed by ESI-MS. 53 mg (60%) of pure 37 could be obtained after HPLC.

CHARACTERIZATION:

ESI-MS (M, 753.0 Da): (752.5 ± 1) Da

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 24.8 min

BsCM(88-93) 11

The peptide BsCM(88-93) 11 has the sequence CIRVMM. It was synthesized on a 0.1 mmol scale using the FM010_MPPSTD4 chemistry. The synthesizer was loaded with 0.167 g NovaSyn Wang-resin preloaded with protected Met (substitution: 0.6 mmol/g) and the coupling of the underlined AAs was performed according to the

6 EXPERIMENTAL SECTION 121 program in Table 6.11. Arginine was used without side chain protection. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing d 2 Met problematic double coupling bCADEgADEIfh 3 Val problematic double coupling bCADEgADEIfh 4 Arg problematic single coupling bADEfh 5 Ile problematic double coupling bCADEgADEIfh 6 Cys problematic tripple coupling 4 h bCADEIGADEIGAFEIGh 7 final peptide deprotection bCDcc Table 6.11. Program used to synthesize BsCM(88-93) 11.

After drying, 0.168 g (expected: 0.239 g) of resin loaded with the side chain protected peptide was obtained. 27.9 mg of resin were deprotected; cleavage of the peptide from the resin gave 5 mg of crude material. The analysis of the sample showed that the desired peptide 11 was not obtained; only AcVMM (mass 422.1 Da) was found.

CHARACTERIZATION:

ESI-MS (M, 752.0 Da): no corresponding mass was found

Fmoc-D102E BsCM(92-127) 38

D102E BsCM(92-127) 38 (MMTVQTDVPQEQIRHVYLEKAVVLRPDLSL TKNTEL) was synthesized on a 0.25 mmol scale using the FM025_MPPSTD4 chemistry. To avoid aspartimide formation, Asp102 was changed to Glu. The synthesizer was loaded with 2.06 g NovaSyn TGA-resin preloaded with protected BsCM(108-127) 35 (substitution: 0.121 mmol/g) and the coupling of the underlined AAs was performed according to the program in Table 6.12. For a description of the modules see the appendix.

6 EXPERIMENTAL SECTION 122

No. AA Cycle Modules 1 first resin washing dDDDD 2 Val problematic double coupling bCADEgADEIfh 3 His problematic double coupling bCADEgADEIfh 4 Arg problematic triple coupling 4 h bCADEIGADEIGADEIGh 5 Ile problematic double coupling bCADEgADEIfh 6 Gln problematic double coupling bCADEgADEIfh 7 Glu problematic double coupling bCADEgADEIfh 8 Gln problematic double coupling bCADEgADEIfh 9 Pro problematic double coupling bCADEgADEIfh 10 Val problematic double coupling bCADEgADEIfh 11 Asp problematic double coupling bCADEgADEIfh 12 Thr problematic double coupling bCADEgADEIfh 13 Gln problematic double coupling bCADEgADEIfh 14 Val problematic double coupling bCADEgADEIfh 15 Thr problematic double coupling bCADEgADEIfh 16 Met problematic double coupling bCADEgADEIfh 17 Met problematic double coupling bCADEgADEIfh 22 final peptide washing Dcc Table 6.12. Program used to synthesize Fmoc-D102E BsCM(92-127) 38.

After drying of 64.6 mg of wet resin, 25.6 mg of dry resin loaded with the Fmoc and side chain protected peptide was obtained. 14 mg of it was side chain deprotected and cleaved from the resin to give 3.8 mg (30%) of crude Fmoc protected 38. The peptide was purified by preparative HPLC (C18, 5/95 to 80/20 in 60 min) and the main peak at 43.1 min was collected and analyzed by ESI-MS. Approximately 30% (w/w) of the crude material was obtained as pure Fmoc protected 38 after HPLC. The remaining loaded resin was stored at 4 °C (or -20 °C for prolonged storage) under N2 until used for further elongations.

CHARACTERIZATION:

ESI-MS (M, 4,419.0 Da): (4,419.5 ± 2) Da

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 39.1 min

6 EXPERIMENTAL SECTION 123

Arg90homo-Lys D102E BsCM(88-127) 12

Arg90homo-Lys D102E BsCM(88-127) 12 (CIhLysVMMTVQTDVPQEQIR HVYLEKAVVLRPDLSLTKNTEL) was synthesized on a 0.05 mmol scale using the FM010_MPPSTD4 chemistry. Homo-Lys was used with Boc as the side chain protecting group (RSP Amino Acid Analogues Inc., Hopkinton MA, USA). The synthesizer was charged with 0.62 g NovaSyn TGA-resin, preloaded with protected D102E BsCM(92-127) 38 (substitution: 0.08 mmol/g), and the coupling of the underlined AAs was performed according to the program in Table 6.13. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing d 18 Val problematic double coupling bCADEgADEIfh 19 hLys problematic double coupling bCADEgADEIfh 20 Ile problematic double coupling bCADEgADEIfh 21 Cys problematic triple coupling 4 h bCADEIGADEIGADEIGh 22 final peptide deprotection bCDcc Table 6.13. Program used to synthesize Arg90homo-Lys D102E BsCM(88-127) 12.

After drying, 0.625 g (expected: 0.645 g) of resin loaded with the side chain protected peptide was obtained. The entire sample was deprotected and cleaved from the resin to give 196 mg of crude 12. The peptide was purified by preparative HPLC

(C18, 20/80 to 50/50 in 50 min) in several batches of 15 to 20 mg each, and the main peak at 30.9 min was collected and analyzed by ESI-MS. After HPLC 47 mg of pure 12, corresponding to about 25% (w/w) of the crude material was obtained.

CHARACTERIZATION:

ESI-MS (M, 4,654.4 Da): (4,654.6 ± 2) Da

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 20.9 min

6 EXPERIMENTAL SECTION 124

Arg90F2Arg D102E BsCM(88-127) 31

Arg90F2Arg D102E BsCM(88-127) 31 (CIF2ArgVMMTVQTDVPQEQIRH VYLEKAVVLRPDLSLTKNTEL) was synthesized on a 0.25 mmol scale using the FM025_MPPSTD4a chemistry. Difluoro arginine was used with Pbf as the side chain protecting group. The synthesizer was charged with 2.9 g NovaSyn TGA-resin preloaded with protected D102E BsCM(92-127) 38 (substitution: 0.086 mmol/g), and the coupling of the underlined amino acids was performed according to the program in Table 6.14. For a description of the modules see the appendix.

No. AA Cycle Modules 1 first resin washing dDDDD 18 Val problematic double coupling bCADEgADEIfh 19 F2Arg F2Arg_aux_double coupling 4 h bCADEIaADEIah 20 Ile problematic double coupling bCADEgADEIfh 21 Cys problematic triple coupling 4 h bCADEIGADEIGADEIGh 22 final peptide deprotection bCDcc Table 6.14. Program used to synthesize Arg90F2Arg D102E BsCM(88-127) 31.

After drying, 2.308 g (expected: 3.090 g) of resin loaded with the side chain protected peptide was obtained. 1.3 g of the resin was deprotected, and the peptide was cleaved from the resin to give 423 mg of crude 31. The peptide was purified by preparative HPLC (C18, 20/80 to 50/50 in 50 min) in several batches of 20 to 25 mg each and the main peak at 30.9 min was collected and analyzed by ESI-MS. After HPLC 123 mg of pure 31, corresponding to about 30% (w/w) of the crude material, was obtained.

CHARACTERIZATION:

ESI-MS (M, 4,704.5 Da): (4,705.0 ± 2) Da (C204H343F2N57O61S3)

HiRes-ESI-MS (M+H, 4,701.462 Da): 4,701.474 Da (M+H)+, 4,723.455 Da

(M+Na)+

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 31.4 min

6 EXPERIMENTAL SECTION 125

6.3 Cloning and Expression

BsCM(1-87) encoded by pTYB1-BsCM87

The construction of pTYB1-BsCM87 encoding BsCM(1-87) fused to the Sce VMA intein and a CBD first required amplification of the the acceptor plasmid pTYB1

(NEB, Sce VMA intein) by transformation into KA12 (CaCl2, page 104). The vector was extracted (page 102) and digested using the restriction enzymes SapI and NdeI (page 103). The restriction digest was done with 38.5 µl plasmid solution (pTYB1, 10 ng/µl), 5 µl buffer 4 (NEB), 5 µl SapI and 1.5 µl NdeI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from the template pKET3-W, using the T7PRO2 forward and the AK-3-SAP-BsCM87 reverse primers. After purification, it was digested with the same restriction enzymes (10 µl insert solution, 31 µl UPW, 5 µl buffer 4 (NEB), 3 µl SapI and 1 µl NdeI at 37 °C for 3 h). Preparative agarose-gels (page 86 and 103) were used to purify the insert (2%) and acceptor (0.8%) fragments. The purified DNA fragments were ligated (page 103) by mixing 0.5 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were purified and transformed into electrocompetent KA12 cells (page 105, time constant 4.12). From single colony streak outs, 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), and the plasmid was extracted (page 102) and sequenced (page 107) using the forward primer T7PRO2 and the reverse primer AK-1-

RPSceI. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 110). The protein was produced as described under general procedures (page 110).

YIELD: 0.5-1 mg per 1 l of culture (after on column cleavage).

CHARACTERIZATION:

BSCM 1-87 NMMAA

6 EXPERIMENTAL SECTION 126

ESI-MS (M, 9,939.7 Da): (9,941 ± 2) Da

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 39.6 min

BsCM(1-87) encoded by pTXB1-BsCM87

The construction of pTXB1-BsCM87 encoding BsCM 1-87 fused to the Mxe GyrA intein and a CBD first required amplification of the acceptor plasmid pTXB1

(NEB) by transformation into KA12 (CaCl2, page 104). The vecxtor was extracted (page 102) and digested using the restriction enzymes SapI and NdeI (page 103). The restriction digest was done with 38.5 µl plasmid solution (pTXB1, 10 ng/µl), 5 µl buffer 4 (NEB), 5 µl SapI and 1.5 µl NdeI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the AK-3-SAP-BsCM87 reverse primers. After purification, it was digested with the same restriction enzymes (25 µl insert solution, 4 µl buffer 4 (NEB), 4 µl SapI and 1 µl NdeI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (2%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 5 µl insert solution, 3 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl

T4-ligase and 9 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), and the plasmid was extracted (page 102) and sequenced (page 107) using the forward primer T7PRO2 and the reverse primer AK-1-

RPMxeI. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 110). The protein was produced as described under general procedures (page 110).

YIELD: 3-5 mg per 1 l of culture (after on column cleavage and HPLC purification).

6 EXPERIMENTAL SECTION 127

CHARACTERIZATION:

BSCM 1-87 NMMAA

ESI-MS (M, 9,939.7 Da): (9,940 ± 2) Da contaminated with approximately ≥

40% 9,835 Da: probably lactam of the C-terminal lysine (same Rt on HPLC).

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 39.6 min

BSCM 1-87 MESNA

ESI-MS (M, 9,976 Da): (9,977 ± 1) Da contaminated with approximately 15%

9,835 Da: probably lactam of the C-terminal lysine (same Rt on HPLC).

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 39.3 min

D102E BsCM encoded by pAK-D102E

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the SVSF-1-D102E forward and the 04-T7TR reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the

6 EXPERIMENTAL SECTION 128

reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 51 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 14,503 Da): (14,505 ± 2) Da

R105K BsCM encoded by pAK-R105K

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the SVSF-2-R105K forward and the 04-T7TR reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

6 EXPERIMENTAL SECTION 129

YIELD: 117 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 14,461 Da): (14,464 ± 3) Da

D102E R105K BsCM encoded by pAK-DERK

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the SVSF-3-DERK forward and the 04-T7TR reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 40 mg per 1 l of culture.

6 EXPERIMENTAL SECTION 130

CHARACTERIZATION:

ESI-MS (M, 14,475 Da): (14,478 ± 3) Da

BsCM(1-124) encoded by pAK-T125*

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the SVSF-4-T125* reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 35 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 14,146 Da): (14,147 ± 2) Da

6 EXPERIMENTAL SECTION 131

BsCM(1-122) encoded by pAK-K123*

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the SVSF-5-K123* reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 142 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 13,903 Da): (13,906 ± 3) Da

6 EXPERIMENTAL SECTION 132

BsCM(1-121) encoded by pAK-T122*

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the SVSF-6-T122* reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 101 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 13,802 Da): (13,804 ± 3) Da

6 EXPERIMENTAL SECTION 133

BsCM(1-120) encoded by pAK-L121*

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (on page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the SVSF-7-L121* reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 40 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 13,689 Da): (13,691 ± 2) Da

6 EXPERIMENTAL SECTION 134

BsCM(1-119) encoded by pAK-S120*

The acceptor plasmid pKET3-W was extracted (page 102) from KA13/pKET3-W and digested using the restriction enzymes BsrGI and XhoI (page 103). The restriction was done with 45 µl plasmid solution (pKET3-W, 10 ng/µl), 5 µl buffer 2 (NEB), 6 µl 10 × BSA, 0.5 µl BsrGI and 0.5 µl XhoI at 37 °C for 2 h. The insert was amplified by PCR (page 106) from pKET3-W, using the T7PRO2 forward and the SVSF-8-S120* reverse primers. After purification, it was digested with the same restriction enzymes (30 µl insert solution, 2 µl UPW, 4 µl buffer 2 (NEB), 4 µl 10 × BSA, 2.5 µl BsrGI and 1.25 µl XhoI at 37 °C for 2 h). Preparative agarose gels (page 86 and 103) were used to purify the insert (3%) and acceptor (0.8%) fragments. The purified DNA fragments were then ligated (page 103) by mixing 1 µl insert solution, 10 µl acceptor fragment solution, 2 µl ligase-buffer, 1 µl T4-ligase and 7 µl UPW. The ligated plasmids were transformed into CaCl2 competent KA12 cells (page 104). From single colony streak outs 5 ml LB-Amp150 cultures were made (37 °C, 230 rpm, overnight), the plasmid was extracted (page 102) and sequenced (page 107) using the reverse primer 04-T7TR. Correct plasmids were transformed into CaCl2 competent KA13 cells (page 104) for expression of the gene (page 109). The protein was produced as described under general procedures (page 109).

YIELD: 8.2 mg per 1 l of culture.

CHARACTERIZATION:

ESI-MS (M, 13,602 Da): (13,603 ± 3) Da

6 EXPERIMENTAL SECTION 135

6.4 Native Chemical Ligation

D102E BsCM

As described under general procedures (page 112), 4 mg (0.4 mmol, 1 eq) of the fragment BsCM(1-87) (N-terminal thioester with MESNA) and 2.4 mg (0.5 mmol, 1.25 eq) of the synthetic fragment D102E BsCM(88-127) 9 were coupled by native chemical ligation. The ligated protein was folded (page 111) and purified on a Mono Q column (page 107). The fractions (#xx) that contained the desired protein as judged by SDS-PAGE were pooled and concentrated by ultrafiltration (MWCO = 10 kDa). The buffer was also changed to 50 mM potassium phosphate (pH 7.5) during the ultrafiltration process.

YIELD: 1.9 mg (#15, 35%)

CHARACTERIZATION:

ESI-MS (M, 14,503 Da): (14,502 ± 2) Da

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 40.2 min

Arg90Cit D102E BsCM

As described under general procedures (page 112), 17.8 mg (1.8 mmol, 1.05 eq) of the fragment BsCM(1-87) (N-terminal thioester with MESNA) and 8 mg (1.7 mmol, 1 eq) of the synthetic fragment Arg90Cit D102E BsCM(88-127) 10 were coupled by native chemical ligation. The ligated protein was folded (page 111) and purified on a Mono Q column (page 107). The fractions (#xx) that contained the desired protein as

6 EXPERIMENTAL SECTION 136 judged by SDS-PAGE were pooled and concentrated by ultrafiltration (MWCO = 10 kDa). The buffer was also changed to 50 mM potassium phosphate (pH 7.5) during the ultrafiltration process.

YIELD: 7 mg (#13, 28%)

CHARACTERIZATION:

ESI-MS (M, 14,504 Da): (14,503 ± 2) Da

RP-HPLC (C18, 0/100 to 60/40 in 45 min): Rt = 40.5 min

Arg90homo-Lys D102E BsCM

As described under general procedures (page 112), 12.0 mg (1.2 mmol, 1.13 eq) of the fragment BsCM(1-87) (N-terminal thioester with MESNA) and 5 mg (1.06 mmol, 1 eq) of the synthetic fragment Arg90homo-Lys D102E BsCM(88-127) 12 were coupled by native chemical ligation. The ligated protein was folded (page 111) and purified on a Mono Q column (page 107). The fractions (#xx) that contained the desired protein as judged by SDS-PAGE were pooled and concentrated by ultrafiltration (MWCO = 10 kDa). The buffer was simultaneously changed to 50 mM potassium phosphate (pH 7.5).

YIELD: 5.2 mg (#13, #14, #15, 35%)

CHARACTERIZATION:

ESI-MS (M, 14,489 Da): (14,489 ± 2) Da

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 39.0 min

6 EXPERIMENTAL SECTION 137

Arg90F2Arg D102E BsCM

As described under general procedures (page 112), 21.7 mg (2.2 mmol, 1 eq) of the fragment BsCM(1-87) (N-terminal thioester with MESNA) and 10.4 mg (2.2 mmol,

1 eq) of the synthetic fragment Arg90F2Arg D102E BsCM(88-127) 31 were coupled by native chemical ligation. The ligated protein was folded (page 111) and purified on a Mono Q column (page 107). The fractions (#xx) that contained the desired protein as judged by SDS-PAGE were pooled and concentrated by ultrafiltration (MWCO = 10 kDa). The buffer was also changed to 50 mM potassium phosphate (pH 7.5) during the concentration.

YIELD: 2.1 mg (#12, #13, #14, #15, 10%)

CHARACTERIZATION:

ESI-MS (C635H1052F2N174O193S9, 14,539 Da): (14,540 ± 2) Da

RP-HPLC (C18 short, 5/95 to 60/40 in 45 min): Rt = 34.9 min

19 F-NMR (376.5 MHz, H2O/10% D2O ): δ = -111.4 (dm, JFF = 226 Hz, Fa),

-101.7 (dm, JFF = 226 Hz, Fb)

6.5 Isotopically Labeled BsCM

15N-labeled BsCM

To prepare the enriched minimal medium used for expressing 15N-labeled BsCM 22 ml 20% glucose, 20 ml 10 × Bioexpress YBN (Cambridge Isotope Labs), 100 ml

6 EXPERIMENTAL SECTION 138

15 10 × M9 salts (with NH4Cl), 0.245 g MgSO4·2H2O, 0.015 g CaCl2·2H2O, 1 ml thiamin HCl (5 mg/ml), 1 ml 1000 × Aro-P-Mix and 0.85 ml ampicillin (100 mg/ml) were mixed, diluted with water to 1 l, and filter sterilized. The medium was inoculated with 2 ml of an overnight preculture of KA13/pKET3-W and grown at 30 °C and

250 rpm until an OD600 of 0.7 was reached (approximately 9 h). Gene expression was induced by adding 10 ml of a 50 mM IPTG solution, and the sample was incubated for another 24 h under the same conditions. The cells were harvested and the protein was purified as described on page 109 under the general procedure for BsCM variants. The osmolysis step was done twice. For NMR measurements the buffer was changed to 25 mM potassium phosphate pH 7.5 by dialysis (MWCO = 10 kDa).

YIELD: 69 mg

CHARACTERIZATION:

ESI-MS (M, 14,664 Da): (14,661 ± 2) Da

D (< 35%), 13C-, and 15N-labeled BsCM

To prepare the enriched minimal medium used for expressing D (< 35%), 13C- and 15N-labeled BsCM 70 ml 10 × Silantes D(<35%)13C15N (Silantes GmbH, Munich, 15 Germany), 50 ml 10 × M9 salts (with NH4Cl), 0.123 g MgSO4·2H2O, 0.0075 g

CaCl2·2H2O, 0.5 ml thiamin HCl (5 mg/ml), 0.5 ml 1000 × Aro-P-Mix and 0.75 ml ampicillin (100 mg/ml) were mixed, diluted with water to 0.5 l, and filter sterilized. The medium was inoculated with 1 ml of an overnight preculture of KA13/pKET3-W and grown at 30 °C and 250 rpm until an OD600 of 1.2 was reached (approximately 10 h). Gene expression was induced by adding 5 ml of a 50 mM IPTG solution. After 1 h, 750 µl of a rifampicin solution (0.1 mg/ml) was added and then the culture was incubated for another 24 h under the same conditions. The cells were harvested and the

6 EXPERIMENTAL SECTION 139 protein was purified as described on page 109 under the general procedure for BsCM variants. The osmolysis step was done twice. For NMR measurements the buffer was changed to 25 mM potassium phosphate (pH 7.5) by dialysis (MWCO = 10 kDa).

YIELD: 24.8 mg

D, 13C-, and 15N-labeled BsCM

In order to grow in deuterated medium (> 98%) cells need to be adapted slowly. A 5 ml LB medium (150 µg/ml Amp) preculture was inoculated with a freshly grown single colony of KA13/pKET3-W and incubated overnight at 230 rpm 37 °C. A filter sterilized second preculture of LB medium (150 µg/ml Amp) in 5 ml of a 1:1 mixture of

D2O and H2O was inoculated with 100 µl of the freshly grown first preculture and incubated overnight at 230 rpm 37 °C. A third preculture of 6 ml filter sterilized 100% deuterated Celton medium (150 µg/ml Amp, Martek Biosciences Corporation, now Spectra Gases) was inoculated with 200 µl of the second preculture. All precultures were incubated for 24 h at 230 rpm and 37 °C. An OD600 of 3.0 was reached. To prepare the enriched minimal medium used for expressing D, 13C- and 15N-labeled BsCM 5.5 g Celton dCN (Martek Biosciences Corporation, now Spectra Gases), 150 ml 10 × M9 15 salts (with NH4Cl, in D2O), 0.368 g MgSO4·2H2O, 0.022 g·CaCl2 2H2O, 1.5 ml thiamin HCl (5 mg/ml), 1.5 ml 1000 × Aro-P-Mix and 1.3 ml ampicillin (100 mg/ml) were mixed, filled up with D2O to 1.5 l and filter sterilized. The medium was divided into six portions of 250 ml and each one was inoculated with 1 ml of the third preculture of KA13/pKET3-W and grown at 30 °C and 250 rpm until an OD600 of 0.9 was reached (approximately 28 h). Protein production was induced by adding 2.5 ml of a 50 mM IPTG solution to each flask. After 1 h, 375 µl of a rifampicin solution (0.1 mg/ml) was added and then the cultures were incubated for another 27 h under the same conditions. The cells were harvested and the protein was purified as described on page 109 under the general procedure for BsCM variants. The osmolysis step was done

6 EXPERIMENTAL SECTION 140 twice. For NMR measurements the buffer was changed to 25 mM potassium phosphate (pH 7.5) by dialysis (MWCO = 10 kDa).

YIELD: 42.5 mg

D (50%), 13C-, and 15N-labeled BsCM

To two sterile 2 l Erlenmeyer flasks each containing 500 ml liquid Celton d(50)CN medium (Martek Biosciences Corporation, now Spectra Gases) was added 800 µl of a 100 mg/ml solution of Ampicillin, 750 µl 1000 × Aro-P-Mix and 750 µl of a 5 mg/ml thiamin-HCl solution. Each flask was inoculated with 1 ml of a freshly grown overnight (230 rpm 37 °C) preculture of KA13/pKET3-W. It was incubated for 13 h

(OD600 = 1.2) at 250 rpm 30 °C. Gene expression was induced by adding 5 ml of a 50 mM IPTG solution to each flask and the cultures were then incubated for another 35 h under the same conditions. The cells were harvested and the protein was purified as described on page 109 under the general procedure for BsCM variants. The osmolysis step was done twice. For NMR measurements the buffer was changed to 25 mM potassium phosphate (pH 7.5) by dialysis (MWCO = 10 kDa). MS showed that in addition to all 174 nitrogens and 635 carbons about 400 hydrogens were isotopically labeled. Considering the total number of exchangeable hydrogens in BsCM this corresponds to about 50%.

YIELD: 76.4 mg

CHARACTERIZATION:

ESI-MS (C635H1052F2N174O193S9, 14,489 Da (unlabeled)): (15,683 ± 4) Da

6 EXPERIMENTAL SECTION 141

6.6 Synthesis of Fmoc-4,4-difluoro-L-arginine(Pbf)-OH

N-Boc-D-Serinemethylester 14 was purchased from Aldrich and PbfCl was purchased from Lancaster (Eastgate, White Lund, Morecambe, England). Compounds 3-(tert-butyl) 4-methyl (4R)-2,2-dimethyl-1,3-oxazolane-3,4-dicarboxylate 15 (222- 230), tert-butyl (4S)-4-(hydroxymethyl)-2,2-dimethyl-1,3-oxazolane-3-carboxylate 16 (225, 228, 230-233), tert-butyl (4R)-4-formyl-2,2-dimethyl-1,3-oxazolane-3- carboxylate ((R)-Garner,s aldehyde) 17 (222-225, 227, 228, 231-233, 236, 238, 346), and N,N,-bis-tert-butoxycarbonylthiourea 38 (251, 254, 255) were synthesized according to published procedures. TLC, 1H-NMR spectra, melting points for solids, 20 and [α] D values for optically active compounds were found to be identical with published values. The synthesis of 13 followed the general strategy of Kim et al. (219, 236) up to 8.

Synthesis of tert-butyl (4R)-4-(3-ethoxy-2,2-difluoro-1-hydroxy-3- oxopropyl)-2,2-dimethyl-1,3-oxazolane-3-carboxylate 18:

OH O Reformatsky reaction of 17 with ethyl bromodifluoroacetate was performed in analogy to O literature procedures for similar compounds (234, O F F N 235, 237) and communications about this or similar O compounds (219, 236, 238-240). To a mixture of zinc O powder (12.54 g, 191.8 mmol, 3 eq) that was stored under N2 and 17 (14.66 g, 63.93 mmol, 1 eq) was added THF (350 ml) and ethyl bromodifluoroacetate (11.47 ml, 89.5 mmol, 1.4 eq) at r.t. under an argon atmosphere. The mixture was sonicated in a water bath for 15 h at 25 to 50 °C. The reaction was quenched with 150 ml sat. NH4Cl (aq) and, after 10 min, extracted twice with 150 ml

EA. The organic layer was washed with 100 ml sat. NaHSO3 (aq) and 100 ml water, separated, dried over anhydrous Na2SO4, and the solvent was evaporated under reduced

6 EXPERIMENTAL SECTION 142 pressure to give a colorless oil. The product was pure enough to be used for the next step. For characterization, a small portion was purified by silica gel column chromatography using DCM:EA 16:1 as eluent. NMR spectra were in agreement with unpublished data shared by P. Meffre from his communication (238).

YIELD: 19.20 g (85% consisting of two diastereomers in a ratio of 9:1)

CHARACTERIZATION:

ESI-MS (M+H, 354.18 Da): 354.2 Da (M+H)+

major diastereomer:

1 H-NMR (300 MHz, CDCl3): δ = 1.36 (t, J = 7.2 Hz, 3 H, OCH2CH3), 1.48 (s,

9 H, C(CH3)3), 1.58 and 1.61 (s, 6 H, C(CH3)2), 4.00 (m, 1 H, CHN), 4.20-4.53

(m, 5 H, CH2O, CHOH, OCH2CH3)

1 H-NMR (300 MHz, CDCl3, 60 °C): δ = 1.36 (t, J = 7.2 Hz, 3 H, OCH2CH3),

1.49 (s, 9 H, C(CH3)3), 1.50 and 1.60 (s, 6 H, C(CH3)2), 4.00 (dd, J = 6.3 Hz and

9.6 Hz, 1 H, CHN), 4.25 (d, J = 9.6 Hz, 2 H, , CH2O), 4.36 (q, J = 7.2 Hz, 2 H,

OCH2CH3), 4.48 (ddd, J = 6.3 Hz and JFH = 20 Hz and 1.2 Hz, 1 H, CHOH,)

13 C-NMR (75 MHz, CDCl3): δ = 14.1 (OCH2CH3), 22.9, 24.4, 26.4 and 26.7

(C(CH3)2), 28.6 (C(CH3)3), 56.0 and 57.3 (CH2O), 63.4 (OCH2CH3), 63.6 and

63.8 (CHN), 71.2 and 72.2 (2 × t, JFC = 26 Hz and 23 Hz, CHOHCF2), 81.0 and

81.5 (C(CH3)3), 94.0 (C(CH3)2), 114.7 (dd, JFC = 252 Hz and 259 Hz, CF2),

151.6 and 153.2 (NCOO), 163.4 (t, JFC = 27 Hz, CF2CO)

19 F-NMR (282.36 MHz, CDCl3): δ = -125.2 (dd, JFF = 263 Hz and JFH = 20 Hz,

st nd 1 rotamer of Fa), -122.8 (dd, JFF = 263 Hz and JFH = 20 Hz, 2 rotamer of Fa),

st -112.8 (dd, JFF = 263 Hz and JFH = 1.2 Hz, 1 rotamer of Fb), -112.5 (dd, JFF =

nd 263 Hz and JFH = 1.2 Hz, 2 rotamer of Fb)

6 EXPERIMENTAL SECTION 143

19 st F-NMR (282.36 MHz, CDCl3, 60 °C): δ = -124.4 (br d, JFF = 270 Hz, 1

nd rotamer of Fa), -122.0 (br d, JFF = 270 Hz, 2 rotamer of Fa), -112.7 (d, JFF =

270 Hz Fb)

20 [α] D (c 1.29, CHCl3): +56.5°

Rf : 0.31 (hexane:EA 4:1; KMnO4)

minor diastereomer:

1 H-NMR (300 MHz, CDCl3): δ = 1.36 (t, J = 7.1 Hz, 3 H, OCH2CH3), 1.48 (s,

9 H, C(CH3)3), 1.59 (s, 6 H, C(CH3)2), 3.9-4.1 (m, 3 H, CHN, CH2O), 4.26-4.38

(m, 3 H, CHOH, OCH2CH3)

13 C-NMR (75 MHz, CDCl3): δ = 13.7 (OCH2CH3), 23.1, 23.8, 26.7, and 27.0

(C(CH3)2), 28.1 (C(CH3)3), 56.5 (CH2O), 62.8 (OCH2CH3), 64.7 (CHN), 73.6 (t,

JFC = 25 Hz, CHOHCF2), 82.1 (C(CH3)3), 93.6 (C(CH3)2), 114.8 (t, JFC =

254 Hz, CF2), 155.9 (NCOO), 162.8 (t, JFC = 33 Hz, CF2CO)

19 F-NMR (282.36 MHz, CDCl3): δ = -126.0 (dd, JFF = 260 Hz and JFH = 15 Hz,

Fa), -109.4 (d, JFF = 260 Hz Fb)

20 [α] D (c 1.20, CHCl3): +29.4°

Rf : 0.42 (hexane:EA 4:1; KMnO4)

Synthesis of tert-butyl (4R)-4-3-ethoxy-2,2-difluoro-1-[(1H-imidazol-1- ylcarbothioyl)oxy]-3-oxopropyl-2,2-dimethyl-1,3-oxazolane-3- carboxylate 19:

6 EXPERIMENTAL SECTION 144

N Thiocarbonylation of 18 with thiocarbonyl

N diimidazole was performed in analogy to literature

S procedures for similar compounds (235-237, 241) and O O communications about this compound (219, 238). To a O mixture of thiocarbonyl diimidazole (25.2 g, 141 mmol, O F N F 2 eq) and 18 (25.0 g, 70.7 mmol, 1 eq) was added dry 1,2- O dichloroethane (250 ml) at r.t. under argon atmosphere. The O mixture was stirred for 20 h and then directly loaded onto a silica gel column using hexane:EA 1:1 as eluent. The product was obtained as a pale yellow oil. NMR spectra were in agreement with unpublished data shared by P. Meffre from his communication (238).

YIELD: 25.3 g (79% only the major diastereomer was isolated)

CHARACTERIZATION:

ESI-MS (M+H, 464.17 Da): 464.2 Da (M+H)+

major diastereomer:

1 H-NMR (300 MHz, CDCl3): δ = 1.25 (m, 3 H, OCH2CH3), 1.26, 1.31, 1.40

and 1.43 (s, 6 H, C(CH3)2), 1.48 and 1.54 (s, 9 H, C(CH3)3), 4.1 (m, 1 H, CHN),

4.24-4.52 (m, 4 H, CH2O, OCH2CH3), 6.68 (dd, JFH = 19.5 Hz and JFH = 6.5 Hz,

st nd 1 H, 1 rotamer of CHOCS), 6.82 (dd, JFH = 16.5 Hz and JFH = 9.0 Hz, 1 H, 2

rotamer of CHOCS), 7.06 (m, 1 H, =NCH), 7.63 (br s, 1 H, NCH=C), 8.35 (br s,

1 H, NCH=N)

13 C-NMR (75 MHz, CDCl3): δ = 14.0 (OCH2CH3), 22.6, 24.0, 26.1 and 26.9

(C(CH3)2), 28.6 (C(CH3)3), 55.5, 56.0 (CH2O), 63.2, 63.6 (CHN), 64.2

(OCH2CH3), 76.5-78.1 (m, CHOCF2), 81.6 (C(CH3)3), 93.7 and 94.2 (C(CH3)2),

112.4 and 112.6 (dd, JFC = 260 Hz and 250 Hz, CF2), 118.5 (=NCH), 131.5

(NCH=C), 137.7 (NCH=N), 151.2 and 152.3 (NCOO), 161.8 (t, JFC = 32 Hz,

6 EXPERIMENTAL SECTION 145

CF2CO), 183.5 (C=S)

19 F-NMR (282.36 MHz, CDCl3): δ = -118.4 (dd, JFF = 267 Hz and JFH =

st nd 19.5 Hz, 1 rotamer of Fa), -116.6 (dd, JFF = 269 Hz and JFH = 16.5 Hz, 2

nd rotamer of Fa), -112.0 (dd, JFF = 269 Hz and JFH = 9 Hz, 2 rotamer of Fb),

st -111.0 (dd, JFF = 267 Hz and JFH = 6.5 Hz, 1 rotamer of Fb)

20 [α] D (c 1.10, CHCl3): +61.5°

Rf : 0.17 (hexane:EA 4:1; KMnO4)

Synthesis of tert-butyl (4S)-4-(3-ethoxy-2,2-difluoro-3-oxopropyl)-2,2- dimethyl-1,3-oxazolane-3-carboxylate 20:

O Barton-McCombie radical deoxygenation of 19 with

O Et3SiH and benzoyl peroxide was performed according to O F literature procedures for similar compounds (235-237, 241- N F O 243) and communications about this compound (219, 238). O To a mixture of benzoyl peroxide (38.9 g, 160.5 mmol,

3 eq) and 19 (24.8 g, 53.5 mmol, 1 eq) was added Et3SiH (130 ml, 805 mmol, 15 eq) at r.t. under nitrogen atmosphere. The mixture was refluxed for 1 h and then quenched by the addition of 300 ml sat. NaHCO3 (aq) (effervescence). The solution was extracted twice with 150 ml EA. The organic layer was washed with 50 ml sat. NaHCO3 (aq), separated, and dried over anhydrous Na2SO4. The solvent was evaporated under reduced pressure and the resulting residue was purified on a silica gel column using hexane:EA 4:1 as eluent. The product was obtained as a colorless oil that was sufficiently clean to be used for the next step, but still contained some Et3SiH. For characterisation, a small sample was purified three times by silica gel chromatography to fully remove the Et3SiH. NMR spectra were in agreement with unpublished data shared by P. Meffre for his communication (238).

6 EXPERIMENTAL SECTION 146

YIELD: 16.2 g (90%)

CHARACTERIZATION:

ESI-MS (M+H, 338.18 Da): 338.2 Da (M+H)+

1 H-NMR (300 MHz, CDCl3): δ = 1.36 (t, 3 H, J = 7.2 Hz, OCH2CH3), 1.47,

1.49, 1.54, and 1.58 (s, 6 H, C(CH3)2), 1.47 (br s, 9 H, C(CH3)3), 2.15-2.75 (m,

2 H, CH2CF2), 3.88-4.02 (m, 2 H, CH2O), 4.12-4.26 (br m, 1 H, CHN), 4.33 and

4.34 (q, J = 7.2 Hz, 2 H, OCH2CH3)

13 C-NMR (75 MHz, CDCl3): δ = 13.8 (OCH2CH3), 22.9, 24.3, 26.7 and 27.4

(C(CH3)2), 28.3 (C(CH3)3), 37.6 (m, CH2CF2), 51.7 (CHN), 62.9 (OCH2CH3),

67.4 (CH2O), 80.2 (C(CH3)3), 92.9 and 93.3 (C(CH3)2), 114.8 (t, JFC = 253 Hz,

CF2), 151.0 and 151.7 (NCOO), 163.5 (t, JFC = 32 Hz, CF2CO)

19 F-NMR (282.36 MHz, CDCl3): δ = -107.2 (ddd, JFF = 262 Hz, JFH = 22 Hz

st and JFH = 15 Hz, 1 rotamer of Fa), -106.6 (ddd, JFF = 240 Hz, JFH = 22 Hz and

nd JFH = 15 Hz, 2 rotamer of Fa), -105.7 (ddd, JFF = 240 Hz, JFH = 23 Hz and JFH

nd = 11 Hz, 2 rotamer of Fb), -104.1 (ddd, JFF = 262 Hz, JFH = 23 Hz and JFH =

st 11 Hz, 1 rotamer of Fb)

20 [α] D (c 1.16, CHCl3): +23.5°

Rf : 0.84 (hexane:EA 1:1; KMnO4)

6 EXPERIMENTAL SECTION 147

Synthesis of tert-butyl (4S)-4-(3-amino-2,2-difluoro-3-oxopropyl)-2,2- dimethyl-1,3-oxazolane-3-carboxylate 21:

O Conversion of the ethyl ester 20 with NH3 to the amide

NH2 was performed in analogy published procedures with similar O F N F compounds (234, 236) and a communication about this O compound (219). A solution of 20 (17.0 g, 50.4 mmol, 1 eq) in O anhydrous Et2O was cooled to -78 °C in an acetone dry ice bath.

NH3 was condensed into it by bubbling the dry gas through the solution for 30 min. The solution was stirred for another 30 min at -78 °C and then the temperature was allowed to rise to r.t.. The solvent and the ethanol formed during the reaction were evaporated under reduced pressure and the resulting residue was purified on a silica gel column using hexane:EA 4:1 as eluent. The product was obtained as a white solid. Alternatively, it was purified by trituration with hexane:EA 2:1 and the brown solution was separated from the white solid by filtration.

YIELD: 11.2 g (72%)

CHARACTERIZATION:

ESI-MS (M+Na, 331.14 Da): 331.2 Da (M+Na)+

1 H-NMR (300 MHz, CDCl3): δ = 1.48 (br s, 9 H, C(CH3)3), 1.48, 1.54, 1.58,

and 1.62 (s, 6 H, C(CH3)2), 2.2-2.7 (m, 2 H, CH2CF2), 3.9-4.0 (m, 2 H, CH2O),

4.1-4.3 (br m, 1 H, CHN), 5.80, 6.26, and 6.73 (br s, 2 H, NH2)

13 C-NMR (75 MHz, CDCl3): δ = 23.3, 24.6, 27.0 and 27.8 (C(CH3)2), 28.6

(C(CH3)3), 36.8 and 37.6 (t, JFC = 21.5 Hz, CH2CF2), 52.1 and 52.4 (CHN), 67.7

(CH2O), 80.5 and 81.0 (C(CH3)3), 93.5 and 93.7 (C(CH3)2), 116.8 (t, JFC =

251 Hz, CF2), 151.5 and 152.4 (NCOO), 166.3 (t, JFC = 29.5 Hz, CF2CO)

19 F-NMR (282.36 MHz, CDCl3): δ = -106.9 (ddd, JFF = 260 Hz, JFH = 22.3 Hz

6 EXPERIMENTAL SECTION 148

st and JFH = 16 Hz, 1 rotamer of Fa), -106.0 (ddd, JFF = 261.5 Hz, JFH = 22.3 Hz

nd and JFH = 16 Hz, 2 rotamer of Fa), -105.2 (ddd, JFF = 260 Hz, JFH = 23 Hz and

st JFH = 13.8 Hz, 1 rotamer of Fb), -104.2 (ddd, JFF = 261.5 Hz, JFH = 23 Hz and

nd JFH = 13.8 Hz, 2 rotamer of Fb)

20 [α] D (c 1.40, CHCl3): +28.4°

Mp. : 120-124 °C

Rf : 0.62 (hexane:EA 1:1; I2 or KMnO4)

Synthesis of tert-butyl (4S)-4-(3-amino-2,2-difluoropropyl)-2,2- dimethyl-1,3-oxazolane-3-carboxylate 22:

Conversion of the amide 21 to the amine with sodium NH2 O F bis(2-methoxyethoxy)aluminum hydride (Red Al) was N F O performed in analogy to published procedures for similar O compounds (244-247) and a communication about this compound (219). A solution of 21 (14.1 g, 45.7 mmol, 1 eq) in dry toluene under nitrogen was cooled to 0 °C in an ice bath. A 3.5 mol/l solution of sodium bis(2- methoxyethoxy)aluminum hydride (39.2 ml, 137.2 mmol, 3 eq) in dry toluene was added slowly (effervescence). Upon complete addition, the mixture was stirred for 30 min at this temperature before stirring another 3 h at r.t.. The reaction was worked up by adding 300 ml sat. KNa-tartrate (aq) to complex the aluminum and 100 ml Et2O.

After stirring for 1 h the organic layer was separated, dried (Na2SO4), and the solvent was evaporated under reduced pressure. Attempts to purify the product on a silica gel column using hexane:EA 1:1 as eluent resulted in partial decomposition of the product. Since it was rather clean as judged by NMR, it was used for the next step without further purification and only partially characterized.

6 EXPERIMENTAL SECTION 149

YIELD: 14.1 g (crude)

CHARACTERIZATION:

ESI-MS (M+H, 295.19 Da): 295.2 Da (M+H)+, 239.2 (M - =<)+

1 H-NMR (300 MHz, CDCl3): δ = 1.48 (br s, 9 H, C(CH3)3), 1.54 and 1.59 (s,

6 H, C(CH3)2), 2.0-2.6 (m, 2 H, CH2CF2), 2.9-3.1 (m, 2 H, CH2NH2), 3.9-4.05

(m, 2 H, CH2O), 4.15-4.2 (br m, 1 H, CHN)

13 C-NMR (75 MHz, CDCl3): δ = 23.3, 24.6, 27.1 and 27.8 (C(CH3)2), 28.6

(C(CH3)3), 36.8 and 38.0 (t, JFC = 23 Hz, CH2CF2), 47.4 and 47.8 (t, JFC =

30 Hz, CF2CH2NH2), 52.5 and 52.9 (CHN), 67.7 and 68.0 (CH2O), 80.3 and

80.7 (C(CH3)3), 93.2 and 93.6 (C(CH3)2), 123.2 and 123.6 (t, JFC = 240 Hz,

CF2), 151.5 and 152.2 (NCOO)

19 F-NMR (282.36 MHz, CDCl3, phenotype !): δ = -109.9 - -109.6 (m), -109.0 -

-108.6 (m), -108.0 - -107.7 (m), -107.6 - -107.25 (m), -106.7 - -106.3 (m), -104.2

- -103.9 (m), -103.3 - -103.0 (m) (2 F as 2 rotamers, all 4 peaks as ddddd)

Rf : 0.16 (hexane:EA 1:1; Ninhydrin)

Synthesis of tert-butyl (4S)-4-(3-[(benzyloxy)carbonyl]amino-2,2- difluoropropyl)-2,2-dimethyl-1,3-oxazolane-3-carboxylate 23:

O Protection of the amine 22 with CbzCl was performed in analogy to a published N O H O F procedure with similar compounds (236) and a N F O communication about this compound (219). To

6 EXPERIMENTAL SECTION 150

a stirred solution of 22 (13.5 g, 45.9 mmol, 1 eq) in 150 ml EA was added NaHCO3. After cooling to 0 °C a solution of CbzCl in 15 ml EA was added drop wise. The solution was stirred for another 45 min at 0 °C. The reaction was worked up by adding

100 ml sat. NaHCO3 (aq) and extracting twice with 100 ml EA each. The organic phase was dried over Na2SO4, and the solvent was evaporated under reduced pressure. The resulting residue was purified on a silica gel column using hexane:EA 3:1 as eluent. The product was obtained as a colorless oil.

YIELD: 16.9 g (86%)

CHARACTERIZATION:

HiRes-ESI-MS (M+Na, 451.2021 Da): 451.2011 Da (M+Na)+, 467.1727 Da

(M+K)+

1 H-NMR (300 MHz, CDCl3): δ = 1.45 and 1.48 (br s, 9 H, C(CH3)3), 1.4 - 1.5,

1.53 and 1.57 (s, 6 H, C(CH3)2), 2.05-2.5 (m, 2 H, CH2CF2), 3.45 – 3.7 (m, 2 H,

CH2NHCbz), 3.85-4.0 (m, 2 H, CH2O), 4.1-4.2 (br m, 1 H, CHN), 5.13 (s, 2 H,

OCH2Ph), 5.1 and 5.7 (br s, 1 H, NHCbz), 7.35 (br s, 5 H, Ph)

13 C-NMR (75 MHz, CDCl3): δ = 23.3, 24.5, 27.0 and 27.7 (C(CH3)2), 28.6

(C(CH3)3), 37.4 and 37.8 (t, JFC = 23.1 Hz, CH2CF2), 45.4 and 46.2 (t, JFC =

30.4 Hz, CF2CH2NHCbz), 52.2 and 52.7 (CHN), 67.2 and 67.6 (CH2OH), 67.9

(OCH2Ph), 80.3 and 81.0 (C(CH3)3), 93.3 and 93.6 (C(CH3)2), 122.0 and 122.4

(t, JFC = 242 Hz, CF2), 128.2, 128.4, 128.6, and 128.7 (o, m ,p-Ph), 136.2 and

136.5 (i-Ph), 151.5 and 152.4 (NCOO-tBu), 156.6 (CH2NHCOOCH2Ph)

19 F-NMR (282.36 MHz, CDCl3): δ = -106.9, -106.3, -104.6, and -100.5 (br d,

JFF = 250 Hz, CF2)

20 [α] D (c 1.04, CHCl3): +19.4°

Rf : 0.72 (hexane:EA 1:1; KMnO4)

6 EXPERIMENTAL SECTION 151

Synthesis of benzyl N-(4S)-4-[(tert-butoxycarbonyl)amino]-2,2- difluoro-5-hydroxypentylcarbamate 24:

O Ring opening of the 1,3-oxazolane 23

N O with TsOH was performed in according to H HO F HN F literature procedures for similar compounds O (236, 248) and a communication about this O compound (219). To a stirred solution of 23 (16.0 g, 37.3 mmol, 1 eq) in 200 ml methanol was added TsOH (213 mg, 1.12 mmol, 0.03 eq). The flask was equipped with a reflux condenser and the reaction stirred for 6 h at 50 °C. The reaction was worked up by adding 100 ml water and 20 ml sat. NaHCO3

(aq) and extracting twice with 100 ml Et2O each. The combined organic phases were washed with sat. NaCl (aq) and dried over Na2SO4. The solvent was evaporated under reduced pressure. The resulting white solid was pure enough to be used for the next step.

YIELD: 13.9 g (96%)

CHARACTERIZATION:

ESI-MS (M+Na, 411.17 Da): 411.2 Da (M+Na)+, 427.2 Da (M+K)+

1 H-NMR (300 MHz, CDCl3): δ = 1.43 (s, 9 H, C(CH3)3), 2.0-2.2 (m, 2 H,

CH2CF2), 3.5 – 3.7 (m, 2 H, CH2NHCbz), 3.63 (d, J = 4.2 Hz, 2 H, CH2OH),

3.85-4.0 (br m, 1 H, CHN), 5.11 (s, 2 H, OCH2Ph), 5.2 and 5.46 (br s, 2 H, NH),

7.34 (br s, 5 H, Ph)

13 C-NMR (75 MHz, CDCl3): δ = 28.5 (C(CH3)3), 35.3 (t, JFC = 20 Hz,

CH2CF2), 45.6 (t, JFC = 31.6 Hz, CF2CH2NHCbz), 47.9 (CHN), 65.1 (CH2OH),

67.5 (OCH2Ph), 80.2 (C(CH3)3), 122.5 (t, JFC = 245 Hz, CF2), 128.3, 128.5,

128.7 (o, m ,p Ph), 136.2 (i-Ph), 156.1 (NCOO-tBu), 156.8

6 EXPERIMENTAL SECTION 152

(CH2NHCOOCH2Ph)

19 F-NMR (282.36 MHz, CDCl3): δ = -103.1 and -101.4 (br d, JFF = 250 Hz,

CF2)

20 [α] D (c 1.02, CHCl3): +1.3

Mp. : 108-110 °C

Rf : 0.33 (hexane:EA 1:1; UV, I2)

Synthesis of tert-butyl N-[(1S)-4-amino-3,3-difluoro-1-(hydroxymethyl) butyl]carbamate 25:

Deprotection of the Cbz protected 24 with Pd(OH)2/C NH2 HO F was performed in analogy to a literature procedure with a HN F O similar compound (249) and a communication about this O compound (219). To a stirred solution of 24 (14.3 g,

36.9 mmol, 1 eq) in 200 ml ethanol:EA 1:1 was added Pd(OH)2/C (1.4 g, 20% Pd). The flask, which was equipped with a septum and a balloon, was flushed three times with N2 and then three times with H2. The reaction was stirred for 6 h at r.t.. The reaction mixture was filtered over celite to remove the catalyst. The solvent from the filtrate was evaporated under reduced pressure. The resulting slightly pink oil was pure enough to be used for the next step.

YIELD: 9.4 g (quantitative)

CHARACTERIZATION:

HiRes-ESI-MS (M+Na, 277.1340 Da): 277.1330 Da (M+Na)+, 531.2800 Da

(2M+Na)+

6 EXPERIMENTAL SECTION 153

1 H-NMR (300 MHz, CDCl3): δ = 1.43 (s, 9 H, C(CH3)3), 2.17 (m, 2 H,

CH2CF2), 2.32 (br s, 2 H, NH2), 3.01 (t, J = 15 Hz, 2 H, CH2NHCbz), 3.66 (d, J

= 4.5 Hz, 2 H, CH2OH), 3.85-3.95 (br m, 1 H, CHN), 5.18 (br d, J = 6.6 Hz, 1 H,

NHBoc)

13 C-NMR (75 MHz, CDCl3): δ = 28.4 (C(CH3)3), 35.3 (t, JFC = 23.5 Hz,

CH2CF2), 46.8 (t, JFC = 28.7 Hz, CF2CH2NHCbz), 48.0 (CHN), 64.8 (CH2OH),

80.1 (C(CH3)3), 123.5 (t, JFC = 241 Hz, CF2), 155.7 (NCOO-tBu)

19 F-NMR (282.36 MHz, CDCl3): δ = -104.5 and -103.2 (d quintet, JFF =

248 Hz, JFH = 15 Hz, CF2)

20 [α] D (c 1.22, CHCl3): -7.1°

Rf : 0.04 (hexane:EA 1:1; UV, I2, Ninhydrin)

Synthesis of tert-butyl N-[(Z)-[(tert-butoxycarbonyl)amino]((4S)-4- [(tert-butoxycarbonyl)amino]-2,2-difluoro-5-hydroxypentylamino) methylidene]carbamate 26:

O Guanidinylation of the amine 25 with di-Boc thiourea was performed in N O accordance with published procedures for N NH H O similar compounds (249-255) and a HO F HN F O O communication about this compound (219).

O To a stirred mint green solution of 25 (8.9 g,

35.0 mmol, 1 eq), di-Boc thiourea (11.6 g, 42 mmol, 1.2 eq) and NEt3 (17 ml,

122.5 mmol, 3.5 eq) in 100 ml DMF was added HgCl2 (11.4 g, 42 mmol, 1.2 eq) at 0 °C. The reaction was stirred for 1 h at 0 °C. After addition of 150 ml EA, the reaction mixture was filtered over celite to remove the HgCl2. The filtrate was washed with

6 EXPERIMENTAL SECTION 154

250 ml water and 100 ml sat. NaCl (aq). The organic layer was separated, dried

(Na2SO4), and the solvent was evaporated under reduced pressure. The resulting residue was purified on a silica gel column using hexane:EA 1:1 as eluent. The product was obtained as a white solid.

YIELD: 13.6 g (78%)

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 497.2787 Da): 497.2777 Da (M+H)+, 519.2600 Da

(M+Na)+

1 H-NMR (300 MHz, CDCl3): δ = 1.44, 1.49, and 1,51 (s, 27 H, 3 × C(CH3)3),

2.1-2.4 (m, 2 H, CH2CF2), 3.65-3.8 and 3.95-4.1 (m, 2 H, CH2NH), 3.59 and

3.68 (dd, J = 4.2 Hz and J = 11.1 Hz, 2 H, CH2OH), 4.0-4.15 (br m, 1 H, CHN),

6.25 (br s, 1 H, CHNHBoc), 8.70 (br s, CH2NHCNN), 11.34 (s, NHCNHBoc)

13 C-NMR (125 MHz, CDCl3): δ = 28.0, 28.2, and 28.5 (3 × C(CH3)3), 34.3 (t,

JFC = 24.0 Hz, CH2CF2), 46.8 (dd, JFC = 28.6 Hz and JFC = 34.6 Hz,

CF2CH2NH), 48.1 (CHN), 65.7 (CH2OH), 79.9 and 83.9 (3 × C(CH3)3), 121.9

(dd, JFC = 243 Hz and JFC = 245 Hz, CF2), 152.9, 156.0, 156.8, and 162.7 (3 ×

NCOO-tBu and NCNN)

19 F-NMR (282.36 MHz, CDCl3): δ = -101.9 and -96.5 (br d, JFF = 260 Hz, CF2)

20 [α] D (c 1.13, CHCl3): -7.5°

Mp. : 65-68 °C

Rf : 0.5 (hexane:EA 1:1; UV, KMnO4)

6 EXPERIMENTAL SECTION 155

Synthesis of (2S)-2-[(tert-butoxycarbonyl)amino]-5-([(tert- butoxycarbonyl)amino][(tert-butoxycarbonyl)imino]methylamino)-4,4- difluoropentanoic acid 27:

O Oxidation of the alcohol 26 with pyridinium dichromate was performed in N O O analogy to published procedures with similar N NH H O HO compounds (236, 256-270) and a F HN F O O communication about this compound (219).

O To a stirred solution of 26 (7.2 g, 14.5 mmol, 1 eq) in 40 ml DMF was added pyridinium dichromate (27.26 g, 72.5 mmol, 5 eq) in the dark. The reaction was stirred for 20 h at r.t.. After the addition of 200 ml Et2O and 10 g

MgSO4, the reaction mixture was filtered over MgSO4 to remove the chromium. The solvent of the filtrate was evaporated under reduced pressure. The resulting residue could not be purified further at this point and was used directly for the next step. It was therefore only partially characterized.

YIELD: 6.3 g (approximately 85%)

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 511.2579 Da): 511.2579 Da (M+H)+, 533.2413 Da

(M+Na)+

1 H-NMR (300 MHz, CDCl3): δ = 1.44, 1.48, and 1,49 (s, 27 H, 3 × C(CH3)3),

2.4-2.6 (m, 2 H, CH2CF2), 3.65-3.85 and 3.95-4.15 (m, 2 H, CH2NH), 4.5-4.65

(br m, 1 H, CHN), 6.18 (br s, 1 H, CHNHBoc), 8.72 (br s, CH2NHCNN), 11.41

(s, NHCNHBoc)

13 C-NMR (75 MHz, CDCl3): δ = 28.2, 28.3, and 28.5 (3 × C(CH3)3), 35.4 (t,

JFC = 24.0 Hz, CH2CF2), 45.3 (dd, JFC = 27.4 Hz and JFC = 31.7 Hz,

CF2CH2NH), 49.3 (CHN), 80.0, 80.2, and 83.9 (3 × C(CH3)3), 121.8 (t, JFC =

6 EXPERIMENTAL SECTION 156

244.8 Hz, CF2), 153.1, 156.0, 156.8, and 163.1 (3 × NCOO-tBu and NCNN),

173.3 (COOH)

19 F-NMR (282.36 MHz, CDCl3): δ = -102.5 and -98.2 (br d, JFF = 260 Hz, CF2)

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 41.3 min

Synthesis of (2S)-2-amino-5-[amino(imino)methyl]amino-4,4- difluoropentanoic acid 8:

NH Deprotection of tris Boc protected 27 with TFA O was performed as previously described in a N NH2 H HO F F communication about this compound (219). To 27 (6.3 g, NH2 12.3 mmol, 1 eq) was added 20 ml TFA at 0 °C. The reaction was shaken for 1 h at r.t..

After the addition of 70 ml Et2O, the resulting precipitate was removed by centrifugation (4 °C, 5000 rpm). The supernatant was concentrated around 10-fold by evaporation under reduced pressure and subjected to another round of precipitation using the same amount of ether as before. The precipitates were purified by preparative HPLC on a C18-100 column using a linear gradient of 0/100 to 40/60 in 60 min. Fractions eluting between 10 and 12 min contained product and were lyophilized to give a white powder of the bis-TFA salt of 8.

YIELD: 2.85 g (52%)

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 211.1006 Da): 211.1005 Da (M+H)+

1 H-NMR (300 MHz, D2O, pH 3): δ = 2.45-2.85 (m, 2 H, CH2CF2), 3.77 (t, JFH

= 14.5 Hz, 2 H, CF2CH2NH), 4.26 (dd, J = 3.5 Hz and J = 8.7 Hz, 1 H, CHN)

6 EXPERIMENTAL SECTION 157

13 C-NMR (75 MHz, D2O, pH 3): δ = 34.1 (t, JFC = 22.9 Hz, CH2CF2), 45.4 (t,

JFC = 27.8 Hz, CF2CH2NH), 48.4 (CHN), 121.9 (t, JFC = 244.0 Hz, CF2), 157.7

(CNN), 171.6 (COOH)

19 F-NMR (282.36 MHz, D2O, pH 3): δ = -104.3 and -102.9 (dm, JFF = 246 Hz,

CF2)

20 [α] D (c 1.12 (2 × TFA), H2O): -0.89° (pH 3), -2.8° (pH 7), +1.07° (pH 10)

20 [α] D (c 0.54 (free base), H2O): -1.87° (pH 3), -5.97° (pH 7), +2.24° (pH 10)

Mp. : 115-118 °C

Synthesis of (2S)-5-[amino(imino)methyl]amino-2- [(benzyloxy)carbonyl]amino-4,4-difluoropentanoic acid 28:

NH Selective protection of the amino group of 8 with O CbzCl was performed according to published procedures N NH2 H HO F for similar compounds (271-274). A solution of 8 (1.4 g, HN F O 3.25 mmol, 1 eq) in 1 M NaHCO3 (aq) was adjusted to O pH 9.2 with 0.66 M NaOH (aq). CbzCl (464 µl, 3.25 mmol, 1 eq) was added at 0 °C and the reaction was stirred for 16 h at 4 °C. The mixture was lyophilized and the resulting residue extracted twice with 50 ml MeOH on a sintered glass funnel. The solvent of the filtrate was evaporated under reduced pressure. The resulting white solid was purified by preparative HPLC on a C18-100 column using a linear gradient of 0/100 to 60/40 in 45 min. Fractions eluting between 27 and 34 min contained product and were lyophilized to give a white powder of the very hygroscopic TFA salt of 28. Fractions eluting between 10 and 11 min contained starting material and were recycled.

6 EXPERIMENTAL SECTION 158

YIELD: 0.91 g (61%)

CHARACTERIZATION:

HiRes-ESI-MS (M+Na, 367.1194 Da): 367.1183 Da (M+Na)+

EA: exp. (C16H19F5N4O6 · 0.5 H2O) C 41.38, H 4.28, N 11.99, O 22.26, F 20.46

found C 41.15, H 4.22, N 11.81, F 20.45

1 H-NMR (300 MHz, D2O, pH 3): δ = 2.3-2.7 (m, 2 H, CH2CF2), 3.63 and 3.64

(t, JFH = 14.7 Hz, 2 H, CF2CH2NH), 4.26 (dd, J = 3.6 Hz and J = 9.3 Hz, 1 H,

CHN), 5.06 and 5.15 (d, J = 13 Hz, 2 H, CH2Ph), 7.4 (m, 5 H, Ph)

13 C-NMR (125 MHz, DMSO): δ = 34.8 (t, JFC = 22.9 Hz, CH2CF2), 44.6 (t, JFC

= 26.7 Hz, CF2CH2NH), 48.2 (CHN), 65.4 (CH2Ph), 121.9 (t, JFC = 240 Hz,

CF2), 127.5, 127.7, 128.2 (o, m , p Ph), 136.8 (i-Ph), 155.6 (NCOO), 157.2

(CNN), 172.5 (COOH)

19 F-NMR (282.36 MHz, D2O, pH 3): δ = -102.9 (quintet, JFH = 15.4 Hz, CF2)

20 [α] D (c 0.50 (TFA), H2O): -12.10° (pH 3.5), -6.45° (pH 12)

Mp. : 148-152 °C (decomp.)

RP-HPLC (C18, 5/95 to 60/40 in 45 min): Rt = 24.3 min

Synthesis of (2S)-2-[(benzyloxy)carbonyl]amino-4,4-difluoro-5- [(imino[(2,2,4,6,7-pentamethyl-2,3-dihydro-1-benzofuran-5- yl)sulfonyl]aminomethyl)amino]pentanoic acid 29:

Protection of the guanidino NH O O group of 28 with PbfCl was N N S O H H HO F O HN F O

O 6 EXPERIMENTAL SECTION 159 performed in analogy to published procedures with similar compounds (272, 275-278). To a solution of the mono TFA salt of 28 (1.03 g, 2.25 mmol, 1 eq) in 20 ml acetone was added PbfCl (1.04 g, 3.56 mmol, 1.6 eq) and 3.3 ml (6 eq) 4 M NaOH (aq) at 0 °C, and the reaction was stirred for 2 h at 0 °C and an additional 2 h at r.t.. For work-up, the reaction was treated with 900 µl sat. citric acid (aq), and most of the acetone was removed by evaporation under reduced pressure. Another 2.6 ml sat. citric acid (aq) were added and the mixture was extracted twice with 50 ml EA. The organic phase was dried over Na2SO4 and the solvent was evaporated under reduced pressure. The resulting sticky residue was purified by preparative HPLC on a C18-100 column using a linear gradient of 0/100 to 80/20 in 60 min. Fractions eluting between 52 and 56 min contained product and were lyophilized to give 29 as a white sticky solid. Fractions eluting between 32 and 40 min contained starting material and were recycled.

YIELD: 0.60 g (45%) plus 40% recovered starting material

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 597.2193 Da): 597.2198 Da (M+H)+, 619.2048 Da

(M+Na)+

1 H-NMR (300 MHz, CD3OD): δ = 1.43 (s, 6 H, C(CH3)2), 2.06 (s, 3 H, m-

ArCH3), 2.49, and 2.54 (s, 6 H,2 × o-ArCH3), 2.2-2.5 (m, 2 H, CH2CF2), 2.97 (s,

2 H, ArCH2), 3.66 (t, JFH = 14.4 Hz, 2 H, CF2CH2NH), 4.42 (dd, J = 3.5 Hz and

J = 9.5 Hz, 1 H, CHN), 5.07 (d, J = 13 Hz, 2 H, CH2Ph), 7.3 (m, 5 H, Ph)

13 C-NMR (125 MHz, DMSO): δ = 12.1 (m-ArCH3), 17.5, and 18.8 (2 × o-

ArCH3), 28.2 (C(CH3)2), 34.9 (t, JFC = 22.4 Hz, CH2CF2), 42.3 (ArCH2), 44.0 (t,

JFC = 26.4 Hz, CF2CH2NH), 48.4 (CHN), 65.4 (CH2Ph), 86.3 (C(CH3)2), 116.2

(m-ArCH3), 122.0 (t, JFC = 241.8 Hz, CF2), 124.3 (m-ArCH2), 127.5, 127.7, and

128.2 (o, m ,p-Ph), 131.5 (o-ArCH3), 133.6 (i-ArSO2), 136.8 (i-Ph), 137.4 (o-

ArCH3), 155.7 (NCOO), 156.1 (CNN), 157.5 (p-ArO), 172.5 (COOH)

6 EXPERIMENTAL SECTION 160

19 F-NMR (282.36 MHz, CD3OD): δ = -104.0 and -102.6 (br d, JFF = 250 Hz,

CF2)

20 [α] D (c 0.81, MeOH): -7.04°

Mp. : 97-99 °C

RP-HPLC (C18, 5/95 to 60/40 in 45 min to 60 min): Rt = 46.8 min

Synthesis of (2S)-2-amino-4,4-difluoro-5-[(imino[(2,2,4,6,7- pentamethyl-2,3-dihydro-1-benzofuran-5-yl)sulfonyl] aminomethyl)amino]pentanoic acid 30:

NH Deprotection of the Cbz O O O N N S protecting group of 29 with H2 and H H HO F F O NH2 Pd/C was performed in analogy to literature procedures for similar compounds (249, 272, 275). To a stirred solution of 29 (1.0 g, 1.68 mmol, 1 eq) in 10 ml methanol was added 100 mg of Pd/C (10%). The flask, which was equipped with a septum and a balloon, was flushed three times with N2 and then three times with H2. The reaction was vigorously stirred for 15 h at r.t.. The reaction mixture was filtered over celite to remove the Pd/C. The solvent from the filtrate was evaporated under reduced pressure. The resulting white solid was pure enough to be used for the next step.

YIELD: 0.79 g (quant.)

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 463.1826 Da): 463.1815 Da (M+H)+, 485.1631 Da

(M+Na)+

1 H-NMR (300 MHz, CD3OD): δ = 1.45 (s, 6 H, C(CH3)2), 2.07 (s, 3 H,m

6 EXPERIMENTAL SECTION 161

ArCH3), 2.50, and 2.55 (s, 6 H, 2 × o ArCH3), 2.25-2.85 (m, 2 H, CH2CF2), 2.99

(s, 2 H, ArCH2), 3.72 and 3.74 (t, JFH = 13.8 Hz, 2 H, CF2CH2NH), 4.01 (dd, J =

3.0 Hz and J = 9.3 Hz, 1 H, CHN)

13 C-NMR (75 MHz, CD3OD): δ = 11.2 (m ArCH3), 17.2, and 18.3 (2 × o

ArCH3), 27.5 (C(CH3)2), 34.9 (t, JFC = 23.6 Hz, CH2CF2), 42.7 (ArCH2), 44.4 (t,

JFC = 26.6 Hz, CF2CH2NH), 47-48.7 (CHN, hidden under solvent peak), 86.6

(C(CH3)2), 117.4 (m ArCH3), 122.3 (t, JFC = 240 Hz, CF2), 124.9 (m ArCH2),

132.4 (o ArCH3), 132.6 (i ArSO2), 138.3 (o ArCH3), 157.0 (CNN), 158.9 (p

ArO), 170.7 (COOH)

19 F-NMR (282.36 MHz, CD3OD): δ = -104.2 and -102.2 (br d, JFF = 250 Hz,

CF2)

20 [α] D (c 0.134, H2O:MeOH:DMF 5:1:1): -6.5° (pH 13.5)

Mp. : 140-145 °C (decomp.)

RP-HPLC (C18, 5/95 to 80/20 in 45 min to 60 min): Rt = 18.6 min

Synthesis of (2S)-2-[(9H-fluoren-9-ylmethoxy)carbonyl]amino-4,4- difluoro-5-[(imino[(2,2,4,6,7-pentamethyl-2,3-dihydro-1-benzofuran-5- yl)sulfonyl]aminomethyl)amino]pentanoic acid 13:

NH Protection of the amino O O group of 30 with FmocOSu was N N S O H H HO performed in as described for F O HN F O similar compounds (272, 275). To a

O solution of 30 (0.49 g, 1.06 mmol, 1 eq) in 4 ml water and 6 ml

6 EXPERIMENTAL SECTION 162

acetonitrile was added NEt3 until a pH of 9 was reached (approximately 200 µl). FmocOSu (357 mg, 1.06 mmol, 1 eq) was added to this mixture at 0 °C, and the mixture was stirred for 1 h at r.t.. During the reaction the pH was kept around 9 by adding NEt3.

After the addition of 10 ml water and 100 µl NEt3 the acetonitrile was evaporated under reduced pressure. The solution was extracted twice with 5 ml Et2O, then acidified to pH 2.5 with sat. citric acid (aq) and extracted twice with 10 ml EA. The organic layer was washed with 10 ml sat. NaCl (aq), dried (Na2SO4), and filtered, and the solvent was evaporated under reduced pressure. The resulting residue was purified by preparative

HPLC on a C8-100 column using a linear gradient of 3/97 to 80/20 in 60 min. Fractions eluting between 51 and 55 min contained product and were lyophilized to give 13 as a white solid, which still contained some TFA. To completely remove the TFA it was necessary to redissolve it in 0.1 M HCl (H2O:CH3CN 7:3) and lyophilize it 6 times (347). The water phase was also lyophilized and purified using the same HPLC conditions. Fractions eluting between 33 and 36 min contained starting material and were recycled.

YIELD: 0.47 g (65%)

CHARACTERIZATION:

HiRes-ESI-MS (M+H, 685.2507 Da): 685.2511 Da (M+H)+, 707.2332 Da

(M+Na)+

EA: exp. (C34H38F2N4O7S · 0.5 H2O · 0.7 HCl):

C 56.77, H 5.56, N 7.78, O 16.68, F 5.28

found:

C 56.81, H 5.57, N 7.85, O 16.10, F 5.29

1 H-NMR (600 MHz, DMSO): δ = 1.39 (s, 6 H, C(CH3)2), 1.98 (s, 3 H,m

ArCH3), 2.42, and 2.47 (s, 6 H,2 × o ArCH3), 2.2-2.55 (m, 2 H, CH2CF2), 2.93

(s, 2 H, ArCH2), 3.59 and 3.60 (t, JFH = 14.8 Hz, 2 H, CF2CH2NH), 4.2-4.3 (m,

4 H, CHN, CHCH2O), 6.59 and 7.00 (br s, 3 H, NHC(=NH)NH, disappears upon

6 EXPERIMENTAL SECTION 163

addition of D2O), 7.32, 7.41, 7.70, and 7.89 (m, 8 H, aromatic protons), 7.80 (d,

1 H, J = 8.5 Hz, CHNHCO, disappears upon addition of D2O)

13 C-NMR (150.9 MHz, CD3OD): δ = 12.1 (m ArCH3), 17.5, and 18.8 (2 × o

ArCH3), 28.2 (C(CH3)2), 34.9 (t, JFC = 22.6 Hz, CH2CF2), 42.3 (ArCH2), 44.1 (t,

JFC = 21.1 Hz, CF2CH2NH), 46.5 (CHN), 48.3 (CHCH2O), 65.6 (CHCH2O),

86.2 (C(CH3)2), 116.2 (m-ArCH3), 120.2, 125.1, 127.0, and 127.5 (aromatic CH

Fmoc), 122.0 (t, JFC = 243 Hz, CF2), 124.3 (m-ArCH2), 131.5 (o-ArCH3), 133.6

(i-ArSO2), 137.4 (o-ArCH3), 140.6 and 143.7 (aromatic C of Fmoc), 155.6,

156.1, and 157.5 (CNN, p-ArO, and NCOO), 172.4 (COOH)

19 F-NMR (282.36 MHz, DMSO): δ = -102.6 (d, J = 22.8 Hz, CF2)

20 [α] D (c 0.354, DMSO): -5.7°

Mp. : 136-138 °C (melting) and 164 °C (decomp.)

RP-HPLC (C18, 5/95 to 80/20 in 45 min to 60 min): Rt = 33.0 min

164

7 Appendix

7.1 Programs for the Peptide Synthesizer

Modifications to the standard programs

For the synthesis of BsCM 88-127 or parts thereof, the preinstalled program modules were altered to the modules described below. Briefly, the first number gives the line number. The second number is an ABI code for the program step, which is explained in the next column. In the forth column, the duration of the corresponding step is given in seconds, or alternatively, if it is a loop, the number of repetitions is shown, or again alternatively for “end loop monitoring” the difference to the previous deprotection peak that has to be reached to end the loop is given in ‰. In the fifth column additions to the value in the forth column are given. One tenth of this value is added per coupled amino acid. The last column gives the total time in seconds (only correct for 0.1 mmol scale and the first cycle without any additions) consumed up to that step. Four different chemistries (FM010_MPPSTD3, FM025_MPPSTD3, FM010_MPPSTD4 or FM025_MPPSTD4) were used, that differ slightly in the combination of the modules and the modules itself. The main difference between STD3 and STD4 is the existence of module G, the only difference to module g is the elongation of the coupling time to 4 h. Alterations for the 0.25 mmol scale (025), where larger amounts of solvents are needed, compared to the 0.1 mmol scale (010) are given in brackets.

Module a Like module G but recovery of excess amino acid Module b Problematic deprotection 10 times 15 min (3 %) (10 min for FM010_MPPSTD4) 7 APPENDIX 165

Module c DCM washes Module d First long resin washing Module f Problematic coupling with monitoring Module g Multi coupling 2 h (or 2.5 h for 0.25 mmol scale) Module h Capping and NMP washing Module i NMP soak 1 h Module A Activation Module B DCM washes Module C Conditional additional deprotection 45 min Module D NMP washes Module E Transfer Module G Multi coupling 4 h (same as g except for the duration) Module H Conditional wait 30 min Module I Wait 30 min

Module a (Like module G but lines 33 to 41 are changed to:) 33 82 Drain reaction vessel to aux. waste 150 1 15351 34 38 Transfer activator to reaction vessel (top open) 10 15361 35 40 Mix reaction vessel 1 15362 36 2 Vortex reaction vessel on 3 15365 37 40 Mix reaction vessel 2 15367 38 99 End loop UPPER 1 15367 39 82 Drain reaction vessel to aux. waste 150 15517 40 86 Flow NMP through reaction vessel to aux. waste 15 15532 41 82 Drain reaction vessel to aux. waste 150 15682 42 118 Flush bottom valve block with NMP to aux.waste 10 15692 43 80 Flush bottom valve block with gas to aux.waste 20 15712 44 22 Drain activator to waste 10 15722 45 3 Vortex reaction vessel off 1 15723 46 42 Drain reaction vessel to waste 15 1 15738

Module b (For FM010_MPPSTD4 in line 24, forth column 900 is changed to 600) 1 1 Wait 1 1 2 58 Interrupt conditional 1 3 135 Monitoring reset 1 1 4 110 Begin loop lower 2 1 5 42 Drain reaction vessel to waste 10 (18) 1(2) 1067 6 98 Begin loop UPPER 3 1067 7 56 Deliver NMP to reaction vessel 4 (13) 1 1147 8 88 Resin sampler to reaction vessel 1 1148 9 91 Deliver NMP to resin sampler 4 1152 10 89 Resin sampler to frac.collector 1 1153 11 93 Deliver gas to resin sampler 2 1155 12 2 Vortex reaction vessel on 5 1160 13 40 Mix reaction vessel 2 1162 14 3 Vortex reaction vessel off 1 1163 15 42 Drain reaction vessel to waste 10 (18) 1 1173 16 50 Flow NMP through reaction vessel to waste 3 (5) 1176 17 42 Drain reaction vessel to waste 5 (10) 1 1181 18 99 End loop UPPER 1 1181 19 56 Deliver NMP to reaction vessel 3 (12) (2) 1184 20 79 Pressurize Pip 10 1194

7 APPENDIX 166

21 51 Deliver Pip to reaction vessel 5 (10) (1) 1199 22 56 Deliver NMP to reaction vessel 4 1203 23 40 Mix reaction vessel 2 1205 24 2 Vortex reaction vessel on 900 2105 25 3 Vortex reaction vessel off 1 2106 26 40 Mix reaction vessel 1 2107 27 130 Monitoring previous peak 1 2107 28 42 Drain reaction vessel to waste 3 2110 29 1 Wait 3 2113 30 131 Monitoring stop 1 2113 31 132 Read monitoring peak 1 2113 32 111 End loop lower 1 2113 33 133 Begin loop monitoring 8 2113 34 42 Drain reaction vessel to waste 8 (18) 1 9765 35 98 Begin loop UPPER 4 9765 36 56 Deliver NMP to reaction vessel 4 (13) 1 9883 37 88 Resin sampler to reaction vessel 1 9884 38 91 Deliver NMP to resin sampler 4 9888 39 89 Resin sampler to frac.collector 1 9889 40 93 Deliver gas to resin sampler 2 9891 41 2 Vortex reaction vessel on 5 9896 42 40 Mix reaction vessel 2 9898 43 3 Vortex reaction vessel off 1 9899 44 42 Drain reaction vessel to waste 10 (18) 1 9909 45 50 Flow NMP through reaction vessel to waste 3 (5) 9912 46 42 Drain reaction vessel to waste 5 (10) 1 9917 47 99 End loop UPPER 1 9917 48 56 Deliver NMP to reaction vessel 3 (12) (2) 9920 49 79 Pressurize Pip 10 9930 50 51 Deliver Pip to reaction vessel 5 (10) (1) 9935 51 56 Deliver NMP to reaction vessel 4 9939 52 40 Mix reaction vessel 2 9941 53 2 Vortex reaction vessel on 900 10841 54 3 Vortex reaction vessel off 1 10842 55 40 Mix reaction vessel 1 10843 56 130 Monitoring previous peak 1 10843 57 42 Drain reaction vessel to waste 3 10846 58 1 Wait 3 10849 59 131 Monitoring stop 1 10849 60 132 Read monitoring peak 1 10849 61 134 End loop monitoring 30 10849 62 10 Flush bottom valve block with gas to waste 3 10852 63 40 Mix reaction vessel 2 10854 64 41 Vent reaction vessel 2 10856 65 88 Resin sampler to reaction vessel 1 10857 66 93 Deliver gas to resin sampler 3 10860 67 41 Vent reaction vessel 1 10861 68 89 Resin sampler to frac.collector 1 10862 69 98 Begin loop UPPER 2 10862 70 13 Flush top valve block with NMP to waste 3 10877 71 14 Flush bottom valve block with NMP to waste 3 10880 72 9 Flush top valve block with gas to waste 3 10883 73 10 Flush bottom valve block with gas to waste 3 10886 74 99 End loop UPPER 1 10886

Module c 1 1 Wait 1 1

7 APPENDIX 167

2 12 Flush bottom valve block with DCM to waste 1 2 3 9 Flush top valve block with gas to waste 2 4 4 10 Flush bottom valve block with gas to waste 2 6 5 98 Begin loop UPPER 6 6 6 55 Deliver DCM to reaction vessel 5 (12) 1 251 7 40 Mix reaction vessel 2 253 8 2 Vortex reaction vessel on 1 254 9 41 Vent reaction vessel 1 255 10 88 Resin sampler to reaction vessel 1 256 11 90 Deliver DCM to resin sampler 2 258 12 93 Deliver gas to resin sampler 2 260 13 90 Deliver DCM to resin sampler 5 265 14 89 Resin sampler to frac.collector 1 266 15 40 Mix reaction vessel 2 268 16 1 Wait 5 273 17 3 Vortex reaction vessel off 1 274 18 42 Drain reaction vessel to waste 5 (10) 1 279 19 41 Vent reaction vessel 2 281 20 49 Flow DCM through reaction vessel to waste 3 (5) 284 21 42 Drain reaction vessel to waste 10 (15) 1 294 22 99 End loop UPPER 1 294 23 10 Flush bottom valve block with gas to waste 2 296 24 93 Deliver gas to resin sampler 5 301 25 41 Vent reaction vessel 2 303 26 88 Resin sampler to reaction vessel 1 304 27 93 Deliver gas to resin sampler 5 309 28 41 Vent reaction vessel 2 311 29 89 Resin sampler to frac.collector 1 312 30 42 Drain reaction vessel to waste 30 342 31 29 Flow DCM through activator to waste 5 347 32 22 Drain activator to waste 30 377 33 11 Flush top valve block with DCM to waste 1 378 34 12 Flush bottom valve block with DCM to waste 1 379 35 10 Flush bottom valve block with gas to waste 10 389 36 9 Flush top valve block with gas to waste 10 399

Module d 1 1 Wait 1 1 2 12 Flush bottom valve block with DCM to waste 1 2 3 9 Flush top valve block with gas to waste 2 4 4 10 Flush bottom valve block with gas to waste 2 6 5 98 Begin loop UPPER 3 6 6 55 Deliver DCM to reaction vessel 5 (12) 1 1287 7 40 Mix reaction vessel 2 1289 8 2 Vortex reaction vessel on 600 1889 9 41 Vent reaction vessel 1 1890 10 40 Mix reaction vessel 2 1892 11 1 Wait 5 1897 12 3 Vortex reaction vessel off 1 1898 13 42 Drain reaction vessel to waste 5 (10) 1 1903 14 41 Vent reaction vessel 2 1905 15 49 Flow DCM through reaction vessel to waste 5 1910 16 42 Drain reaction vessel to waste 10 1 1920 17 99 End loop UPPER 1 1920 18 1920 19 1 Wait 1 1921 20 110 Begin loop lower 2 1921

7 APPENDIX 168

21 3 Vortex reaction vessel off 1 3776 22 42 Drain reaction vessel to waste 10 (18) 1(2) 3786 23 41 Vent reaction vessel 2 3788 24 50 Flow NMP through reaction vessel to waste 2 (3) 3790 25 42 Drain reaction vessel to waste 5 (6) 1 3795 26 56 Deliver NMP to reaction vessel 4 (13) 1 3799 27 2 Vortex reaction vessel on 900 4699 28 40 Mix reaction vessel 4 (6) 4703 29 2 Vortex reaction vessel on 900 5603 30 1 Wait 15 5618 31 3 Vortex reaction vessel off 1 5619 32 42 Drain reaction vessel to waste 10 (15) 1 5629 33 111 End loop lower 1 5629 34 5629 35 98 Begin loop UPPER 6 1 5629 36 41 Vent reaction vessel 2 6836 37 50 Flow NMP through reaction vessel to waste 2 (3) 6838 38 42 Drain reaction vessel to waste 5 (6) 1 6843 39 56 Deliver NMP to reaction vessel 4 (13) 1 6847 40 2 Vortex reaction vessel on 100 6947 41 40 Mix reaction vessel 2 6949 42 2 Vortex reaction vessel on 100 7049 43 1 Wait 15 7064 44 3 Vortex reaction vessel off 1 7065 45 42 Drain reaction vessel to waste 10 (15) 1 7075 46 99 End loop UPPER 1 7075

Module f 1 1 Wait 1 1 2 2 Vortex reaction vessel on 1 2 3 5 Needle down 10 12 4 62 Drain cartridge to waste 10 22 5 98 Begin loop UPPER 3 22 6 67 Deliver NMP to cartridge small needle 2 38 7 62 Drain cartridge to waste 5 43 8 99 End loop UPPER 1 43 9 98 Begin loop UPPER 2 (3) 43 10 65 Deliver NMP to cartridge 18 (24) 119 11 60 Mix cartridge 10 129 12 24 Transfer cartridge to activator 20 149 13 62 Drain cartridge to waste 10 159 14 99 End loop UPPER 1 159 15 98 Begin loop UPPER 2 159 16 67 Deliver NMP to cartridge small needle 2 173 17 62 Drain cartridge to waste 10 183 18 99 End loop UPPER 1 183 19 62 Drain cartridge to waste 10 193 20 60 Mix cartridge 5 198 21 61 Vent cartridge 2 200 22 133 Begin loop monitoring 200 23 98 Begin loop UPPER 60 (120) 2(4) 200 24 2 Vortex reaction vessel on 15 1985 25 3 Vortex reaction vessel off 13 1998 26 41 Vent reaction vessel 2 2000 27 99 End loop UPPER 1 2000 28 134 End loop monitoring 1 2000 29 88 Resin sampler to reaction vessel 1 2001

7 APPENDIX 169

30 91 Deliver NMP to resin sampler 5 2006 31 89 Resin sampler to frac.collector 1 2007 32 98 Begin loop UPPER 3 (4) 2007 33 3 Vortex reaction vessel off 1 2056 34 28 Pressurize activator 4 2060 35 42 Drain reaction vessel to waste 7 (15) 1 2067 36 38 Transfer activator to reaction vessel (top open) 6 (10) 2073 37 40 Mix reaction vessel 1 2074 38 2 Vortex reaction vessel on 3 2077 39 40 Mix reaction vessel 2 2079 40 99 End loop UPPER 1 2079 41 22 Drain activator to waste 5 2084 42 3 Vortex reaction vessel off 1 2085 43 42 Drain reaction vessel to waste 7 (15) 1 2092

Module g and G (For G in line 22, forth column 240 is changed to 480) 1 1 Wait 1 1 2 2 Vortex reaction vessel on 1 2 3 5 Needle down 10 12 4 62 Drain cartridge to waste 10 22 5 98 Begin loop UPPER 3 22 6 67 Deliver NMP to cartridge small needle 2 38 7 62 Drain cartridge to waste 5 43 8 99 End loop UPPER 1 43 9 98 Begin loop UPPER 2 (3) 43 10 65 Deliver NMP to cartridge 18 (24) 119 11 60 Mix cartridge 10 129 12 24 Transfer cartridge to activator 20 149 13 62 Drain cartridge to waste 10 159 14 99 End loop UPPER 1 159 15 98 Begin loop UPPER 2 159 16 67 Deliver NMP to cartridge small needle 2 173 17 62 Drain cartridge to waste 10 183 18 99 End loop UPPER 1 183 19 62 Drain cartridge to waste 10 193 20 60 Mix cartridge 5 198 21 61 Vent cartridge 2 200 22 98 Begin loop UPPER 240 (300) 2(4) 200 23 2 Vortex reaction vessel on 15 7385 24 3 Vortex reaction vessel off 13 7398 25 41 Vent reaction vessel 2 7400 26 99 End loop UPPER 1 7400 27 88 Resin sampler to reaction vessel 1 7401 28 91 Deliver NMP to resin sampler 5 7406 29 89 Resin sampler to frac.collector 1 7407 30 98 Begin loop UPPER 3 (4) 7407 31 3 Vortex reaction vessel off 1 7456 32 28 Pressurize activator 4 7460 33 42 Drain reaction vessel to waste 7 (15) 1 7467 34 38 Transfer activator to reaction vessel (top open) 6 (10) 7473 35 40 Mix reaction vessel 1 7474 36 2 Vortex reaction vessel on 3 7477 37 40 Mix reaction vessel 2 7479 38 99 End loop UPPER 1 7479 39 22 Drain activator to waste 5 7484 40 3 Vortex reaction vessel off 1 7485 41 42 Drain reaction vessel to waste 7 (15) 1 7492

7 APPENDIX 170

Module h 1 1 Wait 1 1 2 3 Vortex reaction vessel off 1 2 3 42 Drain reaction vessel to waste 10 (20) 1 12 4 98 Begin loop UPPER 6 1 12 5 41 Vent reaction vessel 2 399 6 50 Flow NMP through reaction vessel to waste 2 (3) 401 7 42 Drain reaction vessel to waste 5 1 406 8 56 Deliver NMP to reaction vessel 4 (12) 1 410 9 2 Vortex reaction vessel on 30 440 10 40 Mix reaction vessel 2 442 11 88 Resin sampler to reaction vessel 1 443 12 91 Deliver NMP to resin sampler 2 445 13 89 Resin sampler to frac.collector 1 446 14 93 Deliver gas to resin sampler 2 448 15 1 Wait 15 463 16 3 Vortex reaction vessel off 1 464 17 42 Drain reaction vessel to waste 10 (18) 1 474 18 99 End loop UPPER 1 474 19 77 Pressurize #4 10 484 20 10 Flush bottom valve block with gas to waste 2 486 21 52 Deliver #4 to reaction vessel 15 (35) 1(2) 501 22 40 Mix reaction vessel 2 503 23 41 Vent reaction vessel 2 505 24 88 Resin sampler to reaction vessel 1 506 25 93 Deliver gas to resin sampler 3 509 26 91 Deliver NMP to resin sampler 2 511 27 93 Deliver gas to resin sampler 3 514 28 89 Resin sampler to frac.collector 1 515 29 2 Vortex reaction vessel on 300 815 30 3 Vortex reaction vessel off 1 816 31 88 Resin sampler to reaction vessel 1 817 32 91 Deliver NMP to resin sampler 3 820 33 89 Resin sampler to frac.collector 1 821 34 42 Drain reaction vessel to waste 5 826 35 50 Flow NMP through reaction vessel to waste 5 831 36 93 Deliver gas to resin sampler 3 834 37 42 Drain reaction vessel to waste 10 (20) 1(2) 844 38 1 Wait 1 845 39 3 Vortex reaction vessel off 1 846 40 42 Drain reaction vessel to waste 10 (18) 1(2) 856 41 98 Begin loop UPPER 8 1 856 42 41 Vent reaction vessel 2 1397 43 50 Flow NMP through reaction vessel to waste 2 (3) 1399 44 42 Drain reaction vessel to waste 5 1 1404 45 56 Deliver NMP to reaction vessel 4 (13) 1 1408 46 2 Vortex reaction vessel on 30 1438 47 40 Mix reaction vessel 2 (5) 1440 48 88 Resin sampler to reaction vessel 1 1441 49 91 Deliver NMP to resin sampler 2 1443 50 89 Resin sampler to frac.collector 1 1444 51 93 Deliver gas to resin sampler 2 1446 52 1 Wait 15 1461 53 3 Vortex reaction vessel off 1 1462 54 42 Drain reaction vessel to waste 10 (15) 1 1472 55 99 End loop UPPER 1 1472

7 APPENDIX 171

Module i 1 42 Drain reaction vessel to waste 10 1 10 2 98 Begin loop UPPER 3 10 3 56 Deliver NMP to reaction vessel 4 (13) 1 74 4 2 Vortex reaction vessel on 5 79 5 40 Mix reaction vessel 2 81 6 3 Vortex reaction vessel off 1 82 7 42 Drain reaction vessel to waste 10 (15) 1 92 8 50 Flow NMP through reaction vessel to waste 3 95 9 42 Drain reaction vessel to waste 5 1 100 10 99 End loop UPPER 1 100 11 56 Deliver NMP to reaction vessel 12 112 12 40 Mix reaction vessel 2 114 13 2 Vortex reaction vessel on 600 714 14 3 Vortex reaction vessel off 1 715 15 40 Mix reaction vessel 1 716 16 98 Begin loop UPPER 5 716 17 1 Wait 580 3616 18 99 End loop UPPER 1 3616 19 42 Drain reaction vessel to waste 12 (15) 3628

Module A 1 1 Wait 1 1 2 4 Read cartridge 10 1 3 6 Needle up 10 11 4 7 Eject cartridge 10 21 5 8 Advance cartridge 10 31 6 5 Needle down 10 41 7 14 Flush bottom valve block with NMP to waste 1 42 8 9 Flush top valve block with gas to waste 2 44 9 65 Deliver NMP to cartridge 5 49 10 60 Mix cartridge 5 54 11 78 Pressurize manifold 10 64 12 18 Flush bottom valve block with HBTU to waste 1 65 13 94 Deliver HBTU to cartridge 9 74 14 98 Begin loop UPPER 6 74 15 2 Vortex reaction vessel on 1 385 16 60 Mix cartridge 30 415 17 3 Vortex reaction vessel off 1 416 18 60 Mix cartridge 30 446 19 99 End loop UPPER 1 446 20 13 Flush top valve block with NMP to waste 1 447 21 14 Flush bottom valve block with NMP to waste 3 450 22 9 Flush top valve block with gas to waste 2 452 23 10 Flush bottom valve block with gas to waste 5 457

Module C 1 137 Do module if condition not met 1 2 98 Begin loop UPPER 2 3 42 Drain reaction vessel to waste 10 (20) 1 2756 4 56 Deliver NMP to reaction vessel 3 (12) 2759 5 79 Pressurize Pip 10 2769 6 51 Deliver Pip to reaction vessel 5 (10) (1) 2774 7 56 Deliver NMP to reaction vessel 4 2778

7 APPENDIX 172

8 40 Mix reaction vessel 2 2780 9 110 Begin loop lower 9 2780 10 2 Vortex reaction vessel on 280 5460 11 3 Vortex reaction vessel off 17 5477 12 40 Mix reaction vessel 3 5480 13 111 End loop lower 1 5480 14 42 Drain reaction vessel to waste 12 (20) 5492 15 99 End loop UPPER 1 5492

Module D 1 1 Wait 1 1 2 3 Vortex reaction vessel off 1 2 3 42 Drain reaction vessel to waste 10 (18) 1(2) 12 4 98 Begin loop UPPER 4 (5) 1 12 5 41 Vent reaction vessel 2 164 6 50 Flow NMP through reaction vessel to waste 2 (3) 166 7 42 Drain reaction vessel to waste 5 1 171 8 56 Deliver NMP to reaction vessel 4 (13) 1 175 9 2 Vortex reaction vessel on 3 (5) 178 10 40 Mix reaction vessel 2 (5) 180 11 88 Resin sampler to reaction vessel 1 181 12 91 Deliver NMP to resin sampler 2 183 13 89 Resin sampler to frac.collector 1 184 14 93 Deliver gas to resin sampler 2 186 15 1 Wait 15 201 16 3 Vortex reaction vessel off 1 202 17 42 Drain reaction vessel to waste 10 (15) 1 212 18 99 End loop UPPER 1 212

Module E 1 1 Wait 1 1 2 10 Flush bottom valve block with gas to waste 2 3 3 5 Needle down 10 13 4 78 Pressurize manifold 10 23 5 70 Flush bottom valve block with loop contents to waste 2 25 6 98 Begin loop UPPER 2 25 7 68 Deliver DIEA to measuring loop (open) 2 33 8 63 Transfer measuring loop to cartridge 4 37 9 99 End loop UPPER 1 37 10 60 Mix cartridge 5 42 11 98 Begin loop UPPER 4 42 12 41 Vent reaction vessel 2 80 13 96 Transfer cartridge to reaction vessel (top closed) 6 86 14 2 Vortex reaction vessel on 3 89 15 3 Vortex reaction vessel off 1 90 16 99 End loop UPPER 1 90 17 65 Deliver NMP to cartridge 90 18 60 Mix cartridge 90 19 98 Begin loop UPPER 3 90 20 41 Vent reaction vessel 2 106 21 96 Transfer cartridge to reaction vessel (top closed) 5 111 22 99 End loop UPPER 1 111 23 40 Mix reaction vessel 1 (2) 112 24 93 Deliver gas to resin sampler 2 114 25 41 Vent reaction vessel 2 116 26 88 Resin sampler to reaction vessel 1 117

7 APPENDIX 173

27 93 Deliver gas to resin sampler 3 120 28 41 Vent reaction vessel 2 122 29 89 Resin sampler to frac.collector 1 123 30 2 Vortex reaction vessel on 1 124 31 62 Drain cartridge to waste 5 129 32 40 Mix reaction vessel 1 130

Module H 1 137 Do module if condition not met 1 2 98 Begin loop UPPER 6 3 2 Vortex reaction vessel on 290 1790 4 3 Vortex reaction vessel off 10 1800 5 99 End loop UPPER 1 1800

Module I 1 98 Begin loop UPPER 6 2 2 Vortex reaction vessel on 290 1790 3 3 Vortex reaction vessel off 10 1800 4 99 End loop UPPER 1 1800

7 APPENDIX 174

7.2 Analytical Ultracentrifugation Data for BsCM* and Arg90Cit BsCM*

Data for BsCM*

Semisynthetic BsCM* (independent species model)

s3, st ep #62, gridsize = 1000; rmsd = 0.006545 (n = 87500, SSR = 3.748184 M = 41539 | 132; s = 3.1 | 0.3; c = 1.99 | 0.59; Runs test Z (#st ddev) = 236.98 Meniscus = 6.07; Bottom = 7.22; fitt ing 6.01 – 7.05

7 APPENDIX 175

Semisynthetic BsCM* (continuous mass model)

s5, step #15, gridsize = 500 rmsd = 0.006679 (n = 87500, SSR = 3.903491) frict ration = 1.03; Runs test Z (#stddev) = 205.07 Meniscus = 6.06; Bottom = 7.22; fitting 6.10 – 7.06

Data for Arg90Cit BsCM*

Arg90Cit BsCM* (independent species model)

s2, step #70, gridsize = 1000; rmsd = 0.007937 (n = 61100, SSR = 3.848958 M = 43597; s = 3.3; c = 0.9; Runs test Z (#stddev) = 206.71 Meniscus = 6.02; Bottom = 7.20; fitting 6.05 – 7.05

7 APPENDIX 176

Arg90Cit BsCM* (continuous mass model)

s2, step #40, gridsize = 500 rmsd = 0.007996 (n = 611000, SSR = 3.906271 frict ration = 1.26 Runs test Z (#stddev) = 206.83 Meniscus = 6.02; Bottom = 7.21 fitting = 6.05 – 7.05

8 LITERATURE 177

8 Literature

1. Claisen, L. (1912) Über Umlagerung von Phenol-allyl-äthern in C-allyl-phenole, Ber. Dtsch. Chem. Ges. 45, 3157-3166. 2. Ziegler, F. E. (1988) The Thermal, Aliphatic Claisen Rearrangement, Chem. Rev. 88, 1423- 1452. 3. Wick, A. E., Steen, K., Felix, D., and Eschenmoser, A. (1964) Claisensche Umlagerungen Bei Allyl- & Benzylalkoholen Mit Hilfe Von Acetalen des N,N-Dimethylacetamids, Helv. Chim. Acta 47, 2425-2429. 4. Johnson, W. S., Werthema.L, Bartlett, W. R., Brocksom, T. J., Li, T. T., Faulkner, D. J., and Petersen, M. R. (1970) A Simple Stereoselective Version of Claisen Rearrangement Leading to Trans-Trisubstituted Olefinic Bonds . Synthesis of Squalene, J. Am. Chem. Soc. 92, 741-743. 5. Ireland, R. E., and Mueller, R. H. (1972) Claisen Rearrangement of Allyl Esters, J. Am. Chem. Soc. 94, 5897-5898. 6. Barillier, D., Morin, L., Paquar, D., Rioult, P., Vazeux, M., and Andrieu, C. G. (1977) Action of Unsaturated Bromides on Thioketones - Claisen Thio-Rearrangement, Bul. Soc. Chim. Fr., 688- 692. 7. Heimgartner, H., Hansen, H.-J., and Schmid, H. (1979) Iminium Salts in Organic Chemistry Part 2, Wiley-Interscience, New York, 655-731. 8. Meyer, M. P., DelMonte, A. J., and Singleton, D. A. (1999) Reinvestigation of the Isotope Effects for the Claisen and Aromatic Claisen Rearrangements: The Nature of the Claisen Transition States, J. Am. Chem. Soc. 121, 10865-10874. 9. Gajewski, J. J. (1980) Energy Surfaces of Sigmatropic Shifts, Acc. Chem. Res. 13, 142-148. 10. Gajewski, J. J., and Conrad, N. D. (1979) Aliphatic Claisen Rearrangement Transition-State Structure from Secondary α-Deuterium Isotope Effects, J. Am. Chem. Soc. 101, 2747-2748. 11. Gajewski, J. J., and Conrad, N. D. (1979) Variable Transition-State Structure in 3,3-Sigmatropic Shifts from α-Secondary Deuterium-Isotope Effects, J. Am. Chem. Soc. 101, 6693-6704. 12. Hansen, H. J., and Schmid, H. (1974) Stereochemistry of Sigmatropic[3,3]-Rearrangement and Sigmatropic[5,5]-Rearrangement, Tetrahedron 30, 1959-1969. 13. Fráter, G., Habich, A., Hansen, H. J., and Schmid, H. (1969) Zur Stereochemie der aromatischen Claisen-Umlagerung, Helv. Chim. Acta 52, 335-361. 14. Coates, R. M., Rogers, B. D., Hobbs, S. J., Peck, D. R., and Curran, D. P. (1987) Synthesis and Claisen Rearrangement of Alkoxyallyl Enol Ethers - Evidence for a Dipolar Transition-State, J. Am. Chem. Soc. 109, 1160-1170. 15. Denmark, S. E., and Harmata, M. A. (1982) Carbanion-Accelerated Claisen Rearrangements, J. Am. Chem. Soc. 104, 4972-4974. 16. Gajewski, J. J., and Emrani, J. (1984) Origin of the Rate Acceleration in the Ireland-Claisen Rearrangement, J. Am. Chem. Soc. 106, 5733-5734. 17. Burrows, C. J., and Carpenter, B. K. (1981) Substituent Effects on the Aliphatic Claisen Rearrangement I: Synthesis and Rearrangement of Cyano-Substituted Allyl Vinyl Ethers, J. Am. Chem. Soc. 103, 6983-6984. 18. Burrows, C. J., and Carpenter, B. K. (1981) Substituent Effects on the Aliphatic Claisen Rearrangement II: Theoretical-Analysis, J. Am. Chem. Soc. 103, 6984-6986. 19. Gajewski, J. J., Jurayj, J., Kimbrough, D. R., Gande, M. E., Ganem, B., and Carpenter, B. K. (1987) On the Mechanism of Rearrangement of Chorismic Acid and Related-Compounds, J. Am. Chem. Soc. 109, 1170-1186. 20. Brandes, E., Grieco, P. A., and Gajewski, J. J. (1989) Effect of Polar-Solvents on the Rates of Claisen Rearrangements - Assessment of Ionic Character, J. Org. Chem. 54, 515-516.

8 LITERATURE 178

21. Gajewski, J. J., and Brichford, N. L. (1994) Secondary Deuterium Kinetic Isotope Effects in the Aqueous Claisen Rearrangement - Evidence against an Ionic Transition-State, J. Am. Chem. Soc. 116, 3165-3166. 22. Ganem, B. (1996) The Mechanism of the Claisen Rearrangement: Déjà vu all over again, Angew. Chem. Int. Ed. 35, 937-945. 23. Severance, D. L., and Jorgensen, W. L. (1992) Effects of Hydration on the Claisen Rearrangement of Allyl Vinyl Ether from Computer-Simulations, J. Am. Chem. Soc. 114, 10966-10968. 24. Dodziuk, H., Vonvoithenberg, H., and Allinger, N. L. (1982) A Molecular Mechanics Study of Methyl Vinyl Ether and Related-Compounds, Tetrahedron 38, 2811-2819. 25. Woodward, R. B., and Hoffmann, R. (1965) Selection Rules for Sigmatropic Reactions, J. Am. Chem. Soc. 87, 2511-2513. 26. Hoffmann, R., and Woodward, R. B. (1965) Orbital Symmetries and Orientational Effects in a Sigmatropic Reaction, J. Am. Chem. Soc. 87, 4389-4390. 27. Lutz, R. P. (1984) Catalysis of the Cope and Claisen Rearrangements, Chem. Rev. 84, 205-247. 28. Bartlett, P. A., and Hahne, W. F. (1979) Stereochemical Control of the Ynamine-Claisen Rearrangement, J. Org. Chem. 44, 882-883. 29. Nonoshita, K., Banno, H., Maruoka, K., and Yamamoto, H. (1990) Organoaluminum-Promoted Claisen Rearrangement of Allyl Vinyl Ethers, J. Am. Chem. Soc. 112, 316-322. 30. Maruoka, K., Saito, S., and Yamamoto, H. (1995) Molecular Design of a Chiral Lewis-Acid for the Asymmetric Claisen Rearrangement, J. Am. Chem. Soc. 117, 1165-1166. 31. Haslam, E. (1993) Shikimic Acid: Metabolism and Metabolites, John Wiley & Sons, New York. 32. Chook, Y. M., Gray, J. V., Ke, H. M., and Lipscomb, W. N. (1994) The Monofunctional Chorismate Mutase from Bacillus subtilis - Structure Determination of Chorismate Mutase and Its Complexes with a Transition-State Analog and Prephenate, and Implications for the Mechanism of the Enzymatic-Reaction, J. Mol. Biol. 240, 476-500. 33. Lee, A. Y., Karplus, P. A., Ganem, B., and Clardy, J. (1995) Atomic-Structure of the Buried Catalytic Pocket of Escherichia coli Chorismate Mutase, J. Am. Chem. Soc. 117, 3627-3628. 34. Xue, Y. F., Lipscomb, W. N., Graf, R., Schnappauf, G., and Braus, G. (1994) The Crystal- Structure of Allosteric Chorismate Mutase at 2.2-Angstrom Resolution, Proc. Natl. Acad. Sci. USA 91, 10814-10818. 35. Xue, Y. F., and Lipscomb, W. N. (1995) Location of the Active-Site of Allosteric Chorismate Mutase from Saccharomyces Cerevisiae, and Comments on the Catalytic and Regulatory Mechanisms, Proc. Natl. Acad. Sci. USA 92, 10595-10598. 36. Inagaki, E., Kuramitsu, S., Yokoyama, S., Miyano, M., and Tahirov, T. H. (2003) The Crystal Structure of Chorismate Mutase from Thermus thermophilus (1UFY and 1ODE), Direct submission to Protein Data Bank. 37. Helmstaedt, K., Heinrich, G., Merkl, R., and Braus, G. H. (2004) Chorismate Mutase of Thermus thermophilus is a Monofunctional AroH Class Enzyme Inhibited by Tyrosine, Arch. Microbiol. 181, 195-203. 38. Bartlett, P. A., Nakagawa, Y., Johnson, C. R., Reich, S. H., and Luis, A. (1988) Chorismate Mutase Inhibitors - Synthesis and Evaluation of Some Potential Transition-State Analogs, J. Org. Chem. 53, 3195-3210. 39. Mandal, A., and Hilvert, D. (2003) Charge Optimization Increases the Potency and Selectivity of a Chorismate Mutase Inhibitor, J. Am. Chem. Soc. 125, 5598-5599. 40. Romero, R. M., Roberts, M. F., and Phillipson, J. D. (1995) Chorismate Mutase in Microorganisms and Plants, Phytochemistry 40, 1015-1025. 41. Krappmann, S., Pries, R., Gellissen, G., Hiller, M., and Braus, G. H. (2000) HAR07 Encodes Chorismate Mutase of the Methylotrophic Yeast Hansenula polymorpha and is Derepressed upon Methanol Utilization, J. Bacteriol. 182, 4188-4197. 42. Krappmann, S., Helmstaedt, K., Gerstberger, T., Eckert, S., Hoffmann, B., Hoppert, M., Schnappauf, G., and Braus, G. H. (1999) The aroC Gene of Aspergillus nidulans Codes for a Monofunctional, Allosterically Regulated Chorismate Mutase, J. Biol. Chem. 274, 22275-22282. 43. Görisch, H. (1978) Mechanism of Chorismate Mutase Reaction, Biochemistry 17, 3700-3705. 44. Chook, Y. M., Ke, H. M., and Lipscomb, W. N. (1993) Crystal-Structures of the Monofunctional Chorismate Mutase from Bacillus subtilis and Its Complex with a Transition- State Analog, Proc. Natl. Acad. Sci. USA 90, 8600-8603.

8 LITERATURE 179

45. Lee, A. Y., Stewart, J. D., Clardy, J., and Ganem, B. (1995) New Insight into the Catalytic Mechanism of Chorismate Mutases from Structural Studies, Chem. Biol. 2, 195-203. 46. MacBeath, G., Kast, P., and Hilvert, D. (1998) A Small, Thermostable, and Monofunctional Chorismate Mutase from the Archeon Methanococcus jannaschii, Biochemistry 37, 10062- 10073. 47. Sasso, S., Ramakrishnan, C., Gamper, M., Hilvert, D., and Kast, P. (2005) Characterization of the Secreted Chorismate Mutase from the Pathogen Mycobacterium tuberculosis, Febs J. 272, 375-389. 48. Helmstaedt, K., Heinrich, G., Lipscomb, W. N., and Braus, G. H. (2002) Refined Molecular Hinge Between Allosteric and Catalytic Domain Determines Allosteric Regulation and Stability of Fungal Chorismate Mutase, Proc. Natl. Acad. Sci. USA 99, 6631-6636. 49. Helmstaedt, K., Krappmann, S., and Braus, G. H. (2001) Allosteric Regulation of Catalytic Activity: Escherichia coli Aspartate Transcarbamoylase versus Yeast Chorismate Mutase, Microbiol. Mol. Biol. Rev. 65, 404-419. 50. Cload, S. T., Liu, D. R., Pastor, R. M., and Schultz, P. G. (1996) Mutagenesis Study of Active Site Residues in Chorismate Mutase from Bacillus subtilis, J. Am. Chem. Soc. 118, 1787-1788. 51. Ladner, J. E., Reddy, P., Davis, A., Tordova, M., Howard, A. J., and Gilliland, G. L. (2000) The 1.30 Ångstrøm Resolution Structure of the Bacillus subtilis Chorismate Mutase Catalytic Homotrimer, Acta Crystallogr. Sect. D-Biol. Crystallogr. 56, 673-683. 52. Lee, Y. S., Worthington, S. E., Krauss, M., and Brooks, B. R. (2002) Reaction Mechanism of Chorismate Mutase Studied by the Combined Potentials of Quantum Mechanics and Molecular Mechanics, J. Phys. Chem. B 106, 12059-12065. 53. Ranaghan, K. E., Ridder, L., Szefczyk, B., Sokalski, W. A., Hermann, J. C., and Mulholland, A. J. (2004) Transition State Stabilization and Substrate Strain in Enzyme Catalysis: Ab Initio QM/MM Modelling of the Chorismate Mutase Reaction, Org. Biomol. Chem. 2, 968-980. 54. Szefczyk, B., Mulholland, A. J., Ranaghan, K. E., and Sokalski, W. A. (2004) Differential Transition-State Stabilization in Enzyme Catalysis: Quantum Chemical Analysis of Interactions in the Chorismate Mutase Reaction and Prediction of the Optimal Catalytic Field, J. Am. Chem. Soc. 126, 16148-16159. 55. Kast, P., Hartgerink, J. D., AsifUllah, M., and Hilvert, D. (1996) Electrostatic Catalysis of the Claisen Rearrangement: Probing the Role of Glu78 in Bacillus subtilis Chorismate Mutase by Genetic Selection, J. Am. Chem. Soc. 118, 3069-3070. 56. Guo, H., Cui, Q., Lipscomb, W. N., and Karplus, M. (2003) Understanding the Role of Active- Site Residues in Chorismate Mutase Catalysis from Molecular-Dynamics Simulations, Angew. Chem. Int. Ed. 42, 1508-1511. 57. Garcia-Viloca, M., Gao, J., Karplus, M., and Truhlar, D. G. (2004) How Enzymes Work: Analysis by Modern Rate Theory and Computer Simulations, Science 303, 186-195. 58. Liu, D. R., Cload, S. T., Pastor, R. M., and Schultz, P. G. (1996) Analysis of Active Site Residues in Escherichia coli Chorismate Mutase by Site-Directed Mutagenesis, J. Am. Chem. Soc. 118, 1789-1790. 59. Zhang, S., Pohnert, G., Kongsaeree, P., Wilson, D. B., Clardy, J., and Ganem, B. (1998) Chorismate Mutase-Prephenate Dehydratase from Escherichia coli - Study of Catalytic and Regulatory Domains using Genetically Engineered Proteins, J. Biol. Chem. 273, 6248-6253. 60. Zhang, S., Kongsaeree, P., Clardy, J., Wilson, D. B., and Ganem, B. (1996) Site-Directed Mutagenesis of Monofunctional Chorismate Mutase Engineered from the E. coli P-protein, Bioorg. Med. Chem. 4, 1015-1020. 61. Ramakrishnan, C. (2001) Untersuchungen durch evolutionäre Strategien von Mechanismus und Topologie einer ungewöhnlichen Chorismatmutase der AroQ- Klasse aus dem Pathogen Mycobacterium tuberculosis, Diplomarbeit, ETH Zürich. 62. Galopin, C. C., Zhang, S., Wilson, D. B., and Ganem, B. (1996) On the Mechanism of Chorismate Mutases: Clues from Wild-Type E. coli Enzyme and a Site-Directed Mutant Related to Yeast Chorismate Mutase, Tetrahedron Lett. 37, 8675-8678. 63. Schmidheini, T., Mosch, H. U., Evans, J. N. S., and Braus, G. (1990) Yeast Allosteric Chorismate Mutase Is Locked in the Activated State by a Single Amino-Acid Substitution, Biochemistry 29, 3660-3668. 64. Schnappauf, G., Strater, N., Lipscomb, W. N., and Braus, G. H. (1997) A Glutamate Residue in the Catalytic Center of the Yeast Chorismate Mutase Restricts Enzyme Activity to Acidic Conditions, Proc. Natl. Acad. Sci. USA 94, 8491-8496.

8 LITERATURE 180

65. Sträter, N., Schnappauf, G., Braus, G., and Lipscomb, W. N. (1997) Mechanisms of Catalysis and Allosteric Regulation of Yeast Chorismate Mutase from Crystal Structures, Structure 5, 1437-1452. 66. Haynes, M. R., Stura, E. A., Hilvert, D., and Wilson, I. A. (1994) Routes to Catalysis - Structure of a Catalytic Antibody and Comparison with Its Natural Counterpart, Science 263, 646-652. 67. Hilvert, D., and Nared, K. D. (1988) Stereospecific Claisen Rearrangement Catalyzed by an Antibody, J. Am. Chem. Soc. 110, 5593-5594. 68. Hilvert, D., Carpenter, S. H., Nared, K. D., and Auditor, M. T. M. (1988) Catalysis of Concerted Reactions by Antibodies - the Claisen Rearrangement, Proc. Natl. Acad. Sci. USA 85, 4953- 4955. 69. Addadi, L., Jaffe, E. K., and Knowles, J. R. (1983) Secondary Tritium Isotope Effects as Probes of the Enzymic and Non-Enzymic Conversion of Chorismate to Prephenate, Biochemistry 22, 4494-4501. 70. Gustin, D. J., Mattei, P., Kast, P., Wiest, O., Lee, L., Cleland, W. W., and Hilvert, D. (1999) Heavy Atom Isotope Effects Reveal a Highly Polarized Transition State for Chorismate Mutase, J. Am. Chem. Soc. 121, 1756-1757. 71. Wiest, O., and Houk, K. N. (1994) On the Transition-State of the Chorismate-Prephenate Rearrangement, J. Org. Chem. 59, 7582-7584. 72. Copley, S. D., and Knowles, J. R. (1987) The Conformational Equilibrium of Chorismate in Solution - Implications for the Mechanism of the Nonenzymatic and the Enzyme-Catalyzed Rearrangement of Chorismate to Prephenate, J. Am. Chem. Soc. 109, 5008-5013. 73. Copley, S. D., and Knowles, J. R. (1985) The Uncatalyzed Claisen Rearrangement of Chorismate to Prephenate Prefers a Transition-State of Chairlike Geometry, J. Am. Chem. Soc. 107, 5306-5308. 74. Sogo, S. G., Widlanski, T. S., Hoare, J. H., Grimshaw, C. E., Berchtold, G. A., and Knowles, J. R. (1984) Stereochemistry of the Rearrangement of Chorismate to Prephenate - Chorismate Mutase Involves a Chair Transition-State, J. Am. Chem. Soc. 106, 2701-2703. 75. Guilford, W. J., Copley, S. D., and Knowles, J. R. (1987) On the Mechanism of the Chorismate Mutase Reaction, J. Am. Chem. Soc. 109, 5013-5019. 76. Turnbull, J., Cleland, W. W., and Morrison, J. F. (1991) Ph Dependency of the Reactions Catalyzed by Chorismate Mutase Prephenate Dehydrogenase from Escherichia coli, Biochemistry 30, 7777-7782. 77. Westheimer, F. H. (1962) Mechanisms Related to Enzyme Catalysis, Adv. Enzymol. Relat. Subj. Biochem. 24, 441-482. 78. Kast, P., AsifUllah, M., and Hilvert, D. (1996) Is Chorismate Mutase a Prototypic Entropy Trap? Activation Parameters for the Bacillus subtilis Enzyme, Tetrahedron Lett. 37, 2691-2694. 79. Galopin, C. C., Zhang, S., Wilson, D. B., and Ganem, B. (1997) Corrigenda for: On the Mechanism of Chorismate Mutases: Clues from Wild-type E. coli Enzyme and a Site-directed Mutant Related to Yeast Chorismate Mutase (vol 37, page 8675, 1996), Tetrahedron Lett. 38, 1467-1467. 80. Gray, J. V., and Knowles, J. R. (1994) Monofunctional Chorismate Mutase from Bacillus subtilis - FTIR Studies and the Mechanism of Action of the Enzyme, Biochemistry 33, 9953- 9959. 81. Gamper, M., Hilvert, D., and Kast, P. (2000) Probing the Role of the C-terminus of Bacillus subtilis Chorismate Mutase by a Novel Random Protein-Termination Strategy, Biochemistry 39, 14087-14094. 82. Kast, P., AsifUllah, M., Jiang, N., and Hilvert, D. (1996) Exploring the Active Site of Chorismate Mutase by Combinatorial Mutagenesis and Selection: The Importance of Electrostatic Catalysis, Proc. Natl. Acad. Sci. USA 93, 5043-5048. 83. Kast, P., Grisostomi, C., Chen, I. A., Li, S. L., Krengel, U., Xue, Y. F., and Hilvert, D. (2000) A Strategically Positioned Cation is Crucial for Efficient Catalysis by Chorismate Mutase, J. Biol. Chem. 275, 36832-36838. 84. Strajbl, M., Shurki, A., Kato, M., and Warshel, A. (2003) Apparent NAC Effect in Chorismate Mutase Reflects Electrostatic Transition State Stabilization, J. Am. Chem. Soc. 125, 10228- 10237. 85. Crespo, A., Scherlis, D. A., Marti, M. A., Ordejon, P., Roitberg, A. E., and Estrin, D. A. (2003) A DFT-Based QM-MM Approach Designed for the Treatment of Large Molecular Systems: Application to Chorismate Mutase, J. Phys. Chem. B 107, 13728-13736.

8 LITERATURE 181

86. Lyne, P. D., Mulholland, A. J., and Richards, W. G. (1995) Insights into Chorismate Mutase Catalysis from a Combined QM/MM Simulation of the Enzyme Reaction, J. Am. Chem. Soc. 117, 11345-11350. 87. Marti, S., Andres, J., Moliner, V., Silla, E., Tunon, I., and Bertran, J. (2003) Preorganization and Reorganization as Related Factors in Enzyme Catalysis: The Chorismate Mutase Case, Chem.- Eur. J. 9, 984-991. 88. Marti, S., Andres, J., Moliner, V., Silla, E., Tunon, I., and Bertran, J. (2001) Transition Structure Selectivity in Enzyme Catalysis: A QM/MM Study of Chorismate Mutase, Theor. Chem. Acc. 105, 207-212. 89. Marti, S., Andres, J., Moliner, V., Silla, E., Tunon, I., and Bertran, J. (2004) A Comparative Study of Claisen and Cope Rearrangements Catalyzed by Chorismate Mutase. An Insight into Enzymatic Efficiency: Transition State Stabilization or Substrate Preorganization?, J. Am. Chem. Soc. 126, 311-319. 90. Guo, H., Cui, Q., Lipscomb, W. N., and Karplus, M. (2001) Substrate Conformational Transitions in the Active Site of Chorismate Mutase: Their Role in the Catalytic Mechanism, Proc. Natl. Acad. Sci. USA 98, 9032-9037. 91. Guimaraes, C. R. W., Repasky, M. P., Chandrasekhar, J., Tirado-Rives, J., and Jorgensen, W. L. (2003) Contributions of Conformational Compression and Preferential Transition State Stabilization to the Rate Enhancement by Chorismate Mutase, J. Am. Chem. Soc. 125, 6892- 6899. 92. Hur, S., and Bruice, T. C. (2002) The Mechanism of Catalysis of the Chorismate to Prephenate Reaction by the Escherichia coli Mutase Enzyme, Proc. Natl. Acad. Sci. USA 99, 1176-1181. 93. Hur, S., and Bruice, T. C. (2003) The Near Attack Conformation Approach to the Study of the Chorismate to Prephenate Reaction, Proc. Natl. Acad. Sci. USA 100, 12015-12020. 94. Hur, S., and Bruice, T. C. (2003) Just a Near Attack Conformer for Catalysis (Chorismate to Prephenate Rearrangements in Water, Antibody, Enzymes, and their Mutants), J. Am. Chem. Soc. 125, 10540-10542. 95. Hur, S., and Bruice, T. C. (2003) Comparison of Formation of Reactive Conformers (NACs) for the Claisen Rearrangement of Chorismate to Prephenate in Water and in the E. coli Mutase: The Efficiency of the Enzyme Catalysis, J. Am. Chem. Soc. 125, 5964-5972. 96. Hur, S., and Bruice, T. C. (2003) Enzymes do what is Expected (Chalcone Isomerase versus Chorismate Mutase), J. Am. Chem. Soc. 125, 1472-1473. 97. Warshel, A. (1998) Electrostatic Origin of the Catalytic Power of Enzymes and the Role of Preorganized Active Sites, J. Biol. Chem. 273, 27035-27038. 98. Khanjin, N. A., Snyder, J. P., and Menger, F. M. (1999) Mechanism of Chorismate Mutase: Contribution of Conformational Restriction to Catalysis in the Claisen Rearrangement, J. Am. Chem. Soc. 121, 11831-11846. 99. Worthington, S. E., Roitberg, A. E., and Krauss, M. (2001) An MD/QM Study of the Chorismate Mutase-Catalyzed Claisen Rearrangement Reaction, J. Phys. Chem. B 105, 7087- 7095. 100. Cornish, V. W., Mendel, D., and Schultz, P. G. (1995) Probing Protein-Structure and Function with an Expanded Genetic-Code, Angew. Chem. Int. Ed. 34, 621-633. 101. Srinivasan, G., James, C. M., and Krzycki, J. A. (2002) Pyrrolysine Encoded by UAG in Archaea: Charging of a UAG-Decoding Specialized tRNA, Science 296, 1459-1462. 102. Hao, B., Gong, W. M., Ferguson, T. K., James, C. M., Krzycki, J. A., and Chan, M. K. (2002) A new UAG-Encoded Residue in the Structure of a Methanogen Methyltransferase, Science 296, 1462-1466. 103. Fenske, C., Palm, G. J., and Hinrichs, W. (2003) How Unique is the Genetic Code?, Angew. Chem. Int. Ed. 42, 606-610. 104. Robertson, S. A., Ellman, J. A., and Schultz, P. G. (1991) A General and Efficient Route for Chemical Aminoacylation of Transfer-RNAs, J. Am. Chem. Soc. 113, 2722-2729. 105. Noren, C. J., Anthonycahill, S. J., Suich, D. J., Noren, K. A., Griffith, M. C., and Schultz, P. G. (1990) Invitro Suppression of an Amber Mutation by a Chemically Aminoacylated Transfer- RNA Prepared by Runoff Transcription, Nucleic Acids Res. 18, 83-88. 106. Milligan, J. F., Groebe, D. R., Witherell, G. W., and Uhlenbeck, O. C. (1987) Oligoribonucleotide Synthesis Using T7 RNA-Polymerase and Synthetic DNA Templates, Nucleic Acids Res. 15, 8783-8798.

8 LITERATURE 182

107. Stiege, W., and Erdmann, V. A. (1995) The Potentials of the in-Vitro Protein-Biosynthesis System, J. Biotechnol. 41, 81-90. 108. Jermutus, L., Ryabova, L. A., and Plückthun, A. (1998) Recent Advances in Producing and Selecting Functional Proteins by Using Cell-Free Translation, Curr. Opin. Biotechnol. 9, 534- 548. 109. Hames, B. D., and Higgins, S. J. (1984) Transcription and Translation: A Pratical Approach, IRL, Oxford. 110. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K., and Ueda, T. (2001) Cell-free Translation Reconstituted with Purified Components, Nature Biotech. 19, 751- 755. 111. Cook, S. N., Jack, W. E., Xiong, X. F., Danley, L. E., Ellman, J. A., Schultz, P. G., and Noren, C. J. (1995) Photochemically-Initiated Protein Splicing, Protein Eng. 8, 87-87. 112. Cook, S. N., Jack, W. E., Xiong, X. F., Danley, L. E., Ellman, J. A., Schultz, P. G., and Noren, C. J. (1995) Photochemically-Initiated Protein Splicing, Angew. Chem. Int. Ed. 34, 1629-1630. 113. Thorson, J. S., Chapman, E., Murphy, E. C., Schultz, P. G., and Judice, J. K. (1995) Linear Free- Energy Analysis of Hydrogen-Bonding in Proteins, J. Am. Chem. Soc. 117, 1157-1158. 114. Mendel, D., Ellman, J. A., Chang, Z. Y., Veenstra, D. L., Kollman, P. A., and Schultz, P. G. (1992) Probing Protein Stability with Unnatural Amino-Acids, Science 256, 1798-1802. 115. Judice, J. K., Gamble, T. R., Murphy, E. C., Devos, A. M., and Schultz, P. G. (1993) Probing the Mechanism of Staphylococcal Nuclease with Unnatural Amino-Acids - Kinetic and Structural Studies, Science 261, 1578-1581. 116. Mendel, D., Ellman, J., and Schultz, P. G. (1993) Protein-Biosynthesis with Conformationally Restricted Amino-Acids, J. Am. Chem. Soc. 115, 4359-4360. 117. Wang, L., Brock, A., Herberich, B., and Schultz, P. G. (2001) Expanding the Genetic Code of Escherichia coli, Science 292, 498-500. 118. Wang, L., and Schultz, P. G. (2002) Expanding the Genetic Code, Chem. Commun., 1-11. 119. Wang, L., and Schultz, P. G. (2001) A General Approach for the Generation of Orthogonal tRNAs, Chem. Biol. 8, 883-890. 120. Wang, L., Brock, A., and Schultz, P. G. (2002) Adding L-3-(2-Naphthyl)alanine to the Genetic Code of E. coli, J. Am. Chem. Soc. 124, 1836-1837. 121. Santoro, S. W., Wang, L., Herberich, B., King, D. S., and Schultz, P. G. (2002) An Efficient System for the Evolution of Aminoacyl-tRNA Synthetase Specificity, Nature Biotech. 20, 1044- 1048. 122. Wang, L., Xie, J. M., Deniz, A. A., and Schultz, P. G. (2003) Unnatural Amino Acid Mutagenesis of Green Fluorescent Protein, J. Org. Chem. 68, 174-176. 123. Merrifield, R. B. (1963) Solid Phase Peptide Synthesis 1: Synthesis of a Tetrapeptide, J. Am. Chem. Soc. 85, 2149-2154. 124. Dawson, P. E., and Kent, S. B. H. (2000) Synthesis of Native Proteins by Chemical Ligation, Annu. Rev. Biochem. 69, 923-960. 125. Miranda, L. P., and Alewood, P. F. (2000) Challenges for Protein Chemical Synthesis in the 21st Century: Bridging Genomics and Proteomics, Biopolymers 55, 217-226. 126. Kent, S. B. H. (1988) Chemical Synthesis of Peptides and Proteins, Annu. Rev. Biochem. 57, 957-989. 127. Fields, G. B., and Noble, R. L. (1990) Solid-Phase Peptide-Synthesis Utilizing 9- Fluorenylmethoxycarbonyl Amino-Acids, Int. J. Pept. Prot. Res. 35, 161-214. 128. Chan, W. C., and White, P. D. (2000) Fmoc Solid Phase Peptide Synthesis: A Practical Approach, Oxford University Press, Oxford. 129. Bayer, E. (1991) Towards the Chemical Synthesis of Proteins, Angew. Chem. Int. Ed. 30, 113- 129. 130. Bodanszky, M. (1993) Principles of Peptide Synthesis (2nd edition), Springer Verlag, New York. 131. Bodanszky, M. (1994) The Practice of Peptide Synthesis (2nd edition), Springer Verlag, New York. 132. Hiemstra, H. S., Benckhuijsen, W. E., Amons, R., Rapp, W., and Drijfhout, J. W. (1998) A new Hybrid Resin for Stepwise Screening of Peptide Libraries Combined with Single Bead Edman Sequencing, J. Pept. Sci. 4, 282-288.

8 LITERATURE 183

133. Kates, S. A., McGuinness, B. F., Blackburn, C., Griffin, G. W., Sole, N. A., Barany, G., and Albericio, F. (1998) "High-Load" Polyethylene Glycol-Polystyrene (PEG-PS) Graft Supports for Solid-Phase Synthesis, Biopolymers 47, 365-380. 134. Wang, S. S. (1973) Para-Alkoxybenzyl Alcohol Resin and Para- Alkoxybenzyloxycarbonylhydrazide Resin for Solid-Phase Synthesis of Protected Peptide Fragments, J. Am. Chem. Soc. 95, 1328-1333. 135. McKay, F. C., and Albertson, N. F. (1957) New Amine-Masking Groups for Peptide Synthesis, J. Am. Chem. Soc. 79, 4686-4690. 136. Anderson, G. W., and McGregor, A. C. (1957) Tert-Butyloxycarbonylamino Acids and Their Use in Peptide Synthesis, J. Am. Chem. Soc. 79, 6180-6183. 137. Schnölzer, M., Alewood, P., Jones, A., Alewood, D., and Kent, S. B. H. (1992) In Situ Neutralization in Boc-Chemistry Solid-Phase Peptide-Synthesis - Rapid, High-Yield Assembly of Difficult Sequences, Int. J. Pept. Prot. Res. 40, 180-193. 138. Schneider, J., and Kent, S. B. H. (1988) Enzymatic-Activity of a Synthetic 99 Residue Protein Corresponding to the Putative HIV-1 Protease, Cell 54, 363-368. 139. Chang, C. D., and Meienhofer, J. (1978) Solid-Phase Peptide-Synthesis Using Mild Base Cleavage of Nα-Fluorenylmethyloxycarbonylamino Acids, Exemplified by a Synthesis of Dihydrosomatostatin, Int. J. Pept. Prot. Res. 11, 246-249. 140. Meienhofer, J., Waki, M., Heimer, E. P., Lambros, T. J., Makofske, R. C., and Chang, C. D. (1979) Solid-Phase Synthesis without Repetitive Acidolysis, Int. J. Pept. Prot. Res. 13, 35-42. 141. Sarin, V. K., Kent, S. B. H., Tam, J. P., and Merrifield, R. B. (1981) Quantitative Monitoring of Solid-Phase Peptide-Synthesis by the Ninhydrin Reaction, Anal. Biochem. 117, 147-157. 142. Robertson, N., and Ramage, R. (1999) Total Chemical Synthesis of Deglycosylated Human Erythropoietin, J. Chem. Soc. Perkin Trans. 1, 1015-1021. 143. Nicolaou, K. C., Vourloumis, D., Winssinger, N., and Baran, P. S. (2000) The Art and Science of Total Synthesis at the Dawn of the Twenty-First Century, Angew. Chem. Int. Ed. 39, 44-122. 144. Sakakibara, S. (1995) Synthesis of Large Peptides in Solution, Biopolymers 37, 17-28. 145. Jones, J. (1991) The Chemical Synthesis of Peptides, Clarendon Press, Oxford. 146. Kiyama, S., Fujii, N., Yajima, H., Moriga, M., and Takagi, A. (1984) Studies on Peptides 113: Synthesis of the Heptacosapeptide Amide Corresponding to the Entire Amino-Acid-Sequence of Chicken Secretin, Int. J. Pept. Prot. Res. 23, 174-186. 147. Carreno, C., Carrascosa, A. L., Vinuela, E., Giralt, E., and Andreu, D. (1995) Peptides 1994, ESCOM Science Publisher B. V., Netherlands. 148. Abrahmsen, L., Tom, J., Burnier, J., Butcher, K. A., Kossiakoff, A., and Wells, J. A. (1991) Engineering Subtilisin and Its Substrates for Efficient Ligation of Peptide-Bonds in Aqueous- Solution, Biochemistry 30, 4151-4159. 149. Wells, J. A., Chang, T., and Abrahmsen, L. (1992) Subtiligase - an Enzyme Designed for Semisynthesis of Proteins, Biochemistry 31, 2191-2191. 150. Welker, E., and Scheraga, H. A. (1999) Use of Benzyl Mercaptan for Direct Preparation of Long Polypeptide Benzylthio Esters as Substrates of Subtiligase, Biochem. Bioph. Res. Commun. 254, 147-151. 151. Braisted, A. C., Judice, J. K., and Wells, J. A. (1997) Synthesis of Proteins by Subtiligase, Methods Enzymol. 289, 298-313. 152. Suich, D. J., Ballinger, M. D., Wells, J. A., and DeGrado, W. F. (1996) Fmoc-Based Synthesis of Glycolate Ester Peptides for the Assembly of de Novo Designed Multimeric Proteins using Subtiligase, Tetrahedron Lett. 37, 6653-6656. 153. Chang, T. K., Jackson, D. Y., Burnier, J. P., and Wells, J. A. (1994) Subtiligase - a Tool for Semisynthesis of Proteins, Proc. Natl. Acad. Sci. USA 91, 12544-12548. 154. Geoghegan, K. F., and Stroh, J. G. (1992) Site-Directed Conjugation of Nonpeptide Groups to Peptides and Proteins Via Periodate-Oxidation of a 2-Amino Alcohol - Application to Modification at N-Terminal Serine, Bioconj. Chem. 3, 138-146. 155. Gaertner, H. F., Rose, K., Cotton, R., Timms, D., Camble, R., and Offord, R. E. (1992) Construction of Protein Analogs by Site-Specific Condensation of Unprotected Fragments, Bioconj. Chem. 3, 262-268. 156. Gaertner, H. F., Offord, R. E., Cotton, R., Timms, D., Camble, R., and Rose, K. (1994) Chemoenzymatic Backbone Engineering of Proteins - Site-Specific Incorporation of Synthetic Peptides That Mimic the 64-74-Disulfide Loop of Granulocyte-Colony-Stimulating Factor, J. Biol. Chem. 269, 7224-7230.

8 LITERATURE 184

157. Rose, K. (1994) Facile Synthesis of Homogeneous Artificial Proteins, J. Am. Chem. Soc. 116, 30-33. 158. Schnölzer, M., and Kent, S. B. H. (1992) Constructing Proteins by Dovetailing Unprotected Synthetic Peptides - Backbone-Engineered Hiv Protease, Science 256, 221-225. 159. Milton, R. C. D., Milton, S. C. F., and Kent, S. B. H. (1992) Total Chemical Synthesis of a D- Enzyme - the Enantiomers of HIV-1 Protease Show Demonstration of Reciprocal Chiral Substrate-Specificity, Science 256, 1445-1448. 160. Nilsson, B. L., Hondal, R. J., Soellner, M. B., and Raines, R. T. (2003) Protein Assembly by Orthogonal Chemical Ligation Methods, J. Am. Chem. Soc. 125, 5268-5269. 161. Dawson, P. E., Muir, T. W., Clarklewis, I., and Kent, S. B. H. (1994) Synthesis of Proteins by Native Chemical Ligation, Science 266, 776-779. 162. Mandolini, L. (1978) Ring-Closure Reactions .11. Activation Parameters for Formation of 4- Membered to 6-Membered Lactones from Omega-Bromoalkanoate Ions - Role of Entropy Factor in Small-Ring and Common-Ring Formation, J. Am. Chem. Soc. 100, 550-554. 163. Wieland, T., Bokelmann, E., Bauer, L., Lang, H. U., and Lau, H. (1953) Über Peptidsynthesen, 8. Mitteilung: Bildung von S-haltigen Peptiden durch intramolekulare Wanderung von Aminoacylresten, Liebigs Ann. Chem. 583, 129-149. 164. Erlanson, D. A., Chytil, M., and Verdine, G. L. (1996) The Leucine Zipper Domain Controls the Orientation of AP-1 in the NFAT AP-1 DNA Domplex, Chem. Biol. 3, 981-991. 165. Southworth, M. W., Amaya, K., Evans, T. C., Xu, M. Q., and Perler, F. B. (1999) Purification of Proteins Fused to Either the Amino or Carboxy Terminus of the Mycobacterium Xenopi Gyrase A Intein, Biotechniques 27, 110-120. 166. Evans, T. C., Benner, J., and Xu, M. Q. (1999) The in Vitro Ligation of Bacterially Expressed Proteins Using an Intein From Methanobacterium Thermoautotrophicum, J. Biol. Chem. 274, 3923-3926. 167. Mathys, S., Evans, T. C., Chute, I. C., Wu, H., Chong, S. R., Benner, J., Liu, X. Q., and Xu, M. Q. (1999) Characterization of a Self-Splicing Mini-Intein and its Conversion into Autocatalytic N- and C-terminal Cleavage Elements: Facile Production of Protein Building Blocks for Protein Ligation, Gene 231, 1-13. 168. Xu, R., Ayers, B., Cowburn, D., and Muir, T. W. (1999) Chemical Ligation of Folded Recombinant Proteins: Segmental Isotopic Labeling of Domains for NMR Studies, Proc. Natl. Acad. Sci. USA 96, 388-393. 169. Cotton, G. J., and Muir, T. W. (2000) Generation of a Dual-Labeled Fluorescence Biosensor for Crk-II Phosphorylation Using Solid-Phase Expressed Protein Ligation, Chem. Biol. 7, 253-261. 170. Muir, T. W., Sondhi, D., and Cole, P. A. (1998) Expressed Protein Ligation: A General Method for Protein Engineering, Proc. Natl. Acad. Sci. USA 95, 6705-6710. 171. Severinov, K., and Muir, T. W. (1998) Expressed Protein Ligation, a Novel Method for Studying Protein-Protein Interactions in Transcription, J. Biol. Chem. 273, 16205-16209. 172. Evans, T. C., Jr., Benner, J., and Xu, M. Q. (1998) Semisynthesis of Cytotoxic Proteins Using a Modified Protein Splicing Element, Protein Sci. 7, 2256-2264. 173. Li, X. Q., Kawakami, T., and Aimoto, S. (1998) Direct Preparation of Peptide Thioesters Using an Fmoc Solid-Phase Method, Tetrahedron Lett. 39, 8669-8672. 174. Jensen, K. J., Alsina, J., Songster, M. F., Vagner, J., Albericio, F., and Barany, G. (1998) Backbone Amide Linker (BAL) Strategy for Solid-Phase Synthesis of C-terminal-Modified and Cyclic Peptides, J. Am. Chem. Soc. 120, 5441-5452. 175. Backes, B. J., Virgilio, A. A., and Ellman, J. A. (1996) Activation Method to Prepare a Highly Reactive Acylsulfonamide ''Safety-Catch'' Linker for Solid-Phase Synthesis, J. Am. Chem. Soc. 118, 3055-3056. 176. Swinnen, D., and Hilvert, D. (2000) Facile, Fmoc-Compatible Solid-Phase Synthesis of Peptide C-terminal Thioesters, Org. Lett. 2, 2439-2442. 177. Sewing, A., and Hilvert, D. (2001) Fmoc-Compatible Solid-Phase Peptide Synthesis of Long C- Terminal Peptide Thioesters, Angew. Chem. Int. Ed. 40, 3395-3396. 178. Low, D. W., Hill, M. G., Carrasco, M. R., Kent, S. B. H., and Botti, P. (2001) Total Synthesis of Cytochrome b562 by Native Chemical Ligation Using a Removable Auxiliary, Proc. Natl. Acad. Sci. USA 98, 6554-6559. 179. Low, D. W., and Hill, M. G. (1998) Rational Fine-Tuning of the Redox Potentials in Chemically Synthesized Rubredoxins, J. Am. Chem. Soc. 120, 11536-11537.

8 LITERATURE 185

180. Baca, M., Muir, T. W., Schnolzer, M., and Kent, S. B. H. (1995) Chemical Ligation of Cysteine- Containing Peptides - Synthesis of a 22-kDa Tethered Dimer of Hiv-1 Protease, J. Am. Chem. Soc. 117, 1881-1887. 181. Kimmerlin, T., Seebach, D., and Hilvert, D. (2002) Synthesis of β(3)-Peptides and Mixed α/β(3)-Peptides by Thioligation, Helv. Chim. Acta 85, 1812-1826. 182. Shin, Y., Winans, K. A., Backes, B. J., Kent, S. B. H., Ellman, J. A., and Bertozzi, C. R. (1999) Fmoc-Based Synthesis of Peptide-α-thioesters: Application to the Total Chemical Synthesis of a Glycoprotein by Native Chemical Ligation, J. Am. Chem. Soc. 121, 11684-11689. 183. Quaderer, R., Sewing, A., and Hilvert, D. (2001) Selenocysteine-Mediated Native Chemical Ligation, Helv. Chim. Acta 84, 1197-1206. 184. Hondal, R. J., Nilsson, B. L., and Raines, R. T. (2001) Selenocysteine in Native Chemical Ligation and Expressed Protein Ligation, J. Am. Chem. Soc. 123, 5140-5141. 185. Hondal, R. J., Nilsson, B. L., and Raines, R. T. (2001) Selenocysteine in Native Chemical Ligation and Expressed Protein Ligation, Biochemistry 40, 8608-8609. 186. Gieselman, M. D., Xie, L. L., and van der Donk, W. A. (2001) Synthesis of a Selenocysteine- Containing Peptide by Native Chemical Ligation, Org. Lett. 3, 1331-1334. 187. Yan, L. Z., and Dawson, P. E. (2001) Synthesis of Peptides and Proteins without Cysteine Residues by Native Chemical Ligation Combined with Desulfurization, J. Am. Chem. Soc. 123, 526-533. 188. Canne, L. E., Bark, S. J., and Kent, S. B. H. (1996) Extending the Applicability of Native Chemical Ligation, J. Am. Chem. Soc. 118, 5891-5896. 189. Offer, J., Boddy, C. N. C., and Dawson, P. E. (2002) Extending Synthetic Access to Proteins with a Removable Acyl Transfer Auxiliary, J. Am. Chem. Soc. 124, 4642-4646. 190. Offer, J., and Dawson, P. E. (2000) N-α-2-Mercaptobenzylamine-Assisted Chemical Ligation, Org. Lett. 2, 23-26. 191. Noren, C. J., Wang, J. M., and Perler, F. B. (2000) Dissecting the Chemistry of Protein Splicing and its Applications, Angew. Chem. Int. Ed. 39, 450-466. 192. Belfort, M., and Perlman, P. S. (1995) Mechanisms of Intron Mobility, J. Biol. Chem. 270, 30237-30240. 193. Belfort, M., and Roberts, R. J. (1997) Homing Endonucleases: Keeping the House in Order, Nucleic Acids Res. 25, 3379-3388. 194. Klabunde, T., Sharma, S., Telenti, A., Jacobs, W. R., and Sacchettini, J. C. (1998) Crystal Structure of GyrA Intein from Mycobacterium Xenopi Reveals Structural Basis of Protein Splicing, Nat. Struct. Biol. 5, 31-36. 195. Perler, F. B. (1999) InBase, the New England Biolabs Intein Database, Nucleic Acids Res. 27, 346-347. 196. Xu, M. Q., and Perler, F. B. (1996) The Mechanism of Protein Splicing and its Modulation by Mutation, Embo J. 15, 5146-5153. 197. Southworth, M. W., Adam, E., Panne, D., Byer, R., Kautz, R., and Perler, F. B. (1998) Control of protein splicing by intein fragment reassembly, Embo J. 17, 918-926. 198. Mills, K. V., Lew, B. M., Jiang, S. Q., and Paulus, H. (1998) Protein Splicing in Trans by Purified N- and C-Terminal Fragments of the Mycobacterium tuberculosis RecA Intein, Proc. Natl. Acad. Sci. USA 95, 3543-3548. 199. Lew, B. M., Mills, K. V., and Paulus, H. (1998) Protein Splicing in Vitro with a Semisynthetic Two-component Minimal Intein, J. Biol. Chem. 273, 15887-15890. 200. Wu, H., Hu, Z. M., and Liu, X. Q. (1998) Protein trans-Splicing by a Split Intein Encoded in a Split DnaE Gene of Synechocystis sp. PCC6803, Proc. Natl. Acad. Sci. USA 95, 9226-9231. 201. Yamazaki, T., Otomo, T., Oda, N., Kyogoku, Y., Uegaki, K., Ito, N., Ishino, Y., and Nakamura, H. (1998) Segmental Isotope Labeling for Protein NMR using Peptide Splicing, J. Am. Chem. Soc. 120, 5591-5592. 202. Otomo, T., Ito, N., Kyogoku, Y., and Yamazaki, T. (1999) NMR Observation of Selected Segments in a Larger Protein: Central-Segment Isotope Labeling Through Intein-Mediated Ligation, Biochemistry 38, 16040-16044. 203. Wu, W., Wood, D. W., Belfort, G., Derbyshire, V., and Belfort, M. (2002) Intein-Mediated Purification of Cytotoxic Endonuclease I-TevI by Insertional Inactivation and pH-Controllable Splicing, Nucleic Acids Res. 30, 4864-4871. 204. Chong, S. R., Mersha, F. B., Comb, D. G., Scott, M. E., Landry, D., Vence, L. M., Perler, F. B., Benner, J., Kucera, R. B., Hirvonen, C. A., Pelletier, J. J., Paulus, H., and Xu, M. Q. (1997)

8 LITERATURE 186

Single-Column Purification of Free Recombinant Proteins Using a Self-Cleavable Affinity Tag Derived from a Protein Splicing Element, Gene 192, 271-281. 205. Knorr, R., Trzeciak, A., Bannwarth, W., and Gillessen, D. (1989) New Coupling Reagents in Peptide Chemistry, Tetrahedron Lett. 30, 1927-1930. 206. Mutter, M., Nefzi, A., Sato, T., Sun, X., Wahl, F., and Wohr, T. (1995) Pseudo-Prolines (Ψ-Pro) for Accessing Inaccessible Peptides, Peptide Res. 8, 145-153. 207. Sampson, W. R., Patsiouras, H., and Ede, N. J. (1999) The Synthesis of 'Difficult' Peptides using 2-Hydroxy-4-methoxybenzyl or Pseudoproline Amino Acid Building Blocks: A Comparative Study, J. Pept. Sci. 5, 403-409. 208. Wohr, T., and Mutter, M. (1995) Pseudo-Prolines in Peptide-Synthesis - Direct Insertion of Serine and Threonine Derived Oxazolidines in Dipeptides, Tetrahedron Lett. 36, 3847-3848. 209. Bartlett, P. A., and Johnson, C. R. (1985) An Inhibitor of Chorismate Mutase Resembling the Transition-State Conformation, J. Am. Chem. Soc. 107, 7792-7793. 210. Gray, J. V., Golinellipimpaneau, B., and Knowles, J. R. (1990) Monofunctional Chorismate Mutase from Bacillus subtilis - Purification of the Protein, Molecular-Cloning of the Gene, and Overexpression of the Gene-Product in Escherichia coli, Biochemistry 29, 376-383. 211. Smith, R. E., and Macquarrie, R. (1978) Sensitive Fluorometric Method for Determination of Arginine Using 9,10-Phenanthrenequinone, Anal. Biochem. 90, 246-255. 212. Mattei, P., Kast, P., and Hilvert, D. (1999) Bacillus subtilis Chorismate Mutase is Partially Diffusion-Controlled, Eur. J. Biochem. 261, 25-32. 213. Northrop, D. B. (1998) On the Meaning of Km and V/K in Enzyme Kinetics, J. Chem. Ed. 75, 1153-1157. 214. Phillips, M. A., Fletterick, R., and Rutter, W. J. (1990) Arginine-127 Stabilizes the Transition- State in Carboxypeptidase, J. Biol. Chem. 265, 20692-20698. 215. Tissot, A. C., Vuilleumier, S., and Fersht, A. R. (1996) Importance of two Buried Salt Bridges in the Stability and Folding Pathway of Barnase, Biochemistry 35, 6786-6794. 216. Aemissegger, A., Jaun, B., and Hilvert, D. (2002) Investigation of the Enzymatic and Nonenzymatic Cope Rearrangement of Carbaprephenate to Carbachorismate, J. Org. Chem. 67, 6725-6730. 217. Kienhöfer, A., Kast, P., and Hilvert, D. (2003) Selective Stabilization of the Chorismate Mutase Transition State by a Positively Charged Hydrogen Bond Donor, J. Am. Chem. Soc. 125, 3206- 3207. 218. Gray, J. V., Eren, D., and Knowles, J. R. (1990) Monofunctional Chorismate Mutase from Bacillus subtilis - Kinetic and C-13 Nmr-Studies on the Interactions of the Enzyme with Its Ligands, Biochemistry 29, 8872-8878. 219. Kim, K. S., and Qian, L. G. (1993) Synthesis of 4,4-Difluoro-L-Arginine, Tetrahedron Lett. 34, 7195-7196. 220. Tozer, M. J., and Herpin, T. F. (1996) Methods for the Synthesis of gem-Difluoromethylene Compounds, Tetrahedron 52, 8619-8683. 221. Schlosser, M., and Michel, D. (1996) About the ''Physiological Size'' of Fluorine Substituents: Comparison of Sensorially Active Compounds with Fluorine and Methyl Substituted Analogues, Tetrahedron 52, 99-108. 222. Garner, P., and Park, J. M. (1992) 1,1,-Dimethylethyl (S)- or (R)-4-Formyl-2,2-dimethyl-3- oxazolinecarboxylate: A Useful Serinal Derivative, Org. Synth. 70, 18-28. 223. Garner, P., and Park, J. M. (1987) The Synthesis and Configurational Stability of Differentially Protected β-Hydroxy-α-Amino Aldehydes, J. Org. Chem. 52, 2361-2364. 224. Guibourdenche, C., Roumestant, M. L., and Viallefont, P. (1993) Enantiospecific Synthesis of Quisqualic Acid, Tetrahedron-Asymmetr 4, 2041-2046. 225. Meffre, P., Durand, P., Branquet, E., and Legoffic, F. (1994) A Straightforward Synthesis of N- Boc-L-Serinal and N-Boc-L-Threoninal Acetonides, Synthetic Commun. 24, 2147-2152. 226. Merino, P., Franco, S., Merchan, F. L., and Tejero, T. (1998) Nucleophilic Additions of 2- Furyllithium to Carbonyl Derivatives of L-Serine. Formal Synthesis of (2R,3R)β-Hydroxy Aspartic Acid, Molecules 3, 26-30. 227. Moriwake, T., Hamano, S., Saito, S., and Torii, S. (1987) A Straightforward Synthesis of Allyl Amines from α-Amino-Acids without Racemization, Chem Lett, 2085-2088. 228. Williams, L., Zhang, Z. D., Shao, F., Carroll, P. J., and Joullie, M. M. (1996) Grignard Reactions to Chiral Oxazolidine Aldehydes, Tetrahedron 52, 11673-11694.

8 LITERATURE 187

229. Falb, E., Bechor, Y., Nudelman, A., Hassner, A., Albeck, A., and Gottlieb, H. E. (1999) IOOC Route to Substituted Chiral Pyrrolidines. Potential Glycosidase Inhibitors, J. Org. Chem. 64, 498-506. 230. Hoffman, R. V., and Tao, J. H. (1998) A synthesis of D-erythro- and L-threo-Sphingosine and Sphinganine Diastereomers via the Biomimetic Precursor 3-Ketosphinganine, J. Org. Chem. 63, 3979-3985. 231. Roush, W. R., and Hunt, J. A. (1995) Asymmetric Allylboration of 2-N,3-O-Isopropylidene-N- I Boc-L-Serinal - Diastereoselective Synthesis of the Calicheamicin γ1 Amino Sugar, J. Org. Chem. 60, 798-806. 232. Yanagida, M., Hashimoto, K., Ishida, M., Shinozaki, H., and Shirahama, H. (1989) Syntheses of Acyclic Analogs of Kainoids and Neuroexcitatory Activity, Tetrahedron Lett. 30, 3799-3802. 233. Dondoni, A., Perrone, D., and Merino, P. (1995) Chelation- and Non-Chelation-Controlled Addition of 2-(Trimethylsilyl)thiazole to α-Amino Aldehydes - Stereoselective Synthesis of the β-Amino-α-hydroxy Aldehyde Intermediate for the Preparation of the Human- Immunodeficiency-Virus Proteinase-Inhibitor Ro-31-8959, J. Org. Chem. 60, 8074-8080. 234. Schirlin, D., Baltzer, S., Altenburger, J. M., Tarnus, C., and Remy, J. M. (1996) Difluoromethyleneketone Retroamide, a Versatile Concept of Inactivation of Proteolytic Enzymes, Tetrahedron 52, 305-318. 235. Ding, Y., Wang, J. Q., Abboud, K. A., Xu, Y. L., Dolbier, W. R., and Richards, N. G. J. (2001) Synthesis of L-4,4-Difluoroglutamic Acid via Nucleophilic Addition to a Chiral Aldehyde, J. Org. Chem. 66, 6381-6388. 236. Kim, K. S., Moquin, R. V., Qian, L. G., Morrison, R. A., Seiler, S. M., Roberts, D. G. M., Ogletree, M. L., Youssef, S., and Chong, S. (1996) Preparation of Argatroban Analog Thrombin Inhibitors with Reduced Basic Guanidine Moiety, And Studies of Their Cell Permeability and Antithrombotic Activities, Med Chem Res 6, 377-383. 237. Konas, D. W., Pankuch, J. J., and Coward, J. K. (2002) The Synthesis of (2S)-4,4- Difluoroglutamyl γ-Peptides Based on Garner's Aldehyde and Fluoro-Reformatsky Chemistry, Synthesis-Stuttgart, 2616-2626. 238. Meffre, P., Dave, R. H., Leroy, J., and Badet, B. (2001) A Concise Synthesis of L-4,4- Difluoroglutamine, Tetrahedron Lett. 42, 8625-8627. 239. Rieke, R. D., and Uhm, S. J. (1975) Activated Metals. XI. An Improved Procedure for Preparation of β-Hydroxy Esters Using Activated Zinc, Synthesis-Stuttgart, 452-453. 240. Schirlin, D., Baltzer, S., and Altenburger, J. M. (1988) A Convenient Synthesis of α',β- Diamino-α,α-Difluoroketones, New Dipeptide Isosteres, Tetrahedron Lett. 29, 3687-3690. 241. Barton, D. H. R., Jang, D. O., and Jaszberenyi, J. C. (1991) Radical Mono- and Dideoxygenation with the Triethylsilane + Benzoyl Peroxide System, Tetrahedron Lett. 32, 7187-7190. 242. Chatgilialoglu, C., Ferreri, C., and Lucarini, M. (1993) A Comment on the Use of Triethylsilane as a Radical-Based Reducing Agent, J. Org. Chem. 58, 249-251. 243. Gregg, B. T., and Cutler, A. R. (1996) Hydrosilation of the Manganese Acetyl (CO)5MnC(O)CH3 with Monohydrosilanes, J. Am. Chem. Soc. 118, 10069-10084. 244. Bazant, V., Capka, M., Cerny, M., Chvalovs.V, Kochloef.K, Kraus, M., and Malek, J. (1968) Properties of Sodium-Bis-(2-Methoxyethoxy)Aluminiumhydride I: Reduction of Some Organic Functional Groups, Tetrahedron Lett., 3303-3306. 245. Capka, M., Chvalovs.V, Kochloef.K, and Kraus, M. (1969) Properties of Sodium Bis-(2- Methoxyethoxy)Aluminium Hydride II: Reduction of Aldehydes and Ketones, Collect. Czech. Chem. Comm. 34, 118-124. 246. Cerny, M., Malek, J., Capka, M., and Chvalovs.V. (1969) Properties of Sodium Bis-(2- Methoxyethoxy)Aluminium Hydride .4. Reduction of Oximes Amides Lactams Imides and Nitriles, Collect. Czech. Chem. Comm. 34, 1033-1041. 247. Malek, J., and Cerny, M. (1972) Reduction of Organic Compounds by Alkoxyaluminohydrides, Synth-Int. J. Methods, 217-234. 248. Ojima, I., and Vidal, E. S. (1998) Rhodium-Catalyzed Cyclohydrocarbonylation: Application to the Synthesis of (+)-Prosopinine and (-)-Deoxoprosophylline, J. Org. Chem. 63, 7999-8003. 249. Georgieva, A., Hirai, M., Hashimoto, Y., Nakata, T., Ohizumi, Y., and Nagasawa, K. (2003) Stereoselective Synthesis of Novel Ptilomycalin A Analogs via Successive 1,3-Dipolar Cycloaddition Reactions and their Ca2+-ATPase Inhibitory Activity, Synthesis-Stuttgart, 1427- 1432.

8 LITERATURE 188

250. Feichtinger, K., Zapf, C., Sings, H. L., and Goodman, M. (1998) Diprotected Triflylguanidines: A new Class of Guanidinylation Reagents, J. Org. Chem. 63, 3804-3805. 251. Iwanowicz, E. J., Poss, M. A., and Lin, J. (1993) Preparation of N,N'-Bis-tert- Butoxycarbonylthiourea, Synthetic Commun. 23, 1443-1445. 252. Kim, K. S., and Qian, L. G. (1993) Improved Method for the Preparation of Guanidines, Tetrahedron Lett. 34, 7677-7680. 253. Pons, J. F., Fauchere, J. L., Lamaty, F., Molla, A., and Lazaro, R. (1998) A Constrained Diketopiperazine as a new Scaffold for the Synthesis of Peptidomimetics, Eur. J. Org. Chem., 853-859. 254. Poss, M. A., Iwanowicz, E., Reid, J. A., Lin, J., and Gu, Z. X. (1992) A Mild and Efficient Method for the Preparation of Guanidines, Tetrahedron Lett. 33, 5933-5936. 255. Schiavi, B., Ahond, A., Poupat, C., and Potier, P. (2002) Preparation of N-tert- butoxycarbonylthiourea Opens the Way to Protected 2-Aminothiazoles, Synthetic Commun. 32, 1671-1674. 256. Alcon, M., Canas, M., Poch, M., Moyano, A., Pericas, M. A., and Riera, A. (1994) A Concise Enantioselective Synthesis of Allylamines and N-Boc-β-Amino Acids, Tetrahedron Lett. 35, 1589-1592. 257. Beaulieu, P. L., Duceppe, J. S., and Johnson, C. (1991) Synthesis of Chiral Vinylglycines, J. Org. Chem. 56, 4196-4204. 258. Corey, E. J., and Schmidt, G. (1979) Useful Procedures for the Oxidation of Alcohols Involving Pyridinium Dichromate in Aprotic Media, Tetrahedron Lett., 399-402. 259. Jako, I., Uiber, P., Mann, A., Wermuth, C. G., Boulanger, T., Norberg, B., Evrard, G., and Durant, F. (1991) Stereoselective Synthesis of 3-Alkylated Glutamic Acids - Application to the Synthesis of Secokainic Acid, J. Org. Chem. 56, 5729-5733. 260. Kurokawa, N., and Ohfune, Y. (1993) Synthetic Studies on Antifungal Cyclic-Peptides, Echinocandins - Stereoselective Total Synthesis of Echinocandin-D via a Novel Peptide Coupling, Tetrahedron 49, 6195-6222. 261. Lease, T. G., and Shea, K. J. (1993) The Type-2 Intramolecular Imino Diels-Alder Reaction - Synthesis and Structural Characterization of Bicyclo[N.3.1] Bridgehead Olefin Bridgehead Lactams, J. Am. Chem. Soc. 115, 2248-2260. 262. Marshall, J. A., Seletsky, B. M., and Coan, P. S. (1994) Highly Syn-Selective Additions of Allylic Stannanes to Protected α-Amino Aldehydes, J. Org. Chem. 59, 5139-5140. 263. Matsunaga, H., Ishizuka, T., and Kunieda, T. (1997) Chiral Electrophilic ''Glycinal'' Equivalents. New Synthons for Optically Active α-Amino Acids and 4-Substituted 2-Oxazolidinones, Tetrahedron 53, 1275-1294. 264. Sasaki, N. A., Hashimoto, C., and Potier, P. (1987) A Novel-Approach to the Synthesis of Optically Pure Non-Protein α-Amino-Acids in Both L-Configuration and D-Configuration from L-Serine, Tetrahedron Lett. 28, 6069-6072. 265. Sasaki, N. A., Hashimoto, C., and Pauly, R. (1989) Stereoselective Synthesis of Optically Pure β, γ−Unsaturated α-Amino-Acids in both L and D Configurations, Tetrahedron Lett. 30, 1943- 1946. 266. Schmidt, U., Mundinger, K., Mangold, R., and Lieberknecht, A. (1990) Lavendomycin - Total Synthesis and Assignment of Configuration, J. Chem. Soc. Chem. Comm., 1216-1219. 267. Schmidt, U., Zäh, M., and Lieberknecht, A. (1991) The Total Synthesis of Frangulanine, J. Chem. Soc. Chem. Comm., 1002-1004. 268. Sheu, J. H., Yen, C. F., Huang, H. C., and Hong, Y. L. V. (1989) Total Synthesis of Ostopanic Acid, a Plant Cyto-Toxin, Via Cyclopropanation of 2-Normal-Hexylfuran, J. Org. Chem. 54, 5126-5128. 269. Spencer, R. W., Tam, T. F., Thomas, E., Robinson, V. J., and Krantz, A. (1986) Ynenol Lactones - Synthesis and Investigation of Reactions Relevant to Their Inactivation of Serine Proteases, J. Am. Chem. Soc. 108, 5589-5597. 270. Zhang, Q. W., Mohan, R. M., Cook, L., Kazanis, S., Peisach, D., Foxman, B. M., and Snider, B. B. (1993) Asymmetric Induction in Mn(III)-Based Oxidative Free-Radical Cyclizations of Phenylmenthyl Acetoacetates and 2,5-Dimethylpyrrolidine Acetoacetamides, J. Org. Chem. 58, 7640-7651. 271. Nagano, H., Kato, E., Yamamura, S., and Ueda, M. (2003) Fluorescence Studies on Nyctinasty using Fluorescence Labeled cis-p-Coumaroylagmatine, a Leaf-Opening Substance of Albizzia

8 LITERATURE 189

Plants: Existence of Genus-Specific Receptor for Leaf-Movement Factor, Tetrahedron Lett. 44, 2953-2956. 272. Ali, S. S., Khan, K. M., Echner, H., Voelter, W., Hasan, M., and Atta-ur-Rahman. (1995) 2 New Protecting Groups for the Guanidino Function of Arginine, J. Prakt. Chem. 337, 12-17. 273. Goldstein, T. P., Nachlas, M. M., Seligman, A. M., and Plapinger, R. E. (1962) Synthesis of Chromogenic Substrates for Assay of Aminopeptidase Activity, J. Med. Pharmaceut. Chem. 5, 852-857. 274. Moynihan, H. A., and Yu, W. P. (1998) Improved Selective Protections of L-arginine, Synthetic Commun. 28, 17-23. 275. Ramage, R., Green, J., and Blake, A. J. (1991) An Acid Labile Arginine Derivative for Peptide- Synthesis - NG-2,2,5,7,8-Pentamethylchroman-6-sulfonyl-L-arginine, Tetrahedron 47, 6353- 6370. 276. Carpino, L. A., Shroff, H., Triolo, S. A., Mansour, E. M. E., Wenschuh, H., and Albericio, F. (1993) The 2,2,4,6,7-Pentamethyldihydrobenzofuran-5-Sulfonyl Group (Pbf) as Arginine Side- Chain Protectant, Tetrahedron Lett. 34, 7829-7832. 277. Fields, C. G., and Fields, G. B. (1993) Minimization of Tryptophan Alkylation Following 9- Fluorenylmethoxycarbonyl Solid-Phase Peptide-Synthesis, Tetrahedron Lett. 34, 6661-6664.

278. Ramage, R., and Green, J. (1987) Nγ-2,2,5,7,8-Pentamethylchroman-6-Sulfonyl-L-Arginine - a New Acid Labile Derivative for Peptide-Synthesis, Tetrahedron Lett. 28, 2287-2290. 279. Fitzgerald, M. C., Chernushevich, I., Standing, K. G., Kent, S. B. H., and Whitman, C. P. (1995) Total Chemical Synthesis and Catalytic Properties of the Enzyme Enantiomers L-4- Oxalocrotonate and D-4-Oxalocrotonate Tautomerase, J. Am. Chem. Soc. 117, 11075-11080. 280. Peller, L., and Alberty, R. A. (1959) Multiple Intermediates in Steady State Enzyme Kinetics I: The Mechanism Involving a Single Substrate and Product, J. Am. Chem. Soc. 81, 5907-5914. 281. Knowles, J. R. (1976) Intrinsic Pka-Values of Functional-Groups in Enzymes - Improper Deductions from Ph-Dependence of Steady-State Parameters, CRC Crit. Rev. Biochem. 4, 165- 173. 282. Alberty, R. A., and Massey, V. (1954) On the Interpretation of the Ph Variation of the Maximum Initial Velocity of an Enzyme-Catalyzed Reaction, Biochim. Biophys. Acta 13, 347- 353. 283. Wendt, S., McCombie, G., Daniel, J., Kienhöfer, A., Hilvert, D., and Zenobi, R. (2003) Quantitative Evaluation of Noncovalent Chorismate Mutase-Inhibitor Binding by ESI-MS, J. Am. Soc. Mass Spectrosc. 14, 1470-1476. 284. Downing, A. K. (2004) Protein NMR Techniques, Vol. 278, 2nd ed., University of Oxford Press, Oxford, UK. 285. Philippopoulos, M., and Lim, C. (1999) Exploring the Dynamic Information Content of a Protein NMR Structure: Comparison of a Molecular Dynamics Simulation with the NMR and X-ray Structures of Escherichia coli Ribonuclease HI, Proteins 36, 87-110. 286. Pervushin, K., Riek, R., Wider, G., and Wüthrich, K. (1997) Attenuated T-2 Relaxation by Mutual Cancellation of Dipole- Dipole Coupling and Chemical Shift Anisotropy Indicates an Avenue to NMR Structures of very Large Biological Macromolecules in Solution, Proc. Natl. Acad. Sci. USA 94, 12366-12371. 287. Guntert, P., Salzmann, M., Braun, D., and Wüthrich, K. (2000) Sequence-Specific NMR Assignment of Proteins by Global Fragment Mapping with the Program MAPPER, J. Biomol. NMR 18, 129-137. 288. Gardner, K. H., Zhang, X. C., Gehring, K., and Kay, L. E. (1998) Solution NMR Studies of a 42 kDa Escherichia coli Maltose Binding Protein β-Cyclodextrin Complex: Chemical Shift Assignments and Analysis, J. Am. Chem. Soc. 120, 11738-11748. 289. Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wüthrich, K. (2000) NMR Assignment and Secondary Structure Determination of an Octameric 110 kDa Protein using TROSY in Triple Resonance Experiments, J. Am. Chem. Soc. 122, 7543-7548. 290. Eletsky, A., Kienhöfer, A., and Pervushin, K. (2001) TROSY NMR with Partially Deuterated Proteins, J. Biomol. NMR 20, 177-180. 291. Pervushin, K., and Eletsky, A. (2003) A new Strategy for Backbone Resonance Assignment in Large Proteins using a MQ-HACACO Experiment, J. Biomol. NMR 25, 147-152. 292. Hu, K. F., Eletsky, A., and Pervushin, K. (2003) Backbone Resonance Assignment in Large Protonated Proteins using a Combination of new 3D TROSY-HN(CA)HA, 4D TROSY- HACANH and C-13-detected HACACO Experiments, J. Biomol. NMR 26, 69-77.

8 LITERATURE 190

293. Eletsky, A., Heinz, T., Moreira, O., Kienhöfer, A., Hilvert, D., and Pervushin, K. (2002) Direct NMR Observation and DFT Calculations of a Hydrogen bond at the Active Site of a 44 kDa Enzyme, J. Biomol. NMR 24, 31-39. 294. Eletsky, A., Moreira, O., Kovacs, H., and Pervushin, K. (2003) A Novel Strategy for the Assignment of Side-Chain Resonances in Completely Deuterated Large Proteins using C-13 Spectroscopy, J. Biomol. NMR 26, 167-179. 295. Rajagopalan, J. S., Taylor, K. M., and Jaffe, E. K. (1993) C-13 NMR-Studies of the Enzyme Product Complex of Bacillus subtilis Chorismate Mutase, Biochemistry 32, 3965-3972. 296. Luginbuhl, P., and Wüthrich, K. (2002) Semi-Classical Nuclear Spin Relaxation Theory Revisited for use with Biological Macromolecules, Prog. Nucl. Magn. Reson. Spectrosc. 40, 199-247. 297. Tjandra, N., Feller, S. E., Pastor, R. W., and Bax, A. (1995) Rotational Diffusion Anisotropy of Human Ubiquitin from N-15 NMR Relaxation, J. Am. Chem. Soc. 117, 12562-12566. 298. Palmer, A. G. (1996) http://cpmcnet.columbia.edu/dept/gsas/biochem/labs/palmer/software/diffusion.html. 299. Hall, J. B., and Fushman, D. (2003) Characterization of the Overall and Local Dynamics of a Protein with Intermediate Rotational Anisotropy: Differentiating Between Conformational Exchange and Anisotropic Diffusion in the B3 Domain of Protein G, J. Biomol. NMR 27, 261- 275. 300. Luginbuhl, P., Pervushin, K. V., Iwai, H., and Wüthrich, K. (1997) Anisotropic Molecular Rotational Diffusion in N-15 Spin Relaxation Studies of Protein Mobility, Biochemistry 36, 7305-7312. 301. Dellwo, M. J., and Wand, A. J. (1989) Model-Independent and Model-Dependent Analysis of the Global and Internal Dynamics of Cyclosporine-A, J. Am. Chem. Soc. 111, 4571-4578. 302. Orekhov, V. Y., Pervushin, K. V., Korzhnev, D. M., and Arseniev, A. S. (1995) Backbone Dynamics of (1-71) Bacterioopsin and (1- 36) Bacterioopsin Studied By 2-Dimensional H-1-N- 15 NMR- Spectroscopy, J. Biomol. NMR 6, 113-122. 303. Schneider, D. M., Dellwo, M. J., and Wand, A. J. (1992) Fast Internal Main-Chain Dynamics of Human Ubiquitin, Biochemistry 31, 3645-3652. 304. Korzhnev, D. M., Billeter, M., Arseniev, A. S., and Orekhov, V. Y. (2001) NMR Studies of Brownian Tumbling and Internal Motions in Proteins, Prog. Nucl. Magn. Reson. Spectrosc. 38, 197-266. 305. Mandel, A. M., Akke, M., and Palmer, A. G. (1995) Backbone Dynamics of Escherichia coli Ribonuclease-H1: Correlations With Structure and Function in an Active Enzyme, J. Cell. Biochem., 29-29. 306. d'Auvergne, E. J., and Gooley, P. R. (2003) The use of Model Selection in the Model-Free Analysis of Protein Dynamics, J. Biomol. NMR 25, 25-39. 307. Wishart, D. S., and Sykes, B. D. (1994) The C-13 Chemical-Shift Index: A Simple Method for the Identification of Protein Secondary Structure Using C-13 Chemical-Shift Data, J. Biomol. NMR 4, 171-180. 308. Koradi, R., Billeter, M., and Wüthrich, K. (1996) MOLMOL: A Program for Display and Analysis of Macromolecular Structures, J. Mol. Graphics 14, 51-55. 309. Palmer, A. G., Williams, J., and McDermott, A. (1996) Nuclear Magnetic Resonance Studies of Biopolymer Dynamics, J. Phys. Chem. 100, 13293-13310. 310. Lipari, G., and Szabo, A. (1982) Model-Free Approach to the Interpretation of Nuclear Magnetic- Resonance Relaxation in Macromolecules 1: Theory and Range of Validity, J. Am. Chem. Soc. 104, 4546-4559. 311. Lipari, G., and Szabo, A. (1982) Model-Free Approach to the Interpretation of Nuclear Magnetic- Resonance Relaxation in Macromolecules 2: Analysis of Experimental Results, J. Am. Chem. Soc. 104, 4559-4570. 312. Clore, G. M., Szabo, A., Bax, A., Kay, L. E., Driscoll, P. C., and Gronenborn, A. M. (1990) Deviations from the Simple 2-Parameter Model-Free Approach to the Interpretation of N-15 Nuclear Magnetic-Relaxation of Proteins, J. Am. Chem. Soc. 112, 4989-4991. 313. Fischer, M. W. F., Zeng, L., Majumdar, A., and Zuiderweg, E. R. P. (1998) Characterizing Semilocal Motions in Proteins by NMR Relaxation Studies, Proc. Natl. Acad. Sci. USA 95, 8016-8019.

8 LITERATURE 191

314. Pang, Y., Buck, M., and Zuiderweg, E. R. P. (2002) Backbone Dynamics of the Ribonuclease Binase Active Site Area using Multinuclear (N-15 and (CO)-C-13) NMR Relaxation and Computational Molecular Dynamics, Biochemistry 41, 2655-2666. 315. Wang, L. C., Pang, Y. X., Holder, T., Brender, J. R., Kurochkin, A. V., and Zuiderweg, E. R. P. (2001) Functional Dynamics in the Active site of the Ribonuclease Binase, Proc. Natl. Acad. Sci. USA 98, 7684-7689. 316. Vamvaca, K., Vogeli, B., Kast, P., Pervushin, K., and Hilvert, D. (2004) An Enzymatic Molten Globule: Efficient Coupling of Folding and Catalysis, Proc. Natl. Acad. Sci. USA 101, 12860- 12864. 317. Fischer, M. W. F., Zeng, L., Pang, Y. X., Hu, W. D., Majumdar, A., and Zuiderweg, E. R. P. (1997) Experimental Characterization of Models for Backbone Picosecond Dynamics in Proteins. Quantification of NMR Auto- and Cross-Correlation Relaxation Mechanisms Involving Different Nuclei of the Peptide Plane, J. Am. Chem. Soc. 119, 12629-12642. 318. Wang, T. Z., Cai, S., and Zuiderweg, E. R. P. (2003) Temperature Dependence of Anisotropic Protein Backbone Dynamics, J. Am. Chem. Soc. 125, 8639-8643. 319. Boyar, A., and Marsh, R. E. (1982) L-Canavanine, a Paradigm for the Structures of Substituted Guanidines, J. Am. Chem. Soc. 104, 1995-1998. 320. de la Torre, J. G., Huertas, M. L., and Carrasco, B. (2000) Calculation of Hydrodynamic Properties of Globular Proteins from their Atomic-Level Structure, Biophys. J. 78, 719-730. 321. Schuck, P. (2000) Size-Distribution Analysis of Macromolecules by Sedimentation Velocity Ultracentrifugation and Lamm Equation Modeling, Biophys. J. 78, 1606-1619. 322. Laue, T. M., Shah, B. D., Ridgeway, T. M., and Pelletier, S. L. (1992) Analytical Ultracentrifugation in Biochemistry and Polymer Science, Cambridge, England, 90-125; software by Philo, J., http://www.jphilo.mailway.com/. 323. Demeler, B. (2004) UltraScan v. 6.2, Data Analysis Software for Analytical Ultracentrifugation Experiments. The University of Texas Health Science Center at San Antonio. http;//www.ultrascan.uthscsa.edu. 324. Guntert, P., Dötsch, V., Wider, G., and Wüthrich, K. (1992) Processing of Multidimensional NMR Data With the New Software Prosa, J. Biomol. NMR 2, 619-629. 325. Bartels, C., Xia, T. H., Billeter, M., Guntert, P., and Wüthrich, K. (1995) The Program Xeasy for Computer-Supported NMR Spectral-Analysis of Biological Macromolecules, J. Biomol. NMR 6, 1-10. 326. Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wüthrich, K. (1999) [C-13]-Constant- Time [N-15,H-1]-TROSY-HNCA for Sequential Assignments of Large Proteins, J. Biomol. NMR 14, 85-88. 327. Salzmann, M., Wider, G., Pervushin, K., Senn, H., and Wüthrich, K. (1999) TROSY-Type Triple-Resonance Experiments for Sequential NMR Assignments of Large Proteins, J. Am. Chem. Soc. 121, 844-848. 328. Zhu, G., and Bax, A. (1990) Improved Linear Prediction for Truncated Signals of Known Phase, J. Magn. Reson. 90, 405-410. 329. Renner, C., Schleicher, M., Moroder, L., and Holak, T. A. (2002) Practical Aspects of the 2D N- 15-{H-1}-NOE Experiment, J. Biomol. NMR 23, 23-33. 330. Mandel, A. M., Akke, M., and Palmer, A. G. (1995) Backbone Dynamics of Escherichia coli Ribonuclease H1: Correlations With Structure and Function in an Active Enzyme, J. Mol. Biol. 246, 144-163. 331. Palmer, A. G., Rance, M., and Wright, P. E. (1991) Intramolecular Motions of a Zinc Finger DNA-Binding Domain From Xfin Characterized By Proton-Detected Natural Abundance C-12 Heteronuclear NMR-Spectroscopy, J. Am. Chem. Soc. 113, 4371-4380. 332. Kay, L. E., Torchia, D. A., and Bax, A. (1989) Backbone Dynamics of Proteins as Studied by N- 15 Inverse Detected Heteronuclear NMR-Spectroscopy - Application to Staphylococcal Nuclease, Biochemistry 28, 8972-8979. 333. Orekhov, V., Pervushin, K. V., Korzhnev, D. M., and Arseniev, A. S. (1995) Backbone Dynamics of (1-71) and (1-36) Bacterioopsin Studied By 2-Dimensional NMR-Spectroscopy, J. Cell. Biochem., 50-50. 334. Sambrook, J., and Russell, D. W. (2001) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, NY. 335. MacBeath, G., and Kast, P. (1998) UGA Read-Through Artifacts - When Popular Gene Expression Systems Need a pATCH, Biotechniques 24, 789-794.

8 LITERATURE 192

336. Kast, P. (1994) pKSS - a 2nd-Generation General-Purpose Cloning Vector for Efficient Positive Selection of Recombinant Clones, Gene 138, 109-114. 337. Duan, X. Q., Gimble, F. S., and Quiocho, F. A. (1997) Crystal Structure of PI-Scel, a Homing Endonuclease with Protein Splicing Activity, Cell 89, 555-564. 338. Mullis, K., Faloona, F., Scharf, S., Saiki, R., Horn, G., and Erlich, H. (1986) Specific Enzymatic Amplification of DNA in vitro - the Polymerase Chain-Reaction, Cold Spring Harb. Sym. 51, 263-273. 339. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) DNA Sequencing with Chain-Terminating Inhibitors, Proc. Natl. Acad. Sci. USA 74, 5463-5467. 340. Edelhoch, H. (1967) Spectroscopic Determination of Tryptophan and Tyrosine in Proteins, Biochemistry 6, 1948-1954. 341. Gill, S. C., and von Hippel, P. H. (1989) Calculation of Protein Extinction Coefficients from Amino-Acid Sequence Data, Anal. Biochem. 182, 319-326. 342. Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T. (1995) How to Measure and Predict the Molar Absorption-Coefficient of a Protein, Protein Sci. 4, 2411-2423. 343. Grisostomi, C., Kast, P., Pulido, R., Huynh, J., and Hilvert, D. (1997) Efficient in Vivo Synthesis and Rapid Purification of Chorismic Acid using an Engineered Escherichia coli Strain, Bioorg. Chem. 25, 297-305. 344. Todhunter, J. A. (1979) Reversible Enzyme Inhibition, Method. Enzymol. 63, 383-411. 345. Bunin, B. A. (1998) The Combinatorial Index, Academic Press, San Diego. 346. Ohira, S., Akiyama, M., Kamihara, K., Isoda, Y., and Kuboki, A. (2002) Stereoselective Synthesis of (1S,3R)-1-Aminocyclopentane-1,3-dicarboxylic Acid via C-H Insertion of Alkylidenecarbene, Biosci Biotech Bioch 66, 887-891. 347. Brazier, S. P., Ramesh, B., Haris, P. I., Lee, D. C., and Srai, S. K. S. (1998) Secondary Structure Analysis of the Putative Membrane-Associated Domains of the Inward Rectifier K+ Channel ROMK1, Biochem. J. 335, 375-380.

193

Curriculum Vitae

PERSÖNLICHE DATEN: Name: Alexander Kienhöfer

Geburtsdatum: 15.02.1973

Geburtsort: Göppingen, Deutschland

Zivilstand: Ledig

AUSBILDUNG: 8/79 - 7/83 Grundschule Rechberghausen

9/83 - 5/92 Freihof-Gymnasium Göppingen

10/93 Beginn des Chemiestudiums an der Universität Konstanz

8/95 Vordiplom

8/95 - 8/96 Chemiestudium an der Northern Arizona University in Flagstaff (USA)

10/96 Fortsetzung des Chemiestudiums an der Universität Konstanz

1/99 - 9/99 Diplomarbeit in Organischer Chemie in der Gruppe von Prof. U. Groth. Titel: Der Einfluß von Ce(III) auf die Reaktion von Zinkorganylen mit Carbonylgruppen

9/99 Diplom in Chemie

2/00 – 4/05 Doktorarbeit an der ETH Zürich in Organischer Chemie in der Arbeitsgruppe von Prof. D. Hilvert. Während dieser Zeit Tätigkeit als Assistent für Praktika in Biologischer, Anorganischer und Organischer Chemie, Vorlesungsassistent für die Vorlesung und Übungen in Biologischer Chemie, Betreuung zweier Semesterstudenten und eines Diplomanden.

Zürich, im April 2005

Alexander Kienhöfer